Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
sequencelengths 1
91
| keywords
sequencelengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
X chromosome inactivation in eutherian mammals has been thought to be tightly controlled , as expected from a mechanism that compensates for the different dosage of X-borne genes in XX females and XY males . However , many X genes escape inactivation in humans , inactivation of the X in marsupials is partial , and the unrelated sex chromosomes of monotreme mammals have incomplete and gene-specific inactivation of X-linked genes . The bird ZW sex chromosome system represents a third independently evolved amniote sex chromosome system with dosage compensation , albeit partial and gene-specific , via an unknown mechanism ( i . e . upregulation of the single Z in females , down regulation of one or both Zs in males , or a combination ) . We used RNA-fluorescent in situ hybridization ( RNA-FISH ) to demonstrate , on individual fibroblast cells , inactivation of 11 genes on the chicken Z and 28 genes on the X chromosomes of platypus . Each gene displayed a reproducible frequency of 1Z/1X-active and 2Z/2X-active cells in the homogametic sex . Our results indicate that the probability of inactivation is controlled on a gene-by-gene basis ( or small domains ) on the chicken Z and platypus X chromosomes . This regulatory mechanism must have been exapted independently to the non-homologous sex chromosomes in birds and mammals in response to an over-expressed Z or X in the homogametic sex , highlighting the universal importance that ( at least partial ) silencing plays in the evolution on amniote dosage compensation and , therefore , the differentiation of sex chromosomes .
Vertebrates with heteromorphic sex chromosomes have either male heterogamety like humans ( XX female and XY male ) , or female heterogamety like birds ( ZZ male and ZW female ) . Degeneration of the non-recombining Y or W chromosome , central to the evolution of sex chromosomes , left genes on the X or Z as a single copy in the heterogametic sex . This resulted in an imbalance of X or Z gene dosage relative to the autosomes , and between the sexes . Different dosage compensation systems have arisen independently in diverse organisms , suggesting that dosage compensation is critical for the survival of species with differentiated sex chromosomes . Ohno [1] hypothesized that degeneration of the Y/W chromosome would result in under expression from the X/Z in the heterogametic sex ( equivalent to monosomy ) , which would result in pressure to up regulate the single X/Z to restore parity with the autosomes . Because the homogametic sex has two X/Z chromosomes , over expression would result in doubled normal expression ( equivalent to tetrasomy ) , which would result in pressure for global down regulation of the X/Z to again resort parity with the autosome ( reviewed in [2] ) . However , recent data has questioned global over regulation of the X/Z [3] , [4] , sparking considerable debate [5]–[7] and suggestion that dosage compensation evolved in response to a subset of dosage sensitive genes [8] . In eutherian mammals one whole X is transcriptionally silenced in the somatic cells of females , although many genes located on the evolutionarily more recent region escape silencing [9] . X inactivation ( XCI ) is established early in embryogenesis and is somatically heritable . In eutherian mammals the silenced X is chosen at random in a process governed by a large non-coding RNA transcribed from a locus called XIST ( X-inactive-specific transcript ) [10]–[12] . In all eutherian mammals , XIST is transcribed from the X to be inactivated , which it coats in cis during early development , although the timing and regulation of expression varies between species [13] . After one X is chosen for inactivation , a specific signature of epigenetic modifications is established [14] , which appears to be conserved in even the most distantly related eutherians [15] . The increasing availability of genomic data now makes it possible to study dosage compensation in non-traditional model organisms such as marsupial and monotreme mammals . In marsupials , XCI is imprinted , with the paternal X always being the inactivated homologue [16] . X-inactivation in representative Australian marsupials is thought to be incomplete , tissue-specific and gene-specific [17] . RNA-fluorescent in situ hybridization ( RNA-FISH ) showed that within a population of fibroblasts cells derived from female tammar wallaby there was a mixture of 1X-active and 2X-active nuclei , the proportion of which was characteristic of each locus [18] . A similar profile was obtained for human and elephant X genes that escape inactivation [19] . However , it was shown ( with RNA-FISH ) in post-mortem tissue of the South American grey short-tailed opossum that X-inactivation is efficient [20] , and RNA-sequencing revealed that X expression levels are similar between the sexes [21] . The egg-laying monotreme mammals ( platypus and echidna species ) have multiple sex chromosomes that share homology , not with the therian X , but the bird Z [22] , [23] . Recent transcriptome sequencing in platypus showed that compensation is achieved via upregulation of X genes in males , and that global X-inactivation in females is likely unnecessary [21] . However , as for marsupials and escaper genes on the human X , RNA-FISH showed that within a population of platypus fibroblast cells , there was a mixture of 1X-active and 2X-active cells , each locus having a characteristic frequency of inactivation [24] . Data on dosage compensation in birds are fragmentary and contradictory . Real-time PCR showed equivalent expression of most Z-borne genes in ZZ male and ZW female chick embryos ( that is , complete compensation ) for most genes [25] , but global microarray analysis in chicken , and a small cDNA microarray in zebra finch showed that male to female ratios were significantly higher for Z genes than for autosomal genes [26] . A male: female Z gene dosage of approximately 1 . 5 was demonstrated by microarray [27] and RNA-sequencing [21] analyses in chicken , and dosage ratios of 1 . 23 was observed in zebra finch [28] and 1 . 36 in crow [29] . The incomplete dosage compensation of Z-linked genes , at least in chicken , was reported to be regulated locally on a gene-by-gene basis , and is tissue and developmental stage specific [30] . It is difficult to compare dosage compensation between birds and mammals . However , recent comparative transcriptome sequencing [21] indicated that genes on the single Z in female chicken had equivalent expression levels to orthologous genes in outgroup species , where the Z is autosomal ( i . e . expression from one Z equals expression from two proto-Zs ) , providing evidence for global upregulation of the single Z in females . In male chicken ( with two Zs ) the ZZ:proto-ZZ expression ratio was 1 . 13 to 1 . 56 , indicating that Z gene upregulation was not specific to females . A similar pattern , although less clear , was observed for genes on platypus X5 . Less efficient upregulation in the homogametic sex ( i . e . ZZ/X5X5: ZW/X5Y5 expression ratio <2 ) indicated that upregulation was more efficient in the heterogametic sex , or that there was a mechanism to partially reduce expression of Z/X5 in the homogametic sex . Julien et al . [21] suggested that dosage compensation only mildly affects the homogametic sex in platypus and chicken and , as such , there was potentially no requirement for the evolution of Z/X5 inactivation . Here we use RNA-FISH to examined chicken and platypus dosage compensation , which permits detection of transcription from one or both alleles at specific loci in individual nuclei . We found that genes on the chicken Z , as well as on the partially homologous platypus Xs , were expressed from one ( or both ) alleles in characteristic frequencies for different loci , just as on the independently evolved therian X . Our results indicate that silencing mechanisms were exapted multiple times in the homogametic sex ( likely in response to the upregulation of genes on the Z/X in the heterogametic sex ) resulting in transcriptional inactivation on non-homologous Zs and Xs in distantly related species .
As controls , we chose ten BACs containing known autosomal genes , including a BAC containing GAPDH , which was used as a control in all experiments . In RNA-FISH experiments with each of the ten autosomal BACs , we observed two signals in at least 97% of nuclei in both male and female fibroblast cells ( Table S1 and Figure S1 ) . Thus , autosomal genes are generally transcribed from both alleles in cultured chicken fibroblasts . We selected eleven Z-borne BACs , five of which contained a single gene , and six contained two or more genes ( Table S1 ) . We performed two-color RNA-FISH , with the autosomal control ( GAPDH ) and each test gene ( Z-borne and autosomal ) , on male ( ZZ ) and female ( ZW ) fibroblasts ( Table S1 ) , and scored ≥100 cells for each hybridization . To control for polyploidy and accessibility of the probe into an individual nucleus , only nuclei with two signals from the control BAC were scored for the test BAC . The frequency with which expression of the single Z in ZW females was detected as a single signal , controlled for the hybridization efficiency of each Z probe . All Z BAC probes displayed hybridization efficiencies of between 95–100% ( Table S1 ) . Even at the lowest hybridization efficiency of 95% ( p = 0 . 95; q = 0 . 05 ) , hybridization in ZZ male cells would produce few nuclei with 0 ( q2 = 0 . 25% ) or 1 ( 2pq = 9 . 5% ) signal ( see Materials and Methods ) . For all probes hybridized to ZZ male nuclei , the cell population consisted of a mixture of nuclei that were 1Z-active ( one signal ) and 2Z-active ( two signals ) ( Table S1 ) . No locus was completely 1Z-active or completely 2Z-active . Instead , each Z locus was inactivated in a characteristic frequency of cells , ranging from 15% to 51% ( Figure 1 ) , that was significantly different ( p<0 . 01 after Bonferroni correction and estimating experimental error; see Materials and Methods ) from the number of 1-Z active nuclei expected due to inefficient hybridization . This frequency was reproducible between biological replicates for a subset of BACs , in which RNA-FISH was repeated on fibroblasts from a second individual ( Table S1 ) . The activity status of Z loci in a given nucleus did not appear to be clonally inherited ( i . e . if the mother cell was 2Z-active both daughter cells should be 2Z-active; or conversely if the mother cell was 1Z-active both daughter cells should be 1Z-active ) ; we observed daughter cells in our preparations in which one was 1Z-active and the other 2Z-active ( Figure S2 ) . There was no obvious clustering of loci on the Z with particularly high or particularly low frequencies of 1Z-active cells , suggesting that the probability of transcriptional inhibition of a gene is independent of its physical location . The frequency of 1Z-active nuclei for the six BACs within the proposed dosage compensation ‘valley’ ( 129B9 , 110A9 , 164N4 , 87K13 , 57B13 , and 89C2 ) was not lower than the five BACs in the ‘peak’ ( 65D18 , 30H20 , 73F14 , 112C1 , and 163I20 ) regions of the Z chromosome identified in chicken [28] , [33] . However , there did appear to be a correlation of M∶F expression ratio ( calculated from data in [34]; see Materials and Methods ) of Z-genes , with the proportion of 1Z-active nuclei in males ( R2 = 0 . 47 , p = 0 . 03 ) . If a gene was over expressed in males compared to females , there was a greater percentage of 1Z-active nuclei for that locus in males ( Figure S3A ) . From these initial RNA-FISH results it could not be determine if transcriptional inhibition of different loci were on the same Z chromosome , or on different Z chromosomes . Therefore we used two-color RNA-FISH to examine transcription of two pairs of neighbouring Z genes in female and male chicken nuclei: SMARCA2/PTPRD ( ∼2 Mb apart ) , and BNC2/MLLT3 ( ∼1 . 5 Mb apart ) ( Table 1 ) ( Figure 2 ) . In ZW female nuclei , we observed co-location ( close proximity ) of signals in 96% of nuclei that expressed both genes . In male nuclei , for each cell in which both loci were 1Z-active , we observed co-location of signals in 96% of nuclei ( Figure 2A ) , indicating transcription from the same Z chromosome . This is consistent with the presence of an active Z ( Za ) , and an inactive Z ( Zi ) on which genes are prone to silencing . Two-color RNA-FISH experiments with the same gene pairs also provided the opportunity to determine whether the silencing of neighbouring genes on the Zi chromosome was coordinated . In cells in which at least one of the neighbouring gene pairs was 2Z-active , we observed ( for both Z gene pairs ) that the second locus was simultaneously transcribed ( i . e . both loci were 2Z-active in the same nucleus; Figure 2B–C ) at a frequency no greater than expected by chance ( see Materials and Methods; Table 2 ) , consistent with independent silencing of tightly linked genes on Zi . The frequency of nuclei , in which both loci in a gene pair are expected to be 2Z- active , was the product of 2Z-active frequencies of each gene ( from initial RNA-FISH results; Table S1 ) . A total of 40 platypus BACs were analyzed: 19 in X-specific regions ( two on X1 , one on X2 , two on X3 , 14 on X5 ) , nine in pseudoautosomal regions ( PAR ) , and 12 on autosomes . The autosomal BAC bearing HPRT ( on chromosome 6 ) was used as the autosomal control in all experiments . Hybridization efficiencies for each probe were assessed in male cells . These ranged from 94% to 100% ( Table S2 ) . At least 100 cells were scored for each hybridization . In RNA-FISH experiments with each of the 12 autosomal BACs , we observed two signals in at least 95% of nuclei in both male and female fibroblast cells ( Table S2 and Figure S4 ) . Thus autosomal genes are generally transcribed from both alleles in cultured platypus fibroblasts . For all X specific loci we observed 1X-active female nuclei at a significantly greater frequency ( p<0 . 01 after Bonferroni correction and estimating experimental error; see Materials and Methods ) than expected from inefficient hybridization . No loci were completely 1X-active , or 2X-active in every nucleus , and frequencies of 1X-active nuclei ranged from 25% to 62% for different loci . There was no obvious clustering of genes on any X with particularly high , or particularly low frequencies of 1X-active nuclei ( Figure 1 ) . However , for loci that were expressed at a higher level in females compared to males , there did appear to be a correlation of F∶M expression ratio ( in cultured fibroblast calculated from data in [21]; see Materials and Methods ) of X-genes with the proportion of 1X-active nuclei in females ( R2 = 0 . 54 , p = 0 . 004 ) . If a gene was over expressed in females compared to males , there was a greater percentage of 1X-active nuclei for that locus in females ( Figure S3B ) . The activity status of X loci in a given nucleus was not strictly clonally inherited ( i . e . if the mother cell is 2× active both daughter cell should be 2× active; or conversely if the mother cell is 1× active both daughter cell should be 1× active ) ; we observed daughter cells in which one was 1X-active and the other 2X-active ( Figure S2 ) . We therefore conclude that genes on the platypus X chromosomes are subject to inactivation in a proportion of nuclei that is characteristic for each gene . To determine if transcriptional inhibition of adjacent loci was coordinated , we used two-color RNA-FISH to examine transcription of three pairs of neighboring genes on platypus X5: MPDZ/NFIB ( ∼500 kb apart ) , HSD17B4/SEMA6A ( ∼1 . 8 Mb apart ) and SEMA6A/SLC1A1 ( ∼500 kb apart ) ( Table 1 ) ( Figure 2A–C ) . As expected , in male controls we observed co-location of signals in 100% of nuclei for each gene pair . In female nuclei in which both loci were 1X-active , we also observed co-location of signals in 100% of nuclei for each gene pair tested ( Figure 2A ) , consistent with the presence of a single active X5 ( X5a ) , and an X5 ( X5i ) on which genes are prone to silencing . Subsequently , we tested whether inactivation of adjacent loci was synchronized on X5i for these same neighbouring gene pairs by observing , in cells with one gene 2X-active , the frequency at which the second gene was 2X-active . We observed 2X-active nuclei at frequencies that were significantly ( p<0 . 01 ) greater than expected by chance ( Table 2 , Figure 2; see Materials and Methods ) , indicating at least regional coordination of transcription on X5i . It was not possible to assess coordinated transcription over extended regions because FISH signals derived from X5a could not be distinguished from signals derived from X5i ( see Materials and Methods ) . Using RNA-FISH we also examined , in both male and female fibroblasts , transcription from genes located in the PAR of platypus sex chromosomes . We studied one locus in each of PARs X2/Y2 , X3/Y2 , X3/Y3 and X4/Y3 , two loci in X5/Y4 , and three loci in X1/Y1 ( Figure S5 ) . We expected to see nearly 100% 2X- active nuclei in both males and females , as for loci in the human PAR1 , and the autosomal controls . However , all nine pseudoautosomal BACs tested were 1× active in 27–42% of male nuclei , and 16–47% of female nuclei , indicating significant ( p<0 . 01 after Bonferroni correction ) PAR gene inactivation in both sexes for seven of the nine loci tested ( Table S2 , Figure 1 ) . Importantly , two chicken autosomal BACs , orthologous to the platypus PAR genes that are subject to partial silencing , were 2-allele-active . Additionally , we demonstrated that four of the partially inactivated chicken Z/platypus X5 loci tested here are always 2 allele-active in human fibroblasts ( Table S3 ) , where they are autosomal . Finally , two of the biallelically expressed autosomal chicken BACs ( Table S1 ) , and four of the biallelically expressed autosomal platypus BACs ( Table S2 ) , tested here contained genes orthologous to X genes in human where they are subject to silencing . These experiments demonstrate that inactivation was dependent on the type of chromosome a locus was located on ( i . e . sex chromosome or autosome ) , rather than it being a phenomenon unique to the loci examined here .
Using RNA-FISH , we demonstrate for the first time that Z inactivation plays a role in chicken dosage compensation , and confirmed that inactivation is also a feature of orthologous platypus X loci , as was previously shown by Deakin et al . [24] . Our observations conflict with those of a previous study of five chicken Z loci using RNA-FISH . Kuroda et al . [35] reported that most male nuclei expressed both Z alleles in five different tissue types ( two Z loci were tested in liver , and one each in kidney , spleen and retina ) , and suggested that there was no inactivation on the chicken Z chromosome . The inconsistency in findings with this study may be due to the different tissue types used , and different inactivation profiles of chicken Z genes in other tissue types would not be surprising . We used fibroblast cells because they are easily collected and cultured from species that are difficult to sample , and to be consistent with our previous studies of sex chromosome silencing in mammals [18] , [19] , [24] . Additionally , Kuroda et al . [35] concluded that inactivation does not occur on the chicken Z because the frequency of detecting two signals of hybridization was similar for the autosomal probe and the Z probes ( 70–80% ) . However , the detection frequencies ( 1 or 2 signals observed ) in male nuclei for the autosomal and Z probes were low ( which were 66–77% , and were not co-hybridized ) . In this study we co-hybridized our autosomal control and test genes using one cell type , and only cells with two autosomal signals were scored for the test gene , controlling for ploidy and probe accessibility into individual nuclei . As a result we obtained much higher hybridization efficiencies , providing for much simpler interpretation . An exhaustive study of male-female expression ratios using microarrays showed that different loci on the chicken Z are compensated to different extents [26] , [30] , and Julien et al . [21] eloquently demonstrated that the Z/X5 is upregulated in the heterogametic sex in chicken/platypus . These studies assessed transcription across whole cell populations , so could not distinguish between partial expression in all nuclei , and heterogeneity of expression of different cells . In male chicken fibroblast nuclei , we found for all loci a mixture of 1Z-active and 2Z active cells at reproducible frequencies ( 15% to 51% 1Z-active , depending on the locus ) . Therefore , we demonstrate that in the homogametic sex there is partial inactivation of one Z allele . As such , these results demonstrate that an integral step of the partial bird dosage compensation system must include the silencing of one Z allele in males , analogous to the inactivation we observed here ( and previously [24] ) on the orthologous ( although independently evolved ) platypus X5 , and the independently evolved therian mammal X chromosome . We also observe that loci over expressed in the homogametic sex were more likely to be 1Z/X-active , indicating that there is pressure to reduce expression of these genes in the homogametic sex . We conclude that partial inactivation on the Z in birds , and at least on four of the five X chromosomes in platypus , is likely in response to Z/X upregulation ( perhaps of only several dosage sensitive genes ) carrying through to the homogametic sex . Because higher transcript levels in the homogametic sex appear to correlate with a greater percentage of 1Z/X-active nuclei , evolution to down regulate the chicken Z ( and platypus Xs ) is conceivably an ongoing process . Were everything in equilibrium 1Z/1X-activity should correlate with a 1∶1 transcript ratio between the sexes . We found no correlation between the extent of inactivation and the location of the gene on the chicken Z ( Figure 1 ) , and did not observe stronger inactivation of loci in the dosage compensation valley , which was detected on the chicken Z chromosome using microarrays [26] , [28] , [33] , and is differentially methylated on the two Z chromosomes in males [36] . The functional significance of this dosage compensation valley is unclear because it is not conserved in zebra finch [28] . Other studies of chicken concluded that the dosage compensated valley region , concentrated in genes with low male to female expression ratios , simply contains genes with a strong female bias [30] , [37] , although this interpretation has also been challenged [38] , [39] . It is important to note that our data does not necessarily conflict with original reports on the dosage compensation valley region [33] , in which it was suggested that females up-regulate genes , rather than males down-regulating genes . We also observed no clustering of platypus X-borne genes with very low or very high frequencies of 1X-active nuclei ( Figure 1 ) . This implies that the probability of X inactivation is not correlated with gene location , and provides no evidence for an X chromosome inactivation center from which inactivation might spread . We observed that adjacent genes were always expressed from the same Za chromosome in male chicken cells . This is consistent with a hypothesis that all loci on one Z chromosome are active ( Za ) , and loci on the other Z ( Zi ) are prone to inactivation in a proportion of cells . We found that the probabilities of expression of adjacent loci on the Zi were not correlated , suggesting that escape from inactivation of each locus on Zi is independently regulated . In female platypus cells , our analysis of three platypus X5 pairs showed co-location of signals for neighbouring loci ( Table 2; Figure 2A ) , consistent with coordinate activity on one X5 ( X5a ) and inactivation of loci on the other X5 ( X5i ) in a proportion of cells in females . Partial activity of genes on Xi is consistent with the previous demonstration that both alleles were transcribed [24] . Unlike chicken , neighbouring genes on platypus X5 showed concordant 2X-activity ( Figure 2B ) , implying that the regulation of transcription from X5i is at least under regional control . We suggest , therefore , that in cells from male chickens there is one active Za , and one partially inactive ( “inactivatable” ) Zi on which loci have specific probabilities of being inactivated . Likewise , we propose that in cells from female platypus , there is one active Xa and one partially inactive ( “inactivatable” ) Xi on which loci have specific probabilities of being inactivated . Without allelic markers that would distinguish paternal and maternal Z/X chromosomes , we cannot determine whether silencing is random or imprinted . Allelic differences were used to demonstrate that both alleles are expressed , apparently at equivalent levels , in female platypus heterozygous for FBX010 , GLIS3 , and SHB [24] . These loci are all expressed from Xi in about half the cells , so the approximately equal representation of the two alleles might favor a random inactivation hypothesis; however peak intensity on sequencing trace files is at best semi-quantitative , so this question remains open . Because platypus sex chromosomes share homology with the chicken Z chromosome , our observation of inactivation in a proportion of cells for genes on the sex chromosomes in the homogametic sex of a representative bird ( ZZ male chicken ) and monotreme ( XX female platypus ) could be interpreted as the evolution of a silencing mechanism on an ancient bird-like Z in a common reptile-mammal ancestor . However , it is uncertain whether the bird Z/platypus X homology represents identity by descent ( reviewed in [40]–[42] ) , but under either scenario the sex chromosome inactivation observed here is likely independently evolved [43] . Similar patterns of 1X gene expression in a proportion of cells have been observed in marsupials ( tammar wallaby ) [18] and escaper genes on the X in eutherians ( elephant , mouse and human ) [19] . The marsupial and eutherian X shares a large conserved region , which is completely non-homologous with the sex chromosomes of birds and monotremes . Thus , the probabilistic inactivation we observed in birds , monotremes and therian mammals was independently exapted from an ancient toolkit of mechanisms to turn off transcription of one allele . Indeed , this partial silencing system shares characteristics with monoallelic expression of some autosomal mammalian genes , such as interleukins ( IL2 , IL4 , IL5 , IL10 , IL13 ) [44]–[48] . We propose that probabilistic silencing of the Z or X chromosome in the homogametic sex is an early step in the evolution of sex chromosome inactivation . A major difference in the inactivation systems is the extent of coordination . This is most evident for X inactivation in eutherian mammals , which is coordinately controlled by XIST , a locus that is absent outside eutherians [49] . In both chicken and platypus , the co-expression of neighbouring loci on the Za/X5a implies that the choice of which Zi/Xi to inactivate is controlled at least at the regional level , and possibly at the level of the whole chromosome , and in platypus ( but not chicken ) the probability of expression of neighbouring loci on X5i is at least regionally coordinated . In eutherian mammals silencing was augmented by additional molecular changes ( including histone modification and DNA methylation ) , into the stable and complex inactivation system typical of most genes on the conserved region of the eutherian X ( XCR ) , which shows tight 1X-active expression observed at the single cell , as well as the population level [9] , [19] . In contrast , the evolutionarily younger X added region of the human X ( XAR ) contains many escaper genes that show a probabilistic expression pattern similar to that of birds and monotremes . The marsupial X , though homologous to the eutherian XCR , displays a pattern of expression similar to the eutherian XAR , monotreme Xs and bird Z , suggesting that changes to the regulation of the eutherian XCR occurred after the divergence of marsupials from eutherian mammals 150 million years ago . This is consistent with the different epigenetic profiles displayed by the marsupial X and the eutherian XCR [15] , [20] , [50] . PAR inactivation has intriguing implications for sex chromosome evolution . The accepted hypothesis for the evolution of sex chromosomes and dosage compensation is that Y degeneration resulted in loss of Y gene function , which in turn drove the evolution of dosage compensation ( resulting in XCI in therian mammals ) [1] , [51] . However , an alternate hypothesis is that the spread of the XCI signal into undifferentiated X regions preceded , and drove , degradation of homologous Y regions [52] . Genes in the human and mouse PAR are exempt from X-inactivation because the Y copy complements the X [53] . We therefore expected X and Y alleles of platypus PAR genes to be active in all nuclei in both sexes , as they are for genes in the human PAR1 . Surprisingly , we observed a significantly lower frequency of 2X-active cells in both sexes for seven of the nine PAR loci . This confirms the observation [24] that other PAR loci were 1-allele active in platypus , and implies that inactivation of PAR genes regularly occurs in platypus . This contrasts with the expression of all autosome loci from both alleles in nearly all cells . In females , inactivation of genes on the X PAR might be achieved via spreading of inactivation from the X-specific regions , as has been observed in the recently evolved human PAR 2 [54] , [55] . However , PAR inactivation in males cannot be the result of XCI spreading , instead Y PARs could be inactivated by their proximity to heterochromatin on the male-specific region of the Y . It is possible that upregulation of the active X chromosomes in platypus , which maintained parity with autosomal genes , also resulted in upregulation of genes in the PARs . To mitigate this there was selection for partial inactivation of Xi PAR genes in females , and Y PAR genes in males . In conclusion , our studies of dosage compensation on independently evolved amniote sex chromosomes reveals common patterns of inactivation . This suggests that repressive molecular mechanisms were independently exapted to reduce the probability of transcription from one Z or X chromosome in the homogametic sex , in response to upregulation ( perhaps of only a subset genes ) in the heterogametic sex .
The study was approved , and all samples were collected and held under The Australian National University Animal Experimentation Ethics Committee proposal numbers R . CG . 11 . 06 and R . CG . 14 . 08 . Genes were chosen based on three criteria: 1 ) Widespread expression in other species to maximize the chance they were expressed in platypus and chicken fibroblast cells . 2 ) A BAC was preferentially chosen if it contained a single gene . 3 ) Loci were selected that were distributed at roughly even intervals along chicken Z and platypus X5 . On some of the platypus X chromosomes loci selection was limited by the paucity of anchored BACs . BACs containing chicken and platypus genes were ordered from CHORI BACPAC Resources Centre ( http://bacpac . chori . org/ ) . Chicken BACs bearing genes of interest were chosen using the UCSC genome browser BAC track ( http://genome . ucsc . edu/cgi-bin/hgGateway ) . Platypus BACs were from Veyrunes et al . [23] . Additional platypus BACs were identified by blasting the BAC end trace archive ( http://www . ncbi . nlm . nih . gov/Traces/trace . cgi ) . Chicken fibroblast cell lines were established from 8-day-old chicks; metaphase spreads were karyotypically normal . Platypus fibroblast cell lines were established from adult wild animals that were karyotypically normal . Platypus and chicken fibroblast cells were cultured at 32°C and 37°C respectively in 5 . 0% CO2 on gelatin-coated coverslips in 1∶1 AmnioMax C100 medium ( Invitrogen ) /DMEM 10% FCS to a density of 60 to 80% . RNA-FISH was carried out as previously described [18] . Hybridization of the probe to homologous DNA will not occur because there is no DNA denaturation step in the RNA-FISH protocol . For chicken RNA-FISH experiments a BAC from chromosome 1 , which contained the gene GAPDH ( CH261-14L1 ) , was used as the autosomal control . For platypus RNA-FISH experiments a BAC from chromosome 6 , which contained HPRT1 ( OaBb_405M2; GenBank Accession No . AC148426 ) , was the autosomal control . Only diploid nuclei ( two signals from the autosomal BAC ) were scored for the test gene ( X/Z or pseudoautosomal ) . For each test gene , 1X- or 1Z-active nuclei were observed as one signal , and 2X- or 2Z-active nuclei were observed as two signals . Hybridization efficiencies ( p ) were obtained from results in the heterogametic sex ( in which one signal is expected in all nuclei ) . These were used to calculate the expected frequency of nuclei with two signals , one signal , and no signals in the homogametic sex using the formula p2+2pq+q2 = 1; where p2 is the expected frequency of nuclei with two signals , 2pq ( q = 1–p ) is the frequency of nuclei with one signal , and q2 is the frequency with no signal . P-values were calculated using a χ2 test with two degrees of freedom and Bonferroni correction was conducted . For a more rigorous statistical analysis we removed no signal cells from the dataset , and calculated experimental error using missed hybridization events in our autosomal RNA-FISH experiments . In chicken we conducted 16 autosomal RNA-FISH experiments and scored a total of 1756 nuclei , each of which should produce 2 signals , for a total of 3512 signals . We observed a total of 12 nuclei with no signals and 16 nuclei with one signal , which is equal to 40 missed hybridisations out of 3512 ( 1 . 14% ) . Following a Poisson distribution , 95% confidence limits for 40 events gives a minimum of 28 . 58 and a maximum of 54 . 47 . The upper value of the 95% confidence limits represents 1 . 55% experimental error ( i . e . 54 . 47 out of 3512 ) . For platypus , using exactly the same approach ( 18 autosomal RNA-FISH experiments , 1884 nuclei scored in total , 3768 expected signals , 69 missed hybridisations , upper 95% confidence interval of 87 . 32 ) , maximum experimental error was estimated at 2 . 32% . All observed values were adjusted towards the expected values by 1 . 55% and 2 . 32% in chicken and platypus respectively . P-values were recalculated using a χ2 test with one degree of freedom . Significance remained ( after Bonferroni correction ) for all BACs . Taking into account an arbitrary conservative experimental error of 10% ( ±5% ) , and adjusting all the observed values by 5% towards the expected values still results in significant p-values . Expression values were obtained for known genes on each BAC in chicken and platypus ( from [21] , [34] ) , and expression ratios calculated . If a BAC carried more than one gene , expression data from the gene that spanned the largest proportion of the BAC was used ( DMRT1/3 expression data was not included due to its involvement in sex determination ) . For chicken expression data was available for brain , cerebellum , heart , kidney and liver . Because no data were available for cultured fibroblasts , an average M∶F expression ratio was taken across all tissues to best control for tissue specific expression changes . For platypus , data were available for cultured fibroblast cells , which was used to calculate F∶M expression ratios because the present study was conducted in cultured fibroblasts . These expression ratios were plotted against the percentage of 1Z/1X-active nuclei observed for the relevant BAC . If an RNA-FISH experiment was repeated for a BAC , an average observed percentage of 1Z/X activity was used . Scatterplots were drawn ( Figure S3 ) , and R2 and p-values were calculated in Microsoft Excel . RNA-FISH signals co-locate in the nucleus when closely linked and transcribed from the same chromosome , whereas signals from genes transcribed from different chromosomes ( or distantly linked on the same chromosome ) are further apart . We only used pairs of genes physically close enough to each other to give unmistakable results . To determine if the there was a single active Z ( or X for platypus ) , all nuclei that were 1Z ( or 1X ) active for both loci in a pair were scored . Co-location of the two signals was interpreted as a single active Z ( or X ) in that nucleus . The expected frequency of nuclei that were 2Z ( or 2X ) active for both loci in a gene pair was the product of individual 2Z ( or 2X ) frequencies of each gene ( from initial RNA-FISH results ) . To determine if inactivation of neighbouring gene pairs was coordinated , this frequency was compared to the observed frequency of nuclei that were 2Z ( or 2X ) active for both loci . P-values were calculated using a χ2 test with one degree of freedom and Bonferroni correction was conducted . | Dosage compensation is a mechanism that restores the expression of X chromosome genes back to their original level when Y homologues lose function . In placental and marsupial mammals this is achieved by upregulating the single X in males . The carry-through of overexpression to females would result in functional tetraploidy , so there is subsequent inactivation of one X chromosome in the somatic cells of females , leaving males ( XY ) and females ( XX ) with a single upregulated X . In contrast , genes on the five platypus ( a monotreme mammal ) X chromosomes and the chicken Z chromosome ( which are orthologous but independently evolved ) are expressed globally at a higher level in female platypus and male chicken respectively , indicating partial dosage compensation . Here , for the first time , we provide evidence for inactivation of genes on the chicken Z chromosome in ZZ males , and on all five Xs in female platypus . Our results suggest that the silencing of genes on sex chromosomes has evolved independently in birds and mammals , and is , therefore , a critical step in the pathway to dosage compensate independently evolved amniote sex chromosomes systems . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology",
"genetics",
"biology",
"genomics"
] | 2013 | Independent Evolution of Transcriptional Inactivation on Sex Chromosomes in Birds and Mammals |
Caenorhabditis elegans vulval development provides an important paradigm for studying the process of cell fate determination and pattern formation during animal development . Although many genes controlling vulval cell fate specification have been identified , how they orchestrate themselves to generate a robust and invariant pattern of cell fates is not yet completely understood . Here , we have developed a dynamic computational model incorporating the current mechanistic understanding of gene interactions during this patterning process . A key feature of our model is the inclusion of multiple modes of crosstalk between the epidermal growth factor receptor ( EGFR ) and LIN-12/Notch signaling pathways , which together determine the fates of the six vulval precursor cells ( VPCs ) . Computational analysis , using the model-checking technique , provides new biological insights into the regulatory network governing VPC fate specification and predicts novel negative feedback loops . In addition , our analysis shows that most mutations affecting vulval development lead to stable fate patterns in spite of variations in synchronicity between VPCs . Computational searches for the basis of this robustness show that a sequential activation of the EGFR-mediated inductive signaling and LIN-12 / Notch-mediated lateral signaling pathways is key to achieve a stable cell fate pattern . We demonstrate experimentally a time-delay between the activation of the inductive and lateral signaling pathways in wild-type animals and the loss of sequential signaling in mutants showing unstable fate patterns; thus , validating two key predictions provided by our modeling work . The insights gained by our modeling study further substantiate the usefulness of executing and analyzing mechanistic models to investigate complex biological behaviors .
Describing mechanistic models in biology in a formal language , especially one that is dynamic and executable by computer , has recently been shown to have various advantages ( see review [1] ) . A formal language comes with a rigorous semantics that goes beyond the simple positive and negative interaction symbols typically used in biological diagrammatic models . If the language used to formalize the model is intended for describing dynamic processes , the semantics , by its very nature , provides the means for tracing the dynamics of system behavior , which is the ability to run , or execute , the models described therein . Dynamic models can represent phenomena of importance to biological behaviors that static diagrammatic models cannot represent , such as time and concurrency . In addition , formal verification methods can be used to ensure the consistency of such computational models with the biological data on which they are based [2 , 3] . It was previously suggested that by formalizing both the experimental observations obtained from a biological system and the mechanisms underlying the system's behaviors , one can formally verify that the mechanistic model reproduces the system's known behavior [3] . Formal models are used in a variety of situations to predict the behavior of real systems and have the advantage that they can be executed by computers; often at a fraction of the cost , time , or resource consumption that the observation of the real system would require . In addition , formal models have the advantage that they can be analyzed by computers . For example , it may be possible to predict , by analyzing a model , that all possible executions will reach a stable state , independent of environment behavior . The result of such an analysis would not be obtainable by executing the real system , no matter for how long or how many times , as there are often infinitely many possible environment behaviors . This process of computational model analysis , in the case of state-based models , is called model checking [4] . Here we follow the idea that model execution and model checking can be used to test a biological hypothesis: if the execution and analysis of the model conform to the experimental observations of the biological system , then the model may correctly represent the mechanism that underlies the system behavior; otherwise , the model needs modification or refinement . Thus , the model can be seen as a “hypothesis , ” i . e . , an explanation for a biological mechanism and experiments can confirm or falsify the hypothesis . As part of an ongoing effort to model C . elegans vulval development [3 , 5] , we have previously created a formal dynamic model of vulval cell fate specification based solely on the proposed diagrammatic model of Sternberg and Horvitz from 1989 [6] . Our previous work [3] has demonstrated that state-based mechanistic models are particularly well-suited for capturing the level of understanding obtained using the tools and approaches common in the field of developmental genetics , and that creating such executable biological models is indeed beneficial . Since the original model proposed by Sternberg and Horvitz ( 1989 ) , our understanding of the molecular pathways governing vulval fate specification has advanced significantly . In particular , several modes of crosstalk and lateral inhibition between the major signaling pathways specifying the vulval cell fates have been discovered . Here , we report on a dynamic computational model of the more sophisticated understanding of vulval cell fate specification that we have today . In addition , we use model checking to test the consistency of the current conceptual model for vulval precursor cells ( VPC ) fate specification with a large set of observed behaviors and experimental perturbations of the vulval system . The C . elegans vulva is formed by the descendants of three VPCs that are members of a group of six equivalent VPCs named P3 . p–P8 . p ( Figure 1 ) . Each of the six VPCs is capable of adopting one of three cell fates ( termed 1° , 2° , or 3° ) [6–8] . The actual fate a VPC adopts depends upon the integration of two opposing signals that each VPC receives: an inductive signal emanating from the gonadal anchor cell ( AC ) in the form of the LIN-3 epidermal growth factor activates the epidermal growth factor receptor ( EGFR ) /LET-23 in the VPCs . The inductive signal is transduced downstream of the EGFR/LET-23 by the conserved RAS/MAPK signaling cascade to specify the 1° cell fate . In response to the inductive signal , the VPCs produce a lateral signal that counteracts the inductive AC signal in the neighboring VPCs ( lateral inhibition ) by inducing the expression of a set of inhibitors of the EGFR/RAS/MAPK pathway collectively termed lst genes ( for lateral signal targets ) [9–11] . In a second step , the lateral signal induces the 2° cell fate ( lateral specification ) [12] . The lateral signal is encoded by three functionally redundant members of the Delta/Serrate protein family ( dsl-1 , apx-1 , and lag-2 ) and transduced by the LIN-12 / Notch receptor [13 , 14] . Moreover , two functionally redundant inhibitory pathways defined by the synthetic Multivulva ( synMuv ) genes prevent the surrounding hypodermal syncytium ( hyp7 ) from producing the inductive LIN-3 signal , thus allowing the AC to establish a gradient of inductive LIN-3 signal [15] . The fate of the VPCs is influenced by their relative distance from the AC and the lateral signals between the VPCs . The cell closest to the AC ( P6 . p ) receives most of the inductive signal and adopts the 1° fate . The neighboring cells P5 . p and P7 . p receive the stronger lateral signal from P6 . p and hence adopt the 2° fate . The remaining distal VPCs P3 . p , P4 . p , and P8 . p adopt the 3° fate as they do not receive enough inductive signal or lateral signal; and LIN-3 expression in the hyp7 is repressed by the synMuv genes [8 , 15–17] . One remarkable feature of vulval fate specification is its absolute precision . Despite the ability of each cell to adopt any of the three cell fates , the pattern of fates adopted by P3 . p–P8 . p in wild-type animals is always 3°-3°-2°-1°-2°-3° , respectively . This precision is thought to be achieved by multiple modes of crosstalk between the inductive and lateral signaling pathways discovered in recent years [6 , 9–11 , 18 , 19] . Here we use computer simulations and formal verification to investigate whether the known gene interactions are sufficient to produce such patterning precision and to gain insights into the system dynamics . Using the language of Reactive Modules ( RM ) [20] and the Mocha tool [21] , we have constructed a discrete , dynamic , state-based mechanistic model consisting of the key components of the inductive and lateral signaling pathways with their interconnections . By looking analytically at all possible behaviors of a model , we find previously unnoticed dependencies that are present in the data and explained by the model . Specifically , the analysis of our model predicts additional genetic interactions necessary for efficient lateral inhibition and , through the analysis of the behavior of lin-15 mutants , gives new insights into the temporal order of events necessary to achieve a stable pattern of cell fates , which were also validated experimentally .
Models in the language of RM are constructed by defining the objects of the system ( these are the modules ) , and their variables representing semi-independent components of an object . The state of the system is determined by the states of its objects , which in turn are determined by the values of all their variables . Changes in the value of a variable depend on the previous values of the variable and possibly on other variables . A behavior of the system is a sequence of states that the system goes through during execution . Our model consists of a worm module that comprises an AC module and six identical copies of a VPC module ( Figure 2 ) . Additional modules handle the synchronization between VPCs ( i . e . , the scheduler in Figure 2 , which is setting the order of interaction between the VPCs for a particular execution ) and manage the initialization of simulations ( i . e . , the organizer in Figure 2 , which is setting the initial conditions for a particular execution ) . Each VPC module runs its own copy of the same program simultaneously , based on the inputs it receives from its neighboring cells ( AC , hyp7 , and the adjacent VPCs ) ( Figure 3 ) . All VPCs begin with the same conditions determined by the genetic background but may receive different levels of inductive signal depending on their distance from the AC . The AC module contains variables that indicate if the AC is ablated or formed and determine the level of inductive signal sensed by the VPCs according to their distance from the AC . If the AC is ablated , the inductive signal variables in all VPCs are set to the OFF level . If the AC is not ablated , the VPC closest to the AC ( P6 . p ) senses HIGH inductive signal , the next closest ( P5 . p and P7 . p ) sense MEDIUM inductive signal , and the farthest ( P3 . p , P4 . p , and P8 . p ) sense LOW inductive signal ( Figure 3 ) . The VPC module contains variables that represent the behavior of the EGFR/RAS/MAPK pathway , the LIN-12 Notch-mediated lateral signaling pathway , and the lin-15–mediated inhibition of LIN-3 EGF in hyp7 ( Figure 4 ) . In addition , there is a variable for each VPC that follows the temporal progress toward fate acquisition . Each of these variables is now described briefly: The lateral signal variable ( LS ) can be either ON or OFF . The variable starts as OFF and is turned ON upon activation by the EGFR/RAS/MAPK pathway . Once the lateral signal is ON , it is sensed by the immediate neighbors of the respective VPC . The lin-12 variable represents the level of lin-12 / Notch activity . If lin-12 activity is specified as wild-type , its activity level starts as MEDIUM in all VPCs . If lin-12 activity is eliminated [lin-12 ( 0 ) mutation] , then the lin-12 variable is set to OFF ( when using the word set we mean that its value cannot change ) . By contrast , increasing lin-12 activity [lin-12 ( d ) mutation] leads the variable to start as HIGH . Upon activation by the lateral signal , lin-12 activity increases from MEDIUM to HIGH . Upon inhibition of lin-12 activity by the EGFR/RAS/MAPK pathway , lin-12 activity decreases from MEDIUM or HIGH to LOW . The lst genes variable , which can be either ON or OFF , collectively represents the activation state of the lst genes lip-1 , ark-1 , dpy-23 , and lst-1 through lst-4 [9–11] . If lst genes are mutated to an inactive state , the variable is set to OFF . If all lst genes are wild-type , the variable starts as OFF and switches to ON upon activation by lin-12 . We consider all the lst genes either as wild-type or as mutated to an inactive state . The lin-15 variable collectively represents the state of lin-15 and other synMuv genes in hyp7 , which can be either ON or OFF . If lin-15 is wild-type , this variable is set to ON and constitutively inhibits EGFR activation by LIN-3 from hyp7 ( Figure 3 ) . On the other hand , if lin-15 is mutated to an inactive state , the lin-15 variable is set to OFF , and the EGFR/RAS/MAPK pathway is constitutively and uniformly activated in all VPCs . The inductive EGFR/RAS/MAPK pathway is represented by variables describing the status of the following four core components: let-23 ( the EGF receptor ) , sem-5 ( the Grb2-like adaptor ) , let-60 ( the RAS GTP-binding protein ) , and mpk-1 ( the MAP kinase ) . We consider either the wild-type behavior or mutations that completely inactivate a component ( e . g . , let-23 ( 0 ) mutation causing the rest of the pathway never to be activated ) . Before the inductive signal is produced , let-23 egfr is OFF in all VPCs due to the presence of lin-15 and other synMuv genes , which prevent ectopic activation of LET-23 by repressing lin-3 egf expression in hyp7 [15] . Upon receiving the inductive signal from the AC , the variables simulate the activation of let-23 , then sem-5 , then let-60 , and then mpk-1 . The activation by a MEDIUM inductive signal is slower than the activation by a HIGH inductive signal . The EGFR/RAS/MAPK pathway can be counteracted by the lst variable described above during every stage of this activation sequence ( see Figure 3 , middle VPC ) . A simulation starts by setting the type of mutation ( s ) that we would like to examine , and then following an execution of the model by choosing how to schedule the different VPCs . Once the cells assume their fates , we compare the fate assumption versus the desired experimental results . To capture the diverse behavior often observed in biological systems , such as cases where the same genotype leads to different fate patterns , we allow complete freedom in the order of reactions between the different VPCs modules , but restrict the amount of progress each cell makes before its neighbors . The resulting model is highly nondeterministic , allowing many choices of execution without giving priorities or quantities to each choice . Each VPC is treated as a separate process . By adding a mechanism that decides which VPC to advance and for how long , we could get various patterns of VPC fates in different executions . Consequently , the model has approximately 4 , 000 different possible ways to complete one round in which all cells move . Subsequently , there are about 1036 possible executions of the model . In addition , the model has 48 initial states , corresponding to 48 different experimental conditions , and about 92 , 000 different reachable states ( possible assignments to all the variables ) , each corresponding to a snapshot of the system . As the number of possible executions of the model is astronomical , we use formal verification to ensure that all possible runs of the model emanating from a given mutation produce results that match the experimental results . To do that , we have formalized the experimental observations that led to formulate the mechanistic model underlying VPC pattern formation ( e . g . , if the model starts in the wild-type state , the VPCs assume fates according to the following pattern: 3°-3°-2°-1°-2°-3° ) , and used them to formally check whether the mechanistic model reproduces the reported experimental observations . Once we have established a model that reproduces all the experimental data , we can also use simulations to predict the outcome of new experiments that have not been performed yet . In nondeterministic models , simple simulations ( i . e . , testing ) are not sufficient to verify the model's consistency with the experimental data . The reason is that in nondeterministic models the number of possible behaviors resulting form the same initial condition could be enormous . Therefore , to test a nondeterministic model , one would have to run many simulations ( one for each scenario ) . Another way to test nondeterministic models is to use model checking [4] , which allows us to formally check all the different executions of the system against a formal specification . By exploring all the possible states and transitions of a system , we can determine whether some property holds true for the system . In the case that the property does not hold , the model-checking algorithm supplies a “counterexample , ” which is an execution of the system that does not satisfy the given requirement , in the current case an experimentally observed pattern of vulval cell fates . Here , we have used model checking for two purposes . First , to ascertain that our mechanistic model reproduces the biological behavior observed in different mutant backgrounds . For that , we have formalized the experimental results described in a set of papers ( for references see Table 1 ) and verified that all possible executions satisfy these behaviors . That is , regardless of the order of interactions from a given set of initial conditions , different executions always reproduced the experimental observations . Second , we used model checking to query the behavior of the model . By phrasing queries such as which mutations may lead to a stable or an unstable fate pattern , we analyze the behavior of the model . Once an unstable mutation was found , we determined what part of the execution allows this kind of mutation by disallowing different behavioral features of the model and checking when the instability disappears . We have tested the behavior of our model for a set of 48 perturbations corresponding to 24 mutant combinations , which were analyzed in the presence and absence of the AC ( Table 1 ) . For some of the combinations , the outcome has been tested experimentally as indicated in Table 1 by the respective references , but many other combinations have not been tested experimentally , as some of the double , triple , or quadruple mutants might be technically very difficult to generate . For example , complete loss-of-function mutations in most components of the inductive signaling pathway cause early larval lethality , or homozygous lin-12 ( gf ) mutants lack an AC . Forty-four of the 48 conditions tested yielded a stable fate pattern , as all possible executions gave the same result . All four conditions leading to an unstable pattern included the lin-15 knockout mutation . The cause of the unstable pattern in these four cases is discussed in the next section . Twenty-two of the conditions have been tested experimentally and the observed results are reported in the literature ( see references in Table 1 ) . Our model faithfully reproduces the predominant cell fate patterns that had been reported except for the phenotype of lin-12 ( d ) ; lin-15 ( 0 ) double mutants . While in lin-12 ( d ) ; lin-15 ( 0 ) animals , the distal VPCs ( P3 . p , P4 . p , and P8 . p ) adopt either a 1° or a 2° cell fate [6] , our model predicted that the six VPCs would always adopt a 2° fate . This discrepancy was traced to the fact that the high activity of LIN-12 simultaneously induced lst expression in all VPCs , which immediately repressed the transduction of the EGFR signal that was activated in all VPCs due to the lin-15 ( 0 ) mutation ( Figure 5A ) . The high levels of lst gene expression thus prevented the cells from engaging the mechanisms reducing LIN-12 activity , which is necessary for a 1° fate specification . In spite of having represented LIN-12 downregulation by EGFR signaling [18] and lst-mediated lateral inhibition on EGFR signaling [9 , 10] , the model could not reproduce the experimental observations . Therefore , we postulated that some additional regulation is needed to allow primary fates while avoiding adjacent primary fates [22] . One possibility is that EGFR signaling downregulates one or several lst genes in addition to inducing lin-12 degradation ( Figure 5B ) . If this happens before the activation of the lst genes completely blocks EGFR/RAS/MAPK signaling , then at least some VPCs are allowed to adopt a 1° fate . We further suggest that in order to avoid adjacent primary cells in these lin-12 ( d ) ; lin-15 ( 0 ) mutants , the lateral signal can still override the EGFR signal and lst activity prevails . These insights led to a revised model with at least one additional negative feedback loop indicated by the red line in Figure 3 . This refined model reproduces all the experimentally observed cell fate patterns including the lin-12 ( d ) ; lin-15 ( 0 ) double mutants ( Table 1 , rows 21 and 45 ) . Of particular interest are those cases where two signaling pathways specifying different cell fates are simultaneously perturbed . For example , if the lateral signal is constitutively activated and at the same time the transduction of the inductive signal is blocked , then all VPCs are predicted to adopt the 2° cell fate irrespective of the presence or absence of the AC ( Table 1 , rows 19 and 43 ) . Indeed , in lin-12 ( n137gf ) mutants that carry a dominant-negative or strong reduction-of-function mutation in let-60/Ras , all VPCs were found to adopt the 2° cell fate [23] . In another condition we examined the interaction between the inhibitory lin-15 pathway and the lst genes . If both components are inactivated at the same time , all the VPCs are predicted to adopt a 1° cell fate in the presence as well as in the absence of the AC ( Table 1 , rows 6 and 30 ) . As predicted by modeling , in the majority of lin-15 ( n309 ) ; lip-1 ( zh15 ) double mutants , adjacent VPCs adopt the 1° cell fate indicated by the expression of the 1° fate marker egl-17::gfp and by morphological criteria ( [9] and unpublished data ) . An example for a condition that could not be tested experimentally is shown in Table 1 , row 38 . If all three signals , the inductive and lateral signals as well as the lin-15–mediated inhibition of hyp7 , are inactive and the lst genes are mutated , all six VPCs are predicted to adopt the 1° cell fate as long as the EGFR/RAS/MAPK pathway is functional . This suggests that the default fate in the vulval equivalence group is 1° . Conversely , if the inductive and lateral signaling pathways are both constitutively activated ( “lin15kolin12d” in Table 1 , row 21 ) , then the VPCs may adopt a 1° or 2° fate depending on the activity state of the lst genes ( Table 1 , rows 21 , 22 , 45 , and 46 ) . In summary , by using model checking to compare our executable model with existing experimental data , we can predict novel interactions in the regulatory network governing vulval fate specification . In addition , analysis of the model allows us to predict the outcome of perturbations that are difficult to test experimentally . Using model checking , we found that 44 out of 48 perturbations affecting vulval development lead to a stable fate pattern , despite the vast number of possible executions of our model . The only four mutations leading to unstable patterns are lin-15 ( 0 ) , lin-15 ( 0 ) ; lin-12 ( d ) , lin-15 ( 0 ) ; ac- and lin-15 ( 0 ) ; lin-12 ( d ) and ac- ( Table 1 , rows 15 , 21 , 29 , and 45 ) . To determine whether variations in the exact timing of the lateral signaling are the cause of this instability , we asked , using model checking , whether it is possible to get an unstable fate pattern without allowing variations in the timing of the lateral signal and found this not to be the case . We discovered that in order to adopt two different cell fates in two different executions , a VPC has to send the lateral signal before its neighbors in one execution , and after its neighbors in another execution . In the first case , the VPC will adopt a 1° fate and force its neighbors to adopt a 2° fate , while in the second case it is forced by one of its neighbors to adopt a 2° fate before it can adopt the 1° fate . Specifically , we found that in the four cases containing the lin-15 ( 0 ) mutation , perturbation of the intricate timing dependency between the activation of the lateral signal and the inhibition of LIN-12 activity by the EGFR/RAS/MAPK pathway allows VPCs to adopt different fates in different executions of the model . Figure 6 distinguishes between stable and unstable fate patterns according to the ordering of events derived from the analysis of our model . In a stable pattern , the response to the inductive signal is temporally graded in a way that allows one VPC ( e . g . , VPC1 in Figure 6A ) to send the lateral signal always before its neighbors reduce their level of LIN-12 . In unstable patterns , on the other hand , the activation of the EGFR/RAS/MAPK pathway occurs more or less simultaneously in all VPCs , and small , stochastic timing differences result in variable patterns among genetically identical animals ( Figure 6B ) . We note that this instability comes into effect only in AC-ablated animals or in VPCs that are too distant from the AC , suggesting that the AC organizes not only the spatial but also the temporal order of events . To test the predictions made by our model , we examined the expression of cell fate-specific transcriptional reporters in developing animals . Using a strain carrying both the egl-17::cfp and lip-1::yfp transgenes as reporters for the 1° and 2° cell fate , respectively , we could simultaneously observe the activation of the inductive and lateral signaling pathways in the VPCs of individual animals . We first performed a time-course analysis in a wild-type background and quantified the strength of the fluorescent signals of the 1° and 2° fate-specific markers during the critical phase from the mid L2 stage on ( 22 h after starvation-induced L1 arrest ) until the end of the L2 stage just before the VPCs have adopted their fates and start dividing ( 28 h after starvation-induced L1 arrest ) . In all animals except for one case at the 25-h time point , an increase in the expression of the 1° fate marker egl-17::cfp was observed in P5 . p , P6 . p , and P7 . p before a significant upregulation of the 2° fate marker lip-1::yfp occurred in P5 . p and P7 . p ( Figure 7A , 7B , and 7D ) . Thus , the inductive signaling pathway is activated already in mid-L2 larvae ( +22 h ) , while lateral signaling is effective only toward the end of the L2 stage ( +28 h ) . These experimental data provide , for the first time to our knowledge , direct evidence for a sequential activation of the inductive and lateral signaling pathways during vulval induction , as predicted by our model in the case of stable fate patterns . It should be noted that mosaic analysis of let-23 egfr had already suggested a sequential model for vulval fate specification [24] , though the relative timing of the inductive versus lateral signaling events has to date not been investigated . Next , we tested if in lin-15 ( n309 ) mutants that exhibit an unstable fate pattern in the distal VPCs the sequential activation of signaling pathways may be disrupted . Since larval development in lin-15 ( n309 ) animals is significantly delayed ( unpublished data ) , it was not possible to perform the same time-course analysis as shown above for wild-type animals . We therefore staged lin-15 ( n309 ) animals carrying the egl-17::cfp and lip-1::yfp reporters based on the length of their posterior gonad arms and the shape of the VPCs to identify late L2 larvae corresponding approximately to the +28 h time point in wild-type larvae ( see Materials and Methods ) . In 12 out of 22 late L2 lin-15 ( n309 ) larvae , the 1° and 2° fate markers were simultaneously expressed in at least one of the distal VPCs , P3 . p , P4 . p , or P8 . p , which is consistent with the unstable fate pattern predicted for the distal VPCs in lin-15 mutants ( Figure 7C and 7F ) . Moreover , in 21 out of 22 lin-15 ( n309 ) animals , P5 . p and/or P7 . p expressed both 1° and 2° fate markers ( Figure 7F ) . In wild-type animals , on the other hand , co-expression of the 1° and 2° fate markers was never observed in the distal VPCs , but 12 out of 21 animals showed weak 1° fate marker expression in P5 . p and/or P7 . p in addition to the strong 2° marker expression ( Figure 7E ) . Thus , we could experimentally confirm two key predictions provided by our modeling work ( Figure 6 ) ; the temporal gradient in the activation of the inductive and lateral signaling pathways in wild-type animals and the loss of sequential signaling in lin-15 mutants leading to an unstable fate pattern .
Formal executable models have become valuable tools to enhance our understanding of complex biological systems [1 , 3 , 25–30] . Here , we present an up-to-date comprehensive model of C . elegans vulval fate specification and experimental validation of two key predictions made by the model . Our model represents the current understanding of the regulatory signaling network and includes multiple modes of crosstalk between the EGFR/RAS/MAPK and NOTCH signaling pathways such as the LIN-12 / Notch-mediated lateral inhibition [9 , 10 , 22] . Since the model is dynamic and nondeterministic , it allows a very large number of different executions for a given starting condition . By using model checking , which permits us to investigate all possible executions of the model , we identify gaps in the conceptual understanding of the events leading to a stable pattern of vulval cell fates . The insights gained through model checking can then be used to refine an initial model until it fits all the experimental data . There could be several different ways to refine a model , and every conjecture made in the refinement process should then be validated experimentally . For example , our model suggests that the EGFR/RAS/MAPK pathway not only represses lin-12/Notch signaling [18] but also negatively regulates lst gene expression in 1° cells . In the 2° cells , on the other hand , lateral signaling overrides this postulated negative loop and lst activity prevents 1° cell fate specification . Although such a molecular mechanism has not yet been elucidated , our modeling study makes explicit the importance of this putative negative feedback loop . Interestingly , it was previously reported that some lst genes are not only positively regulated by lateral signaling but are also negatively regulated by inductive signaling [10] . In addition , the recently discovered homolog of the mammalian tumor suppressor dep-1 gene might also be part of this postulated negative feedback loop [31] . DEP-1 dephosphorylates the EGFR and thereby inhibits inductive signaling in the 2° cell lineage in parallel with the lst genes , while inductive signaling simultaneously downregulates DEP-1 and LIN-12/Notch expression in the 1° cell lineage , allowing full activation of the EGFR in these cells [18 , 31] . Thus , the reciprocal activation of EGFR/RAS/MAPK signaling and lateral inhibitors in 1° and 2° VPCs , respectively , might in part be mediated by a novel negative feedback loop downstream of the MAP kinase . In mammals , the negative crosstalk between EGFR and Notch signaling may be important to control the balance between stem cell proliferation and differentiation [32] , and alterations in the connections between these two signaling pathways may lead to cancer in humans [11] . Thus , future studies investigating the molecular details of negative feedback loops between EGFR signaling and the lst genes may help elucidate conserved mechanisms underlying EGFR function as an oncogene . Our computational model allows flexibility in the order between different reactions , which resembles variations in the rate of biochemical reactions . This is akin to the robustness of simple biochemical networks that are resistant to variations in their biochemical parameters [33–35] . Despite this variability , we have found by model checking that in a wild-type situation all possible executions reach a stable state , independently of the order of reactions between the VPCs . This behavior of the model closely resembles the remarkable robustness of vulval development observed under various experimental conditions in the laboratory as well as in free living Nematodes . Furthermore , we observed that for most perturbations ( i . e . , mutations in the inductive or lateral signaling pathways ) , all the different executions lead to stable fate patterns . This suggests that the mechanism underlying VPC specification is relatively resistant to genetic variability and might therefore represent a process subject to high selective pressure . A notable exception is the behavior of lin-15 ( lf ) mutants , which—both experimentally and by modeling—exhibit an unstable fate pattern as long as the inductive signaling pathway is functional . We could trace down the cause of this instability to the fact that lin-15 mutations abrogate the temporal order in the activation of the inductive versus lateral signaling pathway among individual VPCs . Interestingly , recent experiments have demonstrated that lin-15 ( lf ) mutations result in the ectopic expression of the inductive LIN-3 EGF signal in the hypodermal syncytium hyp7 [15] . Since all VPCs are in direct contact with hyp7 , it seems reasonable to assume that in lin-15 mutants the EGFR is simultaneously activated in all VPCs , which likely disrupts the relative timing of inductive versus lateral signaling among adjacent VPCs . Our analysis of the behavior of lin-15 mutants thus illustrates how computational modeling can provide a plausible mechanistic explanation for phenotypic instability observed in real life . Through model checking an executable model representing the crosstalk between EGFR and LIN-12 / Notch signaling during C . elegans vulval development , we have gained new insights into the usage of these conserved signaling pathways that control many diverse processes in all animals . While many modeling efforts use simulations that allow us to investigate only a few possible executions , our work emphasizes the power of analyzing all possible executions using model checking . Previous attempts to use model checking in biological modeling have concentrated on adapting model checking to formalisms such as differential equations and probabilistic modeling [36–38] . Our work demonstrates how biological processes can be described and analyzed with the use of formal methods , which enhances our comprehension of complex biological systems . We suggest that combining model-checking analysis with high-level modeling , similar to the level of abstraction used by biologists in describing mechanistic models , can help in many areas of biology to obtain more accurate , formal , and executable models , eventually leading to better understanding of biological processes .
RM is a modeling language for reactive systems [20] . RM is designed to describe systems which are discrete , deadlock-free , and nondeterministic . The elementary particles in RM are variables . We describe the behavior of variables in atoms and combine atoms into modules . Modules can be combined to create more complicated modules ( including combinations of several copies of the same module ) . Each variable ranges over a finite set of possible values . An atom describes the possible updates on variables . An atom can be synchronous , meaning that it updates the variables it controls in every step of the system , or asynchronous , meaning that it updates the variables it controls from time to time . An update of a variable may depend on the value of itself as well as the values of other variables . There can also be dependencies between the mutual update of several variables in the same step . RM enables nondeterminism by allowing multiple overlapping update options . The current RM model does not include probabilities . Mocha is a software tool for the design and analysis of RM [21] . Mocha can simulate a model by following step-by-step evolution of the variables in the model . Simulations show the sequence of values assumed by variables during the simulation . In simulation of nondeterministic models , the user is expected to choose the next step between different nondeterministic options . The simulation engine can highlight the variable values that lead to the assignment of a certain value . Mocha supports invariant model checking directly ( to check that all reachable states satisfy some property that relates to the values of variables in the state ) , as well as model checking of safety properties using monitors ( to check that all executions satisfy some property ) . We use both enumerative and symbolic ( using Boolean Decision Diagrams , BDDs ) model checking; the difference between the two has to do with their performance in practice . Counterexamples are presented as sequences of variable values . Parallelism is an important property of biological systems . In computer science , this is referred to as concurrency ( processes running in parallel and sharing common resources ) . We usually distinguish between two forms of concurrency: synchronous and asynchronous . In synchronous systems , all components move together . That is , there is some basic work unit that all components share . All components do one work unit in parallel simultaneously . Then they all move to the next unit . In asynchronous systems , every component moves separately . Usually , in asynchronous systems , we do not allow components to move together and we cannot guarantee the relative speed of different components . Biological systems , while highly concurrent , are neither completely synchronous nor completely asynchronous . Different molecules , or cells , do not progress in perfect lockstep , and neither does any molecule or cell rest for arbitrary amounts of time . For this reason , we have introduced a new notion of bounded-asynchrony into our computational model . In bounded-asynchrony , the scheduler , which chooses the next component to move , is not completely free in its choices . No component can be neglected more than a bounded number of times . This captures the phenomenon that the components of a biological system ( say , molecules or cells , depending on the level of modeling granularity ) progress neither in lockstep nor completely independently , but that they are loosely coupled and proceed approximately along the same timeline . We find the notion of bounded asynchrony a pragmatic way to model cell–cell interactions in an abstract discrete framework . Further studies are needed to identify the appropriate model for concurrency in different biological contexts . Standard methods were used for maintaining and manipulating C . elegans [39] . The C . elegans Bristol strain , variety N2 , was used as the wild-type reference strain in all experiments . Mutations used: lin-15 ( n309 ) [22]; integrated transgene arrays used: arIs92[egl-17::cfp , tax-3::gfp] [10] , mfIs42[lip-1::yfp] ( gift of M . A . Félix ) . Synchronized populations of L1 larvae were obtained by isolating embryos from gravid adults using sodium hypochlorite treatment and arresting the newly hatched larvae by food starvation . The arrested L1 larvae were then placed on standard NGM growth plates containing E . coli OP50 and collected for microscopic observation at the indicated time points . For observation under Nomarski optics , animals of the indicated stages were mounted on 4% agarose pads with M9 buffer containing 10 mM sodium azide . Fluorescent images were acquired on a Leica DMRA wide-field microscope equipped with a cooled CCD camera ( Hamamatsu ORCA-ER , http://www . hamamatsu . com/ ) controlled by the Openlab 3 . 0 software package ( Improvision , http://www . improvision . com/ ) . For quantification of YFP and CFP intensity in the VPCs , all images were acquired with the same microscopy , camera , and software settings using YFP- and CFP-specific filter sets . The mean intensity of CFP and YFP expression in the nuclei of the VPCs was measured using the measurement tool in the Openlab 3 . 0 software package ( Improvision ) , and each measurement was standardized to the background intensity in the same picture . For each time point , between ten and 12 animals were quantified . Late L2 lin-15 ( n309 ) animals were identified by selecting larvae in which the VPCs had adopted an oval shape and the distal tip of the posterior gonad arm had migrated past P7 . p ( see Figure 7B and 7C ) . | Systems biology aims to gain a system-level understanding of living systems . To achieve such an understanding , we need to establish the methodologies and techniques to understand biological systems in their full complexity . One such attempt is to use methods designed for the construction and analysis of complex computerized systems to model biological systems . Describing mechanistic models in biology in a dynamic and executable language offers great advantages for representing time and parallelism , which are important features of biological behavior . In addition , automatic analysis methods can be used to ensure the consistency of computational models with biological data on which they are based . We have developed a dynamic computational model describing the current mechanistic understanding of cell fate determination during C . elegans vulval development , which provides an important paradigm for studying animal development . Our model is realistic , reproduces up-to-date experimental observations , allows in silico experimentation , and is analyzable by automatic tools . Analysis of our model provides new insights into the temporal aspects of the cell fate patterning process and predicts new modes of interaction between the signaling pathways involved . These biological insights , which were also validated experimentally , further substantiate the usefulness of dynamic computational models to investigate complex biological behaviors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"caenorhabditis",
"cell",
"biology",
"developmental",
"biology",
"computational",
"biology"
] | 2007 | Predictive Modeling of Signaling Crosstalk during C. elegans Vulval Development |
All eukaryotes have the ability to detect and respond to environmental and hormonal signals . In many cases these signals evoke cellular changes that are incompatible and must therefore be orchestrated by the responding cell . In the yeast Saccharomyces cerevisiae , hyperosmotic stress and mating pheromones initiate signaling cascades that each terminate with a MAP kinase , Hog1 and Fus3 , respectively . Despite sharing components , these pathways are initiated by distinct inputs and produce distinct cellular behaviors . To understand how these responses are coordinated , we monitored the pheromone response during hyperosmotic conditions . We show that hyperosmotic stress limits pheromone signaling in at least three ways . First , stress delays the expression of pheromone-induced genes . Second , stress promotes the phosphorylation of a protein kinase , Rck2 , and thereby inhibits pheromone-induced protein translation . Third , stress promotes the phosphorylation of a shared pathway component , Ste50 , and thereby dampens pheromone-induced MAPK activation . Whereas all three mechanisms are dependent on an increase in osmolarity , only the phosphorylation events require Hog1 . These findings reveal how an environmental stress signal is able to postpone responsiveness to a competing differentiation signal , by acting on multiple pathway components , in a coordinated manner .
Eukaryotic cells commonly employ mitogen activated protein kinases ( MAPKs ) to transduce extracellular signals and evoke intracellular responses [1] . MAPKs are a part of an evolutionarily-conserved three-tiered signaling cascade comprised of the MAPK , a MAPK kinase ( MAPKK ) , and a MAPKK kinase ( MAPKKK ) . In mammalian cells MAPKs respond to diverse stimuli including hormones , stresses , and cytokines . These different stimuli will in many cases activate a common MAPK . Conversely a single stimulus will often activate multiple MAPKs . Understanding how each stimulus and each response is coordinated is often obscured by the large number of components and the functional complexity of signaling networks [2] . MAPK pathways are also present in the unicellular eukaryote Saccharomyces cerevisiae ( hereafter , yeast ) . As in higher eukaryotes , yeast use multiple MAPK pathways to respond to a variety of environmental signals [3] . The two best-characterized examples are the mating pathway and the high osmolarity glycerol ( HOG ) response pathway ( detailed in Figure 1A ) [4] , [5] . The mating pathway operates through a cell-surface receptor that activates a canonical G protein heterotrimer . The activated G protein recruits Ste5 , a scaffold protein that assembles and activates three component kinases: Ste11 , Ste7 and the MAPK Fus3 [6] , [7] . Active Fus3 promotes events leading to cell fusion including new gene transcription , cell cycle arrest and cytoskeletal rearrangements [8]–[10] . High osmotic stress activates Ste11 as well as Pbs2 and the MAPK Hog1 [11] . Active Hog1 promotes events leading to stress adaptation including increased glycerol production , cell cycle arrest and a pause in protein translation [12]–[17] . Individually , the two pathways have well-defined components , known points of regulation , and established measures of pathway output . Together , the pathways form a signaling network that is a model for the study of signal coordination . The mating and HOG pathways share several components , yet exhibit remarkable signal fidelity when stimulated individually [3] , [18] ( Figure 1A , shared components highlighted in green ) . Hyperosmotic stress does not activate Fus3 or promote mating , and mating pheromones do not activate Hog1 or the HOG pathway . Such pathway fidelity may be maintained by two mechanisms: ( i ) pathway insulation and ( ii ) pathway cross-inhibition [19] . The pathway insulation model proposes that physical sequestration of components maintains specificity . For example , Ste11 exists in two scaffolded pools , one that selectively activates Fus3 and another that selectively activates Hog1 [20] . The pathway cross-inhibition model proposes that one pathway inhibits signaling by the competing pathway . For example , Hog1 is required to prevent the inadvertent activation of the mating response by hyperosmotic stress . When Hog1 is absent , or rendered catalytically inactive , hyperosmotic stress promotes mating . Thus it appears that Hog1 targets a component of the mating pathway to maintain fidelity [21]–[24] . However , previous studies were unsuccessful in identifying the substrate ( s ) of Hog1 in the mating pathway . Thus the mechanisms that prevent cross-talk remain unresolved . A related and potentially more tractable question is how cells coordinate responses when the mating and HOG pathways are activated simultaneously . To address this question investigators have treated cells simultaneously with mating pheromone and hyperosmotic stress and used pathway-specific transcription reporters to monitor signaling in individual cells [24] , [25] . One group reported that the responses to these inputs are mutually exclusive [25] . However a subsequent analysis identified a potential artifact , wherein cell death can produce a spurious signal in reporter assays that employ the red fluorescent protein [24] . In surviving cells reporters of both pathways are activated in proportion to their respective stimuli . Thus a single cell can respond to both hyperosmotic stress and pheromone , but how these responses are prioritized or coordinated remains to be determined . In this study we establish that the hyperosmotic stress and mating pheromone signals are coordinated . Using a broad array of activity assays , conducted over various time scales , we show that Hog1 delays and dampens the response to pheromone and does so by two distinct mechanisms: ( i ) negative feedback phosphorylation of a shared component ( Ste50 ) and ( ii ) feed-forward phosphorylation of a negative regulator of translation ( Rck2 ) . Thus , activated Hog1 invokes pathway cross-inhibition to delay the mating differentiation response . Mating differentiation resumes once cellular osmotic balance is restored and cross-inhibition is relieved . These studies provide a model of how a cell integrates competing signals to control cell fate .
A hallmark of the mating response is the appearance of a mating projection ( shmoo formation ) , which functions as the eventual site of cell-cell fusion [26] . A hallmark of the osmotic stress response is a rapid but transient reduction in cell volume . This reduction occurs as water leaves the cell in order to equalize internal and external osmolarity . The cell then ramps up glycerol production to restore osmotic balance and cell volume [27] . These signaling pathways are likely to be coordinated as it was reported previously that a decrease in extracellular osmolarity disrupts efficient cell-cell fusion [28] . Here we investigated how an increase in extracellular osmolarity impinges on processes leading to fusion . Recent publications have examined the cell response following co-stimulation with pheromone and hyperosmotic stress , but these papers reached opposing conclusions [24] , [25] . Moreover , the authors of the second paper conclude that the pathway insulation model is operative , but base their conclusion on the absence of evidence for the pathway cross-inhibition model . Both reports relied primarily on transcription-reporter assays conducted over a limited time scale . However , as detailed herein , hyperosmotic stress conditions can have confounding effects on transcription-reporter activity , particularly at early time points . Both the mating and HOG pathways can be activated in a single cell [24] . Because both pathways can be activated simultaneously , it is evident that cross-inhibition does not operate between these pathways . However , several key questions remain unanswered . Most importantly , do cells mate normally under osmotic stress , or is the mating program delayed by the osmotic stress response ? If so , how is the mating program delayed when both pathways can be activated simultaneously at the transcriptional level as shown previously [24] ? Accordingly , we determined the effect of co-stimulation on multiple events and over a period of several hours; these events include MAPK activation , transcription induction , protein expression , cell differentiation and cell fusion . We first examined if hyperosmotic stress interferes with mating . To this end we performed quantitative mating assays in the absence and presence of an osmolyte ( 0 . 5 M KCl ) . As shown in Table 1 , hyperosmotic conditions decrease mating efficiency by about 64% after a 4-hour mating period . The observed decrease may be caused by events during stress adaptation that postpone mating . To allow stress adaptation we extended the mating period to 24-hours . In this case we observed a more modest 8% decrease in mating efficiency . These results suggest that cells are capable of efficient mating during hyperosmotic conditions , but only after a period of adaptation . Mating requires that cells undergo G1 arrest and the formation of a mating projection . Thus we next investigated how hyperosmotic stress affects pheromone-induced shmoo formation . We stimulated MATa cells with a saturating concentration of mating pheromone ( α factor ) , or co-stimulated cells with pheromone and KCl . We then visualized and quantified the appearance of shmoos over time by microscopy ( Figure 1B and Videos S1 , S2 , S3 ) . As shown in Figure 1B , the addition of mating pheromone resulted in detectable shmoo formation by 60 minutes , with 60% of cells forming shmoos by 180 minutes . The simultaneous addition of osmolyte resulted in detectable shmoo formation only after 120 minutes , with just 20% of cells forming shmoos by 180 minutes . Addition of higher concentrations of osmolyte , 0 . 75 M ( Figure 1B ) or 1 M KCl ( data not shown ) , further delayed shmoo formation . The duration of delay is likely a function of the time needed for cells to adapt , and could also account for the delay in mating noted above . The data presented above reveal that salt stress delays shmoo formation and diminishes mating efficiency . We then considered whether there was a delay in other aspects of the pheromone response . To this end we monitored Far1 . Far1 is induced by pheromone only during the G1 phase of the cell cycle , is required for cell polarization during mating , and is quickly degraded as cells exit G1 [29]–[32] . Thus , Far1 is a broad indicator of cellular events leading up to mating . Addition of mating pheromone alone resulted in detectable Far1 by 60 minutes , while co-stimulation with 0 . 75 M KCl delayed the appearance of Far1 to 120 minutes ( Figure 1C ) . These findings indicate that the delay in shmoo formation corresponds with a delay in Far1 induction . Thus osmotic stress triggers a delay in the mating response , and this delay is evident at the molecular level as well as at the level of cellular morphogenesis and mating . Co-stimulation of cells has previously been associated with cytotoxicity [24] . Under our experimental conditions however , nearly all cells survived co-stimulation and were able to resume cell division , as shown in Figure 1B , Videos S2 and S3 . We also quantified cell viability using methylene blue staining . By this approach we observed cytotoxicity in ∼6% of the population after 2 hours of co-stimulation with KCl and pheromone ( n = 2 , 632 ) . These results indicate that the delay in Far1 induction , and the corresponding delay in morphogenesis and mating , is not the result of cell death . In contrast about one-third of cells co-stimulated with sorbitol and pheromone did not survive , as reported previously [24] . Hyperosmotic stress activates Hog1 and induces genes required for adaptation [33] . During the immediate response to stress however , there is transient repression ( <5 min ) of overall gene transcription [34] . Moreover the duration of the delay correlates with the concentration of osmolyte and is prolonged in cells that lack Hog1 [34] , [35] . Thus , transcription is regulated by Hog1-dependent and Hog1-independent mechanisms . We postulated that hyperosmotic stress might delay mating in part through a transient repression of transcription . Indeed , we have already shown that Far1 expression is delayed by salt stress; however Far1 abundance is also subject to stimulus-dependent ubiquitination and degradation [30] . To focus specifically on mating gene induction we used a reporter comprised of the β-galactosidase gene fused to the FUS1 promoter ( FUS1-lacZ ) . The FUS1 gene is among the most strongly induced genes during the mating response [9] . As shown in Figure 2 , cells stimulated with mating pheromone reached half maximum ( t½ max ) β-galactosidase activity at roughly 50 minutes ( Figure 2A and Table S1 ) . The addition of 0 . 75 M KCl increased the t½ max by ∼67% and dampened the maximum response by ∼17% . The effects of salt were dose-dependent; with increasing concentrations the delay and dampening became progressively more pronounced . To distinguish Hog1-dependent and Hog1-independent effects on the transcription response , we measured FUS1 induction in cells lacking HOG1 ( hog1Δ ) as well as in cells expressing a catalytically deficient mutant , Hog1K52R . These cells were then stimulated with pheromone , or co-stimulated with pheromone and KCl . Similar to wild-type cells , co-stimulation of hog1Δ cells increased the t½ max for FUS1 induction ( Figure 2B and Table S2 ) . However , unlike wild-type cells , co-stimulation of hog1Δ cells did not dampen the maximum response . Thus Hog1 contributes to the reduction in transcription response . Also , the change in t½ max was less pronounced in hog1Δ cells compared to wild-type cells , suggesting that Hog1 is at least partly responsible for the delay in mating transcription ( Table S2 ) . Hog1K52R cells showed an intermediate t½ max and increased maximum response ( Figure 2C and Table S3 ) . This result was anticipated given that Hog1K52R cells exhibit an intermediate level of cross-inhibition and sensitivity to osmotic stress [22] , [23] , [36] . The non-ionic osmolyte sorbitol was 5-fold more likely to cause cell death but otherwise acted much like KCl ( Figure S1 and Table S1 ) . Taken together these results support the view that osmotic stress attenuates mating transcription , and does so by Hog1-dependent and Hog1-independent mechanisms . Moreover , no additional effects associated with hyperosmotic stress were observed on the mating response ( Text S1 , Figure S2 , Table S4 , and Figure S3 ) . The transcriptional reporter FUS1-lacZ measures average differences in a population of cells . However this approach could mask larger differences within a subpopulation of cells , such as those in G1 phase where mating occurs [37] . To determine whether salt diminishes the mating response in a cell cycle dependent manner , we monitored transcription activity in single cells using a green fluorescent protein-based reporter ( FUS1-GFP ) . To avoid any confounding effects of cell cycle-arresting agents , we specifically examined unbudded ( G1 phase ) cells in an otherwise asynchronous population . As shown in Figure 2D salt delays both GFP production and shmoo formation . Measurements of average pixel intensity among single-cells showed that the distribution of responding cells was uniform whether stimulated with pheromone or co-stimulated with pheromone and KCl ( Figure 2E ) . Moreover , the salt mediated delay observed for cells in G1 phase ( ∼18 min ) was similar to that of an asynchronous cell population ( ∼12 min ) ( Figure 2A , 2E and Tables S1 , S5 ) . Taken together , single-cell measurements corroborate the observations made using the population-based reporter for mating pathway output . Mating pheromones activate Fus3 and induce the transcription of genes required for haploid cell fusion . We have observed that Hog1 dampens and delays the mating transcription response . To determine how Hog1 limits the activation of Fus3 we monitored its activity directly , by immunoblotting with an antibody that recognizes the dually-phosphorylated , fully-active form of the kinase ( phospho-Fus3 ) [38] . As shown in Figure 3 , co-stimulation with KCl reduced phospho-Fus3 by one-third compared to cells treated with pheromone alone . Pheromone also induces the expression of the FUS3 gene [9] , [39] . To determine the effect of KCl on Fus3 production we quantified Fus3 protein levels with a Fus3-specific antibody . As with phospho-Fus3 , total Fus3 was reduced by one third in co-stimulated cells . Thus hyperosmotic stress leads to dampened induction of Fus3 and a concomitant reduction in phospho-Fus3 . We then conducted the same experiment in cells lacking Hog1 ( Figure 3B ) . In this case , we found no effect of salt co-stimulation on phospho-Fus3 or Fus3 . These data indicate that Hog1 regulates mating by dampening Fus3 production and , consequently , Fus3 activity . Thus Hog1 has a role in limiting gene induction and mating , and may do so by targeting a component downstream of Fus3 . Fus3 is part of a positive feedback loop: the activation of Fus3 by mating pheromone leads to induction of more Fus3 , which is subsequently activated by pheromone . We have shown above that Hog1 acts downstream of Fus3 , by limiting induction of the protein . We then considered whether Hog1 acts upstream of Fus3 , by limiting activation of the kinase . To exclude effects of hyperosmotic stress on Fus3 induction we replaced the native ( pheromone-inducible ) promoter with the galactose-regulated GAL1 promoter . As expected , we found that cells grown in galactose stably express Fus3 , with no induction in the presence of pheromone . However , co-stimulation with salt reduced phospho-Fus3 to nearly one-half of that in cells treated with pheromone alone , even as Fus3 abundance remained unchanged ( Figure 4A ) . In contrast , co-stimulation did not alter phospho-Fus3 in the absence of Hog1 ( Figure 4B ) , except for a small reduction at the earliest ( 5 min ) time point . The effect of salt at 5 minutes was reported by others to be independent of Hog1 . Specifically , cells lacking Hog1 require more time to reach ionic equilibrium [34] . Therefore the reduction seen at 5 minutes in hog1Δ cells is likely associated with the extended time required for adaptation to the mechanical and ionic stress caused by cell shrinking . Thus hyperosmotic stress dampens Fus3 activation . The reduction at early time points is evident with or without Hog1 , while the reduction at later time points ( 30 and 60 min ) is Hog1-dependent ( Figure 2 and Figure 4 ) . Hog1 is activated rapidly following salt stress and then becomes inactive once the cells have adapted [14] . Fus3 is activated by pheromone , but activation in this case is delayed as long as Hog1 is active . Thus it appears that Fus3 cannot fully respond to pheromone until the cells have adapted to osmotic stress conditions . To investigate this behavior further we treated cells with KCl for various times , followed by treatment with pheromone for 15 minutes . Once again we observed that Fus3 activity is restricted as long as Hog1 is active ( Figure 4C ) . Taken together , our findings indicate that Hog1 regulates mating at two points in the pathway , one downstream of Fus3 that limits protein induction and another upstream of Fus3 that limits kinase activation . Cells utilize Hog1-dependent and Hog1-independent mechanisms to adapt to hyperosmotic stress . To focus exclusively on Hog1-dependent mechanisms we activated the kinase directly , without an osmolyte . To this end we introduced a constitutively active MAPKKK , Ssk2ΔN [40] , [41] . Ssk2 is a component of the SLN1-branch of the HOG pathway and is not shared with the mating pathway . Thus expression of Ssk2ΔN activates Hog1 but does not affect Fus3 directly . First , we measured the effect of constitutively-activated Hog1 on pheromone-activated Fus3 over time ( Figure 5A ) . Under these conditions Fus3 activation was reduced by up to 50% , comparable to the reduction observed with KCl ( Figure 3A ) . Hog1 also limited expression of total Fus3 protein . As an additional control we tested mutants lacking Hog1 expression or Hog1 catalytic activity . In this case we observed no change in phospho-Fus3 or Fus3 abundance ( Figure 5B ) . Thus Fus3 can be regulated by Hog1 even in the absence of hyperosmotic stress . Together these results reveal that Hog1 activation is necessary and sufficient to dampen Fus3 activation . As noted above , induction of Fus3 can confound any analysis of Fus3 activation . To distinguish the effects of Hog1 on Fus3 induction and Fus3 phosphorylation , again we replaced the native FUS3 promoter . In this case we used the strong constitutive promoter from ADH1 ( ADH1-FUS3 ) instead of the GAL1 promoter used above , so as to prevent promoter competition and to ensure consistent levels of Hog1 activation . Under these conditions , constitutively active Hog1 reduced phospho-Fus3 by about one-third , somewhat less than the one-half reduction obtained in cells with the native FUS3 promoter ( Figure 5C ) . We obtained similar results using GAL1-FUS3 ( Figure S4 ) instead of ADH1-FUS3 ( Figure 5C ) . Thus Hog1 dampens Fus3 activation , even when expression is permanently elevated . When Fus3 is expressed from the native promoter , activation is dampened even further . These findings confirm that Fus3 induction and Fus3 phosphorylation are diminished by at least two distinct mechanisms that require Hog1 . Each of these mechanisms is considered below . To establish the mechanisms of pathway cross-inhibition we began with the target of Hog1 that limits Fus3 production . The induction of Fus3 requires transcription of the FUS3 gene and translation of the corresponding mRNA . Hog1 phosphorylates and activates the protein kinase , Rck2 [42] . Activated Rck2 phosphorylates the yeast elongation factor , EF2 , and thereby transiently represses translation [43] , [44] . Thus we considered whether Rck2 regulates the production of Fus3 under osmotic stress conditions . To test the hypothesis , we constitutively activated Hog1 in the absence of RCK2 , and in the presence or absence of pheromone ( Figure 5D ) . Under these conditions , constitutively active Hog1 reduced phospho-Fus3 by about one-third , somewhat less than the one-half reduction obtained in cells that express Rck2 ( compare Figure 5B and Figure 5D ) . Thus Rck2 is partially responsible for the diminished Fus3 response . Taken together these results suggest that Rck2 generally represses translation in response to Hog1 , which results in diminished production of Fus3 . More broadly these results provide evidence that the consequent repression of Fus3 translation contributes to decreased pheromone responsiveness following hyperosmotic stress . The data presented above indicate that Hog1 limits Fus3 activity in two ways . First , Hog1 phosphorylates Rck2 and suspends translation of mating pathway components . We have also presented evidence that Hog1 inhibits an upstream activator of the mating MAPK . To identify the second target of Hog1 we employed a genetic epistasis approach . First we determined if constitutively active Hog1 dampens the mating pathway at the level of the three-tiered MAPK cascade . Under normal circumstances the mating signal is initiated by the recruitment of the MAPK scaffold Ste5 to the plasma membrane [45] . Ste5 can be tethered permanently to the plasma membrane via fusion to a carboxy-terminal transmembrane domain ( CTM ) , thus bypassing the need for pheromone , receptor , and G protein in pathway activation ( Figure 6A ) [7] . In cells that co-express GAL1-STE5CTM and GAL1-SSK2ΔN , phospho-Fus3 was dampened , similar to that seen with pheromone and GAL1-SSK2ΔN ( Figure 6B ) . These data indicate that Hog1 acts on a component downstream of the G protein . We likewise observed dampening of phospho-Kss1 , which is also activated by pheromone . These data suggest that the putative Hog1 target is upstream of both Fus3 and Kss1 . Taken together these results narrowed the likely target to a handful of components associated with Ste5: the MAPKKK Ste11 , its adaptor protein Ste50 , its activators Cdc42 and Ste20 , and its substrate Ste7 ( Figure 1A ) . Ste50 is required for full activation of Hog1 , Fus3 , and Kss1 . We and others have demonstrated that Hog1 phosphorylates Ste50 during hyperosmotic conditions . Moreover , the phosphorylation of Ste50 leads to functional downregulation of Hog1 [46] , [47] . Given this precedent , we hypothesized that phosphorylation of Ste50 leads to the downregulation of Fus3 and Kss1 . To test the role of Ste50 , we activated the mating pathway using a truncated form of Ste11; Ste11ΔN lacks the kinase auto-inhibitory domain [48] , and also lacks the Ste50 binding domain ( Figure 6C ) [49] . Thus Ste11ΔN is both constitutively active and refractory to Ste50 . As shown in Figure S5 , Ste50 is not required for pathway activation by Ste11ΔN even while it is required for full activation by Ste5CTM . We had postulated that Fus3 activity is dampened when Ste50 is phosphorylated . Accordingly , Fus3 should not be affected by Hog1 or Ste50 when the pathway is activated through Ste11ΔN ( Figure 6D ) . Under these conditions , Fus3 is fully activated , consistent with our prediction . Presumably Ste7 is also activated under these conditions , although currently we are not able to monitor its activity directly . Taken together these data suggest that Hog1 limits the mating signal at the level of Ste50 . The mating and hyperosmotic stress signals are integrated by Ste50 , which in turn regulates the shared MAPKKK , Ste11 . Next we sought to establish whether phosphorylation of Ste50 by Hog1 was responsible for pathway cross-inhibition . To this end we used a mutant of Ste50 ( Ste505A ) where five MAPK sites have been changed to alanine , thereby abrogating phosphorylation by Hog1 [46] , [47] . Consistent with our prediction Ste505A restored the ability of pheromone to activate Fus3 , even under conditions of constitutive Hog1 activation ( Figure 6E ) . Fus3 was not fully activated however , presumably because Hog1 could still target Rck2 . When we deleted RCK2 from the ste505A strain we were able to attain full activation of Fus3 ( Figure 6F and Figure S6 ) . Thus Hog1 limits mating through the phosphorylation of at least two proteins , Ste50 and Rck2 . More generally , these results reveal that cross-inhibition occurs through a combination of feedback and feedforward phosphorylation events . Finally we aimed to determine the biological significance of Fus3 cross-inhibition by Hog1 . As shown in Table 1 , mating efficiency of wild-type cells is reduced by hyperosmotic conditions , presumably by Hog1-dependent and Hog1-independent mechanisms . To determine the contribution of Hog1 we performed quantitative mating assays in the presence or absence of the two cross-inhibition targets , either alone or in combination . Whereas mating efficiency is diminished in the presence of salt , mating was partially restored in the ste505A rck2Δ mutant strain ( Figure 6G and Table S6 ) . The partial rescue suggests that other mechanisms may be operative , or perhaps mating fails because the cells are still responding to stress . Fus3 activation was likewise restored in these mutant cells ( Figure S7 ) . Together , our results show that Hog1 inhibits Fus3 induction and activation , and these processes serve to delay mating until the cells have fully adapted to osmotic stress conditions .
Prior to our investigations , it was established that Hog1 activation is proportional to the severity of the hyperosmotic stress [50] . Furthermore , the duration of Hog1 activation is tightly correlated with glycerol production and a return to osmotic equilibrium [14] . Thus , hyperosmotic stress and cell adaptation dictate the level and duration of Hog1 activity . Our results support a model where Hog1 suspends the mating response until cells are fully adapted . In particular , we found that Hog1 dampens and delays Fus3 activation , and that the duration of delay is proportional to the severity of the hyperosmotic stress . Just as transient activation of Hog1 leads to transient inhibition of Fus3 , persistent activation of Hog1 leads to persistent inhibition of Fus3 . It was established previously that osmotic stress results in a general inhibition of gene transcription . In cells that lack Hog1 , transcription initiation is delayed further [34] . These results show the broad negative effects of salt on gene transcription and point to Hog1 as the primary mediator of the stress response . Paradoxically , cells that lack Hog1 exhibit a stress-mediated increase in the transcription of mating genes . These findings point to a special function for Hog1 in limiting the mating pathway . A major challenge has been to understand how Hog1 regulates Fus3 , in addition to any Hog1-independent processes that might affect Fus3 induction . This was achieved by ( i ) constitutive expression of Fus3 ( via promoter replacement ) and ( ii ) direct activation of Hog1 ( via Ssk2ΔN ) . Ultimately these approaches allowed us to identify Ste50 and Rck2 as important targets of Hog1 . Phosphorylation of these proteins accounts for delayed mating responses during co-stimulation . However , other targets of Hog1 are likely . In the absence of Hog1 , high osmolarity activates the transcriptional outputs of both the filamentous growth pathway and mating pathway . As yet the relevant substrates of Hog1 in cross-talk suppression have not been identified . Ste50 is a shared component , required for activation of Ste11 , that acts early in the mating and osmotic stress pathways . Thus Ste50 is well positioned to coordinate the activity of both Fus3 and Hog1 . Moreover , Ste50 is phosphorylated by Hog1 and as a consequence of this phosphorylation there is an attenuated response to hyperosmotic stress [46] , [47] . Here we show that as an additional consequence of Ste50 phosphorylation there is an attenuated response to pheromone . On that account , Ste50 is a target of both negative feedback during stress adaptation and cross-inhibition during co-stimulation . The effects of co-stimulation are most evident at the level of the mating MAPKs . Salt-dependent phosphorylation of Ste50 attenuates pheromone-dependent activation of both Fus3 and Kss1 . Whereas the phosphorylation of Ste50 fully accounts for cross-inhibition of Kss1 , it is only partially responsible for cross-inhibition of Fus3 . Consequently we searched for additional mechanisms of signal integration that act on Fus3 but not Kss1 . Given that Fus3 is induced by pheromone - whereas Kss1 is not - we considered whether salt stress inhibits Fus3 transcription or translation . It was established previously that Hog1 directly phosphorylates and activates a repressor of translation elongation , Rck2 [42] . When Rck2 is absent , translation repression is abrogated [43] . Accordingly , we found that Rck2 is needed to inhibit Fus3 accumulation . As with Ste505A , the effect of the rck2Δ mutation was incomplete . However combining both mutations ( rck2Δ ste505A ) eliminated the ability of Hog1 to inhibit Fus3 ( Figure S6 ) . Thus Hog1 phosphorylates components necessary for the activation and induction of Fus3 . Together these phosphorylation events act to limit mating responses as long as Hog1 is active . Once the cells are fully adapted , mating can proceed . While it is clear that Rck2 confers a global inhibition of protein translation [43] , it is important to note that Rck2 also contributes to the induction of distinct gene transcripts necessary for stress adaptation [51] . Thus Rck2 may represent a more general mechanism that ensures competing cellular processes do not interfere with the early translation of stress adaptive genes . It is also possible that Rck2 regulates other components of the pheromone pathway , in addition to Fus3 . However , many of the core components that make up the MAPK cascade are stably expressed , including the scaffold ( Ste5 ) , the MAPKK ( Ste7 ) , the MAPKKK ( Ste11 ) and its adaptor ( Ste50 ) [9] . Therefore Hog1 and Rck2 are not expected to interfere with the ability to sense pheromone; rather Hog1 is likely to arrest signal transduction by those proteins that are induced by pheromone , most of which function downstream in the pathway ( including Fus3 ) . Thus we postulate that early components of the pheromone pathway are unaffected by hyperosmotic stress conditions . We propose that the earliest events in pheromone signaling , those not subject to pheromone mediated transcriptional induction , are unaffected by hyperosmotic stress conditions . These early events include G protein activation and recruitment of Ste5 [45] . Consistent with this view , our epistasis studies indicate that Hog1 acts downstream of the G protein . Moreover recruitment of Ste5 to the plasma membrane occurs even in the face of hyperosmotic stress [24] . This behavior suggests that the mating pathway remains quiescent only as long as conditions are unfavorable to launch a full mating response . Once cells adapt to stress , Hog1 is deactivated and mating can proceed immediately . Taken together , available data support a model where mating and HOG pathways are both initiated in response to pheromones and hyperosmotic stress . However , the activation of Hog1 imparts a “checkpoint” midway in the pheromone signaling pathway , and does so to ensure quiescence of the mating response while cells adapt to stress ( Figure 7 ) . This design ensures that the mating pathway is primed to resume full signaling once Hog1 is no longer activated . Accordingly , Ste50 and Rck2 are both rapidly dephosphorylated upon adaptation [43] , [46] . From these behaviors we can infer that scaffold proteins and shared adaptor proteins have distinct but complementary roles in signaling; scaffold proteins , epitomized by Ste5 , behave as insulators , while shared components , such as Ste50 , behave as dynamic integrators of multiple signals . As genomics and proteomics have defined signal pathway components , attention will turn increasingly to understanding how cells coordinate competing signals . In this regard , our findings reveal that pathway cross-inhibition is not a single process , but rather a network of events that work together to postpone cell differentiation until the cell adapts to stress conditions . More broadly , it is increasingly evident that a complete analysis of signal transduction networks will need to consider multiple inputs , multiple regulatory targets , and multiple mechanisms of action .
Standard procedures for growth , maintenance , and transformation of yeast and bacteria and for the manipulation of DNA were used throughout . Plasmids and strains were constructed as previously described [46] , [52]–[55] . Yeast strains and plasmids used are listed in supplemental Table S7 and supplemental Table S8 , respectively . All mutations were constructed with the QuikChange site-directed mutagenesis kit ( Stratagene ) according to the manufacturer's directions . Cells were grown in synthetic complete medium containing 2% ( w/v ) dextrose ( SCD ) or raffinose followed by the addition of 2% galactose to induce gene expression . Plasmid-transformed cells were grown in synthetic complete medium lacking the appropriate nutrient . Yeast mating efficiency was determined by a quantitative method , as described previously [56] . Cells were grown to OD600∼0 . 6 and counted using a hemocytometer . 5×106 BY4741 ( MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 ) cells were mixed with 5×106 BY4742 ( MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 ) cells in a volume of 10 ml and passed over a nitrocellulose filter ( Millipore ) . Filter disks were incubated for 4 h or 24 h on YPD agar or YPD agar containing 0 . 5 M KCl . Mating efficiency was calculated by dividing the number of diploid cells by the number of total cells after 4 h or 24 h mating period . Cells were grown to A600 nm∼0 . 8 , dispersed by sonication with 10 pulses ( 1 sec , 50% output ) , and collected by centrifugation at 14 , 000× g for 15 seconds . 3 µl of cells were placed on glass slides coated with SCD medium 2% agar ( w/v ) and either α factor pheromone , or α factor and KCl . Cells were visualized every 15 min by differential interference contrast ( DIC ) and fluorescence microscopy using an Olympus Fluoview FV1000 confocal microscope with a 60× objective . GFP fluorescence was imaged using a 488-nm argon laser and 500–550 nm emission filter . Videos were constructed and images were analyzed using ImageJ ( National Institutes of Health ) . Logarithmically growing cells ( A600 nm∼0 . 6 ) were stimulated with 10 µM α factor and 0 . 75 M KCl or 1 M sorbitol . Viability was assessed by methylene blue staining ( 0 . 01% solution w/v ) before and after 2 h of treatment . Prior to counting cells were dispersed by sonication with 5 pulses ( 1 sec , 50% output ) and diluted 1∶10 in SCD with methylene blue and counted using a hemocytometer . Protein extracts were produced by glass bead lysis in TCA as previously described [55] . Protein concentration was determined by Dc protein assay ( Bio-Rad Laboratories ) . Protein extracts were resolved by 7 . 5% or 12 . 5% SDS-PAGE and immunoblotting with HA antibodies ( clone 3F10 , Roche Applied Science ) at 1∶2000 , Phospho-p44/42 MAPK antibodies ( 9101 , Cell Signaling Technology ) at 1∶500 , Fus3 antibodies ( sc-6773 , Santa Cruz Biotechnology , Inc . ) at 1∶500 , phospho-p38 MAPK antibodies ( 9216 , Cell Signaling Technology ) at 1∶500 , Hog1 antibodies ( sc-6815 , Santa Cruz Biotechnology ) at 1∶500 , and glucose-6-phosphate dehydrogenase ( G6PDH ) antibodies ( A9521 , Sigma-Aldrich ) at 1∶50 , 000 . Far1-HA immunoreactive species were visualized by chemiluminescent detection ( PerkinElmer Life Sciences LAS ) of horseradish peroxidase-conjugated antibodies ( sc-2006 , Santa Cruz Biotechnology , Inc . ) at 1∶10 , 000 . All remaining immunoreactive species were visualized by fluorescent detection ( Typhoon Trio+Imager , GE Healthcare ) of AlexaFluor conjugated antibodies ( A21245 , A21424 , A21431 , Invitrogen ) at 1∶2 , 000 . Band intensity was quantified by scanning densitometry using Image J ( National Institutes of Health ) . P-Fus3 and P-Kss1 values were normalized to G6PDH loading control . FUS1-LacZ levels were measured every 30 min after treatment with mating pheromone α factor , or α factor and KCl or sorbitol using a β-galactosidase assay as described previously [54] . Cells were split and diluted 30% with fresh medium containing pheromone alone or pheromone and an indicated concentration of KCl or sorbitol . Aliquots of cells were removed every 30 min , lysed , and β-galactosidase activity was measured . | All cells can detect and respond to signals in their environment . The ability to interpret these signals with accuracy is needed for proper growth and differentiation . Moreover , cells must prioritize responses when confronted with competing signals . However the molecular mechanisms that govern signal prioritization are poorly understood . To address this question , we studied two signaling pathways in the genetic model organism budding yeast . Specifically we focused on the pheromone mating ( differentiation ) pathway and the high osmolarity glycerol ( stress response ) pathway . These pathways respond differently to each stimulus despite sharing pathway components . We find that cells must first adapt to stress before they can mate . At early times , the stress response cross-inhibits and dampens the pheromone response to suspend mating differentiation . Once cells adapt , the stress response ends and the differentiation program resumes . All signaling pathways that regulate cell fate decisions are interconnected to varying degrees . Our study highlights the importance of proper signal coordination in cell fate decisions , and it reveals new mechanisms that govern signal coordination within complex signaling networks . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cellular",
"stress",
"responses",
"mechanisms",
"of",
"signal",
"transduction",
"microbiology",
"feeback",
"regulation",
"cell",
"differentiation",
"developmental",
"biology",
"model",
"organisms",
"crosstalk",
"mapk",
"signaling",
"cascades",
"stress",
"signaling",
"cascade",
"signal",
"initiation",
"biology",
"molecular",
"biology",
"signal",
"transduction",
"signal",
"termination",
"cell",
"biology",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"molecular",
"cell",
"biology",
"cell",
"fate",
"determination",
"signaling",
"cascades"
] | 2012 | Checkpoints in a Yeast Differentiation Pathway Coordinate Signaling during Hyperosmotic Stress |
Peroxiredoxins are a family of antioxidant enzymes critically involved in cellular defense and signaling . Particularly , yeast peroxiredoxin Tsa1p is thought to play a role in the maintenance of genome integrity , but the underlying mechanism is not understood . In this study , we took a genetic approach to investigate the cause of genome instability in tsa1Δ cells . Strong genetic interactions of TSA1 with DNA damage checkpoint components DUN1 , SML1 , and CRT1 were found when mutant cells were analyzed for either sensitivity to DNA damage or rate of spontaneous base substitutions . An elevation in intracellular dNTP production was observed in tsa1Δ cells . This was associated with constitutive activation of the DNA damage checkpoint as indicated by phosphorylation of Rad9/Rad53p , reduced steady-state amount of Sml1p , and induction of RNR and HUG1 genes . In addition , defects in the DNA damage checkpoint did not modulate intracellular level of reactive oxygen species , but suppressed the mutator phenotype of tsa1Δ cells . On the contrary , overexpression of RNR1 exacerbated this phenotype by increasing dNTP levels . Taken together , our findings uncover a new role of TSA1 in preventing the overproduction of dNTPs , which is a root cause of genome instability .
Peroxiredoxins belong to a family of thiol-specific peroxidases widely and abundantly expressed in most living organisms [1] , [2] . Through one or more redox-sensitive cysteines , peroxiredoxins not only scavenge reactive oxygen species ( ROS ) including peroxides and peroxynitrite [3] , [4] , but also function as an ROS sensor to regulate cell signaling [5]–[11] . For many peroxiredoxins , another level of regulation can be achieved through oligomerization [1] , [2] , [12] . In addition to their roles in peroxide reduction , peroxiredoxins are also known to possess chaperone activity [12] , [13] . Loss-of-function studies in mice implicated an essential role of peroxiredoxins in antioxidant defense and tumor suppression [14] . Particularly , peroxiredoxin 1-knockout mice not only suffered from severe anemia due to oxidative stress , but were also susceptible to several types of malignant tumors [15] . Consistent with this , genome-wide screening revealed that yeast peroxiredoxin TSA1 was a strong suppressor of gross chromosomal rearrangements and spontaneous mutations [16] , [17] . In addition , a mutator phenotype was observed in yeast cells lacking one or more peroxiredoxins . The phenotype could be rescued by yeast peroxiredoxin Tsa1p or mammalian Prx1 , but not by their active-site mutants defective for peroxidase activity [18] , [19] . In further support of a role of TSA1 in the maintenance of genome stability , many genetic interaction partners of TSA1 identified through synthetic genetic array analysis were components of DNA repair machinery or DNA checkpoints [20] , [21] . For example , TSA1 was found to interact genetically with REV1/REV3 and OGG1 , which are critically involved in translesion synthesis ( TLS ) and the repair of oxidative DNA damage , respectively [22] , [23] . However , the exact mechanism by which Tsa1p suppresses genome instability remains to be fully understood . Intracellular dNTP levels are one important determinant of cellular response to DNA damage [24] . For yeast cells to survive DNA damage , increased dNTP production would be allowed to facilitate replication , but with a trade-off of high spontaneous mutation rate [25] . In other words , abnormally high dNTP levels are causally associated with genome instability [24] , [26] . We previously demonstrated that yeast Tsa1p is a house-keeping peroxiredoxin which sufficiently suppressed the mutator phenotype [18] . Although both an aberrantly high level of ROS and an imbalance in free radical contents , which is caused by compensational activation of other antioxidants such as Sod1p [27] , could underlie the mutator phenotype of tsa1Δ cells , additional events subsequent to the disruption of TSA1 might also be influential in the induction of genome instability . In this study we asked whether perturbation of dNTP pools might contribute to the mutator phenotype observed in tsa1Δ cells . We then investigated the cause of dNTP pool expansion . Our findings suggested that constitutive activation of the DNA damage checkpoint and consequent overproduction of dNTPs are the root cause of genome instability in tsa1Δ cells .
Yeast peroxiredoxin TSA1 was found to be a strong suppressor of mutations and gross chromosomal rearrangements [16]–[18] . In addition , further deletion of another gene involved in DNA repair or DNA checkpoints caused synthetic growth defect or lethality in tsa1Δ cells [21] , [22] . Bearing these findings in mind , here we sought to dissect the interaction of TSA1 with the DNA damage checkpoint and particularly the machinery of dNTP synthesis , in order to understand the role of TSA1 in the maintenance of genome stability . We first examined the sensitivity of tsa1Δ cells to various DNA damaging agents . tsa1Δ cells were sensitive to hydroxyurea ( HU ) , 4-nitroquinoline 1-oxide ( 4NQO ) and ultraviolet ( UV ) irradiation ( Figure 1A , lanes 1 and 5; Figure 1B; Figure 1C and Figure 1D , lanes 1 and 2 ) . Re-expression of TSA1 in tsa1Δ cells suppressed the sensitivity phenotype ( Figure 1C and Figure 1D , lane 3 ) . This suppression required the catalytic cysteine ( Cys47 ) of Tsa1p , but not the C-terminal cysteine ( Cys170 ) , pointing to the importance of the antioxidant property of Tsa1p in the protection against DNA damage ( Figure 1C , lanes 3–5 ) . The sensitivity of tsa1Δ cells to DNA damage prompted us to investigate further the genetic interactions between TSA1 and components of the DNA damage checkpoint . In light of the finding that TSA1 genetically interacts with DNA damage checkpoint genes DUN1 and SML1 [22] , we chose these two genes and their effector CRT1 for further analysis . Dun1p is a checkpoint kinase that phosphorylates and regulates ribonucleotide reductase ( RNR ) inhibitor Sml1p [28] . Dun1p also inhibits Crt1p , a transcriptional corepressor of RNR , through phosphorylation [29] , [30] . Deletion of DUN1 , SML1 or CRT1 in tsa1Δ cells exerted a significant impact on their sensitivity to HU , 4NQO and UV irradiation . Loss of DUN1 further sensitized tsa1Δ cells to H2O2 , HU , 4NQO and UV ( Figure 1A , lanes 1 , 2 , 5 and 6 ) . In support of the specificity of effect , this sensitization was reversed upon expression of TSA1 or DUN1 in tsa1Δ dun1Δ cells ( Figure 1D , lanes 6–8 ) . Conversely , loss of SML1 or CRT1 rescued the sensitivity phenotype of tsa1Δ cells to 4NQO and UV ( Figure 1A , lanes 5 , 7 and 8 ) . It is noteworthy that such reversion of sensitivity was not observed in cells treated with H2O2 ( Figure 1A , lanes 5 , 7 and 8 ) , suggesting that the effect might be specific to DNA damaging agents and was not caused directly by ROS . These observations supported the notion that TSA1 interacts specifically with the DNA damage checkpoint in a manner that is not mediated directly through ROS . Although the sensitivity pattern of the different mutant strains in the spot assay was highly reproducible , a more quantitative comparison of these strains is desired . Hence , survival curves of strains in the presence of 4NQO and UV were also obtained ( Figure 1B ) . Dose-dependent killing of the strains by 4NQO and UV was observed . At all doses tested , the degrees of sensitivity of different strains to either 4NQO or UV were in the same order as shown in the spot assay . In particular , the survival curves indicated a further enhancement of the sensitivity phenotype in tsa1Δ dun1Δ versus tsa1Δ cells and a suppression of sensitivity in tsa1Δ sml1Δ cells ( Figure 1B ) . Collectively , our results demonstrated that the survival of tsa1Δ cells under DNA damage was decreased upon deletion of DUN1 , but enhanced when either SML1 or CRT1 was genetically disrupted . We also compared the phenotypes of tsa1Δ cells and cells lacking Sod1p , another key antioxidant enzyme [31] . In contrast to the genetic interactions observed in tsa1Δ cells , deletion of DUN1 , SML1 or CRT1 in sod1Δ cells enhanced its sensitivity to HU and 4NQO ( Figure 1E , lanes 6 , 7 and 8 ) . Thus , TSA1 and SOD1 interact with DUN1 , SML1 and CRT1 through different mechanisms . We next investigated whether compromising DNA damage checkpoint genes in tsa1Δ cells might also alter their mutator phenotype . In agreement with previous reports [16]–[18] , tsa1Δ cells exhibited high rates of spontaneous mutations in both canavanine-resistant ( CANR ) and 5FC-resistant ( 5FCR ) assays ( Figure 2A , Figure 2B and Figure 2D , groups 1 and 5 ) . On the other hand , deletion of DUN1 did not significantly affect spontaneous mutation rates ( Figure 2A , Figure 2B and Figure 2D , groups 1 and 2 ) , whereas loss of SML1 or CRT1 caused a mild increase in CANR mutation rates in WT cells ( Figure 2A and Figure 2D , groups 1 , 3 and 4 ) . However , the disruption of DUN1 , SML1 and CRT1 in tsa1Δ cells modulated the mutator phenotype in opposite directions ( Figure 2A and Figure 2B , columns 5–8 ) . Whereas reduction of spontaneous mutation rates was observed in tsa1Δ dun1Δ cells ( Figure 2A and Figure 2B , columns 5 and 6 ) , deletion of SML1 or CRT1 in tsa1Δ cells significantly enhanced the mutator phenotype ( Figure 2A and Figure 2B , columns 1 , 7 and 8 ) . Complementation of the reduction of mutation rate in tsa1Δ dun1Δ cells by re-introduction of DUN1 or TSA1 further verified the specificity of effect ( Figure 2C , columns 6 and 8 ) . Thus , the mutation rates of tsa1Δ cells correlated with the activity of the DNA damage checkpoint . In all individual deletion mutants , tsa1Δ cells displayed the highest mutation rate ( Figure 2 ) . We postulated that this might be attributed either directly or indirectly to the elevation of intracellular ROS levels in these cells [4] . If that is the case , challenging the other DNA damage checkpoint mutants with ROS might have an impact on the mutator phenotype . To test this idea , we treated the cells with low-dose H2O2 and assessed the impact on CANR mutation rates . Interestingly , mutation rates increased in WT and dun1Δ cells to comparable levels ( Figure 2D , groups 1 and 2 ) . In contrast , a further increase in mutability was observed when SML1 or CRT1 was comprised ( Figure 2D , groups 1 , 3 and 4 ) . Although the mutation rates of sml1Δ and crt1Δ cells in the presence of H2O2 were still not as high as that of tsa1Δ cells in the absence of H2O2 ( Figure 2D , groups 3–5 ) , our results did suggest that ROS could differentially modulate the mutator phenotype of different mutants . We next investigated the mechanism that underlies the correlation of DNA damage checkpoint activity in tsa1Δ cells with drug sensitivity and mutator phenotype ( Figure 1 and Figure 2 ) . DUN1 , SML1 and CRT1 are regulators of RNR , the rate-limiting enzyme in dNTP synthesis [32]–[34] . Considered together with the model that elevated dNTP levels are required for surviving DNA damage in yeast at the price of increasing mutation rates [24] , we asked whether the mutator phenotype of tsa1Δ cells would be due to alteration in cellular dNTP production . Thus , we measured dNTP levels in our mutants . Surprisingly , tsa1Δ cells produced significantly more dNTPs than wild type ( WT ) cells ( Figure 3A , groups 1 and 5 ) . The magnitude of dNTP overproduction in tsa1Δ cells was comparable to that in sml1Δ cells ( Figure 3A , groups 3 and 5 ) , in which the removal of Sml1p activates RNR leading to the rise in dNTP levels [28] . To shed further light on the roles of dNTP production in the generation of mutator phenotype , we compared the dNTP levels in other mutant cells . As expected , dun1Δ cells produced less dNTPs than WT cells ( Figure 3A , groups 1 and 2 ) , since Dun1p is required for phosphorylation and subsequent removal of the RNR inhibitor Sml1p [28] . Loss of CRT1 was also found to increase cellular dNTP production ( Figure 3A , groups 1 and 4 ) , as Crt1p is a transcriptional corepressor of RNRs [29] . However , loss of DUN1 reduced cellular dNTP production in tsa1Δ cells ( Figure 3A , groups 5 and 6 ) , whereas deletion of SML1 or CRT1 in tsa1Δ cells further increased dNTP levels ( Figure 3A , groups 5 , 7 and 8 ) . Noteworthily , increased production of dNTPs in tsa1Δ cells could be fully complemented by Tsa1p , but not by its catalytic cysteine mutant C47S ( Figure 3B ) . Thus , the antioxidant property of Tsa1p was likely required for preventing overproduction of dNTPs . We then asked whether the reduction of dNTP pools in tsa1Δ dun1Δ cells would be associated with a further drop in intracellular ROS levels in the absence of DUN1 . Interestingly , tsa1Δ dun1Δ cells exhibited a higher level of intracellular ROS over WT , tsa1Δ or dun1Δ cells ( Figure 3C , columns 1 , 2 , 5 and 6 ) , suggesting that the mutator phenotype in tsa1Δ and tsa1Δ dun1Δ cells correlates directly with dNTP production , but not generation of ROS . On the other hand , loss of SML1 or CRT1 did not alter the ROS levels in either WT or tsa1Δ cells ( Figure 3C , columns 1 , 3 , 4 , 5 , 7 and 8 ) . Thus , in addition to the accumulation of ROS , elevation of dNTP production might also contribute to genome instability in tsa1Δ cells . While deletion of DUN1 , SML1 or CRT1 has an impact on dNTP production , they are multifunctional proteins that might also affect other biological processes [28]–[30] . To address this concern , we modulated the production of dNTP more directly by overexpressing RNR1 gene in tsa1Δ and sod1Δ cells . This overexpression has previously been shown to elevate intracellular dNTP levels substantially [24] , [26] . Indeed , when we induced the expression of Rnr1-3MYCp in WT cells ( Figure 4A ) , the spontaneous mutation rate was increased ( Figure 4B , columns 1 and 2 ) . Furthermore , overexpression of RNR1 also exacerbated the mutator phenotype in tsa1Δ and sod1Δ cells ( Figure 4B , columns 3–6 ) . This lent additional support to the importance of dNTP overproduction in the induction of genome instability . If the mutability of tsa1Δ cells is indeed caused by dNTP overproduction , the mutations generated would primarily be base substitutions rather than large deletion and gross chromosomal rearrangements [24] . With this in mind , we examined the types of mutations arisen in tsa1Δ and other mutants ( Table 1 ) . We noted that the majority of mutations found in tsa1Δ cells were base substitutions ( 83 . 3% ) and frameshifts ( 13 . 3% ) . Large deletions were very rare in tsa1Δ cells ( 3 . 3% ) when compared with WT ( 13 . 3% ) cells . In addition , all of the mutations found in tsa1Δ sml1Δ cells with high dNTP levels ( Figure 3 ) were base substitutions ( 90% ) and frameshifts ( 10% ) , whereas more deletions ( 13 . 3% ) were detected in tsa1Δ dun1Δ cells ( Table 1 ) with low dNTP concentrations ( Figure 3 ) . In keeping with previous findings [16] , relatively more deletions ( 10% ) were also observed in sod1Δ cells ( Table 1 ) . Generally , base substitutions were more prevalent in the strain when dNTP levels were high ( Figure 3 ) , whereas the incidences of deletions correlated negatively with dNTP concentrations . Therefore , the mutation spectra of tsa1Δ and other strains are consistent with the notion that elevation of dNTP levels is the underlying cause of genome instability in the absence of TSA1 . Above we demonstrated the elevation of dNTP levels in tsa1Δ cells ( Figure 3 ) . In addition , our results also indicated the genetic interaction of TSA1 with DNA checkpoint genes ( Figure 1 ) . This led us to further investigate whether elevated production of dNTPs in the absence of TSA1 might be explained by the activation of the DNA damage checkpoint . As a first step , we assessed checkpoint activation by examining the steady-state levels of Rad53p , the yeast ortholog of human CHK2 kinase whose phosphorylation and activation are pivotally involved in the control of checkpoint response to DNA damage [35] , [36] . Particularly , Rad53p is a master regulator of Dun1p , Sml1p and Crt1p [35] . In this analysis we included the sod1Δ control strain , in which the effectors of the Mec1p-dependent DNA damage checkpoint were previously shown to be downregulated [31] . Phosphorylated Rad53p species were more evident in tsa1Δ cells compared to WT and sod1Δ cells ( Figure 5A , lanes 1 , 3 and 5; Figure 5B , lanes 1 and 2; Figure 6A , lanes 1 and 2; and Figure 7A , lanes 1 and 3 ) . This difference became more pronounced in the presence of H2O2 ( Figure 5A , lanes 2 , 4 and 6; Figure 5B , lanes 5 and 6; Figure 6A , lanes 3 and 4; and Figure 7A , lanes 5 and 7 ) . The steady-state levels of phosphorylated Rad53p in DNA damage checkpoint mutants were also compared . Whereas deletion of DUN1 triggered phosphorylation of Rad53p ( Figure 5C , lanes 1 and 2 ) , an observable increase in phosphorylated Rad53p species was not found in sml1Δ or crt1Δ cells ( Figure 5C , lanes 1 , 3 and 4 ) . Notably , loss of DUN1 in tsa1Δ cells further enhanced the activation of Rad53p ( Figure 5C , lanes 5 and 6 ) , whereas tsa1Δ sml1Δ and tsa1Δ crt1Δ cells had similar levels of phosphorylated Rad53p compared to tsa1Δ cells ( Figure 5C , lanes 5 , 7 and 8 ) . In addition to Rad53p , we also checked for the status of Rad9p , a more upstream transducer in the DNA damage checkpoint pathway [37] . Rad9-13MYCp was found to be activated in tsa1Δ cells ( Figure 6B , lanes 1 , 2 , 4 and 5 ) and this activation could not be reversed by the C47S mutant of Tsa1p ( Figure 6B , lanes 2 , 3 , 5 and 6 ) . As a marker for DNA double-strand breaks ( DSBs ) [38] , the level of γH2A was also found to be elevated in tsa1Δ cells as compared to WT ( Figure 5B , lanes 1 and 2; Figure 6A , lanes 1 and 2; and Figure 6B , lanes 1 and 2 ) . This agrees with a recent report that tsa1Δ cells displayed an increased number of Rad52-YFP foci indicative of DNA damage [23] . The levels of γH2A in other DNA damage checkpoint mutant cells were also examined . Among dun1Δ , sml1Δ and crt1Δ cells , an elevation in γH2A level was only found in dun1Δ cells ( Figure 5C , lane 2 ) . In addition , disruption of DUN1 , SML1 or CRT1 in tsa1Δ cells did not affect γH2A levels significantly ( Figure 5C , lanes 5–8 ) . Noteworthily , although phosphorylated Rad53p and γH2A were abundant in dun1Δ and tsa1Δ dun1Δ cells ( Figure 5C , lanes 2 and 6 ) , their mutation rates remained low ( Figure 2A and Figure 2B , columns 2 and 6 ) plausibly due to the low levels of dNTPs ( Figure 3A , groups 2 and 6 ) . In other words , elevation of dNTP levels might be the direct cause of genome instability . Consistent with the activation of Rad53p and Rad9p , the levels of Rad53p target Sml1-3HAp were diminished in tsa1Δ cells in the presence ( Figure 5B , lanes 5 and 6 ) and absence of H2O2 ( Figure 5B , lanes 1 and 2 ) . Importantly , all of the above changes in tsa1Δ cells could be fully complemented by TSA1 ( Figure 5B , lanes 2 and 3 ) , but not by its C47S mutant ( Figure 5B , lanes 2 and 4 ) . On the other hand , in agreement with previous findings [31] , we did not observe a significant change of Sml1p level in sod1Δ cells ( data not shown ) . RNR is an important downstream effector of the DNA damage checkpoint which mediates the production of dNTPs [33] , [34] . Since the expression of RNR genes is transcriptionally activated in response to DNA damage [29] , we used semi-quantitative RT-PCR to determine the relative levels of RNR1/2/3/4 transcripts in the presence and absence of HU . In this analysis we included an additional control termed HUG1 , a target of Mec1p induced highly by DNA damage [39] . As shown in Figure 6C , RNR transcripts were induced to higher levels in tsa1Δ cells than in WT and sod1Δ cells . The induction of RNR1 and RNR3 was greatest in both untreated and HU-treated tsa1Δ cells . The level of HUG1 transcript was also elevated in tsa1Δ cells and this was more pronounced in the presence of HU ( Figure 6C ) . In sharp contrast , sod1Δ cells treated with HU showed a lower magnitude of induction of RNR1 , RNR3 and HUG1 mRNAs ( Figure 6C ) . These results obtained from sod1Δ cells were generally consistent with previous findings [31] . Thus , the pattern of RNR induction in tsa1Δ cells was not ascribed to a general effect caused by the lack of any antioxidant enzyme , but was highly specific . Collectively , our results suggested that loss of TSA1 induces the activation of the DNA damage checkpoint leading to the induction of RNR and consequent overproduction of dNTPs . If activation of the DNA damage checkpoint in tsa1Δ cells is really important to the generation of genome instability caused by dNTP overproduction , genetic disruption of the checkpoint would be able to reverse the mutator phenotype of tsa1Δ cells . To test this hypothesis , we employed a RAD53 mutant termed rad53AA , in which both T354 and T358 in the activation loop of Rad53p had been replaced by alanine , thereby abrogating the autophosphorylation activity in response to DNA damage [40] . This defective rad53AA allele similar to rad53-11 is thought to be associated with reduced dNTP production due to high abundance of Sml1p [40] , [41] . Thus , we set out to characterize the phenotypes of tsa1Δ cells carrying the rad53AA allele . As documented previously [40] , TSA1 rad53AA cells exhibited lower levels of phosphorylated Rad53p and higher abundance of Sml1-3HAp than TSA1 RAD53 cells ( Figure 7A , lanes 1 , 2 , 5 and 6 ) . In response to H2O2 , γH2A was induced to higher levels in all mutant cells ( Figure 7A , lanes 5–8 compared to lanes 1–4 ) . Notably , both tsa1Δ RAD53 and tsa1Δ rad53AA cells showed similar basal levels of γH2A ( Figure 7A , lanes 3 and 4 ) . Although stronger Rad53p activation was observed in tsa1Δ rad53AA cells , a more pronounced Sml1-3HAp protein band was seen ( Figure 7A , lane 4 compared to lane 3 ) , suggestive of a defective DNA damage checkpoint . We next characterized the sensitivity of these mutants towards H2O2 , UV and HU . TSA1 rad53AA cells were sensitive to HU ( Figure 7B , lanes 1 and 2 ) as previously described [40]; while tsa1Δ RAD53 cells were sensitive to H2O2 , UV and HU ( Figure 7B , lanes 1 and 3 ) similar to tsa1Δ cells in BY4741 background ( Figure 1 , lanes 1 and 5 ) . Resembling tsa1Δ dun1Δ cells in BY4741 , tsa1Δ rad53AA cells in W303 background displayed further sensitivity to H2O2 and UV when compared to tsa1Δ RAD53 cells ( Figure 7B , lanes 3 and 4 ) . We then looked at the effect of a defective DNA damage checkpoint on genome instability in tsa1Δ cells . Intriguingly , tsa1Δ rad53AA cells displayed a significantly reduced ( ∼50% ) rate of spontaneous 5FCR mutations over tsa1Δ RAD53 cells ( Figure 7C ) . On the other hand , both tsa1Δ rad53AA and tsa1Δ RAD53 cells had high levels of intracellular ROS over WT cells as measured by DCF fluorescence ( Figure 7D ) . These observations suggested that rad53AA mutation can suppress the mutator phenotype in tsa1Δ cells without affecting cellular redox environment . This generally agrees with the phenotypes of tsa1Δ dun1Δ cells ( Figure 1 and Figure 2 ) , lending further support to the concept that intracellular dNTP levels are an important determinant in the induction of genome instability in tsa1Δ cells .
Here , we provided the first evidence that loss of yeast peroxiredoxin TSA1 causes genome instability through constitutive activation of the DNA damage checkpoint leading to overproduction of intracellular dNTPs . There are two salient points in our work . First , we demonstrated the elevation of dNTP levels in tsa1Δ cells and its direct correlation with the mutator phenotype ( Figure 2 , Figure 3 , Figure 4 ) . Second , we demonstrated the activation of the DNA damage checkpoint in tsa1Δ cells in relation to elevated production of dNTPs ( Figure 1 , Figure 5 , Figure 6 , Figure 7 ) . Our findings suggested a new model for the role of peroxiredoxins in the maintenance of genome integrity , which has implications in the understanding of human diseases including cancer . In agreement with our findings on the accumulation of γH2A and activation of the DNA damage checkpoint in tsa1Δ cells , several lines of evidence in the literature supported the role of Tsa1p and other peroxiredoxins in the protection of cells against DNA damage . First , tsa1Δ cells produce significantly more ROS [4] , which cause DNA and protein damage [27] , [42] , [43] . Second , loss of TSA1 results in increased formation of Rad52-YFP foci , an indicator of DNA DSBs [23] . Third , tsa1Δ cells are highly sensitive to the functional state of DNA repair and checkpoints [22] . In particular , tsa1Δ is synthetically lethal with rad51Δ mutation , indicating that the viability of rad51Δ cells deficient in recombination repair requires TSA1 function [44] . Finally , human peroxiredoxins have been implicated in cellular defense against oxidative DNA lesions [45] . In this context , the activation of the DNA damage checkpoint in tsa1Δ cells demonstrated in our study highlights the pivotal roles of the checkpoint in cell survival and provides an explanation for the synthetic lethality seen in various double deletion mutants involving TSA1 and another DNA repair or checkpoint gene [21] . Deletion of TSA1 in yeast cells has previously been shown to result in both a mutator phenotype and an increase in gross chromosomal rearrangements [16] , [22] , [23] . Although the causes and origin of gross chromosomal rearrangements remain poorly understood , oxygen metabolism and ROS production are implicated in the prevalence of these rearrangements in tsa1Δ cells [23] . Noteworthily , base substitution , but not chromosomal rearrangement , was the predominant type of mutation found in our analysis of mutation rates ( Table 1 ) . Thus , the major type of genome instability analyzed in our study is an increased rate of point mutations , but not gross chromosomal rearrangements involving more complex alterations such as translocations , large deletions and amplifications . Our findings point to a role of dNTP levels in determining the mutation rate of tsa1Δ cells . Strong genetic interactions between TSA1 and four RNR regulators DUN1 , SML1 , CRT1 and RAD53 were observed in the context of sensitivity to DNA damage ( Figure 1 and Figure 7 ) , spontaneous mutability ( Figure 2 and Figure 7 ) and dNTP production ( Figure 3 ) . Although the catalytic cysteine of Tsa1p is required for the suppression of mutator phenotype , the mutability of tsa1Δ cells correlated directly with dNTP concentrations ( Figure 2 , Figure 3 , Figure 4 ) , but not with high ROS levels ( Figure 3 and Figure 7 ) . One plausible explanation is that the loss of TSA1 might cause accumulation of both ROS [4] and DNA damage ( Figure 5 ) . This activates the DNA damage checkpoint through Rad53p , Rad9p and Sml1p ( Figure 5 , Figure 6 , Figure 7 ) leading to transcriptional activation of RNR genes ( Figure 6 ) and elevated production of dNTPs ( Figure 3 ) . Once at high dNTP levels , replicative and TLS polymerases by-pass DNA lesions more efficiently to promote survival , but only at the price of increasing mutation rates [24] , [25] . This model implicates dNTP pool expansion as the major culprit in the induction of genome instability in tsa1Δ cells . Indeed , reducing dNTP levels without affecting ROS production was sufficient to reverse the mutator phenotype of tsa1Δ cells ( Figure 3 ) . In particular , tsa1Δ dun1Δ cells have high levels of ROS ( Figure 3C ) , phosphorylated Rad53p ( Figure 5C ) and γH2A ( Figure 5C ) . However , these cells showed a low mutation rate ( Figure 2C ) because the dNTP levels were also low ( Figure 3A ) . On the contrary , increasing dNTP levels by overexpressing RNR1 aggravated the mutator phenotype ( Figure 4 ) . Furthermore , point mutations but not deletions were predominantly found in tsa1Δ cells ( Table 1 ) , implicating a role for dNTP overproduction in compromising genome stability . In further support of this model , compromise of TLS polymerases also suppressed CANR mutations in tsa1Δ cells [23] . We found that the levels of dNTPs in tsa1Δ cells were as high as those in sml1Δ cells ( Figure 3A ) . This finding revealed an unexpected role of TSA1 in the maintenance of dNTP pools in eukaryotic cells . We further observed transcriptional activation of RNR genes in tsa1Δ cells ( Figure 6 ) , which could be mediated through the activation of Rad53p checkpoint . Although this might provide an explanation for the overproduction of dNTPs , exactly how Tsa1p is mechanistically involved in regulating RNR expression remains to be further investigated . Consistent with previous findings [26] , elevation of intracellular dNTPs over a particular threshold level by overexpressing Rnr1p can sufficiently induce a mutator phenotype ( Figure 4 ) . Plausibly , the dNTP levels in tsa1Δ sml1Δ and tsa1Δ crt1Δ cells might have reached the threshold level causing a dramatically increased mutation rate ( Figure 2 and Figure 3 ) . When the elevation of dNTP levels have not reached the threshold as in the case of sml1Δ , crt1Δ and tsa1Δ cells , accumulation of intracellular ROS might serve to trigger or aggravate the mutator phenotype . In tsa1Δ cells , ROS levels were constantly high ( Figure 3C ) causing severe DNA damage ( Figure 5 ) . In contrast , ROS levels were low ( Figure 3C ) and DNA damage was not detected ( Figure 5C ) in sml1Δ or crt1Δ cells . This might explain the higher mutation rate in tsa1Δ cells versus sml1Δ or crt1Δ cells ( Figure 2 ) . Further exacerbation of the mutator phenotype of sml1Δ and crt1Δ cells by ROS such as H2O2 ( Figure 2D ) lent some credence to this model . We demonstrated the requirement of the catalytic cysteine for the ability of Tsa1p to modulate dNTP production ( Figure 3 ) . Through irreversible hyperoxidation , this residue can act as a redox sensor , which triggers the switch of peroxiredoxin from peroxidase to chaperone activity under stress [12] , [13] . In this connection , it would be of great interest to understand whether and how the chaperone activity of Tsa1p might be involved in the regulation of dNTP production . Activation of Rad53p by upstream kinase Mec1p requires adaptor proteins Rad9p and Mrc1p [46] , [47] . We noted that Rad53p phosphorylation was dramatically increased in tsa1Δ versus WT cells ( Figure 6A and Figure 7A ) . In contrast , the increase in Rad9p phosphorylation in the absence of TSA1 was less pronounced ( Figure 6B ) . Although additional experiments are required to investigate the cause of this difference between Rad53p and Rad9p , one possibility is that the deletion of TSA1 might exert a stronger effect on Mrc1p activity . Hypermutability or genome instability is a hallmark of cancer [48] . Mammalian Prx1 is a candidate tumor suppressor gene [15] , [49] . Because peroxiredoxins are highly evolutionarily conserved proteins , an understanding of the mechanism by which yeast Tsa1p protects cell from genome instability might derive novel insight into the tumor suppressive role of Prx1 in mammalian cells . Our work demonstrates the importance of high dNTP levels in the mutability of tsa1Δ cells . Further analysis of dNTP concentrations of Prx1-null mouse cells will reveal whether increased production of dNTPs might be a general mechanism for the generation of genome instability in higher eukaryotes .
S . cerevisiae strains BY4741 [50] and W303-1a , and their isogenic strains ( Table 2 ) were used . All knockout mutants were constructed by one-step gene deletion method [51] . Primers were listed in Table 3 . Expression vector for DUN1 was derived from pRS415 . Expression plasmids for TSA1 and its mutants have been described [18] . Plasmid pGal-RNR1 kindly provided by Dr . Stephen Elledge has also been described previously [52] . Rates of spontaneous forward mutations to confer CANR or 5FCR were measured as described [18] , [53] . Spectra of CANR mutations were determined by DNA sequencing . Ten independent cultures were analyzed in each experiment . Cell extracts were prepared and dNTP levels were measured with Klenow enzyme and [3H] labeled dATP or dTTP ( PerkinElmer ) as described [54] . Standard curves were used to estimate the cellular dNTP levels . Three independent cultures were analyzed in each experiment . Intracellular ROS levels were measured by fluorimetry using DCF ( Molecular Probes ) as described [4] , [18] . Total RNA was extracted by phenol/freeze RNA preparation method as described [55] . For RT-PCR , 3 µg of total RNA was used for cDNA synthesis . Semi-quantitative PCR was performed and optimized to ensure that the amplification was in the linear range . PCR primers were listed in Table 3 . Western blot analysis was performed essentially as described [31] . Yeast cells were harvested by centrifugation , followed by trichloroacetic acid extraction with the help of glass beads . | Peroxiredoxins are a family of antioxidant enzymes highly conserved from yeast to human . Loss of peroxiredoxin in mice can lead to severe anemia and malignant tumors , but the underlying cause is not understood . One way to derive new knowledge of peroxiredoxins is through genetic analysis in yeast . We have shown that loss of peroxiredoxins in yeast is associated with an increase in mutation rates . Here , we demonstrate that this elevation of mutation rates in yeast cells lacking a peroxiredoxin is due to increased production of deoxyribonucleoside triphosphates ( dNTPs ) , the building blocks of DNA . Our findings suggest a new model in which compromised antioxidant defense causes accumulation of damaged DNA and activation of the DNA damage checkpoint . For yeast cells to survive DNA damage , dNTP production is increased to facilitate DNA replication , but at the price of high mutation rates . This new model could lead to a better understanding of human diseases including cancer . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry/replication",
"and",
"repair",
"genetics",
"and",
"genomics/gene",
"function"
] | 2009 | Loss of Yeast Peroxiredoxin Tsa1p Induces Genome Instability through Activation of the DNA Damage Checkpoint and Elevation of dNTP Levels |
Completion of early stages of retrovirus infection depends on the cell cycle . While gammaretroviruses require mitosis for proviral integration , lentiviruses are able to replicate in post-mitotic non-dividing cells . Resting cells such as naive resting T lymphocytes from peripheral blood cannot be productively infected by retroviruses , including lentiviruses , but the molecular basis of this restriction remains poorly understood . We demonstrate that in G0 resting cells ( primary fibroblasts or peripheral T cells ) , incoming foamy retroviruses accumulate in close proximity to the centrosome , where they lie as structured and assembled capsids for several weeks . Under these settings , virus uncoating is impaired , but upon cell stimulation , Gag proteolysis and capsid disassembly occur , which allows viral infection to proceed . The data imply that foamy virus uncoating is the rate-limiting step for productive infection of primary G0 cells . Incoming foamy retroviruses can stably persist at the centrosome , awaiting cell stimulation to initiate capsid cleavage , nuclear import , and viral gene expression .
After entry into their host cells , retroviruses undergo uncoating and reverse transcription , leading to the formation of pre-integration complexes , which must then gain access to the nuclear compartment and integrate the provirus into host chromosomes [1] . The cell cycle status is a key determinant for completion of these early stages of infection [2] , and retroviruses have been classified based on their ability to productively infect non-cycling cells . Gammaretroviruses , like murine leukemia virus , require mitosis for proviral integration [3] , while lentiviruses such as human immunodeficiency virus type 1 ( HIV-1 ) show almost no difference between dividing and non-dividing cells [4] . Indeed , HIV-1 and other lentiviruses can replicate in terminally differentiated and post-mitotic cells such as neurons or macrophages [5 , 6] . However , like murine leukemia virus , HIV-1 cannot infect naive quiescent CD4-positive T cells or monocytes isolated from peripheral blood that are in the G0 stage of the cell cycle [2 , 7] . Since reverse transcription does occur in these conditions [8] , it is conceivable that other host cell proteins or processes are necessary for the completion of the viral cycle [9] . Foamy viruses ( FVs ) are complex retroviruses isolated from different animal species , mainly in non-human primates . FVs share with all retroviruses the same organization of the genome , which encodes Gag , Pol , and Env proteins . In addition , the FV genome encodes at least two other proteins , Tas and Bet , which are not incorporated into the viral particle . Remarkably , FVs exhibit some features related to the hepatitis B virus [10] . In particular , they reverse transcribe their RNA genome during the late stages of infection , leading to the presence of infectious viral DNA in extracellular virions [11] . Moreover , the structural FV Gag presents specific characteristics that set it clearly apart from other retroviral Gag proteins . In particular , FV Gag maturation by the viral protease does not lead to the formation of the canonical matrix , capsid , and nucleocapsid products . Rather , the Gag precursor is partially cleaved by the viral protease near its C terminus into a mature product , before or during budding . This results in the presence of two Gag proteins of 71 kDa and 68 kDa in extracellular virions [12] . We have previously reported that upon entry into target cells and prior to nuclear translocation , incoming FV capsids traffic along the microtubule network to reach the microtubule-organizing centre ( MTOC ) [13 , 14] , which includes the centrosome in animal cells [15] . A similar route has been described for HIV-1 [16] . Drugs disrupting the microtubule network , such as nocodazole or colchicine , largely prevent early intracellular FV trafficking [13] , as well as that of HIV-1 [16] . Gag also targets the pericentrosomal region for capsid assembly during the late stages of infection [17] . Similar to murine leukemia virus , productive FV infection requires passage through mitosis [18 , 19] . Like all other animal retroviruses , FVs do not productively infect cells arrested in the G0 stage of the cell cycle , such as peripheral T lymphocytes or growth-arrested fibroblasts in vitro [19] . To gain insight into this restriction , we have focused our attention on early stages of FV replication in G0 cells . Here , we demonstrate that incoming FVs stably localize at the vicinity of the MTOC as structured and assembled capsids for several weeks in resting cultures in vitro . Upon cell stimulation , Gag proteolysis and capsid disassembly take place , allowing infection to proceed . Altogether , these data demonstrate that the centrosome represents a cellular site around which incoming viruses persist as a stable pre-integration intermediate in resting cells , and that virus uncoating is the rate-limiting step for FV infection in growth-arrested cells .
Cycling and resting human primary MRC5 fibroblastic cells were infected with the prototypic primate foamy virus ( PFV ) at a multiplicity of infection ( m . o . i . ) of 1 , and cell-free supernatant was collected 48 h later for virus titration . Consistent with previous reports [18 , 19] , we found that PFV does not productively infect G0 resting cells ( unpublished data ) . To investigate the molecular mechanisms involved in the restriction of FV replication cycle in G0 resting cells , we first analyzed the distribution of incoming viral components in resting MRC5 cells . As early as 4 h post-infection ( p . i . ) . , we observed that incoming Gag proteins localized at the centrosome , which was revealed with anti-γ-tubulin antibodies ( abs ) . These findings are consistent with previous observations showing the pericentrosomal concentration of incoming FVs after trafficking along the microtubule network in cycling cells [13] . Remarkably , while in cycling cells , incoming Gag antigens are no longer observed near this organelle 10 h p . i . ( [20]; unpublished data ) ; in resting cells , centrosomal localization of viral antigens was consistently observed up to 30 d p . i . ( Figure 1A ) . We then investigated the localization of the viral genome in infected resting cells . By fluorescent in situ hybridization , using the entire PFV DNA genome as a probe , we found that the incoming viral DNA genome also localized at the centrosome in resting MRC5 cells 15 d p . i . ( Figure 1B ) . Therefore , both incoming FV antigens and the viral DNA genome reside at the MTOC of resting cells for several weeks after infection . To visualize the status ( assembled capsids or not ) of incoming viruses in resting cells , MRC5 cells were analyzed by electron microscopy ( EM ) at different time points after infection . In cycling cells ( Figure 2A ) , incoming FVs were observed at the centrosome 4 h p . i . , mainly as structured and assembled capsids , confirming that uncoating is not complete at this time point [20] . At later time points , these viral capsids completely disassembled in cycling cells as already reported [20] , and viral capsids were never detected in uninfected cells ( Figure 2A ) . Importantly , assembled and structured viral capsids were observed around the centrosome in resting cells 5 or 15 d p . i . ( Figure 2A ) , strongly suggesting that virus uncoating is impaired under these settings . FV uncoating requires the enzymatic activity of both viral and cellular proteases , which cleave the major structural components of viral capsids ( the 71–68 kDa Gag doublet ) , into shorter fragments [20] . Among these cleavage products , a 38-kDa Gag-derived product specifically results from the action of the viral protease [20] . To confirm at the biochemical level that FV uncoating is inhibited in resting cells , the status of the Gag polyprotein was determined . For that purpose , resting and cycling MRC5 cells were infected with PFV at an m . o . i . of 1 , and intracellular viral proteins were analyzed by Western blot using mouse anti-Gag abs . Figure 2B shows that Gag cleavage products , in particular the 38-kDa fragment , were easily detected in cycling cells . On the contrary , FV Gag was not cleaved following infection of resting MRC5 cells and remained as an intact doublet of 71 and 68 kDa ( Figure 2B ) . These results demonstrate the absence of Gag cleavage in resting cells , reflecting the absence of viral uncoating observed by EM . To assess whether incoming FV capsids at the centrosome of resting cells constitute a stable pre-integration intermediate , which can later be reactivated for productive infection , resting PFV-infected MRC5 cell cultures were stimulated to divide by splitting and serum addition . At different time points following this activation , Gag expression and distribution were analyzed by indirect immunofluorescence using mouse anti-Gag abs . Twenty-four hours after cell activation , we observed that , although Gag still associated to the centrosome , it could be detected in both the cytoplasm and the nucleus of 20% of the cells ( Figure 3A ) . Gag was detected in both compartments in the entire culture 48 h after cell activation , and the formation of numerous syncytia was detected 96 h post-activation ( Figure 3A ) . These results demonstrate that viral replication , which was inhibited in resting cells , resumes upon cell activation . To confirm that cell activation actually triggered virus uncoating , the status of the Gag polyprotein was studied by Western blot . To exclusively analyze incoming viral antigens , avoiding contamination from Gag synthesis and degradation , which might occur following reactivation , these experiments were performed under cycloheximide ( CHX ) treatment , a translation inhibitor . At 24 h post-reactivation , several Gag cleavage products , notably the 38-kDa fragment [20] , were clearly detected in CHX-treated cells ( Figure 3B ) . Moreover , under these settings , accumulation of Gag was observed only in untreated reactivated cells . Altogether , these observations demonstrated that entry into the cell cycle triggered virus uncoating , as assessed by Gag cleavage , and productive infection . Several reports have demonstrated that FVs infect lymphocytes in vivo and that infectious particles can be recovered from peripheral T cells of infected animals [21−24] . Therefore , the intracellular distribution of incoming FVs was assessed in one of its natural targets . To this aim , primary resting human CD4-positive T cells were infected with PFV at an m . o . i of 1 and were maintained in a resting state in culture for 5 d without addition of exogenous lymphokines . PFV infection of resting CD4-positive T cells was non-productive ( unpublished data ) and did not trigger cell activation ( Figure 4A ) . In these cells , localization of incoming capsids was analyzed by confocal microscopy . Consistent with our previous observation of infected resting MRC5 cells , incoming Gag strictly localized at the centrosome from day 2 to day 5 p . i . ( Figure 4B ) . On the contrary , Gag was diffusely distributed in the cytoplasm and the nucleus of activated CD4-positive T cells , indicating that productive replication was taking place in these cells ( Figure 4B ) [25] . When resting CD4-positive T cells were stimulated to divide 5 d p . i . , Gag was no longer observed uniquely at the centrosome but localized diffusely in the cytoplasm and the nucleus at 24 h post-activation ( Figure 4B ) . Taken together , these results confirm that a stable PFV intermediate can persist in the vicinity of the MTOC in primary resting CD4-positive T cells and that T cell activation allows completion of the viral replication cycle .
For all retroviruses , completion of the early steps of the replication cycle depends on the cell cycle status ( reviewed in [2 , 9] ) . We show that in vitro FV infection is restricted in G0 resting cells , either from fibroblastic ( MRC5 ) or lymphocytic ( CD4-positive T cells ) origin . We further demonstrate that incoming viral capsids persist in infected G0 resting cells as a stable pre-integration intermediate over a period of at least 30 d in MRC5 cells . Under these conditions , Gag proteolysis , and consequently virus uncoating , is blocked , and incoming viruses are maintained as assembled and structured capsids around the centrosome . Upon cell activation , Gag is cleaved , viral capsids disassemble , and infection proceeds . Post-mitotic cells such as neurons or macrophages are productively infected by lentiviruses . In contrast , resting G0 cultures in vitro , such as naive T lymphocytes isolated from peripheral blood , cannot be productively infected by any classes of retroviruses , including HIV-1 [8 , 26−30] . Since reverse transcription is completed in these cells [8 , 28] , additional blocks seem to occur during the early stages of the virus life cycle . Several hypotheses have been raised to elucidate the molecular basis of this restriction . It has been suggested that APOBEC3G , a cellular antiretroviral protein that is associated with the hypermutation of viral DNA through cytidine deamination [31] , could inhibit HIV replication as part of a low molecular mass ribonucleoprotein complex in resting T cells . This seems to impair the formation of HIV-1 late reverse transcription products [29] . A recent alternative hypothesis suggests that virus uncoating represents the main rate limiting stage in resting T CD4+ cells , since cellular extracts from activated , but not resting cells , support uncoating of HIV cores in vitro [30 , 32] . Clearly , our observations suggest this second scenario in the case of FVs . Interestingly , our observations might explain the efficient in vivo transduction of haematopoietic stem cells by FV-derived vectors in mice [33 , 34] . Indeed , despite the fact that FVs cannot productively infect resting stem cells in vitro , FV vectors can repopulate bone marrow [33−36] . In fact , in vivo implantation of transduced resting stem cells likely triggers their activation , allowing the infection to proceed . The remarkable stability of the FV intermediate in resting cells that we have evidenced here could be related to the particular mode of replication of these viruses . First , virus uncoating is a relatively late event during FV replication , allowing incoming viral antigens to accumulate in close vicinity to the nuclear compartment as assembled and structured capsids [13 , 20] . On the contrary , for HIV-1 , capsid disassembly in activated and cycling cells seems to start as soon as the viral particles enter into the cytoplasm [16 , 37] . Second , in contrast to other retroviruses , the presence of an infectious viral DNA genome in incoming capsids could make FV less dependent on the metabolism of the target cell [11] . Our data also demonstrate that the centrosome is a cellular site around which incoming FVs can stably persist , awaiting further cell stimulation for completion of the viral cycle . Recent studies have shown that the centrosome is not a mere spectator of the cell cycle but can exert significant control over it [38] . By providing a scaffold for many cell cycle regulators and their activities [38−40] , the centrosome influences cell cycle progression , especially during the transition from G1 to S phase [41 , 42] . Therefore , this organelle receives and integrates signals from outside the cell and facilitates conversion of these signals into cellular functions . Maintenance of viral capsids at the vicinity of the centrosome in resting cells could be a strategy that some viruses have evolved to rapidly respond to growth stimuli received by the cell . The cellular signal ( s ) triggering the uncoating process upon cell stimulation remains unclear , but is likely linked to the centrosome cycle .
Human MRC5 fibroblasts were cultured as described [14] . Quiescent MRC5 fibroblasts were generated following confluence , serum starvation , and addition of 10−6 M dexamethasone . The cells were cultured in these conditions for 10 d before infection . The cell cycle status was analyzed by pyronine staining as described by [43] . In resting MRC5 cells , less than 3% of cells remain activated , as already reported ( [44]; unpublished data ) . For cell activation , cells were stimulated to divide by subculture and serum addition at different times after infection . CHX treatment was performed at 150 μg/ml immediately after cell activation and maintained during the entire experiment . Peripheral blood mononuclear cells obtained from healthy donors after informed consent were separated on lymphocytes separation medium ( Eurobio , http://www . eurobio . fr ) . CD4-positive T cells were negatively selected using the CD4-positive T Cell Isolation Kit II ( Miltenyi Biotec , http://www . miltenyibiotec . com ) . Briefly , cells were labeled using a cocktail of biotin-conjugated abs against CD8 , CD16 , CD19 , CD36 , CD56 , CD123 , TCRγδ , and glycophorin A , and anti-biotin magnetic beads . After washing , negative cells were selected by magnetic separation with the autoMACS Separator ( Miltenyi Biotec ) . The cells were then labeled with FITC-anti-CD8 ( clone SK1; BD Biosciences , http://www . bdbiosciences . com ) , anti-CD14 ( Clone TUK4 , Miltenyi Biotec ) , Phycoerythrin-conjugated anti-CD25 ( clone 4E3 , Miltenyi Biotec ) , and anti HLA-DR ( L243 , BD Biosciences ) ab , washed , and sorted on a FACSVantage . The purity was examined by flow cytometry with the same ab and an allophycocyanin-conjugated anti-CD4 ( clone RPA-T4 , BD Biosciences ) . Typically , cells were more than 99% negative for the activation markers ( CD25 and HLA-DR ) and more than 98% positive for the CD4 . Activated CD4-positive T cells were prepared by incubating negatively selected CD4 T cells in RPMI 10% normal human serum and 1 μg/ml PHA-L ( Sigma , http://www . sigmaaldrich . com ) . At day 3 , 150 UI/ml of interleukin 2 ( IL-2; Promocell , http://www . promocell . com ) was added to the cells . Activation status of the control and infected CD4 T cells were checked by flow cytometry with the ab anti-CD4 FITC and an anti-CD25 or HLA DR PE ab . MRC5 cultures or CD4-positive T cells were infected by spinoculation at an m . o . i . of 1 for 1 h 30 min at 30 °C . Titres were determined by infection using FAG cells ( BHK cells stably harbouring the GFP gene under the control of the PFVU3 promoter ) [45] . MRC5 cells grown on glass coverslips were infected with PFV at an m . o . i . of 1 by spinoculation for 1 h 30 min at 30 °C . After different times p . i . , cells were rinsed with PBS , fixed for 10 min at 4 °C with 4% PFA , and permeabilized for 5 min at −20 °C with ice-cold methanol . Cells were incubated successively with mouse polyclonal anti-Gag serum overnight at 4 °C ( 1/250 ) and with rabbit polyclonal γ-tubulin antiserum ( 1/2000; Abcam , http://www . abcam . com ) for 1 h at 37 °C . Cells were washed and incubated for 1 h with a 1/500 dilution of appropriate fluorescent-labeled secondary abs . Nuclei were stained with 4'-6-diamidino-2-phenylindole ( DAPI ) and the coverslips were mounted in Moviol . Confocal microscopy observations were performed with a laser-scanning confocal microscope ( LSM510 Meta; Carl Zeiss , http://www . zeiss . com ) equipped with an Axiovert 200 M inverted microscope using a Plan Apo 63_/1 . 4-N oil immersion objective . Fifteen days following PFV infection ( m . o . i . of 1 ) , arrested MRC5 cells were fixed with 4% PFA , permeabilized with 0 . 2% Triton X-100 , and incubated with γ-tubulin antiserum ( Abcam ) . Then , cells were fixed a second time in 4% PFA for 20 min at room temperature to cross-link bound abs . Incubation with the secondary ab was performed during the fluorescent in situ hybridization detection step , performed as described [20] . Briefly , cells were treated with RNase at 100 μg/ml in PBS for 30 min at 37 °C and incubated with probe ( plasmid p13 containing the entire PFV genome ) overnight at 37 °C . Probes were labeled with FITC-avidin DN ( 1/200; Vector Laboratories , http://www . vectorlabs . com ) and signals were amplified with biotinylated anti-avidin D ( 1/500 , Vector Laboratories ) , followed by another round of FITC-avidin staining . Finally , cells were stained for DNA with DAPI and mounted in Vectashield . Confocal observations were performed as previously described . Cell pellets were lysed in Triton buffer ( 10 mM Tris [pH 7 . 4]; 50 mM NaCl; 3 mM MgCl2; 1 mM CaCl2; orthovanadate , benzamidine , and protease inhibitor cocktail [Roche , http://www . roche . com] at 1 mM each; 10 mM NaF; and 0 . 5% Triton X-100 ) for 30 min at 4 °C and centrifuged for 15 min at 20 , 000g . Resulting pellets were treated with radioimmunoprecipitation buffer ( 10 mM Tris [pH 7 . 4]; 150 mM NaCl; orthovanadate , benzamidine , and protease inhibitor cocktail at 1 mM each; 10 mM NaF; 1% deoxycholate; 1% Triton X-100; and 0 . 1% sodium dodecyl sulfate [SDS] ) during an additional 30 min at 4 °C , centrifuged for 15 min at 20 , 000g , collected , and diluted in Laemmli buffer . Samples were migrated on an SDS–10% polyacrylamide gel , and proteins were transferred onto cellulose nitrate membrane ( Optitran BA-S83; Schleicher-Schuell , http://www . schleicher-schuell . com ) , incubated with appropriated abs , and detected by enhanced chemiluminescence ( Amersham ECL Advance Western Blotting Detection Kit , http://www . gelifesciences . com ) . For EM studies , MRC5 cells were infected at an m . o . i . of 5 as described above and fixed in situ by incubation for 48 h in 4% PFA and 1% glutaraldehyde in 0 . 1 M phosphate buffer ( pH 7 . 2 ) ; they were then post-fixed by incubation for 1 h with 2% osmium tetroxide ( Electron Microscopy Science , http://www . emsdiasum . com ) . Next , MRC5 cells were dehydrated in a graded ethanol series , cleared in propylene oxyde , and then embedded in Epon resin ( Sigma ) , which was allowed to polymerize for 48 h at 60 °C . Ultrathin sections were cut , stained with 5% uranyl acetate/5% lead citrate , and then placed on EM grids coated with collodion membrane . They were then observed with a Jeol 1010 transmission electron microscope ( Jeol , http://www . jeol . com ) . | Naive quiescent CD4-positive T cells or monocytes that are in the G0 stage of the cell cycle cannot be productively infected by retroviruses in vitro , but the molecular basis of this restriction remains poorly understood . In this report , we demonstrate that incoming foamy retroviruses remain around the centrosome as structured and assembled capsids for weeks in resting cultures . Under these conditions , virus uncoating is impaired , but upon cell activation , viral capsids undergo proteolysis and disassembly , allowing infection to proceed . Maintenance of incoming viral capsids at the centrosome in resting cells could be a strategy that viruses have evolved to rapidly respond to stimuli received by the cell . The cellular signal triggering the uncoating process upon cell stimulation remains unclear , but is likely linked to the centrosome cycle . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"viruses",
"cell",
"biology",
"microbiology",
"virology"
] | 2007 | Centrosomal Latency of Incoming Foamy Viruses in Resting Cells |
The pathogenesis of bacteraemia after challenge with one million pneumococci of three isogenic variants was investigated . Sequential analyses of blood samples indicated that most episodes of bacteraemia were monoclonal events providing compelling evidence for a single bacterial cell bottleneck at the origin of invasive disease . With respect to host determinants , results identified novel properties of splenic macrophages and a role for neutrophils in early clearance of pneumococci . Concerning microbial factors , whole genome sequencing provided genetic evidence for the clonal origin of the bacteraemia and identified SNPs in distinct sub-units of F0/F1 ATPase in the majority of the ex vivo isolates . When compared to parental organisms of the inoculum , ex-vivo pneumococci with mutant alleles of the F0/F1 ATPase had acquired the capacity to grow at low pH at the cost of the capacity to grow at high pH . Although founded by a single cell , the genotypes of pneumococci in septicaemic mice indicate strong selective pressure for fitness , emphasising the within-host complexity of the pathogenesis of invasive disease .
Streptococcus pneumoniae , one of the major human bacterial pathogens , is also part of the normal upper respiratory tract flora , where nasopharyngeal colonisation with one or more strains often lasts weeks to months with seasonal peaks in late winter [1] , [2] . Carriage of S . pneumoniae ( pneumococci ) may result in disease as the consequence of contiguous spread from the nasopharynx to other sites in the upper or lower respiratory tract causing , for example , otitis media or pneumonia . More rarely , there is hematogenous dissemination of pneumococci resulting in septicaemia and metastatic disease such as meningitis [1] , [3]–[5] . In experimental models of pneumococcal infection , the challenge dose required to induce disease is dependent on the route of infection , the genetic background of the host and the virulence of the infecting strain [6] and may vary from a very few to millions of organisms [7] . Following intravenous inoculation of mice with laboratory grown pneumococci , a hallmark of experimental bacteraemic infections is the rapid and efficient clearance of most of the inoculated bacteria [8]–[10] . In non-immune rodents , major factors mediating this clearance are splenic macrophages and complement mediated opsonisation [11]–[14] . A challenge dose of about one million virulent , encapsulated pneumococci is generally needed to induce bacteraemia in about half of challenged animals ( the effective dose or ED50 ) and which is the dose generally used to address investigations into the early events shaping an infectious process . Most work on the pathogenesis of infectious disease focuses on specific virulence determinants which are generally presented as the cause , either alone or in combination with other factors , of the events leading to the infection of the host where the microbial population is considered to be a uniform entity . However , several investigations have addressed the within host population dynamics , especially on the early phases of host-pathogen interactions [15] . There are different models which address these early events that include: ( i ) the model of independent action , which postulates that at the LD50 ( lethal dose for 50% for the hosts ) the hosts develop infection “following the multiplication of only one of the inoculated bacteria , regardless of the total number of bacteria inoculated” [16] , ( ii ) the hypothesis of synergy which “postulates that inoculated bacteria co-operate and that fatal infections will be initiated by more than one bacterium and that this will lead to the predominance of several variants” [16] , and models which introduce time as a factor into the process and propose a two-stage model where a birth–death phase would be responsible for generation of the heterogeneity within the population later during the infection [17] . Both for viral and bacterial infections it has been shown that the effective number of infectious agents which actually start the disease is generally many orders of magnitude below the actual dose used for challenge [16] , [16] , [18]–[21] . In particular a series of reports , generally based on experimental challenge using an inoculum containing an approximately equal mixture of two isogenic variants at the LD50 , has shown experimentally that systemic infections may be initiated by the multiplication of as few as a single organism [22]–[27] . Models differ widely , and due to the nature of pathogenesis of many infections , they rely on experimental challenge at a site different to that investigated for disease implying that multiple bottlenecks occur and that potentially a series of invasive events could be enucleated [16] , [18]–[27] . We here have investigated the host and microbial determinants that underpin the occurrence of the single cell bottleneck in the pathogenesis of pneumococcal septicaemia following inoculation of mice with three isogenic variants by the intravenous route . This route has an advantage over other experimental infection models as there are less biological events between the initial challenge and full blown disease so that rigorous analysis of the events is facilitated .
CD1 mice were inoculated intravenously ( i . v . ) with a mixture of pneumococci comprising approximately equal numbers of each of three isogenic TIGR4 mutants ( FP122 , FP321 and FP318 ) with different resistance markers ( Table 1 ) [28] . Following inoculation , blood samples were collected at different times and spread on selective plates . Colony counts allowed quantitation of the distribution of the different mutants making up the pneumococcal population in the blood ( Figure 1 ) . Two hours after bacterial challenge , blood samples from all mice , with one exception , grew all three of the variants that had been included in the challenge dose . Samples of the second group of mice , sampled one hour thereafter , showed a mixed population of the variants: two in 6 mice , three in 17 mice and one mouse had negative blood culture ( Figure 1A ) . At 7 h after challenge , this pattern was distinctly different: there were 10 positive and 2 negative blood cultures and the numbers of bacteria were significantly reduced . At 8–9 h post-infection , most blood cultures ( 25/29 mice ) were negative . The remaining 4 mice had monoclonal blood cultures in that each grew colonies of only one mutant ( Figure 1A ) . In the subsequent hours of infection , bacteria were detected in the blood at high concentrations ( up to 1×106 CFU ) in 17/55 ( 31% ) mice at 24 h , 16/43 ( 37% ) mice at 48 h and 15/24 ( 62% ) mice at 72 h ( Figure 1B ) . At all these time points , most blood samples yielded colonies of only one of the variants: 12 out of 17 blood cultures at 24 h , 12/16 at 48 h and 8/16 at 72 h . Among the 32 single-variant blood samples each variant was more or less equally represented as the progenitor , although owing to the smaller challenge dose of strain FP318 in one experiment , this strain was recovered at lower density from the blood and caused fewer single-variant infections ( Table S1 ) . In three of twelve mice for which three serial blood cultures were taken we observed an increase in the number of variants . One of these mice showed evidence of infection with 3 variants at 48 h after earlier having a monoclonal infection and a further two mice had three variants infection at 72 h after having previously had infection with only one or two variants respectively . Seven of the single-variant bacteraemia isolates were checked for colony morphology and each was found to have the opaque phenotype , in contrast to the challenge strains ( not-mouse-passaged ) that yielded a mixed population of about similar proportions of opaque and translucent colonies [29] . The bacterial counts from cultures of spleen tissue were assayed in twelve mice at each time point ( Figure S1 ) . All mice which had positive blood cultures showed bacterial counts also in splenic samples . In addition , small numbers of organisms were cultured from spleen tissue both at 24 and 48 h in two mice each which presented with sterile blood cultures . Similarly at 72 h , bacteria were only detected in the spleen in one mouse ( Figure S1 ) . These data show that infectious foci can be detected in the spleens of mice that have negative blood cultures , indicating that the spleen is the probable site where the infection originates . Our data indicate that the near totality of bacteria in the challenge with three isogenic pneumococcal variants is cleared by the immune system ( predominantly by splenic macrophages; see below ) and that few bacterial cells remain viable within a defined site of the host ( i . e . the spleen ) . This small number of bacteria may start to grow and re-invade the host giving rise to bacteraemia . In our experimental infection most of the mice challenged were not bacteraemic and , of those becoming bacteraemic , most were infected by a single bacterial variant . In addition , we detected in some mice an increase in variants within blood cultures over time . These data indicate that over time more than one invasion event may occur . Theoretically bacteraemia may be generated in two ways: ( i ) by a single bacterium establishing a population in the blood in a single invasion event or several bacteria each independently establishing a population in distinct invasion events ( independent action ) ; ( ii ) by a defined number ( more than one ) of bacteria acting together to invade once or several times ( co-operative synergism ) [22] . We hypothesise that the former explanation pretains . To statistically evaluate the number of bacteria involved in founding the blood population in each invasion event , we construct a model that assumes that bacteria invade and establish a population in the blood at random ( Supplementary text ) [23] . In this model , the number of invasion events in each mouse is assumed to follow a Poisson distribution so the expected number can be estimated from the proportion of the non-bacteraemic mice . Then we determined the expected numbers of mice infected with one , two or three variants , assuming that the number of bacteria ( w ) responsible for establishing the blood population in each invasion event were 1 , 2 , 3 , etc ( Table 2 ) . Given that some mice were culled during the course of the experiment and some got multiple samplings , we limited the statistical analysis to the observations at the 24 h time point . In table 2 , we report the comparison between the expected versus the observed numbers of the infected mice with one , two and three variants for different numbers of bacteria ( w ) potentially responsible for founding a population blood . The statistical analysis shows that the most probable number of bacteria responsible for establishing a blood population is 1 ( Table 2 ) . Note that the p-value was calculated by combining data for blood cultures with two or three variants because of the small expected frequencies in these two categories . Given that w is equal to one , we conclude that polyclonal blood infections are the result of more than one invasion event , each event founded by a single bacterium , consistent with the observed time-dependent increase in the frequency of polyclonal bacteremia over the 72 h of the experiment . Prior to the assessment of the impact of host factors in the control of bacteraemia , we compared bacterial counts in the blood of two mouse strains known to be resistant to pneumococcal infection ( outbred CD1 mice and inbred BALB/c mice ) and a susceptible mouse strain CBA/Ca [30] . The mice were inoculated i . v . with a mixture of three encapsulated pneumococcal strains of different serotypes , D39 ( type 2 ) , TIGR4 ( type 4 ) and G54 ( type 19F ) . Bacterial clearance in CD1 and BALB/c mice showed similar kinetics ( Figure 2 A–B ) , while CBA/Ca mice were less able to reduce the initial number of bacteria ( Figure 2 C ) . All mouse strains cleared G54 bacteria immediately ( no positive blood culture 10 min after infection ) and showed a first phase of rapid clearance also for both D39 and TIGR4 . Only the two resistant mouse strains showed bacterial numbers in the blood that were less than the limit of detection . In contrast , bacterial numbers increased in the susceptible strain after the first phase ( Figure 2 A–C ) . This indicates that , depending on which host-pathogen pair was investigated , the bottlenecks may vary considerably . Since BALB/c mice showed a more uniform clearance of bacteria , subsequent experiments were conducted with this mouse strain in order to keep experimental groups to a minimal size . To identify the host immune cells responsible for the initial clearance of bacteria from the blood , we performed a set of experiments in BALB/c mice depleted either of macrophages or neutrophils . Macrophage depletion was achieved by intraperitoneal ( i . p . ) injection of clodronate liposomes and neutrophil depletion by using anti-GR-1 monoclonal antibody [31]–[33] . Control groups were treated either with PBS-containing liposomes as control for clodronate experiments or with isotype-matched antibody in the case of experiments with anti-GR-1 . The results obtained for the control groups were comparable to the untreated control mice and differed from the groups of mice treated with clodronate ( Figure S3 A–C ) or anti-GR-1 ( Figure S3 E–G ) . To verify macrophage or neutrophils depletion , spleen samples were analyzed by flow cytometry . The reduction of macrophages in the spleen of clodronate-treated mice was 61%±14 . 2 measure by anti-F4/80 and 47%±10 . 4 by anti-CD11b compared to naïve mice . Similar results were obtained when liver samples were analyzed ( Figure S3 D ) . In anti-GR-1-treated mice , the neutrophil number was reduced by 83%±2 . 7 as compared to control mice ( Figure S3 H ) . To check for anti-pneumococcal antibodies in naïve mice , we evaluated the reactivity of mouse serum towards whole pneumococcal cells . Mice had no detectable serum antibodies to any of the pneumococcal serotypes ( Figure S3 I–K ) . A result supported by the observation that addition of type specific rabbit serum to the blood from naïve mice conferred specific bactericidal activity ( P<0 . 01 ) . To analyse the role of macrophages and neutrophils , mice were divided into three groups: untreated ( Figure 3 A1–A4 ) , clodronate-treated ( Figure 3 B1–B4 ) , and anti-GR1-treated ( Figure 3 C1–C4 ) . After i . v . challenge with a mixture of four different strains ( TIGR4 , D39 , DP1004 and G54 ) , time course of bacterial counts was monitored by sampling blood , spleen , lung , liver and kidney ( Figure 3 and S2 ) . Analysis of control mice allowed categorisation of the pneumococcal strains into two groups: TIGR4 and D39 , which were slowly cleared ( virulent strains ) , and G54 and DP1004 , which were cleared from the blood within minutes . The counts of TIGR4 and D39 were higher in the blood than in the spleen at 5 min and at 4 h compared to the other two strains ( P<0 . 05 ) . Bacterial loads in the other organs were similar to those found in the spleen ( Figure S2 ) . In contrast , mice infected with strains DP1004 and G54 showed higher CFU counts in the spleen than in the blood and other organs ( P<0 . 05 at 5 min for both strains , P<0 . 01 at 4 and 8 h for DP1004 ) ( Figure 3 A3–4 ) . The groups of mice depleted of macrophages showed significantly reduced ability to clear bacteria from the bloodstream . An increase in bacterial numbers in blood from 5 min to the later time points was observed in mice infected with TIGR4 ( Figure 3 B1 ) and D39 ( Figure 3 B2 ) ( P<0 . 01 ) . Blood bacterial counts were significantly higher in the clodronate-treated mice than in the control group ( P<0 . 05 for all time points for both D39 and TIGR4 ) . Bacterial counts of TIGR4 and D39 in liver and spleen were lower but with a similar trend , over time , to those in the blood ( Figure 3 B1–B2 and Figure S2 B1–B2 ) . In clodronate-treated mice , the numbers of non-virulent bacteria ( strains G54 and DP1004 ) were higher in the blood than in the spleen ( P<0 . 05 at 5 min for DP1004 and P<0 . 05 at 5 and 4 h for G54 ) and paralleled the trend observed for the virulent strains TIGR4 and D39 in untreated animals ( Figure 3 B3–B4 compared to A1–A2 ) . In neutrophil-depleted mice , bacterial counts of both TIGR4 ( Figure 3 C1 ) and D39 ( Figure 3 C2 ) in blood and spleen decreased in the first 4 h after challenge with a similar trend to that observed in untreated animals . Thereafter , blood and organ counts remained stable ( Figure 3 C1–C2 and Figure S2 C1–C2 ) . For both virulent strains , the number of bacteria were higher in the blood than in the spleen ( P<0 . 01 at 5 min ) . Interestingly , strain G54 had a peculiar behaviour in neutrophil-depleted mice , as it persisted in the spleen at high levels throughout the whole experiment despite being cleared from the blood within a few min of infection ( P<0 . 001 at 4 and 8 h ) , ( Figure 3 C4 ) . At 4 h post-infection G54 bacteria reappeared in the blood ( Figure 3 C4 ) . The experiment was repeated for the later time points , and the pattern of counts was identical . The rough DP1004 strain was cleared from each body site as well as in untreated mice ( Figure 3 C3 ) . To determine more precisely the early events occurring in the clearance of pneumococci , we have plotted separately the data on blood bacterial counts obtained 5 min and 30 min after challenge ( Figure 2 D–E ) . In untreated mice , bacterial blood counts of the invasive strains D39 and TIGR4 were respectively 6 . 2×104 and 5 . 4×104 CFU/ml , while those of the non-invasive strains , DP1004 and G54 , were three times lower ( reduction of 60 to 75% ) . Differences between the virulent and non-virulent strains were statistically significant ( P<0 . 01 ) . Given the key role of splenic macrophages in pneumococcal clearance , we evaluated the capacity of splenic macrophages to internalise pneumococci . Splenic BALB/c macrophages were grown as primary cell cultures , washed and re-cultured for seven days in M-CSF supplemented medium . Cytoflourimetric data showed expression of the characteristic markers of splenic macrophages , CD11b , CD11c , F4/80 and SIGLEC-1 ( Figure 2 H ) [34] . Adhesion was evaluated by counting pneumococci after 45 min and phagocytosed bacteria were enumerated by plating after a further 30 min of antibiotic treatment ( viable intracellular bacteria ) . Despite similar values in adherent cells ( Figure 2 F ) , our data show higher numbers of intracellular bacteria for the rough DP1004 and for the G54 strain and less for the virulent D39 and TIGR4 ( Figure 2 G ) . Essentially identical data where obtained when performing the experiment with splenic macrophages from C57BL/6 mice , while in contrast bone marrow macrophages from BALB/c mice and RAW264 . 7 macrophages showed different patterns of surface markers expression to the spleen macrophages and their phagocytosis of pneumococci showed no correlation to the extent of early clearance in the host ( Figure S4 ) . These data emphasise the importance in the choice of cell lines for performing phagocytosis assays in vitro to assess pneumococcal clearance in vivo . Pneumococci grown from the blood of 6 mice were subjected to whole genome sequencing ( Table 3 ) . In each case , the isolates had the identical antibiotic resistance phenotype and were therefore presumptively monoclonal . Two of the blood cultures were obtained from the same mouse , but at 24 or 48 h respectively , ( mouse 3 . 1 . 5; Table 3 ) . To identify possible mutations characterising the founding cell of the monoclonal blood culture , we searched for single nucleotide polymorphisms ( SNP ) present in all cells isolated from a given blood culture . Such a SNP would demonstrate that the re-expanded population arose from a single cell . We identified one or two SNPs in 100% of the bacterial populations from four out of six mice ( 3 . 1 . 5 , 4 . 1 . 4 , 4 . 2 . 2 and 4 . 2 . 6 ) , when compared to bacteria from the challenge inoculum ( Table 3 ) . The identification of a SNP common to all bacteria of a given sample is conclusive genetic evidence that the pneumococcal populations were monoclonal . Further , the SNPs were either inter-genic , silent or in regions not predicted to be functional in pathogenesis . Thus , we conclude that these mutations were unlikely to be associated with changes in within-host fitness . This argues strongly that bacteremia was founded as the result of a stochastic process rather than the selection of fitter variants . However , further analysis identified a second set of SNPs in pneumococci of 5 of 6 blood cultures . Crucially , this second set of SNPs were only found in a proportion of the bacteria obtained from mouse blood and therefore must have occurred after the bottleneck . Further , these SNPs differed between isolates of different mice , but all were located within distinct sub-units of the pneumococcal F1/F0 ATPase operon ( Table 3 ) . In three bloods , more than one SNP was detected . To determine if more than one SNP in the ATPase operon occurred in a single cell , we sequenced single colony isolates of these populations . In all cases , where a multi-SNP profile would have been possible according to the genomic data , only clones with a single SNP within the F1/F0 ATPase operon were recovered . These isolates included FP490 , a 3 . 2 . 4 derivative with a SNP in atpA , FP487 , a 3 . 1 . 5 derivative with a SNP in atpC , and FP489 and FP498 , two 4 . 1 . 6 derivatives with different SNPs in atpD ( Table 1 and 3 ) . In few cases subpopulations with mutations in other genes were detected ( pilus sortase , potassium uptake protein , metE , and SP0760 ) ( Table 3 ) , but no confirmation by direct sequencing was performed for these genomic data and we do not think that these mutations are of major biological relevance . Phenotypic analysis of eight independent ex vivo blood isolates each having a mutation ( SNP ) in the ATPase ( Table 1 and 3 ) , showed normal colony morphology on agar plates and no significant change in their susceptibility to optochin . In liquid culture , the mutants showed normal or more efficient growth in Todd Hewitt Yeast Extract ( Figure 4 A ) , but were unable to grow in other media ( Tryptic Soy Broth ) ( Figure 4 B ) . Given that the F1/F0 ATPase is involved in multiple aspects of proton trafficking , we investigated the impact of pH , buffer composition and salt concentration on bacterial growth . Using a phenotype microarray for osmotic susceptibility using Biolog microtiter plates PM9 , we compared the phenotype of parental strains derived from strain TIGR4 to the ex vivo mutants . The mutants had acquired a series of metabolic characteristics , also shared by strain D39 ( Figure S5 ) . Growth experiments performed in serial buffer and salt dilutions showed that TIGR4 and its isogenic derivatives used in the challenge experiments had a restricted pH optimum when compared to D39 , which limited growth at potassium phosphate concentrations below 10 mM and pH below 6 . 8 ( Figure 4 D ) . Interestingly many of the mutants had gained this capacity , making them equally able to grow at low pH as D39 . In contrast , high buffer concentrations ( 80 mM K2HPO4 and pH 8 ) , inhibited growth of all mutants ( Figure 4 C ) . To investigate effects on intracellular pH homeostasis of the ATPase mutations we transformed the frame-shift in the atpC gene into the non-encapsulated strain DP1004 . Using in vivo NMR , the atpC mutant and its parental strain were both shown to have an identical intracellular pH of 6 . 52 to 6 . 56 during active metabolism of glucose . No differences in susceptibility to neutrophil killing were observed when mutants were assayed in an opsonophagocytosis assay in the presence of type specific antibodies ( Figure 4 E ) . Also , data of macrophage phagocytosis were unaltered in primary cultures of splenic macrophages ( Figure 4 F ) . To check for any fitness cost in vivo , the encapsulated atpC mutant was compared to the challenge strain FP321 in our i . v . mouse sepsis model . At early time points both strains showed comparable blood counts ( data not shown ) . Also at 72 h post-challenge , bacterial counts in blood were similar , but bacterial spleen counts for the atpC mutant were significantly increased when compared to the wild-type ( Figure 4 G ) .
We have investigated the pathogenesis of pneumococcal bacteraemia following intravenous inoculation of mice with three isogenic clones ( variants ) . In our model , the infection followed the classic , three phase pattern in which a majority of pneumococci are cleared in the first minutes post-challenge . This leads to an “eclipse phase” of several hours in which bacterial numbers decline further or are undetectable . This is followed by the emergence of sustained and high density bacteraemia in a proportion of the challenged animals [8] , [10] . By analysing the survival in the blood of three isogenic variants of S . pneumoniae , we observed that the majority of blood cultures arose from only one of the three variants . We used a statistical model to characterise the infection dynamics in which the number of bacteria starting the infection in each invasion event is w and the number of times this happens is k [23] . From the model , we could infer that the number of bacteria at the origin of infection is below 2 ( w = 1 ) . Thus , it follows that bacteraemia was generated by either ( a ) a single bacterium establishing a population in the blood in a single invasion event or ( b ) several bacteria each of which independently established a population in distinct invasion events . The probability of ( b ) is small ( about 5% in our data , because the probability of two or more invasion events occurring is about 5% ) . Genome sequencing provided genetic evidence in 4/6 cases that monoclonal bacteraemia did actually start from a single bacterial cell ( w = 1 ) confirming the first statement . For the remaining 2/6 cases we could not determine w = 1 by genome sequencing as we could not distinguish several invasions of a single bacterial variant from one invasion of several cells of the same variant without any SNPs . More complex is the experimental observation of invasive events . For this we could document bacteraemia in mice with previous negative blood samples ( k≥1 ) and in other mice the increase of variants in serial blood samples ( k>1 ) . Since after the first 24 h the observed numbers of both these types of invasion events are similar , this strongly favours the occurrence of polyclonal infections resulting from independent , not cooperative action . In the case of H . influenzae it had been hypothesised that the single cells giving rise to the monoclonal infection might be selected by within-host evolution [23] . Our work now tests this hypothesis by whole genome sequencing . The data show in two cases absence of any SNPs and in four cases SNPs that apparently do not indicate selection for virulence . Despite the low numbers , it suggests that the single cells at the origin of infection apparently have no advantage ( higher virulence ) over the other cells in the population . Such results show that the bacteria in the challenge dose act independently to give rise to infection , that each has a similar probability of causing infection and that a dose near the LD50 , a single cell may initiate disease . These criteria satisfy the theory of independent action [16] , [17] , [22] , [23] . As such our investigation provides strong evidence that the single founding cell of an invasive infection is the result of a stochastic event . However , it must be emphasised that epigenetic variations would have eluded our genetic and genomic analysis . Previously published studies have shown a major role for splenic macrophages in the initial clearance of pneumococci . In the seminal investigations of Brown et al . [35] , they conclude: “… it appears that an anatomically normal spleen plays a unique role in the clearance of experimental pneumococcal bacteraemia , and that this role is of increasing importance as the pathogenicity of the invading organism increases” . Our data provide evidence that splenic macrophages have properties not found in those derived from other tissue sites , with respect to their efficiency to ingest and kill pneumococci . It is worth noting that the impressive efficiency of splenic clearance in vivo in the non-immune host is somewhat at odds with the relatively inefficient ingestion and killing of pneumococci in standard in vitro phagocytosis assays [34] . The innate host factors that result in the removal of the vast majority of bacteria within 45 minutes of challenge [11] , [35] , [36] deserve further attention . Despite the efficiency of splenic macrophages in clearance , sustained bacteraemia eventually occurs after an eclipse phase of several hours during which bacteria are largely undetectable in blood . Similar data were obtained in work based on intranasal inoculation of H . influenzae , where also mixed blood cultures were detected in the first minutes after challenge and before the eclipse period [25] . We propose for our intra venous injection model that during this time , a fraction of the inoculated bacteria are sequestered in extravascular tissues , most probably in the spleen , in accordance with our data on positivity of bacterial spleen counts also in mice with negative blood cultures ( Figure S1 ) . This emergence of a clone from the potential splenic focus into the “sterile” bloodstream can be viewed as equivalent to the “invasive events” described for models which consider more than one organ system [16] , [18] , [21] , [23]–[25] . Sustained bacteraemia is initiated from replication of one bacterial cell , perhaps a stochastic event in which the first replicon to reach a threshold biomass sufficient to seed the blood “wins the day” . The exponential increase in the number of bacteria in the blood is consistent with contributions from both intravascular and extravascular replication of pneumococci . We favour a scenario in which , at a challenge dose below the LD50 , the rate of replication occurring in the extra-vascular site , followed by seeding of bacteria to the blood , exceeds host clearance rates thereby resulting in progressively more severe bacteraemia . Our data do not infer that only one pneumoccocus survives the initial host clearance , but rather that from those that do survive; only one cell initiates bacteraemia . The observed increase of polyclonal infections over time , as predicted also by the independent action hypothesis , is in accordance with the doubling of the ratio of infected mice at those time points ( 0 . 31 at 24 h to 0 . 67 at 72 h ) [22] . The strong positive selection which drives the emergence of the ATPase-SNP subclones during the later phase of the infection is novel with respect to previous models ( i . e . the live-death model ) , which postulates a neutral selection during this phase [16] , [17] , [37] . The in depth genomic analysis , in contrast to previous works [16] , [17] , [37] , shows evidence for a more dynamic behaviour of the infecting bacterial population with an increase in heterogeneity of the monoclonal population over time due to a strong positive selection after the single cell bottleneck . However , we observed added complexity; the residual , but inadequate , innate clearance mechanisms exert a selective pressure resulting in the emergence of adaptive mutants . Sequencing of bacteria from blood revealed that in most mice the bacterial clones had each acquired SNPs in different sub-units of the pneumococcal F1/F0 ATPase gene . This apparently high frequency of mutations , given the relatively small biomass of pneumococci in each animal , is consistent with the estimated mutation rates of up to 5×10−4 per genome described recently for pneumococci during one-cell bottleneck in vitro passages [38] . The selection for altered function of the ATPase , was found only in a proportion of the bacteria making up the population obtained from blood , compelling evidence that the ATPase mutations must have occurred after the single cell bottleneck . As stated above , the observation of subclones being selected during the bacteremic phase underlines a highly dynamic situation , which extends over the neutral two stage infection models [16] , [17] , [37] . In pneumococci , it has been recognised that ATPase mutations occur at high frequency during pneumococcal infection in humans , possibly in response to oxidative stress [39] , and have been described both in vitro and in clinical isolates [40]–[44] . Polymorphisms in F0_atpA and F0_atpC ( the trans-membrane part of the ATPase ) were found to confer phenotypes of reduced susceptibility to optochin , quinine and mefloquine [40] , [42] , [43] . In particular , the detection of optochin resistant pneumococci in clinical samples is well described [44] , as it has a practical impact on pneumococcal identification in the diagnostic laboratory [41] . None of the ex vivo ATPase mutants in our investigation were optochin resistant and the SNPs accordingly did not map to the optochin resistance conferring regions . The F1/F0 ATPase is encoded by a highly conserved eight-gene operon and , as in aero-tolerant anaerobes , it is involved in the maintenance of intracellular pH through the generation of a membrane proton gradient [45] . In some of the mutants we were able to identify a clear metabolic benefit of the mutations which enabled growth at pH lower than 6 . 8 , albeit all mutants showed that loss of capacity to grow at pH above 7 . 8 . Interestingly the phenotype of TIGR4 mutants recovered from blood was not different from other virulent pneumococcal stains , such as D39 . The high frequency of mutation observed here , given by the many different sites mutated , strongly suggests within-host adaptation through selective pressure during sepsis . While in vitro susceptibility of the ATPase mutants to antibody mediated neutrophil killing and macrophage phagocytosis was essentially unaltered , the phenotypic consequence of the ATPase mutations may be linked to a gain in fitness related to the increased survival of bacteria within the splenic , extravascular focus that provides the source of pneumococci re-seeding the blood and sustaining the progressively escalating and ultimately lethal bacteraemia . In agreement with this hypothesis is the recent description of inhibition of the own F1F0 ATPase by both Salmonella enterica and Mycobacterium tuberculosis as strategy to withstand phagolysosomal activity [46] . In summary , we propose that after the majority of the bacteria of the challenge inoculum have been removed , a few bacteria survive the predominantly lethal activity of splenic macrophages and neutrophils . From these rare survivors , single pneumococcal cells may start to replicate and initiate seeding of the blood resulting in a steady state bacteraemia in which efficient host clearance is off-set by re-seeding from the original , persisting extravascular reservoir of bacteria . These extravascular bacteria are subjected to strong selection for adaptive mutations . Later during infection , selected subpopulations of the initial clone may become part of the bacterial population causing disease . These observations are in accordance with a two stage model of infection where independent action generating the initial stochastic event is followed by a dynamic birth-death phase which increases heterogeneity due to strong selection [16] , [17] , [24] . In the case of the model organism S . pneumoniae , our data show different selective pressures shaping the invasive bacterial population during different phases of infection [16] . Given the demonstration that pneumococci are independent in generating disease in our rodent model and that less than twenty per cent of human pneumonia cases are bacteraemic [3] , we hypothesize that human pneumococcal bacteraemia is generally monoclonal originating from a single cell in analogy to the monoclonal meningitis case recently described [47] . Presentation of a model which foresees development of invasive disease from a single bacterium and strong selection during outgrowth represents an important example on which to model fitness selection during invasive infection .
Three isogenic zmpC knock-out mutants of TIGR4 ( FP122 , FP318 and FP321 ) that differed only for the resistance marker , ermB ( erythromycin resistance ) , aad9 ( spectinomycin resistance ) and aphIII ( kanamycin resistance ) , respectively were constructed for co-infection studies with isogenic clones [28] , [48] . The experiments with mice depleted of macrophages and neutrophils were done with four different strains: TIGR4 ( serotype 4; strain FP321 zmpC::aphIII ) , G54 ( serotype 19F; erythromycin and tetracycline resistant ) [49] , [50] , D39 ( serotype 2; strain FP335 bglA::aad9; gift of Hasan Yesilkaya , Leicester ) , and the streptomycin resistant non-encapsulated D39 derivative DP1004 [51] , [52] . The transfer of the atpC frame-shift into DP1004 was performed by transformation of a marker flanked by two PCR fragments , one of which containing the frame-shift . This was possible since the atpC SNP is only 76 bp from the end of the operon . Two representative transformants FP499 and FP500 were confirmed by sequencing . The series of ATPase mutants isolated are described in Table 3 , while all other strains are listed in Table 1 . Strains were cultured in Tryptic Soy Broth ( TSB , Liophilchem , Teramo ) or Todd Hewitt ( THY , Oxoid , Milano ) supplemented with 0 . 5% Yeast Extract ( Liophilchem ) . Solid media were blood agar plates ( Tryptic soy agar , Difco ) supplemented with 3% horse blood ( Biotech , Grosseto ) . The colony morphology was checked on Todd-Hewitt agar plates containing 200 units/ml of catalase ( Sigma-Aldrich , Milano , Italy ) [53] , [54] . Antibiotics were used at the following concentrations: 1 mg/L erythromycin , 500 mg/L kanamycin , 100 mg/L spectinomycin and 500 mg/L streptomycin ( all from Sigma-Aldrich ) . The intracellular pH was determined by Nuclear magnetic resonance ( NMR ) . In brief , 400 ml of mid log pneumococcal cells grown in Todd Hewitt broth were pelleted and mixed with 1 ml of sodium alginate 6% ( w/v 0 . 9‰ NaCl ) . Mixture were extruded manually through 25G needle on a surface of 0 . 25 M CaCl2 solution . The small drops were washed and transferred in the 10 mm NMR tube . NMR 31P spectra were recorded on a Bruker DRX 600 instrument operating at 242 . 9 MHz . 31P spectra were recorded with a 1 . 5 s repetition time and 45°flip angle . Line broadening of 10 Hz were applied before Fourier Transform . 31P chemical shift were determined by comparison with external standard Trisodium trimetaphospate at −20 . 80 ppm . Intracellular pH was determinate by Pi ( intracellular phosphate ) chemical shift in phosphate-free perfusion model [55] . Active metabolism of pneumococci was confirmed by acidification of the extracellular medium during the experiment carried out at 37°C . Growth profiles of wild type strains and ATPase mutants were assayed both in standard laboratory media and in defined media . Standard laboratory media included TSB ( Liophilchem ) and Todd-Hewitt broth supplemented with Yeast Extract ( 0 . 5% ) ( Oxoid ) . Defined media were prepared in CAT medium by adding serial concentration of potassium phosphate buffer with different range of pH ( 6 to 8 ) and by adding several concentration of K2HPO4 as source of salt . CAT medium was composed by: Casitone ( 10gr/l ) ( Becton Dickinson ) , Tryptone ( 10 gr/l ) ( Oxoid ) , Yeast Extract ( 1 gr/l ) ( Liophilchem ) , NaCl ( 5 gr/l ) ( Panreac , Milano , Italy ) , Catalase ( 200 U/ml ) ( Sigma-Aldrich ) and Glucose ( 0 . 2% ) ( J . T . Baker , Milano , Italy ) . Metabolism of pneumococcal strains including wild type and ATPase mutants were assayed by Phenotype MicroArray ( PM ) microplate PM9 containing a total of 96 different osmolyte sources . PM technology measures active metabolism by recording the irreversible reduction of tetrazolium violet to formazan as an indirect evidence for NADH production . PM procedures were carried out as previously described ( Viti C 2009 ) . Quantitative colour change were recorded automatically every 15 min for a period of 72 h . Analyses were performed by the Omnilog-PM Software ( Biolog , inc . ) and data were filtered using average height as a parameter . Animal experimentation in Italy is regulated by Decreto Legislativo 116/92 and Directive 210/63/EU . The animal protocol was approved by the “Comitato Etico Locale” of the Azienda Universitaria Ospedaliera Senese and received thereafter the relative project licence issued by the Italian Ministry of Health ( 193/2008-B ) . Six to seven-weeks old female CD1 , BALB/c , and CBA/Ca mice were purchased from Charles River Italia ( Lecco , Italy ) . For the bottleneck experiments , outbred CD1 mice were used , while BALB/c mice were chosen for both in vivo macrophage and neutrophil depletion and ex vivo experiments . CBA/Ca data are shown only for comparison of the dynamics of the early phases of infection . Animals were sacrificed by intraperitoneal ( i . p . ) injection of xylazine hydrochloride and zolazepam tiletamine cocktail ( Xilor 2% , Bio 98 S . r . l . , Bologna , Italy and Zoletil 20 , Virbac S . r . l . , Milano , Italy ) . Mice were kept at the animal facility of the LAMMB , University of Siena , according to its guidelines for the maintenance of laboratory animals [48] , [56]–[58] . Blood samples from mice were collected by sub-mandibular vein or cardiac puncture under terminal anaesthesia . To prevent blood coagulation , 100 U/ml of heparin ( MS Pharma , Milano , Italy ) was added . All the collected organs ( spleen , lung , liver and kidney ) were homogenized in 1 ml of TSB , and then frozen at −80°C after making to 10% v/v of glycerol . Two series of experiments were performed in order to define the bottleneck for invasive pneumococcal infection with a total of 68 CD1 mice . Mice were challenged intravenously ( i . v . ) as described [48] , [56]–[58] with a mixture of the three isogenic TIGR4 derivatives ( FP122 , FP318 and FP321 ) at 3 . 3×105 CFU each/mouse . At pre-set time points blood samples were collected and selected groups were sacrificed for obtainment of spleen samples . Two blood samples from each animal , taken at different time points , are reported in Figure 1 . Bacteria were enumerated by plating on selective media . The dose of the experiment was decided after having observed in a preliminary experiment 5/8 mixed and 3/8 monoclonal infections using two pneumococcal clones at a dose of 2×10∧6 ( data not shown ) . A pilot experiment for comparison of virulence in CD1 , BALB/c and CBA/Ca mice was carried out by infecting i . v . four mice each with a mixture of G54 , D39 and TIGR4 ( 3×105 CFU each/mouse ) . Three blood samples per mouse were obtained . For depletion of macrophages , BALB/c mice were treated 24 h prior to challenge by i . p . injection with 750 µl of a suspension of clodronate ( CL2MBP ) liposomes . One control group received PBS-containing liposomes [31] and the other was untreated . Clodronate was encapsulated in liposomes , as described earlier [31] and was a gift of Roche Diagnostics ( Mannheim , Germany ) . Neutrophil depletion was performed by a single i . p . injection of 150 µg/mouse of anti-GR-1 antibody ( Ly6G and Ly6C , clone RB8-8C5; Becton Dickinson ) 24 h prior to infection [32] , [33] . Two control groups were either left untreated or administered with a rat isotype control antibody IgG2b K ( kappa ) ( Becton Dickinson ) . Groups of mice depleted of macrophages or neutrophils were infected i . v . with 1×106 CFU/mouse containing 2 . 5×105 CFU of each TIGR4 , D39 , DP1004 and G54 . Bacterial viable counts were determined at preset time points . The virulence of the ATPase mutant FP487 ( atpC mutant ) was assayed in parallel with TIGR4 . BALB/c mice ( n = 6 ) were infected with 1×106 CFU/mouse i . v . and blood and spleen samples collected at 72 h . Spleen and bone marrow macrophages were isolated from mice using a modified protocol previously described [34] . Cells were cultured in medium supplemented with 25 ng/ml of recombinant M-CSF ( Invitrogen ) and re-seeded at day 7 at the concentration of 2×105 cell/ml . After 24 h , 0 . 1 ml of pneumococci cultured to OD590 0 . 25 were added . After 45 min plates were washed and reincubated with 10 mg/L of penicillin and 200 mg/L of gentamicin for 30 min . Intracellular bacteria were enumerated after lysis with saponin 1% . Phagocytosis of RAW264 . 7 macrophages followed the same protocol , but in addition samples were reincubated after removal of the antibiotics for an additional hour in fresh medium . Flow cytometric analysis was conducted on bacteria suspended in 1% v/v paraformaldehyde in PBS on a FACScalibur machine ( Becton Dickinson , California , USA ) . To verify macrophage and neutrophil depletion , homogenised organ samples were washed in DMEM ( Sigma-Aldrich ) and non-specific binding was blocked with FcR blocking agent [59] . Cells were incubated 30 min with 1 µg of specific fluorochrome-conjugated antibodies per 106 cells . Neutrophils were stained using a rat anti-GR-1 antibody ( MACS , Bologna , Italy ) . Macrophages were detected with rat anti-F4/80 mAb ( BM8 clone; Abcam , Milano , Italy ) , and a rat anti-mouse CD11b mAb ( Becton-Dickinson ) . Surface markers of macrophages were analysed using the following antibodies: anti-F4/80 mAb , anti-mouse CD11b mAb , anti-CD11c mAb ( eBioscience ) , anti-mouse SIGNR1/CD209b Ab , goat IgG control Ab , anti-mouse Siglec-1 mAb , rat IgG2A Isotype control Ab , anti-mouse MARCO mAb and rat IgG1 isotype control Ab ( R&D Systems ) . To assay for the presence of anti-pneumococcal antibodies in mouse sera , the four pneumococcal strains TIGR4 , G54 , D39 and DP1004 were blocked in PBS-BSA 1% v/v for 30 min at 37°C and incubated for 1 h at 37°C with sera ( 1∶100 ) obtained from BALB/c mice and the positive anti-serotype 2 control serum ( Staten Serum Institute , SSI , Copenhagen , DK ) . Samples were marked with anti-mouse IgG ( 1∶64 ) or anti-rabbit IgG ( 1∶160; Sigma-Aldrich ) . In order to evaluate the capacity of whole blood to kill or inhibit the multiplication of pneumococci and to investigate the effect of specific antibodies , ex vivo experiments were set up . Blood from BALB/c mice was collected into tubes containing heparin and infected with pneumococci . For the assay of opsono-phagocytosis of ATPase mutants 1×104 CFU/ml of parental and ATPase mutant were inoculated in blood and incubated in rotation . The anti-type 4 serum ( SSI ) was used at 1∶50 dilution . For the evaluation of growth of pneumococci in blood 3×105 CFU/ml of G54 , D39 and TIGR4 were inoculated in rotating blood . The efficacy of type 2 anti-serum ( SSI ) on D39 and its non-encapsulated derivative DP1004 was assayed as above using a inoculum of 3×105 CFU/ml and a 1∶100 dilution of the serum . Chromosomal DNA was extracted using the High Pure PCR Template preparation kit ( Roche ) . Whole genome sequencing was performed by the Institute of Applied Genomics and IGA Technology Services srl ( University of Udine , Italy ) using an Illumina ( Solexa ) Genome Analyzer II platform [60] . Reads of both , parent and mutant strains , were aligned to the reference genome of TIGR4 ( accession NC_003028 ) using the Mosaik Assembler suite ( The MarthLab , Boston College , Massachusetts , USA ) . Single nucleotide polymorphisms ( SNPs ) , insertions and deletions ( INDELs ) were retrieved with VarScan software [61] . SNPs and INDELs of the challenge strains were subtracted from those found by aligning the blood isolates . All F1/F0 ATPase mutations were re-sequenced by the Sanger method and deposited in GenBank ( accession KF705516 to KF705525 ) . In order to evaluate the number of bacteria at the origin of blood infection , a model derived from that previously described [23] , was developed . A full description of the statistical model is given in the supplementary materials . Statistical analysis of bacterial counts in blood and organs was performed by the Student's t-test for data reported in Figures 3 , 4 , S2 and S3 . The analysis of different bacterial blood clearance at 5 and 30 min and the differences in bacterial phagocytosis and data of phenotype microarray were performed using Kruskal-Wallis and Dunn's multiple comparison post test ( Figure 2 C–F and S5 ) . Values of P<0 . 05 were considered statistically significant . The Fluorescence Index ( Figure S3 K ) was calculated by multiplying the percentage of positive events with the geometric mean fluorescence intensity ( GeoMean ) . | Decades of research on bacterial sepsis have been devoted to analysing the steps that lead from a local event , either carriage or a localised infection , to systemic disease . Our work analyses in depth the events determining systemic infection by one of the main human pathogens , Streptococcus pneumoniae . Consistent with similar findings on the pathogenesis of bacteraemia due to other commensal pathogens , our results show that after an intravenous inoculum of a million pneumococci , the resulting septicaemia is often founded by a single bacterial cell . Investigation into the nature of this monoclonal infection identified strong within-host selective pressure for metabolic fitness during outgrowth of the bacterial population . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"bacterial",
"diseases",
"microbial",
"metabolism",
"infectious",
"diseases",
"microbial",
"mutation",
"streptococci",
"microbial",
"evolution",
"microbial",
"pathogens",
"host-pathogen",
"interaction",
"biology",
"microbiology",
"infectious",
"disease",
"modeling",
"bacterial",
"pathogens",
"gram",
"positive"
] | 2014 | The Role of Host and Microbial Factors in the Pathogenesis of Pneumococcal Bacteraemia Arising from a Single Bacterial Cell Bottleneck |
Plant disease resistance is often mediated by nucleotide binding-leucine rich repeat ( NLR ) proteins which remain auto-inhibited until recognition of specific pathogen-derived molecules causes their activation , triggering a rapid , localized cell death called a hypersensitive response ( HR ) . Three domains are recognized in one of the major classes of NLR proteins: a coiled-coil ( CC ) , a nucleotide binding ( NB-ARC ) and a leucine rich repeat ( LRR ) domains . The maize NLR gene Rp1-D21 derives from an intergenic recombination event between two NLR genes , Rp1-D and Rp1-dp2 and confers an autoactive HR . We report systematic structural and functional analyses of Rp1 proteins in maize and N . benthamiana to characterize the molecular mechanism of NLR activation/auto-inhibition . We derive a model comprising the following three main features: Rp1 proteins appear to self-associate to become competent for activity . The CC domain is signaling-competent and is sufficient to induce HR . This can be suppressed by the NB-ARC domain through direct interaction . In autoactive proteins , the interaction of the LRR domain with the NB-ARC domain causes de-repression and thus disrupts the inhibition of HR . Further , we identify specific amino acids and combinations thereof that are important for the auto-inhibition/activity of Rp1 proteins . We also provide evidence for the function of MHD2 , a previously uncharacterized , though widely conserved NLR motif . This work reports several novel insights into the precise structural requirement for NLR function and informs efforts towards utilizing these proteins for engineering disease resistance .
One of the main plant disease resistance mechanisms is mediated by dominant resistance ( R ) genes with a major effect [1] . Many R genes have been cloned from diverse plant species , most of which encode nucleotide binding leucine-rich-repeat ( NB-LRR or NLR ) proteins [2 , 3] . Based on the secondary structure of their N-termini , NLRs can largely be subdivided into two classes: one containing a Toll-interleukin 1 receptor ( TIR ) domain ( TIR-NB-LRR , TNL hereafter ) , and the other a putative coiled-coil ( CC ) domain ( CC-NB-LRR , CNL hereafter ) . Both TNL and CNL proteins have been identified in dicots , while only CNL proteins have been found in monocots [4] . Each NLR is capable of directly or indirectly recognizing the presence of at least one effector protein , usually produced by a subset of isolates of a single pathogen species [5] . Recognition generally leads to activation of the NLR protein and initiation of signal transduction , resulting in the hypersensitive response ( HR ) , which is often accompanied by the induction of a rapid localized cell death at the point of pathogen penetration [6] . HR can contribute to the halting of pathogen growth . When the corresponding effector is not present , wild-type NLRs are held in an inactive state , largely through self-inhibitory intra-molecular interactions , which may also be facilitated or reinforced by additional host proteins [5 , 7 , 8] . In general , plant NLR proteins contain three separable structural domains , an N-terminal CC or TIR domain , a C-terminal LRR domain and a central NB-ARC ( ARC: APAF1 , R gene products and CED-4 ) domain . The ARC subdomain in plant NLRs is further divided into two separated structural units: ARC1 and ARC2 [9 , 10] . In the NB-ARC domain , several conserved motifs are recognized; in linear order: P-loop/Kinase-1a/Walker A ( henceforce called the P-loop ) , RNBS ( Resistance Nucleotide Binding Site ) -A , Kinase-2/Walker B , RNBS-B , RNBS-C , GLPL , RNBS-D , MHD [11] . Among the different motifs , the P-loop and MHD motifs are known to be very important for NLR function . In their inactive form , NLRs likely have ADP bound to the NB-ARC domain . Activation of NLRs involves the exchange of ADP for ATP through a structural change that is thought to result in an open structure of the proteins [12 , 13] . The P-loop motif within the NB-ARC domain is thought to be involved in binding ATP/ADP [13] . P-loop mutations in the NLR resistance genes I-2 from tomato , M from flax and RPM1 from Arabidopsis impair ATP binding and ability to confer HR-based resistance [14 , 15 , 16] . Mutations in the conserved MHD-motif lead to autoactivation of NLRs , which is thought to be due to weakened ADP-binding and resulting structural change into an open conformation , favoring nucleotide exchange [15 , 17] . Autoactivate NLRs , i . e . NLRs that can be activated without the need for a recognition event , can be generated in vivo or experimentally via recombination between different NLRs , or through specific point mutations . One example of an in vivo recombination event leading to an autoactive NLR is from the maize Rp1 locus . Rp1 is a complex locus that carries multiple NLR paralogs . The number of these paralogs varies widely between haplotypes , some carrying more than 50 [18] . The Rp1-D haplotype consists of 9 NLR paralogs; Rp1-dp1 to Rp1-dp8 which have no known function , and Rp1-D itself , which confers resistance against the biotrophic fungal pathogen Puccinia sorghi , the causal agent of maize common rust [19 , 20] . These Rp1 paralogs are more than 90% identical in nucleotide sequence , allowing them to undergo unequal crossing over and occasional intragenic recombination . An intragenic recombination between paralogs Rp1-D and Rp1-dp2 produced the chimeric gene Rp1-D21 , which was identified on the basis of its ‘lesion mimic’ phenotype in the absence of pathogen infection [20 , 21] . We previously demonstrated that many of the hallmarks of pathogen-induced immune response , such as H2O2 accumulation , increased expression of the defense-related genes PR1 , PRms and WIP1 , are associated with Rp1-D21-mediated lesion phenotype in maize . We also demonstrated that the Rp1-D21 lesion phenotype is genetic background- , temperature- and light-dependent [22 , 23] . Recently , we deployed Rp1-D21 as a tool in a novel enhancer/suppressor screen to identify natural modifiers of the HR [23 , 24 , 25 , 26] . Different NLR R-genes appear to have a variety of different structural requirements for proper activation and functioning [27 , 28 , 29 , 30 , 31 , 32] . The autoactive nature of Rp1-D21 and the fact that its ‘parental’ proteins , Rp1-D and Rp1-dp2 , are not autoactive makes it a very useful tool to explore the molecular requirements for NLR regulation , specifically the switch between inactive and active states . Here we report the characterization of the activity of Rp1 proteins in maize and in Nicotiana benthamiana , a model system widely used for characterization of R-gene mediated HR for both dicot and monocot NLRs [27 , 29 , 33 , 34] . Using a combination of genetic , molecular biological , biochemical and computational techniques we derive a model for Rp1 activity which provides a better understanding of how NLRs regulate the switch between the resting and the activation states .
Rp1-D21 is derived from the recombination of two NLR paralogs at the Rp1 locus , Rp1-D , and Rp1-dp2 that are 90% identical at the amino acid level [19 , 20 , 21 , 35] . Sequencing of Rp1-D21 showed that it encodes a protein of 1290 amino acids ( AAs ) and is a typical CNL with an N-terminal coiled-coil ( CC ) domain ( AAs 1–189 ) , an NB-ARC domain ( AAs 190–527 ) and an LRR domain at the C-terminus ( AAs 528–1290 ) . The N-terminus of Rp1-D21 derives from Rp1-dp2 ( up to AA 770–778 ) , with the remainder deriving from Rp1-D ( the precise breakpoint is impossible to define since nucleotides 2310–2333 , corresponding to AAs 771–777 are identical in the two progenitor genes , Fig . 1 ) . Thus , Rp1-D21 is comprised of the CC , NB-ARC and the N-terminus of the LRR domain from Rp1-dp2 and the C-terminus of the LRR domain from Rp1-D . A conserved EDVID motif ( EDLLD in Rp1-D21 , Fig . 1 ) can be identified in the CC domain of Rp1-D21 . All of the important motifs present in the NB-ARC domain of typical NLR proteins can be found in Rp1 proteins ( Fig . 1; [11] ) . We performed a targeted ethyl methanesulfonate ( EMS ) mutagenesis screen in a maize line harboring Rp1-D21 to identify mutations that lost the autoactive HR phenotype conferred by Rp1-D21 . Putative suppressor mutants were easily identified due to their robust growth compared to the stunted , lesion-mimic siblings heterozygous for Rp1-D21 ( Fig . 2A ) . From about 23 , 000 EMS-mutated M1 plants , 12 missense and 2 nonsense intragenic mutants were identified ( Fig . 2B; Table 1 ) . Among the 12 missense mutants , five had mutations in the CC domain , three in the NB-ARC domain and four in the LRR domain ( Fig . 2B ) . Notably , one mutant ( nucleotide G244A , thus D82N in amino acid ) was in the conserved EDVID motif of the CC domain , and one ( C1193T , P398L ) in the conserved GLPL motif of the NB-ARC domain ( Table 1; Figs . 1 and 2B ) . The mutants T260I and E312K were adjacent to the conserved RNBS-A and Walker B motifs , respectively ( Table 1; Figs . 1 and 2B ) . We used Agrobacterium-mediated transient expression in N . benthamiana to investigate the structure/function of Rp1-D21 . Rp1-D , Rp1-dp2 and Rp1-D21 were either fused to the N-terminus of enhanced green fluorescent protein ( EGFP ) , a 3×HA ( influenza hemagglutinin ) tag or not tagged for subsequent functional analysis . When the three fusion proteins were transiently expressed using the cauliflower mosaic virus 35S promoter , a HR phenotype was observed 3 days post-infiltration ( dpi ) only with Rp1-D21 , but not with Rp1-D , Rp1-dp2 or the empty vector ( EV ) control . The same phenotypes were obtained regardless of the tags used ( EGFP , 3×HA or no tag ) and regardless of the differing levels of protein accumulation observed for the three proteins ( Fig . 3; S1A Fig . ) . To further confirm that the phenotype conferred by Rp1-D21 in our N . benthamiana system conformed to the phenotypes observed in maize , we constructed Rp1-D21 expression vectors carrying the missense Rp1-D21 suppressor mutations identified in maize . When transiently expressed in N . benthamiana , Rp1-D21 produced a strong HR phenotype , rating 4 based on a 0–5 scale rating with 0 being no cell death or chlorosis at all and 5 being confluent cell death [36] . The HR rating is based on the average results from at least 10 individual leaves . All the mutants we identified as non-functional in maize , except for H59Y and G850D , abolished or greatly reduced the level of Rp1-D21-induced HR ( Table 1; S2A Fig . ) . All constructs conferred high levels of protein expression ( S2B Fig . ) . The concordance between the maize and N . benthamiana systems is very good and we expect that our conclusions based on the transient expression N . benthamiana system are generally relevant to the endogenous maize system . Thus , we used the transient expression of Rp1-D21 and its derivatives in N . benthamiana to further analyze the molecular mechanism of activation/auto-inhibition of Rp1 proteins . The MHD motif is a highly conserved region of the NB-ARC domain involved in nucleotide binding [15] . Mutations in the MHD domain confer autoactivity to a number of NLRs . Two MHD motifs , here termed MHD1 and MHD2 ( actually LHD in Rp1-D and Rp1-dp2; Fig . 4A ) , separated by a single amino acid are apparent in Rp1-D , Rp1-dp2 and Rp1-D21 , and in a number of other CNLs [37] . To investigate whether mutations in either of the MHD motifs could cause autoactivity in the non-autoactive parental proteins , Rp1-D and Rp1-dp2 , we generated the following mutations: Rp1-D ( H517A ) , Rp1-D ( D518V ) , Rp1-dp2 ( D513V ) and Rp1-dp2 ( H512A/D513V ) in the MHD1 motif , and Rp1-D ( H521A ) , Rp1-D ( D522V ) , Rp1-dp2 ( H516A ) , Rp1-dp2 ( D517V ) in the MHD2 motif . Transient expression of the Rp1-dp2 ( D517V ) MHD2 mutant conferred an obvious HR , with symptoms appearing at 3 . 5 days , about 12 h later than that observed with Rp1-D21 ( Fig . 4B ) . HR was not induced by transient expression of any of the other MHD mutants , including the Rp1-D ( D522V ) MHD2 mutation ( S3 Fig . ) . However , Rp1-D ( D522V ) expression was not detectable by western blot analysis using anti-HA antibody . Protein expressed from all the other constructs could be detected ( Fig . 4B; S3 Fig . ) . We also generated the MHD2 mutations V1 ( D517V ) and V16 ( D522V ) in V1 and V16 , two recombinant constructs that were not autoactive ( see below ) and found both of them induced HR ( Fig . 4C ) . The P-loop is a highly conserved motif in NLRs and mutations in this domain often result in loss of function and loss of autoactivity [14 , 15 , 27 , 34 , 38 , 39] . P-loop mutations ( K225R ) introduced into Rp1-D21 and the autoactive Rp1-dp2 ( D517V ) mutant abrogated their ability to confer HR ( Fig . 4D ) . This indicated that the Rp1-D21 autoactivity and presumably also the activity of the Rp1 proteins are P-loop dependent . To examine which region of Rp1-D21 was required for triggering HR , different domains or domain combinations of Rp1-D21 , including CC , CC-NB-ARC , NB-ARC , NB-ARC-LRR and LRR domains ( hereafter CCD21 , CC-NB-ARCD21 , NB-ARCD21 , NB-ARC-LRRD21 and LRRD21; Fig . 5A ) , were fused to the N-terminus of EGFP . CCD21 and NB-ARCD21 were derived from ( and are therefore identical to ) the corresponding domains of Rp1-dp2 while LRRD21 is recombinant between the LRRs of Rp1-D and Rp1-dp2 . The transient expression of CCD21 , but of no other Rp1-D21 domains or domain combinations , conferred HR ( Fig . 5B ) . CCD21 , NB-ARCD21 and CC-NB-ARCD21 produced higher protein accumulation than NB-ARC-LRRD21 and LRRD21 ( Fig . 5B ) , excluding the possibility that lack of HR phenotype seen with most of the domain constructs was solely due to low protein accumulation . We also tested the HR phenotype conferred by different Rp1-D domains and found that the CC domain , but no others , induced HR when fused with EGFP ( Fig . 5A ) . Surprisingly , no untagged or HA-tagged domains from either Rp1-D or Rp1-D21 induced HR ( Fig . 5; S1B Fig . ) . CCD21 , but not CC-NB-ARCD21 , induced HR , which suggested that NB-ARCD21 can inhibit CCD21-induced HR in cis ( when the two domains were fused in the same molecule ) . Consistent with this result , no HR was observed when EGFP-tagged CCD21 and NB-ARCD21 were transiently co-expressed in trans ( when the two domains were co-expressed as separate molecules ) in N . benthamiana ( Table 2 ) . LRRD21 did not restore HR when co-expressed with CCD21 and NB-ARCD21 separately , or with CC-NB-ARCD21 in trans . However , co-expressing CCD21 with NB-ARC-LRRD21 retained HR in trans , suggesting that LRRD21 might interact with NB-ARCD21 to regulate CCD21–induced HR in cis but not in trans ( Table 2 ) . However , since LRRD21 or NB-ARC-LRRD21 are expressed at substantially lower levels than CCD21 or CC-NB-ARCD21 ( Figs . 5B and 5C ) , we cannot exclude the possibility that the concentration of LRRD21 or NB-ARC-LRRD21 is simply too low to affect the NB-ARC suppression of CCD21–induced HR or CCD21–induced HR itself when co-expressed in trans . In contrast to the in cis result of CC-NB-ARCD , we found that NB-ARCD did not suppress CCD-induced HR in trans ( Table 2 ) , suggesting NB-ARCD might interact with CCD in cis but not in trans . To further investigate the inhibition region of NB-ARC in CC-induced HR , we generated a series of deletion constructs from CC-NB-ARC ( S4 Fig . ) . We found that extension including AAs 190–260 from the NB domain was sufficient to suppress CC-induced HR . Consistent with the data , the nonsense EMS mutant ( Q346* ) suppressed the Rp1-D21 lesion-mimic and stunted growth phenotype in maize ( Fig . 2B; Table 1 ) . The C-terminus of the LRR domain has been demonstrated to be important for NLR activity [37] . We therefore investigated whether C-terminal deletions of the LRR domain from Rp1-D21 affected the HR phenotype . None of the six C-terminal deletion constructs ( ranging from 34 to 639 AAs ) induced HR after transient expression , even D21-LRR27 that lacked only the last 34 AAs of the acidic tail in C-terminus ( S5 Fig . ) . This is despite the fact that all the constructs conferred high levels of protein expression ( S5B Fig . ) . To delineate the functional regions that cause Rp1-D21 to be autoactive and keep its progenitor proteins auto-inhibited , we generated a series of chimeric constructs recombinant at different positions between Rp1-D and Rp1-dp2 ( Fig . 6 ) . These proteins , both with and without a 3×HA C-terminal tag , were tested in the N . benthamiana transient expression system . Previously , the chimeric construct Rp1-dp2-D2 , with the N-terminus deriving from Rp1-dp2 and the C-terminus from Rp1-D , with a recombination point at 980 AA ( Fig . 6A ) , had been shown to cause HR in transgenic maize [21] . The identical chimeric construct , hereafter named Hd2 , induced a strong HR when transiently expressed in N . benthamiana ( Fig . 6A ) , consistent with the transgenic maize result . The results of the transient expression of a series of chimeric proteins are summarized in Fig . 6 . The HR phenotype conferred by the constructs fused with 3×HA was largely similar to that conferred by the constructs without any tag ( Fig . 6A ) . All the tagged proteins accumulated to high and broadly comparable levels ( Figs . 7 and 8; S6 Fig . ) . Constructs V1 through V7 were generated with recombination points between Rp1-dp2 and Rp1-D ranging from AAs 1200 to 488 ( Fig . 6A ) . V1 with a recombination point at AA 1200 did not confer HR while V2 with a recombination point just 30 additional AAs N-terminal to AA 1170 produced a strong HR , indicating that AAs 1170–1200 from Rp1-dp2 are important for auto-inhibition of V1 . V4 and V5 ( recombination point at 690 and 651 , respectively ) also produced strong HR phenotype , while V6 and V7 ( recombination point 575 , 488 respectively ) conferred very weak or no HR ( rating 1 . 5 and 0 , respectively , Fig . 6A ) . These results suggest that the AAs 488–651 from Rp1-dp2 are important for the autoactivity of Rp1-D21 . Together these data suggest that proteins with a combination of AAs 488–651 from Rp1-dp2 and AAs 1170–1292 from Rp1-D replacing the corresponding sequences of either parental protein are unable to maintain a self-inhibited state . To test this hypothesis , we investigated whether swapping the N-terminal region ( CC , NB-ARC or AAs 1–488 ) in Rp1-D21 affected its autoactivity . We generated 8 more chimeric constructs in which we exchanged the Rp1-dp2-derived N-terminal parts of Rp1-D21 with the corresponding parts from Rp1-D ( Fig . 6B ) . Construct V8 , in which the CC-NB-ARC ( AAs 1–527 ) of Rp1-dp2 was replaced with the corresponding Rp1-D sequence , did not confer an HR . Similar results were obtained by exchanging the NB-ARC domain in V9 . However , V10 , exchanging the CC domain of Rp1-D into Rp1-D21 conferred a strong HR ( Fig . 6B ) . These results suggest that the ‘parental origin’ of the CC domain ( AAs 1–189 ) in Rp1-D21 plays only a minor or no role in its autoactive phenotype . V11 and V12 were generated by replacing AAs 190–370 or 190–458 ( NB-ARC1 plus part of ARC2 , respectively ) of V10 with the corresponding Rp1-D sequence . They differed from V10 by only 2 and 5 single amino acid polymorphisms ( SAAPs ) , respectively . V11 conferred a slightly weaker HR than V10 and the HR conferred by V12 was weaker still ( Fig . 6B ) . Construct V13 , of which the CC-NB-ARC domain was ‘reciprocal’ of V12 , did not confer HR , confirming that AAs 458–527 ( ARC2 region ) from Rp1-dp2 were important for the autoactivity of Rp1-D21 ( comparing V13 , V12 and V8 ) . Constructs V14 and V15 were generated by exchanging AAs regions 690–775 and 651–775 respectively of V10 with the corresponding Rp1-D sequence . Both constructs showed strong HR ( Fig . 6B ) , similar with the corresponding AAs exchanges in Rp1-D21 ( constructs V4 and V5 in Fig . 6A ) . V16 and V17 were ‘reciprocals’ of Rp1-D21 and V3 , respectively and did not confer an HR phenotype ( Fig . 6C ) . In summary , this set of experiments indicated that the combination of AAs 458–651 from Rp1-dp2 and AAs 1170–1200 from Rp1-D is central to the deficiency of self-inhibition of Rp1-D21 . There are only 17 single amino acid polymorphisms ( SAAPs ) between AAs 1170–1292 of Rp1-D and Rp1-dp2 , and only 5 SAAPs in the AAs 1170–1200 region that differentiate constructs V1 and V2 ( Figs . 6A and 7A ) . However , V2 was autoactive , while V1 was not ( Fig . 6 ) . To analyze which SAAP is important for the self-inhibition of Rp1 proteins , we generated several mutants based on these 5 SAAPs in the autoinhibited construct V1 , including V1 ( V1181A ) , V1 ( K1184N ) , V1 ( V1181A/K1184N ) and V1 ( V1186T/F1188L/C1189D ) . When transiently expressed in N . benthamiana , V1 ( V1181A ) was not autoactive , while V1 ( K1184N ) and V1 ( V1181A/K1184N ) produced a strong HR ( rating 4 ) , and V1 ( V1186T/F1188L/C1189D ) a weak HR ( rating 2; Fig . 7A ) . The results indicated that residue K1184 from Rp1-dp2 plays an important role in self-inhibition , and the residues V1186/F1188/C1189 play a minor role . Interestingly , one of the Rp1-D21 suppressor EMS mutants had a mutation at residue 1180 ( P to S ) , only 2 AAs away from N1182 in Rp1-D21 ( equivalent to N1184 in Rp1-D ) , emphasizing the importance of this region for regulating the activity of the Rp1 proteins . To investigate whether mutating K1184 was sufficient to convert Rp1-dp2 into an autoactive NLR , we generated Rp1-dp2 ( K1184N ) . However , this single point mutation was not able to ‘activate’ Rp1-dp2 ( Fig . 7B ) . This result suggests that additional SAAPs in the region of AAs 1200–1292 are also important for self-inhibition of Rp1-dp2 . There are 12 additional SAAPs in the last 92 AAs between Rp1-D and Rp1-dp2 and of these six are found in the last 16 AAs ( Fig . 7B ) . Therefore , we constructed the chimeric protein V18 by replacing the C-terminal 16 AAs in V1 ( K1184N ) , the autoactive V1 protein , with the equivalent AAs from Rp1-dp2 . Interestingly , this chimeric protein V18 lost the ability to induce HR ( Fig . 7B ) , demonstrating that the very C-terminal 16 AAs are very important for the self-inhibition of the Rp1 proteins . In conclusion , these experiments further defined the combination of AAs required for the autoactivity of Rp1-D21: a combination of AAs 458–651 from Rp1-dp2 and AAs 1184–1292 ( especially N1184 and the last 16 AAs ) from Rp1-D are central to the autoactivity ( see model in Fig . 7C ) . As shown in Fig . 6B ( construct V13 ) , AAs 458–527 from Rp1-dp2 were very important for the autoactivity of Rp1-D21 . There are two major regions of polymorphism , which we termed Patch 1 and 2 , between Rp1-D and Rp1-D21 within the ARC2 sub-domain , from AAs 458–527 ( Figs . 1 and 8 ) . Homology modeling based on the crystal structure of the NB-ARC domain of the ADP-bound human apoptosis regulator APAF1 ( Protein data bank entry 1z6t; [40] ) predicted that Patch 1 and 2 were surface-exposed ( Figs . 8A and 8B ) . Surface electropotential predictions indicated that the surfaces of both patches in Rp1-D were mostly negatively charged , while for Rp1-D21 Patch 1 was slightly positively charged and Patch 2 was mostly positively charged ( Fig . 8C ) . To investigate whether Patch 1 and 2 played roles in the autoactivity of Rp1-D21 , we constructed another two chimeric proteins , V19 and V20 , by replacing AAs 458–463 ( Patch 1 ) and 488–527 ( Patch 2 ) in Rp1-D21 with the corresponding AAs from Rp1-D , respectively ( Fig . 8D ) . Both V19 and V20 conferred a greatly reduced HR compared to Rp1-D21 , while V13 ( containing Patch 1+2 from Rp1-D ) did not cause HR ( Fig . 8D ) , suggesting that these patches have an additive effect on the autoactivity of Rp1-D21 . Self-association has been observed in both CNL ( eg . barley MLA , Arabidopsis RPS5 ) and TNL ( eg . tobacco N , flax L6 ) proteins and is important for the activity of NLRs [38 , 41 , 42 , 43] . To test whether Rp1-D21 , Rp1-D or Rp1-dp2 could self-associate , we transiently co-expressed EGFP- and 3×HA-tagged proteins in N . benthamiana and performed co-immunoprecipitation ( Co-IP ) analyses . We observed that full-length Rp1-D21 , Rp1-D and Rp1-dp2 all self-associated ( Fig . 9A ) . We further showed that all three of the domains ( CCD21 , NB-ARCD21 and LRRD21 ) of Rp1-D21 , were able to self-associate ( Fig . 9B ) . To further investigate whether Rp1-D or Rp1-dp2 could form heteromers with Rp1-D21 , we transiently co-expressed Rp1-D21:EGFP and 3×HA-tagged Rp1-D or Rp1-dp2 proteins in N . benthamiana and performed Co-IP analyses . We found that Rp1-dp2 interacted with Rp1-D21 ( Fig . 9C ) . We did not detect interaction between Rp1-D and Rp1-D21 , which might have been due to the low expression of Rp1-D ( Fig . 3B ) . Consistent with the interaction data , we observed that Rp1-dp2 could partially suppress Rp1-D21-induced HR ( Fig . 9C ) . To investigate whether different intra-molecular interactions are correlated with the different activities of Rp1-D21 and Rp1-D , we co-expressed pair-wise combinations of different domains in N . benthamiana and analyzed by Co-IP . We observed that NB-ARCD21 interacted with CCD21 , but not with LRRD21 or LRRdp2 ( Fig . 10; S7A–S7B Fig . ) . Additionally , we did not observe interactions between NB-ARCD and CCD or between NB-ARCD and LRRD under our conditions ( Fig . 10; S7A–S7B Fig . ) . Since the CC and the NB-ARC domains in Rp1-D21 and Rp1-D have different interactions , we investigated the regions required for this interaction . As noted above , the ARC2 domain ( including Patch 1 and 2 ) is the major difference in the NB-ARC domain between Rp1-D21 and Rp1-D . In order to test whether this region affects the interaction between the CC and NB-ARC domains , we performed a Co-IP assay with transiently co-expressed CCD21/CCD and NB-ARCV12/NB-ARCV13 , and observed that NB-ARCV12 interacted with both CCD and CCD21 , while NB-ARCV13 did not interact with CCD21 ( Fig . 10; S7A Fig . ) . The results suggested that ARC2D21 ( AAs 458–527 from Rp1-dp2 ) is important for the interaction between CC and NB-ARC domains . In other words , ARC2D can suppress the interaction . We also observed that LRRD21 interacted with NB-ARCV12 , but not NB-ARCV13 ( Fig . 10; S7B Fig . ) , suggesting that NB-ARC1D is required for the interaction between NB-ARC and LRRD21 , which is consistent with previous report that ARC1 is required for binding to LRR of potato CNL protein , Rx [44] . To determine whether , as suggested by the recombination studies , N1184 and the C-terminal 16 AAs are involved in the intra-molecular interaction , we performed Co-IP assays with NB-ARCD21 and LRRV1 , LRRV1 ( K1184N ) or LRRV18 and observed that NB-ARCD21 interacted with LRRV1 ( K1184N ) , but not with LRRV1 and LRRV18 ( Fig . 10; S7C Fig . ) , indicating that N1184 and the last 16 AAs from Rp1-D are required for the interaction between NB-ARCD21 and LRRV1 ( K1184N ) . We investigated the effect of the five CC domain missense EMS mutations on CCD21-induced HR . Consistent with the results from full length proteins , CCR125W abolished HR induction and CCL89F and CCS108F reduced HR , while CCH59Y had no obvious effect compared to CCD21 ( Table 1 ) . Surprisingly , CCD82N induced stronger HR than CCD21 , while CCD82N in the full length Rp1-D21 did not induce HR ( Table 1 ) . Co-IP experiments were performed to explore whether the CC domain mutations affect the self-association of CCD21 . CCD82N but not the other four point mutants reduced the strength of CCD21 self-association ( S8A Fig . ) . To further test whether the mutations affect the interaction between CCD21 and NB-ARCD21 , we chose CCD82N and CCR125W , two mutants that completely abolished HR in full length , to test the inter-domain interaction . We found that both CCD82N and CCR125W greatly reduced the interaction between CCD21 and NB-ARCD21 ( S8B Fig . ) .
While 10 of the 12 Rp1-D21 loss of function EMS mutants isolated from maize also reduced or abolished the HR phenotype when transiently expressed in N . benthamiana , two , H59Y and G850D , did not . It is possible that these two mutations reduced the activity of Rp1-D21 to a level below the threshold for signaling in maize , but not in N . benthamiana due to the higher expression . One mutation ( D82N ) was in the conserved EDVID motif . Previous reports have indicated an important role for the EDVID motif in NLR function . In Rx , the EDVID motif mediates the intra-molecular interaction between the CC and the NB-ARC-LRR domains [30 , 45] and the residues flanking the EDVID motif affect inter-molecular interactions with RanGAP2 , which is required for Rx function [30 , 46 , 47] . In this study , we found that D82N abolished HR induced by full length Rp1-D21 , but not by CCD21 ( Tables 1 and 2 ) , similar to what was observed for a mutation in the EDVID motif of MLA10 [27] . We further showed that CCD82N and CCR125W reduced the interaction with NB-ARCD21 ( S8B Fig . ) . The “mousetrap model” for NLR activation [48 , 49] suggests that , like a mousetrap , the NLR protein must be ‘set’ in a primed state , ready to be activated . The apparent paradox that the D82N mutation causes loss of autoactivity in full length Rp1-D21 but not in the CC domain alone may be explained by the reduced association of CCD21 and NB-ARCD21 and the consequent loss of ability to set up this initial primed state in the full-length protein . It is also likely that the mutations in the CC domain might disrupt the inter-molecular interactions of Rp1-D21 with other co-factors . Three of the missense mutations were within or next to the ADP binding pocket of the NB-ARC domain when the NB-ARC was modeled onto the APAF-1 structure; T260I was close to the RNBS-A motif , E312K was next to Walker B motif and P398L was in the GLPL motif ( Table 1; Figs . 1 and 8B ) , suggesting that these mutations might change the ADP binding state . Consistent with our results , important roles for the RNBS-A , Walker B and GLPL motifs on modulating HR have been reported in other NLRs [14 , 42 , 44] . Mutations in the Walker B and GLPL motifs can also affect the intra-molecular interaction between CC and NB-ARC-LRR domains as evidenced in Rx [30] . Thus , we infer that the three mutations in the NB-ARC domain might also affect the interaction between CCD21 and NB-ARCD21 . Finally , four of the suppressor mutations ( S737L , S794F , G850D and P1180S ) were in the LRR domain ( Fig . 2B ) . Loss-of-function mutations in the LRR domain have been reported in several NLRs [16 , 50 , 51 , 52 , 53 , 54] . According to our model ( see below ) , the mutations likely abolish the ability of the LRRD21 domain to destabilize the interaction between NB-ARCD21 and CCD21 . Mutations from H ( histidine ) to A ( alanine ) or D ( aspartate ) to V ( valine ) in the highly conserved MHD domain result in constitutive activity of multiple NLRs from multiple species [17 , 27 , 37 , 41 , 55 , 56] . In the Flax NLR , M , ADP is bound to wild-type NB-ARC domain , while ATP is bound to the autoactive MHD mutant M ( D555V ) [15] , providing direct evidence that the inactive “off” NLR binds ADP while the active “on” NLR binds ATP . It appears therefore that the MHD motif is important for maintaining the NLR protein in its appropriate state of activity . Rp1-dp2 and Rp1-D as well as several other CNLs , but not TNLs , have two MHD motifs [17 , 37] . We have termed these motifs MHD1 and MHD2 ( Fig . 4A ) . MHD1 is more conserved and is the functional MHD motif for many NLRs as defined in previous studies [15 , 17 , 27 , 38 , 56] . The aspartate ( D ) in MHD2 is quite widely conserved throughout most CNLs [37] . Of the CNLs that contained two MHD motifs [37] , the effects of mutations in the MHD1 domain only have been reported for tomato NLR Mi-1 . 2 ( H840A ) and Mi-1 . 2 ( D841V ) . Both mutations activate these proteins [34] . The possible function of the MHD2 domain has recently been examined in rice CNL-RGA5 , but no functional effect was observed [57] . Transient expression of Rp1-dp2 ( D517V ) and of V1 ( D517V ) and V16 ( D522V ) , in which the MHD2 was mutated , conferred autoactive HR , while no MHD1 mutation had this effect ( Figs . 4B and 4C; S3 Fig . ) . It is possible that the MHD2 rather than the MHD1 domain is functional in Rp1-dp2 , or that the MHD1 and MHD2 domains coordinate the activity . This is the first report showing the functional significance of the MHD2 domain of a CNL for its activity . The P-loop motif in the NB-ARC domain regulates nucleotide binding . P-loop mutations abolish the ability to confer disease resistance or HR induction in multiple NLR proteins [14 , 15 , 27 , 34 , 38 , 39 , 56] . As expected , the HR induced by Rp1-D21 and the MHD2 mutant Rp1-dp2 ( D517V ) was abrogated by the introduction of a P-loop mutation ( Fig . 4D ) , indicating that the activity of the Rp1 proteins is P-loop dependent . There are several examples in the literature of recombinations between the TIR-NB-ARC or CC-NB-ARC and LRR domains of different NLRs resulting in autoactive proteins . The recombination of CC-NB-ARC from Gpa2 with LRR from Rx1 produces a gene conferring autoactive HR; The combination of Gpa2-ARC2 and the first two repeats of Rx-LRR region is essential for autoactivity [36 , 44] . Domain swaps between RPS5 and RPS2 [29] and between Mi-1 . 1 and Mi-1 . 2 [58] also gave rise to genes conferring an autoactive phenotype . In contrast to these artificially-constructed autoactive NLRs characterized in transient assays , Rp1-D21 is an autoactive protein that occurred via recombination and was identified in its endogenous genetic background . To identify the precise structural requirement for its activity , we performed a systematic structural analysis of Rp1-D21 using a set of artificial recombinants between the two ‘parental’ alleles ( Fig . 6 ) and showed that the combination of AAs 458–651 ( the ARC2 and N-terminus of the LRR region ) from Rp1-dp2 and the C-terminal LRR region ( especially N1184 and the C-terminal 16 AAs ) from Rp1-D was critical for the autoactivity of Rp1-D21 . This combination either destabilized the Rp1-D21 intra-molecular interactions that cause the inactive resting state , or stabilized interactions resulting in the active state ( see below ) . In other words , these two regions appear to be important for maintaining the parental proteins in an inactive ‘resting’ state . Thus , in light of the “mousetrap” model [48] , this region of the Rp1 family NB-ARC and LRR domains is involved in a precarious autoinhibiting conformation that is easily broken by alterations including exchange of a very small C-terminal region of the LRR domain , or presumably , in a wild-type context by the action of the cognate effector protein . It seems evident that the NB-ARC and LRR domains within each NLR must co-evolve to maintain the NLRs in a suitably auto-inhibited resting state in the absence of pathogen infection while maintaining the ability to respond to the cognate effector via intra-molecular interactions [36 , 38 , 55] , but that precise mechanisms of auto-inhibition and activation may vary between NLRs . The ARC2 domain is important for function in several NLRs [36] . We identified two major polymorphic regions localized in ARC2 that differentiated the NB-ARCD21 and NB-ARCD domains , Patch 1 and Patch 2 , and showed that they are important for the autoactivity of Rp1-D21 ( Figs . 1 and 8 ) . Replacing either Patch 1 or Patch 2 of Rp1-D21 with the region from Rp1-D almost completely suppressed the autoactive phenotype ( constructs V13 , V19 and V20; Fig . 8D ) . In agreement with our structural modeling , these two patches were surface-exposed and carried largely opposite charges in NB-ARCD21 compared to NB-ARCD ( Fig . 8B ) . The prevailing model for the activation of NLRs is that the off state binds ADP while the on state binds ATP [36 , 59] . Patch 2 is located three AAs N-terminal to the MHD1 motif and is localized in the exposed surface next to the nucleotide binding pocket [40] , suggesting it might affect the state of nucleotide binding . Patch 1 is adjacent to a conserved RNBS-D motif in the ARC2 domain ( Figs . 1 and 3A ) . Mutations in or next to the RNBS-D motif of PM3 , RPM1 and Rx affected their function [16 , 36 , 45 , 55] . Thus , the sequence differences of Patch 1 between NB-ARCD21 and NB-ARCD might also affect the protein activity through disturbing the function of the RNBS-D motif . Interestingly , we found that ARC2D21 is critical for the interaction between NB-ARCD21 and CCD21 or CCD ( Fig . 10 , compare V13 with V1 and Rp1-D with V12 , and see S7 Fig . ) . This is , to our knowledge , the first demonstration of the role of ARC2 in determining NB-ARC and CC interaction in plant NLRs . Our modeling data also suggests why we detected interaction between NB-ARCD21 and CCD21 but not between NB-ARCD and CCD ( Fig . 10; S7 Fig . ) . In the CC domain , the side chain of the EDVID motif of CNLs is largely negatively charged , and this motif is thought to mediate intra-molecular interactions of the CC domain with the NB-ARC-LRR domain of Rx [30 , 60] . In Rp1 proteins , the negatively charged EDVID and the positively-charged ARC2D21 are apparently very important for the interaction between CCD21 and NB-ARCD21 while the ARC2D is more negatively charged ( Fig . 8C ) and thus may not interact as strongly with the CC . Interestingly , we observed that AAs 190–260 from the NB domain were sufficient to suppress CCD21-induced HR ( S4 Fig . ) , suggesting that ARC2D21 is not the only region which can regulate Rp1-D21 autoactivity . CCD21 and CCD domains alone were sufficient to induce HR when fused with EGFP , but not with 3×HA or on their own ( Fig . 5; S1 Fig . ) . This phenomenon of “tag-dependent activity” has been observed in other NLR studies [30 , 61] . The functional domains required for HR induction vary in different NLR proteins . The CC domain alone is sufficient for the HR phenotype triggered by MLA10 , and also by NRG1 and ADR1 , which contain atypical CC domains [28 , 41] , while the NB-ARC domain of Rx can trigger HR [30] , and for RPS5 the CC-NB-ARC is sufficient [29] . The fact that transient expression of CC-NB-ARCD21 or CC-NB-ARCD did not induce HR suggested that their respective NB-ARC domains repressed the signaling by the CC domain . Since full length Rp1-D21 induced HR and full length Rp1-D did not ( Fig . 3 ) , we inferred that the LRRD21 or LRRD domains are structurally incompatible with , respectively , repression of CCD21 autoactivity or with CCD autoactivity ( Fig . 5 ) . While interaction between the LRR and NB-ARC domains was detected in some combinations ( e . g . , see proteins V12 , V1 ( K1184N ) in Fig . 10 ) , we did not detect any interaction between the LRR and NB-ARC domains of Rp1-D , Rp1-dp2 or Rp1-D21 in trans ( Fig . 10; S7 Fig . ) . It is possible though that these domains interact in the full length context , or that their interactions are weak or transient and cannot be detected in our Co-IP conditions . Consistent with this , LRRD21 was unable to alter the suppression of CCD21 autoactivity by NB-ARCD21 in trans ( Table 2 ) . However , we cannot exclude the possibility that the lack of suppression in trans was due to the relatively low expression of LRRD21 ( Table 2; Fig . 5B ) . A similar finding of cis but not trans interaction/autoactivity was reported for potato Rx and autoactive MHD mutants of tomato Mi-1 . 2 [34 , 44 , 62] . It is notable also that the CC-NB-ARC fusion from MLA10 and Rx can still trigger HR [27 , 30] , but not in Rp1-D and Rp1-D21 . LRR domains have multiple reported roles in NLR activation . They are essential for the activation of the NLRs Rx and Mi-1 . 2 [34 , 62] . In the case of RPS5 , the first four LRRs are the minimum region sufficient to inhibit the autoactive phenotype conferred by the CC-NB-ARC domain [29] . Conversely , the activation of RPS5 in response to disease requires the entire LRR domain [29] . Our analyses of Rp1-D21 C-terminal deletion constructs and LRR domain swap data ( Fig . 7B; S5 Fig . ) further confirmed the importance of the C-terminal LRR domain in regulating the activity of the Rp1 proteins . Similarly , the importance of C-terminal LRR domain has also been observed in flax TNL , L proteins [37] . The interaction patterns observed between NB-ARC and LRR from V1 , V1 ( K1184N ) and V18 suggested that K1184N and the last 16 AAs from Rp1-D are required for the inter-domain interaction ( Fig . 10; S7 Fig . ) . No interaction was detected between the NB-ARC and LRR domains from V1 or between those from V18 when they were expressed in trans , but a trans-interaction was detected between these domains using the V1 ( K1184N ) construct . Assuming that these interactions are maintained in the full length proteins , this result implies that the ability of V1 ( K1184N ) to induce HR is due to the interaction of the LRR with the NB-ARC , which apparently is then locked into an activated state allowing CC-dependent HR . This also requires the C-terminal 16 amino acids of the Rp1-D21 LRR domain ( Fig . 10; S7 Fig . ) . A further inference is that the C-terminal 16 AAs and N1182 in Rp1-D21 ( corresponding to N1184 in Rp1-D ) are crucial for the inhibition of NB-ARC activity that is required for CC-dependent HR . Alternate “on” or “off” states mediated through intra-molecular interactions are thought to be one of the major mechanisms regulating NLR activity [51 , 62] . For example , intra-molecular interactions were detected between CC and NB-ARC-LRR , and CC-NB-ARC and LRR domains of Rx in the absence but not in the presence of its cognate effector , the PVX coat protein [62] . Similarly the differentially-activated states of Rp1-D , Rp1-dp2 , Rp1-D21 and our additional recombinant proteins are likely due to the specific intra-molecular interactions found within each protein . While Rp1-D21 and V3 confer HR , their reciprocal constructs , V16 and V17 , do not ( Fig . 6C ) , indicating that the autoactivity of the recombinant protein is triggered by specific combination of sequences from Rp1-dp2 and Rp1-D . Our model ( Fig . 11 ) explaining the intra-molecular interactions underlying the activity of the Rp1 proteins is as follows: While we have not identified a cognate effector of Rp1-D and do not know how it is detected , we hypothesize that its presence is necessary for activating Rp1-D via direct or indirect interaction with the LRR domain . This model is based both upon the data presented here and a review of the related literature . While plant NLR resistance genes share similar structures and appear to perform very similar functions , it is clear that the molecular mechanisms underlying their function and their appropriate activation , while sharing certain similarities , vary substantially . This variation likely reflects the intimate and unique co-evolutionary processes each NLR has undergone both with respect to the interactions among their different domains and with their cognate pathogen effector proteins , and/or the host targets or decoys of those effectors . The molecular mechanisms underlying the auto-inhibition of the maize NLR common rust resistance protein Rp1-D and its autoactive derivative Rp1-D21 show both broad similarities to and distinct differences from what is known of other NLRs . Furthermore , we have identified several unique , or previously un-observed features , including: the AAs required for HR induction identified through EMS mutagenesis screening; the functional characterization of MHD2 motif; the involvement of the ARC2 domain in the interaction of the NB-ARC domain with the CC domain; the observation that N1184 and the C-terminal 16 AAs are involved both in the LRR/NB-ARC physical interaction and in regulating activity . Recently , several NLR genes that have been transferred between plant species were demonstrated to still confer expected disease resistance specificities without obvious fitness effects , eg . Arabidopsis RPS4 ( Resistance to Pseudomonas syringae 4 ) , RRS1 ( Resistance to Ralstonia solanacearum 1 ) and barley MLA1 [64 , 65] . Genetic manipulation of Rx indicates stepwise artificial evolution can be used to reduce the costs associated with disease resistance of NLRs [66] . Rp1-D21 is an autoactive mutant conferring nonspecific response to multiple maize rust species , including P . sorghi and P . triticina [19 , 21] . Presumably the resistance conferred by Rp1-D21 might extend beyond rusts to other biotrophic and hemi-biotrophic pathogens , however , the severe growth penalties associated with the expression of this gene make its application in agricultural production impractical . In conjunction with suitable promoters , it may be possible to engineer maize or even other plants with some of the chimeric constructs characterized here that confer a weaker HR phenotype to achieve an elevated disease resistance with fewer fitness consequences .
Wild type N . benthamiana plants were grown at 23°C with a cycle of 16 hrs light/8 hrs dark . Maize line Rp1-D21-H95 [25] was used to isolate the genomic DNA sequence of Rp1-D21 . To generate material for EMS mutageneis , Rp1-D21 was first introgressed into A632 by 7 backcrosses ( BC7 ) . A few of these BC7 Rp1-D21 heterozygous plants were self-pollinated to generate Rp1-D21 homozygotes in the A632 background which were identified on the basis of their enhanced Rp1-D21 severity compared to the heterozygotes . Pollen from the homozygous Rp1-D21 plants in an A632 background was collected and treated with EMS for 45 min before using it to fertilize the ears of an inbred line H95 . The resulting M1 population was all heterozygous for Rp1-D21 and showed an HR phenotype with a relatively stunted growth stature . In this background , Rp1-D21-suppressed plants due to EMS mutagenesis were easily distinguished from the rest of the M1 siblings because of their non-lesioned and highly robust growth phenotype ( Fig . 2A ) . About 23 , 000 M1 progenies were screened in field to identify intragenic suppressor mutants that had lost the Rp1-D21 phenotype . In total , 32 mutants which lacked the HR phenotype of Rp1-D21 were identified and 14 of them were characterized in detail . Gene specific primers for Rp1-D21 ( S1 Table ) were used to sequence the Rp1-D21 gene in these mutants . All primers used in this study are listed in S1 Table . In Rp1-D and all its paralogs , no intron is found in the open reading frame ( ORF ) region [20 , 35] , thus we amplified the ORFs of Rp1-D , Rp1-dp2 and Rp1-D21 from the plasmids gifted by Dr . Scot Hulbert ( Washington State University ) , and cloned them into pENTR directional TOPO cloning vector ( D-TOPO , Invitrogen ) . After sequencing , they were transferred into gateway vectors by LR reactions: pGWB2 ( no tag ) , pGWB14 ( with a 3×HA epitope tag in the C-terminus ) or pSITEII-N1-EGFP ( with EGFP epitopic tag in the C-terminus ) [67 , 68] . Rp1-D21 was also isolated from maize line Rp1-D21-H95 using the primers listed in S1 Table . The different domains ( CC , CC-NB-ARC , NB-ARC , NB-ARC-LRR , and LRR ) of Rp1-D21 or Rp1-D were amplified using primer pairs listed in S1 Table . The resulting PCR products were cloned into D-TOPO and sequenced , then transferred into pGWB2 , pGWB14 or pSITEII-N1-EGFP by gateway LR reactions . Overlapping extension PCR primers ( S1 Table ) were designed for generating the site-directed mutations: Rp1-D21 ( K225R ) , Rp1-D ( H517A ) , Rp1-D ( D518V ) , Rp1-D ( H521A ) , Rp1-D ( D522V ) , Rp1-dp2 ( D513V ) , Rp1-dp2 ( H512A/D513V ) , Rp1-dp2 ( H516A ) , Rp1-dp2 ( D517V ) , Rp1-dp2 ( K225R/D517V ) and 12 missense EMS mutations listed in Table 2 . The site-directed mutations were cloned into D-TOPO and verified by sequencing and sub-cloned into the gateway vector pGWB14 by LR reaction . NLR sequences were aligned by ClustalW ( www . ebi . ac . uk ) , and edited by BioEdit software . Homology modeling of the NB-ARC domain was performed with Phyre 2 [69] based on the crystal structure of human APAF1 ( PDB: 1z6t ) . The three-dimensional structure and the surface electropotential were mapped using PyMOL ( http://www . pymol . org/ ) . Agrobacterium tumefaciens strain GV3101 ( pMP90 ) transformed with binary vector constructs was grown at 28°C overnight in 5ml L-broth medium supplemented with appropriate antibiotics . The bacteria were collected at 4 , 000g by centrifugation and resuspended in 2 ml resuspension buffer ( 10 mM MES pH5 . 6 , 10 mM MgCl2 and 200 μM acetosyringone ) . The final concentration of the bacteria was diluted to the OD600 of 0 . 5 using the same resuspension buffer . To prevent the onset of post-transcriptional gene silencing and improve the efficiency of transient expression , a strain containing p19 protein was included at OD600 of 0 . 2 to all the bacteria strains [70] . The solution was left at room temperature for 3 hrs on bench before infiltration into the abaxial side of N . benthamiana leaves . After infiltration , plants were put at room temperature with 16h-light/8h-dark . At least 15 individual leaves were infiltrated by different constructs , and each experiment was repeated at least three times . For protein expression analysis , three leaf discs ( 1 . 2 cm diameter ) from different single plants were collected at 30 hours post infiltration ( hpi ) . The samples were ground with prechilled plastic pestles in liquid nitrogen , and total protein was extracted in 160 μl extraction buffer [20 mM Tris·HCl ( pH 8 . 0 ) , 150 mM NaCl , 1 mM EDTA ( pH 8 . 0 ) , 1% Triton X-100 , 0 . 1% SDS , 10 mM DTT , 40 μM MG132 , and 1× plant protein protease inhibitor mixture ( Sigma-Aldrich ) ] . Samples were centrifuged at 14 , 000 g for 15 min at 4°C , and 12 μl supernatants were mixed with 2× Laemmli buffer and loaded for SDS-PAGE . For Co-IP assay , EGFP- and 3×HA-tagged constructs were transiently co-expressed in N . benthamiana . Agrobacterium carrying each construct were diluted to a final concentration of OD600 = 0 . 4 plus p19 with OD600 = 0 . 2 . Samples were collected at 30 hpi , and proteins were extracted by grinding 0 . 8 g of leaf tissues in 2 . 4 ml extraction buffer [50 mM HEPES ( pH 7 . 5 ) , 50 mM NaCl , 10 mM EDTA ( pH 8 . 0 ) , 0 . 5% Triton 100 , 4 mM DTT and 1× plant protein protease inhibitor mixture ( Sigma-Aldrich ) ] . Extracts were centrifuged at 14 , 000 rpm for 20 min at 4°C , and 2 ml supernatant was mixed with 30 μl anti-GFP microbeads ( Miltenyi Biotec ) and rotated for 3 hrs at 4°C . The samples were passed through pre-equilibrated MACS separation columns , and washed four times ( 1 ml , 1 ml , 500 μl and 500 μl ) by washing buffer [50 mM HEPES ( pH 7 . 5 ) , 150 mM NaCl , 10 mM EDTA ( pH 8 . 0 ) , 0 . 2% Triton 100 , 4 mM DTT and 1× plant protein protease inhibitor mixture ( Sigma-Aldrich ) ] . The proteins were eluted by 100 μl pre-heated elution buffer and separated by SDS-PAGE . Proteins were transferred to nitrocellulose membrane ( Fisher ) , and analyzed by western blot . HA detection was performed using a 1:350 dilution of anti-HA-HRP ( horseradish peroxidase ) ( Cat# 12013819001 , Roche ) . GFP detection was performed using a 1:8 , 000 dilution of primary mouse monoclonal anti-GFP ( Cat# ab1218 , Abcam ) , followed by hybridization with a 1:15 , 000 dilution of anti-mouse-HRP second antibody ( Cat# A4426 , Sigma ) . The HRP signal was detected by ECL substrate kit ( Supersignal west femto chemiluminescent substrate , Thermo Scientific ) . | The plant hypersensitive defense response ( HR ) is a rapid , localized cell death , usually occurring upon the recognition of specific pathogen-encoded molecules and consequent activation of nucleotide binding-leucine rich repeat ( NLR ) proteins . Rp1-D21 , a naturally-occurring mutant caused by the recombination of two NLR genes , confers a ‘lesion mimic’ , HR-like phenotype in the absence of pathogen infection and provides a powerful tool to investigate the molecular mechanisms of NLR regulation . Here we report the results of a genetic screen in maize that identified novel mutations abrogating Rp1-D21-induced HR . To characterize the function of Rp1-D21 , we transiently expressed Rp1-D21 and various derivatives in Nicotiana benthamiana to observe the resulting levels of HR . We furthermore examined the protein-protein interactions between and within different Rp1-D21 derivatives . We report novel insights into the precise structural requirements for NLR function and determine the function of a previously undefined motif . These insights enable a better understanding of how NLRs regulate the switch between the resting and the active states . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Molecular and Functional Analyses of a Maize Autoactive NB-LRR Protein Identify Precise Structural Requirements for Activity |
To help learn how phytopathogens feed from their hosts , genes for nutrient transporters from the hemibiotrophic potato and tomato pest Phytophthora infestans were annotated . This identified 453 genes from 19 families . Comparisons with a necrotrophic oomycete , Pythium ultimum var . ultimum , and a hemibiotrophic fungus , Magnaporthe oryzae , revealed diversity in the size of some families although a similar fraction of genes encoded transporters . RNA-seq of infected potato tubers , tomato leaves , and several artificial media revealed that 56 and 207 transporters from P . infestans were significantly up- or down-regulated , respectively , during early infection timepoints of leaves or tubers versus media . About 17 were up-regulated >4-fold in both leaves and tubers compared to media and expressed primarily in the biotrophic stage . The transcription pattern of many genes was host-organ specific . For example , the mRNA level of a nitrate transporter ( NRT ) was about 100-fold higher during mid-infection in leaves , which are nitrate-rich , than in tubers and three types of artificial media , which are nitrate-poor . The NRT gene is physically linked with genes encoding nitrate reductase ( NR ) and nitrite reductase ( NiR ) , which mobilize nitrate into ammonium and amino acids . All three genes were coregulated . For example , the three genes were expressed primarily at mid-stage infection timepoints in both potato and tomato leaves , but showed little expression in potato tubers . Transformants down-regulated for all three genes were generated by DNA-directed RNAi , with silencing spreading from the NR target to the flanking NRT and NiR genes . The silenced strains were nonpathogenic on leaves but colonized tubers . We propose that the nitrate assimilation genes play roles both in obtaining nitrogen for amino acid biosynthesis and protecting P . infestans from natural or fertilization-induced nitrate and nitrite toxicity .
Successful pathogens must efficiently exploit the nutrients of their hosts to support their growth . Pathogens accomplish this through mechanisms that include manipulating hosts to shift nutrients to sites of infection , generating feeding structures , expressing transporters for the nutrients , and making metabolic enzymes for assimilating the nutrients [1–3] . Studies in fungi and bacteria have identified transporters and metabolic pathways that contribute to pathogenesis [4–10] . Understanding how genes for such functions are expressed during infection yields insight into what nutrients are available and preferred by the pathogen . The nutritional strategies used by a plant pathogen may vary depending on the metabolic condition of its host , stage of infection , and lifestyle . Biotrophic pathogens typically feed from living cells , absorbing nutrients from the apoplast or cell-penetrating structures such as haustoria [1] . While a role of oomycete haustoria in nutrition is not proved , sugar and amino acid transporters expressed specifically in haustoria have been identified from rusts and powdery mildew fungi [5 , 7] . In contrast with biotrophs , necrotrophs are believed to primarily acquire nutrients from damaged host cells . Hemibiotrophs shift strategies during the disease cycle . The oomycete Phytophthora infestans , for example , behaves as a biotroph during much of its interaction with its potato and tomato hosts , but may shift to necrotrophy late in infection [11] . Although pathogens must obtain carbon , nitrogen , phosphate , and sulfate to synthesize biosubstances for growth , they vary in their preferred raw materials . For example , many bacteria prefer inorganic nitrogen over amino acids as a nitrogen source [12]; most use nitrate and nitrite reductases to convert nitrate to ammonium , from which the nitrogen can be incorporated into amino acids [13] . In contrast , eukaryotes such as fungi and oomycetes typically prefer amino acids over nitrate [14] . Reflecting these trends , mutations in nitrate reductase were shown to reduce pathogenicity of the bacteria Pseudomonas and Ralstonia [15 , 16] while not affecting Fusarium [17] . Most fungi and oomycetes have nevertheless retained the nitrate assimilation pathway , perhaps since it benefits them during survival in debris or some conditions of growth [18] . Plants also encode a similar pathway for using nitrate , which may reach near-molar levels in certain tissues when fertilizers are applied [19 , 20] . Animals in contrast lack these enzymes and are thus prone to nitrate toxicity , which results primarily from the oxidation of hemoglobin but also involves damage to other proteins and membranes [21–23] . This study focuses on nutrient assimilation pathways in P . infestans , which causes the devastating late blight diseases of potato and tomato . This is an interesting system for studying nutrient utilization since P . infestans infects diverse host tissues with distinct chemical compositions , such as leaves and potato tubers . Also , while functional studies of nutrient transporters and associated metabolic genes have been performed in filamentous fungi , as of yet there are limited data on these in any oomycete . We therefore mined the P . infestans genome for transporters potentially involved in nutrient transport , and measured their expression during growth on tomato leaves , potato tubers , and rich , semidefined , and defined minimal media . We observed dynamic changes in mRNA levels of the transporters , including patterns specific to different host organs and each media . The finding that a nitrate transporter was expressed at very high levels in leaves but not in tubers or media led to a detailed study of the entire nitrate assimilation pathway . This included silencing genes for the nitrate transporter ( NRT ) , nitrate reductase ( NR ) , and nitrite reductase ( NiR ) , which abolished the ability of P . infestans to colonize leaves while having only a minor effect on tuber infection . We also observed a spreading effect associated with gene silencing by DNA-directed RNAi , which has implications for gene function studies in oomycetes .
P . infestans genes encoding transporters that are potentially involved in nutrition were identified through domain and protein similarity searches . Included in the analysis were transporters for organic compounds ( sugars , amino acids , nucleosides , etc . ) and for ammonium , nitrate , phosphate and sulfate . Excluded were ABC transporters and proteins that serve primarily as ion channels . This identified 453 genes that encode proteins representing 19 conserved transporter families ( Fig 1 , S1 Table ) . Only 168 of these proteins were described as transporters in the original genome annotation . The largest group was the Major Facilitator Superfamily ( MFS ) , with 111 proteins . At the other size extreme was the Nucleobase Cation Symporter ( NCS ) family , which had one member . Most of the families , representing 341 of the 453 proteins , are found at the plasma membrane of other eukaryotes and thus may participate in nutrient uptake during plant colonization by P . infestans . A few are likely to play other roles , such as intracellular trafficking or efflux . These include 58 mitochondrial proteins ( Mitochondrial Carriers and Mitochondrial-Plastid Porin ) , 19 Multidrug/Oligosaccharidyl-lipid/Polysaccharide Flippases which translocate oligosaccharides destined for protein glycosylation [24] , and 34 Drug/Metabolite Transporters which in other taxa include both intracellular solute carriers that transport nucleotide sugars between the ER and Golgi , as well as plasma membrane efflux proteins [25] . To compare the transporter complement of P . infestans with those of other plant pathogens , we examined a second oomycete , Pythium ultimum , and the ascomycete Magnaporthe oryzae . These species exhibit necrotrophic and hemibiotrophic lifestyles , respectively . P . infestans , Py . ultimum var . ultimum , and M . oryzae were predicted to encode 453 , 452 , and 381 transporters which correspond to 2 . 6 , 3 . 0 , and 3 . 1% of their genes , respectively ( Fig 1 ) . The most dramatic differences between the two oomycetes and the fungus were the absence from the latter of Phosphate/Sodium Symporters , Folate-Biopterin , and SWEET Transporters . This was not entirely surprising since these families lack a very broad taxonomic distribution . Folate-Biopterin transporters , for example , are found in bacteria , plants , and some protists but not animals [27 , 28] . Other major differences between the fungus and the oomycetes were a five-fold reduction in the number of Amino Acid/Auxin Permeases ( AAAP ) , a doubling of the size of the MFS family , and a reduction of the size of the Sulfate Permease family in M . oryzae . That the sulfate and amino acid permeases are over-represented in oomycetes compared to other eukaryotes has been reported previously by Seidl et al . [29] . Few major changes were observed between the two oomycetes . One modest difference was the expansion of the Dicarboxylate Amino Acid-Cation Symporter ( DAACS ) family in Py . ultimum , which had 14 members compared to only 11 in P . infestans . The Folate-Biopterin family was also expanded in Py . ultimum , with 55 encoded predicted proteins versus 41 in P . infestans . Also , only P . infestans encoded a predicted Nucleobase Cation Symporter ( NCS ) . RNA-seq experiments were designed to study the expression of transporters during growth in planta compared to artificial media , on different host tissues ( leaves and tubers ) , and rich versus defined minimal media . As a goal was to examine cultures during the biotrophic stage of growth , in preliminary experiments we observed that haustoria were abundant in infected tomato leaves and potato tuber slices at two to three days after infection ( Fig 2A ) . Based on these observations , a preliminary RNA-seq experiment ( Experiment One; Table 1 ) was performed in which tomato leaves ( cv . New Yorker ) and potato tubers ( cv . Yukon Gold ) were sampled at three and six days post-infection ( dpi ) with P . infestans isolate 1306 , using two to four biological replicates per sample . Sporulation was first observed at 5 dpi , therefore the 3 and 6 dpi timepoints were taken about 2 days before and one day after initial sporulation , respectively . The fraction of reads mapping to P . infestans ranged from 4 . 1 to 39 . 8% ( Fig 2B; Table 1 ) . To help assess these preliminary tests , we examined the expression of the genes encoding effector Avr3A and haustorial protein Hmp1 which are markers of biotrophic growth , Npp1 which is a marker of necrotrophic growth , and a flagella-associated centrin which is a marker of sporulation [3 , 30] . In both leaves and tubers , Avr3A and Hmp1 mRNAs were >10-fold higher in the leaf and tuber samples at 3 dpi than 6 dpi ( Fig 2B ) . Conversely , mRNAs for Npp1 and the flagella-associated centrin were >10-fold higher at 6 than 3 dpi . Haustoria were also evident in leaves and tubers at 3 dpi . Thus , 3 dpi appeared to represent a biotrophic growth stage in leaves and tubers under our infection and incubation conditions . Using a minimum FPKM cut-off of 1 . 0 , expression was detected for 403 of the 453 transporter genes in at least one sample . A second RNA-seq experiment was then performed ( Experiment Two ) which again used tomato leaves and potato tubers , but included earlier time points to focus more on the biotrophic stage . These were 2 , 3 , 4 , and 5 dpi in leaves and 1 . 5 , 2 . 5 , and 4 dpi in tubers . The experiment also used Russet potato tubers , which seemed to be more easily colonized by P . infestans isolate 1306 than Yukon Gold . The experiment also included 3 day-old nonsporulating cultures in rye-sucrose media ( RS; rye A in reference [31] ) , and defined and semidefined media . The latter two were based on the recipe by Xu et al . [32] and used glucose and fumarate as carbon sources and either ammonium sulfate ( MNH ) or amino acids from a casein hydrolysate as the nitrogen source ( MAA ) . In the RS and MAA media , sporulation began after about four days except for MNH media , which did not support sporulation . On the leaf and tuber samples , sporulation was first observed at 4 dpi . In Experiment Two , the percentage of reads mapping to P . infestans ranged from 3 . 9 to 76 . 4% in the plant samples compared to an average of 96% in the media samples ( Table 1 ) . In leaves , the expression of both Hmp1 and Avr3a were high at 2 dpi and declined through 5 dpi , and similar results were obtained on tubers ( Fig 2C ) . In leaves and tubers , Npp1 expression was first detected at low levels at days 3 and 2 . 5 , respectively , and rose in each succeeding timepoint . Expression of the flagella-associated centrin gene was first detected at 3 and 2 . 5 dpi in leaves and tubers , respectively . It is notable that this was at least a full day before the first sporangia were present . In Experiment Two , 411 of the 453 transporter genes were expressed with a FPKM>1 . 0 in at least one sample . A heatmap comparing expression of the transporters in the artificial media and plant samples from Experiment Two is shown in Fig 3A . The leaf and tuber samples clustered with each other , separate from all three artificial media samples . The data used to construct the heatmap , along with the results from Experiment One , are shown in S1 Table . Based on a FDR cut-off of 0 . 05 , 56 genes were up-regulated significantly in the earliest leaf or tuber timepoints compared to the artificial media in Experiment Two , while 207 were down-regulated . Nineteen genes were upregulated by at least 4-fold in both leaves and tubers compared to each artificial media , while 66 were down-regulated by at least 4-fold ( Fig 3B ) . It is notable that distinct expression patterns were observed in the three artificial media . Conclusions about the number of infection-induced or repressed genes would thus have varied if drawn from comparisons with fewer types of media . The 17 genes that were most-expressed in the early infection stage in Experiment Two are of interest since they may encode haustoria-associated transporters or may be induced by a plant signal . These included six members of the Amino Acid/Auxin Permease ( AAAP ) family , three Folate-Biopterin Transporters , one Mitochondrial Carrier protein , and seven Major Facilitator Superfamily ( MFS ) proteins . To confirm the validity of these results , the expression levels of the 17 genes in Experiments One and Two are compared in Fig 4A . Although there are some quantitative differences , 16 of the 17 genes were also expressed at much higher levels at the 3 dpi versus 6 dpi time points in Experiment One . The sole exception was PITG_02409 , which had similar mRNA levels in tubers at 3 dpi and 6 dpi . Experiment Three ( which used different plant material from that used in Experiments One and Two ) was performed to further validate the data , focusing on three AAAPs that were expressed at higher levels in planta than in artificial media . RNA was extracted from purified sporangia and from leaves at 1 . 5 , 2 . 5 , 3 . 5 , 4 . 5 and 5 . 5 dpi , and analyzed by RT-qPCR . The results demonstrated that PITG_12808 , PITG_17803 , PITG_20230 mRNAs were very low in sporangia , peaked between 1 . 5 and 2 . 5 dpi in the leaves , and then dropped precipitously ( Fig 4B ) . The three genes were expressed at the same time , or slightly earlier , than the biotrophic stage marker Avr3A . Transcripts of the necrotrophic stage marker Npp1 increased as expression of the three AAAPs fell . These patterns match that seen in Experiments One and Two . To help place the expression patterns in context with the levels of transporter substrates , the concentrations of soluble sugars ( sucrose , glucose , fructose ) , NO3− , NH4+ , and free amino acids were determined in plant and media samples ( Fig 5 ) . This showed , for example , that the levels of free amino acids in tubers and MAA media were similar to each other , but 10-fold higher than in leaves or RS media . Also , NO3− levels were much higher in leaves than tubers or the artificial media . It is recognized that these are tissue averages , and levels may vary between different regions of a leaf or tuber , between plant cells and the apoplast , and at different infection timepoints [33 , 34] . Nevertheless , the results may help explain some of the expression differences that are described in the next section , and provide a basis for the media manipulation studies that are presented later . A heatmap showing the data from Experiment Two in which genes are classified by transporter type is shown in Fig 6 . Most families contain members that vary significantly in mRNA level between different host tissues , in planta timepoint , growth in planta compared to artificial media , or type of artificial media . It was uncommon for the majority of genes in a family to be coregulated , however , and a simple relationship did not usually exist between mRNA level , media composition , and the predicted substrate of the transporter . Infection-upregulated genes generally exhibited higher amplitudes of expression in leaves than tubers . This and other patterns of expression were similar between Experiments One and Two ( S1 Table ) . One group in which the diversity in expression patterns was evident is the Amino Acid/Auxin Permease ( AAAP ) family . In Experiment Two , using a 2-fold cut-off , 23 of its 55 expressed members ( 41% ) were expressed more in tubers or leaves compared to any artificial media ( e . g . PITG_17803 ) , 19 were expressed more on artificial media ( e . g . PITG_04400 ) , and 13 showed little change . About 19% had higher mRNA in early compared to later infection timepoints , 56% were expressed more later versus early , 20% were expressed more in tubers than leaves , and 14% more in leaves versus tubers . Similar patterns were seen in Experiment One ( S1 Table ) . Such diverse expression profiles might be explained by the response of each gene to specific metabolic cues , since a given AAAP is often specific for certain amino acids [35] . This may also explain why all AAAPs did not behave similarly on each type of artificial medium . For example , three AAAPs were >2-fold lower and eight >2-fold higher ( FDR<0 . 05 ) in NH4+-based minimal media compared to media in which NH4+ was substituted by amino acids . mRNA levels of five of these eight ( e . g . PITG_11283 ) were also >2-fold higher in leaves than tubers; these genes may be exhibiting a shared response to amino acid limitation , since tubers have ~10-fold more free amino acids than leaves . None of the eight AAAP genes were up-regulated in the relatively amino acid-poor rye grain media , but it should be noted that amino acid levels at the periphery of hyphae might be higher than shown in Fig 5 due to proteases secreted into the medium . Diverse patterns were also seen in two other families that include amino acid transporters , the Amino Acid-Polyamine-Organocation ( APC ) and Dicarboxylate/ Amino Acid:Cation Symporters ( DAACSs ) . In both groups , about half were expressed at higher levels in media and half at higher levels in planta , particularly at the later timepoints . For example , APC gene PITG_03725 and DAACSs PITG_09295 and PITG_17951 had higher average mRNA levels in leaves than media , while APCs such as PITG_11024 were expressed more in media than in leaves or tubers at any timepoint . Interpreting these patterns is complicated since some APCs and DAACSs have diverse substrates; for example , some DAACSs move fumarate , which is a component of the minimal media [36] . Also , some APCs function in efflux as well as uptake [37] . Similar patterns were seen in Experiment One . Diverse transcriptional profiles were also seen in a family that primarily includes sugar transporters , the Major Facilitator Superfamily ( MFS ) . In Experiment Two , about 15% were expressed at >2-fold higher levels in the early infection timepoints compared to artificial media , 24% were higher in the later infection timepoints compared to artificial media , and 31% were at similar levels in planta and in media but rose in both leaves and tubers as infections proceeded; similar patterns were observed in Experiment One . Significant variation was also detected between rye-sucrose media and the two other media . Other families that primarily include sugar transporters also showed diverse patterns of expression . While about half of Glycoside-Pentoside-Hexuronide:Cation Symporters ( GPH ) showed similar expression under all conditions , several were expressed at their highest levels in leaves , which was the tissue with the lowest level of soluble sugar; a similar pattern had also been seen in Experiment One . This provides an interesting contrast to another family of sugar transporters , the SWEETs . Of the 19 expressed genes in the family , 9 and 6 were up- or down-regulated by >2-fold or more , respectively , in leaves and/or tubers compared to artificial media; a similar pattern had also been seen in Experiment One . While the majority of MFSs are annotated as sugar transporters , their substrates are more diverse . The PANTHER database [38] places 41 of the 111 P . infestans MFSs into functional groups , of which about 80% or 34 proteins are sugar transporters . The rest include transporters of carboxylic acids , lipids , and nitrate ( 5 , 2 , and 2 proteins , respectively ) . There was not a strong relationship between the predicted substrate of the MFS and its expression pattern . For example , while one nitrate transporter ( PITG_09342 ) was up-regulated by >10-fold in the artificial media compared to tubers and leaves , a second nitrate transporter ( PITG_13011 ) exhibited the opposite pattern , with high mRNA in leaves and low mRNA in artificial media . The roles of PITG_13011 and the rest of the nitrate assimilation pathway are addressed in detail later in this paper . Divergent , dynamic patterns of expression were not limited to families participating in amino acid , sugar , or nutritive ion transport . For example , subsets of both the Folate-Biopterin ( FBT ) and Choline Transporter ( CTL ) groups showed patterns of expression that were higher in planta than on media or vice versa , higher in leaves than tubers and vice versa , increased or fell with time of infection , or were similar between all growth conditions . This diversity was also observed within the Equilibrative Nucleoside Transporter ( ENT ) family , although a smaller minority of their members were plant-induced . Similar patterns were seen in Experiments One and Two . Several families exhibited more coherent patterns of expression than those listed above . The majority of the primarily intracellular transporters were expressed in all tissues , such as the Mitochondrial Carrier ( MC ) , Mitochondrial Porin ( MP ) , and Drug/Metabolite Transporter ( DMT ) . A consistent pattern of plant-induced expression was also displayed by each of the three and five expressed members of the Phosphate Symporter ( PNS ) and Ammonium Transporter ( AMT ) families . While all three PNS genes were expressed nearly exclusively in tubers , different AMT genes were expressed primarily in leaves ( e . g . PITG_20291 ) or tubers ( e . g . PITG_07458 ) . Similar results were observed in Experiments One and Two . By integrating the analysis of the transcription of the genes with their genomic organization , it was observed that genes in a family that were physically linked were usually expressed coordinately . Of the expressed transporters , 61 were immediately adjacent to a gene from the same family and 73 were separated by less than two genes . This was most prevalent in the SWEET and AAAP families , where 65% and 38% of genes were adjacent to a relative , respectively . Based on the data from Experiment One , the median correlation coefficient R for the expression patterns of all linked transporter gene pairs in the five growth conditions was +0 . 59 , and the distribution of R values among linked genes was distinct ( P = 10-5 ) from unlinked genes by a Kolmogorov-Smirnov two-sample test , which compares population distributions . There was also moderate conservation of expression level ( R = +0 . 56 ) between each cluster member based on the normalized number of RNA-seq reads mapped per gene . This however did not hold for gene pairs that lacked positively correlated expression patterns , where expression levels were negatively correlated ( R = -0 . 18 ) One interesting finding from above concerned the nitrate transporter from the MFS family , PITG_13011 . Like all other nitrate transporters in P . infestans , this belongs to the high-affinity NRT2 family [18] . The gene , abbreviated hereafter as NRT , was induced in leaves compared to artificial media by an average of 70-fold in Experiment Two and 200-fold in Experiment One . With a FPKM value of 146 in leaves , NRT was the 4th most highly-expressed MFS gene . Possibly , NRT was induced in leaves due to their high NO3− levels compared to tubers and media , or was repressed in tubers due to their more abundant free amino acids or ammonium ( Fig 5 ) . The NRT gene was selected for further analysis due to its in planta expression pattern and the possibility of illuminating why late blight is reported to worsen in high-nitrogen fertilization regimes [39–42] . As noted by others [18] , NRT is part of a cluster that also encodes nitrate reductase ( NR , PITG_13012 ) and nitrite reductase ( NiR , PITG_13013 ) . The three genes allow for the uptake and conversion of NO3− to NH4+ from which nitrogen can be moved into amino acids . NRT and NR are transcribed from a common promoter region of 479 nt , while NR and NiR are separated by 403 nt and transcribed in the same direction ( Fig 7A ) . NRT , NR , and NiR are regulated in concert and expressed preferentially during leaf infection , as seen in the RNA-seq data from Experiment Two ( Fig 7 , panel B1 ) . In addition to the samples shown in Figs 6 and 7 includes RNA-seq data from germinated zoospore cysts and 5 dpi leaves . All three genes have low levels of mRNA in germinated cysts , lower expression in 2 dpi tomato leaves , higher levels in 3 dpi leaves , and very high levels in leaves at 4 dpi , which fall quickly by 5 dpi . Based on comparisons with the expression of Avr3a , Hmp1 , and Npp1 as shown in Fig 2C , it appears that NRT , NR , and NiR are expressed mainly in what may represent a transition between the biotrophic and necrotrophic stages in leaves . In contrast , low levels of expression were observed in tubers at any timepoint , in the complex RS media , in the semidefined minimal media containing amino acids , or the defined ammonium-based minimal media . A similar trend was seen in Experiment One ( Fig 7 , panel C1 ) , where high expression was observed in 3 dpi leaves , low expression in 6 dpi leaves , and virtually no expression in tubers at 3 or 6 dpi . The sharp spike in expression is not due to an error in data analysis since three control genes ( PITG_11766 , PITG_15117 , and PITG_01946 , which encode RpS3A , Actin A , and a RAS GTPase , respectively ) had similar expression in all samples in both Experiment One ( Fig 7 , panel C2 ) and Experiment Two ( Fig 7 , panel B2 ) . The genes were also expressed much more highly in infected potato leaves than tubers , indicating that the leaf-specific pattern of expression described in the preceding paragraph was due to variation in plant organs ( leaf vs . tuber ) and not species ( tomato vs . potato ) . This is shown in Fig 7D , where NR mRNA was measured during infection of the leaves of potato cultivar Atlantic and Russet tubers ( Experiment Four ) ; high expression was observed at 2 and 3 dpi in leaves , but little expression was observed in tubers at any timepoint . Higher expression in potato leaves was also observed in a comparison of isolate 1306 infecting leaves and tubers of Russet potato , where NR levels at 2 dpi were at least 50-fold higher in leaves in three biological replicates than tubers ( Experiment Five ) ; the mean Ct value for leaves was 33 . 3 , while no amplification ( Ct>40 ) was observed using RNA from tubers . Higher expression in potato leaves was also observed in a comparison of isolate 88069 infecting leaves of potato cultivar Bintje and tubers of cultivar Yukon Gold , where NR and NRT RNA levels were 15 . 4 and 7 . 3-fold higher in 2 dpi leaves than tubers , respectively . This experiment lacked biological replicates , but the results are consistent with the data from Experiment One , Experiment Two , and the two potato leaf-tuber comparisons ( Experiments Four , Five ) described in the prior paragraph . Interestingly , the peak of NR expression in the experiments performed on potato was 2 dpi , which was one day earlier than in tomato leaves . The difference may be related to the fact that tomato leaf infection by the isolate tested ( 1306 ) is relatively biotrophic with little necrosis or water-soaking , while necrosis and water-soaking are apparent by 2 dpi during its infection of potato leaflets . To help understand what physiological conditions regulate the three genes , their mRNAs were quantified by RT-qPCR in rye-sucrose media supplemented with 10 or 50 mM NO3− , levels similar to that measured earlier in leaves . It was not possible to test media with NO3− as the nitrogen source , since as noted in prior studies and verified by our laboratory for several isolates , such media does not support the growth of P . infestans [43] . We also tested rye-sucrose supplemented with 1 mM NH4+ , which resembles its concentration in leaves and tubers , and combinations of NO3− and NH4+ . Interestingly , NO3− failed to induce NRT , NR , or NiR and in general caused their mRNAs to decline in abundance ( Fig 8A ) . The failure of NO3− to induce NR was surprising since this contradicted results from another group [45] . However , their study used a different medium , the amino acid-based Henninger medium [44] . We therefore performed the experiment using that medium ( Fig 8B ) . In Henninger , NO3− caused a modest increase in NR mRNA , as opposed to the decrease seen in rye-sucrose media . This suggests that a complex balance of metabolites may regulate the gene cluster . We also considered other conditions that may influence expression of the nitrate assimilation cluster and explain the differences between leaves ( high expression ) and tubers ( low expression ) . First , we tested whether light affected NR expression . This is because the infected leaves had been incubated in a 12 hr light/dark cycle and the tubers in continuous darkness , in order to mimic the conditions of typical natural infections . Using rye-sucrose cultures grown in constant darkness , constant light , or a 12 hour light/dark cycle , we observed that in each case NR levels rose as cultures aged , peaking at about 4 days , about 24 hr after sporulation began ( Fig 9A ) . NR levels were lower in the cultures exposed to continuous light , which partially suppresses sporulation [46] . Indeed , in a prior study , we reported that NRT was induced 6-fold during sporulation in artificial media [47] . The possibility of a connection between NR induction and sporulation was , however , weakened by the consideration of additional evidence . First , NR mRNA was low in sporulating tuber infections ( i . e . the 4 and 6 dpi samples in Experiments One and Two ) . Second , NR mRNA fell 1–2 days before sporulation in infected potato and tomato leaves . Third , NR levels stayed low when P . infestans grew and sporulated in green pea media ( Fig 9B ) . Fourth , NR mRNA levels were similar in 5-day cultures grown at high humidity , which allows sporulation , and low humidity , which totally blocks sporulation ( Fig 9C ) . Fifth , NR mRNA was not higher in sporulating cultures of wild-type P . infestans than in strains silenced for Cdc14 , which blocks sporulation ( [48]; Fig 9D ) . Overall , the results are consistent with a model in which changing nutrient levels in cultures or plant infections regulate the NR gene cluster , and not sporulation itself . We also considered whether differences in porosity ( air content ) of leaves and tubers might be responsible for the higher level of NR mRNA in leaves . The porosity of tomato leaves and tubers are reported to be about 48% and 1% , respectively [49 , 50] . To assess if air content influenced NR expression , we compared 3-day cultures of P . infestans grown submerged in rye-sucrose broth and on the surface of rye-sucrose agar ( Fig 9E ) . Average NR levels were higher in the broth cultures , although the difference was not significant ( P = 0 . 27 ) and the relationship trend between air content and NR mRNA levels was opposite that seen in the plant material . To test the function of the nitrogen assimilation pathway , we expressed a sense copy of NR in stable transformants of P . infestans . Previous studies indicated that expressing sense , antisense , or hairpin RNAs can silence a target genes through a process that involves heterochromatinization of the target locus [51] . Three transformants were obtained which exhibited <10% of wild-type NR mRNA levels based on RT-qPCR ( Fig 10A ) . These were also no longer sensitive to chlorate ( Fig 10C ) . This compound is toxic to organisms with active nitrate reductases , which convert chlorate to the highly reactive molecule chlorite [52] . Curiously , NRT and NiR were also strongly down-regulated in those transformants . A likely explanation is that the chromatin alterations were regional and not limited to NR . To check the extent to which silencing had spread , we also tested the nearest flanking genes , PITG_13010 and PITG_13014 , which reside 4 and 11 kb from the cluster , respectively . These genes showed wild-type levels of expression in the NRT/NR/NiR-silenced strains ( Fig 10B ) . Each of the three silenced strains was unable to colonize tomato leaves in a detached leaflet assay . This is illustrated for Sil1 in Fig 11A . While wild-type P . infestans grew through the leaflet and sporulated by 6 dpi , the silenced strains either yielded no symptoms or a few necrotic lesions . Microscopic examination of leaves challenged with the silenced strains identified cysts forming appressoria on leaf surfaces and some hyphae on the surface of the leaf . However , few hyphae were seen spreading within the leaf tissue . In contrast , both wild-type and the silenced strains were able to complete their life cycles on tuber slices , with hyphae emerging on the surface and sporulating by 6 dpi ( Fig 11B ) . These observations were confirmed and extended by extracting DNA from plant tissue challenged with wild-type , Sil1 , and Sil2 , and quantifying the relative amount of P . infestans per gram of plant tissue by qPCR ( Fig 11C ) . On leaflets inoculated with the silenced strains , the amount of P . infestans DNA was <1000-times less than that measured with wild-type . On tubers , the silenced strains proliferated slightly less than wild-type . In contrast to their severe defect in leaf infection , the silenced transformants underwent the life cycle in a normal manner . For example , in experiments performed on rye-sucrose agar , they formed normal numbers of sporangia , which produced normal-looking zoospores . The zoospores also encysted with normal efficiencies , and produced the same number of appressoria as did wild-type . An experiment was performed on artificial media to test the hypothesis that the reason for the growth arrest of the silenced strains in leaves was that NO3− in that organ was toxic . On unamended rye-sucrose media , the growth rate of the silenced strains was identical to that of the wild-type progenitor , an empty vector transformant , and a GUS-expressing control ( Fig 12A ) . In contrast , the growth of the silenced strains was reduced by half when the media was amended with 50 mM NO3− , a concentration similar to that measured in leaves ( Fig 12B ) . Silencing NRT was not expected to reduce NO3− uptake dramatically since other transporters are still being expressed . These include the other MFS NO3− transporter mentioned earlier , PITG_09342 , as well as nine Chloride Channel Family proteins , which in other taxa are known to transport NO3− in addition to other anions [53] . That the silenced strains still acquired NO3− was confirmed by measuring that compound in hyphae . On rye sucrose media with 50 mM NO3− , the intracellular NO3− concentrations in wild-type and Sil1 were 16 mM and 33 mM , respectively . The lower level in wild-type was presumably due to the ability of its higher levels of NR and NiR to convert NO3− to NH4+ .
By mining P . infestans for genes encoding metabolite transporters and measuring their expression by RNA-seq , we observed that members of virtually all families exhibited dynamic changes in mRNA levels when growth on tubers , leaves , or artificial media varying in composition was compared . Transporters are known to be regulated by mechanisms that include nutrient limitation , substrate induction , and negative feedback [54–56]; such processes likely explain many of the patterns seen in P . infestans . The many amino acid transporters induced in leaves , for example , may reflect a response to nutrient limitation consistent with the lower levels of free amino acids that we observed in that tissue . The versatility of P . infestans as a pathogen of both leaves and tubers is reflected in the organ-specific expression patterns of many of its transporters , although some differences are not readily explained . For example , while AMT ( ammonium ) transporters are induced both during tuber and leaf infection , there is little overlap between those induced in the two organs . Such genes are possibly wired into regulatory networks that respond to multiple biosubstances or developmental cues . A link to development has been suggested for some transporter genes in fungi , where some family members have similar substrate specificities and Km values yet are expressed at distinct stages of the life cycle [57] , or have acquired additional roles in regulating metabolism [58] . Complex patterns of expression were observed within the majority of the transporter groups . Interpreting these is particularly challenging for families that have diverse substrates , such as MFS transporters . Bioinformatics has only limited utility in predicting the substrates of such transporters . Moreover , some of the proteins may participate in efflux in addition to uptake . That multiple factors regulate transporters was also evident from our detailed studies of NRT in P . infestans , along with NR and NiR . The strong induction of these three genes in leaves compared to tubers parallels the levels of NO3− in those organs . The low expression of the genes in the first few days of leaf infection may indicate that other nitrogen sources such as amino acids or ammonium are preferred , and that NO3− utilization only becomes important later . A corollary is that the low level of expression in tubers could result both from a dearth of NO3− and their higher levels of amino acids or ammonium compared to leaves . It should be noted that the coordinate expression of these three physically linked genes is atypical for P . infestans , as adjacent genes are expressed usually with independent patterns [59] . The inability of P . infestans to use NO3− as a sole nitrogen ( N ) source hindered our attempts to manipulate artificial media to test how NRT , NR , and NiR are regulated . In most other species , the genes are induced by NO3− and often repressed by more preferred nitrogen sources [60 , 61] . Depending on the media employed , the in P . infestans genes were either repressed slightly or induced by NO3− while NH4+ either had little effect or was slightly inhibitory . Based on those results and the patterns observed in planta , we propose that the gene cluster in P . infestans is regulated by the balance between several different nutrients , of which NO3− is only one . These need not be limited to nitrogenous compounds , considering that some NRs are known to be regulated by sucrose [62] . Regardless , that the P . infestans genes were not repressed strongly by NH4+ in all media was surprising , since this inhibits the orthologous genes in plants and filamentous fungi [60 , 61] . The contrast with fungi was surprising since the gene clusters of fungi and oomycetes are thought to have a shared ancestry [18] . The diversification of the regulatory schema may reflect the fact that unlike fungi , Phytophthora spp . lack a significant growth phase in soil , where NH4+ is more abundant than amino acids [12] . Other oomycetes , such as Pythium , do persist in soil as saprophytes [63] . Whether the genes are regulated similarly in Phytophthora and Pythium is an interesting question for future studies . In addition to our inability to grow P . infestans on media containing only NO3− as the nitrogen ( N ) source , further complexities in understanding the regulation of the gene cluster are that the developmental and nutritional status of P . infestans changes with time , and artificial media cultures and infected tissue of the same age may not be developmentally equivalent . The importance of making appropriate comparisons is illustrated by the failure of a prior study [64] to identify NRT , NR , and NiR as infection-induced . In that work , leaves from 2 to 5 dpi were compared to 12 day-old cultures in artificial media . Extrapolating from our rye-sucrose media timecourses where NR mRNA rose as cultures aged ( Fig 9A ) , the genes may have had high mRNA levels in the 12 day-old media and would thus not have been recognized as plant-induced . In our studies , the genes were plant-induced at every timepoint compared to the rye-sucrose control , but the degree of induction varied: 3 , 10 , 88 , and 10-fold at 2 , 3 , 4 , and 5 dpi , respectively . The choice of timepoint is therefore crucial . The inability of P . infestans to use NO3− as a sole nitrogen ( N ) source raises intriguing questions about the role of the pathway in oomycetes . It is not surprising that NO3− is an unfavored N-source , since its reduction to the level of an α-amino group requires ten electrons ( e . g . NAD[P]H ) plus one ATP , making it a less efficient substrate than NH4+ or amino acids . The pathway may nevertheless still benefit P . infestans when amino acids and NH4+ are limiting . Our gene silencing results also highlight a potentially valuable secondary role of the gene cluster in alleviating NO3− toxicity , either by using NRT to increase the efflux of NO3− or using NR and NiR to convert NO3− into ammonium . The ability of P . infestans to consume NO3− may also lessen levels of NO , which plants generate from NO3− and use to signal defense responses [65] . Particularly in commercial agriculture , fertilizer applications can raise NO3− within plants to high levels [19 , 20] . In potato and tomato production , fertilizers typically include various mixtures of NH4+ , NO3− , or urea , with the latter being converted by soil microbes into NH4+ . P . infestans may benefit under such circumstances from having NO3− as a supplementary N-source , but also needs protection against the deleterious effects of the compound , which include membrane and protein oxidation [22 , 23] . That fertilization affects the incidence of late blight is well known , and growers are cautioned against applying too much nitrogen to their fields . A predominant theory is that heavy fertilization promotes a dense canopy that favors the spread and survival of P . infestans spores [39–41] . However , fertilization was also reported to promote lesion expansion , which suggests that it directly boosts the growth of P . infestans [42] . Some other oomycete diseases are also believed to be stimulated by fertilization , as are some fungal diseases [66–68] . Whether all oomycetes respond similarly is unknown , as there is diversity in their nitrate assimilation pathways . Obligate biotrophs such as white rusts and downy mildews are unlikely to benefit directly from nitrate since they have lost the assimilatory gene cluster . On the other hand , some Phytophthora spp . can use nitrate as a sole N-source , which suggests that their assimilation pathways are more active than that of P . infestans [43] . This study contributes to a growing body of data about how plant pathogens adapt to growth on their hosts . Studies in fungi pathogenic to plants or animals identified transporters of sugars , amino acids , and other biosubstances that are infection-specific [4 , 69–72] and in the case of rusts , haustorium-specific [7] . We also found that many sugar and amino acid transporters in P . infestans display infection-specific patterns of expression , but it remains to be determined if any localize to haustoria . These most likely reside somewhere on the plasma membrane , where they can uptake host biosubstances . The transporters that respond to infection but are intracellular , such as members of the mitochondrial carrier family , would still contribute to the ability of P . infestans to exploit host nutrients by moving plant biosubstances into subcellular compartments where they can be metabolized . Determining the substrates of these and the other transporters may shed more insight into the nutrients preferred by P . infestans during in planta growth . Our work can also be related to prior reports from oomycetes . An RNA-seq study of P . infestans on tomato leaves detected SWEET transporter PITG_04999 , which in our data was expressed highly in leaves , particularly at 4 dpi [73] . A microarray analysis of Phytophthora parasitica on Arabidopsis thaliana roots identified six transporters induced by >4-fold during infection [74] . One was the ortholog of P . infestans folate-biopterin transporter PITG_07565 , which in our study was infection-induced but only in one tuber timepoint . That study also reported that the P . parasitica ortholog of AAAP protein PITG_17804 was root-induced compared to media , which contradicts our finding that it was expressed at lower levels in planta than media . Also , a microarray study on Phytophthora capsici on tomato [30] showed that the ortholog of folate-biopterin transporter PITG_01211 was induced in the biotrophic stage compared to media , while the P . infestans gene showed little change between growth conditions in our study . Differences between species or hosts are not surprising , but the likelihood that each study is biased by the type of media ( or culture age ) used for comparison must be recognized . Another outcome of this work relates to the application of functional genomics tools in oomycetes . DNA-directed RNAi is currently the most common strategy for silencing genes , and has been applied to multiple loci involved in pathogenesis and development . Our work here with the nitrate assimilation cluster has demonstrated that silencing can extend from the target locus to adjacent genes . That this might occur in P . infestans was proposed in a prior study that silenced a cluster of transcriptional regulators [51] . The phenomenon should not be surprising since silencing involves the modification of chromatin [51 , 75] and most oomycete genes reside within a few hundred bases of each other [59] . The ability of heterochromatinized domains to spread has also been observed in other systems [76 , 77] . Depending on one's perspective , this may be a blessing since multiple genes can be silenced coordinately , or a curse since it may obfuscate the connection between a targeted locus and the resulting phenotype . We suggest that future gene silencing studies include tests of genes that are physically close to the targeted locus .
P . infestans gene models were obtained from the database formerly maintained by the Broad Institute of Harvard and MIT , which is now accessible through Fungidb . org . Potential transporters were identified from the legacy annotations , from Fungidb . org based on InterPro annotations [78] , and by using those sequences as queries in command line BLAST searches against all P . infestans genes using an E value of 10-5 as a cutoff . Proteins were selected for the final transporter list ( after excluding ABC transporters and ion channels ) if searches using the Conserved Domain Database identified PFAM domains for transporters ( or TIGR domains for families lacking such a definition ) using an E value threshold of 10-5 , and if they also matched transporters in the TransportDB database [79] with the same E-value threshold . Py . ultimum var . ultimum and M . oryzae sequences were obtained from Fungidb . org and transporters identified through BLAST and domain searches as described above . Strain 1306 ( isolated from tomato ) was used for all analyses except for one study that used strain 88069 ( isolated from potato ) as described in Results . Cultures were maintained in the dark on rye-sucrose agar [31] at 18°C . Conditions for RNA analysis employed rye-sucrose agar , a defined minimal medium based on the recipe of Xu [32] , the latter with ( NH4 ) 2SO4 omitted and replaced by 1% casamino acids , or Henninger medium [44] . Some cultures were amended with KNO3 or ( NH4 ) 2SO4 as described in Results . Cultures for RNA-seq analysis were harvested at 2 . 5 to 3 days after inoculation , prior to sporulation . Germinated zoospore cysts of strain 1306 were harvested 6 hr after encystment as described [80] . Transformations of P . infestans were performed using the zoospore method [81] with G418 as a selective marker . The vector for gene silencing expressed the 2 . 8 kb open reading frame of NR . This was constructed by obtaining the gene by PCR from 1306 cDNA , and cloning the sequences into ClaI-SfiI sites of pSTORA [81] . Silenced strains were identified by RT-qPCR as described below , and were confirmed with a minimum of three biological replicates . Some experiments used transformants expressing pMCherryN [82] . Leaves and tubers infected with the latter were viewed by confocal microscopy using a water-dipping objective . Growth rate studies were performed by placing a 4 × 4 mm plug of inoculum at the edge of 100-mm rye-sucrose agar plate and measuring the colony radius every other day . Measurements of asexual sporulation , zoospore release , encystment , and appressorium development were performed as described [83] . Infections for RNA-seq analysis were performed using tomato plants ( cvs . New Yorker or Pieraline as described in Results ) , or potato plants and tubers ( cvs . Atlantic , Yukon Gold , Atlantic , or Russet Burbank , as described in Results ) . The plants were grown with a 12 hr light/dark cycle ( 25°C day , 350 μmol·m-2·s-1 fluorescent light; 18°C night ) for 4–5 weeks before infection . The soil for Experiments One , Two , Four , and Five contained an equal mix of peat moss , silica sand , and 16 kg/m3 of Ca ( H2PO4 ) 2 . H2O , KNO3 , and dolomite , while plants for Experiment Three were grown in a commercial mix comprised of bark , peat , sand , and fertilizer . For infections , whole plants were dipped in suspensions ( 104/ml ) of zoospores of P . infestans isolate 1306 , and incubated at 18°C with a 12 hr light/dark cycle with 115 μmol·m-2·s-1 illumination in a clear plastic bag to maintain high humidity . For tuber infections , slices were obtained with a mandoline slicer , washed in sterile water , cut into disks with a 1-cm diameter , dipped in zoospores as described above , placed on a metal rack , and incubated in sealed boxes at 18°C in the dark . Experiment One used 3 . 5-mm tuber slices while the later experiments used 3-mm slices . Disks from separate tubers or leaves from separate plants were used as biological replicates . Infection conditions for phenotyping transformants involved detached leaflet assays , in which the leaflets were laid on 0 . 8% water agar in a sealed box , or infections on 3 . 5-mm whole tuber slices which were prepared by soaking in 5% bleach for 15 min followed by a water rinse . The tuber infections used for characterizing transformants employed a 12 hr light/dark cycle . RNA was isolated using Sigma or Thermo kits for plant RNA . For RT-qPCR , the RNA was DNase-treated and cDNA synthesized using the SuperScript III ( Invitrogen ) or Maxima ( Thermo ) First-Strand RT-PCR kits . PCR was then performed using the primers shown in S3 Table , which were targeted to the 3' end of the genes . Primers were tested using a dilution series of template and accepted if efficiencies were above 94% . Amplifications were performed using a Bio-Rad iCycler IQ or CFX Connect system using the Dynamo SYBR Green qPCR kit ( Thermo ) with the following program: 95°C for 15 min , followed by 40 cycles of 94°C for 30 sec , 58°C for 30 sec , and 72°C for 30 sec . At the end of the run , melt curves were generated to evaluate the fidelity of amplification . Expression levels were calculated using the ΔΔCT method , using a constitutive gene ( ribosomal protein S3a , PITG_11766 ) as a control; prior studies demonstrated that this gene is expressed at similar levels during the life cycles of both P . infestans and P . parasitica [84 , 85] . Three technical and three biological replicates were used . RNA-seq was performed using indexed libraries prepared using the Illumina Truseq kit , which were sequenced on a Hiseq 2500 or Hiseq 3000 . Reads were aligned and mapped to P . infestans gene models using Bowtie version 2 . 2 . 5 and Tophat version 2 . 0 . 14 , allowing for 1 mismatch . Expression and differential expression calls were made with edgeR , using TMM normalization , a generalized linear model , and false discovery rate calculations based on the Benjamini-Hochberg method [86] . Data were trimmed to exclude unreliably-expressed genes using a RPKM threshold of 1 . 0 . Heatmaps were generated in R using Heatmap2 . The growth of P . infestans in tomato leaflets and tuber slices was measured by qPCR using O8-1 and O8-2 primers [87] . A minimum of five leaflets or tuber slices were pooled , weighed , ground under liquid nitrogen , and DNA isolated at 3 dpi using GeneJET Genomic DNA Purification Kit ( Thermo ) . qPCR was then performed using Sybr Green , with three technical replicates . The relative amounts of P . infestans DNA per gram of plant tissue were then calculated from the resulting Ct values . For calculations of nitrate , ammonium , and soluble sugars from plant and artificial media samples , materials were weighed , lyophilized , ground into a fine powder using an electric mill , and provided to the Analytical Lab of the University of California , Davis for analysis . Nitrate and ammonium were assayed using a diffusion-conductivity method followed by conductivity detection [88] . Free amino acids were analyzed in-house from similar materials using a ninhydrin assay method against extracts made using 10 mM HCl and with glycine as a standard , after eliminating proteins by sodium tungstate precipitation [89 , 90] . | Little is known of how plant pathogens adapt to different growth conditions and host tissues . To understand the interaction between the filamentous eukaryotic microbe Phytophthora infestans and its potato and tomato hosts , we mined the genome for genes encoding proteins involved in nutrient uptake and measured their expression in leaves , tubers , and three artificial media . We observed dynamic changes between the growth conditions , and identified transporters expressed mainly in the biotrophic stage , leaves , tubers , or artificial media . When we blocked the expression of a nitrate transporter and two other genes involved in assimilating nitrate , we observed that those genes were required for successful colonization of nitrate-rich leaves but not nitrate-poor tissues , and that nitrate had become toxic to the silenced strains . We therefore hypothesize that the nitrate assimilation pathway may help the pathogen use inorganic nitrogen for nutrition and/or detoxify nitrate when its levels may become damaging . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"plant",
"anatomy",
"oomycetes",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"microbiology",
"nitrates",
"fungi",
"plant",
"science",
"nutrition",
"crops",
"potato",
"plants",
"vegetables",
"crop",
"science",
"gene",
"expression",
"chemistry",
"tomatoes",
"leaves",
"fruits",
"agriculture",
"tubers",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"nutrients",
"organisms",
"solanum"
] | 2016 | Gene Expression and Silencing Studies in Phytophthora infestans Reveal Infection-Specific Nutrient Transporters and a Role for the Nitrate Reductase Pathway in Plant Pathogenesis |
Because of the development of resistance in trypanosomes to trypanocidal drugs , the livelihood of millions of livestock keepers in sub-Saharan Africa is threatened now more than ever . The existing compounds have become virtually useless and pharmaceutical companies are not keen on investing in the development of new trypanocides . We may have found a breakthrough in the treatment of resistant trypanosomal infections , through the combination of the trypanocide isometamidium chloride ( ISM ) with two affordable veterinary antibiotics . In a first experiment , groups of mice were inoculated with Trypanosoma congolense strains resistant to ISM and either left untreated or treated with ( i ) tetracycline , ( ii ) ISM or ( iii ) the combination of the antibiotic and the trypanocide . Survival analysis showed that there was a significant effect of treatment and resistance to treatment on the survival time . The groups treated with ISM ( with or without antibiotic ) survived significantly longer than the groups that were not treated with ISM ( P<0 . 01 ) . The group treated with the combination trypanocide/antibiotic survived significantly longer than the group treated with ISM ( P<0 . 01 ) . In a second experiment , groups of cattle were inoculated with the same resistant trypanosome strain and treated with ( i ) ISM , ( ii ) ISM associated with oxytetracycline or ( iii ) ISM associated with enrofloxacine . All animals treated with ISM became parasitaemic . In the groups treated with ISM-oxytetracycline and ISM-enrofloxacine , 50% of the animals were cured . Animals from the groups treated with a combination trypanocide/antibiotic presented a significantly longer prepatent period than animals treated with ISM ( p<0 . 001 ) . The impact of the disease on the haematocrit was low in all ISM treated groups . Yet , it was lower in the groups treated with the combination trypanocide/antibiotic ( p<0 . 01 ) . After optimization of the administration protocol , this new therapeutic combination could constitute a promising treatment for livestock infected with drug resistant T . congolense .
African Animal Trypanosomiasis affects about 10 million km2 of sub-Saharan Africa and is a primary cause of rural poverty and food insecurity as explicitly recognized by the African Union , FAO and others [1] . Tsetse and the disease they transmit will continue to be a considerable threat to livestock and rural development [2] . Over the years , a large arsenal of vector control tools has been developed but they are difficult to sustain at the smallholder level . Hence , the control of animal trypanosomiasis ( mainly Trypanosoma congolense ) and zoonotic Human African Trypanosomiasis ( mainly T . brucei rhodesiense ) in poor rural communities has and will continue to rely heavily on the use of trypanocidal drugs . However , the development of trypanocidal drug resistance in T . congolense was reported in 17 countries of sub-Saharan Africa [3] and is becoming a huge threat for the cattle breeders in many regions . On the Adamaoua plateau in Cameroon , for example , up to 100% of the tested trypanosome isolates were found resistant to isometamidium chloride ( ISM ) and to diminazene aceturate ( DA ) leaving farmers helpless [4] . Unfortunately , no new drug is expected to be available in the near future and resistance is spreading very rapidly . For example , a five fold increase in DA resistance within a seven years interval was observed in the Eastern Province of Zambia [5] . Hence , alternatives are urgently needed to circumvent trypanocidal drug resistance . Reversal of drug resistance or chemosensitization was successfully achieved , among others , in yeast [6] , Plasmodium [7] , [8] , cancer cells [9] and Leishmania [10] . Such strategies could bring a much needed relief to African livestock breeders if they could be implemented at a reasonable price by shortcutting the development of new compound , toxicity studies and long clinical trials . Many bacterial secondary multidrug resistance transporters belonging to the two major families , i . e . the Major Facilitator Superfamily ( MFS ) and the Multi Antimicrobial Extrusion Family ( MatE ) are described as having affinity for ethidium bromide ( Homidium ) as well as for many different compounds such as plant alkaloids , noxious metabolic products ( such as fatty acids or bile salts ) , organic solvents and diverse antibiotics [11] . At least eight representatives of those transporters families are present in the genome of T . congolense . Homidium is part of the ISM molecule , the structural relatedness of both molecules being thus obvious ( Figure 1 ) . Furthermore , in the field , cross-resistance is observed between ethidium bromide and ISM [12] suggesting that uptake and extrusion of the drug within and from the trypanosome are mediated by the same mechanisms for both compounds . Hence , our working hypothesis is that chemical compounds could interfere ( compete ) with the extrusion of ISM from the drug resistant trypanosome allowing a prolonged trypanocidal action . The objective of this work was to bring some indirect evidence confirming this working hypothesis . Preliminary experiments conducted in vitro would have allowed a more precise definition of the role of those secondary transporters in trypanocidal drug resistance but research in this domain is hampered by the fact that except for some atypical laboratory strains , bloodstreams forms of T . congolense do not grow properly in vitro [13]–[15] . Seeking a commercially available chemical compound that could be used for treating livestock , a number of antibiotics were selected and screened in a mouse model . The criterion for inclusion in this study was the affinity of the medications for bacterial efflux systems as described for β-lactams [16] , [17] , tetracycline ( TC ) , oxytetracycline ( OTC ) [18] , nalidixic acid ( quinolone ) [19] and the fluoroquinolone enrofloxacine ( FQE ) [20] . The nalidixic acid and β-lactam ( Penicilline G ) were rejected after preliminary experiments ( see table 1 ) , no difference being observed between groups treated with ISM alone or treated with ISM associated with one of the two compounds . The absence of curative effects of the antibiotics used alone was consistently checked in a mouse model ( see table 1 ) . After this preliminary screening , OTC was selected for the experiment in cattle as it is available as an injectable long acting form allowing for a reduction of the number of injections . For the experiment in mouse , TC was chosen as the easiest and cheapest commercial preparation for oral administration by dilution in drinking water . Enrofloxacin was not pre-tested in combination with ISM in mice but immediately used in cattle .
This is to certify that the experiments carried out at the Institute of Tropical Medicine in the framework of the hereunder mentioned study were approved by the Ethics Committee of the Institute of Tropical Medicine and that the study was conducted adhering to the institutional guidelines for animal husbandry . In Belgium protection of experimental animals is regulated by the Royal Decision of 14/11/1993 . Article 3bis paragraph 1 of this Royal Decision stipulates that: Every laboratory that keeps vertebrates with a view to perform experiments that may cause pain , suffering or lesions , has to establish an Ethics Committee . The Ethics Committee is composed of at least 6 members . The laboratory director or his representative , the leaders of the experiments , some laboratory assistants and the veterinary surgeon or the expert charged with the supervision of the health and the well-being of the animals are part of the Ethics Committee . Moreover one or more independent members , not belonging to the laboratory staff , will be member of the Committee . A veterinary inspector of the Ministry of Agriculture will also have a seat on the Ethics Committee . Identification of the experiment: DG008-VD-M-Tryp Title of the project: Study on the genetic basis and improved detection methods of resistance against isometamidium and diminazene in animal trypanosomes . Date of reception of the application: 03/11/2008 Date of approval by the Ethical Commission: 23/12/2008 ( extension of a similar application DG006-VD-M-Tryp approved in 2004 ) Validity of this approval: from 23/12/2008 until 22/12/2012 . The cloned T . congolense savannah type strain IL3343 was identified as resistant to ISM when tested in mice ( CD50 = 1 . 7 mg/kg ) [21] with the CD50 defined as the curative dose that gives complete cure in 50% of the animals . The T . congolense savannah type strain TRT57C10 was isolated from cattle in Eastern Zambia in 1996 , cloned and conserved as a stabilate in liquid nitrogen . It was identified as highly resistant to ISM when tested in mice . Three doses of ISM , i . e . 0 . 1 , 1 and 10mg/kg were tested according to the protocol described by Eisler et al . [22] . When treated with 10mg/kg ISM , 100% of the mice relapsed ( CD50>10mg/kg ) . The stabilates of the cloned isolates were reactivated in mice . When the parasitaemia reached 8 on Herbert and Lumsden's scale [23] , blood was collected under terminal anaesthesia by heart puncture and 4 groups of 16 adult OF1 mice weighing on average 30g each were inoculated with one of the two trypanosome clones ( 5*105 trypanosomes/mouse through intraperitoneal injection ) . Twenty four hours after inoculation and for each clone , group 1 was left untreated and served as control , group 2 was treated for 30 days , per os , with 125mg/kg/day tetracycline , group 3 was treated with 1mg/kg ISM injected once intraperitoneally and finally , group 4 was treated with 1mg/kg ISM injected once intraperitoneally and was treated per os with 125mg/kg/day TC for 30 days . Mean water consumption of the mice was determined before and during the experiment and was on average of 3ml/day/mouse at 18°C . This water intake was not affected by the presence of TC in the drinking water . All mice were monitored three times a week for survival and presence of trypanosomes by microscopic examination of a wet film made from fresh blood sampled from the tail of each mouse for a period of 140 days . Mice were euthanized when their health status , determined by clinical examination , was deteriorating ( prostration , lateral decubitus , hyperventilation , unconsciousness and/or PCV≤20 ) . At day 140 , all surviving mice were euthanized and between 1 . 5 and 2ml blood was collected . The DNA of the whole blood sample was then extracted using a routine phenol–chloroform–isoamyl alcohol method [24] . To confirm the presence or absence of trypanosomes , the PCR technique on the 18S small subunit of the ribosomal DNA ( Ssu-rDNA ) was used [25] , [26] . 5Three groups of 6 adult crossbred zebus weighing on average 158 kg each ( extremes 140 and 201kg ) were inoculated with 5×105 trypanosomes ( cloned isolate IL3343 ) each by intra-jugular injection 30 days after treatment with DA ( 7mg/kg ) to clear all trypanosomal infections and deworming . One non-treated control group of 2 cattle was inoculated in the same way . The 20 cattle were housed in fly-proof facilities . From day 7 after the inoculation , all animals were monitored 2 times a week during 95 days . Their PCV was measured and jugular blood was examined for the presence of parasites by microscopic examination of the buffy coats and by PCR [25] performed on buffy coats collected on on Whatman 4 filter paper ( Whatman ) . The DNA was obtained using a routine chelex-based extraction method [24] . At the first parasitaemia , group A was treated with one single administration of 0 . 5mg/kg ISM by intramuscular ( IM ) injection , group B with one single administration of 0 . 5mg/kg ISM and with 20mg/kg OTC ( Terramycin LA ) IM every 3 days for 30 days and group C with one single administration of 0 . 5mg/kg ISM and with 5mg/kg FQE ( Baytril 100 ) IM every 2 days for 30 days . For each animal , the injection sites of the drugs were alternatively selected in forehand and hindquarters , shaved and coloured with methylene blue and picric acid for OTC and FQE respectively . A minimal distance of 6 cm between injection sites was respected . The survival of the mice in the 8 groups and of the cattle in the three groups was analysed in two separate survival models in Stata 10 ( Copyright 1996–2009 StataCorp LP ) using groups as an explanatory variable . A log-normal distribution was used in a parametric model ( Text S1 ) . The start of the model corresponded to the day of inoculation and the experiment was short enough to ignore natural mortality of animals . The cattle's PCV values were analyzed using a cross-sectional linear regression , accounting for repeated measures from individual animals . Explanatory variables were the animal groups , post-treatment periods and the interactions between them . Three post-treatment periods each containing the same number of samplings were defined as follows: day 1–21 , day 22–54 and day 55–98 . The interaction term between the groups and the third period ( using the first period as a baseline ) was used as indicator of the impact of the disease on the PCV .
A significant effect of treatment and resistance to treatment on the survival time of the mice was observed . The data are summarized in table 2 . Groups 3 and 4 survived significantly longer than group 1 ( control without treatment; P<0 . 01 ) , unlike group 2 ( received only TC as treatment; P>0 . 1 ) . The longer survival time of the mice treated with ISM with or without potentiator is confirming our field observations that even when trypanocidal drug resistance is present , ISM seems to impair the development of the parasite , reducing the impact of the disease on the health of the infected animal . For both strains , resistant ( IL3343 ) and highly resistant ( TRT57C10 ) , there was a significant difference between groups 3 ( ISM treatment only ) and 4 ( treated with ISM and TC; P<0 . 01 ) ( Figure 2 and Figure 3 ) . When considering the efficacy of the compounds against the trypanosomes , the complete ineffectiveness of TC alone and the increased efficacy of ISM in presence of TC , provides strong arguments in favor of the hypothesis that the two compounds compete for the same efflux system . Despite the unusual high degree of resistance of T . congolense TRT57C10 , the survival times were significantly higher after treatment with the association of ISM and TC . One mouse survived the infection for 140 days . Such a high survival time was never observed before in laboratory experiments using this strain . Furthermore , the PCR analysis of the blood sample at day 140 was negative suggesting that the trypanosomal infection was cleared completely . Moreover , the blood sample used for diagnosis was between 1 . 5 and 2ml of blood representing the average total amount of blood that can be collected from a mouse . Since the sensitivity of the diagnostic method is 25 trypanosomes/ml [25] , the complete absence of trypanosomes and thus , the complete clearance of the parasites from the host can reasonably be assumed . For the resistant strain ( IL3343 ) , in group 3 , among the 13 surviving mice , 3/13 were microscopically positive for the presence of trypanosomes and 6/13 were positive for the presence of trypanosomes by PCR . In group 4 , among the 15 surviving mice , 1/15 was microscopically positive for the presence of trypanosomes and 3/15 were positive for the presence of trypanosomes by PCR . The two untreated control animals became parasitaemic 11 days after inoculation and were treated with DA ( 7 mg/kg ) on day 30 because their PCV reached the critical value of 25 . All 6 animals of group A ( ISM ) became positive between days 24 and 46 post-inoculation . The data are summarized in table 3 . When ISM was used in combination with either OTC ( group B ) or FQE ( group C ) , the prepatent period was significantly longer ( p<0 . 001; Figure 4 ) . 50% of the cattle became infected ( between days 46 and 82 ) and 50% completely cleared the infection . In the groups B ( ISM-OTC ) and C ( ISM-FQE ) , the parasitaemia remained very low , below the detection level of the microscopic examination , i . e . 450 trypanosomes/ml [27] . The PCR results were fluctuating with animals being detected parasitaemic every 2 to 3 weeks , indicating a parasitaemia oscillating just above and below the detection limit of the PCR test , i . e . 25 trypanosomes/ml blood [25] . The impact of the infection on the PCV was not very pronounced , even in group A ( average PCV reduction 8 to 14 weeks after treatment: 5 . 9%; 95% CI: 4 . 5–7 . 3 ) . However , this impact was lower in groups B ( ISM-OTC ) and C ( ISM-FQE ) compared to group A ( ISM ) ( p<0 . 01 ) . These observations indicate that even in the case of ISM-resistant trypanosomes , farmers still seem to benefit from the use of the trypanocide because of the significant decrease of the effect of the infection on the health status of the animals as represented in the PCV values . Although resistance to DA and ISM , is developing quickly [5] , [28] , [29] , controlling the parasite in livestock using drugs remains the control method of choice for small-scale livestock breeders . Localised tsetse control is usually not effective [30] and a vaccine is not yet available , leaving little choice to control the disease . Trypanosomiasis not only affects livestock production ( milk , meat ) but also impacts greatly on crop production through the inability to keep draft animals in tsetse-infested areas [31] . Notwithstanding the alarming levels of trypanocidal drug resistance that have been reported in the cotton belt of West Africa [32] and in some regions of southern Africa ( including Zambia ) [4] , [5] , new trypanocidal drugs for animal use are not expected to become available in the near future . Pharmaceutical companies do not invest in research and development of new veterinarian trypanocidal compounds for a too specific , limited African market with poor benefit perspectives [33] . Hence , potentiating the available trypanocidal drugs may represent a powerful alternative to the current problems associated with the control of trypanosomes in livestock . Research in the field of non-competitive inhibitors of efflux pumps in bacteria is being conducted [34]–[36] and may ultimately represent an immense hope for future control of trypanosomiasis using drugs . In the meantime , TC and some derivatives are cheap drugs , registered for use in livestock , widely available on the African market and with an expired patent , now in the public domain . More importantly , TC is commonly used by African farmers and will not require elaborate new chemistry and safety tests . Hence , assuming that further trials confirm the effectiveness of the antibiotics in potentiating the activity of trypanocidal drugs in cattle under natural tsetse challenge , the new control approach can be implemented rapidly . It is likely that the combination ISM–TC/OTC can also be made more cost effective after adjusting dosage and the duration of the treatment . Furthermore , several analogues of TC/OTC and FQE are available albeit somewhat more expensive as patents are still in force . These compounds are currently being screened with the aim of optimizing the delivery system to increase the specificity of the treatment , to boost the intracellular concentration of the chemosensitizer within the trypanosome and to reduce the dose . Obviously , the current treatment schedule cannot be used under field conditions . The repeated administration of a high dose of antibiotics is far too expensive for the rural communities and would certainly render the treated animals unsuitable for human consumption . Further research is thus ongoing to identify the best galenic solution , the optimal combination of chemosensitizer with ISM ( qualitative and quantitative ) and to test this combination in livestock under controlled and field conditions in areas with high tsetse challenge and high trypanocidal drug resistance . An effective combination of ISM and chemosensitizer ( s ) should result in ( i ) a decrease in the proportion of circulating strains resistant to ISM and ( ii ) a decrease in the impact of the disease on the health status of the cattle . Strategic use of this approach may result in an increased efficacy of currently available trypanocidal drugs in extensive areas of sub-Saharan Africa where their use is severely curtailed as a result of the development of resistance in trypanosomes . | African Animal Trypanosomiasis causes the death of 3 million head of cattle each year . The annual economic losses as a result of the disease are estimated to be 4 . 5 billion US dollars . Trypanosomes are transmitted by tsetse flies and can infect a wide range of hosts from wildlife to domestic animals . This study is dealing with Trypanosoma congolense , which is one of the very prevalent parasites affecting livestock of poor African rural communities , decreasing the milk and meat production but also reducing the fitness of cattle that is used as draught power . Infected animals can only be treated by three compounds , i . e . , diminazene , isometamidium and ethidium . These three products have been in use for more than a half century and it is thus not surprising to observe treatment failures . In some areas , the trypanosomes circulating have developed resistance to the three drugs leaving the farmers with no further options . As pharmaceutical companies are not keen on investing efforts and money in the development of new veterinary drugs for this low-budget market , our idea was to render an old ineffective drug effective again by combining it with existing potentiating compounds that are available and affordable for the livestock keeper . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
] | 2010 | Chemosensitization of Trypanosoma congolense Strains Resistant to Isometamidium Chloride by Tetracyclines and Enrofloxacin |
Transmembrane channel proteins play pivotal roles in maintaining the homeostasis and responsiveness of cells and the cross-membrane electrochemical gradient by mediating the transport of ions and molecules through biological membranes . Therefore , computational methods which , given a set of 3D coordinates , can automatically identify and describe channels in transmembrane proteins are key tools to provide insights into how they function . Herein we present PoreWalker , a fully automated method , which detects and fully characterises channels in transmembrane proteins from their 3D structures . A stepwise procedure is followed in which the pore centre and pore axis are first identified and optimised using geometric criteria , and then the biggest and longest cavity through the channel is detected . Finally , pore features , including diameter profiles , pore-lining residues , size , shape and regularity of the pore are calculated , providing a quantitative and visual characterization of the channel . To illustrate the use of this tool , the method was applied to several structures of transmembrane channel proteins and was able to identify shape/size/residue features representative of specific channel families . The software is available as a web-based resource at http://www . ebi . ac . uk/thornton-srv/software/PoreWalker/ .
Transmembrane channel proteins play pivotal roles in maintaining the homeostasis and responsiveness of cells and the cross-membrane electrochemical gradient by mediating the transport of ions and molecules through biological membranes [1] . For instance , aquaporins facilitate the flux of water and small uncharged solutes across cellular membranes and , in humans , are involved in several diverse functions , like concentrating urine in kidneys and participating in forming biological fluids [2]–[5] . In contrast , potassium channels are fundamental regulators of cell membrane potential and control the action potential waveform and the secretion of hormones and neurotransmitters [6]–[8] . Moreover , a family of transmembrane proteins , known as translocons , have been found to mediate protein transfers between different cellular compartments and consequently to be involved in the folding of membrane and secretory proteins [9] . Understanding the structure and function of transmembrane channel proteins and studying their properties and biochemical mechanisms is therefore a very important task in biological and pharmaceutical research [10] , [11] . Transmembrane channel proteins usually show a cavity spanning the whole protein , herein designated as the pore , which forms the path used by ions and/or molecules to cross the membrane . The pore has two openings , one on each side of the membrane , and it has been hypothesized ( and in some cases shown ) that the specificity and selectivity to different solutes is strongly dependent on the particular structural or amino acid composition features of the channel [5] , [8] , [12] . Consequently , computational methods for the identification and description of pores in transmembrane protein 3D-structures represent key tools to gain insights into how these proteins function . To our knowledge , several methods for the analysis of protein surface and cavities have been developed [13]–[19] but the only currently available method for the structural analysis and visualisation of transmembrane channels is HOLE , developed in 1993 and still widely used [20] , [21] . This elegant algorithm implements a Monte Carlo simulated annealing approach to find the path that a sphere of variable radius can use to go through a channel and also provides pore anisotropy analysis and conductance prediction tools . The path is optimised so that it can be considered as the route of a plastic sphere squeezing through the channel , i . e . at each point of the path the channel can accommodate the largest possible sphere . Three more recent methods , developed for the detection of internal cavities and tunnels in any protein structure , CAVER [22] , its improved version MOLE [23] and MolAxis [24] can be applied to identify pores in transmembrane proteins . CAVER explores the protein inner space using a grid-based approach , while MOLE implements an algorithm based on Voronoi polyhedra . Both approaches use an optimality criterion based on the minimization of a cost function , which depends on reciprocal atomic distances , and calculates the optimal way out from a user-specified starting point inside the protein to the outside environment . MolAxis exploits computational geometry techniques , in particular the alpha shapes theory and the medial axis concept , to detect possible routes that small molecules or ions can take to pass through channels and cavities . It is worth highlighting here that all the four programs , to be applied to transmembrane proteins , require user-defined specific information about the geometry of the channel that necessitate a fairly good knowledge of the location of the pore and/or of key residues lining the pore walls , like a starting point for the path search through the channel or a vector approximating the location of the pore within the protein 3D-structure . Moreover , they provide only a limited description of the channel geometry mainly consisting of diameter values and some of the residues lining the pore walls . Herein we present PoreWalker , a method to provide a detailed description of the three dimensional geometry of a channel ( or pore ) through a transmembrane protein , given the coordinates of the protein structure . These 3D pore descriptors provide a quantitative description , including the size , shape and regularity of the pore , which we hope will help to explain pore specificity , the critical biological function of these molecules . PoreWalker is fully automated , requiring only the 3D protein coordinates from the PDB file , and so can be applied to any new structure or across all transmembrane proteins in the PDB . The method was applied to several structures of transmembrane channel proteins and was able to identify shape/size/residue features representative of specific channel families . The software is implemented as a web-based resource at http://www . ebi . ac . uk/thornton-srv/software/PoreWalker/ and its source codes will soon be available upon request to the authors .
In transmembrane proteins , the channel runs approximately perpendicular to the membrane plane and parallel to the bundle or barrel that makes up the transmembrane portion of the pore . The first step of the program consists in re-orienting the protein structure so that the origin lies at the centre of gravity of the transmembrane portion of the protein and the bundle/barrel lies perpendicular to the membrane plane . The main axis of the transmembrane bundle/barrel is calculated according to the position of the secondary structure elements that putatively form it . Each secondary structure element in the protein is identified from the separation of sequential C-alphas as described in Supplementary Text S1 and , if the helix or the strand is longer than 15 or 10 amino acids , respectively , it is approximated by a vector , which starts at its centre of mass and points toward the centre of mass of the terminal four and two amino acids of the helix or strand , respectively . The length threshold was applied because , on average , transmembrane helices and strands used for this calculation need to be sufficiently long to cross the membrane . This excludes small helices which often do not lie perpendicular to the membrane plane . The sign of all the vectors is selected so that they point in approximately the same direction and the averaged vector is calculated . However , outlying secondary structures found to be more perpendicular than parallel to the bundle/barrel axis are excluded from the averaging at this stage so that the transmembrane portion of the structure is orientated as parallel to the membrane axis as possible . The whole protein 3D structure is then re-oriented so that its calculated main axis overlaps with the x-axis of the current 3D system and the centre of gravity of its transmembrane portion lies at the origin . In this way , the structure is moved into a new reference frame that approximately aligns the transmembrane secondary structure elements perpendicular to the membrane . The pore axis is then approximated as coinciding with the protein main axis ( see Figure 2 , step 2 ) . This starting assumption , despite its crudeness , simplifies and speeds up the following steps of the method . The centre of the pore is defined by iteratively maximising the number of detected putative pore-lining residues , i . e . water-accessible amino acids pointing towards the pore axis . At the beginning of the process , the centre of the pore and the pore axis , i . e . the linear vector going through the middle of the pore , are assumed to correspond to the centre of mass of the protein and to the x-axis , respectively . Putative pore-lining amino acids around the pore axis are then selected to satisfy three criteria: ( 1 ) the relative sidechain solvent accessibility calculated by NACCESS ( [25] , downloadable at http://www . bioinf . manchester . ac . uk/naccess/ ) must be higher than 5%; ( 2 ) the vector defined by the C-alpha-C-beta bond must point towards the pore axis; and ( 3 ) the distance of the C-alpha atom from the pore axis must be below a given threshold . The distance threshold is calculated at each iteration as the smallest distance between any pore-lining residue C-alpha and the current pore axis plus 6 Å . This prevents the inclusion of “second shell” residues in the selection of putative pore-lining residues and in the calculation of the final centre of the pore . Glycines lack C-betas and are therefore treated differently . For each Gly , a dummy atom is defined as the average of 3D-coordinates of its backbone carbonyl carbon and amide nitrogen . This atom can be considered a mirror image of the C-betas of a virtual side chain located between the two hydrogen atoms bound to its C-alphas and can therefore be used to evaluate the orientation of Gly backbone atoms . Glycines with a total relative accessibility higher than 5% and with the dummy atom pointing away from the pore axis are defined as pore-lining . A new centre of the pore is then calculated from the selected putative pore-lining amino acids and the protein structure is translated so that the new pore centre and the x-axis corresponds to the origin of the 3D-system and to the new pore vector , respectively . The above procedure is performed iteratively and stops when the number of newly selected putative pore-lining residues converges to its maximum , indicating that the pore centre has reached its optimal position . As a result of this first process , the protein structure is translated in space so that the x-axis goes through the current best-guess of the centre of the pore and a preliminary list of putative pore-lining residues is generated ( see Figure 2 , step 3 ) . The effectiveness of this step of the method was assessed by monitoring the distance of the selected pore-lining residues from the pore centre , as described in Supplementary Text S1 and shown in Figure S1 . To derive the best possible axis and cavity of the pore an iterative slice-based approach is used , in which the centre of the pore is systematically optimised for each slice and therefore eventual irregularities in the cavity can be detected . At each iteration , the protein structure is mapped onto a 3D-grid of 1Å steps and then sliced along the x-axis ( i . e . the current pore axis ) in 1Å thick layers . The pore centre of each slice is then identified by a grid-based approach so that it lies at the centre of the sphere with the maximum radius that the slice can accommodate . The maximum sphere and its centre are derived by expanding the sphere from the current centre until it clashes with a pore-lining atom , and systematically shifting the centre on the vertices of a 2D-grid so that the centre of the sphere of maximum volume for that slice can be identified . The pore centre of the slice is initially set as the average of C-alpha and C-beta atoms of the putative pore-lining amino acids belonging to the slice selected in the previous step of the program , and the corresponding maximum sphere is calculated . A square 2D-grid perpendicular to the current pore axis ( x-axis ) is then built and used to optimize the location of the pore centre . The grid has 0 . 1 Å squares , it is centred at the pore centre , and its size depends on the sequence length of the protein and on the size of the pore . Grid vertices not surrounded by atoms in all the possible y and z directions are taken as located outside the pore and excluded from the optimization process . The sphere of maximum volume at a given centre is calculated by increasing its radius by 0 . 1 Å until it hits a vertex of the 3D-grid occupied by a backbone or C-beta atom . The current sphere radius is adjusted by subtraction of the atomic van der Waals radius ( 1 . 8 Å , corresponding to the average radius of all types of heavy atoms found in protein structures as in the AMBER united force field [26] ) or approximate residue side chain radius ( as in Levitt's amino acid ‘lollypop model’ [27] ) if a backbone atom or a C-beta is hit , respectively . If the radius value is above any previously calculated radius , the current radius and corresponding sphere centre are taken as the maximum radius and pore centre for that slice . At the end of the iteration , coordinates of the last four consecutive sphere centres at each end of the pore , that represent the two pore openings , are averaged to generate two points , which define the new pore axis . The structure is then re-oriented to align with the new vector ( see Figure 2 , steps 4–8 ) . The last four consecutive spheres are used because the ends of the channels can be very irregular in term of shapes and therefore pore axes derived from the two very last sphere centres ( one per end ) often do not cross correctly one or both the pore entrances ( the value 4 was derived on a trial-and-error basis in the range of values from 1 to 5 ) . The refinement process stops when the new pore vector “overlaps” to the old pore axis ( i . e . when their angle is lower than 0 . 5 degrees ) and the current pore axis and maximum sphere radii ( i . e . those calculated in the previous iteration ) are retained as optimal and used in the further analysis of the pore shape . The last step of the method is the analysis and calculation of three main pore descriptors: the pore-lining atoms and residues ( Section 4 . 1 ) , and the shape of the pore cavity ( Section 4 . 2 ) and its regularity ( Section 4 . 3 ) .
Pore descriptors calculated by PoreWalker for a submitted structure are summarised in the corresponding output webpage , which shows the features of the channel cavity and several visualizations of the pore based on the identified pore-lining residues . As an example , the output of the bovine aquaporin-1 ( PDB code 1j4n ) is summarised in Figure 3 . The 3D shape of the pore is simplified in 2D as a stack of building blocks shaped as trapezia for funnel-like shapes ( Figure 3B ) going from the most negative to the most positive coordinate along the pore axis . In addition , the pore cavity is represented as a series of consecutive straight and wiggly lines representing channel areas where pore centres can ( straight ) or cannot ( wiggly ) be fitted to a line , respectively ( Figure 3E ) . It is worth highlighting here that the approach does not take into account any chemistry ( e . g . H-bonds ) but just calculates the path of the pore centres . In practice , ions/molecules may well hop between low energy off-centre sites , within the channel , that optimize their interactions with pore residues during their passage through the channel . Vertical and horizontal visualizations of the pore help to provide a better understanding of the channel features . Vertical sections ( Figure 3A , D ) are generated halving the protein structure along the pore axis , while horizontal sections ( Figure 3G , I ) are produced as 5Å slices of the protein structure perpendicular to the pore axis . Amino acids are coloured according to whether they are classified as pore-lining and red spheres represent optimal pore centres . PoreWalker was tested on the 19 structures from the “Membrane Proteins of Known 3D Structure” resource ( http://blanco . biomol . uci . edu/Membrane_Proteins_xtal . html ) listed in Table 1 , that include both ion and small molecule channels with straight and curve pores . Results are shown in Table 1 , Figure 4 and Supplementary Figure S2 . Although there is no fully comprehensive experimental data to assign with certainty the location and residue composition of channels in transmembrane protein 3D-structures , the position of the pore axis and of the pore centres , visually analysed in relation to the protein structure , and the minimum diameter value give a hint of the effectiveness of the method . From visual inspection , PoreWalker seems able to locate correctly the pore axis and the pore centres in most of the cases and therefore to identify fairly correctly the amino acids that line the pore walls with one or more atoms . In fact , the pore axis seems wrongly located only for the Amt-B ( Figure 4K ) , Amt-1 ( Supplementary Figure S2E ) and the SecYE-beta translocon ( Figure 4H ) channels ( PDB codes 1xqf , 2b2f and 2yxr , respectively ) . Both Amt-B and Amt-1 channels share a common hour-glassed shape with multiple exits at one of the pore gates and can therefore be thought to include more than one transmembrane tunnel of different length ( Figure 5B ) . Likewise , the SecY-beta translocon shows two flexible loops at both sides of the channel that make a further narrower but longer cavity crossing the protein structure . Despite the misassignments of pore axis and pore centres , in these three examples most of the pore-lining residues still seem to be identified correctly because the calculated optimal cavities , indicated by red spheres , partially overlap with the “true” cavities , indicated by the black arrows . In terms of pore shape , PoreWalker seems to recognise common sub-shapes across channel families . For instance , all aquaglyceroporins show a DU-like string shape ( where D and U represent funnel-like shape of decreasing and increasing diameter , respectively ) , which represents a hour-glasssed shape confirmed by a few published data [5] , [28] , [29] . Likewise , potassium channels present a shared sub-shape , a DUD sub-string shape at the cytoplasmic side of the channel , that is in agreement with the channel features reported by Mackinnon et al . , i . e . a constriction at the cytoplasmic side , the internal pore , widening into a larger water-filled void , the internal cavity , which leads towards the narrow selectivity filter located at the periplasmic side of the channel [12] . In addition , the linearity of the cavity seems to give some insights on the pore selectivity to different types of solutes ( Table 1 ) . In fact , 10 of the 13 channels for inorganic ions in the set showed a very regular cavity , with average percentage of co-linear pore centres of 91 . 9% ( SD = 7 . 0% ) and organic small molecule/ion channels had less regular pores , with percentage of co-linear centres lower than 60% . For completeness , PoreWalker output was also compared with results obtained using HOLE and MolAxis on the same set of structures . A systematic comparison with MOLE results could not be performed because , probably due to the intrinsic looseness of some structures , like the MthK and the ASIC1 channels , many of the tunnels identified by MOLE lie parallel and not perpendicular to the membrane axis and could not be considered as transmembrane . Within the set of pore features produced by PoreWalker and HOLE , the only comparable quantitative measure is the diameter , calculated along the pore at given heights . Diameter profiles obtained at 1Å steps for the 19 transmembrane proteins in the set were compared using the R2 correlation coefficient ( see Supplementary Text S1 , Table 1 , and Figures 6 , S3 , S4 and S5 ) . Pore diameter analyses performed with the two methods showed good agreement for 12 of the 19 diameter profiles , with R2 higher than 0 . 75 . However , the remaining 7 profiles showed very poor correlation coefficients , with R2 very close or equal to zero . This behaviour seemed to be strongly affected by the regularity of the cavity . In fact , R2 values showed a good correlation with the number of co-linear pore centres ( Supplementary Figure S5 ) with a R2 of 0 . 70 and only one strong outlier , the sodium-potassium channel ( PDB code 2ahy ) . The disagreement between the two profiles in this case was due to a completely different pore exit at the top channel side identified by HOLE that seems visually incorrect and makes the diameter trend in that area very peculiar . As for MolAxis , the program does not calculate diameter values at given heights along the channel axis but provide a partial list of the amino acids that contribute to the pore surface . Therefore , minimum diameters and pore lining residues were used to compare PoreWalker and MolAxis results . MolAxis could not identify a channel for 9 of the 19 test protein structures ( Table 1 ) , the water , glycerol and ammonia channels and three potassium channels . For the remaining 10 proteins , minimum diameter values derived from the two methods gave poor correlation ( R2 = 0 . 46 ) . The exclusion of the SecYE-beta translocon , incorrectly characterised by PoreWalker , lead to an R2 of 0 . 69 ( corresponding MolAxis-HOLE R2 were 0 . 60 and 0 . 57 , respectively ) . Minimum diameters calculated by HOLE and PoreWalker gave a better correlation , with R2 of 0 . 54 and 0 . 90 , respectively ( the overall R2 on the 19 structure set was 0 . 67 ) . In term of pore-lining residues , MolAxis provides a list of the amino acids responsible for the calculated diameters , i . e . a subset of the amino acids that make the surface pore . MolAxis pore-lining residues were fully included in PoreWalker pore-lining residue list in all the compared proteins but the SecYE-beta translocon . In this case , 23 of the 24 pore-lining residues detected by MolAxis were included in the list generated by PoreWalker , showing that the method can reliably identify amino acids which build a channel despite mis-placements of its pore vector . Finally , transmembrane pores identified by PoreWalker were found to coincide well with molecules of solute found in the 3D structure . Figure 5C–F shows the SoPIP2;1 plant aquaporin ( 1z98 ) and the sodium-potassium channel ( 2ahy ) filled with water molecules and sodium and calcium cations , respectively . In both cases the cavities generated by PoreWalker completely surround and include water molecules and ions , which provide good evidence for the location and shape of the pore . Interestingly , PoreWalker is also able to identify the two main choke points in the water channel of the SoPIP2;1 reported to be in a closed state -the canonical Ar/R constriction site near the top of the pore and a narrower restriction close to the bottom of the channel ( Figure 5D ) . The method can therefore analyse and characterise both “open” and “closed” transmembrane protein channels and transmembrane transporters . The KcsA potassium channel is a homotetrameric integral membrane protein with high sequence similarity to all the potassium channels , particularly in the pore region . Its channel includes three elements: 1 ) a narrow entrance , known as the internal pore , starting at the intracellular side of the membrane; 2 ) an internal cavity , about 10Å in diameter , at the middle of the membrane; 3 ) a further narrowing , the selectivity filter , which leads to the extracellular environment [30] . The KcsA channel is therefore a good target to assess the ability of PoreWalker to detect constrictions , gates and internal cavities in the 3D-structure of a channel protein . The 3D structures of the Kcsa potassium channel in the presence of low ( 3 mM , Figure 7A ) and high ( 200 mM , Figure 7C ) K+ concentrations are available at the wwPDB ( codes 1k4c [30] and 1bl8 [8] , respectively ) and their pore features were derived and analysed using PoreWalker ( Figures 7–9 ) . The diameter profile of the low-K+ channel ( Figure 7B , solid line ) shows that PoreWalker can neatly identify the three main features of the channel: first a ∼3Å narrowing corresponding to the internal pore , the internal ∼9 . 0Å bigger cavity and a second narrower ( ∼1Å ) constriction corresponding to the selectivity filter , highlighted in the Figure in orange , blue and red , respectively . It is interesting to notice here that diameter values calculated at 1Å steps by both HOLE ( dotted line ) and PoreWalker ( dashed line ) at the maximum width of the internal cavity ( ∼4Å ) were significantly smaller than those reported in the description of the 3D-structure [30] ( ∼10Å ) and found using the standard PoreWalker protocol at 3Å steps ( ∼9Å ) . The calculated diameters of the internal pore and cavity also strongly agree with the proposed mechanism of ion conductance through the pore . In fact , potassium cations are thought to move through the internal pore and cavity in a hydrated form and to be dehydrated at the selectivity filter . The internal pore detected by PoreWalker is ∼3Å in diameter and could allow through one water molecule per time ( the average diameter of a water molecule is usually taken as 2 . 8Å ) . Therefore , K+ ion could move through it alternating with water molecules . On the other side , the selectivity filter has a predicted diameter of ∼1Å and could therefore let through only dehydrated K+ cations . The comparison of the diameter profiles of the channel in presence of low and high quantity of ions ( Figure 7D , solid and dotted line , respectively ) showed that besides expected differences at the cytoplasmic side of the pore , where a gate mechanism is known to operate , the entrance of the selectivity filter is ∼2 . 5Å wider at high concentrations of K+ . According to PoreWalker , the pore lining residues , which define access to the selectivity filter , are the Thr75s from the four chains making up the pore . The difference in pore diameters at this point seems mainly to be due to different Thr sidechain conformations ( Figure 7E–F ) . A significant difference in the two conformations of the KcsA selectivity filter had been previously highlighted at the level of residues Val76 and Gly77 . A deeper analysis of the whole selectivity filter ( Figure 8A ) showed that the periplasmic side of the filter ( at the top of the Figure ) varies very slightly , while a major change is hinged at Gly77 and extends through Val76 to Thr75 , where a pincher-like shutting mechanism could reasonably be hypothesized ( RMSDs of all-atom superpositions were 0 . 33Å , 0 . 58Å and 0 . 99Å for Gly77 ( Figure 8B ) , Val76 ( Figure 8C ) and Thr75 ( Figure 8D ) , respectively ) . Besides , the internal cavity accommodates K+ ions as hydrated by eight water molecules . The 3D-structure of the low-K+ channel cavity ( Figure 9 ) shows that the four water molecules facing the filter are aligned to the sidechain oxygens of Thr75s and can make hydrogen bonds with them ( inter-oxygen distances are 3 . 9Å ) . Moreover , their distances from the corresponding K+ ion are close to optimal ( 3 . 4Å versus 2 . 8Å [31] ) . Therefore , it might be reasonably thought that the pinching mechanism could be aimed at weakening the water-K+ hydration complex by increasing the distance between the water molecules and the ion to facilitate its way into the pore . We developed PoreWalker , a novel web-available method for the detection and characterisation of channels in transmembrane proteins from their three-dimensional structure . PoreWalker is fully automated and very user-friendly , requiring as input only the 3D coordinates of a transmembrane protein structure . A key prerequisite of the submitted structure is the presence of a transmembrane helix bundle or beta-barrel creating the pore , which is needed for the geometrical identification of the main protein axis . If this condition is not met , the detection/description cannot be performed with the current version of the software . In term of outputs , in addition to diameter profiles , PoreWalker describes several specific pore features , in particular the shape and the regularity of the channel cavity , the atoms and corresponding amino acids lining the pore wall , and the position of pore centres along the channel . These features can be very helpful to gain further insights into the functional and structural properties of transmembrane protein channels by triggering specific in silico or experimental analyses , as shown from the recent characterization of the bacterial TolC channel [32] . PoreWalker is based on the assumption that , in a transmembrane channel protein , the pore is made by the longest cavity crossing the protein along the main axis of its transmembrane portion and therefore detects the longest widest cavity in a transmembrane protein structure . However , there are cases , as in the Amt-B and the SecYE-beta translocon , where the longest widest cavity does not correspond to the most likely “true” channel and in such cases the method assigns incorrectly one or both the pore gates . Interestingly , for these examples , calculated optimal cavities partially overlapped with the “true” cavities and most of the pore-lining residues were anyway identified properly . In summary , PoreWalker provides a robust and automated resource to interpret , coordinate data and derive quantitative descriptors , which help to provide a deeper understanding and classification of membrane protein structures . | Transmembrane channel proteins are responsible for the transport of ions and molecules through biological membranes and are pivotal for the physiology of the cell . In fact , their incorrect functioning is involved or related to several diseases ( diabetes , myotonia , Parkinson's disease , etc . ) . Moreover , their specificity and selectivity to different ions or molecules have been hypothesized and sometimes shown to strongly depend on the shape and size or amino acid composition of the channel . Therefore , computational methods to identify and quantitatively characterise channel geometry in transmembrane protein structures are key tools to better understand how they function . We have developed PoreWalker , a new method to detect and describe the geometry of these channels in transmembrane proteins from their 3D structures . The method is fully automated , very user-friendly , identifies the location of the channel and derives a number of channel features: diameter profiles at given heights along the channel , all the residues lining the channel walls , size , shape and regularity of the channel . These features can be very helpful in the study of how these channels might function . We have applied PoreWalker to several channel protein structures and were able to identify shape/size/residue features that were representative of specific channel families . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results/Discussion"
] | [
"computational",
"biology/macromolecular",
"structure",
"analysis"
] | 2009 | PoreWalker: A Novel Tool for the Identification and Characterization of Channels in Transmembrane Proteins from Their Three-Dimensional Structure |
Calcineurin governs stress survival , sexual differentiation , and virulence of the human fungal pathogen Cryptococcus neoformans . Calcineurin is activated by increased Ca2+ levels caused by stress , and transduces signals by dephosphorylating protein substrates . Herein , we identified and characterized calcineurin substrates in C . neoformans by employing phosphoproteomic TiO2 enrichment and quantitative mass spectrometry . The identified targets include the transactivator Crz1 as well as novel substrates whose functions are linked to P-bodies/stress granules ( PBs/SGs ) and mRNA translation and decay , such as Pbp1 and Puf4 . We show that Crz1 is a bona fide calcineurin substrate , and Crz1 localization and transcriptional activity are controlled by calcineurin . We previously demonstrated that thermal and other stresses trigger calcineurin localization to PBs/SGs . Several calcineurin targets localized to PBs/SGs , including Puf4 and Pbp1 , contribute to stress resistance and virulence individually or in conjunction with Crz1 . Moreover , Pbp1 is also required for sexual development . Genetic epistasis analysis revealed that Crz1 and the novel targets Lhp1 , Puf4 , and Pbp1 function in a branched calcineurin pathway that orchestrates stress survival and virulence . These findings support a model whereby calcineurin controls stress and virulence , at the transcriptional level via Crz1 , and post-transcriptionally by localizing to PBs/SGs and acting on targets involved in mRNA metabolism . The calcineurin targets identified in this study share little overlap with known calcineurin substrates , with the exception of Crz1 . In particular , the mRNA binding proteins and PBs/SGs residents comprise a cohort of novel calcineurin targets that have not been previously linked to calcineurin in mammals or in Saccharomyces cerevisiae . This study suggests either extensive evolutionary rewiring of the calcineurin pathway , or alternatively that these novel calcineurin targets have yet to be characterized as calcineurin targets in other organisms . These findings further highlight C . neoformans as an outstanding model to define calcineurin-responsive virulence networks as targets for antifungal therapy .
Cryptococcus neoformans is an environmental fungus found ubiquitously throughout the world [1 , 2] . Both spores and desiccated yeast cells are infectious propagules and exposure via inhalation causes life-threatening disease [2 , 3] . C . neoformans is primarily an opportunistic pathogenic fungus that causes meningoencephalitis , frequently in immunocompromised patients with HIV/AIDS , organ transplants , or autoimmune diseases [4 , 5] . Key virulence attributes of Cryptococcus are its ability to adapt to stressful host environments , including the elevated body temperature of the mammalian host [6 , 7] . Calcineurin is necessary for Cryptococcus to effectively survive host thermal stress [8] . Calcineurin is a virulence factor conserved in human fungal pathogens across species including C . neoformans , Candida albicans , Aspergillus fumigatus , and Mucor circinelloides [8–12] . Two immunosuppressive natural products , FK506 and cyclosporin A ( CsA ) , bind to intracellular proteins , cyclophilin A and FKBP12 respectively , and the protein-drug complexes then inhibit calcineurin [13–17] . Because calcineurin is conserved from fungi to humans , FK506 and CsA exhibit broad antifungal and immunosuppressive activities [18–21] . Calcineurin is a Ca2+/calmodulin-activated serine/threonine-specific protein phosphatase consisting of two subunits: a catalytic A subunit and a regulatory B subunit [22 , 23] . In response to internal or external stress-derived signals intracellular Ca2+ levels increase , Ca2+ binds to calmodulin , and Ca2+-calmodulin then binds to the catalytic A subunit of calcineurin , leading to calcineurin activation [24] . Activated calcineurin dephosphorylates target proteins , which in turn modulate various biological processes [25] . In mammals such as mice and humans , calcineurin dephosphorylates the NFAT ( nuclear factor of activated T cell ) family of transcription factors that controls transcription of genes required for T cell activation , cardiac hypertrophy , and development [26–28] . In the model budding yeast Saccharomyces cerevisiae , Crz1 ( calcineurin-responsive zinc finger 1 ) was identified as a calcineurin-activated transcription factor inducing transcription of genes involved in stress responses [29–32] . Recent studies identified Crz1 orthologues in multiple ascomycetous fungal species and other eukaryotes [31] . Recently , three groups characterized a candidate Crz1/Sp1 transcription factor in the human pathogenic fungus C . neoformans and demonstrated that it plays crucial roles in cell wall integrity and virulence; however , only one of these studies concluded that this was a bona fide Crz1 ortholog [33–35] . Calcineurin is required for growth at 37°C , virulence , and sexual reproduction in the fungal pathogen C . neoformans [9 , 36 , 37] . Our previous studies found that calcineurin is re-localized from the cytoplasm to puncta and the mother-bud neck in response to heat and other stresses [38 , 39] . Calcineurin was found to co-localize in puncta with components of PBs/SGs , which are known to contain non-translating mRNPs ( messenger ribonucleoprotein complexes ) [40–42] . These cytoplasmic structures consist of mRNAs associated with translation initiation factors , RNA binding proteins , and the mRNA decay machinery , and function to control translation initiation , mRNA degradation , and siRNA function [43–45] . These findings lead to the hypothesis that PBs/SGs may contain putative calcineurin targets whose functions are governed by calcineurin to promote thermal stress survival . To identify calcineurin targets , we compared the phosphopeptide profiles obtained from calcineurin-activated ( WT cells exposed to 37°C ) and calcineurin-deficient conditions ( cna1Δ mutant or WT cells exposed to FK506 ) . In total , 56 calcineurin-dependent dephosphorylation targets were identified , including Cna1 , the Crz1 ortholog , and proteins whose orthologs have roles in stress responses , mRNA binding/stability , protein translation , and vesicular trafficking . Intriguingly , several proteins , including Pbp1 ( Poly ( A ) binding protein-Binding Protein ) and Puf4 ( PUmilio-homology domain Family ) were identified as calcineurin targets and upon exposure to 37°C , co-localize in PBs/SGs with calcineurin . We demonstrate that Crz1 is a bona fide calcineurin substrate whose localization and transcriptional activity is controlled by calcineurin , while Pbp1 functions as a calcineurin target involved in calcineurin-dependent sexual reproduction . Moreover , mutation of the genes encoding the calcineurin targets Crz1 , Puf4 , and Lhp1 conferred hypersensitivity to thermal and other stresses . Epistasis analysis revealed that Crz1 and the RNA binding proteins Puf4 , Pbp1 , and Lhp1 function in a branched pathway controlled by calcineurin . These results support a model in which calcineurin governs growth at high temperature , virulence , and sexual reproduction by controlling both DNA- and RNA-binding proteins in transcriptional and post-transcriptional circuits .
Calcineurin is essential for survival at 37°C and virulence; however , no calcineurin targets have been identified in C . neoformans . We sought to identify calcineurin targets by employing a quantitative phosphoproteomic approach comprising TiO2-based affinity chromatography enrichment of phosphopeptides and quantitative analysis using 2-dimentional ultraperformance liquid chromatography coupled to high-resolution accurate-mass spectrometry ( 2D-LC/LC-MS/MS ) . To identify calcineurin-dependent dephosphorylation events at 37°C , we performed three differential expression phosphoproteomic screens . In the first and second screens , we aligned and compared phosphopeptide profiles from WT cells grown at 25°C and shifted to 37°C for 1 hour and either exposed or not to FK506 15 minutes prior to and during the shift to 37°C . For the third screen , phosphopeptide patterns of cna1Δ mutant cells grown at 25°C and shifted to 37°C for 1 hr were compared to similarly grown and treated WT cells . In summary , these screens identified 2 , 016 total peptides ( 1 , 398 unique phosphorylated peptides ) and 796 total proteins ( 576 unique phosphorylated proteins ) across all samples ( S1 Table ) . We performed a T-test analysis on log2 transformed phosphopeptide intensities in the datasets of the different conditions compared in the screens . Phosphopeptides that were overrepresented by more than 2-fold ( T-test p-value <0 . 05 ) in the calcineurin-deficient conditions relative to the calcineurin-activated condition were considered as potential calcineurin targets ( S2 and S3 Tables ) . A total of 59 phosphorylated peptides ( 44 proteins ) were more abundant in calcineurin-deficient cells ( S4 and S5 Tables ) . We submitted these 44 ORFs to BLAST searches for functional domain homologies against protein databases and best fits were assigned based on the S . cerevisiae genome database ( SGD ) . Importantly , we found that only 12 proteins containing phosphopeptides that changed in abundance in calcineurin-deficient strains overlapped between the two different conditions ( cna1Δ mutant or WT cells exposed to FK506 ) ( Fig 1A , S5 Table ) . Moreover , more potential calcineurin targets were identified when calcineurin was blocked by mutation ( 36 targets ) than by inhibition with FK506 ( 20 targets ) . We interpreted these finding as being the result of a more profound physiological remodeling triggered by permanent calcineurin inhibition in the cna1 mutant as opposed to the potentially less severe effects imposed by transient inhibition of calcineurin with FK506 ( Fig 1A , S5 Table ) . Similar results were found by a study that identified the calcineurin substrates in S . cerevisiae [46] . The identified functional categories included proteins involved in calcium signaling or stress responses such as calcineurin A ( Cna1 ) itself , and an ortholog of the transcription factor Crz1 ( Fig 1B and 1C ) . Intriguingly , the screen also identified several proteins whose S . cerevisiae orthologs are localized to P-bodies and stress granules , including Pbp1 , Pab1 , Puf4 , Vts1 , and Gwo1 ( Gis2 ortholog ) [47–49] as well as other components whose orthologs have functions in protein synthesis and mRNA binding/stability . Another significant functional category that was identified involves vesicular trafficking , in accord with our previous studies that COPI and COPlI members interact and co-localize with calcineurin in ER-associated stress granules [38] . Although it is generally accepted that protein phosphatases do not dephosphorylate specific amino acid sequence motifs , we analyzed the 59 calcineurin-dependent phosphopeptide sequences employing the PhosphoSitePlus software [50] to test if they exhibit any characteristic features . Interestingly , the calcineurin-dependent phosphopeptide containing a phosphoserine residue showed a significant enrichment of Pro at the +1 position or Arg at the -3 position ( 27 or 16 phosphopeptides out of 47 , respectively ) with 7 featuring both Arg at -3 and Pro at +1 ( S1A Fig and S5 Table ) . However , the calcineurin-dependent phosphopeptides containing a phosphothreonine residue exhibited a weak potential signature in that 3 or 2 out of 10 had an Arg at the -2 or -3 position , respectively ( S1B Fig and S5 Table ) . Previous studies demonstrated that calcineurin recognizes specific substrate docking sites such as PxIxIT motifs [51] . To examine if the potential calcineurin substrates identified by our screens contain PxIxIT motifs , we searched for the verified PxIxIT motif ( P[^PG][IVLF][^PG][IVLF][TSHEDQNKR] ) characterized in S . cerevisiae [46] . In C . neoformans , 2 , 530 predicted ORFs contain the PxIxIT motif . Among the 44 putative calcineurin targets , 13 of them , including Cna1 , Crz1 , Puf4 , and Gcd2 feature a PxIxIT motif ( S5 Table ) . In particular , the Crz1 ortholog contains two potential PxIxIT motifs ( PALSIS and PMICIQ ) , suggesting that Crz1 could represent an authentic calcineurin substrate in C . neoformans . However , the PxIxIT motif is poorly conserved amongst different fungal species ( 46 ) and occurs rather frequently ( 2536 out of a predicted total 6957 proteins or 36% ) in the Cryptococcus proteome . Thus , it remains to be tested whether the identified PxlxlT motifs in fact function as calcineurin docking sites in Cryptococcus . To confirm the effectiveness of our phosphoproteomic approach , we first tested if Crz1 is a calcineurin substrate . Cells expressing a C-terminally FLAG-tagged Crz1 protein were grown at 24°C and shifted from 24°C to 37°C for 1 hour in the presence or absence of FK506 . The Crz1-FLAG tagged protein exhibited increased electrophoretic mobility when isolated from cells shifted from 24°C to 37°C as compared to that isolated from the culture grown at 24°C ( Fig 2A ) . In contrast , Crz1-FLAG displayed reduced mobility in cells treated with FK506 at both temperatures ( Fig 2A ) . To test if the reduction in mobility of Crz1-FLAG is caused by phosphorylation , we immune-isolated the Crz1-FLAG protein from WT cells exposed to FK506 , and the isolated protein was incubated with λ phosphatase in the presence or absence of phosphatase inhibitor . The reduced mobility of the Crz1-FLAG protein in FK506 was reversed by treatment with λ phosphatase but not with λ phosphatase plus a phosphatase inhibitor ( Fig 2B ) . Moreover , the mobility pattern of Crz1-FLAG in cells exposed to FK506 was similar to the mobility of Crz1-FLAG when expressed in a cna1 mutant background ( Fig 2C ) . We next tested if calcineurin dephosphorylates Crz1 in vitro . The immunopurified Crz1-FLAG protein was isolated from WT and cna1Δ strains and treated with λ phosphatase or human calcineurin . The Crz1-FLAG protein from the cna1Δ strain exhibited decreased mobility compared with the protein from the WT strain and this decreased mobility was reversed by treatment with either human calcineurin or λ phosphatase ( Fig 2D ) . These results demonstrate that Crz1 is a phosphoprotein and a bona fide substrate of calcineurin and support the validity of our phosphoproteomic approach . Previous studies have concluded based on indirect evidence that calcineurin dephosphorylates Crz1 and that this triggers Crz1 translocation into the nucleus in several fungi [31 , 52 , 53] . In C . neoformans , Lev et al . presented findings that Crz1 is nuclear localized in response to several stresses , and that this relocalization activates the transcriptional activity of Crz1 [33] . To verify the subcellular localization of Crz1 , we examined localization of Crz1-mCherry in WT and cna1Δ strains co-expressing the nucleolar marker GFP-Nop1 [54] . The Crz1-mCherry protein was distributed throughout the cytosol in cells grown at room temperature . Crz1 translocated into the nucleus upon calcineurin activation by exposure to 37°C . In contrast , Crz1-mCherry failed to translocate to the nucleus in cna1Δ mutant cells in response to temperature shift ( Fig 2E ) . The chitin synthase genes are known targets of the calcineurin-Crz1 cascade in several fungi [33 , 55–57] . We therefore examined expression of three chitin synthase genes ( CHS5 , CHS6 , and CHS7 ) in the WT , and the cna1Δ , and crz1Δ mutant strains at 24°C and 37°C . Expression of all three chitin synthase genes was induced upon shift from 24°C to 37°C ( Fig 2F ) . Importantly , the induced mRNA expression of CHS5 and CHS6 , but not CHS7 , was prevented by the cna1Δ or crz1Δ mutation , confirming that expression of the CHS5 and CHS6 genes is controlled by the calcineurin-Crz1 pathway . Taken together , these results demonstrate that calcineurin directly dephosphorylates Crz1 and thereby controls Crz1 sub-cellular localization and transcriptional activity in response to elevated temperature . To determine the role of calcineurin-dependent phosphosites in Crz1 function , we tested the effect of mutations in these phosphosites on Crz1 nuclear localization and transactivation activity . Substitutions of serine with the non phosphorylatable amino acid alanine were introduced in transgenic Crz1 constructs , which then were fused at their C-termini to the mCherry fluorescent protein and integrated into the genome of the crz1Δ mutant . Because individual or concomitant substitution of S288 and S508 , which are the Crz1 phosphosite residues originally identified by the phosphoscreen , with alanine resulted in only 23% nuclear localization of Crz1 ( S2C Fig ) , we hypothesized the occurrence of additional calcineurin-dependent phosphosites . Therefore , we performed an independent phosphoscreen analysis of the immunoprecipitated Crz1-mCherry protein from cells shifted from 25°C to 37°C in the absence or the presence of FK506 . This analysis revealed a total of 12 calcineurin-dependent phospho-serine residues ( including the above S288 and S508 residues originally identified ) ( S2A and S2B Fig ) , which were systematically mutated in various combinations . Combined substitution of 3 ( S563 , S565 , S569 ) , 4 ( S288 , S291 , S294 , S298 ) and ( S288 , S329 , S508 , S569 referred as Crz14S-A in Fig 3C and 3D ) , or 6 ( S288 , S329 , S508 , S569 , S765 , S810 , referred as Crz16S-A mutant in Fig 3C–3E ) , serine residues to alanine , all resulted in about 20% Crz1 nuclear localization at 24°C ( S2C Fig and Fig 3B and 3C ) . Strikingly , concurrent substitution of the seven serine residues ( S103 , S288 , S329 , S508 , S569 , S765 , S810 ) with alanine ( depicted in Fig 3A and referred to as Crz17S-A mutant in Fig 3B–3F ) elicited Crz1 nuclear localization in 81% of the cells observed at 24°C . In contrast , at 24°C , only 5 . 09% of Crz1WT cells exhibited Crz1 nuclear localization , and the majority of cells displayed diffuse cytoplasmic localization of Crz1 ( Fig 3B and 3C ) . Comparing the mobility shift of the recombinant Crz1 proteins to wild-type Crz1 when calcineurin is inhibited , we observed that concurrent mutations of four , six , or seven predicted sites did not result in complete abolishment of the mobility shift . However , at 37°C the reduction in the mobility shift of the Crz17S-A mutant is more evident in comparison to the Crz1WT strain and remarkably further addition of FK506 did not alter the mobility of the major slowest migrating protein fraction ( Fig 3D ) . Moreover , the Crz17S-A mutant exhibited stronger induction in the expression of the Crz1-dependent CHS6 gene at 24°C compared to the induction shown by the WT strains , and remarkably this induction was partially resistant to FK506 ( Fig 3E ) . Despite the robust nuclear localization and transactivation activity of the Crz17S-A mutant , when the phenotypes of the Crz1WT and Crz17S-A mutants were compared , we did not observe differences in growth during thermal stress or on media containing 0 . 4 M calcium or cell-wall damaging agents ( Fig 3F ) . Therefore , increased Crz1 nuclear localization alone is not sufficient to improve stress resistance . Taken together , this data demonstrates that dephosphorylation of Crz1 at S103 in combination with the other 6 serine residues mutated in the Crz17S-A is critical to elicit Crz1 nuclear localization and transactivation activity resistant to FK506 . These results also suggest that in addition to those mutated in the Crz17S-A mutant , other sites are acted upon by calcineurin , most likely S563 and S565 , which are important to trigger complete Crz1 nuclear localization , or that another phosphatase may be acting in parallel with calcineurin . In the budding yeast S . cerevisiae Pbp1 ( Pab1-Binding Protein ) localizes in PBs/SGs under stress conditions and is essential for RNA processing including polyadenylation , splicing , and degradation [47 , 58 , 59] . In our phosphoscreen , four Pbp1 phosphopeptides were more abundant in both FK506 treated and cna1Δ mutant cells ( Fig 1 , S4 Table ) . To test whether Pbp1 is regulated by calcineurin , in vitro phosphatase assays were performed with Pbp1-FLAG protein immune-isolated from WT cells or WT cells treated with FK506 . We observed that the Pbp1-FLAG protein exhibited increased mobility when treated with λ phosphatase and this mobility was reduced by treatment with phosphatase inhibitors ( Fig 4A ) , strongly suggesting that Pbp1 is a potential calcineurin substrate . As calcineurin A mainly co-localizes with components of PBs/SGs at 37°C [39] , we next examined the localization of Pbp1 under thermal-stress conditions . Pbp1-mCherry fluorescence was observed in cytosolic puncta at 24°C and upon shift to 37°C these Pbp1-mCherry puncta co-localized with the P-body marker Dcp1 , ( Fig 4B ) or with the stress granule component Pub1 ( Fig 4C ) . Interestingly , Pbp1 also co-localized with the calcineurin catalytic subunit Cna1 in PBs/SGs in response to thermal stress ( Fig 4D ) . In S . cerevisiae , Pbp1 is involved in the regulation of mating-type switching in mother cells by post-transcriptionally controlling expression of the HO endonuclease [60] . As calcineurin is necessary for hyphal elongation during sexual reproduction in C . neoformans [36] , we tested whether Pbp1 is involved in this process . Both cna1Δ and pbp1Δ mutant strains exhibited sexual reproduction defects ( Fig 4E ) in bilateral mating assays . The defect in sexual reproduction caused by the cna1 mutation was more severe than the one observed in the pbp1 mutant . Moreover , similar to the cna1 mutant strain , the pbp1 mutant showed a marked decrease in pheromone gene ( MFα1 ) expression levels during sexual reproduction ( Fig 4F ) , suggesting that Pbp1 is required for activation of pheromone gene expression or mRNA stability . Collectively , these results provide evidence that calcineurin positively regulates sexual reproduction in part by controlling the phosphorylation state of Pbp1 . Calcineurin controls responses to stress conditions , and we reasoned that calcineurin may promote survival during stress by dephosphorylating key targets . Mutants were generated in several of the identified potential calcineurin substrates , including Puf4 , Pbp1 , Vts1 , Anb1 , Gcd2 , and Gwo1 ( Fig 1C ) , and these mutants were then subjected to phenotypic analysis under a variety of stress conditions ( Fig 5 ) . We confirmed previous reports that the puf4Δ mutant exhibits a pattern of sensitivity to multiple stresses , including heat , FK506 , Congo red , SDS , or dithiothreitol ( DTT ) , but in contrast to a prior study it was resistant to tunicamycin [61 , 62] . The lhp1Δ mutant exhibited increased sensitivity to heat , FK506 , or SDS , whereas the pbp1Δ mutant showed more resistance to heat stress or FK506 as compared to the WT . Finally , mutation of TIF3 caused an increased sensitivity to ER stress imposed by exposure to DTT ( Fig 5 ) . Taken together , these results indicate that Puf4 , Lhp1 , Pbp1 , and Tif3 contribute to stress survival . In C . neoformans , calcineurin is also required for sexual reproduction , implying that calcineurin targets may also be involved in sexual development . To assess this , we tested the phenotypes conferred by mutating calcineurin targets and found that tif3Δ mutants , similar to pbp1Δ mutants , exhibit a defect in dikaryotic hyphal production ( Fig 6A ) . The tif3 mutation also resulted in reduced expression levels of the pheromone gene ( MFα1 ) during mating , suggesting that Tif3 is required for proper sexual reproduction . The puf4Δ mutants exhibited a hyper-filamentation phenotype and reduced expression levels of the pheromone gene ( Fig 6A and 6B ) . In addition , puf4Δ mutants showed defective basidiospore formation , implying that Puf4 is essential for basidiospore chain formation during bisexual reproduction . These results suggest that in addition to Pbp1 , Tif3 and Puf4 are also involved in sexual reproduction . Calcineurin is relocalized from the bulk cytoplasm to PBs/SGs in response to heat and other stresses [39] , and we hypothesize that a fraction of the calcineurin targets may co-localize with calcineurin in PBs/SGs . To test this hypothesis , we generated mCherry-tagged alleles of selected calcineurin target proteins known to exhibit RNA binding activity and examined their localization . We found that Puf4 , Vts1 , Tif3 , and Gwo1-mCherry all re-localized to P-bodies following shift from 24°C to 37°C in C . neoformans ( S3 Fig ) . Other potential calcineurin substrates including Lhp1 , Gcd2 , and Anb1-mCherry did not co-localize with Dcp1 . Taken together , we propose that calcineurin re-localizes to P-bodies and may regulate the function of potential targets at this location including Puf4 , Pbp1 , Vts1 , Tif3 , and Gwo1 . While Crz1 is a major calcineurin target , it contributes to but is not essential for survival at high temperature . Additionally , we found that mutation of three calcineurin targets , pbp1 , puf4 , and lhp1 , conferred defects ( puf4 , lhp1 ) or enhanced ( pbp1 ) thermotolerance ( Fig 5 ) , which were complemented by expression of the corresponding WT allele ( S4 Fig ) . Therefore , we hypothesize that the calcineurin network governing cell stress responses is branched with additional targets including Puf4 , Lhp1 , and Pbp1 that are regulated by calcineurin and function in parallel with Crz1 . To test this model , we conducted epistasis analysis by generating double mutants and examining thermosensitivity and virulence phenotypes . As presented above , the pbp1Δ mutant , unlike the puf4Δ and lhp1Δ mutants , is resistant to heat stress ( Fig 5 ) . Genetic analysis revealed that the pbp1Δ crz1Δ double mutant exhibited phenotypes that are intermediate between the pbp1Δ and crz1Δ single mutant phenotypes ( Fig 7A ) , suggesting that Crz1 and Pbp1 play opposite roles in heat stress responses . We also found that growth of the puf4Δ crz1Δ or lhp1Δ crz1Δ double mutant strains was more severely impaired at 37°C or 38°C compared to either the crz1Δ or the lhp1Δ and puf4Δ single mutant strains ( Fig 7A ) . These results demonstrate that Crz1 and Lhp1 , and Crz1 and Puf4 together play additive roles in thermoresistance . Overall , these results validate that Crz1 functions in pathways parallel to the RNA binding proteins Puf4 , Lhp1 , and Pbp1 in responses to heat stress . Virulence tests were conducted assessing the single and double mutants in the murine inhalation model ( Fig 7B and 7C ) . First , as reference we found that the crz1Δ mutant shows attenuated virulence compared to WT , but not to the same degree as the cna1Δ mutant , which is avirulent . Second , mutation of PUF4 alone did not result in a virulence defect under these conditions . In an insect larval model , however , the puf4Δ mutant was attenuated compared to the wild type strain ( S5A Fig ) . In addition , when mice were infected via intranasal instillation with a ten-fold lower inoculum ( 5×104 CFU ) , puf4Δ mutants were very modestly attenuated compared to the wild type and complemented strains , and crz1Δ puf4Δ double mutants were avirulent up to 60 days post-infection and thus more attenuated than either single mutant alone ( S5B Fig ) . Third , mutation of LHP1 alone did not result in a virulence defect , whereas lhp1Δ crz1Δ double mutant strains were attenuated compared to the lhp1Δ and crz1Δ single mutants ( Fig 7B ) . Fourth , the pbp1Δ single mutant strain showed attenuated virulence in a murine model and the pbp1Δ crz1Δ double mutant was avirulent in the murine inhalation model compared to the partial attenuation of either single mutant parent ( Fig 7B ) . The phenotypes of the crz1 mutation combined with the puf4 , lhp1 , and pbp1 mutations are additive , and this supports a model in which Puf4 , Lhp1 , and Pbp1 operate in a pathway parallel to Crz1 . These findings are consistent with models whereby Crz1 and the RNA binding proteins Puf4 , Lhp1 , and Pbp1 function in a branched pathway controlled by calcineurin to promote high temperature growth and virulence .
In this study , we employed a phospho-proteomic approach and identified calcineurin targets that are associated with thermal stress responses and virulence of C . neoformans . More targets where identified when calcineurin was inactivated by mutation than by drug inhibition with FK506 , and we attribute this to permanent in opposed to transient inhibition effects . However , irrespective of the differences in the phosphorylation profiles perturbed by calcineurin mutation versus inhibition , which are only partially overlapping , either mutation or inhibition of calcineurin is fully sufficient to render Cryptococcus neoformans sensitive to temperature and to other stress conditions that occur in the host . These observations fully validate calcineurin as an antifungal drug target . One of the key calcineurin targets identified is the Cryptococcus ortholog of the calcineurin-activated zinc finger transcription factor Crz1 . Crz1 is a bona fide calcineurin target in yeasts and other ascomycetous fungi , but because of limited sequence homology conservation , whether or not it is conserved in basidiomycetous fungi had been unclear . It also had been unclear whether the identified Crz1 ortholog is a direct calcineurin target in basidiomycetes . Two conflicting results on Crz1 have been reported in C . neoformans [33 , 34] . Adler et al . reported that GFP-Crz1 constitutively localizes in the nucleus , suggesting that Crz1 is not a calcineurin substrate [34] . Djordjevic and colleagues found that localization of Crz1-GFP is calcineurin-dependent and proposed that Crz1 may be a calcineurin target; however , neither study addressed whether Crz1 is a direct calcineurin substrate [33] . Here we demonstrate that dephosphorylation , nuclear localization , and transcriptional activity of Crz1 are directly dependent upon calcineurin ( Figs 2 and 3 ) . Other calcineurin targets identified in our study include the RNA-binding proteins Lhp1 , Puf4 , and Pbp1 . These RNA-binding proteins recognize and bind specific RNA sequences or structures to control multiple processes such as mRNA localization , stability , and function [63] . Among these RNA-binding proteins , Puf4 and Pbp1 together with Cna1 are concentrated in PBs/SGs in response to a shift to high temperature growth , suggesting that re-localization of calcineurin is a critical process for interaction with substrates [39] . In contrast , at either 24°C or 37°C , Lhp1-mCherry localizes to large cytoplasmic puncta that do not co-localize with Dcp1 or Cna1 , suggesting that Lhp1 is either an indirect target of calcineurin or that calcineurin that has not been re-localized to PBs/SGs acts on Lhp1 . However , with the exception of Pbp1 , the remaining RNA binding proteins , including Lhp1 , Puf3 , Vts1 , Anb1 , Tif3 , Gcd2 , and Gwo1 , should be considered as candidate calcineurin targets until they be subjected to further phosphorylation studies . The Puf proteins are members of a conserved family of sequence-specific RNA binding proteins that bind to the 5’ or 3’ UTRs of mRNA transcripts in eukaryotes [64 , 65] . This protein family controls translation of specific mRNA targets [66] . Recently , Lee and colleagues found that phosphorylation of Puf3 controls the translation of Puf3 target mRNAs that are associated with mitochondrial biogenesis in S . cerevisiae [67] . In C . neoformans , Puf4 is required for splicing and decay of the mRNA encoding the unfolded protein response transactivator Hxl1 [61] . Thus , mutation of PUF4 results in delayed splicing and expression of HXL1 mRNA in response to temperature or ER stresses . Interestingly , calcineurin is also involved in thermotolerance mediated by the unfolded protein response ( UPR ) [68] . Thus , we considered the intriguing possibility that calcineurin controls HXL1 expression via Puf4 phosphorylation and thereby contributes to responses to ER stress . However , in contrast to a previous report [61] and to the tunicamycin hypersensitivity exhibited by the cna1 mutant , puf4 mutation did not alter sensitivity to tunicamycin . However , it is possible that calcineurin and Puf4 are involved in the UPR via independent pathways . In agreement with previous studies , the puf4Δ mutant strain exhibited a similar degree of virulence to that seen with the WT in a murine infection model ( S5A Fig ) . However , our results showed that puf4Δ mutants were very modestly attenuated compared to wild type in insect larval and mouse models ( S5B Fig ) . The puf4Δ crz1Δ double mutant was more attenuated than either single mutant in either mice or insects ( S5 Fig ) . These results support the conclusion that Puf4 contributes to the virulence of Cryptococcus . While calcineurin is globally conserved , previous studies have proposed that calcineurin targets are remarkably different across divergent fungal species [46] . We found that the calcineurin targets identified in this study share little overlap with known calcineurin substrates in S . cerevisiae , with the exception of Crz1 . In particular , the mRNA binding proteins Lhp1 , Puf4 and Pbp1 comprise a cohort of novel potential calcineurin targets , which have not been previously linked to calcineurin in mammals or in S . cerevisiae . Thus , while calcineurin dephosphorylates different targets in different fungi , calcineurin plays a similar role in stress survival in S . cerevisiae and C . neoformans , but a unique role in virulence of C . neoformans and other pathogenic fungi . Taken together , our results suggest that calcineurin cellular networks have been extensively rewired throughout the evolution of ascomycetes and basidiomycetes from their last common ancestor . However , further analysis identifying calcineurin targets in other fungal species will be required to test this model . In summary , our results support a model whereby activated calcineurin functions in a branched pathway to activate transcriptional and post-transcriptional processes that coordinate stress survival , virulence , and sexual reproduction ( Fig 8 ) . At the transcriptional level , calcineurin dephosphorylates Crz1 , which in turn , translocates into the nucleus to activate mRNA expression of genes linked to cell wall integrity , stress responses , and virulence . Second , calcineurin translocates into PBs/SGs where it targets post-transcriptional processes via Pbp1 and Puf4 to control thermotolerance , virulence , and sexual reproduction .
Strains used in this study were derived from C . neoformans laboratory reference strain H99 and are listed in S6 Table . C . neoformans strains were maintained on YPD agar ( 1% yeast extract , 2% bactopeptone , and 2% dextrose ) supplemented with the relevant antibiotics and grown at 30°C unless otherwise indicated . For analysis of cell viability at different temperatures , 5 μl of 10-fold serial dilutions from overnight cultures were spotted on YPD agar ( BD Difco ) and incubated at a range of temperatures ( 30 , 37 , 38 and 39°C ) . To test stress-related phenotypes , 2 . 5 to 10 μl of cell cultures that had been grown in liquid YPD medium overnight were 10-fold serially diluted and spotted on YPD medium containing the indicated concentrations ( see Fig legends ) of the following compounds: calcium chloride , Congo red , tunicamycin , DTT ( Sigma ) , FK506 ( Astellas Pharma ) , and SDS ( Teknova ) . Gene disruption cassettes were generated using an overlap-PCR strategy as previously described [69] . The oligonucleotides used in this study are listed in S7 Table . The 5’ and 3’-flanking regions of the target genes were amplified using primers 5F and 5R and primers 3F and 3R , respectively , and C . neoformans H99 genomic DNA as a template . Primers M13F and M13R were used to amplify the selectable markers conferring resistance to nourseothricin ( NAT1 ) ( Werner BioAgents ) and G418 ( NEO ) ( Gold Biotechnology ) [70] . The final constructs for the gene disruption cassettes were generated by means of overlap PCR performed using primers NF and NR and the 5’ and 3’-flanking regions and markers as templates . The amplified gene disruption cassettes were purified using the QIAquick Gel Extraction kit ( Qiagen ) and then combined with gold microcarrier beads ( Bio-Rad ) and introduced into the H99 or crz1Δ mutant strains using the biolistic transformation method [71] . Multiple stable transformants were selected on YPD medium containing nourseothricin or G418 , and then confirmed by diagnostic PCR for the 5’ and 3’ junctions followed by restriction enzyme digestion in each case . To generate MATa calcineurin-mutant strains , each MATα calcineurin mutant and the WT strain KN99a were mixed in equal proportions on V8 media ( pH = 5 ) and incubated in the dark at room temperature for 7 days . Basidiospores on the edges of a mating colony were removed and transferred onto YPD agar plates , and individual basidiospores were transferred onto a fresh YPD plate with G418 as described [72] . Isolated basidiospores were incubated at 30°C for 2 to 3 days and the resulting colonies were transferred to fresh YPD plates . Mating type determination was performed by PCR analysis using specific primers for SXI1α ( MATα; JOHE41440/JOHE41441 ) and SXI2a ( MATa; JOHE41442/JOHE41443 ) [73] . To generate the Pbp1-FLAG construct , the PBP1 gene region including its predicted promoter was PCR amplified using the primer pair JOHE41154 and JOHE41155 . The PCR product was then digested with NotI and cloned into pHP1 , which contains a 4x FLAG tag , the terminator of the HOG1 gene , and the hygromycin B-resistant gene . The resulting plasmid pHP4 was then introduced into the recipient pbp1Δ strains . The plasmids used in this study are listed in S8 Table . The calcineurin target genes were replaced with calcineurin target-mCherry-NEO genes by homologous recombination as described previously [74] . Calcineurin target protein-mCherry chimera cassettes were generated using a modified overlap-PCR strategy . The 5’ ( an ~1 kb region of the sequence immediately upstream of the start codon ) and 3’ ( an ~1 kb fragment of the sequence immediately downstream of the stop codon ) flanking regions of the target genes were amplified using primers 5CF and 5CR and primers 3CF and 3R , respectively . The mCherry-NEO cassette was PCR amplified from plasmid pLKB25 with primers JOHE40808 and JOHE40809 . The final PCR constructs were amplified with primers CNF and NR and were introduced into strain H99 . To construct the Crz1-mCherry fused strains for the site-directed mutagenesis , the Crz1 ORF was PCR amplified and fused with the mCherry fusion protein by splice overlap PCR , cloned into pCR2 . 1-TOPO ( Invitrogen ) to yield plasmid pXW15 which was sequenced . The 7 . 96 kb BamHI Crz1-mCherry fusion fragment was subcloned into the BamHI site of the safe haven plasmid pSDMA25 , to generate plasmid pEC13 which was sequenced . Site-directed mutations were introduced into Crz1 ORF ( pXW15 was used as the template ) by PCR mutagenesis and then subcloned into pEC13 using the Gibson Assembly Master-mix ( NEB ) . Plasmid clones of each mutagenic Crz1 allele were sequenced to ensure they contained the mutations . The recombinant Crz1-mCherry plasmid clones were linearized using the restriction enzyme AscI , and the crz1Δ::NAT deletion mutant was biolistically transformed . Targeted integration of each construct at the safe heaven locus was confirmed by PCR amplification ( 5’ junction: JOHE40956/JOHE40957; 3’ junction: JOHE40958/JOHE41562 ) . Tandem array integration was ruled out by PCR using inverse primers ( JOHE41450/JOHE41451 ) that would only yield a product if two or more copies of the construct were tandemly integrated at the safe haven site . The NAT selectable marker in pSL04 , which contains the nucleolar fluorescence marker GFP-Nop1 , was replaced with the HYG selectable marker from pJAF15 . The GFP-Nop1 construct was introduced into the recombinant Crz1-mCherry mutant strains by ectopic integration . All primers used are listed in S8 Table . To analyze sexual reproduction phenotypes , the MATα WT ( H99 ) or calcineurin-mutants and MATa WT ( KN99a ) or calcineurin mutant strains were grown overnight in liquid YPD media , washed with sterile water , mixed at equal density of cells ( 1×108 ) , spotted on MS ( Sigma ) or V8 ( pH = 5 ) solid media , and incubated in the dark for 7 to 14 days at room temperature . To examine pheromone gene expression , the MATα WT ( H99 ) or mutants and strain KN99a were mixed at equal cell densities , spread on V8 ( pH = 5 ) solid media , and incubated in the dark for 24 hour at room temperature . All samples were collected , quick-frozen , and stored at -80°C . Total RNA was isolated from each sample using Trizol reagent ( Thermo ) . Complementary DNA was synthesized using the AffinityScript QPCR cDNA Synthesis Kit ( Agilent ) . Quantitative real-time PCR was performed with each gene-specific primer set and Brilliant III Ultra-Fast SYBR QPCR Mix ( Agilent ) and using a StepOnePlus Real-Time PCR system ( ABI ) . Cryptococcus strains were grown overnight in YPD medium at 30°C . The cells were collected , washed twice with sterile PBS , counted with a hemocytometer , and the final density was adjusted to 1×107 CFU/ml . Groups of 10 female A/J mice ( Jackson Labs or NCI/Charles River Laboratories , 16~20 g ) per strain were anesthetized with pentobarbital ( Lundbeck Inc . Deerfield , IL ) , and inoculated with 5 × 104 or 5 × 105 CFU in a volume of 50 μl via intranasal inhalation as previously described [75] . Survival was monitored daily , and moribund mice were sacrificed with CO2 . Survival curves were generated according to the Kaplan-Meier method by the Prism 4 . 0 ( GraphPad software ) and statistical significance ( p values ) assessed with the log-rank test . All experiments and animal care were conducted in accordance with the ethical guidelines of the Institutional Animal Care and Use Committee ( IACUC ) of Duke University Medical Center ( DUMC ) . The DUMC IACUC approved all of the vertebrate studies under protocol number A245-13-09 . Mice studies were conducted in the Division of Laboratory Animal Resources ( DLAR ) facilities at DUMC and animals were handled according to the guidelines defined by the United States Animal Welfare Act and in full compliance with the DUMC IACUC . WT ( H99 ) and cna1Δ ( KK1 ) cells were grown in YPD at 25°C to an optical density OD600 = 0 . 8~1 ( log phase ) . Next , the WT cell culture was divided in thirds whereas the cna1 culture was split in half , and one culture for each strain was incubated at 25°C while the rest of the cultures were transferred to 37°C for 1 hour . For FK506 treatment , one of the WT cultures was exposed to 2 μg/ml FK506 for 15 minutes prior to and during the shift to 37°C . Note that all cultures were conducted with 2 biological replicates for each condition analyzed . Following 1 hr incubation , cells were treated with 6% TCA ( Sigma ) and incubated on ice for 30 minutes to stop metabolic activity . Cells were collected by centrifugation , washed twice with cold acetone , and dried under vacuum . The dried pellets were resuspended in 500 μl of lysis buffer ( 50 mM Tris-Cl pH 7 . 5 , 7 M urea , 10 mM NaF , 10 mM p-nitrophenylphosphate , 10 mM NaP2O4 , 10 mM glycerophosphate , 1X Roche protease inhibitor cocktail and 1 mM PMSF ) using a mini-bead beater ( BioSpec ) for 8 cycles ( 60 sec homogenization with 2 min rest ) . Protein was quantitated by a Bradford assay ( Bio-Rad ) . Cell lysates were centrifuged at 14 , 000 rpm for 15 min at 4°C , supernatants were recovered , and protein was quantitated by a Bradford assay ( Bio-Rad ) . For trypsin digestions 1 . 0 mg of protein from each sample was diluted 4 . 67 fold with 50 mM ammonium bicarbonate ( Sigma ) to achieve a final urea concentration of 1 . 5 M . Samples were treated with 20 mM DTT for 30 min at 70°C and alkylated with 40 mM iodoacetamide for 45 min at room temperature . Trypsin digestion ( at a 1:50 w/w ) was allowed to occur for 18 hr at 37°C . Digested samples were then acidified and a 600 μg total protein aliquot was removed from each sample . The remainder of the sample ( total of 3 . 49 mg ) was pooled and divided into three 600 μg aliquots to access TiO2 enrichment reproducibility ( Enrichment quality control ( QC ) ) and analytical reproducibility ( Analytical QC ) . Each sample was desalted using a 50 mg Sep-Pack C18 Solid-Phase Extraction ( SPE ) cartridge . Briefly , samples were loaded onto pre-equilibrated SPE samples in acidified digestion buffer and washed twice 1 ml with 0 . 1% trifluoroacetic acid ( TFA ) in water . Samples were eluted in 1 ml 80% acetonitrile ( MeCN ) , 0 . 1% TFA , partially dried using vacuum centrifugation and then brought to complete dryness using lyophilization . The crz1Δ + Crz1WT strain ( AFA3-3-3 ) was grown in liquid YPD media at 24°C overnight . Cells were divided into three aliquots and diluted with fresh media ( 50 mL each ) to an optical density of OD600 = 0 . 3 . Cultures were further incubated at 24°C until an optical density of OD600 = 0 . 8 was reached . To one aliquot , FK506 ( 1 mg/mL final concentration ) was added incubated at 24°C for 30 minutes prior to the temperature shift to 37°C . Two cultures , including the FK506-treated aliquot , were transferred to shaking water bath equilibrated to 37°C , and incubated for 30 minutes , while the third culture was kept at 24°C . After the 30 minute incubation , the cells from each treatment condition were collected by centrifugation , and washed once with ice-cold water , and then twice with ice-cold lysis buffer ( 10 mM Tris/HCl , pH 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA , 1 mM PMSF , 1% Triton X100 , 1X Roche protease inhibitor ) . Cells were pelleted and snap-frozen at -80°C . Frozen pellet was thawed on ice , with 500 μL of lysis buffer and 700 μL of acid-washed glass beads added . Cells were lysed using a mini-bead beater for 10 cycles ( 90 seconds homogenization with 2 minutes rest intervals ) and the cell lysates cleared by centrifugation at 3 , 000 rpm for 10 minutes , at 4°C . The cell lysates were then further cleared of membrane-associated material by a second round of centrifugation at 14 , 000 rpm for 20 minutes . The RFP-Trap® beads ( Chromotek , Planegg , Germany ) were equilibrated following manufacturer’s protocol . 30 μL aliquots of the equilibrated beads were added to the cell lysates , and the lysate + bead suspensions were incubated with constant rotation at 4°C for 2 hours . The beads were pelleted by centrifugation at 700 rpm for 1 minute , and the supernatant removed . The beads were washed twice with lysis buffer , twice with wash buffer ( 10 mM Tris/HCl , pH 7 . 5 , 500 mM NaCl , 0 . 5 mM EDTA ) , and twice with lysis buffer without the protease inhibitor cocktail . The beads were resuspended in ice-cold PBS with 1X Roche protease inhibitor cocktail , and 1 mM PMSF added and submitted to the Duke Proteomic Core Facility for phosphopeptide enrichment and phosphosite determination . Each sample was resuspended in 150 μl of 80% MeCN , 1% TFA , 1 M glycolic acid and 30 fmol/μg of pre-trypsin digested bovine alpha casein . TiO2 enrichments were performed on GL Sciences p200 TiO2 spin columns following the manufacturer’s directions . Eluents were acidified with ~6 μl of formic acid . Samples were then frozen and lyophilized . Prior to LC-MS analysis , samples were resuspended in 10 μl of 200 mM ammonium format , 10 mM sodium citrate . A separate Analytical quality control ( QC sample ) ( performed in triplicate ) was generated by removing 5 μl from each of the Enrichment QC samples . Quantitative two-dimensional liquid chromatography-tandem mass spectrometry ( LC/LC-MS/MS ) was performed on 5 μl ( 50% ) of each enriched sample . The method used two-dimensional liquid chromatography in a high-low pH reversed phase/reversed phase configuration on a nanoAcquity UPLC system ( Waters Corp ) coupled to a Synapt G2 HDMS high resolution accurate mass tandem mass spectrometer ( Waters Corp . ) with nanoelectrospray ionization as previously described [76–78] . Peptides were first trapped at 2 μl/min at 97/3 v/v water/MeCN in 20 mM ammonium formate ( pH = 10 ) on a 5 μm XBridge BEH130 C18 300 um × 50 mm column ( Waters ) . A series of step-elutions of MeCN at 2 μl/min was used to elute peptides from the first dimension column . Three steps of 4 . 7% , 9 . 4% , and 30% MeCN were employed; these percentages were optimized for delivery of an approximately equal load to the second dimension column for each fraction . For the second dimension separation , the eluent from the first dimension was first diluted 10-fold online with 99 . 8/0 . 1/0 . 1 v/v/v water/MeCN/formic acid and trapped on a 5 μm Symmetry C18 180 μm × 20 mm trapping column ( Waters ) . The second dimension separations were performed on a 1 . 7 μm Acquity BEH130 C18 75 μm x 150 mm column ( Waters ) using a linear gradient of 3 to 30% MeCN with 0 . 1% formic acid over 36 min , at a flow rate of 0 . 5 μl/min and column temperature of 35°C . Data collection on the Synapt G2 mass spectrometer was performed in a data-dependent acquisition ( DDA ) mode , using 0 . 6 second MS scans with three subsequent 0 . 3 MS/MS scans in Resolution mode . CID fragmentation settings were charge state dependent . A dynamic exclusion list of 120 sec was employed to increase unique MS/MS triggers . The total analysis cycle time for each sample injection was approximately 4 hours . Two additional DDA acquisitions were performed on the pooled samples operating in sensitivity mode ( two with 0 . 3s MS/MS scans and two with 0 . 6s MS/MS scans ) . Data was imported into Rosetta Elucidator v3 . 3 ( Rosetta Biosoftware , Inc ) , and all LC/LC-MS runs were aligned based on the accurate mass and retention time of detected ions ( “features” ) using PeakTeller algorithm ( Elucidator ) . The relative peptide abundance was calculated based on area-under-the-curve ( AUC ) of aligned features across all runs and LC-MS feature maps ( phosphorylation profiles ) were generated for each analyzed condition . The overall dataset had 720 , 304 quantified features , and high collision energy ( peptide fragment ) data was collected in 105 , 784 spectra for sequencing by database searching . This MS/MS data was searched against the Cryptococcus neoformans H99 ( http://www . broadinstitute . org/annotation/genome/cryptococcus_neoformans/MultiHome . html ) database ( n = 6954 entries ) , which also contained a reversed-sequence “decoy” database for false positive rate determination . Included in the database searches were variable modifications on methionine ( oxidation , indicated by a M[147 . 035] ) , Asn/Qln ( deamidation , indicated by a N[115 . 0269] or Q[129 . 0426] ) , and Ser/Thr/Tyr ( phosphorylation , indicated by a S[166 . 9984] , T[181 . 014] or Y[243 . 0297] ) . Individual peptide scoring using a Mascot Ion Score of 13 . 8 resulted in a 1 . 0% peptide false discovery rate . To assess the workflow and technical reproducibility of this study we used three different approaches . First , based on the intensity data for six quality control samples ( QC ) samples ( spiked bovine albumin in each of the two duplicates , which subsequently were analyzed in three technical replicates ) included in this analysis , we calculated an average coefficient variation ( CV ) of intensity of 40 . 4% . In the second approach , we accessed the variability in each phosphopeptide peak area intensities across the three technical replicates analyzed from the same sample . This resulted in an average relative standard deviation ( RSD ) of 24% . The third assessment was to calculate the average variability in all of the phosphopeptides peak area intensities across a pooled sample , which was subjected to three separate TiO2 enrichments . This resulted in an average RSD in phosphopeptide peak area of 24 . 2% . A 2-fold cutoff is suitable unless the %CV of the QC samples is substantially lower or higher than 36% . There was not a power calculation involved . The strains expressing Crz1-FLAG or Pbp1-FLAG were grown in YPD at 24°C to an optical density OD600 0 . 6–0 . 8 with or without FK506 ( 1 μg/ml ) . The culture was divided into two halves , and one half was incubated at 24°C and the other half at 37°C for 1 hour . Cells were collected by centrifugation , rapidly chilled using dry ice , and disrupted in lysis buffer ( 50 mM Tris-HCl pH = 7 . 5 , 150 NaCl , 0 . 5 mM EDTA , 0 . 5% Triton X-100 supplemented with protease inhibitor tablet ( Roche ) and phosphatase inhibitor cocktails ( Thermo ) ) using a mini-bead beater for 10 cycles ( 90 sec homogenization with 2 min rest ) . Cell extracts were centrifuged for 15 min at 14 . 000×g , the supernatant was recovered and protein concentration was determined by employing the BioRad Bradford reagent . The Crz1-FLAG protein was immunoprecipitated from the supernatant by incubating with anti-FLAG M2 affinity gel beads ( Sigma ) according to the manufacturer’s instructions . The immunoprecipitates were collected by centrifugation and washed with PBS ( Sigma ) supplemented with protease inhibitors . For lambda phosphatase assays , the Crz1-FLAG immunoprecipitates bound to beads were incubated with PMP buffer ( 50 mM HEPES pH 7 . 5 , 100 mM NaCl , 2 mM DTT , 0 . 01% Brij 35 , 1 mM MnCl2 ) and 400 units of lambda protein phosphatase ( New England BioLabs ) ( with or without phosphatase inhibitor cocktail ) for 1 hour at 30°C . For calcineurin assays , the Crz1-FLAG beads were incubated in PMP buffer containing 400 units of human calcineurin ( Enzo Life Sciences Inc . ) at 30°C for 1 hour . The Crz1-FLAG beads were resuspended in Laemmli sample buffer ( Bio-Rad ) , boiled for 10 min , briefly centrifuged , and the supernatant was resolved by SDS-PAGE and transferred to PVDF membranes ( Bio-Rad ) . Membranes were assayed by western blot employing mouse monoclonal anti-FLAG M2 antibodies ( Sigma ) , followed by anti-mouse antibody conjugated to horseradish peroxidase , and ECL western blotting detection reagent ( GE Healthcare ) . Images of sexual hyphae were captured by using a Nikon Eclipse E400 microscope equipped with a Nikon DXM1200F camera . For fluorescence imaging of cells , unless otherwise indicated , the cell suspension was placed on a slide containing a 2% agar patch and covered with a coverslip . Fluorescence images were obtained using Deltavision system ( Olympus IX-71 base ) equipped with a Coolsnap HQ2 high resolution SSD camera . Images were processed using FIJI software . | Calcineurin is a Ca2+/calmodulin-dependent protein phosphatase essential for stress survival , sexual development , and virulence of the human fungal pathogen Cryptococcus neoformans and other major pathogenic fungi of global human health relevance . However , no calcineurin substrates are known in pathogenic fungi . Employing state-of-the-art phosphoproteomic approaches we identified calcineurin substrates , including calcineurin itself and the conserved Crz1 transcriptional activator known to function in calcium signaling and stress survival . Remarkably , our study also identified novel calcineurin targets involved in RNA processing , stability , and translation , which colocalize together with calcineurin in stress granules/P-bodies upon thermal stress . These findings support a model whereby calcineurin functions in a branched pathway , via Crz1 and several of the identified novel targets , that governs transcriptional and posttranscriptional circuits to drive stress survival , sexual development , and fungal virulence . Our study underscores C . neoformans as an experimental model to define basic paradigms of calcineurin signaling in global thermostress responsive virulence networks that can be targeted for fungal therapy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"phosphorylation",
"cryptococcus",
"neoformans",
"medicine",
"and",
"health",
"sciences",
"cryptococcus",
"classical",
"mechanics",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"microbiology",
"sexual",
"reproduction",
"mechanical",
"stress",
"enzymology",
"phosphatases",
"developmental",
"biology",
"fungi",
"mutation",
"model",
"organisms",
"fungal",
"pathogens",
"saccharomyces",
"research",
"and",
"analysis",
"methods",
"mycology",
"mutant",
"strains",
"proteins",
"medical",
"microbiology",
"thermal",
"stresses",
"microbial",
"pathogens",
"modes",
"of",
"reproduction",
"yeast",
"physics",
"biochemistry",
"post-translational",
"modification",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"physical",
"sciences",
"saccharomyces",
"cerevisiae",
"organisms"
] | 2016 | Calcineurin Targets Involved in Stress Survival and Fungal Virulence |
Reward-modulated spike-timing-dependent plasticity ( STDP ) has recently emerged as a candidate for a learning rule that could explain how behaviorally relevant adaptive changes in complex networks of spiking neurons could be achieved in a self-organizing manner through local synaptic plasticity . However , the capabilities and limitations of this learning rule could so far only be tested through computer simulations . This article provides tools for an analytic treatment of reward-modulated STDP , which allows us to predict under which conditions reward-modulated STDP will achieve a desired learning effect . These analytical results imply that neurons can learn through reward-modulated STDP to classify not only spatial but also temporal firing patterns of presynaptic neurons . They also can learn to respond to specific presynaptic firing patterns with particular spike patterns . Finally , the resulting learning theory predicts that even difficult credit-assignment problems , where it is very hard to tell which synaptic weights should be modified in order to increase the global reward for the system , can be solved in a self-organizing manner through reward-modulated STDP . This yields an explanation for a fundamental experimental result on biofeedback in monkeys by Fetz and Baker . In this experiment monkeys were rewarded for increasing the firing rate of a particular neuron in the cortex and were able to solve this extremely difficult credit assignment problem . Our model for this experiment relies on a combination of reward-modulated STDP with variable spontaneous firing activity . Hence it also provides a possible functional explanation for trial-to-trial variability , which is characteristic for cortical networks of neurons but has no analogue in currently existing artificial computing systems . In addition our model demonstrates that reward-modulated STDP can be applied to all synapses in a large recurrent neural network without endangering the stability of the network dynamics .
Numerous experimental studies ( see [1] for a review; [2] discusses more recent in-vivo results ) have shown that the efficacy of synapses changes in dependence of the time difference Δt = tpost−tpre between the firing times tpre and tpost of the pre- and postsynaptic neurons . This effect is called spike-timing-dependent plasticity ( STDP ) . But a major puzzle for understanding learning in biological organisms is the relationship between experimentally well-established rules for STDP on the microscopic level , and adaptive changes of the behavior of biological organisms on the macroscopic level . Neuromodulatory systems , which send diffuse signals related to reinforcements ( rewards ) and behavioral state to several large networks of neurons in the brain , have been identified as likely intermediaries that relate these two levels of plasticity . It is well-known that the consolidation of changes of synaptic weights in response to pre- and postsynaptic neuronal activity requires the presence of such third signals [3] , [4] . In particular , it has been demonstrated that dopamine ( which is behaviorally related to novelty and reward prediction [5] ) gates plasticity at corticostriatal synapses [6] , [7] and within the cortex [8] . It has also been shown that acetylcholine gates synaptic plasticity in the cortex ( see for example [9] and [10] , [11] contains a nice review of the literature ) . Corresponding spike-based rules for synaptic plasticity of the form ( 1 ) have been proposed in [12] and [13] ( see Figure 1 for an illustration of this learning rule ) , where wji is the weight of a synapse from neuron i to neuron j , cji ( t ) is an eligibility trace of this synapse which collects weight changes proposed by STDP , and d ( t ) = h ( t ) −h̅ results from a neuromodulatory signal h ( t ) with mean value h̅ . It was shown in [12] that a number of interesting learning tasks in large networks of neurons can be accomplished with this simple rule in Equation 1 . It has recently been shown that quite similar learning rules for spiking neurons arise when one applies the general framework of distributed reinforcement learning from [14] to networks of spiking neurons [13] , [15] , or if one maximizes the likelihood of postsynaptic firing at desired firing times [16] . However no analytical tools have been available , which make it possible to predict for what learning tasks , and under which parameter settings , reward-modulated STDP will be successful . This article provides such analytical tools , and demonstrates their applicability and significance through a variety of computer simulations . In particular , we identify conditions under which neurons can learn through reward-modulated STDP to classify temporal presynaptic firing patterns , and to respond with particular spike patterns . We also provide a model for the remarkable operant conditioning experiments of [17] ( see also [18] , [19] ) . In the simpler ones of these experiments the spiking activity of single neurons ( in area 4 of the precentral gyrus of monkey cortex ) was recorded , the deviation of the current firing rate of an arbitrarily selected neuron from its average firing rate was made visible to the monkey through the displacement of an illuminated meter arm , whose rightward position corresponded to the threshold for the feeder discharge . The monkey received food rewards for increasing ( or in alternating trials for decreasing ) the firing rate of this neuron . The monkeys learnt quite reliably ( within a few minutes ) to change the firing rate of this neuron in the currently rewarded direction . Adjacent neurons tended to change their firing rate in the same direction , but also differential changes of directions of firing rates of pairs of neurons are reported in [17] ( when these differential changes were rewarded ) . For example , it was shown in Figure 9 of [17] ( see also Figure 1 in [19] ) that pairs of neurons that were separated by no more than a few hundred microns could be independently trained to increase or decrease their firing rates . Obviously the existence of learning mechanisms in the brain which are able to solve this extremely difficult credit assignment problem provides an important clue for understanding the organization of learning in the brain . We examine in this article analytically under what conditions reward-modulated STDP is able to solve such learning problem . We test the correctness of analytically derived predictions through computer simulations of biologically quite realistic recurrently connected networks of neurons , where an increase of the firing rate of one arbitrarily selected neuron within a network of 4000 neurons is reinforced through rewards ( which are sent to all 142813 synapses between excitatory neurons in this recurrent network ) . We also provide a model for the more complex operant conditioning experiments of [17] by showing that pairs of neurons can be differentially trained through reward-modulated STDP , where one neuron is rewarded for increasing its firing rate , and simultaneously another neuron is rewarded for decreasing its firing rate . More precisely , we increased the reward signal d ( t ) which is transmitted to all synapses between excitatory neurons in the network whenever the first neuron fired , and decreased this reward signal whenever the second neuron fired ( the resulting composed reward corresponds to the displacement of the meter arm that was shown to the monkey in these more complex operant conditioning experiments ) . Our theory and computer simulations also show that reward-modulated STDP can be applied to all synapses within a large network of neurons for long time periods , without endangering the stability of the network . In particular this synaptic plasticity rule keeps the network within the asynchronous irregular firing regime , which had been described in [20] as a dynamic regime that resembles spontaneous activity in the cortex . Another interesting aspect of learning with reward-modulated STDP is that it requires spontaneous firing and trial-to-trial variability within the networks of neurons where learning takes place . Hence our learning theory for this synaptic plasticity rule provides a foundation for a functional explanation of these characteristic features of cortical network of neurons that are undesirable from the perspective of most computational theories .
In this section , we derive a learning equation for reward-modulated STDP . This learning equation relates the change of a synaptic weight wji over some sufficiently long time interval T to statistical properties of the joint distribution of the reward signal d ( t ) and pre- and postsynaptic firing times , under the assumption that the weight and correlations between pre- and postsynaptic spike times are slowly varying in time . We treat spike times as well as the reward signal d ( t ) as stochastic variables . This mathematical framework allows us to derive the expected weight change over some time interval T ( see [21] ) , with the expectation taken over realizations of the stochastic input- and output spike trains as well as stochastic realizations of the reward signal , denoted by the ensemble average 〈·〉 E ( 5 ) where we used the abbreviation . If synaptic plasticity is sufficiently slow , synaptic weights integrate a large number of small changes . In this case , the weight wji can be approximated by its average 〈wji〉 E ( it is “self-averaging” , see [21] ) . We can thus drop the expectation on the left hand side of Equation 5 and write it as . Using Equation 1 , this yields ( see Methods ) ( 6 ) This formula contains the reward correlation for synapse ji ( 7 ) which is the average reward at time t given a presynaptic spike at time t−s−r and a postsynaptic spike at time t−s . The joint firing rate νji ( t , r ) = 〈Sj ( t ) Si ( t−r ) 〉 E describes correlations between spike timings of neurons j and i , i . e . , it is the probability density for the event that neuron i fires an action potential at time t−r and neuron j fires an action potential at time t . For synapses subject to reward-modulated STDP , changes in efficacy are obviously driven by co-occurrences of spike pairings and rewards within the time scale of the eligibility trace . Equation 6 clarifies how the expected weight change depends on how the correlations between the pre- and postsynaptic neurons correlate with the reward signal . If one assumes for simplicity that the impact of a spike pair on the eligibility trace is always triggered by the postsynaptic spike , one gets a simpler equation ( see Methods ) ( 8 ) The assumption introduces a small error for post-before-pre spike pairs , because for a reward signal that arrives at some time dr after the pairing , the weight update will be proportional to fc ( dr ) instead of fc ( dr+r ) . The approximation is justified if the temporal average is performed on a much longer time scale than the time scale of the learning window , the effect of each pre-post spike pair on the reward signal is delayed by an amount greater than the time scale of the learning window , and fc changes slowly compared to the time scale of the learning window ( see Methods for details ) . For the analyzes presented in this article , the simplified Equation 8 is a good approximation for the learning dynamics . Equation 8 is a generalized version of the STDP learning equation in [21] that includes the impact of the reward correlation weighted by the eligibility function . To see the relation between standard STDP and reward-modulated STDP , consider a constant reward signal d ( t ) = d0 . Then also the reward correlation is constant and given by D ( t , s , r ) = d0 . We recover the standard STDP learning equation scaled by d0 if the eligibility function is an instantaneous delta-pulse fc ( s ) = δ ( s ) . Furthermore , if the statistics of the reward signal d ( t ) is time-independent and independent from the pre- and postsynaptic spike statistics of some synapse ji , then the reward correlation is given by Dji ( t , s , r ) = 〈d ( t ) 〉 E = d0 for some constant d0 . Then , the weight change for synapse ji is . The temporal average of the joint firing rate 〈νji ( t−s , r〉 T is thus filtered by the eligibility trace . We assumed in the preceding analysis that the temporal average is taken over some long time interval T . If the time scale of the eligibility trace is much smaller than this time interval T , then the weight change is approximately , and the weight wji will change according to standard STDP scaled by a constant proportional to the mean reward and the integral over the eligibility function . In the remainder of this article , we will always use the smooth time-averaged weight change , but for brevity , we will drop the angular brackets and simply write . The learning Equation 8 provides the mathematical basis for our following analyses . It allows us to determine synaptic weight changes if we can describe a learning situation in terms of reward correlations and correlations between pre- and postsynaptic spikes . We now apply the preceding analysis to the biofeedback experiment of [17] that were described in the introduction . These experiments pose the challenge to explain how learning mechanisms in the brain can detect and exploit correlations between rewards and the firing activity of one or a few neurons within a large recurrent network of neurons ( the credit assignment problem ) , without changing the overall function or dynamics of the circuit . We show that this phenomenon can in principle be explained by reward-modulated STDP . In order to do that , we define a model for the experiment which allows us to formulate an equation for the reward signal d ( t ) . This enables us to calculate synaptic weight changes for this particular scenario . We consider as model a recurrent neural circuit where the spiking activity of one neuron k is recorded by the experimenter ( Experiments where two neurons are recorded and reinforced were also reported in [17] . We tested this case in computer simulations ( see Figure 2 ) but did not treat it explicitly in our theoretical analysis ) . We assume that in the monkey brain a reward signal d ( t ) is produced which depends on the visual feedback ( through an illuminated meter , whose pointer deflection was dependent on the current firing rate of the randomly selected neuron k ) as well as previously received liquid rewards , and that this signal d ( t ) is delivered to all synapses in large areas of the brain . We can formalize this scenario by defining a reward signal which depends on the spike rate of the arbitrarily selected neuron k ( see Figure 3A and 3B ) . More precisely , a reward pulse of shape εr ( r ) ( the reward kernel ) is produced with some delay dr every time the neuron k produces an action potential ( 9 ) Note that d ( t ) = h ( t ) −h̅ is defined in Equation 1 as a signal with zero mean . In order to satisfy this constraint , we assume that the reward kernel εr has zero mass , i . e . , . For the analysis , we use the linear Poisson neuron model described in Methods . The mean weight change for synapses to the reinforced neuron k is then approximately ( see Methods ) ( 10 ) This equation describes STDP with a learning rate proportional to . The outcome of the learning session will strongly depend on this integral and thus on the form of the reward kernel εr . In order to reinforce high firing rates of the reinforced neuron we have chosen a reward kernel with a positive bump in the first few hundred milliseconds , and a long negative tail afterwards . Figure 3C shows the functions fc and εr that were used in our computer model , as well as the product of these two functions . One sees that the integral over the product is positive and according to Equation 10 the synapses to the reinforced neuron are subject to STDP . This does not guarantee an increase of the firing rate of the reinforced neuron . Instead , the changes of neuronal firing will depend on the statistics of the inputs . In particular , the weights of synapses to neuron k will not increase if that neuron does not fire spontaneously . For uncorrelated Poisson input spike trains of equal rate , the firing rate of a neuron trained by STDP stabilizes at some value which depends on the input rate ( see [24] , [25] ) . However , in comparison to the low spontaneous firing rates observed in the biofeedback experiment [17] , the stable firing rate under STDP can be much higher , allowing for a significant rate increase . It was shown in [17] that also low firing rates of a single neuron can be reinforced . In order to model this , we have chosen a reward kernel with a negative bump in the first few hundred milliseconds , and a long positive tail afterwards , i . e . we inverted the kernel used above to obtain a negative integral . According to Equation 10 this leads to anti-STDP where not only inputs to the reinforced neuron which have low correlations with the output are depressed ( because of the negative integral of the learning window ) , but also those which are causally correlated with the output . This leads to a quick firing rate decrease at the reinforced neuron . The mean weight change of synapses to non-reinforced neurons j≠k is given by ( 11 ) where νj ( t ) = 〈Sj ( t ) 〉 E is the instantaneous firing rate of neuron j at time t . This equation indicates that a non-reinforced neuron is trained by STDP with a learning rate proportional to its correlation with the reinforced neuron given by νkj ( t−dr−r′ , s−dr−r′ ) /νj ( t−s ) . In fact , it was noted in [17] that neurons nearby the reinforced neuron tended to change their firing rate in the same direction . This observation might be explained by putative correlations of the recorded neuron with nearby neurons . On the other hand , if a neuron j is uncorrelated with the reinforced neuron k , we can decompose the joint firing rate into νkj ( t−dr−r′ , s−dr−r′ ) = νk ( t−dr−r′ ) νj ( t−s ) . In this case , the learning rate for synapse ji is approximately zero ( see Methods ) . This ensures that most neurons in the circuit keep a constant firing rate , in spite of continuous weight changes according to reward-modulated STDP . Altogether we see that the weights of synapses to the reinforced neuron k can only change if there is spontaneous activity in the network , so that in particular also this neuron k fires spontaneously . On the other hand the spontaneous network activity should not consist of repeating large-scale spatio-temporal firing patterns , since that would entail correlations between the firing of neuron k and other neurons j , and would lead to similar changes of synapses to these other neurons j . Apart from these requirements on the spontaneous network activity , the preceding theoretical results predict that stability of the circuit is preserved , while the neuron which is causally related to the reward signal is trained by STDP , if is positive . We tested these theoretical predictions through computer simulations of a generic cortical microcircuit receiving a reward signal which depends on the firing of one arbitrarily chosen neuron k from the circuit ( reinforced neuron ) . The circuit was composed of 4000 LIF neurons , with 3200 being excitatory and 800 inhibitory , interconnected randomly by 228954 conductance based synapses with short term dynamics ( All computer simulations were also carried out as a control with static current based synapses , see Methods and Suppl . ) . In addition to the explicitly modeled synaptic connections , conductance noise ( generated by an Ornstein-Uhlenbeck process ) was injected into each neuron according to data from [26] , in order to model synaptic background activity of neocortical neurons in-vivo ( More precisely , for 50% of the excitatory neurons the amplitude of the noise injection was reduced to 20% , and instead their connection probabilities from other excitatory neurons were chosen to be larger , see Methods and Figure S1 and Figure S2 for details . The reinforced neuron had to be chosen from the latter population , since reward-modulated STDP does not work properly if the postsynaptic neuron fires too often because of directly injected noise ) . This background noise elicited spontaneous firing in the circuit at about 4 . 6 Hz . Reward-modulated STDP was applied continuously to all synapses which had excitatory presynaptic and postsynaptic neurons , and all these synapses received the same reward signal . The reward signal was modeled according to Equation 9 . Figure 3C shows one reward pulse caused by a single postsynaptic spike at time t = 0 with the parameters used in the experiment . For several postsynaptic spikes , the amplitude of the reward signal follows the firing rate of the reinforced neuron , see Figure 3B . This model was simulated for 20 minutes of biological time . Figure 4A , 4B , and 4D show that the firing rate of the reinforced neuron increases within a few minutes ( like in the experiment of [17] ) , while the firing rates of the other neurons remain largely unchanged . The increase of weights to the reinforced neuron shown in Figure 4C can be explained by the correlations between its presynaptic and postsynaptic spikes shown in panel E . This panel shows that pre-before-post spike pairings ( black curve ) are in general more frequent than post-before-pre spike pairings . The reinforced neuron increases its rate from around 4 Hz to 12 Hz , which is comparable to the measured firing rates in [15] before and after learning . In Figure 9 of [17] and Figure 1 of [19] the results of another experiment were reported where the activity of two adjacent neurons was recorded , and high firing rates of the first neuron and low firing rates of the second neuron were reinforced simultaneously . This kind of differential reinforcement resulted in an increase and decrease of the firing rates of the two neurons correspondingly . We implemented this type of reinforcement by letting the reward signal in our model depend on the spikes of the two randomly chosen neurons ( we refer to these neurons as neuron A and neuron B ) , i . e . , where is the component that positively rewards spikes of neuron A , and negatively rewards spikes of neuron B . Both parts of the reward signal , and , were defined as in Equation 9 for the corresponding neuron . For we used the reward kernel εr as defined in Equation 29 , whereas for we used εr− = −εr ( note that the integral over εr− is still zero ) . At the middle of the simulation ( simulation time t = 10 min ) , we changed the direction of the reinforcements by negatively rewarding the firing of neuron A and positively rewarding the firing of neuron B ( i . e . , ) . The results are summarized in Figure 2 . With a reward signal modeled in this way , we were able to independently increase and decrease the firing rates of the two neurons according to the reinforcements , while the firing rates of the other neurons remained unchanged . Changing the type of reinforcement during the simulation from positive to negative for neuron A and from negative to positive for neuron B resulted in a corresponding shift in their firing rate change in the direction of the reinforcement . The dynamics of a network where STDP is applied to all synapses between excitatory neurons is quite sensitive to the specific choice of the STDP-rule . The preceding theoretical analysis ( see Equations 10 and 11 ) predicts that reward-modulated STDP affects in the long run only those excitatory synapses where the firing of the postsynaptic neuron is correlated with the reward signal . In other words: the reward signal gates the effect of STDP in a recurrent network , and thereby can keep the network within a given dynamic regime . This prediction is confirmed qualitatively by the two panels of Figure 4A , which show that even after all excitatory synapses in the recurrent network have been subject to 20 minutes ( in simulated biological time ) of reward-modulated STDP , the network stays within the asynchronous irregular firing regime . It is also confirmed quantitatively through Figure 5 . These figures show results for the simple additive version of STDP ( according to Equation 3 ) . Very similar results ( see Figure S3 and Figure S4 ) arise from an application of the more complex STDP-rule proposed in [22] where the weight-change depends on the current weight value . The preceding model for the biofeedback experiment of Fetz and Baker focused on learning of firing rates . In order to explore the capabilities and limitations of reward-modulated STDP in contexts where the temporal structure of spike trains matters , we investigated another reinforcement learning scenario where a neuron should learn to respond with particular temporal spike patterns . We first apply analytical methods to derive conditions under which a neuron subject to reward-modulated STDP can achieve this . In this model , the reward signal d ( t ) is given in dependence on how well the output spike train of a neuron j matches some rather arbitrary spike train S* ( which might for example represent spike output from some other brain structure during a developmental phase ) . S* is produced by a neuron μ* that receives the same n input spike trains S1 , … , Sn as the trained neuron j , with some arbitrarily chosen weights , . But in addition the neuron μ* receives n′−n further spike trains Sn+1 , … , Sn′ with weights . The setup is illustrated in Figure 6A . It provides a generic reinforcement learning scenario , when a quite arbitrary ( and not perfectly realizable ) spike output is reinforced , but simultaneously the performance of the learner can be evaluated clearly according to how well its weights wj1 , … , wjn match those of the neuron μ* for those n input spike trains which both of them have in common . The reward d ( t ) at time t depends in this task on both the timing of action potentials of the trained neuron and spike times in the target spike train S* ( 12 ) where the function κ ( r ) with describes how the reward signal depends on the time difference r between a postsynaptic spike and a target spike , and dr>0 is the delay of the reward . Our theoretical analysis ( see Methods ) predicts that under the assumption of constant-rate uncorrelated Poisson input statistics this reinforcement learning task can be solved by reward-modulated STDP for arbitrary initial weights if three constraints are fulfilled: ( 13 ) ( 14 ) ( 15 ) The following parameters occur in these equations: ν* is the output rate of neuron μ* , is the minimal output rate , is the maximal output rate of the trained neuron , is the integral over the eligibility trace , is the integral over the STDP learning curve ( see Equation 2 ) , is the convolution of the reward kernel with the shape of the postsynaptic potential ( PSP ) ε ( s ) , and is the integral over the PSP weighted by the learning window . If these inequalities are fulfilled and input rates are larger than zero , then the weight vector of the trained neuron converges on average from any initial weight vector to w* ( i . e . , it mimics the weight distribution of neuron μ* for those n inputs which both have in common ) . To get an intuitive understanding of these inequalities , we first examine the idea behind Constraint 13 . This constraint assures that weights of synapses i with decay to zero in expectation . First note that input spikes from a spike train Si with have no influence on the target spike train S* . In the linear Poisson neuron model , this leads to weight changes similar to STDP which can be described by two terms . First , all synapses are subject to depression stemming from the negative part of the learning curve W and random pre-post spike pairs . This weight change is bounded from below by for some positive constant α . On the other hand , the positive influence of input spikes on postsynaptic firing leads to potentiation of the synapse bounded from above by . Hence the weight decays to zero if , leading to Inequality 13 . For synapses i with , there is an additional drive , since each presynaptic spike increases the probability of a closely following spike in the target spike train S* . Therefore , the probability of a delayed reward signal after a presynaptic spike is larger . This additional drive leads to positive weight changes if Inequalities 14 and 15 are fulfilled ( see Methods ) . Note that also for the learning of spike times spontaneous spikes ( which might be regarded as “noise” ) are important , since they may lead to reward signals that can be exploited by the learning rule . It is obvious that in reward-modulated STDP , a silent neuron cannot recover from its silent state , since there will be no spikes which can drive STDP . But in addition , Condition 13 shows that in this learning scenario , the minimal output rate —which increases with increasing noise—has to be larger than some positive constant , such that depression is strong enough to weaken synapses if needed . On the other hand , if the noise is too strong also synapses i with wi = wmax will be depressed and may not converge correctly . This can happen when the increased noise leads to a maximal postsynaptic rate such that Constraints 14 and 15 are not satisfied anymore . Conditions 13–15 also reveal how parameters of the model influence the applicability of this setup . For example , the eligibility trace enters the equations only in the form of its integral and its value at the reward delay in Equation 15 . In fact , the exact shape of the eligibility trace is not important . The important property of an ideal eligibility trace is that it is high at the reward delay and low at other times as expressed by the fraction in Condition 15 . Interestingly , the formulas also show that one has quite some freedom in choosing the form of the STDP window , as long as the reward kernel εκ is adjusted accordingly . For example , instead of a standard STDP learning window W with W ( r ) ≥0 for r>0 and W ( r ) ≤0 for r<0 and a corresponding reward kernel κ , one can use a reversed learning window W′ defined by W′ ( r ) ≡W ( −r ) and a reward kernel κ′ such that εκ′ ( r ) = εκ ( −r ) . If Condition 15 is satisfied for W and κ , then it is also satisfied for W′ and κ′ ( and in most cases also Condition 14 will be satisfied ) . This reflects the fact that in reward modulated STDP the learning window defines the weight changes in combination with the reward signal . For a given STDP learning window , the analysis reveals what reward kernels κ are suitable for this learning setup . From Condition 15 , we can deduce that the integral over κ should be small ( but positive ) , whereas the integral should be large . Hence , for a standard STDP learning window W with W ( r ) ≥0 for r>0 and W ( r ) ≤0 for r<0 , the convolution εκ ( r ) of the reward kernel with the PSP should be positive for r>0 and negative for r<0 . In the computer simulation we used a simple kernel depicted in Figure 6B , which satisfies the aforementioned constraints . It consists of two double-exponential functions , one positive and one negative , with a zero crossing at some offset tκ from the origin . The optimal offset tκ is always negative and in the order of several milliseconds for usual PSP-shapes ε . We conclude that for successful learning in this scenario , a positive reward should be produced if the neuron spikes around the target spike or somewhat later , and a negative reward should be produced if the neuron spikes much too early . In order to explore this learning scenario in a biologically more realistic setting , we trained a LIF neuron with conductance based synapses exhibiting short term facilitation and depression . The trained neuron and the neuron μ* which produced the target spike train S* both received inputs from 100 input neurons emitting spikes from a constant rate Poisson process of 15 Hz . The synapses to the trained neuron were subject to reward-modulated STDP . The weights of neuron μ* were set to for 0≤i<50 and for 50≤i<100 . In order to simulate a non-realizable target response , neuron μ* received 10 additional synaptic inputs ( with weights set to wmax/2 ) . During the simulations we observed a firing rate of 18 . 2 Hz for the trained neuron , and 25 . 2 Hz for the neuron μ* . The simulations were run for 2 hours simulated biological time . We performed 5 repetitions of the experiment , each time with different randomly generated inputs and different initial weight values for the trained neuron . In each of the 5 runs , the average synaptic weights of synapses with and approached their target values , as shown in Figure 7A . In order to test how closely the trained neuron reproduces the target spike train S* after learning , we performed additional simulations where the same spike input was applied to the trained neuron before and after the learning . Then we compared the output of the trained neuron before and after learning with the output S* of neuron μ* . Figure 7B shows that the trained neuron approximates the part of S* which is accessible to it quite well . Figure 7C–F provide more detailed analyses of the evolution of weights during learning . The computer simulations confirmed the theoretical prediction that the neuron can learn well through reward-modulated STDP only if a certain level of noise is injected into the neuron ( see preceding discussion and Figure S6 ) . Both the theoretical results and these computer simulations demonstrate that a neuron can learn quite well through reward-modulated STDP to respond with specific spike patterns . Equations 13–15 predict under which relationships between the parameters involved the learning of particular spike responses through reward-modulated STDP will be successful . We have tested these predictions by selecting 6 arbitrary settings of these parameters , which are listed in Table 1 . In 4 cases ( marked by light gray shading in Figure 8 ) these conditions were not met ( either for the learning of weights with target value wmax , or for the learning of weights with target value 0 . Figure 8 shows that the derived learning result is not achieved in exactly these 4 cases . On the other hand , the theoretically predicted weight changes ( black bar ) predict in all cases the actual weight changes ( gray bar ) that occur for the chosen simulation times ( listed in the last column of Table 1 ) remarkably well . We examine here the question whether a neuron can learn through reward-modulated STDP to discriminate between two spike patterns P and N of its presynaptic neurons , by responding with more spikes to pattern P than to pattern N . Our analysis is based on the assumption that there exist internal rewards d ( t ) that could guide such pattern discrimination . This reward based learning architecture is biologically more plausible than an architecture with a supervisor which provides for each input pattern a target output and thereby directly produces the desired firing behavior of the neuron ( since the question becomes then how the supervisor has learnt to produce the desired spike outputs ) . We consider a neuron that receives input from n presynaptic neurons . A pattern X consists of n spike trains , each of time length T , one for each presynaptic neuron . There are two patterns , P and N , which are presented in alternation to the neuron , with some reset time between presentations . For notational simplicity , we assume that each of the n presynaptic spike trains consists of exactly one spike . Hence , each pattern can be defined by a list of spike times: , , where is the time when presynaptic neuron i spikes for pattern X∈{P , N} . A generalization to the easier case of learning to discriminate spatio-temporal presynaptic firing patterns ( where some presynaptic neurons produce different numbers of spikes in different patterns ) is straightforward , however the main characteristics of the learning dynamics are better accessible in this conceptually simpler setup . It had already been shown in [12] that neurons can learn through reward-modulated STDP to discriminate between different spatial presynaptic firing patterns . But in the light of the analysis of [27] it is still open whether neurons can learn with simple forms of reward-modulated STDP , such as the one considered in this article , to discriminate temporal presynaptic firing patterns . We assume that the reward signal d ( t ) rewards—after some delay dr—action potentials of the trained neuron if pattern P was presented , and punishes action potentials of the neuron if pattern N was presented . More precisely , we assume that ( 16 ) with some reward kernel εr and constants αN<0<αP . The goal of this learning task is to produce many output spikes for pattern P , and few or no spikes for pattern N . The main result of our analysis is an estimate of the expected weight change of synapse i of the trained neuron for the presentation of pattern P , followed after a sufficiently long time T′ by a presentation of pattern Nwhere 〈·〉 E |X is the expectation over the ensemble given that pattern X was presented . This weight change can be estimated as ( see Methods ) ( 17 ) where νX ( t ) is the postsynaptic rate at time t for pattern X , and the constants for X∈{P , N} are given by ( 18 ) As we will see shortly , an interesting learning effect is achieved if is positive and is negative . Since fc ( r ) is non-negative , a natural way to achieve this is to choose a positive reward kernel εr ( r ) ≥0 for r>0 and εr ( r ) = 0 for r<0 ( also , fc ( r ) and εr ( r ) must not be identical to zero for all r ) . We use Equation 17 to provide insight on when and how the classification of temporal spike patterns can be learnt with reward-modulated STDP . Assume for the moment that . We first note that it is impossible to achieve through any synaptic plasticity rule that the time integral over the membrane potential of the trained neuron has after training a larger value for input pattern P than for input pattern N . The reason is that each presynaptic neuron emits the same number of spikes in both patterns ( namely one spike ) . This simple fact implies that it is impossible to train a linear Poisson neuron ( with any learning method ) to respond to pattern P with more spikes than to pattern N . But Equation 17 implies that reward-modulated STDP increases the variance of the membrane potential for pattern P , and reduces the variance for pattern N . This can be seen as follows . Because of the specific form of the STDP learning curve W ( r ) , which is positive for ( small ) positive r , negative for ( small ) negative r , and zero for large r , has a potentiating effect on synapse i if the postsynaptic rate for pattern P is larger ( because of a higher membrane potential ) shortly after the presynaptic spike at this synapse i than before that spike . This tends to further increase the membrane potential after that spike . On the other hand , since is negative , the same situation for pattern N has a depressing effect on synapse i , which counteracts the increased membrane potential after the presynaptic spike . Dually , if the postsynaptic rate shortly after the presynaptic spike at synapse i is lower than shortly before that spike , the effect on synapse i is depressing for pattern P . This leads to a further decrease of the membrane potential after that spike . In the same situation for pattern N , the effect is potentiating , again counteracting the variation of the membrane potential . The total effect on the postsynaptic membrane potential is that the fluctuations for pattern P are increased , while the membrane potential for pattern N is flattened . For the LIF neuron model , and most reasonable other non-linear spiking neuron models , as well as for biological neurons in-vivo and in-vitro [28]–[30] , larger fluctuations of the membrane potential lead to more action potentials . As a result , reward-modulated STDP tends to increase the number of spikes for pattern P for these neuron models , while it tends to decrease the number of spikes for pattern N , thereby enabling a discrimination of these purely temporal presynaptic spike patterns . We tested these theoretical predictions through computer simulations of a LIF neuron with conductance based synapses exhibiting short-term depression and facilitation . Both patterns , P and N , had 200 input channels , with 1 spike per channel ( hence this is the extreme where all information lies in the timing of presynaptic spikes ) . The spike times were drawn from an uniform distribution over a time interval of 500 ms , which was the duration of the patterns . We performed 1000 training trials where the patterns P and N were presented to the neuron in alternation . To introduce exploration for this reinforcement learning task , the neuron had injected 20% of the Ornstein-Uhlenbeck process conductance noise ( see Methods for further details ) . The theoretical analysis predicted that the membrane potential will have after learning a higher variance for pattern P , and a lower variance for pattern N . When in our simulation of a LIF neuron the firing of the neuron was switched off ( by setting the firing threshold potential too high ) we could observe the membrane potential fluctuations undisturbed by the reset mechanism after each spike ( see Figure 9C and 9D ) . The variance of the membrane potential did in fact increase for pattern P from 2 . 49 ( mV ) 2 to 5 . 43 ( mV ) 2 ( Figure 9C ) , and decrease for pattern N ( Figure 9D ) , from 2 . 34 ( mV ) 2 to 1 . 33 ( mV ) 2 . The corresponding plots with the firing threshold included are given in panels E and F , showing an increased member of spikes of the LIF neuron for pattern P , and a decreased number of spikes for pattern N . Furthermore , as Figure 9A and 9B show , the increased variance of the membrane potential for the positively reinforced pattern P led to a stable temporal firing pattern in response to pattern P . We repeated the experiment 6 times , each time with different randomly generated patterns P and N , and different random initial synaptic weights of the neuron . The results in Figure 9G and 9H show that the learning of temporal pattern discrimination through reward-modulated STDP does not depend on the temporal patterns that are chosen , nor on the initial values of synaptic weights . A longstanding open problem is how a biologically realistic neuron model can be trained in a biologically plausible manner to extract information from a generic cortical microcircuit . Previous work [31]–[35] has shown that quite a bit of salient information about recent and past inputs to the microcircuit can be extracted by a non-spiking linear readout neuron ( i . e . , a perceptron ) that is trained by linear regression or margin maximization methods . Here we examine to what extent a LIF readout neuron with conductance based synapses ( subject to biologically realistic short term synaptic plasticity ) can learn through reward-modulated STDP to extract from the response of a simulated cortical microcircuit ( consisting of 540 LIF neurons ) , see Figure 10A , the information which spoken digit ( transformed into spike trains by a standard cochlea model ) is injected into the circuit . In comparison with the preceding task in simulation 4 , this task is easier because the presynaptic firing patterns that need to be discriminated differ in temporal and spatial aspects ( see Figure 10B; Figure S10 and S11 show the spike trains that were injected into the circuit ) . But this task is on the other hand more difficult , because the circuit response ( which creates the presynaptic firing pattern for the readout neuron ) differs also significantly for two utterances of the same digit ( Figure 10C ) , and even for two trials for the same utterance ( Figure 10D ) because of the intrinsic noise in the circuit ( which was modeled according to [26] to reflect in-vivo conditions during cortical UP-states ) . The results shown in Figure 10E–H demonstrate that nevertheless this learning experiment was successful . On the other hand we were not able to achieve in this way speaker-independent word recognition , which had been achieved in [31] with a linear readout . Hence further work will be needed in order to clarify whether biologically more realistic models for readout neurons can be trained through reinforcement learning to reach the classification capabilities of perceptrons that are trained through supervised learning .
In our theoretical analysis , we use a linear Poisson neuron model whose output spike train is a realization of a Poisson process with the underlying instantaneous firing rate Rj ( t ) . The effect of a spike of presynaptic neuron i at time t′ on the membrane potential of neuron j is modeled by an increase in the instantaneous firing rate by an amount wji ( t′ ) ε ( t−t′ ) , where ε is a response kernel which models the time course of a postsynaptic potential ( PSP ) elicited by an input spike . Since STDP according to [12] has been experimentally confirmed only for excitatory synapses , we will consider plasticity only for excitatory connections and assume that wji≥0 for all i and ε ( s ) ≥0 for all s . Because the synaptic response is scaled by the synaptic weights , we can assume without loss of generality that the response kernel is normalized to . In this linear model , the contributions of all inputs are summed up linearly: ( 19 ) where S1 , … , Sn are the n presynaptic spike trains . Since the instantaneous firing rate R ( t ) is analogous to the membrane potential of other neuron models , we occasionally refer to R ( t ) as the “membrane potential” of the neuron . In the following , we denote by the ensemble average of a random variable x given that neuron k spikes at time t and neuron i spikes at time t′ . We will also sometimes indicate the variables Y1 , Y2 , … over which the average of x is taken by writing . We assume that a reward with the functional form εr is delivered for each postsynaptic spike with a delay dr . The reward as time t is therefore The reward correlation for a synapse ki afferent to the reinforced neuron is If we assume that the output firing rate is constant on the time scale of the reward function , the first term vanishes . We rewrite the result asThe mean weight change for weights to the reinforced neuron is therefore ( 20 ) We show that the second term in the brackets is very small compared to the first term:The last approximation is based on the assumption that fc ( s ) ≈fc ( s−r′ ) and 〈νki ( t−r′ , r ) 〉 T ≈〈νki ( t , r ) 〉 T for r′∈[−TW−Tε , TW] . Here , TW is the time scale of the learning window ( see above ) , and Tε is time scale of the PSP , i . e . , we have ε ( s ) ≈0 for s≥Tε . Since by definition , we see that this is the first term in the brackets of Equation 20 scaled by wki . For neurons with many input synapses we have wki≪1 . Thus the second term in the brackets of Equation 20 is small compared to the first term . We therefore have The reward correlation of a synapse ji to a non-reinforced neuron j is given byWe havefor which we obtainIn analogy to the previous derivation , we assume here that the firing rate νj ( t−s ) in the denominator results from many PSPs . Hence , the single PSP wjiε ( r ) is small compared to νj ( t−s ) . Similarly , we assume that with weights wki , wji≪1 , the second term in the nominator is small compared to the joint firing rate νkj ( t−dr−r′ , s−dr−r′ ) . We therefore approximate the reward correlation byHence , the reward correlation of a non-reinforced neuron depends on the correlation of this neuron with the reinforced neuron . The mean weight change for a non-reinforced neuron j≠k is thereforeThis equation deserves a remark for the case that νj ( t−s ) is zero , since it appears in the denominator of the fraction . Note that in this case , both νkj ( t−dr−r′ , s−dr−r′ ) and νji ( t−s , r ) are zero . In fact , if we take the limit νj ( t−s ) →0 , then both of these factors approach zero at least as fast . Hence , in the limit of νj ( t−s ) →0 , the term in the angular brackets evaluates to zero . This reflects the fact that since STDP is driven by pre- and postsynaptic spikes , there is no weight change if no postsynaptic spikes occur . Below , we will indicate the variables Y1 , Y2 , … over which the average of x is taken by writing . From Equation 12 , we can determine the reward correlation for synapse i ( 21 ) where denotes the instantaneous firing rate of the trained neuron at time t , and ν* ( t ) = 〈S* ( t ) 〉 E denotes the instantaneous rate of the target spike train at time t . Since weights are changing very slowly , we have wji ( t−s−r ) ≈wji ( t ) . In the following , we will drop the dependence of wji on t for brevity . For simplicity , we assume that input rates are stationary and uncorrelated . In this case ( since the weights are changing slowly ) , also the correlations between inputs and outputs can be assumed stationary , νji ( t , r ) = νji ( r ) . With constant input rates , we can rewrite Equation 21 aswith . We use this results to obtain the temporally smoothed weight change for synapse ji . With stationary correlations , we can drop the dependence of νji on t and write νji ( t , r ) = νji ( r ) . Furthermore , we define and obtainWe assume that the eligibility function fc ( dr ) ≈fc ( dr+r ) if |r| is on the time scale of a PSP , the learning window , or the reward kernel , and that dr is large compared to these time scales . Then , we havewhere is the convolution of the reward kernel with the PSP . Furthermore , we findWith these simplifications , and the abbreviation we obtain the weight change at synapse jiwhere . For uncorrelated Poisson input spike trains of rate and the linear Poisson neuron model , the input-output correlations are . With these correlations , we obtain where , and . The weight change at synapse ji is then ( 22 ) We will now bound the expected weight change for synapses ji with and for synapses jk with . In this way we can derive conditions for which the expected weight change for the former synapses is positive , and that for the latter type is negative . First , we assume that the integral over the reward kernel is positive . In this case , the weight change given by Equation 22 is negative for synapses i with if and only if , and . In the worst case , wji is wmax and is small . We have to guarantee some minimal output rate such that even if wji = wmax , this inequality is fulfilled . This could be guaranteed by some noise current . Given such minimal output rate , we can state the first inequality which guarantees convergence of weights wji with For synapses ji with , we obtain two more conditions . The approximate weight change is given byThe last term in this equation is positive and small . We can ignore it in our sufficient condition . The second to last term is negative . We will include in our condition that the third to last term compensates for this negative term . Hence , the second condition iswhich should be satisfied in most setups . If we assume that this holds , we obtainwhich should be positive . We obtain the following inequalityAll three inequalities are summarized in the following:where is the maximal output rate . If these inequalities are fulfilled and input rates are positive , then the weight vector converges on average from any initial weight vector to w* . The second condition is less severe , and should be easily fulfilled in most setups . If this is the case , the first Condition 13 ensures that weights with w* = 0 are depressed while the third Condition 15 ensures that weights with w* = wmax are potentiated . We assume that a trial consists of the presentation of a single pattern starting at time t = 0 . We compute the weight change for a single trial given that pattern X∈{P , N} was presented with the help of Equations 1 , 3 , and 4 asWe can compute the average weight change given that pattern X was presented:If we assume that fc is approximately constant on the time scale of the learning window W , we can simplify this toFor the linear Poisson neuron , we can write the auto-correlation function aswhere νX ( t ) = 〈Spost ( t ) 〉 E |X is the ensemble average rate at time t given that pattern X was presented . If an experiment for a single pattern runs over the time interval [0 , T′] , we can compute the total average weight change of a trial given that pattern X was presented as ( 23 ) By definingwe can write Equation 23 asWe assume that eligibility traces and reward signals have settled to zero before a new pattern is presented . The expected weight change for the successive presentation of both patterns is thereforeThe equations can easily be generalized to the case where multiple input spikes per synapse are allowed and where jitter on the templates is allowed . However , the main effect of the rule can be read off the equations given here . We describe here the models and parameter values that were used in all our computer simulations . We will specify in a subsequent section the values of other parameters that had to be chosen differently in individual computer simulations , in dependence of their different setups and requirements of each computer simulation . For the computer simulations LIF neurons with conductance-based synapses were used . The membrane potential Vm ( t ) of this neuron model is given by: ( 24 ) where Cm is the membrane capacitance , Rm is the membrane resistance , Vresting is the resting potential , and ge , j ( t ) and gi , j ( t ) are the Ke and Ki synaptic conductances from the excitatory and inhibitory synapses respectively . The constants Ee and Ei are the reversal potentials of excitatory and inhibitory synapses . Inoise represents the synaptic background current which the neuron receives ( see below for details ) . Whenever the membrane potential reaches a threshold value Vthresh , the neuron produces a spike , and its membrane potential is reset to the value of the reset potential Vreset . After a spike , there is a refractory period of length Trefract , during which the membrane potential of the neuron remains equal to the value Vm ( t ) = Vreset . After the refractory period Vm ( t ) continues to change according to Equation 24 . For a given synapse , the dynamics of the synaptic conductance g ( t ) is defined by ( 25 ) where A ( t ) is the amplitude of the postsynaptic response ( PSR ) to a single presynaptic spike , which varies over time due to the inherent short-term dynamics of the synapse , and {t ( k ) } are the spike times of the presynaptic neuron . The conductance of the synapse decreases exponentially with time constant τsyn , and increases instantaneously by amount of A ( t ) whenever the presynaptic neuron spikes . In all computer simulations we used the following values for the neuron and synapse parameters . The membrane resistance of the neurons was Rm = 100 MΩ , the membrane capacitance Cm = 0 . 3 nF , the resting potential , reset potential and the initial value of the membrane potential had the same value of Vresting = Vreset = Vm ( 0 ) = −70 mV , the threshold potential was set to Vthresh = −59 mV and the refractory period Trefract = 5 ms . For the synapses we used a time constant set to τsyn = 5 ms , reversal potential Ee = 0 mV for the excitatory synapses and Ee = −75 mV for the inhibitory synapses . All synapses had a synaptic delay of tdelay = 1 ms . We modeled the short-term dynamics of synapses according to the phenomenological model proposed in [37] , where the amplitude Ak = A ( tk+tdelay ) of the postsynaptic response for the kth spike in a spike train with inter-spike intervals Δ1 , Δ2 , … , Δ k −1 is calculated with the following equations ( 26 ) with hidden dynamic variables u∈[0 , 1] and R∈[0 , 1] whose initial values for the 1st spike are u1 = U and R = 1 ( see [38] for a justification of this version of the equations , which corrects a small error in [37] ) . The variable w is the synaptic weight which scales the amplitudes of postsynaptic responses . If long-term plasticity is introduced , this variable is a function of time . In the simulations , for the neurons in the circuits the values for the U , D and F parameters were drawn from Gaussian distributions with mean values which depended on whether the type of presynaptic and postsynaptic neuron of the synapse is excitatory or inhibitory , and were chosen according to the data reported in [37] and [39] . The mean values of the Gaussian distributions are given in Table 2 , and the standard deviation was chosen to be 50% of its mean . Negative values were replaced with values drawn from uniform distribution with a range between 0 and twice the mean value . For the simulations involving individual trained neurons , the U , D , and F parameters of these neurons were set to the values from Table 2 . We have carried out control experiments with current-based synapses that were not subject to short-term plasticity ( see Figure S5 , Figure S8 , and Figure S9; successful control experiments with static current-based synapses were also carried out for computer simulation 1 , results not shown ) . We found that the results of all our computer simulations also hold for static current-based synapses . To reproduce the background synaptic input cortical neurons receive in vivo , the neurons in our models received an additional noise process as conductance input . The noise process we used is a point-conductance approximation model , described in [26] . According to [26] , this noise process models the effect of a bombardment by a large number of synaptic inputs in vivo , which causes membrane potential depolarization , referred to as “high conductance” state . Furthermore , it was shown that it captures the spectral and amplitude characteristics of the input conductances of a detailed biophysical model of a neocortical pyramidal cell that was matched to intracellular recordings in cat parietal cortex in vivo . The ratio of average contributions of excitatory and inhibitory background conductances was chosen to be 5 in accordance to experimental studies during sensory responses ( see [40]–[42] ) . In this model , the noisy synaptic current Inoise in Equation 24 is a sum of two currents: ( 27 ) where ge ( t ) and gi ( t ) are time-dependent excitatory and inhibitory conductances . The values of the respective reversal potentials were Ee = 0 mV and Ei = −75 mV . The conductances ge ( t ) and gi ( t ) were modeled according to [26] as a one-variable stochastic process similar to an Ornstein-Uhlenbeck process:with mean ge0 = 0 . 012 µS , noise-diffusion constant De = 0 . 003 µS and time constant τe = 2 . 7 ms for the excitatory conductance , and mean gi0 = 0 . 057 µS , noise-diffusion constant Di = 0 . 0066 µS , and time constant τi = 10 . 5 ms for the inhibitory conductance . χ1 ( t ) and χ2 ( t ) are Gaussian white noise of zero mean and unit standard deviation . Since these processes are Gaussian stochastic processes , they can be numerically integrated by an exact update rule:where N1 ( 0 , 1 ) and N2 ( 0 , 1 ) are normal random numbers ( zero mean , unit standard deviation ) and Ae , Ai are amplitude coefficients given by: For the computer simulations we used the following parameters for the STDP window function W ( r ) : A+ = 0 . 01wmax , A−/A+ = 1 . 05 , τ+ = τ− = 30 ms . wmax denotes the hard bound of the synaptic weight of the particular plastic synapse . Note that the parameter A+ can be given arbitrary value in this plasticity rule , since it can be scaled together with the reward signal , i . e . multiplying the reward signal by some constant and dividing A+ by the same constant results in identical time evolution of the weight changes . We have set A+ to be 1% of the maximum synaptic weight . We used the α-function to model the eligibility trace kernel fc ( t ) ( 28 ) where the time constant τe was set to τe = 0 . 4 s in all computer simulations . For computer simulations 1 and 4 we performed control experiments ( see Figure S3 , Figure S4 , and Figure S7 ) with the weight-dependent synaptic update rule proposed in [22] , instead of the purely additive rule in Equation 3 . We used the parameters proposed in [22] , i . e . μ = 0 . 4 , α = 0 . 11 , τ+ = τ− = 20 ms . The w0 parameter was calculated according to the formula: where wmax is the maximum synaptic weight of the synapse . is equal to the initial synaptic weight for the circuit neurons , or to the mean of the distribution of the initial weights for the trained neurons . The synaptic weights of excitatory synapses to the trained neurons in experiments 2–5 were initialized from a Gaussian distribution with mean wmax/2 . The standard deviation was set to wmax/10 bounded within the range [3wmax/10 , 7wmax/10] . All computer simulations were carried out with the PCSIM software package ( http://www . lsm . tugraz . at/pcsim ) . PCSIM is a parallel simulator for biologically realistic neural networks with a fast c++ simulation core and a Python interface . It has been developed by Thomas Natschläger and Dejan Pecevski . The time step of simulation was set to 0 . 1 ms . For all computer simulations , both for the cortical microcircuits and readout neurons , the same parameters values for the neuron and synapse models and the reward-modulated STDP rule were used , as specified in the previous section ( except in computer simulation 3 , where the goal was to test the theoretical predictions for different values of the parameters ) . Each of the computer simulations in this article modeled a specific task or experimental finding . Consequently , the dependence of the reward signal on the behavior of the system had to be modeled in a specific way for each simulation ( a more detailed discussion of the reward signal can be found in the Discussion section ) . The parameters for that are given below in separate subsections which address the individual simulations . Furthermore , some of the remaining parameters in the experiments , i . e . the values of the synaptic weights , the number of synapses of a neuron , number of neurons in the circuit and the Ornstein-Uhlenbeck ( OU ) noise levels were chosen to achieve different goals depending on the particular experiment . Briefly stated , these values were tuned to achieve a certain level of firing activity in the neurons , a suitable dynamical regime of the activity in the circuits , and a specific ratio between amount of input the neurons receive from the input synapses and the input generated by the noise process . We carried out two types of simulations: simulations of cortical microcircuits in computer simulations 1 and 5 , and training of readout neurons in computer simulations 2 , 3 , 4 , and 5 . In the following we discuss these two types of simulations in more detail . The values of the initial weights of the excitatory and inhibitory synapses for the cortical microcircuits are given in Table 3 . All synaptic weights were bounded in the range between 0 and twice the initial synaptic weight of the synapse . The cortical microcircuit was composed of 4000 neurons connected randomly with connection probabilities described in Details to computer simulation 1 . The initial synaptic weights of the synapses and the levels of OU noise were tuned to achieve a spontaneous firing rate of about 4 . 6 Hz , while maintaining an asynchronous irregular firing activity in the circuit . 50% of all neurons ( randomly chosen , 50% excitatory and 50% inhibitory ) received downscaled OU noise ( by a factor 0 . 2 from the model reported in [26] ) , with the subtracted part substituted by additional synaptic input from the circuit . The input connection probabilities of these neurons were scaled up , so that the firing rates remain in the same range as for the other neurons . This was done in order to observe how the learning mechanisms work when most of the input conductance in the neuron comes from a larger number of input synapses which are plastic , rather than from a static noise process . The reinforced neurons were randomly chosen from this group of neurons . We chose a smaller microcircuit , composed of 540 neurons , for the computer simulation 5 in order to be able to perform a large number of training trials . The synaptic weights in this smaller circuit were chosen ( see Table 3 ) to achieve an appropriate level of firing activity in the circuit that is modulated by the external input . The circuit neurons had injected an Ornstein-Uhlenbeck ( OU ) noise multiplied by 0 . 4 in order to emulate the background synaptic activity in neocortical neurons in vivo , and test the learning in a more biologically realistic settings . This produced significant trial-to-trial variability in the circuit response ( see Figure 10D ) . A lower value of the noise level could also be used without affecting the learning , whereas increasing the amount of injected noise would slowly deteriorate the information that the circuit activity maintains about the injected inputs , resulting in a decline of the learning performance . The maximum values of the synaptic weights of readout neurons for computer simulations 2 , 4 , and 5 , together with the number of synapses of the neurons , are given in Table 4 . The neuron in computer simulation 2 had 100 synapses . We chose 200 synapses for the neuron in computer simulation 4 , in order to improve the learning performance . Such improvement of the learning performance for larger numbers of synapses is in accordance with our theoretical analysis ( see Equation 17 ) , since for learning the classification of temporal patterns the temporal variation of the voltage of the postsynaptic membrane turns out to be of critical importance ( see the discussion after Equation 17 ) . This temporal variation depends less on the shape of a single EPSP and more on the temporal pattern of presynaptic firing when the number of synapses is increased . In computer simulation 5 the readout neuron received inputs from all 432 excitatory neurons in the circuit . The synaptic weights were chosen in accordance with the number of synapses in order to achieve a firing rate suitable for the particular task , and to balance the synaptic input and the noise injections in the neurons . For the pattern discrimination task ( computer simulation 4 ) and the speech recognition task ( computer simulation 5 ) , the amount of noise had to be chosen to be high enough to achieve sufficient variation of the membrane potential from trial to trial near the firing threshold , and low enough so that it would not dominate the fluctuations of the membrane potential . In the experiment where the exact spike times were rewarded ( computer simulation 2 ) , the noise had a different role . As described in the Results section , there the noise effectively controls the amount of depression . If the noise ( and therefore the depression ) is too weak , w* = 0 synapses do not converge to 0 . If the noise is too strong , w* = wmax synapses do not converge to wmax . To achieve the desired learning result , the noise level should be in a range where it reduces the correlations of the synapses with w* = 0 so that the depression of STDP will prevail , but at the same time is not strong enough to do the same for the other group of synapses with w* = wmax , since they have stronger pre-before-post correlations . For our simulations , we have set the noise level to the full amount of OU noise . The cortical microcircuit model consisted of 4000 neurons with twenty percent of the neurons randomly chosen to be inhibitory , and the others excitatory . The connections between the neurons were created randomly , with different connectivity probabilities depending on whether the postsynaptic neuron received the full amount of OU noise , or downscaled OU noise with an additional compensatory synaptic input from the circuit . For neurons in the latter sub-population , the connection probabilities were pee = 0 . 02 , pei = 0 . 02 , pie = 0 . 024 and pii = 0 . 016 where the ee , ei , ie , ii indices designate the type of the presynaptic and postsynaptic neurons ( e = excitatory or i = inhibitory ) . For the other neurons the corresponding connection probabilities were downscaled by 0 . 4 . The resulting firing rates and correlations for both types of excitatory neurons are plotted in Figure S1 and Figure S2 . The shape of the reward kernel εr ( t ) was chosen as a difference of two α-functions ( 29 ) one positive α-pulse with a peak at 0 . 4 sec after the corresponding spike , and one long-tailed negative α-pulse which makes sure that the integral over the reward kernel is zero . The parameters for the reward kernel were , , , , and dr = 0 . 2 s , which produced a peak value of the reward pulse 0 . 4 s after the spike that caused it . We used the following function for the reward kernel κ ( r ) ( 30 ) where and are positive scaling constants , and define the shape of the two double-exponential functions the kernel is composed of , and tκ defines the offset of the zero-crossing from the origin . The parameter values used in our simulations were , , , and tκ = −1 ms . The reward delay was equal to dr = 0 . 4 s . We used a linear Poisson neuron model as in the theoretical analysis with static synapses and exponentially decaying postsynaptic responses . The neuron had 100 excitatory synapses , except in experiment #6 , where we used 200 synapses . In all experiments the target neuron received additional 10 excitatory synapses with weights set to wmax . The input spike trains were Poisson processes with a constant rate of rpre = 6 Hz , except in experiment # 6 where the rate was rpre = 3 Hz . The weights of the target neuron were set to for 0≤i<50 and for 50≤i<100 . The time constants of the reward kernel were , whereas had different values in different experiments ( reported in table 1 ) . The value of tκ was always set to an optimal value such that the . The time constant τ− of the negative part of the STDP window function W ( r ) was set to τ+ . The reward signal was delayed by τd = 0 . 4 s . The simulations were performed for varying durations of simulated biological time ( see the tsim-column in Table 1 ) . We used the reward signal from Equation 16 , with an α-function for the reward kernel , and the reward delay dr set to 300 ms . The amplitudes of the positive and negative pulses were αP = −αN = 1 . 435 and the time constant of the reward kernel was τ = 100 ms .
The theoretical analysis of this model is directly applicable to the learning rule considered in [12] . There , the network behavior of reward-modulated STDP was also studied some situations different from the ones in this article . The computer simulations of [12] operate apparently in a different dynamic regime , where LTD dominates LTP in the STDP-rule , and most weights ( except those that are actively increased through reward-modulated STDP ) have values close to 0 ( see Figure 1b and 1d in [12] , and compare with Figure 5 in this article ) . This setup is likely to require for successful learning a larger dominance of pre-before-post over post-before-pre pairs than the one shown in Figure 4E . Furthermore , whereas a very low spontaneous firing rate of 1 Hz was required in [12] , computer simulation 1 shows that reinforcement learning is also feasible at spontaneous firing rates which correspond to those reported in [17] ( the preceding theoretical analysis had already suggested that the success of the model does not depend on particularly low firing rates ) . The articles [15] and [13] investigate variations of reward-modulated STDP rules that do not employ learning curves for STDP that are based on experimental data , but modified curves that arise in the context of a very interesting top-down theoretical approach ( distributed reinforcement learning [14] ) . The authors of [16] arrive at similar learning rules in a supervised scenario which can be reinterpreted in the context of reinforcement learning . We expect that a similar theory as we have presented in this article for the more commonly discussed version of STDP can also be applied to their modified STDP rules , thereby making it possible to predict under which conditions their learning rules will succeed . Another reward based learning rule for spiking neurons was recently presented in [49] . This rule exploits correlations of a reward signal with noisy perturbations of the neuronal membrane conductance in order to optimize some objective function . One crucial assumption of this approach is that the synaptic plasticity mechanism “knows” which contributions to the membrane potential arise from synaptic inputs , and which contributions are due to internal noise . Such explicit knowledge of the noise signal is not needed in the reward-modulated STDP rule of [12] , which we have considered in this article . The price one has to pay for this potential gain in biological realism is a reduced generality of the learning capabilities . While the learning rule in [49] approximates gradient ascent on the objective function , this cannot be stated for reward-modulated STDP at present . Timing-based pattern discrimination with a spiking neuron , as discussed in the section “Pattern discrimination with reward-modulated STDP” of this article , was recently tackled in [50] . The authors proposed the tempotron learning rule , which increases the peak membrane voltage for one class of input patterns ( if no spike occurred in response to the input pattern ) while decreasing the peak membrane voltage for another class of input patterns ( if a spike occurred in response to the pattern ) . The main difference between this learning rule and reward-modulated STDP is that the tempotron learning rule is sensitive to the peak membrane voltage , whereas reward-modulated STDP is sensitive to local fluctuations of the membrane voltage . Since the time of the maximal membrane voltage has to be determined for each pattern by the synaptic plasticity mechanism , the basic tempotron rule is perhaps not biologically realistic . Therefore , an approximate and potentially biologically more realistic learning rule was proposed in [50] , where plasticity following error trials is induced at synapse i only if the voltage within the postsynaptic integration time after their activation exceeds a plasticity threshold κ . One potential problem of this rule is the plasticity threshold κ , since a good choice of this parameter strongly depends on the mean membrane voltage after input spikes . This problem is circumvented by reward-modulated STDP , which considers instead the local change in the membrane voltage . Further work is needed to compare the advantages and disadvantages of these different approaches . Reward-modulated STDP is a very promising candidate for a synaptic plasticity rule that is able to orchestrate local synaptic modifications in such a way that particular functional properties of larger networks of neurons can be achieved and maintained ( we refer to [12] and [27] for discussion of potential biological implementations of this plasticity rule ) . We have provided in this article analytical tools which make it possible to evaluate this rule and variations of this rule not just through computer simulations , but through theoretical analysis . In particular we have shown that successful learning is only possible if certain relationships hold between the parameters that are involved . Some of these predicted relationships can be tested through biological experiments . Provided that these relationships are satisfied , reward-modulated STDP turns out to be a powerful rule that can achieve self-organization of synaptic weights in large recurrent networks of neurons . In particular , it enables us to explain seemingly inexplicable experimental data on biofeedback in monkeys . In addition reward-modulated STDP enables neurons to distinguish complex firing patterns of presynaptic neurons , even for data-based standard forms of STDP , and without the need for a supervisor that tells the neuron when it should spike . Furthermore reward-modulated STDP requires substantial spontaneous activity and trial-to-trial variability in order to support successful learning , thereby providing a functional explanation for these ubiquitous features of cortical networks of neurons . In fact , not only spontaneous activity but also STDP itself may be seen in this context as a mechanism that supports the exploration of different firing chains within a recurrent network , until a solution is found that is rewarded because it supports a successful computational function of the network . | A major open problem in computational neuroscience is to explain how learning , i . e . , behaviorally relevant modifications in the central nervous system , can be explained on the basis of experimental data on synaptic plasticity . Spike-timing-dependent plasticity ( STDP ) is a rule for changes in the strength of an individual synapse that is supported by experimental data from a variety of species . However , it is not clear how this synaptic plasticity rule can produce meaningful modifications in networks of neurons . Only if one takes into account that consolidation of synaptic plasticity requires a third signal , such as changes in the concentration of a neuromodulator ( that might , for example , be related to rewards or expected rewards ) , then meaningful changes in the structure of networks of neurons may occur . We provide in this article an analytical foundation for such reward-modulated versions of STDP that predicts when this type of synaptic plasticity can produce functionally relevant changes in networks of neurons . In particular we show that seemingly inexplicable experimental data on biofeedback , where a monkey learnt to increase the firing rate of an arbitrarily chosen neuron in the motor cortex , can be explained on the basis of this new learning theory . | [
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
] | [
"neuroscience/animal",
"cognition",
"neuroscience/theoretical",
"neuroscience"
] | 2008 | A Learning Theory for Reward-Modulated Spike-Timing-Dependent
Plasticity with Application to Biofeedback |
Bacterial microcompartments ( BMCs ) are proteinaceous organelles involved in both autotrophic and heterotrophic metabolism . All BMCs share homologous shell proteins but differ in their complement of enzymes; these are typically encoded adjacent to shell protein genes in genetic loci , or operons . To enable the identification and prediction of functional ( sub ) types of BMCs , we developed LoClass , an algorithm that finds putative BMC loci and inventories , weights , and compares their constituent pfam domains to construct a locus similarity network and predict locus ( sub ) types . In addition to using LoClass to analyze sequences in the Non-redundant Protein Database , we compared predicted BMC loci found in seven candidate bacterial phyla ( six from single-cell genomic studies ) to the LoClass taxonomy . Together , these analyses resulted in the identification of 23 different types of BMCs encoded in 30 distinct locus ( sub ) types found in 23 bacterial phyla . These include the two carboxysome types and a divergent set of metabolosomes , BMCs that share a common catalytic core and process distinct substrates via specific signature enzymes . Furthermore , many Candidate BMCs were found that lack one or more core metabolosome components , including one that is predicted to represent an entirely new paradigm for BMC-associated metabolism , joining the carboxysome and metabolosome . By placing these results in a phylogenetic context , we provide a framework for understanding the horizontal transfer of these loci , a starting point for studies aimed at understanding the evolution of BMCs . This comprehensive taxonomy of BMC loci , based on their constituent protein domains , foregrounds the functional diversity of BMCs and provides a reference for interpreting the role of BMC gene clusters encoded in isolate , single cell , and metagenomic data . Many loci encode ancillary functions such as transporters or genes for cofactor assembly; this expanded vocabulary of BMC-related functions should be useful for design of genetic modules for introducing BMCs in bioengineering applications .
Membrane-bound organelles for compartmentalization of specific functions are the hallmark feature of all eukaryotic cells . Bacteria also have organelles , but they are not ubiquitous throughout the domain; instead they are sporadically distributed and frequently provide functions that are key to niche specialization . For example , anammoxosomes are lipid-bound compartments that enable certain planctomycetes to obtain energy from anaerobic ammonium oxidation ( reviewed in [1] ) , and magnetosomes are invaginations of the inner membrane that allow magnetotactic bacteria to orient along the Earth's magnetic field to search for microaerobic environments ( reviewed in [2] ) . Another type of organelle , composed entirely of protein , is the bacterial microcompartment ( BMC ) ( reviewed in [3]–[5] ) . BMCs were discovered initially in electron micrographs as polyhedral bodies within members of the Cyanobacteria [6] and later in chemoautotrophs [7] . Subsequently , it was shown that these inclusions encapsulate enzymes required for carbon fixation , and they were termed carboxysomes [8] . X-ray crystallographic studies of subunits of the carboxysome shell provided the first structural examples of each of the three main types of BMC shell proteins [9]–[11] . These led to an icosahedral model of the shell: facets formed by ( pseudo ) hexameric building blocks capped by pentameric vertices ( Fig . 1 ) . Hexameric ( BMC-H ) shell subunits are formed by proteins that contain a single copy of the Pfam [12] domain PF00936 , while BMC-T proteins are a fusion of two PF00936 domains and form trimers ( pseudohexamers ) [10] , [13] . Pentagonal vertices are formed by BMC-P proteins , which contain a single PF03319 domain [9] , [14] , [15] . Several studies have shown that all three components are required for the construction of fully functional carboxysomes [16]–[18] . Carboxysomes were thought to be a peculiarity confined to some autotrophic organisms until operons involved in vitamin B12-dependent propanediol ( PDU ) and ethanolamine utilization ( EUT ) in Salmonella enterica were sequenced and found to encode genes for BMC-H , BMC-T , and BMC-P proteins [19]–[22] . The formation of carboxysome-like structures was subsequently confirmed by electron microscopy of cells grown on 1 , 2-propanediol or ethanolamine [22] , [23] . These studies provided the first examples of catabolic BMCs used by heterotrophic bacteria to degrade specific compounds , in contrast to the anabolic carboxysomes . More recently , the availability of genomic sequence data has enabled the discovery of the prevalence of BMCs among the Bacteria . Shell protein genes are typically found clustered with genes for putative enzymes , allowing prediction of the potential to form BMCs to be made from sequence data . Recently , the first functional characterizations of such bioinformatically-predicted BMCs were reported: one that degrades propanediol via a glycyl radical enzyme ( termed the fucosome ) [24] and a second that degrades fucose and rhamnose directly via an aldolase ( termed the PVM BMC ) [25] . For visualization of catabolic BMCs , knowledge of the substrate is necessary in order to induce the operon and organelle formation for subsequent visualization by electron microscopy . This contrasts with historic organelle discovery , when light microscopy of eukaryotic cells was sufficient to reveal new organelles . Whether anabolic or catabolic , experimentally characterized BMCs share a common functional theme: the shell sequesters enzymes and provides a diffusion barrier for volatile and/or toxic reaction intermediates . For example , in carboxysomes , CO2 is generated by carbonic anhydrase ( CA ) and then fixed within the compartment by RuBisCO; the shell helps to confine the CO2 near RuBisCO [26] ( Fig . 1A ) . Likewise , the PDU and EUT BMC shells prevent the leakage of propionaldehyde [27] and acetaldehyde [28] , respectively , which are toxic and/or volatile intermediates of the encapsulated biochemical pathway . These experimental observations were used to infer the type of aldehyde intermediate in recently characterized BMCs: propionaldehyde in the fucosome [24] and lactaldehyde in the PVM BMC [25] . Recently , comparison of experimentally characterized catabolic BMCs led to the identification of their common core biochemistry ( Fig . 1B ) [25] . The metabolic function of the BMC ( e . g . propanediol utilization ) is defined by the aldehyde-generating enzyme ( e . g . propanediol dehydratase ) which we refer to as the “signature” enzyme ( Fig . 1 ) . The aldehyde is acted on by an NAD+ and CoA-dependent aldehyde dehydrogenase ( AldDH ) , forming NADH and an acyl∼CoA product [29] , [30] . Because this reaction is encapsulated , the cofactors must be recycled within the metabolosome [31] , [32] . In order to regenerate NAD+ , an alcohol dehydrogenase ( AlcDH ) reduces a second aldehyde , forming an alcohol [28] , [31] . A phosphotransacetylase ( PTAC ) then acts on the acyl∼CoA to replace the CoA moiety with a phosphate [32] , [33] , which can then participate in substrate-level phosphorylation . Because the biochemical core is a requirement for cofactor recycling , we hypothesize that all BMCs that utilize an acylating AldDH also require the other core enzymes , and we refer to the BMCs that encapsulate these core biochemical transformations as metabolosomes . Accordingly , two functional paradigms for BMCs have emerged: anabolic carboxysomes , which encapsulate RuBisCO and carbonic anhydrase , and catabolic metabolosomes , which encapsulate the aforementioned core enzymes for the specific purpose of recycling cofactors ( Fig . 1 ) . In experimentally characterized alpha-carboxysome and metabolosome loci , the essential BMC components ( shell proteins , signature and core enzymes; Fig . 1 ) are encoded within a genetic locus . This has implications for horizontal gene transfer . The PDU locus has been suggested to instantiate the concept of a selfish operon: a group of contiguous genes that functions together [34] . Such a self-sufficient genetic and metabolic unit is likely to confer a beneficial skillset to a host organism in a single horizontal gene transfer event [34] . Substantiating this hypothesis , it has been shown that the PDU locus alone is necessary and sufficient to form a fully functional organelle that provides the host organism with new metabolic potential [35] . This attribute of BMC loci is also significant for synthetic biology; engineered BMC loci can be envisioned as both genetic and metabolic modules for plug and play introduction of biochemical pathways of industrial interest [36] . The increasing availability of genomic sequence data enabled the discovery that many bacteria potentially form organelles . The identification of BMC shell protein genes , gene cluster conservation and genomic context were used in the first surveys of putative BMC functions and their distribution [3] , [37] . As the number of available microbial genomes increased , a correlation network for co-occurrence of protein functional groups within BMC loci was devised [38] . A major limitation of this method was its gene-centric approach; loci could only be predicted manually after visualizing the resulting co-occurrence network of genes . Although this analysis effectively recovered a number of loci that had been experimentally characterized , the reliance of the method on co-occurrence of protein functional groups made it insensitive to rare BMC locus types . As a result , several BMC types that had been previously described based on genomic context-based approaches [3] , [37] were missed . Furthermore , the presence of highly abundant functional groups in multiple different BMC locus types ( such as alcohol dehydrogenases ) posed a problem because the clustering step sought to assign these functional groups to a single BMC locus type . This resulted in co-occurrence networks that do not represent the true co-occurrence of genes within a given locus type and underestimation of BMC diversity . The most recent gene-centric bioinformatic survey of BMC loci divided them into different groups based on pfams of known signature enzymes but did not divide them into functional sub-types , predict new signature enzymes , or define novel BMC locus types [39] . Here we present the results of a new approach , LoClass ( Locus Classifier ) , to surveying and classifying BMC loci with an emphasis on discovery of novel BMC locus types and variants of the paradigmatic types . It is a novel locus-centric method for predicting and sorting loci through the generation of a locus similarity network . To compare loci , we approximate the functional attributes of a locus by representing it as the set of pfam domains encoded by genes in the locus . While the use of pfams to visualize the functionality in a locus is relatively low-resolution , in that one pfam may correspond to a variety of homologous sequences of different functions , this coarse resolution allows for the recognition of pfams corresponding to the functionally similar biochemical cores . Moreover , the lost resolution is regained through the multiplicity of pfams conferring organelle-supporting functions present in any given locus; the diversity among these pfams aids in distinguishing different BMC locus types and sub-types . By focusing on the regions flanking BMC shell protein genes instead of being confined to presumed operons , we are able to circumvent the lack of transcriptional data for the majority of sequenced genomes that contain BMCs . LoClass also captures the genomic neighborhood of BMC shell protein genes , revealing that genes encoding the organelles are frequently situated in the context of other genes that provide ancillary functions , such as regulation , co-factor synthesis , or transport for BMC substrates . Recognition of these gene products and their roles in supporting BMC function will be useful both in functionally characterizing diverse BMCs and in the design of BMC locus modules that are ready for “plug and play” applications . Finally , LoClass is able to perform direct comparisons and classification of loci , granting high sensitivity to rare locus types . The result is a comprehensive taxonomy of BMC locus ( sub ) types and the identification of several novel putative BMC locus types , including one that we predict extends BMC functions beyond the paradigmatic metabolosomes and carboxysomes .
The standard PF00936 hidden markov model ( HMM ) is often not robust for the identification of N-terminal cryptic BMC domains of CsoS1D/CcmP-type BMC-T proteins , which have only weak sequence similarity to other BMC domains [10] , [13] . Accordingly , a modified PF00936 HMM was built using the seed alignment ( 128 sequences ) in the Pfam database [12] with five cryptic BMC-T domain sequences ( NCBI protein accessions: YP_007073159 . 1 , YP_007144647 . 1 , YP_473976 . 1 , YP_637369 . 1 , YP_925211 . 1 ) added to the alignment ( a total of 133 sequences ) to render it more robust for detecting these cryptic BMC domains ( Dataset S3 , Dataset S4 ) . This expanded PF00936 HMM and the PF03319 HMM were searched using hmmsearch in the hmmer [40] package against a local copy of the NCBI Non-redundant Protein Database ( NR ) downloaded from NCBI on July 3rd , 2013 . All hits with an e-value less than or equal to 1e-05 that corresponded to genomic records from the Genbank , RefSeq , EMBL , and DDBJ databases were accepted as BMC shell proteins homologs . This cutoff was chosen based on manual inspection of results for spurious hits relative to various e-value thresholds . Identical proteins to each of these hits in the genomes containing these hits were retrieved from NCBI using NCBI Entrez Programming Utilities [41] because NR frequently stores identical proteins in a single record . Where at least one BMC protein in a given genome was non-identical to all other proteins included in this analysis , that genome was included , and this step retrieved all other BMC proteins in its genome that might be identical to other proteins included in this analysis . Where every BMC protein in a given genome was 100% identical to a protein already included in this analysis , the BMC proteins from this genome were not included . This reduced the computational and analytical burden introduced by the analysis of likely identical loci from very closely related genomes . The genome of the model cyanobacterium Synechococcus elongatus PCC 7942 ( Syn7942 ) was not included in this analysis due to the formatting irregularities of its sequence records . The closely related organism Synechococcus elongatus PCC 6301 ( Syn6301 ) , whose BMC proteins are identical to those of Syn7942 , was included . We examined the similarity of these genomes using NUCmer in the MUMmer 3 . 1 [42] package . These genomes are highly syntenic , the greatest differences being a 188 . 9 kb inversion in the genome sequence and a 231 bp sequence in Syn6301 that is absent from Syn7942 , confirming a previous genomic comparison [43] . Sequence identity between the two genomes over all aligned regions was 99 . 9% . Syn6301 was therefore used as a proxy for Syn7942 . Prospective BMC Loci were initially defined as the region on the chromosome 5 kb upstream and downstream of each BMC shell protein gene , analogous to the five open reading frame distance used by Jorda et al [38] to define BMC-related gene co-occurrence . However , the conserved locus of several characterized BMC locus types ( e . g . the alpha-carboxysome locus of Halothiobacillus neapolitanus ) contain a gap greater than 10 kb between BMC-related genes and the nearest BMC shell protein gene . This resulted in several characterized loci being incomplete . Prospective BMC Loci were then redefined as the region 10 kb upstream and downstream of BMC shell protein genes ( Fig . 2A ) . Wherever these BMC loci overlapped with each other , they were merged into one Prospective BMC Locus ( Fig . 2B ) . We defined the envelope as the largest region in the locus flanked by BMC shell protein genes ( Fig . 2B ) . All genes that at least partially overlapped a given locus range were deemed part of the locus . All loci that were truncated by the beginning or end of the scaffold were designated as potentially incomplete loci and were not included in subsequent analyses . A local copy of the Pfam 27 . 0 database [12] was searched against the proteins corresponding to all non-shell protein genes found in a given BMC locus using hmmsearch with a loose e-value cutoff of 0 . 01 ( Fig . 2C ) . Where a pfam alignment in a protein sequence overlapped a pfam alignment in the same sequence by over 50% of the pfam length , the alignment with the higher e-value was discarded . The pfam sets found in each locus were then used to compare it to each other locus using a novel scoring mechanism ( Fig . 2D ) . A novel scoring method we name LoClass , the Locus Classifier , was developed to determine the relative similarity of each locus to every other locus . The scoring mechanism used has a positive component , determined by the pfam domains any two loci share in common , and a negative component , determined by the pfam domains present in only one of the two loci . Let L represent the set of all BMC loci , where PI is the set of all pfams found in the locus I ( Fig . 2D ) and p represents a given pfam domain . The set of pfams common to loci I and J can then be represented as ( 1 ) while the set of pfams found in I but not J can be represented as ( 2 ) ( Fig . 2E ) . Various weights comprise the positive and negative scoring components . First , since the shared pfams between two loci are more important for determining their similarity than their disjoint pfams , a weight k is applied to the negative score . Testing various values for k showed that for these loci , a weight of 0 . 5 yielded the best results , as judged by the recapitulation of the experimentally confirmed PDU , EUT , and PVM locus types . If d represents the minimum number of open reading frames between the envelope and a gene that contains the pfam , then a distance weight can be applied to account for the decreasing likelihood that genes in a locus are related to the function of the BMC the further they are from the envelope ( Fig . 2D ) . This distance weight can be represented by ( 3 ) Some pfams may be used to identify BMC loci described in the literature: PF12288 ( CsoS2 ) and PF08936 ( CsoSCA ) are found specifically in alpha-carboxysomal loci [44] , and PF00132 ( CcmM ) and PF00132 ( CcmN ) are found specifically in beta-carboxysomal loci [45] . Likewise , PF06751 and PF05985 ( ethanolamine ammonia lyase subunits ) comprise the signature enzyme ethanolamine ammonia lyase for the ethanolamine utilization locus , while PF02286 , PF02287 , and PF02288 ( propanediol dehydratase subunits ) identify the signature enzyme for the propanediol utilization locus . PF00596 corresponds to an aldolase predicted to be the signature enzyme for the Planctomycetes and Verrucomicrobia microcompartment ( PVM ) locus type [25] . This prior knowledge is incorporated into the score by designating these pfams as identifying pfams and creating an identifying pfam weight ( 4 ) Some pfams , such as PF00171 representing aldehyde dehydrogenase , PF06130 representing PduL phosphotransacetylase [46] , and PF00465 representing iron-containing alcohol dehydrogenase are extremely common among BMCs and therefore do not reveal much about the similarity of two loci . Likewise , if two loci both contain a pfam that is found in very few BMC loci , this more strongly indicates that these two loci are similar , while if only one contains the rare pfam , this may indicate that they are quite different types . We therefore add a rare pfam weight ( 5 ) Since R ( p ) represents the proportion of BMC loci that do not contain the pfam p , the weight approaches but never reaches 1 for very rare pfams and 0 for very common pfams . Finally , when a pfam is only found in one of the loci , it is necessary to consider whether that pfam is frequently found within loci that also contain the set of pfams that the two loci share in common . This aids in down-weighting secondary pfams that are not necessary for the core function of a BMC , are not always localized in the BMC loci of a given type , or are only occasionally found near the periphery of a BMC locus while not actually being related to the BMC function . This co-occurring pfam weight can be represented by ( 6 ) Applying the appropriate weights to the positive and negative score components , we define a similarity score between any two loci as ( 7 ) This score was generated for each locus pairwise comparison in this analysis , and these scores were then used to construct the edge lengths for a BMC locus similarity network . Analysis and clustering of the locus similarity network was complicated by the presence of what we term satellite loci , loci that encode an insufficient complement of BMC shell protein genes ( only encoding either PF03319 or PF00936 domains , but not both ) to form a BMC; genomes containing satellite loci putatively encode additional shell proteins and possibly other BMC components for a Candidate/Confirmed BMC in another locus . We define satellite BMC loci as those that meet three conditions: the locus contains an insufficient shell protein gene complement , all the shell protein genes in the locus are contiguous to each other , and there is at least one other non-satellite BMC locus in the genome . In order to prevent false positives , predicted satellite loci were manually inspected for the presence of genes with common BMC-associated pfams immediately nearby the BMC shell protein genes . Where such genes were found or where the locus met all requirements except that there was more than one locus in the genome , these loci were labeled as satellite-like loci . Where a signature enzyme was present or a putative function had already been assigned to the locus in the literature , as with the putative ethanol utilization locus , type names were assigned to these loci ( see below ) [47] . To simplify the analysis of the locus similarity network , predicted satellite loci were excluded from further steps of the analysis , while satellite-like loci , which we predict encode a structurally incomplete BMC locus , were included . This greatly reduced the spurious results in the clustering step . The locus similarity network of BMC loci was visualized in Cytoscape 2 . 8 . 3 [48] at a score cut-off of −20 using the Force-directed layout , where the edge lengths corresponded to the locus similarity score determined above and normalized within Cytoscape . The resulting network was then clustered using multiple iterations of the Markov Cluster Algorithm ( MCL ) [49] at increasing stringency using the clusterMaker [50] plug-in in Cytoscape . MCL has been used successfully in clustering various types of biological networks , such as sequence similarity , protein expression profiles , and scientific article similarity [51] and is best applied to undirected networks where edges represent similarity [49] . Stringency in MCL is controlled by manipulating the score cut-off and the inflation value , where the score represents the metric used to calculate edge length and the inflation value is a property of MCL; a higher inflation value will effect a more fine-grained clustering [49] . Resulting clusters were then assigned numbers ( Fig . 3 ) . Whenever a cluster still contained obvious sub-groups , another iteration of MCL sub-clustering was performed at more stringent cut-offs until the resulting clusters were either too small to be informative , representing nearly identical loci , or until the cluster was so tight that the score cut-off necessary to split up the cluster exceeded 40 . This process of varying the inflation value and score cut-off to achieve relevant clusterings has been well described [51] . At each level of sub-clustering , an additional numeral was added to the cluster number . For example , the first cluster resulting from the sub-clustering of Cluster 1 was called Cluster 1 . 1 . The cut-offs used and resulting clusters are shown in Figure S1 and its legend , and the resulting network is included in Dataset S5 . At each level of clustering , the loci were manually compared with previously studied or predicted BMC loci . Where all of the loci in a large cluster appeared to be of the same type as a previously studied locus ( e . g . EUT ) , those loci were first assigned that type name as a super-type . Where that cluster was further sub-clustered once or twice , the loci in those clusters were additionally assigned a numeral followed by an alphabetic character ( e . g . EUT1 , EUT1A ) . In our taxonomy , all loci with a different numeral are deemed different types , while those whose type names are differentiated only by the alphabetic character ( e . g . EUT2A , EUT2B ) are deemed different sub-types . Where a locus ( sub ) type clustered with loci of a known type but either was lacking the signature enzyme in the pathway or bore significant differences from the type , that locus was designated to be “like” that BMC type ( e . g . PDU-like ) . Where a locus type had not previously been named in the literature , if all genomes containing that type came from a distinct set of taxa that do not contain any other types , we assigned that type a three-letter name based on those taxa ( e . g . RMM1 ) . Where multiple loci of unknown function did not meet any of these prior criteria , they were designated Metabolosome of Unknown Function ( MUF ) , Metabolosome with an Incomplete Core ( MIC ) , or BMC of Unknown Function ( BUF ) type , depending on whether they contained all core metabolosome genes , only the AldDH but not a complete core , or none of the metabolosome core enzymes , respectively . Where multiple MIC loci clustered together , an additional numeral was assigned to the name to designate that specific type ( MIC1 ) . For each locus ( sub ) type , a representative locus that best captures the pfam diversity of all loci of that ( sub ) type was chosen ( Fig . 4; Dataset S1 ) . For each pfam that appears in any of the loci of a given ( sub ) type , the proportion of loci of that ( sub ) type that contain that pfam was calculated ( Fig . S2 ) . Then , for each locus , a representative score was calculated by adding these proportions for all pfams it contains and subtracting the proportions for all pfams it lacks . The locus with the highest score was chosen as the representative locus for that ( sub ) type; where multiple loci corresponded to the highest representative score , a characterized BMC locus or a BMC locus from a reference genome , in that order of preference , was selected as the representative . If one or more pfams present in the majority of the loci were not encoded in the highest scoring locus but were encoded in the next highest scoring locus , then the second locus was selected as the representative . These loci generally exemplify the consistent trends across a locus ( sub ) type , as well as the unique trends present in a subset of the loci . However , where there is a great deal of diversity within a given locus ( sub ) type , as is the case with alpha- and beta-carboxysomes , many loci of that ( sub ) type may differ substantially from the representative locus . In order to identify BMC loci in genomes from candidate phyla not archived in NR , we examined the single-cell genomes from the recent GEBA-MDM project [52] for the presence of BMC shell protein genes using the Integrated Microbial Genomes ( IMG ) Database [53] . In addition , we searched for the presence of BMCs in all unclassified bacterial phyla in IMG . These loci were then manually inspected and compared to the BMC locus taxonomy constructed by LoClass from data in NR . Sequences for aldehyde dehydrogenase genes within BMC loci ( accession numbers in Dataset S1 ) were aligned using MUSCLE [54] , [55] . The alignment was visualized and edited using Jalview [56] . All sequences that did not contain the catalytic cysteine ( see Results and Discussion ) were removed from the alignment . In addition , sequences from Shigella sonnei ( NCBI protein accessions: YP005456972 . 1 , YP310962 . 1 ) and Roseburia inulinivorans ( EEG94445 . 1 ) were significantly shorter than the rest of the sequences and were also removed ( Dataset S6 ) . The alignment was manually curated by removing all gapped positions . A maximum likelihood phylogenetic tree was constructed by using PhyML [57] in the phylogeny . fr web server [58] , [59] with 100 bootstraps ( Fig . 5 ) . Four sequences ( IMG Gene OID: 2264936618; NCBI protein accessions: YP_004405262 . 1 , YP_007299416 . 1 , CCK75300 . 1 ) were significantly divergent , forming extremely long branches , and were removed from the alignment for the final tree . Genome assemblies of 2 , 025 bacteria were scanned for homologs of a set of 38 universally conserved single-copy proteins present in Bacteria and Archaea [52] . The assemblies were translated into all six reading frames , and marker genes were detected and aligned with hmmsearch and hmmalign included in the HMMER3 [60] package using HMM profiles obtained from PhyloSift [61] . Extracted marker protein sequences were used to build concatenated alignments of up 38 markers per genome . The phylogenetic inference method used was the maximum likelihood based FastTree2 [62] with CAT approximation with 20 rate categories and Jones-Taylor-Thorton ( JJT ) for FastTree2 . Sequences were grouped into clades and manually corrected in ARB [63]; newick trees were exported from ARB and beautified with iTOL [64] . BMC locus ( sub ) types were then mapped onto the resulting phylum tree ( Fig . 6 ) .
The numbers of genomes and loci analyzed using LoClass , as well as definitions of terms used in this Discussion are given in Table 1 . BMC loci are prevalent among Bacteria , found in a total of 23 bacterial phyla ( Fig . 6 ) . Sixteen phyla were identified by searching against NR and analyzing using LoClass; seven additional phyla were found after inclusion of BMC loci identified in candidate phyla and single-cell genomes from IMG ( Fig . 6 ) . The greatest diversity of BMC locus ( sub ) types is in the Firmicutes and Gammaproteobacteria; however , these phyla also contain the highest numbers of BMC-containing genomes . Many of these ( sub ) types also appear in distantly related phyla ( Fig . 6 , Fig . S3 ) . This distribution is consistent with the hypothesis that BMC loci are frequently horizontally transferred . Six genomes contain the highest number ( five ) of total BMC loci in one genome ( Table S2 ) . Subtracting the satellite and satellite-like loci in these genomes reduces this number to two or fewer Candidate/Confirmed Loci . Among all genomes surveyed , as many as three functionally distinct Candidate/Confirmed BMC Loci are found within a single genome; this occurs in 12 genomes surveyed ( Table S2 ) . Nearly every outlier in our analysis , including known and predicted cyanobacterial satellite BMC loci [16] , fulfill all three of the criteria for satellite loci ( see Materials and Methods; Dataset S1 ) . Based on the experimental data [65] , [66] and structural models of BMCs , the number of BMC-H ( a major component of the shell ) genes is expected to be larger than that for BMC-T ( presumably minor shell components ) and BMC-P ( only required to cap the vertices ) . We observed that each genome contains 3 . 5 BMC-H genes , 1 . 4 BMC-T genes , and 1 . 2 BMC-P genes on average per Candidate/Confirmed BMC Locus . The largest number of BMC-H genes predicted in any one genome ( fifteen ) is in Clostridium saccharolyticum WM1 ( Table S2 ) , distributed across three Candidate/Confirmed Loci . Four genomes contain the highest count ( five ) BMC-T genes in any genome: Desulfosporosinus orientis DSM 765 , Desulfosporosinus meridiei DSM 13257 , and two strains of Clostridium kluyveri ( Table S2 ) ; these are distributed across two to three Candidate/Confirmed BMC Loci . Melioribacter roseus P3M-2 exceeds every other genome in BMC-P count , with a total of seven genes spread out across five loci , four of which are satellite/satellite-like loci . No other genome contains more than three BMC-P genes . While we found that BMC-H genes typically outnumber BMC-T and BMC-P genes , this is not always the case . The most extreme example is Haliangium ochraceum DSM 14365 , which contains three BMC-T , three BMC-P genes , and only one BMC-H . Despite the unusual proportion of BMC shell protein gene types in this genome , recent expression of the seven H . ochraceum shell proteins in E . coli resulted in remarkably homogeneous and stable BMC shells that could be readily purified in large quantities [67] . The two types of carboxysomes are named for the form of encapsulated RuBisCO; alpha-carboxysomes encapsulate form 1A RuBisCO , beta-carboxysomes encapsulate form 1B RuBisCO [45] . Alpha-carboxysome loci are found in the phylum Cyanobacteria , as well as in some chemoautotrophs from the phyla Actinobacteria , Alphaproteobacteria , Betaproteobacteria , and Gammaproteobacteria ( Fig . 6 , Table S1 ) . The common core of all alpha-carboxysome loci consists of genes for the RuBisCO large ( CbbL ) and small ( CbbS ) subunits , a beta carbonic anhydrase CsoSCA [68] , a protein of unknown function CsoS2 , and an accessory protein , a pterin dehydratase-like RuBisCO assembly factor [69] ( Fig . 4 , Dataset S1 ) . The alpha-carboxyome loci of chemoautotrophs and cyanobacteria are distinguished by differences in the genes flanking the conserved core ( Fig . 4 ) . For example , most chemoautotrophic loci encode a protein 27–39% identical to the LysR family regulator CbbR ( UniProtKB: P52690 ) [70] , while no cyanobacterial alpha-carboxysome loci do , presumably because the carboxysome is constitutively expressed . This gene is usually encoded immediately upstream of cbbL and , in two of these loci , has been shown experimentally to regulate and be divergently transcribed from the carboxysome operon [71] , [72] ( Fig . 4 , Dataset S1 ) . Other proteins encoded in over half of the 20 chemoautotroph carboxysome loci are bacterioferritin , the accessory proteins CbbO and CbbQ involved in RuBisCO activation , a UPF0753 family protein of unknown function , and a homolog to the chromosome partitioning protein ParA ( Dataset S1 ) . ParA has been implicated in spatial arrangement of beta-carboxysomes in the cyanobacterium Synechococcus elongatus PCC 7942 [73] . Approximately half of the loci in chemoautotrophs encode a protein 34–39% identical to NdhF ( UniProtKB: P31971 ) , part of the NDH-1 complex; some paralogs of NdhF are involved in CO2 uptake ( reviewed in [74]; [75] ) . Cyanobacterial alpha-carboxysome loci are found in the Prochlorococcus and marine Synechococcus genomes ( Table S2 ) . In addition to the core alpha-carboxysome proteins , these loci always encode a Ham1 family protein of unknown function on the opposite strand . ( Dataset S1 ) . Furthermore , nearly all of these loci encode the light-independent ( dark-operative ) protochlorophyllide reductase ( DPOR ) subunits ChlB , ChlN , and ChlL and the light-dependent protochlorophyllide reductase ( LPOR ) ( Dataset S1 ) . However , the benefit of co-localizing the LPOR and DPOR chlorophyll biosynthesis systems with the carboxysome operon is unclear . The NDH-13/NDH-14 CO2 uptake system , which includes NdhD , NdhF , and ChpX/Y ( reviewed in [74] ) , is encoded in the marine Synechococcus loci ( Dataset S1 ) . Beta-carboxysome loci vary significantly from one another , as depicted by the loose locus similarity network ( Fig . S1 ) . In addition to the shell proteins CcmK and CcmL , the only proteins encoded in every beta-carboxysome locus are the γ-carbonic anhydrase homolog , CcmM [76] , and CcmN [37] ( Fig . 4 , Dataset S1 ) ; both are essential for beta-carboxysome formation [17] , [37] , [77] . Interestingly , the carboxysome signature genes , the RuBisCO large and small subunits ( RbcL and RbcS ) , are only encoded in approximately a quarter of beta-carboxysome loci ( Fig . S2 ) , almost always with the RuBisCO chaperone RbcX ( Dataset S1 ) . In contrast , the presumed accessory genes for inorganic carbon uptake , the NDH-13/NDH-14 gene cluster present in the marine Synechococcus alpha-carboxysome loci , are encoded in 70% of the beta-carboxysome loci ( Fig . 4 , Dataset S1 ) . Certain pfams are highly abundant across many different heterotrophic BMC loci . Notably , the majority of Candidate/Confirmed BMC Loci analyzed with LoClass ( excluding carboxysome loci ) contain the metabolosome core genes ( Fig . 1B ) for AldDH ( PF00171; 94% ) , AlcDH ( PF00465; 76% ) , and PduL ( PF06130; 66% ) , a PTAC , as well as PduV/EutP proteins of unknown function ( PF10662; 68% ) . 57% of satellite-like loci encode a PduV/EutP pfam nearby the BMC shell protein gene ( s ) ( Dataset S1 ) . The PDU1A-D loci are relatively syntenic ( Fig . 4 ) ; the sub-types generally group phylogenetically but are confined to two phyla ( Fig . 6 , Table S1 ) . PDU1A loci include the experimentally characterized propanediol utilization operon found in Salmonella enterica [22] ( Table S2 ) . The PDU1A and C loci contain part of the cob operon , which encodes the accessory function of synthesizing cobalamin ( Dataset S1 ) . Cobalamin , or vitamin B12 , is a required cofactor for propanediol dehydratase , the signature enzyme of the PDU loci [78] . Other PDU1 sub-types are typically distinguished by the absence of PDU1A ancillary genes . For example , PduS oxidoreductase , which is involved in cobalamin biosynthesis [79] , is not found in PDU1C loci . PduW , a propionate kinase , is absent from PDU1B loci ( Fig . S2 ) . The PocR regulatory protein is not present in PDU1C and PDU1D sub-types . An interesting difference between PDU1D loci and other PDU1 loci is the presence of a putative two component regulatory system , in which the histidine kinase contains a PocR-domain ( Dataset S1 ) . The PocR domain is also found in the AraC family regulatory protein of the PDU1A loci as well as in several other uncharacterized regulatory proteins [80] . This is a previously undescribed regulatory mechanism for the PDU metabolosome , though a similar PocR-domain containing two-component system has been described for the 1 , 3-propanediol synthesis operon in Clostridium butyricum [81] . Cluster 2 . 1 also contains a PDU-like locus , found in Clostridium kluyveri . This appears to be a reduced PDU locus , lacking the medium subunit ( PduD ) of the diol dehydratase , and the core AldDH and AlcDH enzymes ( Fig . 4 , Fig . S2 ) . Another unique feature is the presence of four PduS homologs ( Fig . 4 ) . The additional AlcDH and AldDH necessary to form a metabolosome core enzyme set may be supplied by the ethanol utilizing ( ETU ) locus ( Fig . 4; discussed below ) elsewhere in the genome . Additional observations and comparisons of the various PDU locus ( sub ) types are included in Text S1 . EUT1 loci contain the experimentally characterized ethanolamine utilization operon in Salmonella enterica . All EUT1 loci except for one are found in the Gammaproteobacteria ( Table S1 , Table S2 ) , the majority in pathogenic enterobacteria . The EUT2 loci form a much looser similarity network than those of EUT1 ( Fig . S1 ) , indicating a great deal more variety between the loci . While nearly all EUT1 loci are found in one phylum , EUT2 sub-types are found in organisms inhabiting diverse environments; representatives are found in four phyla , the majority in Firmicutes ( Table S1 , Fig . S3 ) . EUT3 loci are found only in two Desulfitobacterium hafniense strains . All three of the EUT locus types encode both subunits of the signature enzyme ethanolamine ammonia lyase as well as its reactivating factor , but they differ in presence and type of core metabolosome components , regulatory proteins , and genes that encode ancillary functions . Three features distinguish nearly every EUT2 locus from the EUT1 loci . First , while EUT2 loci usually encode a PduL-like PTAC , EUT1 loci uniquely among BMC loci encode the pta-like PTAC EutD ( Fig . 4 , Fig . S2 ) . Additionally , nearly half of the EUT2 loci lack a complete metabolosome core; only EUT2A encodes the EutG AlcDH ( Fig . 4 ) . Moreover , in place of the EutR regulatory protein [82] found in EUT1 loci , a two-component regulatory system [83] is encoded in 87% of the EUT2 loci ( Fig . S2 ) . By contrast , EUT3 loci lack all of the metabolosome core enzymes but encode the signature enzyme subunits . D . hafniense also encodes two additional BMC loci which could contribute the AldDH and PTAC enzymes for the EUT BMC , but none of these loci encode an obvious AlcDH , indicating that EUT3 may not be a functional BMC or that the requisite AlcDH function is one encoded elsewhere in the genome . Notably , LoClass detected several genes that have not previously been linked to ethanolamine utilization in 90% of EUT1 loci ( Fig . 4 , Dataset S1 ) . For example , LoClass underscored a connection between the eut operon in EUT1 and a gene in the locus encoding malic enzyme MaeB , but its significance to the function of the EUT1 BMC is unknown , although there are hints of a connection . The C-terminus of MaeB contains a non-functional EutD-like PTAC domain [84] . MaeB is inhibited by acetyl-CoA [84] , the substrate of the PTAC reaction . Also encoded in the EUT1 locus are the accessory proteins HemF coproporphyrinogen III oxidase and YfeX porphyrinogen oxidase [85] , which may be involved in cobalamin metabolism . Only EUT2C loci lack the accessory protein EutT ( cobalamin adenosyltransferase ) and EutQ ( a member of the cupin family of unknown function [86] ) , otherwise found in EUT1 and all other EUT2 locus sub-types ( Dataset S1 ) . Both EUT2B and 2D loci encode the PduS oxidoreductase , a flavoprotein , and an acetate kinase , which could convert acetyl-phosphate to acetate , generating ATP ( Fig . 1B ) . Instead of the two-component regulatory system found in other EUT2 loci , EUT2D encodes a PocR domain-containing regulatory protein . Additional observations and comparisons of the various EUT locus ( sub ) types are included in Text S1 . LoClass identified the first examples of fusions of BMC loci . With one exception ( a PDU/GRM fusion , Dataset S1 ) , these are combinations of various PDU and EUT loci . They are found primarily in the genus Listeria , as well as in the genomes of the actinobacterium Propionibacterium sp . oral taxon 192 str . F0372 , the firmicute Streptococcus sanguinis SK36 , and the fusobacterium Sebaldella termitidis ATCC 33386 ( Table S2 ) . This locus has been previously described in Listeria species [87]–[89] . Although the EUT portion of the fusion was assumed to be comparable to the canonical eut operon ( EUT1 in the LoClass taxonomy ) , it is instead more closely related to EUT2A loci ( Fig . 4 ) . This is fused to a rearranged PDU1A locus that contains additional AlcDH genes at its terminus ( Fig . 4 ) . Despite the merging of the PDU1A and EUT2A loci in Listeria , there is no evidence to suggest that they are co-transcribed . Indeed , the regulatory protein PocR is present in the PDU region of the fusion locus , and the two-component regulatory proteins common to EUT2A loci are encoded in the EUT2 region ( Fig . 4 ) , suggesting that these are independently regulated . Both regions include ancillary genes related to cobalamin synthesis ( Dataset S1 ) . The non-Listeria PDU/EUT loci are quite different from those of Listeria and from each other . The most significant examples are in S . sanguinis and Propionibacterium , where the order in which the PDU and EUT loci appear in the genome is inverted compared to the order in Listeria ( Dataset S1 ) , suggesting that they may have originated from fusion events distinct from that which generated the Listeria fusion locus . Several selective pressures may potentially have driven the repeated fusion events of these two particular classes of loci , such as the benefit of coregulation with the cobalamin biosynthesis genes found in the merged locus , since both PDU and EUT require vitamin B12 as a cofactor . A PDU/EUT merger would be significantly more effective if the host organisms are commonly exposed to environments that contain both propanediol and ethanolamine . For example , the PDU and EUT loci have been implicated in improving the intracellular growth ( and , as a result , virulence ) of S . enterica [90] and L . monocytogenes [88] , implying that propanediol and ethanolamine are relevant to pathogenesis . However , the PDU and EUT loci are separate in S . enterica , suggesting that other factors may have driven a fusion event in L . monocytogenes , such as evolutionary pressures for a reduced genome; L . monocytogenes has a genome of 3 Mbp , while S . enterica has a genome of almost 5 Mbp . Similar evolutionary forces may have been at play to fuse the PDU and EUT loci in the non-pathogenic strains . LoClass identified five distinct types of loci that contain the metabolosome core enzymes and a glycyl radical enzyme ( Fig . S1 ) , which , with its activating enzyme , we predict to be the signature enzymes of this class of metabolosome ( Fig . 4 ) . GRM loci are widespread , found in members of the phyla Actinobacteria , Firmicutes , Proteobacteria ( Table S1 ) , and are differentiated by their complement of shell proteins and accessory genes , such as regulators , transporters , and other genes that could encode ancillary functions . The GRM1 locus ( Cluster 2 . 2 ) is mainly found in the Firmicutes but also in some species of the Deltaproteobacteria and Olsenella uli , a member of the Actinobacteria ( Table S2 ) . This locus encodes one additional AldDH ( Fig . 4 , Dataset S1 ) , suggesting that this metabolosome may degrade several different aldehydes . Accordingly , we aligned all AldDH sequences found in metabolosome loci to identify differences in the active site ( Dataset S6 ) . Surprisingly , we observed that in all GRM1 loci with two AldDHs , one of the genes contains a mutation in the catalytic cysteine to either a serine or proline , indicating that the enzyme cannot efficiently catalyze a dehydrogenation reaction [91] , [92] . Catalytically defunct enzyme domains are likewise found in the carboxysome [37] , [68] , [93] . Frequently , nonfunctional enzyme domains act as scaffolds or regulators in various systems ( reviewed in [94] ) ; this second AldDH could have similar functions in the GRM1 loci . Apart from the signature and core enzymes , other notable ancillary proteins in the GRM1 loci include homologs to PduS , PduV/EutP , EutQ , EutJ , a multidrug resistance ( MDR ) efflux transporter , and a predicted transcription factor ( Dataset S1 ) . In the PDU metabolosome , PduS is involved in the biosynthesis of vitamin B12 , an essential cofactor for propionaldehyde dehydrogenase [79] , [95] , [96] . Its conservation in GRM1 loci is unexpected , as there are no B12-dependent enzymes present in the locus . PduS has an iron sulfur cluster-binding domain , as does PduT ( a BMC-T shell protein ) and the pair of proteins has been proposed to be involved in electron transport across the shell [35] , [97] . The cysteine coordinating the 4Fe-4S cluster in PduT is conserved in the BMC-T protein in the GRM1 loci . Therefore , it is plausible that the PduS homolog could accept electrons from the PduT homolog , either for catalytic purposes , and/or for shuttling electrons out of the shell . PduV , EutP , EutQ , and EutJ are all proteins of unknown function but are conserved members of their respective Confirmed ( PDU1 or EUT1 ) Loci . The glycyl radical enzyme of the GRM1 locus has been experimentally characterized; it is a choline lyase , producing trimethylamine and acetaldehyde [98] , similar to the EUT BMC in which the signature enzyme produces ammonia and acetaldehyde [19] , [99] . The GRM2 locus ( Cluster 2 . 3 ) is only found in the Gammaproteobacteria and mainly in pathogens . The only ancillary genes encoded in this locus are three distinct transcription factors ( with only one found in all GRM2 loci ) and two multi-drug resistance proteins ( Fig . 4 ) . Due to their conservation within the GRM2 locus , the transporters are likely related to the function of the metabolosome , perhaps as transporters for substrate . Interestingly , this locus does not contain any genes for BMC-T shell proteins , while the other GRM loci do ( Fig . 4 ) , suggesting GRM2 metabolosomes differ from the other GRM BMCs in some aspect of metabolite flux across the shell . Although functional metabolosomes lacking BMC-T proteins are known [25] , the majority of BMC loci ( 27 of 30 Candidate/Confirmed BMC Locus ( sub ) types; Fig . 4 , Dataset S1 ) encode one or more BMC-T proteins . The GRM3 locus ( Cluster 2 . 4 ) is found in both innocuous and pathogenic species of the Clostridium , Desulfosporosinus , and Oscillibacter genera of the Firmicutes , as well as in various alpha- and gammaproteobacteria ( Table S1 , Table S2 ) . Ancillary proteins encoded in the loci include acetate kinase , a peptidase , a flavoprotein , a EutJ homolog , S-adenosylmethionine synthetase , a pair of two-component signaling proteins , and a protein with a domain of unknown function ( DUF336; Dataset S1 ) . The potential role of a peptidase in the context of a BMC is unclear; perhaps it plays a regulatory role . The flavoprotein could be involved in a peripheral enzymatic step ( including a possible role as an electron shuttle ) . S-adenosylmethionine ( SAM ) is a required cofactor for glycyl radical enzyme chemistry ( reviewed in [100] ) , and the presence of SAM synthetase indicates that the locus encodes the accessory function of synthesizing SAM from ATP and methionine . Along with an adenosyltransferase domain , DUF336 is one of two domains found in PduO , an enzyme involved in the synthesis of vitamin B12 in the PDU BMC [101] . The GRM4 locus ( Cluster 2 . 5 ) is restricted to two Shewanella species ( Table S2 ) . Without additional examples for comparison , we confined predictions of ancillary genes to those found within the region bounded by the core AlcDH and shell proteins ( Fig . 4 ) . In this set of genes , there is a major intrinsic protein ( 65 . 59% identity to the PduF transporter in S . enterica ) , as well as a protein containing the PduO component domain DUF336 ( see above; Dataset S1 ) . The GRM5 locus ( Cluster 4 . 2 ) is only found in firmicutes of the Ruminococcus genus , Roseburia inulinivorans , and Clostridium phytofermentans ( Table S2 ) . Ancillary genes within this locus consist of a transcriptional regulator in the DeoR family , a PduS homolog , a class II aldolase , a hydrolase , and a protein with DUF336 ( Dataset S1 ) . Transcriptional profiling of this locus suggests that its function is the anaerobic degradation of L-fucose and L-rhamnose , similar deoxy sugars [24] , [102] . The aldolase is expected to cleave the hexose , one product being lactaldehyde that is further converted to propanediol , which the glycyl radical enzyme is expected to dehydrate to propionaldehyde [24] . The predicted AlcDH [24] in this locus contains the zinc-binding dehydrogenase pfams , which are different than the typical AlcDH pfam found in BMC loci ( iron-binding ) , indicating that the specific type of a core component can be plastic as long as the function is maintained . One BMC locus ( Cluster 4 . 1 ) is almost exclusively restricted to species in the phyla Planctomycetes and Verrucomicrobia ( Table S1 ) . The metabolosome encoded by this locus has recently been experimentally characterized; it is involved in the aerobic degradation of L-fucose and L-rhamnose [25] . Its signature enzyme is a class II aldolase ( homologous to that in GRM5 ) , and ancillary genes include a DeoR-family transcriptional regulator and an acetate kinase ( Dataset S1 ) . Again , the hypothesis that some plasticity is tolerated in the core enzyme composition is supported by the PVM BMC; a gene with the lactate/malate dehydrogenase pfams was predicted to provide core AlcDH function for the PVM BMC , as it is the only gene ( other than AldDH ) present in the locus that could regenerate NAD+ [25] . PVM-like loci ( Cluster 4 . 2 ) are compositionally distinct yet cluster with the PVM loci ( Fig . 3 ) . The PVM-like loci are found in diverse members of the Firmicutes , the actinobacterium Candidatus Solibacter usitatus , and the ignavibacterium Melioribacter roseus ( Table S2 ) . A class II aldolase is found in each these loci , and a DeoR family transcriptional regulator is found in most , but only one encodes the core AlcDH ( Dataset S1 ) . The complement of shell protein genes is also peculiar; most of the PVM-like loci encode more BMC-H than BMC-P proteins; only M . roseus and one locus in S . usitatus have ratios characteristic of the PVM loci , which generally have more BMC-P than BMC-H proteins ( Table S2 ) . One locus completely lacks BMC-P genes , but this is one of two PVM-like loci in the S . usitatus genome; the other locus does encode BMC-P genes , and the gene products of the two loci could constitute a functional organelle ( Dataset S1 ) . Given that the GRM5 locus also contains an aldolase , these similarities in shell protein complement may indicate that most of the PVM-like loci are more closely related to the GRM5 loci , while the loci in M . roseus and S . usitatus could be more closely related to the PVM loci . Indeed , by comparing the genetic organization between the GRM5 , PVM-like , and PVM representative loci in Fig . 4 , the PVM-like representative locus is more syntenic to the GRM5 locus than to the PVM locus , where the gene order is completely different . Considering these observations , the PVM-like loci may be fragments of a related locus , perhaps GRM5 . One BMC locus , present in two Clostridium kluyveri species , forms a distinct subgroup of Cluster 2 ( Fig . S1 ) . Although this locus fits our definition of a satellite locus , this set of genes has been implicated in ethanol degradation , where ethanol is oxidized to acetyl-CoA in two enzymatic steps [47] , [103] , [104] . The locus contains an incomplete metabolosome core , only encoding two AldDHs with identical sequences and three AlcDHs ( Fig . 4 ) , which can assemble into an aldehyde-alcohol dehydrogenase complex [103] , [104] . The proposed biochemical pathway for ethanol utilization , ethanol to acetaldehyde to acetyl-CoA , consumes two NAD+ and one CoA . However , it is not obvious how these cofactors are recycled; the locus lacks a PTAC , and using any of the AlcDHs to recycle NAD+ would result in a futile cycle . It is appealing to solve this conundrum by positing that a BMC shell is not formed , but polyhedral structures have been observed in this species when fermenting ethanol and acetate [105] . Interestingly , the locus also lacks genes for BMC-P proteins , which are required to complete the diffusional barrier formed by the shell; shells deficient in BMC-P proteins are more “leaky” [18] , [28] . It may be that the ETU BMC shell is open enough that cofactor recycling within the organelle is not required . Although a leaky shell may not effectively sequester acetaldehyde , there may be another benefit to spatially clustering the enzymes , such as substrate channeling . Alternatively , genes from the PDU-like locus present in this species ( Fig . 4; see above ) could be co-opted , providing the BMC-T , BMC-P , and PTAC genes to form a complete metabolosome . In addition to Candidate/Confirmed Loci that we could classify and for which we could infer putative BMC-related reactions , we also identified five locus types for which we could not easily predict a function . Four of these are distinctly different from metabolosomes in that they do not encode a complete biochemical core . Two of these locus types are found exclusively in members of the phylum Actinobacteria – specifically in several species of the genus Mycobacterium and in Rhodococcus jostii ( Table S2 ) ; accordingly , we refer to these as the Rhodococcus and Mycobacterium Microcompartment ( RMM ) loci . Two more locus types were observed in phylogenetically distinct organisms; these are designated Metabolosomes of Unknown Function ( MUF ) or Metabolosomes with an Incomplete Core ( MIC ) . The major difference between the two types of RMM loci is the presence ( RMM2; Cluster 2 . 1 . 2 ) or absence ( RMM1; Cluster 5 ) of diol dehydratase genes ( Fig . 4 ) . Putative ancillary proteins encoded in the RMM1 locus include multiple hydrolases , a phosphotransferase , a short chain dehydrogenase , an amino acid permease , and a regulatory protein of the GntR family ( Dataset S1 ) . Amino-2-propanol has been shown to be an inducer of the short chain dehydrogenase via GntR [106] , and the short chain dehydrogenase has been shown to convert amino-2-propanol to aminoacetone [107] . If the aminotransferase were to act on aminoacetone , methylgyloxal would be produced , which is extremely toxic ( reviewed in [108] ) . This toxicity could be alleviated by conversion of methylglyoxal to pyruvyl-CoA by the core AldDH . However , as mentioned above , both the NAD+ and CoA used in this step must be regenerated somehow within the compartment , and the core metabolosome genes to do so are not obviously present . Perhaps the aforementioned ancillary enzymes form a circuit that recycles the required cofactors . For example , a hydrolase could regenerate CoA to form lactate . Alternatively , it may be that some BMC shells allow cofactors across , or that the AldDH may not be acylating , at least not requiring CoA to be recycled . The RMM2 locus contains the same set of genes as RMM1 but has several additional genes and diol dehydratase homologs , while the hydrolases present in RMM1 are absent ( Dataset S1 ) . Based on their annotations , none of the constituent genes appear to catalyze the regeneration of NAD+ or CoA . The subset of genes in the locus spanning from the permease to the transcription factor are highly syntenic between RMM1 and RMM2 ( Fig . 4 ) , suggesting that RMM1 and RMM2 metabolosomes function similarly but have different peripheral reactions . Furthermore , the diol dehydratase genes are on the opposite strand from these ancillary genes and would not be on the same polycistronic message as the rest of the genes in the locus . We interpret these observations to indicate that the RMM2 BMC does not utilize the diol dehydratase genes and is instead involved in amino alcohol degradation . Alternatively , the BMC may be bifunctional , capable of using both the diol dehydratase and aminotransferase as signature enzymes . The MUF locus ( Cluster 2 . 6 ) is present in Clostridium botulinum B str . Eklund 17B and Clostridium botulinum E3 str . Alaska E43 ( Table S2 ) and includes all metabolosome core enzymes but lacks a prospective signature enzyme ( Fig . 4 ) . This locus also encodes BMC-H , BMC-T , and BMC-P as well as PduS- and EutJ-like proteins ( Dataset S1 ) . These observations indicate that MUF is likely a functional metabolosome , possibly involved in degradation of an unspecified aldehyde , or that its signature enzyme is recruited from elsewhere in the genome . The MIC1 ( Cluster 7 ) metabolosome locus contains a dehydrogenase in addition to AldDH and two putative AlcDHs , as well as a hydrolase , a phosphatase , and an MreB-like protein ( Dataset S1 ) , but lacks a PTAC and a signature enzyme , which may be encoded elsewhere in the genome . One dehydrogenase contains the lactate/malate dehydrogenase pfams , similar to the predicted core AlcDH in the PVM locus , while one contains the zinc-binding dehydrogenase pfams found in the GRE5 locus , so we could not resolve which provides the core recycling function . Due to the low number of conserved genes in the locus , it is not possible to predict how CoA is recycled . However , as for the RMM loci , there are possible alternatives for circumventing this necessity ( discussed above ) . Four additional MIC loci ( Dataset S1 ) , each with incomplete metabolosome cores , were identified , and there is only one representative of each . These include loci from the actinobacterium Verrucosispora maris AB-18-032 and the firmicute Mahella australiensis 50-1 BON , each of which encodes a class II aldolase , which could serve as the signature enzyme ( Dataset S1 ) . Other MIC loci are found in the deltaproteobacterium Haliangium ochraceum and in Synergistetes bacterium SGP1 ( Dataset S1 ) . Based on the presence of an AldDH , the RMM and MIC BMCs likely encapsulate an aldehyde , but the absence of AlcDH and/or PTAC to recycle CoA and NAD+ indicates cofactor recycling is not absolutely required . If this hypothesis is correct , these BMCs would be inaugural members of a new class of aldehyde-processing metabolosomes . We also identified loci found in candidate phyla mostly comprised of incomplete single-cell genomes and compared them to the BMC taxonomy ( Fig . 4 ) . Many of these loci are incomplete , occurring at the end of a scaffold . However many fragments are long enough to show synteny to other loci of the same type and to differentiate them from Candidate/Confirmed BMC Loci identified in NR ( Dataset S2 ) . The majority of these divide into two main types , discussed below . One Candidate BMC Locus type found in single-cell genomes belonging to the Atribacteria , Gemmatimonadetes , and Marinimicrobia genera ( Fig . 6 , Dataset S2 ) contains an aldolase 52–57% identical to DeoC deoxyribose-phosphate aldolase ( EC: 4 . 1 . 2 . 4; UniProtKB: Q9X1P5 ) , a class I aldolase that cleaves 2-deoxyribose 5-phosphate into glyceraldehyde-3-phosphate and acetaldehyde [109] . These loci also contain an AldDH , possibly forming acetyl-CoA from acetaldehyde . Moreover , all but Marinimicrobia contain a PduL PTAC ( Dataset S2 ) . All of these loci contain a sugar isomerase 48–53% identical to RpiB ribose-5-phosphate isomerase ( EC: 5 . 3 . 1 . 6; UniProtKB: A3DIL8 ) , which isomerizes ribose-5-phosphate to ribulose-5-phosphate [110] , [111] . A number of RpiB proteins exhibit low substrate specificity , isomerizing a wide variety of sugars [112] . In Marinimicrobia , the aldolase and isomerase occur as a fusion protein . Of these loci , only those in Atribacteria contain a PduS homolog , the oxidoreductase associated with cobalamin biosynthesis . However , cobalamin is not a required cofactor for DeoC type aldolases . As AlcDH is the only core enzyme absent from these loci , it is possible that PduS may generate an alcohol from an aldehyde , as predicted in one GRM5 locus [24] . PduS also may be present in these loci as a remnant of common ancestry with the pdu operon . The Atribacteria bacterium JGI 0000079-F20 locus additionally encodes a protein 44% identical to triosephosphate isomerase ( TIM ) ( EC: 5 . 3 . 1 . 1; UniProtKB: P00943 ) , which catalyzes the reversible conversion of glyceraldehyde-3-phosphate to dihydroxyacetone phosphate . When acted on by an aldolase , sugar phosphates can be metabolized to either or both of these products , in some cases resulting in another product such as an aldehyde , which could be metabolized according to the canonical metabolosome model . Thus , based on the presence of TIM and a DeoC-like aldolase , we name these BMCs putative sugar phosphate utilizing ( SPU ) metabolosomes with an incomplete core; most of the enzymes encoded in the SPU loci fit into the deoxyribulose/deoxyribose 5-phosphate degradation pathway for DNA catabolism . The second Candidate BMC Locus type that appears in single-cell genomes is found in two representatives from the candidate phylum Latescibacteria . This locus is similar to the SPU type , containing genes encoding an AldDH , PduL , PduS , an RpiB-like protein , and an aldolase ( Dataset S2 ) , so we refer to it as SPU-like . Interestingly , the aldolase is different than in SPU; instead , a class II aldolase like those found in several other BMC loci is present . Despite this difference , the enzyme could still generate an aldehyde from a sugar phosphate as in the GRM5 and SPU locus types . In addition , this locus contains a protein 40–44% identical to acetate kinase ( EC: 2 . 7 . 2 . 1; UniProtKB: Q9WYB1 ) . Neither the SPU nor SPU-like loci had been observed prior to the analysis of these single-cell genomes from candidate phyla , indicating that with continued sequencing of microbial “dark matter” [52] many more BMC loci types may be discovered . Several additional Candidate and satellite-like loci found in single-cell genomes could only be classified as metabolosome loci with incomplete cores . These were identified in members of the candidate phyla Caldithrix , Poribacteria [113] and BRC1 ( Dataset S2 ) . Caldithrix abyssi contains a locus which encodes AldDH , a pta type PTAC , and PduS . However , the absence of an apparent signature enzyme necessitates the designation of this locus as a MIC . Each sequenced locus from Poribacteria and BRC1 meets all of the criteria of a satellite-like locus ( Dataset S2 ) which may indicate that the region of the genome encoding the core locus has not been sequenced , or that BMC-related genes are constitutively expressed and so are not required to be proximal to other BMC-related genes; several of these satellite loci encode a PduL PTAC or an AldDH , indicating that these BMCs may function as metabolosomes . Based on our observation that the majority of heterotrophic BMC loci encode an AldDH ( Fig . 4 , Fig . S2 ) , we hypothesized that these protein sequences may be used to construct a phylogeny of this large subset of BMCs , which could be used to validate the LoClass method . We excluded the presumably catalytically defunct AldDHs found in the GRM1 loci ( discussed above ) , but we did include other duplicate AldDHs that were not obviously altered in any catalytic residues and could remain functional ( i . e . we could not resolve which one was the core AldDH ) . In the resulting phylogenetic tree , we observed two major “trunks” that were separated by a relatively long branch ( Fig . 5A ) and rooted the tree at this branch ( Fig . 5B ) . All PDU , RMM , PVM/PVM-like , and SPU/SPU-like associated AldDHs cohere with their respective trunks , while the other BMC ( sub ) types do not initially seem as polarized; sequences from EUT and GRM loci were found on both trunks ( Fig . 5B ) . However , when mapping LoClass's sub-clustering results onto the phylogeny , we observed EUT and GRM subgroups to also display bifurcated localization: EUT1 , GRM3 , GRM4 , and GRM5 sequences predominantly localized to one trunk , while EUT2 , GRM1 , and GRM2 sequences clustered on the other trunk . Furthermore , sequences from most metabolosome loci formed clades corresponding to the locus ( sub ) types shown in Figures 3 and 4 ( Fig . 5B ) . The only major exceptions to this are two GRM1 taxa that branch close to EUT1 , two GRM3 taxa that branch close to the AldDH from the PDU/EUT locus in Listeria , and one GRM3 sequence adjacent to ETU taxa ( asterisks in Fig . 5 ) . The outlier GRM1 sequences belong to a second GRM1 locus in their respective organisms ( Desulfotalea psychrophila LSv54 , Desulfitobacterium hafniense DCB-2 ) ; this may indicate that this additional locus has a distinct function . The GRM3 outliers near the PDU/EUT sequences are secondary AldDH genes in their respective loci ( Clostridium cf . saccharolyticum , Clostridium beijerinckii ) . The GRM3 AldDH near the ETU branch belongs to Clostridium novyi; its proximity to ETU may indicate an evolutionary link between GRM3 and ETU . In addition , some GRM3 taxa were found nested within the large GRM1 clade , which may indicate that they share common origin with GRM1 loci . Thus , apart from these few outlier taxa , we observed congruence between the molecular phylogeny of one gene ( AldDH ) and our clustering results based on total pfam complement of a locus ( Fig . 3 , Fig . S1 ) , validating our LoClass-based metabolosome taxonomy . In contrast to all other metabolosome loci , including MIC loci , Cluster 6 , found only in four different species of Firmicutes , lacks all metabolosome core enzymes but does encode BMC-H , BMC-T , and BMC-P genes ( Fig . 4 ) , suggesting that a complete shell is formed . Accordingly , we refer to this locus type as BMC of Unknown Function ( BUF ) . Many of the genes encoded within the BUF locus are putative enzymes , including amidohydrolases , deaminase , dehydrogenase , carboxy-lyase , carbon monoxide hydrogenase , isochorismatase , and formiminotransferases ( Dataset S1 ) , making it difficult to predict a function . However , the presence of amidohydrolases and deaminases suggests that nitrogen-containing compounds are processed via this Candidate BMC Locus . Alternatively , it is possible that this BMC encapsulates enzymes from the core metabolic model that are scattered throughout the genome . However , since the enzymes in the locus are conserved across these loci , and no other loci besides the carboxysome maintain a conserved locus while lacking all core metabolosome enzymes , these observations , and notably the absence of the core AldDH pfam , suggest that this BMC encapsulates a distinctly different metabolite . This locus appears to encode a new type of BMC that does not conform to the metabolosome or carboxysome functional paradigms . Methods for detection and classification of conserved loci typically approach the problem with the goal of operon detection [114] , [115] , require analysis of multiple whole genomes [115]–[117] , are sensitive to genomic rearrangements and gene order [114] , [116] , or through computationally-driven mergers produce extended gene clusters that may not exist in any one genome [116] . Even after a conserved gene cluster is identified , none of these methods enable the direct comparison and classification of these loci by an automated scoring method . LoClass circumvents all of these issues . By “seeding” with BMC shell protein genes , we confined the scope of our analysis to genomic regions most likely to encode BMC-related functions . Then , based on the assumption that genes encoding other BMC structural components and supporting functions would be proximal to the shell protein genes , as in experimentally characterized BMCs , we were able to compare loci based entirely upon the subset of constituent genes that did not encode BMC shell proteins . Because this assumption will likely be valid for other gene clusters encoding macromolecular complexes or metabolic pathways , the LoClass strategy may be useful for detecting contiguous groups of genes that functionally cohere . LoClass enabled us to compare and classify hundreds of BMC loci . Moreover , it allowed detection of functional sub-types , prediction of novel BMC types , and identification of genes that support the function of the BMC that are not structurally part of the organelle . Among these ancillary pfams are , for example , actin-like proteins ( PF06723 and PF11104 ) and ParA ( PF01656 ) ( Fig . 4 , Dataset S1 ) . MreB , a bacterial actin , and ParA have been implicated in spatial localization of carboxysomes [73] . PduV has also been suggested to play a role in the spatial positioning of the PDU BMC [118] , and we found homologs ( containing PF10662 ) to be encoded in most PDU , EUT , and GRM1 loci , as well as in a quarter of all satellite-like loci . The frequent observation of these pfams in BMC loci suggests that spatial positioning within the cell is important to the formation and function of BMCs . Likewise , we found that many Candidate/Confirmed BMC Loci contain genes for transporters; their distinctive types suggest that they may be specific to the substrate ( s ) processed by the BMC . Regulatory genes were also frequently observed; for example , EUT2D , some GRM1 , GRM3 , GRM4 , and ETU loci , in addition to the PDU loci , encode regulatory proteins that contain the PocR domain . Considering that all experimentally characterized metabolosomes are only induced in the presence of their substrate , these putatively BMC-associated genes are not likely to be “genomic hitchhikers , ” genes unrelated to the function encoded by the locus but merely similarly transcriptionally regulated [116]; rather , they are likely genes that provide important supporting functions for the BMC . Recognition of these cryptically associated accessory genes should prove useful in efforts to functionally characterize diverse BMCs . In addition to expanding the vocabulary of pfams associated with bacterial organelle function , LoClass unveiled BMC-associated puzzles , pfams for which we could not readily infer a reason for their conservation in various locus types . For example , GRM1 , GRM3 , GRM4 , GRM5 , MUF , SPU , and SPU-like loci , which do not encode cobalamin-dependent enzymes , encode various genes that have been associated with cobalamin synthesis , such as PduS and PduO ( both full length and a single DUF336 domain ) . These may be relics of the evolutionary history of these BMC loci or , more likely , these genes have been co-opted to perform different functions useful to the BMC metabolism , as has been predicted in the GRM5 locus in C . phytofermentans [24] . On the other hand , the conserved co-localization of the light-dependent and light-independent protochlorophyllide reductase genes with the cyanobacterial alpha-carboxysome loci and of the MaeB malic enzyme with EUT1 loci cannot yet be explained . One intriguing trend uncovered by LoClass is the frequent presence of domains associated with flavin binding or utilization ( PF02441 , PF01593 , PF03358 , PF00941 , PF01494 , and PF00258; Dataset S1 ) . Furthermore , PduS has been shown to bind flavin mononucleotide [95] , [96] . The frequent presence of flavin-binding pfams and proteins in BMC loci suggests that these cofactors could play a previously unrecognized and important role in BMC biochemistry and redox . LoClass also highlighted the diversity within numerically abundant locus types ( Fig . 4 ) ; this has only been attempted previously for EUT loci [119] . Additionally , capturing genes for functions ancillary to the BMC and their contribution to the clustering led to detection of unanticipated relationships . A striking example of this is seen when comparing the carboxysome loci . Due to the low similarity between the core structural components ( i . e . between the encapsulated carbonic anhydrases/CcmN and CsoS2 ) of the alpha- and beta-carboxysome , it was unexpected that they would cluster together . Investigation revealed that the principle cause of this clustering is the NDH-13/NDH-14 CO2 transport gene cluster , which is conserved across cyanobacterial carboxysome loci of both types . Several striking observations from our classification enlarge our view of the diversity of BMCs . New locus types identified in this study are the PDU/EUT and PDU/GRM fusions , GRM2 , GRM3 , GRM4 , PVM-like , RMM2 , MUF , MIC1 , SPU , SPU-like , and BUF locus types . While many of these BMC loci adhere to the complete biochemical model of the metabolosome paradigm ( aldehyde utilization and the regeneration of cofactors , catalyzed by a PTAC [32] and an AlcDH [31] ) , other metabolosome loci ( those containing AldDH but missing PTAC and/or AlcDH ) apparently regenerate or acquire cofactors in as yet unknown ways . LoClass also enabled us to identify new types of metabolosomes ( e . g . multiple types of GRM loci ) . Furthermore , LoClass identified the BUF locus , a novel BMC type that does not conform to either the carboxysome or metabolosome paradigms , presenting a potentially new model of BMC metabolism . The novel locus types we identified are typically found in poorly characterized clades of the Bacterial domain , hinting that there are bacterial organelles of unknown function yet to be discovered . Likewise , the observation of GRM , PVM , PVM-like , MIC , and SPU-like loci in diverse phyla that nevertheless contain homologous aldolases is interesting , given that class II aldolases are generally quite promiscuous [120] . Considering the presumed promiscuity of the core enzymes of metabolosomes , their colocalization with the class II aldolase suggests that these loci may be descendants of a common ancestor , one with a functionally malleable core due to substrate ambiguity of its component enzymes; this would render it readily able to confer new catabolic capabilities by horizontal gene transfer . In addition to providing evolutionary insights , this descriptive , domain-based taxonomy of BMC loci is a guide to uncovering the functional diversity of BMCs and their roles as modules of metabolic specialization in bacteria; this parallels the historical discovery of eukaryotic organelles , in which observation and description laid the foundation for experimental elucidation of function . Domains are the structural , functional , and evolutionary units of proteins; analogously , LoClass captures ( re ) combinations of groups of domains that constitute loci and contribute to BMC function and evolution . Such a comprehensive view of the requisite building blocks for diverse BMC functions can likewise inform the design of BMC loci for the introduction of genetic and metabolic modules for applications in synthetic biology . | Some enzymatic transformations have undesirable side reactions , produce toxic or volatile intermediates , or are inefficient; these shortcomings can be alleviated through their sequestration with their substrates in a confined space , as in the membrane-bound organelles of eukaryotes . Recently , it was discovered that bacteria also form organelles–bacterial microcompartments ( BMCs ) –composed of a protein shell that surrounds functionally related enzymes . BMCs long evaded detection because they typically form only in the presence of the substrate they metabolize , and they can only be visualized by electron microscopy . A few BMCs have been experimentally characterized; they have diverse functions in CO2 fixation , pathogenesis , and niche colonization . While the encapsulated enzymes differ among functionally distinct BMCs , the shell architecture is conserved . This enables their detection computationally , as genes for shell proteins are typically nearby genes for the encapsulated enzymes . We developed a novel algorithm to comprehensively identify and categorize BMCs in sequenced bacterial genomes . We show that BMCs are often encoded adjacent to genes that play supporting roles to the organelle's function . Our results provide the first glimpse of the extent of BMC metabolic diversity and will inform design of genetic modules encoding BMCs for introduction of new metabolic functions in a plug-and-play approach . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"bacterial",
"genomics",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology",
"microbial",
"genomics"
] | 2014 | A Taxonomy of Bacterial Microcompartment Loci Constructed by a Novel Scoring Method |
How pathogenic bacteria infect and kill their host is currently widely investigated . In comparison , the fate of pathogens after the death of their host receives less attention . We studied Bacillus thuringiensis ( Bt ) infection of an insect host , and show that NprR , a quorum sensor , is active after death of the insect and allows Bt to survive in the cadavers as vegetative cells . Transcriptomic analysis revealed that NprR regulates at least 41 genes , including many encoding degradative enzymes or proteins involved in the synthesis of a nonribosomal peptide named kurstakin . These degradative enzymes are essential in vitro to degrade several substrates and are specifically expressed after host death suggesting that Bt has an active necrotrophic lifestyle in the cadaver . We show that kurstakin is essential for Bt survival during necrotrophic development . It is required for swarming mobility and biofilm formation , presumably through a pore forming activity . A nprR deficient mutant does not develop necrotrophically and does not sporulate efficiently in the cadaver . We report that necrotrophism is a highly regulated mechanism essential for the Bt infectious cycle , contributing to spore spreading .
Saprophytism , probably one of the most common lifestyle for micro-organisms , involves living in dead or decaying organic matter . For most pathogens , saprophytism is limited to necrotrophism ( the use of the host cadaver ) . This step of the infection process is essential for the proliferation and horizontal transmission of these microorganisms ( transfer of infection within a single generation ) [1] . However , there have been very few studies addressing this major issue . The transition from a pathogenic to a necrotrophic lifestyle implies substantial metabolic changes for microorganisms [2] . The death of the host is a critical event which compels the micro-organisms to cope with a new series of challenges: competition with the commensal organisms and opportunistic incomers , stress , and nutrient deficiencies . Therefore , necrotrophism is likely to be highly regulated . The insect pathogen Bacillus thuringiensis ( Bt ) is a suitable model for studying the time course of the infection process , including necrotrophism in the insect cadaver . Bt is an ubiquitous spore-forming bacterium belonging to the Bacillus cereus ( Bc ) group [3] . Its spores are found in a large variety of environments , such as soils , dead and living insects and plant phylloplane [4] . However , Bt probably does not grow in soil and reports of natural epizootic episodes are very rare [5] , [6] . Unlike soil bacteria , such as Streptomyces spp and B . subtilis , Bc group genomes contain a large number of genes involved in nitrogen metabolism [3] . It is therefore likely that Bt multiplies in the host cadaver [1] , [6] . Bt carries plasmids encoding specific insecticidal toxins responsible for their insecticidal properties [7] . Bt spores and toxins are ingested by larvae , and the toxins bind to specific receptors on the midgut epithelial cells , inducing cell lysis and creating favorable conditions for the development of the bacteria [8] . The vegetative bacteria multiply in the insect hemocoel and cause septicemia [1] , [9] . Bt also harbors genes encoding exported virulence factors including enterotoxins , hemolysins , phospholipases and proteases [10] . The transcription of most of these virulence genes in bacteria growing in a rich medium is activated at the onset of stationary phase by the quorum-sensing system PlcR-PapR [11] , [12] . PlcR-regulated factors account for about 80% of the extracellular proteome of Bt during early stationary phase in rich medium [13] . In sharp contrast , the expression of the PlcR-regulated genes is repressed when the bacteria enter sporulation [14] and the stationary phase secretome of Bt and B . anthracis ( Ba ) growing in a sporulation medium is mainly composed of the metalloprotease NprA [15] , [16] . NprA ( also designated NprB and Npr599 in Ba ) cleaves tissue components such as fibronectin , laminin and collagen , thus displaying characteristics of pathogenic factors [17] . Transcription of nprA is activated during the late stationary phase by the regulator NprR [16] . NprR is a quorum sensor activated by its cognate signaling peptide , NprX . NprR-NprX functions as a typical Gram-positive quorum-sensing system: the pro-signaling peptide NprX is exported from the cell , and after being processed to its active form is reimported , and binds to NprR allowing the recognition of its DNA target and the activation of nprA transcription [16] . The first stages of Bt infection are relatively well documented , but the fate of the bacteria after death of the host remains unclear . Here , we report evidence that the necrotrophic lifestyle of Bt is a specific and highly regulated process . The quorum-sensing system NprR-NprX controls at least 41 genes some of which are required for Bt to survive in the insect cadaver and to complete its development in vivo ending with the production of spores .
We tested whether NprR , the activator of nprA transcription [16] , is involved in the pathogenicity of Bt . The LD50 s of the Bt 407 Cry− ( wt ) strain and of the nprR-deficient ( ΔRX ) strain in the insect model Galleria mellonella were measured in two ways: by feeding larvae with spores mixed with the insecticidal toxin Cry1C and by injection of vegetative bacteria into the insect hemocoel ( Table S1 ) . The LD50 s of two strains did not differ significantly in either of the two conditions indicating that NprR was not required for pathogenicity . Consistently , an nprA-deficient strain was similarly found not to be affected in pathogenicity ( not shown ) . We investigated the involvement of NprR in the infection process by comparing , in vivo , the expression kinetics of nprA with that of the protease gene mpbE , reflecting the transcriptional activities of NprR and PlcR , respectively [16] , [18] . The reporter strains grew similarly in insect larvae and a constitutively expressed PaphA3-lacZ fusion was used as the reference standard ( Figure 1A and Figure S1A ) . Transcription of mpbE increased between 0 h and 24 h after injection and gradually decreased thereafter . In contrast , nprA transcription was low between 0 h and 24 h , increased between 24 h and 48 h and then decreased sharply ( Figure 1B ) . Thus , NprR is active later in the infection process than PlcR , and after the death of the host . To investigate the role of NprR during the late stage of infection , we compared the growth of the wt and ΔRX strains in insect larvae ( Figure 2A ) . The total population of the two strains increased between 0 h and 24 h to reach about 1×108 cfu/mL . From 24 h to 96 h , the population of the wt strain remained stable , whereas the population of the ΔRX strain decreased sharply: 96 h post infection , the total population of the ΔRX strain was 6-log lower than that of the wt strain . Complementation of the ΔRX strain by pHT304-RX restored the wt phenotype . These findings indicate that NprR substantially improves the survival of Bt in insect cadavers . In sporulating microorganisms , sporulation is generally regarded as the key process ensuring survival in unfavorable conditions . We therefore investigated i ) whether NprR was involved in the sporulation process of Bt in the insect cadaver , and ii ) whether sporulation is responsible for the survival of the bacteria in the insect cadaver . We compared the sporulation efficiencies of the wt and ΔRX strains in both LB and sporulation-specific medium ( HCT ) ( Table S2 ) . In HCT , the sporulation efficiencies of the two strains were similar . However , in LB medium , the total number of viable spores of the ΔRX strain was half that for the wt strain ( 8 . 30×107 vs 1 . 58×108 ) , suggesting that NprR is involved in the sporulation of Bt in rich medium . Next , we monitored the counts of wt and ΔRX strain spores in insect larvae over 96 h ( Figure 2B and Table S2 ) . For the wt strain , heat-resistant spores were detected 24 h after injection and their number increased until 48 h . From 48 h to 96 h , the number of spores remained stable and represented one third of the total bacterial population . The large number of non sporulated bacteria 96 h after the death of the insect suggests that sporulation was not the main mechanism allowing Bt to survive . For the ΔRX strain , less than one percent of the bacterial population was heat-resistant spores throughout the infection process . The decrease in the number of heat-resistant spores from 48 h to 72 h is likely due to the germination of the spores . We suspected that the low number of spores is not a cause but a consequence of the inability of the ΔRX strain to survive in the insect cadaver . To test this idea , we tested the survival of a sigK-deficient strain ( Figure 2C ) : SigK is a sigma factor involved in the transcription of late sporulation genes in the mother cell , and sigK-deficient strains are not able to form viable spores [19] , [20] . The total population of the sigK strain in the insect cadaver was similar to the total population of the wt strain , indicating that NprR ensures the survival of Bt by a process independent of sporulation . The only gene described as being controlled by NprR was nprA . Therefore , we monitored the survival of a ΔnprA strain in infected larvae ( Figure S2 ) . The survival of the wt and ΔnprA strains was similar throughout the experiment , suggesting that other NprR-regulated genes are involved in bacterial survival . Microarray analysis was used to identify other NprR-regulated genes . Gene expression ratios between the wt and the ΔRX strains were determined 3 h after the onset of stationary phase ( t3 ) , when nprA transcription increases sharply [16] . For 107 genes , this expression ratio was greater than 2 ( p<0 . 05 ) ( http://www . ebi . ac . uk/arrayexpress/experiments/E-TABM-790 ) , suggesting that NprR has a direct or indirect effect on their transcription . Thirty-nine genes , with a relative expression ratio greater than 4 , and a significance value ( p ) smaller than 0 . 01 , were considered for subsequent analysis . The genes matching probes for BC2622 , a macrolide glycosyltransferase , and BC3725 , an exochitinase , were also investigated due to their functional similarity to the genes fulfilling these criteria . Quantitative RT-PCR confirmed that these 41 genes were at least four times up- or downregulated . Fusions to lacZ were constructed for nine of these genes and used to confirm that they are differentially regulated in the ΔRX mutant and wt strains ( Figure S3 ) . The expression kinetics of these genes were similar , with a sharp increase of expression after the onset of stationary phase . The final list of NprR-regulated genes is presented in Table S3 . Of the 41 genes directly or indirectly regulated by NprR , 37 were down-regulated in the ΔRX strain , suggesting that NprR primarily acts as a transcriptional activator . The subcellular localizations of the products of these genes were assessed: 46% were cytoplasmic , 20% were associated with the membrane and 34% were extracellular or associated with the cell wall . The NprR-regulated genes can be distributed into four functional groups . The first group is composed of genes encoding proteins potentially involved in stress resistance: they include the genes for cytochrome P450 ( BC2613 ) , cysteine dioxygenase ( BC2617 ) , and Transporter Drug/Metabolite exporter family members ( BC1063 ) . The second group is a four-gene locus encoding the oligopeptide permease system Opp required for the import of small peptides into the cell . The third group is a five-gene locus encoding a nonribosomal peptide synthesis ( NRPS ) system showing similarities with the systems involved in the synthesis of secreted factors like toxins and antibiotics . The last group codes for degradative enzymes ( metalloproteases , esterases and chitinases ) and for proteins which can bind organic material ( chitin-binding protein and collagen adhesion protein ) . The role of NprR in the degradation of lipids , proteins and chitin was analyzed by growing the ΔRX and wt strains on specific culture media ( Figure 3A ) . The lipolytic , proteolytic and chitinolytic activities of the ΔRX strain were significantly lower than those of the wt strain . We monitored the expression kinetics in infected insect larvae of two NprR-regulated genes encoding degradative enzymes ( BC0429 and BC2167 ) ( Figure 3B ) . The two genes were specifically expressed after the insect death , from 24 h to 96 h , suggesting that Bt displays a necrotrophic lifestyle in the insect cadaver . To identify the NprR-dependent survival factor we first tested whether this putative factor was secreted . Insect larvae were co-infected with two different ratios of wt and ΔRX strains: 90% of wt bacteria with 10% of ΔRX bacteria ( 90∶10 ) , and 10% of wt bacteria with 90% of ΔRX bacteria ( 10∶90 ) ( Figure 4A and Figure 4B ) . In insects infected with the ratio 10∶90 , the total population of the wt and the ΔRX strains decreased after 24 h . This may result from the ΔRX strain capturing NprX without being able to express NprR-regulated genes , but nevertheless removing the peptide from the environment . Consequently , the amount of signaling peptide was insufficient to activate NprR-regulated genes in the wt , resulting in clearance of both populations . Co-infection with the ratio 90∶10 led to the survival of the two subpopulations during the 96 h of the experiment . In this condition , the concentration of NprX in the host was presumably sufficient to maintain the expression of the NprR-regulated genes in the wt subpopulation , and this expression allowed the survival of the ΔRX population . Therefore , the wt strain may produce a secreted factor that enables the ΔRX strain to survive in the insect cadaver . NprR-dependent extracellular factors are degradative enzymes and the factor synthesized by the NRPS system . NprA , the major degradative enzyme produced during late stationary phase , is not required for bacterial survival in insect cadaver ( Figure S2 ) . The NRPS locus consists of seven open reading frames annotated BC2450 to BC2456 in the genome of the strain Bc ATCC 14579 used for designing the microarrays [3] . In silico analysis of all available sequenced Bt and Bc genomes , including that of strain Bt 407 used in this study , reveals that in all cases , this locus includes only four genes ( http://www . ncbi . nlm . nih . gov/bioproject/29717 ) . Several studies suggest that these four genes ( designated krsA , B , C , E; Figure 5A ) are involved in the production of the lipopeptide kurstakin [21] , [22] , [23] . KrsE is a presumed efflux protein and KrsA , B , C are the peptide synthetase subunits . The genes krsA , krsB and krsC were deleted from a wt strain and the survival of the mutant ( ΔkrsABC ) in insects was monitored for 96 h ( Figure 5B ) . The total population of the ΔkrsABC strain declined from 2 . 107 cfu/ml at 24 h falling to 1 . 102 cfu/ml at 96 h . To test whether this effect was specifically dependent on the krsABC genes , we introduced a constitutive promoter upstream from these genes in the ΔRX strain . This NprR-independent expression of krsABC partially and significantly restored the survival of the ΔRX strain in the insect cadaver . These observations implicate the krsABC genes in the necrotrophic properties of Bt . We used MALDI-ToF-MS analysis to determine whether the krsABC genes are responsible for the production of kurstakin . Peaks characteristic of kurstakin were found for whole cells of the wt strain and not for those of the ΔkrsABC mutant ( Figure 5C ) . This confirms that the krsABC genes are involved in kurstakin synthesis . We compared properties of the wt and ΔkrsABC strains on swimming plates ( LB Agar 0 . 3% ) and on swarming plates ( LB Agar 0 . 7% and EPS Agar 0 . 7% ) ( Figure 6A ) . The wt and the ΔkrsABC strains grown on LB 0 . 3% agar covered the plates , indicating that both strains were swimming proficient . However , unlike the wt , the ΔkrsABC strain was unable to swarm or to form dendrites indicative of swarming mobility [24] , [25] . Lipopeptides are known to enhance biofilm formation [26] and it has been shown that the Bt 407 strain forms a thick biofilm at the air / liquid interface in glass tubes [27] . We tested the ability of the two strains to form biofilm in glass tubes ( Figure 6A ) . The wt strain produced a significant ring at the air / liquid interface , whereas biofilm formation was abolished for the ΔkrsABC strain . Kurstakin is therefore necessary for swarming and biofilm formation . Lopez and coll . have shown that swarming mobility in B . subtilis is triggered by surfactin , a lipopeptide , which acts as a pore-forming molecule causing potassium leakage across the cytoplasmic membrane [26] . We tested the swarming mobility of the ΔkrsABC strain on swarming plates with nystatin ( a pore-forming molecule ) and with nystatin plus K2HPO4 ( Figure 6B ) . Nystatin restored the swarming mobility of the ΔkrsABC strain and the addition of K2HPO4 reversed this phenotype . These results suggest that kurstakin is a pore forming molecule causing a potassium leakage across the cytoplasmic membrane of Bt .
PlcR is the main virulence regulator in Bt and Bc [10] , [12] and is required for the early steps of the infection process [9] , [28] . We show here that another quorum sensor , NprR , is active after host death and is necessary for Bt to survive in the insect cadaver . NprR is a pleiotropic regulator directly or indirectly affecting the expression of at least 41 genes during the stationary phase . About 30% of the NprR-regulated genes encode extracellular or cell wall-associated proteins involved in the degradation of proteins , lipids and chitin . We report that nprA and two other NprR-regulated genes encoding degradative enzymes were expressed after death of the host . Therefore , it is likely that these enzymes allow Bt to use the content of the host , indicating that Bt displays a necrotrophic lifestyle in the insect cadaver . This nutrient acquisition may support the developmental program of Bt until sporulation . The degradative enzymes may also have other functions . Insect cuticles are made of chitin filaments arranged within a protein matrix which constitutes a physical barrier to the outside environment . Degradative enzymes may degrade this barrier and facilitate spore and toxin release into the environment . Degradative enzymes may also participate in cell protection against competitors . For example , the endochitinase ChiCW was reported as having antifungal properties [29] , and InhA3 ( BC2984 ) is a member of the Immune Inhibitor A metalloprotease family , which plays a key role in the resistance to the host immune defenses by degrading antimicrobial peptides [30] , [31] , [32] . In addition , two substrate-binding proteins ( BC2827 and the BC3526 ) may increase the efficiency of these enzymes . A large locus of three NprR-regulated genes ( krsABC ) codes for a NRPS system involved in the synthesis of a secreted lipopeptide called kurstakin . At least three suggestions could explain the important and surprising function of kurstakin: Some NprR-regulated genes encoding cytoplasmic or membrane-associated proteins may participate in the necrotrophic development of Bt . A putative efflux system ( BC1063 ) , two macrolide glycosyl transferases ( BC2066 and BC2622 ) and a N-hydroxyarylamine O-acetyltransferase could be involved in resistance to antimicrobial molecules , and cytochrome P450 ( BC2613 ) may be involved in resistance to reactive oxygen species . The membrane-associated proteins are mainly components of an oligopeptide permease system ( Opp ) involved in the uptake of PapR , the signaling peptide required for PlcR activation [36] . The operon encoding this Opp system is downregulated by NprR suggesting that NprR controls PlcR expression negatively through the Opp system . Interestingly , nprA expression was only slightly reduced in a oppB mutant strain suggesting that another oligopeptide permease system is involved for the uptake of NprX ( Figure S4 ) . Consequently , a down-regulation of the opp genes could not have a significant effect on the expression of the NprR-regulated genes . Possibly , the extracellular concentration of NprX is a signal triggering the transition from a pathogenic to a necrotrophic lifestyle . During the early stage of infection , the PlcR regulon is expressed and the extracellular concentration of NprX might be low . Indeed , recent results indicate that nprX transcription starts in late stationary phase ( TD , SP , DL , unpublished data ) . In view of these data , we hypothesize that after the host death , the extracellular concentration of NprX increases , leading to the expression of NprR-regulated genes and repression of the PlcR regulon via the Opp transporter . These various observations indicate that the necrotrophic lifestyle of Bt is a complex developmental stage , not limited to simple feeding on the host contents . They also imply that the transition from a pathogenic to a necrotrophic lifestyle is associated with significant metabolic changes . It is becoming clear that the infectious cycle of Bt can be divided into four distinct and sequential phases starting with toxemia caused by the Cry proteins , followed by the action of PlcR in virulence , necrotrophism and the completion of the sporulation process involving NprR , and finally the dissemination of spores ( Figure 7 ) . We describe a nprR-nprX mutant that does not develop necrotrophically and is unable to sporulate efficiently , demonstrating that the necrotrophic properties of Bt are essential for horizontal transmission . It is remarkable that the developmental process through the complete infection cycle in the insect host is coordinated by three regulator-signaling peptide cell-cell communication systems: PlcR-PapR , NprR-NprX and the Rap-Phr complexes which control the phosphorylation of Spo0A [37] , [38] , [39] . Members of the Bc group may be able to follow two different life cycles: an infectious cycle ( described above ) and an endosymbiotic cycle in which the bacteria live in a symbiotic relation with their invertebrate hosts [6] , [40] . Here , we report that Bt multiplies efficiently in the insect cadaver and has a genetic developmental program to live in this biotope . Consequently , we propose the existence of a strictly necrotrophic life cycle in which Bt colonizes a wide variety of dead insects , and uses the cadaver as a bioreactor to multiply and to produce spores and toxins . In silico analysis reveals that nprR is found in all strains of the Bc group , except that in Bc ATCC14579 nprR is disrupted by a transposon [3] . The published genome sequences of Ba and Bc strains provide no evidence for any loss of genetic determinants , which might be crucial for saprophytic survival [41] . Therefore , the function of NprR is probably conserved in the Bc group , and it would be interesting to determine whether the necrotrophic development of Bc and Ba in their mammalian hosts requires the same quorum-sensing regulated network . It is also important to characterize kurstakin to determine how it promotes survival in insect cadavers . Moreover , the properties of this molecule may indicate possibilities for the development of phytosanitary products or adjuvants to improve both the ecological fitness and the efficacy of biopesticides .
The Bt strain 407 Cry− is an acrystalliferous strain cured of its cry plasmid [42] . This strain shares high phylogenic similarity with Bc [43] . Bacillus 407 oppB::tet , Bacillus 407 sigK::aphA3 , Bacillus 407 nprRX::tet ( ΔRX ) , Bacillus 407 nprA::lacZ and Bacillus 407 nprRX::tet nprA::lacZ mutant strains were described previously [16] , [20] , [36] . Escherichia coli K-12 strain TG1 was used as host for the construction of plasmids and cloning experiments . Plasmid DNA for Bacillus electroporation was prepared from the Dam− Dcm− E . coli strain ET12567 ( Stratagene , La Jolla , CA , USA ) . E . coli and Bt cells were transformed by electroporation as described previously [42] , [44] . E . coli strains were grown at 37°C in Luria Broth ( LB ) . Bacillus strains were grown at 30 or 37°C in LB or in HCT , a sporulation-specific medium [45] . The following concentrations of antibiotic were used for bacterial selection: 100 µg/ml ampicillin for E . coli; 200 µg/ml kanamycin , 10 µg/ml tetracycline , 200 µg/ml spectinomycin and 10 µg/ml erythromycin for Bacillus . Numbers of viable cells were counted as total colony-forming units ( cfu ) on LB plates . Numbers of spores were determined as heat-resistant ( 80°C for 12 min ) cfu on LB plates . Force-feeding and intrahemocelic injection experiments with G . mellonella were carried out as described previously [9] . LD50 data were analyzed using the program StatPlus 2007 of Analysoft . Bt cells in living and dead insects were counted as follows . For each strain , each larva was injected with 2 . 104 bacteria and kept at 30°C for 96 h; 24 h after injection , surviving insects were eliminated . At the injection time and every 24 h for the 96 h of the experiment , two larvae were crushed and homogenized in 10 ml of physiological water and dilutions were plated onto LB agar plates containing appropriate antibiotics . To follow the spore population , bacterial colony-forming units were determined before and after treatment of the insect homogenate for 12 min at 80°C . At least four independent replicates were performed for each time and for each strain tested . In vivo ß-galactosidase activity was assayed from 2 ml aliquots of the insect homogenate as described previously [16] . At least three independent measurements were performed for each time and for each transcriptional fusion tested . Chromosomal DNA was extracted from Bt cells using the Puregene Yeast/Bact . Kit B ( QIAgen , France ) . Plasmid DNA was extracted from E . coli using QIAprep spin columns ( QIAgen , France ) . Restriction enzymes ( New England Biolabs , USA ) and T4 DNA ligase ( New England Biolabs , USA ) were used in accordance with the manufacturer's recommendations . Oligonucleotide primers ( Table S4 ) were synthesized by Sigma-Proligo ( Paris , France ) . PCRs were performed in a Applied Biosystem 2720 Thermak cycler ( Applied Biosystem , USA ) . Amplified fragments were purified using the QIAquick PCR purification Kit ( QIAgen , France ) . Digested DNA fragments were separated on 1% ( w/V ) agarose gels after digestion and extracted from gels using the QIAquick gel extraction Kit ( QIAgen , France ) . Nucleotide sequences were determined by Beckman Coulter Genomics ( Takeley , UK ) The plasmid pRN5101 [46] was used for homologous recombination . The low-copy-number plasmid pHT304 was used for complementation experiments with wild-type nprR-nprX genes under their own promoters [16] . Transcriptional fusions were constructed in pHT304-18Z [47] . All the plasmids used in this study are described in Table S5 . The krsABC genes were disrupted by inserting a spectinomycin resistance gene into the coding sequence . The thermosensitive plasmid pRN5101ΩkrsABC::spc was used to disrupt the chromosomal wild-type copy of the krsABC genes in the Bacillus 407 wt strain by homologous recombination as described previously [46] . The recombinant strain , designated Bacillus 407 ΔkrsABC , was resistant to spectinomycin and sensitive to erythromycin . The thermosensitive plasmid pRN5101ΩPkrsABC::aphA3 was used to replace the natural promoter region of the krsABC genes in the Bacillus 407 ΔRX strain by aphA3 and its constitutive promoter . In the resulting Bacillus recombinant strain , the krsABC genes were transcribed from the aphA3 promoter; it was designated Bacillus 407 ΔRX PaphA3-krsABC , and was resistant to kanamycin and sensitive to erythromycin . The methods used to study the proteolitic activity , the chitinolytic activity and the lipolytic activity have been described previously [32] , [48] . Swimming and swarming were evaluated using LB 0 . 3% agar plates and LB 0 . 7% agar plates , respectively . Biofilm formation was assayed in LB medium and in glass tubes as described previously [27] . Dendrite formation was evaluated on EPS 0 . 7% agar . Strains were cultured in LB medium at 37°C until the beginning of stationary phase and 2 . 106 bacteria were spotted onto the center of the agar plate . Plates were incubated at 37°C for 24 h to 96 h . For in vitro ß-galactosidase activity measurements , Bt cells containing lacZ transcriptional fusions were cultured in LB medium at 37°C . In vivo ß-galactosidase activity was assayed from 2 ml aliquots of insect homogenate ( see in vivo experiments ) . ß-Galactosidase activities were measured as described previously [49] . The specific activities are expressed in units of ß-galactosidase per milligram of protein ( Miller units ) . Prewarmed 500 ml baffled erlenmeyer flasks with 50 ml LB medium were inoculated with 1 ml overnight cultures of Bacillus 407 nprA::lacZ or Bacillus 407 ΔRX nprA::lacZ , and incubated at 37°C and 250 rpm . Samples for microarray analysis were taken three ( t3 ) hours after the onset of the stationary phase . Samples were harvested as described previously [12] mixed with RLT buffer from the RNeasy midi kit ( Qiagen , France ) and frozen at −70°C . After thawing samples at 37°C for 15 min , RNA isolation , cDNA synthesis , labeling and purification were performed as described [12] . The microarray slides were printed , prehybridized and hybridized as described previously [12] , except that hybridization was extended to 17 hours . The slides were scanned on an Axon 4000B scanner ( Molecular Devices ) . Gridding , spot annotation and calculation of raw spot intensities was done with the GenePix Pro 6 . 1 software ( Molecular Devices ) . The LIMMA package [50] , [51] , [52] on the R 2 . 7 . 1 platform [53] was used for filtering , normalization and further analysis . The raw data were filtered and weighted by quality [54] , and the four technical replicates on each slide were averaged to increase robustness . P-values were computed using a false discovery rate of 0 . 05 . The analysis was based on hybridization to three slides , all employing biological replicates . Gene expression was investigated in Bacillus 407 nprA::lacZ and Bacillus 407 ΔRX nprA::lacZ . Reverse transcription was performed according to the SuperScript III reverse transcriptase protocol from Invitrogen , but RNaseOUT was replaced with 0 . 1 µl SUPERase-In ( Ambion ) . A negative control without reverse transcriptase was included . In all samples , the reaction volume was adjusted to 20 µl with DEPC-treated water ( Ambion ) before reverse transcription . The reaction product was diluted ( 1 µl in 39 µl ) with water , and 8 µl applied to each well ( 2 µl for 5 s rRNA samples ) . Primers were added to a final concentration of 0 . 56 µM . A volume of 9 µl LightCycler 480 DNA SYBR Green I Master ( Roche ) was added , and the volume was adjusted to 18 µl . Primers ( available on request ) were designed to give PCR products of around 100 bp . The reference genes , gatB ( BC4306 ) and 5 s rRNA , were included on every plate . The samples were analyzed on a Roche Lightcycler 480 ( Roche Diagnostics GmbH , Mannheim , Germany ) . Cycling conditions were 95°C for 5 minutes followed by 45 cycles at 95°C for 10 seconds , 58°C for 10 seconds , and 72°C for 8 seconds . Cp values were determined using 2nd derivative max , and are averages of two technical replicates . The results were calculated by the delta-delta Ct approximation . The log2 expression ratios of Bacillus 407 ΔRX nprA::lacZ over Bacillus 407 nprA::lacZ in Table S3 are averages for three biological replicates . Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry ( MALDI-ToF MS ) was used to screen kurstakin production from whole bacterial cells on solid media . Cultures were performed on AK agar plates incubated at 30°C for 24 h or 48 h . A saturated solution of α-cyano-4-hydroxy-cinnamic acid was prepared in 1∶2 ( v/v ) solution of CH3CN and H2O containing 0 . 1% TFA . Measurement was performed using UV laser MALDI-ToF spectrometer ( Bruker UltraFlex TOF; Bruker Daltonics ) equipped with a pulsed nitrogen laser ( λ = 337 nm ) . The ions were extracted from the ionization source with an acceleration voltage of 20 kV . Samples were measured in the reflector mode , positive mode . The equivalent of about 1 µl of cell material was picked from agar plates with an automatic pipette . The tip with the culture was deposited in an eppendorf tube . 20 µl of matrix solution ( saturated solution of α-cyano-4-hydroxycinnamic acid in a 1∶2 v/v solution of CH3CN/H2O with 0 . 1% TFA ) are added . The eppendorf with the tip and the matrix solution was vortexed for 30 seconds . 1 µl of this sample solution was deposited on the MALDI target and let dry at room temperature . The spectrum was obtained with 5×30 shots on the sample . Analyses were performed on two different samples . | Bacillus thuringiensis ( Bt ) is a well known entomopathogenic bacterium successfully used as a biopesticide for fifty years . The insecticidal properties of Bt are mainly due to specific toxins forming a crystal inclusion associated with the spore . After ingestion by susceptible insect larvae , toxins could induce favorable conditions for spore germination . The bacteria multiply in the insect and coordinate their behavior using signaling molecules involved in quorum sensing . The activation of the quorum sensor PlcR leads to the production of virulence factors allowing the bacteria to kill the insect host . Here we show that , in the cadaver , Bt shifts from a virulent to a necrotrophic lifestyle during which a second quorum sensor ( NprR ) becomes functional . NprR activates genes encoding degradative enzymes ( proteases , lipases and chitinases ) and a lipopeptide ( kurstakin ) involved in swarming and biofilm formation . The kurstakin is also essential for the survival of Bt after insect death . This suggests that NprR allows the bacteria to survive and eventually to sporulate in the host cadaver , thus improving their ability to disseminate in the environment . Altogether these results show that the pathogenic and necrotrophic lifestyles of Bt are tightly controlled by two quorum-sensing systems acting sequentially during the infection process . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biotechnology",
"bacteriology",
"small",
"molecules",
"enzymes",
"gene",
"regulation",
"microbiology",
"host-pathogen",
"interaction",
"molecular",
"genetics",
"microbial",
"growth",
"and",
"development",
"bacterial",
"pathogens",
"applied",
"microbiology",
"microbial",
"physiology",
"gene",
"expression",
"microbial",
"pathogens",
"biology",
"pathogenesis",
"microbial",
"ecology",
"gram",
"positive",
"biochemistry",
"bacterial",
"physiology",
"bacterial",
"biofilms",
"gene",
"identification",
"and",
"analysis",
"genetics",
"genetics",
"and",
"genomics"
] | 2012 | Necrotrophism Is a Quorum-Sensing-Regulated Lifestyle in Bacillus thuringiensis |
Previous studies suggest that protective immunity against Schistosoma haematobium is primarily stimulated by antigens from dying worms . Praziquantel treatment kills adult worms , boosting antigen exposure and protective antibody levels . Current schistosomiasis control efforts use repeated mass drug administration ( MDA ) of praziquantel to reduce morbidity , and may also reduce transmission . The long-term impact of MDA upon protective immunity , and subsequent effects on infection dynamics , are not known . A stochastic individual-based model describing levels of S . haematobium worm burden , egg output and protective parasite-specific antibody , which has previously been fitted to cross-sectional and short-term post-treatment egg count and antibody patterns , was used to predict dynamics of measured egg output and antibody during and after a 5-year MDA campaign . Different treatment schedules based on current World Health Organisation recommendations as well as different assumptions about reductions in transmission were investigated . We found that antibody levels were initially boosted by MDA , but declined below pre-intervention levels during or after MDA if protective immunity was short-lived . Following cessation of MDA , our models predicted that measured egg counts could sometimes overshoot pre-intervention levels , even if MDA had had no effect on transmission . With no reduction in transmission , this overshoot occurred if protective immunity was short-lived . This implies that disease burden may temporarily increase following discontinuation of treatment , even in the absence of any reduction in the overall transmission rate . If MDA was additionally assumed to reduce transmission , a larger overshoot was seen across a wide range of parameter combinations , including those with longer-lived protective immunity . MDA may reduce population levels of immunity to urogenital schistosomiasis in the long-term ( 3–10 years ) , particularly if transmission is reduced . If MDA is stopped while S . haematobium is still being transmitted , large rebounds ( up to a doubling ) in egg counts could occur .
Urogenital schistosomiasis ( caused by the blood fluke Schistosoma haematobium ) remains a prevalent tropical disease , infecting over 100 million people in sub-Saharan Africa [1] , [2] . Recent control efforts have focussed upon mass drug administration ( MDA ) using the antihelminthic drug praziquantel [3] , [4] , with the principal aim of reducing morbidity , although MDA can significantly reduce both population infection levels [5] , [6] and transmission rates [5] , [7] . To maintain low infection levels treatments must be repeatedly administered for an indefinite time period [8] , [9] . MDA reduces infection levels directly through killing worms , and indirectly through reducing transmission . Acquired immunity enhances treatment efficacy , and influences subsequent infection dynamics [10] . Previous modelling work , which assumed protective immunity was stimulated by live worms , suggested that repeated population-level treatment would disrupt the development of acquired immunity by removing the antigenic stimulus [10] , [11]; if treatment ceased , then under some circumstances , infection levels could ‘overshoot’ to exceed pre-treatment levels [10] . Protective immunity to schistosomes appears to develop slowly , with children in endemic areas experiencing repeated re-infection while adults experience much lower levels of infection , even with high exposure [12] , [13] . Infection intensity peaks at an earlier age in areas with more intense transmission [14] , and this is mirrored by immune responses associated with protection [15] , suggesting that protective immunity is related to cumulative exposure to infection rather than age-related physiological factors . Earlier studies have shown that age-related changes in reinfection rates are explained by protective antibody levels [16] , and that the development of resistance is dependent upon exposure history [17] . Several studies have demonstrated that praziquantel treatment boosts schistosome-specific antibody responses to S . mansoni and S . haematobium [18]–[20] , and accelerates isotypic changes which occur more gradually with age [21] , [22] . Praziquantel kills adult worms , enhancing serological recognition of S . haematobium antigens [23] . Increased exposure to antigens released from dying worms is thought to be responsible for stimulating these immunological changes following praziquantel treatment . Several of the responses boosted by praziquantel treatment , including IgE , IgG1 , and cytokines IL-4 and IL-5 , have been associated with protection against re-infection in other studies [16] , [20] , [24] , [25] , and some studies have shown that responses boosted by treatment are associated with protection against re-infection in the same population [26] , [27] , suggesting that treatment enhances protective immunity . Recent mathematical modelling for S . haematobium showed that post-treatment boosts in antibody responses associated with protection are most consistent with protective antibody being stimulated by dying worms and reducing worm fecundity [28] . This study suggested that if protective antibody were mainly stimulated by antigens from other life stages ( including cercariae , live worms , or eggs ) then a boost in antibody would not be seen following treatment [28] . No models have previously looked at long-term effects of MDA upon the dynamics of protective immunity and measured egg output when such immunity is stimulated by dying ( rather than live ) worms . While treatment is expected to increase antigenic exposure and boost protective immune responses in the short term through worm killing , a period of reduced exposure to dying worms will follow the initial reduction in worm burden since treatment causes worms to die sooner than they naturally would . Exposure will be further reduced if population-wide treatment reduces transmission rates , decreasing re-infection . The long-term implications of mass treatment for the development of protective immunity are not fully clear [29] . Here , using a model with protective immunity stimulated by antigens released from dying S . haematobium worms , we assess the expected impact of MDA upon the development of acquired immunity , and upon measured egg output , both during and after a mass treatment campaign .
We used a stochastic individual-based model which describes changes in worm burden , egg output and a protective antibody response with age for people living in an area with endemic schistosome infection . This model has been fully described previously [28] . Briefly , the model tracks the number of worms an individual harbours between their birth and 34 years of age . Individuals acquire new worms through contact with water containing infective larvae . As suggested by field studies , rates of water contact change with age [30] and vary between individuals [31] . The number of cercariae acquired per water contact is independent of population infection levels and remains constant over time ( unless reduced transmission is assumed during MDA ) . Note that transmission of parasites between humans and snail intermediate hosts is not explicitly modelled . Acquired cercariae develop into adult worms ( with approximately Gaussian-distributed survival , following earlier modelling studies [32] ) , which produce eggs . The number of eggs within the host is assumed to be proportional to current worm burden but reduced by protective antibody . It is also assumed that egg output per worm is constant regardless of worm age . Measured egg output is calculated as the average of three ‘samples’ drawn from a negative binomial distribution around the number of within-host eggs . The protective antibody response is relatively long-lived ( decay rate of 0 . 008–0 . 8 year−1 , equivalent to a half-life of 10 months–87 years ) , as suggested by earlier model fitting [28]; no direct estimates are available for the longevity of protective immunity against schistosome parasites in humans , but these estimates fall between the decay rates estimated for antibody responses to other pathogens [33] , and for memory B cells [34] . Protective antibody is stimulated by antigens from dying worms and reduces worm fecundity , as suggested by previous comparison of model output with field data [28] , and as demonstrated for the leading schistosome vaccine candidate , a 28 kDa glutathione S-transferase [35] . Most of the models used here include an ‘antigen threshold’ , a level of cumulative antigen exposure which must be exceeded before a protective antibody response is mounted , as suggested by previous model fitting , but we include models without an antigen threshold which have also been found to fit the data [28] . In earlier work , this model was parameterised using data from studies in Zimbabwe and elsewhere , and fitted to population data on pre- and post-treatment S . haematobium egg counts and specific antibody responses from several rural sites in Zimbabwe with endemic infection [28] . A grid-search of parameter space was performed to identify parameter combinations which were consistent with field data , varying the following parameters simultaneously: mean population infection rate , worm life span , antibody strength , immune response decay rate , and antigen threshold level . This grid-search was repeated , varying each of the following parameters separately: aggregation of contacts , rate of changing contact rate , aggregation of acquired cercariae , number of eggs per worm , and aggregation of egg output . Parameter combinations from all of these grid-searches which were able to reproduce cross-sectional egg output and antibody patterns and short-term post-treatment egg output and antibody dynamics were used in the current analysis to estimate the long-term impact of treatment . A population of 175 individuals was simulated , with 5 individuals in each yearly age group from 1 to 34 years old at the time of the baseline survey . Individuals were simulated up to their respective ages before the initial baseline egg count and antibody levels were recorded and the first round of treatment applied . Individuals were then simulated for a further 15 years after this initial survey , during and after MDA ( see next section for treatment schedules ) . When individuals reached the age of 34 , they were replaced by 1 day old infants with no worms or antibody , to maintain a constant population size . Six treatment schedules were used which vary treatment frequency , target population , coverage and reduction in transmission ( table 1 ) . For the standard treatment schedule ( schedule 1 ) , treatment was given to school-aged children , defined as those aged 6–15 years old , as recommended by the WHO and implemented by the Schistosomiasis Control Initiative ( SCI ) [3] , [4] . Treatment was applied annually for five years ( five treatments in total ) . Annual treatment is advised for , and used in , high-prevalence communities [3] , [4] . In the standard treatment schedule , coverage was assumed to be 75% , in line with WHO targets and achieved coverage in several countries [3] , [36] , [37] , and it was assumed that treatment did not affect transmission . In each of the other treatment schedules , one parameter was changed from the standard schedule ( table 1 ) . In schedule 2 , biennial treatment ( i . e . treatment every two years ) was given over a five year period ( three treatments in total ) . Biennial treatment is advised for , and used in , areas with moderate prevalence [3] , [4] . Schedule 3 had 90% treatment coverage ( as achieved in some countries [38] , [39] ) . In schedule 4 , the whole population over the age of 5 was treated , as recommended and implemented for high risk populations [3] , [4] . In schedules 5 and 6 , it was assumed that treatment reduced transmission by 100% or 50% respectively . For simplicity , treatment was assumed to reduce transmission as a step change to a fixed level , from the day after the first treatment up until one year after the final treatment , when transmission returned to its original level . For all treatment schedules ( 1–6 ) , treatment was applied over a five year period and each treatment was applied the day after egg counts and antibody levels were recorded . Treatment was applied randomly across the eligible population at the required coverage level ( 75% or 90% ) at each round of treatment , meaning that an individual's chance of being treated in each round was independent of whether they had been treated in previous rounds . Treatment was assumed to be given independently of worm burden or egg output , in line with the usual MDA strategy of giving treatment to all school-aged children [3] . For all schedules , a treatment efficacy of 90% was assumed ( 90% of worms were killed ) , which gave reductions in egg output of 87–98% , in line with field studies [40] . For each parameter set , 200 repeat simulations of the whole population were run , and mean levels of egg output and antibody for the whole population aged 6–34 years old were calculated pre-treatment and at yearly intervals during and after the simulated treatment regime , averaged over the 200 repeat simulations . This age range was used in order to capture the changes in egg counts and antibody levels in treated individuals as they aged over the long follow-up period . Egg output and antibody dynamics were studied to see how quickly they returned to pre-treatment levels . The conditions ( parameter values or treatment schedules ) under which protective antibody levels fell below pre-treatment levels or egg counts overshot pre-treatment levels were identified .
The impacts of separately varying the frequency of treatment and the coverage and age-range of the target population ( treatment schedules 2–4 ) are shown for single parameter combinations ( figure 4 ) , but demonstrate trends seen for all parameter sets . Results are shown for one parameter set that did and one that did not give an overshoot in egg output for treatment schedule 1 . With biennial treatment ( schedule 2 ) , protective antibody declined and egg output increased following non-treatment years , but their levels approached those seen with annual treatment ( schedule 1 ) one year after each treatment ( figure 4 ) . The overshoot in egg output was less pronounced for biennial versus yearly treatment ( figure 4d ) . Changing the level of coverage of the school-aged population ( 90% coverage ( schedule 3 ) vs . 75% ( schedule 1 ) ) made little difference; it gave a slightly greater increase in antibody and greater reduction in egg output during treatment , and a slightly more pronounced overshoot in egg output ( figure 4 ) . Treating both adults and children ( aged 6–34 years old , schedule 4 ) rather than just school-aged children ( 6–15 years old , schedule 1 ) gave a much larger antibody boost and greater reduction in egg output during the treatment programme and greater overshoot in egg output ( figure 4 ) . The effects of assuming that transmission is reduced during MDA are shown for one parameter set ( figure 5 ) , but these trends were seen for all of the parameter sets examined . If 100% reduction in ( i . e . no ) transmission was assumed during MDA ( schedule 5 ) , protective antibody always fell below pre-treatment levels at some point , before treatment ceased if there was rapid immune decay ( 0 . 8 year−1; figure 5a ) . Egg output was always reduced to below 5% of pre-treatment levels after five treatment rounds and overshoots in egg output were always predicted ( figure 5b ) . Higher infection rates gave rise to higher and earlier overshoots in egg output ( data not shown ) . With 50% transmission reduction ( schedule 6 ) , antibody levels always dropped below pre-treatment levels , but more slowly and to a lesser extent than when 100% reduction in transmission was assumed ( figure 5a ) , and for most parameter sets , egg output was still predicted to overshoot pre-treatment levels , to a lesser extent than with 100% reduction in transmission ( figure 5b ) .
Several MDA programs for schistosomiasis are currently underway in Africa [41] . While their immediate impact on infection and morbidity in affected individuals is unequivocal , their long-term effects on infection dynamics are not yet fully understood . Our models predict that population levels of S . haematobium infection will be substantially reduced by repeated MDA , while levels of protective antibody will be initially boosted by treatment , in agreement with patterns seen in the field . We predict that , in the long-term , levels of antibody could fall below pre-treatment levels after or even during MDA . More rapid declines in protective antibody levels are predicted with more rapid immune decay , shorter worm life span or reduced transmission . After the initial increased exposure to dying worms that treatment brings about , the reduced worm burden leads to a subsequent reduction in exposure to dying worms , leaving antibody levels strongly influenced by immune decay rates . Reduced transmission further reduces antigen exposure . Baseline antigenic exposure rates are expected to be lower in models with a longer worm life span , and so the reduced antigenic exposure following treatment will have a more rapid effect in models with short worm life span , leading to more rapid declines in antibody . We found that measured egg output could rebound to levels exceeding pre-treatment levels after cessation of MDA . Our finding that this was very likely to happen if treatment temporarily reduced transmission confirms findings from earlier work using different models [10] . Importantly , we found that it could also occur in the absence of any reduction in transmission , and was more likely to occur if the immune response decayed rapidly . The fact that , without reduced transmission , a rebound in infection only happened for a restricted set of parameters , highlights how important it is to estimate these parameters to improve the accuracy of model predictions . The rate of decay of protective immunity is particularly important . Some studies ( mainly on S . mansoni ) suggest that schistosome-specific antibody levels may decline below pre-treatment levels following an initial boost , behaviour predicted for medium- or short-lived antibody responses in our model [42] , [43] , but this is not always seen and more accurate estimates are required . Our results suggests that MDA might disrupt the build-up of protective immunity ( or may disrupt existing immunity ) against schistosomes , despite short-term boosting of this protective response . Interestingly , a reduction in antibody levels below pre-treatment levels during MDA did not necessarily correspond with overshooting of egg output after treatment ceased . It should be noted that , even when overshoots in egg output occur after treatment stops , the overall impact of the intervention on egg output ( taking the overshoots into account ) may still be positive; the reductions in egg output during the control programme may be sufficient to offset the increases seen after treatment stops . We found that increasing the coverage of treatment of school children from 75% to 90% only increased population antibody levels and decreased measured infection levels by a very small amount . The random allocation of treatment at each round meant that even at 75% coverage , the chances of an individual never being treated over the five rounds of treatment were very small , which may account for the comparatively small coverage effect . Two previous modelling studies looking at S . haematobium in Ghana [8] and S . japonicum in China [44] also found little difference in long-term infection dynamics between biennial and annual treatment . However , other modelling studies have suggested different impacts of biennial versus annual treatment [9] , [45] . Treating the whole population rather than just school-aged children gave a more pronounced boost to population-level protective antibody and a greater reduction in egg output during MDA , but meant that any overshooting of egg output after treatment ceased became more pronounced , suggesting that infection rebounds could be more serious following more intensive control efforts . Previous modelling studies which considered the effects of acquired immunity on the impact of MDA suggest that the strength and duration of protective immune responses play an important role in determining infection dynamics [10] , which was also found here . Our results suggest that , without any reduction in transmission post-treatment , an overshoot in measured infection levels after treatment stops is most likely to occur with relatively rapid immune decay rates ( half-life of 10 months ) ; in contrast , Chan et al . ( 1996 ) [10] reported overshoots with slow immune decay rates ( half-life of 7 years ) , and not with more rapid decay . This discrepancy may arise because they compensated for slow immune decay rates with higher infection rates [10] . In the current analysis , when treatment was assumed to reduce transmission , higher infection rates gave rise to more pronounced overshoots in egg output . Our results support the long-term maintenance and monitoring of existing MDA programmes , to ensure that treatment continues while transmission is still ongoing . In addition to MDA , other measures to reduce transmission should also be strengthened , including the provision of safe water and sanitation facilities , and good health education [46]–[48] . In conclusion , our models predict that , with protective immune responses stimulated by dying S . haematobium worms , repeated MDA will boost protective immunity initially , but antibody levels could decline below pre-treatment levels during or after MDA . In some circumstances , we also predict that post-MDA egg output could exceed pre-intervention levels . Field data are not currently available to test these predictions , but they have been made using a calibrated model which reproduces robust patterns seen in short-term pre- and post-treatment studies of S . haematobium infection [28] . While MDA programmes have had substantial impact upon schistosomiasis infection levels , this analysis highlights the potential negative consequences of ceasing a mass treatment programme . | Urogenital schistosomiasis , caused by schistosome blood flukes , infects more than 100 million people in sub-Saharan Africa . Current control efforts involve regularly treating all school-aged children with the drug praziquantel , which kills schistosome worms . Earlier work by our group suggests that protective immunity against schistosomes is mainly stimulated by dying worms , and that in the short term , praziquantel treatment boosts immunity through killing worms . The longer-term impact upon the development of protective immunity is unknown . In this paper , we used a mathematical model which was able to replicate short-term patterns of infection and antibody to predict the long-term changes in antibody and infection levels that would occur during and after a 5-year treatment programme . We found that the longevity of protective immunity was particularly influential . Short-lived protective immunity was associated with levels of protective antibody declining below pre-treatment levels in the long term , and also with an increase in measured infection levels ( eggs in urine ) to peak above pre-treatment levels after the treatment programme finished . Antibody declines and infection peaks post-treatment were also predicted if treatment programmes reduced schistosome transmission . These results highlight the possible negative consequences of ceasing mass treatment programmes once they have commenced . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"humoral",
"immunity",
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"applied",
"mathematics",
"immunity",
"to",
"infections",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"mathematics",
"population",
"modeling",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"epidemiology",
"infectious",
"disease",
"modeling",
"helminth",
"infections",
"schistosomiasis",
"immunity",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology"
] | 2014 | Predicted Impact of Mass Drug Administration on the Development of Protective Immunity against Schistosoma haematobium |
The conformational dynamics of proteins is rarely used in methodologies used to predict the impact of genetic mutations due to the paucity of three-dimensional protein structures as compared to the vast number of available sequences . Until now a three-dimensional ( 3D ) structure has been required to predict the conformational dynamics of a protein . We introduce an approach that estimates the conformational dynamics of a protein , without relying on structural information . This de novo approach utilizes coevolving residues identified from a multiple sequence alignment ( MSA ) using Potts models . These coevolving residues are used as contacts in a Gaussian network model ( GNM ) to obtain protein dynamics . B-factors calculated using sequence-based GNM ( Seq-GNM ) are in agreement with crystallographic B-factors as well as theoretical B-factors from the original GNM that utilizes the 3D structure . Moreover , we demonstrate the ability of the calculated B-factors from the Seq-GNM approach to discriminate genomic variants according to their phenotypes for a wide range of proteins . These results suggest that protein dynamics can be approximated based on sequence information alone , making it possible to assess the phenotypes of nSNVs in cases where a 3D structure is unknown . We hope this work will promote the use of dynamics information in genetic disease prediction at scale by circumventing the need for 3D structures .
A 3D structure is still required to computationally obtain protein dynamics , drastically limiting the extent to which conformational dynamics can be incorporated into genomic analysis . The reason for this is that there are exponentially more sequences than experimental structures . Currently , UniProtKB contains more than 100 million sequence entries , whereas the PDB reports the number of known 3D structures to be around 140 , 000 [1] . Furthermore , the number of known sequences is increasing at an exponential rate , compared to the much slower addition of new experimental PDB structures . This is due to the advent of high-throughput genomic sequencing , which is providing an unprecedented amount of data for genomic analysis . The vast amount of sequence data has driven the rapid classification of novel genetic variations through genome-wide association studies [2 , 3] . A large catalogue of non-synonymous single nucleotide variants ( nSNVs ) occurs in coding regions that can severely impact protein function , potentially leading to disease [4] . There are many in silico methods developed using evolutionary methodologies such as positional conservation and phylogeny and those that combine evolutionary approaches with biochemical and structural properties to diagnose neutral and disease associated nSNVs [5–11] . However , the accuracy of the majority of these in silico prediction methods is significantly lower for predicting the impact of nSNVs at highly evolving sites [12–16] . Protein dynamics can also be used to elucidate the functional impact of nSNVs and mechanisms of disease [5 , 17] . Our previous studies have evinced that a site-specific conformational dynamics analysis is capable of diagnosing nSNVs irrespective of evolutionary conservation [5 , 18 , 19] and recently has been incorporated as an additional feature for in silico prediction tools [20] . However , only a small fraction of the catalogued nSNVs in the coding regions ( i . e . missense variants ) have 3D experimental structures , [20] , impeding broad application of protein dynamics in in silico tool predictions . Coevolution , on the other hand , has become a valuable tool for its ability to predict structural contacts of 3D structures , particularly using global information through Potts models [21–27] . Coevolving residues are inferred from a multiple sequence alignment ( MSA ) of a given protein family , whereby if two given amino acids exhibit concordant patterns of evolution throughout the MSA then they are assumed to be in close spatial proximity in the folded 3D structure . This evolutionary principle can be leveraged so that sequence information can be used to describe protein topology , making de novo structure prediction possible [24 , 27] . It has been reported that only one correct contact for every 12 residues in a protein is necessary for accurate topology-level modeling [28] . In addition to structure prediction , coevolution analysis has also been used to identify critical interactions between protein complexes [22] important functional sites [24] and allosteric response [29] . The use of coevolution for structure prediction is largely possible for two reasons . First , the amount of sequence data for different protein families is sufficient to be leveraged by this technique to make predictions . Second , the methods for inferring coevolving residues from an MSA are becoming increasingly robust [30–34] . Inferring evolutionary couplings from an MSA are based on two primary approaches categorized as local [35–37] and global approaches [37–39] . The global approaches detangle direct evolutionary couplings from indirect couplings which enables them to capture spatial contacts [40] . Regardless of the method , the accuracy of detecting coevolving residues that correspond to structural contacts is fundamentally limited by the number of sequence homologs in the MSA . While most of the current methods use only the sequence homologs of the protein family belonging to target sequence , integrating multiple orthology protein families ( i . e . families that share similar phylogeny and retain similar functions ) was used to increase the number of homologs to produce a more accurate statistical inference [41] . RaptorX , leverages this joint family methodology; it uses an ultra-deep neural network combining coevolution information with sequence conservation information to infer 3D contacts and has produced higher accuracy than other methods [42–44] . In this paper , we will demonstrate the efficacy of our novel sequence-based GNM approach , called Seq-GNM , to estimate the dynamics profile of a protein with no a priori knowledge of its 3D structure . This de novo approach based on a Gaussian network model ( GNM ) enables the prediction of the magnitude of mean-square fluctuations of residues , which are proportional to the B-factors determined by X-ray crystallography experiments . However , instead of using a cutoff distance to determine 3D contacts as does the original structure-based GNM , we use coevolving residues ( evolutionary couplings ) in our model . We show that the theoretical predictions from our Seq-GNM are in reasonable agreement with experimental crystallographic B-factors as well as the values obtained from the structure GNM models that use spatial contacts . We also extend this analysis to determine the capacity of our model to assess the functional impact of nSNVs . We will demonstrate that the dynamics predicted by Seq-GNM can adequately classify disease and benign nSNVs across the proteome .
We considered a high-resolution protein ( 2 . 25 Å ) that is involved in amino acid catabolism , acyl-CoA dehydrogenase ( 1JQI ) , as an example case to examine the B-factor profiles and predicted contact maps using Seq-GNM . Coevolution analysis using direct coupling analysis ( DCA ) has been shown to recapitulate accurate structural contact maps for a wide range of proteins [21 , 23 , 24 , 27 , 31 , 45] . As expected , the contact maps of Seq-GNM and structural GNM are similar ( Fig 1 ) . In a comparison of their B-factor profiles , both Seq-GNM and structural GNM exhibit good agreement with observed B-factors , capturing flexible and rigid positions . Using evolutionary coupling ( EC ) values obtained from RaptorX , the correlation between the Seq-GNM and observed B-factors is 0 . 77 , whereas the correlation between the structural GNM and observed B-factors is 0 . 57 ( Fig 1A ) . Similarly , using EC values obtained by EVcouplings produced a correlation of 0 . 60 between the Seq-GNM and observed B-factors ( Fig 1B ) . The scores obtained from EVcouplings are still reasonable , yet relatively lower correlations compared to those obtained by the RaptorX . This is likely due to the relatively noisy contact map predictions by EVcouplings compared to the more reliable contact maps produced by RaptorX ( we think this is due to their inclusion of multiple orthology protein families ) [42] . The Seq-GNM produces a correlation with crystallographic B-factors of 0 . 60 , which is within the same range as those produced by the GNM from structure of 0 . 57 . Moreover , theoretical B-factor profiles obtained from both methods were able to identify the catalytic sites on all of the proteins . As a further test of the efficacy of the Seq-GNM , we superimposed the predicted B-factors onto the structures of three diverse proteins– 5' ( 3' ) -deoxyribonucleotidase ( 2JAO ) , acyl protein thioesterase ( 1FJ2 ) , and NADH-cytochrome b ( 5 ) reductase ( 1UMK ) –to visually contrast the predicted B-factors with that of experiment . Fig 2 shows each protein color-coded according to their B-factor profile on a spectrum of blue–white–red , where blue represents the lowest B-factors ( less mobility ) and red represents the highest B-factors ( more mobility ) . The left panel shows the experimental B-factors for each protein , while the right panel shows the theoretical values predicted by the Seq-GNM . We investigated whether secondary structure was a factor in how the B-factors were distributed across the protein , and if certain secondary structure domains would exhibit less agreement with experiment . In this context , the proteins were selected so that they had a variety of secondary structure components–2JAO contains primarily alpha helices , 1UMK is mainly composed of beta-sheets , and 1F2J is a combination of alpha helices and beta-sheets . For 2JAO , the exterior helices that are flexible ( red ) in the observed structure are all reproduced in the predicted structure . The one highly rigid ( blue ) helix in the observed structure was more flexible in the predicted structure but was still in overall agreement . There is a surprising amount of similarity between the observed and predicted structure of 1F2J , considering that it contains both alpha-helix and beta-sheet elements . Similarly , 1UMK showed good agreement , except for some miniscule differences . This gives further evidence that the magnitudes of residue fluctuations predicted by the Seq-GNM model is representative of the crystallographic B-factor profiles for many proteins . In order to compare predicted B-factors with crystallographic B-factors , we extracted a subset of 39 structures that had a resolution better than 2 . 0Å to obtain more realistic crystallographic B-factors ( unreliable B-factors are common for many PDB structures ) [18 , 46] . The same cutoff of 2 . 0Å was used in an earlier study to compare GNM predicted B-factors with those determined by crystallography [47] . For all 39 structures , the Seq-GNM ( using EC values from RaptorX ) and structure GNM were used to estimate their B-factors , which were then compared with the observed B-factors by calculating the correlation for each protein . The mean correlation coefficient for the Seq-GNM was 0 . 53 while the mean correlation coefficient for the structure GNM was 0 . 58 . The correlation of 0 . 58 for structural GNM of our smaller data set is consistent with the findings of Kundu et al . where 113 high-resolution structures ( resolution <2 . 0 Å ) were used and , the mean correlation coefficient with observed B-factors was 0 . 59 [47] . As shown in Fig 3A , boxplot distributions reveal that correlations are not significantly different between the sequence and structure GNM ( p = 0 . 055 in a student t-test ) . The structure GNM appears to perform only slightly better than the Seq-GNM . Fig 3B shows the same distribution separated into 10 individual bins of size 0 . 1 . The overall shapes of the two distributions are similar , except for the exaggerated relative lower second peak of the Seq-GNM at 0 . 4 . It should also be noted that for these cases where Seq-GNM had low correlations , the EC threshold could be tuned to yield much higher correlations . If this were done on a case-by-case basis , the overall correlation distributions would be even more similar . Thus , the EC threshold may be used as a tuning parameter to enhance the correlation coefficient for purposes of model optimization . Interestingly , for the cases where predicted B-factors by Seq-GNM yielded significantly better correlations with the experimental B-factors than those obtained by GNM from structures , we observed that biological units of these proteins are assigned as oligomeric forms . While predicted B-factors obtained using Seq-GNM does not retain this information , it successfully predicts the experimentally low B-factor values of interface positions as shown for protein 5' ( 3' ) -deoxyribonucleotidase ( 2JAO ) and protein aldehyde Dehydrogenase 7A1 ( 2J6L ) in Fig 4 . It is indeed shown in earlier work of direct contact analysis that co-evolution can identify positions of protein interfaces and protein-protein interaction partners and successfully reconstruct protein complexes and interaction network [23 , 30 , 48] . Thus , it is not surprising to see that it yields good correlations with the experimental B-factors . Conversely , predicted B-factors from structure can only improve when the oligomeric structure is used for the GNM analysis . Even when using high-resolution X-ray structures , there is still some uncertainty about the realistic nature of crystallographic B-factors . For this reason , we thought a more plausible way to determine the efficacy of the Seq-GNM was to compare it directly with the structure GNM . The structure GNM is a robust method to describe thermal fluctuations in a protein , and in many cases , it performs as good or better than the ANM or MD [47 , 49] . We systematically evaluated the performance of the Seq-GNM and structure GNM for the entire set of 139 structures and obtained the correlation coefficients for each protein ( Fig 5 ) . The average correlation of B-factors between the Seq-GNM and structure GNM model is 0 . 63 when using EC contacts from RaptorX and 0 . 43 when using contacts from EVcouplings . As seen in Fig 5A , the distribution of correlation coefficients increases until 0 . 8 , and then subsequently decreases . Interestingly , there are still an appreciable number of sequences yielding high correlations from 0 . 8 to 1 . 0 . A distinguishing feature of the distribution is the pronounced peak in the bin from 0 . 7 to 0 . 8 , indicating that significant fraction of our data set yields high correlations between 0 . 7 and 0 . 8 . This is evidence that the Seq-GNM is efficiently capturing protein dynamics and supports the theory that ECs can be used as a substitute to 3D structure contacts in the GNM and still produce reliable dynamics profiles . The results of Seq-GNM based on contacts predicted by RaptorX usually yields B-factors that are closer to experimental B-factors as it uses structural information in its neural networks leading to better EC values and correlations with structure [44] . Crystallographic B-factors have previously been used to assess the impact of nSNVs on protein function [18 , 50–54] . A study [51] found that mutations on lysozyme that impaired function exhibited lower than average temperature factors , suggesting that rigid sites on the protein are more susceptible to destabilizing nSNVs than flexible sites [55] . Another study revealed a relationship between crystallographic B-factors and the impact of nSNVs on protein function [56] . A commonly used tool to diagnose neutral and disease associated nSNVs , PolyPhen-2 , uses evolutionary information , structural information , and crystallographic B-factors in its prediction model [49] . These studies indicate that crystallographic B-factors can be used to predict the tolerance of a given residue to an nSNV ( i . e . , whether or not the occurrence of an nSNV would impact function ) . We investigated whether B-factors predicted by the Seq-GNM were indicative of biological phenotype for nSNVs in the human population . A total of 738 nSNVs were mapped to the 139 enzymes , where 436 are disease-associated and 302 are neutral . S1 Table shows the number of disease and neutral nSNVs that occur on each protein . The Seq-GNM ( using EC contacts from RaptorX and EVcouplings ) was computed systematically for all 139 enzymes to obtain their dynamics profiles . The theoretical B-factors scores were converted into a percentile rank so that the values could be compared across different proteins . We initially looked at two human enzymes , human lysozyme ( PDB: 1C7P ) and human cytochrome reductase ( PDB: 1UMK ) . They were chosen because they were short proteins that each contain a disease and neutral nSNV . Human lysozyme is a glycoside hydrolase that functions in the immune system by causing damage to cell walls of bacteria . Human cytochrome b5 reductase is involved in many oxidation/reduction reactions including converting methemoglobin to hemoglobin [55] . Each structure is color-coded according to its theoretical B-factor profile on a spectrum of blue–white–red . Sites that exhibit high mobility ( flexible ) are red , and sites that have low mobility ( rigid ) are blue . Regions that are characterized by low mobility are usually important for maintaining stability and function , thus a mutation could act to destabilize the protein and impair its function . Fig 6A show the disease mutation I56T occurring on a rigid site with a B-factor of 0 . 0075 . The neutral mutation T70N has a B-factor of 0 . 96 indicating that it is a highly mobile site . Both I56T and T70N occur on loop regions . Although loops are generally more flexible , three alpha-helical domains encompass the loop containing I56T , which implies that it may be involved in interactions that contribute to stabilizing the functional conformation . Thus , the I56T mutation may disrupt these critical interactions and impair the enzymatic function . In the case of cytochrome reductase ( Fig 6B ) , the disease mutation R57Q is also on a rigid site with a B-factor of 0 . 14 . Instead of being located near the core , R57Q is highly exposed protruding outwardly from a beta-barrel . However , since beta-barrels often harbor functional residues , the R57Q mutation may disrupt certain interactions critical for modulating function . The neutral mutation T116S is located on a loop and has a B-factor of 0 . 96 , indicating that is it has a high mobility . In our earlier proteome wide study of over 100 human protein structures , we have shown that sites that are highly flexible ( e . g . , loop regions , or superficial sites ) are typically more robust to mutations . Conversely , rigid sites are more susceptible to mutations that may disrupt function [18 , 19] . For these two cases , the B-factors produced by Seq-GNM successfully distinguished between the disease and neutral nSNVs , without using the 3D structures . These findings prompted us to analyze the proteome-wide set of 139 enzymes to determine if the B-factors were indicative of phenotype for all 436 disease and 302 neutral nSNVs . The raw B-factor values were converted into a percentile rank ( %B-factor ) and then binned into 5 bins of size 0 . 2 . We computed the observed-to-expected ratio of B-factors , where the expected values were based on the B-factor distribution of all 51 , 618 sites across all 139 proteins , and the observed values were based on the B-factors of the 436 disease sites . The same process was done for the 302 neutral nSNVs . Under the null hypothesis that predicted B-factor of the disease associated nSNVs yields similar distribution of all the positions gathered from 139 enzyme sequences , the ratio of expected and observed sites harboring disease mutations for each %B-factor bin should be close to 1 , which would imply that B-factor does not distinguish sites that are prone to disease . This is the null hypothesis that disease sites are distributed uniformly between sites with low and high mobility . However , the null hypothesis was rejected for the 436 disease nSNVs ( p <0 . 001 ) . Fig 7 shows the observed-to-expected ratio plot of disease and neutral nSNVs , which indicates that disease nSNVs are overabundant at low %B-factor sites ( <0 . 4 ) and under abundant at high %B-factor sites . Conversely , neutral nSNVs are overabundant at high %B-factor sites ( >0 . 6 ) and under abundant at low %B-factor sites . This evidence suggests that the occurrence of an nSNV on a site with a low B-factor is likely damaging based on the position irrelative of the substitution . This is in agreement with our previous proteome-wide study showing that substitutions at rigid sites are more often associated with diseases [18] . Conversely , an nSNV on a high B-factor site is usually benign . Low B-factors usually signify a residue that is crucial for modulating functional motions ( e . g . , a hinge ) . Thus , mutations on these sites can severely impact function . High B-factor sites are more flexible ( e . g . , loops ) and more robust to mutations . Fig 7 suggest that it is possible to use the predicted B-factors to discriminate between disease and neutral nSNVs using co-evolution obtained from only multiple sequence alignment . Moreover , it can be used as an additional feature for in silico predictions [12] . Predictive models were created using logistic regression as the classification algorithm , 80% of the data was used for training and 20% for testing for 10 randomized sets . Models were evaluated based on ROC curves and their respective area under curve ( AUC ) , the best performance is labeled as AUC_max and average performance as AUC . Theoretical B-factors obtained by Seq-GNM , experimental B-factors , and evolutionary parameters were used as predictive variables for training and testing ( Fig 8 ) . Seq-GNM and experimental B-factors have similar performance ( maximum AUC of best 0 . 76 and 0 . 75 , respectively ) , with Seq-GNM overshadowing experimental B-factors on average ( AUC of 0 . 69 and 0 . 60 , respectively ) . The ~0 . 70 AUC of B-factors obtained from Seq-GNM is impressive , as it has been shown that majority of state-of-art methods also yields similar AUC in independent tests [5 , 13] . Moreover , incorporation of Seq-GNM as an additional feature with evolutionary parameters resulted in higher prediction performance . While the AUC scores obtained using the evolutionary features for classification gives 0 . 76 , this is increased to 0 . 81 after including the B-factors of Seq-GNM ( Fig 8C and 8D ) . This result also demonstrates the efficacy of Seq-GNM in disease prediction as a complementary metric to other metrics used as features in classifiers . We also compared the performance of Seq-GNM with common in silico prediction tools like Polymorphism Phenotyping v2 ( PolyPhen-2 ) , and Sorting Intolerant from Tolerant ( SIFT ) [6 , 58] . The accuracy , sensitivity , and selectivity of disease predictions for nSNVs with experimental B-factors , B-factors from SIFT , PolyPhen-2 , evolutionary parameters , and Seq-GNM are tabulated in Table 1 . The accuracy of Seq-GNM using both EC values from EVcouplings and RaptorX is ~0 . 70 . This accuracy is similar to using experimental B-factors for prediction ( 0 . 69 ) and also very close to prediction with evolutionary parameters ( 0 . 75 ) , suggesting that Seq-GNM allows us to incorporate protein dynamics in nSNV predictions when the 3D experimental structures are not available . Moreover , accuracy of Seq-GNM approach is greater than SIFT ( 0 . 65 ) and PolyPhen-2 ( 0 . 64 ) . Interestingly , Seq-GNM obtained by EVcouplings and RaptorX yields similar accuracies indicating that evolutionary couplings without the inclusion of structure could be utilized to predict B-factors to include as a feature to in silico prediction tools . Seq-GNM sensitivity ( ~0 . 90 ) surpasses other methods ( 0 . 80 for SIFT , 0 . 63 for PolyPhen-2 , and 0 . 85 for evolutionary parameters ) , but it has a shortcoming in selectivity ( ~0 . 36 ) as other methods reach higher ( ~0 . 59 ) . Conversely , training Seq-GNM combined with evolutionary parameters enhances the selectivity ( 0 . 66 ) to its highest value compared to others . Seq-GNM with evolutionary parameters predicted disease related nSNVs with accuracy 0 . 78 and sensitivity of 0 . 84 , reaching beyond predictions of other metrics solely . These results suggest the incorporation of Seq-GNM with other prediction metrics can augment accuracy , sensitivity , and selectivity of prediction . Prediction accuracy of Seq-GNM is further tested using 323 nSNVs ( 187 disease-associated , 136 neutral ) of 22 proteins where their 3D experimental structures are not available ( S2 Table ) . We used the trained classifier model of Seq-GNM B-factors for this test . While the B-factors obtained solely from Seq-GNM are used , it reached an accuracy , sensitivity , and selectivity of 0 . 82 , 0 . 82 , 0 . 83 , respectively . This result further suggests that Seq-GNM allows us to incorporate protein dynamics as additional feature in in silico prediction tools without a known 3D structure .
While we and others [5 , 19 , 59–63] have shown that the integration of conformational dynamics into genomic analysis will help next generation of approaches to predict the impact of novel missense mutations on the human proteome , the inherent limitations in the availability of 3D structures compared to the vast number of sequences must be addressed . This begs the question: how can protein dynamics be used in genome-wide analysis to predict functional impacts of nSNVs ? There is , therefore , a need to be able to obtain protein dynamics by leveraging only sequence information , without a priori knowledge of a 3D structure . For this reason , we have developed this novel method to estimate the dynamics profile of a protein by using only a sequence as input . The method uses the coevolution of amino acids through multiple sequence ( which tend to be spatially close in the 3D tertiary structure ) and a simple Gaussian network model ( GNM ) to obtain dynamics . The original GNM based on the 3D structure is well-known for its ability to describe residue dynamics profiles due to thermal motions in proteins ( i . e . , B-factors ) . We showed that our sequence-based GNM model is able to adequately reproduce the mean-square fluctuations ( B-factors ) calculated by the original GNM , particularly outperforms for the cases where biological functional state is oligomeric . Our estimates of B-factors for a proteome-wide set of proteins exhibited good correlation with the structure GNM . Moreover , our estimated B-factors were in reasonable agreement with crystallographic B-factors for many cases . To address the issue of how protein dynamics can determine the impact of nSNVs across the genome where there are no known 3D structures , we tested the ability of our predicted dynamics from the Seq-GNM to assess nSNV phenotypes . A plot of the observed-to-expected ratio of the predicted B-factors revealed distributions of disease and neutral nSNVs that are similar to those in a previous protein dynamics analysis work [18] . The predicted B-factors using the Seq-GNM was able to discriminate between disease and neutral nSNVs with an accuracy of 0 . 70 and incorporating the Seq-GNM predicted B-factors with evolutionary parameters increased overall accuracy to 0 . 78 . This analysis demonstrates that the Seq-GNM makes it possible to obtain estimates of dynamics without using a 3D structure , which will allow for the integration of conformational dynamics into large-scale analysis of genomic variants .
A curated set of 139 structures was selected for several reasons . First , they have high query coverage ( >80% ) and sequence identity ( >80% ) as found from a BLAST search , and the structures had already been modeled using the Modeller software package [64] to account for any missing residues . Second , genetic variants were previously mapped onto these structures , such that the positions containing known nSNVs were already determined , enabling us to easily compare our results using sequence coevolution with the genetic variation data . A total of 738 genetic variants were obtained from the HumVar database [58] , which was comprised of 436 disease and 302 neutral nSNVs . Finally , the structures were either monomers or the single-chain unit of a multimer with <600 residues , allowing for tractable calculations of residue coevolution using the RaptorX web server [42 , 44] , and EVfold ( EVcouplings ) [21] . A table summarizing the dataset is presented in S1 Table . The amino acid sequence from each of the 139 structures was used as input for the evolutionary coupling ( EC ) analysis . The choice of taking the amino acid sequence from the structure was done so that the predicted EC contacts could be compared directly to the experimentally observed structure contacts as verification that the model was producing realistic contact maps . Moreover , the theoretical B-factors predicted by our sequence-based model could be directly compared to the experimental B-factors for each protein . If the structure was unknown , however , sequence databases ( e . g . UniProt , PFAM , etc . ) could be used . The PDB sequences were given to the RaptorX web server [42 , 43] , which computed the relative probability of each residue pair i , j of being in 3D contact based on their coevolution strength . The sequences were also used to generate MSAs using phmmer [65] . Using MSAs , DI values are calculated by EVcouplings . In order to ensure consistency between different proteins of varying lengths , we converted the raw scores into percentile ranks . We then used a threshold value , taking only the top scoring evolutionary couplings ( i . e . , the strongest couplings are more likely to be in spatial contact ) . An optimized threshold value was systematically evaluated and is discussed in the Methods . The Gaussian network model ( GNM ) is an isotropic approach based on the contact topology of a crystal protein structure to obtain the equilibrium fluctuations of residues due to thermal motion . It uses a specified cutoff distance to define interacting pairs that are connected by springs with a single-parameter harmonic potential . In this structure-based GNM , the interacting residue pairs within the cutoff range are represented as contacts in the Kirchhoff ( connectivity matrix ) . In the proposed sequence-based GNM ( Seq-GNM ) approach we will instead use coevolving residue pairs ( evolutionary couplings ) as contacts in the Kirchhoff . In this way , the 3D structure is no longer a prerequisite to form a GNM . To construct the Kirchhoff , a threshold is defined where any evolutionary coupling scores above that threshold are sufficiently coupled such that they are spatially close in 3D structure . If a given evolutionary coupling pair meets the threshold criteria , it is assigned a value in the Kirchhoff for non-bonded contacts of –1 multiplied by its evolutionary coupling score ( i . e . , –1×ECscore ) . This will permit that the strength of each connection will attenuate proportionally to the evolutionary coupling strength . The Kirchhoff can be decomposed into the individual contributions from the bonded contacts representing the chain connectivity ( Rouse chain ) and that from the non-bonded contacts [56] . In the Seq-GNM the contribution of non-bonded contacts to the Kirchhoff is constructed according to Γijnb={−1×ECscore , i≠jevolutionarycoupling0 , i≠jnocoupling−∑i , i≠jΓij , i=j ( 1 ) For the local chain connectivity ( Rouse chain ) , we don’t take into account evolutionary couplings , and matrix was constructed such that every residue pair i , i ± 1 to i , i ± 3 is in contact as Γijcc={−1 , i≠jand∑i|k=1 , 2 , 3Li , i±k0 , i≠jelse–∑i , i≠jΓij , i=j ( 2 ) Then the overall Kirchhoff is the combination of the two contributions Γij=Γijcc+Γijnb . The vibrational dynamics due to thermal fluctuations can then be evaluated in the same way as the original GNM by inverting the Kirchhoff matrix . The magnitude of mean-square fluctuations is then written in terms of the inverse Kirchhoff as ⟨ ( ΔRi ) 2⟩≅[Γ−1]ii ( 3 ) This is proportional to the Debye-Waller temperature factors or B-factors , which describe the attenuation of X-ray scattering due to the thermal motions of atoms ( Bi = 8π2⟨ ( ΔRi ) 2⟩/3 ) . Here there is no single-parameter force constant as in the GNM obtained from structure [52] , and the pair-wise interactions are simply the strength of the evolutionary couplings as given by their ranked scores . The theoretical predictions of our Seq-GNM can be compared to the predictions of the original GNM obtained from structure as well as observed crystallographic B-factors . A general workflow of our method is presented as a flow diagram in Fig 9 . To ensure consistency when analyzing different proteins with varying lengths , we converted the raw scores of evolutionary couplings ( EC ) into a percentile rank . We computed the Seq-GNM for all 139 structures using a constant threshold percentile rank EC value to assign contacts and measured the correlation between the B-factors predicted by our Seq-GNM to the GNM obtained from structure . We used only the top percentile EC scores predicted by RaptorX and EVcouplings as predicted contacts , because only certain fraction of high EC scores are true native contacts in 3D structure , largely due to noisy artifacts in the MSA such as the transitivity of correlations and phylogeny . To determine the optimal threshold value , we tested a range of threshold values from 0 . 92 to 0 . 99 . A threshold value ≤0 . 92 yields superfluous contacts leading to a noisy contact map , and thus , a lower overall correlation ( Fig 10 ) . Conversely , a threshold value ≥0 . 99 gives a deficient number of contacts , which yields an excessively sparse contact map and a lower overall correlation . As Fig 10 shows , a threshold value of 0 . 98 produced the best overall correlation with the GNM from structure and , thus , was taken to be the optimal threshold value used in the analysis . | Proteins are dynamic machines that undergo atomic fluctuations , side chain rotations , and collective domain movements that are required for biological function . There is , therefore , a need for quantitative metrics that capture the dynamic fluctuations per position to understand the critical role of protein dynamics in shaping biological functions . A limiting factor in incorporating structural dynamics information in the classification of non-synonymous single nucleotide variants ( nSNVs ) is the limited number of known 3D structures compared to the vast number of available sequences . We have developed a new sequence-based GNM method , termed Seq-GNM , which uses co-evolving amino acid positions based on the multiple sequence alignment of a given query sequence to estimate the thermal motions of C-alpha atoms . In this paper , we have demonstrated that the predicted thermal motions using Seq-GNM are in reasonable agreement with experimental B-factors as well as B-factors computed using 3D crystal structures . We also provide evidence that B-factors predicted by Seq-GNM are capable of distinguishing between disease-associated and neutral nSNVs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"crystal",
"structure",
"split-decomposition",
"method",
"condensed",
"matter",
"physics",
"multiple",
"alignment",
"calculation",
"protein",
"structure",
"prediction",
"protein",
"structure",
"crystallography",
"structural",
"genomics",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"solid",
"state",
"physics",
"protein",
"structure",
"determination",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"molecular",
"biology",
"physics",
"protein",
"structure",
"comparison",
"biochemistry",
"computational",
"techniques",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"macromolecular",
"structure",
"analysis"
] | 2018 | Coevolving residues inform protein dynamics profiles and disease susceptibility of nSNVs |
Microbes have an astonishing capacity to transform their environments . Yet , the metabolic capacity of a single species is limited and the vast majority of microorganisms form complex communities and join forces to exhibit capabilities far exceeding those achieved by any single species . Such enhanced metabolic capacities represent a promising route to many medical , environmental , and industrial applications and call for the development of a predictive , systems-level understanding of synergistic microbial capacity . Here we present a comprehensive computational framework , integrating high-quality metabolic models of multiple species , temporal dynamics , and flux variability analysis , to study the metabolic capacity and dynamics of simple two-species microbial ecosystems . We specifically focus on detecting emergent biosynthetic capacity – instances in which a community growing on some medium produces and secretes metabolites that are not secreted by any member species when growing in isolation on that same medium . Using this framework to model a large collection of two-species communities on multiple media , we demonstrate that emergent biosynthetic capacity is highly prevalent . We identify commonly observed emergent metabolites and metabolic reprogramming patterns , characterizing typical mechanisms of emergent capacity . We further find that emergent secretion tends to occur in two waves , the first as soon as the two organisms are introduced , and the second when the medium is depleted and nutrients become limited . Finally , aiming to identify global community determinants of emergent capacity , we find a marked association between the level of emergent biosynthetic capacity and the functional/phylogenetic distance between community members . Specifically , we demonstrate a “Goldilocks” principle , where high levels of emergent capacity are observed when the species comprising the community are functionally neither too close , nor too distant . Taken together , our results demonstrate the potential to design and engineer synthetic communities capable of novel metabolic activities and point to promising future directions in environmental and clinical bioengineering .
Microbes have a remarkable capacity to transform their environments , converting key nutrients and energy into accessible forms that are essential for the survival of all other organisms [1] , [2] . This capacity is mediated by a plethora of interactions with the environment and a complex web of metabolic reactions occurring within each microbial cell . In nature , however , most microbes do not typically exist in isolation but rather form complex and diverse communities , joining forces to accomplish tasks that may be energetically unfavorable if performed by a single species [3] , [4] . Such communities play a central role in ecosystem dynamics , agriculture , environmental stewardship , and human health [5]–[9] . The various species comprising each community often form tight relationships and metabolic dependencies [10] , [11] , which can affect the overall stability of the community and its activity . Synergistic relationships endow microbial consortia with enhanced metabolic capacities including nitrification in soil and marine environments [12] , methane oxidation [13] , and pesticide degradation [14] , further affecting the interplay of the community with its environment [15] . Considering these dependencies and synergistic relationships , the metabolic capacity of a given microbial community clearly cannot be described simply as the aggregated capacity of its member species , and a deeper understanding of how the activity of each community member impacts the activity of the others and ultimately the behavior of the community as a whole is required . Of specific interest are cases where the biosynthetic activity of the community is fundamentally different from the biosynthetic activity of the member species . This is often the result of a simple niche construction process [16] , wherein the secretion or uptake of metabolites by one species modifies the composition of the environment and consequently modulates the metabolic activity of another species , causing it to produce and secrete metabolites it would not have produced if growing in isolation [3] , [17] . Characterizing such phenomena , which we term here emergent biosynthetic capacity , is crucial for understanding how microbes jointly construct their environment . More importantly , understanding the determinants of emergent biosynthetic capacity and ultimately designing microbial communities that exhibit specific metabolic capabilities , is a promising research avenue with many industrial and clinical applications , ranging from biofuel production to personalized microbiome-based therapy [3] , [17]–[22] . Here , we therefore set out to characterize emergent biosynthetic capacity on a large-scale and to inform future efforts to design microbial communities that perform desired metabolic tasks . One approach to study the metabolic capacity of microorganisms and to obtain a systems-level predictive understanding of microbial metabolism is through metabolic modeling . Specifically , genome-scale metabolic models have been instrumental in providing insights into the metabolism of various microbial species , their ecology , and their behavior in different settings [23]–[30] . Of the various genome-scale modeling frameworks , constraint-based modeling ( CBM ) methods , such as Flux Balance Analysis ( FBA ) , are perhaps the best-established and most commonly used methods [31] , [32] . Such methods aim to model cellular metabolism as a set of mass balance , thermodynamic , and flux capacity constraints , and to predict the growth rate of the organism as well as the specific distribution of fluxes across the metabolic network by optimizing a cellular objective such as cellular growth or energy production . This modeling framework was shown to correctly capture various factors that govern microbial metabolic processes , providing mechanistic insights into microbial metabolism [28] , [32] . Most importantly , such models proved extremely successful in accurately predicting microbial behavior and activity in multiple environments and under various perturbations [33] , with numerous clinical , environmental , and industrial applications [34] , [35] . With the increased availability of high-quality , manually curated single species models [36] and the recent introduction of automated model reconstruction pipelines [37] , [38] , constraint-based methods provide a unique opportunity to model multi-species ecosystems and to study their metabolic capacities [39]–[42] . However , integrating multiple single species models and developing a framework for modeling diverse microbial communities is not a simple and straightforward undertaking , and to date relatively few CBM-based multi-species models have been presented . One critical problem is how to properly define an objective function at a community level . Stolyar et al . [43] introduced the first two-species FBA model to study a methanogenic syntrophic system , using an objective function that maximizes a fixed combination of biomass from two organisms . A similar approach , maximizing the sum of individual species growth as an overall community objective , has been applied to capture metabolic interactions between ecologically associated species [44] , and to predict measured phenotypes of representative gut microbiome species [45] . Notably , however , the overall community growth objective inherently assumes that member species cooperate and act for the common good of the community , which may potentially lead to biased predictions , wherein , for example , one species barely grows ( although nutrients are available ) to enable the growth of another . One approach to relax this overall community growth objective using a multi-layer optimization algorithm to introduce trade-offs between individual and community level optimization criteria has recently been proposed , and applied to study syntrophic interactions in a few well-characterized , multi-species microbial systems [46] . Alternative optimization methods have also been used to study synthetic cooperation between single gene deletion mutants [47] , environments that induce species cooperation [48] , and diet dependent changes in uptake and secretions between a host and a dominant gut microbe [49] . Yet , the methods above often assume a predefined community composition , a certain level of optimality for each species , or a well characterized species interaction pattern , and may not be easily generalized to predict the consequences that the introduction of one species may have on the metabolism of another or to systematically study the metabolic capacity of microbial ecosystems . One fundamentally different approach to tackle this challenge is to incorporate temporal dynamics into these modeling frameworks . Previously , temporal dynamics have been successfully incorporated into single-species models to predict metabolic reprogramming , growth , and secretion rates [50]–[53] . Recently , a few preliminary studies have similarly used this approach to study microbial co-cultures composed of sub-populations of strains or of multiple species [54]–[58] . In several such dynamics-based studies each species aims to maximize its own growth on a short time-scale and the overall community dynamics is a long time-scale integration of species dynamics . This modeling framework is not dependent on maximizing community growth and more importantly , is especially suited for studying the biosynthetic and secretion capacities of microbial communities over time . Here we introduce a comprehensive computational framework tailored specifically for studying emergent biosynthetic capacity in simple microbial communities . We extend recently introduced dynamical modeling frameworks , presenting a multi-scale model that tracks both the metabolic activity of each species over time and the effect of this activity on the concentration of various metabolites in the environment . This framework therefore allows us to examine how environmental shifts induced by the activity of one species may impact the activity of another . Using this framework to model the growth of both single- and two-species microbial systems , we aim to identify instances wherein a two-species community secretes certain metabolites that cannot be secreted by any of the member species when grown in isolation . Notably , to obtain confident predictions of emergent capacity , our framework further incorporates flux variability-based techniques to account for multiple alternative FBA predictions . We first use a simple toy ecosystem model to demonstrate the ability of our framework to detect emergent biosynthetic capacity . Next , we utilize high-quality , manually curated , and previously validated genome-scale models of six microbial species to systematically characterize emergent biosynthetic capacity across a large array of growth media . We specifically set out to examine how common emergent capacity is , within which growth phase does it most frequently occur , and what combinations of microbial species are most likely to exhibit emergent capacity . We further characterize several typical mechanisms underlying emergent biosynthetic capacity and explore frequent emergent metabolites . Finally , we apply this framework to a large collection of automatically reconstructed models of >100 microbial species to validate the observed patterns on a broader scale . Taken together , our results highlight promising directions for studying unique metabolic capacities of microbial communities , facilitating future efforts to steer complex ecosystems and their environments towards beneficial states .
To systematically characterize the emergent biosynthetic capacity of microbial communities , we focus here on simple two-species microbial ecosystems . Formally , given two species and a predefined growth medium , we define emergent metabolites as metabolites that are secreted and consequently accumulate in the environment when the two species grow in co-culture but that are not secreted by either of the two species when they grow in mono-culture ( Figure 1 ) . A two-species system in a given medium is then said to exhibit emergent biosynthetic capacity if it secretes at least one emergent metabolite . Notably , this is a very strict definition , as we do not consider the potentially many cases wherein the secretion rate of some metabolite is higher in co-culture than in mono-culture , but rather focus on cases wherein the co-culture system secretes metabolites that are completely absent in the two mono-culture systems . This definition allows us to study the prevalence and determinants of fundamentally novel behavior of microbial ecosystems , rather than quantitative and potentially minor differences . Furthermore , as described below , this definition may be less sensitive to parameter selection or other inaccuracies in the underlying model ( e . g . , in the bounds set for the uptake rate of nutrients ) . Moreover , we only consider “neutral” growth media that allow each species to grow in mono-culture , rather than media that explicitly induce commensal or mutualistic interactions ( and see also ref . [48] ) . Accordingly , we go beyond studies of species symbiosis ( e . g . , [59] ) and specifically target scenarios where emergent capacity is not simply the outcome of one or both species surviving only due to the association with the other . Following these definitions , we developed a computational framework for detecting emergent metabolites and emergent biosynthetic capacity . Our framework integrates genome-scale metabolic models , temporal dynamics , ecosystem settings , and various optimization schemes . To simulate the growth of a two-species system in a given growth medium over time and its impact on the medium , we followed previous studies [54] , [57] , using a multi-scale dynamic FBA-based framework . Rather than defining an arbitrary community objective that governs the flux distribution of the system , this framework assumes that each species in the community seeks maximum growth . Briefly , in this framework , we first predict the behavior of each species ( including flux activity and growth rate ) in the initial medium within a short time interval , Δt , using FBA [50] . We use Michaelis-Menten equations and scale metabolites by the total cell density and time interval , to estimate nutrient availability and the allocation of nutrients among the species . We then determine the cell density of each species according to the predicted growth rate and update the concentration of metabolites in the medium at the end of this time interval based on the predicted uptake and secretion fluxes of the various species . Notably , as we are using a shared medium , species can then utilize metabolites secreted by other species . We perform these two steps repeatedly , each time using the updated medium composition as the initial medium for the following iteration . Simulations continue until nutrients are exhausted and all species stop growing . Notably , our framework simulates a batch culture condition , following typical dynamic FBA studies [50]–[53] . Throughout the process , we record the concentration of metabolites in the medium , obtaining a full characterization of the co-culture and medium over time . A detailed description of this ecosystem model is provided in the Methods . To detect emergent metabolites , we simulate the growth of each species in mono-culture in a similar manner , again recording the composition of the medium at each time step . We then mine these records to identify metabolites that occur in the co-culture medium at some point throughout the growth of the species , but that never occur in the medium of either mono-culture . Importantly , FBA provides only a single flux solution , whereas many alternative solutions with equally optimal growth rates may exist . Considering our definition above , we therefore wish to confirm that candidate emergent metabolites are not only absent from the specific solution obtained , but are absent from any possible solution ( under the optimal growth criterion ) . Notably , characterizing all possible solutions over time is a challenging task as we need to account not only for alternative solutions in a specific time point but rather for alternative solutions in all time points and their potential impact on subsequent time points . The space of alternative solutions may therefore expand exponentially with time as the set of solutions in each time point may depend on the solution employed in previous time points . To address this challenge , we developed an iterative flux variability analysis ( see Methods ) as a more stringent protocol for detecting emergent metabolites that accounts for this potentially expanding set of secreted metabolites . We then classify metabolites as emergent only if they do not appear in any of the alternative solutions obtained by this iterative analysis . We confirmed that this protocol filters out potentially spurious results that rely on assuming specific alternative solutions during the growth period ( see Methods ) . The results presented below are accordingly obtained using this stringent protocol . As described below , we applied this framework to several sets of two-species ecosystems growing on a large set of neutral media . For single-species models , we used both manually-curated , high quality genome-scale reconstructions and automatically generated reconstructions obtained from previously published studies ( Methods ) . Neutral media were obtained from previous studies or generated through an optimization algorithm ( see Methods ) . To illustrate the settings that can give rise to emergent biosynthetic capacity and the ability of our computational framework to detect it , we first present a simple toy ecosystem in which an environmental shift induced by one species promotes a second species to activate an alternative pathway and consequently secrete an emergent metabolite . Specifically , consider the two species illustrated in Figure 2A–B ( and see Methods for a full model description ) . Each of these species can successfully grow in mono-culture on the same simple medium ( containing nutrients A and B ) . In the process of converting exogenously acquired nutrients to biomass , the red species secretes metabolite W to the medium as a byproduct ( Figure 2A ) , while the blue species secretes metabolite C ( Figure 2B ) . However , when grown in co-culture ( and on the same medium as in mono-culture ) , the red species has access to metabolite C that was secreted by the blue species ( Figure 2C , dashed arrow ) , allowing it to utilize an alternative pathway for synthesizing metabolite Z – a precursor of biomass . If this alternative pathway is favorable for optimal growth ( see , for example , the stoichiometric details of this model in Methods ) , the red species may activate this pathway to produce more biomass and grow faster , secreting metabolite Y as a byproduct in the process . Since this metabolite Y cannot be secreted by either of the mono-cultures , it is classified as an emergent metabolite and this system is classified as exhibiting emergent biosynthetic capacity . Notably , the definition of emergent biosynthetic capacity is medium-dependent; for example , in this toy ecosystem , Y will not be classified as an emergent metabolite if the growth medium already contains metabolite C since in such a scenario , metabolite Y can be secreted also by the red species in mono-culture . Applying our framework to this ecosystem , this emergent capacity was clearly observed ( Figure 3 ) . Evidently , the availability of metabolite C in co-culture allowed the red species to grow faster than it grew in mono-culture ( Figure 3A ) . Tracking the concentration of various metabolites in co-culture and in mono-cultures ( Figure 3B ) , several differences were further observed . For example , the improved growth of the red species in co-culture led to a faster depletion of metabolite A , which was accordingly exhausted earlier , preventing further growth . Most notably , however , metabolite Y ( Figure 3B , lower panel ) , which was completely absent in either of the mono-cultures , quickly accumulated in co-culture , owing to the activation of the alternative pathway in the red species . Comparing the concentration of metabolites in co-culture with those obtained in mono-culture and applying our variability analysis to examine alternative solutions ( Methods ) , our framework therefore identified Y as an emergent metabolite . The toy ecosystem above was specifically designed to promote emergent biosynthetic capacity . While such synergistic capacities can clearly occur in communities of real microorganisms , it is not clear a priori whether it is common , how likely it is to occur , and which factors may contribute to it . Next , we therefore set out to examine how prevalent emergent biosynthetic capacity is in natural systems and what the determinants of such capacity are in simple microbial communities . To this end we obtained high-quality genome-scale metabolic models of 6 species with potential health and environmental applications ( Methods ) . Importantly , each of these models was manually curated , experimentally tested , and used in multiple previous studies of microbial metabolism [48] . We then applied our framework to characterize emergent biosynthetic capacity in all possible pairwise species communities . Since , as discussed above , emergent capacity is media-dependent , we simulated the growth of each of these two-species communities in 100 random minimal neutral media ( see Methods ) and recorded the cases in which emergent capacity occurred . Interestingly , our analysis demonstrated that emergent biosynthetic capacity is fairly common . Almost all pairwise species combinations analyzed ( 13 out of 14 ) exhibited emergent capacity in at least one of the 100 media , with some species pairs ( e . g . , E . coli and B . subtilis ) exhibiting emergent biosynthetic capacity in 67 of the 100 tested media ( Figure 4 ) . Overall , 30% ( 421 ) of the 1400 community/medium settings analyzed demonstrated emergent capacity . Notably , in most cases ( 343 of 421 ) communities with emergent capacity secreted only 1 emergent metabolite , but in a few cases ( 12 of 421 ) 3 different emergent metabolites were secreted simultaneously . Interestingly , in 16 community/medium settings ( mostly involving E . coli and B . subtilis ) , both species secreted emergent metabolites that were consequently consumed by the other species via cross-feeding , exhibiting an intriguing mutual emergence scenario . Focusing on the emergent metabolites that were identified in each simulation , we found in total 28 metabolites that were classified as emergent in at least one community/medium . Of these , ethanol and urea were the most frequently secreted emergent metabolites ( 37 . 5% and 21 . 9% of the cases in which emergent biosynthetic capacity was detected , respectively ) . Figure 4 lists the complete set of emergent metabolites , ranked by their frequency , and the communities in which they were detected . Perhaps not surprisingly , the most frequently secreted emergent metabolites include common byproducts of microbial metabolism , such as ethanol , succinate , and acetate ( from fermentation ) , and urea ( from nitrogen metabolism ) . Several of these metabolites ( e . g . ethanol , acetate , urea , glycolate ) have been previously highlighted as important cross-feeding metabolites that can substantially impact the composition of various microbial ecosystems [60]–[62] . To better understand the contribution of each species in the community to the observed emergent biosynthetic capacity , we further examined each community with an emergent metabolite , to determine which of the two species actually secreted that metabolite to the environment ( the producer ) and which species was the non-secreting community member ( the partner ) . Focusing on the most frequently secreted emergent metabolites , we found that emergent metabolites were often secreted by a single dominant producer species , whereas many other species could serve as partners ( Figure S1 ) . Critically , many dominant producer species have been shown experimentally to secrete the compound in question given the proper environmental conditions and media . For example , succinate secretion has been demonstrated in B . subtilis [63] and E . coli [64] . Similarly , the primary producers of acetate in our study , S . typhimurium , M . barkeri , and E . coli , are capable of secreting appreciable quantities of this compound under certain conditions [65]–[67] . This analysis further demonstrated that B . subtilis was almost solely responsible for the secretion of several frequent emergent metabolites , including urea , nitrite , glyoxylate , and fumarate , partnering with almost any other species . Notably , B . subtilis was the only species included in our analysis that has a complete urea cycle , providing metabolic flexibility for nitrogen and amino acid metabolism and potential utilization of varied nutrient sources . Indeed , some of the emergent metabolites secreted by B . subtilis are associated with the operation of the urea cycle ( see also Figure S3 ) and have been shown to be the end product of this process in other microorganisms with a complete urea cycle [68] . This promiscuity of B . subtilis may also reflect its adaption to diverse habitats including soil , plant roots , aquatic environments , and the gastrointestinal tract of animals [69] . In other cases , however , only a very specific combination ( or combinations ) of species led to the secretion of an emergent metabolite ( e . g . , Sulfate; Figure S1 ) . Interestingly , comparing the total biomass produced in co-culture to the combined biomass produced by the two mono-cultures , we found that most ( but not all ) communities benefit from such emergent capacity , even though overall community growth was not the optimization objective in our framework ( see Text S1 and Figure S8 ) . The two species comprising the toy ecosystem discussed above demonstrate a simple mechanism of emergent capacity . Specifically , in this example , one organism constructed its niche , converting nutrients or energy sources ( metabolites A and B ) into forms accessible to other species ( metabolite C ) . Species that share this niche may consequently “reprogram” their metabolic activity ( e . g . , via regulation ) to obtain optimal growth and differentially impact their environment ( e . g . , by secreting metabolite Y ) . Clearly , however , in real organisms , which can catalyze hundreds and thousands of reactions , metabolic reprogramming can be markedly more complex , involving differential activation of numerous reactions in response to environmental shifts . Such reprogramming may , for example , activate ( or enhance ) some reactions while deactivating ( or repressing ) other reactions , to support multiple nutrient requirements , energy production , and redox balance . Here , we therefore set out to examine whether similar cross feeding-based mechanisms were responsible for the prevalence of emergent capacity observed above in models of real microbial species and to characterize metabolic reprogramming patterns that may be associated with this capacity . We analyzed the metabolic fluxes in each community that exhibited emergent biosynthetic capacity , comparing the predicted fluxes in co-culture with those predicted in the two mono-cultures . We identified in each two species system the species that secreted the emergent metabolite ( i . e . , the producer; and see also Figure S1 ) and the time point at which the emergent metabolite was first secreted . To examine whether a simple cross-feeding behavior could account for the observed emergent capacity , we first aimed to identify cross-feeding fluxes at this time point that might have prompted the producer to secrete the emergent metabolite . Specifically , we identified metabolites taken up by the producer that were not provided in the initial growth medium and that were not secreted by the producer itself in earlier time points . To confirm that these cross-feeding fluxes were sufficient to induce the secretion of an emergent metabolite , we simulated the growth of the producer in mono-culture again , with small amounts of the detected cross-feeding metabolites added to the medium . We found that in almost all cases ( 99 . 4% ) the addition of these metabolites indeed resulted in a reprogramming event , altering the activity of the producer and causing it to secrete the emergent metabolite . The few cases ( 3 ) where this did not occur may require the presence of additional emergent metabolites that were secreted by the producer at earlier time points ( and that were therefore not added to the medium ) or the availability of the cross-feeding metabolite at a higher concentration . Further analysis also demonstrated that in most cases ( ∼98% ) emergent metabolites could in fact be secreted by the producer also in mono-culture but that such secretions were suppressed when biomass production was maximized ( Text S1 and Figure S9 ) . As this growth penalty was often relatively small , it appears that the role of the partner in many communities amounted to reducing the cost of specific emergent secretions by allowing the producer to shift its metabolic activity toward a pattern that was suboptimal when growing in isolation , and only in some cases did the partner actually provide metabolic capabilities necessary for emergent secretion . Clearly , one of the benefits of a modeling framework is that it allows us to comprehensively characterize flux distributions in each organism and to fully characterize complex reprogramming behaviors . Since ethanol and urea were the two most frequently secreted emergent metabolites observed in our analysis , we examined several common cases of metabolic reprogramming that resulted in the secretion of these metabolites in more detail and illustrated two typical examples of such reprogramming patterns . In the first example , acetate was identified as a cross-feeding metabolite from Shewanella oneidensis to Methylobacterium extorquens , resulting in emergent ethanol production ( Figure 5 ) . Specifically , the uptake of acetate allowed M . extorquens to enhance energy production by providing increased carbon flow into the TCA cycle . Acetate influx additionally enhanced acetyldehyde dehydrogenase and activated alcohol dehydrogenase fluxes through which NAD+ was generated . This excess NAD+ production allowed reactions catalyzed by malate dehydrogenase and 2-oxoglutarate dehydrogenase to regenerate NADH in the TCA cycle as a source of ATP as well as reducing power that were necessary in many other reactions . Notably , a similar phenomenon , wherein exogenous acetate induces ethanol secretion to maintain intracellular redox balance , has been previously documented in Lactobacillus casei [70] . Emergent secretion of ethanol to maintain redox balance was also observed in our analysis in many other communities and in various growth media , often involving markedly more complex reprogramming patterns . For example , in the case illustrated in Figure S2 , ethanol was produced as an emergent metabolite by E . coli in the presence of B . subtilis following a complex reprogramming behavior that involved not only enhancement and activation of various reactions but also flux repression in several branch points to adjust carbon flow into the TCA cycle . Furthermore , the optimization of NADH production appeared to be a common strategy that resulted in emergent ethanol secretion . As a second example we focused on a community that secreted urea and succinate by B . subtilis and E . coli respectively ( Figure S3 ) . In this example , the metabolite fumarate cross-fed from B . subtilis to E . coli and promoted energy production as well as amino acid biosynthesis in E . coli . Fumarate uptake by E . coli was partly mediated by a dicarboxylic acid transporter that resulted in succinate secretion . Utilization of a dicarboxylic acid transporter for fumarate uptake that is accompanied by succinate secretion has been previously documented in E . coli in certain media [71] . Meanwhile , acetate cross-feeding from E . coli to B . subtilis allowed a larger carbon flow into the TCA cycle in B . subtilis , resulting not only in more energy but also in more 2-oxaloglutarate production by B . subtilis . Consequently , the amount of glutamate required to generate 2-oxaloglutarate and NDPH was decreased ( Glutamate:NADP oxidoreductase ) , and excess glutamate was rerouted to the urea cycle to facilitate arginine biosynthesis . Arginase was activated in this new metabolic route , ultimately resulting in the secretion of urea . In determining the prevalence of emergent capacity above and in characterizing various underlying mechanisms of this capacity , we compared the activity of the co-culture and mono-cultures during the initial phase of growth . Specifically , for the results reported in the previous sections we focused on all time points up to the 1 hr time point . During this period organisms in our simulations were still exhibiting exponential growth and were not limited by nutrient availability . However , as cell density increases and nutrients get depleted , organisms may further modulate their activity to optimize their growth in a nutrient limited environment . This may potentially lead to different modes of emergent capacity and to the secretion of emergent metabolites that may not be observed during early growth . To test this hypothesis , we repeated the analysis above considering the entire growth period ( i . e . , until some required nutrients were totally exhausted and growth ceases ) and for each emergent biosynthetic event recorded the first time point the emergent metabolite was secreted . Examining the distribution of such events over time , we indeed found two waves of emergent capacity , one that occurred mostly during the early growth period , and one that occurred toward the end of the growth period ( Figure 6 ) . While the first wave represents emergent capacity that arose as soon as the two species were introduced into the same medium , the second wave may reflect emergent capacity that was dependent upon organisms' activity in a nutrient depleted media . In total , when the entire growth period was considered , 52% ( 729 ) of the 1400 community/medium settings analyzed demonstrated emergent capacity ( compared to the 30% reported above when only the early growth period was considered ) . During this full period , all 14 pairwise species combinations analyzed exhibited emergent capacity in at least several different media , with some species combinations exhibiting such capacity in >80 of the 100 media tested ( see Figure S4 ) . Again , in most cases ( 414 of 729 ) only a single emergent metabolite was secreted , but in one case as many as 11 emergent metabolites were secreted in the same community/medium combination . To further characterize the differences and similarities between these two waves of emergent capacity , we compared the set of emergent metabolites reported above for the initial growth period ( 1 hr ) to the larger set of emergent metabolites obtained when the entire growth period was considered . As expected , this latter set was markedly larger , with 64 different metabolites ( compared to the 28 identified in the 1 hr set ) classified as emergent in at least one community/medium ( Figure S4 ) . Specifically , among the metabolites that were detected as emergent only in the late growth period , glycolaldehyde , malate , and propionate were the most prevalent . Interestingly , propionate was previously observed as a late onset emergent metabolite in a co-culture of Megasphaera elsdenii and Streptococcus bovis [72] . In that system , propionate was produced by M . elsdenii as a result of secondary fermentation of lactate that was secreted to the medium by S . bovis and was detected only 3 hours after the simultaneous inoculation of both species once lactate accumulation was sufficient [72] . Other emergent metabolites that were detected only in late growth may reflect similar processes , wherein one species shifted to metabolize secondary byproducts secreted by the partner . In other cases , however , late detection of emergent metabolites may depend on specific thresholds used . For example , glycolaldehyde , which was produced as an emergent metabolite by M . barkeri , was not detected in the early growth period due to the slow growth of this species and consequently the long time required for glycolaldehyde to accumulate in the medium and to pass the threshold used . More generally , however , while certain emergent events could only be observed in late growth as described above , the results obtained for the 1 hr time point and those obtained for the entire growth period were consistent ( see also Figure S5 ) , with ethanol , succinate , and urea being the most frequently secreted emergent metabolites . Above , we demonstrated that emergent biosynthetic capacity is generally prevalent in simple two-species communities . Here , we set out to detect properties of the community that may be associated with this capacity and that can help us to determine a-priori which community compositions are more likely to exhibit emergent capacity . Identifying such properties can inform future design efforts , allowing them to focus the search for novel metabolic activity on specific species combinations . We specifically examined whether the functional and phylogenetic distance between community members was a potential determinant of emergent biosynthetic capacity and of the ability of two species to interact and exhibit a novel behavior . To this end , we compared the average number of emergent metabolites secreted by each two-species community across random neutral media to the functional and phylogenetic distance between the two species ( Methods ) . As demonstrated in Figure 7A , we found that communities which consisted of functionally close species tended to exhibit low levels of emergent capacity . For example , communities where both species were gamma-proteobacteria ( red triangles in Figure 7A ) produced on average only 0 . 08 emergent metabolites . Interestingly , however , communities in which the two species were functionally distant from one another similarly tended to secrete only few emergent metabolites . Specifically , in our dataset , communities containing a bacterium paired with the archaea M . barkeri ( blue diamonds in Figure 7A ) produced on average only 0 . 21 emergent metabolites . It was only when the two species that comprised the community were at some intermediate functional distance that higher levels of emergent biosynthetic capacity were observed ( black dots in Figure 7A ) . Specifically , the numbers of emergent metabolites in this intermediate group were significantly higher than those observed in both the functionally close species group and the functionally distant species group ( p<10−18; two sample t-test ) . This “Goldilocks” principle of emergent biosynthetic capacity , that emergent capacity of a community is maximized when member species are not too functionally close , nor too functionally distant , may reflect the ability of species to metabolically interact with one another and beneficially exchange metabolites . Specifically , as discussed further below , functionally similar species may exhibit very comparable metabolic strategies in any given environment , such that the by-products of one species do not provide any novel benefit to the second . In contrast , two functionally distant species may apply markedly different metabolic strategies , such that the metabolites secreted by one species are not compatible with the enzymatic capacity of , and therefore cannot be utilized by , the second species . A similar relationship between distance and the likelihood of emergent capacity could also be observed when using phylogenetic distance rather than functional distance to measure the similarity between the two species ( see Figure S6A ) . Finally , to disentangle the role of species composition from that of the growth media , we repeated the above analysis using a set of 500 ‘universally neutral’ media for all 6 species ( Text S1 ) . This analysis further confirmed that the observed Goldilocks principle stems from the functional distance between the two species and is not an artifact of the specific media used for each species pair ( Text S1 and Figure S10 ) . The analysis above , associating optimum emergent capacity levels with an intermediate functional distance between the community members , was demonstrated with high-quality , manually curated metabolic models , but relied on a limited number of species combinations . To confirm the general applicability of this Goldilocks principle , we applied our framework to a markedly larger collection of automatically generated models . These models are potentially less accurate than the manually curated models used above , but represent a significantly higher species diversity and a wider range of functional distances . Specifically , we obtained 116 SEED-based FBA models used in [44] , allowing us to examine 6670 two-species communities . Since this dataset did not include a collection of minimal neutral media for each two-species community ( as those that were available for the 6 species dataset from [48] ) , minimal neutral media were computed using a mixed-integer linear programming algorithm ( see Methods ) . Moreover , as simulating the growth of each of the 6670 communities on 100 different growth media is too computationally expensive ( Methods ) , we investigated the level of emergent capacity for each community in one minimal medium ( computed specifically for this community ) and binned communities into 10 groups representing species pairs with similar functional distances . We recorded the fraction of communities in each bin that exhibited metabolic capacity . Our results mirrored and confirmed those observed in the smaller 6 species dataset ( though a significance analysis is challenging with just one medium per community ) , with an optimum level of emergent capacity obtained by communities in which the two species were at some intermediate functional distance ( Figure 7B ) . Again , a similar pattern held when phylogenetic distance rather than functional distance was used ( Figure S6B ) .
Above we introduced a comprehensive computational framework for exploring enhanced metabolic capacities in simple two-species microbial systems . Investigating a large number of communities and growth media , our results suggest that emergent biosynthetic capacity is relatively prevalent , and can be observed in many , if not all , communities , under certain environmental settings . Notably , while many community/medium combinations exhibited emergent capacities , typically our analysis detected only very few emergent metabolites in each such combination . This may be partly due to our stringent flux variability-based analysis ( Methods ) that likely filters potentially real emergent metabolites , and thus underestimates the level of species-species interaction and emergent metabolic capacity in nature . Another important factor in determining the prevalence of such capacity is our focus on minimal media that probably provide relatively low nutrient diversity compared to more complex media commonly found in natural habitats . Such naturally-occurring complex environments may allow organisms to utilize multiple resources and potentially produce additional byproducts . Importantly , however , such complex , non-minimal environments may instead allow cohabiting species to partition available resources [73] , reducing the impact that one species may have on the other , and consequently the prevalence of emergent biosynthetic capacity . Moreover , as discussed above , our work considered only neutral media , in which both species could survive in isolation . Natural ecosystems , in contrast , often exhibit high levels of obligate symbiosis , wherein one species ( or both ) requires the presence of the other to grow . Such non-neutral media , by definition , promote emergent metabolism , as the activity of the co-culture ( in which both species grow ) is fundamentally different from that of the two mono-cultures ( in which one or both species are not growing ) . While such scenarios are clearly interesting , here we focused on minimal and neutral media ( in which emergent capacity is not a direct outcome of obligate symbiosis ) to systematically and comprehensively characterize the limits of emergent biosynthetic capacities without the confounding effects of niche partitioning or obligate symbiosis , laying the foundation for investigating how species jointly influence their shared environment . Interestingly , our analysis further suggests that emergent biosynthetic capacity is especially likely when community members are neither too similar functionally and phylogenetically , nor too different . This finding is perhaps not surprising: When cohabiting species are too similar functionally , each species is not likely to introduce any metabolite into the environment that the second species cannot already produce by itself , and consequently , the activity of the first species is not likely to modulate the activity of the second . Similarly , if the two species are functionally very different , each species may produce many metabolites that the second species is not capable of producing , but these may be too remote from the metabolism of the second species for it to utilize them . This Goldilocks principle is also in agreement with observations obtained through a simple network expansion analysis [74] or a pathway overlap calculation [75] , examining the overall potential of combined metabolic networks . By further characterizing frequently secreted emergent metabolites and examining typical mechanisms of emergent capacity , we were also able to obtain insights into various principles that govern such species-species and species-environment interactions and to point to many potential modes by which microbial species jointly impact their environments . Taken together , our results provide a first systems-level characterization of emergent biosynthetic capacities across simple microbial communities . As described in the Introduction , several preliminary attempts to model multi-species systems have been previously introduced , some of which have utilized constraint-based modeling approaches or relied on the integration of multiple single-species FBA models ( e . g . , [44]–[48] ) . Importantly , the main aim of our study was not necessarily the development of a new modeling framework , but rather the investigation of emergent biosynthetic capacity in microbial communities . Many elements in our framework were therefore adopted from previous studies or adapted from advanced single-species modeling techniques . Yet , as a whole , our framework includes several key innovations that make it especially fitting for studying emergent capacity or , more generally , ecosystems' metabolic activity . For example , similar to Zhuang et al . [57] , our framework takes a dynamics-based modeling approach . However , Zhuang et al . focused on recapitulating a specific , well-characterized community in specific environmental settings , using measured parameters and a species-specific cell death rate . Here , in contrast , we wish to characterize universal principles and large-scale trends of emergent capacity , and therefore focus on a systematic investigation of numerous communities and a plethora of media , for which such detailed information is clearly not available . Our framework , therefore , has integrated an exhaustive Flux Variability Analysis with an iterative temporal modeling approach to account for the potentially many alternative solutions such uncharacterized communities may exhibit . This integration of two constraint-based techniques ( namely , Flux Variability Analysis and dynamic-FBA ) is essential for a reliable detection of emergent capacity and has not been implemented before . It is also worthwhile to emphasize again the importance of a dynamics-based approach for studying and discovering emergent community metabolism , since to date , only very few studies have employed this technique whereas most other FBA-based models of multi-species systems have applied a non-dynamic , joint-model approach ( e . g . , [76] ) . In such joint-model methods , the stoichiometric matrices of the various species are combined into a single matrix , often with the introduction of explicit exchange reactions . Indeed , such methods are computationally less expensive as they rely on a single ( or only a handful of ) optimization task ( s ) , without an iterative optimization procedure for tracking the system's dynamics . Yet , in contrast to our method ( which optimizes each species' model separately ) , the size of the stoichiometric joint matrix grows with the number of species , potentially exceeding the capacity of available solvers . More importantly , such methods often require the definition of a universal objective function to represent the optimization objective of the community , such as the total growth of the community or some other community-level feature . Notably , however , a community-level objective may unjustly promote species cooperation , potentially pushing species toward altruistic behavior that could benefit the community . This may be an appropriate approach in cases where cooperation is expected and well-characterized , such as in stable communities of obligate symbiotic pairs ( e . g . , [76] ) , but may not be suitable for more general settings in which cooperation is not expected a priori . In our framework we therefore instead assume that each species aims to maximize its own growth without regarding the benefit of the community and accordingly optimize each model ( representing each species ) separately . With this assumption , species impact other community members only by modifying the shared environment as part of this selfish growth process , and a temporal dynamics approach is used to allow environmental shifts induced by the activity of one species to potentially affect the behavior of other species in subsequent time points . Moreover , in the context of our study , this dynamics-based approach further allowed us to obtain insights crucial for understanding emergent capacity . For example , cross-feeding metabolites may be secreted at a slow rate , and a long period may be required for such metabolites to accumulate in the environment and reach high enough levels to affect the activity of other species and to induce emergent secretion . Similarly , nutrient availability in the environment can drop over time , pushing community members to shift their flux activity and secrete emergent metabolites only when cell density becomes high . Emergent events can therefore occur at many different time points and may involve different mechanisms , as also demonstrated by our findings on early and late onset of emergent metabolism . Such temporal patterns can only be characterized and investigated with a dynamics-based simulation . More generally , considering the dynamic nature of many natural environments and our focus on emergent secretion events that may further shift the composition of the environment , a dynamical modeling approach capable of tracking these environmental shifts and their temporal impact on the species inhabiting the environment is a natural choice . Ultimately , however , the various approaches for studying community metabolism and the multiple frameworks developed to date are all essential for gaining a comprehensive , principled understanding of microbial communities and of species' metabolic interactions . Future efforts to integrate such different modeling frameworks and to develop a multi-scale framework capable of capturing the many facets of ecosystem metabolism could be especially exciting . Our modeling framework may clearly have some important limitations . To avoid an arbitrary definition of community objective , we used dynamic FBA-based methods to model community metabolism , assuming that each species aims to maximize its growth and that community dynamics is a second order consequence of species behavior . However , in many natural ecosystems , some species may exhibit sub-optimal growth due to constraints that are currently beyond the scope of FBA . Sub-optimal growth can be observed , for example , when an organism is introduced to a non-natural habitat that it may not have encountered during its evolution . One way to address this challenge is to incorporate additional omic data to augment an FBA-based framework and to guide FBA prediction [53] , [77]–[79] . With recent advances in high throughput meta-omic technologies [80] , ecosystem-level meta-transcriptomic and meta-proteomic data are continually becoming available and efforts are needed to develop computational methods for incorporating such large-scale meta-omic data into a community-level FBA framework . Other factors that may lead to sub-optimal growth include inhibitory mechanisms such as the bacterial toxin-antitoxin system [81] and inter-species communication mechanisms such as quorum sensing [82] . Such mechanisms are currently not accounted for by genome scale metabolic models . More generally , although efforts have been made to take into account potential sub-optimal growth of community members [46] , the rationale and the extent of sub-optimality remains unclear . Other factors may further impact community growth and species interaction . For example , some organisms may have adapted to harsh habitats , such as high temperature or salt concentrations that are stressful to other organisms . Spatial structure could also constrain inter-species metabolic flow and has been shown to induce species cooperation within a remarkably short time of laboratory evolution [83] . Integrating such physical-chemical factors into an enzymatic- or stoichiometric-base framework is a challenging task and will require further developments . Recently introduced efforts to model multiple cellular processes on a whole cell level [84] or to mathematically model spatial constraints among interacting partners [85] are promising advances toward this goal . Ultimately , however , any computational or mathematical model aiming to capture the activity and dynamics of microbial communities or their impact on the environment is bound to be incomplete and may fail to incorporate various factors that could affect the behavior of the community . Environmental attributes , the induction of stress , pH , alcohol , antibiotics , and signaling may all steer the behavior of a specific community away from our metabolic model-based prediction . In this study we therefore aimed to identify large-scale patterns of emergent capacity and to generate hypotheses concerning the universal principles that govern emergent behavior , rather than to predict the metabolic activity of a specific species pair in a specific medium . These robust large-scale patterns ( such as the prevalence of emergent capacity or the Goldilocks principle ) are potentially less sensitive to model incompleteness and allow us to obtain fundamental insights concerning the capacity , timing , and likelihood of emergent biosynthesis . In this context , a metabolic modeling-based framework further provides a tool for studying an ideal , well-controlled metabolic system , ignoring other confounding processes and focusing on mapping of boundaries and first principles of emergent behavior that is governed by metabolism alone . One of the promises underlying microbial ecology research is the potential for therapeutic or bioengineering applications [41] . In contrast to traditional efforts aimed to genetically engineer a single species toward desired metabolic tasks , engineering microbial communities by constructing specialized combinations of already existing strains is a cost-effective solution [3] . Many clinical and industrial applications may be difficult to address at the single species level but could be potentially attainable at the community level . Of specific interest are microbiome-based therapy applications and gut microbiome transplantation efforts aiming to restore healthy phenotypes or endow the host with some metabolic capabilities [20] . Such transplantation efforts currently utilize complete microbiome transfers from healthy donors or simple synthetic microbiomes constructed from a small collection of carefully handpicked strains [19] , [21] , [22] . Yet , to allow the construction of such engineered communities on a large scale and to enable researchers to search the vast space of possible community compositions , a more comprehensive design framework is clearly needed [42] . The development of predictive community models is a critical first step toward such rational design of microbial communities , and will enable researchers , for example , to move from complete microbiome transplantations toward a targeted and personalized microbiome-based therapy [41] , [86] . Such a predictive comprehensive framework for modeling microbial communities , however , cannot be gained without a principled understanding of how the joint activity of multiple species influences their environment and vice versa . A recent study , for example , demonstrated how different environmental conditions may induce different forms of species interactions and that it may be the environment , rather than the gain or loss of genes that have larger impact on the specific type of interaction two species will have [48] . Conversely , our study highlights the prevalence of species interactions that will impact the environment via emergent biosynthetic capacities . Our framework not only allows exploring the boundaries of the metabolic tasks microbial consortia can accomplish but also provides mechanistic insights on the pathway level and accounts for the impact of the abundances of species in the community . Moreover , our findings suggest potentially universal principles that have important bearing on community design efforts . For example , the Goldilocks principle observed above points to potentially more promising starting points in the search for communities that exhibit novel biosynthetic capabilities even when relatively little is known about the participating species . Similarly , the contribution of B . subtilis to multiple emergent biosynthetic processes demonstrated by our analysis suggests it may be a preferred partner in designed communities , in agreement with its role in promoting plant growth or in maintaining healthy gut communities [69] , [87] . Clearly , much work is still ahead before a complete , predictive , and multi-scale framework for modeling microbial communities can be fully realized . Fast and accurate metabolic reconstructions , multi-omics data integration , and the development of novel community-level integration techniques will all contribute tremendously to our ability to model naturally occurring complex ecosystems . Furthermore , in the context of host-associated communities , it will likely be equally important to incorporate a model of the host and its interaction with the community . Yet , our framework and other modeling frameworks of simple microbial communities are an important first step toward the construction of such a comprehensive model . Moreover , these frameworks are already capable of generating testable hypotheses on niche construction and on microbial interaction , further elucidating the forces that govern the assembly , function , and dynamics of microbial ecosystems [25] . Multiple such frameworks could be integrated and ultimately coupled with optimization and design algorithms , resulting in a comprehensive framework for designing novel microbial communities with desired metabolic activities or clinical manipulation of environmental- and host-associated communities .
Flux Balance Analysis ( FBA ) describes cellular fluxes at steady state with the mass balance equation: ( 1 ) where x is the vector of metabolite concentrations , v is the vector of reaction fluxes , and S is the stoichiometric matrix describing how many molecules are consumed or produced by each reaction . Under this steady state assumption , nutrients acquired from the environment are used to produce biomass or byproducts , with no accumulation of metabolites in the cell . Additional constraints , including reversibility of reactions and measured exchange fluxes , can be incorporated to further limit the possible solution space of metabolic fluxes . Given the complete set of constraints , FBA predicts a specific flux distribution by optimizing a given objective function , typically maximum growth [88] . Growth rate is approximated by a vgro reaction that consumes energy and a predefined set of nutrients at some relative proportion to form biomass: ( 2 ) This maximum growth objective has proved successful in providing predictions that are consistent with experimental data [33] , [50] , [89] . Given the set of species that comprise some microbial community and the composition of an initial growth medium , we used a computational framework to characterize the growth of the community on this medium over time . Our framework is a multi-species extension of the dynamic FBA method [50] , [51] , which aims to predict the temporal behavior of microbial systems ( and see also refs [54] , [57] ) . Briefly , we divided the entire growth period into short time intervals ( 0 . 1 hr ) , assuming a steady state solution in each interval and using FBA to obtain the flux distribution and growth rate of each species during this interval . We then calculated the impact of this predicted activity on the community and on the environment at the end of this interval and updated environmental attributes before simulating the next time interval . Specifically , each time step includes these three basic steps: The concentration of metabolite j in the medium , xj , was updated according to the biomass biok ( t1 ) , the growth rate μk , and the exchange flux vkj of all species k in the community over Δt , using an equation derived by integrating the differential equation dxj/dt = ∑ vkj *biok ( t1 ) over time ( see detailed derivation of this equation in Text S1 ) : ( 6 ) Steps 1–3 were repeated until all of the species in the ecosystem stopped growing in this batch condition . The dynamics of growth , exchange fluxes of all species , as well as the concentrations of all the metabolites in the medium were recorded . A similar protocol was used to simulate the growth of the two single-species mono-culture systems . The initial biomass concentration in the two mono-cultures was set to be the same as that of the co-culture ( 0 . 01 g/Liter ) such that the initial nutrient uptake rates are comparable . Once the dynamics of the co-cultures and mono-cultures were simulated , the obtained concentrations of the various metabolites over time were used to detect emergent metabolites . Given a specific growth period ( e . g . , from beginning of growth to the 1 hr time point ) , an emergent metabolite was defined as a secreted metabolite whose concentration in the medium exceeds a detection threshold ( 0 . 001 mM ) at some point during this period in co-culture , but not in either of the two mono-cultures . Note that this definition also naturally excluded metabolites that were present in the initial growth medium . As described in the main text , FBA provides only a single solution , whereas multiple alternative solutions may exist . Assuming a single solution in each time point while simulating the growth of a given mono-culture can clearly underestimate the scope of metabolites that this mono-culture may secrete to the medium , and consequently lead to spurious prediction of emergent metabolites . To account for all alternative solutions in each time point and for their impact on the behavior of the organism in subsequent time points , we developed an Iterative Flux Variability Analysis , extending the previously introduced Flux Variability Analysis ( FVA ) [94] . Briefly , FVA infers the full set of metabolites that can be secreted while still satisfying the maximum growth criterion . This is done by attempting to maximize the secretion rate of each metabolite subject to the optimal growth rate calculated via FBA and examining whether the obtained maximized flux is positive . As an iterative extension of this method , we performed FVA in each time point , and took into account possible secretions identified by FVA when updating the concentration of metabolites in the growth medium before simulating the next time interval . To avoid simulating the exponentially growing set of possibilities , we assume that all possible metabolic secretions were in fact secreted simultaneously to the medium at the maximal rate without compromising growth . This protocol therefore provided an upper bound for the scope of metabolites that may be secreted by the mono-culture throughout the growth period . Using this large set of metabolites potentially secreted by the mono-cultures to exclude candidate emergent metabolites accordingly provided a stringent criterion for emergent biosynthetic capacity , potentially underestimating the set of emergent metabolites . In contrast , results obtained with the simple FBA-based protocol described above could overestimate the prevalence of emergent capacity as some identified emergent metabolites are solution-dependent and may not be inferred when alternative solutions are considered . Figure S7 illustrates the differences and similarities between the set of predicted emergent metabolites obtained by these two protocols , highlighting emergent events that were identified by the simple FBA-based protocols and that were removed when the more stringent FVA-based protocol was used . Interestingly , many events of emergent succinate secretions were filtered out by the FVA-based protocol , suggesting that succinate secretion was often found as an alternative optimal solution in mono-culture with potentially higher enzymatic cost . Overall , however , the prevalence level of emergent metabolites was consistent across the two protocols . Throughout the text we aimed to focus on high-confidence predictions of emergent capacity and reported results obtained by the stringent FVA-based protocol . The toy ecosystem model described in the main text ( and see Figure 2 ) was defined as follows: Red Species: Mass balance constraints dA/dt = −R1−J1 = 0 dD/dt = R1−3R2−R4 = 0 dZ/dt = R2−10R3+R4 = 0 dW/dt = 2R2−J4 = 0 dC/dt = −R4−J2 = 0 dY/dt = R4−J3 = 0 Flux constraints LB1<J1<∞ LB2<J2<∞ LB3<J3<∞ LB4<J4<∞ −∞<R1<∞ −∞<R2<∞ −∞<R3<∞ −∞<R4<∞ Blue species: Mass balance constraints dB/dt = −R5−J6 = 0 dX/dt = R5−50R6 = 0 dC/dt = R5−J7 = 0 dA/dt = −R5−J5 = 0 Flux constraints LB5<J5<∞ LB6<J6<∞ LB7<J7<∞ −∞<R5<∞ −∞<R6<∞ In the definition above , the Ri denote internal reactions whereas the Ji denote transport reactions . The uptake limits , LBi for all nutrients were determined as described in eq ( 3 ) at each time point . Notably , as indicated by the expressions for dD/dt and dZ/dt in the red species , R2 consumes 3 unit of D to generate 1 unit of product Z whereas R4 requires only 1 unit of D to generate 1 unit of Z , making R4 a more favorable reaction over R2 as it allows a higher biomass yield per unit of metabolite D . Note also that Y is secreted as an emergent metabolite once R4 is active ( and see also Figure 2 ) . We used FBA models from two previously published studies on microbial metabolism . The first set was studied in ref [48] and contained 6 manually curated , high quality species models that span a wide phylogenetic range . This set included models of Escherichia coli [95] , Salmonella typhimurium [96] , Bacillus subtilis [97] , Methanosarcina barkeri [98] , Shewanella oneidensis [99] , and an extended genome scale model of Methylobacterium extorquens provided by Stephen Van Dien [100] . To ensure compatibility with computationally derived media ( see below ) , we obtained all these models directly from ref [48] , using the same versions of the models as in this previous study . We additionally obtained from ref [48] a large scale dataset of computationally derived growth media for each pair of species . These media were classified as either mutualism-inducing media ( i . e . , media that allow for the growth of both species in co-culture but do not support growth of either species when grown in isolation ) , commensalism-inducing ( i . e . , sustain growth of one of the two species but not the other ) , or neutral ( i . e . , sustain growth of each species individually ) . From this dataset , we randomly selected 100 neutral media for each of the two-species communities except for the Shewanella oneidensis/Methanosarcina barkeri pair for which no neutral media was found in ref [48] . Helicobacter pylori , which was also studied in ref [48] had very few neutral media available and was therefore not included in our analysis . The complexity of these media may depend on the models included in each community , but overall , these minimal neutral media contained on average 25 . 5±5 . 9 metabolites . Roughly 12∼15 of these compounds were inorganic ions ( e . g . Zn2+ , cobalt2+ ) that are essential and defined as biomass components . The second set included 116 automatically reconstructed models from the SEED pipeline [37] and was obtained from ref [44] ( Table S1 ) . Since a large set of neutral media for communities composed of these SEED models was not available , we applied a mixed-integer linear programming ( MILP ) method , similar to the one described in ref [44] , to infer minimal neutral media for each community . This was done by finding a minimal set of metabolites that supports the growth ( above some minimal rate ) of each species in isolation , out of a large initial set of nutrients composed of the union of all transportable metabolites of the two species . Conceptually , the objective of this MILP problem was to remove as many nutrients as possible from the initial set ( and hence leaving as few as possible in the medium ) , while maintaining the primary constraint of allowing the growth rate of each species to be larger than some minimal threshold ( set here to be μk≥0 . 3 ) . To formulate this objective , we first defined a binary decision variable θ for each nutrient , denoting whether this nutrient was removed from the medium or not ( i . e . , when θ = 1 the nutrient was removed from the medium and when θ = 0 the nutrient was included ) . Next , we added a constraint for the uptake flux of nutrient i , vki , connecting the MILP objective with the minimal growth constraint of each species k . The aim of this constraint was to guarantee that when i is removed from the medium ( i . e . , θki = 1 ) , the nutrient uptake flux vki was non negative ( note that a non-negative vki constraint means this transporter could not serve as an uptake flux ) and that when nutrient i was included in the medium ( θki = 0 ) , vki could take any value and species k was able to take up and utilize metabolite i . Mathematically , this concept was implemented with the constraint vki+Lθki≥L , where L is a large negative number ( here set to −1000 ) . This constraint then became vki≥0 when θki = 1 , indicating that this transport flux cannot be used for uptake . When θki = 0 , this constraint became vki≥L , allowing this flux to serve as an uptake flux ( e . g . vki = −10≥L ) . The MILP problem was solved with a single optimization procedure considering the minimal growth constraints of the two species simultaneously . This was done using a stoichiometric matrix that included both species and that described their growth independently ( i . e . , no cross-feeding fluxes were included in the model ) . Formally , the MILP problem could then be formulated as a maximization of the sum of decision variables as follows: ( 7 ) Given the solution of this MILP problem , the set of nutrients for which the decision variable θ equals zero in at least one species represented the minimal neutral media . The solution to this MILP problem was determined using the GLPK solver through GLPKmex ( see above ) . Notably , simulating the growth of all pair-wise communities from the first dataset , each on 100 different media , took ∼2 days on a 144 node cluster . The second dataset included roughly 500 times more pair-wise communities and therefore each community was simulated on one medium as described above . For consistency , in analyzing this second dataset , we only considered communities in which both species were still growing at 1 hr . All models and media used in this study are available for download on our website ( http://elbo . gs . washington . edu/download . html ) . In order to determine the functional and phylogenetic distance among pairs of species , KEGG orthology ( KO ) annotations [101] and 16S rRNA gene sequence information were obtained from the Integrated Microbial Genome Database ( IMG ) [102] . Functional distance was calculated as the pairwise Jaccard distance between the KO presence/absence profiles of the two species . Phylogenetic distance was calculated as described in [25] and [103] . Briefly , 16S rRNA sequences were first aligned using NAST [104] . When multiple sequences in a single species passed the NAST filter , a single sequence was chosen at random . Lane mask was applied to this alignment , and percent identity was calculated with Clearcut [105] using the Kimura two-parameter distance correction . | Microbes constantly change their environment , consuming some compounds from their surroundings and secreting others . This microbial activity plays a crucial role in many important environmental cycles , ultimately making all life possible . These processes , however , are often not accomplished by a single species but rather by a diverse community of interacting microorganisms . Characterizing these interactions and their impact is essential not only for understanding global ecosystem metabolism , but also for uncovering the tremendous potential of microbial communities in industrial , environmental , and clinical applications . In this paper , we present a computational framework for modeling , exploring , and tracking such enhanced metabolic capacities in simple two-species communities . We demonstrate that emergent biosynthetic capacity – the ability of multiple species growing together to produce and secrete metabolites that none of the member species secretes when growing alone – is common , and identify typical reprogramming mechanisms and temporal patterns that underlie this capacity . Importantly , we show that emergent capacity is most likely when the species comprising the community are neither too functionally similar nor too distant . Overall , our findings lay the foundation for a comprehensive and predictive understanding of synergistic microbial activity and highlight promising routes for designing , engineering , and manipulating microbial communities toward desired metabolic capabilities . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"systems",
"biology",
"computer",
"and",
"information",
"sciences",
"ecosystem",
"modeling",
"ecology",
"network",
"analysis",
"biology",
"and",
"life",
"sciences",
"metabolic",
"networks",
"computational",
"biology",
"microbial",
"ecology"
] | 2014 | Emergent Biosynthetic Capacity in Simple Microbial Communities |
Neurons are particularly vulnerable to perturbations in endo-lysosomal transport , as several neurological disorders are caused by a primary deficit in this pathway . In this report , we used positional cloning to show that the spontaneously occurring neurological mutation teetering ( tn ) is a single nucleotide substitution in hepatocyte growth factor-regulated tyrosine kinase substrate ( Hgs/Hrs ) , a component of the endosomal sorting complex required for transport ( ESCRT ) . The tn mice exhibit hypokenesis , muscle weakness , reduced muscle size and early perinatal lethality by 5-weeks of age . Although HGS has been suggested to be essential for the sorting of ubiquitinated membrane proteins to the lysosome , there were no alterations in receptor tyrosine kinase levels in the central nervous system , and only a modest decrease in tropomyosin receptor kinase B ( TrkB ) in the sciatic nerves of the tn mice . Instead , loss of HGS resulted in structural alterations at the neuromuscular junction ( NMJ ) , including swellings and ultra-terminal sprouting at motor axon terminals and an increase in the number of endosomes and multivesicular bodies . These structural changes were accompanied by a reduction in spontaneous and evoked release of acetylcholine , indicating a deficit in neurotransmitter release at the NMJ . These deficits in synaptic transmission were associated with elevated levels of ubiquitinated proteins in the synaptosome fraction . In addition to the deficits in neuronal function , mutation of Hgs resulted in both hypermyelinated and dysmyelinated axons in the tn mice , which supports a growing body of evidence that ESCRTs are required for proper myelination of peripheral nerves . Our results indicate that HGS has multiple roles in the nervous system and demonstrate a previously unanticipated requirement for ESCRTs in the maintenance of synaptic transmission .
In neurons , endosomal transport and sorting of internalized cargo affects the abundance of plasma membrane proteins and regulates a diverse group of cellular processes such as signal transduction and synaptic vesicle recycling [1–3] . A group of evolutionarily conserved cytosolic proteins referred to as the endosomal sorting complexes required for transport ( ESCRT ) associate with the endosomal membrane and direct the sorting of internalized cargo [4 , 5] . Four distinct ESCRT complexes , ESCRT-0 , -I , -II and—III , act sequentially to contribute to the sorting of membrane proteins , thus determining if endocytosed receptors are recycled back to the cell surface or are further sorted into intraluminal vesicles ( ILVs ) within specialized endosomes called multivesicular bodies ( MVBs ) for their eventual degradation by lysosomes [6] . Receptor degradation rates affect the duration of receptor signaling , and alterations in the turnover of activated receptors have been implicated in a variety of disease processes [7–10] . The ESCRT-0 complex , composed of HGS and signal transducing adaptor molecule 1 ( STAM1 ) [11–13] , is essential for the initial recognition of ubiquitinated cargo that will be sorted at the endosomal membrane and degraded upon endo-lysosomal fusion [14 , 15] . HGS is believed to serve as a master regulator of endo-lysosomal trafficking by binding ubiquitinated cargo and initiating the recruitment of the ESCRT-I , -II , and—III components [16–19] . Given its role in the early stages of endosomal sorting of internalized cell surface receptors , it has been suggested that HGS also plays an important role in the recycling of cargo to the plasma membrane [20–22] . In addition , HGS can bind synaptosomal associated protein 25 ( SNAP-25 ) and prevent SNARE complex formation , thereby inhibiting neurotransmitter secretion [23] and endosomal fusion [24] . Studies in Drosophila have also implicated HGS in synaptic protein homeostasis and synaptic vesicle rejuvenation [25] . However , despite the insights into HGS function gained from the studies described above , the precise role of HGS in endosomal trafficking has not been determined in the mammalian nervous system . Disruption of endosomal transport is associated with many neurological disorders [26] . In some instances , a direct link has been demonstrated between the genetic defect and endosomal dysfunction . For example , mutations in both ESCRT components and endosomal-associated factors have been attributed to neurodegeneration [26] . Mutations in multivesicular body protein 2B ( CHMP2B ) , a component of the ESCRT-III complex , are linked to both frontotemporal dementia and amyotrophic lateral sclerosis [7 , 27 , 28] . In addition , mutations in kinesin family member 5A ( KIF5A ) and SPARTIN , which facilitate endosomal trafficking , are associated with motor neuron dysfunction in hereditary spastic paraplegia [29–31] . Together , these findings suggest that motor neurons are particularly vulnerable to deficits in endosomal trafficking and that genes in this pathway may be excellent candidate genes for inherited motor neuron diseases . In addition to its association with neurodegeneration in the central nervous system , ESCRT dysfunction is also associated with peripheral demyelinating diseases . Mutations in lipopolysaccharide-induced tumor necrosis factor-alpha factor ( LITAF ) /SIMPLE , that interacts with HGS and STAM1 in the ESCRT-0 complex [32 , 33] , cause a demyelinating form of Charcot Marie Tooth disease [34 , 35] , suggesting that ESCRT-0 is required for myelin formation or stability . By directing activated receptor tyrosine kinases to the lysosome , ESCRT-mediated endosomal sorting provides a mechanism to prevent aberrant signaling that is believed to result in demyelination and degeneration of peripheral axons . In this study , we used a positional cloning approach to demonstrate that the neurological phenotypes observed in the tn mice are due to a point mutation in Hgs that leads to hypomorphic expression of HGS . The subsequent reduction of HGS caused motor and sensory deficits that were accompanied by peripheral nerve hypermyelination and dysmyelination . Further analysis of motor function demonstrated that the tn mice had reduced muscle development and both structural and functional alterations at the NMJ . These deficits in neuromuscular function were associated with the accumulation of ubiquitinated proteins at the synapse . This report represents the first demonstration that a mutation in an ESCRT-0 component causes a motor and sensory neuropathy and suggests that HGS is required for the sorting of ubiquitinated proteins at the synapse to maintain synaptic transmission at the NMJ .
The tn mutation arose spontaneously in a C3H/HeJ inbred mouse line , resulting in a progressive neurodevelopmental disorder that first presents at 3 weeks of age [36] . Homozygous tn mice exhibit reduced growth , ataxia , hypokinesis and premature death at 4 to 5 weeks of age ( Fig 1A–1C ) [36] . These deficits are thought to be due to dysgenesis of the brainstem and spinal cord [36] . To identify the minimal chromosomal region harboring the tn mutation , a congenic mouse line was generated by backcrossing the tn mutation onto the C57BL/6J genetic background . Analysis of single nucleotide polymorphisms localized the tn mutation to a 2 . 8 Mb region on distal chromosome 11 ( Fig 1D ) . We then performed transcriptome analysis of brain RNA to identify nucleotide changes in the coding sequences of the genes located within the tn critical region . Sequence analysis of the 87 genes within the tn critical region only revealed a single nucleotide substitution in the Hgs gene . Genomic sequencing of the Hgs gene from the tn mice confirmed an adenine to guanine transition at position 265 , which resulted in a methionine to valine substitution at amino acid 89 ( M89V ) ( Fig 1E ) . The methionine residue at position 89 of HGS is highly conserved throughout eukaryotic phylogeny ( Fig 1F ) . The M89V substitution is located in the N-terminal Vps27-HGS-STAM ( VHS ) domain of HGS that is involved in substrate binding [37 , 38] . To confirm that this mutation in HGS is responsible for the tn phenotype , a genetic complementation assay was performed by breeding mice that were heterozygous for the tn allele ( Hgstn/+ ) to mice that were heterozygous for a knockout allele of Hgs ( Hgstm1S/+ , henceforth referred to as HgsKO/+ ) [39] . The resulting HgsKO/tn offspring exhibited perinatal lethality by postnatal day 40 ( Fig 1C ) and a neurological phenotype similar to the tn mice , confirming that the tn gene was allelic to Hgs . Since the HgsKO/KO mice die by embryonic day 11 [39] , the enhanced survival of the HgsKO/tn mice indicated that the tn allele was not a null allele of Hgs . To investigate the distribution of Hgs expression in vivo , we performed quantitative polymerase chain reaction ( qPCR ) on reverse-transcribed RNA isolated from 4-week-old Hgs+/+ wild type mice . Hgs was expressed in all tissues examined , with the highest levels of Hgs detected in the brain and spinal cord ( Fig 2A ) . Consistent with the qPCR data , immunoblot analysis also revealed high levels of HGS in the nervous system of wild type mice ( Fig 2B ) . While the tn mutation resulted in a significant reduction of HGS in the nervous system of the Hgstn/tn mice , it differentially affected HGS levels in non-neuronal tissues ( Fig 2B ) . Examination of the heart , liver , spleen , kidney , adrenal gland , thalamus , testis , and ovaries from the Hgstn/tn mice revealed a striking absence of pathology in these tissues , suggesting that HGS provides an essential function that is unique to the nervous system . Because Hgstn/tn mice display a progressive neurological disease starting around 3 weeks of age , it was important to establish a developmental profile of Hgs expression in the brain . We found that Hgs levels undergo a normal down regulation in the Hgs+/+ mice during early postnatal development ( Fig 2C ) . Immunoblot analysis also showed the highest levels of HGS during embryonic development , with abundance decreasing throughout postnatal development ( Fig 2D and 2E ) . The M89V mutation led to reduced expression of HGS at all time points examined ( Fig 2D–2F ) , with the most severe loss of HGS occurring during the time when the neurological phenotypes were most pronounced in the Hgstn/tn mice . In addition to the primary immunoreactive band that migrated at approximately 110 kDa , a second higher molecular weight band migrating at approximately 120 kDa , which may represent a post-translationally modified form of HGS [40–42] , was also detected early in development . A previous study of a neuronal-specific knockout of Hgs demonstrated multiple deficits in the hippocampus that included increased ubiquitin staining of CA3 pyramidal cells , increased CA3 pyramidal cell death , and reduced CA3 pyramidal cell numbers [43] . To determine whether the Hgstn/tn mice exhibited any alterations in the number of CA3 pyramidal cells or an increase in cell death , we measured the number of CA3 cells and looked for evidence of increased cell death by performing Nissl and activated caspase-3 staining , respectively , in the hippocampi of wild type and Hgstn/tn mice . Unlike what was reported for the neuronal-specific Hgs knockout mice , we did not detect any change in the number of CA3 pyramidal cells or an increase in cell death markers in the Hgstn/tn hippocampus ( Fig 3A ) . In addition , the levels of hippocampal myelin basic protein were similar in the Hgstn/tn and Hgs+/+ mice . Moreover , while increased glial fibrillary acid protein ( GFAP ) abundance is often associated with neurodegeneration , there was no increase in GFAP immunoreactivity in the hippocampus of the Hgstn/tn mice compared to wild type controls ( Fig 3A ) . Similar to what we observed in total brain extracts , there was a significant reduction in both HGS and STAM1 levels in the hippocampus of the Hgstn/tn mice . Since mutations in CHMP2B are associated with neurodegeneration in humans , we examined the effect of loss of HGS on CHMP2B levels and found no detectable differences in the hippocampus between the Hgstn/tn or wild type controls . Examination of the levels of the receptor tyrosine kinases TrkA and TrkB , which are putative substrates for HGS , in hippocampal extracts from the Hgstn/tn and Hgs+/+ mice also did not reveal any significant differences in expression ( Fig 3B and 3C ) . As a component of the endosome , HGS has been implicated in the maturation of autophagosomes , as the loss of HGS expression results in increased levels of markers of autophagy in mammalian cell lines [44] . To investigate if HGS is required for autophagy in neurons , we examined the expression of the autophagosome marker microtubule-associated protein 1A/1B-light chain 3 ( LC3 ) and the autophagy substrate p62 in hippocampal lysates from wild type and Hgstn/tn mice . In contrast to the increased levels of LC3 and p62 that are observed when autophagy is impaired [44] , we detected a reduction in the levels of both LC3 and p62 in the hippocampus of the Hgstn/tn mice as compared to controls ( Fig 3B and 3C ) . Together , these data indicate that loss of HGS does not appear to result in increased cell death or a significant impairment of autophagy in the hippocampus of the Hgstn/tn mice . Several neurological diseases are caused by mutations that lie within genes involved in the endosomal sorting of membrane proteins [45–48] . To examine whether HGS deficiency resulted in motor and sensory deficits commonly seen in patients with inherited neuropathies , we performed a series of behavioral assays on our mutant Hgs mouse lines . By 3 weeks of age , the Hgstn/tn mice exhibited significant motor deficits as demonstrated by the presence of clawed paws , decreased locomotion in an open field assay , impaired rotarod performance and an increased time to transverse an elevated beam compared to controls ( Fig 4A–4D ) . In addition , the Hgstn/tn mice also demonstrated increased tactile sensitivity that was consistent with a heightened pain response when examined using the von Frey assay ( Fig 4E ) and had reduced forelimb muscle strength as compared to age-matched controls ( Fig 4F ) . The initial description of the tn mice indicated that the heterozygous tn mice displayed motor abnormalities in a prolonged swimming assay [36] . We therefore investigated whether heterozygous Hgstn/+ and HgsKO/+ mice also exhibited motor and sensory abnormalities in our assays . Both the HgsKO/+ mice and the Hgstn/+ mice displayed increased pain sensitivity and reduced muscle strength ( Fig 4E and 4F ) , indicating that the tn mutation is a loss-of-function mutation and that a 50% loss of HGS is sufficient to cause peripheral nervous system dysfunction . Alterations in axon number or peripheral nerve myelination are common features of inherited peripheral neuropathies and are thought to be important contributors to motor and sensory dysfunction . To examine whether loss of HGS expression results in axonal loss , demyelination , or dysmyelination that could contribute to the behavioral deficits in the Hgstn/tn mice , we compared the sciatic nerves of 4-week-old Hgs+/+ and Hgstn/tn mice by transmission electron microscopy . Although no change in the density of myelinated and unmyelinated axons was found in the Hgstn/tn mice ( Fig 5A and 5B ) , there was a significant increase in the diameter of myelinated axons and a significant decrease in the diameter of unmyelinated axons in the 4-week-old Hgstn/tn mice as compared to controls ( Fig 5C ) . This increase in the average axonal diameter of myelinated fibers was attributed to a shift in distribution towards a greater number of large diameter axons in the Hgstn/tn mice ( Fig 5D ) . Morphometric analysis of myelin structure also revealed a small but significant decrease in the G-ratio ( the ratio of the inner axonal diameter to the total fiber diameter ) of sciatic nerves from the Hgstn/tn mice ( Fig 5E ) . This decrease was attributed to a 22% increase in the myelin thickness of small-diameter axons that are between 1 . 0 to 2 . 0 μm in diameter ( Fig 5F , circled region ) . Several alterations in myelin structure were also observed in the sciatic nerves of the Hgstn/tn mice ( Fig 5G ) . These alterations included the presence of tomacula , a prominent thickening of compact myelin with redundant loops , which likely contributed to the decreased G-ratio observed in the Hgstn/tn mice . Additionally , many regions of the myelin sheaths were disorganized in the Hgstn/tn mice ( Fig 5G1–5G2 ) and contained structural alterations such as myelin infoldings ( Fig 5G3–5G5 ) as compared to controls ( Fig 5G6 ) . The changes in myelination that were observed in the Hgstn/tn mice could be attributed to loss of HGS in either axons or Schwann cells . However , when we examined HGS expression in the sciatic nerves of wild type mice by indirect immunofluorescence ( Fig 6 ) , we found that HGS staining did not overlap with the neurofilament staining in sciatic nerve axons . Rather , when Schwann cells were visualized with antibodies to the cytoplasmic protein S100β , HGS localized to the external boundary of the Schwann cell body . These findings are in agreement with a previous study detecting HGS transcripts in Schwann cells [49] . In cell culture models , the stability of STAM1 appears to be dependent upon HGS expression in a transcript-independent manner [16 , 50] . However , this relationship between ESCRT-0 components has not been investigated in the nervous system . We therefore examined the effect of the tn mutation on the expression of ESCRT components and their putative substrates in the sciatic nerves of the Hgstn/tn mice . While the HGS levels were reduced by 50% in the sciatic nerves of 4-week-old Hgstn/tn mice compared to controls , there were no significant differences in the levels of STAM1 , tumor susceptibility protein 101 ( TSG101 ) or epidermal growth factor receptor substrate 15 ( EPS15 ) in the sciatic nerve extracts from the Hgstn/tn and Hgs+/+ mice ( Fig 7A–7C ) . Analysis of receptor tyrosine kinases ( RTKs ) sorted by the ESCRT pathway showed that loss of HGS resulted in a significant reduction in both the full length ( TrkB . FL ) and truncated ( TrkB . T1 ) isoforms of TrkB in the sciatic nerves of the Hgstn/tn mice that was not associated with a corresponding reduction in TrkB mRNA levels ( Fig 7D–7F ) . In contrast , reduction of HGS did not affect the level of either the epidermal-growth factor receptor ( EGFR ) or the receptor tyrosine-protein kinase erbB-2 ( ERBB2 ) ( Fig 7D and 7E ) . Since loss of HGS resulted in a dramatic motor phenotype in the Hgstn/tn mice , but had only a modest effect on peripheral nerve myelination , we examined the effect of reduced HGS expression on the abundance of ESCRT-0 components in the spinal cords of 4-week-old mice ( Fig 8A and 8B ) . Despite the significant increase in Hgs transcript levels in the Hgstn/tn mice ( Fig 8C ) , HGS levels were reduced 70% in the spinal cords of the Hgstn/tn mice compared to controls ( Fig 8A and 8B ) . Similarly , we also observed an 80% reduction in the levels of STAM1 in the spinal cords of the Hgstn/tn mice with no corresponding decrease in Stam1 transcript abundance ( Fig 8A–8C ) . When we examined whether the loss of HGS affected the abundance of other proteins reported to interact with HGS [51–53] , there was no difference in the levels of EPS15 or TSG101 in the spinal cords of the Hgstn/tn and Hgs+/+ mice ( Fig 8A and 8B ) . Although HGS appears to influence the abundance of some receptor tyrosine kinases in immortalized cell lines [22 , 53–58] and TrkB in the sciatic nerves ( Fig 7 ) , the levels of EGFR , TrkA and TrkB found in the spinal cords of the Hgstn/tn mice were similar to those observed in the Hgs+/+ control mice ( Fig 8A and 8B ) . To determine if the movement disorder in the Hgstn/tn mice was associated with a loss of motor neurons , we compared motor neuron numbers in the ventral horn of 4- to 5-week-old Hgstn/tn and Hgs+/+ mice . Consistent with the axonal density studies on the sciatic nerves , there was no significant difference in the number of motor neurons ( Fig 8D ) in the lumbar 4/5 region between the Hgstn/tn and Hgs+/+ mice . However , loss of HGS in the Hgstn/tn mice did result in increased GFAP staining in the lumbar 4/5 region of the spinal cord ( Fig 8E ) . The Hgstn/tn mice exhibit several signs of neuromuscular disease , including muscle weakness and decreased motor performance . When we compared the muscle mass between the Hgstn/tn mice and the Hgs+/+ controls , we detected a significant difference in gastrocnemius weights at 4 weeks of age ( Fig 9A and 9B ) . While we did not detect angular muscle fibers or centrally located nuclei in the muscle sections of the Hgstn/tn mice , which are common features of motor neuron disease , there was a 32% reduction in muscle fiber size in the Hgstn/tn mice ( Fig 9C ) . These changes were consistent with decreased motor neuron input onto the muscle fibers . Since acetylcholine receptor ( AChR ) abundance is inversely correlated with motor neuron input onto muscle fibers , we performed qPCR on gastrocnemius muscle RNA isolated from 4-week-old wild type and Hgstn/tn mice . Both the embryonic and adult acetylcholine receptor subunit mRNAs were significantly increased in the gastrocnemius muscles of the Hgstn/tn mice ( Fig 9D ) . Alterations in motor endplate structure are found in several models of neuromuscular disease [59–62] . Using mice that express yellow fluorescent protein ( YFP ) in the motor axons to visualize motor neuron terminals from tibialis anterior ( TA ) muscles , we found that every postsynaptic AChR cluster in the wild type mice was innervated by a motor neuron axon ( Fig 9E ) . Although all of the AChR clusters from the 4-week-old Hgstn/tn mice were also innervated , we observed a significant increase in the number of terminals that exhibited either terminal swellings or sproutings ( Fig 9F ) , which are phenotypes that are often observed in mice with motor neuron disease [59 , 62 , 63] . In addition , there was a significant increase in endplate size in the Hgstn/tn mice ( Fig 9G ) , which resembles the increase in endplate area observed in mice lacking the vesicular acetylcholine transporter [61] . These results are consistent with the idea that reduced HGS expression results in decreased neurotransmitter release onto Hgstn/tn muscles . By sorting internalized membrane proteins from the endosome to the lysosome , the ESCRT complexes are thought to regulate the turnover of internalized cargo [26] . To investigate whether loss of HGS altered endosome or MVB abundance at motor neuron terminals , we compared the NMJs from wild type and Hgstn/tn TA muscles by electron microscopy ( Fig 10A ) . Loss of HGS expression resulted in an increase in the number of endosomal-like structures in the Hgstn/tn mice ( Fig 10A and 10B ) . While we did not observe any MVBs in the electron micrographs taken from wild type NMJs , we observed several MVBs ( 0 . 23 MVBs/μm2 ) in the motor neuron terminals of the Hgstn/tn mice . No significant difference was observed in the number of autophagosomes in the NMJs of the Hgstn/tn and Hgs+/+mice . HGS is thought to have a role in regulating exocytosis through its interactions with components of presynaptic nerve terminals [23 , 52] . To test whether HGS is required to maintain neurotransmitter release , we used two-electrode voltage-clamp analysis to study the release of acetylcholine at the NMJs of 3-week-old Hgstn/tn and Hgs+/+ mice and observed a pronounced effect of HGS reduction on neurotransmitter release in the TA muscles of the Hgstn/tn mice ( Fig 10C ) . The amplitude of evoked neurotransmitter release , measured as endplate currents ( EPCs ) and obtained by stimulating sciatic nerves , was decreased by 50% in the Hgstn/tn mice as compared to controls ( Fig 10C ) . The EPC amplitude is determined by both the number of synaptic vesicles released after nerve stimulation ( quantal content ) and the amplitude of the muscle response to the neurotransmitter released from a single vesicle ( quantal amplitude ) [60 , 64] . The amplitude of miniature endplate currents ( MEPCs ) , representing a response to a single quanta or vesicle , was also reduced in the Hgstn/tn mice compared to controls ( Fig 10D ) . This finding suggests that there was either a decrease in the amount of neurotransmitter per vesicle or that there was a reduction in the sensitivity or abundance of post-synaptic receptors in the Hgstn/tn mice . However , the increase in AChR expression observed in the Hgstn/tn mice compared to controls , and the similarites between the phenotypes of the Hgstn/tn mice and mice lacking the vesicular ACh transporter , both suggest that there was a reduction in the level of acetylcholine found in the synaptic vesicles of the Hgstn/tn mice . The finding of decreased MEPC amplitudes in the Hgstn/tn mice also suggests that HGS may be required post-synaptically for stable AChR expression . In addition , a reduction in quantal content ( the number of vesicles released by evoked stimulation ) , which is calculated by dividing EPC amplitude by MEPC amplitude , was also observed in the Hgstn/tn mice compared to controls ( Fig 10E ) . To further examine the effect of HGS loss on presynaptic function , we measured the frequency of spontaneous neurotransmitter release and determined that there was a 66% decrease in MEPC frequency in the Hgstn/tn mice compared to controls ( Fig 10F ) . These findings indicate that loss of HGS severely affects synaptic transmission by reducing acetylcholine release at the NMJ and that these changes in presynaptic function likely contribute to the motor deficits seen in the Hgstn/tn mice . HGS is involved in the sorting of ubiquitinated proteins , and our studies indicate that loss of HGS has a profound effect on synaptic transmission at the NMJ . To investigate whether HGS is required for the sorting of ubiquitinated proteins at the synapse , we examined the level of ubiquitinated proteins in both the total and synaptosomal fractions prepared from cerebral cortices of wild type and Hgstn/tn mice . While the level of ubiquitin conjugates in the spinal cord , sciatic nerve and cortex were similar between the Hgstn/tn mice and controls , we observed a 2-fold increase in the level of ubiquitinated proteins isolated from the cortical synaptosomes of the Hgstn/tn mice ( Fig 11A and 11B ) . This compartment-specific effect of HGS reduction on the levels of ubiquitin conjugates suggests that HGS is involved in the sorting of ubiquitinated proteins at the synapse and that proper endosomal sorting at the NMJ is required to maintain synaptic transmission .
These studies demonstrate that the neurological deficits observed in the tn mice are caused by a hypomorphic mutation in the ESCRT-0 component Hgs which significantly reduced HGS levels in the central nervous system . By 4 weeks of age , the Hgstn/tn mice show significant defects in motor and sensory function as well as reduced muscle development , increased muscle AChR expression , terminal sproutings and swellings of motor axons , and increased numbers of endosomes and MVBs at the motor neuron axon terminals . In addition to these structural changes in the motor axon terminals , we also observed a significant reduction in both spontaneous and evoked vesicular release of acetylcholine at the NMJs of the Hgstn/tn mice , suggesting that HGS-dependent sorting of proteins at the endosome is required to maintain synaptic transmission at the NMJ . Analysis of the level of ubiquitinated proteins revealed a significant increase in ubiquitin conjugates specifically in the synaptosomal fraction of the Hgstn/tn mice as compared to controls . While the numbers of myelinated and unmyelinated peripheral nerves in the Hgstn/tn mice were similar to the numbers observed in wild type mice , the reduced HGS expression resulted in an increase in myelin thickness and dysmyelination of the sciatic nerve axons in the Hgstn/tn mice . These results demonstrate that HGS is required to support both the myelination of peripheral nerves and synaptic transmission at the NMJ . HGS regulates the sorting and trafficking of ubiquitinated , internalized receptors on the endosome [42 , 65 , 66] . Our finding that the M89V mutation in the VHS domain had variable effects on HGS expression in the different non-neuronal tissues examined suggests that HGS stability is dictated , at least in part , by its interaction with specific binding partners [54] . Alternatively , a variable demand for the ESCRT pathway may also be responsible for the differential effects of the M89V mutation on HGS expression in the various Hgstn/tn tissues . There are several similarities between the myelination defects detected in the Hgstn/tn mice and those observed in the brain-derived neurotrophic factor ( BDNF ) -overexpressing mice [67] . Although we found no evidence for altered BDNF levels in the Hgstn/tn mice , there was a significant reduction in TrkB in the sciatic nerves of Hgstn/tn mice as compared to controls . However , there was no difference in the levels of TrkB mRNA detected in the Hgstn/tn mice and controls , suggesting that HGS promotes TrkB stability by facilitating recycling to the plasma membrane and preventing trafficking to the lysosome . This conclusion is consistent with previous reports that HGS can regulate the recycling of TrkB . FL in neuronal culture [55] . It is possible that the limited effects of loss of HGS on TrkB expression may reflect a partial loss of HGS in the Schwann cells of the Hgstn/tn mice . Our finding that STAM1 expression is not altered in Hgstn/tn sciatic nerve extracts , but is reduced in spinal cord and brain extracts , may indicate that STAM1 associates with other proteins in Schwann cells to stabilize its expression or that the level of HGS expression was not sufficiently reduced in Schwann cells to destabilize STAM1 . In addition , since HGS interacts with the NF2 gene MERLIN in Schwann cells [68 , 69] , loss of this complex may contribute to the changes observed in the levels of TrkB . Since the observed changes in myelin thickness are very subtle , they are unlikely to be a major contributor to the neurological deficits observed in the tn mice . Although the Hgstn/tn mice represent the first report linking ESCRT-0 to neuromuscular disease , mutations in the ESCRT-III component CHMP2B have been linked to amyotrophic lateral sclerosis [27 , 28] , and patients with CHMP2B mutations exhibit phenotypes consistent with lower motor neuron disease [28] . The Hgstn/tn mice also show signs of dysfunction of the lower motor neurons , including motor incoordination , muscle weakness , muscle atrophy , hypokinesis , and reduced synaptic transmission at the NMJ . These deficits are not associated with loss of motor neuron cell bodies or axons in the Hgstn/tn mice , or with denervation of the NMJ . In immortalized cell lines , expression of mutant CHMP2B resulted in the formation of dysmorphic endosomes [7] , increased levels of autophagy markers , and the formation of ubiquitin-positive inclusions [28] , suggesting that the motor neuron damage results from a block in autophagic clearance of ubiquitinated proteins . Although data from mammalian cell lines suggests that HGS is required for the maturation of autophagic vesicles , as depletion of HGS resulted in increased LC3 levels and decreased numbers of LC3/lysosome-associated membrane glycoprotein 1 positive structures in cell culture [44] , loss of HGS resulted in reduced levels of the autophagy markers LC3 and p62 in the hippocampus of the Hgstn/tn mice , indicating that loss of HGS does not appear to result in a blockade in autophagy in neurons . Our studies demonstrate that HGS is required for synaptic transmission at the NMJ . The reductions in MEPC and EPC amplitudes and quantal content are consistent with reduced vesicle number and/or release at the NMJ of the Hgstn/tn mice . These findings are also consistent with a previous report that showed that mutation of the neuronal Rab35 GTPase activating protein skywalker in Drosophila leads to an increase in the readily releasable pool of vesicles , enhanced neurotransmitter release and increased HGS-dependent protein sorting at the NMJ [25] , suggesting that the ESCRT machinery is required for synaptic vesicle rejuvenation and removing defective synaptic vesicle proteins at the NMJ . The increased levels of ubiquitinated synaptic proteins detected in the Hgstn/tn mice suggests that an ESCRT-dependent mechanism for clearing ubiquitinated proteins at synapses is also important to maintaining synaptic transmission in mammals . Alternatively , the increase in ubiquitinated conjugates may be an indirect effect caused by disruption of the endosomal sorting pathway . HGS is known to interact with synaptic vesicle proteins and could therefore directly affect vesicle trafficking [52 , 56] . For example , by binding to SNAP25 [23 , 51] , HGS can displace vesicle-associated membrane protein 2 and inhibit SNARE complex formation , thereby blocking neurotransmitter secretion and endosomal fusion [23 , 70] . While the presence of enlarged vesicles would be the predicted effect of HGS loss , we did not detect an increase in the size of the endosomes at motor neuron axon terminals in the Hgstn/tn mice , but instead found an increase in the number of endosomes and MVBs at the NMJs of the Hgstn/tn mice . Previous reports from a proteomic screen identified mammalian uncoordinated-18 , a component of the synaptic vesicle fusion protein complex , as a protein that could interact with the ubiquitin-interacting motif of HGS [71] . HGS has also been shown to inhibit the fusion of vesicles with the early endosome [24] . Together , these results provide evidence that HGS interacts with components of the synaptic vesicle release machinery and plays a direct role in synaptic transmission . Given that loss of HGS results in a predominantly neurological phenotype in the Hgstn/tn mice , it is surprising that a neuronal-specific deletion of HGS [43] leads to such a mild phenotype compared to what we observed in our studies . Whereas Hgstn/tn mice die by 5 weeks of age and present with ataxia , decreased muscle size , and reduced strength , the neuronal-specific deletion of HGS did not affect gait or grip strength , and there was no reported effect of HGS loss on viability [43] . However , it is possible that the conditional knockout allele used by Tamai et al . [43] , which utilized a rat synapsin Cre driver [72] , may not have led to a reduction of HGS in the motor neurons , or that the level of HGS was not reduced as much as what is observed in the Hgstn/tn mice , thereby resulting in a milder phenotype . Identification of the tn mutation in HGS has provided a new model for investigating the role of the ESCRT pathway in the nervous system . Our identification of motor and sensory phenotypes in both the Hgstn/+ and HgsKO/+ mice demonstrates that loss of HGS results in an autosomal dominant inheritance pattern similar to that observed in human cases of amyotrophic lateral sclerosis and Charcot-Marie-Tooth disease [27 , 28 , 35] . Hereditary neuralgic amyotrophy , an autosomal dominant form of recurrent focal neuropathy , is the only neurological disorder that has been mapped near HGS on chromosome 17q25 [73] . Mutations in Septin 9 have been found in many , but not all , cases of hereditary neuralgic amyotrophy , making HGS a good candidate gene for other cases of this disease . With our demonstration that HGS is required for motor neuron function , it will now be possible to determine if HGS is involved in other neurological disorders in humans .
The tn mutation spontaneously arose on the C3H/HeJ inbred mouse strain at The Jackson Laboratory in 1959 . We generated a congenic line of B6 . C3Hgstn mice by backcrossing the tn mutation onto the C57BL/6J background for at least 12 generations . The tn mice were maintained by intercrossing tn sibling heterozygotes ( Hgstn/+ ) . Hgs knockout heterozygotes ( Hgstm1Sor/+ , stock number 003539 ) were obtained from The Jackson Laboratory , and sibling pairs were intercrossed to maintain the strain . Research was conducted using equal numbers of male and female mice . Wild type littermates or age-matched C57BL/6J mice ( Hgs+/+ ) were used as controls . All mice were maintained in our breeding colony at the University of Alabama at Birmingham , which is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International . All research was approved by the UAB IACUC committee and complied with the United States Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals , and adhered to principles stated in the Guide for the Care and Use of Laboratory Animals , United States National Research Council . The tn mutation was previously mapped to distal chromosome 11 [74] . Small nucleotide polymorphism ( SNP ) analysis localized the tn mutation to a 2 . 8 Mb region on the distal end of chromosome 11 flanked by rs27043138 and rs3675603 . Total RNA was isolated from brain lysates of 4-week-old Hgstn/tn and Hgs+/+ mice using RNA-STAT60 ( Tel-Test , Friendswood , TX ) and subsequently purified using an RNeasy Mini Kit ( Qiagen Sciences , Valencia , CA ) . RNA-seq experiments were carried out at the Hudson Alpha Genome Services Laboratory ( Huntsville , Alabama ) . Two μg of total RNA underwent quality control ( Bioanalyzer; all RIN values > 9 . 5 ) , and was prepared for directional RNA sequencing at Hudson Alpha using NEBNext reagents ( New England Biolabs ) according to manufacturer’s recommendations with minor modifications ( including the use of custom library adapters and indexes ) . RNA libraries were quantified with the Kapa Library Quant Kit ( Kapa Biosystems ) , and underwent sequencing ( 50 bp paired-end directional reads; ~70M reads/sample ) on an Illumina sequencing platform ( HiSeq2000 ) . Exon 4 of Hgs was amplified from genomic DNA of Hgstn/tn and Hgs+/+ mice . The PCR products were sequenced by The Genomics Core Facility of the Heflin Center for Genomic Science ( Birmingham , AL ) . Total RNA was isolated from specified tissues using RNA-STAT60 and reverse transcribed using the Superscript VILO cDNA synthesis kit ( Life Technologies ) . Individual gene assays were purchased from Applied Biosystems for each of the RNAs analyzed . ΔΔCt values were generated using Hgs ( Mm00468635_m1 ) , Stam ( Mm00488457_m1 ) , AChRα ( Mm00431629_m1 ) , AChRβ ( Mm00680412_m1 ) , AChRδ ( Mm00445545_m1 ) , AChrδ ( Mm00437419_m1 ) , AChRε ( Mm00437411_m1 ) , Bdnf ( Mm04230607_s1 ) and TrkB Mm00435422_m1 ) . Taqman gene assays with 18S ( Mm03928990_g1 ) and β-actin ( Mm00607939_s1 ) served as internal standards . qPCR results are shown as the average of three different amplifications of cDNAs that were generated from at least three different mice . Unpaired Student’s t tests were conducted on ΔΔCt values from each genotype to determine their significance . Mice of appropriate age and genotype were asphyxiated with CO2 or deeply anesthetized via isoflurane prior to rapid decapitation . Tissues were removed and homogenized in a modified RIPA buffer containing 50 mM Tris , pH 7 . 5 , 150 mM NaCl , 5 mM MgCl2 , 0 . 5 mM EGTA , 1 mM EDTA , 0 . 5% SDS , 1% Triton X-100 , and 1% sodium deoxycholate . Complete protease inhibitors ( Roche , Indianapolis , IN ) , phosphatase inhibitor cocktail I ( Sigma Aldrich , St . Louis , MO ) , and 50 μM PR-619 ( Life Sensors , Malvern , PA ) were added to the homogenization buffer . Samples were homogenized in 1X Laemmli buffer , sonicated , and boiled . After homogenization , tissues were centrifuged at 17 , 000 x g for 10 min at 4°C , and supernatants were removed and immediately frozen at—80°C . Protein concentrations were determined by using the bicinchoninic acid ( BCA ) protein assay kit from Pierce ( Rockford , IL ) . Proteins were resolved on either 10% Tris-glycine gels or 4–20% Novex Tris-glycine gels ( Life Technologies ) and transferred onto nitrocellulose or PVDF membranes . Immunoblots were probed for HGS , STAM1 , TSG101 , EPS15 , EGFR ( Santa Cruz , Dallas , TX ) , TrkA and TrkB ( Millipore , Billerica , MA ) , CHMP2B ( Abcam , Cambridge , MA ) , ERBB2 , p62 , LC3 ( 4108 ) and cleaved caspase 3 ( Cell Signaling , Danvers , MA ) , mylein basic protein ( Biolegend , Dedham , MA ) , GFAP ( Dako , Carpinteria , CA ) and ubiquitin ( UAB Hybridoma Facility , Birmingham , AL ) . β-tubulin and β-actin ( Developmental Hybridoma Bank , Iowa City , IA ) were used for loading controls . Primary antibodies were diluted in 1X phosphate buffered saline containing 0 . 1% NP-40 and either 2% BSA or 1% non-fat dry milk , and proteins were detected by using an anti-mouse or anti-rabbit HRP-conjugated secondary antibody ( Southern Biotechnology Associates , Birmingham , AL ) and SuperSignal West Pico Chemiluminescent Substrate ( Thermo Scientific , Rockford , IL ) . Blots were scanned using a Hewlett-Packard Scanjet 3970 ( Palo Alto , CA ) and quantitated using ImageJ ( NIH , Bethesda , MD ) or UN-SCAN-IT software ( Orem , Utah ) . Each value represents the mean and standard error from at least two blots using at least three different animals per genotype . Unpaired t-tests , corrected for multiple comparisons using the Holm- Šídák method , were utilized to determine significant effects between genotypes . One-way ANOVAs with the Geisser-Greenhouse correction were used to analyze the developmental expression patterns of HGS . Gastrocnemius muscles were collected from 4-week-old Hgs+/+ and Hgstn/tn mice . Muscle and body weights were determined for at least six animals per genotype , and values are reported as the average muscle or body mass ± SE . Brain sections were prepared and stained as previously described [75] . Sciatic nerves were dissected from 4-week-old wild type mice and immediately submersed in ice cold 4% PFA for 1 h . Sciatic nerves were cryopreserved in PBS containing 30% sucrose overnight at 4°C . Nerves were embedded in OCT ( Tissue-Tek; Sakura Finetek USA ) and 20 μM sections were cut for analysis . Sections were blocked in PBS containing 10% normal goat serum for 1 h at room temperature . Primary antibodies ( HGS , S100β and Neurofilament ) were diluted 1:200 in PBS containing 10% normal goat serum and incubated with sections overnight at 4°C . Sections were washed two times with PBS at room temperature and then incubated with secondary antibodies labeled with Alexa Fluor 488 or 568 dye ( Invitrogen ) . Confocal imaging was performed on a Nikon C2 laser scanning confocal microscope using the NIS element advance imaging software package version 4 . 2 ( Nikon , Melville , NY ) . Images were captured at 40x using a 0 . 3 μM z-step . Motor and sensory performance was assayed at 3 to 4 weeks of age with Hgs+/+ , Hgstn/tn , Hgstn/+ , and HgsKO/+ mice ( n ≥ 4 ) . Before each behavioral assay , animals were habituated to the testing room for 30 min . Unpaired Student’s t-tests were performed on open field , grip strength and von Frey data . Two-way ANOVA was utilized to determine significance between genotypes on rotarod and elevated beam assays . Unpaired Student’s t-tests were utilized to determine the significance of each trial . Animals were handled at least three days prior to open field testing . Locomotor activity was measured in an open-field arena ( 43 . 2 cm x 43 . 2 cm x 30 . 5 cm ) for 15 min by an automatic video tracking system ( Med Associated , St . Albans , VT ) . The first 5 min were not analyzed to account for habituation to the open field chamber . Motor coordination and balance were tested by placing mice on an accelerating rotarod ( ENV-575 , Med Associates ) and recording latency to fall . The rotarod initially started rotating at 3 . 5 rpm and accelerated to 35 rpm over a 5 min period . Each mouse performed 3 trials separated by 1 h . Motor coordination and proprioception were tested by the elevated beam assay as previously described [76] . Four trials were performed with each trial consisting of 3 repetitions of the assay . The San Diego Instrument Grip Strength System ( San Diego , CA ) was used to assay mouse grip strength . The maximum amount of force generated from forelimbs was recorded . Each trial consisted of 12 repetitions of the assay with the two highest and two lowest data points dropped from final analysis . Animals were habituated to an open gridded floor chamber for 5 min . A series of 10 von Frey fibers varying from 0 . 4 g to 60 g of force ( Ugo Basile , Comerio , Italy ) was applied from below the wire mesh chamber in ascending order beginning with the smallest fiber . The fiber was applied to the central region of the plantar surface to avoid the foot pads . The hind paw withdrawal threshold was determined by Dixon’s formula . TA muscle tissues were fixed in 4% paraformaldehyde in 0 . 1 M sodium cacodylate buffer for 1 h . Fixed muscle fibers were stained with α-bungarotoxin and dissected to isolate regions of muscles containing NMJs . Muscles were washed in 0 . 1 M sodium cacodylate buffer and then post-fixed in 1% osmium tetroxide for 1 h in the dark . After rinsing with sodium cacodylate , the samples were dehydrated in increasing concentrations of acetone ( 50% for 10 min , 75% for 10 min , 90% for 10 min , 95% for 10 min , 100% for 4 x 10 min ) . Muscles were transitioned to epoxy resin ( Electron Microscopy Sciences , Hatfield , IL ) by rotating the samples overnight in a 50:50 solution of epoxy:acetone . The samples were then immersed in fresh 100% resin several times throughout the next day and baked at 65°C overnight . Ultra-thin cross-sections were collected using a Leica EM-UC6 ultramicrosome ( Buffalo Grove , IL ) and stained for contrast with uranyl acetate and lead citrate . Samples were viewed using an FEI Tecnai T-12 electron microscope ( Delmont , PA ) with a Hamamatsu digital camera ( Bridgewater , N . J ) . For quantitation of motor neuron numbers , transverse frozen sections of lumbar spinal cords ( L4 /L5 ) from 7-week-old wt and axJ mice were prepared and stained with cresyl violet as previously described [77] . Ultra-thin sections of muscle tissues were scanned for the presence of NMJs . Micrographs were taken at 4400-6500x magnification . A total of 14 synapses were analyzed from Hgs+/+ mice ( n = 3 ) , and 27 synapses from Hgstn/tn mice ( n = 3 ) were quantified . Endosomes were defined as single membrane intracellular vesicles that were at least 100 nm in feret diameter , and MVBs were defined as endosomes with at least 2 ILVs within the lumen . Endosomes and MVBs were quantified and normalized to the size of the synapse . Sciatic nerves were excised from 4-week-old Hgs+/+ and Hgstn/tn mice and immersed in 6% gluteraldehyde/2% paraformaldehyde in 0 . 1 M sodium cacodylate buffer for at least 1 . 5 h at room temperature . Samples were processed as described above for NMJ analysis . Photomicrographs for morphometric analysis were obtained by systematically covering adjacent but non-overlapping fields to measure axon and myelin diameter , myelin thickness , and myelin defects . The feret diameter was calculated to determine axon and myelin caliper . The G-ratio was also calculated by dividing the axon diameter by the myelin diameter . Quantitation was performed using ImageJ software . A total of 595 myelinated axons from Hgs+/+ mice ( n = 3 ) and 574 axons from Hgstn/tn mice ( n = 3 ) were quantified . Micrographs taken at 240x magnification were utilized to calculate axon density of myelinated fibers , while micrographs taken at 1100x were used to quantitate the density of unmyelinated fibers . All experiments were carried out in vitro at room temperature on the sciatic nerve and TA muscles from 3- to 4-week-old mice . For dissection , mice were deeply anesthetized with isofluorane . The TA muscle was dissected free from the extensor digitorum longus ( EDL ) muscle , partially bisected , and folded apart to flatten the muscle before pinning it down . Because the sciatic nerve innervates both EDL and TA muscles , we transected the EDL and posterior portion of the TA muscle to eliminate excessive muscle contraction . The TA muscle was perfused with Tyrode's solution ( 137 mM NaCl , 2 . 8 mM KCl , 1 . 8 mM CaCl2 , 1 . 1 mM MgCl2 , 11 . 9 mM NaHCO3 , 0 . 33 mM NaH2PO4 , and 11 . 2 mM dextrose , pH 7 . 4 , when bubbled with a mixture of 95% O2 and 5% CO2 ) at 30°C . Intracellular potentials and currents were measured using an Axoclamp-2A amplifier system and glass microelectrodes filled with 3 M KCl ( resistance , 10–15 MΩ ) . Using the two-microelectrode voltage-clamp system , synaptic currents , including miniature endplate currents ( MEPCs ) and endplate currents ( EPCs ) , were obtained at -60 mV holding potential . The currents were digitized at 50 μs per point and were stored , captured , and analyzed using pClamp 9 . 0 software ( Molecular Devices ) . EPCs were elicited by stimulating the corresponding sciatic nerve with rectangular pulses of 0 . 05 ms duration . During single shock stimulation , the nerve was stimulated at 0 . 2 Hz . The amplitude decay and time to peak of the EPCs and MEPCs were captured and subsequently analyzed with pClamp 9 . 2 software and the Mini Analysis Program ( Synaptosoft ) . Quantal content was determined by dividing MEPC amplitude with EPC amplitude from the same endplate [78–80] . Synaptosomes were prepared from 4-week-old cerebral cortices of wild type and tn mice as described [81] . | Endocytic trafficking involves the internalization , endosomal sorting and lysosomal degradation of cell surface cargo . Many factors involved in endosomal sorting in mammalian cells have been identified , and mutations in these components are associated with a variety of neurological disorders . While the function of endosomal sorting components has been intensely studied in immortalized cell lines , it is not known what role these factors play in endosomal sorting in the nervous system . In this study , we show that the teetering ( tn ) gene encodes the hepatocyte growth factor regulated tyrosine kinase substrate ( Hgs ) , a core component of the endosomal sorting pathway . The tn mice exhibit several signs of motor neuron disease , including reduced muscle mass , muscle weakness and motor abnormalities . Although HGS is predicted to be required for the lysosomal degradation of receptor tyrosine kinases , there was no change in the levels of receptor tyrosine kinases in the spinal cords of the tn mice . Instead , we found that HGS is required for synaptic transmission at the neuromuscular junction and for the proper myelination of the peripheral nervous system . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Motor and Sensory Deficits in the teetering Mice Result from Mutation of the ESCRT Component HGS |
During thrombosis , thrombin generates fibrin , however fibrin reversibly binds thrombin with low affinity E-domain sites ( KD = 2 . 8 μM ) and high affinity γ’-fibrin sites ( KD = 0 . 1 μM ) . For blood clotting on collagen/tissue factor ( 1 TF-molecule/μm2 ) at 200 s-1 wall shear rate , high μM-levels of intraclot thrombin suggest robust prothrombin penetration into clots . Setting intraclot zymogen concentrations to plasma levels ( and neglecting cofactor rate limitations ) allowed the linearization of 7 Michaelis-Menton reactions between 6 species to simulate intraclot generation of: Factors FXa ( via TF/VIIa or FIXa ) , FIXa ( via TF/FVIIa or FXIa ) , thrombin , fibrin , and FXIa . This reduced model [7 rates , 2 KD’s , enzyme half-lives~1 min] predicted the measured clot elution rate of thrombin-antithrombin ( TAT ) and fragment F1 . 2 in the presence and absence of the fibrin inhibitor Gly-Pro-Arg-Pro . To predict intraclot fibrin reaching 30 mg/mL by 15 min , the model required fibrinogen penetration into the clot to be strongly diffusion-limited ( actual rate/ideal rate = 0 . 05 ) . The model required free thrombin in the clot ( ~100 nM ) to have an elution half-life of ~2 sec , consistent with measured albumin elution , with most thrombin ( >99% ) being fibrin-bound . Thrombin-feedback activation of FXIa became prominent and reached 5 pM FXIa at >500 sec in the simulation , consistent with anti-FXIa experiments . In predicting intrathrombus thrombin and fibrin during 15-min microfluidic experiments , the model revealed “cascade amplification” from 30 pM levels of intrinsic tenase to 15 nM prothrombinase to 15 μM thrombin to 90 μM fibrin . Especially useful for multiscale simulation , this reduced model predicts thrombin and fibrin co-regulation during thrombosis under flow .
The reaction network and kinetics of human blood clotting impact diseases such as coronary thrombosis , stroke , deep vein thrombosis , hemophilia , disseminated intravascular coagulopathy ( DIC ) , and traumatic bleeding . Numerous therapeutics are designed to either inhibit or catalyze reactions of the coagulation cascade . Despite decades of study , new reaction modulators ( eg . platelet polyphoshate [1] and reaction pathways ( eg . direct conversion of FVIIIa by TF/FVIIa/Xa [2] ) are still being discovered . In some cases , the significance of a particular reaction studied in a purified system may be difficult to resolve since pM-levels of factors formed transiently in whole blood are challenging to measure directly . Excluding platelet metabolism other than the availability of anionic phospholipid , isotropic kinetic models of plasma coagulation in a closed system can include 50 to 100 reactions , 1 to 3 kinetic rate coefficients per reaction , and about 10 initial conditions for zymogen or cofactor concentrations [3–5] . Fortunately , these large ODE models can be parameterized and solved with minor computational expense . In these models , a trigger at t = 0 is required such as 1 to 10 pM tissue factor ( TF ) along with 1% of FVII being in a cleaved yet zymogen-like state as free FVIIa . Alternatively , if no TF is present , a source term for FXIIa generation or non-zero levels of cleaved factors is required to drive clotting [5] . In closed systems , the concentration of substrates and products can undergo >103-fold changes as clotting proceeds non-linearly through initiation , propagation/amplification , and exhaustion ( inhibition and substrate consumption ) . Calibrated automated thrombinography ( CAT ) assay reports these dynamics for platelet-poor or platelet-rich plasma with typical time lags of 3 . 1 and 8 . 1 min , peak thrombin levels of 458 and 118 nM at 10 min , and reaction completion by 25 min [6] . As clotting progresses , the extrinsic tenase/IXase ( TF/FVIIa ) converts FX to FXa and FIX to FIXa . Along with conversion of activated cofactors FVIIIa and FVa , the intrinsic tenase ( FIXa/FVIIIa ) dramatically amplifies production of FXa , while prothrombinase ( FXa/FVa ) generates thrombin ( releasing fragment F1 . 2 ) . Thrombin cleaves platelet PAR-1 and PAR-4 and converts fibrinogen to fibrin monomer by release of fibrinopeptides A and B ( FPA/B ) . The reaction of thrombin and antithrombin to form thrombin-antithrombin ( TAT ) is relatively slow ( ~1 min ) unless catalyzed by heparin . Fibrin monomers associate into protofibrils that laterally aggregate into bundles . Thrombin also activates FXIIIa , a transglutaminase that crosslinks fibrin . Plasmin-mediated fibrinolysis of crosslinked fibrin releases various fibrin degradation products ( FDP ) including D-dimer . These reactions can be studied in closed systems , ± fluid mixing and ± spatial gradients . To mimic thrombosis at a specific wall location ( an open system ) , blood treated with the FXIIa inhibitor corn trypsin inhibitor ( CTI ) can be perfused over a defined thrombotic surface containing TF . Clotting on a surface under flow includes mathematically complex physical phenomenon such as platelet margination to the wall [7] , convective/diffusive transport , concentration boundary layers , pressure-driven permeation , and moving boundaries [8 , 9] . Solving large sets of partial differential equations ( PDEs ) for coagulation species transport and reaction is expensive and non-trivial [10] . Generally , enzyme-substrate interactions at the single molecule level are considered unaffected by macroscopic flow forces . Fibrin has ‘antithrombin-I activity’ via thrombin binding to the low affinity site in the E domain and the high affinity site in the D-domain of the alternative splice variant , γ’-fibrin ( ogen ) . The γ’-fibrinogen splice variant represents about 6–8% of total γ-chains , with γA/γ’ heterodimer representing 12–16% of total fibrinogen [11] . γ′ fibrinogen level is associated with cardiovascular disease . [12] . During thrombosis under flow , thrombin co-localizes on fibrin [13 , 14] . A recent observation is that little thrombin ( detected as TAT ) leaks out of a growing clot unless fibrin polymerization is inhibited with Gly-Pro-Arg-Pro ( GPRP ) [15] . By immunoassays for TAT and F1 . 2 ( ± fibrin inhibitor , GPRP ) and D-dimer ( post-plasmin treatment ) , the dynamics of thrombin and fibrin generation have only recently been measured for flow of human whole blood over defined collagen/TF surfaces [13] . To our knowledge , no model has calculated intrathrombus thrombin generation and fibrin polymerization under flow conditions where fibrin is being formed dynamically and local thrombin is reversibly binding fibrin through the weak ( E-domain ) and strong binding sites ( γ’-variant ) . We present a reduced model where key assumptions are supported by direct experiment measurements . This reduced model deploys a thin film assumption for the clot core ( thickness ~ 15 microns ) where zymogen levels in the clot are set to be identical to those in the flowing plasma . This assumption did not hold for fibrinogen transport , which is not surprising given that fibrinogen ( 340 kDa ) is considerably larger than the other coagulation factors . For a set of prevailing plasma concentrations for Factors FVIIa , FIX , FX , FXI , prothrombin , fibrinogen as well as initial surface [TF]o , the reduced model makes quantitatively accurate predictions of thrombin and fibrin levels under venous flow conditions . Fibrin appears to allow for explosive but feedback-inhibited production of thrombin . After a clotting episode , the large amount of fibrin-bound thrombin was predicted to take a few hours to elute into the circulation to form TAT . This reduced model may be particularly useful for multiscale simulations of thrombosis over vessel length scales of mm to cm .
For perfusion of CTI-treated whole blood across a 250-μm long patch of collagen/TF ( 1 molecule-TF/μm2 ) , platelets rapidly accumulate and create a sheltered reaction environment triggered by TF for production of thrombin and fibrin [15 , 16] . The effluent can be sampled and subjected to immunoassays to determine the measured species flux for a 250-long x 250 μm-wide patch of collagen/TF for TAT and F1 . 2 , in the presence and absence of fibrin assembly ( ± GPRP ) ( Fig 2A and 2C ) . The dynamic accumulation of fluorescent fibrin in the experiment was converted to a fibrin concentration by end-point immunoassay of D-dimer , post-plasmin treatment ( Fig 2E ) . For the 7 reaction rate coefficients ( α1-α7 ) ( Fig 1B , Table 1 ) , only 3 rates required adjustment ( η4 , η5 , η6 ) from their literature values in order to simulate thrombin and F1 . 2 elution and fibrin polymerization in the presence and absence of GPRP . The adjustments for prothrombinase activity ( η4 = 0 . 18 ) and thrombin activation of FXIa were modest ( η6 = 0 . 36 ) and could involve either transport rate limits or just as possible the difference of the reaction in the whole blood milieu in comparison to dilute buffer conditions used in enzyme studies . The adjustment in thrombin mediated activation of fibrinogen was markedly pronounced , requiring a 20-fold reduction in the rate ( η5 = 0 . 05 ) . This 20-fold reduction in rate corresponds either to a ~80-fold increase in Km ( unlikely ) or an 80-fold decrease in the intraclot level of fibrinogen substrate relative to plasma levels . In the experimental measurement , the generation of fibrin per thrombin molecule was unexpectedly low , given the known speed of FPA release by thrombin ( kcat = 80 s-1 ) . In considering the value η5 as an effectiveness factor ( actual rate/ideal rate in the absence of transport limits ) , the penetration of fibrinogen ( 340 kDa ) into the dense fibrin-rich core of the clot is hypothesized , and required in the model to be diffusion-limited . The model clearly predicts that thrombin has difficulty eluting from the fibrin due to fibrin binding ( Fig 2A and 2B ) . Once fibrin is prevented from forming or binding thrombin ( α5 = 0 or setting KD>10 M ) in the simulation or in the experiment ( +GPRP ) , the TAT flux increases linearly with time for the first 500s and then increases even faster from 500 to 800s . As thrombin is generated , a small fragment F1 . 2 is released as a result of the prothrombinase activity . In both experiment and simulation , F1 . 2 elutes from the clot even in the presence of fibrin ( Fig 2C and 2D ) . Additionally , more F1 . 2 is made than TAT , since thrombin can be inhibited by other inhibitors such as C1 and α2-macroglobulin; the simulation accounts for this ( Note the value of 0 . 7 in Eq 1 for J-F1 . 2 ) . With GPRP to eliminate fibrin’s antithrombin-I activity and facilitate FXIa-mediated feedback pathway , more F1 . 2 is detected both in the experiment and in the simulation ( Fig 2C and 2D ) . Under flow conditions , fibrin reached a concentration that was 10-fold greater than plasma fibrinogen concentration ( 3 mg/mL , 9 μM ) ( Fig 2E ) . The dynamics of intrinsic tenase generation , prothrombinase production , and thrombin binding to fibrin were explored in the model under various conditions . In the model , intrinsic tenase ( “IXa” = FIXa/FVIIIa ) reaches a level of 30 pM by 200 sec . By turning off thrombin-feedback activation of FXIa ( setting α6 = α7 = 0 ) , the model demonstrates that most of the intrinsic tenase is generated in the first 200 sec is from tissue factor ( curve c , Fig 3A ) while after 500 sec , most of the intrinsic tenase is a result of the feedback activation of FXIa by thrombin as seen in curve b = ( a–c ) ( Fig 3A ) . FXIa reaches a level of only 5 pM in the simulation ( dashed line , Fig 3A ) demonstrating how potent FXIa can be for FIXa production and thrombin production . Similarly , the intrinsic tenase can be turned off ( i . e . severe hemophilia ) by setting α2 = α3 = 0 such that all of the prothrombinase is the direct result of the extrinsic tenase ( Fig 3B ) . In this case , very little prothrombinase is generated , as expected for extreme hemophilia A/B . The role of FXIa in prothrombinase generation can be seen , especially after 500 sec , where most of the prothrombinase is a downstream result of the generation FXIa ( Fig 3B ) . In the simulation , little thrombin is made when the extrinsic tenase/FIXase ( TF* ) cannot generate FIXa ( α2 = 0 ) , again consistent with the circumstances of severe hemophilia . As expected from the dynamics for prothrombinase , the majority of thrombin made at times >500 sec was the result of thrombin-feedback activation of FXIa ( Fig 3C ) . Thrombin reached 18 μM-levels after 800 sec of clotting with almost all of it bound to the weak ( E-domain ) and strong ( γ’ ) site and about only 1% of the thrombin ( ~100 nM ) existing as a free species ( Fig 3D ) . By 800 sec of clotting , the full effect of cascade amplification is seen in that an initial surface concentration of [TF]o = 1 molecule-TF/μm2 ( 2 . 2 pM TF/VIIa = TF* in the core ) results in the generation of 30 pM intrinsic tenase , ~15 μM prothrombinase , ~18 μM thrombin ( 100 nM free thrombin ) , and ~90 μM fibrin ( 30 mg/mL ) . The range of γ’ fibrinogen concentrations can vary in healthy individuals [22] , with a reference range of 0 . 088 to 0 . 551 mg/mL . Additionally , fibrinogen is an acute response gene and the fraction of splice variant can change . The concentration of γ’ fibrinogen concentrations and the γ’ fibrinogen/total fibrinogen ratio have been reported to be relevant in thrombosis , and different in different stages of disease , potentially with some protectant effect in venous thrombosis [23] . In the simulation , we varied the γ’ fibrinogen concentration to explore the effect on the co-regulation of fibrin and free thrombin concentration . With more γ’ fibrinogen , there was slightly more high-affinity sites for thrombin , therefore , sequestering more thrombin and decreasing the fibrin and free thrombin concentration ( Fig 4A and 4B ) . In contrast , a 50% reduction in γθ caused a slight increase in the level of free thrombin and the amount of fibrin made . However , the effect of γ’-fibrinogen levels were not particularly marked , a reasonable result given the excess fibrin that is formed relative to thrombin , but still suggestive of a protective or regulating contribution in venous thrombosis . As reported previously , platelet contraction can alter protein transport [24] with soluble proteins retained longer in the core of the clot than the less dense outshell shell . In the laser injury mouse model , albumin half-life in the clot core has been measured to be about 2 sec . In the simulation , we artificially adjusted the escape time between 1 sec and 4 sec to explore how intrathrombus diffusion influences local free thrombin and , consequently , fibrin production . A longer escape time of 4 sec resulted in dramatically higher intrathrombus concentration of fibrin and free thrombin , indicating the model was very sensitive on escape time . The concentration of thrombin increased more than 3-fold with a doubled escape time to 4 sec ( Fig 4C and 4D ) . To simulate dynamic concentrations of intrathrombus thrombin in the core , the velocity field and convective-diffusive transport of thrombin was calculated by COMSOL ( Fig 5A and 5B ) . The empirically measured flux of thrombin JIIa ( t ) |Y = 0 ( via F1 . 2 ELISA ) was set at the bottom boundary condition of the core region . The empirically measured time-varying fibrin concentration fibrin ( t ) ( calibrated by end-point D-dimer ELISA ) was set uniformly in the core region . The concentration of Eθ and γθ sites were set to 1 . 6x fibrin ( t ) and 0 . 3x fibrin ( t ) , respectively . After 800 sec , the thrombin flux entering the domain was set to zero in order to explore long term thrombin elution from the clot . The time-averaged flux into and out of the clot outlet ( Fig 5C and 5D ) revealed that most of the thrombin was captured by the fibrin , via both sites . By 500 sec , the concentration of intrathrombus thrombin was only 61 nM , only about 1% of total thrombin ( 5 . 5 μM ) in the clot ( Fig 5C and 5D ) , indicating that the literature KD values for binding were consistent with actual independent measurements of TAT elution . The transient concentrations of total thrombin , intrathrombus free thrombin , and bound thrombin to each site , are shown in Fig 5E . After 800 sec , the thrombin flux from the bottom plate was set to zero and the thrombin in the clot was allowed to be eluted by diffusion under prevailing flow conditions . The binding of thrombin by fibrin was sufficiently strong under a venous shear rate with an apparent half-life in the clot of 1 . 1 hour ( Fig 5F ) . Thrombin eluted slowly into the flow field , relative to its half-life in the presence of antithrombin , such that its concentration would not be expected to perturb the hemostatic balance in the circulation . Thus , circulation levels of TAT can accumulate over hours and be measured in patients , even when most of the thrombin made in the first 800 sec is fibrin bound .
By assuming plasma zymogens can enter a thin clot at a rate significantly greater than their consumption , a highly reduced and essentially linearized ODE model provided a reaction topology suitable for accurate prediction of blood clotting on collagen/TF under venous flow . The thin film assumption was first formalized in Kuharsky-Fogelson model [25] to generate a large systems of ODEs describing clotting under flow . With the well mixed , thin film approximation , we were able to simplify clotting under flow to 8 ODEs and 19 parameters . A total of 16 parameters were from literature and only 3 were adjusted in order to fit the measured TAT and F1 . 2 and fibrin generation data ( ± GPRP ) . Of the 3 adjustable parameters , only the rate of fibrinogen activation by thrombin appeared to be strongly diffusion-limited ( η5 = 0 . 05 ) . This result was not particularly surprising given the enormous size of fibrinogen in comparison to the other coagulation factors . While ignoring cofactor activation of FVa and FVIIIa as non-rate limiting appeared to be compatible with predicting clotting of healthy blood , the generation of FIXa was absolutely required for robust thrombin production . As an ODE model , the actual transport physics were mainly parameterized by the rate of free thrombin elution from the clot , guided by experimental measurements of ~2-sec half-life of flash-activated albumin in a clot subjected to flow along its outer boundary . In the presence of thrombin binding to fibrin , the elution rate of free thrombin from the clot appears to be an important regulator of clotting ( Fig 4C and 4D ) . The 2-sec elution half-life for proteins was consistent with ( i ) in vivo mouse measurement , ( ii ) the human blood microfluidic measurements , and ( iii ) the average time it takes a protein to diffuse an average distance of 15 microns . The roles of FXIIa in mouse thrombosis models [26] and platelet released polyphosphate to amplify thrombin-mediated feedback activation of FXIa [27] have motivated the pharmaceutical development of FXIIa , FXIa , and polyphosphate inhibitors . In Fig 3A , FXIa reaches ~5 pM by 500 sec of clotting and the amount of thrombin generated between ~500 and 800 sec is largely FXIa-dependent ( Fig 3C ) . This is exactly consistent with microfluidic experiments conducted with anti-FXI antibody that blocked the increase of TAT and F1 . 2 flux and fibrin deposition after 500 sec [13 , 15 , 27] . By calibrating the model on measured thrombin and fibrin generation rates , the simulation provides insights , based on Table 1 kinetics , into pathways proximal to thrombin . The concentrations of FIXa/FVIIIa and FXa/FVa and FXIa were particularly low and would be difficult to measure directly inside the clot under flow conditions . Over 800 sec of clotting , the model revealed “cascade amplification” from 30 pM levels of intrinsic tenase to 15 nM prothrombinase to 15 μM thrombin to 90 μM fibrin , with FXIa pathways contributing significantly after 500 sec . Interestingly , little thrombin results directly from the FXa produced by TF/FVIIa , consistent with severe hemophilic blood producing little fibrin following perfusion over TF surfaces [28] . For the thin core region within the rapidly formed platelet deposit , the kinetics of thrombin and fibrin production are largely sheltered from the prevailing flow on the outer boundary of the clot . The current model may have some applicability to core dynamics during TF-driven arterial thrombosis since the core thickness ( thrombin and fibrin and P-selectin positive region ) has been measured to be relatively similar between the venous ( 100 s-1 ) and arterial ( 2000 s-1 ) condition [29] . As a model analyzing dynamics limited to the thin , core region using ODEs , the model was not designed to predict clot growth and spatial dynamics over distances of 100s or 1000s of microns . However , the TF-dominated thrombin generation rate in the core region could be coupled to spatial models of platelet deposition and FXIa-enhanced thrombin generation . The reduced model focuses on concentration changes of thrombin and fibrin in the thin “core” region which is only 15-microns thick . Since proteins can diffuse this short distance in ~10 sec , the well-mixed assumption of this thin region is reasonable , although not suited for predicting longer distance wave propagation such as found for clotting in stagnant plasma over millimeter distances [30–32] . The full biochemical and spatial complexity of human coagulation ( typically involving >50 reactions in reality and in simulation ) may relate more to the kinetic demands of hemostasis and its strong selective pressure ( eg . surviving child birth ) , rather than to the complexities of thrombosis in older adults of modern times . The reduced model exploits the thin-film approximation under flow to emphasize a few key zymogen activation events at constant zymogen concentration . Despite its simplicity , this reduced model may have a useful implementation within more complex spatial thrombosis models that include platelet activation and accumulation [33 , 34] . The reduced model is not directed at describing the full progression of a thrombotic event over large spatial distances in large vessels , particularly where platelet accumulation dominates the growth process . However , this model may be very useful in establishing a surface TF-dependent thrombin flux and fibrin regulation that is a time-dependent boundary condition to a larger multi-species PDF model where platelets continue to accumulate and thrombin production transitions to platelet polyphosphate/FXIa-dependent . Few simulations of clotting under flow include the role of anti-thrombin-I activity of fibrin or γ’-fibrinogen levels . The ratio of γ’-fibrinogen to total fibrinogen may be clinically relevant . Reduced γ’-fibrinogen levels have been associated with an increased venous thrombosis risk [23 , 35] Here , we demonstrated a simplified ODEs model to simulate the thrombin and fibrin generation . This reduced model may be particularly useful in multiscale simulations that seek to account for single platelet phenomenon [36] , microscopic attributes of a wound site [16] , and whole vessel dynamics [17 , 34] .
By direct imaging , blood clotting over collagen/tissue factor surfaces generates a 15-micron thick “core” region ( δ = 15 μm , porosity ~0 . 5 ) that contains P-selectin-positive platelets with concomitant thrombin generation and dense fibrin deposit [13 , 37 , 38] . This is the thin film reaction compartment in which species are considered to be pseudo-homogeneous and of uniform concentration . The porosity used in the model is an estimate of the spatially averaged porosity over the entire 250 μm x 250 μm clotting region of the microfluidic assay , recognizing that this averages over both platelet/fibrin dense regions surrounded by fibrin dense regions . It is expected that the lower porosity decreases the effective diffusion of thrombin . Substrate concentrations in the clot core are considered to be constant at plasma levels ( So ) . Thus , the Michaelis-Menton reactions to generate product [P] become linearized with respect to [E , enzyme] as dP/dt = [kcat ( So/ ( Km+So ) ) ] • E ( t ) = α• [E ( t ) ] ( Table 1 ) . For species without diffusion limitations , the effectiveness factor ƞ = 1 ( actual rate/ideal rate without diffusion limits ) . If a species experience transport limits , then ƞ < 1 and dP/dt = ƞ • α• [E ( t ) ] . This approach is supported by direct measurements of TAT , F1 . 2 , FPA , and D-dimer that indicate local clot associated product levels ( thrombin and fibrin ) are in excess of plasma levels ( prothrombin and fibrinogen ) demonstrating continual substrate delivery into the core of the clot [13] . Just as substrates can enter the core , free thrombin and FXIa were considered to escape the clot . The escape time was set to the measured half-life of 2 sec for albumin within the clot core [24] where kelute = ln ( 2 ) /2 sec . This half-life in the thin-film for product escape = kelute • [E] is conceptually and mathematically similar to the use of a mass transfer coefficient kc with units of 1/time as defined in [25] . Although the binding characteristics of F1 . 2 to fibrin are unknown , we hypothesize that the observation that fibrin suppresses F1 . 2 elution may be consistent with fibrin inhibiting the thrombin-feedback pathway involving FXIa which in turn results in less prothrombin conversion . The small Fragment F1 . 2 was considered to leak out of the clot core as fast as thrombin was generated in the core . For TAT , 70% of the thrombin eluted from the clot was considered complexed with antithrombin with the remaining 30% of eluted thrombin complexed with other inhibitors [13] . Based upon all thrombin and F1 . 2 begin generated in the pore space Vpore of the clot core ( Fig 1A ) , the flux J-F1 . 2 and the flux J-TAT leaving the clot were calculated as: Flux , J‐F1 . 2 ( t ) =η4•α4[Xa ( t ) ] ( Vpore/area ) =η4•α4[Xa ( t ) ]δ Eq 1 Flux , J‐TAT ( t ) =0 . 7•kelute[IIa ( t ) ] ( Vpore/area ) =0 . 7•kelute[IIa ( t ) ]δ Eq 2 For healthy non-hemophilic blood , the generation of cofactors ( FVa , FVIIIa ) was treated as non-rate limiting . Thus , the intrinsic tenase FIXa/FVIIIa = “FIXa” , and prothrombinase FXa/Va = “FXa” . The availability of FIXa and FXa ( not FVa or FVIIIa ) controlled the enzymatic cleavage of their substrates according to the reactions parameterized in Table 1 . FVII and FVIIa in plasma were assumed to instantaneously equilibrate with surface TF such that [TF*] = TF/FVIIa = 1% of [TF]0 where [TF]0 = 1 molecule/μm2 set experimentally . No additional TF* was allowed to be generated in the model , equivalent to the quenching dynamics via platelet coverage invoked by Kuharsky and Fogelson . [25] . The inhibition mechanisms of coagulation proteases via TFPI , ATIII , C1-inhibitor and α2-macroglobulin are complex and diverse and not fully resolved . Inhibition was treated uniformly to be a pseudo-first order reaction . Enzyme half-lives were set to be on the order of 1-min for FXa , FIXa , FXIa , FIIa ( and 3 min for TF* since FVIIa generation was neglected ) . In other words , inhibition was clearly not as rapid as 0 . 1 min and clearly not as slow as 10 min . Thus , ki = ln ( 2 ) /60s as a first approximation . Reversible thrombin binding to the weak E-domain site ( EKD = 2 . 8 μM ) and the strong γ’-site ( γKD = 0 . 1 μM ) was consider to have diffusion-limited association ( kf = 100 μM-1 s-1 ) . Fibrin-bound thrombin was considered to be fully resistant to inhibition . All fibrin monomer generated in the clot core was assumed to be fully incorporated into fibrin . The delivery of plasma zymogens and platelets to the surface of a growing clot would be even faster at arterial flow conditions , however the increased shear forces tend to enhance platelet removal . Unfortunately , it is difficult to measure eluted FPA , F1 . 2 , or TAT under arterial conditions due to their 10-20X greater dilution in the exit flow stream compared to the venous measurement [15] . Importantly , arterial syndromes tend to be drugged with anti-platelet agents , not anticoagulants . For the reaction topology shown in Fig 1 , these assumptions result in a reduced clotting model with only 8 ODEs for 6 reactive species undergoing 7 reactions ( Table 1 ) and 2 fibrin sites for reversible binding of thrombin . These 8 ODEs were solved in Matlab R2016b using the ODE solver ode15s . ODE 1 . dTF*dt=−ki , TF∙TF*forki , TF=ln ( 2 ) /180s ODE 2 . dXadt=α1∙TF*+α3∙IXa−ki∙Xa ODE 3 . dIXadt=α2∙TF*+α7∙XIa−ki∙IXa ODE 4 . dXIadt=η6∙α6∙IIa−kelute∙XIa−ki∙XIaforη6=0 . 36 ODE 5 . dFibrindt=η5∙α5∙IIaforη5=0 . 05 where: Eθtotal= ( 1 . 6 ) ∙fibrin and γθtotal= ( 0 . 3 ) ∙fibrin ODE 6 . dIIadt=η4∙α4∙Xa− ( dESdt+dγSdt ) −kelute∙IIa−ki∙IIaforη4=0 . 18 ODE 7 . dESdt=Ekf∙IIa∙ ( Eθtotal−ES ) −Ekr∙ES ODE 8 . dγSdt=γkf∙IIa∙ ( γθtotal−γS ) −γkr∙γS This reduced model for blood clotting on a collagen/TF surface under flow uses 19 parameters , only 3 of which were adjusted to fit the experimental data: For simulations of thrombin equilibrated to fibrin ( no thrombin generation ) followed by desorption-controlled elution of thrombin , a full PDE simulation was solved for a 2D rectangular domain ( 1000 μm long x 60 μm high ) representing a channel of the microfluidic device [13] . At a location 150 μm downstream of the entrance , a porous fibrin reaction zone ( 250-μm long x 15-μm high ) was defined as the clot core , with fibrin concentration set by D-dimer ELISA experiment . A thrombin flux was imposed along the bottom plate to allow thrombin diffusion through the fibrin ( in the presence of two binding sites ( 1 . 6 Eθ-sites per fibrin monomer; 0 . 3 γθ−sites per fibrin monomer ) with binding kinetic parameters given Table 1 . COMSOL was used to solve the convection-diffusion-reaction equation for thrombin transport in two steps . First , the Free and Porous Media Flow module was first solved with a Stationary study step to get the velocity field ( u ) . Second , the mass transport was solved by the Transport of Diluted Species in Porous Media ( thrombin ) coupled with General Form PDE for weak and tight thrombin binding was solved with a time-varying time-step , with a relative tolerance of 0 . 0001 . The thrombin binding by fibrin was described by the following equations: ∂CIIa∂t=−∇∙ ( −D∇CIIa ) −u∙∇CIIa−∂ES∂t−∂γS∂t ∂ES∂t=Ekf∙CIIa ( Eθtotal−ES ) −Ekr∙ES ∂γS∂t=γkf∙CIIa ( γθtotal−γS ) −γkr∙γS | During blood clotting events , a complex series of reaction are involved . Simulation gives insights to the concentration of different enzymes which are at too low of concentration to be detected . However , the models are often large and difficult to solve for clotting under flow conditions . With a thin film approximation , we were able to simplify clotting under flow with parameters from literature , with only 3 adjusted in order to fit the experimental data . This model gave insights into the dynamics of the species involved , and the roles of γ’-fibrin and thrombin feedback activation . This reduced model may be useful in further multiscale simulations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"fibrinogen",
"fibrin",
"cardiovascular",
"medicine",
"thrombosis",
"elution",
"platelets",
"glycoproteins",
"research",
"and",
"analysis",
"methods",
"separation",
"processes",
"coagulation",
"disorders",
"thrombin",
"animal",
"cells",
"proteins",
"blood",
"plasma",
"hematology",
"biochemistry",
"blood",
"coagulation",
"blood",
"anatomy",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"cardiovascular",
"diseases",
"vascular",
"medicine",
"glycobiology"
] | 2019 | Reduced model to predict thrombin and fibrin during thrombosis on collagen/tissue factor under venous flow: Roles of γ’-fibrin and factor XIa |
Aspergillus fumigatus is an environmental fungus that causes invasive aspergillosis ( IA ) in immunocompromised patients . Although -CC-chemokine receptor-2 ( CCR2 ) and Ly6C-expressing inflammatory monocytes ( CCR2+Mo ) and their derivatives initiate adaptive pulmonary immune responses , their role in coordinating innate immune responses in the lung remain poorly defined . Using conditional and antibody-mediated cell ablation strategies , we found that CCR2+Mo and monocyte-derived dendritic cells ( Mo-DCs ) are essential for innate defense against inhaled conidia . By harnessing fluorescent Aspergillus reporter ( FLARE ) conidia that report fungal cell association and viability in vivo , we identify two mechanisms by which CCR2+Mo and Mo-DCs exert innate antifungal activity . First , CCR2+Mo and Mo-DCs condition the lung inflammatory milieu to augment neutrophil conidiacidal activity . Second , conidial uptake by CCR2+Mo temporally coincided with their differentiation into Mo-DCs , a process that resulted in direct conidial killing . Our findings illustrate both indirect and direct functions for CCR2+Mo and their derivatives in innate antifungal immunity in the lung .
The incidence of fungal infections has been on the rise for several decades due to increased use of immunosuppressive and myeloablative therapies for malignant and non-malignant diseases [1] , [2] , [3] . Invasive aspergillosis ( IA ) , most commonly caused by A . fumigatus , is a frequent cause of infectious morbidity and mortality in patients with leukemia and in allogeneic hematopoietic cell transplant ( HCT ) recipients [4] , [5] , [6] , [7] . Previous studies have determined that innate and adaptive components of the immune system play essential roles in defense against IA [2] , [4] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] . Neutrophils have long been recognized as an essential innate cell in defense against IA , as neutropenia represents an important clinical risk factor [18] . Human susceptibility to IA in patients with defective neutrophil function ( e . g . patients with chronic granulomatous disease ) underscores the functional role of neutrophils in host defense . These findings are recapitulated in animal models of IA in which antibody-mediated depletion of neutrophils leads to uncontrolled fungal growth in the lung and to mortality from IA [19] , [20] , [21] , [22] , [23] . In addition to neutrophils , protective immune functions have been ascribed to a variety of innate cells that include macrophages , NK cells , myeloid DCs and plasmacytoid DCs [17] , [19] , [21] , [24] , [25] . While alveolar macrophages are capable of conidial killing in vitro [26] and in vivo [27] , and likely contribute to innate defense , clodronate-mediated alveolar macrophage ablation did not lead to IA , suggesting that AM fungicidal activity can be functionally compensated by other leukocytes in vivo [21] . Similarly , the contributions of NK cells and myeloid DCs to antifungal defense against aspergillosis have been examined only in neutropenic mouse models of IA [24] , [25] . Thus , despite the important contributions of other innate cells subsets in antifungal immunity , previous studies suggest that neutrophils are the sole indispensable innate effector cell in host defense against IA [19] , [20] , [21] , [22] . In contrast to their essential role against respiratory fungal infection , neutrophils have been found to be dispensable for defense against the intracellular pathogens Listeria monocytogenes and Toxoplasma gondii [28] , [29] . In both infection models , CCR2+Ly6Chi inflammatory monocytes ( CCR2+Mo , throughout this manuscript CCR2+Mo is used as an abbreviation for inflammatory monocytes , defined as CD45+CCR2+Ly6ChiCD11b+Ly6G− leukocytes ) were identified as essential innate effector cells that mediate bacterial and parasitic eradication [28] , [29] , [30] , [31] . In these models , the formation of monocyte-derived TNF- and inducible nitric oxide synthase-producing dendritic cells ( Tip-DCs ) correlated with microbial clearance [31] , [32] , [33] . Since bacterial uptake by Tip-DCs during systemic listeriosis and salmonellosis appears to be an infrequent event ( <1% of Tip-DCs ) [34] , [35] , [36] , it remains unknown whether inflammatory monocytes and their derivatives exert relevant antimicrobial activity by pathogen engulfment and killing at the portal of infection . In fungal infection models , the role of CCR2+Mo has largely been understood in the context of adaptive CD4 T cell responses . In a respiratory A . fumigatus infection model CCR2+Mo are rapidly recruited to the lung and differentiate into CCR2+CD11c+MHCII+CD11b+CD103− monocyte-derived DCs ( Mo-DC ) that are essential for the induction and maintenance of A . fumigatus-specific Th1 CD4 T cell responses [37] , [38] . Mo-DCs have also been found to be important for initiation of fungus-specific T cell responses in the context Blastomyces vaccination and Histoplasma capsulatum infection in the lung [39] , [40] , [41] , [42] . In vivo studies with human blood monocytes have shown that these cells have fungistatic activity ex vivo and elaborate cytokines and chemokines following stimulation with A . fumigatus conidia [43] , [44] , [45] , [46] . Although emerging evidence indicates that CCR2+Mo and their derivatives contribute to innate defense against systemic candidiasis [47] , [48] , it remains unclear whether CCR2+Mo act to control the influx and activity of other effector cell populations or directly contribute fungicidal capacity at sites of infection . One possible model is that CCR2+Mo and their derivatives control antifungal activity in the lung by regulating neutrophil influx , as suggested in LPS-induced models of pulmonary inflammation [49] . A second model of CCR2+Mo antifungal activity during respiratory fungal infection may involve the release of pro-inflammatory mediators [25] to enhance the fungicidal activity of resident or recruited leukocytes . A third model of antifungal activity involves direct antimicrobial effects of CCR2+Mo and derivative cells . In the present study we set out to elucidate the mechanisms by which CCR2+Mo contribute to innate antifungal immunity in the lung . To this end , we employed genetically engineered mice that express a diphtheria toxin receptor ( CCR2 depleter mice ) or a GFP transgene ( CCR2 reporter mice ) under the control of the endogenous CCR2 promoter [29] , [38] and fluorescent Aspergillus reporter ( FLARE ) conidia that trace the outcome of CCR2+Mo and Mo-DC interactions with conidia in the lung with single-encounter resolution [27] . We found that sustained depletion of CCR2+Mo and Mo-DCs led to the development of IA and a reduction in neutrophil conidiacidal activity . Beyond their impact on neutrophil conidiacidal responses , CCR2+Mo and Mo-DCs formed a TNF and iNOS-producing effector cell population in the lung that exerted rapid and effective conidiacidal activity similar in magnitude to neutrophil fungicidal activity . In aggregate , our studies suggest that CCR2+Mo and their derivatives mediate an essential role in antifungal defense in the lung by directly containing conidial germination and by enhancing neutrophil antifungal activity .
To examine the contributions of CCR2+ Mo and their derivatives to respiratory fungal defense , we monitored the outcome of intratracheal A . fumigatus conidial challenge in CCR2 depleter mice [38] that express a functional diphtheria toxin receptor ( DTR ) under control of the CCR2 promoter . CCR2 depleter mice were treated with diphtheria toxin ( DT ) on day −1 , +1 , and +3 to ablate CCR2-expressing cells during respiratory fungal infection . We included two control groups: non-transgenic C57BL/6J ( B6 ) littermates that received the same DT administration regimen as CCR2 depleter mice and B6 mice that were depleted of neutrophils by administration of anti-Ly6G antibodies . Consistent with previous studies using a different neutrophil-depleting antibody [20] , [21] , [22] , anti-Ly6G-treated mice rapidly succumbed to IA ( Figure 1A ) . Non-transgenic B6 control animals treated with DT did not develop disease symptoms throughout the duration of the experiment . Strikingly , CCR2 depleter mice treated with DT uniformly succumbed to infection when challenged with inocula that ranged from 4–8×107 conidia ( Figure 1A and 1B ) . To determine whether mortality was associated with fungal tissue invasion , Gomori methenamine silver ( GMS ) -stained lung sections were examined from CCR2 depleter mice and control animals at various time points post-infection . Lung sections from CCR2 depleter mice showed extensive and progressive hyphal growth ( Figure 1C ) starting at day +3 post infection ( p . i ) . Extensive lung parenchymal destruction and obliteration of bronchoalveolar architecture was apparent at later time points . In contrast , lung sections from DT-treated B6 mice only showed evidence of conidia that failed to germinate at all time points examined . This is consistent with our previous studies in which B6 mice were able to effectively prevent conidial germination [37] , [50] , [51] . In aggregate these findings demonstrate that CCR2+ cells are essential for early host defense against A . fumigatus and that their ablation leads to the development of IA . Previous studies have shown a protective role for NK cells in a neutropenic model of IA [24] . Since a subset of NK cells express CCR2 , we explored whether the phenotype observed in CCR2 depleter mice could be linked to a defect in NK cells . We examined the recruitment of NK cells to the lung of CCR2 depleter mice and to control non-transgenic littermates during respiratory fungal infection . We observed that DT treatment significantly depleted NK cells in the lung of infected CCR2 depleter mice at 24 and 48 h p . i . ( Figures 2A and 2B ) . To examine whether this reduction in NK cells could be linked to the development of IA in CCR2 depleter mice , we examined the progression of A . fumigatus infection in mice that lack all lymphocytes , including NK cells , iNKT cells , and innate lymphocytes ( recombination activating gene [RAG-2] and common gamma chain [γC] double deficient mice; RAG−/−γC−/− ) . NK1 . 1+ cells were absent from the lungs of A . fumigatus-infected RAG−/−γC−/− mice ( Figure 2A ) but the mice showed normal neutrophil and enhanced monocyte recruitment to infected lungs ( Figures 2C and 2D ) . Despite a total lack of NK cells , RAG−/−γC−/− mice controlled A . fumigatus conidial inocula at 24 and 48 h p . i . , as judged by the recovery of viable fungal cells from the lungs of RAG−/−γC−/− compared to control mice ( Figures 2E and 2F ) . In addition , we did not observe invasive fungal growth in infected RAG−/−γC−/− mice by lung histopathology ( Figure 2G ) and RAG−/−γC−/− mice did not develop disease symptoms within the one week observation period . In contrast , CCR2 depleter mice showed a significant increase in the number of viable fungal cells in the lung at 24 and 48 h p . i . ( Figures 2E and 2F ) which preceded invasive fungal growth at 3 days p . i . ( Figure 1C ) . In aggregate , these results indicate that the development of IA in CCR2-depleter mice cannot be explained by DT-induced ablation of CCR2+ NK cells . Given the crucial role of neutrophils in defense against IA , we examined the impact of CCR2+Mo ablation on neutrophil chemotactic responses and recruitment to the lung . Although previous studies have clearly established that neutrophils are not directly eliminated by DT administration in CCR2 depleter mice [29] , [38] we hypothesized that CCR2+Mo ablation could interfere with lung neutrophil recruitment due to their role as producers or amplifiers of chemotactic mediators , as has been observed in a LPS-induced model of lung inflammation [49] . To test this possibility , CCR2 depleter mice were treated with DT , infected with A . fumigatus conidia , and euthanized at various time points after infection to measure the production of neutrophil-recruiting chemokines and to enumerate and analyze lung homogenates by flow cytometry . CCR2 depleter mice treated with DT had similar lung levels of chemokine ( C-X-C ) motif ligand 1 ( CXCL1 ) and CXCL2 as control non-transgenic littermates treated with DT , suggesting that CCR2+ cells are not required for the production of these chemokines during respiratory fungal infection ( Figures 3A and 3B ) . Although DT administration clearly eliminated all CCR2+Mo in infected mice ( Figures 3C and 3D ) , DT administration did not deplete lung neutrophils ( identified as CD45+CD11b+Ly6G+Ly6C+ cells ) ( Figure 3C ) . Furthermore , similar numbers of neutrophils were present in the lung of CCR2 depleter and control mice at various times after infection ( Figure 3E ) . Although there was a modest trend towards lower numbers of neutrophils in CCR2 depleter mice these differences did not reach statistical significance . In contrast , B6 mice treated with anti-Ly6G antibodies had preserved lung CCR2+Mo recruitment ( Figure 3F ) , but were depleted of neutrophils ( Figure 3G ) . In aggregate , these findings indicate that CCR2 depleter mice produce wild-type levels of CXCL1 and CXCL2 during respiratory fungal infection and display preserved neutrophil recruitment to the site of infection , though these processes per se are insufficient to prevent the development of IA . Since neutrophil recruitment was not affected by CCR2+Mo depletion , we hypothesized that neutrophil function may be altered , resulting in a reduction in neutrophil fungicidal activity in CCR2 depleter mice . To test this hypothesis , we utilized a recently developed fluorescent Aspergillus reporter strain ( FLARE ) to monitor and quantify neutrophil-mediated uptake and killing of A . fumigatus conidia in vivo [27] . The FLARE strain distinguishes live and dead conidia by incorporation of a tracer ( Alexa Fluor 633; AF633 ) and a viability ( DsRed ) fluorophore . Host leukocytes that engulf live DsRed+AF633+ conidia emit two fluorescent signals , one of which ( DsRed ) is extinguished when leukocytes kill engulfed conidia . Using the FLARE strain , we quantified neutrophil conidial uptake and killing in CCR2 depleter and control mice . Infection of DT-treated CCR2 depleter mice with FLARE conidia revealed that CCR2+Mo ablation did not alter the frequency of neutrophils with engulfed conidia at 12 or 36 hours p . i . compared to non-transgenic , DT-treated littermate controls ( Figure 4B and data not shown ) , indicating that ablation of CCR2+Mo does not decrease neutrophil conidial uptake . However , the frequency of neutrophils with live conidia was substantially increased in DT-treated CCR2 depleter mice compared to control mice ( Figures 4A and 4C ) . In other words , conidia engulfed by neutrophils were more likely to be killed in control mice than in CCR2 depleter mice ( Figure 4C ) . Neutrophil expression of Toll-like receptor 2 and 4 and of the C-type lectin receptor Dectin-1 was similar in CCR2 depleter and in control mice at 36 p . i . ( data not shown ) . To extend these observations , we compared bone marrow neutrophil conidiacidal activity in vitro in the absence and presence of CCR2+ Mo , using bone marrow cells harvested from DT-treated CCR2 depleter and non-transgenic littermate controls . When CCR2+ Mo were absent from neutrophil–conidia co-culture experiments , neutrophil conidial viability was higher than in co-cultures that included CCR2+ Mo , though neutrophil conidial uptake was similar in both cases ( Figures 4D–4F ) . Addition of flow-sorted bone marrow monocytes ( identified as CCR2 ( GFP+ ) , CD11b+CD11c−NK1 . 1− cells ) restored the conidiacidal function of neutrophils to baseline levels ( Figure S1 ) . These findings indicate that CCR2+ Mo and derivative cells enhance neutrophil conidiacidal activity when these leukocytes are combined as purified cellular components in the test tube or are found in the complex inflammatory context within the lungs . To determine additional mechanisms by which CCR2+Mo and/or Mo-DC mediate protection against A . fumigatus , we performed a transcriptome analysis on sorted cell populations with RNA-seq . To this end we infected CCR2 reporter mice with A . fumigatus and sorted CCR2+Mo and Mo-DC ( identified as CCR2 ( GFP+ ) , CD11b+CD11c+NK1 . 1− ) 48 h p . i . to >97% purity . CCR2+Mo present in the lung of naïve CCR2 reporter mice were sorted as a control population . We performed three independent experiments and found consistent upregulation of multiple cytokines and chemokines in response to fungal infection ( Figure 5A ) , with the highest expression of these genes in the Mo-DC subset , as confirmed by qRT-PCR ( Figure 5B ) . Cells isolated in the GFP+CD11b+CD11c+ fraction expressed genes identified as part of the core DC signature ( Figure 5A ) [52] , consistent with their designation as dendritic cells ( Mo-DCs ) . CCR2+Mo and Mo-DCs were not only capable of producing IL-12 , Nos2 and TNF upon infection but appeared to act as essential sources for these inflammatory mediators during respiratory fungal infection , since ablation of these cells in CCR2 depleter mice resulted in significantly diminished production of these factors ( Figures 5C–E ) . These findings thus suggest that CCR2+Mo and Mo-DC recruited to the lung during A . fumigatus infection express soluble factors , including cytokines ( e . g . TNF ) and effector molecules ( e . g . pentraxin-3 ) that enhance neutrophil antifungal activity . To examine whether CCR2+Mo and Mo-DCs play a direct role in conidial killing we infected CCR2 reporter with FLARE conidia to track the dynamics of pulmonary CCR2+Mo recruitment , their differentiation into Mo-DC , and their conidiacidal activity . CCR2+ cells in the lung are comprised primarily of CCR2+CD11b+Ly6C+ inflammatory monocytes ( CCR2+Mo ) that are present in the naïve lung ( Figure S2 ) and are rapidly recruited from bone marrow stores during respiratory fungal infection [38] . CCR2+Mo rapidly upregulate CD11c and MHC class II expression levels in the inflamed lung ( Figure S2 , [38] ) . To determine whether CCR2+Mo and Mo-DCs are capable of conidial killing in vivo , we first performed imaging cytometry of GFP+ cells isolated from FLARE-infected CCR2 reporter mice . We found evidence of GFP+ cells that contain viable DsRed+AF633+ conidia as well as GFP+ cells that contain killed AF633+ conidia ( Figure 6A ) . To define the relative contribution of CCR2+Mo and their derivative Mo-DCs to conidial killing in vivo , we determined the kinetics of cell recruitment ( Figure 6B ) , conidial uptake ( Figure 6C ) , and killing by flow cytometry ( Figures 6D and 6E ) . This analysis revealed that although similar numbers of CCR2+Mo and Mo-DCs were present in the lung at 36 h p . i . ( Figure 6B ) , Mo-DCs were far more likely to engulf conidia and contain killed conidia compared to CCR2+Mo ( Figures 6C and 6E ) . Analysis of conidiacidal activity on a per cell basis revealed that once conidia were internalized , CCR2+Mo and Mo-DCs were as efficient in mediating conidial killing as neutrophils ( Figures 6F–6H ) . The efficiency of conidial killing was determined by examining different fungus-engaged leukocyte populations ( neutrophils , CCR2+Mo , Mo-DCs ) and by comparing the frequencies of fungus-engaged leukocytes that contain either viable conidia or killed conidia . To examine the requirement for NADPH oxidase in Mo-DC conidiacidal activity , we generated mixed bone marrow chimeric mice that contained equal numbers of congenically marked NADPH oxidase-deficient ( p47phox ( −/− ) and –sufficient ( p47phox ( +/+ ) ) hematopoietic cells . In this host setting , NADPH oxidase-deficient and –sufficient leukocytes are isolated from and analyzed in the same inflammatory context . Similar to neutrophils , Mo-DCs employ reactive oxygen species ( ROS ) as a conidiacidal mechanism , since NADPH-deficient Mo-DCs kill conidia less effectively than NADPH-oxidase sufficient counterparts ( Figure S3 ) . Analysis of FLARE killing by Mo-DCs showed that the frequency of viable conidia in p47phox−/− Mo-DCs was higher compared to p47phox+/+ Mo-DCs ( Figure S3 ) . Despite the superior conidiacidal acitivity in p47phox+/+ Mo-DCs , there was significant killing preserved in p47phox−/− cells , indicating that conidial killing by Mo-DCs is only partially dependent on NADPH oxidase . When neutrophils and Mo-DCs were analyzed side-by-side , neutrophil conidiacidal activity was more dependent on NADPH oxidase activity than Mo-DC conidiacidal activity ( data not shown and [27] ) . These findings indicate that Mo-DCs , similar to neutrophils , employ NADPH oxidase activity as a conidiacidal mechanism . The total number of viable fungal cells in the lung of CCR2 depleter mice was significantly elevated at ( Figure 6I ) , demonstrating that lung conidiacidal activity is significantly reduced at early time points p . i . when CCR2+Mo and Mo-DCs are ablated , consistent with an essential role in innate antifungal defense in the lung . Although essential , CCR2+Mo and Mo-DCs per se are not sufficient for conidial containment since monocyte-sufficient , neutropenic mice ( anti-Ly6G treated mice ) also showed enhanced conidial survival and fungal germination in the lung ( Fig . 6I ) . In aggregate our findings are consistent with a model in which CCR2+Mo and Mo-DC derivatives are essential in preventing IA development via a non-redundant role in conidial clearance by direct killing and by regulation of neutrophil conidiacidal activity .
In this study , we uncover novel and essential functions for CCR2+ inflammatory monocytes and their derivative Mo-DCs in innate antifungal defense in the lung . The protective role of CCR2+Mo and their derivatives against A . fumigatus is not compensated by neutrophil antifungal activity . Similarly , our findings confirm the long-standing tenet that neutrophil function is essential for host defense against IA [19] , [20] , [21] , [22] . Thus , CCR2+Mo and derivative Mo-DC as well as neutrophils represent essential innate immune cells that prevent the formation of tissue-invasive hyphae and IA in the murine lung . In contrast , NK cells and other common gamma chain-dependent innate lymphocyte populations were not essential to mediate innate defense against inhaled A . fumigatus conidia , since mice deficient in these leukocyte populations contained conidial germination and did not develop invasive disease . Previous studies showed that neutrophil depletion leads to increased pulmonary recruitment of CD11b+CD11c+ TNF-producing DCs [25] . The TNF-producing DC population described by Park et al . [25] was recruited in response to enhanced CCL2 production and appears similar to Mo-DCs described in our study . The finding that ablation of CD11c-expressing cells diminished fungal clearance in this model was consistent with a protective role of one or several CD11c-expressing myeloid cell subsets in the context of neutropenia [25] . Similarly , the accumulation of CD11b+CD11c+ myeloid DCs in the lung was greater in CCR7 ( −/− ) neutropenic mice than in CCR7 ( +/+ ) neutropenic mice . This finding correlated with reduced susceptibility to IA , consistent with a protective role of CD11b+CD11c+ DCs at the site of respiratory fungal infection in neutropenic animals [53] . In our experiments , we examined the relationship between CCR2+ Mo and their derivatives and neutrophil recruitment and function in the lung . DT-treated CCR2 depleter and control mice showed similar kinetics and magnitude of neutrophil recruitment during the early phases of infection . In the respiratory A . fumigatus infection model , conidial clearance is a hallmark of the first 24 hours post-infection . In both mouse strains , the number of viable fungal cells is reduced by a factor of five to ten during this time period , with a more effective reduction in monocyte-sufficient mice compared to monocyte-depleted mice . This early difference in conidial clearance occurs despite the preserved synthesis of neutrophil chemotactic factors in the lung and the rapid accumulation of neutrophils at the site of infection . Thus , the difference in fungal CFUs among the groups likely reflects three factors: the early absence of CCR2-Mo and derivative cell conidiacidal activity , the reduction in neutrophil conidiacidal activity on a per-cell basis , and neutrophil recruitment that may be considered suboptimal since the number of viable fungal cells in the lung of CCR2 depleter mice is on average twice as high as in control mice . Essential protective functions for monocyte-derived DCs subsets have been demonstrated in other infection models [28] , [30] , [31] , [54] , [55] . In the context of systemic Listeria monocytogenes infection CCR2-dependent , TNF- and iNOS-producing DCs ( Tip-DC ) were found to play an essential role in innate defense against intracellular bacteria [31] , a finding that has been extended to other intracellular pathogens , for example Leishmania and Toxoplasma [28] , [30] , [55] , [56] . Our studies now show that the protective function for CCR2+Mo and their derivative cells is not restricted to intracellular bacteria and parasites but is also essential for innate antifungal defense . In response to A . fumigatus infection , CCR2+ Mo-DCs produced TNF and iNOS and are likely comparable to Tip-DCs induced by L . monocytogenes infection . Given the unique composition of fungal pathogens it will be important to examine how the recruitment and differentiation of CCR2+Mo into Tip-DCs is regulated by innate receptors specialized in fungal recognition . During systemic listeriosis , Tip-DCs mediate protective effects due to their role as major producers of TNF and nitric oxide [31] . The association of defects in TNF signaling with murine susceptibility to IA and of TNF inhibitors with human susceptibility to IA indicates that TNF plays a critical role in antifungal immunity in the lung . Although CCR2+Mo are a major source of TNF early during respiratory A . fumigatus infection , the precise function of TNF during conidial clearance remains to be established . It is unclear whether TNF-producing CCR2+Mo represent an important target of TNF signaling to enhance cell-intrinsic conidiacidal activity . In an ocular model of fungal keratitis , iNOS activity was dispensable for host defense against A . fumigatus in the cornea . The role of CCR2+Mo-derived nitric oxide during pulmonary fungal infection remains undefined . In both instances , the development of cell type-specific gene knockout strategies [57] will enable researchers to address these questions . Besides their role as producers of inflammatory mediators our data shows that CCR2+ Mo and Mo-DCs are crucial for direct conidial containment . Although both populations kill conidia efficiently , the frequency of Mo-DCs with engulfed conidia is far higher than that of CCR2+Mo . Thus , Mo-DCs kill a significant larger number of conidia than CCR2+ Mo . Unlike alveolar macrophages [58] , the conidiacidal activity of CCR2+Mo and their derivatives was partially dependent on NADPH oxidase . Thus , CCR2+Mo and their derivatives contribute to ROS-dependent mechanisms that are implicated in human defense against Aspergillus sp . , i . e . the susceptibility of patients with chronic granulomatous disease to IA . [23] , [26] , [59] . These findings are similar to observations in leishmaniasis , in which CCR2+Mo mediate elimination of parasites via the production of reactive oxygen species ( ROS ) [55] . During secondary responses to L . monocytogenes infection , inflammatory monocytes also represent a significant source of protective ROS [60] . Our work extends previous studies on the role of CCR2+Mo and their derivatives in trafficking fungal antigen to lung-draining lymph nodes , priming Aspergillus-specific CD4 T cells , and in inducing the development of Th1 effector cells . Taken together , our findings suggest that CCR2+Mo are required in antifungal defense as innate conidiacidal effectors and precursors of inflammatory Mo-DCs; the latter cells provide a significant reservoir of conidiacidal activity in the lung and elicit Th1 responses [37] , [38] that perpetuate a protective immune response [37] . Whether lung-resident Aspergillus-elicited Tip-DCs described in this study are identical to migratory Mo-DCs required for fungus-specific CD4 T cell priming is not clear at this time . It is possible that a subset of Tip-DCs migrate to the lung-draining lymph node for antigen transport and CD4 T cell priming or that a subset of Mo-DCs that do not produce TNF and iNOS are responsible for fungal antigen trafficking . Further studies will be required to dissect these possibilities . Although the current study addresses the role of inflammatory monocytes in a murine model , it is possible that human monocytes similarly carry out an important role in defense against IA . The antifungal capacity of human monocytes against A . fumigatus has long been recognized [46] and exogenous cytokines , including M-CSF , IFN-γ and IL-12 , enhance antifungal effects of these cells in vitro [46] , [61] , [62] . More detailed analysis of human monocyte subsets showed that CD14+CD16− monocytes could prevent conidial germination [45] . In contrast , CD14+CD16+ monocytes mounted robust inflammatory responses to conidia but did not prevent germination in vitro [45] , suggesting distinct contributions of human monocyte subsets to antifungal defense . Interestingly , CD14+CD16− monocytes express CCR2 and have been proposed to be analogous to murine CCR2+ Mo [63] . Thus , the direct conidiacidal activity observed in murine CCR2+ Mo and their derivatives in the lung is likely functionally conserved in human CD14+CD16− monocytes and their derivatives . In human neutropenic pulmonary aspergillosis there is significant pulmonary recruitment of CD1a+ DCs , which represent monocyte-derived cells [25] . Patients with autosomal dominant or sporadic deficiency in monocytes , DCs , and NK cells ( termed MonoMAC syndrome ) due to mutations in the transcription factor GATA2 are prone to disseminated nontuberculous mycobacterial infections ( incidence ∼80% ) , invasive fungal infections ( incidence ∼30% ) , primarily histoplasmosis but also aspergillosis , and to viral infections ( e . g . human papilloma virus; incidence ∼80% ) . The clinical manifestations of patients with MonoMAC syndrome support the notion that circulating myeloid cells , independent of neutrophils and tissue-resident macrophages , play an essential role in antifungal defense [64] , [65] , [66] . The ablation of circulating monocytes and monocyte-derived DCs as well as the partial loss of NK cells in CCR2 depleter mice is similar to the quantitative defects in circulating monocytes , DCs , and NK cells observed in MonoMAC patients and in both instances , hosts are vulnerable to invasive fungal disease ( this work and [21] ) . Although neutropenia has long been considered the most important risk factor for IA development in patients with hematologic malignancies and in allogeneic HCT patients , there is clinical evidence that monocytopenia represents an additional risk factor for IA development [67] , [68] . In aggregate , these lines of evidence suggest that the importance of CCR2+Mo in antifungal defense is likely not exclusive to murine models of IA , but reflective of a conserved essential function of these cells in antifungal defense .
The CCR2 depleter ( CCR2-DTR ) and CCR2 reporter ( CCR2-GFP ) strains were generated on the C57BL/6 background as previously described [38] , [69] . Control animals for CCR2+Mo-depletion experiments were sex and age-matched , non-transgenic littermates . For antibody depletion experiments , sex and age-matched C57BL/6 mice were purchased from Jackson Laboratories . RAG−/−γC−/− ( RAG-2−/−IL2rg−/− ) lymphopenic mice were purchased from Taconic . All strains were maintained and bred in the Rutgers-NJMS Cancer Center Research Animal Facility or in the Fred Hutchinson Cancer Research Center Animal Health Resources Facility under specific pathogen-free conditions . Mixed bone marrow chimeric mice were generated as described in ( Jhingran et al . , 2012 ) [27] by transferring an equal mixture of CD45 . 1+ p47phox ( +/+ ) and CD45 . 2+ p47phox ( −/− ) bone marrow cells using lethally irradiated CD45 . 1+CD45 . 2+ recipients . Recipient mice were rested for 6 weeks prior to experimental infection . Animal studies were performed following biosafety level 2 ( BSL-2 ) protocols approved by the Institutional Animal Care and Use Committee ( IACUC ) of Rutgers University and of Fred Hutchinson Cancer Research Center . For these studies , we employed an Aspergillus fumigatus-DsRed expressing strain ( Af293 . 1RFP ) [70] , a generous gift from Dr . Michelle Momany . For lung ELISA studies , we used Aspergillus fumigatus strain Af293 . A . fumigatus was cultured on Sabouraud dextrose agar ( SDA ) for 7–10 days prior to infection . Mice were challenged with 4–8×107 live conidia per mouse using a non-invasive intratracheal ( i . t . ) infection procedure as previously described [51] . The viability of A . fumigatus conidia in the inoculum was confirmed by plating serial dilutions on SDA . For assessment of fungal burden in infected mice lung single-cell suspensions were serially diluted and plated on SDA at various times after infection . For histological examination , lungs were perfused with 10 ml of PBS to remove blood and fixed in 10% buffered formalin . Fixed lung tissue was paraffin embedded and stained with modified GMS stain at the Histology Core Facility ( Rutgers-NJMS ) . For the selective removal of neutrophils , mice were injected daily with 1A8 monoclonal antibodies ( anti-Ly6G ) . Mice were injected with 500 µg i . p together with another dose of 100 µg i . t of 1A8 antibodies in order to achieve significant depletion of Ly6G+ neutrophils in the lung as previously reported [71] . Highly concentrated , purified 1A8 antibodies were isolated from ascites fluid following IACUC approved protocols ( Rutgers-RWJMS ) . For depletion of CCR2+ cells , CCR2-DTR mice and control CCR2-DTR negative littermates received 250 ng of diphtheria toxin i . p . one day prior to infection and every other day thereafter in order to maintain depletion . Diphtheria Toxin was purchased from List Biological Laboratories ( Campbell , CA ) , and reconstituted at 1 mg/ml in PBS . Aliquots were stored in −80°C . The specificity and efficiency of depletion in the lung was confirmed by flowcytometric analysis . Lung samples were minced in PBS with 3 mg/ml collagenase type IV ( Worthington ) , and were incubated at 37°C for 45 min to obtain single cells suspensions . After digestion , lung suspensions underwent RBC lysis . All antibodies were purchased from BD Biosciences . The staining protocols included combinations of the following antibodies: Gr-1 ( RB6-8C5 FITC ) , Ly6C ( AL-21 PE ) , Ly6G ( 1A8 APC ) , CD11b ( M1/70 , PerCP Cy5 . 5 ) , CD11c ( N418 Pacific Blue ) , MHC Class II I-A/I-E ( M5/11 . 415 . 2 , Alexa Fluor 700 ) , and CD45 ( 30-F11 APC-Cy7 ) . Samples were collected zon a BD LSRII Flow Cytometer and analyzed using FlowJo software . Total RNA from lungs was extracted with Trizol ( Invitrogen ) . Relative mRNA levels were determined by qRT-PCR . One microgram of total RNA was reverse transcribed using High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) . Taq Man Fast Universal PCR Master Mix ( 2× ) No Amp and TaqMan probes ( Applied Biosystems ) for each gene were used , and normalized to GAPDH . Gene expression was calculated using ΔΔCT method relative to naïve sample . For cytokine and chemokine measurements we performed ELISAs on lung homogenates according to the manufacturer's instructions . Mouse CXCL1 and CXCL2 ELISA kits were purchased from R&D systems . IL-12p70 and TNF ELISAs were obtained from BD Bioscience . CCR2GFP+CD45+CD11b+NK1 . 1−CD11c− ( CCR2+Mo ) and CCR2+CD45+CD11b+NK1 . 1−CD11c+ ( Mo-DC ) populations were isolated to more than 97% purity using a BD FACS ARIA II cell sorter dedicated for the processing of BSL-2 samples ( Flowcytometry core facility NJMS ) . Cell subsets were sorted from lung single cell suspensions obtained from A . fumigatus-infected CCR2-GFP mice that were challenged 2 days earlier . CCR2GFP+CD45+CD11b+NK1 . 1−CD11c− cells ( Mo-naïve ) were sorted from uninfected CCR2-GFP mice . DAPI was used as a viability control during sort . Immediately after sorting RNA was extracted using Qiagen RNeasy kit . One microgram of total RNA was rRNA depleted using the Ribominus Human/Mouse depletion module . Library generation and sequencing was performed by the Molecular Resource Facility at Rutgers-NJMS . Briefly , The SOLiD™ Total RNA-Seq Kit ( P/N 4445374 ) was used to convert rRNA-depleted RNA into a cDNA library for analysis on the Applied Biosystems SOLiD™ Sequencing System . The RNA was fragmented using RNase III to produce 100 to 300 base fragments which were then size selected and purified using the Purelink RNA micro kit ( Applied Biosystems , Foster City , CA ) . The yield and size distribution of fragmented RNA was confirmed using the RNA 6000 Pico Chip kit on a Bioanalyzer ( Agilent Technologies , Santa Clara , CA ) . The fragmented RNA was hybridized and ligated to Solid™ oligonucleotide adaptors and RNA ligation reagents . Reverse transcription was done using ArrayScript Reverse Transcriptase to generate the cDNA which is was then purified and size-selected using Agencourt® AMPure® XP Reagent ( Beckman Coulter , Inc . , Brea , CA ) , to ensure capture and size-selection of cDNA greater than 150 bp . The cDNA was amplified and purified using Invitrogen Purelink PCR Micro kit . The library size and concentration was confirmed using the Bioanalyzer DNA1000 kit and was used to generate template for sequencing using emulsion PCR . Three independent cell sorting and RNA sequencing reactions were performed . RNA seq results of representative genes were confirmed by qRT-PCR . The SOLiD reads were aligned to the mm9 mouse reference genome using Tophat [72] 2 . 0 . 8b and expression levels were determined using Cufflinks [73] 2 . 1 . 1 and the UCSC genome annotation . FLARE conidia were generated as described in [27] . Briefly , to generate FLARE conidia , 5×108 Af293-dsRed conidia were rotated in 0 . 5 mg/ml Biotin XX , SSE ( B-6352; Invitrogen ) in 1 ml of 50 mM carbonate buffer ( pH 8 . 3 ) for 2 hr at 4°C and labeled with 0 . 02 mg/ml AF633-streptavidin ( S-21375; Invitrogen ) in 1 ml PBS for 30 min at RT , and resuspended in PBS and 0 . 025% Tween 20 for use within 24 h . In all experiments , leukocyte conidial uptake refers to the frequency of fungus-engaged neutrophils ( dsRed+AF633++dsRed−AF633+ ) . Conidial viability within a specific leukocyte subset refers to the frequency of leukocytes that contains live conidia ( dsRed+AF633+ ) among all fungus-engaged leukocytes of the particular subset . For in vitro studies of neutrophil conidiacidal activity , neutrophils were isolated from the bone marrow of CCR2 depleter mice treated with DT for 24 hours or from DT-treated transgene-negative , littermate controls . Bone marrow cells were obtained by flushing the femurs and tibia bone cavities with PBS . Bone marrow cell suspensions were enriched for neutrophils using a density gradient-centrifugation protocol as described by Swamydas , et al [74] . BM neutrophils were cultured in the presence or absence of monocytes together with FLARE conidia at a multiplicity of infection of 1∶4 conidia to cell ratio . FLARE conidia killing was assessed at 24 hours post culture initiation as described above . For in vitro reconstitution , BM neutrophils were FACS sorted from BM of CCR2 depleter mice and cultured in the absence or presence of BM monocytes that were FACS sorted from CCR2-GFP as GFP+CD11b+Ly6C+Ly6G−NK1 . 1− cells . Monocytes were cultured at 1∶4 ratio relative to neutrophils numbers to reflect the ratios of these cells seen in vivo . The studies performed were governed by protocol 10094E1213 as approved by the IACUC committee of New Jersey Medical School and by protocol 1813 as approved by the IACUC committee at the Fred Hutchinson Cancer Research Center . Animal studies were compliant with all applicable provisions established by the Animal Welfare Act and the Public Health Services ( PHS ) Policy on the Humane Care and Use of Laboratory Animals . | Despite the significant impact of fungal infections to human health our understanding of immunity to these pathogens remains incomplete . Human mycoses are associated with high morbidity and mortality , even with modern antifungal therapies . Aspergillus fumigatus is the most common etiologic agent of invasive aspergillosis ( IA ) , a serious infection that develops in immunodeficient patients . In this study we employ a combination of cell ablation strategies to examine the role of CCR2+Ly6C+ inflammatory monocytes ( CCR2+Mo ) in innate responses against a pulmonary infection with A . fumigatus conidia . We find that CCR2+Mo and their derivative dendritic cells ( Mo-DCs ) are required for defense against IA and that mice lacking these cells succumb to infection with A . fumigatus . Our studies indicate that CCR2+Mo and Mo-DCs exert crucial innate antifungal defense by two main mechanisms: 1 ) CCR2+Mo and Mo-DCs are a significant source of inflammatory mediators that augment the killing capacity of neutrophils and 2 ) conidial uptake by CCR2+Mo is coincident with their differentiation into Mo-DCs that directly kill fungal conidia via partially NADPH oxidase-dependent mechanisms . In aggregate , our studies find a novel essential function for CCR2+Mo in innate defense against a pulmonary fungal pathogen by mediating indirect and direct containment of fungal cells at the portal of infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"aspergillosis",
"immune",
"cells",
"monocytes",
"immunity",
"innate",
"immunity",
"immune",
"defense",
"immunology",
"immunity",
"to",
"infections",
"biology",
"fungal",
"diseases"
] | 2014 | Inflammatory Monocytes Orchestrate Innate Antifungal Immunity in the Lung |
There is unequivocal evidence in the literature that epidemics adversely affect the livelihoods of individuals , households and communities . However , evidence in the literature is dominated by the socioeconomic impacts of HIV/AIDS and malaria , while evidence on the impact of the Ebola virus disease ( EVD ) on households’ livelihoods remains fragmented and scant . Our study investigates the effect of the EVD epidemic on the livelihoods of Liberian households using the Sustainable Livelihood Framework ( SLF ) . The study also explores the effect of the EVD epidemic on agricultural production and productive efficiency of farm households using Spatial Stochastic Frontier Analysis ( SSFA ) . We collected data from 623 households across Liberia in 2015 , using a systematic random sampling design . Our results indicated that the annual income of sample households from communities where EVD occurred did not differ from the annual income of households from communities where EVD did not occur . Nonetheless , the majority of sample households reported a decrease in their income , compared to their income in the year before the survey . This suggests that the impact of the EVD epidemic might not only have been limited to communities directly affected by the epidemic , but also it may have indirectly affected communities in areas where EVD was not reported . We also found that the community-level incidence of EVD negatively affected crop production of farm households , which may have exacerbated the problem of food insecurity throughout the country . Moreover , we found that the EVD epidemic weakened the society’s trust in Liberian institutions . In a nutshell , our results highlight that epidemics , such as the recent EVD outbreak , may have long-lasting negative effects on the livelihoods of a society and their effect may extend beyond the communities directly affected by the epidemics . This means that the nation’s recovery from the impact of the epidemic would be more challenging , and the social and economic impacts of the epidemic may extend well beyond the end of the health crisis .
There is a plethora of evidence in the literature that epidemics such as HIV/AIDS and malaria have profound implications for the livelihoods of the affected society . The impact of HIV/AIDS on livelihoods has been intensively investigated and there is universal consensus that HIV/AIDS adversely affects the livelihoods of individuals , households and communities [1–2] , and has macro-level implications for poverty , economic growth , unemployment and political stability [3–7] . Similarly , malaria has been found to have a strong negative effect on the socioeconomic status of households [8–11] . In contrast , the socioeconomic impacts of the Ebola virus disease ( EVD ) epidemic , specifically the most recent and largest ever EVD epidemic recorded in West Africa from 2014–2016 , have not been systematically analyzed . Due to the differences in transmission mechanisms , latency , and mortality rate between EVD and other infectious tropical diseases , such as HIV/AIDS and malaria [12–14] , EVD outbreaks likely impact the livelihoods of a society differently . For example , EVD can wipe out an entire family or village within a relatively short period of time . In areas affected by EVD , economic activities may cease completely , as people no longer work on their fields , nor trade or even travel ( because of check points and travel restrictions ) [15] . HIV/AIDS infections and resultant mortalities , on the other hand , occur over a longer time period; and as such , their effects on livelihoods and the economy are more subtle at first . HIV/AIDS and malaria result in higher costs in terms of the opportunity cost of the time spent caring for the sick household member [16–18] , medical expenses and , for the unlucky ones , funeral expenses . In contrast , EVD lowers livelihood outcomes by weakening the ability of the households to earn their living rather than by increasing the expenditure on the sick person and funeral ceremonies . For example , the medical costs of the EVD epidemic in Liberia were mostly covered by the government and the international community , as the epidemic presented a global health emergency . These and other disease-specific characteristics necessitate specific research to investigate the effect of different infectious diseases on the livelihood of the affected societies . In light of this , we analyze the impact of the recent EVD epidemic on the livelihoods of the Liberian society . There are few studies that explored the impact of the EVD epidemic on the agricultural sector in Liberia [19–20] . These studies reported that the EVD epidemic negatively affected employment in the agricultural sector . At the peak of the epidemic , almost half of the country’s labor force was out of work [19–20] . Farmers were less likely to work on their farms during the EVD epidemic [19] . These studies found that most of the households returned to their farms during the survey , which was conducted from October , 21 to November , 7 , 2014 , and concluded that the impact of the EVD crisis on the agricultural sector may not have been as severe as predicted [19–20] . Nonetheless , these studies used phone surveys to collect data , which could have resulted in selection bias , as households without mobile phones were systematically excluded from the survey . In addition , the emphasis of the studies was focused on the impacts of the EVD epidemic on the employment in the agricultural sector . However , employment and whether or not households were working on their farm tell only part of the story . Even if households were working on their farms during the EVD epidemic , the productivity of their agricultural inputs , and hence their efficiency may have been compromised by the epidemic . Therefore , our study provides evidence on how the Ebola crisis affected the efficiency of farm households , and the concomitant effects on agricultural production in Liberia . Moreover , our study explores the impact of the EVD epidemic on livelihoods of the Liberian society using the sustainable livelihood framework ( SLF ) [21–22] . The EVD epidemic affects the livelihoods of individuals , households and communities by weakening the household assets upon which the households’ ability to enhance their livelihood , depends [21] . These assets can broadly be categorized into five categories: physical capital ( e . g . , infrastructure , tools , equipment ) , human capital ( e . g . , knowledge and ability to work ) , financial capital ( e . g . , available stocks , access to financial services , regular inflows of money ) , social capital ( e . g . , networks for cooperation , trust , support ) and natural capital ( e . g . , land , forests , water ) [21–24] . Shocks , such as epidemics , that weaken some or all of these household assets , negatively affect livelihood [2 , 25] . Therefore , a complete understanding of the effects of epidemics , such as EVD , on the livelihood of households requires the investigation of their influence on the assets owned by the households . The EVD epidemic may have weakened the resource base of the Liberian society for various reasons . First , the incidence of the epidemic in the households and/or their communities may have weakened the different categories of assets owned by households directly [2 , 24] . Second , measures taken by the Liberian government to contain the spread of the disease may have further dampened the households’ assets and affected their welfare [15 , 26 , 27] . For example , the government declared a state of emergency and established quarantine zones in most of the affected communities . Schools and markets were closed in several districts and communities . Restrictions on domestic and international travels were imposed [15 , 19 , 26 , 27] . Thus , the mobility restrictions and complete closure of markets might have considerably hampered the livelihoods of individuals , households and communities by reducing their access to different livelihood assets [15 , 26] . Third , fear of contracting the disease may have coerced people into avoiding social gatherings and participation in different activities and organizations , thereby weakening the social capital that the Liberian society possesses [27] . Furthermore , during crises , people may increasingly seek support from different social networks such as friends , families , the community , and the government , but the prospect of receiving the needed help may have been significantly hampered by an epidemic , which strains the social cohesion . Stigmatization of survivors of the disease may also contribute to the degradation of the social capital . Our study uses a systematic nationwide random sampling design to explore the effect of the EVD epidemic on the livelihood assets possessed by Liberian households , and the livelihood outcomes they achieved during the EVD epidemic . In addition , we emphasize the effect of the epidemic on the agricultural sector . We are interested in the effect of the EVD epidemic on agricultural production for two main reasons . First , the agricultural sector plays a crucial role in Liberia and contributes more than 35% of the country’s GDP [28] . Moreover , the majority of Liberians live in rural areas ( 50% ) , and are primarily engaged in the agricultural sector to earn a living [20] . Almost 80% of rural households and 18% of urban households are agricultural households in Liberia [29] . Second , we are not aware of any study that examined the effect of the recent EVD epidemic on agricultural production in Liberia or elsewhere accounting for the productive efficiency of farm households .
Liberia is one of the poorest countries in Africa with per capita GDP of $457 . 9 as of 2014 [30] . Nevertheless , after the end of the civil war in 2003 , the country’s economy has been steadily growing . For instance , the economy grew by an annual growth rate of 8% in 2006 and 8 . 7% in 2013 [30] . Although most of the population lives in poverty ( 68 . 6% of the population lives on less than $1 . 9 a day ) , Liberia had one of the fastest growing economies in Africa over the past 10 years [31] . This gain of momentum in terms of macro-economic performance was disrupted in 2014 by the epidemic of EVD [32] , which affected the country from 2014 to 2016 and resulted in the tragic loss of 4 , 809 lives ( 45% of reported cases ) [33] . All 15 counties reported incidences of the disease , but the severity of the epidemic varied from place to place [33] . For example , based on the number of EVD cases per 1 , 000 inhabitants , counties such as Margibi , Montserrado , Grand Cape Mount , and Bomi were more severely affected by the epidemic than Grand Gedeh , River Gee , Sineo and Maryland [33] ( Fig 1 ) . We conducted a systematic nationwide household survey from February to June , 2015 . The survey was administered in person by trained Liberian enumerators . Sample households were randomly selected following a “random walk” procedure [e . g . , 34 , 35] . Starting from the center of a village/town/city ( hereafter ‘interview location’ ) , the enumerators walked in different directions and randomly selected households to be interviewed . We aimed at interviewing 5–10 households per location , depending on their size ( i . e . , more interviews in larger locations ) . The random walk technique was used to reduce non-response rates , as the enumerators would walk until they found enough households that were willing to participate in the interview . This method is particularly useful in sensitive times like the EVD crisis , when people may have been more reluctant to interact with strangers out of fear of contracting the EVD . Nonetheless , we acknowledge that in larger locations , the random walk method may have resulted in some biases , as households closer to the center of locations were more likely to be sampled than those living farther away from the centers . Whenever possible , household heads were selected and interviewed . A total of 623 sample households were interviewed . We collected data through face-to-face interviews with sample households across Liberia using a questionnaire ( S6 Appendix ) . We were granted permission to conduct the survey by Liberian authorities after careful evaluation of staff safety , data collection procedures and agreements on data sharing ( see S7 Appendix ) . Additional data on EVD deaths and cases were obtained from the World Health Organization ( WHO ) and the Liberia Institute of Statistics and Geo-information Services ( LISGIS ) . To explore the impact of EVD on the livelihoods of Liberian society , we applied the Sustainable Livelihoods Framework ( SLF ) used by the Department for International Development-United Kingdom ( DFID-UK ) [21–22] . We used the SLF as it enables us to understand not only the effects of the EVD epidemic on livelihood outcomes , but also the mechanisms driving the effects of the epidemic [21–22] . Thus , our study provides important insight into policies aiming to avert or reduce the impact of future epidemics on livelihoods by addressing the important factors that drive these effects . Although there is a scarcity of studies that applied this framework to investigate the impact of EVD , several studies employed the SLF to explore the livelihood effects of other epidemics , mainly HIV/AIDS [1 , 2] . Traditionally , the SLF has been applied to understand differential capabilities of rural families to cope with stresses or shocks [36] and their ability to achieve sustainable livelihoods . A livelihood is defined as a means of living , and the assets required to achieving it [21–22] . The different types of assets that a household needs to achieve better livelihood outcomes are broadly categorized as human , physical , financial , social and natural capitals [21 , 22 , 24] . Hence , the likelihood of a household to achieve improved livelihood outcomes ( such as income , food security and others ) depends on its access to different livelihood assets . The livelihood of a household is deemed sustainable when it copes with , and recovers from stresses without compromising the abilities of future generations [21] . Factors that weaken some or all of the livelihood assets of households , adversely affect their livelihood [2 , 25] . Therefore , a complete understanding of the effect of shocks , such as epidemics , on the livelihood of households necessitates the investigation of their influence on the assets owned by the households . Employing the SLF , we investigate the impact of the EVD epidemic on different categories of assets possessed by Liberian households , and the livelihood outcomes they achieved during the EVD epidemic ( i . e . , in the 12 months preceding our survey ) . Here , we focused on the effect of the EVD epidemic on three categories of household capitals ( natural , financial , and social capitals ) , and their total income and agricultural production . See Table 1 for the definition of the household assets included in our study . To analyze the data , we used descriptive statistics , a factor analysis and regression models . We used Stata 10 . 1 [53] and R 3 . 2 . 4 [54] for our analysis . In our analysis self-reported incidence of EVD in the community of the respondent was used as a proxy for EVD occurrence . Respondents were asked whether they knew anybody in their own or nearby communities who had contracted EVD . We used self-reported incidence of EVD in one’s community instead of EVD cases reported by WHO , as the publicly available WHO reports are at the spatial resolution of the county-level and not community-level . In our understanding , community-level incidence might be more relevant for household-level decisions , because the EVD incidence at the community spatial scale may have a larger impact on household livelihood than the EVD occurrence at the county spatial scale . Our analysis is divided into two parts . First , we analyzed the effect of the EVD epidemic on livelihood outcomes , such as total annual household income and agricultural production , using descriptive statistics and a regression analysis . Second , we analyzed the impact of the epidemic on the households’ assets . Here , we used descriptive statistics to summarize the effect of the EVD epidemic on the resource bases of households , and a regression analysis to analyze changes in social capital . For the summary of the methods used see S1 Appendix Table A . In our regression analysis , we employed spatial econometric models to account for potential spatial dependencies in the data using the “spdep” package in R [55] . Before we ran the models , we tested for the significance of the spatial correlation of the residuals obtained from the mixed-effects models ( with random slopes and intercepts ) using Moran’s test ( see S1 Appendix Table D ) . The test results suggested that there was a significant spatial correlation in our data , justifying the use of spatial econometric models . We further checked for the suitability of spatial-lagged models and spatial-error models , but found no significant difference between these models . Hence , we report results from spatial-lagged models here and include the results from the spatial-error models in the appendix ( see S1 Appendix Table E Models 3 and 6; and Table G model 3 ) . To estimate the effect of the EVD crisis on agricultural production , we used a spatial stochastic production frontier model with the “ssfa” package in R [39] . The application of the stochastic frontier model ( instead of the classical linear regression model ) was justified after testing for the significance of the existence of inefficiencies in our agricultural production data using the likelihood ratio test ( LR test: chi-bar square ( 1 ) = 6 . 396 , p = 0 . 006 ) . Finally , we conducted power analyses and found that the power of our tests ranged from 73–98% , indicating a relatively low probability of conducting Type II error .
The majority of our respondents ( 90% ) were male with an average age of 43 years . Almost 64% of the respondents were literate , which is comparable to 66 . 7% reported by LISGIS [56] . Respondents had attended school for an average of six years . Sample farm households owned , on average , 1 . 53 hectares of farm land , which is similar to the average farm size of 1 . 54 hectares reported by the FAO [57] . Forty-two percent of the sample respondents were from urban areas and 58% from rural areas ( for more detail see S1 Appendix Table B ) . Thirty-one percent of the sample households reported the incidence of EVD in their community or nearby communities and almost 20% of these were households that relied heavily on farming for their livelihoods . Our study offers important insights on the effect of the EVD epidemic on the livelihoods of Liberian society and the mechanisms underlying them . We found that the incidence of EVD did not influence the total annual household income depending on whether or not the households were located in/near communities where EVD occurred . However , the majority of the sample households reported that their income was lower during the EVD crisis , as compared to their income before the outbreak . This suggests that the effects of the EVD epidemic were not limited to the communities where EVD occurred , but that the EVD crisis affected communities throughout the country . These results are in line with the findings of Bowles et al . ( 2016 ) [63] who reported that during the EVD epidemic , there was a remarkable decline in economic activities across Liberia , but that in most cases , there was little association between the decline in economic activities and the number of Ebola cases . Thus , post-epidemic rehabilitation measures should not only be limited to communities directly affected by EVD , but should also target those indirectly affected by the epidemic . We also found that the incidence of Ebola significantly reduced total crop production of farm households , which is in line with other studies [see also 15 , 26 , 58 , 64] . As farm households in our sample had consumed about 85% of their own production , they heavily depended on their own agricultural production for survival , which is a typical characteristic of subsistence farmers . Thus , the reduction in their agricultural production likely had an adverse effect on their food security , which is in line with the existing literature [15 , 20 , 26 , 35 , 58 , 64] . Most of these studies reported that the agricultural sector was one of the most severely affected sectors by EVD , and food security was significantly hampered by the epidemic in Liberia . Our results also revealed that the incidence of EVD negatively affected the trust of the citizens in Liberian institutions . Respondents who reported the incidence of EVD in/near their community , were more likely to report a decrease in their trust in the government and village chief ( s ) . Our results suggest that in the long term , the deterioration in the social capital resulting from the EVD epidemic , may have adverse effects on the stability of the country’s political system [60] . Degradation of social capital may increase the likelihood of social conflict and crime . Moreover , distrust in state institutions may render recovery efforts more challenging and make the country more susceptible to future outbreaks , as citizens may no longer comply with the recommendations of the state institutions [27] . Finally , although we controlled for most of the relevant factors in our analysis , there may still be some confounding effects , as our results were based on cross-sectional data collected at a single point in time . In addition , as it is customary to the surveys of our type , there could be limitations associated with retrospective memory , as respondents may not have accurately recalled information from the previous year , though we believe that one year is short enough for respondents to accurately report the events in their household . We believe that our results could help Liberia and other countries in the developing world with similar socioeconomic conditions , as well as the international community , to be better prepared for future crises and distribute livelihood rehabilitation efforts more effectively , thereby facilitating the affected nation’s speedy recovery after such crises . | Epidemics such as HIV/AIDS , malaria and Ebola virus disease ( EVD ) may adversely impact the livelihoods of the society affected by the epidemics . Nonetheless , the mechanism behind the effects of the epidemics may differ depending on different factors , such as the transmission mechanisms , latency , and mortality rates associated with the diseases , which requires specific research to investigate the effect of each epidemic . In light of this , we analyzed the impact of the recent EVD epidemic on the agricultural production of farm households and its impact on the livelihoods of Liberian society . We collected data from 623 households throughout Liberia during the EVD crisis in 2014–2016 , and found that there was no significant difference in the annual income of sample households from communities where EVD occurred and did not occur . Nonetheless , the majority of the sample households reported a decrease in their income compared to the year before our survey . We also found that the community level incidence of EVD had a significant negative effect on crop production of farm households , which might have exacerbated food insecurity in the country . Moreover , the EVD epidemic negatively affected the Liberian society’s trust in Liberian institutions . Our results underline that epidemics , like EVD , might have long-lasting negative effects on the livelihoods of a society , and they may have adverse effect beyond the communities directly affected by the epidemics . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"and",
"discussion"
] | [
"livestock",
"medicine",
"and",
"health",
"sciences",
"geographical",
"locations",
"tropical",
"diseases",
"social",
"sciences",
"parasitic",
"diseases",
"social",
"epidemiology",
"farms",
"crops",
"africa",
"agricultural",
"production",
"crop",
"science",
"epidemiology",
"economics",
"agriculture",
"people",
"and",
"places",
"finance",
"liberia",
"biology",
"and",
"life",
"sciences",
"malaria"
] | 2018 | The impact of the Ebola virus disease (EVD) epidemic on agricultural production and livelihoods in Liberia |
We use stochastic neural field theory to analyze the stimulus-dependent tuning of neural variability in ring attractor networks . We apply perturbation methods to show how the neural field equations can be reduced to a pair of stochastic nonlinear phase equations describing the stochastic wandering of spontaneously formed tuning curves or bump solutions . These equations are analyzed using a modified version of the bivariate von Mises distribution , which is well-known in the theory of circular statistics . We first consider a single ring network and derive a simple mathematical expression that accounts for the experimentally observed bimodal ( or M-shaped ) tuning of neural variability . We then explore the effects of inter-network coupling on stimulus-dependent variability in a pair of ring networks . These could represent populations of cells in two different layers of a cortical hypercolumn linked via vertical synaptic connections , or two different cortical hypercolumns linked by horizontal patchy connections within the same layer . We find that neural variability can be suppressed or facilitated , depending on whether the inter-network coupling is excitatory or inhibitory , and on the relative strengths and biases of the external stimuli to the two networks . These results are consistent with the general observation that increasing the mean firing rate via external stimuli or modulating drives tends to reduce neural variability .
Consider a pair of mutually coupled ring networks labeled j = 1 , 2 . Let uj ( θ , t ) denote the activity at time t of a local population of cells with stimulus preference θ ∈ [−π , π ) in network j . Here θ could represent the direction preference of neurons in area-middle temporal cortex ( MT ) [10] , the orientation preference of V1 neurons , after rescaling θ → θ/2 [26 , 27] , or a coordinate in spatial working memory [22 , 28 , 29] . For concreteness , we will refer to θ as a direction preference . The variables uj evolve according to the neural field equations [21 , 22 , 30 , 31] τ d u 1 ( θ , t ) = [ - u 1 ( θ , t ) + ∫ - π π J 1 ( θ - θ ′ ) f ( u 1 ( θ ′ , t ) ) d θ ′ + ϵ ∫ - π π K 1 ( θ - θ ′ ) f ( u 2 ( θ ′ , t ) ) d θ ′ + ϵ h 1 ( θ ) ] d t + 2 ϵ d W 1 ( θ , t ) ( 1a ) τ d u 2 ( θ , t ) = [ - u 2 ( θ , t ) + ∫ - π π J 2 ( θ - θ ′ ) f ( u 2 ( θ ′ , t ) ) d θ ′ + ϵ ∫ - π π K 2 ( θ - θ ′ ) f ( u 1 ( θ ′ , t ) ) d θ ′ + ϵ h 2 ( θ ) ] d t + 2 ϵ d W 2 ( θ , t ) ( 1b ) where ϵ is a constant scale factor ( see below ) , Jj ( θ − θ′ ) is the distribution of intra-network connections between cells with stimulus preferences θ′ and θ in network j , Kj ( θ − θ′ ) is the corresponding distribution of inter-network connections to network j , and hj ( θ ) is an external stimulus . The firing rate function is assumed to be a sigmoid f ( u ) = f 0 1 + e - γ ( u - η ) , ( 2 ) with maximal firing rate f0 , gain γ and threshold η . The final term on the right-hand side of each equation represents external additive noise , with Wj ( θ , t ) a θ-dependent Wiener process . In particular , E [ d W j ( θ , t ) ] = 0 , E [ d W i ( θ , t ) d W j ( θ ′ , s ) ] = δ i , j C j ( θ - θ ′ ) δ ( t - t ′ ) d t d t ′ , ( 3 ) where δ ( t ) is the Dirac delta function and δi , j is the Kronecker delta function . For concreteness , we will take C ( θ ) = aδ ( θ ) + b cos ( θ ) for constants a , b . For b ≠ 0 , the noise is colored in θ ( which is necessary for the solution to be spatially continuous ) and white in time . ( One could also take the noise to be colored in time by introducing an additional Ornstein-Uhlenbeck process . For simplicity , we assume that the noise processes in the two networks are uncorrelated , which would be the case if the noise were predominantly intrinsic . Correlations would arise if some of the noise arose from shared fluctuating inputs . For a discussion of the effects of correlated noise in coupled ring networks see [22] . ) The external stimuli are taken to be weakly biased inputs of the form ϵ h j with h j = h ¯ j cos ( θ - θ ¯ j ) , ( 4 ) where θ ¯ j is the location of the peak of the input ( stimulus bias ) and h ¯ j is the contrast . Finally , the time-scale is fixed by setting the time constant τ = 10 msec . The maximal firing rate f0 is taken to be 100 spikes/sec . The weight distributions are 2π-periodic and even functions of θ and thus have cosine series expansions . Following [21] , we take the intra-network recurrent connections to be J j ( θ - θ ′ ) = J ¯ j cos ( θ - θ ′ ) , ( 5 ) which means that cells with similar stimulus preferences excite each other , whereas those with sufficiently different stimulus preferences inhibit each other . It remains to specify the nature of the inter-network connections . As we have already mentioned , we consider two different network configurations ( see Fig 1 ) : ( A ) a vertically connected two layer or laminar model and ( B ) a horizontally connected single layer model . In model A , the inter-network weight distribution is taken to have the form K j ( θ - θ ′ ) = E j + K ¯ j cos ( θ - θ ′ ) , ( 6 ) which represents vertical coupling between the layers . We also assume that both layers receive the same stimulus bias , that is , θ ¯ 1 = θ ¯ 2 = θ ¯ in Eq ( 4 ) . In model B , the inter-network weight distribution represents patchy horizontal connections , which tend to link cells with similar stimulus preferences [32–35] . This is implemented by taking K j ( θ - θ ′ ) = K ¯ j δ ( θ - θ ′ ) . ( 7 ) Now the two networks can be driven by stimuli with different biases so that θ ¯ 1 ≠ θ ¯ 2 . Note that in order to develop the analytical methods of this paper , we scale the internetwork coupling , the noise terms and the external stimuli in Eq ( 1 ) by the constant factor ϵ . Taking 0 < ϵ ≪ 1 ( weak noise , weak inputs and weak inter-network coupling ) will allow us to use perturbation methods to derive explicit parameter-dependent expressions for neural variability . We do not claim that cortical networks necessarily operate in these regimes , but use the weakness assumption to obtain analytical insights and make predictions about the qualitative behavior of neural variability . In the case of weak inter-network connections , the validity of the assumption is likely to depend on the source of these connections . For example , in model B , they arise from patchy horizontal connections within superficial or deep layers of cortex , which are known to play a modulatory role [36] . On the other hand , vertical connections between layers are likely to be stronger than assumed in our modeling analysis , at least in the feedforward direction [37] . Finally , the weak stimulus assumption depends on a particular view of how cortical neurons are tuned to stimuli , which is based on the theory of ring attractor networks , see the Discussion .
Let us begin by considering stimulus-dependent neural variability in a single ring network evolving according to the stochastic neural field equation d u ( θ , t ) = [ - u ( θ , t ) + ∫ - π π J ( θ - θ ′ ) f ( u ( θ ′ , t ) ) d θ ′ + ϵ h ( θ ) ] d t + 2 ϵ d W ( θ , t ) , ( 12 ) where E [ d W ( θ , t ) ] = 0 , E [ d W ( θ , t ) d W ( θ ′ , t ′ ) ] = C ( θ - θ ′ ) δ ( t - t ′ ) d t d t ′ , ( 13 ) with J ( θ ) = J ¯ cos θ , h ( θ ) = h ¯ cos θ , C ( θ ) = a δ ( θ ) + b cos θ . A clear demonstration of the suppressive effects of an external stimulus can be seen from direct numerical simulations of Eq ( 12 ) , see Fig 2 . In the absence of an external stimulus , the center-of-mass ( phase ) of the bump diffuses on the ring , whereas it exhibits localized fluctuations when a weakly-biased stimulus is present . Clearly , the main source of neural variation is due to the wandering of the bump , which motivates the amplitude phase decomposition given by Eq ( 9 ) . Applying the perturbation analysis of Materials and Methods yields a one-network version of the phase Eq ( 10 ) , which takes the form d β ( t ) = - ϵ Λ sin β ( t ) d t + 2 ϵ D d w ( t ) , ( 14 ) with Λ = h ¯ / A and D = C ¯ / 2 A 2 , where A is the amplitude of the bump for ϵ = 0 . Eq ( 14 ) is known as a von Mises process , which can be regarded as a circular analog of the Ornstein-Uhlenbeck process on a line , and generates distributions that frequently arise in circular or directional statistics [45] . The von Mises process has been used to model the trajectories of swimming organisms [46 , 47] , oscillators in physics [48] , bioinformatics [49] , and the data fitting of neural population tuning curves [50] . ( Nonlinear stochastic phase equations analogous to ( 14 ) also arise in models of ring attractor networks with synaptic heterogeneities , which have applications to spatial working memory [23 , 51 , 52] ) . Introduce the probability density p ( β , t | β 0 , 0 ] d β = P [ β < β ( t ) < β + d β | β ( 0 ) = β 0 ] . This satisfies the forward Fokker-Planck equation ( dropping the explicit dependence on initial conditions ) ∂ p ( β , t ) ∂ t = ∂ ∂ β [ ϵ Λ sin ( β ) p ( β , t ) ] + ϵ D ∂ 2 p ( β , t ) ∂ β 2 ( 15 ) for β ∈ [−π , π] with periodic boundary conditions p ( −π , t ) = p ( π , t ) . It is straightforward to show that the steady-state solution of Eq ( 15 ) is the von Mises distribution p ( β ) = M ( β ; 0 , κ ) , κ = h ¯ ϵ A D , ( 16 ) with M ( β ; β * , κ ) ≔ 1 2 π I 0 ( κ ) exp ( κ cos ( β - β * ) ) . ( 17 ) Here I0 ( κ ) is the modified Bessel function of the first kind and zeroth order ( n = 0 ) , where I n ( κ ) = 1 2 π ∫ - π π exp ( κ cos θ ) cos ( n θ ) d θ . Sample plots of the von Mises distribution are shown in Fig 3 . One finds that M ( β; β* , κ ) → 1/2π as κ → 0; since κ ∼ h ¯ this implies that in the absence of an external stimulus one recovers the uniform distribution of pure Brownian motion on the circle . On the other hand , the von Mises distribution becomes sharply peaked as κ → ∞ . More specifically , for large positive κ , M ( β ; β * , κ ) ≈ 1 2 π σ 2 e - ( β - β * ) 2 / 2 σ 2 , σ 2 = κ - 1 . ( 18 ) We thus have an explicit example of the noise suppression of fluctuations by an external stimulus , since σ 2 ∝ 1 / h ¯ . ( We are assuming that the time for the distribution of the stochastic phase variable to reach steady-state is much shorter than the time for the amplitude-phase decomposition ( 9 ) to break down . This can be proven rigorously using variational methods for sufficiently small ϵ , since the time for a large transverse fluctuation becomes exponentially large [44] ) . Moments of the von Mises distribution are usually calculated in terms of the circular moments of the complex exponential x = eiβ = cos β + i sin β . The nth circular moment is defined according to μ n = ⟨ z n ⟩ κ , β * = ∫ - π π z n M ( β ; β * , κ ) d β = I n ( κ ) I 0 ( κ ) e i n β * . ( 19 ) In particular , ⟨ cos β ⟩ κ , β * = I 1 ( κ ) I 0 ( κ ) cos β * , ⟨ sin β ⟩ κ , β * = I 1 ( κ ) I 0 ( κ ) sin β * . ( 20 ) We can use these moments to explore stimulus-dependent variability in terms of the stochastic wandering of the bump or tuning curve . That is , consider the leading order approximation u ( θ , t ) ≈ A cos ( θ + β ( t ) ) , with β ( t ) evolving according to the von Mises SDE ( 14 ) . Trial-to-trial variability can be captured by averaging the solution with respect to the stationary von Mises density ( 16 ) . First , ⟨ U ⟩ ( θ ) = A ∫ - π π cos ( θ + β ) M ( β , 0 , κ ) d β = A [ ⟨ cos β ⟩ κ , 0 cos θ - ⟨ sin β ⟩ κ , 0 sin θ ] ≔ A ( κ ) cos θ , A ( κ ) = A I 1 ( κ ) I 0 ( κ ) . ( 21 ) from Eq ( 20 ) . Hence , the mean amplitude A ( κ ) is given by the first circular moment of the von Mises distribution , see inset of Fig 3 . When κ = 0 ( zero external stimulus ) , the amplitude vanishes due to the fact that the random position of the bump is uniformly distributed around the ring . As the stimulus contrast h ¯ increases the wandering of the bump is more restricted and A ( κ ) monotonically increases . Second , ⟨ U 2 ⟩ ( θ ) = A 2 ∫ - π π cos 2 ( θ + β ) M ( β , 0 , κ ) d β = A 2 2 ∫ - π π ( 1 + cos ( 2 [ θ + β ] ) ) M ( β , 0 , κ ) d β = A 2 2 [ 1 + ⟨ cos 2 β ⟩ κ , 0 cos 2 θ - ⟨ sin 2 β ⟩ κ , 0 sin 2 θ ] = A 2 2 [ 1 + I 2 ( κ ) I 0 ( κ ) cos 2 θ ] . It follows that the variance is var ( U ) = A 2 2 [ 1 + I 2 ( κ ) I 0 ( κ ) cos 2 θ - 2 ( I 1 ( κ ) I 0 ( κ ) cos θ ) 2 ] = A 2 2 { 1 - ( I 1 ( κ ) I 0 ( κ ) ) 2 - [ ( I 1 ( κ ) I 0 ( κ ) ) 2 - I 2 ( κ ) I 0 ( κ ) ] cos 2 θ } ( 22 ) In Fig 4 ( a ) , we show example plots of the normalized variance var ( U ) /A2 as a function of the parameter κ , which is a proxy for the input amplitude h ¯ , since κ ∝ h ¯ . It can be seen that our theoretical analysis reproduces the various trends observed in [10]: ( i ) a global suppression of neural variability that increases with the stimulus contrast; ( ii ) a directional tuning of the variability that is bimodal; ( iii ) a peak in the suppression of cells at the preferred directional selectivity . One difference between our theoretical results and those of [10] is that , in the latter case , the directional tuning of the variance is not purely sinusoidal . Part of this can be accounted for by noting that we consider the variance of the activity variable u rather than the firing rate f ( u ) . Moreover , for analytical convenience , we take the synaptic weight functions etc . to be first-order harmonics . In Fig 4 ( b ) we show numerical plots of the variance in the firing rate , which exhibits the type of bimodal behavior found in [10] when the ring network operates in the marginal regime . We now turn to a pair of coupled ring networks that represent vertically connected layers as shown in Fig 1 ( a ) ( model A ) , with inter-network weight distribution ( 6 ) . For analytical tractability , we impose the symmetry conditions A1 = A2 = A and K ¯ 1 = K ¯ 2 = K ¯ . However , we allow the contrasts of the external stimuli to differ , h ¯ 1 ≠ h ¯ 2 . Also , without loss of generality , we set θ ¯ 1 = θ ¯ 2 = 0 . Eq ( 10 ) then reduce to the form , see Materials and methods d β 1 = - ϵ Λ 1 sin ( β 1 ) d t - ϵ K ¯ sin ( β 1 - β 2 ) d t + 2 ϵ d w 1 ( t ) , ( 23a ) d β 2 = - ϵ Λ 2 sin ( β 2 ) d t - ϵ K ¯ sin ( β 2 - β 1 ) d t + 2 ϵ d w 2 ( t ) , ( 23b ) with E [ d w j ( t ) ] = 0 , E [ d w j ( t ) d w k ( t ′ ) ] = δ j k D j δ ( t - t ′ ) d t ′ d t . ( 24 ) Given our various simplifications , we can rewrite Eq ( 23 ) in the more compact form d β j = - ϵ ∂ Φ ( β 1 , β 2 ) ∂ β j d t + 2 ϵ d w j ( t ) , j = 1 , 2 ( 25 ) where Φ is the potential function Φ ( β 1 , β 2 ) = - Λ 1 cos ( β 1 ) - Λ 2 cos ( β 2 ) - K ¯ cos ( β 1 - β 2 ) . ( 26 ) Introduce the joint probability density p ( β 1 , β 2 , t | β 1 , 0 , β 2 , 0 , 0 ] d β 1 d β 2 = P [ β j < β j ( t ) < β j + d β j , j = 1 , 2 | β j ( 0 ) = β j , 0 , j = 1 , 2 ] . This satisfies the two-dimensional forward Fokker-Planck equation ( dropping the explicit dependence on initial conditions ) ∂ p ( β 1 , β 2 , t ) ∂ t = ϵ ∑ j = 1 , 2 ∂ ∂ β j [ ∂ Φ ( β 1 , β 2 ) ∂ β j p ( β 1 , β 2 , t ) ] + ϵ ∑ j = 1 , 2 D j ∂ 2 p ( β 1 , β 2 , t ) ∂ β j 2 ( 27 ) for βj ∈ [−π , π] and periodic boundary conditions p ( −π , β2 , t ) = p ( π , β2 , t ) , p ( β1 , −π , t ) = p ( β1 , π , t ) . The existence of a potential function means that we can solve the time-independent FP equation . Setting time derivatives to zero , we have ∑ j = 1 , 2 ∂ J j ∂ β j = 0 , J j = ϵ ∂ Φ ∂ β j p + ϵ D j ∂ p ∂ β j , where Jj is a probability current . In the stationary state the probability currents are constant , but generally non-zero . However , in the special case D1 = D2 = D , then there exists a steady-state solution in which the currents vanish . This can be seen by rewriting the vanishing current conditions as J j = ϵ p ∂ ∂ β j [ Φ + ϵ D ln p ] = 0 . This yields the steady-state probability density , which is a generalization of the von Mises distribution , p ( β 1 , β 2 ) = N - 1 e - Φ ( β 1 , β 2 ) / ϵ D = N - 1 exp ( κ 1 cos ( β 1 ) + κ 2 cos ( β 2 ) + χ cos ( β 1 - β 2 ) ) ≔ M 2 ( β 1 , β 2 ; κ 1 , κ 2 , χ ) , ( 28 ) where κ j = h ¯ j ϵ A j D ≥ 0 , χ = K ¯ ϵ D , and N is the normalization factor N ( κ 1 , κ 2 , χ ) = ∫ - π π ∫ - π π exp ( κ 1 cos ( β 1 ) + κ 2 cos ( β 2 ) + χ cos ( β 1 - β 2 ) ) d β 1 d β 2 . ( 29 ) The distribution M2 ( β1 , β2;κ1 , κ2 , χ ) is an example of a bivariate von Mises distribution known as the cosine model [49] . The normalization factor can be calculated explicitly to give N ( κ 1 , κ 2 , χ ) = ( 2 π ) 2 [ I 0 ( κ 1 ) I 0 ( κ 2 ) I 0 ( χ ) + 2 ∑ s = 1 ∞ ( - 1 ) s I s ( κ 1 ) I s ( κ 2 ) I s ( χ ) ] . ( 30 ) The corresponding marginal distribution for β1 is p ( β 1 ) = ∫ - π π p ( β 1 , β 2 ) d β 2 = N ( κ 1 , κ 2 , χ ) - 1 2 π I 0 ( κ 13 ( β 1 ) ) exp ( κ 2 cos ( β 1 ) ) , ( 31 ) where κ 13 ( β ) 2 = κ 1 2 + χ 2 + 2 κ 1 χ cos β . An analogous result holds for the marginal density p ( β2 ) . We now summarize a few important properties of the cosine bivariate von Mises distribution [49]: We will assume that the vertical connections are maximal between neurons with the same stimulus preference so that K ¯ ≥ 0 and χ ≥ 0 . It then follows that p ( β1 , β2 ) is unimodal . Moreover , from Eq ( 32 ) we have Σ = 1 κ 1 κ 2 + χ ( κ 1 + κ 2 ) ( κ 2 + χ χ χ κ 1 + χ ) . ( 33 ) For zero inter-network coupling ( χ = 0 ) , we obtain the diagonal matrix Σ = diag ( κ 1 - 1 , κ 2 - 1 ) and we recover the variance of the single ring networks , that is , var ( β j ) = κ j - 1; there are no interlaminar correlations . On the other hand , for χ > 0 we find two major effects of the interlaminar connections . First , the vertical coupling reduces fluctuations in the phase variables within a layer . This is most easily seen by considering the symmetric case κ1 = κ2 = κ for which Σ = 1 κ ( κ + 2 χ ) ( κ + χ χ χ κ + χ ) . ( 34 ) Clearly , var ( β j ) = 1 κ κ + χ κ + 2 χ < κ - 1 . ( 35 ) ( This result is consistent with a previous study of the effects of inter-network connections on neural variability , which focused on the case of zero stimuli and treated the bump positions as effectively evolving on the real line rather than a circle [22] . In this case , inter-network connections can reduce the variance in bump position , which evolves linearly with respect to the time t . ) The second consequence of interlaminar connections is that they induce correlations between the phase β1 ( t ) and β2 ( t ) . Having characterized the fluctuations in the phases β1 ( t ) and β2 ( t ) , analogous statistical trends will apply to the trial-to-trial variability in the tuning curves . This follows from making the leading-order approximation uj ( x , t ) ∼ A cos ( θ + βj ( t ) ) , and then averaging the βj with respect to the bivariate von Mises density M2 ( β1 , β2; κ1 , κ2 , χ ) . In the large κj regime , this could be further simplified by averaging with respect to the bivariate normal distribution under the approximations cos ( β ) ≈ 1 − β2/2 and sin β ∼ β . Both the mean and variance of the tuning curves are similar to the single ring network , see Eqs ( 21 ) and ( 22 ) : ⟨ U 1 ⟩ ( θ ) = A ⟨ cos β 1 ⟩ cos θ , ( 36 ) and var ( U 1 ) = A 2 2 [ 1 - ⟨ cos β 1 ⟩ 2 - [ ⟨ cos β 1 ⟩ 2 - ⟨ cos 2 β 1 ⟩ ] cos 2 θ ] ( 37 ) Their dependence on the coupling strength χ and input parameter κ1 = κ2 = κ is illustrated in Fig 5 . Finally , ⟨ U 1 ( θ ) U 2 ( θ ′ ) ⟩ = A 2 ∫ - π π ∫ - π π cos ( θ + β 1 ) cos ( θ ′ + β 2 ) M 2 ( β 1 , β 2 ; κ , κ , χ ) d β 1 d β 2 = A 2 ∫ - π π ∫ - π π ( cos θ cos β 1 - sin θ sin β 1 ) × ( cos θ ′ cos β 2 - sin θ ′ sin β 2 ) M 2 ( β 1 , β 2 ; κ , κ , χ ) d β 1 d β 2 = A 2 ( cos θ cos θ ′ ⟨ cos β 1 cos β 2 ⟩ + sin θ sin θ ′ ⟨ sin β 1 sin β 2 ⟩ ) = A 2 ( cos θ cos θ ′ ⟨ cos β 1 cos β 2 ⟩ + sin θ sin θ ′ ⟨ sin β 1 sin β 2 ⟩ ) . so that inter-network covariance take the form ⟨ U 1 ( θ ) U 2 ( θ ′ ) ⟩ - ⟨ U 1 ( θ ) ⟩ ⟨ U 2 ( θ ′ ) ⟩ = A 2 cos θ cos θ ′ [ ⟨ cos β 1 cos β 2 ⟩ - ⟨ cos β 1 ⟩ ⟨ cos β 2 ⟩ ] + A 2 sin θ sin θ ′ ⟨ sin β 1 sin β 2 ⟩ . In particular , for θ = θ′ we have ⟨ U 1 ( θ ) U 2 ( θ ) ⟩ - ⟨ U 1 ( θ ) ⟩ ⟨ U 2 ( θ ) ⟩ = A 2 2 [ ⟨ sin β 1 sin β 2 ⟩ + ⟨ cos β 1 cos β 2 ⟩ - ⟨ cos β 1 ⟩ ⟨ cos β 2 ⟩ ] - A 2 2 [ ⟨ sin β 1 sin β 2 ⟩ - ⟨ ( cos β 1 cos β 2 ⟩ - ⟨ cos β 1 ⟩ ⟨ cos β 2 ⟩ ) ] cos ( 2 θ ) . ( 38 ) The resulting correlation tuning curve behaves in a similar fashion to the variance , see Fig 5 ( c ) , where corr ( U 1 , U 2 ) = ⟨ U 1 ( θ ) U 2 ( θ ) ⟩ - ⟨ U 1 ( θ ) ⟩ ⟨ U 2 ( θ ) ⟩ var ( U 1 ) var ( U 2 ) . ( 39 ) ( Note that our definition of the cross-correlation function differs from that used , for example , by Churchland et al [9] . These authors consider the covariance matrix of simultaneous recordings of spike counts obtained using a 96-electrode array . The matrix is then decomposed into a network covariance matrix and a diagonal matrix of private single neuron noise . Our definition involves pairwise correlations between the activity of two distinct populations . Nevertheless , consistent with the findings of Churchland et al . [9] , we find that the cross-correlations decrease in the presence of a stimulus ) . The above qualitative analysis can be confirmed by numerical simulations of the full neural field Eq ( 1 ) , as illustrated in Fig 6 ( a ) –6 ( d ) for a pair of identical ring networks . In Fig 6 ( e ) –6 ( h ) , we show corresponding results for the case where network 2 receives a weaker stimulus than network 1 ( h ¯ 1 = 2 and h ¯ 2 = 0 . 5 ) . In the absence of interlaminar connections , the phase of network 2 fluctuates much more than the phase of network 1 . When interlaminar connections are included , fluctuations are reduced , but network 2 still exhibits greater variability than network 1 . This latter result is consistent with an experimental study of neural variability in V1 [24] , which found that neural correlations were more prominent in superficial and deep layers of cortex , but close to zero in input layer 4 . One suggested explanation for these differences is that layer 4 receives direct feedforward input from the LGN . Thus we could interpret network 1 in model A as being located in layer 4 , whereas network 2 is located in a superficial layer , say . Our final example concerns a pair of coupled ring networks that represent horizontally connected hypercolumns within the same superficial layer , say , as shown in Fig 1 ( b ) ( model B ) , with inter-network weight distribution ( 7 ) . Again , for analytical tractability , we impose the symmetry conditions A1 = A2 = A and K ¯ 1 = K ¯ 2 = K ¯ . However , unlike model A , we take the contrasts to be the same , h ¯ 1 = h 2 ¯ = h ¯ , but allow the biases of the two inputs to differ , θ ¯ 1 ≠ θ ¯ 2 . Eq ( 10 ) become , see Materials and methods d β 1 = - ϵ Λ sin ( β 1 + θ ¯ 1 ) d t + ϵ K ( β 1 - β 2 ) d t + 2 ϵ d w 1 ( t ) , ( 40a ) d β 2 = - ϵ Λ sin ( β 2 + θ ¯ 2 ) d t + ϵ K ( β 2 - β 1 ) d t + 2 ϵ d w 2 ( t ) , ( 40b ) with wj ( t ) given by Eq ( 24 ) and K ( β ) = 2 K ¯ Γ ∫ - π π f ′ ( A cos ( θ - β ) ) sin ( θ - β ) f ( A cos ( θ ) ) d θ . ( 41 ) We can rewrite K ( β ) in the form K ( β ) = - 2 K ¯ A | Γ | ∂ ϕ ∂ β , ϕ ( β ) = f ( A cos ( θ - β ) ) f ( A cos ( θ ) ) . ( 42 ) Note that ϕ ( −β ) = ϕ ( β ) and thus ϕ′ ( −β ) = −ϕ′ ( β ) . A sample plot of the potential ϕ ( β ) is shown in Fig 7 ( a ) , together with an approximate curve fitting based on a von Mises distribution . For the given firing rate parameters η = 0 . 5 and γ = 4 , the unperturbed bump amplitude is A ≈ 1 . 85 . As in the case of model A , we can rewrite Eq ( 40 ) in the more compact form d β j = - ϵ ∂ Ψ ( β 1 , β 2 ) ∂ β j d t + 2 ϵ d w j ( t ) , j = 1 , 2 ( 43 ) where Ψ is the potential function Ψ ( β 1 , β 2 ) = - Λ cos ( β 1 + θ ¯ 1 ) - Λ cos ( β 2 + θ ¯ 2 ) - K ¯ ϕ ( β 1 - β 2 ) , ( 44 ) and we have absorbed the factor 2/ ( A|Γ| ) into the constant K ¯ . The corresponding two-dimensional forward Fokker-Planck equation is ∂ p ( β 1 , β 2 , t ) ∂ t = ϵ ∑ j = 1 , 2 ∂ ∂ β j [ ∂ Ψ ( β 1 , β 2 ) ∂ β j p ( β 1 , β 2 , t ) ] + ϵ ∑ j = 1 , 2 D j ∂ 2 p ( β 1 , β 2 , t ) ∂ β j 2 ( 45 ) for βj ∈ [−π , π] and periodic boundary conditions p ( −π , β2 , t ) = p ( π , β2 , t ) , p ( β1 , −π , t ) = p ( β1 , π , t ) . Following the analysis of model A , if D1 = D2 = D then the stationary density takes the form p ( β 1 , β 2 ) = M - 1 e - Ψ ( β 1 , β 2 ) / ϵ D = M - 1 exp ( κ cos ( β 1 + θ ¯ 1 ) + κ cos ( β 2 + θ ¯ 2 ) + χ ϕ ( β 1 - β 2 ) ) , ( 46 ) where κ = h ¯ ϵ A D ≥ 0 , χ = K ¯ ϵ D , and M is a normalization factor . Long-range horizontal connections within superficial layers of cortex are mediated by the axons of excitatory pyramidal neurons . However , they innervate both pyramidal neurons and feedforward interneurons so that they can have a net excitatory or inhibitory effect , depending on stimulus conditions [36 , 53 , 54] , More specifically , they tend to be excitatory at low contrasts and inhibitory at high contrasts . Suppose that ring network 1 represents a hypercolumn driven by a stimulus h ¯ cos θ and network 2 represents a hypercolumn driven by a stimulus h ¯ cos ( θ - θ ¯ ) , see Fig 1 ( b ) . In Fig 7 ( b ) and 7 ( c ) we plot how the normalized maximal mean and variance of network 1 ( at θ = ±π/2 ) varies with the directional bias θ ¯ of the input to network 2 . We also show the baseline mean and variance in the absence of horizontal connections ( χ = 0 ) . It can be seen that the mean and variance covary in opposite directions . In particular , for inhibitory horizontal connections ( χ < 0 ) the variance is facilitated relative to baseline when the two stimuli have similar biases ( θ ¯ ≈ 0 ) and is suppressed when they are sufficiently different ( θ ¯ ≈ ± π ) . The converse holds for excitatory horizontal connections ( χ > 0 ) . In the Discussion , these results will be explored within the context of surround modulation .
In order to utilize perturbation methods , we assumed that the ring networks were driven by weakly biased stimuli . This assumption depends on a particular view of how cortical neurons are tuned to stimuli . Consider the most studied example , which involves orientation tuning of cells in V1 . The degree to which recurrent processes contribute to the receptive field properties of V1 neurons has been quite controversial over the years [55–58] . The classical model of Hubel and Wiesel [59] proposed that the orientation preference and selectivity of a cortical neuron in input layer 4 arises primarily from the geometric alignment of the receptive fields of thalamic neurons in the lateral geniculate nucleus ( LGN ) projecting to it . ( Orientation selectivity is then carried to other cortical layers through vertical projections ) . This has been confirmed by a number of experiments [60–64] . However , there is also significant experimental evidence suggesting the importance of recurrent cortical interactions in orientation tuning [65–71] . One issue that is not disputed is that some form of inhibition is required to explain features such as contrast-invariant tuning curves and cross-orientation suppression [58] . The uncertainty in the degree to which intracortical connections contribute to orientation tuning of V1 neurons is also reflected in the variety of models . In ring attractor models [26 , 27 , 72 , 73] , the width of orientation tuning of V1 cells is determined by the lateral extent of intracortical connections . Recurrent excitatory connections amplify weakly biased feedforward inputs in a way that is sculpted by lateral inhibitory connections . Hence , the tuning width and other aspects of cortical responses are primarily determined by intracortical rather than thalamocortical interconnections . On the other hand , in push-pull models , cross-orientation inhibition arises from feedforward inhibition from interneurons [62 , 74] . Finally , in normalization models , a large pool of orientation-selective cortical interneurons generates shunting inhibition proportional in strength to the stimulus contrast at all orientations [75] . In the end , it is quite possible that are multiple circuit mechanisms for generating tuned cortical responses to stimuli , which could depend on the particular stimulus feature , location within a feature preference map , and cortical layer [58] . Surround modulation ( SM ) refers to the phenomenon in which stimuli in the surround of a neuron’s receptive field ( RF ) modulate the neuron’s response to stimuli simultaneously presented inside the RF . SM is a fundamental property of sensory neurons in many species and sensory modalities , and is thought to play an important role in contextual image processing . As with mechanisms of orientation tuning , there is considerable debate over whether feedforward or intracortical circuits generate SM , and whether this results from increased inhibition or reduced excitation [19 , 36 , 53 , 54 , 76–82] . SM has been characterized in many species , commonly using circular grating patches of increasing radius or grating patches confined to the RF surrounded by annular gratings , and varying systematically the grating parameters . Modulatory effects are typically quantified in terms of changes in the mean firing rates of single neurons recorded from the center . Some of the main features of SM in V1 are as follows ( see [36] and references therein ) : ( i ) SM is spatially extensive . For example , in primates , modulatory effects from the surround ( both facilitatory and suppressive ) can be evoked at least 12 . 5 degrees away from a neuron’s RF center . ( ii ) SM is tuned to specific stimulus parameters . The strongest suppression is induced by stimuli in the RF and surround of the same orientation , spatial frequency , drift direction , and speed , and weaker suppression or facilitation is induced by stimuli of orthogonal parameters ( e . g . , orthogonally oriented stimuli or stimuli drifting in opposite directions ) . ( iii ) SM is contrast dependent . Surround stimulation evokes suppression when the center and surround stimuli are of high contrast , but can be facilitatory when they are of low contrast . One way to interpret the results of model B is to treat networks 1 and 2 as hypercolumns driven by center and surround stimuli , respectively . SM is then mediated by the horizontal connections that can have a net excitatory or inhibitory effect , depending on stimulus conditions . Here , for simplicity , we impose the sign of the horizontal connections by hand . However , one could develop a more detailed model that implements the switch between excitation and inhibition using , for example , high threshold interneurons [54] . The major prediction of our analysis is that whenever the surround modulation suppresses ( facilitates ) the center firing rate , the corresponding variance is facilitated ( suppressed ) . One of the main simplifications of our neural field model is that we do not explicitly distinguish between excitatory and inhibitory populations . This is a common approach to the analysis of neural fields , in which the combined effects of excitation and inhibition are incorporated using , for example , Mexican hat functions [83–85] . In the case of the ring network , the spontaneous formation of population orientation tuning curves or bumps is implemented using a cosine function , which represents short-range excitation and longer-range inhibition around the ring . We note , however , that the methods and results presented in this paper could be extended to the case of separate excitatory and inhibitory populations , as well as different classes of interneuron , as has been demonstrated elsewhere for deterministic neural fields [27 , 54] . One major difference between scalar and E-I neural fields is that the latter can also exhibit time-periodic solutions , which would add an additional phase variable associated with shifts around the resulting limit cycle . The effects of noise on limit cycle oscillators can be analyzed in an analogous fashion to wandering bumps [86 , 87] . We also note that neural variability in a two-population ( E-I ) stabilized supralinear network has been analyzed extensively using linear algebra [20] . Another possible extension of our work would be to consider higher-dimensional neural fields . For example , one could replace the ring attractor on S1 by a spherical attractor on S2 . In the latter case , marginally stable modes would correspond to rotations of the sphere . ( Mathematically speaking , this corresponds to the action of the Lie group SO ( 3 ) rather than SO ( 2 ) for the circle . ) One could generalize the Fourier analysis of the ring network by using spherical harmonics , as previously shown for deterministic neural field models of orientation and spatial frequency tuning in V1 [88 , 89] . One could also consider a planar neural field with Euclidean-symmetric weights , for which marginally stable modes would be generated by the Euclidean group of rigid body transformations of the plane ( translations , rotations and reflections . ) However , this example is more difficult since the marginally stable manifold is non-compact , and one cannot carry out a low-dimensional harmonic reduction . In order to obtain analytical results , one has to use Heaviside rate functions [30 , 90] . A third possible extension would be to develop a more detailed model of the laminar structure of cortex . Roughly speaking , cortical layers can be grouped into input layer 4 , superficial layers 2/3 and deep layers 5/6 [37 , 91–93] . They can be distinguished by the source of afferents into the layer and the targets of efferents leaving the layer , the nature and extent of intralaminar connections , the identity of interneurons within and between layers , and the degree of stimulus specificity of pyramidal cells . In previous work , we explored the role of cortical layers in the propagation of waves of orientation selectivity across V1 [94] , under the assumption that deep layers are less tuned to orientation . This suggests considering coupled ring networks that differ in their tuning properties . Another modification would be to consider asymmetric coupling between layers , both in terms of the range of coupling and its strength . Interestingly , the properties of SM also differ across cortical layers , suggesting different circuits and mechanisms generating SM in different layers . More specifically , surround fields in input layer 4 are smaller than in other layers , and SM is weaker and untuned for orientation . Moreover , SM is stronger and more sharply orientation-tuned in superficial layers compared to deep layers [36] . Therefore , it would be interesting to consider coupled ring networks that combine models A and B . One final comment is in order . Neural variability in experiments is typically specified in terms of the statistics of spike counts over some fixed time interval , and compared to an underlying inhomogeneous Poisson process . Often Fano factors greater than one are observed . In this paper , we worked with stochastic firing rate models rather than spiking models , so that there is some implicit population averaging involved . In particular , we focused on the statistics of the variables uj ( x , t ) , which represent the activity of local populations of cells rather than of individual neurons , with f ( uj ) the corresponding population firing rate [30] . This allowed us to develop an analytically tractable framework for investigating how neural variability depends on stimulus conditions within the attractor model paradigm . In order to fit a neural field model to single-neuron data , one could generate spike statistics by taking f ( uj ) to be the rate of an inhomogeneous Poisson process . Since f ( uj ) is itself stochastic , this would result in a doubly stochastic Poisson process , which is known to produce Fano factors greater than unity [95] . Moreover , the various phenomena identified in this paper regarding stimulus-dependent variability would carry over to a spiking model , at least qualitatively . However , one should not expect a mean-field reduction to capture everything in a spiking model . For example , multivariate doubly stochastic Poisson processes can have correlations between their spike times in addition to the correlations induced by shared rate fluctuations . Spiking network models typically do produce these spike timing correlations that are not captured by most mean-field reductions , even those that account for correlated firing rate fluctuations [13 , 15 , 96–98] . These correlations could , in turn , affect auto-correlation and firing rates in the network .
First , suppose that there are no external inputs , no inter-network coupling ( J12 = J21 = 0 ) , and no noise ( ϵ = 0 ) . Each network can then be described by a homogeneous ring model of the form ∂ u ( θ , t ) ∂ t = - u ( θ , t ) + ∫ - π π J ( θ - θ ′ ) f ( u ( θ ′ , t ) ) d θ ′ . ( 47 ) Let J ( θ ) = J ¯ cos θ and consider the trial solution u ( θ , t ) = U ( θ ) with U ( θ ) an even , unimodal function of θ centered about θ = 0 . This could represent a direction tuning curve in MT ( ( in the marginal regime ) or a stationary bump encoding a spatial working memory . It follows that U ( θ ) satisfies the integral equation U ( θ ) = J ¯ ∫ - π π cos ( θ - θ ′ ) f ( U ( θ ′ ) ) d θ ′ . ( 48 ) Substituting the cosine series expansion cos ( θ - θ ′ ) = cos ( θ ) cos ( θ ′ ) + sin ( θ ) sin ( θ ′ ) ( 49 ) into the integral equation yields the even solution U ( θ ) = A cos θ with the amplitude A satisfying the self-consistency condition A = J ¯ ∫ - π π cos ( θ ) f ( U ( θ ) ) d θ = J ¯ g ( A ) . ( 50 ) The amplitude Eq ( 50 ) can be solved explicitly in the large gain limit γ → ∞ , for which f ( u ) → H ( u − κ ) , where H is the Heaviside function [21] . That is , A = 1 + κ ± 1 - κ , corresponding to a marginally stable large amplitude wide bump and an unstable small amplitude narrow bump , consistent with the original analysis of Amari [90] . On the other hand , at intermediate gains , there exists a single stable bump rather than an unstable/stable pair of bumps , see Fig 8 . Linear stability of the stationary solution can be determined by considering weakly perturbed solutions of the form u ( θ , t ) = U ( θ ) + ψ ( θ ) eλt for |ψ ( θ ) | ≪ 1 . Substituting this expression into Eq ( 47 ) , Taylor expanding to first order in ψ , and imposing the stationary condition ( 48 ) yields the infinite-dimensional eigenvalue problem [27] ( λ + 1 ) ψ ( θ ) = ∫ - π π J ( θ - θ ′ ) f ′ ( U ( θ ′ ) ) ψ ( θ ′ ) d θ ′ . ( 51 ) This can be reduced to a finite-dimensional eigenvalue problem by applying the expansion ( 49 ) : ( λ + 1 ) ψ ( θ ) = A cos ( θ ) + B sin ( θ ) , ( 52 ) where A = J ¯ ∫ - π π cos ( θ ) f ′ ( U ( θ ) ) ψ ( θ ) d θ , B = J ¯ ∫ - π π sin ( θ ) f ′ ( U ( θ ) ) ψ ( θ ) d θ . ( 53 ) Substituting Eqs ( 52 ) into ( 53 ) then gives the matrix equation [21] ( λ + 1 ) ( A B ) = J ¯ ( I [ cos 2 θ ] I [ cos θ sin θ ] I [ cos θ sin θ ] I [ sin 2 θ ] ) ( A B ) , ( 54 ) where for any periodic function v ( θ ) I [ v ( θ ) ] = ∫ - π π v ( θ ) f ′ ( U ( θ ) ) d θ . ( 55 ) Integrating Eq ( 50 ) by parts shows that for A ≠ 0 I [ sin 2 θ ] = ∫ - π π sin 2 θ f ′ ( U ( θ ) ) d θ = 1 / J ¯ . Hence , exploiting the fact that I is a linear functional of v , I [ cos 2 θ ] = I [ 1 - sin 2 θ ] = I [ 1 ] - I [ sin 2 θ ] = I [ 1 ] - 1 / J ¯ . Finally , integration by parts establishes that I [ cos θ sin θ ] = ∫ - π π cos θ sin θ f ′ ( U ( θ ) ) d θ = - ∫ - π π sin θ f ( U ( θ ) ) d θ = 0 , since U ( θ ) is even . Eq ( 54 ) now reduces to ( λ + 1 ) ( A B ) = J ¯ ( I [ 1 ] - 1 / J ¯ 0 0 1 ) ( A B ) , ( 56 ) which yields the pair of solutions λ 0 = 0 , λ e = 2 [ J ¯ ∫ 0 π f ′ ( U ( θ ) ) d θ - 1 ] . ( 57 ) The zero eigenvalue is a consequence of the fact that the bump solution is marginally stable with respect to uniform shifts around the ring; the generator of such shifts is the odd function sinθ . The other eigenvalue λe is associated with the generator , cosθ , of expanding or contracting perturbations of the bump . Thus linear stability of the bump reduces to the condition λe < 0 . This can be used to determine the stability of the pair of bump solutions in the high-gain limit [21] . ( Note that there also exist infinitely many eigenvalues that are equal to −1 , which form the essential spectrum . However , since they lie in the left-half complex λ-plane , they do not affect stability ) . A variety of previous studies have shown how breaking the underlying translation invariance of a homogeneous neural field by introducing a nonzero external input stabilizes wave and bump solutions to translating perturbations [21 , 99–102] . For the sake of illustration , suppose that h ( θ ) = h ¯ cos ( θ ) in the deterministic version of Eq ( 1 ) . This represents a weak θ-dependent input with a peak at θ = 0 . Extending the previous analysis , one finds a stationary bump solution U ( θ ) = A cos θ + ϵ h ¯ cos θ , with A satisfying the implicit equation A = J ¯ ∫ - π π cos θ f ( A cos θ + ϵ h ¯ cos θ ) d θ . Again , this can be used to determine both the width and amplitude of the bump in the high-gain limit . Furthermore , the above analysis can be extended to establish that , for weak inputs , the bump is stable ( rather than marginally stable ) with respect to translational shifts [21] . The amplitude phase decompositions ( βj , vj ) defined by Eq ( 9 ) are not unique , so additional mathematical constraints are needed , and this requires specifying the allowed class of functions of vj ( the appropriate Hilbert space ) . We will take take vj ∈ L2 ( S1 ) , that is , vj ( θ ) is a periodic function with ∥ v j∥ 2 = 〈 v j , v j 〉 = ∫ - π π v j ( θ ) 2 d θ < ∞ . Substituting the decomposition into the stochastic neural field Eq ( 1 ) and using Ito’s lemma gives [103] U 1 ′ ( θ + β 1 ) d β 1 + 1 2 U 1 ′ ′ ( θ + β 1 ) d β 1 2 + ϵ d v 1 ( θ , t ) = [ - U 1 ( θ + β 1 ) - ϵ v 1 ( θ , t ) + ∫ - π π J 1 ( θ - θ ′ ) f ( U 1 ( θ ′ + β 1 ) + ϵ v 1 ( θ ′ , t ) ) d θ ′ ] d t + [ ∫ - π π ϵ K 1 ( θ - θ ′ ) f ( U 2 ( θ ′ + β 2 ) + ϵ v 2 ( θ ′ , t ) ) d θ ′ + ϵ h 1 ( θ ) ] d t + 2 ϵ d W 1 ( θ , t ) . and U 2 ′ ( θ + β 2 ) d β 2 ( t ) + 1 2 U 2 ′ ′ ( θ + β 2 ) d β 2 2 + ϵ d v 2 ( θ , t ) = [ - U 2 ( θ + β 2 ) - ϵ v 2 ( θ , t ) + ∫ - π π J 2 ( θ - θ ′ ) f ( U 2 ( θ ′ + β 2 ) + ϵ v 2 ( θ ′ , t ) ) d θ ′ ] d t + [ ∫ - π π ϵ K 2 ( θ - θ ′ ) f ( U 1 ( θ ′ + β 1 ) + ϵ v 1 ( θ ′ , t ) ) d θ ′ + ϵ h 2 ( θ ) ] d t + 2 ϵ d W 2 ( θ , t ) . Introduce the series expansions v j = v j , 0 + ϵ v j , 1 + O ( ϵ ) , Taylor expanding the nonlinear function F , imposing the stationary solution ( 48 ) , and dropping all O ( ϵ ) terms . This gives [21 , 30] , after dropping the zero index on vj , 0 , ϵ d v 1 ( θ , t ) = ϵ L β 1 1 v 1 ( θ , t ) d t + ϵ K ^ 1 ( θ + β 2 ) d t + ϵ h 1 ( θ ) d t + 2 ϵ d W 1 ( θ , t ) , - U 1 ′ ( θ + β 1 ) d β 1 ( 58a ) ϵ d v 2 ( θ , t ) = ϵ L β 2 2 v 2 ( θ , t ) d t + ϵ K ^ 2 ( θ + β 1 ) + ϵ h 2 ( θ ) d t + 2 ϵ d W 2 ( θ , t ) - U 2 ′ ( θ + β 2 ) d β 2 , ( 58b ) where L β j are the following linear operators L β j v ( θ , t ) = - v ( θ , t ) + ∫ - π π J j ( θ - θ ′ ) f ′ ( U j ( θ ′ + β ) ) v ( θ ′ , t ) d θ ′ , ( 59 ) and K ^ 1 ( θ + β ) = ∫ - π π K 1 ( θ - θ ′ ) f ( U 2 ( θ ′ + β ) ) d θ ′ , K ^ 2 ( θ + β ) = ∫ - π π K 2 ( θ - θ ′ ) f ( U 1 ( θ ′ + β ) ) d θ ′ . ( 60 ) It can be shown that the operator L 0 j has a 1D null space spanned by U j ′ ( θ ) . The fact that U j ′ ( θ ) belongs to the null space follows immediately from differentiating Eq ( 48 ) with respect to θ . Moreover , U j ′ ( θ ) is the generator of uniform translations around the ring , so that the 1D null space reflects the marginal stability of the bump solution . ( Marginal stability of the bump means that the linear operator L 0j has a simple zero eigenvalue while the remainder of the discrete spectrum lies in the left-half complex plane . The spectrum is discrete since S1 is a compact domain . ) This then implies a pair of solvability conditions for the existence of bounded solutions of Eq ( 58a ) , namely , that dvj is orthogonal to all elements of the null space of the adjoint operator L β j j † . The corresponding adjoint operator is L β j † v ( θ , t ) = - v ( θ , t ) + f ′ ( U j ( θ + β ) ) ∫ - π π J j ( θ - θ ′ ) v ( θ ′ , t ) d θ ′ . ( 61 ) Let V j ( θ ) span the 1D adjoint null space of L 0 † . Now taking the inner product of both sides of Eq ( 58a ) with respect to V j ( θ + β j ) and using translational invariance then yields the following SDEs to leading order: d β 1 = ϵ H 1 ( β 1 ) d t - ϵ K 1 ( β 1 - β 2 ) d t + 2 ϵ d w 1 ( t ) , ( 62a ) d β 2 = ϵ H 2 ( β 1 ) d t - ϵ K 2 ( β 2 - β 1 ) d t + 2 ϵ d w 2 ( t ) , ( 62b ) where H j ( β ) = Γ j - 1 ∫ - π π V j ( θ ) h j ( θ - β ) d θ , ( 63 ) for Hj ( β + 2π ) = Hj ( β ) , K j ( β ) = Γ j - 1 ∫ - π π V j ( θ ) K ^ j ( θ + β ) d θ , ( 64 ) and Γ j = ∫ - π π V j ( θ ) U j ′ ( θ ) d θ , ( 65 ) Here wj ( t ) are scalar independent Wiener processes , E [ d w j ( t ) ] = 0 , E [ d w j ( t ) d w k ( t ′ ) ] = δ j , k D j δ ( t - t ′ ) d t ′ d t , with D j = 1 Γ j 2 ∫ - π π ∫ - π π V j ( θ ) V j ( θ ′ ) C j ( θ - θ ′ ) d θ ′ d θ . ( 66 ) Note that stochastic phase equations similar to ( 62 ) were previously derived in [21 , 22] , except that the functions Hj ( β ) and K j ( β ) were linearized , resulting in a system of coupled Ornstein-Uhlenbeck ( OU ) processes: d β 1 = ϵ ν 1 β 1 d t - ϵ r 1 ( β 1 - β 2 ) d t + 2 ϵ d w 1 ( t ) , ( 67a ) d β 2 = ϵ ν 2 β 2 d t - ϵ r 2 ( β 2 - β 1 ) d t + 2 ϵ d w 2 ( t ) , ( 67b ) for constant coefficients ν1 , ν2 , r1 , r2 . Properties of one-dimensional OU processes were then used to explore how the variance in the position of bump solutions depended on inter-network connections and statistical noise correlations . However , it should be noted that the variables βj ( t ) are phases on a circle ( rather than positions on the real line ) , so that the right-hand side of Eq ( 67 ) should involve 2π-period functions . Therefore , the linear approximation only remains accurate on sufficiently short times scales for which the probability of either of the phases winding around the circle is negligible . In order to illustrate this point , consider an uncoupled OU process evolving according to d β j = ϵ ν j β j d t + 2 ϵ d w j ( t ) . A standard analysis shows that [103] ⟨ β j ( t ) ⟩ = β 0 e - ν j t , ⟨ β j ( t ) 2 ⟩ - ⟨ β j ( t ) ⟩ 2 = ϵ D j ν j [ 1 - e - 2 ν j t ] . In particular , the variance approaches a constant ϵD/2νj in the large t limit . The corresponding density is given by the Gaussian ρ ( β , t | β 0 , 0 ) = 1 2 π ϵ D j [ 1 - e - 2 ν j t ] / ν j exp ( - ( β - β 0 e - k t ) 2 2 ϵ D j [ 1 - e - 2 ν j t ] / ν j ) . Although the linear approximation is sufficient if one is interested in estimating the diffusivity Dj , which was the focus of [21 , 22] , it does not yield the correct steady-state distribution on the ring in the limit t → ∞ . Indeed , for vj → 0 , the density of the OU process converges point-wise to zero , whereas ρ ( β , t ) → 1/2π on the ring . In our paper , we are interested in the full steady-state densities rather than just the diffusivities Dj . In order to determine the functions Hj and K j we need to obtain explicit expressions for the null vectors V j . We will take h j ( θ ) = h ¯ j cos ( θ - θ ¯ j ) . Applying the expansion ( 49 ) to the adjoint equation L 0 j † V j = 0 with L 0 j † defined by Eq ( 61 ) , we can write [21] V j ( θ ) = f ′ ( U j ( θ ) ) [ C j cos θ + S j sin θ ] , with C j = J ¯ j ∫ - π π cos θ V j ( θ ) d θ , S j = J ¯ j ∫ - π π sin θ V j ( θ ) d θ . Substituting the expression for V j ( θ ) into the expressions for Cj and Sj then leads to a matrix equation of the form ( 56 ) with λ = 0 . Since I [ 1 ] ≠ 1 , it follows that Cj = 0 so that , up to scalar multiplications , V j ( θ ) = f ′ ( U j ( θ ) ) sin θ , U ( θ ) = A j cos θ . ( 68 ) Now substituting V ( θ ) into Eq ( 63 ) , we have H j ( β ) = Γ j - 1 ∫ - π π V j ( θ ) h j ( θ - β ) d θ = h ¯ j Γ ∫ - π π f ′ ( U j ( θ ) ) sin θ cos ( θ - θ ¯ j - β ) d θ = h ¯ j Γ ∫ - π π f ′ ( U j ( θ ) ) sin θ [ cos θ cos ( β + θ ¯ j ) + sin θ sin ( β + θ ¯ j ) ] d θ = - Λ j sin ( β + θ ¯ j ) , ( 69 ) with Λ j = - h ¯ j Γ j ∫ - π π f ′ ( U j ( θ ) ) sin 2 θ d θ . ( 70 ) We have used the fact that f″ ( Uj ( θ ) ) is an even function of θ , so that the coefficient for cos ( β + θ ¯ j ) is zero . The constant Γj can be calculated from Eq ( 65 ) : Γ j = ∫ - π π V j ( θ ) U j ′ ( θ ) d θ = - A j ∫ - π π f ′ ( U j ( θ ) ) sin 2 θ d θ < 0 . ( 71 ) It follows that Λ j = h ¯ j A j > 0 . ( 72 ) The calculation of K j ( β ) depends on whether we consider model A or model B , see Fig 1 . From Eqs ( 6 ) , ( 60 ) and ( 64 ) , we have for model A K 1 ( β ) = Γ 1 - 1 ∫ - π π V 1 ( θ ) [ ∫ - π π K 1 ( θ - θ ′ ) f ( U 2 ( θ ′ + β ) ) d θ ′ ] d θ , = Γ 1 - 1 ∫ - π π f ′ ( U 1 ( θ ) ) sin θ [ ∫ - π π [ E 1 + K ¯ 1 cos ( θ - θ ′ ) ] f ( U 2 ( θ ′ + β ) ) d θ ′ ] d θ = Γ 1 - 1 [ ∫ - π π f ′ ( U 1 ( θ ) ) sin 2 θ d θ ] [ ∫ - π π K ¯ 1 sin θ ′ f ( U 2 ( θ ′ + β ) ) d θ ′ ] = - 1 A 1 [ ∫ - π π K ¯ 1 sin ( θ ′ - β ) f ( U 2 ( θ ′ ) ) d θ ′ ] = K ¯ 1 A 1 sin β [ ∫ - π π cos θ ′ f ( A 2 cos ( θ ′ ) ) d θ ′ ] ≡ K ¯ 1 A 2 A 1 sin β , ( 73a ) where we have used the stationary condition ( 8 ) , and K 2 ( β ) = K ¯ 2 A 2 sin β [ ∫ - π π cos θ ′ f ( A 1 cos ( θ ′ ) ) d θ ′ ] ≡ K ¯ 2 A 1 A 2 sin β . ( 73b ) Similarly , from Eqs ( 7 ) , ( 60 ) and ( 64 ) , we have for model B K 1 ( β ) = Γ 1 - 1 ∫ - π π f ′ ( U 1 ( θ ) ) sin θ [ ∫ - π π K ¯ 1 δ ( θ - θ ′ ) f ( U 2 ( θ ′ + β ) ) d θ ′ ] d θ = 2 K ¯ 1 Γ 1 ∫ - π π f ′ ( U 1 ( θ ) ) sin θ f ( U 2 ( θ + β ) ) d θ = 2 K ¯ 1 Γ 1 ∫ - π π f ′ ( U 1 ( θ - β ) ) sin ( θ - β ) f ( U 2 ( θ ) ) d θ . ( 74a ) Similarly , K 2 ( β ) = 2 K ¯ 2 Γ 2 ∫ - π π f ′ ( U 2 ( θ - β ) ) sin ( θ - β ) f ( U 1 ( θ ) ) d θ . ( 74b ) Finally , from Eq ( 66 ) , the diffusion coefficients Dj become D j = 1 Γ j 2 ∫ - π π ∫ - π π C j ( θ - θ ′ ) f ′ ( U j ( θ ) ) f ′ ( U j ( θ ′ ) ) sin θ sin θ ′ d θ ′ d θ . ( 75 ) One finds that the diffusivities decreases as the spatial correlation lengths increase . For example , in the case of spatially homogeneous noise ( C j ( θ - θ ′ ) = C ¯ j ) , Dj = 0 since f′ ( Uj ( θ ) ) is even . On the other hand , for spatially uncorrelated noise ( C j ( θ - θ ′ ) = C ¯ j δ ( θ - θ ′ ) ) , we have D j = C ¯ j Γ j 2 ∫ - π π sin 2 θ [ f ′ ( U j ( θ ) ] 2 d θ > 0 . ( 76 ) In Results we take C j ( θ - θ ′ ) = C ¯ j cos ( θ - θ ′ ) so that D j = 1 Γ j 2 ∫ - π π ∫ - π π C ¯ j cos ( θ - θ ′ ) f ′ ( A j cos ( θ ) ) f ′ ( A j cos ( θ ′ ) ) sin θ sin θ ′ d θ ′ d θ = C ¯ j Γ j 2 [ ∫ - π π f ′ ( A j cos ( θ ) ) sin 2 θ d θ ] 2 = C ¯ j 2 A j 2 . ( 77 ) All numerical simulations were performed in Matlab . One dimensional numerical simulations were performed using a forward Euler method scheme in time and a trapezoidal rule for integration in θ . Time steps were taken to be Δt = 0 . 001 , and orientation steps Δθ = 0 . 01π . | A topic of considerable current interest concerns the neural mechanisms underlying the suppression of cortical variability following the onset of a stimulus . Since trial-by-trial variability and noise correlations are known to affect the information capacity of neurons , such suppression could improve the accuracy of population codes . One of the main candidate mechanisms is the suppression of noise-induced transitions between multiple attractors , as exemplified by ring attractor networks . The latter have been used to model experimentally measured stochastic tuning curves of directionally selective middle temporal ( MT ) neurons . In this paper we show how the stimulus-dependent tuning of neural variability in ring attractor networks can be analyzed in terms of the stochastic wandering of spontaneously formed tuning curves or bumps in a continuum neural field model . The advantage of neural fields is that one can derive explicit mathematical expressions for the second-order statistics of neural activity , and explore how this depends on important model parameters , such as the level of noise , the strength of recurrent connections , and the input contrast . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"perturbation",
"theory",
"neural",
"networks",
"neuroscience",
"learning",
"and",
"memory",
"cognitive",
"neuroscience",
"mathematics",
"cognition",
"network",
"analysis",
"algebra",
"memory",
"interneurons",
"quantum",
"mechanics",
"neuronal",
"tuning",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"physics",
"working",
"memory",
"cellular",
"neuroscience",
"cell",
"biology",
"linear",
"algebra",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"cognitive",
"science",
"eigenvalues"
] | 2019 | Stochastic neural field model of stimulus-dependent variability in cortical neurons |
The malaria parasite Plasmodium falciparum exports several hundred proteins into the infected erythrocyte that are involved in cellular remodeling and severe virulence . The export mechanism involves the Plasmodium export element ( PEXEL ) , which is a cleavage site for the parasite protease , Plasmepsin V ( PMV ) . The PMV gene is refractory to deletion , suggesting it is essential , but definitive proof is lacking . Here , we generated a PEXEL-mimetic inhibitor that potently blocks the activity of PMV isolated from P . falciparum and Plasmodium vivax . Assessment of PMV activity in P . falciparum revealed PEXEL cleavage occurs cotranslationaly , similar to signal peptidase . Treatment of P . falciparum–infected erythrocytes with the inhibitor caused dose-dependent inhibition of PEXEL processing as well as protein export , including impaired display of the major virulence adhesin , PfEMP1 , on the erythrocyte surface , and cytoadherence . The inhibitor killed parasites at the trophozoite stage and knockdown of PMV enhanced sensitivity to the inhibitor , while overexpression of PMV increased resistance . This provides the first direct evidence that PMV activity is essential for protein export in Plasmodium spp . and for parasite survival in human erythrocytes and validates PMV as an antimalarial drug target .
Each year malaria parasites cause several hundred million infections and over 650 , 000 deaths [1] . Plasmodium falciparum causes the most lethal malaria and is endemic in Africa [2] . Plasmodium vivax causes most malarial deaths outside Africa and is associated with liver-stage hypnozoites [3] . Although chloroquine and artemisinin have been effective antimalarials , their decreasing efficacy [4] , [5] emphasizes the need for therapies against novel targets shared by both Plasmodium spp . Malaria parasites develop in erythrocytes within a parasitophorous vacuole and export over 450 proteins to the cell ( reviewed in [6] , [7] ) . Export utilizes an N-terminal motif called the Plasmodium export element ( PEXEL; RxLxE/Q/D ) [8] or Vacuolar transport signal ( VTS ) [9] . Exported proteins are cleaved in the PEXEL after Leu ( RxL↓ ) in the endoplasmic reticulum ( ER ) [10] , which requires the conserved Arg and Leu residues [11] . PEXEL cleavage is performed by the aspartyl protease Plasmepsin V ( PMV ) [12] , [13] . PEXEL-containing proteins and PMV are conserved in all Plasmodium spp . [8] , [14]–[16] . Repeated attempts to disrupt the PMV gene have failed , suggesting it is essential [12] , [13] , [16] , but direct and decisive proof is still lacking . A functional survey of P . falciparum exported proteins indicated that 25% or more are essential for parasite survival in human erythrocytes [17] . The current P . falciparum PEXEL exportome is predicted to be 463 proteins [18]; thus , possibly 100 or more exported parasite proteins are required for development in erythrocytes . Some exported proteins lack a PEXEL , for example , skeleton binding protein 1 ( SBP1 ) and the major virulence adhesin family known as P . falciparum erythrocyte membrane protein 1 ( PfEMP1 ) . PfEMP1 is expressed on the erythrocyte surface and mediates cytoadherence to microvascular endothelia , causing severe malaria [19] . PfEMP1 is thought not to be cleaved by PMV [18] , but its transport to , and expression on , the erythrocyte surface requires exported PEXEL and PEXEL-negative proteins ( reviewed in [7] , [20] ) . Aspartyl proteases can be inhibited by transition-state isosteres in which the scissile bond is replaced by a noncleavable moiety . Examples include statine ( Sta ) -containing inhibitors and several are now in clinical use [21] . Here , we developed a transition-state inhibitor that potently blocks PMV from P . falciparum and P . vivax . The inhibitor demonstrates that PMV activity is essential for protein export , PfEMP1 surface display , cytoadherence , and parasite survival in human erythrocytes .
The PMV gene is present in all Plasmodium spp . ; however , only the P . falciparum enzyme ( PfPMV; Pf3D7_1323500 ) has been characterized . A multiple alignment of PfPMV with putative P . vivax PMV ( PvPMV; PVX_116695 ) indicated that they share 82 . 2% similarity , 54 . 7% identity ( Figure S1 ) . Both proteins are predicted to contain a signal peptide , an aspartyl protease domain with DTG and DSG residues defining the catalytic dyad , and a C-terminal transmembrane domain ( Figure 1A ) . Due to four insertions , PfPMVHA is predicted to be approximately 7 . 5 kDa larger than PvPMVHA ( Figure 1A ) ; however , following signal peptide removal , PfPMVHA is predicted to be 8 kDa larger than PvPMVHA . To determine whether PvPMV is an ortholog of PfPMV , we expressed it in P . falciparum fused to 3× hemagglutinin ( HA ) tags ( Figure 1A ) . As a positive control , we expressed PfPMV fused to 3× HA tags ( Figure 1A ) [13] . Expression of PfPMVHA and PvPMVHA was confirmed by immunoblot using anti-HA antibodies ( Figure 1B ) . PvPMVHA was ∼8 kDa smaller than PfPMVHA , as predicted ( Figure 1B ) . PfPMV was previously localized to the ER using a mouse anti-PfPMV antibody that colocalizes with BiP [16] and ERC [13] . To further study PMV , we developed a rabbit antibody that was specific for PfPMV ( Figure 1C , compare lanes 1 and 2 ) and colocalizes with the ER signal from the mouse PfPMV antibody ( Figure 1D , top ) but does not cross-react with PvPMVHA ( Figure S2A ) . Using anti-HA antibodies , a strong perinuclear signal was observed in parasites expressing PvPMVHA or PfPMVHA ( Figure 1D , middle panels , red ) . Both proteins colocalized with rabbit anti-PfPMV antibodies , indicating the location was the ER ( Figure 1D ) . PvPMVHA also colocalized with ERC ( Figure 1D , bottom ) , as shown previously for PfPMVHA [13] . To investigate PvPMVHA activity , we affinity purified it , as well as the PfPMVHA control , using anti-HA agarose , as previously described [13] , [18] . Immunoblot with anti-HA and anti-PfPMV antibodies showed that the purified proteins were species-specific ( Figure S2A ) . The proteins were incubated with a fluorogenic peptide of nine amino acids that contained the PEXEL sequence from knob-associated histidine-rich protein ( KAHRP ) , and efficient processing was observed by both enzymes ( Figure 1E ) . Cleavage of KAHRP by PfPMVHA was previously shown to occur after Leu ( RTL↓ ) by mass spectrometry [13] , and this position was also confirmed for PvPMVHA ( Figure S2B–E ) . Km values of 9 . 7 ( ±3 . 0 ) and 11 . 7 ( ±1 . 8 ) µM were calculated for PvPMVHA and PfPMVHA , respectively ( Figure S2F ) . In contrast , no processing was observed when the PEXEL Arg and Leu residues were mutated to Ala ( ATAAQ ) , consistent with the substrate specificity of PfPMV ( Figure 1E ) . To verify that processing was due to HA-tagged PMV rather than other co-precipitated proteases , we expressed in P . falciparum a mutant of PvPMVHA or PfPMVHA [13] , where one or both catalytic Asp residues were mutated to Ala ( see Figure 1A , B ) . Following affinity purification , the mutant enzymes were incubated with KAHRP PEXEL peptides , but no processing was observed ( Figure 1E; blue , white ) , confirming that the PEXEL-dependent cleavage activity observed for each protease was attributable only to HA-tagged PMV . The substrate specificity of PfPMV is restricted , such that even minor changes to the PEXEL sequence markedly reduces cleavage efficiency—that is , Arg to Lys ( R>K ) or Leu to Ile ( L>I ) [18] . We assessed whether PvPMVHA shares this feature . Indeed , PvPMVHA poorly cleaved peptides possessing R>K or L>I mutations ( Figure 1E; KTLAQ , RTIAQ ) . Collectively , these data show that PvPMV localizes to the ER and cleaves the PEXEL with the same restricted specificity as PfPMV . While maintaining P . falciparum cultures overexpressing PfPMVmutHA , we noticed a delay in growth , suggesting a possible dominant negative phenotype , which has been reported previously with a different PfPMV catalytic mutant [12] . A flow cytometry-based growth assay revealed that parasites expressing PfPMVmutHA with WR99210 selection grew to a parasitemia 2 . 6-fold less than parasites expressing a similar episomal construct , encoding a mini PfEMP1 reporter fused to 3× HA tags ( miniVarHA ) with WR99210 selection ( p< . 0001; Figure 1F ) . This demonstrated that overexpression of inactive enzyme conveyed a growth disadvantage , providing evidence that PMV is important for parasite survival . Analysis of PMV protein levels in parasites overexpressing the PMVHA transgenes indicated that PvPMVHA and PvPMVmutHA had no effect on endogenous PfPMV levels; however , overexpression of inactive PfPMVmutHA caused a clear decrease in expression of the endogenous enzyme ( Figure 1C , compare lanes 1 and 4 , and Aldolase loading controls ) , indicating a negative feedback mechanism occurs in these parasites . The conserved P3 Arg and P1 Leu residues in the PEXEL ( see Figure 2A for a description of nomenclature ) are crucial for PMV activity [12] , [13] . We developed a homology model and designed compounds with a transition-state isostere that mimics the natural PEXEL substrate with the aim of inhibiting PMV . One mimetic , WEHI-916 ( Figure 2B ) , consisted of Arg that could bind in the S3 pocket of PfPMV , Val that would position in the S2 pocket , and Leu-Statine ( Leu-Sta ) , to engage the S1 pocket and inhibit both catalytic Asp residues of PMV ( Figure 2C ) . As control compounds , we synthesized analogs similar to 916 but that mimic noncleavable PEXEL mutant substrates , with the aim that they would be poor PMV inhibitors; the first replaced the P3 Arg with Lys ( R>K; WEHI-024 ) and the second replaced the P1 Leu with Ile ( L>I; WEHI-025; Figure 2B ) . These analogs were designed on the basis that mutations of the conserved PEXEL residues R>K or L>I almost completely inhibit cleavage by PMV ( Figure 1E and [18] ) and should therefore have lower affinity for PMV . Each compound was incubated with PfPMVHA in the presence of KAHRP PEXEL peptides . WEHI-916 ( henceforth 916 ) potently inhibited PEXEL cleavage by PfPMVHA with a 50% inhibitory concentration ( IC50 ) of 20 nM ( Figure 2D ) . In contrast , WEHI-024 and WEHI-025 ( henceforth 024 and 025 , respectively ) had weak activity ( IC50>100 µM and 1 . 11 µM , respectively; Figure 2D ) . 916 inhibited PvPMVHA with an IC50 of 24 nM , whereas 024 and 025 again had weak activity ( Figure 2D ) . To investigate potential off-target activity against human aspartyl proteases , the compounds were assessed against beta-secretase ( BACE-1 ) , for which PMV has distant relatedness [12] , Cathepsin D , and two human cell lines; the compounds displayed poor activity ( IC50>100 µM for BACE-1; 25 µM for Cathepsin D ) and had negligible toxicity against human HEpG2 and fibroblast cell lines ( Figure 2D ) . Collectively , this demonstrated that 916 potently inhibited PMV and had low off-target activity against BACE-1 and Cathepsin D , whereas the closely related analogs 024 and 025 were poorly active . To assess whether 916 could inhibit PMV in P . falciparum–infected erythrocytes , parasites expressing the PEXEL protein P . falciparum erythrocyte membrane protein 3 ( PfEMP3 ) fused to green fluorescent protein ( GFP ) [13] were treated with increasing concentrations of inhibitor , and PEXEL processing was evaluated by immunoblot . A dose-dependent increase in unprocessed PfEMP3-GFP was observed ( black arrow , Figure 3A ) , which was the same size as uncleaved PEXEL R>A mutant PfEMP3-GFP ( Figure 3A ) [13] . The level of PEXEL-cleaved protein ( blue arrow , Figure 3A ) did not quantitatively reflect the degree of PMV inhibition , as inhibitor was added well after PEXEL processing and export of PfEMP3-GFP had initiated . A GFP-only band , representing degraded chimera in the food vacuole , was also observed at ∼26 kDa ( Figure 3A ) . Together , this demonstrated that PEXEL processing was impaired by 916 treatment and that engagement of PMV occurred in P . falciparum–infected erythrocytes . To understand the timing required for PMV inhibition in P . falciparum , parasites were treated with 916 for 1–5 h , and cleavage was evaluated by immunoblot . No effect was seen after 1 h; however , uncleaved PfEMP3-GFP increased between 2 and 5 h ( Figure 3B ) , indicating 916 accessed the parasite ER slowly . Inhibition of PfEMP3-GFP cleavage by 916 was rescued following culture in inhibitor-free medium , to approximately 50% after 2 h ( Figure S3A ) , indicating that cleavage inhibition was reversible or that additional active PMV was synthesized during the experiment . We next assessed whether the control analogs 024 and 025 , which were poor inhibitors of PMV in vitro , had an effect on PEXEL cleavage in parasites . Although a dose-dependent effect was again observed with 916 , analogs 024 and 025 had no effect on PEXEL processing of PfEMP3-GFP or KAHRP-GFP , even at 50 µM ( Figure 3C ) . In the case of KAHRP-GFP , 916 treatment caused accumulation of both uncleaved ( black arrow ) and signal peptide-cleaved ( red arrow ) protein , which were the same size as bands observed for PEXEL R>A mutant KAHRP-GFP ( Figure 3C , right ) . These bands were shown previously to be uncleaved and signal peptide-cleaved KAHRP-GFP , respectively , by mass spectrometry [11] . As 916 treatment caused accumulation of both uncleaved and signal peptide-cleaved species of PEXEL proteins , the potential for off-target effects against signal peptidase was investigated using parasites expressing SERA5s-GFP . This protein contains a signal peptide but lacks a PEXEL and is efficiently secreted to the parasitophorous vacuole ( Figure S3B ) . 916 treatment did not impair processing of the signal peptide from SERA5s-GFP ( Figure 3D , see position of black arrow ) , indicating it cannot inhibit signal peptidase . Taken together , this shows that 916 can effectively inhibit PMV , but not signal peptidase , in P . falciparum–infected erythrocytes and that 024 and 025 have no affect on PMV or signal peptidase activity at concentrations up to 50 µM . The rate of PEXEL protein synthesis , ER import , and processing by PMV in P . falciparum is unknown . We evaluated these processes by radiolabeling parasite proteins in culture for 0 . 5–15 min before immunoprecipitating PfEMP3-GFP with anti-GFP agarose , visualizing bands by autoradiography and quantifying them by densitometry . Labeled PfEMP3-GFP became evident 1 min after addition of label to the culture medium and increased exponentially throughout the experiment ( Figures 3E and S3D ) . Uncleaved PfEMP3-GFP ( black arrow ) was faint and the major species was a doublet of approximately 33 kDa ( red arrow; signal peptide-cleaved ) and 29 kDa ( blue arrow; PEXEL-cleaved; Figure 3E ) . This showed that signal peptidase cleaves PfEMP3 within seconds ( <1 min ) of protein synthesis ( i . e . , cotranslationaly ) , but this molecular species may be transient , as it was not detected by immunoblot of 916-treated parasites , or in PEXEL R>A mutant protein ( Figure 3A–C ) . The radiolabeled bands on the 35S-membrane were confirmed to be GFP-specific by immunoblot ( Figure S3C ) . PEXEL-cleaved PfEMP3-GFP ( blue arrow ) was also evident 1 min after addition of label to the culture medium and increased exponentially , indicating PMV cleavage was also rapid and likely cotranslational ( Figures 3E and S3D ) . The proportion of PEXEL-cleaved protein increased slightly as signal peptide-cleaved protein decreased ( Figure 3E ) , suggesting signal peptidase cleaves before PMV and that PMV may cleave after signal peptidase . Addition of 916 to parasites for 5 h prior to radiolabeling caused accumulation of uncleaved PfEMP3-GFP in parasites ( black arrow ) , which was evident 1–2 min postlabeling ( Figure 3F ) . After a total of 15 min , the degree of PMV inhibition in parasites was quantified by densitometry , and a 13-fold decrease in PEXEL cleavage was observed compared to labeling without 916 ( Figure 3G ) , indicating PMV was inhibited . The signal peptide-cleaved and PEXEL-cleaved species were only weakly detectable throughout the experiment and were not visible until 3–5 min postlabeling ( Figure 3F ) , compared to a stronger signal within 1–2 min of labeling in the absence of inhibitor when the same quantity of parasites was used ( Figure 3E ) . This indicated that 916 significantly blocked PMV in P . falciparum and caused a delay in protein synthesis , ER import , or N-terminal processing of PfEMP3-GFP . Processing of uncleaved PfEMP3-GFP within the PEXEL was rescued after 15 min of culture in inhibitor-free medium ( Figure S3E , F ) , indicating that PMV is able to process full-length PfEMP3-GFP . Having demonstrated that 916 can directly engage PMV in P . falciparum–infected erythrocytes , the effect of the inhibitor on parasite viability was examined by treating early ring parasites for 72 h and assessing parasitemia by flow cytometry . 916 killed parasites with half maximal effective concentration ( EC50 ) of 2 . 5–5 µM ( Figure 4A ) . Analogs 024 and 025 had negligible effect on parasite viability at concentrations up to 20 µM , where 916 completely killed parasites; however , they had EC50 values of 66 µM and 30 µM , respectively , indicating they adversely affected parasite growth at high concentrations . Because PEXEL processing in parasites was unaffected by 024 and 025 treatment , even at 50 µM ( Figure 3C ) , we conclude that the analogs impart toxicity at high concentrations independent of PMV . To determine the stage of the parasite lifecycle that 916 exerted its toxic effects , ring parasites were treated with 15 µM 916 for increasing times through the 48 h cycle and then cultured in inhibitor-free medium to a total of 72 h to see if parasites could recover . Parasites grown in 916 for 1–20 h completely recovered and grew like DMSO-treated controls; however , treatment for >23 h adversely affected growth ( Figure 4B ) , indicating the timing of killing began after 20 h of age , at the ring-trophozoite transition . Analogs 024 and 025 did not affect growth at any parasite stage at the concentration used ( 15 µM ) , whereas chloroquine and artemisinin killed parasites when added to rings for only 30 min before addition of inhibitor-free medium ( Figure 4B , see 0 h ) . This indicated the controls either killed rapidly ( artemisinin ) or were retained inside parasites and killed later , as chloroquine is reported to kill trophozoites [22] , [23] . Both profiles were clearly different than that observed for 916 . Toxicity by 916 after the ring-trophozoite transition was then investigated by adding compound to parasites at different time points through the 48 h cycle . At 48 h , inhibitor-free medium was added and parasitemia for all conditions was determined at 72 h . Toxicity decreased when 916 was added to parasites aged beyond 24 h and schizonts were resistant , indicating 916 did not affect merozoite egress or reinvasion ( Figure 4C ) . As expected , the addition of 916 to rings or early trophozoites was lethal ( Figure 4C ) . 024 and 025 did not have any effect on parasite growth at the concentration used , whereas all parasite stages were sensitive to chloroquine and artemisinin ( Figure 4C ) . Light microscopy of parasites following treatment of early ring stages with 916 revealed a normal ring-stage morphology after 16 h; however , treatment for 32 h revealed a blockage in the ring-trophozoite transition and the majority of parasites appeared pyknotic and abnormal ( Figure 4D ) . As greater than 50% of parasites could not recover from this treatment condition ( refer to Figure 4B ) , the majority of parasites with this appearance were dying or dead . Treatment with DMSO , 024 , or 025 had no effect on development by 32 h at 15 µM ( Figure 4D ) . The morphology of parasites treated with 916 , 024 , 025 , and DMSO was distinctly different from that observed for E-64–treated parasites , which contained swollen food vacuoles from inhibition of haemoglobin breakdown [24] ( Figure 4D , arrow ) . This indicated that parasite toxicity to 916 was unlikely due to off-target inhibition of those food vacuole proteases . Collectively , the toxicity profile seen in the above experiments defines the window of parasite death as between 20 and 30 h , consistent with perturbed protein export and erythrocyte remodeling [7] . To gain further insight into the effects of 916 , 024 , and 025 on parasites , we assessed global protein synthesis following drug treatment by radiolabeling parasite proteins . Treatment of trophozoites with inhibitor for 5 h prior to radiolabeling had no detectable effect on translation , even at 50 µM concentrations ( Figure S4A ) , indicating the compounds are not direct inhibitors of the translation machinery . We next assessed protein synthesis following 23 h of drug treatment of ring stage parasites ( aged 1–3 h old at the initiation of treatment ) . A minor reduction in translation was observed following 916 treatment , but not 024 or 025 treatment , even at 50 µM ( Figure S4B ) . A small but concomitant decrease in the cytosolic protein , Aldolase , was also evident by immunoblot following 916 treatment but not 024 or 025 treatment ( Figure S4B ) , suggesting that parasites were beginning to die from the 916 treatment ( parasites ranged from 24–26 h old at this time point ) . Indeed , Giemsa smears following treatment identified a small proportion of pyknotic parasites in the population ( results not shown ) . Treatment with Brefeldin A , which prevents retrograde trafficking and ER exit , for 23 h severely impaired translation and parasites appeared as dying rings ( i . e . , had not progressed to trophozoites ) . It is possible that PMV inhibition by 916 treatment has a similar but weaker effect to BFA , in that it causes the accumulation of uncleaved PEXEL precursors in the ER , perturbing ER transport , and that this negatively affects translation by an ER stress response [25] . An alternative possibility is that translation was decreased slightly as a result of parasites dying from 916-mediated impairment of erythrocyte remodeling . Either way , the profile for 916 was different to that seen for 024 and 025 , even at 50 µM , indicating the latter analogs likely kill parasites via a different mechanism to 916 . Treatment of P . falciparum with 916 impaired PEXEL cleavage and killed parasites , strongly suggesting that PMV is essential . To investigate this phenotype further , conditional protein knockdown was attempted in P . falciparum using the RNA-degrading glmS ribozyme , which utilizes glucosamine ( GlcN ) as a cofactor [26] . DNA encoding 3× HA epitopes , a stop codon , and glmS was incorporated in frame at the 3′ of the PMV locus by homologous recombination ( Figure S5A ) . Correct genomic integration was confirmed by immunoblot with anti-HA and anti-PfPMV antibodies ( Figure 5A ) . To activate glmS , GlcN was titrated into the culture medium of trophozoites . From 75% to 90% PMV knockdown was achieved after 48 h using 4–6 mM GlcN , but higher concentrations adversely affected parasite HSP70 levels and were subsequently avoided ( Figure S5B ) . Addition of 5 mM GlcN to trophozoites for 24 h reduced PMV levels in subsequent rings by approximately 80% and trophozoites by ∼90% but caused little knockdown in parasites expressing inactive glmS ( M9 ) ( Figure 5B ) [26] . Protein export predominates in rings , when knockdown reached ∼80%; surprisingly , this substantial degree of knockdown did not significantly affect PEXEL processing or parasite growth rate ( p = . 6250; Figure 5C ) , indicating that the remaining PMV levels were sufficient to enable export and sustain parasite development . This demonstrates that PMV activity is potent in P . falciparum and that knockdown to approximately 20% of wild-type levels could not facilitate the characterization of PMV essentiality . As 916 inhibited PMV in parasites ( for example , by 13-fold , Figure 3G ) , the additive effect of PMV knockdown plus 916 treatment was investigated . Parasites expressing PMVHA-glmS were transfected with a construct encoding PfEMP3-GFP , and PEXEL processing was assessed by immunoblot . PEXEL processing of PfEMP3-GFP was barely affected by 48 h of PMV knockdown alone ( Figure 5D , see “+GlcN , ” 0 µM 916 ) ; however , addition of 916 to parasites for 5 h impaired PEXEL cleavage and this was significantly enhanced following knockdown of PMV ( e . g . , by 50% at 20 µM; Figure 5D ) . The quantity of PfEMP3-GFP expressed in the PMV knockdown ( +GlcN ) appeared slightly less than in parasites without knockdown ( −GlcN ) , whereas the loading control HSP70 did not vary appreciably between conditions ( Figure 5D ) . Parasites expressing PMVHA-glmS were next assessed for toxicity to 916 . The EC50 of 916 was reduced by 3 . 3-fold following PMV knockdown compared to no knockdown ( Figure 5E ) . As a control , parasites expressing PMVHA-M9 were treated with 916 in the presence or absence of GlcN; the EC50 reduced by 1 . 4-fold in the presence of GlcN , indicating it had a minor effect . However , the enhancement of PEXEL cleavage inhibition and 3 . 3-fold sensitization of parasites to inhibitor following knockdown indicated that PMV is a direct target of 916 and that PMV inhibition is lethal to parasites . We next investigated the possible effects of PMV overexpression on parasite sensitivity to 916 . Although parasites expressing PfPMVHA do not overexpress enzyme , due to integration of the construct at the endogenous PMV locus [13] , parasites expressing PvPMVHA from episomes also express wild-type levels of endogenous enzyme ( Figure 1B , C ) and therefore contain additional , active PMV in the ER . To control for the carriage of episomes and selection on WR99210 , sensitivity to 916 was compared between parasites overexpressing a similar construct on episomes on WR99210 selection ( encoding miniVarHA; see Figure 1F ) . The EC50 of 916 was 1 . 9-fold greater for parasites overexpressing PvPMVHA compared to parasites overexpressing the control construct , and 1 . 4-fold greater than wild-type 3D7 parasites without WR99210 selection , indicating that PMV overexpression had increased parasite resistance to 916 ( Figure 5F ) . Localization of parasite proteins in Maurer's clefts ( MCs ) , which are parasite-induced membranous structures in the erythrocyte that facilitate protein trafficking , enables accurate quantification of export by immunofluorescence microscopy as the signal is concentrated in puncta [27] . To study export in P . falciparum , we investigated a novel PEXEL-containing protein with two transmembrane domains and unknown function , called Hyp8 ( MAL13P1 . 61/PF3D7_1301700 ) [14] , [28] , that we hypothesized may localize to MCs . Transgenic parasites expressing Hyp8-GFP or Hyp8-HA were generated ( Figure S6A ) . Immunoblotting revealed that Hyp8 is expressed in rings ( Figure S6B ) , and immunofluorescence microscopy showed it is exported ( Figure S6C ) and colocalizes with SBP1 in MCs ( Figure 6A ) . Immunoelectron microscopy confirmed that Hyp8 localizes in MCs ( Figure 6A , right ) . Three independent attempts to delete the hyp8 gene were unsuccessful in this study , in addition to earlier reported attempts [17] , suggesting that Hyp8 may be an essential exported protein . The effect of 916 treatment on export in P . falciparum–infected erythrocytes was then examined . Early ring parasites were treated with inhibitor , and the subcellular Hyp8-GFP fluorescence was quantified by immunofluorescence microscopy ( Figure 6B ) . 916 treatment caused a dose-dependent decrease of Hyp8-GFP in MCs ( the GFP signal in puncta outside the EXP2-labelled parasitophorous vacuole membrane ) compared to DMSO and 024 treatment ( p< . 0001; Figure 6C ) . A small but significant increase in nonexported GFP signal ( the signal inside the EXP2 labeling ) was observed as puncta of fluorescence internal to the parasitophorous vacuole membrane following 916 treatment ( p< . 0001; Figure 6D , see also arrows in B ) . We next examined whether 916 treatment affected protein secretion in P . falciparum–infected erythrocytes by measuring the quantity of EXP2 signal at the parasitophorous vacuole membrane following treatment . There was no statistical difference in EXP2 signal between treatments ( p = . 0977; Figure 6E ) , indicating 916 specifically affected export but not secretion , under the conditions used . As Hyp8 localizes to MCs , we quantified the number of GFP-positive MCs in infected cells following the drug treatments . The mean number of clefts was significantly reduced by 916 treatment , by up to 39% , compared to DMSO ( p< . 0001; Figure 6F ) . Collectively , these data demonstrated that 916 dramatically reduced the export of the PEXEL protein , Hyp8 , and that MC development was impaired following treatment . An important function of exported proteins in remodeling and virulence is to assemble the cytoadherence complex at the erythrocyte surface [17] . Because export of Hyp8 and MC formation was decreased following PMV inhibition , we investigated whether trafficking of PfEMP1 was also affected by quantifying its display on the surface of infected erythrocytes . Ring-stage CS2-GFP parasites [29] selected for expression of the PfEMP1 var2csa gene were treated with one of two sublethal doses of 916 ( Figure S7A ) , and surface-expressed PfEMP1 was measured 24 h postinvasion using PAM1 . 4 antibodies [30] that specifically recognize VAR2CSA [31] by flow cytometry . PfEMP1 surface expression decreased in a dose-dependent manner , by up to 55% , following 916 treatment , but addition of DMSO and 025 had no effect ( Figure 6G ) . Parasitemia across all treatment conditions was measured as GFP fluorescence by flow cytometry and was approximately equal at 24 h , confirming parasite viability ( Figure S7B ) . To evaluate whether decreased PfEMP1 surface expression affected cytoadherence of infected erythrocytes , static binding assays with purified CSA were performed [32] , [33] . Adhesion to CSA was reduced by almost 50% following treatment with 916 compared to DMSO ( p< . 0001 ) and 025 had no effect ( Figure 6H ) . Collectively , this experimentally validates PMV activity as essential for export of PEXEL-containing proteins , resulting in correct MC formation and PfEMP1 assembly at the erythrocyte surface and cytoadherence .
Protein export allows malaria parasites to remodel their cellular niche , and the protein machineries involved are obvious targets for the development of inhibitors . PMV acts by cleaving the PEXEL in the parasite ER and represents one such target . We developed a PEXEL-mimetic compound that potently inhibits the activity of PMV and , combined with protein knockdown or overexpression , used it to demonstrate the essentiality of PMV for parasite survival and its function for export . The PMV inhibitor 916 mimics the transition-state of amide bond proteolysis for PEXEL substrates using statine . Our PfPMV structural model in complex with 916 outlines key interactions that are likely necessary for inhibitor binding: the guanidine side chain of Arg forms salt bridges with the acid of Glu179 and 215 , and a π-stacking interaction with Tyr177 in the S3 pocket of PMV . This in part explains the necessity for Arg at P3 for PEXEL processing . The P1 Leu side chain is encased by hydrophobic residues in the S1 pocket formed in part by Ile116 , Tyr177 , and Val227 , explaining the importance of Leu in PMV binding . 916 potently blocked PMV activity in vitro and reduced PEXEL cleavage , by up to 13-fold ( Figure 3G ) , in cultured parasites , demonstrating the inhibitor directly engaged PMV in the ER . However , inhibition was time-dependent and incomplete at even 50 µM , indicating the inhibitor has suboptimal qualities . This may be due to a combination of poor diffusion across membranes , suboptimal final concentration in the ER , and the potent activity of PMV in parasites , revealed in this study by knockdown of PMV protein levels . Although further work is required to develop an inhibitor with enhanced properties , 916 has proven sufficiently active in parasites to examine PMV function and essentiality . Previously , it has been shown that overexpression of a PfPMV D118A mutant produced a dominant-negative effect on parasite growth rate and protein export [12] . When we overexpressed an alternate PfPMV mutant ( D118A , D365A , F370A ) on episomes in P . falciparum , we observed a similar defect in parasite growth rate and subsequent down-regulation of endogenous PfPMV expression levels , suggesting a negative feedback effect . A similar negative feedback effect has been described for Toxoplasma gondii myosin A [34] . Collectively , these PMV dominant-negative mutants provide evidence that the enzyme is important for parasite survival . The effects of 916 were amplified when PMV was knocked down and decreased when PMV was overexpressed , demonstrating that PMV is a target and that its inhibition is toxic to parasites . Addition of 916 to ring stages arrested their transition to trophozoites , between 20 and 30 h postinvasion , and parasites could not recover . Although this phenotype is consistent with death from impaired export and cellular remodeling , it is possible that ER stress due to accumulation of uncleaved PEXEL proteins in the organelle , and a decrease in translation , contributed [25] . The morphology of parasites treated with 916 was distinctly different to E-64–treated parasites , which contained swollen food vacuoles from inhibition of haemoglobin breakdown [24] . This indicates that parasite toxicity to 916 was unlikely due to off-target effects on food vacuole proteases . Although aspartyl protease inhibitors are known to kill P . falciparum , it is not entirely clear which aspartyl proteases ( Plasmepsins ) are essential ( reviewed in [35] ) . In Plasmodium , there are 10 Plasmepsins; I–IV are important enzymes for haemoglobin degradation by P . falciparum , but the genes encoding each enzyme can be deleted from the genome , indicating they are not essential [36] . The death of P . falciparum following 916 treatment is therefore unlikely to be due to inhibition of these Plasmepsins . A survey of P . falciparum transcriptomes [37] suggests that , of the remaining five Plasmepsins ( VI–X ) , only VII , IX , and X are expressed by asexual blood-stage parasites; however , VII and X are expressed at very low levels . Thus , Plasmepsin IX ( PMIX ) is considered the primary possible off-target in our study . However , we have shown previously that HA-tagged PMIX does not cleave the PEXEL motif [13] , and although the enzyme itself possesses a PEXEL motif , its function and essentiality is currently unknown . Analogs of 916 that mimic noncleavable PEXEL mutant sequences ( R>K , L>I ) were ineffective inhibitors of PMV in vitro and had no discernable effect on parasites at concentrations that 916 inhibited PEXEL cleavage in parasites and was lethal ( i . e . , <20 µM ) . However , at concentrations above 20 µM they were toxic to parasites and possessed EC50 values between 6- and 26-fold less potent than 916 . At 50 µM , we saw no evidence of PMV or signal peptidase inhibition , export or secretion defects , or global effects on translation caused by 024 or 025 . This suggests that they hit a target ( s ) distinct from 916 . The rapid rate of protein synthesis , ER import , signal peptide processing , and PEXEL cleavage in P . falciparum was determined for the first time . The rate was within immeasurable seconds after translation , consistent with both signal peptidase and PMV activity occurring cotranslationaly . The full-length PEXEL is thus only present in the proprotein very transiently , making its function in export even more remarkable . This underscores the importance of the remaining PEXEL residues ( xE/Q/D ) in export following processing [11] . The PEXEL has been suggested to function independent of PMV by binding PI3P in the ER , and ER-derived transport vesicles , via the PEXEL Arg [38] . The rate at which PEXEL processing occurred in our experiments is inconsistent with this hypothesis , as the PEXEL Arg is cleaved off during , or soon after , ER entry . It is also challenging to envisage how the PEXEL Arg could dock within the S3 pocket of PMV , where it is required for proteolytic cleavage , if it is bound to the ER membrane via an interaction with PI3P . This work has characterized a novel PEXEL protein , Hyp8 , which is exported in the early ring-stage to MCs . The function of Hyp8 is unknown , but the hyp8 gene was refractory to deletion and may be essential . 916 treatment dramatically impaired Hyp8 export , resulting in some accumulation in the parasite and possibly degradation . It is unknown whether the reduced export of Hyp8 directly contributed to parasite death; however , the phenotype demonstrates the importance of PMV in export of PEXEL-containing cargo . This is further supported by the decrease in MC numbers observed following 916 treatment; MC formation is known to require exported proteins ( reviewed in [20] ) . The secretion of EXP2 to the parasitophorous vacuole membrane was unaltered by 916 treatment , demonstrating that the effects of 916 were specific to export . A clear defect in PfEMP1 surface exposure and cytoadherence was also observed following 916 treatment . PfEMP1 is unlikely to be a PMV substrate [18] , but its trafficking through the erythrocyte and onto the surface requires at least six PEXEL-containing proteins [17]; thus , the demonstration that PMV activity is essential for PfEMP1 surface expression and cytoadherence is consistent with the current literature and validates the specificity of the inhibitor . Further , it directly demonstrates the importance of PMV at the first step in the export pathway for cellular remodeling that leads to virulence . P . vivax is an important global pathogen that cannot be cultured in the laboratory , and novel therapeutic targets for this enigmatic parasite are urgently needed . This work has characterized PMV from P . vivax for the first time . PvPMVHA possesses the trafficking information to localize to the ER and has similar PEXEL cleavage activity and specificity to PfPMV . This indicates that PMV function is to cleave the PEXEL motif of exported proteins across Plasmodium spp . , and future compounds that block PMV are likely to affect multiple Plasmodium spp . Protein export also occurs in gametocytes [28] and liver stages [39] , and 916 may aid the characterization of PMV in these stages . A putative PMV homolog , ASP5 , is present in Toxoplasma and localizes to the Golgi [40] . Recent evidence suggests that some exported T . gondii proteins contain a PEXEL [41] , and some are cleaved in a manner that requires the conserved PEXEL residues [42] . The PEXEL protease may therefore be conserved beyond the Plasmodium genus , and PMV and its homologs may therefore represent multistage , multispecies antiparasitic targets of the future .
P . falciparum 3D7 parasites expressing PfPMVHA , PfPMVmutHA , and PfEMP3-GFP were generated previously [13] , as was KAHRP-GFP [18] and CS2-GFP [29] . DNA encoding PvPMV or PvPMVmut fused to 3× HA tags was synthesized ( Epoch Biosciences ) and cloned into pGlux . 1 [11] with XhoI and PacI , removing GFP . DNA encoding miniVarHA [PfEMP1 NTS ( residues 1–51 ) fused to SVL-TM-ATS ( residues 2640–2734 ) of IT4 VAR2CSA] was synthesized ( Epoch Biosciences ) and cloned into pGlux . 1 with XhoI and PacI . DNA encoding the signal peptide of SERA 5 ( PFB0340c ) ( residues 1–25 ) or the entire hyp8 gene ( MAL13P1 . 61 ) was amplified from P . falciparum gDNA and cloned in frame with GFP in pGlux . 1 using XhoI and XmaI . For HA tagging Hyp8 , the 3′ 800 bp of hyp8 was cloned into p1 . 2-SHA [13] ( also called pHA3; [43] ) using BglII and PstI . For tagging PfPMV with HA-glmS in P . falciparum NF54 , the 3′ 1144 bp of PMV was cloned into pPTEX150-HA-glmS , which consisted of the glmS riboswitch from pGFP_glmS [26] cloned into pHA3 using BglII and PstI to replace the PTEX150 gene with PMV , generating pPMVHA-glmS . For tagging PfPMV with HA-M9 , the M9 insert from pGFP_M9 [26] was cloned into pPMVHA-glmS to generate pPMVHA-M9 . To express PfEMP3-GFP in P . falciparum NF54 harboring PMVHA-glmS , the dihydrofolate reductase selection cassette in pPfEMP3Glux . 1 [13] was replaced with blasticidin deaminase using BamHI and HindIII prior to transfection . P . falciparum transfectants were selected with 5 nM WR99210 ( Jacobus Pharmaceuticals ) and/or 2 µg/ml Blasticidin S ( Calbiochem ) and grown in O+ human erythrocytes as described [18] . CS2-GFP parasites preferentially expressing the var2csa gene ( PFL0030c/PF3D7_1200600 ) were selected every 2 wk by enriching for knob-positivity with gelatin [44] and panning for CSA-binding [45] . Rα-PfMV antibodies were generated by immunization of rabbits with recombinant PfPMV generated previously [13] and collecting serum during four boost immunizations . Affinity-purified polyclonal rabbit α-Hyp8 antibodies were generated by Genscript using the peptide N-55ETEQSTPAKPEPTE68-C . PMV-agarose was prepared by adding α-HA-agarose ( Sapphire Bioscience ) to parasite lysates , prepared by sonication in 1% Triton X-100/PBS , for 1 h before extensive washing in same [12] , [13] . PEXEL cleavage assays ( 20 µl total volume ) consisted of 0 . 2 µl PMV-agarose in digest buffer ( 25 mM Tris , 25 mM MES , pH 6 . 4 ) with 1 . 5 µM FRET peptide substrate ( DABCYL-RNKRTLAQKQ-E-EDANS , DABCYL-RNKATAAQKQ-E-EDANS , LifeTein; DABCYL-RNKKTLAQKQ-E-EDANS , DABCYL-RNKRTIAQKQ-E-EDANS; Mimotopes ) ± inhibitor . Samples were excited at 340 nm and fluorescence emission measured at 492 nm using an Envision fluorescence plate reader ( Perkin-Elmer ) heated to 37°C for 150 min . Samples were shaken between measurements . For determination of the peptide cleavage position by PvPMVHA , the fluorogenic peptide ( DABCYL-RNKRTLAQKQ-E-EDANS ) , representing the wild-type KAHRP PEXEL sequence , was incubated with and without PvPMVHA at 37°C for 48 h . Products of the incubation were detected by a molecular formula algorithm using an Agilent 6200 TOF/6500 series mass spectrometer . Parasite growth assays were performed in 96-well plates by incubating highly synchronous ring-stage P . falciparum 3D7 or NF54 parasites with compounds solubilized in DMSO at the indicated concentrations for the indicated times . In the case of dose-response curves , medium was kept for the entire experiment; in the case of curves in Figure 4B , C , medium was replaced with inhibitor-free medium at 48 h postinfection . Parasitaemia was always determined at 72 h by flow cytometry . To knock down PMV , GlcN ( Sigma ) was added to trophozoites and drug curves initiated by adding compound at subsequent rings ( 24 h ) for 24–72 h . Trophozoites ( 30–34 h ) expressing PfEMP3-GFP , KAHRP-GFP , or SERA5s-GFP were magnet-purified ( Miltenyi Biotech ) , incubated with inhibitors in 400 µl total volume at 37°C for 1–5 h , treated with 0 . 09% saponin containing inhibitor , and washed pellets were solubilized in Laemmli's buffer , boiled for 3 min , and frozen at −20°C . Proteins were separated by SDS-PAGE , transferred to nitrocellulose and blocked in 10% skim milk/PBS-T , and probed with rat α-HA ( Roche 3F10; 1∶1 , 000 ) , mouse α-GFP ( Roche; 1∶1 , 000 ) , rabbit α-Aldolase ( 1∶1 , 000 ) , rabbit α-HSP70 ( 1∶4 , 000 ) , rabbit α-PfPMV ( 1∶1 , 000 ) , or rabbit α-Hyp8 ( 1∶500 ) primary antibodies followed by horseradish peroxidase-conjugated secondary antibodies ( Silenius ) and detected by enhanced chemiluminescence ( Amersham ) . Whole parasite proteins were radiolabeled by culturing magnet-purified trophozoites ( wild-type 3D7 or expressing PfEMP3-GFP ) in Met/Cys-free medium for 30 min at 37°C before addition of 800 µCi/ml 35S-Met/Cys ( Perkin/Elmer ) to the medium for the indicated times . Pellets were snap frozen in ethanol/dry ice bath and stored at −80°C . For radiolabeling in the presence of PMV inhibitor , parasites were treated with 20 µM WEHI-916 for 5 h before labeling commenced . For pulse-chases , proteins were radiolabeled in the presence or absence of inhibitor , as above , before further culture in radiolabel-free , inhibitor-free complete medium for the indicated times at 37°C before snap freezing . Frozen samples were either solubilized in Laemmli's buffer ( Figure S4 ) or solubilized in 1% Triton X-100/PBS with protease inhibitor cocktail ( Roche ) and PfEMP3-GFP species immunopurified with α-GFP agarose ( MBL ) at 4°C for 2 h ( Figures 3 and S3 ) , and proteins were resolved by SDS-PAGE , visualized by autoradiography ( 7-d exposures ) , and quantified using a GS-800 Calibrated Densitometer ( Bio-Rad ) . For immunofluorescence microscopy , smears were fixed in cold acetone∶methanol ( 90∶10 ) and probed with rabbit Rα-PfPMV ( 1∶750 ) , mouse Mα-PfPMV ( 1∶25 ) , rabbit α-EXP2 ( 1∶200 ) , rat α-HA ( Roche 3F10; 1∶50 ) , mouse α-GFP ( Roche; 1∶500 ) , or rabbit α-Hyp8 ( 1∶200 ) antibodies followed by Alexa Fluor 488- or 594-conjugated secondary antibodies ( Molecular Probes; 1∶1 , 000 ) . DNA was stained with 4′-6-Diamidino-2-phenylindole ( DAPI ) at 0 . 2 µg/ml . Samples were viewed on a Deltavision Elite microscope and images collected with a Coolsnap HQ2 CCD camera through an Olympus 100× UPlanSApo NA1 . 4 objective with SoftWorx software . Images were assembled with ImageJ Fiji 1 . 47d and Adobe Photoshop CS6 v13 . 0 x64 . Light and immunoelectron microscopy was performed as described in [46] . For quantification of events in cells infected with parasites expressing Hyp8-GFP , highly synchronous ring-stage parasites engineered to express Hyp8-GFP from the CRT promoter were obtained by incubation of erythrocytes with viable merozoites for 15 min [47] and treated with 20 or 50 µM 916 30 min postinvasion for 13 h ( until Hyp8-GFP expression from the CRT promoter had occurred for 1 h ) . Smears were fixed in 90∶10 acetone∶methanol , labeled with anti-GFP and anti-EXP2 antibodies , and Z-stacks captured on a Deltavision Elite microscope using a 100× objective . Over 40 Z-stacks per condition were imaged using the same exposure settings to allow quantitative analysis between groups . As 916 treatment has adverse affects on parasites after 24 h , a sublethal dosing regime was developed ( see Figure S7 ) to maximize the inhibitor effect while ensuring parasites remained viable 24 h postinvasion when surface-exposed PfEMP1 was measured . Treatment with >15 µM for 23 h or with 50 µM for >12 h postinvasion prior to decreasing to 15 µM adversely affected parasite growth and was avoided . To measure PfEMP1 display , highly synchronous ring-stage CS2-GFP parasites preferentially expressing VAR2CSA ( see Plasmids , Parasites , and Antibody Production ) were obtained by incubation of erythrocytes with viable merozoites for 15 min [47] , and parasites were treated with 916 , 025 , or DMSO using the dosage regime above . At 24 h postinvasion ( presence of inhibitor for no more than 23 h ) , erythrocytes were incubated with human monoclonal PAM1 . 4 serum [30] ( 1∶200 ) to label VAR2CSA followed by goat anti-human IgG Biotin-conjugated secondary antibodies ( Invitrogen ) ( 1∶200 ) and Alexa Fluor-633 Streptavidin-conjugated tertiary antibodies ( Invitrogen ) ( 1∶500 ) for 30 min each . Labeled cells were washed with 0 . 1% casein/PBS and analyzed on a FACSCalibur cytometer ( Becton-Dickinson , USA ) . Fluorescence in channel FL1 was used to measure parasite-infected erythrocytes ( GFP ) , and fluorescence in channel FL4 was used to measure bound PAM1 . 4 IgG antibodies ( Alexa 633 ) for each sample . The geometric mean fluorescence of uninfected erythrocytes ( treated with secondary and tertiary but not primary antibodies ) was deducted from the geometric mean fluorescence of infected erythrocytes using >100 , 000 cells per condition . Experiments were conducted in duplicate . Analyses were performed using FlowJo 8 . 8 . 7 ( Tree Star , USA ) . Adhesion assays were performed as described previously [32] , [33] . Briefly , CSA ( Sigma ) was spotted at 50 µg/ml in triplicate into petri dishes , incubated overnight at 4°C , and blocked in 1% casein/PBS for 2 h . Inhibitor-treated erythrocytes were added to CSA-coated dishes and incubated for 45 min at 37°C . Dishes were washed four times with 5 ml warm RPMI-HEPES , fixed in 2% paraformaldehyde for 2 h , stained with 10% Giemsa for 15 min , and the number of adherent erythrocytes per mm2 quantified by light microscopy counts . Assays were performed in triplicate . This information is presented in Materials and Methods S1 . | To survive within human red blood cells , malaria parasites must export a catalog of proteins that remodel the host cell and its surface . This enables parasites to acquire nutrients from outside the cell and to modify the cell surface in order to evade host defenses . Protein export involves proteolytic cleavage of the Plasmodium Export Element ( PEXEL ) by the aspartyl protease Plasmepsin V . We report here the development of a small molecule inhibitor that closely mimics the natural PEXEL substrate and blocks the activity of Plasmepsin V from the malarial parasites Plasmodium falciparum and Plasmodium vivax . The inhibitor impairs export and cellular remodeling and kills P . falciparum at the ring-trophozoite transition , providing direct evidence that Plasmepsin V activity is essential for export of PEXEL proteins and parasite survival within the host . These findings validate Plasmepsin V as a highly conserved antimalarial drug target . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"enzymology",
"microbiology",
"plasmodium",
"falciparum",
"chemical",
"biology",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"parasitology",
"gene",
"function",
"mutation",
"mathematical",
"computing",
"plasmodium",
"vivax",
"protozoans",
"mathematics",
"molecular",
"genetics",
"pharmacology",
"infectious",
"disease",
"control",
"enzyme",
"kinetics",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"malarial",
"parasites",
"medical",
"microbiology",
"chemistry",
"computing",
"methods",
"pathogenesis",
"membrane",
"trafficking",
"drug",
"discovery",
"mutagenesis",
"biochemistry",
"cell",
"biology",
"drug",
"research",
"and",
"development",
"host-pathogen",
"interactions",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"malaria",
"physical",
"sciences",
"molecular",
"cell",
"biology",
"organisms"
] | 2014 | Inhibition of Plasmepsin V Activity Demonstrates Its Essential Role in Protein Export, PfEMP1 Display, and Survival of Malaria Parasites |
Up to 1 . 45 billion people currently suffer from soil transmitted helminth infection , with the largest burden occurring in Africa and Asia . Safe and cost effective deworming treatment exists , but there is a debate about mass distribution of this treatment in high prevalence settings . While the World Health Organization recommends mass administration of anthelmintic drugs for preschool and school-aged children in high ( >20% ) prevalence settings , and several long run follow up studies of an influential trial have suggested large benefits that persist over time , recent systematic reviews have called this recommendation into question . This paper analyzes the long-term impact of a cluster-randomized trial in eastern Uganda that provided mass deworming treatment to preschool aged children from 2000 to 2003 on the numeracy and literacy skills of children and young adults living in those villages in 2010-2015 . This study uses numeracy and literacy data collected seven to twelve years after the end of the deworming trial in a randomly selected subset of communities from the original trial , by an education-focused survey that had no relationship to the deworming study . Building on an earlier working paper which used data from 2010 and 2011 survey rounds , this paper uses an additional four years of numeracy and literacy data ( 2012 , 2013 , 2014 , and 2015 ) . Aggregating data from all survey rounds , the difference between numeracy scores in treatment versus control communities is 0 . 07 standard deviations ( SD ) ( 95% CI -0 . 10 , 0 . 24 , p = 0 . 40 ) , the difference in literacy scores is 0 . 05 SD ( 95% CI -0 . 16 , 0 . 27 , p = 0 . 62 ) , and the difference in total scores is 0 . 07 SD ( 95% CI -0 . 11 , 0 . 25 , p = 0 . 44 ) . There are significant differences in program impact by gender , with numeracy and literacy differentially positively affected for girls , and by age , with treatment effects larger for the primary school aged subsample . There are also significant treatment interactions for those living in households with more treatment-eligible children . There is no evidence of differential treatment effects on age at program eligibility or number of years of program eligibility . Mass deworming of preschool aged children in high prevalence communities in Uganda resulted in no statistically significant gains in numeracy or literacy 7-12 years after program completion . Point estimates were positive but imprecise; the study lacked sufficient power to rule out substantial positive effects or more modest negative effects . However , there is suggestive evidence that deworming was relatively more beneficial for girls , primary school aged children , and children living in households with other treated children . As this analysis was conducted on secondary data which is publicly available , no research approval was sought or received . All individual records were anonymized by the data provider prior to public release .
The theory that deworming could improve educational outcomes goes back at least as far as the 1930s [18] , but there was little rigorous research on the topic until a series of trials in the 1990s testing the impact of deworming on cognitive outcomes over short periods ( 1-7 months ) [19 , 20 , 21] , followed by a series of mass treatment trials which recorded educational or cognitive outcomes . The Cochrane and Campbell reviews [11 , 12] identify a relatively modest number of multiple dose mass deworming trials with cognition outcomes . Both reviews note their limited ability to aggregate these trials due to incomparable outcome measures , but find that there is no evidence of positive effect from deworming on these outcomes . However , these trials largely studied school age children , while the deworming program examined in this paper targeted pre-school aged children . There is a priori reason to believe that improved early childhood nutrition can have larger effects on a range of outcomes , including cognition , than improved nutrition for school aged children [22 , 23 , 24] . Going back to the Barker hypothesis [25] , evidence has accumulated that early child nutritional deprivation , such as from famine , has long run consequences for long run health , well-being , and cognition [26 , 27 , 28 , 29] . This implies that interventions that mitigate early childhood nutritional deficits could have long run benefits [24 , 30 , 31] . Another theme is the broad range of shocks that appear to have important effects , including some seemingly mild shocks [32] . For this age group , the findings from the deworming meta-analysis literature are sparser . A smaller group of studies in the Campbell review examine cognitive outcomes for preschool aged children [33 , 34] . Neither finds a significant benefit , although both papers note that the studies in question were not powered for cognitive outcomes . Evidence of short run benefit on education-related outcomes from deworming of school-aged children comes from Miguel and Kremer [5] , who use a cluster randomized design to show increases in school attendance resulting from deworming . In addition , they identify spillovers both within treatment schools , and at schools up to 3 km away from treatment schools . Replication of this paper and subsequent exchanges with Aiken et al . [35] confirmed their finding of within-school spillovers and geographic spillovers up to 3 km , but showed that spillover benefits did not extend beyond 3 km range as the original paper argued . High levels of STH prevalence in this setting could explain the benefits; by contrast a recent trial in a low prevalence setting did not find comparable educational gains [36] . Evidence for longer run benefits of deworming comes from follow up data collection on the cohorts from the Miguel and Kremer study . Baird et al . [7] find marked gender-related patterns to gains in their long run follow up of the original trial: while men see employment related gains , they do not gain more education . By contrast , women are one third more likely to attend secondary school , and have improved self-reported health and increased body mass index . Ozier [6] examines whether younger siblings of children treated in the original Miguel and Kremer trial show cognitive benefit , conditional on the age at which they were exposed to the spillover . Children whose communities were treated before they were 1 year old show large benefit ( via spillovers from sibling treatment ) , while children treated after that age show no evidence of benefit . Furthermore , the effects were twice as large for those likely to have had treated older siblings in their household , suggesting that children may benefit more from deworming when their siblings are also dewormed .
With 36 clusters and an intra-cluster correlation ( ICC ) of 0 . 02 for numeracy scores , we calculate an intention-to-treat minimum detectable effect of 0 . 18 standard deviations ( for the numeracy outcome ) if all children were treated in the treatment group and none were treated in the control group ( the ICC is calculated from Uwezo data in study clusters , and the power calculation is implemented using Stata’s “clustersampsi” command ) . Inflating this to account for incomplete take up in the treatment group ( 74% of the treatment group reported attending Child Heath Days ) implies a minimum detectable effect ( MDE ) for numeracy of 0 . 33 standard deviations ( SD ) . ( Some control group deworming was reported from private drug shops . However without information about the frequency or intensity of this deworming we cannot precisely adjust power calculations to reflect it ) . As this analysis was conducted on secondary data which is publicly available , no research approval was sought or received . All individual records were anonymized by the data provider prior to public release . Outcome data in this study comes from the Uwezo initiative , which conducts large-scale annual surveys to test basic literacy and numeracy in Kenya , Tanzania , and Uganda . Every year , Uwezo randomly samples 30 villages in every district in all three countries , and all children between ages 6 and 16 in 20 households per village are tested on basic literacy and numeracy [38] . Further details about Uwezo’s sampling procedures are available in Supporting Information Section 1 . In the Uwezo survey , children are given a numeracy test with seven skill areas of ascending difficulty , and a literacy test with six skill areas , for a maximum of 13 points . Following standard practice in the education literature , standardized math , literacy , and total ( math plus literacy ) scores are created , with mean zero and standard deviation of 1 , as outcome variables . In all rounds , the basic math test consists of seven categories , representing counting objects up to 10 , number recognition of numbers from 10-999 , addition , subtraction , multiplication , and division . There are bonus questions across all years but since these vary in content substantially ( from practical questions about buying items in market in some years , to place value questions in other years ) , we omit them . The literacy test includes five categories , from letter recognition , to the ability to read words , sentences , and paragraphs . There are two comprehension questions , for which one point is awarded , in early rounds , while in later survey rounds two points are awarded for the two questions . Since the meaning of a comprehension bonus point appears to change over time ( and is missing for most respondents in 2010 ) , we omit the bonus comprehension questions for all rounds and simply use the 5 point literacy scale which remains fundamentally unchanged across all six survey rounds . The Uwezo survey captures numeracy and literacy outcomes for a small fraction of the population exposed to the original program ( the Uwezo sample is 2 , 210 while there were 27 , 995 individuals in the original trial ) . However , Uwezo’s sampling method makes it likely that the 2 , 210 individuals surveyed are in effect a random subsample of individuals from the communities in the original sample . This is because according to its procedures , the Uwezo survey sampled a random subset of villages from the original trial districts , and a random group of households from within these villages . Exploiting the random allocation of treatment in the original trial , and the random sampling of a subset of original trial communities by Uwezo , the effect of the deworming program is estimated by regressing test scores on a treatment indicator variable ( unadjusted model ) and on the treatment variable together with dummy variable controls for each age category ( ages 6 through 16 ) , each survey round ( 2010 , 2011 , 2012 , 2013 , 2014 , and 2015 ) , and gender , and all interactions of gender , age , and survey round ( adjusted model ) . The specifications are: y i c = β 0 + β 1 t r e a t c + ε i , c ( 1 ) for unadjusted specifications ( odd numbered columns ) , and for adjusted specifications ( presented in even numbered columns ) , y i c = β 0 + β 1 t r e a t c + χ i c ′ · γ + ε i , c ( 2 ) where i is for example an individual and c is a community ( parish ) . yi , c is the learning outcome of interest , treatc is a dummy for treatment communities from the original trial , and χ i ′ is a vector of indicator variables for each age , gender , and survey round , and all interactions of these age , gender , and survey round indicators . The coefficient of interest is β1 . εi , c is the error term , clustered at the parish level . Based on reported age , a child’s age in the years of 2000 , 2001 , 2002 , and 2003 is calculated . If the child was between age 1 and age 7 during any one of those years , he or she is considered to be in the age group potentially exposed to the program . Motivated by the broader deworming literature , we also examine potential modifiers of effect . We seek to do so in ways which are robust to concerns about multiple hypothesis testing . Therefore we first conduct the main analysis , and the heterogeneity analysis , using the specifications used in the 2014 working paper which was mentioned previously . In effect we treat this 2014 working paper as analogous to a pre-registered analytical plan for the full paper . Then , in more exploratory vein , we examine additional hypotheses which we believe are appropriate given changes in the sample across rounds , as well as hypotheses which have been generated by other recent papers in the literature . We examine the following dimensions of heterogeneity on the treatment effect of deworming: primary school vs . secondary school age , potential exposure to other dewormed children , and age at first eligibility for the program . Given the exploratory nature of these analyses , we calculate q-values for the coefficients of interest , limiting the false discovery rate using the method proposed by Benjamini and Hochberg [39] . ( We also examine these effect modifiers on school enrolment and attendance in Supporting Information S5 Table ) . For hetereogeneity analysis , we interact the treatment variable with variables created to represent poverty , gender , and cumulative years of program eligibility ( for the pre-specified analyses ) , and age at program initiation , primary school age vs . post-primary school age , and number of other deworming-eligible children in the household ( for the exploratory analyses ) . As a proxy for poverty , we calculate the first principal component of a wealth index made up of the household asset variables which are measured consistently across all survey rounds ( access to electricity , and ownership of a television , radio , and phone ) , and create a “low assets” indicator which represents the bottom two wealth quintiles . To represent households with multiple treatment-eligible children we create a variable which represents the number of other treatment-eligible children in the household ( excluding the respondent him or herself ) . The “years of program eligibility” variable is calculated by summing up the number of years that a respondent was of program-eligible age ( age 1 to age 7 ) between 2000-2003 , and creating a binary variable equal to one for respondents who were eligible for 2 , 3 , or 4 years of the program . The primary school age variable is binary , equal to one for respondents aged 7-13 at the time of survey , and equal to 0 for respondents aged 14-16 . The “age at first eligibility” variable is a binary variable equal to 1 for respondents who were age 1 when the program was initiated ( in 2000 ) , and equals 0 for those who were older than age 1 in when the program was first initiated . We also conduct a range of robustness checks for the main results . Since survey non-response could bias results , we re-run the main analysis using predicted scores ( for respondents who did not take the Uwezo test ) generated by the Uwezo team based on observable respondent characteristics [40] , as well as using inverse probability weighting to minimize any bias due to non-response . In addition , we use a different dataset , the Uganda National Panel Survey , to examine whether migration out of the study sample is correlated with being of post-primary school age , and with literacy . We also conduct all additional robustness checks which were implemented in the 2014 working paper . Finally , we also run the main analysis using population weights provided by Uwezo . Since Uwezo conducted stratified sampling ( with equal numbers of communities sampled from both densely and thinly population districts ) , observations can be re-weighted to be more representative of the population [40] . All of these robustness checks are presented in the Supporting Information .
Across the three outcome variables , there is no statistically significant treatment effect of deworming in the full sample . In adjusted specifications , the coefficients for numeracy ( 0 . 07 SD , standard error ( s . e . ) 0 . 08 ) , literacy ( 0 . 05 SD , s . e . 0 . 11 ) and total score ( 0 . 07 SD , s . e . 0 . 09 ) are not significantly different from zero ( Table 2 ) . The other educational outcomes captured in the Uwezo dataset relate to whether the child is currently in school , and whether he or she ever enrolled in school . There is no significant relationship between deworming program exposure and attendance or school enrollment ( Table 3 ) ; the coefficients in adjusted specifications on being in school at the time of the survey or never having enrolled are -0 . 01 ( s . e . 0 . 01 ) and 0 . 01 ( s . e . 0 . 01 ) . In this section we focus on several dimensions of heterogeneity in treatment effect which were examined in the 2014 working paper: gender , poverty ( proxied by asset ownership ) , and number of years eligible for treatment , as well as three additional dimensions of heterogeneity ( presence of other program-eligible children in the household , primary vs . post-primary school age , and age at first program eligibility ) , which are exploratory in nature .
This study addresses the question of whether a mass deworming program in a high prevalence setting had a measurable impact on basic academic skills , such as numeracy and literacy , for children who live in villages which were part of this program during their preschool years . While an earlier working paper using the 2010 and 2011 Uwezo data suggested that deworming can have important long run impact on these skills , analysis incorporating all data from 2010–2015 shows that there is no statistically significant effect of early childhood deworming on numeracy or literacy 7-12 years later . However , there is suggestive evidence of differential impact along several dimensions . We find that the program differentially affected female respondents , and that treatment effects for females alone , although imprecisely estimated and not statistically significant at conventional levels , are substantively large . In exploratory heterogeneity analysis , we note that treatment effects for the primary school age ( 7-13 year old ) respondents are significantly larger than for the post-primary ( 14-16 year old ) cohorts . In addition , motivated by recent findings on spillovers in deworming programs , we find that the treatment effect increases as the number of treatment-eligible children in a household increases . However , we do not find evidence that initiation of program eligibility at age 1 results in larger effects than initiation after age 1 , and we do not find evidence of differential gains for age cohorts who had more years of program eligibility . What factors might explain the difference between the findings of the original working paper and this analysis ? A published critique of the 2014 working paper [41] criticized it for only having information about a subsample of original trial communities , and for potential imbalance . However , as noted by several participants in 2016 International Journal of Epidemiology symposium about this and related papers , the Uwezo sampling procedure leaves no reason to expect imbalance in which of the trial communities were sampled by the Uwezo survey [42 , 43 , 44] , and the full sample demonstrates no significant imbalance between the treatment and control groups across a broad range of covariates . There are no statistically significant differences on any of the variables tested , and an F test of a regression of all the variables in the table below on treatment has a p value of 0 . 85 . A potential explanation is suggested by the heterogeneity analysis , which showed larger effects for primary school aged children . A key difference between this paper and the original working paper is the age profile of the sample . Compared to the sample used in the original working paper , the 2010-2015 sample is now more heavily composed of post-primary school age cohorts . In the first three years of the sample ( 2010-2012 ) , 34% of the sample were aged 14-16 , compared to 54% in the 2013-2015 surveys . If 14 , 15 , and 16 years olds are more likely to migrate away from their home community than primary school aged children , it is more likely that they are interviewed in a community in which they did not live at the time the program was implemented ( 2000-2003 ) . This introduces measurement error into the independent variable ( potential exposure to the deworming program in early childhood ) , and therefore biases any potential treatment effects towards zero [45] . It could also bias effects downward if more literate and numerate individuals have higher propensity to leave their home communities . Differential migration rates by age and skill are impossible to document in the Uwezo sample , since Uwezo household rosters only include children and young adults living at home; those who have migrated are simply not observed in the data . However , using the 2009-2010 Uganda National Panel Survey ( which includes both household members living both at home , as well as those living outside the household for education or work purposes ) , we document differential migration among the aged 14-16 cohorts , and among the more literate . This is consistent which evidence from other developing countries: For example , Young [46] aggregates DHS data from 65 developing countries to shows that education is correlated with rural-urban migration , while Hicks et al . [47] use panel surveys in Kenya and Indonesia to show that cognitive ability is strongly associated with rural-urban migration , even after adjusting for education . In Uganda , Mensah and O’Sullivan [48] show associations between education and migration in the Uganda National Panel Survey . A weakness of this study is that we do not know specifically that the children and young adults surveyed by Uwezo were directly exposed to the deworming program . We do know that over 70% of children in treatment communities attended Child Health Days and were dewormed , but some unknown fraction of the respondents may have been born in different communities than the ones they were surveyed in , or else did live in the communities but never attended the deworming events . Another limitation of the study is that , after the trial from 2000-2003 , the government of Uganda incorporated deworming into Child Health Days , meaning that some percentage of children in the control group received some deworming treatments after the original trial period concluded [49] . This would attenuate any benefits experienced by children in the original treatment group , since both treatment and control communities were potentially exposed to national deworming programs after 2003 . How should this pattern of results be interpreted , in light of the broader long run deworming literature ? In the main treatment effect specifications , much depends on the approach taken to treatment estimates which are positive but do not reach conventional levels of significance . In this case , the power of the study and the 95% confidence intervals of the estimated treatment effects should also be considered . The main regressions for numeracy have 95% confidence intervals of ( -0 . 1 , 0 . 24 ) , meaning that the study lacks sufficient power to rule out large positive effects , as well as more modest negative effects . These wide confidence intervals are a function of the fact that this study took key design parameters ( especially number of clusters ) as fixed by the original trial and the subsequent Uwezo data collection exercise . While p values above or below 0 . 05 are often interpreted in binary terms , this has long been criticized by epidemiologists and statisticians [50 , 51] . In this “vote counting” paradigm , Baird et al . [7] and Ozier [6] count as evidence for the long run benefits of deworming ( p<0 . 05 ) while the results presented here count as evidence against ( p>0 . 05 ) . A better approach would be to formally aggregate these results via meta-analysis . However , whereas short-run deworming trials have typically measured a small set of outcomes , such as height and weight , consistently across trials , making meta-analysis very feasible , comparable meta-analysis is more challenging for long run studies . The treatments differ ( direct deworming versus exposure via spillovers ) , as do the populations of interest ( all exposed cohorts versus children under age 1 , as in [6] ) , and the primary outcomes studied differ ( labor market outcomes [7] versus generalized cognitive ability [6] versus numeracy and literacy ( this paper ) . While the deworming treatments differ and the outcomes measured are not directly comparable , like Baird et al . [7] we find educational gains concentrated among females , and like Ozier [6] and Miguel and Kremer [5] we find suggestive evidence of very local spillovers . The smaller point estimates of treatment effect are also consistent with this literature , given that the worm burden in the Ugandan setting , while high in absolute terms , was notably lower than the burden in western Kenya where the original deworming trial ( on which Ozier [6] and Baird et al . [7] are based ) took place . Relating this back to policy , a frequentist policymaker might dismiss the evidence from this setting as too imprecise to use . A more Bayesian-minded policymaker , with imprecise information about potential benefits , but needing to decide about a deworming program on the basis of this information , might compare the expected benefit of the policy to the expected cost [52] . This policymaker might be primarily concerned about the effect of deworming on the health of the STH-infected , but in a high prevalence setting , it also makes sense to consider potential long run benefits to learning . A Bayesian policymaker would not consider a point estimate that is statistically indistinguishable from zero as effectively a zero ( evidence of no effect ) , but would consider it as the most likely of a distribution of potential effect sizes which also includes zero and negative numbers . Then the decisionmaker would compare the expected costs and benefits of this intervention to other policies to increase learning . To give context for cost-effectiveness , McEwan [53] estimates cost effectiveness ratios for a subset of learning interventions . Most studies require between 10 USD and 100 USD ( PPP adjusted ) to generate 0 . 2 SD of learning gain . A version of Uganda’s Child Health Day program which only included Vitamin A supplementation and deworming cost $0 . 22 per child served , which includes both financial outlays , staff time , and in-kind contributions [54] . These costs as were 73% driven by the deworming component ( $0 . 16 per recipient ) and 27% by Vitamin A supplementation ( $0 . 06 per recipient ) . Since CHDs are conducted two times per year , this implies $0 . 32 per child per year for the deworming component , or $0 . 96 total per child for a three year program like the one studied here . A cost of $0 . 96 for 0 . 07 SD for learning implies an expenditure of $2 . 74 per 0 . 2 SD of learning . Even allowing for imprecision or differential assumptions between the cost estimates from Uganda’s Child Health Day deworming [54] and the various education trials [53] , these cost ratios are favorable to deworming . In conjunction with the policymaker’s prior views on the value of deworming and the local prevalence and intensity of worm infection , a policymaker who significantly discounted the probability that 0 . 07 SD was the true effect on numeracy might still choose to invest in preschool deworming [55] . More generally , how should these finding affect interpretation of the long run impact of deworming ? Clearly , the estimated treatment effects here are smaller than the large effects estimated in the original working paper using 2010 and 2011 data , and are no longer statistically significant . The case for meaningful long run effects is therefore at least somewhat less convincing than it would be if the results of this study , like that of Baird et al . [7] and Ozier [6] , still provided strong evidence of large effects of mass deworming . Yet this study should be interpreted in the context of the serious challenges involved in generating evidence about the long run effects of interventions such as deworming . The ideal research design for long run effects involves well-powered trials followed by detailed longitudinal data collection . This is the model of the original Miguel and Kremer [5] trial and the Baird et al . [7] follow up , and studies stemming from this original experiment have generated all of the experimental evidence on the long run effect of deworming to date . Outside of this setting , and apart from conducting new trials and waiting a decade or more for new long-run results , another approach is to identify experimental interventions , such as the one on which this paper is based , for which existing data can be used to assess long run impact . The benefit of this approach is the limited marginal cost of the exercise , and the ( relative ) speed with which results can be generated . The downside is that in this type of research design , control over study parameters such as measurement , survey design , and statistical power are limited . Results should therefore be interpreted in this light . | Mass deworming is recommended by the World Health Organization for health benefits to communities where soil-transmitted helminth infection is endemic . In addition to health benefits , several recent studies find long run educational or economic benefits for cohorts dewormed as children . In this paper , treatment and control communities which formed part of a cluster randomized deworming trial in eastern Uganda from 2000-2003 were surveyed from 2010-2015 to measure children’s basic numeracy and literacy . We analyze this data to see if there are detectable improvements in basic academic skills among children in the dewormed communities . We find that mass deworming of preschool aged children in high prevalence communities in Uganda resulted in no statistically significant gains in numeracy or literacy 7-12 years after program completion . However , there is suggestive evidence that deworming is relatively more effective for girls , primary school aged children , and children living in households with other treatment-eligible children . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"children",
"medicine",
"and",
"health",
"sciences",
"education",
"sociology",
"geographical",
"locations",
"social",
"sciences",
"numeracy",
"uganda",
"neuroscience",
"parasitic",
"diseases",
"pediatrics",
"age",
"groups",
"academic",
"skills",
"research",
"design",
"cognitive",
"psychology",
"literacy",
"surveys",
"africa",
"families",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"schools",
"child",
"health",
"people",
"and",
"places",
"helminth",
"infections",
"psychology",
"survey",
"research",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"cognitive",
"science"
] | 2019 | The long run impact of early childhood deworming on numeracy and literacy: Evidence from Uganda |
Co-occurrence of malaria and filarial worm parasites has been reported , but little is known about the interaction between filarial worm and malaria parasites with the same Anopheles vector . Herein , we present data evaluating the interaction between Wuchereria bancrofti and Anopheles punctulatus in Papua New Guinea ( PNG ) . Our field studies in PNG demonstrated that An . punctulatus utilizes the melanization immune response as a natural mechanism of filarial worm resistance against invading W . bancrofti microfilariae . We then conducted laboratory studies utilizing the mosquitoes Armigeres subalbatus and Aedes aegypti and the parasites Brugia malayi , Brugia pahangi , Dirofilaria immitis , and Plasmodium gallinaceum to evaluate the hypothesis that immune activation and/or development by filarial worms negatively impact Plasmodium development in co-infected mosquitoes . Ar . subalbatus used in this study are natural vectors of P . gallinaceum and B . pahangi and they are naturally refractory to B . malayi ( melanization-based refractoriness ) . Mosquitoes were dissected and Plasmodium development was analyzed six days after blood feeding on either P . gallinaceum alone or after taking a bloodmeal containing both P . gallinaceum and B . malayi or a bloodmeal containing both P . gallinaceum and B . pahangi . There was a significant reduction in the prevalence and mean intensity of Plasmodium infections in two species of mosquito that had dual infections as compared to those mosquitoes that were infected with Plasmodium alone , and was independent of whether the mosquito had a melanization immune response to the filarial worm or not . However , there was no reduction in Plasmodium development when filarial worms were present in the bloodmeal ( D . immitis ) but midgut penetration was absent , suggesting that factors associated with penetration of the midgut by filarial worms likely are responsible for the observed reduction in malaria parasite infections . These results could have an impact on vector infection and transmission dynamics in areas where Anopheles transmit both parasites , i . e . , the elimination of filarial worms in a co-endemic locale could enhance malaria transmission .
Malaria and lymphatic filariasis ( LF ) are two of the most important mosquito-borne diseases . Currently , there are 2 . 5 billion people at risk of contracting malaria , and on average , there are 300–500 million clinical cases of malaria each year causing between one and three million deaths [1]–[3] . Human LF is caused by several species of mosquito-borne filarial nematodes , including Brugia malayi , Brugia timori , and Wuchereria bancrofti , but W . bancrofti is responsible for 90% of LF infections worldwide . It is estimated that 120 million people in the world have LF , with ∼1 . 1 billion at risk of becoming infected . Although LF is rarely fatal , severe morbidity ( including adverse economic and psychosexual effects ) occurs in 40% of infected individuals and involves disfigurement of the limbs and male genitalia ( elephantiasis and hydrocele , respectively ) [2]–[4] . Both malaria and LF are co-endemic in many areas of the tropics and in certain areas are transmitted by the same Anopheles mosquitoes [5]–[7] . Co-infection of multiple species of malaria parasites or a combination of malaria and filarial worm parasites in humans have been reported [7] , [8] , and in some cases can be quite frequent [9] . Mixed infections of these two parasites within individual mosquitoes also can occur in areas where more than one species of parasite is endemic and where Anopheles mosquitoes transmit both Plasmodium and filarial worm parasites , e . g . , Papua New Guinea ( PNG ) , rural sub-Saharan Africa , Asia , etc . ( for a comprehensive review see [10] ) . But the interaction of co-infection between parasites and the effects on the fitness and survival of vectors is poorly known or incomplete and based only on a handful of studies ( e . g . , [5] , [6] , [8] ) . Moreover , traditional methods used to screen mosquitoes for the presence of parasites relied on morphological criteria for vector identification and dissection of individual mosquitoes for pathogen recovery . These approaches , however , have limitations because they are time consuming , can only provide estimates of prevalence when actual prevalence is high , do not allow the distinction of mosquito species within species complexes , and cannot differentiate species of Plasmodium or conclusively separate human filarial worms from those found in other animals . Recently , alternative tests have been developed for the accurate assessment of the prevalence of these pathogens in human and vector populations [10] . Therefore , an understanding of the impact mixed infections has on parasite-vector and parasite-parasite interactions are both necessary for accurate measurements of vector infection and transmission in endemic areas . PNG is endemic for the four major species of human malaria parasites , malaria transmission occurs in all 20 PNG provinces , and malaria intensity ranges from unstable low levels of endemicity to stable malaria with year-round transmission [11] . Over one million residents of PNG are estimated to be infected with W . bancrofti , and the occurrence of microfilaremia , chronic disease , and acute disease is higher in PNG than in any other filariasis-endemic country [11] . Both malaria and LF are transmitted by mosquitoes in the Anopheles punctulatus complex in PNG [5] , [11] and co-occurrence of multiple species of malaria and LF in humans is common [9] and has been shown to occur in mosquito hosts [5] . In fact , a recent study in East Sepik Province , PNG showed that 29% of individuals examined using a molecular-based assay for the simultaneous detection of the four major human Plasmodium spp . and W . bancrofti harbored both filarial worms and one or more species of malaria parasites [9] . Interactions among parasites within the same human host are known to alter disease severity [12] , and the fact that two parasites might interact within the human host suggests that dual infection may alter the course of disease development and the dynamics of transmission [13] , [14] . Within the vector , both Plasmodium and filarial worm parasites share the common developmental step of traversing the mosquito midgut , but little is known about how these two parasites interact within the vector when they share the same midgut environment . To achieve successful results in the ongoing campaigns for malaria and LF control , it is important to understand the interactions of different species of parasite with their shared hosts [3] , because it has been hypothesized that eliminating W . bancrofti from co-endemic areas has the potential to improve the capacity of Anopheles to transmit Plasmodium [12] , [15] . Accordingly , we initiated experiments to evaluate the interaction between W . bancrofti and An . punctulatus in PNG . We found that An . punctulatus utilizes a melanization-based immune response as a natural mechanism of resistance to filarial worms in PNG . Based on these results and the fact that An . punctulatus also transmits the parasites that cause malaria , we initiated laboratory experiments to test the hypothesis that immune system activation and/or development by filarial worms in mosquitoes play a role in reducing the intensity of Plasmodium transmission in areas where they are co-endemic . To test this hypothesis , we conducted studies using two mosquito species ( Armigeres subalbatus and Aedes aegypti ) , three filarial worm species ( Brugia malayi , Brugia pahangi , and Dirofilaria immitis ) , and an avian malaria parasite ( Plasmodium gallinaceum ) . It was necessary to utilize model mosquito-parasite combinations to evaluate this hypothesis because there are no suitable animal models available for W . bancrofti and P . falciparum , and there is no laboratory colonized species of Anopheles that utilizes melanization as a natural mechanism of resistance to filarial worms . However , the melanization immune response does function as a natural mechanism of resistance to the filarial worm B . malayi in Ar . subalbatus [16] , mimicking the scenario observed with An . punctulatus and W . bancrofti in PNG . Additionally , Ar . subalbatus used in this study are natural vectors of P . gallinaceum [17] and B . pahangi [16] and they are naturally refractory to D . immitis . Therefore , this unique mosquito-parasite system provides a means to assess the relationship between filarial worms and malaria parasites with the same vector in the presence or absence of an immune response or in the presence or absence of midgut penetration . The data presented herein demonstrate that when a mosquito imbibes a bloodmeal containing both malaria and filarial worm parasites , there is a significant reduction in malaria parasite development in co-infected mosquitoes regardless of whether the mosquito has an immune response to the invading filarial worms or not . However , it needs to be determined whether these results apply to Anopheles vectors of human malaria and lymphatic filariasis in areas of co-endemicity .
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All animals and animal facilities were under the control of the School of Veterinary Medicine with oversight from the University of Wisconsin Research Animal Resource Center and the protocol was approved by the University of Wisconsin Animal Care and Use Committee ( Approval #A3368-01 ) . Ar . subalbatus and Ae . aegypti , black-eyed Liverpool ( LVP ) strain , used in this study were maintained at the University of Wisconsin-Madison as previously described [18] , [19] . Four- to five-day-old female mosquitoes were sucrose starved for 14 to 16 hours prior to bloodfeeding . P . gallinaceum infection was maintained by chicken ( Gallus gallus ) and mosquito passage . P . gallinaceum-infected blood was harvested from infected chickens via cardiac puncture , mixed , and split equally into two aliquots: experimental and control . Brugia and D . immitis mf were obtained from the Filariasis Research Reagent Repository Center ( FR3 ) ( Athens , Georgia , USA ) , filtered from cat or dog blood as described previously [20] , and mixed with the experimental P . gallinaceum-infected blood . Microfilariae concentrations for all species of filarial worm used were approximately 50–175 mf/20 µl of blood , and Plasmodium gametocytemias ranged from 1–3% from biological replicate to biological replicate . Mosquitoes were exposed to bloodmeals via water-jacketed membrane feeders maintained at 36 . 5°C [21] . Mosquitoes that fed to repletion were separated into cartons and maintained on 0 . 3 M sucrose in an environmental chamber at 26 . 5°±1° C , 75±10% relative humidity , and with a 16 hour ( h ) photoperiod with a 90 minute crepuscular period . At 6 days ( d ) post ingestion ( PI ) , mosquitoes were dissected , oocysts were counted , and Plasmodium mean intensity and prevalence was calculated . Midguts were excised in a drop of saline , transferred to a clean slide , stained with mercurochrome , and oocysts were visualized using phase contrast optics on an Olympus BH2 compound microscope at 200X magnification ( Olympus America Inc . , Center Valley , PA ) . For each biological replicate , a separate group of mosquitoes were dissected over the course of Plasmodium development to verify that oocysts were not being melanized . Stained and unstained midguts were examined using phase contrast optics on an Olympus Provis compound microscope at 200X and 400X magnification ( Olympus America Inc . , Center Valley , PA ) . Ar . subalbatus were exposed to a single bloodmeal containing a mixture of B . malayi mf and P . gallinaceum or a mixture of B . pahangi mf and P . gallinaceum . Controls were mosquitoes from the same cohort exposed to P . gallinaceum-infected blood . Control P . gallinaceum-infected blood had an equivalent amount of saline added to it to control for the saline that was added with mf to the dually infected blood even though the proportion was small . Both the concurrent ingestion of B . malayi mf and P . gallinaceum and B . pahangi mf and P . gallinaceum experiments were performed four times with separate cohorts of mosquitoes to account for stochastic variations . An additional five mosquitoes were dissected at 6 d PI to verify B . pahangi development and at 24 h PI to verify B . malayi melanization . Ar . subalbatus were exposed to a single bloodmeal containing a mixture of D . immitis mf and P . gallinaceum . Experimental conditions mimicked those described for concurrent ingestion of Brugia and Plasmodium , and an additional five mosquitoes were dissected at 24 h PI to confirm ingestion of D . immitis mf and at 6 d PI to verify that D . immitis was not developing . This experiment was performed three times with separate cohorts of mosquitoes . Ae . aegypti were exposed to a single bloodmeal containing a mixture of B . pahangi mf and P . gallinaceum . Experimental conditions mimicked those described for concomitant infections in Ar . subalbatus . This experiment was performed three times . An additional two replicates were performed to assess Plasmodium zygote formation in Ae . aegypti approximately 20 h after ingestion of a co-infected bloodmeal . A single midgut was excised in a drop of saline , transferred to a clean microscope slide , homogenized in three µl of fetal bovine serum , smeared on the slide , and stained with Giemsa . Zygotes were visualized using bright field optics on an Olympus Provis compound microscope at 400X magnification ( Olympus America Inc . , Center Valley , PA ) . P . gallinaceum zygotes were identified as described by [17] . Ar . subalbatus were exposed to an initial infection with P . gallinaceum in their first bloodmeal by feeding on ketamine/xylazine anesthetized chickens ( gametocytemias = 2-4% ) . Six days later , following oviposition , they were exposed to a B . malayi infective bloodmeal in their second feeding by feeding on ketamine/xylazine anesthetized gerbils , Meriones unguiculatus ( microfilaremias = 100 mf/20 µl ) . Controls were mosquitoes from the same cohort exposed to the same P . gallinaceum infected chicken in their first bloodmeal and six days later , bloodfed on uninfected gerbils . This experiment was repeated ( n = 2 biological replicates ) with separate cohorts of mosquitoes . At 2 days following their second feeding ( 8 days post P . gallinaceum exposure ) , mosquitoes were dissected to determine Plasmodium mean intensity and prevalence . B . malayi microfilariae ( 50 mf/20 µl ) were cultured in serum-free RPMI 1640 ( GIBCO ) supplemented with five g/L glucose and antibiotic-antimycotic ( Invitrogen , 100 U/ml penicillin , 100 µg/ml streptomycin , and 0 . 25 µg/mL amphotericin B ) . Spent media was collected and replaced with fresh media every 24 hours to a maximum time of 3 days . The medium collected was filtered through 0 . 2 µM filters ( Millipore ) , pooled and concentrated using Amicon Ultrafilters with 3 kDa cut-off membranes , and stored at −80°C until use . Excretory/secretory ( E/S ) product concentrations were estimated based on OD280 using a Nanodrop ND-1000 Spectrophotometer ( Thermo Fisher Scientific , San Jose , CA ) [22] . P . gallinaceum-infected blood was supplemented with B . malayi E/S product to a final concentration of 0 . 025 mg/ml and fed to mosquitoes via a water-jacketed membrane feeder . Controls were mosquitoes from the same cohort exposed to the same P . gallinaceum-infected blood supplemented with an equivalent amount of media added as was added with E/S product . At 6 d PI , 50 mosquitoes were dissected to determine Plasmodium mean intensity and prevalence . Ar . subalbatus were initially exposed to P . gallinaceum by feeding on a ketamine/xylazine anesthetized chicken , with a gametocytemia of 3 . 2% . Immediately following exposure to P . gallinaceum , fully blood fed mosquitoes were intrathoracically injected with approximately 200 D . immitis mf . Controls were mosquitoes from the same cohort exposed to the same P . gallinaceum-infected chicken . Immediately , following blood feeding , approximately 0 . 5 µl of Aedes saline , without mf , was intrathoracically injected into mosquitoes of the control group . At 6 d PI , 38 mosquitoes were dissected to determine oocyst mean intensity in the control group and 21 mosquitoes in the experimental group . An additional five mosquitoes were dissected at 24 h PI to verify D . immitis melanization . Mean intensity is here defined as the mean number of oocysts per infected mosquito . Prevalence is defined as the number of infected hosts per the number of hosts examined . Comparisons of prevalence were analyzed using an Exact unconditional test , and comparisons of mean intensity were analyzed using a Bootstrap t-test as described in [23] and [24] . Statistical tests were run using Quantitative Parasitology 3 . 0 , a software package designed to analyze the highly aggregated frequency distributions exhibited by parasites [23] .
In PNG , the interaction of W . bancrofti with its Anopheles vectors has generally been considered one of facilitation , i . e . , the proportional conversion of infective-stage larvae ( L3 ) in mosquito vectors increases as the density of circulating microfilariae ( mf ) increases from very low numbers ( e . g . , ∼10 mf/ml blood ) to intermediate levels ( e . g . , ∼100 mf/ml blood ) . When mf densities are relatively high ( e . g . , ∼1000 mf/ml blood ) , however , there is a reduction in the intensity of mosquito infections [25] . It has been reported previously that prevalence of W . bancrofti ranged from 2% to 11 . 7% in An . punctulatus in East Sepik Province [26] , and sporozoite rates of the An . punctulatus group of mosquitoes seldom exceed 3% [27] . We conducted experiments at the field station facilities of the PNG Institute of Medical Research in Maprik to evaluate the interaction between W . bancrofti and An . punctulatus . Mosquitoes ( n = 418 ) were collected in the village complex of Drekikire in the early morning as they rested inside village homes . Seventy-two of the dissected mosquitoes harbored some stage of W . bancrofti ( 17 . 2% prevalence ) and a total of 242 parasites were recovered ( 101 mf , 71 L1 , 56 L2 , and 14 L3 ) . Of the 72 infected mosquitoes , nearly 50% ( 35/72 ) employed an innate immune response called melanization against these parasites ( see [28] ) , and a total of 54 parasites were melanized and killed . In addition , 14 of the infected mosquitoes had killed all of their parasites , providing an estimated resistance rate of 19 . 4% . This is one of the few instances where melanization has been shown to function as a primary mechanism controlling resistance in a natural vector population ( see [28] ) , and it seems that this response is the primary factor controlling facilitation in this mosquito-parasite interaction . Based on these data and previous reports , we hypothesized that immune system activation and/or development by filarial worms in mosquitoes play a role in reducing the intensity of Plasmodium transmission in areas where they are co-endemic [12] , [15] . When Ar . subalbatus ingests mf in a bloodmeal , penetration of the mosquito midgut epithelium occurs shortly after ingestion ( within minutes ) [29] . If Ar . subalbatus ingests mf of B . pahangi , migration to the thoracic musculature follows and is complete by approximately 12 h PI . If the mosquito ingests mf of B . malayi , midgut penetration occurs , but mf are rapidly melanized in the hemocoel [30] , [31] . At 24 to 48 h PI mf begin to die , and by 72 h PI , the response is all but complete [18] , [32] . In contrast , P . gallinaceum penetration into the mosquito midgut is comparatively a much longer process than filarial worm penetration , i . e . , filarial worms penetrate in a matter of minutes whereas malaria parasites penetrate many hours after ingestion . Ingestion of P . gallinaceum gametocytes by Ar . subalbatus during a bloodmeal activates the formation of gametes in the mosquito midgut lumen , which undergo syngamy to form a zygote . The zygotes transform into motile ookinetes 30 h later , move out of the blood bolus , and migrate across the peritrophic matrix [17] , [33]–[35] . Ookinetes exit the midgut epithelium through the basal end and transform into sessile oocysts , which are evident on the midgut approximately 2 d PI . Therefore , P . gallinaceum development was assessed 6 d PI , i . e . , a time when development was well established and easily visualized . Our first goal was to determine if a melanization-based immune response activated by B . malayi had any effect on P . gallinaceum development in Ar . subalbatus . Mosquitoes that ingested blood containing P . gallinaceum alone ( control ) or both P . gallinaceum and B . malayi ( experimental ) were assessed for Plasmodium development , and there was a significant reduction in the intensity ( Bootstrap t-test ) of Plasmodium infection in mosquitoes exposed to both parasites as compared to P . gallinaceum alone ( Figure 1A–D ) , and there was no evidence of melanization against oocysts in any replicate . In three of the replicates , there also was a significant reduction in the prevalence ( Exact unconditional test ) of infection ( Figure 1B–D ) and for the other there was a close to statistically significant reduction ( p = 0 . 052 ) ( Figure 1A ) . These results supported our initial hypothesis that immune system activation by filarial worms in mosquitoes negatively affects Plasmodium development , but it was not clear if activation of the mosquito's immune system by filarial worms was in fact mediating the reduction and not some other phenomenon . To ascertain if the melanization immune response was mediating the reduction in Plasmodium development , we activated this immune response in the absence of midgut penetration by filarial worms . Melanization was activated by intrathoracic inoculation of D . immitis mf , which stimulates an extremely robust melanization immune response in the hemocoel of Ar . subalbatus [36] . Mosquitoes that ingested blood containing P . gallinaceum plus an intrathoracic inoculation of D . immitis mf were assessed for Plasmodium development and compared with control mosquitoes inoculated with saline without mf . Microscopic examination of midguts from each group indicated no difference in the intensity or prevalence of Plasmodium infection ( Figure 2 ) . We then conducted experiments to test if melanization had a negative effect on an established P . gallinaceum infection . Mosquitoes with an established P . gallinaceum infection were exposed to a subsequent bloodmeal containing B . malayi mf or an uninfected bloodmeal 6 d following the initial exposure to a bloodmeal containing P . gallinaceum gametocytes . Mosquito midguts were analyzed 48 h after the subsequent bloodmeal and there was no difference in the intensity or prevalence of Plasmodium infection ( Figure 3A and B ) . These results suggested that filarial worm activation of a melanization immune response was not mediating the reduction in Plasmodium development . We then investigated whether or not filarial worm development ( in the absence of melanization ) had a negative effect on P . gallinaceum development in Ar . subalbatus . Mosquitoes that fed on blood containing P . gallinaceum alone or containing both P . gallinaceum and B . pahangi ( Ar . subalbatus supports the complete development of B . pahangi ) were assessed for Plasmodium development , and there was a significant reduction in the intensity ( Figure 4A , B , and D ) and prevalence ( Figure 4B–D ) of Plasmodium infection . In one group ( Figure 4C ) there was no difference in intensity; however , within this group there were two mosquitoes with dual infections that harbored 33% of the total oocysts recovered , and if these two mosquitoes are removed from the data set , there is a significant reduction in the intensity of Plasmodium infection ( Figure 4C ) . These results suggested that midgut penetration , regardless of the melanization-based immune response , was mediating the reduction in Plasmodium development in co-infected mosquitoes; therefore , we postulated that the reduced infectivity of mosquitoes for P . gallinaceum is directly , or indirectly , related to filarial worm penetration . Our next goal was to verify that midgut penetration by mf was contributing to the reduction in P . gallinaceum development in co-infected Ar . subalbatus . Mosquitoes that fed on blood containing P . gallinaceum alone or containing both P . gallinaceum and D . immitis ( mf present in the bloodmeal but no midgut penetration ) were assessed for Plasmodium development . D . immitis is a filarial worm that does not penetrate the midgut of mosquitoes , rather it develops in the Malpighian tubules . In Ar . subalbatus , D . immitis travels to the Malpighian tubules but does not develop past the mf stage . This failure to develop is probably due to a physiological incompatibility and seems to be independent of an active immune response [37] . There was no difference in Plasmodium development in mosquitoes exposed to both parasites ( Figure 5A–C ) as compared to P . gallinaceum alone . Additionally , filarial worm excretory/secretory ( E/S ) products released in the mosquito midgut were not found to reduce P . gallinaceum development in mosquitoes that ingested P . gallinaceum-infected blood supplemented with B . malayi E/S products as compared to mosquitoes that fed on blood infected with P . gallinaceum alone ( Figure 6 ) . These results strongly suggest that midgut penetration by filarial worms is directly , or indirectly , responsible for a reduction in Plasmodium development in co-infected mosquitoes . Finally , we tested if the reduction in P . gallinaceum development was mediated by the specific physiology of the Ar . subalbatus midgut or if this phenomenon could be repeated in another species of mosquito using the same parasites . Ae . aegypti , black-eyed Liverpool strain ( which supports the complete development of B . pahangi and P . gallinaceum ) , that fed on blood containing both B . pahangi and P . gallinaceum or P . gallinaceum alone were assessed for Plasmodium development , and there was a significant reduction in the intensity and the prevalence of Plasmodium infection ( Figure 7A–C ) in co-infected mosquitoes 6 d post bloodfeeding . Additionally , no observed difference in the intensity or the prevalence of Plasmodium infection or in zygote morphology at 20 h post infection in the same mosquitoes suggested that the presence of mf does not affect Plasmodium syngamy or zygote formation in co-infected bloodmeals . These results also demonstrate that the reduced infectivity of P . gallinaceum in the presence of filarial worms could be repeated in another mosquito species . In sum , concurrent ingestion of Brugia mf and P . gallinaceum gametocytes significantly affects the development of P . gallinaceum in co-infected mosquitoes . This was demonstrated by a significant reduction in both malaria parasite intensity and prevalence in Ar . subalbatus mosquitoes with double infections and is independent of whether the mosquito has an immune response to the filarial worm ( B . malayi ) or not ( B . pahangi ) . These results lead to our belief that the reduction is related ( either directly or indirectly ) to microfilarial penetration through the mosquito midgut . Consistent with this belief is the fact that we did not observe a significant effect on Plasmodium development in mosquitoes that concurrently ingested P . gallinaceum gametocytes and D . immitis mf , P . gallinaceum gametocytes and B . malayi E/S products , P . gallinaceum gametocytes followed by intrathoracic inoculation of D . immitis mf ( melanization activated but no midgut penetration ) , or no effect on zygote formation in Ae . aegypti that ingested P . gallinaceum gametocytes and B . pahangi mf . In addition , Albuquerque and Ham ( 1995 ) showed no difference in oocyst numbers ( using their untransformed data ) in Plasmodium-infected Ae . aegypti when B . pahangi mf were inoculated into the hemocoel at 4 d post P . gallinaceum infection [38] , thereby enabling filarial worm development or immune activation without midgut penetration by mf . There are several mechanisms associated with midgut penetration by filarial worms that could account for this reduction in Plasmodium infectivity . One possibility is that damage to midgut tissue could interfere with the ability of ookinetes to traverse the midgut epithelium . In Ae . aegypti , when Brugia mf penetrate the midgut , pathology extends across two to four adjacent cells ( e . g . , the cytoplasm of adjacent cells contains vacuolated mitochondria and pycnotic nuclei ) surrounding the point of penetration and disrupts the full depth of the midgut wall resulting in the destruction of cellular integrity ( i . e . , the basal plasma membrane is disrupted and the underlying musculature is torn and partially dislodged ) [39] , and this could result in the destruction of the intracellular junctions necessary for ookinete entry into midgut cells [40]–[42] . Similar pathological consequences have been observed in An . gambiae and Ae . aegypti following W . bancrofti infection , i . e . , microfilarial penetration caused the cytoplasm of affected cells to become basophilic and their nuclei to become pycnotic [43] . The only major difference was that W . bancrofti-infected An . gambiae midgut cells showed evidence of hypertrophy , a phenomenon that has not been observed in Ae . aegypti infected with either Brugia or W . bancrofti [39] , [43] . And it has been shown that pathology associated with P . gallinaceum invasion into Ae . aegypti midgut cells persists for at least 24 h post infection [44]; therefore , the pathology associated with filarial worm penetration persists for a period of time that is long enough to have an influence on ookinete migration out of the midgut . The suggestion that midgut damage might interfere with ookinete migration through the midgut also was proposed by Kala and Gunasekaran ( 1999 ) , in studies where Ae . aegypti co-infected with P . gallinaceum and Bacillus thuringiensis ssp . israelensis ( Bti ) had a significant reduction in Plasmodium development as compared to controls . These authors suggested that the Bti toxin disrupted the midgut epithelium and interfered with the ability of ookinetes to invade midgut epithelial cells [45] . A second mechanism is that midgut penetration by filarial worms activates alternative immune-mediated mechanisms against invading mf- even if the mosquito supports the development of filarial worms- that are also active against Plasmodium parasites ( e . g . , reactive intermediates of nitrogen and oxygen , antimicrobial peptides , etc . ) . Both Ar . subalbatus and Ae . aegypti support the development of B . pahangi , but parasite tolerance may involve immunological mechanisms directed at tissue damage or other harmful substances resulting from infection with filarial worms , or may even reflect the filarial worm's ability to persistently evade the host's defenses to remain inside the host to achieve eventual transmission [46] . A number of transcripts implicated in innate immunity showed significantly different transcriptional behavior as a result of B . pahangi infection vs . uninfected blood in a study previously conducted by our laboratory [31] , and similar results were shown in a study examining the infection response of Ae . aegypti to B . malayi [47] . And these immune mechanisms could be detrimental to Plasmodium development ( especially considering that their induction loosely coincides with the time Plasmodium parasites are most vulnerable ) in concomitantly infected mosquitoes , i . e . , a particular gene may be involved in both tolerance and resistance to filarial worms but also may an have anti-Plasmodium effect , because resistance and tolerance can be mutually exclusive , interchangeable , or complementary components of a mixed strategy of defense [48] depending on the pathogens involved . A third possible mechanism involves the physical disruption of the midgut that could facilitate leakage of mosquito midgut bacteria into the hemocoel in a manner similar to what has been observed with concomitant infection involving filarial nematodes and arbovirus , i . e . , physical disruption of the midgut facilitates virus penetration into the hemocoel and enhances the vector's susceptibility to the arbovirus . In contrast , bacterial leakage into the hemocoel could be inducing a suite of antimicrobial factors that also are detrimental to Plasmodium development [49] . An additional mechanism could be related to mf-induced pathology and the subsequent repair of the midgut having a detrimental effect on Plasmodium development . In our laboratory's previous transcriptomic analyses of filarial worm associated gene expression , a number of transcripts previously implicated in apoptosis showed significantly different transcriptional behavior ( e . g . , cathepsin , calcium-independent phospholipase , etc . ) [30] , [31] , and cell death in vertebrates has been shown to trigger both innate and adaptive immune responses [50] , [51] . The destruction of basal and apical plasma membranes by penetrating mf likely results in cell death [39] and the resultant restitution of the midgut could negatively impact Plasmodium development . In Drosophila , midgut homeostasis is maintained following pathogenic bacterial infection or physical stress via the induction of cytokines in the Unpaired family . In intestinal stem cells , these cytokines activate the Jak/Stat signaling pathway , which promotes proliferation of intestinal stem cells , and also causes a gut-specific immune response in enterocytes , leading to the production of antimicrobial peptides [52] . Although there is no experimental evidence to support this phenomenon in mosquitoes , similar processes do occur ( e . g . , [53] ) , and there are transcriptomic data from our previous studies implicating the possible involvement of a number of different signal transduction molecules , cell cycle regulators , and antimicrobial peptides in both susceptibility and refractoriness of filarial worms to mosquitoes [30] , [31] that could be having inadvertent negative consequences on Plasmodium development . Regardless of the mechanism involved in mediating this phenomenon , if this laboratory model of concomitant infection is representative of what occurs naturally in areas where both malaria parasites and filarial worms are transmitted to humans by the same Anopheles vector , then the possibility exists that the elimination of filarial worms in a co-endemic locale could enhance malaria transmission . A study conducted on the Kenyan coast by Muturi et al . ( 2006 ) came to a similar conclusion even though their results showed higher sporozoites rates in W . bancrofti- and P . falciparum-infected An . gambiae . They suggested that enhancement of malaria transmission also could occur as a result of the reduction in filarial worm-induced mosquito mortality in co-endemic areas following elimination of LF [8] . It is important to consider that transmission intensity is a function of both the prevalence and intensity of infection . And our results showed a statistically significant decrease in both measures of infection in co-infected mosquitoes , which further supports the possibility of inadvertent enhancement of malaria with the elimination of LF . Additionally , the evidence provided here is consistent with studies that have examined malaria and LF co-infection in mosquitoes ( for review see [3] , [10] ) , i . e . , despite the two parasites sharing common vectors and environmental factors necessary for development , one parasite tends to dominate the other in a mixed infection [5] , [6] , [8] , [12] . At the very least , these results warrant further exploration , both in the laboratory and in the field , of the interaction of human malaria and filarial worm parasites when they co-infect an Anopheles species that functions as a natural vector for both parasites . Such studies would help to determine if attempts to control one parasite may inadvertently lead to a change in prevalence of the other [12]; because , the control of either disease depends on sufficient epidemiological knowledge before being able to propose and implement a sound intervention strategy [10] . This becomes increasingly important considering that the main aim of the Global Programme to Eliminate Lymphatic Filariasis ( GPELF ) is to achieve worldwide elimination of the parasites that cause this disease through mass drug administration by the year 2020 [54] . More importantly , the evidence provided here supports the argument for the expansion of vector control based on integrated control strategies targeting both LF and malaria [55]–[57] . Resources can be limited in many countries endemic for malaria and LF; therefore , integrating control efforts for these two diseases should be a priority . Integrated vector control has been extremely successful in the past ( e . g . , PNG , Kenya , and the Solomon Islands ) , and it is the most cost-effective approach to achieving simultaneous malaria and LF reduction or outright elimination [10] . In fact , in many situations , the timeline of the GPELF might be achieved more rapidly by incorporating vector control strategies into their program [58] . | The parasites that cause malaria and human lymphatic filariasis are both transmitted by mosquitoes , and often times in areas where these two diseases are co-endemic , mosquitoes in the genus Anopheles transmit both parasites . Currently , it is unknown how parasite transmission is effected when malaria and filarial worm parasites share the same vector . Here , we show that when these two parasites share the same mosquito host , there is a significant reduction in the intensity and prevalence of Plasmodium infections . This reduction occurs regardless of the mosquito having a melanization-based immune response activated by filarial worms or when filarial worms successfully develop within the mosquito host . We also observed that filarial worm penetration of the mosquito midgut was necessary for malaria parasite reduction to occur . Our study provides new insight into the relationship between malaria and filarial worm parasites with their mosquito host , which could impact transmission dynamics in areas where both parasites are transmitted by the same mosquito species . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"microbiology/parasitology",
"immunology/innate",
"immunity"
] | 2011 | Filarial Worms Reduce Plasmodium Infectivity in Mosquitoes |
Although the tprK gene of Treponema pallidum are thought to play a critical role in the pathogenesis of syphilis , the profile of variations in tprK during the development of human syphilis infection have remained unclear . Through next-generation sequencing , we compared the tprK gene of 14 secondary syphilis patients with that of 14 primary syphilis patients , and the results showed an increased number of variants within the seven V regions of the tprK gene in the secondary syphilis samples . The length of the sequences within each V region also presented a 3-bp changing pattern . Interestingly , the frequencies of predominant sequences within the V regions in the secondary syphilis samples were generally decreased compared with those found in the primary syphilis samples , particularly in the V7 region , where a frequency below 60% was found in up to 57% ( 8/14 ) of all secondary samples compared with 7% ( 1/14 ) of all primary samples . Moreover , the number of minor variants distributed between frequencies of 10 and 49 . 9% was increased . The alignment of all amino acid sequences within each V region of the primary and secondary syphilis samples revealed that some amino acid sequences , particularly the amino acid sequences IASDGGAIKH and IASEDGSAGNLKH in V1 , were highly stable . Additionally , the amino acid sequences in V6 also exhibited notable intrastrain heterogeneity and were likely to form a strain-specific pattern at the interstrain level . The identification of different profiles of the tprK gene in primary and secondary syphilis patients indicated that the tprK gene of T . pallidum undergoes constant variation to result in the best adaptation to the host . The highly stable peptides found in V1 are likely promising potential vaccine components . The highly heterogenetic regions ( e . g . , V6 ) could help to understand the role of tprK in immune evasion .
Syphilis is a complex chronic disease caused by infection with Treponema pallidum subsp . pallidum ( T . pallidum ) . Specifically , the disease has a series of highly distinct clinical stages [1] , which usually includes the localized chancre primary stage , the disseminated secondary stage , and the late tertiary stage in untreated individuals [2] . This pattern of successive episodes during infection evokes other similar chronic diseases in which antigenic variation explains this characteristic long-term infection [3 , 4] . Previous studies have indicated that antigenic variation in outer membrane antigens is a hallmark of many multistage infectious diseases [5–7] . Investigations of tprK from a 12-member gene family ( tpr ) of the T . pallidum have revealed that tprK is highly heterogeneous at both inter- and intrastrain levels . The sequence diversity of this gene is restricted to seven discrete variable ( V ) regions ( V1–V7 ) , which are separated by conserved sequences [8–10] . Although further investigations are needed to determine whether TprK is an outer membrane antigen [11–13] , researchers have found that infection-induced antibody responses are directly related to V regions of TprK , where sequence variations would abrogate specific antibody binding [14 , 15] . Therefore , it has been hypothesized that antigenic variations in TprK would facilitate T . pallidum to escape immune clearance and thereby allow the pathogen to persist in the host . Remarkable results using rabbit models support this hypothesis [14–16] . In our previous study [17] , we employed a more sensitive and reliable approach , next-generation sequencing ( NGS ) , to explore the tprK gene of T . pallidum directly from primary syphilis patient samples instead of rabbit-derived samples and found that variations in the V regions of tprK generally exhibited a sequence pool containing a high-proportion sequence ( frequency above 80% ) and many low-frequency minor variants ( frequency below 20% ) . Only specific V region sequences appeared at frequencies of 20–80% . Based on these findings , we were interested on the variations in the tprK gene in secondary syphilis samples . Comparisons of the variations between primary and secondary syphilis infection could provide notable information on the association of genetic variations in tprK with disease progression , help researchers gain insights into the processes underlying immune evasion by the pathogen and aid the identification of potential vaccine components for human immunology study .
The subjects included in this study were adults , and all of the subjects provided written informed consent in accordance with the institutional guidelines prior to the study . This study was approved by the Institutional Ethics Committee of Zhongshan Hospital , School of Medicine , Xiamen University , and complied with national legislations and the Declaration of Helsinki guidelines . 25 skin lesion samples ( erythema or condylomata lata ) were collected from patients with secondary syphilis . The clinical diagnosis of syphilis was based on the US Centers for Disease Control and Prevention ( CDC ) [18] and the European CDC ( ECDC ) guidelines [19] . The lesions were placed into a sterile Petri dish containing 1 mL of saline ( containing 20% normal rabbit serum ) , minced into very small pieces and squeezed into the liquid [20] . The samples were then examined by dark field microscopy , and the positive samples were used for subsequent DNA extraction . DNA extraction was performed using the QIAamp DNA Mini Kit ( Qiagen , Inc . , Valencia , CA , USA ) as previously described [21] . qPCR targeting tp0574 was performed to determine whether each DNA sample contained treponemal DNA . 14 secondary syphilis samples ( S-1~14 ) were ultimately included in this study . The samples were then subjected to molecular typing using the Enhance CDC system [22] and amplification of tp0136 to determine whether they belonged to the Nichols-like group or SS14-like group [23] . Segmented amplification of the tprK gene was conducted as described previously [17] . Briefly , the extracted DNA was directly used for amplification of the tprK gene open reading frame ( ORF ) , and the amplicons were gel purified . The purified tprK amplicons ( diluted 100× ) were used as the segmented amplification template for partial amplification of four fragments of 400–500 bp ( overlapping by at least 20 bp ) covering the tprK ORF . All the products were verified by 2% agarose gel electrophoresis and gel purified . A high-fidelity PCR polymerase , KOD FX Neo polymerase ( Toyobo , Osaka , Japan ) , was used for the amplification , and the amplification primers are shown in S1 Table . The four subfragment amplicons corresponding to each sample were mixed in equimolar amounts into one pool to produce a separate library , and a barcode was used to distinguish each sample . Library construction and sequencing were performed by Sangon Biotech Company ( Shanghai , China ) using the MiSeq platform ( Illumina , San Diego , CA , USA ) in the paired-end sequencing ( 2×300 bp ) mode . The FastQC and FASTX tools were applied to check and improve the quality of the raw sequence data , respectively . The final reads of tprK were compared with the tprK of the Seattle Nichols strain ( GenBank Accession Number AF194369 . 1 ) using Bowtie 2 ( version 2 . 1 . 0 ) to estimate the sequencing depth and coverage . Based on a previously described principle for the extraction of sequence data [17] , an in-house Perl script was applied to specifically capture DNA sequences within seven regions of the tprK gene from the raw data , both forward and reverse . Thus , the exact number of distinct sequences within seven variable regions of the tprK gene from each sample was acquired . The intrastrain heterogeneous sequences were valid if the following conditions were simultaneously verified: ( 1 ) supported by at least fifty reads and ( 2 ) with a frequency above 1% . The relative frequency of the sequences within each variable region was then calculated . To systemically present the variation characteristics of tprK at different clinical stages , we included previous data for tprK in primary syphilis patients ( X-1~14 ) for comparison purposes [17] . All statistical analyses were performed using SPSS version 22 . 0 ( SPSS , Chicago , IL , USA ) . To compare the frequencies below 60% in V7 in secondary versus primary samples , odds ratios were estimated by logistic regression . The Chi-square test was used to identify differences in the amount of variants captured in seven variable regions of the tprK gene between primary syphilis samples and secondary syphilis samples , and the distribution of minor variants between the samples at two different stages comprised three ranges ( 1–5% , 5–10% and 10%-49 . 9% ) . A two-sided P value < 0 . 05 was considered statistically significant . The raw data of tprK obtained in this study were deposited in the SRA database ( BioProject ID: PRJNA512914 ) under the following BioSample accession numbers: SAMN10690826- SAMN10690839 for S-1~14 . The data of tprK in previous studies were deposited in the SRA database ( BioProject ID: PRJNA498982 ) under the following BioSample accession numbers: SAMN10340238-SAMN10340251 for X-1~14 .
The 14 secondary syphilis samples ( S-1~14 ) were collected at Zhongshan Hospital , Xiamen University . The clinical information for all 14 patients is shown in Table 1 . The data obtained from the qPCR analysis of the tp0574 gene showed that each DNA sample contained a certain amount of treponemal DNA for amplification of full-length tprK . Molecular typing using the ECDC system detected seven different genotypes , and genotype 16d/f was the most prevalent in these 14 samples ( S2 Table ) . Based on the sequencing data for the tp0136 gene , most strains belonged to the SS14-like group , and only three strains ( S-7 , S-9 and S-12 strains ) belonged to the Nichols-like group . The median sequencing depth of the tprK segment samples ranged from 9810 . 91 to 52366 . 84 , and the coverage ranged from 99 . 34% to 99 . 61% , indicating high identity with the tprK gene of the Seattle Nichols strain ( S2 Table ) . To more clearly present the results , background data for the 14 primary syphilis samples ( X-1~14 ) obtained in previous studies were also included in Table 1 . Using the extraction strategy , distinct nucleotide sequences in the individual V regions of tprK were captured from each sample , and 491 sequences were obtained for the 14 secondary syphilis samples ( Fig 1 ) . Calculation of the relative frequencies of distinct sequences within each V region in a single strain revealed that the tprK gene in secondary syphilis samples also contained a predominant sequence within the regions ( Fig 2 ) . However , the distribution of the sequences within seven V regions presented some dispersity: the predominant variants had broader frequency spectra ( almost between 20–80% ) , and the minor variants reached higher frequencies ( above 20% ) . Compared with the sequences captured within each V region from primary syphilis samples in previous studies ( 335 in total ) ( S3 Table ) , the secondary syphilis samples presented a higher number of variants within each V region . The Chi-square test was used to identify differences in the amounts of variants captured within each V region of the tprK gene . The trend found for the sequence variability in the V regions of tprK showed no significant differences between the samples at the two different stages ( P = 0 . 767 ) , and the highest and lowest sequence variability was found in V6 and V1 , respectively ( Fig 1 ) . Compared to the frequencies of predominant sequences ( frequency almost above 80% ) within the V regions among primary syphilis samples , the frequencies of the sequences within the V regions among secondary syphilis samples were generally lower and this finding was particularly true for the V7 region , where a frequency below 60% was found in up to 57% ( 8/14 ) of the secondary samples compared with 7% ( 1/14 ) of the primary samples . We used logistic regression to estimate the odds radio for frequencies below 60% appearing in V7 among secondary versus primary samples . The odds ratio for frequencies below 60% appearing in V7 among secondary samples were 17 . 3-fold higher than those found among primary samples ( OR = 17 . 3 [95% confidence interval , 1 . 75 to 171 . 78]; P = 0 . 015 ) . Notably , the frequencies of predominant sequences in V1 among all 28 samples remained almost above 80% . In the secondary syphilis samples , tprK still contained a pool of minor variants within each V region . As shown in Fig 3 , most of the minor variants were concentrated in the frequency range of 1–5% in both groups , and the proportions in the other two frequency ranges ( 5–10% and 10–49 . 9% ) among the secondary syphilis samples were reversed relative to the distribution pattern in primary syphilis samples ( 9 . 4% and 14 . 0% , 14 . 3% and 9 . 3% , respectively ) . However , the Chi-square test was used to investigate the distribution of minor variants in these three ranges , and no significant difference was found between the primary and secondary syphilis samples ( P = 0 . 053 ) . Additionally , the length of variable sequences within the V regions in secondary syphilis samples corroborated the finding that the length of these variants within the V regions differed in multiples of 3 bp ( S4 Table ) . Compared with the lengths within the regions in primary syphilis samples , V3 and V5 in the secondary syphilis samples maintained the same forms , but the other regions showed the appearance of some new lengths ( V1 , V4 , V6 and V7 ) or the disappearance of some lengths ( V2 and V7 ) . We translated the variable nucleotide sequences within each V region in silico . No early terminations or changes in the reading frames in tprK were found among the secondary samples , and synonymous sequences were rare and also only found in V2 and V5 ( S5 Table ) . Similar to the phenomenon found among the primary syphilis samples , substantial interstrain sequence redundancy was found in each V region . Altogether , V1 , V2 and V4 showed strong shared sequence ability , and V6 showed the least shared ability region ( Table 2 ) . Furthermore , we determined whether a specific V region sequence found in the secondary syphilis samples also presented in the primary syphilis samples . After aligning the amino acid sequences that were unique to the samples at one of the two stages , we found that a number of sequences in V1 , V2 and V4 that were specific to the secondary samples were also found in the primary samples ( Fig 4A ) . Notably , the predominant sequences of the V regions also presented overlapping ( Fig 4B ) . Among the seven V regions , V2 and V5 showed more identical predominant sequences between samples at the two different stages . However , the sequences were only found in a few samples . The analysis of the sequences in V1 and V4 showed that although V1 and V4 only presented two identical predominant sequences ( IASDGGAIKH and IASEDGSAGNLKH in V1 and DVGHKKENAANVNGTVGA and DVGRKKDGAQGTVGA in V4 ) , the identical sequences showed high interstrain sharing . In addition , the frequencies of the two shared sequences in V1 reached 80% in the strains . As previously described , V6 was the most variable region of tprK . We further corroborated this feature in the context of primary and secondary syphilis infection . As shown in Fig 4A , only eight identical sequences were found in the samples at the two different stages , and the proportions of these overlapping sequences relative to the unique sequences in the primary and secondary samples were 12 . 7% ( 8/71 ) and 7 . 8% ( 8/110 ) , respectively . Moreover , none of the predominant sequences were identical ( Fig 4B ) . The levels of nucleotide diversity in V6 between each sample ( Dxy ) were calculated using DnaSP v . 6 . 12 . 01 . The Dxy nucleotide diversity in V6 across each sample was almost above 0 . 15 ( S1 Fig ) , which was in agreement with the proposed view that V6 presents high diversity among most T . pallidum strains .
With the identification of a 12-member gene family ( tpr ) in the Nichols strain of T . pallidum [24] , the antigen-coding tprK had been extensively studied because of its highly variable antigenic profile [9 , 10 , 14 , 25 , 26] . Similar to known mechanisms through which many pathogens undergo antigenic variation to evade the immune system and establish chronic infection in the host [5 , 27] , tprK is believed to play an essential role in the pathogenesis of syphilis [15 , 16] . Hence , efforts to understand tprK diversity in the context of human infection , particularly at different clinical stages , would be beneficial to the clinical elucidation of the role of tprK in successive episodes of this chronic infection and would contribute to a more in-depth understanding of the pathogenesis of syphilis . In this study , NGS was used in combination with an in-house Perl script to confirm the features characterizing the diversity of the tprK gene during natural human infection: tprK contained a predominant sequence and numerous minor variants within each V region , and most variants were found at low frequency in the range of 1 to 5% . Interestingly , in primary syphilis samples , the frequencies of predominant variants were almost above 80% , and those of minor variants were almost below 20% . However , the predominant variants in secondary syphilis samples had broader frequency spectra ( almost between 20 and 80% ) , and more minor variants reached higher frequencies with a broader range of frequencies ( above 20% ) . Combining these two different profiles of the tprK gene , it seems that the variants within the V regions in secondary syphilis samples fill in the middle zone which is almost empty in primary syphilis pattern . This finding suggested that the variations in the tprK gene might follow a logical fitness-based evolution . An analysis of T . pallidum infection at the primary stage showed that the sequences within each V region of the tprK gene presented a two-level distribution ( above 80% and below 20% ) , suggesting the high frequency sequences may be better associated with the avoidance of immune recognition . Changes in the immune environment ( development into secondary syphilis infection ) could cause the original fitted sequences within each V region to no longer facilitate the survival of T . pallidum . The original predominant sequences in V regions need to change to obtain a new better TprK epitope for T . pallidum . At present , the frequencies of the predominant sequences are lower in the populations , and certain minor variants might be selected and exhibit higher frequencies in the populations . As a result , the original sequences might disappear , and new advantageous sequences would emerge [17 , 28] . Due to the continuous evolution of the sequences of tprK , the TprK antigen in the infection process becomes increasingly diverse , which would enable T . pallidum to successively evade the antibody response and thereby establish chronic infection [15 , 26] . In addition , we demonstrated that V6 might be the first region to change in primary syphilis samples [17] . In this study , we noted that the predominant sequences in V7 among secondary syphilis samples appeared at frequencies almost below 60% ( P = 0 . 015 ) , which might suggest that variations in V7 evolved following V6 and that the region might be important for the development of secondary syphilis infection [29] . Additionally , a strict 3-bp changing pattern in each variable region was further confirmed in the secondary syphilis samples , and no frame shifts have been found [9 , 10 , 29] , which demonstrates the existence of an elaborate system for the regulation of tprK sequence variation . Substantial interstrain sequence redundancy was observed in tprK among the samples at the two different stages . Among all V regions , the amino sequences IASDGGAIKH and IASEDGSAGNLKH in V1 and the amino sequences DVGHKKENAANVNGTVGA and DVGRKKDGAQGTVGA in V4 showed strong interstrain sharing ability across 28 clinical strains . Moreover , the sequences in V1 were those that presented a relatively high frequency ( above 80% ) in the populations . In fact , the two sequences were also found to be the most stable amino acid sequences among the samples in the investigation of Pinto et al . [29] . As described previously [17] , a high-frequency amino acid sequence for antigen-coding tprK highly reflects the immune response of the host . Moreover , this sequence presented high-level interstrain sharing , that is , the sequence was found in several syphilitic patients , which indicates that TprK has a better-fitted epitope profile for allowing T . pallidum to adapt to its host . Among the seven V regions in tprK , V1 was found to be relatively stable among the samples at the two different stages , and the two sequences of V1 were most stable among the 28 clinical strains , which suggested that maintaining V1 relatively stable would be essential for the pathogen and that the stable peptides in V1 would be a promising vaccine component for future research [30] . Notably , the promising vaccine peptides found in this study might target the majority but not all of the strains . This problem should be considered further to explore the function of these peptides . Additionally , we confirmed high heterogeneity at the intrastrain level in V6 throughout the infection process . The existence of a highly diverse region in this antigen-coding gene of an isolate of T . pallidum might greatly enable the pathogen to resist binding by existing opsonic antibodies and might make the pathogen less likely to be recognized by activated macrophages [15] . The amino acid sequences of V6 also presented high diversity at the interstrain level , showing a strain-specific pattern for the sequences , which might explain why the protection of TprK was compromised and a lack of heterologous protection [26] . Currently , it is very difficult to distinguish between treatment failure ( relapse ) and reinfection in clinical practice . Myint et al . [31] used a molecular method by analyzing tprK sequences to distinguish relapse from reinfection in a patient with recurrent secondary syphilis . Based on the results , whether there is speculation that the sequences in V6 retain a high homology in a relapse case , but the sequences are highly strain-specific in a reinfection case ? This speculation requires further supporting evidence from additional future experiments . Finally , the limitations of our study should be discussed . First , the limited sample size did not support us to draw further definitive conclusions , and the study did not explore the function of the sequences within each V region . A future study could investigate the different peptides in tprK regions observed between primary and secondary syphilis to explore the potential importance of these differences in host interactions and immune evasion . Moreover , the potential promising vaccine components identified in this study could be synthesizes to investigate the immune function of these peptides and thereby lay a foundation for vaccine development . Second , the study provided information on individual V regions instead of information on a single tprK ORF . Using a novel PacBio sequencing pipeline to obtain full length of the tprK sequence could be optimal . And the data might provide useful insights into the structure and function of TprK . Third , because the samples used in this study were from one lesion rather than different lesions , we cannot completely exclude the possibility that the individual patient was infected with different strains resulting in the initial diversity in tprK , even though we verified that the samples presented a single genetic background , as demonstrated by molecular typing ( ECDC system and sequencing of tp0136 locus ) . Additionally , the same genetic background of the tested samples may require similar studies to explore the potential relationship between the genetic background and the variations in tprK . In this study , we revealed that the characteristic profiles of tprK in the context of primary and secondary infection were different , which indicated that throughout the development of the disease , T . pallidum might constantly undergo variations in its tprK gene to achieve its best adaptation to the host . Interestingly , tprK maintains a contradictory scenario during the course of infection , that is , having a relatively conserved region ( V1 ) and a highly diverse region ( V6 ) . The stable sequences in V1 and the highly heterogeneous sequences in V6 could provide important information for exploring promising potential vaccine components and the role of tprK in persistent syphilis infection . | Antigenic variation of the TprK antigen has been acknowledged to explain the persistence of Treponema pallidum in the host , however , the profile of variations in tprK during the development of human syphilis infection has not been well characterized . Here , we performed next-generation sequencing to compare the variations in tprK between primary and secondary syphilis samples . The profiles of tprK in the samples at different stages showed differences . A higher amount of pool variants within seven V regions was found in the secondary syphilis samples , and the frequencies of their predominant sequences generally decreased with increases in the number of minor variants with frequencies in the range of 10 to 49 . 9% . However , the length of variable sequences within the V regions of tprK in the secondary syphilis samples also presented a 3-bp changing pattern . Notably , the amino acid sequences IASDGGAIKH and IASEDGSAGNLKH in V1 not only presented a high proportion of interstrain sharing but also were found at a relatively high frequency ( above 80% ) in the populations . The sequences in V6 of the samples demonstrated substantial variability at the intra- and interstrain levels . These findings could provide insights into the potential syphilis vaccine components and the role of TprK in immune evasion . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"urology",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"antigenic",
"variation",
"treponematoses",
"bacterial",
"diseases",
"next-generation",
"sequencing",
"sexually",
"transmitted",
"diseases",
"genome",
"analysis",
"neglected",
"tropical",
"diseases",
"molecular",
"biology",
"techniques",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"genomics",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"antigens",
"microbial",
"pathogens",
"molecular",
"biology",
"nucleotide",
"sequencing",
"biochemistry",
"treponema",
"pallidum",
"genitourinary",
"infections",
"physiology",
"database",
"and",
"informatics",
"methods",
"transcriptome",
"analysis",
"biology",
"and",
"life",
"sciences",
"genetics",
"computational",
"biology",
"dna",
"sequencing",
"protein",
"sequencing",
"syphilis"
] | 2019 | Insights into the genetic variation profile of tprK in Treponema pallidum during the development of natural human syphilis infection |
Histone lysine ( K ) residues , which are modified by methyl- and acetyl-transferases , diversely regulate RNA synthesis . Unlike the ubiquitously activating effect of histone K acetylation , the effects of histone K methylation vary with the number of methyl groups added and with the position of these groups in the histone tails . Histone K demethylases ( KDMs ) counteract the activity of methyl-transferases and remove methyl group ( s ) from specific K residues in histones . KDM3A ( also known as JHDM2A or JMJD1A ) is an H3K9me2/1 demethylase . KDM3A performs diverse functions via the regulation of its associated genes , which are involved in spermatogenesis , metabolism , and cell differentiation . However , the mechanism by which the activity of KDM3A is regulated is largely unknown . Here , we demonstrated that mitogen- and stress-activated protein kinase 1 ( MSK1 ) specifically phosphorylates KDM3A at Ser264 ( p-KDM3A ) , which is enriched in the regulatory regions of gene loci in the human genome . p-KDM3A directly interacts with and is recruited by the transcription factor Stat1 to activate p-KDM3A target genes under heat shock conditions . The demethylation of H3K9me2 at the Stat1 binding site specifically depends on the co-expression of p-KDM3A in the heat-shocked cells . In contrast to heat shock , IFN-γ treatment does not phosphorylate KDM3A via MSK1 , thereby abrogating its downstream effects . To our knowledge , this is the first evidence that a KDM can be modified via phosphorylation to determine its specific binding to target genes in response to thermal stress .
Histone modifications , such as methylation and acetylation , regulate RNA synthesis [1] , [2] . Unlike the activating impact of acetylation , the methylation of lysine residues in histones can exert either an activating or a repressive effect on genes , depending on the number of methyl groups that are added and the position of the lysine residue in the histone tail [3] . For example , the di- or tri-methylation of lysine ( K ) 9 on histone ( H ) 3 ( H3K9me2/3 ) , H3K27me2/3 , and H4K20me3 is repressive , whereas that of H3K4me3 and H3K36me3 enhances the transcription of their target genes [4]–[6] . A major breakthrough in this field was the discovery that the methylation of histone tails is a reversible process . This discovery was based on the identification of two classes of histone lysine demethylases ( KDMs ) , namely the FAD-dependent amine oxidase LSD1 [7] and the Jumonji C ( JmjC ) domain demethylases , a family of Fe2+- and 2-oxoglutarate-dependent KDMs [8] . Among the JmjC domain demethylases , KDM3A ( also known as JHDM2A or JMJD1A ) was first identified as a testis-enriched zinc finger protein that is highly expressed in male germ cells and is involved in germ cell development [9] . KDM3A was later identified as an H3K9me2/1 demethylase that activates the expression of the androgen receptor ( AR ) gene via an androgen-dependent pathway [10] . Furthermore , KDM3A has been demonstrated to regulate genes that are involved in spermatogenesis [11] , [12] , metabolism [13] , and cell differentiation [14] . With such a broad functional diversity , the mechanism by which KDM3A regulates the appropriate gene ( s ) in vivo at the appropriate time and targets the appropriate element is of great interest . Post-translational protein modification is very important for determining the function of proteins , including JmjC domain-containing proteins such as PHF8 , which is phosphorylated by cyclin-dependent kinases ( CDK ) , inducing the dissociation of PHF8 from chromatin [15] . PHF2 is enzymatically inactive in isolation , but PKA-phosphorylated PHF2 in complex with ARID5B displays H3K9Me2 demethylase activity [16] . PKCα–phosphorylated LSD1 forms a complex with CLOCK:BMAL1 to facilitate E-box-mediated transcriptional activation [17] . However , it is unknown whether KDM3A is phosphorylated , and the consequences of such a modification are also unknown . In this study , we demonstrate that MSK1 is activated and specifically phosphorylates KDM3A at Ser264 under heat shock . The phosphorylated KDM3A ( p-KDM3A ) is enriched at the regulatory regions of gene loci and co-localizes with Stat1 in the human genome . Extensive experiments indicate that p-KDM3A directly interacts with and is recruited by Stat1 to mediate chromatin remodeling and the expression of its target genes in response to heat shock .
Histone modifications are recognized by specific proteins , including transcription factors ( TFs ) , thereby mediating functional signaling to affect chromatin condensation or remodeling near target genes [2] , [18] , [19] . Methylated H3K9 , a repressive histone mark , must be recognized and demethylated during the initiation of gene activation . Among the identified KDMs , KDM3A was the only KDM that targeted an IFNγ-activated sequence ( GAS ) in heat-shocked Jurkat cells ( S1 Figure ) . Using an antibody against pan-phosphorylated serine ( p-Ser ) to detect the proteins immunoprecipitated for phosphorylated KDM3A , we found that KDM3A was phosphorylated after 30 or 60 min of heat shock at 42°C ( the treatment of cells at 42°C for 60 min is generally defined as “heat shock” or abbreviated as “HS” in this study; it should be otherwise indicated when a shorter incubation time is applied ) ( Fig . 1A ) . This phosphorylation occurred within the first 661 aa of the N-terminus of KDM3A ( Fig . 1B ) . Analysis of mutants in which serine was substituted with alanine at 264 , 265 , 445 , and 463 aa of KDM3A revealed that only the S264A mutant abrogated the HS-induced phosphorylation of KDM3A ( Fig . 1C ) . Next , we generated an antibody against a serine-phosphorylated peptide ( cVKRK ( p ) SSENNG ) and verified its efficacy via western blot ( S2 Figure ) . Phosphorylated Ser264-KDM3A ( p-KDM3A ) was confirmed to be specifically induced under HS ( Fig . 1D ) . To explore the upstream kinase responsible for KDM3A phosphorylation under heat shock , mitogen- and stress-activated protein kinase 1 ( MSK1 ) was considered as the most likely candidate because Jil1 , the Drosophila ortholog of human MSK1 , is activated in response to heat shock [20] . Because the activation of MSK1 can be identified based on its phosphorylation at S376 ( p-MSK ) [21] , an antibody against p-MSK was used . An increased level of p-MSK was detected following extended incubation of the cells under HS ( Fig . 1E ) . In co-IP assays with antibody targeting either MSK1 or KDM3A , co-IP of KDM3A and MSK1 in their phosphorylated forms was found only under HS . In contrast , the non-phosphorylated forms of MSK1 and KDM3A were unable to interact with one another under physiological condition ( Fig . 1F ) . Furthermore , this interaction in heat-shocked cells was not affected by introducing either a dominant negative mutant of MSK1 or the S264A mutant of KDM3A ( S3 Figure ) . Next , we analyzed the specificity of activated MSK1 for KDM3A via an in vitro kinase assay using γ-32P-ATP to label the phosphorylated substrate . We demonstrated that only the GST-fused wild-type N-terminal KDM3A ( 1-394 aa ) , but not the S264A mutant ( S/A ) , was phosphorylated by MSK1 based on 32P labeling ( central panel of Fig . 1G ) . Then , MSK1 was incubated in the two GST-fused KDM3A protein fragments as described above , resulting in the specific phosphorylation of wild-type but not mutant KDM3A in vitro ( Fig . 1H ) . Furthermore , we performed an in vitro kinase assay followed by mass spectrometric analysis to determine the specific target serine of MSK1 between the two successive serine residues at 264 and 265 aa in the synthesized KDM3A peptide ( Fig . 1I ) . These in vitro data demonstrated that MSK1 specifically phosphorylates S264 of KDM3A . To determine the effect of S264 phosphorylation on KDM3A , the demethylase activity of this enzyme was examined in vitro . However , no clear changes in the activity of KDM3A with or without S264 phosphorylation were detected ( S4 Figure ) . Then , chromatin immunoprecipitation sequence ( ChIP-seq ) was performed to determine the global occupancy of p-KDM3A . Chromatin fragments were immunoprecipitated using an antibody against p-KDM3A from Jurkat cells subjected to HS ( + ) or not ( - ) or using a native KDM3A antibody from Jurkat cells not subjected to HS . A heat map containing more than 25 , 000 elements ( gene promoters ) was generated using seqMINER [22] , and the results presented in four rows based on the antibody used and the heat-shock status . These elements were separated into three clusters , consisting of 12 , 719 elements in cluster 1 ( top ) , 5 , 304 elements in cluster 2 ( middle ) , and 7 , 120 elements in cluster 3 ( bottom ) ( right panel , Fig . 2A ) . The MetaGene profiles indicated that the reads were enriched at the transcription start site ( TSS ) in cluster 1 genes , whereas both the TSS and the body of the genes were enriched in those of cluster 2 ( top and middle , left panel , Fig . 2A ) . We analyzed all of the significant peaks in each sequencing sample using SICER V1 . 1 [23] . The percentages of the peaks of p-KDM3A that occupied the 2 , 700-MB mappable genome were 0 . 49% ( HS- ) and 0 . 42% ( HS+ ) , and their distributions across the genome are shown in a pie chart ( Fig . 2B and S1 Table ) . The peaks were significantly enriched in the upstream regulatory region ( approximate 10-fold , all p<1×10−100 ) . By screening the differential SICER intervals near gene promoters ( from −5 kb to approximately +2 kb ) ( FDR threshold 10−20 ) , KDM3A and the non-treated or heat-shocked p-KDM3A target genes were identified , as shown in the Venn diagrams ( Fig . 2C and listed in S2 Table ) . Gene Ontology ( GO ) and MSigDB Pathway analyses were performed on the target genes using GREAT 2 . 0 . 2 [24] ( Fig . 2D and S5 Figure ) . Next , we performed a TF motif analysis of the p-KDM3A-binding regions under HS using MEME [25] , [26] and found that two of the three most common motifs ( RGRAA and CSDGGA ) correspond to Stat1-binding sites , indicating the genomic co-localization of p-KDM3A with Stat1 ( Fig . 2E , S6 Figure , and S3 Table ) . Then , we determined the nearest gene locus in the top 68 sites of p-KDM3A binding that displayed the most significant difference between the HS and control conditions ( S4 Table ) to determine the binding peaks of p-KDM3A at four gene loci , DNAJB1 , SERPINH1 , SMIM20 , and RNASEK , each of which is on a distinct chromosome in Jurkat cells ( Fig . 2F , bottom panel ) . In addition , profiles of the Stat1-binding peaks in HeLa S3 cells treated with or without IFN-γ [27] were used as a reference ( top panel ) . To further illustrate the relationships between p-KDM3A occupancy and the expression of selected genes , ChIP-quantitative PCR ( ChIP-qPCR ) and reverse transcription quantitative PCR ( RT-qPCR ) were performed . The data demonstrated that the occupancy of p-KDM3A at all four gene loci examined ( top panel , Fig . 2G ) and the mRNA expression of all of these genes were enhanced under HS ( bottom panel , Fig . 2G ) , suggesting a correlation between these two events in heat-shocked cells . To determine the interaction between p-KDM3A and Stat1 , we used antibodies targeting each protein to immunoprecipitate ( IP ) cell extracts for co-IP assays . We demonstrated that KDM3A and Stat1 interacted with one another only under HS ( Fig . 3A ) . Based on a GST pull-down assay , MSK1 initially bound and phosphorylated KDM3A in vitro , but only p-KDM3A interacted with GST-Stat1 ( Fig . 3B ) . By introducing S/A point mutations into KDM3A , we demonstrated that KDM3A-S264A , but not KDM3A-S265A , lacked this binding between KDM3A and Stat1 under HS ( Fig . 3C ) , indicating that phosphorylation of KDM3A at S264 is critical for Stat1 binding . Next , we mutated S264 of KDM3A to aspartate ( S/D ) to mimic the phosphorylation of KDM3A at S264 in these cells . KDM3A-S/D co-immunoprecipitated with Stat1 even without HS ( Fig . 3D ) , suggesting that although HS induces phosphorylation of both the Y701 and S727 residues of Stat1 [28] , this phosphorylation was not required for Stat1 to interact with either p-KDM3A or KDM3A-S264D . Then , we determined which region of Stat1 is required for its interaction with KDM3A-S264D in these cells . Among the Stat1 fragments S1 , S2 , S4 , and S5 that interacted with KDM3A-S/D ( Fig . 3E , top of right panel ) , the fragment S5 ( residues 129-317 , left panel ) were the least required for this interaction . Based on GST pull-down assays , only the recombinant 1-394 fragment of KDM3A in its S264D form pulled down S5-Stat1 ( Fig . 3F ) . Based on co-IP assays , HA-tagged Stat1 ( 129-317 ) interacted with full length S/D-KDM3A ( Fig . 3G ) and the shorter fragment S/D-KDM3A ( 214-306 ) ( Fig . 3H ) , indicating that this 93-aa fragment of KDM3A interacts with Stat1 . By performing another co-IP using an antibody against FLAG to detect FLAG-tagged KDM3A ( 214-306 ) , we identified the 231-317 aa fragment of Stat1 was co-precipitated ( Fig . 3I ) ; this interaction between S264D-KDM3A ( 214-306 ) and Stat1 ( 231-317 ) was further confirmed in Fig . 3J . Data from Fig . 1 and Fig . 3 revealed that p-MSK1 only interacted with p-KDM3A under HS , and p-KDM3A interacted with Stat1 even in its non-phosphorylated form . To address the detail correlations of MSK1 , KDM3A , and Stat1 in heat-shocked cells , we further showed that p-MSK1 can be co-precipitated by a 214/306aa fragment of KDM3A under HS , suggesting a likely kinase versus substrate interaction for the phosphorylation of KDM3A at S264 ( S7A Figure ) . Furthermore , the interaction of Stat1 and p-KDM3A was enhanced by extended incubation under HS , but not the interaction with p-MSK1 in the same cells and was not in the least enhanced ( S7B Figure ) . However , the fact that the 93aa fragment of p-KDM3A could be co-precipitated by a 213/317aa fragment of Stat1 under HS indicates that the phosphorylated Y701 and S727 of Stat1 were not required for its interaction with p-KDM3A ( Fig . 3J ) . Taken together , these results suggest these three factors do not exist in a complex , but sequentially take parts in the two functional stages: ( 1 ) activated MSK1 interacts and phosphorylates KDM3A-S264 under HS and ( 2 ) the recruitment of p-KDM3A via Stat1 to the promoter of target gene for HS inducing activation . Next , we analyzed the MetaGene profile of p-KDM3A at the gene locus encoding hsp90α ( hsp90aa1 ) under HS , which indicated the reads were enriched around the TSS of a cluster 1 gene . p-KDM3A under HS was markedly enriched at the TSS that is dominant over either non-heat shock p-KDM3A or non-phosphorylated KDM3A without HS ( Fig . 4A ) . Interestingly , the p-KDM3A-enriched TSS region coincidently displays IFNγ-induced Stat1 binding at the hsp90α gene locus in HeLa S3 cells ( Fig . 4A , top panel ) according to Robertson et al [27] . Therefore , hsp90α is appropriately selected as a representative gene to further evaluate the mechanism underlying the targeting and functions of p-KDM3A in the human genome . ChIP assays were then performed to examine the occupancy of p-KDM3A in the upstream sequences , its impact on the H3K9me2 level and in chromatin remodeling of hsp90α . We demonstrated that p-KDM3A was gradually enriched near the GAS element of hsp90α over time under HS ( Fig . 4B ) , while the level of endogenous H3K9me2 decreased ( Fig . 4C ) . This result suggests that p-KDM3A is directly involved in the demethylation of H3K9me2 . Interestingly , once Stat1 was knocked down using a specific shRNA , the heat-shock-induced occupancy of p-KDM3A was abrogated in these cells ( Fig . 4D ) , moreover , KDM3A-S/D mimic was no longer occupied even without HS ( S8 Figure ) . In contrast , Stat1 binding remained following KDM3A knockdown ( S9C Figure ) . ChIP/reChIP assays also demonstrated that p-KDM3A occupancy at the GAS element is Stat1-dependent ( Fig . 4E ) . For DNase I hypersensitivity analysis , we set the sensitivity level without DNase I to 1 . 00 on the y-axis , representing a 100% “resistance” to this enzyme . As the amount of DNase I increased , the resistance to DNase I digestion significantly decreased in the upstream region of hsp90α in mock shRNA-transfected cells under HS ( Fig . 4F , filled bars in left panel ) . In contrast , the HS-mediated changes in DNase I sensitivity at the GAS element were absent from KDM3A shRNA-transfected cells ( Fig . 4F , right panel ) . Furthermore , in non-functional KDM3A H1120Y mutant ( DN-KDM3A ) -transfected cells [10] , a similar profile lacking any clear changes in HS-dependent DNase I sensitivity was found ( Fig . 4G ) . These data indicate that HS-mediated DNase I sensitivity at the GAS element is dependent on KDM3A demethylase activity . The HS-induced activation of hsp90α , as revealed by RT-qPCR analysis of its mRNA expression , was markedly reduced in KDM3A-knockdown cells ( Fig . 4H ) and in DN-KDM3A-transfected cells ( Fig . 4I ) . Jil1 , the Drosophila ortholog of human MSK1 , is activated in response to heat shock [20] and phosphorylates H3 to elicit chromatin relaxation , facilitating the binding of additional regulatory proteins [21] . In this study , we demonstrated that MSK1 is also activated in heat-shocked cells , as shown in Fig . 1E . To further address the detailed functions of MSK1 in KDM3A , we transfected the cells with either shRNA ( i-MSK1 ) or a dominant negative ( DN ) mutant of MSK1; the phosphorylation of KDM3A at S264 under HS was blocked in these cells compared to the wild-type control cells ( Fig . 5A and 5B and S10A–C Figure ) . However , similar to KDM3A knockdown , MSK1 knockdown did not affect the occupancy of Stat1 upstream of hsp90α ( S10D Figure ) . i-MSK1 and DN-MSK1 also significantly impaired the mRNA expression of hap90α under HS ( Fig . 5C ) , similar to the results using i-KDM3A and DN-KDM3A ( Fig . 4H and 4I ) . These results indicate that MSK1 is the critical kinase that is responsible for the phosphorylation of KDM3A at S264 under HS . Then , we demonstrated that these reduced expression profiles in the presence of i-MSK1 and DN-MSK1 were based on a change in the occupancy of KDM3A at the GAS of hsp90α ( Fig . 5D ) ; a high expression level of H3K9me2 was detected ( Fig . 5E ) . Furthermore , using the S264A mutant of KDM3A , the MSK1-mediated occupancy of KDM3A at the GAS was abolished ( Fig . 5F ) , the expression levels of H3K9me2 remained elevated ( Fig . 5G ) , and HS-induced mRNA gene expression was markedly reduced ( Fig . 5H ) . In contrast , using the S265A mutant of KDM3A , identical results were obtained compared to wild-type KDM3A , as shown in the respective figures . Additionally , the importance of residue S264 of KDM3A was further demonstrated in KDM3A-S264A-transfected cells , which exhibited strongly reduced HS-induced DNase I hypersensitivity at the GAS region of hsp90α ( Fig . 5I ) . It is , therefore , notable that the occupancy of p-KDM3A at GAS is required for KDM3A to display its demethylase activity on H3K9me2 and elicit chromatin remodeling at the GAS to activate the hsp90α gene . MSK1 is a major kinase responsible for the phosphorylation of histone H3 , including at S10 and S28 [29] , and the phosphorylation of H3S10 facilitates the accessibility and transcriptional competence of a specific chromatin region in the genome [18] , [30] , [31] . Next , we demonstrated via western blot that the expression of phosphorylated H3S10 ( p-H3S10 ) increased in heat-shocked Jurkat cells and was inhibited by transfection with specific MSK1 shRNA ( Fig . 5J and 5K ) . A ChIP assay also verified the inhibitory effect of this shRNA on the occupancy of p-H3S10 at the GAS region under HS ( Fig . 5L ) . In addition , the ChIP assay revealed that HP1α , the only HP1 isoform in the GAS region of hsp90α , is expressed at high levels preceding HS and reduced rapidly to minimal level within the first 30 min of HS treatment in Jurkat cells ( Fig . 5M and 5N ) . Because the expression of p-H3S10 at the GAS was accompanied by an increase in acetylation of H3K9 but not H3K14 upon HS treatment [28] , the phosphorylation of H3S10 by MSK1 may provide an open chromatin structure to recruit p-KDM3A via Stat1 , thus facilitating the binding of additional regulatory proteins . This explained why the HS-induced DNase I hypersensitivity was severely impaired by the knockdown of MSK1 ( Fig . 5O ) . Although the outcome elicited by MSK1 was similar with that of the KDM3A-S264A transfected ( Fig . 5I ) , it may indicate that a novel aspect of MSK1 functioned on human chromatin remodeling under heat shock . We previously reported that in contrast to HS treatment , IFN-γ treatment does not induce the expression of hsp90α or other related genes , such as CIITA-pIV , in Jurkat cells [28] . In this study , we demonstrated that p-KDM3A occupied at the GAS region of hsp90α ( Fig . 4B ) , and its expression is efficiently induced under HS ( Fig . 4H and 4I ) . IFN-γ did not induce the mRNA expression of this gene , independent of the presence of KDM3A in these cells ( Fig . 6A ) . Unlike HS treatment , as shown in Fig . 1D and 1E , IFN-γ treatment did not induce the expression of MSK1 or activate the kinase activity of MSK1 ( Fig . 6B ) , thus preventing the specific phosphorylation of KDM3A at S264 in IFN-γ-treated cells ( Fig . 6C ) . These data indicate that only HS treatment activates MSK1 to phosphorylate KDM3A at S264 , but this pathway is not activated in IFN-γ–treated cells . Therefore , we conclude that the expression level of p-KDM3A is the critical difference between the impact of HS and IFN-γ on the activation of their target genes in Jurkat cells . To determine the mechanism by which p-KDM3A differentially functions in cells under different treatments , we transfected the cells with mutant KDM3A-S264D to mimic the phosphorylation of the critical S264 of KDM3A . We demonstrated that KDM3A-S264D occupied the GAS element of hsp90α either with or without HS treatment ( Fig . 6D ) and strongly reduced the H3K9me2 expression to the basal level ( Fig . 6E ) . In contrast , hsp90αmRNA expression and DNase I hypersensitivity for the KDM3A-S264D mutant were similar to those for the wild-type enzyme under HS but not the control conditions ( Fig . 6F and 6G ) . Then , the aforementioned transfected cells were treated with IFNγ . The ectopically expressed KDM3A-S264D was efficiently recruited to the GAS region of hsp90α and the expression level of H3K9me2 was markedly reduced in the presence or absence of IFN-γ . However , wild-type and S264A mutant KDM3A did not bind to the GAS in IFNγ-treated cells and did not display any demethylase activity on H3K9me2 ( Fig . 6H and 6I ) . Notably , KDM3A-S264D , but not the wild-type or S/A mutant counterparts , rendered hsp90α to be susceptible to IFN-γ treatment , as that shown under HS ( Fig . 6J , slanted line-filled bars compared to the open bars ) . The above results indicate that in untreated Jurkat cells , the ectopic KDM3A S/D mutant occupied the GAS and decreased the H3K9me2 level , but for an unknown reason , hsp90αmRNA expression was not induced . Therefore , we transfected wild-type and S/D mutant KDM3A into Jurkat cells to examine the occupancy of the Brg1 chromatin remodeling complex at the GAS before and after HS treatment or after IFNγ treatment . The ChIP data indicated that only when KDM3A-S/D was transfected did Brg1 efficiently occupy the GAS following both HS ( Fig . 6K ) and IFNγ treatment ( Fig . 6L ) , but this binding was never constitutive at the GAS . However , transfected KDM3A and its S/A , S/D mutants did not affect Stat1 binding at the GAS ( S11 Figure ) . This result agrees with our previous report that Brg1 is only recruited by p-Stat1 that is induced in response to HS treatment [28] . In IFNγ-treated cells , p-Stat1 also occupied the GAS [32] , possibly providing a docking site for KDM3A-S/D and activating hsp90α . Therefore , it is conceivable that Stat1-mediated p-KDM3A recruitment is necessary but not sufficient for gene activation ( Fig . 7 ) . Our data indicate that the level of gene activation under HS or IFN-γ treatment is determined by the potential for an external stimulus to activate MSK1 , which phosphorylates KDM3A . The two-step model in Fig . 7 shows that , first , MSK1-phosphorylated KDM3A is recruited by Stat1 to remove the repressive mark H3K9me2 , and second , p-Stat1 mediates Brg1 complex recruitment to fully activate the target gene .
KDM3A is the second identified JmjC domain lysine demethylase ( JHDM2A ) that is specific for the demethylation of H3K9me2/me1 . This demethylase contains a JmjC domain at 1058-1281 aa and a zinc finger domain at 662-687 aa [10] . Although certain TFs can induce KDM3A expression [13] , [33]–[35] or interact with KDM3A [11] , [14] , [36] , our understanding of the relationship between its modification and function has not been fully elucidated since its discovery . In this study , we demonstrate that KDM3A is phosphorylated at S264 by MSK1 under heat shock . Specifically , S264 of KDM3A is approximately 400 residues from the N-terminus of the zinc finger domain , which performs no known function [10] . We then perform ChIP-Seq analysis to determine the genome-wide distribution of HS-induced p-KDM3A in Jurkat cells . To our surprise , ChIP-Seq data have shown that either with or without HS , the peaks of p-KDM3A could occupy the mappable genome at a comparable percentage . We then analyze the MetaGene profiles of p-KDM3A under HS , which shows the reads are enriched around the TSS at all of the five gene loci encoding the hsp90α ( Fig . 4A ) and the other genes ( Fig . 2F ) ; while those of the constitutive p-KDM3A only show much lower or minimal occupancy at these loci ( the fourth versus the third rows in the bottom panels of Fig . 4 and 2 ) . This finding suggests the p-KDM3As , either induced under stress ( HS ) or expressed in the normal life cycle of the cells , are functionally diverse through distribution to each distinct gene locus in the genome . In addition , the occupancy of p-KDM3A on Myo7B-Lims2 site is reduced under HS . The p-KDM3A in non-HS cells is likely phosphorylated by other kinase ( s ) or even the constitutively expressed MSK1 ( Fig . 1E ) . These kinases can be activated via specific signaling pathway ( s ) , such as IFNα [21] , and exhibit their own function ( s ) on the specific constitutively expressed genes in the cells . The TF motifs from ChIP-Seq data indicate that the p-KDM3A-bound sites are similar to those of some TFs , including Stat1 . The phosphorylation of S264-KDM3A is a prerequisite for its efficient interaction with the TF Stat1 , and residues 231-317 in the coiled-coil domain of Stat1 interact with the p-KDM3A in vitro . We suggest that this Stat1/p-KDM3A interaction represents a TF that directs KDM3A to an appropriate upstream element of its target gene to demethylate H3K9me2 . Because MSK1 is activated in response to a vast array of environmental stress stimuli via the p38 or ERK pathway to phosphorylate histone and HMG proteins [37] , [38] , MSK1 is involved in chromatin remodeling [21] , [39] . We demonstrate that MSK1 is activated by HS but not IFNγ treatment and that p-KDM3A efficiently reduces the level of H3K9me2 at the GAS of hsp90α and renders this region susceptible to DNase I treatment . Our data suggest that the p-KDM3A-mediated reduction in H3K9me2 expression is a major step of gene activation in Jurkat cells . Because no gene expresses efficiently in the presence of high level of H3K9me2 in Jurkat and Raji cells in response to either HS or IFNγ treatment ( S12 Figure and ref . [28] ) . Hence the outcome of gene activation under HS or IFN-γ treatment is determined by the potential for the stimulus to activate MSK1 to phosphorylate KDM3A . KDM3A-S264D was used in this study to mimic the function of phosphorylated KDM3A-S264 in vivo . We demonstrate that this S264D mutant directly interacts with Stat1 to occupy the GAS element regardless of heat shock . Although the KDM3A-S264D mutant constitutively binds to the GAS element , H3K9me2 remains at a basal level under IFN-γ treatment , similar to the results under HS treatment; in contrast , non-phosphorylated KDM3A does not interact with Stat1 , is not recruited to the GAS element , and does not reduce the level of H3K9me2 when exposed to IFN-γ . H1120 in the JmjC domain is indispensable for the demethylase activity of KDM3A [10] . However , the phosphorylation of KDM3A-S264 exerts the same effects , including H3K9me2 reduction and DNase I hypersensitivity at Stat1 target genes . Therefore , it is logical to propose that the Stat1-mediated recruitment of the p-KDM3A represents a specific pathway by which the demethylase activity of KDM3A is regulated under heat shock . In summary , heat shock is a physical stimulus that broadly affects the expression of a variety of genes in human cells , likely in a general manner . In addition to the activation of the well-accepted heat shock factor and heat shock element ( HSF/HSE ) pathways to induce expression of heat-shock-related genes , we present a novel , generalized heat-shock-induced activation mechanism that is centered on the phosphorylation of KDM3A . ( 1 ) p-KDM3A-S264 is enriched genome-wide at the promoter region of several genes , including heat-shock-related genes , under heat shock; ( 2 ) p-KDM3A is guided by a TF to the binding element of TF in the genome; ( 3 ) the genomic occupancy of p-KDM3A at its target genes is a prerequisite for the demethylase activity of KDM3A in situ; and ( 4 ) the phosphorylation of KDM3A is specifically dependent on the upstream stimulus-dependent kinase activity of MSK1 in HS- but not IFN-γ-treated Jurkat cells .
Antibodies against KDM3A , p-MSK1 , GAPDH , H3K9me2 , and H3K9me3 and recombinant activated MSK1 were purchased from Millipore Biotech ( Billerica , MA , United States ) . The FLAG and M2 antibodies were purchased from Sigma . The GST , MSK1 , MSK2 , HA , and Stat1 antibodies were purchased from Santa Cruz Biotechnologies ( Santa Cruz , CA , US ) . The anti-phosphorylated serine ( p-Ser ) ( antibody catalog number AB1603 ) was purchased from Merck ( Darmstadt , Germany ) . A specific antibody against p-S264-KDM3A was produced by Beijing B&M Biotech ( Beijing , China ) using the synthesized peptide VKRKSSENNG , corresponding to residues 260–269 of KDM3A , as an antigen . The FLAG-tagged MSK1 eukaryotic expression plasmid was constructed by cloning MSK1 into the pcDNA6-FLAG vector using a PCR product from a Jurkat cell cDNA library . We inserted point mutations at amino acids 165 ( D to A ) and 565 ( D to A ) in full-length FLAG-MSK1 to produce DN-MSK1 [40] . The FLAG-tagged KDM3A eukaryotic expression plasmid was a gift from Dr . Zhong-Zhou Chen of China Agricultural University . We inserted a point mutation at amino acid 1120 ( H to Y ) to produce DN-KDM3A [10] , and we generated five individual point mutants of KDM3A: S264A , S265A , S445A , S463A , and S264D . The KDM3A fragment from 214-306 was subcloned using the PCR product of full-length FLAG-KDM3A . The MSK1 and KDM3A shRNA oligonucleotide sequences were designed by OriGene Technologies , Inc . ( Rockville , MD , USA ) and inserted into the HindIII/BamHI site of the pRS vector . shRNA-Stat1 was purchased from OriGene Technologies , Inc . The truncation mutants of Stat1 ( S2 and S4-S6 ) were described previously [28] . A new construct of S3 ( 317–750 aa ) was subcloned using the PCR product of full-length HA-Stat1 ( S1 ) . We constructed Stat1 ( 129–235 ) and Stat1 ( 231–317 ) . The primers that were used to generate the MSK1 , KDM3A , and Stat1 mutant plasmids are listed in S5 Table . RT-qPCR was performed as described previously [41] , [42] . The relative expression levels of DNAJB1 , SERPINH1 , SMIM20 , RNASEK , and HSP90AA1 ( hsp90α ) were normalized to those of GAPDH using the comparative CT method according to the manufacturer's instructions ( Rotor-Gene RG-3000A Real-Time PCR System , Corbett Research , Australia ) . The specific primers corresponding to the above genes are listed in S6 Table . The experiments were repeated at least three times , and statistical analysis was performed on the individual experimental sets . All of the values in the experiments are expressed as the means ± SD . The ChIP assays were performed as described previously [41] , [42] . The primers used for DNAJB1 , SERPINH1 , SMIM20 , RNASEK , and HSP90AA1 ( hsp90α ) are listed in S7 Table . The percentage of ChIP DNA relative to the input was calculated and expressed as the mean ± SD of three independent experiments [43] . For ChIP-reChIP analysis [28] , first , Jurkat cells were transiently transfected with FLAG-tagged Stat1 expression plasmids prior to further treatment . The chromatin fragments from the sonicated cells with or without HS treatment were used as the input , which was then immunoprecipitated using an anti-Flag M2 affinity gel ( F1 ) . Aliquots of the F1 chromatin fragments were reverse cross-linked to obtain DNA for qPCR assays or were saved for re-IP using an antibody against KDM3A or p-KDM3A for reChIP assays ( F2 ) . The DNA that was extracted from the chromatin fragments subjected to reChIP was re-amplified using the primer sets used for qPCR . The amount of KDM3A or p-KDM3A that was recruited by the antibody against Stat1 at 42°C was quantified relative to that recruited at 37°C , which was normalized to 1 . For ChIP-Seq , the chromatin fragments of 1×107 Jurkat cells with or without HS treatment were immunoprecipitated using IgG or an antibody against KDM3A or p-KDM3A . The DNA fragments were end-repaired , adenylated , ligated to adaptors , and PCR-amplified for 18 cycles . The PCR products corresponding to bp 250-450 were gel-purified , quantified and stored at −80°C until use for sequencing . For high-throughput sequencing , the libraries were prepared according to the manufacturer's instructions , and to the samples were analyzed using an Illumina GAIIx system for 80-nt single-end sequencing ( ABLife , Wuhan , China ) . The data were analyzed using Active Motif; the flow chart of analysis is shown in S13 Figure . After removing the adaptors and low-quality bases , the reads ( 36 bp in length ) were mapped to the human genome ( hg19 ) using the BWA algorithm with the default settings . The clean reads that passed through the Illumina purity filter and aligned with less than two mismatches and without duplicates were saved as BED files for use in subsequent analyses . The mapped reads were inserted into seqMINER to obtain the Meta Gene distribution profile , and the genes were distributed into three clusters based on their distribution profiles . The reads files were converted to Wig files , which were inserted into the IGV 2 . 3 Genome Browser with the peak height set at 4–24 to determine the peak binding profiles . For peak calling , the mapped BED files were inserted into SICER V1 . 1 [23] ( estimated false discovery rate [FDR] threshold = 1×10−10; window size: 200 bp; fragment size: 200 bp; gap size: 200 bp; hg19 genome database ) and MACS 1 . 4 . 2 ( p-value cutoff = 1×10−7; tag size: 36 bp; band width: 150 bp; model fold = 8 , 24 ) [44] using the pooled input ( control/heat shock ) and IgG experiment reads files as backgrounds . The NCBI Gene Expression Omnibus ( http://www . ncbi . nlm . nih . gov/geo/ ) accession number for the ChIP-seq data is GSE62309 . The GO and MSigDB Pathway analyses were conducted using GREAT 2 . 02 on the SICER intervals data limited to the regulator regions ( from −5 kb to approximately +2 kb of the TSS ) . The pathway analysis database in GREAT is the MSigDB from the Gene Set Enrichment Analysis . The binomial p-value reflects the significance of the targeted genes enriched in a GO term . To identify the genome sites with more p-KDM3A after heat shock , we used the p-KDM3A HS ( + ) MACS interval peaks in Active Regions ( in locations where only one sample had an interval , which defines the Active Region ) to perform a sample comparison with peak metrics against the p-KDM3A HS ( − ) . The unique intervals were annotated into genes ( between 10 kb upstream and 10 kb downstream ) . The GO analysis of these genes was described above . Transcription factor motifs were identified around p-KDM3A SICER islands ( FA files ) after heat shock using MEME ( version 4 . 9 . 1 ) [45] . The database JASPAR_CORE_2014_vertebrates was used . Jurkat cells were transiently transfected with shRNA-MSK1 or shRNA-KDM3A . A total of 1×107 cells were washed twice in PBS , and the nuclei were extracted as described above and digested with DNase I ( ranging from 0 to 80 units/ml ) on ice for 10 min . The DNase I digestion was terminated by incubating in stop buffer ( Promega , M6101 ) at 65°C for 10 min . Then , the nuclei were digested with 50 µg/ml RNase A at 37°C for 60 min and 50 µg/ml proteinase K at 50°C overnight . The genomic DNA was purified via phenol/chloroform extraction and ethanol precipitation [46] , [47] . Aliquots of 10 µg DNA were purified for qPCR using the primers described for the ChIP-qPCR assays . The GST-Stat1 fusion protein was expressed in Escherichia coli ( BL21 DE3 ) and purified using glutathione-sepharose . GST and GST-Stat1 were bound to glutathione-sepharose , and 10 µl packed beads containing 5 µg the GST or GST-Stat1 fusion protein were incubated in the product of the kinase assay for MSK1 and KDM3A . After overnight incubation at 4°C , the beads were washed three times , and the bound proteins were analyzed via western blot . The Co-IP analyses were performed using approximately 500 µg protein samples that were incubated in a specific antibody for 2 hr at 4°C . In total , 20 µl Protein A ( or G ) -agarose were added , and the samples were incubated at 4°C overnight . Then , the pellets were washed with RIPA buffer , followed by the addition of 40 µl 1× Laemmli buffer . Then , the samples were resuspended and boiled . The samples were separated via SDS-PAGE and analyzed via sequential western blot using individual antibodies [48] . Recombinant MSK1 ( Millipore Biotech ) was incubated in 1 µg purified wild-type or mutant KDM3A ( 1-394 ) in the presence of 50 µM ATP or 5 µCi [γ-32P]ATP in kinase buffer ( 10 mM Tris , pH 7 . 4; 10 mM MgCl2 , 150 mM NaCl ) for 30 min at 30°C . The reaction products were resolved via SDS–PAGE for western blot using specific antibodies; alternatively , the 32P-labeled proteins were visualized via autoradiography . Recombinant MSK1 was incubated in 1 µg of the synthesized peptide cVKRKSSENNG , corresponding to residues 260-269 of KDM3A , in the presence of 50 µM ATP in kinase buffer for 30 min at 30°C . The reaction products were purified for mass spectrometric analysis ( Institute of Microbiology , CAS , China ) . Recombinant MSK1 was incubated in full-length GST-KDM3A for the kinase assay; then , 2 µg histone from HeLa cells was added to demethylation buffer ( 50 mM Tris , pH 8 . 0 , 50 mM NaCl , 2 mM L-ascorbic acid , 1 mM α-ketoglutarate , 50 µM Fe ( NH4 ) 2 ( SO4 ) 2 ) at 37°C for 2 hr , and the reaction was terminated by adding SDS-PAGE loading buffer . The results were analyzed via western blot using specific antibodies . The numerical data in all figures are included in S1 Data . | Histone methylation regulates gene expression and can have drastic consequences for health if the process is defective . Histone lysine demethylases ( KDMs ) counteract the activity of methyl-transferases and remove methyl group ( s ) from histones . KDM3A is a H3K9me2/1 demethylase that performs diverse functions via the regulation of its target genes , which are involved in spermatogenesis , metabolism , and cell differentiation . However , the mechanisms underlying KDM3A regulation of specific genes at specific times are largely unknown . Here we found that a physiological stress—elevated temperature—induces KDM3A phosphorylation in human cells via the MSK1 kinase . This phosphorylated form of KDM3A directly interacts with the transcription factor Stat1 , which enables Stat1 to recruit KDM3A to Stat1-binding sequences at the promoters of specific target genes . KDM3A then acts to demethylate H3K9me2/1 at these targets , thereby causing specific gene expression in response to the thermal stress . We conclude that heat shock can affect the expression of many genes in human cells via a novel activation mechanism that is centered around the phosphorylation of KDM3A . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"proteins",
"gene",
"expression",
"genetics",
"biology",
"and",
"life",
"sciences",
"epigenetics",
"dna-binding",
"proteins",
"histones",
"histone",
"modification"
] | 2014 | Specific Phosphorylation of Histone Demethylase KDM3A Determines Target Gene Expression in Response to Heat Shock |
Sox10 is a dynamically regulated transcription factor gene that is essential for the development of neural crest–derived and oligodendroglial populations . Developmental genes often require multiple regulatory sequences that integrate discrete and overlapping functions to coordinate their expression . To identify Sox10 cis-regulatory elements , we integrated multiple model systems , including cell-based screens and transposon-mediated transgensis in zebrafish , to scrutinize mammalian conserved , noncoding genomic segments at the mouse Sox10 locus . We demonstrate that eight of 11 Sox10 genomic elements direct reporter gene expression in transgenic zebrafish similar to patterns observed in transgenic mice , despite an absence of observable sequence conservation between mice and zebrafish . Multiple segments direct expression in overlapping populations of neural crest derivatives and glial cells , ranging from pan-Sox10 and pan-neural crest regulatory control to the modulation of expression in subpopulations of Sox10-expressing cells , including developing melanocytes and Schwann cells . Several sequences demonstrate overlapping spatial control , yet direct expression in incompletely overlapping developmental intervals . We were able to partially explain neural crest expression patterns by the presence of head to head SoxE family binding sites within two of the elements . Moreover , we were able to use this transcription factor binding site signature to identify the corresponding zebrafish enhancers in the absence of overall sequence homology . We demonstrate the utility of zebrafish transgenesis as a high-fidelity surrogate in the dissection of mammalian gene regulation , especially those with dynamically controlled developmental expression .
The neural crest is a transient migratory population of cells that arise at the dorsal aspect of the neural tube as it closes and that give rise to myriad structures including , but not limited to , melanocytes , enteric nervous system ( ENS ) , and myelinating Schwann cells [1] . SOX10 ( SRY-box containing gene 10 ) encodes a critical transcription factor in neural crest development [2] , [3] . All neural crest cells express SOX10 upon emigration throughout the early stages of their respective journeys . Subsequently , expression is down-regulated in all neural crest–derived cells with population-specific timing . SOX10 expression is maintained in melanocytes ( mouse ) and Schwann cells ( mouse and zebrafish ) [2] , [3] , [4] , [5] , [6] , [7] , [8] . Additionally , SOX10 is expressed in presumptive oligodendrocytes; signal increases in concert with the onset of myelination and is maintained thereafter [2] , [6] , [9] , [10] . Naturally occurring and induced mutations in animal models and spontaneous mutations in the human population exemplify the developmental requirement for Sox10 during development . Human loss-of-function SOX10 mutations result in Waardenburg-Shah Syndrome ( WS4; OMIM Accession No . 277580 ) , a disorder characterized by impaired pigmentation and aganglionic megacolon consistent with Hirschsprung's Disease [11] , [12] , [13] , [14] . Furthermore , dominant-negative SOX10 mutations have been identified in patients with PCWH ( OMIM Accession No . 609136 ) , a complex syndrome characterized by the co-presentation of a WS4-like phenotype with central dysmyelinating leukodystrophy and demyelinating peripheral neuropathy consistent with Charcot-Marie-Tooth disease type I [11] , [15] . Similarly , studies in zebrafish and mice demonstrate the broad need for SOX10 during vertebrate development [4] , [8] , [16] , [17] . A missense mutation in the zebrafish sox10 coding sequence results in dramatic hypopigmentation accompanied by a reduction in enteric nervous system neuronal and Schwann cell populations [16] , [17] . In a similar fashion , mutations in the mouse Sox10 gene yield mice with impaired pigmentation and megacolon , and such models have been critical for understanding the expression and function of Sox10 [4] , [18] . Indeed , in a recent report we identified a deletion of approximately 16 kb residing approximately 47 kb upstream of Sox10 , resulting in the Sox10Hry mouse model of WS4 [19] . Although these mice present with enteric and pigmentary deficits , this deletion severely disrupts , but does not ablate , developmental expression of Sox10 . These data suggested that the deletion harbors a subset of cis-acting transcriptional regulatory elements necessary for Sox10 expression , and our functional analyses revealed at least one highly conserved segment within the deleted region that directs reporter gene expression in cultured melanocytes [19] . Comprehensive dissection of the cis-acting transcriptional regulatory sequences at SOX10 will be important for understanding the dynamic control of this transcription factor in specific neural crest derivatives and in oligodendrocytes , as well as for assessing the role of non-coding mutations in disease susceptibility in which SOX10 may play a role . Furthermore , these efforts will provide a significant first step in establishing the sequence substrates necessary for the development of cell-specific regulatory vocabularies . Here , we report a comprehensive analysis of highly conserved , non-coding sequences at the mouse Sox10 locus , integrating in vitro cell studies in multiple cell lines and transgenesis in developing zebrafish and mice . Our results indicate that regulatory control at this locus coordinates information from a wide array of sequences whose independent regulatory contributions range from discrete sub-populations of crest derivatives to near pan-neural crest or pan-Sox10 expression . Critically , our data highlight the power of our zebrafish transgenic assay , demonstrating that it provides a high fidelity read out for the regulatory function of conserved non-coding Sox10 mouse sequences even in the absence of overt sequence conservation . Our results are largely consistent with recent preliminary analysis of a subset of Sox10 sequences in mouse , including the identification of regulatory elements that direct expression in peripheral neuronal populations and non-ectomesenchymal crest derivatives . Where differences are evident , they are largely accounted for by incomplete overlap among assayed sequences or the ability of dynamic analyses in zebrafish to uncover reporter expression that is not evident in Sox10 positive populations during mammalian development . Consequently , we also provide the first report of sequences directing expression in developing melanocytes and myelinating glia , for which no previous report exists . Additionally , our data also revealed regulatory segments that function in the same cell types but within different developmental time intervals . Finally , we demonstrate that two identified regulatory sequences with pan-neural crest regulatory control harbor dimeric SoxE binding sites that account for a significant fraction of their regulatory integrity , and use this data to identify two regulatory sequences at the zebrafish sox10 locus that we propose to be their functional orthologs .
Evolutionary sequence conservation is a robust metric for the identification of putatively functional sequences [20] , [21] , [22] . In a recent study , we identified nine highly-conserved genomic segments 5′ to the mouse Sox10 coding exons [19] . We termed these Sox10-MCSs ( for Multiple-species Conserved Sequences; Figure 1A and Table 1 ) . To determine their regulatory potential , we have assayed each segment in vitro , and in vivo through transgenesis in vertebrate systems . In addition , to more closely examine the genomic region surrounding the Sox10 transcriptional start site ( TSS ) , we also assayed a genomic segment that harbors a conserved intron one element ( Sox10-MCS1 , Figure 1B ) , another that extends 5′ to the TSS ( Sox10-MCS1B , Figure 1B ) , and another harboring only sequences upstream of the TSS ( Sox10-MCS1C , Figure 1B ) . We first tested the ability of each Sox10-MCS to direct reporter gene expression in two relevant cell lines . Briefly , we cloned each mouse genomic segment ( Figure 1A and 1B ) upstream of a minimal promoter ( pe1B ) directing basal luciferase expression and transfected them independently into two cell lines that express Sox10 [cultured melanocytes ( melan-a cells ) and cultured Schwann cells ( S16 cells ) ] , and one in which Sox10 is not expressed ( NIH-3T3 cells ) . To enrich for authentic enhancers , we defined “enhancer activity” as segments that enhance reporter gene expression at a level 10-fold or higher compared to the control ( pe1B in Figure 1C through E ) . One segment ( Sox10-MCS7 ) showed enhancer activity in all three cell lines ( Figure 1C through E ) , three segments ( Sox10-MCS4 , Sox10-MCS5 , and Sox10-MCS9 ) showed enhancer activity only in the neural crest–derived cultured melanocytes and Schwann cells ( Figure 1C and D ) , and three segments ( Sox10-MCS2 , Sox10-MCS3 , and Sox10-MCS6 ) showed enhancer activity only in cultured Schwann cells ( Figure 1D ) . Interestingly , while neither Sox10-MCS1 nor Sox10-MCS1B displayed enhancer activity , Sox10-MCS1C directed reporter gene expression in both melanocytes and Schwann cells ( Figure 1F ) . These data suggest that Sox10-MCS1C may represent a proximal Sox10 enhancer element . In total , 8 of the 11 genomic fragments studied displayed enhancer activity in Sox10 expressing cell lines , suggesting that the Sox10 locus does harbor multiple , coordinated cis-acting regulatory elements . The expression of Sox10 in crest-derived populations and in oligodendrocytes of mammals and fish has been well documented [2] , [3] , [4] , [6] , [8] , [10] , [16] , [17] , [18] , [23] . To determine whether the mouse sequences direct expression in a manner consistent with endogenous sox10 , we studied the ability of each Sox10-MCS to direct reporter gene expression in developing transgenic zebrafish . Importantly , these sequences are not overtly conserved between mammals and teleosts . Furthermore , the limited number of publicly available teleost genomic sequences and the evolutionary distance among them often makes it challenging even to identify sequences conserved within this evolutionary branch ( infraclass ) . Briefly , each mouse segment was cloned upstream of the c-Fos minimal promoter and the enhanced green fluorescent protein ( EGFP ) coding sequence . The resulting constructs were injected into two-cell zebrafish embryos and evaluated for EGFP reporter expression at selected time points ( 1 to 5 days post-fertilization; dpf ) . G0 embryos showing consistent reporter gene expression were raised for germline transmission and reporter expression patterns were further evaluated in subsequent generations in two or more founder lines . In agreement with our in vitro analyses , neither Sox10-MCS1 nor Sox10-MCS1B directed reporter gene expression in zebrafish G0 embryos ( n = 400; data not shown ) . By contrast , however , Sox10-MCS1C not only directed reporter expression in crest-derived cell lines but also directed tissue-specific reporter expression within neural crest populations in vivo , including premigratory neural crest cells , early migrating melanoblasts , cranial ganglia , lateral line ganglia , Schwann cells of the peripheral motor neurons , the sympathetic ganglia , the enteric nervous system , as well as central nervous system oligodendrocytes ( Figure 2A through 2C; some data not shown ) . Sox10-MCS2 directed expression in neural crest–derived populations of the craniofacial skeleton ( data not shown ) , the otic vesicle ( data not shown ) , developing melanocytes ( data not shown ) , and in glial cells of the peripheral and central nervous systems ( Figure 2D ) . While Sox10-MCS3 displayed enhancer activity in cultured Schwann cells , analyses in zebrafish revealed only weak expression in the anterior portion of the lateral line nerve ( data not shown ) . By contrast , and consistent with our in vitro data , Sox10-MCS4 directed reporter gene expression in a manner that recapitulates endogenous zebrafish sox10 expression , except for the otic vesicle ( Figure 3 ) . Additionally , EGFP expression was also detected in oligodendrocytes at 48 hpf ( Figure 2E ) , as well as in dorsal root ganglia , sympathetic ganglia , and the enteric nervous system at 5 dpf ( Figure 2F ) . Sox10-MCS5 also directed EGFP expression in Sox10 positive populations , including the dorsal root ganglia and myelinating glia of both the peripheral and central nervous systems in a manner overlapping temporally with Sox10-MCS4 ( Figure 2G and H ) . Interestingly , while Sox10-MCS6 directed reporter gene expression in cultured Schwann cells , we did not observe any EGFP expression in zebrafish embryos injected with this construct . Indeed , this was the only identified in vitro enhancer that did not drive reporter gene expression in developing zebrafish between 24 hpf and 5 dpf . The three distal-most assayed segments ( Sox10-MCS7 , Sox10-MCS8 , and Sox10-MCS9 ) are deleted in the Sox10Hry mouse model of WS4 , which is characterized by distal intestinal aganglionosis and severe hypopigmentation [19] . We hypothesized that these deleted sequences direct expression in affected tissues . Consistent with this hypothesis , Sox10-MCS7 directs reporter gene expression in early migrating melanoblasts ( Figure 2I ) and the enteric nervous system ( data not shown ) . This element also directs expression in dorsal root ganglia and glial cell populations , including oligodendrocytes ( Figure 2J ) and Schwann cells ( data not shown ) . Although the former cells are not grossly affected in mutant mice , these data are consistent with developmental expression of Sox10 in mouse [7] , [10] , [18] , [24] and zebrafish [16] , [17] , [23] . Sox10-MCS8 directs reporter gene expression in a subpopulation of the enteric nervous system ( Figure 2K ) , Schwann cells of the peripheral nervous system ( data not shown ) , and the sympathetic ganglia ( data not shown ) . But unlike the other constructs that we assayed , which displayed consistent reporter expression patterns among independent founder lines , Sox10-MCS8 showed notable spatial expression differences among multiple lines ( n = 4 ) . This was likely a result of position effects . However , while there was variability in expression patterns from the different founders , reporter expression remained limited to sox10 expressing tissues . Moreover , reported sites of expression were consistent among two or more lines . The reporter expression for this construct alone reflects a composite profile from the expression patterns of identified founders . Both Sox10-MCS7 and Sox10-MCS8 also directed expression in forming otic and craniofacial connective tissue . Finally , Sox10-MCS9 directs reporter gene expression in developing melanocytes ( data not shown ) , the lateral line ganglia ( data not shown ) , and cranial ganglia ( Figure 2L ) . Thus , deletion of Sox10-MCS7 , Sox10-MCS8 , and Sox10-MCS9 likely explain , at least in part , the impaired pigmentation and aganglionic megacolon observed in Sox10Hry mice . Interestingly , among the elements driving reporter expression in oligodendrocytes , some direct reporter expression during incompletely overlapping developmental windows . This temporally distinct control is perhaps most notable for sequence elements Sox10-MCS5 and Sox10-MCS7 , as they direct overlapping reporter expression in the forming ventral white matter fiber tracts of the brain and spinal cord at time points between 2 dpf and 5 dpf . Although Sox10-MCS5 is detected early ( 2 dpf ) in immature oligodendrocytes ( Figure 2G ) , signal driven by Sox10-MCS7 is not readily detected in these populations until more than 24 hours later ( data not shown ) ; reporter expression driven by this sequence ultimately then overlaps with persistent expression driven by Sox10-MCS5 by 5 dpf ( Figure 2H and J ) . These observations highlight the power of temporally dynamic transgenic analyses in zebrafish , as myelination in mammals is a post-natal event and would escape mid-gestational analyses . Our analyses represent the first report of mammalian non-coding sequences at Sox10 directing reporter gene expression in zebrafish and display success consistent with recent analyses of the RET locus . To determine the fidelity of these observations , we assayed a subset of the identified enhancers in mice , prioritizing two genomic segments whose regulatory control can independently account for a significant fraction of , if not all , Sox10 expression . First , Sox10-MCS4 modulated expression in nearly every sox10 expressing structure in zebrafish . Second , Sox10-MCS7 , which is deleted in the Sox10Hry mouse model of WS4 [19] , directed similarly broad sox10-appropriate expression , including in developing melanocytes and the ENS . We therefore examined the ability of Sox10-MCS4 and Sox10-MCS7 to direct reporter gene expression in developing mouse embryos . Briefly , each genomic segment was cloned upstream of a minimal promoter ( Hsp68 ) and LacZ coding sequence , and the resulting constructs were injected into mouse pronuclei . Reporter gene expression was then studied at specific developmental time points ( E11 . 5 and E13 . 5 ) , and compared to the known expression of Sox10 , via a LacZ reporter gene inserted at the Sox10 locus [18] ( Figure 4A and B ) . Sox10-MCS4 directed reporter gene expression in nearly every tissue in which Sox10 is expressed at E11 . 5 ( n = 2 ) , at the resolution of whole mount embryo analysis ( Figure 4C and D ) . Consistent with our observations in zebrafish , we detected reporter gene expression in the cranial , dorsal root and sympathetic ganglia ( Figure 4C ) , the enteric nervous system ( data not shown ) , and in melanoblasts ( Figure 4D , inset ) . The only relevant structure in which reporter gene expression was not observed was the otic vesicle , consistent with our observations in zebrafish . Similarly , at E13 . 5 reporter expression directed by Sox10-MCS4 provided a near complete recapitulation of the endogenous Sox10 gene product ( n = 3 ) ; for example , expression was observed in sensory neurons , sympathetic ganglia , the enteric nervous system , and melanoblasts ( data not shown ) . Given the absence of sequence conservation between the corresponding sequences from these species , data resulting from the analysis of the mouse Sox10-MCS4 sequence in mice and fish display a remarkable concordance . Sox10-MCS7 likewise directed reporter gene expression consistent with both endogenous Sox10 expression at E11 . 5 ( n = 1 ) and with our analyses in transgenic zebrafish . LacZ expression was detected in the cranial ganglia ( Figure 4E ) , dorsal root ganglia ( Figure 4E ) , otic vesicle ( Figure 4E ) , melanoblasts ( Figure 4F ) , sympathetic ganglia ( Figure 4E ) , and in the enteric nervous system ( Figure 4F , inset ) . These data underscore the potential role for Sox10-MCS7 , among others , in the developmental regulation of Sox10 transcription . In particular , they suggest that this segment plays a role in Sox10 expression during melanocyte migration and development , and support the hypothesis that deletion of this element may contribute to the impaired pigmentation observed in the Sox10Hry mouse model of WS4 ( see above ) . Furthermore , Sox10-MCS7 also directs expression in the mouse enteric nervous system , suggesting that deletion of this region may also contribute to the aganglionic megacolon observed in Sox10Hry mice . Taken together , these data suggest that analysis of mouse Sox10 regulatory sequences in zebrafish serves as a reliable surrogate for their analysis in mice . These studies revealed two enhancers with near pan-neural crest regulatory control , including sequences directing expression in melanoblasts . Members of the SoxE transcription factor family ( Sox8 , Sox9 , and Sox10 ) bind DNA as either monomers , wherein they bind to single SRY-like consensus sequences ( e . g . , ACAAA ) , or as dimers that bind to two SRY-like binding sites oriented in a head-to-head fashion ( hereafter referred to as dimeric SoxE consensus sequences ) [25] . Our previous computational and functional analyses revealed highly-conserved dimeric SoxE consensus sequences in Sox10-MCS7 [19] . Furthermore , it has previously been hypothesized that SoxE proteins ( e . g . , SOX9 ) directly regulate the transcription of SOX10 [26] . To determine if SoxE proteins might regulate other genomic segments at Sox10 , we searched each Sox10-MCS for similar consensus sequences . Interestingly , both Sox10-MCS4 and Sox10-MCS7 contain dimeric SoxE consensus sequences that are perfectly conserved between mammals and chicken ( Figure 5A ) , consistent with their demonstrated biological functions and with the suggestion that these sequences have an important role in the transcriptional regulation of Sox10 . To test this hypothesis , we established a deletion series across Sox10-MCS4 ( Figure 5B ) to identify sequence intervals that are essential for aspects of its regulatory control . We assayed each resulting genomic segment for the ability to direct reporter gene expression in vitro and in vivo as before . In cultured melanocytes and Schwann cells , the assayed constructs modulated reporter gene expression with each deletion ( Figure 5C ) . Specifically , removal of 5′ sequences to create Sox10-MCS4 . 1 caused a ∼75% and ∼60% reduction of enhancer activity in melanocytes and Schwann cells , respectively . By contrast , removal of additional 5′ sequences to create Sox10-MCS4 . 2 resulted in a ∼4 . 5-fold and ∼2-fold increase in enhancer activity compared to Sox10-MCS4 in melanocytes and Schwann cells , respectively . This significant increase in activity may result from the removal of sequences adjacent to the dimeric SoxE consensus sequences in Sox10-MCS4 . 2 ( Figure 5B ) . This observation is consistent with analysis of Sox10-MCS7 , wherein in vitro activity is increased upon the deletion of sequences flanking a core 95-basepair fragment containing the dimeric SoxE consensus sequences [19] . Further deletion of sequences from the 5′ end of Sox10-MCS4 . 2 to create Sox10-MCS4 . 3 reduced enhancer activity below levels observed for Sox10-MCS4 . 2 in both cell lines . Interestingly , this deletion removes the dimeric SoxE consensus sequence , but does not disrupt a highly conserved , monomeric SoxE consensus sequence ( Figure 5B ) , suggesting that the interval containing the dimeric sequence contributes significantly to the observed activity . The remaining SoxE consensus sequence may also explain the less dramatic effect of deleting the dimeric SoxE consensus sequences in Sox10-MCS4 compared to Sox10-MCS7 ( see below ) . Finally , the removal of sequences encompassing this consensus sequence to create Sox10-MCS4 . 4 severely reduced enhancer activity below that of Sox10-MCS4 , especially in cultured melanocytes ( Figure 5C ) . These data suggest that the intervals containing SoxE binding sites play a critical role in the activation of reporter expression controlled by Sox10-MCS4 and Sox10-MCS7 in these cell lines . We thus directly assayed their necessity through independent deletion and mutation of the dimeric SoxE consensus sequences in both genomic segments ( Figure 5G ) and compared the ability of each wild-type and mutated genomic segment to direct reporter gene expression in the above cell lines . Consistent with this postulate , deletion or mutation of the dimeric SoxE consensus sequences significantly reduced the enhancer activity in both cell lines ( Figure 5G; p-values between 5 . 98×10−11 and 2 . 11×10−7 ) . This effect was highly significant for both deletion and mutation of these sequences within both elements , albeit more severe for Sox10-MCS7 ( ≥80% reduction in enhancer activity , p≤5 . 42×10−7 ) than for Sox10-MCS4 ( 50–80% reduction in activity , p≤2 . 11×10−7 ) . Together these data support the hypothesis that the dimeric SoxE consensus sequences are important for the function of both genomic segments in these two cell types ( see above ) . They further suggest that although Sox10-MCS7 appears to be highly dependent on the dimeric SoxE consensus sequences for its function , domains in addition to this motif also contribute aspects of the biological potential demonstrated by Sox10-MCS4 . To determine the in vivo functional requirement for each of the Sox10-MCS4 sequence intervals described above , including the dimeric SoxE sequences , we studied each Sox10-MCS4 fragment ( wild type and mutant ) in transgenic zebrafish embryos . Once again all constructs were passed through the germline and evaluated in multiple founder lines to ensure the integrity of our analyses . Sox10-MCS4 . 1 directed expression in a similar manner to that observed for Sox10-MCS4 , however with significantly reduced intensity . For example , low levels of expression were observed in Schwann cells surrounding spinal motor neurons , and almost no signal was observed in the sympathetic ganglia ( Figure 5D ) . In contrast , Sox10-MCS4 . 2 directed reporter gene expression at higher levels in sympathetic chain ganglia and at reduced levels in glia surrounding spinal motor neurons when compared to Sox10-MCS4 . 1 ( Figure 5E ) . Interestingly , this observed flux in signal intensity is completely consistent with our in vitro observations . A more dramatic loss of reporter gene expression was evident upon analysis of Sox10-MCS4 . 3; expression was severely reduced in populations such as the cranial and sympathetic ganglia at 24 hpf ( data not shown ) , as well as in oligodendrocytes and Schwann cells at 72 hpf ( Figure 5F ) . This is similarly consistent with our in vitro observations and with the hypothesis that the dimeric SoxE motif is important for Sox10-MCS4 function . Finally , injection of over 400 G0 zebrafish embryos with Sox10-MCS4 . 4 failed to direct detectable reporter gene expression in any structures , indicating that all sequences necessary to direct tissue-specific reporter gene expression had been deleted . These data are also consistent with our in vitro analyses and may be explained by the removal of the monomeric SoxE consensus sequence ( see above ) . To determine whether the observed reporter deficits in fish injected with Sox10-MCS4 . 3 were the result of removing the dimeric SoxE consensus sequences , we next evaluated transgene reporter expression directed by Sox10-MCS4 harboring a deletion of the dimeric consensus sequences alone . The resulting embryos displayed a decrease in EGFP-positive enteric ganglia ( compare Figure 5H and I to Figure 2F ) ; reporter gene expression was only detected in dorsally located ganglia , suggesting that the consensus sequence may be critical for a subset of cells in the enteric nervous system . Expression was also severely reduced in oligodendrocytes ( compare Figure 5H and I to Figure 2E and F ) . Interestingly , expression in Schwann cells was not disrupted ( compare Figure 5H and I to Figure 2F ) . These data were consistent upon analysis of at least two independent founder lines ( Figure 5H and I ) and suggest that the dimeric SoxE consensus sequences in Sox10-MCS4 play a more critical role regulating Sox10 expression during oligodendrocyte and ENS development than in Schwann cell development at the evaluated time points . We have shown that highly conserved mouse sequences at Sox10 direct tissue-specific reporter gene expression in developing transgenic zebrafish . These observations belie the fact that there is no observable non-coding sequence conservation between mammals and zebrafish at Sox10 ( Figure 1A ) , and are consistent with previous analyses at the RET locus [27] . To identify the zebrafish orthologs of Sox10-MCS4 and Sox10-MCS7 , we exploited the observation that these two genomic segments harbor highly conserved dimeric SoxE consensus sequences . Briefly , we computationally identified all dimeric SoxE consensus sequences in a 100 kb window upstream of the zebrafish sox10 locus . These analyses identified three such consensus sequences ( Figure 6A ) . We hypothesized that sequences encompassing these motifs would function as in vitro and in vivo enhancers in a manner overlapping Sox10-MCS4 and/or Sox10-MCS7 . Thus , we PCR amplified 500 bp surrounding each of these consensus sequences ( zf-sox10-E1 , zf-sox10-E2 , and zf-sox10-E3; Figure 6A and Table 1 ) from zebrafish genomic DNA and tested each in vitro and in vivo as above . zf-sox10-E1 displays strong enhancer activity in Schwann cells , directing reporter gene expression ∼12-fold higher than the control vector ( Figure 6B ) . In contrast , neither zf-sox10-E2 nor zf-sox10-E3 displays strong enhancer activity in either cell line . However , it is important to note that these cells do not fully represent the range of cell types or stages in cell maturation in which sox10 is expressed during development . Indeed , when assayed in vivo , both zf-sox10-E1 and zf-sox10-E3 but not zf-sox10-E2 direct EGFP expression in cell populations consistent with endogenous sox10 expression . zf-sox10-E1 directed reporter expression in scattered cells in zebrafish G0 embryos ( 24 hpf ) in a manner consistent with migrating neural crest , appearing to be positioned at multiple points along the anterior posterior axis both dorso-laterally and along a ventral path ( Figure 6C ) . Additionally , although zf-sox10-E3 did not display enhancer activity in Schwann and melanocyte cell lines , it did direct reporter gene expression in forming oligodendrocytes of the central nervous system ( Figure 6D ) of mosaic G0 embryos . Importantly , these observations underscore the benefit of evaluating regulatory control within an intact organism . Taken together , the presence of such a small number of conserved dimeric SoxE consensus sequences in the evaluated interval and the data from in vitro and in vivo functional assays strongly suggest that zf-sox10-E1 and zf-sox10-E3 are the zebrafish orthologs of Sox10-MCS4 and Sox10-MCS7 , respectively . This is further supported by the relative position of each genomic segment in the mouse and zebrafish genomes ( compare Figure 1A to 6A ) and their associated regulatory control . However , one might expect that the simple identification of conserved dimeric SoxE consensus sequences would yield a high number of elements with in vitro enhancer activity , and that these data may reflect a high false positive rate . However , note that zf-sox10-E2 did not direct reporter gene expression in cultured cells ( Figure 6B ) or developing zebrafish ( data not shown ) . To determine the frequency with which sequences encompassing a highly conserved SoxE consensus sequence will identify enhancers that function in a subset of SoxE positive tissues , we computationally identified all dimeric SoxE consensus sequences in the mouse genome . We then identified the subset of these consensus sequences that have a conservation profile consistent with Sox10-MCS4 and Sox10-MCS7 and that reside within intronic regions of genes expressed in melanocytes ( see methods for details ) . We chose to examine intronic sequences because: ( 1 ) they provide a definable genomic element in which to search; ( 2 ) introns are known to frequently harbor enhancer sequences; and ( 3 ) functional intronic SoxE binding sites have been identified within introns of other SOX10 target genes [e . g . , myelin protein zero ( MPZ ) [28]] . These analyses provided a dataset of 44 genomic segments ( Table S1 ) , which were then tested for enhancer activity in cultured melanocytes and Schwann cells as described above . Only two of the 44 genomic segments ( 4 . 5% ) increased reporter gene expression at least 10-fold higher in either cell line; both directed reporter gene expression in cultured melanocytes ( Figure 6E ) and Schwann cells ( Figure 6F ) , suggesting they may be true SoxE target enhancers . Although we anticipate that some degree of the low frequency of success may arise from the amplification and assay of incomplete regulatory modules or from their assay in inappropriate crest derivatives , we anticipate that many are simply false negatives . Thus , our data suggest that implementing this approach at orthologous loci may enrich for correctly identified crest enhancers , and on a genomic scale , greater information regarding the composition and configuration of critical binding site motifs will be necessary .
SOX10 encodes a developmentally critical transcription factor whose roles in development and disease have been well described . Mutations in SOX10 underlie defects in several neural crest–derived populations , resulting in epidermal hypopigmentation , enteric aganglionosis , and demyelinating peripheral neuropathy; outside neural crest–derived populations , they can also result in dysmyelinating pathologies of the central nervous system oligodendrocytes [11] , [13] , [14] , [15] . The pleiotropic roles of Sox10 are facilitated by tightly controlled and dynamic regulatory control , encoded by sequences at this locus but until recently , little was known of the cis-acting sequences that modulate its dynamic expression in vertebrates [29] , [30] . The first step in deciphering this regulatory lexicon is the identification of cis-acting regulatory sequences within and flanking SOX10 . We directly addressed this challenge , assaying the regulatory control exerted by highly conserved , non-coding sequences at the mouse Sox10 locus in cultured cells and in vivo through transgenesis in zebrafish and mice . Data generated in vitro and in vivo was highly concordant , and critically identified regulatory elements were consistent with all aspects of Sox10 expression during embryogenesis , in both NC-derived and non-NC-derived structures . For example , six genomic segments ( Sox10-MCS1C , Sox10-MCS2 , Sox10-MCS4 , Sox10-MCS5 , Sox10-MCS7 , and Sox10-MCS9 ) directed EGFP expression in developing melanocytes , and six genomic segments ( Sox10-MCS1C , Sox10-MCS2 , Sox10-MCS4 , Sox10-MCS5 , Sox10-MCS7 , and Sox10-MCS8 ) directed EGFP expression in Schwann cells during zebrafish embryogenesis . These are the first data showing that non-coding sequences at Sox10 direct expression in these NC derivatives . Furthermore , six genomic segments ( Sox10-MCS1C , Sox10-MCS2 , Sox10-MCS4 , Sox10-MCS5 , Sox10-MCS7 , and Sox10-MCS8 ) directed EGFP expression in oligodendrocytes; these represent the first data showing that non-coding sequences at Sox10 direct expression in this non-NC-derived cell population . Importantly , these data were generated in teleosts using mouse genomic sequence with which we detected no overt sequence conservation . In an effort to evaluate the fidelity of our observations in zebrafish , we likewise assayed two genomic segments using transgenesis in developing mice , prioritizing sequences that displayed broad Sox10-appropriate regulatory control in teleosts ( Sox10-MCS4 and Sox10-MCS7 ) . The resulting data strongly suggest that zebrafish transgenesis serves as a reliable surrogate for the assayed sequences; Sox10-MCS4 and Sox10-MCS7 both directed reporter gene expression in nearly all endogenous sites of Sox10 expression during key stages of mouse development ( Figure 4 ) . Taken together , one of the most striking observations is the degree of functional overlap shared among identified regulatory sequences . This appears to be an increasingly common observation at developmental genes [31] , [32] , [33] , [34] and suggests a degree of functional redundancy in their regulation . However , in this case , we know that this is unlikely; our recent Sox10Hry mouse model of Waardenburg-Shah Syndrome demonstrated that homozygous mutant mice harboring carry a deletion that encompasses Sox10-MCS7 , Sox10-MCS8 , and Sox10-MCS9 display severe hypopigmentation and enteric aganglionosis . Data presented in this study strongly suggests that the deletion of these three regulatory elements contribute to the enteric and pigmentary phenotype of the Sox10Hry mice [19] . However , these are not the only regulatory elements contributing to expression in the affected cell types; Sox10-MCS1C , Sox10-MCS4 , and Sox10-MCS5 , among others , are not deleted in the Sox10Hry mice yet they also direct reporter gene expression in developing melanocytes and/or the enteric nervous system ( Figure 3 ) . Thus , a parsimonious explanation for the observed data is that multiple regulatory sequences are coordinately required for the proper spatial and temporal expression of Sox10 during development , each contributing to expression in certain subsets of tissues in an additive or interactive manner . To fully address this issue , it will be necessary to perform cell- and tissue-specific ChIP , as well as deletion and mutation analysis of each Sox10-MCS within the context of the entire gene ( e . g . , in BAC transgenic mice ) . Consistent with our data , Werner et al . [30] made similar observations in their preliminary evaluation of a small number of overlapping fragments in mice . However , there are a number of notable discrepancies , especially when considering regulatory control in developing melanocytes , sympathetic ganglia and glial cells . First , our analyses revealed six genomic segments ( Sox10-MCS1C , Sox10-MCS2 , Sox10-MCS4 , Sox10-MCS5 , Sox10-MCS7 , and Sox10-MCS9 ) that direct reporter gene expression in developing melanocytes , consistent with the known role of Sox10 in these cells . Werner and colleagues [30] assayed three genomic segments that partially overlap with sequences that we assayed . ( Their sequences , termed U1 , U2 , and U3 correspond to Sox10-MCS7 , Sox10-MCS5 and Sox10-MCS4 respectively ) . None of the “U” elements directed reporter gene expression in melanoblasts , including U1 and U3 , which correspond to Sox10-MCS7 and Sox10-MCS4 , respectively . However , our analyses of these genomic segments in developing mouse embryos revealed clear expression in melanoblasts of mice at E11 , E11 . 5 and E13 . 5 ( Figure 4 and data not shown ) , consistent with our observations in transgenic zebrafish embryos ( Figure 2 ) . Importantly , U3 lacks 170 bp present in Sox10-MCS4 , suggesting that the additional sequences in Sox10-MCS4 may be important for expression in developing melanocytes . In contrast , U1 and U2 both encompass and are 38% and 49% larger than Sox10-MCS7 and Sox10-MCS5 , respectively , yet in our assays both genomic segments directed reporter gene expression in developing melanocytes; importantly , in the case of Sox10-MCS7 , this was substantiated by reporter expression in melanocyte cell lines as well as reporter expression in melanoblasts of transgenic zebrafish and mouse embryos . One possible explanation is that these U elements also harbor local repressor sequences that are active in developing melanocytes . Another possibility is that position-related effects altered the activity of U1 , U2 , and U3 such that expression in melanoblasts was not detectable . Importantly , our data strongly suggest that Sox10-MCS4 and Sox10-MCS7 are important for expression during mammalian melanocyte development , and represent the first report of transcriptional regulatory elements important for Sox10 expression in melanoblasts . Second , similarly in our evaluation of Sox10-MCS4 in transgenic zebrafish and mouse , we observed expression in the sympathetic ganglia . Once again , this contrasts with observations for U3 [30] . Interestingly , U3 ceases sequence overlap with Sox10-MCS4 at a point where our assayed deletion fragment Sox10-MCS4 . 4 begins; critical motifs contained within this fragment may work in conjunction with other transcription factor binding sites found throughout the Sox10-MCS4 sequence but are not sufficient to independently direct reporter gene expression to the sympathetic ganglia . Another possibility is that U3 harbors a local repressor within the 32 bp that it extends beyond our Sox10-MCS4 element . Third , Sox10 expression in developing oligodendrocytes and Schwann cells of the central and peripheral nervous system , respectively , is critical for the formation of all white matter fiber tracts [6] , [7] , [9] , [10] , [18] . Indeed , six of the genomic segments at Sox10 analyzed in zebrafish revealed expression in oligodendrocytes and Schwann cells ( Sox10-MCS1C , Sox10-MCS2 , Sox10-MCS4 , Sox10-MCS5 , Sox10-MCS7 , and Sox10-MCS8 ) . However , although U1-3 overlap Sox10-MCS4-7 as described above , the authors were unable to detect reporter expression in these populations because myelination is a post-natal developmental event in mammals [6] , [8] , [10] . By contrast , myelination in zebrafish begins at 2–3 days post fertilization in zebrafish and thus , the in vivo enhancer activity of genomic segments important for expression in myelinating cells can be readily detected . Our data underscore another major strength of employing zebrafish to study the transcriptional regulation of Sox10 . Indeed , this is the first report of transcriptional regulatory elements important for Sox10 expression in Schwann cells and oligodendrocytes . Additionally , we demonstrate that the same elements exert regulatory control in both populations . Additionally , in a recent report of BAC-based transgenic rescue of Sox10 deficient mouse phenotypes , Deal and colleagues [29] report the evaluation of a BAC transgene termed DelnA , wherein most sequence 5′ to the Sox10 coding sequence has been deleted . Although , this BAC appears to retain the regulatory sequences we have termed Sox10-MCS1C and Sox10-MCS2 , it directs reporter expression only in the otic canals and melanoblasts , and weakly in the superior cervical ganglia . Given our observation that when assayed independently these sequences direct expression in many neural crest derivatives , one parsimonious explanation is that , in context , these sequences must behave coordinately and/or overcome local repression in order to elicit their requisite control . Taken in combination , these observations may explain previous failures to generate effective Sox10 reporter transgenes using conventional “promoter bashing” strategies . Consequently , Sox10-MCS1C may perhaps be more readily considered a proximal enhancer than a regulatory promoter of Sox10 . Our computational and functional analyses of conserved , non-coding sequences at Sox10 have illuminated the potentially important role of SoxE regulation of Sox10 . We have previously demonstrated that deletion of a large region harboring a dimeric SoxE consensus sequence in mouse gives rise to a phenotype similar to Sox10 coding mutations [19] . Here , we also report that two genomic segments ( Sox10-MCS4 and Sox10-MCS7 ) harbor highly-conserved dimeric SoxE consensus sequences that are essential for the appropriate control of reporter gene expression in relevant cells and tissues . However , while the dimeric SoxE consensus sequences analyzed in this study are likely important for Sox10 expression , our data also indicate that they are not essential for all regulatory control exerted by the sequences containing them and , indeed , other sequences within these segments are likely to be similarly important . Consistent with this idea we also provide deletion of sequences upstream of the SoxE consensus sequences in Sox10-MCS4 also significantly reduce reporter gene expression ( Sox10-MCS4 . 1 , Figure 5C ) . Thus , it is likely that other sequences within the studied genomic segments also contribute to the appropriate regulatory control of Sox10 . Elucidation of this underlying vocabulary will be critical for a full understanding of Sox10 expression during vertebrate development . Although we successfully implemented the zebrafish transgenic assays using mouse genomic DNA with conservation restricted to mammalian and avian species , the absence of corresponding sequence conservation between mammals and fish prevented us from directly identifying the zebrafish orthologs of each mouse Sox10-MCS . We were able , however , to uncover potential zebrafish orthologs of Sox10-MCS4 and Sox10-MCS7 by cataloging dimeric SoxE consensus sequences upstream of the zebrafish sox10 gene ( n = 3 ) . Of the three fragments that harbor these consensus sequences , two ( zf-sox10-E1 and zf-sox10-E3 ) directed reporter gene expression in Sox10 relevant cells and tissues . Furthermore , we demonstrated that the likelihood of stumbling upon a Sox10-like regulatory element by chance based only on the presence of dimeric SoxE consensus sequences was very low; only two of 44 genomic segments that harbor such sequences and that reside within non-coding sequences at genes expressed in melanocytes directed reporter expression in a corresponding cell line . Thus , we propose that zf-sox10-E1 and zf-sox10-E3 are the zebrafish orthologs of Sox10-MCS4 and Sox10-MCS7 , respectively . These observations suggest that motif-based strategies may be useful for identifying the functional orthologs of cis-acting transcriptional regulatory elements in the absence of overt non-coding sequence conservation . In characterizing Sox10-MCS4 and Sox10-MCS7 , we have described the first pan-neural crest enhancers; these genomic segments direct reporter gene expression in nearly all tissues in which Sox10 is expressed , including peripheral nervous system neurons , developing melanocytes , and glial cells . As such , the constructs and transgenic zebrafish strains we have developed provide critical tools for studying the role of Sox10 , and more generally , the neural crest in human development and disease . Additionally , these genomic segments will now be scrutinized for variants that may cause or convey susceptibility to disease . The studies reported here further our knowledge of the transcriptional regulation of the Sox10 locus and underscore the complex regulatory control of this developmentally critical transcription factor . Specifically , we make several notable observations . First , multiple genomic segments direct reporter gene expression in overlapping neural crest derivates and glial cells . Second , at least two genomic segments direct expression in a pan-neural crest manner and SoxE consensus sequences are required for the integrity of their regulatory control . Third , genomic sequences derived from the mouse Sox10 locus direct appropriate reporter gene expression in developing zebrafish , even in the absence of observable conservation and demonstrate that zebrafish provide a high fidelity surrogate for the evaluation of mammalian Sox10 regulatory sequences . Finally , we provide the first report of Sox10 regulatory sequences that direct expression in developing melanocytes and glia . These findings have important implications for studying neural crest development and the etiology of related diseases , and , more generally , provide a paradigm for dissecting the regulatory control of genes implicated in human disease .
Expression constructs were generated using Gateway technology ( Invitrogen , Carlsbad , CA ) . Briefly , PCR primers containing flanking attB sites were designed to amplify each genomic region of interest . Subsequent to PCR reactions , purified products were recombined into the pDONR221 vector according to the manufacturer's specifications ( Invitrogen ) . Each insert was sequenced to ensure the absence of PCR-induced errors . Mutagenesis of MCS4 and MCS7 was performed using the QuikChange Site-Directed Mutagenesis Kit ( Stratagene , La Jolla , CA ) and appropriate mutation-bearing oligonucleotides . Subsequently , plasmid DNA isolated from each entry clone was recombined with either pLGF-E1b ( for luciferase reporter gene expression in culture cells ) [19] , pT2cfosGW ( for EGFP expression in developing zebrafish embryos ) [27] , or Hsp68-LacZ ( for LacZ expression in developing mouse embryos ) [35] destination vectors according to the manufacturer's specifications ( Invitrogen ) . Each resulting construct underwent restriction enzyme digest with BsrGI , EcoRV , or SalI , respectively , to confirm the presence of the appropriately sized insert . Immortalized melanocytes ( melan-a ) [36] , Schwann cells ( S16 ) [37] , and NIH-3T3 cells were cultured under standard conditions . 5×104 cells were placed into each well of a 96-well culture plate and transfected with luciferase reporter vectors ( see above ) using lipofectamine 2000 reagent ( Invitrogen ) according to the manufacturer's instructions . For each reaction , 0 . 25 µl of lipofectamine 2000 and 25 µl of OptiMEM I minimal growth medium ( Invitrogen ) were combined and incubated at room temperature for 10 minutes . Purified luciferase reporter vector ( 200 ng ) or the equivalent volume of water ( in the case of DNA-negative controls ) , and 2 ng of the internal control pCMV-RL renilla expression vector ( Promega , Madison , WI ) were diluted in 25 µl of OptiMEM I and combined with the lipofectamine-OptiMEM I mixture . The ∼50 µl reactions were incubated at room temperature for 20 minutes and then added to a single well of the 96-well culture plate containing cells . After a 3 hour incubation at 37°C , the medium was aspirated and normal growth medium added . After a 48 hour incubation at 37°C , cells were washed with 1X PBS and lysed at room temperature using 1X Passive Lysis Buffer ( Promega ) . A total of 4 µl of the resulting cell lysate were transferred to a white polystyrene 96-well assay plate ( Corning Inc . , Corning , NY ) . Luciferase and renilla activities were determined using the Dual Luciferase Reporter 1000 Assay System ( Promega ) and a model Centro LB 960 luminometer ( Berthold Technologies , Bad Wildbad , Germany ) . Each experiment was performed at least 16 times , and the ratio of luciferase to renilla activity , and the fold increase in this ratio over that observed for pLGF-E1b with no insert were calculated . The mean ( bar height in figures ) , standard deviation ( error bars in figures ) , and p-values ( asterisks in figures ) were determined using standard calculations . Zebrafish were raised and bred in accordance with standard conditions [38] , [39] . Embryos were maintained at 28°C and staged in accordance with standard methods [38] , [39] . EGFP expression constructs ( see above ) were injected into AB background G0 embryos ( n≥200 ) , as previously described [27] , [40] . Injected embryos were evaluated for reporter expression between 24 hpf and 5 dpf . Embryos displaying consistent EGFP expression were selected and allowed to mature , facilitating germline transmission and evaluation of reporter expression . Embryos were analyzed and imaged using a Carl Zeiss Lumar V12 Stereo microscope with AxioVision version 4 . 5 software . Embryos selected for in situ hybridization were raised in embryo medium containing 0 . 003% phenylthiocarbamide to prevent pigmentation and fixed for in situ hybridization using standard protocols . Probes were labeled with digoxigenin using DIG RNA Labeling Kit ( Roche Applied Science , Indianapolis , IN ) and detected with the appropriate antibody . A blue precipitate was formed by incubating with BCIP and NBT ( Roche Applied Science ) . Green fluorescent protein ( GFP ) probe was used as previously described [27] . Linearized plasmid DNA for microinjections was isolated using gel electrophoresis followed by electro-elution into 1X TAE buffer . DNA was further purified using Elutip-d column ion exchange chromotography ( Schleicher & Schuell , Keene , NH ) . The resulting elutant was precipitated and resuspended into injection buffer ( 7 . 5 mM Tris-HCl , 0 . 15 mM EDTA , pH 8 . 0 at a concentration of 2 µg/ml ) . Microinjections into FVB-N/Tac pronuclei were performed as previously described [41] . Subsequent to injections and embryo dissection , embryos were fixed on ice in 1X PBS , 1% formaldehyde , 0 . 2% glutaraldehyde , and 0 . 02% NP40 for 2 hours , followed by 3 15-minute room-temperature washes in 1X PBS , 2 mM MgCl2 , and 0 . 02% NP40 . Embryos were then stained overnight in 1X PBS , 12 mM K-Ferricyanide , 12mM K-Ferrocyanide , 0 . 002% NP40 , 4 mM MgCl2 , and 320 µg/ml 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside in N , N-dimethyl formamide at 37°C . Embryos were then washed twice in 1X PBS and 0 . 2% NP40 for 30 minutes at room temperature and transferred for analysis and storage into 4% formaldehyde , 10% methanol , and 100 mM sodium phosphate . To identify dimeric SoxE consensus sequences throughout the mouse genome , we wrote a Perl computer program to scan mouse genomic sequences [the February 2006 assembly ( mm8 ) at the UCSC Genome Browser] and report any occurrence of ACAAA ( N2-10 ) TKTGT , where N2–10 represents between 2 and 10 N's ( any nucleotide ) and K represents a G or T nucleotide . Subsequently , we determined the genomic context of each identified sequence ( intronic , exonic , upstream of gene , downstream of gene ) using the Table Browser at the UCSC Genome Browser ( http://genome . ucsc . edu/cgi-bin/hgTables ) . To identify genomic sequences similar to Sox10-MCS4 and Sox10-MCS7 , we examined the conservation profile of these two genomic segments at the UCSC genome browser . This revealed that each genomic segment had a PhastCons conservation score of at least 400 ( genome . ucsc . edu ) . The Table Browser at the UCSC Genome Browser was then employed to find genomic segments that: ( 1 ) harbor dimeric SoxE consensus sequences ( see above ) ; ( 2 ) have a conservation score of at least 400; and ( 3 ) reside within an intron . The gene name and RefSeq identifier for each gene was then extracted from the Table Browser . This criteria was subsequently compared to a melanocyte expressed gene list , which was generated as defined by >2-fold higher expression in melan-a as compared to 3T3 on custom printed Operon ( Huntsville , Al ) Mv3 . 0 oligonucleotide probe set representing over 16 , 000 genes ( data not shown ) . The non-coding genomic segments at each of the identified loci were then assessed for enhancer activity in cultured melanocytes and Schwann cells as described above . | The neural crest is a population of embryonic migratory stem cells . They form atop the future spinal cord and migrate throughout developing embryos and form many different cells , including the epidermal pigment cells , bone cells in the head , and nerve cells of the peripheral nervous system . In this study , we studied the genome elements responsible for expression of SOX10 , a dynamically expressed gene that is essential for neural crest development . We isolated candidate regulatory elements for SOX10 by identifying the small percentage of genomic DNA around the gene that did not vary as avian and mammalian genomes changed though evolution . We tested these fragments for their ability to regulate gene expression in zebrafish , a model system that is highly efficient for DNA-mediated expression studies and embryology . We found that even though the genome sequences were not similar to the SOX10 gene in fish , the genomic fragments were able to recapitulate the dynamic expression of SOX10 during development . Through computational analysis of the sequences , we identified a transcription factor binding site signature that identified the corresponding zebrafish SOX10 regulatory elements . This study describes a paradigm for dissecting regulation of essential genes that display complex expression patterns during development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"developmental",
"biology/molecular",
"development",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"neuroscience/neurodevelopment",
"genetics",
"and",
"genomics/functional",
"genomics"
] | 2008 | Identification of Neural Crest and Glial Enhancers at the Mouse Sox10 Locus through Transgenesis in Zebrafish |
Integrins may undergo large conformational changes during activation , but the dynamic processes and pathways remain poorly understood . We used molecular dynamics to simulate forced unbending of a complete integrin αVβ3 ectodomain in both unliganded and liganded forms . Pulling the head of the integrin readily induced changes in the integrin from a bent to an extended conformation . Pulling at a cyclic RGD ligand bound to the integrin head also extended the integrin , suggesting that force can activate integrins . Interactions at the interfaces between the hybrid and β tail domains and between the hybrid and epidermal growth factor 4 domains formed the major energy barrier along the unbending pathway , which could be overcome spontaneously in ∼1 µs to yield a partially-extended conformation that tended to rebend . By comparison , a fully-extended conformation was stable . A newly-formed coordination between the αV Asp457 and the α-genu metal ion might contribute to the stability of the fully-extended conformation . These results reveal the dynamic processes and pathways of integrin conformational changes with atomic details and provide new insights into the structural mechanisms of integrin activation .
Integrins are αβ heterodimeric transmembrane receptors for cell-cell and cell-extracellular matrix adhesions [1] . The overall shape of an integrin ectodomain is that of a large head supported by two long legs [2] , [3] . The head of αA ( or αI ) domain-lacking integrins , including the integrin αVβ3 studied here , consists of the β-propeller domain of the α subunit and the βA ( or βI ) domain of the β subunit ( Fig . 1A ) . The two legs contain the thigh domain and the calf-1 and -2 domains of the α subunit and the hybrid , plexin-semaphorin-integrin ( PSI ) , epidermal growth factor ( EGF ) 1–4 domains and the β tail domain ( βTD ) of the β subunit . Under physiological conditions , integrins may assume an inactive state with low affinities for ligands . Upon activation by extracellular or intracellular stimuli , they may change conformations and bind ligands with high affinities and transduce signals across the plasma membrane [1] . A bent conformation , where the legs are bent at the knees , or genua , between the thigh and calf-1 domains of the α subunit and between the EGF1 and EGF2 domains of the β subunit to allow the N-terminal headpiece ( from the N-termini to the knees ) to contact the C-terminal tailpiece ( from the knees to the C-termini ) ( Fig . 1A ) , was observed in all published crystal structures of complete integrin ectodomains , including αVβ3 [4]–[7] , αIIbβ3 [8] , and αXβ2 [9] . In contrast to the “closed” headpieces observed in these bent integrins , crystal structures of an integrin αIIbβ3 headpiece adopt an open conformation with the hybrid domain swinging out ∼60° [10] . Electron microscopy ( EM ) studies have observed different global conformations for integrins under different conditions , including a bent conformation , an extended conformation with a closed headpiece , and an extended conformation with an open headpiece [11]–[13] . A switchblade model has been proposed such that integrins undergo large global conformational changes from bent to extended during activation [11] , [14] . Alternatively , a deadbolt model suggests much smaller conformational changes such that the CD loop ( the β hairpin loop between the β strands C and D ) of the βTD acts as a regulable deadbolt , the relief of which unlocks the βA domain from the inactive state [15] . As transmembrane mechanical links between intracellular and extracellular environments [1] , integrins are often subjected to tensile forces applied externally at the head , where ligands bind , or internally at the cytoplasmic tail , where the cytoskeleton attaches . It seems intuitive that such pulling force may straighten a bent integrin and even induce other forms of conformational changes , leading to integrin activation . Indeed , force has been shown to strengthen integrin-mediated adhesion [16] , [17] and prolong integrin-ligand bond lifetimes [18] , [19] . Computational approaches , especially molecular dynamics ( MD ) , have been used to study force-induced integrin conformational changes . These included steered MD ( SMD ) to simulate the opening of the αVβ3 headpiece [20] , [21] , targeted MD ( TMD ) and normal mode analysis ( NMA ) to study the conformational change of the β3 headpiece [22] , TMD to simulate the hybrid domain swing-out in the headpieces of both integrin αVβ3 and αIIbβ3 [23] , NMA to investigate the unbending of the ectodomain of integrin αVβ3 that lacks the PSI , EGF1 , and EGF2 domains at the knee region [24] , and SMD to simulate the conformational change between the closed and open legs on an extended model of the complete ectodomain of integrin αIIbβ3 [8] . Extending from these studies , we performed all-atom , explicit-solvent MD simulations for the complete ectodomain of integrin αVβ3 in both unliganded and liganded forms . Constant-velocity pulling of the head gradually unbent the integrin without major distortions of individual domains . Constant-force pulling induced a rapid transition from the bent to extended conformation after a waiting time , which increased exponentially with decreasing force to extrapolate to ∼1 µs at zero force . Two groups of interactions between the headpiece and tailpiece of the β3 subunit have been identified as the major barrier along the unbending pathway . A partially-extended conformation was observed to spontaneously bend back . A fully-extended conformation remained straight , which might be stabilized at least partly by a new coordination between Asp457 of the αV subunit and the metal ion at the α-genu . Pulling the bound ligand also unbent the integrin , suggesting that force can activate integrins .
The present work includes a total of >600 ns all-atom , explicit-solvent MD simulations for four complete integrin αVβ3 ectodomains ( Table S1 ) : two unliganded ( U1 and U2 ) and two liganded forms with a cyclic RGD peptide ( L1 and L2 ) . During equilibration in a small water box ( Fig . 1B ) , the root-mean-square deviations ( RMSDs ) of all Cα atoms ( Cα RMSDs ) first increased and then leveled off at ∼4 Å ( Fig . 1C ) , indicating that equilibrium had been reached . All four equilibrated structures assumed a bent conformation with many contacts between the headpiece ( αV residues 1–600 and β3 residues 1–472 ) and tailpiece ( αV residues 601–956 and β3 residues 473–690 ) . The calculated buried solvent-accessible surface areas ( SASAs ) between the headpiece and tailpiece indicate much more extensive contacts of the β3 subunit than the αV subunit ( Fig . 1D ) . The major barrier to unbending is therefore expected to come from the β3 subunit . We used constant-velocity SMD simulations [25] to accelerate integrin unbending . The enlarged systems with bigger water boxes to accommodate unbending ( Fig . 2A ) were first equilibrated using 4–8 ns free dynamics . A force was then applied to the head of the unliganded integrin αVβ3 ( U1 or U2 ) , where ligand-binding sites are , by a spring moving at a constant velocity while the βTD was constrained to mimic the membrane anchorage ( Fig . 2B ) . The α-leg was not constrained to allow separation of two legs , should it occur . Regardless of the initial pulling directions , force became aligned along the direction from the constraint location to the pulling location after the integrin underwent a rigid-body rotation . Remarkably , pulling readily caused gradual unbending of integrin αVβ3 without major distortions of its individual domains , resulting in a fully-extended conformation with a closed headpiece and closed legs ( Fig . 2C and Videos S1 & S2 ) . Setting the pulling speed v = 2 nm ns−1 and the spring constant k = 0 . 5 kcal mol−1 Å−2 , we varied the pulling direction by loading force at different locations: βA , β-propeller , or both . Interestingly , regardless of where force was loaded , a major force peak occurred at ∼2 nm extensions ( defined as the increase of distance between the constraint and pulling locations ) ( Fig . 2D ) . Setting the pulling direction by pulling the βA domain , we next varied the pulling speed from 0 . 5 to 10 nm ns−1 and the spring constant from 0 . 1 to 1 kcal mol−1 Å−2 ( Fig . 2E ) . At v = 10 nm ns−1 and k = 1 kcal mol−1 Å−2 , the pulling force rapidly increased to ∼1 , 000 pN at the peak level , indicating a large viscous resistant to high speed extension . At v = 0 . 5 nm ns−1 , the force fluctuations with k = 0 . 1 kcal mol−1 Å−2 were much smaller than those with k = 0 . 5 kcal mol−1 Å−2 , as expected . Importantly , a major force peak was observed at similar extensions regardless of the pulling parameters . After the major peak , force dropped to a much lower level until the integrin was straightened . These results suggest that the unbending pathway was unchanged for the different pulling parameters tested . In the following simulations , we applied force to the βA domain and chose a medium pulling speed of 2 nm ns−1 and a medium spring constant of 0 . 5 kcal mol−1 Å−2 unless otherwise indicated . To assess the robustness of the results , we repeated constant-velocity SMD simulations three times ( U1 SMD 1–3 ) with different equilibrated U1 structures and observed well-aligned force-extension curves with force peaks of comparable heights and at similar extensions ( Fig . 2F ) . We also performed one constant-velocity SMD simulation for an equilibrated U2 structure ( U2 SMD ) and observed two major force peaks at the same extension range where the single force peaks were seen in the U1 simulations ( Fig . 2F ) . To elucidate the cause of the major force peak ( s ) , we analyzed the buried SASAs and hydrogen bonds ( H-bonds ) between the headpiece and tailpiece for different integrin domains ( Fig . 3 ) . The decrease in the buried SASAs of different domains during integrin unbending revealed the disruption of interdomain interactions as the originally buried surfaces were exposed . The drops of the buried SASAs of the hybrid , βTD , and EGF4 domains were found to coincide with the major force peak ( Figs . 3 and S1 ) , identifying the interactions at the hybrid/βTD and hybrid/EGF4 interfaces as the major barrier to unbending . Interactions between the βA domain and βTD also contributed to the major force peak for U2 , but these interactions were broken spontaneously during the equilibration of U1 and thus not seen in the U1 SMD simulations ( cyan curves in Figs . 3 and S1 ) . In the U1 SMD 1 & 2 , interactions at the two interfaces were broken at the same time . However , in the U1 SMD 3 and the U2 SMD , interactions at the hybrid/βTD interface were broken before those at the hybrid/EGF4 interface , giving rise to a small bump after the major force peak in the U1 SMD 3 and splitting the major force peak into two in the U2 SMD ( gray curves in Figs . 3 and S1 ) . The hybrid/βTD interface was bound mainly by polar interactions , including a salt bridge between Asp393 and Arg633 ( Figs . 4 A & B ) as well as H-bonds whose number decreased from 8 to 0 as the simulations passed the major force peak ( Figs . 3 and S1 ) . In contrast , the hybrid/EGF4 interface was bound mainly by hydrophobic interactions , involving several nonpolar residues ( Leu375 , Ile380 , Leu383 , Met387 , Met568 , Leu573 , and Leu574 ) ( Figs . 4 C & D ) . In addition to the hydrophobic interactions , several H-bonds were observed at the hybrid/EGF4 interface for U2 ( Fig . S2 ) . To examine whether the forced extension of the two knees occurred cooperatively , we analyzed three hinge angles: the thigh/calf-1 hinge angle for the α-knee and the EGF1/EGF2 and EGF2/EGF3 hinge angles for the β-knee ( Figs . 5 and S3 ) . Both the EGF1/EGF2 and EGF2/EGF3 hinges are bent in the starting structure of U1 ( where the EGF1 and EGF2 domains were modeled ) ( Fig . S4 ) ; hence , these hinge angles opened significantly during integrin extension . In the starting structure of U2 , by comparison , the β-knee is mainly bent at the EGF1/EGF2 hinge and hence this was where the major angle changes occurred . The hinge angles for the α- and β-knees increased with increasing head-tail extension concurrently ( Figs . 5 and S3 A & B ) . To examine their cooperation , different hinge angles were plotted against each other , which reveal linear relationships until the integrin was over-stretched beyond the point where the integrin was fully straightened and the flexible EGF domains began to rotate about different hinges ( Figs . 5 and S3 C & D ) . These results indicate that the interactions that hold the α/β knees in the bent conformation are cooperative . To complement and compare with the constant-velocity SMD simulations , we performed a series of constant-force SMD simulations for U1 and U2 , where constant forces were loaded to the βA domain along the same direction with the βTD constrained as before . Instead of repeating the simulations at the same force , we used five forces from 97–195 pN for U1 and a 195 pN force for U2 . Upon application of force , the head-tail distance was held at 4–5 nm initially , indicating the stability of the bent conformation , then increased suddenly , indicating a rapid conformational transition , and finally leveled off , indicating full extension ( Fig . 6A ) . The rapid transition occurred after disrupting the interactions that gave rise to the major force peaks in the constant-velocity SMD simulations , confirming that these interactions provided the major barrier to unbending . In most cases , the interactions at both the hybrid/βTD and the hybrid/EGF4 interfaces were disrupted simultaneously . In the 146 and 97 pN constant-force simulations , however , the interactions at the hybrid/βTD interface were disrupted first , resulting in an initial slow-increase phase in the head-tail distance until the interactions at the hybrid/EGF4 interface were also broken . In the 122 pN constant-force simulation , the unbending process was slowed down in the middle of the conformational transition because of the interactions at the knees between the EGF1 and EGF2 domains and between the calf-1 and calf-2 domains . Interestingly , during the simulation with the lowest force of 97 pN , an intermediate state was observed , manifesting as a short plateau with a head-tail distance of ∼12 nm ( Fig . 6A , rightward arrow ) . In the intermediate state the interactions at both the hybrid/βTD and hybrid/EGF4 interfaces were disrupted , but the interactions near the knees were still held . Modeling integrin unbending as a rapid transition between two stable states separated by a barrier , the transition rate constant should be inversely proportional to the time required to overcome the major barrier . For the five U1 simulations , it is evident that this waiting time is shortened by increasing force ( Fig . 6A , downward arrows ) , demonstrating that force accelerates unbending . This is consistent with the Bell model , which assumes that force lowers the barrier to exponentiate the transition rate [26] . Indeed , a plot of the natural log of the waiting time ( lnt ) vs . pulling force ( F ) appeared as a straight line ( Fig . 6B ) well-fitted by the Bell equation [26] , t = t0exp ( −FΔx/kBT ) , where t0 = 728±539 ns is the waiting time at zero force and Δx = 1 . 2±0 . 2 Å is the width of the energy well that kinetically traps the integrin in the bent conformation . A similar head-tail distance vs . simulation time curve was obtained for the U2 simulation at 195 pN , although a longer waiting time was required to overcome the major barrier than the U1 simulation with the same force ( Fig . 6 ) . Taking at the face value , the t0 value suggests that integrin αVβ3 could spontaneously overcome the major barrier to unbending on a time scale of 1 µs simply by thermally-excited Brownian motions . We performed two sets of free dynamics to examine the stabilities of integrin conformations along the unbending pathway after overcoming the major barrier . The first set studied a partially-extended conformation obtained right after the major force peak . Two structures from the respective trajectories of the U1 SMD 1 and U2 SMD 2 at ∼5 nm extensions were selected as starting structures for free MD 1 & 2 , respectively ( Fig . 2F ) . In the free MD simulations , the force at the βA domain was turned off . The constraint at the βTD was released in the free MD 1 but maintained in the free MD 2 . In both simulations , the integrin gradually bent back ( Fig . 7A and Video S3 ) as indicated by the decreased Cα RMSD relative to the equilibrated bent conformation and the reduced head-tail extension ( Fig . 7B ) . This is particularly evident in the free MD 1 where the Cα RMSD dropped to ∼4 Å within 21 ns and the integrin returned nearly completely to its bent conformation . Not only did the overall structure bend back , but some important interactions between the headpiece and tailpiece that were broken during the simulated unbending were also recovered . Two groups of interactions , including the polar interactions between the hybrid domain and βTD and the nonpolar interactions between the hybrid and EGF4 domains , were gauged by the respective distances between the center of mass ( COM ) of the two sidechain oxygen atoms of Asp393 and the COM of the three sidechain nitrogen atoms of Arg633 and between the COM of the sidechains of Leu375 , Ile380 , and Leu383 and the COM of the sidechains of Met568 , Leu573 , and Leu574 . These two distances were reduced when the integrin bent back ( Fig . 7C ) . In free MD 1 , the distance of the nonpolar interactions suddenly dropped to ∼4 Å and persisted to the end of the simulation , indicating that the nonpolar interactions between the hybrid and EGF4 domains were recovered . These results demonstrated again the importance of the nonpolar interactions in stabilizing the bent conformation . By comparison , the polar interactions were not recovered in the simulations because the Asp393-Arg633 distance was still much larger than that in the bent integrin ( Fig . 7C ) . The second set of stability analyses was performed on a fully-extended conformation . Two simulations , free MD 3 & 4 , were run with their starting structures selected from the trajectories of the U1 SMD 1 & 2 , respectively , at ∼16 nm extensions ( Fig . 2F ) . Free MD began after turning off the pulling force on the βA domain . The constraint at the βTD was released in the free MD 3 but maintained in the free MD 4 . In both simulations , the integrin was relaxed and slightly bent back , but remained in a globally extended conformation for >20 ns ( Fig . 8A and Video S4 ) . Both Cα RMSDs measured from the equilibrated bent conformation and from the starting extended conformation reached plateaus after ∼5 ns ( Fig . 8B ) . Interestingly , in both simulations , Asp457 of the thigh domain moved to coordinate with the Ca2+ ion at the genu of the αV subunit ( Fig . 8C ) , as observed by a sudden drop in the distance between the COM of the two Asp457 sidechain oxygens and this Ca2+ ion to <4 Å at ∼3 or ∼18 ns , respectively , which persisted throughout the remaining simulations ( Fig . 8E ) , indicating the stability of the new coordination once it was formed . It should be noted that in the bent conformation , Asp457 was too far away ( ∼14 Å ) to coordinate with the αV-genu Ca2+ ( Fig . 8D ) . Only after integrin extension was it physically possible to interact with the αV-genu Ca2+ . These results suggest a role for this newly-formed coordination involving Asp457 in stabilizing the extended conformation . We next performed constant-velocity SMD simulations to unbend the liganded integrin αVβ3 ( L1 or L2 ) by pulling its bound cyclic RGD ligand away from the constrained βTD domain ( Fig . 9A ) . In the L1 SMD 1 and L2 SMD , the integrin readily unbent with the ligand remained bound within 10 ns ( Fig . 9A and Video S5 & S6 ) . Compared to the unliganded integrin , the liganded integrin also reached a fully extended conformation with a closed headpiece , but the legs showed a greater degree of relative rotation . In the initial structure , the RGD peptide interacted with integrin αVβ3 via the ligand Asp sidechain oxygen coordinating with the metal ion in the βA domain metal ion-dependent adhesion site ( MIDAS ) and via the ligand Arg sidechain nitrogen forming a bifurcate H-bond with the Asp218 sidechain oxygen of the αV β-propeller domain ( Fig . 9A ) . The ligand Asp-MIDAS interaction remained intact throughout both simulations for L1 and L2 , as indicated by the ∼3 Å stable distance between the COM of the two ligand Asp sidechain oxygens and the MIDAS ( Fig . 9B ) . By comparison , the bifurcate ligand Arg-Asp218 H-bond was intact in the L2 SMD simulation ( <4 Å stable distance between the COM of the three ligand Arg sidechain nitrogens and the COM of the two Asp218 sidechain oxygens throughout the simulation ) but was disrupted in the L1 SMD 1 simulation . Nevertheless , the successful unbending of integrin αVβ3 by force applied via a bound ligand suggests that the ligand-integrin αVβ3 interaction is stronger than the interaction between the integrin headpiece and tailpiece . The force-extension curves of the SMD simulations of L1 and L2 displayed a major force peak similar to that observed in the SMD simulations of U1 and U2 ( Fig . 9C ) . Additionally , we observed two smaller peaks in the L1 SMD 1 and one smaller peak in the L2 SMD . Inspection of the changes of buried SASAs between the headpiece and tailpiece of integrin αVβ3 revealed that the second peak in the L1 SMD 1 was due to interactions between the hybrid and EGF3 domains while the third peak in the L1 SMD 1 and the second peak in the L2 SMD were both due to interactions at the α/β knees ( Fig . S5 ) . Finally , we performed one more constant-velocity SMD simulation on L1 ( L1 SMD 2 ) by pulling the same βA domain as in the SMD simulations of U1 and U2 ( instead of pulling on the RGD ligand as in the L1 SMD 1 and L2 SMD ) . This time only the first peak was seen on the force-extension curve ( Fig . 9C ) . These results suggest that the force direction plays a role in the determination of the unbending pathway after the first major force peak . To examine the stability of the liganded integrin after extension , we performed two free MD simulations for the final structures of the L1 SMD 1 and L2 SMD but did not observe the formation of the αV Asp457-Mg2+ coordination at the α-knee ( Ca2+ in the unliganded integrin was replaced by Mg2+ in the liganded integrin ) . This may be due to the different post-equilibration structures of the unliganded and liganded integrins . Despite that the βA domain was pulled the same way in the L1 SMD 2 as the previous simulations of the unliganded integrin , the αV Asp457 was far away from the α-knee Mg2+ in the resulting extended liganded integrin . By comparison , the αV Asp457 was very close to the α-knee Ca2+ in the unliganded integrin after extension even before further equilibration . We note that without the αV Asp457/α-knee Mg2+ coordination , the α-knee tended to rebend because the thigh/calf-1 hinge angle decreased significantly in one of the two free dynamics simulations of the liganded integrin ( L1 free MD ) ( Fig . S6 ) . With the αV Asp457/α-knee Ca2+ coordination , by comparison , the thigh/calf-1 hinge angle remained stable at certain levels in the two free dynamics simulations of the extended unliganded integrin ( U1 free MD 3 & 4 ) . The differential stabilities of the α-knee in the absence and presence of the αV Asp457/α-knee metal ion coordination supports the assertion that this metal ion coordination stabilizes , at least partly , the extended conformation .
X-ray crystallographic [4]–[10] , electron microscopic [11]–[13] , and antibody epitope mapping [27] studies have shown different static conformations of integrins corresponding to their different affinity states . The MD simulations reported here have connected the static conformations with dynamic processes , added time and force information , and provided structural insights to the continuous conformational changes of the integrin αVβ3 ectodomain . Our results suggest that tensile forces applied to either the βA and/or β-propeller domains ( Fig . 2 ) or the bound cyclic RGD ligand ( Fig . 9 ) can easily induce the change of integrin αVβ3 from a bent to an extended conformation . This work represents the first study in which the dynamic process of unbending of a complete integrin ectodomain was simulated in atomic details . Because no major distortions of individual domains resulted from pulling , our simulated structures of extended integrins should be more accurate than models obtained by simply rotating the bent crystal structures at the knees [8] or fitting various domains from the bent crystal structures into low resolution EM images of extended integrins [28] . Furthermore , that force can easily extend an integrin suggests an intriguing possible mechanism of force-induced integrin activation . Indeed , shear stress applied to lymphocytes promotes robust integrin-mediated adhesion [16] . In addition , either intracellular force by the cytoskeleton or extracellular force by shear strengthens adhesions mediated by integrin α5β1-fibronection bonds [17] . Moreover , recent single-bond experiments have demonstrated catch bond behavior ( such that force prolongs bond lifetime ) for bonds between integrin α5β1 and fibronectin [18] and between integrin αLβ2 and intercellular adhesion molecule 1 [19] . Although tensile forces readily unbent integrin αVβ3 , we did not observe hybrid domain swing-out , a key conformational change in models of integrin activation [10] , [11] , in the SMD simulations ( Figs . 2 and 9 ) or free dynamics simulations of the extended integrin with closed legs ( Fig . 8 ) . This result differs from previous SMD simulations on the αVβ3 headpiece [20] , [21] , where pulling the ligand binding site at the βA domain and the hybrid domain C-terminus induced the hybrid domain to swing out . Intuitively , when tensile forces apply to the V-shape αVβ3 headpiece , the hybrid domain would naturally swing out to open the headpiece . Pulling the whole ectodomain would be quite different . In the SMD simulations of an extended model of αIIbβ3 , a lateral force opened the closed α/β legs and induced the hybrid domain to swing out , but an axial force pulling along the straight integrin kept the hybrid domain closed [8] . In our SMD simulations with only axial forces , the legs remained closed and the hybrid domain did not swing out . It is not clear whether axial and lateral forces would apply to integrins simultaneously or sequentially under physiological situations . However , we observed extensive contacts between the two α/β legs that were stronger than those between the headpiece and tailpiece ( data not shown ) , suggesting that opening the closed legs may cost higher energy than unbending . Therefore , different forces may be required to first unbend an integrin and then open the legs and the hybrid domain . There may be mechanisms other than unbending and hybrid domain swing-out for force to increase binding affinity and/or bond lifetime of integrins with ligands . Indeed , lifetimes of α5β1-fibronectin [18] and αLβ2-intercellular adhesion molecule 1 [19] bonds have been observed to be prolonged by axial instead of lateral forces regardless of whether the integrins were unbent ( and/or the hybrid domain was swung out ) by force or the integrins had already been extended ( and/or the hybrid domain had already been swung out ) prior to the application of force . We identified two groups of critical interactions between the headpiece and tailpiece of the β3 subunit ( Fig . 4 ) that form the major energy barrier to the unbending of integrin αVβ3 . One group mainly consists of polar interactions between the hybrid domain and the βTD , including several H-bonds and one salt bridge between Asp393 and Arg633 . Indeed , Arg633 has been identified as a key residue to keep integrin αVβ3 in the bent conformation by NMA [24] . To verify the importance of Arg633 , the authors further showed experimentally that its deletion or substitution by Ala enhanced αIIbβ3-fibrinogen binding . The other group of critical interactions includes hydrophobic interactions between the hybrid and EGF4 domains . Not only did these interactions contribute to the major force peak resisting unbending , but they were also seen to reform during rebending of the partially-extended integrin ( Fig . 7 ) . These results predict that disrupting the hydrophobic interactions by mutating the residues involved , including Leu375 , Ile380 , Leu383 , Met387 , Met568 , Leu573 , and Leu574 , to polar residues will promote extension of β3 integrins . Besides the two groups of interactions identified from our simulations , there may be other interactions that play roles in the conformational changes of integrins . Indeed , interactions between a β hairpin protruding from the β-propeller domain and a region in the middle of the EGF3 and EGF4 domains have been shown as a clasp in restraining integrin activation [29] . This is consistent with our simulations showing that these interactions were maintained until the integrin was unbent . Our constant-force SMD simulations indicate that the time required to overcome the major barrier to unbending follows the Bell model [26] ( Fig . 6 ) . The value of Δx = 1 . 2±0 . 2 Å suggests a narrow energy well not very sensitive to force . The extrapolated zero-force waiting time of t0 = 728±539 ns suggests that the major barrier could be spontaneously overcome very rapidly on a time scale of 1 µs , allowing the integrin to reach a partially-extended conformation . It should be noted that these results are based on one simulation per force only , so the inferred zero-force waiting time may not be accurate . We also cannot exclude the possibility of a different unbending time-force relationship at lower forces . On the other hand , our free MD simulations suggested that the partially-extended integrin could bend back on a much shorter time scale of 10 ns ( Fig . 7 ) , so the rebending rate may be much faster than the unbending rate . Therefore , although unbending may occur spontaneously , the integrin may still far more likely assume the bent conformation at zero force . It may also undergo fast transitions between the bent and partially-extended conformation like “breathing” . Springer and coworkers propose that during integrin conformation “breathing” , epitopes on the β-leg may expose for antibody binding [11] , [14] . The binding of an antibody acts as a wedge to keep the interface between the headpiece and tailpiece open , which stabilizes the extended conformation and thus activates integrins . Indeed , EM studies observe that a small group of integrin αVβ3 under inactive conditions has a wider separation between the headpiece and tailpiece while the majority exhibits headpiece-tailpiece contacts [11] . Upon stimulations , the conformational equilibrium in a population of integrins may readily shift so that the extended conformation becomes more popular . Another interesting observation is the formation of a new coordination between the thigh domain Asp457 and the metal ion at the α-genu when the unliganded integrin became fully extended ( Fig . 8 ) , although this was not observed in the unbending simulations of the liganded integrin probably due to the different starting structure and insufficient simulation time . Because the new coordination was very stable once it was formed ( Fig . 8E ) and the α-knee tended to rebend when the new coordination was not formed ( Fig . S6 ) , the new coordination likely plays a role in stabilizing the extended conformation . In addition , both Asp457 and the α-genu metal ion site are conserved across species ( Fig . S7A ) , suggesting a functional role . While future mutagenesis experiments are required to provide definitive tests , the putative coordination between Asp457 and the α-genu metal ion upon integrin extension may explain why the majority of αVβ3 integrins adopt bent conformations in Ca2+ but extended conformations in Mn2+ [11] . The explanation assumes different propensities for distinct metal ions to coordinate with Asp457 such that distinct metal ion conditions favor different fractions of bent vs . extended integrins . It should be noted that among all 18 known human α subunits , residue 457 has wide variations , although the α-genu metal ion site is largely conserved ( Fig . S7B ) . Only in several α subunits , including α5 , α7 , and α8 , is Asp457 conservatively substituted by Glu . Therefore , the putative Asp457/α-genu metal ion coordination is likely a specific property for a subset of integrins . It remains debatable whether the deadbolt or switchblade model describes the mechanism of integrin activation [2] . Our simulations may seem to favor the switchblade model because force induces a switchblade-like extension . However , initial ligand binding before force loading could be regulated by a “deadbolt” . Future experimental and computational studies are needed to clarify the mechanism of integrin activation .
Four systems were set up for MD simulations: U1 , U2 , L1 , and L2 ( Table S1 ) . U1 was modeled after the early crystal structure of the unliganded integrin αVβ3 ectodomain ( PDB code 1U8C ) [6] available at the time when we started this study , in which the EGF1 and EGF2 domains at the genu of the β3 subunit were unresolved . To complete the ectodomain , we used MODELLER [30] to build a homology model for the two missing domains based on the crystal structure of a fragment of the β2 subunit ( PDB code 2P28 ) [31] and used TMD to fit it to the crystal structure ( Text S1 and Fig . S8 ) . The modeled EGF1 and EGF2 domains plus the PSI domain from the 1U8C structure were then added to the liganded integrin αVβ3 ectodomain ( PDB code 1L5G ) [5] to obtain L1 . U2 is the recently released crystal structure of the unliganded integrin αVβ3 ectodomain plus a short α/β transmembrane fragment , where the EGF1 and EGF2 domains are visible ( PDB code 3IJE ) [7] . Wherever resolvable , the 1U8C structure is very close to the 3IJE structure . Our homology model of the EGF1 and EGF2 domains are similar to those of the 3IJE structure in orientation and majority of the tertiary structures , with noticeable variations only in the loops and the linker between the two domains ( Fig . S4 ) . L2 was obtained by extracting the EGF1 , EGF2 , and PSI domains from the 3IJE structure and adding them to the 1L5G structure . Therefore , U1 and U2 are highly comparable except for the EGF1 and EGF2 domains; so are L1 and L2 . For L1 and L2 , we replaced nonstandard N-Methylvaline and D-form Phenylalanine in the cyclic-RGD ligand to standard Valine and L-form Phenylalanine , respectively . We also replaced Mn2+ ions with Mg2+ because force field parameters for Mn2+ were not available . We employed LEaP in AMBER8 package [32] to prepare solvated structures with TIP3P [33] water boxes ( Fig . 1B ) . The closest distance between the walls of the water boxes and the proteins was 15 Å while the closest distance between water molecules and proteins was set to 1 Å . Four water boxes with different dimensions were set up for the four structures ( Table S1 ) . Na+ and Cl− ions were added to neutralize the systems at a 150 mM salt concentration . After equilibration , the water boxes were further enlarged for SMD simulations ( Fig . 2A ) and more Na+ and Cl− ions were added accordingly to maintain the 150 mM salt concentration . MD simulations were performed using NAMD [34] . Duan et al . [35] and GLYCAM04/06 [36] , [37] force fields were used for amino acids and carbohydrates attached to the proteins , respectively . We used 2 fs time step , 12 Å cutoff for non-bonded force , and 10 Å switching distance for smoothing functions of non-bonded force . Bonds involving hydrogen were set rigid by using the SHAKE algorithm [38] for protein and the SETTLE algorithm [39] for water . Periodic boundary conditions were applied and the Particle Mesh Ewald method [40] was used to calculate full electrostatic interactions every 4 fs . Atomic coordinates of the systems were saved every 1 ps . First , each system was energy-minimized for three consecutive 10 , 000 conjugate-gradient steps: first with all protein atoms fixed , second with only the backbone atoms fixed , and third with all atoms free . Then , the systems were gradually heated from 0 to 300 K in 120 ps with constraints of 1 kcal mol−1 Å−2 spring constants applied on all protein atoms under constant volume . Next , pressure was adjusted to 1 atm in 100 ps with the Langevin piston method [41] under constant temperature controlled with Langevin dynamics . After that , the constraints on the protein atoms were gradually released in 100 ps under constant volume and constant temperature ( NVT ) . Finally , 40 ( or 50 ) ns equilibrations were performed for U1 and L1 ( or U2 and L2 ) under constant pressure and constant temperature ( NPT ) . The pressure was maintained at 1 atm by the Langevin piston method while the temperature was maintained at 300 K by Langevin dynamics with a damping coefficient of 5 ps−1 . After equilibration in the small water boxes ( cf . Fig . 1B ) , more water molecules were added to enlarge the water boxes ( cf . Fig . 2A ) . For equilibration of the enlarged water boxes , the systems were first energy-minimized for 10 , 000 conjugate-gradient steps with all protein atoms fixed , next heated from 0 to 300 K in 120 ps with spring constraints of 1 kcal mol−1 Å−2 spring constant on all protein atoms under constant volume followed by 2–4 ns equilibration in NPT ensembles , then set free by gradually releasing constraints on the protein atoms in 100–400 ps followed by 2–4 ns final equilibration in NPT ensembles . The enlarged systems were then ready for SMD simulations . In all production SMD and free MD simulations , the controls on pressure and temperature were turned off so as to reduce disturbance on the dynamics . In SMD simulations , a group of Cα atoms were selected . Then force or constraint was exerted on the COM of the selected atoms . When different domains were pulled or constrained , the residues were selected as follows: βA residues 113–117 , 151–156 , 190–197 , 244–250 , 306–310 , and 329–332; β-propeller residues 22–26 , 97–101 , 160–164 , 225–229 , 279–283 , 343–347 , and 407–411; βTD residues 610–620 , 639–642 , 656–658 , 665–670 , and 679–682; and all five residues of the cyclic RGD ligand . While force was on the pulling COM , a harmonic potential with a spring constant of 10 kcal mol−1 Å−2 was added to the constraint COM . In the constant-velocity SMD simulations [25] , the pulling COM was harmonically constrained with a force , F = k ( vt−x ) , where k is the spring constant , v is the pulling velocity , t is time , and x is the coordinate along the force direction . A total of ∼468 ns production simulations were performed , including 8 constant-velocity and 5 constant-force SMD simulations for U1 , 1 constant-velocity and 1 constant-force SMD for U2 , 2 constant-velocity SMD for L1 , 1 constant-velocity SMD for L2 , 4 free MD for U1 , 1 free MD for L1 , and 1 free MD for L2 ( Table S1 ) . VMD [42] was employed to analyze simulations , render molecular graphics , and generate trajectory videos . SASA was calculated with 1 . 4 Å probe radius . Buried SASA between two contacting portions was calculated as the difference between the SASAs of one portion without and with the other portion . Restricted to one domain during SASA calculation , the buried SASA of that domain was obtained . An H-bond was defined by a <3 . 5 Å donor-acceptor distance and a >120° donor-hydrogen-acceptor angle . Hinge angles were measured using Hingefind [43] . The various αVβ3 domains are defined as follows: β-propeller , residues 1–438; thigh , residues 439–600; calf-1 , residues 601–738; calf-2 , residues 739–956; PSI , residues 1–57; hybrid , residues 58–110 and 354–434; βA , residues 111–353; EGF1 , residues 435–472; EGF2 , residues 473–522; EGF3 , residues 523–559; EGF4 , residues 560–599; β-ankle , residues 600–605; βTD , residues 606–690 . | Proteins can regulate their functions via conformational changes . One example is integrins , which are transmembrane receptors mediating cell-cell and cell-matrix adhesions . Inactive integrins may assume a bent conformation with low affinities for ligands unable to support adhesions . Intracellular or extracellular stimuli induce large scale changes from the bent to an extended conformation , resulting in active integrins with high affinities for ligands to mediate strong adhesions . We used molecular dynamics simulations to reveal the dynamics and pathways of integrin unbending in atomic details . Critical interactions in this process were identified . This study not only sheds light on the structural mechanisms of integrin activation , but also exemplifies allosteric regulations of protein functions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/molecular",
"dynamics",
"cell",
"biology/cell",
"adhesion",
"biophysics/theory",
"and",
"simulation"
] | 2011 | Molecular Dynamics Simulations of Forced Unbending of Integrin αVβ3 |
Piriformospora indica is an endophytic fungus that colonizes roots of many plant species and promotes growth and resistance to certain plant pathogens . Despite its potential use in agriculture , little is known on the molecular basis of this beneficial plant-fungal interaction . In a genetic screen for plants , which do not show a P . indica- induced growth response , we isolated an Arabidopsis mutant in the OXI1 ( Oxidative Signal Inducible1 ) gene . OXI1 has been characterized as a protein kinase which plays a role in pathogen response and is regulated by H2O2 and PDK1 ( 3-PHOSPHOINOSITIDE-DEPENDENT PROTEIN KINASE1 ) . A genetic analysis showed that double mutants of the two closely related PDK1 . 1 and PDK1 . 2 genes are defective in the growth response to P . indica . While OXI1 and PDK1 gene expression is upregulated in P . indica-colonized roots , defense genes are downregulated , indicating that the fungus suppresses plant defense reactions . PDK1 is activated by phosphatidic acid ( PA ) and P . indica triggers PA synthesis in Arabidopsis plants . Under beneficial co-cultivation conditions , H2O2 formation is even reduced by the fungus . Importantly , phospholipase D ( PLD ) α1 or PLDδ mutants , which are impaired in PA synthesis do not show growth promotion in response to fungal infection . These data establish that the P . indica-stimulated growth response is mediated by a pathway consisting of the PLD-PDK1-OXI1 cascade .
The endophytic fungus Piriformospora indica , a cultivable basidiomycete of Sebacinales , colonizes the roots of many plant species including Arabidopsis [1] , [2] . Like other members of Sebacinales , P . indica is found worldwide in association with roots [3] , and stimulates growth , biomass and seed production of the hosts [1] , [2] , [4]–[11] . The fungus promotes nitrate and phosphate uptake and metabolism [6] , [12] , [13] . P . indica also confers resistance against abiotic [7] , [14] , [15] and biotic stress [2] , [16] . The broad host range of P . indica indicates that the beneficial interaction may be based on general recognition and signalling pathways . Little is yet understood about the molecular steps leading to P . indica-induced growth promotion . Plant growth can be induced by a fungal exudate component [9] , suggesting the involvement of specific receptors at the plant cell surface . In support of this hypothesis , an atypical receptor kinase with leucine-rich repeats was identified to be required for the growth response in Arabidopsis [5] . Moreover , a rapid increase in the intracellular calcium concentration in the root cells indicates that an intracellular signalling cascade is triggered early upon plant-fungal interaction [9] . So far , however , no further components of the signalling pathway have been identified . In mammals , the phospholipid-binding 3-PHOSPHOINOSITIDE-DEPENDENT PROTEIN KINASE1 ( PDK1 ) sustains and regulates the balance between growth , cell division and apoptosis [17]–[19] . PDK1 is a member of the cAMP-dependent protein kinase A / protein kinase G / protein kinase C ( AGC ) kinase family [17] and the Arabidopsis homolog AtPDK1 is regulated by binding to the lipid phosphatidic acid ( PA ) [20] , [21] . Phospholipase D ( PLD ) α1 is the main producer of PA in Arabidopsis roots [22] . In plants , PA is a second messenger [23] , [24] that links lipid signalling to oxidative stress signalling [25] , e . g . during abscisic acid-induced stomatal closure or defense against pathogens [26]–[28] . PDK1 is the only AGC kinase in plants with an identifiable lipid-binding domain [20] , [21] , [29] , [30] . OXIDATIVE SIGNAL INDUCIBLE1 ( OXI1 ) is a serine/threonine kinase necessary for oxidative burst-mediated signalling in Arabidopsis roots [20] , [31] . OXI1 is a member of the AGC protein kinase family ( also called AGC2-1 [30] ) and its expression is induced by H2O2 [31] . OXI1 is required for full activation of the two mitogen-activating protein kinases 3 and 6 ( MPK3 and MPK6 ) after treatment with reactive oxygen species ( ROS ) or elicitors and for different ROS-mediated processes including basal resistance to Hyaloperonospora arabidopsidis ( previously known as Peronospora parasitica ) infection and root hair growth [31] . Among all AGC kinases in Arabidopsis [30] , AGC2-2 might be considered as an OXI1 homolog , however this kinase has not yet been investigated . The active OXI1 phosphorylates and thus activates the downstream serine/threonine kinase PTI1-2 in response to ROS and phospholipid signals [21] , and many of these signals derive from microbial pathogens or elicitors , such as cell wall fragments or specific protein factors released by pathogens [32] , [33] . Besides ROS , OXI1 is also activated by PDK1 [20] . In this work , we report on the results of a genetic screen for Arabidopsis mutants , which do not respond to P . indica . By positional cloning , we have identified OXI1 as the responsible gene for the growth phenotype induced by P . indica . Since OXI1 is an AGC protein kinase that can be activated by H2O2 and PDK1 , we also tested whether mutants in PDK1 . 1 and PDK1 . 2 are defective in the P . indica-induced growth phenotype . We found that pdk1 . 1 pdk1 . 2 double knock out mutants do not respond to P . indica . The fungus stimulates PA , but not H2O2 synthesis in Arabidopsis plants . PA is produced by several pathways including by PLD . When PA synthesis was reduced by inactivation of phospholipase D ( PLD ) α1 or PLDδ , the P . indica-induced growth promotion was compromised . These results suggest that P . indica stimulates growth by PA-mediated activation of PDK1 , which subsequently activates OXI1 .
Arabidopsis plants co-cultivated with P . indica are taller than the uncolonized controls [1] , [2] . On the basis of this growth phenotype , we searched for ethylmethane sulfonate-induced mutants , which grow like uncolonized plants or are smaller in the presence of the fungus . One of these mutants , called Piriformospora indica-insensitive12 ( pii12 ) , was smaller in the presence of the fungus ( Figure 1 ) and mapped to a region on chromosome 3 that included oxi1 . Moreover , the pii12 mutant had reduced root hair lengths and reduced oxi1 mRNA levels in roots and shoots when compared to the wild-type ( Figure S1 in Text S1 ) . Sequence analysis uncovered that the mutant lacks a 19 bp segment upstream of the putative translation start site , while the coding region was intact . To clarify whether OXI1 is responsible for the absence of the P . indica-induced growth response in Arabidopsis , pii12 was complemented with the full-length cDNA of OXI1 . Three independent transformants had higher OXI1 mRNA levels when compared to pii12 and showed a growth response to the fungus , which was comparable to the wild type ( Figure S2 in Text S1 ) . An independent T-DNA insertion line for oxi1 was used for further analysis , because it completely lacked OXI1 mRNA ( Figure S3A in Text S1 ) . Like pii12 , growth promotion by P . indica was inhibited in oxi1 plants ( Figure 1 and Figure S2 in Text S1 ) . These results confirm that a deletion in the OXI1 promoter region is responsible for the absence of the growth response of Arabidopsis plants to P . indica . We conclude that P . indica-induced growth promotion in Arabidopsis requires OXI1 . Previously , it was shown that OXI1 is induced by H2O2 in the roots [31] . However , H2O2 measurements and staining of colonized wild type roots with nitrobluetetrazolium chloride ( NBT ) uncovered that P . indica does not induce H2O2 accumulation [9] . Under growth promoting conditions , we even observed a repression of H2O2 accumulation in the roots ( Figure S4C in Text S1 ) . Also high concentrations of fungal hyphae , which are no longer beneficial for the plants , did not result in H2O2 production in the roots ( H2O2 levels , no fungal treatment: 17 . 4±2 . 1 nmol/g fresh weight; non-beneficial interaction: 17 . 1±1 . 7 nmol/g fresh weight; n = 9 independent experiments ) . The inability of oxi1 plants to respond to P . indica might be caused by their shorter root hairs [31] . However , mRNA levels for the P . indica translation elongation factor1 ( Pitef1 ) were comparable in oxi1 and wild-type roots ( Figure S5 in Text S1 ) , indicating that root colonization does not differ from the wild-type in oxi1 . We also investigated the interaction of P . indica with two other mutants with reduced root hair phenotypes: the AGC kinase ire and the NADPH oxidase rhd2 ( [34] , [35] Figure S3B in Text S1 ) . Growth of these mutants was promoted by P . indica ( Figure 1 ) , and the degree of root colonization was again comparable to the wild-type ( Figure S5 in Text S1 ) . Therefore , the root hair phenotype does not seem to be responsible for the impaired interaction of oxi1 with P . indica . Furthermore , among the RHD genes expressed in Arabidopsis roots , RHD2 shows the highest expression level and RHD2 is responsible for most of the H2O2 production in the roots [35] . Thus , the lower H2O2 production in rhd2 roots does not compromise the beneficial plant-fungal interaction . AGC2-2 ( At4g13000 ) is the closest homolog of OXI1 ( see phylogenetic tree in [30] ) and shares >60% sequence identity to OXI1 . Both kinases contain an aspartic acid residue in their active site ( D149 in OXI1 and D146 in AGC2-2 ) and share a conserved PDK1 binding site , the FxxF motif , at their C-terminal ends [20] . However , in contrast to the OXI1 mRNA level , the AGC2-2 mRNA level is not regulated by ROS ( https://www . genevestigator . com ) . agc2-2 plants did not show any visible phenotype , produced the same amount of seeds , and – in contrast to oxi1 [31] - root hairs of agc2-2 plants were not shorter than those of wild-type plants ( Figure S1B in Text S1 ) . However , despite the fact that root colonization was not affected by the agc2-2 mutation ( Figure S5 in Text S1 ) , agc2-2 plants were compromised in the growth response to the fungus ( Figure 1 ) . Thus , besides OXI1 , the so far uncharacterized AGC2-2 is important for P . indica-mediated growth promotion in Arabidopsis . Attempts to generate homozygous oxi1 agc2-2 double knock out lines failed: among 98 F2 plants obtained from crosses of the two mutants , all plants , which were homozygote for either oxi1 or agc2-2 were heterozygote for the other kinase gene . This suggests that both OXI1 and AGC2-2 might play a role in embryogenesis in Arabidopsis . We next tried to identify the upstream components of the OXI1 cascade that is responsible for the fungal growth effect in plants . Previously , it was shown that PDK1 and H2O2 can activate OXI1 in Arabidopsis [20] , [21] . Because P . indica infection did not alter H2O2 levels in Arabidopsis , we turned our attention to the two closely related PDK1 genes , PDK1 . 1 and PDK1 . 2 ( 92% homology at the amino acid level ) , which are present in the Arabidopsis genome ( cf . phylogenetic tree of AGC kinases in [30] ) . Both PDK1 genes are expressed in roots . We generated a pdk1 . 1 pdk1 . 2 double knock out line . RT-PCR analysis confirmed that neither PDK1 . 1 nor PDK1 . 2 transcripts can be detected in the double mutant line ( Figure 2A ) . A phenotypic analysis revealed that pdk1 . 1 pdk1 . 2 plants are smaller than the wild-type ( Figure 2C ) , have shorter siliques ( Figure 2B ) and produce only 41%±6 . 8% ( n = 23 ) of the seeds of the wild-type . Importantly , fungal induced growth promotion in pdk1 . 1 pdk1 . 2 plants was clearly compromised ( Figure 1 ) , whereas root colonization was comparable to the wild-type ( Figure S5 in Text S1 ) . Therefore , besides general functions in growth regulation , the combination of PDK1 . 1 and PDK1 . 2 is required for P . indica-induced growth promotion in Arabidopsis . After having established that PDK1 is an important component of the P . indica-induced growth response pathway , we tried to go even further up in the cascade to identify the regulator of the PDK1s . PDK1 in Arabidopsis is activated by PA . PA is synthesized by PLD and by PLC/diacylglycerol kinase . PA in roots is mainly generated by PLD activity [28] , [36] . The Arabidopsis genome contains 12 genes for PLDs , which are classified into six types , PLDα ( 1–3 ) , β ( 1 and 2 ) , γ ( 1–3 ) , δ , ε and ζ ( 1 and 2 ) [37] . The most abundantly expressed pld genes in roots are pldα1 and pldδ [22] , [38] . PLDα1 is responsible for most of the PA production in roots , and the PA content is severely reduced in the roots of pldα1 knock out mutants [22] . Furthermore , wounding-induced PA production is completely eliminated in the pldα1 pldδ double knock out line [39] . Application of a P . indica exudate fraction , which promotes plant growth [9] stimulates PA accumulation in a time- and dose-dependent manner in the roots ( Figure 3 ) . Furthermore , the growth response of pldα1 and pldδ insertion lines to P . indica was severely impaired ( Figure 4 ) . In comparison , the response of pldα3 und pldε ( Figure 4 , Figure S3C in Text S1 ) plants to P . indica was similar to wild type . These results indicate that signals from the fungus activate PA synthesis via PLDα1 and PLDδ in the roots . Compared to uncolonized roots , the two PDK1 mRNA levels were ∼2-fold higher and the OXI1 and AGC2-2 mRNA levels increased ∼4-fold in P . indica-colonized roots ( Figure 5A ) . In contrast , three classical defense genes , which are targets of PDK1 and OXI1 signalling after pathogen infections ( PR3 , PDF1 . 2 , ERF1 [20] , [21] , [31] ) , are downregulated in P . indica-colonized wild-type roots ( Figure 5B ) . Thus , upregulation of the PDK1 and OXI1 mRNA by P . indica does not result in the activation of the three defense genes . The expression level of defense genes is also downregulated in the colonized pdk1 . 1 pdk1 . 2 , oxi1 and agc2-2 plants . PR2 is mildly upregulated by the fungus , but this occurs also in the colonized mutants ( Figure 5B ) . Thus , the regulation of the defense genes occurs independently of the OXI1 pathway under beneficial co-cultivation conditions of the two symbionts . To test whether the PDK1 , OXI1 and AGC2-2 kinases activate defense processes under non-beneficial conditions , we inoculated Arabidopsis plants with high doses of P . indica . Seven days after transfer to a dense fungal lawn , the seedlings still continued to grow ( Figure S4A in Text S1 ) , but visible accumulation of anthocyanin in the aerial parts were indicative of a stress response in the plants . No H2O2 accumulation would be detected under these co-cultivation conditions ( Figure S4B in Text S1 ) , however the PDK1 , OXI1 and AGC2-2 mRNA levels were moderately upregulated ( Figure 6A ) . In contrast to beneficial co-cultivation conditions , also defense genes , and in particular PDF1 . 2 , were upregulated . However , this response was similar in wild type , oxi1 , agc2-2 and pdk1 . 1 pdk1 . 2 mutants ( Figure 6B ) . Therefore , upregulation of defense genes under non-physiological co-cultivation conditions is not mediated by the OXI1 pathway as well ( Figure 6B ) .
The pii12 and oxi1 mutants are impaired in P . indica-induced growth promotion ( Figure 1 , Figure S2 in Text S1 ) . The OXI1 kinase was shown to be induced by H2O2 and to activate defense responses against pathogen infections [20] , [21] , [31] , [42] . However , H2O2 production is repressed in P . indica-colonized roots under beneficial co-cultivation conditions ( Figure S4C in Text S1 ) and some defense genes are downregulated under beneficial conditions ( Figure 5B ) . Exposure of Arabidopsis seedlings to high doses of the fungal hyphae induces a mild defense response , which occurs also in oxi1 mutants ( Figure 6B ) . Thus , OXI1 is required for the growth response but is not involved in defense gene activation in this beneficial interaction ( cf . below ) . Interestingly , the OXI1 overexpressor lines behaved like the wild-type ( Figure S2 in Text S1 ) suggesting that wild-type amounts of the kinase are sufficient for the beneficial interaction . Furthermore , AGC2-2 , a so far uncharacterized homolog of OXI1 , is also required for the beneficial interaction . AGC2-2 is not induced by H2O2 , but by P . indica in wild-type roots ( Figure 5A ) . Since attempts to isolate a homozygote oxi1 agc2-2 double mutant failed and since the two single knock out lines fail to respond to P . indica , the two kinases have important and presumably different functions . Interestingly , this highly related pair of protein kinases resembles the OXI1-activated MAPKs MPK3 and MPK6 , for which MPK3 is inducible by pathogens , while MPK6 is constitutively expressed and mpk3 mpk6 double mutants are embryo-lethal [9] , [43] . In mammalian systems , AGC kinases play important roles in growth and proliferation . The activation mechanism of AGC kinases from both kingdoms by lipids and their conserved epitopes [17] support the idea that OXI1 and AGC2-2 play a crucial role in regulating cell growth , division and/or elongation in response to the signals from P . indica . Because oxi1 mutants are also compromised in root hair growth , we tested two mutants with shorter root hairs , ire and rhd2 . However , none of these mutants were impaired in the growth response to the fungus ( Figure 1 ) . Moreover , because rhd2 is also impaired in full production of H2O2 in roots , the inability of oxi1 to respond to P . indica is not caused by the reduced root hair phenotype or lower H2O2 levels in the roots . PA is an important second messenger and is involved in regulating plant growth , proliferation , biomass production , cell expansion , as well as responses to biotic and abiotic stresses [23] , [24] , [26]-[28] , [36] , [44]–[48] . In response to stresses , PA balances and fine-tunes the appropriate plant response to environmental signals [28] , [36] . PA accumulation is induced by exudate preparations from P . indica in a dose- and time-dependent manner ( Figure 3 ) , suggesting that the roots sense signalling molecules released from the fungus . The requirement of the PA-activated PDK1s for the beneficial interaction suggests a participation in growth regulation , similar to mammalians [49]–[51] . Nitrate and phosphate uptake and metabolism is stimulated by P . indica and required for growth promotion [6] , [12] , [13] . PA also plays important roles in nitrogen [48] , [52]–[54] and phosphate signalling [55] , [56] . These results might provide a link between the P . indica-induced positive growth phenotype and the primary metabolism . Further experiments are necessary to investigate a role of PDK1 , OXI1 and AGC2-2 in this respect . Interestingly , in mammals and yeast , PDK1 is a central regulatory kinase , which phosphorylates and thus activates AGC kinases in response to rises in the levels of the second messenger phosphatidylinositol 3 , 4 , 5-trisphosphate [19] , [57] . pdk1 knock-out mice are embryo-lethal [58] . Since the Arabidopsis pdk1 . 1 pdk1 . 2 double knock-out line is viable , activation of AGC kinases might be different in plant and mammalian systems [19] , [57] , [58] . PA is synthesized by PLD or phospholipase C/diacylgycerol kinase ( PLC/DAG ) [36] . PLDα1 and PLDδ are abundantly expressed in roots . We observed that their inactivation severely reduces P . indica-induced growth promotion ( Figure 4 ) . pldα1 was shown previously to contain lower PA levels in the roots [22] , has reduced wounding-induced PA production , and this response is completely eliminated in the pldα1 pldδ double knock out line [39] . PLDα1 and PA have also been implicated in regulating NADPH oxidase activity and the production of H2O2 in ABA-mediated stomatal closure [25] . The plasma-membrane-bound PLDδ is activated in response to H2O2 [59] . However , since H2O2 is not accumulating in response to P . indica , the lipases might have a different function and are differently regulated in this beneficial interaction . PLDα1 and PLDδ expression is not induced by P . indica . PLDα1 activity is regulated by dynamic changes in intracellular Ca2+ levels ( cf . [28] ) , and the Ca2+ levels in the root cytoplasm increases even faster in response to the same exudate fraction from P . indica that induces PA accumulation ( Figure 3; [9] ) . These results suggest that signals from P . indica are decoded via the two intracellular second messengers PA and Ca2+ . It remains to be determined how PA and Ca2+ cooperate to induce the appropriate plant responses , and which mechanisms determine whether they activate responses leading to a beneficial interaction or defense activation . In conclusion , we demonstrate that in the beneficial interaction between P . indica and Arabidopsis the OXI1 pathway constitutes a protein kinase signalling pathway that confers growth stimulation ( Figure 7 ) . We propose a model whereby roots sense signals derived from P . indica by activating a signalling pathway that results in PA-mediated activation of PDK1 , which subsequently activates the OXI1 and AGC2-2 protein kinases . Since MPK6 is a downstream target of OXI1 [31] and required for P . indica-mediated growth promotion [9] , it is possible that MPK6 might be an additional component of this pathway . Future studies on the targets of the OXI1 pathway should help to clarify by which mechanism growth promotion occurs in plants and how this knowledge could be used to improve yield and productivity in agriculture . It also remains to be determined whether promotion of plant growth by mycorrhizal fungi or plant-growth promoting bacteria requires the same pathway , and how the Arabidopsis mutants analysed in this study respond to pathogens .
Wild-type Arabidopsis thaliana seeds and seeds from the homozygote T-DNA insertion lines were surface-sterilized and placed on Petri dishes containing MS nutrient medium [60] . After cold treatment at 4°C for 48 h , plates were incubated for 7 days at 22°C under continuous illumination ( 100 µmol m−2 sec−1 ) . P . indica was cultured as described previously [1] , [4] on Kaefer medium . Nine day-old A . thaliana seedlings were transferred to nylon disks ( mesh size 70 µm ) and placed on top of a modified PNM culture medium ( 5 mM KNO3 , 2 mM MgSO4 , 2 mM Ca ( NO3 ) 2 , 0 . 01 µM FeSO4 , 70 µM H3BO3 , 14 µM MnCl2 , 0 . 5 µM CuSO4 , 1 µM ZnSO4 , 0 , 2 µM Na2MoO4 , 0 . 01 µM CoCl2 , 10 . 5 g L−1 agar , pH 5 . 6 ) , in 90 mm Petri dishes . Fungal plugs of 5 mm in diameter were placed at a distance of 1 cm from the roots . Control seedlings remain untreated . Plates were incubated at 22°C under continuous illumination from the side ( 80 µmol m−2 sec−1 ) . The following homozygote T-DNA insertion lines were used: rhd2 ( At5g51060; [35] obtained from Prof . V . Zársky , Prague , Czech Republic ) , ire ( At5g62310 ) Salk_043276 , oxi1 ( At3g25250 ) Gabi_355H08 , agc2-2 ( At4g13000 ) Salk_083220 , pdk1 . 1a ( At5g04510 ) Salk_113251 , pdk1 . 1b ( At5g04510 ) Salk_007800 , pdk1 . 2 ( At3g10540 ) Sail_450_B01 , pldα1-1 ( At3g15730 , [61] ) Salk_067533 , pldδ ( At4g35790 , [61] ) Salk_023247 , pldα3 ( At5g25370 ) Salk_122059 , pldε ( At1g55180 ) Koncz68434 . pdk1 . 1 pdk1 . 2 was generated by crosses between pdk1 . 1 and pdk1 . 2 . 6 week-old adult plants were used for interaction studies with P . indica . Arabidopsis seedlings , grown for 14 days on MS media , were transferred to vermiculite ( rather than soil ) , because this allowed to harvest the intact roots including the lateral roots . The growth response of the plants to P . indica on soil and on vermiculite is comparable ( data not shown ) . The vermiculite was mixed with the fungus ( 1% , w/v ) which was dissolved in PNM medium . 70 ml of liquid PNM medium or inoculated PNM medium was used per plant . The fungal mycelium was obtained from two weeks old liquid cultures after the medium was removed and the mycelium was washed with an excess of distilled water . Cultivation occurred in pots in a temperature-controlled growth chamber at 22°C under short-day conditions ( light intensity: 80 µmol m−2 sec−1 ) . The sizes of the plants were monitored weekly and after six weeks the fresh weights of the shoots were determined and the roots harvested for RNA or DNA extraction . 12-day-old seedlings were directly transferred from MS medium to a plate with a fungal lawn . The fungal lawn was obtained by placing a fungal plug on Kaefer medium and the fungus was allowed to grow for 14 days at 24°C in the dark , before the seedlings were transferred to the plate . Control seedlings were transferred to Kaefer medium without the fungus . The plates were incubated for 7 days at 22°C under continuous illumination ( 80 µmol m−2 sec−1 ) from above . Fresh weights were determined ( data not shown ) and RNA was extracted of the root material . RNA was isolated from the roots with an RNA isolation kit ( RNeasy , Qiagen , Hilden , Germany ) . For quantitative RT-PCR , RNA from Arabidopsis roots grown in the absence or presence of P . indica was used . Reverse transcription of 1 µg of total RNA was performed with oligodT Primer . First strand synthesis was performed with a kit from Qiagen ( Omniscript , Qiagen , Hilden , Germany ) . RT-PCR was conducted with the primer pairs given in Figure S6 in Text S1 . P . indica was monitored with a primer pair for the translation elongation factor 1 ( Pitef1; [62] ) . The colonized ( and control ) roots were removed from vermiculite , rinsed 6 times with an excess of sterile water and were frozen in liquid nitrogen for RNA or DNA extraction . One of the two plant genes ( GAPC2 and UBQ5 ) was used as housekeeping genes for Arabiopsis roots . Semiquantitative analysis was performed after 27 PCR cycles: the products were analysed on 2% agarose gels , stained with ethidium bromide , and visualized bands were quantified with the ImageQuant 5 . 0 ( GE Healthcare Life Sciences ) . Real-time quantitative RT-PCR was performed using the iCycler iQ real-time PCR detection system and iCycler software version 2 . 2 ( Bio-Rad , Munich , Germany ) . For the amplification of the PCR products , iQ SYBR Supermix ( Bio-Rad ) was used according to the manufactureŕs instructions in a final volume of 23 µl . The iCycler was programmed to 95°C 2 min , 35× ( 95°C 30 s , 55°C 40 s , 72°C 45 s ) , 72°C 10 min followed by a melting curve programme ( 55–95°C in increasing steps of 0 . 5°C ) . All reactions were repeated twice . The mRNA levels for each cDNA probe were normalized with respect to the GAPC2 and UBQ5 message levels . Fold induction values were calculated with the ΔΔCP equation of Pfaffl ( 2001 ) [63] . The ratio of a target gene was calculated in the treated sample versus the untreated control in comparison to a reference gene . The primer pairs are given in Figure S6 in Text S1 . H2O2 was determined by an assay coupled to the peroxidase [64] . Roots ( 0 , 1 g ) were homogenized in 1 mL 1 M HClO4/insoluble PVP ( 5% ) . The supernatant was clarified by centrifugation , adjusted to pH 5 . 6 with 5 M K2CO3 solution and incubated with 1U ascorbate oxidase for 10 min to oxidize the ascorbate . The reaction in 0 . 1 M phosphate buffer ( pH 6 . 5 ) , 3 . 3 mM 3- ( dimethylamino ) benzoic acid , 0 . 07 mM 3-methyl-2-benzothiazoline hydrazone and 0 . 3 U peroxidase was started by adding the oxidized extracts and followed by absorbance change at 590 nm and 25°C . NBT staining has been described previously [9] . Arabidopsis seedlings ( 5-days-old ) were labeled overnight in 400 µL buffer ( 2 . 5 mM MES-KOH , 1 mM KCl , pH 5 . 7 ) containing 10 µCi of carrier-free PO43− . Samples ( 3 seedlings each ) were treated by adding 100 µL water with or without elicitor for the times and concentrations indicated . Treatments were stopped by adding 50 µL 50% perchloric acid ( w/v ) and shaking the samples vigorously for 5 min . Liquid was then removed and replaced by 375 µL of CHCl3/MeOH/HCl [50∶100∶1 ( v/v ) ] followed by 100 µL 0 . 9 % NaCl ( w/v ) , to extract the lipids while shaking ( 10 min ) . A two-phase system was induced by the addition of 375 µL of CHCl3 and 200 µL of 0 . 9% ( w/v ) NaCl . The remainder of the extraction was performed as described before [32] . For quantitative analysis , lipids were separated by thin-layer chromatography ( TLC ) using heat-activated , potassium oxalate/EDTA-impregnated , silica TLC plates ( Merck , 20×20×0 . 1 cm ) and an alkaline solvent system of CHCl3/MeOH/25%NH4OH/H2O [90∶70∶4∶16 ( v/v ) ] , essentially as described in [65] . Phospholipids were visualized and quantified by phosphoimaging ( Molecular Dynamics , Sunnyvale , CA , USA ) . All data were analysed with one-side , unpaired students t-Test ( p≤0 . 05 ) in Excel . OXI1 ( other names: AGC2; AGC2-1; OXIDATIVE SIGNAL-INDUCIBLE1; ATOXI1; MJL12 . 22 ) , At3g25250 , NP_189162 . 1; AGC2-2 ( other names: F25G13 . 90; F25G13_90 ) , At4g13000 , NP_193036 . 1; PDK1 . 1 ( other names: 3'-PHOSPHOINOSITIDE-DEPENDENT PROTEIN KINASE 1; ATPDK1; PDK1; T32M21 . 110 ) , At5g04510 , NP_568138 . 1; PDK1 . 2 ( other names: PDK2; F13M14 . 18 ) , At3g10540 , NP_187665 . 2; RHD2 ( other names: A . THALIANA RESPIRATORY BURST OXIDASE HOMOLOG C; ATRBOHC; K3K7 . 25; RBOHC; ROOT HAIR DEFECTIVE 2 ) , At5g51060 , NP_199919 . 1; IRE ( other names: INCOMPLETE ROOT HAIR ELONGATION ) , At5g62310 , NP_201037 . 1; PLDα1 ( other names: MSJ11 . 13; PHOSPHOLIPASE D ALPHA 1; PLD ) , At3g15730 , NP_188194 . 1; PLDδ ( other names: ARABIDOPSIS THALIANA PHOSPHOLIPASE D DELTA; ATPLDDELTA; F4B14 . 60; PLDDELTA ) , At4g35790 , NP_849501 . 1; PLDα3 ( other names: F18G18 . 110; PHOSPHOLIPASE D ALPHA 3; PLDALPHA3 ) , At5g25370 , NP_197919 . 1; PLDε ( other names: F7A10 . 25; PHOSPHOLIPASE D ALPHA 4; PLDALPHA4; PLDEPSILON ) , At1g55180 , NP_175914 . 1 Information from http://www . ncbi . nlm . nih . gov/ and http://www . arabidopsis . org/ | Like many root-colonizing microbes , the primitive Basidiomycete fungus Piriformospora indica colonizes the roots of many plant species and promotes their growth . The lack of host specificity suggests that the plant response to this endopyhte is based on general signalling processes . In a genetic screen for Arabidopsis plants , which do not show a P . indica-induced growth response , we isolated a mutant in the OXI1 ( Oxidative Signal Inducible1 ) gene . Previously , this protein kinase has been shown to play a role in pathogen response and is regulated by H2O2 and PDK1 ( 3-PHOSPHOINOSITIDE-DEPENDENT PROTEIN KINASE1 ) . A genetic analysis showed that deletion of PDK1 also abolishes the growth response to P . indica . PDK1 is activated by phosphatidic acid ( PA ) . P . indica triggers PA synthesis and mutants impaired in PA synthesis do not show growth promotion in response to fungal infection . Since defense processes are repressed by P . indica , we propose that a pathway consisting of the PLD-PDK1-OXI1 cascade mediates the P . indica-induced growth response . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"science",
"plant",
"biology",
"biology",
"molecular",
"cell",
"biology"
] | 2011 | The OXI1 Kinase Pathway Mediates Piriformospora indica-Induced Growth Promotion in Arabidopsis |
Scorpion venom induces systemic inflammation characterized by an increase in cytokine release and chemokine production . There have been few experimental studies assessing the effects of scorpion venom on adipose tissue function in vivo . To study the adipose tissue inflammation ( ATI ) induced by Androctonus australis hector ( Aah ) venom and to assess possible mechanisms of ATI , mice ( n = 6 , aged 1 month ) were injected with Aah ( 0 . 45 mg/kg ) , toxic fraction of Aah ( FTox-G50; 0 . 2 mg/kg ) or saline solution ( control ) . Inflammatory responses were evaluated by ELISA and cell sorting analyses in adipose tissue 45 minutes and 24 hours after injection . Quantitative real-time PCR was used to assess the regulation of genes implicated in glucose uptake . The titers of selected inflammatory cytokines ( IL-1β , IL-6 and TNF-α ) were also determined in sera and in insulin target tissues . The serum concentration of IL-1β rose 45 minutes after envenomation and returned to basal level after 24 hours . The pathophysiological effects of the venom after 24 hours mainly involved M1-proinflammatory macrophage infiltration in adipose tissue combined with high titers of IL-1β , IL-6 and TNF-α . Indeed , TNF-α was strongly induced in both adipose tissue and skeletal muscle . We studied the effects of Aah venom on genes implicated in insulin-stimulated glucose uptake . Insulin induced a significant increase in the expression of the mRNAs for hexokinase 2 and phosphatidylinositol 3-kinase in both skeletal muscle and adipose tissue in control mice; this upregulation was completely abolished after 24 hours in mice envenomed with Aah or FTox-G50 . Our findings suggest that Aah venom induces insulin resistance by mechanisms involving TNF-α-dependent Map4k4 kinase activation in the adipose tissue .
Scorpion venoms induce systemic inflammation associated with an increase in cytokine release and chemokine production [1]–[3] . Androctonus australis hector ( Aah ) venom induces high plasma concentrations of proinflammatory cytokines including interleukin 1 beta ( IL-1β ) , interleukin 6 ( IL-6 ) and tumor necrosis factor alpha ( TNF-α ) [4] , and sympathetic tone is activated by experimental envenomation [5] . Several studies report that the sympathetic nervous system regulates the expression of several adipo-cytokines through adipocyte beta-adrenergic receptor [6] , [7] . Adipose tissue secretes various cytokines including TNF-α , IL-6 and adipokines such as leptin and adiponectin involved in glucose metabolism and insulin resistance [8] . Overproduction of TNF-α in both adipose tissue and skeletal muscle contributes to insulin resistance [9] . Furthermore , TNF-α can stimulate the production of other cytokines and chemokines , such as IL-6 and Monocyte Chemoattractant Protein 1 ( MCP1 ) , which can induce insulin resistance [10] , [11] . TNF-α selectively stimulates the expression of a key component of its own signaling pathway , Mitogen-activated protein 4 kinase isoform 4 ( Map4k4 ) , through a TNFR1-dependent mechanism to induce insulin resistance in adipose tissue [12] . Hyperglycemia and hyperinsulinemia have been reported in scorpion envenomed animals [13] . Although the biological activity of scorpion venom on insulin resistance is clearly established , the mechanisms involved are unknown . We have investigated the effects of scorpion venom on glucose uptake in adipose tissue . We tested the contribution , if any , of TNF-α to the modulation of insulin sensitivity after envenomation . We found that following venom injection , TNF-α increases Map4k4 expression in adipose tissue , promoting insulin resistance . The use of a chemical inhibitor ( etanercept ) of TNF-α binding to its receptor reduced Map4k4 expression and restored the glucose uptake in adipose tissue following envenomation .
Mice were injected with Aah venom or the toxic fraction ( FTox-G50 ) , fasted for 6 hours and injected intraperitoneally with 25% glucose ( 1 . 5 g/kg ) for glucose tolerance test ( GTT ) or with insulin ( 0 . 75 units/kg , Actrapid , Novo Nordisk , Denmark ) for insulin tolerance test ( ITT ) . Tail-vein blood was sampled at baseline and various times thereafter and the glucose concentration determined with a blood glucose meter ( Accu-Check , Roche , Dublin , Ireland ) . Blood glucose was similarly determined in control and injected mice in non-fasting conditions . Insulin concentrations in serum samples were assayed by mouse ELISA ( Linco Research , Inc . , St Charles , MO ) according to the manufacturer's instructions . Tissue extracts from animals injected with Aah venom or FTox-G50 were immersed in 4% formol for 24 hours at room temperature and processed according to standard procedures for hematoxylin/eosin ( HE ) staining . Paraffin-embedded sections ( 5 µm thick ) were prepared , washed in xylene and dehydrated with a series of ethanol washes . Adipose tissue sections spaced 200 µm apart were processed with HE staining . The sections were visualized with a bright field microscope and the pictures analyzed with Motic Software ( Motic Images 2000 , Version 3 . 2 . 0 ) . Photographs were obtained using a ×40 objective for measuring adipocyte cell area and were analyzed using Image J software . Measurements were made on 15–25 adipocytes per section and on 3–4 sections covering the entire adipose tissue of each mouse . Epididymal adipose tissues were rinsed three times in Phosphate-Buffered Saline ( PBS ) , and then minced in fresh Hank's Balanced Salt Solution ( HBSS ) containing 1 mg/ml collagenase D and 2 ng/ml DNase I ( Roche , France ) , incubated for 15 min at 37°C and subjected to vigorous pipetting . The resulting cell suspensions were filtered through a 70 µm-pore size mesh and centrifuged at 500 g for 5 min . The pellet ( SVC ) preparation , was then incubated with erythrocyte lysis buffer for 5 min , centrifuged ( 600 g; 5 min ) and resuspended in FACS buffer . The SVCs were incubated with Fc block for 15 min at 4°C and stained with fluorescently labeled primary antibodies for 15 min at 4°C . F4/80-biotinylated FACS antibodies were purchased from AbD Serotec ( Raleigh , NC ) ; CD11b-PE and CD11c-FITC FACS antibodies were from BD Biosciences . The cells were gently washed twice and resuspended in FACS buffer , then analyzed using a FACS Aria flow cytometer ( BD Biosciences , Switzerland ) . Total RNA was extracted from adipose tissue using TRIzol ( Invitrogen , Carlsbad , CA ) . cDNA was synthesized from 1 µg RNA using random hexamer primers and Superscript II ( Invitrogen , Carlsbad , CA ) . Quantification by PCR was performed using the iCycler iQ Real-Time PCR Detection System ( Bio-Rad , Philadelphia , USA ) and iQ SYBR green Supermix ( Bio-Rad , Philadelphia , USA ) . Results were quantified by comparison to a standard curve generated with serial dilutions of a reference cDNA preparation and were normalized with respect to TATA-binding protein ( TBP ) mRNA . All PCR experiments were repeated at least three times . The primers used are listed in Table 1 . Accession numbers for all sequences used are listed in Table S1 . Adipose tissue from controls and Aah venom- or FTox-G50- envenomed mice were placed in 24-well plates ( 100 mg of tissue/well ) with 1 mL of PBS+0 . 2% Bovine Serum Albumin ( BSA ) , and then stimulated with insulin ( 100 nmol/L ) for 1 hour . The tissues were then treated with TRIzol ( Invitrogen , Carlsbad , CA ) to lyse cells for total RNA extraction . Quantitative RT-PCR was used to measure the mRNAs for selected genes implicated in insulin-stimulated glucose uptake . Serum samples were collected from the retro-orbital sinus under anesthesia by injection of sodium pentobarbital . Adipose tissue extracts ( 0 . 5 g of tissue ) were homogenized with a polytron homogenizer in 1 ml buffer containing PBS and 0 . 04% Tween 80 . These tissue samples were centrifuged at 10 , 000 g for 10 min and the supernatants collected . TNF-α , IL-1β and IL-6 concentrations in the tissue supernatants or in the serum samples were determined with ELISA kits ( eBioscience , San Diego , CA ) according to the manufacturer's instructions . Data are reported as means ± s . e . m . Two-way ANOVA tests or unpaired t tests and GraphPad Prism 5 ( GraphPad Software , San Diego , CA ) were used for comparisons between groups . Area under the curve analysis was performed on GTT and ITT curves using Graphpad Prism 5 Software . Statistical significance is indicated * for P<0 . 05 , **P<0 . 01 , and ***P<0 . 001 .
We tested whether total Aah venom and its toxic fraction ( FTox-G50 ) caused glucose intolerance . Young adult mice ( 1 month-old ) were injected with Aah venom or FTox-G50 , and fasted for 6 hours during the day . Fasting blood glucose levels were significantly higher in Aah venom- and FTox-G50-treated mice than controls ( Figures 1A , 1C ) . Following glucose injection , blood glucose levels were twice as high in envenomated mice as in control mice . The glucose level in mice injected with Aah venom returned to the basal level after 120 min whereas blood glucose levels remained elevated in mice injected with FTox-G50 ( Figures 1A , 1C; P<0 . 001 ) . Glucose intolerance can result from the absence of glucose-stimulated insulin secretion or a decreased action of insulin in the peripheral tissues , or both . We therefore assayed plasma insulin in mice injected with Aah venom and FTox-G50 after 45 min and 24 hours ( Table 2 ) . Forty-five min after envenomation , the plasma insulin concentration was significantly higher in mice injected with Aah venom or FTox-G50 ( 0 . 6 to 1 . 66 µg/l and 0 . 61 to 2 . 41 µg/l , respectively ) , than in controls ( 0 . 28 to 0 . 48 µg/l ) . This hyperinsulinemia persisted for 24 hours after envenomation ( Table 2 ) . Thus , Aah venom and FTox-G50 block the action of insulin , resulting in glucose intolerance . The ability of peripheral tissues to take up glucose in response to exogenous insulin was tested . Glucose uptake in mice treated with Aah venom or FTox-G50 was significantly higher ( P<0 . 05 ) to that in control mice ( Figures 1B , 1D ) . These results indicate that mice injected with Aah venom or FTox-G50 were not deficient in glucose-stimulated insulin secretion supporting the hypothesis of insulin resistance . Plasma IL-1β concentrations 45 min post-envenomation were higher in both Aah venom- and FTox-G50–injected mice than controls ( P<0 . 05 ) , but returned to basal values after 24 hours ( Table 2 ) . This finding agrees well with the higher blood glucose concentration in both Aah venom- and FTox-G50 injected mice than control mice ( Table 2 ) . Adipose tissue inflammation was analyzed in control mice and mice injected with Aah venom or FTox-G50 after 45 min and 24 hours . Significant mRNA upregulation of several proinflammatory cytokines ( IL-1β , IL-6 and TNF-α ) was observed ( Figure 2A ) . Similar to previous study [4] , mRNA down-regulation of IL-10 was observed rapidly after envenomation ( Figure 2A ) . To evaluate whether increased adipose cytokine mRNA expression resulted in increased protein production , we have measured IL-1β , IL-6 and TNF-α cytokine release from Aah venom and FTox-G50 adipose tissue ex vivo compared with control tissue ( Figure 2B ) . IL-1β , IL-6 and TNF-α concentrations in adipose tissue 24 hours after envenomation were significantly higher than those in controls ( P<0 . 001 for IL-1β; P<0 . 01 for IL-6 and TNF-α ) , although the circulating IL-1β concentration returned to its basal value ( Figure 2 B , Table 2 ) . These findings demonstrate that there is a local inflammatory profile in adipose tissue 24 hours following injection of Aah venom and FTox-G50 . Quadriceps skeletal muscle inflammation was analyzed in explants from control , Aah venom- and FTox-G50-injected mice 45 min and 24 hours after envenomation . The TNF-α concentration was higher in animals injected with venom or its toxic fraction 24 hours post-envenomation than in controls ( Figure 3 ) . No significant changes were observed for the production of either IL-1β or IL-6 in mice following Aah or FTox-G50 injection ( Figure 3 ) . Adipose tissue morphology was analyzed 24 hours following Aah venom and FTox-G50 injection . HE staining of adipose tissue sections revealed no difference in adipocyte size or morphology between treated and control animals ( Figures 4A , 4B ) ; total number of nuclei was unaffected in the injected mice ( Figure 4C ) . Adipose tissue inflammation is associated with an increased number of adipose tissue macrophage ( ATMs ) , referred to as M1-proinflammatory macrophages in which ATMs express high levels of CD11c , F4/80 and CD11b [15] . Flow cytometry analysis revealed that the total number of triple positive ( F4/80 high , CD11b high and CD11c+ ) cells in adipose tissue was high 24 hours after envenomation , indicating that M1 proinflammatory macrophages accumulated in adipose tissue rapidly after envenomation ( Figure 5 A ) . The number of CD11c-negative ATMs ( F4/80 high , CD11b high , CD11c− ) , referred to as M2-non-inflammatory macrophages , after Aah envenomation was lower than that in controls ( Figure 5B ) . These results indicated that proinflammatory M1 ATMs infiltrate adipose tissue in mice injected with Aah venom . We investigated whether the inflammation of adipose tissue resulting from FTox-G50 injection contributed to adipose-specific insulin sensitivity . We studied the expression of two genes involved in glucose metabolism , Hexokinase 2 ( Hk2 ) and Phosphatidylinositol 3-kinase , regulatory subunit , polypeptide 2 ( Pik3r2 ) . Insulin stimulation resulted in the increased expression of Hk2 mRNA in the adipose tissue of control mice ( Figure 6A ) ; this correlated with an increased insulin-induced Pik3r2 expression ( Figure 6B ) . In contrast , Aah venom or FTox-G50 administration completely blocked the induction by insulin of both Hk2 and Pik3r2 mRNA expression ( Figures 6A and 6B ) . Therefore , the venom causes insulin resistance on glucose metabolism . We tested whether TNF-α inhibition could prevent insulin resistance in adipose tissue . We treated 1-month-old mice with a chemical inhibitor directed against TNF-α binding ( etanercept ) . In control adipose explants , insulin increased Hk2 mRNA expression by 2 fold ( Figure 7A; P<0 . 01 ) . Anti-TNF-α treatment in control mice had no affect on insulin-induced Hk2 mRNA expression by contrast , FTox-G50 injection completely blocked insulin-induced Hk2 mRNA expression ( Figures 6A; 7A ) . However , FTox-G50-induced insulin resistance on glucose metabolism was abolished by anti-TNF-α treatment , suggesting that insulin resistance induced by venom is TNF-α dependent . We tested whether anti-TNF-α treatment of FTox-G50 mice affected Map4k4 kinase expression in adipose tissue . Map4k4 mediates the effects of TNF-α in adipose tissue and skeletal muscle [12] , [16] . As expected , in control mice , basal and insulin-stimulated Map4k4 expression were unaffected by anti-TNF-α treatment ( Figure 7B ) . Injection of FTox-G50 resulted in a 3-fold increase in Map4k4 expression 24 hours later ( P<0 . 01 ) , and anti-TNF-α treatment of FTox-G50-injected mice prevented this Map4k4 upregulation ( Figure 7B ) . Our findings for the effects of the TNF-α inhibitor etanercept and for Map4k4 expression strongly suggest that the venom-induced insulin resistance is TNF-α-dependent and was mediated at least in part by Map4k4 activation in adipose tissue ( Figure 7C ) . We tested whether FTox-G50 caused insulin resistance in skeletal muscle and whether anti-TNFα treatment could prevent any such effect . In skeletal muscle explants , insulin treatment increased Hk2 mRNA expression ( 1 . 7 fold more than in controls; Figure 8A ) . In control mice treated with anti-TNF-α , insulin treatment slightly but not significantly increased in Hk2 expression ( Figure 8A ) . FTox-G50-induced insulin resistance in skeletal muscle was not abolished by anti-TNF-α treatment and Map4k4 expression was not activated in skeletal muscle of mice injected with FTox-G50 ( Figure 8B ) . Therefore , FTox-G50-induced insulin resistance in skeletal muscle appeared to be TNF-α-independent , and is presumably mediated by other factors ( Figure 8C ) .
Scorpion venoms contain a diversity of neurotoxins , including two major polypeptide populations . One consists of several classes of long-chain peptides ( 60–70 amino acid residues ) affecting Na+ channels [17] , and the other includes short-chain toxins affecting K+ [18] , Cl− [19] and Ca2+ channels [20] . All these toxins have direct effects on the ion permeability of excitable cells . The venom of the scorpion Buthus occitanus tunetanus ( Bot ) also contains compounds that activate other cell functions in non-excitable cells , such as adipocytes [21] , [22] . The addition of Bot venom to the culture media of 3T3-L1 adipocytes or freshly dissociated rat adipocytes rapidly increases lipolysis , as indicated by glycerol release , and does so in a dose-dependent manner [22] . In this work , we demonstrate that scorpion venom can reduce insulin sensitivity in mice . This further strengthens the idea that venom may cause insulin resistance , as described previously [13] . Our findings confirm present reports that scorpion venom induces systemic and local inflammation . In particular , we demonstrate that following envenomation , the expression pattern of proinflammatory cytokines ( IL-1β , IL-6 , TNF-α ) changes substantially in adipose tissue concomitant with infiltration by pro-inflammatory macrophages . Interestingly , TNF-α treatment reduces Map4k4 expression and restores glucose uptake in adipose tissue following envenomation . These observations suggest that decreased insulin sensitivity in mice injected with venom is mainly driven by TNF-α . Hyperglycemia and hyperinsulinemia have been described in scorpion envenomed animals [13] . We observed increased in fed glucose levels 45 min after Aah envenomation . This hyperglycemia did not worsen over the subsequent hours following envenomation , although the hyperinsulinemia persisted many hours after envenomation . Possibly this persistence of hyperinsulinemia , most likely as a result of increasing β-cell hyperplasia , serves to compensate and regulate the glucose level such that it returns to normal values . The effects of inflammation and infiltration of macrophages on adipose tissue function and insulin resistance have been extensively studied [23] , [24] . Macrophage recruitment into adipose tissue plays a key role in the etiology of diet-induced insulin resistance [24] . The phenotypes exhibited by tissue macrophages correspond to a M1–M2 polarization state: M1 cells are defined as activated pro-inflammatory macrophages and M2 cells comprise an anti-inflammatory macrophage population . We observed that the total number of M2 cells that are positive for F4/80 and CD11b but negative for CD11c expression in l adipose tissue did not increase following envenomation , whereas the number of M1-like macrophages ( F4/80 high , CD11b high , CD11c+ ) increased significantly . These results are consistent with the view that these proinflammatory CD11c+ macrophages are the cause of the macrophage-linked component of inflammation/insulin resistance; indeed genetic deletion of these cells is sufficient to normalize obesity-induced inflammation , glucose tolerance and insulin resistance [25] . Therefore , it is possible that the increased numbers of M1-like macrophages in adipocyte tissue in mice injected with venom explains the elevated secretion of TNF-α , IL-6 and IL-1β and thereby contributes to the low grade inflammation and insulin resistance . Adipocytes secrete a number of molecules , including leptin , TNF-α , IL-6 , and resistin , that modulate peripheral insulin sensitivity [8] , [26]–[28] . Consistent with this , we found that TNF-α concentrations in adipose tissue and skeletal muscle were increased following injection of Aah venom or the FTox-G50 fraction . TNF-α stimulates the expression of key components of its own signaling pathway , notably Map4k4 , through a TNFR1-dependent mechanism to induce insulin resistance in adipose tissue [12] . Another study has shown that insulin resistance can be abolished by Map4k4 silencing in skeletal muscle [16] . Here , we show that venom injection significantly increased Map4k4 gene expression and that inhibition of TNF-α significantly reduced Map4k4 gene expression in adipose tissue . The specificity of TNF-α action on Map4k4 is due to the unique phosphorylation of JNK1/2 and p38 SAP kinase that leads to activation of the transcription factors c-JUN and ATF2 , which in turn are required for the regulation of Map4k4 expression [12] . Our observations are consistent with these findings and indicate that the decreased insulin sensitivity observed following FTox-G50 injection is mediated by an increase in the TNF-α concentration in adipose tissue , which selectively stimulates the expression of Map4k4 to cause insulin resistance . It may be an oversimplification to attribute adipose tissue inflammation to the effect of a single cytokine: it is likely that several cytokines act collectively to amplify the inflammatory response of adipose tissue . Indeed , we have demonstrated that the concentrations of TNF-α , IL-6 and IL-1β in adipose tissue were increased by Aah venom and FTox-G50 treatment . However , note that although TNF-α increased in skeletal muscle after 24 hours of envenomation , no significant changes in IL-1β and IL-6 levels were detected . The major adipocytokines IL-1β and TNF-α act synergistically to enhance NFkB activation and secretion of IL-6 in adipose tissue [28] . Furthermore , the ability of TNF-α to induce IL-6 secretion is blunted in IL-1RI−/− adipose tissue , suggesting that TNF-α-induced IL-6 is in part mediated by IL-1 [28] . The ability of TNF-α to induce IL-6 secretion has also been demonstrated in adipose tissue and skeletal muscle [16] . It is therefore plausible that increased IL-6 expression in adipose tissue after envenomation may be due to TNF-α production , suggesting that TNF-α and IL-1β work in concert to cause insulin resistance . The mechanisms of activation of pro-inflammatory cytokines in adipose tissue following envenomation remain unclear . Adipose tissue in rodents is innervated by the sympathetic nervous system , which can regulate lipolysis , fat cell number , and the secretion of some adipocytokines , such as TNF-α and MCP1 [6] , [7] . Furthermore , the activity of the sympathetic nervous system in mice increases following envenomation , an effect mediated by catecholamines [6] , [7] . In addition , the activity of sympathetic nervous system can contribute to insulin resistance through effects of catecholamines on adipocytes [6] , [7] . Nevertheless , our results do not rule out the possibility that the expression of adipocytokines is regulated through β-adrenergic receptors . The pharmacological profiles of molecules acting more selectively on β-adrenergic receptor subtypes strongly suggest that the lipolytic action of Bot venom mainly involves the β2/β1 subtype of adrenergic receptors [21] . In conclusion , we report that Aah venom and its toxic fraction induce M1-like macrophages accumulation , inflammation and insulin resistance in adipose tissue . We demonstrate an increase in TNF-α release causing upregulation of Map4k4 expression that disrupts the normal metabolic function of adipose tissue and thereby leads to insulin resistance . These findings suggest that pharmacological inhibition of TNF-α in animals injected with scorpion venom may restore metabolic function and subsequently improve insulin sensitivity . This study provides consistent evidence linking adipose tissue inflammation to the insulin resistance induced by Aah venom . We believe that it would be useful to assess the value of TNF-α inhibitors for the complementary treatment of scorpion envenomation . | Androctonus australis hector ( Aah ) is the scorpion most frequently causing serious human envenomation . In Algeria , Aah is responsible for approximately 50 , 000 cases of scorpion envenomation per year . The Aah sting causes multi-system failure that may be fatal; the manifestations include cardiopulmonary abnormalities , lung edema and inflammation . In addition , hyperglycemia and hyperinsulinemia have been described in scorpion-envenomed animals . The mechanisms causing systemic and local inflammation are poorly understood . Here , we report that Aah venom causes pronounced upregulation of TNF-α , IL1-β and IL-6 expression in the adipose tissue , exacerbating inflammation . As the inflammatory state intensifies , 24 hours after envenomation , TNF-α and other factors are upregulated , and Map4k4 expression increases , blunting the insulin response in adipocytes by decreasing Hexokinase 2 expression . Administration of TNF-α inhibitor following the envenomation reduces Map4k4 expression and restores glucose uptake in adipose tissue . These findings provide coherent evidence linking Aah venom-induced adipose tissue inflammation to insulin resistance . The value of TNF-α inhibitors as a treatment complementary to anti-scorpion venom immunotherapy should be evaluated clinically . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"biology"
] | 2012 | TNF-α Involvement in Insulin Resistance Induced by Experimental Scorpion Envenomation |
We present a new approach to the handling and interrogating of large flow cytometry data where cell status and function can be described , at the population level , by global descriptors such as distribution mean or co-efficient of variation experimental data . Here we link the “real” data to initialise a computer simulation of the cell cycle that mimics the evolution of individual cells within a larger population and simulates the associated changes in fluorescence intensity of functional reporters . The model is based on stochastic formulations of cell cycle progression and cell division and uses evolutionary algorithms , allied to further experimental data sets , to optimise the system variables . At the population level , the in-silico cells provide the same statistical distributions of fluorescence as their real counterparts; in addition the model maintains information at the single cell level . The cell model is demonstrated in the analysis of cell cycle perturbation in human osteosarcoma tumour cells , using the topoisomerase II inhibitor , ICRF-193 . The simulation gives a continuous temporal description of the pharmacodynamics between discrete experimental analysis points with a 24 hour interval; providing quantitative assessment of inter-mitotic time variation , drug interaction time constants and sub-population fractions within normal and polyploid cell cycles . Repeated simulations indicate a model accuracy of ±5% . The development of a simulated cell model , initialized and calibrated by reference to experimental data , provides an analysis tool in which biological knowledge can be obtained directly via interrogation of the in-silico cell population . It is envisaged that this approach to the study of cell biology by simulating a virtual cell population pertinent to the data available can be applied to “generic” cell-based outputs including experimental data from imaging platforms .
Multiparameter flow cytometry is widely used to study the cell cycle and its perturbation in the context of both basic research and in routine clinical analysis [1]–[6] . Such analyses may use a wide range of fluorescent reporters that correlate to the expression of key molecular components of the cell cycle , such as cyclins and cyclin dependent kinases ( CDK ) , [1] or quantify DNA content [5] . Regardless of the particular fluorophores used the quantitative methodology and the ensuing synthesis of biological knowledge is based on statistical analyses of the experimental data sets . For single variable distributions these may include calculations of moments of increasing orders to provide the mean , variance , skewness etc . or cumulative indices such as the Kolmogorov-Smirnov ( K-S ) test [7]–[9] . More complex , multi-variate approaches may involve discriminant function , cluster or principal component analysis in an n-dimensional space [10]–[12] . In all of these approaches there is a common procedural thread: acquisition of data is followed by a statistical parameterisation of the measurement set to which biological form or function can be correlated . In this work , we present an alternative , based on computational simulation of the experiment . A stochastic simulation of the cell cycle dynamics within a large population is initialised with reference to a flow cytometry data set and then evolved , using evolutionary computer algorithms , with assessment of fitness measures derived from comparisons to subsequent data sets . The cell-cycle information is then read directly from the in-silico populations . The development of a simulated cell population approach has been driven by a requirement to track the evolution of large numbers of cells over multiple generations through the cell cycle and provide a means to track progression of both the whole cell population and distinct sub-groups [13] , [14] . This is in the context of mapping the heterogeneity of cell cycle response to perturbation events e . g . effects on cell proliferation of anticancer therapeutics designed to block cell division . In this report we present the conceptual basis of this simulated cell cytometry and detail of the methodology adopted . To demonstrate the application of the technique and validate its potential we use it to quantify cell cycle perturbation in a tumour cell line by a topoismerase II inhibitor which causes endocycle routing in the late cell cycle . The aim of the simulation is to predict the dynamic evolution of a large population of virtual cells ( vcells ) through a life cycle corresponding to that prescribed by their real-life counterparts ( as reported via flow cytometry experiments ) . Furthermore , the model seeks to account for perturbations in the cell cycle progression of the virtual population ( vpopulation ) . The spatial position of the vcells within the cell cycle is initially determined from a real flow data set . From this information , each vcell is assigned a temporal position within the mean inter-mitotic time ( μIMT ) , allowing cell cycle events such as DNA replication and cell division to be stochastically predicted . After the vpopulation has evolved for a given period , they may be compared with a further experimental data set to enable important simulation parameters , governing their evolution , to be optimised and constrained so that correlations between the respective data sets are maximised . Standard approaches to studying cell cycle involve statistical analysis of distributions either 1D involving nuclear content reporters [5] or 2D when further cell cycle molecular reporters are also included [1] , [15] . Thus these are inherently ‘whole population’ measures and can only describe cell variability via global parameters such as the standard deviation from the mean . Whilst automated analytical approaches have been developed in order to reduce user subjectivity [16]–[18] the majority of flow analyses still involve user-defined gating of the as-measured data set to identify and segment a sub-population of cells . Subsequent mapping of this population onto 2D dot plots of fluorescence provides temporal snapshots ( typically with a 24 hour sampling period ) and further partitioning of cells to different compartments , G1 , S , G2/M within normal and polypoid cycles ( see Figure 1 ( a ) ) . These apparent quantitative assessments become inaccurate and , to varying extent , subjective , as they are based either on user identification of the various components in the dot plot by fitting of Gaussian distributions , representing the G1 , S , G2/M fractions , to the DNA content histogram [5] . The challenge of the current investigation is to adopt a computational approach , where the analytical and interpretive steps are implemented at the simulated biology stage and not on the raw data outputs . No new data is added in this approach and the computer simulations could be viewed as an elaborate form of data analysis . However , the methodology does deliver new insight on process , delivering a continuous simulation of the dynamic evolution of the cellular system between fixed sampling points . In this respect , it provides a physical validation when applying various hypotheses to interpret the experimental data . It also goes some way to visualising the variation between individual cells that gives rise to biological heterogeneity as the stochastic simulation delivers a report on population dynamics in which each and every cell can be tracked .
Experimental data is obtained using well-established bi-variate cytometric methods for study of the cell cycle: U-2 OS ( ATCC HTB-96 ) cells were transfected with a G2M Cell Cycle Phase Marker ( GE Healthcare , UK ) , yielding stable expression of a GFP-cyclin B1 . This provides a green fluorescence signal the intensity of which correlates to position in the cell cycle with a minimum signal at G0 and a peak during the G2/M phase . [19] . The culture was maintained under G418 selection in McCoy's 5a medium supplemented with 10% foetal calf serum ( FCS ) , 1mM glutamine , and antibiotics and incubated at 37°C in an atmosphere of 5% CO2 in air . To obtain fluorescence read-out of DNA content an anthraquinone derivative , DRAQ5™ ( 20 µM Biostatus Ltd . , UK ) was used [14] . This binds to DNA providing a fluorescence intensity that can be related to DNA content and thus it reports on cell cycle progression through the S phase to G2/M ( >4N ) or , in the presence of external perturbing agents , progression through polyploid states as the mitotic stage is by-passed [20] ( see Figure 1 ( a ) ) . To obtain a model system in which we can test the simulated cell population approach we have used a cell division by-pass agent: ICRF-193 [bis ( 2 , 6-dioxopiperazine ) ] , a kind gift from Dr A . M . Creighton ( ICRF , London , UK ) . This is a reversible catalytic inhibitor of topoisomerase II that blocks the ability of the enzyme to resolve interlinked DNA replication products [21] . The decatenation of chromosomal replication products is vital for the completing of segregation and hence normal division . ICRF-193 , was prepared in DMSO at 2 mg/ml and used at a peak concentration of 2 µg/ml ( equivalent to 7 . 2 µM ) . To determine the cell population distribution of fluorescence intensity a FACScan flow cytometer was used ( Becton Dickinson Inc . , Cowley , UK ) which was equipped with an air-cooled argon ion laser ( with 488 nm output only ) . GFP-cyclin B1 data was collected using a 30 nm bandpass emission filter centred at 530 nm and the DRAQ5 signal with a 670 nm long pass filter . CELLQuest software ( Becton Dickinson Immunocytometry Systems ) was used for data acquisition . Flow cytometric analysis was used sample sets of 10 , 000 cells and the data presented represents the signal peak height . Typically , tracking of the population was carried out at 24 hour intervals . The computer simulation consists of two principal components; a cell population model ( CPM ) and an evolutionary algorithm - Differential Evolution ( DE ) [22] . The CPM generates a virtual population of cells ( vcells ) , which is initialised using a flow data set . The vpopulation is then evolved and compared to a subsequent flow data set . The CPM evolves each vcell and generates any cell cycle processes deemed relevant to explain the laboratory experiment . A DE algorithm is employed to optimise important ensemble parameters used in the CPM e . g . cell cycle time , enabling the vpopulation to be evolved such that it maximises correlation with the data . A detailed description of the cell population simulation , complete with a full account of the various numerical algorithms and techniques used is given in Text S1 . A brief outline of the main components of the cell population model is given in the following sections with reference to the simulation flowchart shown in Figure 2 . All numerical algorithms have been written in the MATLAB environment ( MathsWorks UK ) ; fragments of pseudo-code for important aspects of the CPM are given in Text S1 . The vpopulation is initialised by reference to a gated 2D flow cytometric data set composing of the cell cycle reporter cyclin B1 ( GFP-cyclin B1 ) and DNA content determination ( DRAQ5 ) ( see Figures 1 ( b ) and S1 ) . The data is gated using a simple cell density cut-off technique , where a region is labelled active if its cell density is above a set threshold , ( see Text S1 ) . Cells within a contour encapsulating the gated fraction serve to initialise the vpopulation position in the intensity space . The same gating procedure ( and threshold value ) is also applied to subsequent experimental data sets at later time points . More sophisticated gating techniques could be applied such as the expectation-maximisation algorithms presented by Boedigheimer et al [18]; however , this simple approach is adequate to establish the validity of our methodology . The gated data is now used to initialise the fluorescence intensities of the modelled cell population , which correspondingly inherits the biological variation seen in Figure 1 ( b ) ( each gated data point initialises one cell ) . The temporal position of each of the virtual cells within the cell cycle is unknown as the flow data ( consisting only of only fluorescence intensities ) contains no direct cell cycle time information . The time-based information , necessary to model the cell cycle dynamics , is extracted from the intensity signal of the biological markers obtained from the experimental data ( see Text S1 ) . The first approximation to assigning a time to each vcell is obtained by considering the DRAQ5 fluorescence intensity , the histogram of which is shown in Figure 3 ( a ) . In our approach , we use the DRAQ5 , nuclear content indicator to position each vcell in the cell cycle making the following assumptions ( i ) the vcells are randomly distributed throughout their cycle and ( ii ) that their DRAQ5 signal is monotonically increasing through the cell cycle as the nuclear content is duplicated . This infers that the minimum and maximum DRAQ5 intensities correlate to the start and finish of the cell cycle and allows us to assign relative position in time to each vcell ( see Text S1 and figure S2 ) . The experimental dataset for DRAQ5 intensity is sorted into ascending order and fitted with a polynomial function ( see Figure 3 ( b ) ) . Because of the inherent digitisation produced by data binning , of the measured intensity , by a flow cytometer several cells will be recorded with the same DRAQ5 signal . The intensity sorting procedure assigns increasing sort indices to vcell sets with the same intensity ( i . e . all cells within a given bin ) the median index value is therefore used when implementing the polynomial fit ( see inset in Figure 3 ( b ) ) . Finally , the polynomial x-axis values are scaled to a range of zero to the inter-mitotic time ( IMT - time between successive mitotic events ) , this gives an absolute time for each cell within its cycle . To relate the GFP signal to cell cycle time the intensity for each cell is plotted against the cell number index obtained from the DRAQ5 sort procedure , again fitted with a polynomial and scaled to give an x-axis running from 0 to the same IMT value ( Figure 3 ( c ) ) . The use of a stoichiometric nuclear content marker ( such as DRAQ5 ) to estimate DNA content and hence cell cycle position is well established and both deterministic and stochastic models have been used previously to obtain continuous temporal descriptions [23]–[25] . Our approach differs from the previous studies in that through the creation of the virtual cell population we model at the level of single cells rather than using population level parameters . The two fitting polynomials describe the evolution of the fluorescence intensities from cell birth to division as a function of time and are used to produce a median path through the cell cycle shown as the solid black line in the 2D GFP-DRAQ5 intensity plot displayed in Figure 1 ( b ) . It is obvious from the plot that many of the cells lie some distance from the median line this is due to natural variability in the measured signals caused by heterogeneity in reporter loading , noise , variation in collection efficiency etc . Each individual vcell is therefore assigned a cell cycle time by choosing a point on the 2D polynomial median line , that minimises the sum difference of the DRAQ5 and GFP intensity values ( see Text S1 and figure S3 ) . Therefore vcells at the same point within the cell cycle will display a heterogeneity in fluorescence signal value ( corresponding to the width of the population plot in x and y-directions in Figure 1 ( b ) . Therefore , to calculate the time-dependent trajectory of each vcell through the 2D intensity space we update the DRAQ5 and GFP intensity values using the median line as time is incremented ( see Text S1 ) . To summarise , the experimentally measured data is used to establish a virtual cell population with exactly the same heterogeneity in fluorescence signal as seen in the experiment . This vpopulation is then evolved within a stochastic simulation allowing for variability in fluorescence intensity and IMT using a pair of polynomial functions that describe the cell cycle dependence of the signal , i . e . the absolute fluorescence is stochastic but the time evolution function is the same for all cells . Once initialised each member of the population has three discriminating properties corresponding to: ( i ) a time in the cell cycle , ( ii ) DRAQ5 fluorescence intensity and ( iii ) GFP-cyclin B1 levels . In order to mimic DNA synthesis and replication , a supplementary parameter , DRAQ5DNA2 , is required; DRAQ5DNA2 details the DRAQ5 magnitude at which each vcell has doubled its DRAQ5 intensity ( see Text S1 and figures S4 and S5 ) . The CPM directly relates this to the point at which a real cell has doubled its DNA content , i . e . a phase transition to G2 . The value of DRAQ5DNA2 is deduced by calculating the initial DRAQ5 intensity of each vcell at the start of the cell cycle ( see Figure 1 ( b ) , black curve ) , which from the above is estimated at the effective intensity value just after a mitotic event , then assessing the time at which this initial intensity doubles using the polynomial functions shown in Figures 3 ( b ) and ( c ) . Monitoring of the simulated DRAQ5 intensity then allows identification of cells that have multiplied their DNA content allowing placement of each into the following sub-groups: normal cycle – DNA index , DI = 2N ( G1 ) or 4N ( G2/M ) ; polyploidy cycle - DI = 4Np ( G1p ) or 8Np ( G2p/Mp ) . Once vcells have entered the G2/M phase the probability of them entering the M phase and undergoing cell division is calculated . This is achieved using a simple stochastic decision process [13] , where we define a cumulative Gaussian probability distribution which scales between 0 and 1 , defined in terms of a mean inter-mitotic time , with an associated standard deviation , over the cell cycle time . Both these parameters are to be optimised via the evolutionary algorithm to best fit the second set of flow data . At each time step a random number , uniformly distributed in the interval [0 1] is generated and is compared with the cumulative probability distribution value at that time . If the random number is less than the probability distribution value calculated then mitosis is deemed to occur and the simulation generates two daughter cells at t = 0 in G1/S with the DRAQ5 and GFP-cyclin B1 associated with the parent cell . Otherwise , the cell remains in the G2/M phase for re-analysis at the following time step , which will increase both of its intensity coordinates resulting in a higher probability of mitosis ( the cumulative Gaussian distribution tends to 1 with increasing time ) . The implementation of this mitotic variability produces further heterogeneity in the IMT of the individual vcells . The cell population model is defined by a set of parameters specific to the flow cytometry experiment conducted . Optimisation of the fit between simulation and experiment is dependent upon selection and minimisation of the population variables , in our case: the mean inter-mitotic time , its standard deviation and a parameter detailing the presence of a drug in the vpopulation . There are several different methods , which could be used to determine the best fit to the experimental data; we choose to use a differential evolutionary technique to optimise these cell cycle parameters . The quality of fit associated with a set of CPM parameters is determined by calculation of the ratio of evolved vcells to that measured experimentally within a numerically deduced gated region . This simple maximisation strategy , works well for both therapeutically ( un ) perturbed systems , although newer versions of the CPM will explore more sophisticated 2D cross correlative algorithms to infer fitness . Convergence of the differential evolution algorithm is determined true when the quality of fit varies by less than 1% over five subsequent generations ( see Text S1 ) .
To illustrate the evolution of the vcell population , we generate a series of snap-shots derived at different temporal intervals ( Figure 4 - green population ) demonstrating the simulated intensity dot plot at 6 , 12 , 18 and 24 hours respectively after initialisation by an experimental data set . Here , the vcell population has a mean IMT of 22 hours and an associated standard deviation of 6 hours; a small subpopulation of vcells can be depicted ( red dots ) also a contour ( dashed black line ) is displayed , indicating the extent of the gated experimental data set at initialisation . Given that the cells in these experiments are randomly distributed within their cycle and a statistically relevant data set is sampled the acquired plots appear identical for a control sample with an unperturbed biology . The advantage of the simulated population approach is therefore evident in Figure 4 , as a discrete sub-set of cells is identified and its dynamics tracked over a period of time . Despite using a single experimental sample information is obtained across the whole of the cell cycle due to the assumption of random temporal distribution . The fundamental insight gained here is the adoption of a simulated cell approach and subsequently the visualisation of the temporal dimension encoded in the fluorescence intensity distributions . To test the ability of the simulation to capture more complex dynamics associated with aberrant cell cycle progression and variance of response across sub-populations a cell cycle perturbation experiment was undertaken using a mitotic by-pass agent ICRF-193 . Cells treated with this agent progress through multiple replication cycles without undergoing mitosis , therefore doubling DNA content [21] . This leads to an evolving polyploid population that is identified using the nuclear dye , DRAQ5 to obtain an optical read-out of DNA content . Perturbation of the cell cycle and rerouting of cells in this manner provides a system in which the population dynamics of diverted sub-groups within the normal and polyploidy cycles can be analysed . The challenge for the cell population simulation is to track the inter-related pharmacodynamics , taking full account of the detailed evolution of the accompanying fluorescence data . A block and chase experiment was conducted in which cells were continuously treated with ICRF-193 for 24 hours ( Figure 5 ( a–c ) ) . A 2D dot plot of the cell cycle ( GFP-cyclin B1 ) and nuclear content ( DRAQ5 ) reporters at the 24 hour time point shows a sub-population of cells with low GFP-cyclin B1 expression and a DNA index of 4N i . e . polyploid cycle cells in the G1/S phase ( Figure 5 ( b ) ) . Compared to control conditions , where all cells were engaged in the normal cell cycle . Following the 24 hour drug treatment with ICRF-193 , wash-out allows cells to further cycle unperturbed under normal conditions for a further 18 hours ( including cell division ) . ICRF-193 is a reversible topoisomerase II blocker and so removal of this agent enabled the sub-population of cells within G2/M of the normal cycle to be routed back into normal cycle ( i . e . to G1/S ) . Hence , the 42 hour data shows two distinct population groups describing cells within the normal and polyploid cycle ( Figure 5 ( c ) ) . To include the effect of the ICRF-193 in the CPM we include a further optimisation parameter Nbp , which describes the fraction of vcells that have doubled their DNA content ( G2/M phase ) but have bypassed mitosis . These are selected stochastically and inhibited from undergoing cell division when under drug ‘dosing’ conditions . This assignment is undertaken at each time step , until the required percentage of vcells in the population have by-passed mitosis . In the drug ‘wash-out’ conditions , the reduction in Nbp is modelled with a half-life , t1/2 , corresponding to the temporal persistence of the drug-induced perturbation . Thus depending on drug administration or wash-out the CPM has three optimisation variables to be minimised through the evolutionary methods described previously . The comparison of real ( red population displayed in Figure 5 ( b ) ) and vcell populations for selection of the variable parameter values is made 24 hours after initialisation . The optimised vcell population together with the real data contour is shown in Figure 6 ( a ) together with a contour illustrating the position of the initial data set . The simulation clearly captures the key features of the population evolution and given the stochastic nature of both real and virtual cell populations they are well correlated . At this point , following 24 hours of continuous drug treatment , there are large fractions of 4n cells in the normal and 4np polyploid phases as well as a sub-population of polyploid cells progressing to 8np phase . A small population sub-set ( located within the black dashed contour in Figure 6 ( a ) ) represents the vcells yet to be influenced by the drug . As mentioned above , at each discrete time throughout the simulation the vcells are stochastically tested to see if they have been drug treated , hence , a finite time must elapse before all vcells can be influenced by the action of the drug . The vcell dynamical parameters corresponding to the fits shown in Figure 6 are indicated in Table 1 . In the presence of the drug the simulation indicates a mean inter-mitotic time of 36 hours with a standard deviation of 4 hours . In comparison , the IMT value from fitting to a control set of data is 22±4 hours . Multiple runs ( 1 , 000 simulations of the experimental data ) of the model indicate that the variation in the tabulated values , due to stochastic variation and evolutionary selection , is less than 5% . The prediction of an extended IMT within drug treated cells is in agreement with previous studies on the effects of ICRF-193 , showing delays in progression to the mitotic phase plus extension in the duration of mitosis once initiated . Although , the CPM cannot elucidate on the persistence of individual phase duration it does accurately estimate their combined effect . During the chase phase of the assay subsequent to drug wash-out , the simulation evolves from 24 to 42 hours in a similar manner to that above , with the difference that the Nbp parameter is indirectly optimised using a half-life to describe its temporal decay; i . e . the fraction of vcells that retain drug-induced division-bypass is where and t is the time since wash out . The intensity coordinates of the vpopulation at 42 hours after ICRF-193 washout are displayed in Figure 6 ( b ) . The simulation has captured the important features of both the normal and polyploid cycle dynamics . That is , there is a significant sub-population of vcells in each of the four DNA indexed phases . For the 2D fit shown in Figure 6 ( b ) the simulation uses a mean inter-mitotic time of 22±7 hours respectively . This agrees remarkably well with that measured through microscopic techniques for an unperturbed real populace . Furthermore , the simulation gives an insight to the temporal persistence of the drug on the virtual population , indicating that a significant sub-population retain or are committed to the division bypass over the course of a few hours following wash-out . The fact that an effective continuum of intensities straddling the 4np and 8np phases in both real and virtual populations is evident reinforces the simulation result which highlighting of temporal persistence of ICRF-193 post washout . The evolution and perturbation of the vcell population is shown in Video S1 . The continuous population dynamics provided by the simulation are shown in Figure 7 . At the initialisation point ( t = 0 hours ) we see that a significant fraction of vcells are present in the 2n phase compared to that in the 4n phase ( blue and green curves respectively ) , ∼4∶1 ratio . Over the first 24 hours , the drug perturbation re-routes cells from the normal into the polyploid cycle . Thus , the 4n population is stable as equal numbers of move in and out of it producing linearly decreasing 2n and linearly increasing 4np sub-populations . The percentage of mitotic-bypass cells therefore increases over time , but due to dynamical constraints and the optimised mean inter-mitotic time of 36 hours , this does not reach 100% ( maximum of ∼85% ) before washout . The vertical dotted line in Figure 7 indicates the initiation of the washout phase of the simulated experiment . Following drug washout at 24 hours the fraction of mitotic-bypass vcells decreases exponentially with an optimised half-life of 3 hours , thus it takes the full 18 hours following drug removal to achieve something near to normality . This same dynamic inevitably affects the re-creation of a 2n population . This gives an insight to the temporal persistence of drug on the vpopulation indicating that a sub-population retains the bypass commitment for a few hours post washout .
The use of stochastic computing approaches plus evolutionary algorithms to evolve a simulated cell population provides a new approach to the analysis of multi-variate data sets obtained by flow cytometry . In using this simulated biology process to analyse cell cycle perturbation we have obtained detailed information cell cycle time and the detailed dynamics of cell division and proliferation . Furthermore , we have shown that a subpopulation or cohort can be defined and tracked throughout the time course of the experiment without the need for further molecular markers , this can be essentially viewed as an in silico representation of the pulse chase experimental methods such as those incorporating two-parameter flow cytometry analysis: with DNA content and BrdUrd [26] . When applying the technique to drug-treated populations the pharmacodynamic indicators can be tracked and sub-populations within normal and polyploid cycles differentiated . Further , the temporal continuity inherent in the computational assessment also highlights details un-resolvable in the experimental sampling , such as cell cycle traverse ( inter-mitotic time variation ) , cell cycle delays ( persistence of drug-induced effects ) and has also identified the occurrence and location of cell cycle restriction points , which with additional molecular mapping can be further defined [27] . Also , the simulated experiment permits individual in addition to ( sub ) population cell tracking allowing single cell lineage tracking and the ensuing generational patterns and relationships to be continually analysed . This is a systems approach to whole tumour population evolution leading to lineages , in contrary to tracking individual lineages and extracting a global population response [28] . We envisage that this approach would be much more easily applied to a screening approach appropriate for sampling tumours both in vitro and in vivo . In this initial implementation of the technique , we use a cell cycle marker that reports on relative cycle time and a nuclear marker which allows us to discriminate between normal and polyploid cell populations , therefore no further information of the intricate details of the cell cycle ( apart from mean IMT distribution ) can be deduced . In this respect , the simulated cell methodology provides a framework , describing the relationships between cells within a population , at a system level i . e . in the context of progression through a unitary cycle with associated genetic replication and cell division . Importantly this structure can enhance existing approaches by linking detailed molecular level models of cellular evolution through specific cell cycle phases [26] , [27] , [29] to cell heterogeneity and its influence on population level dynamics . We have adopted an approach of minimised complexity in order to clearly demonstrate the concept without the obfuscations of detailed algorithm structures and data filtering . A simple dot density cut-off filter is applied to gate the data , the number of variable parameters within the genetic algorithms is reduced to a minimum of three and goodness of fit assessed by a straightforward maximisation of simulated cells within an experimental data contour . Whilst future work will explore the potential of more sophisticated computational techniques , the simple conceptual base presented here already provides automated , objective data analysis that encapsulates the fundamental biology and delivers statistically robust results . Given the stochastic nature of the simulation it could be argued that a statistical approach should be maintained and increased simulation runs be used to acquire added certainty rather than increased model complexity . The large data sets collected in flow cytometry and the stochastic variation associated with biological systems naturally lead to statistical analysis techniques for data interpretation [23]–[25] . These have proven to be powerful tools in cell biology , however when focussing on individual cell behaviour and heterogeneity expressed at the single cell level the integrative measures of statistics are limiting . The development of a simulated biology , twinned to a real cell population , by fitting experimental data sets , maintains the statistical relevance and provides discrimination via individual cell recognition . The creation of in-silico cells brings the potential for interpolation and extrapolation thus a continuous temporal report of complex population dynamics can be produced from discrete measurements and cellular behaviour predicted beyond the limited time frame imposed by experiment and environment . The temporal continuity inherent in the computational assessment also highlights details of the pharmacodynamics , un-resolvable in the experimental sampling , such as inter-mitotic time variation and persistence of drug-induced effects . Perhaps the most beneficial aspect of the simulated cell approach is its ability to provide direct knowledge of biological state allowing a computational systems approach to inform the biology . This contrasts with traditional flow analysis , which provides information that is primary in relation to data but secondary in relation to cells; i . e . a choice can be made between direct data analysis with interpretation to translate to cell behaviour or direct read-out of cellular information from a data-directed simulation . By ensuring interoperability of the modelling algorithm with experimental cytometry outputs , the simulation provides emergent features of the cell cycle and the functional operation of molecular restrictions and checkpoints; providing further the foundation for considering the evolving asymmetric and symmetric patterns of a dynamic cellular system . | One of the key challenges facing cell biologists today is understanding the influence of molecular controls in shaping and controlling cell growth and proliferation . There is growing recognition that abnormal progression through the cell cycle and the associated effects on the growth of cell populations has a major impact on a wide range of biological and clinical problems , including: tumour growth , developmental control , origins of chromosomal instability and drug resistance . Multiparameter flow cytometry is frequently used to assess proliferation dynamics of cellular populations using fluorescent reporters generating large data sets that can inform simulation models . We have developed stochastic computing approaches allied to evolutionary algorithms to produce simulated cell populations—providing a new approach to the analysis of real multi-variate data sets obtained by flow cytometry . The methodology delivers new insight on biological processes in delivering a continuous simulation of the dynamic evolution of a cellular system between fixed sampling points , hence , converting static data into dynamic data revealing the effective traverse of the cell cycle , restriction points and commitment gateways . The approach also permits the visualisation of the variation between individual cells reflecting biological heterogeneity and potentially Darwinian fitness , given that the simulation delivers a report on population dynamics in which each and every cell can be tracked . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"mathematics",
"biophysics/theory",
"and",
"simulation",
"computational",
"biology/systems",
"biology",
"physics/interdisciplinary",
"physics"
] | 2010 | Flow-Based Cytometric Analysis of Cell Cycle via Simulated Cell Populations |
Paramutation involves homologous sequence communication that leads to meiotically heritable transcriptional silencing . We demonstrate that mop2 ( mediator of paramutation2 ) , which alters paramutation at multiple loci , encodes a gene similar to Arabidopsis NRPD2/E2 , the second-largest subunit of plant-specific RNA polymerases IV and V . In Arabidopsis , Pol-IV and Pol-V play major roles in RNA–mediated silencing and a single second-largest subunit is shared between Pol-IV and Pol-V . Maize encodes three second-largest subunit genes: all three genes potentially encode full length proteins with highly conserved polymerase domains , and each are expressed in multiple overlapping tissues . The isolation of a recessive paramutation mutation in mop2 from a forward genetic screen suggests limited or no functional redundancy of these three genes . Potential alternative Pol-IV/Pol-V–like complexes could provide maize with a greater diversification of RNA–mediated transcriptional silencing machinery relative to Arabidopsis . Mop2-1 disrupts paramutation at multiple loci when heterozygous , whereas previously silenced alleles are only up-regulated when Mop2-1 is homozygous . The dramatic reduction in b1 tandem repeat siRNAs , but no disruption of silencing in Mop2-1 heterozygotes , suggests the major role for tandem repeat siRNAs is not to maintain silencing . Instead , we hypothesize the tandem repeat siRNAs mediate the establishment of the heritable silent state—a process fully disrupted in Mop2-1 heterozygotes . The dominant Mop2-1 mutation , which has a single nucleotide change in a domain highly conserved among all polymerases ( E . coli to eukaryotes ) , disrupts both siRNA biogenesis ( Pol-IV–like ) and potentially processes downstream ( Pol-V–like ) . These results suggest either the wild-type protein is a subunit in both complexes or the dominant mutant protein disrupts both complexes . Dominant mutations in the same domain in E . coli RNA polymerase suggest a model for Mop2-1 dominance: complexes containing Mop2-1 subunits are non-functional and compete with wild-type complexes .
Paramutation , an interaction between specific alleles that leads to a heritable change of expression of one allele , was first described for the maize red1 ( r1 ) gene [1] . Subsequently three more regulatory genes of the flavonoid biosynthetic pathway , b1 ( Booster1 ) , pl1 ( plant color1 ) , and p1 ( pericarp color1 ) [2]–[4] , and a gene involved in phytic acid biosynthesis [5] were shown to undergo paramutation in maize . Paramutation-like phenomena have also been reported in other plants , fungi , and animals [for a review , see [6]–[8]] . Paramutation terminology defines alleles that induce silencing as paramutagenic and alleles that become silenced as paramutable . Once silenced ( paramutated ) , alleles are designated with an apostrophe to signify their paramutant state . In addition to becoming heritably silenced , paramutant alleles also acquire the ability to silence naïve paramutable alleles . Paramutant and paramutable states often have different stabilities , which can potentially be reversible depending on the locus [for a review , see [7] , [9] , [10]] . Most alleles of a locus do not participate in paramutation . Key sequences mediating paramutation have been identified for two systems , b1 [11] , [12] and p1 [4] , [13] . Recombination mapping between alleles that do and do not participate in b1 paramutation defined a specific sequence that when tandemly repeated is absolutely required for paramutation [11] , [12] . Characterization of these repeats revealed that the paramutable and paramutagenic alleles have identical DNA sequences and numbers of repeats , but differ in their chromatin structure demonstrating that paramutation is epigenetic and associated with changes in chromatin [11] . Transgenic approaches were used to identify sequences within p1 sufficient to mediate paramutation . These sequences lie within a direct repeat flanking the p1 alleles that participate in paramutation [4] , [13] . At the r1 locus , paramutagenic alleles contain direct and inverted repeats and the strength of paramutation correlates with repeat number [14]; paramutable alleles have inverted repeats [15]–[18] . Thus , while there are no sequence similarities between the regions that mediate paramutation at these distinct loci , a common theme is the presence of direct or inverted repeat sequences . Mutations that alter paramutation have been isolated using screens with either the b1 or pl1 paramutation systems [19]–[21] . Several of the genes identified in these paramutation screens have been cloned and to date all share homology with genes in Arabidopsis that mediate RNAi transcriptional silencing of transgenes or endogenous genes . The first cloned gene required for paramutation was mediator of paramutation 1 ( mop1 ) , which encodes a RNA dependent RNA polymerase most similar to Arabidopsis RDR2 [22] that mediates heterochromatic silencing of repeats through 24 nt siRNAs [23] . In addition to preventing paramutation at multiple loci and increasing the transcription of paramutated alleles [20] , mop1 mutations also reactivate Mutator transposons [24]–[26] and transcriptionally silenced transgenes [27] . The second gene cloned , required to maintain repression1 ( rmr1 ) , encodes a SNF2-like ATPase [28] , a factor similar to , but distinct from Arabidopsis DRD1 ( Defective in RNA Directed DNA methylation1 ) involved in RNAi-mediated transcriptional silencing [29] and CLSY1 ( CLASSY1 ) involved in RNA signal spreading [30] . The rmr1 mutation increases the expression of previously silenced pl1 and b1 alleles , but does not prevent paramutation at pl1 [19] , [28] or b1 ( V . Chandler , unpublished data ) , suggesting it is involved in maintaining the silenced epigenetic states . While rmr1 mutations can also reactivate transcriptionally silenced transgenes , these transgenes are efficiently resilenced upon the introduction of a wild type allele [27] , in contrast to mop1 mutations in which the reactivated transgenes can remain heritably active even when the wild type allele is reintroduced [27] . The third gene cloned , rmr6 , encodes the largest subunit of the plant specific DNA-dependent RNA polymerase most similar to Arabidopsis NRPD1 [31] , the largest subunit of the Pol-IV complex required for primary siRNA biogenesis [32] , [33] . Mutations in rmr6 cause dramatic developmental phenotypes and prevent paramutation at pl1 , b1 , and r1 as well as relieve silencing of paramutant alleles [21] . The genes cloned to date , when mutated , show a dramatic reduction in 24 nt siRNAs normally associated with heterochromatic silencing of repeated sequences [31] , [34] . In Arabidopsis the related RNAi heterochromatic silencing pathway is often referred to as RdDM ( RNA directed DNA Methylation ) ; the transcriptional silencing requires small RNA biogenesis and targets homologous promoters with DNA methylation and repressive histone modifications [35]–[37] . This pathway in Arabidopsis mediates transcriptional silencing by transgenes in which inverted repeats of promoters are transcribed to generate dsRNA ( referred to a pIR transgenes ) [38] . The dsRNA is processed to 24 nt siRNAs , leading to DNA and chromatin modifications and silencing of any endogenous gene or transgene sharing homologous promoter sequences [36]–[39] . This pathway also mediates de novo DNA methylation and silencing of several endogenous genes associated with tandem repeats [40] . The current model for the Arabidopsis RNAi heterochromatic silencing pathway involves the genes identified in maize discussed above as well as other factors . Pol-IV is required for siRNA biogenesis and is thought to mediate the synthesis of non-coding RNAs at multiple repetitive endogenous loci using either double-stranded DNA , or single- or double-stranded RNA as a template [for a review , see [37]] . The resulting single stranded RNA is postulated to be used by the RNA dependent RNA polymerase , encoded by RDR2 [23] , to generate double stranded RNA molecules that are then diced into double stranded 24 nt siRNAs by an RNAseIII-like endonuclease , encoded by Dicer-like 3 ( DCL3 ) [23] . Transgenes or endogenous genes that can produce dsRNA via strong Pol-II promoters , such as pIR transgenes , do not require Pol-IV or RDR2 for silencing [41] . Another plant-specific Pol-II related polymerase complex , Pol-V [42] , associates with target DNA with the help of the SNF2-like ATP-dependent chromatin remodeler , encoded by DRD1 ( RNA-directed DNA methylation1 ) [43] , and a hinge protein , DMS3 [44] . Pol-V produces transcripts regardless of the presence or absence of the 24 nt siRNA signal [45] , [46] . Nascent Pol-V RNA transcripts associate with an RNA binding protein KTF1 ( KOW domain containing transcription factor 1 ) [47] and recruit the ARGONAUTE4 ( AGO4 ) protein [48] , which is complexed with a guiding strand of complementary 24 nt siRNA . The AGO4 complex then recruits the de-novo DNA methylation enzyme DRM2 ( domain rearranged methytrasferase2 ) [45]–[47] , [49] , and histone modification factors such as HDA6 ( Histone deacetylase6 ) , and a histone methyltransferase , KYP ( Kryptonite ) , to establish and reinforce silencing at target loci [50] , [51] . The requirements of a RDR2-like RNA dependent RNA Polymerase encoded by mop1 [22] , a NRPD1-like large subunit of Pol-IV encoded by rmr6 [31] , and the SNF2-like factor encoded by rmr1 [28] , coupled with the requirements for transcribed tandem repeats to mediate b1 paramutation [11] , [22] , loss of siRNAs in several paramutation mutants [31] , [34] , and associated chromatin differences between paramutagenic and paramutable b1 and p1 alleles [4] , [52] , [53] , suggest that paramutation involves a mechanism similar to the RdDM pathway in Arabidopsis . However , paramutation has properties that are distinct from RdDM [8] , [40] . Most dramatically , the silencing associated with paramutation is highly heritable after the paramutagenic allele is segregated away and the newly silenced allele itself becomes paramutagenic in subsequent generations . These characteristics do not occur with RdDM in Arabidopsis; in most instances expression of the targeted loci returns to normal after the inducing transgene ( or locus ) is segregated away . Even in the examples of heritable silencing at the FWA locus , the silenced allele is not paramutagenic; it can not silence an active allele as reviewed in [40] . In addition , maize contains significant levels of a new class of 22 nt heterochromatic RNAs [34] , suggesting greater complexity with these processes in maize . Clearly , further studies of paramutation in maize are needed to understand how the paramutagenic and paramutable alleles communicate to set up and heritably maintain RNA-mediated transcriptional silencing . In this report we describe the identification of mop2 and show that it is required for paramutation at multiple loci . Map-based cloning demonstrated that mop2 encodes a second largest subunit of plant specific RNA polymerases similar to the NRPD2/NRPE2 subunit shared by Pol-IV and Pol-V in Arabidopsis . Unlike Arabidopsis , which encodes a single functional gene , maize encodes three closely related genes , all of which appear to encode full length proteins and show significant overlapping expression in a variety of tissues . We report on a number of additional gene silencing phenotypes of an EMS-induced dominant mutant allele of mop2 , Mop2-1 , which has a single nucleotide change in a domain highly conserved among all polymerases ranging from E . coli to higher eukaryotes . Our results suggest that Mop2-1 is disrupting siRNA biogenesis and may be disrupting chromatin targeting properties of small RNAs in maize , and that its ability to disrupt epigenetic processes varies with dosage . Models for Mop2-1 dominance and implications of our findings for mechanisms of b1 paramutation are discussed .
Paramutation at b1 involves two alleles , paramutable B-I ( B-Intense ) and paramutagenic B' [54] . Phenotypes of B-I and B' are easily distinguished; the highly expressed paramutable B-I allele specifies high levels of purple anthocyanin pigment in most of the above ground organs ( sheath , husk , and tassel ) , while the low expressed paramutagenic B' allele confers light speckled plant pigmentation ( Figure 1A ) . The B-I allele is unstable and spontaneous paramutation to the low expressed B' state occurs at variable frequencies ranging from 0 . 1 to >50% , depending on the stock . In contrast to B-I , the silenced B' state is very stable and no change to higher expression has been observed in wild types backgrounds in many thousands of plants examined [12] , [54] , [55] . In addition to being very stable , the B' state is highly paramutagenic , when B-I and B' are combined in an heterozygote , B' always paramutates B-I resulting in all F1 progeny having light plant pigment [54] . In addition , the newly paramutated B' allele ( B-I in the previous generation ) is as efficient as the parental B' allele at causing paramutation of naïve B-I alleles [54] . The absolute penetrance of b1 paramutation in wild type backgrounds , such that B' always changes the B-I allele to B' , was exploited to identify mutations required for b1 paramutation . For this genetic screen B' pollen was treated with the chemical mutagen , ethyl methanesulfonate ( EMS ) , and used to fertilize B-I ears ( Figure 1A and Materials and Methods ) . The resulting M1 progeny were screened for rare dark purple plants , which would appear if B' failed to paramutate B-I , presumably due to the presence of a dominant mutation that prevented paramutation . One such exceptional dark plant was found among ∼7300 M1 plants ( Figure 1A ) and the putative mutation this plant carried was named Mediator of paramutation2-1 ( Mop2-1 ) . The B-I allele that escaped paramutation was designated B-I* to indicate its exposure to B' , but that it remained B-I ( Figure 1A ) . In subsequent generations the presence of a single copy of the Mop2-1 mutation continued to protect B-I from paramutation by a newly introduced B' allele and this protection occurred independent of whether Mop2-1 was transmitted through the male or female ( Figure S1 and Figure S2; data not shown ) . From the initial experiments it was apparent that when Mop2-1 is heterozygous B' silencing is not relieved , as these plants are lightly pigmented ( Figure 1B and Figure S1 ) . It was also apparent that Mop2-1 was loosely linked to b1 ( Figure S1 ) . To examine whether the Mop2-1 mutation when homozygous might relieve B' silencing , a family segregating Mop2-1 B'/Mop2-1 B' and Mop2-1 B'/+ B' plants was developed ( Figure S2 ) . If B' silencing was relieved in Mop2-1 homozygous plants , then such plants would be expected to be darker relative to Mop2-1 B'/+ B' siblings . This expectation was met as the majority of homozygous Mop2-1 plants ( 69% ) showed increased pigmentation ( Chart in Figure 1B ) , but none were as dark as B-I ( Photo in Figure 1B ) . The remaining homozygous Mop2-1 plants had medium dark ( 28% ) or light ( 2% ) pigment ( Figure 1B ) suggesting that the ability of Mop2-1/Mop2-1 to relieve B' silencing was not fully penetrant . To determine whether the increased expression of the B' allele observed in homozygous Mop2-1 plants was a heritable change in the absence of Mop2-1 , darkly pigmented Mop2-1 B' homozygous plants were crossed with the B-I/B-P tester ( Figure S2 ) . The B-Peru ( B-P ) allele does not participate in paramutation and confers essentially no plant color , which provides an excellent background for scoring B' and B-I pigmentation . Examination of Mop2-1 B'/+ B-P progeny revealed no dark plants ( Figure 1C ) . The majority of plants had light B' pigment ( 87% ) , indicating that in these individuals B' was efficiently re-silenced in the presence of the wild type allele . The presence of some medium dark plants ( 13% ) suggested that increased expression could be weakly heritable . Taken together , these results demonstrate that only when the Mop2-1 mutation is homozygous is B' silencing relieved , unlike preventing paramutation where a single copy of the Mop2-1 mutation was sufficient to prevent B' from silencing B-I . The crosses described in Figure S2 were also used to assess the penetrance of the Mop2-1/+ effects on b1 paramutation in a larger population of plants . Analysis of the Mop2-1 B'/+ B-I * progeny demonstrated that 96% of these plants were darkly pigmented indicative of no paramutation and demonstrating that Mop2-1 acts in a dominant and highly penetrant manner to prevent paramutation ( Figure 1D ) . There were a few ( 5/126 ) Mop2-1 B'/+ B-I * plants that had a medium dark phenotype . These could result from either spontaneous paramutation of B-I* to B' , incomplete penetrance of Mop2-1/+ in preventing paramutation , or both . To confirm that B-I* segregates phenotypically unchanged from B' in Mop2-1/+ plants , a backcross to the B-I/B-P stock was performed ( Figure S2 ) . Phenotypic and molecular markers were used to identify the + B-I*/+ B-P plants that are informative for B-I* heritability . If B-I* escaped paramutation in the previous generation , then , after accounting for recombination between the linked b1 and mop2 loci , assuming full Mop2-1 penetrance , and absence of spontaneous paramutation ( Figure S2 ) , 73% should be parental ( + B-I*/+ B-P; dark plants ) and 27% recombinant ( +B'/+B-P; light plants ) . Results presented in Figure 1E demonstrate that light plants ( 22% ) , likely representing the B'/B-P progeny , were observed at a frequency close to the expected 27% ( χ2 = 0 . 53 , P = 0 . 46 ) . The frequency of dark plants ( 53% ) was lower than expected ( χ2 = 7 . 7 , P = 0 . 005 ) , with medium dark plants observed at 24% . As there are no molecular markers to distinguish B' from the B-I , the medium dark plants could theoretically represent reduced expression of B-I* ( because of spontaneous paramutation of B-I* to B' ) or increased expression of +B'/+B-P recombinants . Independent of these hypotheses , the significant number of dark B-I* plants segregating demonstrates that the presence of one Mop2-1 mutant allele in the previous generation can prevent paramutation . This finding is in sharp contrast to wild type backgrounds in which exceptional B-I-like plants have never been observed in thousands heterozygous B'/B-I plants grown over decades of experiments [12] , [54] , [55] . The B' and B-I stocks were differentially marked with two genes linked to b1 ( Figure 1A ) , which enabled following the original chromosomes carrying B' and B-I in subsequent generations . Presence of these markers enabled the initial observation that the Mop2-1 mutation was loosely linked to B' , distal of glossy2 ( Figure S1 ) . Screening of a large mapping population ( Materials and Methods ) further located the Mop2-1 mutation to the 3 . 6 cM ( 13 BACs ) interval spanning FPC Contigs 69 and 70 ( Release 3b . 50 , February , 2009 ) ( Figure 2A ) . Using additional molecular markers , the interval was further reduced to 1 . 5 cM , which consisted of two BAC clones ( ∼400 kb ) on FPC Contig 69 ( Figure 2B ) . Analysis of putative genes in this interval revealed a strong candidate , a nrpd2/e2 gene , closely related to the Arabidopsis NRPD2/NRPE2 gene encoding the second largest subunit of the Pol-IV and Pol-V plant specific RNA polymerases . Arabidopsis NRPD2/NRPE2 is involved in regulating several epigenetic gene silencing phenomena [32] , [33] , [41] , [42] . Sequencing of the nrpd2/e2 gene from this interval in the Mop2-1 mutant revealed a transition mutation of guanine to adenine ( G to A ) relative to the progenitor allele , consistent with an EMS-induced mutation ( Figure 2C ) . This change in DNA sequence led to a missense mutation of glutamic acid to lysine ( E1079K ) , within the GEME motif , which is absolutely conserved [56] in Pol-I , Pol-II , Pol-III , and Pol-IV/Pol-V related polymerases from E . coli to higher eukaryotes ( Figure 3A ) . The high conservation of the mutated residue strongly suggested that this change in Mop2-1 would produce a mutant phenotype . The hypothesis that mop2 is a nrpd2/e2 gene was supported by the isolation of a second allele of mop2 from an independent screen ( Materials and Methods ) . The second allele , designated mop2-2 , carries a G to A transition mutation relative to its progenitor , which is consistent with an EMS induced mutation . This mutation changes a glycine to arginine , ( G1026R , Figure 2C ) within another highly conserved domain ( Figure 3B ) , distinct from that mutated in Mop2-1 . Co-segregation analysis revealed that all plants homozygous for the mop2-2 lesion had a dark plant phenotype consistent with the hypothesis that mop2-2 disrupts B' silencing as a homozygote , similar to Mop2-1 . Further experiments demonstrate that unlike Mop2-1 , mop2-2 is a recessive mutation as the establishment of paramutation is not prevented in heterozygotes ( data not shown ) . Two mutations isolated in Arabidopsis NRPD2/E2 are in the same domains as Mop2-1 and mop2-2 ( Figure 3A ) , but as only homozygous phenotypes are reported [41] , it is not clear if the Arabidopsis mutations also have dominant or semi-dominant phenotypes . BLAST searches of the maize genome revealed that maize encodes three nrpd2/e2 genes . In addition to the mop2 nrpd2/e2 gene , designated nrpd2/e2a , located on chromosome 2S , there are two genes on chromosome 10: nrpd2/e2b on 10L , FPC Contig 418 ( 94% identity and 97% similarity to nrpd2/e2a ) ; and nrpd2/e2c on 10S , FPC Contig 401 ( 67% identity and 79% similarity to nrpd2/e2a ) . Phylogenetic analysis demonstrated that the nrpd2/e2a and nrpd2/e2b genes are more similar to the presumed rice ortholog OsNrpd2a , while maize nrpd2/e2c is more similar to the other rice gene OsNrpd2b ( Figure 3C and 3D ) . The more similar genes , nrpd2/e2a and nrpd2/e2b are located in recently duplicated blocks within the maize genome , while the more diverged nrpd2/e2c gene is located in a more anciently duplicated block [57] . High conservation within all of the critical polymerase domains ( Figure S3 ) suggested that all three nrpd2/e2 genes are likely to encode functional proteins . BLAST analysis indicated that nrpd2/e2a and nrpd2/e2b have multiple EST hits , whereas nrpd2/e2c had no significant EST hits in current databases . Lack of nrpd2/e2c ESTs could be either because of low expression , or because it is expressed in tissues under represented in the public EST datasets . To further explore the expression of all three genes , we carried out quantitative RT-PCR experiments using gene-specific primers . We detected expression of all nrpd2/e2 genes in a wide range of tissues , but there was quantitative variation among the genes ( Figure 4 ) . For all three genes , the highest expression was in immature tassel and the lowest expression was in endosperm . The expression of nrpd2/e2c was more elevated in pollen and two callus samples ( HiII and BMS ) relative to nrpd2/e2a and nrpd2/e2b genes . Taken together these results demonstrate that all three maize nrpd2/e2 genes are likely to be functional , in contrast with Arabidopsis where only one functional nrpd2/e2 gene exists . Mutations in the largest and second largest subunit of the Arabidopsis Pol-IV RNA polymerase cause dramatically reduced siRNA production [32] , [58] , [59] . Similarly , mutations in rmr6 , which encodes the maize large subunit most similar to NRPD1 within the Pol-IV complex in Arabidopsis , show a dramatic reduction in siRNAs [31] . To determine if Mop2-1 might reduce the function of a Pol-IV-like complex in maize , siRNA levels in Mop2-1 were tested both globally and from the tandem repeats that mediate b1 paramutation . The small RNA fraction was isolated from immature ears , which are a rich source of RNA , separated on gels , and stained with SyberGold . Staining revealed that global siRNA levels were dramatically reduced in both heterozygous ( Mop2-1/+ ) and homozygous ( Mop2-1/Mop2-1 ) samples ( Figure 5A ) , consistent with the dominant phenotype of Mop2-1 . We next asked whether levels of siRNAs from the 853 bp tandem repeats ( Arteaga-Vazquez et al . , in preparation ) mediating b1 paramutation [11] were altered in Mop2-1 . In the wild type ( +/+ ) background , two siRNA bands ( prominent ∼25 nt and faint ∼35 nt ) were detected in the B' allele ( Figure 5B ) , which carries seven tandem 853 bp repeats and causes paramutation . In contrast , in the B' Mop2-1 samples there was a dramatic reduction of the 24 nt band , while the ∼35 nt siRNAs appeared to increase ( Figure 5B ) . Consistent with the dominant phenotypes that occur in Mop2-1 , reduced levels of siRNAs were seen in both heterozygous and homozygous Mop2-1 individuals , although there was more reduction in 24 nt siRNAs in homozygotes . Future experiments to examine whether the reduction of 24 nt siRNAs in Mop2-1 is associated with reduced asymmetric ( CHH ) DNA methylation , as observed in Arabidopsis will be important to carry out . Small RNAs larger than 24 nt have been reported in multiple species . In the ciliate protozoan , Tetrahymena thermophila , 27–30 nt RNAs direct developmentally directed DNA elimination [60] , in mammals and zebrafish 26–31 nt PIWI-interacting RNAs are present in the germline [61] , [62] , and in Drosophila repeat associated RNAs direct retrotransposon and repetitive sequence silencing [63] . In Arabidopsis , the role of a specific 30 nt siRNA reported for Flowering Locus C is unknown [64] and 30–40 nt siRNAs are induced in response to pathogen infection or under specific growth conditions [65] . In our studies , the ∼35 nt b1 tandem repeat RNAs were only observed when using the VC1658 LNA probe , one of four LNA probes for the b1 tandem repeats that we have used ( data not shown ) . One possibility is that transcripts from the LNA-VC1658 region are stable enough to detect alternative processing that increases when the predominant 24 nt pathway is disrupted . Further studies will be required to determine if the presence of the ∼35 nt siRNA class is significant . The reduction in tandem repeat 24 nt siRNA levels in Mop2-1 plants could theoretically be because Mop2-1 is causing a reduction in transcription of the tandem repeats or a defect in processing . Previously , we showed that the 853 bp tandem repeats that mediate b1 paramutation are transcribed [22] . To test whether transcription from the 853 bp repeats was altered in Mop2-1 , we conducted nuclear run-on analyses from nuclei isolated from young ears , the same tissue used for siRNA analyses . The results presented in Figure 5D revealed no significant differences in transcription from the 853 bp repeats between wild type ( + B' ) and homozygous Mop2-1 B' samples . This result indicated that Mop2-1 did not disrupt transcription from the b1 tandem repeats and suggested that the lack of 24 nt siRNAs is caused by a defect downstream of transcription . The dramatic reduction in 24 nt siRNAs is consistent with mutations in the large subunit of the Pol-IV complex in Arabidopsis and maize supporting the hypothesis that the Mop2-1 mutation is disrupting the function of a Pol-IV-like complex in maize . Our observations that there is no change in the transcription from the b1 tandem repeats in Mop2-1 mutants , further suggests that the major polymerase ( s ) responsible for this transcription is unlikely to be Pol-IV . Consistent with that hypothesis , other experiments suggest that the polymerase responsible for the bulk of the b1 tandem repeat transcription is likely to be Pol-II as transcription is dramatically reduced with levels of alpha-amanitin that inhibit Pol-II ( see Discussion ) . Paramutation has been well characterized at three other maize loci: pl1 , p1 and r1 , all encoding transcription factors that activate pigment synthesis [4] , [7] , [9] , [66] . We were interested in determining whether Mop2-1 disrupted paramutation at these loci and , if so , whether disruption was similar to b1 paramutation , i . e . , when heterozygous Mop2-1 prevented paramutation and when homozygous it increased the expression of silenced alleles . Table 1 summarizes the alleles for each gene used in each experiment and the results . For each of these paramutation systems , genetic backgrounds were available that enabled the monitoring of paramutation through changes in pigment levels ( Materials and Methods ) . To investigate paramutation at each locus , Mop2-1 was introduced into the appropriate genetic background ( Figure 6 , Table 2 , Figure S4 , and Figure S5 ) and pigment was monitored in the appropriate tissues . The results at the pl1 locus resembled b1 paramutation; Mop2-1 was dominant for preventing pl1 paramutation ( Figure 6A ) and it relieved the silencing of the paramutated Pl' allele ( Figure 6B ) . However , the levels of increased expression of the paramutant Pl' allele were variable suggesting Mop2-1 was partially dominant ( Figure 6B ) . In contrast to what was observed at the b1 and pl1 loci , Mop2-1 prevented p1 paramutation only when homozygous and there was no change on the expression of the silenced allele even after multiple generations of exposure to Mop2-1 ( Table 1 and Table 2 ) . At the third locus tested , r1 , Mop2-1 was semi-dominant for preventing paramutation ( Figure 6C ) . These results demonstrated that the Mop2-1 mutation differentially alters paramutation at these loci . Potential reasons for this are explored in the Discussion . Roles for the Arabidopsis Pol-IV and Pol-V polymerase complexes in transcriptional silencing mediated by siRNAs generated from the expression of promoters in inverted repeats ( pIR ) are well documented [67] , [68] . To test whether nrpd2/e2a might be involved in a similar transgene-mediated silencing system in maize , we tested whether the Mop2-1 mutation could prevent the silencing of two pIR-targeted loci ( Figure 7A ) , each required for male fertility [67] . One of these genes , Ms45 is expressed in the anther tapetum during early vacuolate stage of pollen development and is required for microspore development [67] . In wild type backgrounds , targeting the Ms45 gene by the Ms45Δ1pIR inverted repeat transgene results in complete sterility; 100% of tassels do not extrude anthers and they fail to produce any pollen [67] . To test whether Mop2-1 disrupts silencing , transgenic Ms45Δ1pIR/- plants were used as females and pollinated with the homozygous Mop2-1 B' stock ( Figure S6 ) . In the first generation , all plants were heterozygous ( Mop2-1/+ ) ; if Mop2-1 disrupted this process as a dominant or semi-dominant mutation , then full or partial restoration of male fertility would be expected . Examination of the Mop2-1/+ plants revealed that although all plants remained male sterile , small shriveled anthers developed in many of the plants ( data not shown ) . Because Ms45Δ1pIR leads to complete absence of anthers in wild type plants , improved anther development in Mop2-1/+ plants was a significant finding , and suggested Mop2-1 was semi-dominant . To test whether homozygous Mop2-1 plants would show a more dramatic relief of Ms45 silencing , transgenic plants were pollinated with the Mop2-1 B' stock ( Figure S6 ) . The resulting herbicide resistant plants segregating heterozygous and homozygous Mop2-1 individuals were examined . Because the B' allele was introduced together with the Mop2-1 mutation , dark plant pigmentation was used to initially identify Mop2-1 homozygous plants in the segregating families . If Mop2-1/Mop2-1 prevents pIR transgene-induced silencing of Ms45 , dark plants would be expected to exhibit partial or complete restoration of male fertility . This expectation was met as 72% of the dark plants were fertile , 19% had small anthers that did not shed pollen ( referred to as breakers ) , and only 9% were sterile ( Figure 7B ) . These results indicated that Ms45Δ1pIR-induced silencing of the Ms45 gene was disrupted in the majority of the darkly pigmented plants , likely representing Mop2-1/Mop2-1 homozygotes . The majority of light B' plants ( 60% ) , mostly representing Mop2-1/+ , were completely sterile , but a few were fertile ( 9% ) or had the breaker phenotype of extruded sterile anthers ( 31% ) ( Figure 7B ) . The detection of fertile plants among light B' plants could result from cumulative effects of carrying Mop2-1/+ for two generations , or incomplete correspondence between the Mop2-1 genotype and release of B' silencing . Both explanations are likely to be occurring because molecular genotyping revealed that 28/30 dark plants were homozygous and 28/33 of light plants were heterozygous for Mop2-1 . Altogether , these results demonstrated that the Mop2-1 mutation can prevent the Ms45Δ1pIR transgene from silencing the endogenous Ms45 gene in a semi-dominant manner , with the strongest phenotypes observed in homozygotes . A second pIR-targeted locus was tested for whether Mop2-1 could disrupt silencing . The 5126 gene has also been demonstrated to be expressed during microsporogenesis [67] . The 5126pIR-induced silencing of the 5126 gene results in complete male sterility , but a portion of tassels do show the breaker phenotype , i . e . , 30% of the flowers on a tassel extrude small shrunken anthers that do not contain pollen [67] . Similar to the results with the Ms45Δ1pIR/− transgenes , some improvement of anther development of the 5126pIR/− transgenic plants was noted in the first generation of plants heterozygous for Mop2-1/+; unlike wild type backgrounds some plants extruded anthers that produced small amounts of pollen ( unpublished data ) . After a backcross with the homozygous Mop2-1 stock ( Figure S6 ) , families segregating Mop2-1/Mop2-1 and Mop2-1/+ progeny were examined . Almost half of the darkly pigmented Mop2-1 homozygous plants had fertile tassels ( 48% ) , while the other half had the breaker phenotype ( Figure 7C ) . There were fertile plants ( 19% ) among light B' Mop2-1/+ heterozygous plants , although most showed the breaker phenotype ( 81% ) . These results indicate that the Mop2-1 mutation can disrupt pIR silencing at the 5126 locus in a semi-dominant manner , with increased activity as a homozygote . We also tested whether Mop2-1 disrupted pIR-mediated silencing of a third transgene , pg47pIR , which targets the pg47 locus , a gene that is highly expressed during pollen development [68] , [69] . In contrast to the results with Ms45Δ1pIR and 5126pIR , the Mop2-1 mutation did not reverse pg47pIR-mediated silencing of the pg47 gene family ( Cigan M , unpublished data ) . The pg47 gene is expressed in pollen , a tissue where nrpd2/e2a is expressed at a very low level and nrpd2/e2c is expressed highly ( Figure 4 ) . Thus , we favor the hypothesis that nrpd2/e2a is not used extensively in the pollen , and therefore a mutation in it has no phenotype in this tissue . However , an alternative explanation that silencing induced by different pIR transgenes may use distinct mechanisms can not be eliminated . Taken together , these results suggest that a mutation in nrpd2/e2a disrupts pIR-mediated silencing , but not at all loci . In Arabidopsis , Pol-IV is required for siRNA production , whereas Pol-V primarily acts downstream of siRNA production generating non-coding transcripts and helping to direct chromatin modifying enzymes to target loci [41] , [42] , [45] , [70] . To gain insight into possible mechanisms by which the Mop2-1 mutation disrupts pIR-mediated silencing in maize , we assayed transcript levels of the Ms45Δ1pIR and 5126pIR transgenes and the targeted endogenous genes in Mop2-1 plants . Results demonstrated that endogenous Ms45 and 5126 transcripts were absent in sterile +/+ or Mop2-1/+ plants consistent with silencing and that they were present in fertile Mop2-1/Mop2-1 plants ( upper panels in Figure 7D and 7E ) , consistent with Mop2-1 preventing silencing . However , the levels of the endogenous genes' transcripts in Mop2-1/Mop2-1 plants were not as high as in non-transgenic plants ( upper panels in Figure 7D and 7E ) , indicating that partial silencing was still occurring even in homozygous Mop2-1 plants , although clearly enough transcription of the endogenous genes was occurring to restore fertility . Hybridization with the Nos spacer fragment probe [67] , which is unique to the Ms45pΔ1IR and 5126pIR transgenes , revealed that transgene transcript levels , while variable , were similar between sterile ( +/+ or Mop2-1/+ ) and fertile ( Mop2-1/Mop2-1 ) plants ( Figure 7D and 7E ) . Thus , Mop2-1 is not acting to reduce the pool of transgenic transcripts that serve as precursors for siRNA production and silencing . Because the Mop2-1 mutation reduced levels of total endogenous siRNAs and the b1 853 bp repeat specific siRNAs , we tested whether Mop2-1 might also dramatically reduce the siRNAs produced from the Ms45pΔ1IR and 5126pIR transgenes . Because we did not have wild type transgenic lines that were isogenic with the Mop2-1 heterozygotes and homozygotes , we compared the transgene siRNA levels in sterile and breaker heterozygous Mop2-1/+ plants to those in fertile homozygous Mop2-1/Mop2-1 plants . If the ability of Mop2-1 to relieve silencing was due solely to reduced siRNA biogenesis , we would expect few transgenic siRNAs to be seen in fertile homozygotes . Results in Figure 7F demonstrated that 24 nt siRNAs are produced at similar levels from the Ms45Δ1pIR transgene in both fertile Mop2-1 homozygous plants and sterile Mop2-1/+ heterozygous plants . In the 5126pIR transgenic plants transgene siRNAs were also detected in both breaker Mop2-1 heterozygotes and fertile homozygotes , although two out of three Mop2-1 heterozygotes had approximately two fold higher siRNA levels relative to homozygotes ( Figure 7G ) . These results demonstrated that the ability of the Mop2-1 mutation to restore male fertility in the Ms45pΔ1IR and 5126pIR transgenic plants was not simply because it eliminated transgene siRNAs . One possibility is that Mop2-1 can act downstream of siRNA production to relieve sterility . As Pol-V in Arabidopsis acts downstream of siRNA biogenesis , this hypothesis is consistent with Mop2-1 also disrupting a Pol-V-like complex in maize . Alternative explanations are presented in the Discussion . Mutations in Arabidopsis Pol-IV/Pol-V complexes have not been reported to display major developmental phenotypes , except for a delay in flowering [32] , [41] . In contrast , Mop2-1 displays a number of developmental phenotypes , but these are less dramatic and more variable than the developmental phenotypes observed with mutations in the rmr6 gene [21] , which encodes the large subunit of a Pol-IV like complex in maize [31] . When propagating the Mop2-1 mutation through numerous generations , a number of phenotypes were routinely observed including reduced transmission , altered flowering time and abnormal developmental phenotypes . This was similar to phenotypes of mop1 mutations [20] , but different from rmr6 mutations , as the aberrant phenotypes were not observed in every plant carrying the Mop2-1or mop1 mutations , but were observed in all plants homozygous for rmr6 mutations [21] . The types of abnormal morphological and development phenotypes that we observed in stocks segregating Mop2-1 were reduced plant height , skinny plant stature , tassel seed , failure to develop an ear , and poor seed set in ears that did develop . These phenotypes are variable and they occur in both heterozygous or homozygous Mop2-1 plants , but they are more frequent and more severe in homozygous Mop2-1/Mop2-1 plants . For example , in one experiment Mop2-1/Mop2-1 plants were on average 12 cm shorter and flowered 4 . 2 days later then heterozygous siblings , and 22% of Mop2-1/Mop2-1 versus 10% of Mop2-1/+ siblings failed to differentiate ears . These negative pleiotropic phenotypes on plant health and reproduction influenced how the stocks could be maintained such that self pollinations were rarely successful and large numbers of crosses were required to obtain sufficient numbers of homozygous Mop2-1 plants with mature ears and reasonable seed set for our experiments . To determine the transmission of Mop2-1 , 231 plants were genotyped for the Mop2-1 mutation from a family that would be expected to segregate equal numbers of Mop2-1 homozygous and heterozygous plants if there was no reduction in Mop2-1 transmission . This experiment revealed that the number of Mop2-1 homozygotes was reduced ( 39% ) ( χ = 211 . 2 , P = 0 . 0001 ) . To test whether this was due to reduced transmission or reduced germination frequency , we genotyped 94 seeds directly and observed a similar reduced number of Mop2-1 homozygotes , only 39% instead of the 50% expected for normal transmission . This strongly suggested reduced transmission was contributing to the reduced number of Mop2-1 homozygotes . The observation of developmental phenotypes with Mop2-1 differs from the recessive loss-of-function truncation mutations , which lack developmental phenotypes ( Stonaker et al . , this issue ) . This is not simply because Mop2-1 is dominant and the mutations isolated in the Hollick lab are recessive as we also see developmental phenotypes with our recessive mop2-2 allele . Although we haven't grown mop2-2 plants for as many generations as Mop2-1 plants , homozygous mop2-2 plants are very skinny and rarely set seed . One possibility is that our two missense mutations are having broader effects relative to the null mutations isolated in the Hollick lab . Another possibility is that the genetic background of the null mutants is suppressing developmental phenotypes , or the genetic background of our mutants is enhancing developmental phenotypes . It has long been known that different genetic backgrounds can enhance or suppress developmental phenotypes in maize [71] . A third possibility is that the more extreme environmental growth conditions in Arizona relative to Northern California are enhancing the developmental phenotypes .
Our results demonstrate that mop2 , a key gene involved in paramutation at multiple loci , encodes a protein closely related to the second largest subunit of the plant-specific RNA polymerase complexes , Pol-IV/Pol-V , first described in Arabidopsis . Unlike Arabidopsis , which encodes only one functional protein ( NRPD2/E2 ) , which is in both the Pol-IV and Pol-V complexes [32] , [41] , [42] , [58] , maize encodes three closely related genes . These three genes are likely to encode functional proteins as they are full length , have all of the polymerase domains conserved and are expressed in multiple tissues . This observation suggests that multiple Pol-IV/Pol-V-like complexes and potentially even novel complexes may exist in maize . Potentially these could have diversified to function at different loci , different developmental stages , or under different environmental stresses . Multiple Pol-IV/Pol-V related complexes could confer greater complexity and epigenetic regulatory capacity to maize as compared to Arabidopsis . Our expression analyzes revealed that all three maize genes are widely expressed in multiple organs and tissues , similar to that of NRPD2/E2 in Arabidopsis [42] , [70] . All three maize genes are most highly expressed in maize reproductive organs such as tassels and immature ears . The most diverged gene , nrpd2/e2c , is the major gene expressed in pollen and it is also more highly expressed in callus relative to the other two genes . Given this difference in expression and that the phylogenetic analyses suggested nrpd2/e2c represents a more ancient duplication , it is the most likely candidate for a distinct function relative to nrpd2/e2a , b . It is striking that the two most similar genes , nrpd2/e2a and nrpd2/e2b share similar expression patterns across all tissues tested , yet our results demonstrate that a recessive mutation in nrpd2/e2a , mop2-2 , has paramutation defects . This demonstrates that at least with respect to paramutation , the nrpd2/e2a gene is unlikely to be functionally redundant with nrpd2/e2b . This hypothesis is supported by studies from the Hollick lab in which recessive mutations in the same gene were isolated in forward genetic screens for pl1 paramutation ( Stonaker et al , this issue ) . These three genes may not be functionally redundant either because they function in distinct complexes or because there are differences in their expression patterns not detectable with quantitative RT-PCR in complex multi-cellular tissues . Expression differences among different cell types within organs , distinct cell layers within the same tissue , differences in subcellular location , or even locus specific distribution are all possible explanations for the lack of functional redundancy . Given the high degree of similarity among all three genes , the generation of transgenic plants with differentially tagged genes as well as mutations in the other two subunits will enable these hypotheses to be distinguished . The loss of siRNAs from repetitive elements throughout the genome and from the tandem repeats that are required for b1 paramutation is consistent with Mop2-1 disrupting a Pol-IV like complex . When siRNAs are produced independently of Pol-IV as they are in many pIR transgene experiments [41] , i . e . potentially in our case when the strong Pol-II promoter from the maize ubiquitin1 gene is used to produce inverted repeat transcripts , one can also examine the potential for Mop2-1 to reduce Pol-V like function , downstream of siRNA biogenesis . In unpublished data we have shown that the pIR transgenes we used do not require the endogenous siRNA biogenesis pathway for silencing , as the null mop1-1 mutation ( in the RDR orthologous to RDR2 ) , which dramatically reduces 24 nt siRNAs does not prevent pIR silencing ( M . Cigan and V . Chandler , unpublished data ) . In Mop2-1 homozygotes pIR-mediated silencing is relieved , yet there are similar levels of siRNAs from the pIR transgenes in fertile homozygotes relative to sterile heterozygotes . This result is consistent with Mop2-1 also acting downstream of siRNA biogenesis , potentially by disrupting a Pol-V like complex . However , it is possible that Mop2-1 may only be acting through a Pol-IV-like complex . For example , Mop2-1 may partially impair secondary siRNA production from the pIR transgenes , which might partially reduce transcriptional silencing resulting in fertility . This model can account for Mop2-1 effects entirely through Pol-IV deficiencies . If Mop2-1 does alter both Pol-IV and Pol-V functions , it could be because like Arabidopsis , the wild type nrpd2/e2a encoded subunit functions in both Pol-IV-like and Pol-V-like RNA polymerase complexes . Alternatively , the Mop2-1 mutation may confer a gain of function phenotype that enables the mutant subunit to interact with and disrupt complexes the wild type subunit normally does not form . The observation that loss of function mutations in nrpd2/e2a also cause a dramatic reduction in global siRNAs ( Stonaker et al , this issue ) , suggests that at a minimum NRPD2/E2a functions in a Pol-IV like complex . The Mop2-1 mutation's effects on paramutation at multiple loci , pIR-mediated silencing , and plant growth and development depend on the dosage of the Mop2-1 allele with some phenotypes more variable than others . Certain phenotypes were observed with high penetrance when Mop2-1 was heterozygous ( dominant for prevention of b1 and pl1 paramutation , reduction in global and b1 tandem repeat siRNAs ) ; some phenotypes required Mop2-1 to be homozygous ( recessive for release of B' silencing and prevention of p1 paramutation ) ; and still other phenotypes were seen in Mop2-1 heterozygotes , but were much stronger in Mop2-1 homozygotes ( semi-dominant for prevention of r1 paramutation , disrupting pIR-mediated silencing of Ms45 and 5126 , release of Pl' silencing , further reduction of b1 tandem repeat siRNAs , and many developmental phenotypes ) . There were also phenotypes that Mop2-1 did not alter ( no release of P1-rr' silencing and no disruption of pIR-mediated silencing of pg47 ) . Below we discuss a model for Mop2-1 dominance , based on data from similar mutations in the second largest subunit of E . coli RNA Polymerase , and suggest explanations for the dependence of specific phenotypes on Mop2-1 dosage . The GEME motif mutated in Mop2-1 is nearly invariant among all second largest subunits in all polymerases from organisms ranging from bacteria , fungi , animals , and plants . The Mop2-1 mutation changes the second glutamic acid residue ( E1076 ) of the GEME motif to a lysine . In RpoB ( the second largest subunit of E . coli RNA polymerase ) the GEME motif is located within the “anchor” region required for interaction with the clamp fold within the largest polymerase subunit [72] . The clamp swings open to produce a larger opening of the cleft that permits entry of promoter DNA and subsequent initiation of transcription . Extensive mutagenesis of the second largest subunit , rpoB in E . coli revealed that mutations within all four GEME motif amino acids result in dominant phenotypes when the mutant subunit is produced from a plasmid at similar levels to the chromosomal encoded non-mutant copy [56] . Substitutions in this region produced a RNA polymerase that competed with the wild type RNA polymerase complex potentially because the mutant polymerase was blocked after transcription initiation [56] , [73] . The strength of the dominant phenotypes varied depending on the specific substitution . The extreme conservation of the GEME motif [56] and the results in E . coli , lead us to hypothesize that a similar molecular mechanism contributes to the dominant phenotype of the Mop2-1 mutation . Our model is that NRPD2/E2a proteins with the Mop2-1 mutation associate normally with the same largest subunit ( s ) ( NRPD1 and potentially NRPE1 , or both ) that the wild type subunit associates with , forming mutant polymerase complexes that are functionally defective , but that efficiently compete with wild type polymerase complexes . This model predicts that the relative dosage of mutant and wild type subunits would influence the number of functional complexes available . Assuming that the Mop2-1 encoded protein is expressed equivalently to the wild type , a Mop2-1 heterozygote should have equal amounts of wild type and mutant proteins , such that phenotypes that are particularly sensitive to Pol-IV or Pol-V dosage would be altered in the heterozygote . Moreover , processes that require both Pol-IV and Pol-V complexes might show more dramatic phenotypes than processes that require either Pol-IV or Pol-V alone , if there are additive consequences of partial loss of each complex . This model further predicts that a Mop2-1 homozygote , which would have no wild type NRPD2/E2 , would produce a stronger phenotype relative to the heterozygote , potentially equivalent to a loss of function mutation . The presence in maize of two other second largest subunits , including NRPD2/E2b that is 94% identical to NRPD2/E2a , provides a further complication and could also contribute to different phenotypes between Mop2-1 heterozygotes and homozygotes if the Mop2-1 encoded NRPD2/E2a protein competes with the other subunits for complex binding . This could lead to gain of function phenotypes , in which the Mop2-1 encoded NRPD2/E2a subunit is partially poisoning complexes that normally carry the NRPD2/E2b , c subunits . Competition would be postulated to be most effective in homozygous Mop2-1 plants where the dosage of the mutant subunit is highest . As stated previously , detailed biochemical analyses with each of the subunits differentially tagged will be necessary to begin to distinguish among possible models . As b1 paramutation is the most extensively characterized system at the molecular level , we will limit our discussion to the b1 system . The Mop2-1 mutation prevents b1 paramutation when heterozygous and releases silencing of B' when homozygous indicating Pol-IV/Pol-V-like RNA polymerases are required for b1 paramutation . While the requirement for a Pol-IV-like polymerase for b1 paramutation was previously demonstrated in studies with a mutation in the large subunit most similar to Pol-IV in Arabidopsis [21] , [31] , our results provide further clarification on potential roles for Pol-IV/Pol-V-like complexes at distinct steps in paramutation . Mop2-1 reduces siRNA production , characteristic of a Pol-IV-like mutation , but does not reduce transcription from the b1 tandem repeats , suggesting that NRPD2/E2a containing RNA polymerase complexes do not significantly contribute to b1 repeat transcription . Transcription from the b1 repeats is sensitive to actinomycinD , indicating that these transcripts are produced from DNA templates ( Arteaga-Vazquez et al . , in preparation ) . In Arabidopsis , Pol-V has been shown to use DNA as a template to produce non-coding transcripts [45] . Transcripts produced from Pol-V in Arabidopsis are rare [45] , and if a similar situation exists in maize , differences between presumed Pol-V mediated transcription in wild type and homozygous Mop2-1 plants might be difficult to detect with nuclear run-ons , especially if most of the b1 repeat transcription is mediated by Pol-II . Transcription from the b1 repeats is highly sensitive to alpha-amanitin consistent with Pol-II being the best candidate for the polymerase performing the major transcription of the non-coding b1 repeats ( Arteaga-Vazquez et al . , in preparation ) . Further molecular and biochemical characterization of b1 repeat transcripts will determine whether these transcripts are polyadenylated , and chromatin immunoprecipitation assays with tagged Pol IV and Pol V complexes will determine if either physically interact with the b1 tandem repeats mediating paramutation . Our observation that silenced alleles are not up-regulated in Mop2-1 heterozygous plants , in spite of the dramatic reduction in b1 tandem repeat siRNAs , suggests that the major role for the tandem repeat siRNAs is not to maintain silencing . Instead , we hypothesize that the tandem repeat siRNAs may mediate the allele communication that establishes the heritable silent state; a process that is fully disrupted in Mop2-1 heterozygotes . Use of the Mop2-1 plants in future experiments may enable us to separate mechanisms operating at the initial establishment of paramutation from those operating subsequently to maintain silencing . Previously this has not been possible as establishment can only be observed if it is heritably maintained , so mutations defective in maintaining silencing will also appear defective in establishment . However , mutations that do not relieve silencing , but do prevent the establishment of paramutation such as Mop2-1 ( when heterozygous ) provide a system to investigate mechanisms for establishment .
The Mop2-1 mutation was generated by treating Gl2 B' Wt; Pl-Rh; r-g ( inbred W23 background ) pollen with ethyl methanesulfonate [74] . Treated pollen was used to pollinate gl2 B-I wt; Pl-Rh; r-g [W23/K55] ears , producing M1 seed ( Figure 1A ) . In wild-type stocks , B' will always paramutate B-I , and all progeny will be light . Thus , rare plants with B-I pigmentation levels may indicate the presence of a dominant mutation preventing paramutation . Presence of the recessive gl2 and wt markers on the B-I chromosome enabled rapid identification of self-pollination contaminant offspring . Pollen from the exceptional dark plant ( KK1238-1 ) was crossed onto gl2 B' wt ears ( Figure S1 ) . The mop2-2 mutation was isolated in an independent EMS screen in which B-I pollen was treated with EMS [74] and placed on silks of B' plants . Resulting F1 plants were screened for dominant mutation phenotypes ( none found ) and self pollinated , and F2 progeny were screened for dark plants . Initial linkage of Mop2-1 relative to the b1 , gl2 and wt loci was noticed in the test cross of the original Mop2-1 dark plant with + gl2 B' wt tester ( Figure S1 ) . Of the 15 progeny plants , 12 plants inherited parental 2S chromosomes from the original mutant plant , three plants exhibited phenotypes consistent with recombination , two between Mop2-1 and Gl2 , and one between B-I* and wt ( Figure S1 ) . This result indicated that the Mop2-1 mutation is distinct from b1 and located distal of Gl2 . The map position of the Mop2-1 mutation was subsequently confirmed and refined using simple sequence repeat ( SSR ) molecular markers ( www . maizegdb . org ) and a larger number of plants ( data not shown ) . Two markers tightly linked to Mop2-1 , bnlg1117 and bnlg1338 , were used to routinely follow Mop2-1 segregation prior to cloning . PCR products were resolved using 4% Super Fine Resolution agarose gels . To further refine the location of Mop2 , a large mapping population was generated by crossing homozygous Mop2-1 plants with B73 , the inbred sequenced for the maize genome project that is also highly polymorphic relative to the Mop2-1 stock . The F1 was then backcrossed to homozygous Mop2-1 , the resulting seed planted , and DNA was extracted from 1308 dark plants , the phenotype expected for homozygous Mop2-1 . The resulting samples were screened with polymorphic markers on chromosome 2S ( available upon request ) to first determine the boundaries of the Mop2-1 interval ( 13 BACs ) and then to further map it to within a two BAC interval on FPC contig 69 ( Figure 2 ) . Examination of gene models within the two BAC interval ( NCBI accessions AC1911113 . 2 and AC213986 . 2 ) revealed the presence of 21 putative protein encoding genes ( Release 3b . 50 , February , 2009 ) , including the AC191113 . 2_FGT037 gene model , which is predicted to encode the second largest subunit of a RNA polymerase most similar to rice gene SJNBa0063C18 . 1 ( OsNrpd2a ) . The gene model AC191113 . 2_FGT037 was refined using FGENESH+ and Arabidopsis NRPD2/NRPE2 and OsNrpd2a genes as guides . The refined model corresponded well to the NRPDB101 gene model from www . chromdb . org , and was experimentally verified by PCR amplification of the full length transcript and sequencing ( NCBI GQ453405 ) . The corrected AC191113 . 2_FGT037 gene model was renamed nrpd2/e2a . Please note that the order of fragments within the AC191113 in Figure 2 is different from that shown www . maizesequence . org at the time of publication . We reordered fragments based on additional sequence information that indicated positions of overlapping fragments within the AC191113 and neighboring BACs ( not shown ) . A custom BAC library was constructed from Mop2-1 genomic DNA with the assistance of the Arizona Genomic Institute ( Tucson , AZ ) . Nylon filters with printed DNA from this library were hybridized with a probe unique to the 3' UTR of nrpd2/e2a ( available upon request ) . One of the positive clones that contained the full length nrpd2/e2a gene was used as a template to PCR amplify all predicted exons ( primer sequences available upon request ) . The resulting PCR fragments were sequenced in both directions using the core sequencing facility at University of Arizona ( Tucson , AZ ) . Consistent with an EMS-induced mutation , a G to A transition was identified in the nrpd2/e2a gene , within an absolutely conserved motif in exon 7 . To identify additional mutations , we sequenced the exons of the nrpd2/e2a gene in three newly isolated EMS-induced b1 paramutation mutants . Exons of the nrpd2/e2a gene were PCR amplified and amplicons were sequenced in both directions for each candidate mutant . One of the new mutants was found to carry a G to A transition in a conserved domain in exon 6 , consistent with an EMS induced mutation . This mutation was named mop2-2 . The nrpd2/e2a gene was sequenced in a total of 12 dark plants ( mop2-2 homozygotes ) and all carried the same lesion , indicating that the lesion segregated with the mutant paramutation phenotype . Sequence alignment was carried out using MUSCLE [75] , manually edited in GENEDOC 2 . 6 . 04 . Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 4 software [76] . Bootstrap neighbor-joining method with 1000 replicates was used to generate the phylogenic trees . Protein sequences of the maize second largest subunits of Pol-I ( ZmNRPA2 ) and Pol-II ( ZmNRPB2a , ZmNRPB2b ) were predicted using FGENESH+ software and corresponding Arabidopsis proteins as guides ( Figure S7 ) . To obtain genomic sequence suitable for protein prediction of maize nrpd2/e2b , the gap in the AC212557 sequence was PCR amplified and sequenced . FGENESH+ ( http://linux1 . softberry . com/berry . phtml ) was used to generate the gene model , which is equivalent to GRMZM2G146935_T02 at www . maizesequence . org . The maize nrpd2/e2c gene model was predicted from the AC203335 . 4 BAC sequence using FGENESH+ and rice OsNrpd2b as a guide , which was equivalent to GRMZM2G133512_T01 at www . maizesequence . org . The quality of the resulting gene models was inspected using the ClustalX multiple sequence alignment program [77] . For the alignment shown in Figure S3 and the phylogenetic tree shown in Figure 3C and 3D , sequences of the second largest subunits were provided by the Pikaard lab or retrieved from NCBI . The complete list of the genes used for the phylogenetic analysis is in Table S1 . For expression analysis of the maize nrpd2/e2 genes total RNA was extracted from tissues flash frozen in liquid nitrogen using the Trizol protocol as described by manufacturer ( Invitrogen ) . Total RNA was treated with DNAseI ( Invitrogen ) and acid phenol ( Ambion ) to remove DNA contamination . First stand synthesis was carried out using oligo ( dT ) primers and 10 ug of total RNA . Superscript III First Strand Synthesis System ( Invitrogen ) was used according to manufacturer's recommendations at 55 C for 1 hour . About 200 ng of cDNA were used for each quantitative PCR assay on the Bio-Rad MyIQ Real-Time PCR machine and quantified using My-IQ software ( Bio-Rad ) . Expression of nrpd2/e2 genes was normalized to actin1 expression . Sequences of gene-specific primers are available upon request . All plant materials were from the B73 inbred with the exception of the Black Mexican Sweet ( BMS ) tissue culture cells and the HiII callus cells . All tissues were collected in the morning , 2–4 hours after sunrise . At the time of collection all plant tissues were immediately frozen in liquid nitrogen . Seedlings were germinated for 6 days after imbibing in paper towels at 16 day/8 night photo period at 20C and tissue before the first leaf emerged from the coleoptile . Root tissue was from seedlings grown in a vermiculite/soil mixture , 10 days after emergence ( 2–3 leaf stage ) . Seedling roots were washed before freezing . Husk , silks , immature ears , immature tassels , and pollen were collected from field grown plants . Husk and silks were collected simultaneously on the first day of silk emergence . Immature ears and tassels were collected when they were 1 . 5 cm to 2 . 5 cm in length . Pollen was collected in the morning from tassels on the first day of shedding . Pollen was filtered through fine metal mesh filters to remove debris before freezing . Endosperm and embryos were collected from greenhouse grown plants at 12 and 14 days after pollination ( DAP ) . HiIIAxB type II embryogenic callus was maintained on N6 media with 1 . 0 mg/L of 2 , 4-dichlorophenoxyacetic acid with sub-culturing every two weeks as described [78] . The BMS callus suspension culture used in these experiments was acquired from C . Armstrong ( Monsanto Company , St . Louis , MO ) in 2001 . BMS cultures were maintained in N6 media supplemented with 1 . 5 mg/L of 2 , 4-dichlorophenoxyacetic acid with monthly subcultures to a fresh media . Both the HiII and BMS cultures were cultured in the dark at 26–28°C . The gl2 b wt , Pl , r-g ( inbred K55 background ) , the B-I Pl r-g ( inbred W23 background ) and the B' Pl r-g ( inbred K55 background ) stocks were originally obtained from E . H . Coe , Jr . ( University of Missouri , Columbia ) . Paramutagenic Pl' allele and paramutable Pl-Rhoades ( Pl-Rh ) were previously described [3] . The phenotypes and paramutation properties of the R-st , R-r and r alleles were previously described [79] , [80] . Paramutation at r1 is typically assayed in the aleurone , the outer cell layer of the endosperm , where purple anthocyanin pigments accumulate in the highly expressed R-r allele; reduced pigmentation is observed when R-r is paramutated to R-r' by R-st . The fully colored R-sc allele [79] , [81] was used as a positive seed color control ( Figure 6C ) . Both heterozygous and homozygous Mop2-1 plants were assayed for effects on r1 paramutation ( Figure S5 ) , although only a small number of Mop2-1 homozygotes could be assayed because these plants have reduced fertility . To quantify changes in seed color occurring during paramutation , the relative color was determined as described [82] . The paramutable P1-rr stock , the standard P1-rr4B2 allele of the p1 gene [83] , has red pericarp and red cob pigmentation , while the silenced P1-rr' allele has lightly patterned or colorless pericarp and pink cob pigment [4] . The highly paramutagenic P1-rr'/P1-rr'; P1 . 2b::GUS/- stock was used to assess Mop2-1 effects on p1 paramutation , by crossing with a Mop2-1 B' P1-rr stock . The P1 . 2b::GUS transgene ( transgenic event P2P147-37 ) carried the P1 . 2 enhancer fragment that is sufficient for paramutation , fused to the basal P1-rr promoter , the Adh1 ( maize Alcohol dehydrogenase1 gene intron 1 ) , the E . coli GUS gene and the PinII ( potato Proteinase InhibitorII ) 3' , along with a resistance gene for the BASTA herbicide [13] . Details of the crosses are presented in Figure S4 . The Ms45Δ1pIR and 5126pIR transgenic constructs and phenotypes upon silencing were previously described [67] . To assay whether the Mop2-1 mutation could disrupt pIR transgene induced silencing of the endogenous Ms45 and 5126 genes , Mop2-1 plants were crossed with four independent transgenic events each for Ms45Δ1pIR and 5126pIR . Herbicide resistant F1 plants were backcrossed to the Mop2-1 stock to generate a family segregating heterozygous or homozygous Mop2-1 plants . These were grown in the field and sprayed with herbicide to remove non transgenic plants , while the remaining transgenic plants were visually scored for plant color and male fertility at anthesis . Plants homozygous for the Mop2-1 mutations were initially identified using the dark plant color associated with increased expression of the B' allele . For a subset of the Ms45Δ1pIR plants , the presence of the Mop2-1 mutation was confirmed by genotyping with molecular markers . Northern blot analysis of Ms45 and 5126 transcript levels were as previously described [67] , [84] . Conditions for small RNA Northern blots were the same as described below , using previously described probes for MS45 and 5126 transgene siRNA detection [67] . RNA was extracted from 3 g of immature ears ( 3–5 cm long ) using TRIZOL reagent ( Invitrogen ) , and the large RNA fraction was precipitated using 5% polyethyleneglycol MW 8000 [85] . The aqueous phase , enriched for the small RNA fraction , was subjected to phenol:chlorophorm:isoamyl-alcohol ( 24∶1∶1 ) extraction followed by ethanol precipitation . The pellet was resuspended in DEPC treated water . Approximately 100 ug of the small RNA fraction was loaded in each lane . RNA was electrophoresed on 15% denaturing UREA polyacrylamide gels , electroblotted onto GeneScreen Nylon membrane and immobilized using UV crosslinking . Blots were hybridized with a 32P end labeled DNA:LNA ( DNA::Locking Nucleic Acid ) oligonucleotide [86] . The DNA::LNA oligo ( vc1658F , TGAA+CATCTT+GTCCA+GTTAAAT+CACTGG+ACACC+GTGAC+AGCC+ACA; “+” precedes an LNA base ) was synthesized by Sigma-Proligo . For the U6 probe , DNA oligo ( vc1969F , AGACATCCGATAAAATTGGAACGATACAGA ) was end labeled with 32P . Hybridization and image was processed using QuantityOne software ( BioRad ) . Approximately ∼5 g of immature ears were used to extract nuclei as described [20] . The nuclei isolations and run on reactions were as described [22] . To prepare the b1 tandem repeat probes , PCR fragments carrying T3 promoter tails were used as templates for in vitro transcription with T3 RNA polymerase ( Invitrogen ) , as recommended by the manufacturer . Sequences of the primers used to produce the b1 RNA probes are available upon request . The positive control , the Ubiqutin2 RNA probe , was as described [27] . Lambda phage genomic DNA , 100 ng per slot , was used as a negative control . | How an individual's genes are activated or silenced is an essential question impacting all fields of biology . Usually gene expression patterns , i . e . , which genes are on and which are off in different tissues and during development , are highly reproducible; and those patterns are efficiently reset in the next generation of progeny . Paramutation represents an exception to these genetic rules , in that for certain genes the silencing that is established in an individual is efficiently transmitted to their progeny . Importantly , in these subsequent generations , the silenced gene continues to silence active versions of that gene . Prior work has demonstrated that these heritable gene expression changes are not accompanied by changes in DNA sequence: they are epigenetic . Understanding mechanisms for heritable changes in gene expression has major implications for researchers studying complex traits , including diseases . In this manuscript we demonstrate that a subunit of a RNA polymerase is required for paramutation in maize and other gene silencing processes that also involve RNA–mediated chromatin changes . We show that the multiple , closely related , plant-specific RNA polymerases mediating gene silencing have diverged functions in maize . Results from our experiments suggest testable models for the role of these polymerases in multiple gene-silencing processes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/epigenetics"
] | 2009 | A Dominant Mutation in mediator of paramutation2, One of Three Second-Largest Subunits of a Plant-Specific RNA Polymerase, Disrupts Multiple siRNA Silencing Processes |
Over 200 million people have , and another 600 million are at risk of contracting , schistosomiasis , one of the major neglected tropical diseases . Transmission of this infection , which is caused by helminth parasites of the genus Schistosoma , depends upon the release of parasite eggs from the human host . However , approximately 50% of eggs produced by schistosomes fail to reach the external environment , but instead become trapped in host tissues where pathological changes caused by the immune responses to secreted egg antigens precipitate disease . Despite the central importance of egg production in transmission and disease , relatively little is understood of the molecular processes underlying the development of this key life stage in schistosomes . Here , we describe a novel parasite-encoded TGF-β superfamily member , Schistosoma mansoni Inhibin/Activin ( SmInAct ) , which is key to this process . In situ hybridization localizes SmInAct expression to the reproductive tissues of the adult female , and real-time RT-PCR analyses indicate that SmInAct is abundantly expressed in ovipositing females and the eggs they produce . Based on real-time RT-PCR analyses , SmInAct transcription continues , albeit at a reduced level , both in adult worms isolated from single-sex infections , where reproduction is absent , and in parasites from IL-7R−/− mice , in which viable egg production is severely compromised . Nevertheless , Western analyses demonstrate that SmInAct protein is undetectable in parasites from single-sex infections and from infections of IL-7R−/− mice , suggesting that SmInAct expression is tightly linked to the reproductive potential of the worms . A crucial role for SmInAct in successful embryogenesis is indicated by the finding that RNA interference–mediated knockdown of SmInAct expression in eggs aborts their development . Our results demonstrate that TGF-β signaling plays a major role in the embryogenesis of a metazoan parasite , and have implications for the development of new strategies for the treatment and prevention of an important and neglected human disease .
Amongst the Bilateria , transforming growth factor–β ( TGF-β ) signaling is recognized as playing an essential role in embryogenesis in deuterostomes and in arthropod protostomes , but its role in lophotrochozoan protostomes is unclear [1] . Schistosomes , the causative agents of schistosomiasis , one of the major neglected tropical diseases [2 , 3] , are metazoan parasites that belong to the lophotrochozoan phylum Platyhelminthes . Components of TGF-β signaling have been molecularly characterized in metazoans throughout the animal kingdom . Activation of this pathway begins at the cell surface when a dimeric ligand binds a complex consisting of types I and II receptor serine/threonine kinases [4] . Upon ligand binding , the constitutively active type II receptor phosphorylates and activates the type I receptor , which then phosphorylates cytoplasmic Smad proteins that translocate to the nucleus , where they mediate gene expression [4] . Components of a functional TGF-β pathway ( s ) , including one type I receptor [5] ( Schistosoma mansoni receptor kinase-1 [SmRK1] , S . mansoni transforming growth factor–β type I receptor [SmTβ RI] ) , one type II receptor [6 , 7] ( SmRK2 , SmTβ RII ) , and three Smads [8–10] , have been identified in S . mansoni , with nearly all components localized to either the surface of the worm or reproductive tissues of the female [5–9 , 11] . Nevertheless , while nearly the entire transcriptome of S . mansoni has been examined with the identification of 163 , 000 expressed sequence tags ( ESTs ) [12] , a ligand of parasite origin for the TGF-β pathway ( s ) has remained elusive . This has led to the hypothesis that the ligands for schistosome TGF-β receptors are of host origin [5 , 13 , 14] , and a suggestion that host TGF-β , signaling through SmRK2 , plays a role in the pairing of male and female parasites [7] . Sexually mature S . mansoni live within the mesenteric vasculature , where each female produces approximately 300 eggs each day . Transmission of schistosomiasis depends upon the release of parasite eggs from the human host . Development of an immature egg into a mature egg containing a miracidium , the stage of the parasite that invades the intermediate fresh water snail host , occurs outside of the female worm , and takes approximately 5 d . Many of the eggs produced by schistosomes fail to reach the external environment , but instead become trapped in host tissues , where pathological changes caused by the immune responses to secreted egg antigens cause disease [15] . Despite the central importance of egg production in transmission and disease , and recent advances in proteomics and transcriptomics [12 , 16–18] , essentially nothing is known of the molecular pathways involved in embryogenesis in schistosomes . In this study , we describe the cloning and characterization of a S . mansoni TGF-β homolog , S . mansoni Inhibin/Activin ( SmInAct ) . Although we found SmInAct to be expressed in adult male and female parasites , and in eggs , the localization of SmInAct expression to the reproductive organs of female parasites focused our attention on the role of this gene in egg production . A role for SmInAct in reproduction was supported by analyses of female parasites recovered from infertile infections , in which we found that SmInAct protein was undetectable . Confirmation of the importance of this TGF-β superfamily member in the reproductive process was obtained from RNA interference ( RNAi ) studies , in which targeted knockdown of SmInAct in female worms or directly in the eggs that they produce resulted in a marked cessation of embryogenesis .
SmInAct was identified through a tblastn search of the Wellcome Trust's Sanger Institute's S . mansoni genome sequence using the C-terminal region of the Drosophila melanogaster dActivin sequence . We were unable to identify SmInAct in EST databases regardless of whether we searched using the coding or 3′–untranslated region ( UTR ) sequences . The 5′ and 3′ ends of SmInAct were amplified via rapid amplification of cDNA ends ( RACE ) using primers designed from within putative coding sequence and adult S . mansoni cDNA as template . The 1 . 3-kb , full-length SmInAct transcript contains 10 base pairs ( bp ) of 5′UTR , 808 bp of 3′UTR , and a poly-A tail . The deduced amino acid sequence of SmInAct is 161 residues long and contains many of the molecular hallmarks for a TGF-β , including a putative basic proteolytic cleavage site located at position 32 as RQRR where the bioactive , C-terminal domain ( 126 amino acids ) is enzymatically separated from the N-terminal pro-domain . Nine invariant cysteine moieties , and invariant proline and glycine residues ( Figure 1A ) essential for the proper dimerization and tertiary structure of a TGF-β homolog , are all predicted in SmInAct . The deduced amino acid sequence of SmInAct contains one putative N-linked glycosylation site at position 110 . Within the bioactive domain , SmInAct is 27% identical to both DAF-7 from Caenorhabditis elegans and dActivin from D . melanogaster , and 29% identical to human TGF-β 1 ( Figure 1A ) . Phylogenetic analysis of SmInAct among other TGF-β superfamily members groups this homolog with members of the TGF-β/Activin subfamily ( Figure 1B ) , and further clusters SmInAct phylogenetically with TGF-β homologs from the free-living nematode C . elegans ( DAF-7 ) and the parasitic nematodes Brugia malayi ( Bm-TGH-2 ) and Strongyloides stercoralis ( Ss-TGH-1 ) . To determine the expression of SmInAct at the transcript level , real-time reverse transcriptase–polymerase chain reaction ( RT-PCR ) was performed on cDNA from eggs , adult male parasites , and adult female parasites from mixed-sex infections . As seen in Figure 2A , SmInAct is expressed in all stages tested at relatively similar levels . Western analyses using polyclonal antibodies against recombinant SmInAct were used to determine the protein expression profile of SmInAct . The anti-SmInAct serum recognized a 28-kDa protein in egg antigen extracts and a doublet ( 32 kDa and 28 kDa ) in adult male and female extracts ( Figure 2B , lanes 1–3 ) ; these bands presumably represent the unprocessed ( 32 kDa ) inactive and processed ( 28 kDa ) active forms of the molecule . The relative molecular weights of the two bands recognized by anti-SmInAct antiserum in parasite extracts are larger than that predicted by the sequence , presumably due to detergent and reducing agent-resistant dimerization , and/or to glycosylation at amino acid 110 . Glycosylation plays an important role in the solubility and secretion of other members of the TGF-β superfamily [19 , 20] . Eggs appear to contain only the lower molecular weight , putatively active form of SmInAct . To localize SmInAct within the parasite , we performed in situ hybridization on sections of adult worms . Anti-sense probes localized SmInAct transcripts to the reproductive tissues of the adult female , with strong signals in the vitellaria and ovary ( Figure 2C ) , whereas in adult males , SmInAct transcripts localized to various subtegumental regions ( Figure 2D ) . The expression pattern in the female suggested a role for SmInAct in egg production . We focused on this possibility , and reasoned that if this were the case , SmInAct expression might be diminished in unfertile females . In vivo , successful oogenesis requires the presence of male schistosomes [21] , and , for reasons that have remained unclear , an intact CD4+ T lymphocyte compartment within the host [22] . Therefore , we analyzed SmInAct expression in female parasites from mice harboring single-sex infections , and in parasites from severely lymphopenic interleukin-7 receptor knockout ( IL-7R−/− ) mice carrying mixed-sex infections , which produce a significant number of dead eggs [23 , 24] . Real-time RT-PCR demonstrated that SmInAct mRNA levels were significantly decreased , but not absent , in females from these infections ( Figure 2E ) . Of particular interest , SmInAct protein was undetectable by Western analyses in females from single-sex infections as well as from infections of IL-7R−/− mice ( Figure 2B ) . While the localization of SmInAct transcripts to the male subtegumental region is not immediately informative in terms of function in the male , we nevertheless noted that male parasites recovered from infertile infections in IL-7R−/− mice were similar to female parasites in terms of transcriptional and post-transcriptional regulation of SmInAct expression ( Figure 2B and 2F ) . Moreover , this was also the case for male parasites recovered from male single-sex infections ( Figure 2B and 2F ) . To gain a better understanding of the function of SmInAct and the signaling pathway it activates , this TGF-β homolog was targeted for knockdown via RNAi [25–27] . Pairs of adult males and females recovered from infected mice were soaked in double-stranded RNA ( dsRNA ) corresponding to SmInAct ( 1 μg/ml ) or an irrelevant control dsRNA ( luciferase ) for 1 wk in vitro , followed by RNA extraction and real-time RT-PCR analyses . SmInAct dsRNA–treated worms showed a consistent and significant decrease in SmInAct expression of >40% when compared to SmInAct expression in worms soaked in the irrelevant control dsRNA ( Figure 3A ) . No consistently significant difference in the numbers of eggs produced by control versus SmInAct dsRNA–treated worm pairs was observed , suggesting that SmInAct is not important for egg production per se . However , in examining these cultures , we noted that eggs produced by SmInAct dsRNA–treated parasites failed to develop ( unpublished data ) . To specifically address the role of SmInAct in egg development , we treated eggs directly with SmInAct dsRNA . Approximately 20% of eggs laid by adult parasites during the first 2 d of in vitro culture will develop over the ensuing 5 d to contain miracidia [28] , with a typical progression of development through six stages illustrated in Figure 3B . Therefore , eggs produced by worm pairs for the first 2 d ex vivo were collected and soaked in dsRNA ( 1 μg/ml ) corresponding to SmInAct or an irrelevant dsRNA for 5 d , and their development was scored . Relative to eggs soaked in an irrelevant dsRNA , where ∼20% of the eggs developed through stage 6 , eggs treated with SmInAct dsRNA aborted development at stage 2 ( Figure 3C and 3D ) . An absence of SmInAct transcripts ( Figure S1 ) , and a nearly 10-fold decrease in SmInAct protein ( Figure 3E ) , were associated with the failure of SmInAct dsRNA–treated eggs to develop . This phenotype was not observed when eggs were treated with dsRNA corresponding to luciferase , a sequence not encoded in the schistosome genome ( Figure 3C and 3D ) , or to S . mansoni cathepsin B1 ( SmCB1 ) , a cathepsin B detectable in eggs ( Table 1 ) .
Multiple components of a TGF-β signaling pathway have been characterized in S . mansoni , but a ligand of parasite origin for the pathway has remained elusive . Additionally , while functions in host–parasite interactions have been proposed based on the expression of receptors on the parasite surface , and on the responsiveness of the parasite receptors to host TGF-β [5–7 , 14] , the function that TGF-β signaling plays in S . mansoni has remained unclear . In this study , we report the expression of SmInAct , a TGF-β–like ligand in the parasitic flatworm S . mansoni , the production of which is coupled to the reproductive potential of the worms . We provide evidence that SmInAct plays a crucial role in embryogenesis . Understanding of the developmental processes regulated by TGF-β in invertebrates is based largely on data from the model organisms D . melanogaster and C . elegans . Decapentaplegic , a bone morphogenetic protein ( BMP ) –like homolog in D . melanogaster , acts as a morphogen by determining cell fate along the dorsal–ventral axis in a gradient-dependent manner [29] . Also in D . melanogaster , a type I receptor , baboon , stimulates cellular proliferation and is essential for normal embryonic development [30] . Presumably , SmInAct could be fulfilling functions in the schistosome egg analogous to these known roles for decapentaplegic and/or baboon . None of the three characterized TGF-β homologs in C . elegans are important for patterning or growth of the embryo [31–33]; however , two TGF-β homologs have yet to be examined ( tig-2 and Y46E12BL . 1 ) , and , intriguingly , serial analysis of gene expression ( SAGE ) tags for both homologs have been found in the C . elegans embryo [34] . Like the other C . elegans TGF-β homologs that are resistant to RNAi affects , tig-2 and Y46E12BL . 1 have no phenotype in genome-wide RNAi screens [35 , 36]; therefore , direct mutagenesis will likely be required to determine the function of these genes . The identification of SmInAct , a TGF-β superfamily member , as a key component of egg development in S . mansoni , a member of the Platyhelminthes , the earliest branch of the Bilateria [37] , underscores the central role played by this pathway in embryogenesis . While one type I and one type II TGF-β receptor have been characterized for S . mansoni , there appears to be at least three type I receptors and two type II receptors present in the genome based on a preliminary blast search for homologs . It will be important to delineate which of the S . mansoni type I and type II TGF-β receptors are involved in SmInAct signaling and to identify the Smads important for transmitting the signal induced by this growth factor . Furthermore , identifying the genes regulated by SmInAct signaling will provide information regarding the precise function that this growth factor serves in egg maturation , as well as the functions the pathway may serve in other life stages of the parasite , including the adult male . SmInAct protein was not detectable in infertile females recovered from single-sex infections or from IL-7R−/− mice , despite the fact that these parasites contained SmInAct transcripts ( although at lower levels than in fecund parasites ) . This strongly indicates that SmInAct is both transcriptionally and post-transcriptionally regulated by worms of the opposite sex as well as by signals from the host . It is well established that parasites recovered from hosts lacking CD4+ T cells are developmentally stunted and produce significantly fewer fertile eggs than those recovered from mixed-sex infections of immunocompetent hosts . Translation of SmInAct mRNA is the first identified molecular process downstream of the effect of the host immune system on schistosome development [22–24] , and as such , could open the way towards an increased understanding of this unusual feature of schistosome biology . The finding that the production of SmInAct in males is under the same constraints as in females is curious and perhaps indicates an additional function ( s ) for SmInAct in S . mansoni . We are unaware of a link between the site of expression of SmInAct in the male schistosome and reproductive events , and further work is required to elucidate the function of SmInAct in male worms . In other settings , the uncoupling of transcription and translation is linked to the activation of the integrated stress response [38–41] . This mechanism , conserved in eukaryotes , re-programs cells to conserve energy in response to stress signals such as amino acid deficiency and oxidative stress by restricting the translation of transcripts requiring an active translation initiation complex [38–41] . Limited cellular energy is then used for the expression of genes necessary to maintain cell viability [42] . In this context , parasites in single-sex infections and in mice lacking CD4+ T cells may be considered stressed due to the lack of signals received from the opposite sex and immunocompetent host , thereby restricting the translation of non-essential transcripts . SmInAct protein expression may be considered expendable considering the role it plays in embryogenesis rather than in crucial cellular functions linked to the survival of the adult worm . A more thorough investigation of the S . mansoni homologs of translation factors involved in the stress response and of the regulation of other transcripts and protein expression will be required to evaluate this possibility . Post-transcription regulation of eukaryotic transcripts is controlled in part by the 3′UTR [43] . This region can bind elements ( including microRNAs and proteins ) that inhibit the translation and/or decrease mRNA stability . For example , 3′UTRs of several mammalian cytokines contain adenosine- and uridine-rich elements ( AREs ) that bind ARE-binding proteins ( ARE-BPs ) ( reviewed in [44] ) . The binding of ARE-BPs to these transcripts causes either rapid decay or inhibits their translation . While AREs are somewhat divergent in sequence , they often contain the consensus “AUUUA” and are found in a uridine-rich environment . Interestingly , the long 3′UTR of SmInAct has two exact repeats of “UUUCTAUUUA” that contain the consensus “AUUUA” ARE ( underlined ) . Furthermore , the 3′UTR of SmInAct is U-rich ( 43% uridines ) . It will be interesting to determine whether these repeats , or other regions of the long 3′UTR , play a role in the post-transcriptional regulation of SmInAct expression . It is of interest when considering the relationship of schistosomes with their mammalian hosts to note that in other systems , TGF-β superfamily members have been shown to function across phylum boundaries [45 , 46] . For example , the Drosophila BMP homologs DPP and 60A are able to induce bone development when injected into the skin of rats [45] , and mammalian BMP-4 can rescue Drosophila DPP mutants [46] . Consequently , we believe that it is feasible that SmInAct could act as a ligand to initiate signaling in host cells . It is clear that proteins produced by eggs have distinct immunomodulatory functions [47] , and SmInAct could conceivably participate in these effects if secreted/excreted from the schistosome egg . Our identification of SmInAct as a cytokine that is molecularly conserved between host and parasite , coupled with the description of an effective method for altering gene expression in the schistosome egg , allows these and other issues to now be addressed . Despite recent advances in vaccine design [48] , a solution for schistosomiasis remains an elusive goal . Current attempts to control schistosomiasis depend on repeated administration of one drug , praziquantel , with no replacements waiting in the wings should resistance develop . Understanding how schistosome eggs develop could provide targets for intervention in the schistosome life cycle and for blocking disease progression .
The Puerto Rican/NMRI strain of S . mansoni was used in all experiments . Adult schistosomes were recovered by hepatic-portal perfusion from C57BL/6 female mice or B6 IL-7R−/− ( The Jackson Laboratory , http://www . jax . org ) that had each been percutaneously exposed to ∼60 cercariae 8 wk earlier . Adult parasites and eggs laid were maintained in vitro in M199 ( Gibco , http://www . invitrogen . com ) , 10% fetal calf serum , 1% Antibiotic/Antimycotic ( Gibco ) , and 1% HEPES in a 37 °C/5% CO2 atmosphere as previously described [11 , 28] . The C-terminal , translated region of the Drosophila activin homolog ( dActivin ) ( amino acids 565–669 ) was used to search the Wellcome Trust's Sanger Institute's S . mansoni genome assembly using the tblastn algorithm . A contig ( 0020320 ) with significant similarity to dActivin was identified . Full-length cDNA corresponding to SmInAct was isolated using total RNA ( 1 μg ) from adult parasites and the SuperScript III GeneRacer 5′ and 3′ RACE kit ( Invitrogen , http://www . invitrogen . com ) as per manufacturer's instructions . Gene-specific primers were designed for isolation of the 5′-end ( 5′-GGTTCAAAACTTTTCGGGTGTA-3′ ) and 3′-end ( 5′-AATCTTGTTGTCATCCAACTCAA-3′ ) of SmInAct and used in RT-PCR with GeneRacer 5′ and 3′ primers according to manufacturer's suggestions . Resulting amplicons were cloned into the TOPO cloning vector ( Invitrogen ) and sequenced . To verify the full-length sequence of SmInAct , primers designed from the 5′ and 3′ ends of the transcript were used in RT-PCR , and the resulting fragment was cloned and sequenced . Sequence similarities between the deduced amino acid sequence of SmInAct and other members of the TGF-β superfamily were determined through multiple sequence alignments using the ClustalW algorithm , as well as the Align 2 sequences ( bl2seq ) program at the National Center for Biotechnology Information ( http://www . ncbi . nlm . nih . gov ) . An unrooted phylogram was drawn using amino acids within the conserved C-terminal domain of SmInAct , and known TGF-β superfamily members and distances were drawn using the Dayhoff Pam matrix and neighbor-joining algorithm in the PHYLIP software package developed by J . Felsenstein , University of Washington , Seattle , Washington , United States ( http://evolution . genetics . washington . edu/phylip . html ) . Percentages at branch points are based on 1 , 000 bootstrap runs . Total RNA was extracted from parasites using Qiagen's RNeasy Mini kit ( http://www . qiagen . com ) , and contaminating genomic DNA was removed by DNase treatment using the Turbo DNA-free endonuclease ( Ambion , http://www . ambion . com ) . First-strand cDNA was synthesized using 500 ng of RNA , SuperScript II reverse transcriptase ( Invitrogen ) , and oligo dT as a primer . RT-minus controls were performed to confirm the absence of genomic DNA ( unpublished data ) . SmInAct transcript levels in egg and adult stages were quantified relative to α-tubulin using Applied Biosystems' 7500 real-time PCR system and SYBR green PCR Master Mix ( Applied Biosystems , http://www . appliedbiosystems . com ) . Total reaction volume was 10 μl with 300 nM of each primer , 5 μl of SYBR green PCR Master Mix , and 0 . 5 μl of cDNA as template ( or water as a negative control ) . SmInAct primers were: forward 5′-AATCTTGTTGTCATCCAACTCAA-3′ and reverse 5′-AACTACAAGCACATCCTAAAACAA-3′ . α-Tubulin primers were: forward 5′-CCAGCAAATCAGATGGTGAA-3′ and reverse 5′-TTGACATCCTTGGGGACAAC-3′ . PCR efficiency ( E ) was determined for both primer sets by plotting cycle thresholds from a 10-fold serial dilution of cDNA and inputting the slope in the equation E = 10 ( −1/slope ) . For expression analyses , quantification of SmInAct transcript relative to α-tubulin was calculated using the equation: ratio = ( ESmα-tubulin ) CT/ ( ESmInAct ) CT where ESmα-tubulin is the PCR efficiency of the reference gene , ESmInAct is the PCR efficiency of target gene , and CT is the cycle threshold . For analysis of RNAi-induced knockdown , quantification of SmInAct transcript relative to paramyosin ( paramyosin primers were: forward 5′-CGTGAAGGTCGTCGTATGGT-3′ and reverse 5′-GACGTTCAAATTTACGTGCTTG-3′ ) was calculated using the 2−ΔΔCt method . Dissociation curves were generated for each real-time RT-PCR to verify the amplification of only one product . Eco RI ( forward ) and Xho I ( reverse ) tagged primers were designed to amplify the C-terminal bioactive region of SmInAct ( forward 5′-GGAATTCTCATTAACTAAAGGAGATGA-3 and reverse 5′-CCGCTCGAGTTAACTACAAGCACATCCTA-3′ ) . The amplified product was cloned into the expression vector pET28a+ ( Novagen , http://www . emdbiosciences . com ) and sequenced to verify the absence of any mutations . Expression of recombinant SmInAct was induced in Escherichia coli BL21 ( DE3 ) by addition of 1 mM IPTG when cultures reached an OD600 of 0 . 5 at 37 °C , followed by 3 hours of shaking at room temperature . Recombinant SmInAct was expressed in bacteria as insoluble inclusion bodies . Exhaustive attempts to refold the protein using gluathione and reduced glutathione proved unsuccessful . We therefore purified the protein via nickel column chromatography under denaturing conditions ( 6 M urea ) as per the manufacturer's protocol ( Novagen ) . Antiserum was generated by Cocalico Biologicals ( http://www . cocalicobiologicals . com ) through subcutaneous inoculation of a rabbit with 100 μg of purified protein in complete Freund's adjuvant , followed by three boosts of 50 μg in incomplete Freund's adjuvant on days 14 , 21 , and 49 , followed by exsanguinations on day 64 . For detection of SmInAct protein , 10 μg of protein extracted from eggs , adult males , and adult females via Dounce homogenizing in lysis buffer ( 1% Triton-X 100 , 20 mM HEPES , 10% glycerol , 150 mM NaCl ) supplemented with a protease inhibitor cocktail ( Sigma , http://www . sigmaaldrich . com ) were separated by SDS-PAGE , electroblotted , and probed with anti-SmInAct antiserum ( 1:10 , 000 ) , pre-immune serum ( 1:10 , 000 ) , or a monoclonal antibody ( 4B1 ) against paramyosin . Affinity purified HRP-conjugated goat anti-rabbit IgG ( Cell Signaling Technology , http://www . cellsignal . com ) was used to detect bound rabbit antibodies , while an affinity purified HRP-conjugated horse anti-mouse IgG ( Cell Signaling Technology ) was used to detect the anti-paramyosin monoclonal antibody . The secondary antibodies were detected using ECL reagents as per manufacturer's instructions ( GE Healthcare , http://www . gehealthcare . com ) . Localization of SmInAct in 5-μm sections of adult S . mansoni was performed as previously described [49] . DIG-labeled sense and anti-sense transcripts were generated using Roche's DIG RNA labeling mix ( http://www . roche . com ) as per manufacturer's instructions with T7-tagged amplicons as template ( sense: forward 5′-TAATACGACTCACTATAGGGTTGATCCAAAAAAGGTTGTTATGG-3′ , reverse 5′-TTAACTACAAGCAGCTCCTA −3′; anti-sense: forward 5′-ATAATATGTAATAATTGTGA −3′ reverse 5′- TAATACGACTCACTATAGGGAACTACAAGCACATCCTAAAACAA-3′ ) . The hybridized DIG-probes were detected using an alkaline-phosphatase conjugated anti-DIG antibody ( Roche ) , and visualized using NBT ( 337 . 5 μg/ml ) and BCIP ( 175 μg/ml ) in 0 . 1M Tris-HCl , 0 . 1M NaCl , 0 . 05 MgCl2 . Worm sections were photographed using a Leica DMIRB microscope and DC500 camera ( Leica , http://www . leica . com ) . dsRNA was synthesized using the T7 Megascript kit ( Ambion ) as per manufacturer's instructions . T7-tagged primers were used to generate a 381-bp SmInAct-dsRNA template encompassing the active ligand domain ( forward 5′-TAATACGACTCACTATAGGGCGATCATTAACTAAAGGAGATGAG-3′ , reverse 5′-TAATACGACTCACTATAGGGAACTACAAGCACATCCTAAAACAA-3′ ) . Luciferase and SmCB1 dsRNAs ( negative controls ) were generated as described [25] . For dsRNA treatment of worms , five adult pairs were cultured in the presence of 1 μg/ml dsRNA for 7 d with medium and dsRNA changes occurring every other day . For dsRNA treatment of eggs , five adult pairs were cultured as above ( in the absence of dsRNA ) for 2 d , worms were removed , and dsRNA was added at 1 μg/ml . Eggs were photographed using a Leica DMIRB microscope and DC500 camera . Student t-test was used for statistical analyses of dsRNA-induced knockdown of SmInAct expression , change in expression of SmInAct in single-sex and IL-7R−/− mice , and egg developmental phenotypes ( control versus SmInAct dsRNA ) . Chi-square analyses were used to test the statistical significance of the egg developmental phenotype . The Yates correction was applied because we specified only two categories: undeveloped and developed ( Table 1 ) .
Sequence data reported in this manuscript are available from GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) under accession number DQ863513 . Other GenBank accession numbers of genes and sequences used in this study include: B . malayi TGH-1 ( AAB71839 ) ; B . malayi TGH-2 ( AAD19903 ) ; C . elegans DAF-7 ( NP_497265 ) ; C . elegans DBL-1 ( NP_504709 ) ; Danio rerio Activinβ A isoform 2 ( AAX68505 ) ; D . melanogaster Activin ( NP_651942 ) ; D . melanogaster dActivin ( AF454392 ) ; D . melanogaster decapentaplegic ( NP_477311 ) ; Homo sapiens Activinβ E ( NP_113667 ) ; H . sapiens BMP-2 ( NP_001191 ) ; H . sapiens BMP-3 ( NP_001192 ) ; H . sapiens BMP-4 ( NP_031580 ) ; H . sapiens BMP-5 ( NP_066551 ) ; H . sapiens BMP-6 ( NP_001709 ) ; H . sapiens BMP-7 ( NP_001710 ) ; H . sapiens BMP-8 ( NP_861525 ) ; H . sapiens GDF-5 ( NP_000548 ) ; H . sapiens GDF-6 ( NP_001001557 ) ; H . sapiens GDF-7 ( NP_878248 ) ; H . sapiens GDF-10 ( NP_004953 ) ; H . sapiens Inhibinβ A precursor ( NP_002183 ) ; H . sapiens Inhibinβ B ( NP_002184 ) ; H . sapiens Inhibinβ C ( NP_005529 ) ; H . sapiens TGF-β 1 ( NP_000651 ) ; H . sapiens TGF-β 2 ( NP_003229 ) ; H . sapiens TGF-β 3 ( NP_003230 ) ; Mus musculus BMP-2 ( NP_031579 ) ; M . musculus BMP-3 ( NP_775580 ) ; M . musculus BMP-4 ( NP_031580 ) ; M . musculus GDF-10 ( NP_665684 ) ; M . musculus Inhibinβ A ( NP_032406 ) ; M . musculus Inhibinβ B ( NP_032407 ) ; M . musculus TGF-β 1 ( NP_035707 ) ; M . musculus TGF-β 2 ( NP_033393 ) ; M . musculus TGF-β 3 ( NP_033394 ) ; S . mansoni α-tubulin ( M80214 ) ; S . mansoni paramyosin ( M35499 ) ; and Strongyloides stercoralis TGH-1 ( AAV84743 ) . | Schistosomes are parasitic worms that infect hundreds of millions of people in developing countries . They cause disease by virtue of the fact that the eggs that they produce , which are intended for release from the host in order to allow transmission of infection , can become trapped in target organs such as the liver , where they induce damaging inflammation . Egg production by female schistosomes is critically dependent on the presence of male parasites , without which females never fully develop , and ( counterintuitively ) on the contribution of signals from the host's immune system . Very little is understood about the molecular basis of these interactions . Here , we describe a newly discovered schistosome gene , which is expressed in the reproductive tract of the female parasite and in parasite eggs . The protein encoded by this gene is made only when females are paired with males in an immunologically competent setting . Using recently developed tools that allow gene function to be inhibited in schistosomes , we show that the product of this gene plays a crucial role in egg development . Examining how the expression of this gene is controlled has the potential to provide insight into the molecular nature of the interactions between male and female parasites and their hosts . Moreover , the pivotal role of this gene in the egg makes it a potential target for blocking transmission and disease development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"animals",
"in",
"vitro",
"infectious",
"diseases",
"microbiology"
] | 2007 | TGF-β Signaling Controls Embryo Development in the Parasitic Flatworm Schistosoma mansoni |
Mobile bacterial group II introns are evolutionary ancestors of spliceosomal introns and retroelements in eukaryotes . They consist of an autocatalytic intron RNA ( a “ribozyme” ) and an intron-encoded reverse transcriptase , which function together to promote intron integration into new DNA sites by a mechanism termed “retrohoming” . Although mobile group II introns splice and retrohome efficiently in bacteria , all examined thus far function inefficiently in eukaryotes , where their ribozyme activity is limited by low Mg2+ concentrations , and intron-containing transcripts are subject to nonsense-mediated decay ( NMD ) and translational repression . Here , by using RNA polymerase II to express a humanized group II intron reverse transcriptase and T7 RNA polymerase to express intron transcripts resistant to NMD , we find that simply supplementing culture medium with Mg2+ induces the Lactococcus lactis Ll . LtrB intron to retrohome into plasmid and chromosomal sites , the latter at frequencies up to ~0 . 1% , in viable HEK-293 cells . Surprisingly , under these conditions , the Ll . LtrB intron reverse transcriptase is required for retrohoming but not for RNA splicing as in bacteria . By using a genetic assay for in vivo selections combined with deep sequencing , we identified intron RNA mutations that enhance retrohoming in human cells , but <4-fold and not without added Mg2+ . Further , the selected mutations lie outside the ribozyme catalytic core , which appears not readily modified to function efficiently at low Mg2+ concentrations . Our results reveal differences between group II intron retrohoming in human cells and bacteria and suggest constraints on critical nucleotide residues of the ribozyme core that limit how much group II intron retrohoming in eukaryotes can be enhanced . These findings have implications for group II intron use for gene targeting in eukaryotes and suggest how differences in intracellular Mg2+ concentrations between bacteria and eukarya may have impacted the evolution of introns and gene expression mechanisms .
Mobile group II introns are retrotransposons that also function as self-splicing introns [1] . They are found in bacteria , archaea , and in the bacterial endosymbiont-derived mitochondrial and chloroplast genomes of some eukaryotes , particularly fungi and plants [2] . Despite their prokaryotic origin , mobile group II introns are believed to have strongly impacted eukaryotic nuclear genomes as evolutionary ancestors of spliceosomal introns , the spliceosome , LINEs and other non-LTR retrotransposons , and telomerase [3 , 4] . Mobile group II introns insert into new DNA sites by a ribozyme-based site-specific DNA integration mechanism called retrohoming , which is thought to have enabled mobile group II introns or their close relatives to proliferate within the nuclear genomes of early eukaryotes before evolving into spliceosomal introns [4 , 5] . In addition to its evolutionary significance , retrohoming underlies the use of group II introns as gene targeting vectors ( “targetrons” ) , which use intron RNA/DNA target site base-pairing interactions to achieve high and programmable DNA target specificity [6–8] . Targetrons are widely used for gene targeting in bacteria , where retrohoming frequencies are high enough to identify targeting events by colony PCR screening without using genetic markers [9] . By contrast , mobile group II introns and targetrons derived from them function inefficiently in eukaryotes [10–12] , and group II introns appear to be completely absent from the nuclear genomes of present-day eukaryotes [13] . The reasons for the different behavior of group II introns in prokaryotes and eukaryotes and factors that dictated their conversion into spliceosomal introns and exclusion from eukaryotic nuclear genomes remain incompletely understood . Mobile group II introns consist of a catalytically active intron RNA ( a ribozyme ) and an intron-encoded reverse transcriptase ( RT ) , which function together to promote both RNA splicing and retrohoming [1] . The intron RNA catalyzes its own splicing from a precursor RNA via two sequential transesterification reactions that result in ligated exons and an excised intron lariat RNA , identical to the splicing reaction mechanism used by spliceosomal introns in higher organisms [4 , 14] . To catalyze splicing , the intron RNA folds into a conserved tertiary structure that consists of six interacting secondary structure domains ( DI-DVI ) , with three distinct structural subclasses of group II introns , IIA , IIB , and IIC , distinguished by secondary and tertiary structure features [1] . This folded RNA forms a ribozyme active site that includes nucleotide residues highly conserved in all three group II intron subclasses and utilizes site-specifically bound Mg2+ ions to catalyze RNA splicing and reverse splicing reactions [15–17] . The group II intron RT contributes to splicing by binding to the intron RNA and promoting formation of this catalytically active RNA structure [18–20] . After splicing , the RT remains bound to the excised intron lariat RNA in a ribonucleoprotein particle ( RNP ) that initiates retrohoming by recognizing a DNA target site [21] . DNA target site recognition is primarily by base pairing of sequence elements within the intron RNA to DNA sequences spanning the intron-insertion site , with only a small contribution of the group II intron RT , which helps promote local DNA melting [22] . The intron RNA then uses its ribozyme activity to insert directly into the retrohoming site , where it is reverse transcribed by the intron-encoded RT into an intron cDNA that is integrated into the genome by host enzymes [5 , 23–26] . Early findings that group II introns use the same splicing reaction mechanism as spliceosomal introns and that some organellar group II introns have been fragmented by DNA recombination into two or three unlinked segments that reassociate to promote RNA splicing suggested an evolutionary relationship to spliceosomal introns and a possible evolutionary origin for present-day snRNAs [27] . Recently , these hypotheses have been strongly supported by group II intron RNA crystal structures and biochemical studies , which demonstrate striking structural and functional similarities between group II intron domains and three key snRNAs ( U4 , U5 , and U6 ) that comprise the catalytic core of the spliceosome [17 , 28–31] . The similarities include identical RNA-catalyzed splicing reactions based on similarly positioned catalytic Mg2+ ions at the RNA active site [15 , 16 , 30 , 31] . Moreover , recent structural and bioinformatic studies indicate that the conserved spliceosomal core protein Prp8 was derived from a group II intron-like RT and functions similarly as a structural scaffold for an RNA catalytic core [32 , 33] . Considered together with the phylogenetic distribution of group II introns , these findings support a scenario in which mobile group II introns entered ancestral eukaryotes along with bacterial endosymbionts that gave rise to mitochondria , invaded the nucleus , proliferated to high copy number , and then degenerated into snRNAs [34] . Further , this proliferation of introns in eukaryotic nuclear genes is hypothesized to have been a major driving force for the evolution of eukaryotes themselves , including for features such as ( i ) the nuclear membrane to separate transcription and splicing from translation , thereby limiting mistranslation of intron-containing RNAs; ( ii ) nonsense-mediated decay ( NMD ) to degrade unspliced or misspliced intron-containing transcripts that escape to the cytosol; and ( iii ) large-scale alternative splicing , enabling greater organismal complexity within constraints on genome size [3] . Several factors have been identified that limit group II intron function and their ability to propagate in eukaryotes . First , studies in Saccharomyces cerevisiae showed that RNA polymerase II ( Pol II ) transcripts containing the Lactococcus lactis Ll . LtrB group II intron , which belongs to subgroup IIA , are subject to both NMD and translational repression , leading to their accumulation in cytoplasmic foci [11] . This translational repression appears to reflect strong intermolecular base-pairing interactions between the ligated-exon junction sequence in the spliced mRNA and the excised intron or intron-containing precursor RNAs , which may impede translating ribosomes and/or target the RNA for degradation [35] . A second factor affecting group II intron propagation in eukaryotes appears to be suboptimal intracellular Mg2+ concentrations , which limit group II intron ribozyme activity [10] . Group II intron splicing and retrohoming both require relatively high Mg2+ concentrations compared to other cellular processes , and Mg2+ concentrations appear to be significantly lower in eukaryotes than in bacteria [10 , 36–38] . Studies of S . cerevisiae mtDNA introns by Schweyen and coworkers showed that mutations in a mitochondrial Mg2+ transporter inhibit the splicing of all four group II introns , including both subgroup IIA and IIB introns , while having minimal effect on the transcription or splicing of group I introns , which use a different ribozyme-based splicing mechanism that is less sensitive to Mg2+ concentration [37] . Further , microinjection assays in Xenopus laevis oocyte nuclei or Drosophila and zebrafish embryos showed that in vitro reconstituted Ll . LtrB group II intron RNPs could retrohome efficiently into plasmid target sites only if additional Mg2+ was co-injected with the plasmid DNA [10] . An attempt to overcome this limitation in human cells by using an algal mitochondrial group IIB intron ( Pl . LSU/2 ) that self-splices at physiological Mg2+ concentrations in vitro , was unsuccessful [12] , perhaps because efficient self-splicing of this intron at low Mg2+ concentrations requires the presence of 1 M NH4Cl [39] . Recently , we selected variants of the Ll . LtrB group II intron with mutations in catalytic core domain V ( DV ) that retrohome 10- to 20-fold more efficiently than the wild-type intron in a Mg2+-deficient E . coli strain [36] . These findings suggested that it might be possible to overcome the high Mg2+ requirement that prevents efficient group II intron retrohoming in eukaryotes by mutations at a few critical sites within the intron RNA . Here , we developed a mobile group II intron expression system for human cells that utilizes an Ll . LtrB group II intron RNA expressed by using T7 RNA polymerase ( T7 RNAP ) to overcome NMD and a separately expressed human codon-optimized group II intron RT . By using this expression system , we found that simply supplementing the cell culture medium with 20–80 mM Mg2+ enables the Ll . LtrB intron to retrohome into plasmid and genomic target sites , the latter at frequencies of up to ~ 0 . 1% , in viable human cells . Further , we performed multiple rounds of in vivo selection of the intron ribozyme , analyzed the fitness landscape using Pacific Biosciences deep sequencing , and identified positively selected mutations that were used for synthetic shuffling to generate Ll . LtrB variants that show enhanced retrohoming in human cells . However , the maximum enhancement was <4-fold and still required extra Mg2+ in the culture medium . These findings indicate that low Mg2+ concentrations constitute a natural barrier to efficient retrohoming in eukaryotes that is not readily overcome by mutational variation and selection , and they have implications for the use of group II introns for gene targeting in higher organisms and the evolution of introns and gene expression mechanisms .
The mobile group II intron expression system that we developed for human cells consists of three plasmids ( Fig 1A ) . The first plasmid , denoted phLtrA , expresses the Ll . LtrB group II intron RT ( denoted LtrA protein ) with humanized codon usage and a C-terminal SV40 nuclear localization sequence ( NLS ) ( NCBI Genbank , accession number KP851976 ) [40] . The humanized LtrA ORF is expressed from a constitutive RNA polymerase II ( Pol II ) promoter , the cytomegalovirus immediate early ( CMV ) promoter [41] , and is followed by a polyadenylation signal ( pA ) . An early version of this plasmid , phLtrA1 , has a small artificial spliceosomal intron ( IVS ) inserted after the initiation codon [42] that was later found to be unnecessary for expression of hLtrA . The second plasmid , pLl . LtrB , uses a phage T7 promoter to express the Ll . LtrB intron with the LtrA ORF deleted ( denoted Ll . LtrB-ΔORF ) and short flanking 5’- and 3’-exon sequences ( denoted E1 and E2 , respectively ) . Finally , the third plasmid , pT7-NLS , expresses phage T7 RNAP with a fused N-terminal SV40 NLS driven by a CMV promoter . Previous work showed that T7 RNAP can produce high levels of uncapped , non-polyadenylated transcripts in human cells [43] and that its subcellular localization can be controlled , with nearly complete cytoplasmic or nuclear localization when expressed without or with an appended SV40 NLS , respectively [44] . The group II intron expression plasmids were not toxic when transfected by themselves or together into HEK-293 cells ( Fig 1B ) . We first compared expression of the LtrA protein with and without human optimized codons in HEK-293 cells . As shown in Fig 2A , the plasmid expressing the human-codon optimized LtrA ORF produced hLtrA protein that was readily detected by immunoblotting ( lane 5 and 6 ) , whereas an identical plasmid with a native non-codon-optimized LtrA ORF produced no detectable LtrA protein ( lane 4 ) . Further , nuclear lysates from HEK-293 cells transfected with the plasmid expressing hLtrA but not untransfected cells showed a high level of RT activity with a substrate that is efficiently used by purified LtrA protein ( Ll . LtrB/E2+10 RNA; an Ll . LtrB intron-containing transcript with a DNA primer annealed downstream of the intron; Fig 2B ) . Immunofluorescence microscopy showed that the hLtrA expressed with a C-terminal NLS localized to the nucleus in HEK-293 cells and COS-7 cells , whereas hLtrA expressed without an added NLS ( ΔNLS ) localized to the cytoplasm ( Fig 2C–2E ) . The requirements of LtrA for codon optimization and addition of an NLS to localize to the nucleus differ from recent findings for the Sinorhizobium meliloti RmInt1 group II intron RT , which does not require codon optimization and localizes to nucleoli in Arabidopsis thaliana protoplasts without an added NLS [46] . Together , our results show that optimization toward human codon usage overcomes a barrier to the expression of the Ll . LtrB group II intron RT in eukaryotes and that an appended NLS is required to localize this protein to the nucleus . Previous studies in S . cerevisiae showed that RNA polymerase II ( Pol II ) transcripts containing the Ll . LtrB intron are subject to both NMD and translational repression [11] . To determine the effect of NMD on group II intron-containing transcripts in human cells , we constructed plasmids that use a CMV ( Pol II ) promoter to express blue fluorescent protein ( BFP ) with or without the Ll . LtrB-ΔORF intron and short flanking exon sequences ( E1 and E2 ) inserted directly after the BFP start codon ( Fig 3A ) . We transfected the plasmids into HeLa cells that were pre-treated with siRNAs targeted against UPF1 mRNA , which encodes an essential component of the NMD complex [48] , or a scrambled siRNA control , and then quantified BFP transcript levels by RT-qPCR at 48 h after plasmid transfection . As shown in Fig 3B , transcript levels for the uninterrupted BFP ORF remained high in the presence of both the UPF1 and scrambled control siRNA , with little if any significant effect of NMD knockdown . By contrast , the inclusion of the Ll . LtrB intron in the BFP ORF led to a strong decrease in transcript level in the presence of the control siRNA , but not in the presence of the UPF1 siRNA to block NMD , irrespective of co-expression of the LtrA protein . UPF1 knockdown was confirmed by immunoblotting ( Fig 3C ) . These findings indicate that the NMD pathway degrades Pol II transcripts containing the Ll . LtrB intron in human cells as it does in S . cerevisiae [11] . The finding that Pol II transcripts containing the Ll . LtrB intron are subject to NMD in human cells led us to test whether this barrier could be overcome by using T7 RNAP for Ll . LtrB expression . T7 RNAP transcripts are not capped , polyadenylated , or subject to pre-mRNA processing in the same way as Pol II transcripts and thus are not expected to be subject to NMD [43] . For these experiments , we constructed two T7-promoter-driven GFP expression plasmids , one denoted pGFP-Ll . LtrB containing the Ll . LtrB intron and short flanking exon sequences inserted within the GFP ORF , and the other denoted pGFP containing the ligated-exon sequences that would result from Ll . LtrB intron splicing inserted at the same location ( Fig 4A ) . In both plasmids , the GFP ORF is preceded by an internal ribosome entry site ( IRES ) to enable GFP expression if the Ll . LtrB intron is spliced . Paralleling the protocol used for BFP-encoding Pol II transcripts in Fig 3 , we transfected these GFP-encoding plasmids together with pT7-NLS , which expresses T7 RNAP , into HEK-293 cells that had been pre-treated with the UPF1 siRNA or a scrambled control siRNA , and we measured GFP transcript levels by RT-qPCR at 48 h after transfection of the plasmids . In this case , the T7-GFP ORF control and T7-Ll . LtrB-GFP transcript containing the Ll . LtrB intron were present at similar levels with either the scrambled or UPF1 siRNA , with UPF1 knockdown by the UPF1 siRNA again confirmed by immunobotting ( Fig 4B ) . These findings indicate that a T7-transcript containing the Ll . LtrB intron is not subject to NMD pathway-related degradation in human cells . The Pol II-transcripts with the Ll . LtrB intron inserted in BFP ORF described in the preceding sections were not spliced in human cells , and this was also the case for the T7 transcripts with the Ll . LtrB intron inserted in the GFP ORF . As we suspected that splicing of the Ll . LtrB-intron might be limited by low Mg2+ concentrations in human cells , we tested whether splicing of the T7 transcript containing the Ll . LtrB intron might be induced simply by growing cells in culture medium containing elevated concentrations of MgCl2 ( Fig 4C ) . In these experiments , we transfected the three expression plasmids phLtrA , pT7-NLS , and pGFP-Ll . LtrB into HEK-293 cells in culture medium with or without added 80 mM MgCl2 and assayed Ll . LtrB intron splicing by RT-PCR of cellular RNAs at 48-h post-transfection . In standard culture medium , the GFP-Ll . LtrB transcript by itself showed no detectable splicing ( lane 2 ) , while co-expression of hLtrA led to low levels of splicing ( lane 4 ) . Surprisingly , the addition of MgCl2 to the culture medium by itself led to a large increase in splicing even in the absence of hLtrA ( lane 3 ) . Splicing levels in the presence of both exogenous MgCl2 and hLtrA appeared to be somewhat lower than with MgCl2 alone ( lane 5 ) . Accurate splicing was confirmed by sequencing of the ligated-exon junction in the PCR product . Notably , although the Ll . LtrB intron was spliced under these conditions , we detected no expression of GFP from the spliced transcript , whereas GFP was expressed efficiently from the control transcript containing the ligated-exon junction sequence inserted at the same location in the GFP ORF ( Fig 4D ) . Together , these findings indicate that exogenous MgCl2 can by itself induce splicing of the Ll . LtrB intron in human cells , even in the absence of LtrA protein , which is required for Ll . LtrB intron splicing in bacteria [49] . However , T7 transcripts from which the Ll . LtrB intron had been spliced in human cells still appear to be subject to a translational block similar to what was found for RNAP II transcripts in S . cerevisiae [11] . In vitro , the LtrA protein can be reconstituted with excised intron lariat RNA to generate RNPs that are active in retrohoming [21] . Thus , we tested whether the excised intron RNA resulting from Ll . LtrB splicing in culture medium containing added MgCl2 ( 80 mM ) could be combined with expressed hLtrA to promote retrohoming in human cells . To assess retrohoming in human cells , we used sensitive Taqman qPCR-based assays that quantify both the 5'- and 3'-integration junctions resulting from integration of the Ll . LtrB intron into the wild-type DNA target site ( Fig 5A ) . We tested for retrohoming into a single genomic copy of the wild-type Ll . LtrB homing site in HEK-293 Flp-in cells and in the same cells after co-transfection of a recipient plasmid ( pFRT ) carrying the same target site . As the transfected plasmid is expected to be present in much higher copy number ( ~104 ) [50] than the genomic target site , this protocol enables direct comparison of plasmid and genomic targeting in parallel transfections of the same cells . In these experiments , a 24-h period of polyethylenimine ( PEI ) -mediated transfection of the expression plasmids was followed by an additional 24-h period in which cells were incubated in growth medium containing 80 mM MgCl2 . After MgCl2 treatment , the cells ( both adherent and non-adherent ) were collected , and DNA was extracted for qPCR analysis . Cells receiving all three expression plasmids and 80 mM MgCl2 showed significant retrohoming into both genomic and plasmid retrohoming assays ( Fig 5B ) . In three separate experiments , the average retrohoming frequency and standard deviation for the genomic target site measured by qPCR of RNase-treated whole-cell DNA was 0 . 23 ± 0 . 02% for the 3'-integration junction and about 7-fold lower , 0 . 033 ± 0 . 002% , for the 5'-integration junction [note the different scales of the y-axis for 5’-junctions ( blue bars ) and 3’ junctions ( red bars ) in Fig 5B and 5C ) . ] The retrohoming frequencies for cells co-transfected with the recipient plasmid , which is present at ~104 copies per cell and expected to be largely cytosolic [50 , 51] , were substantially higher ( 1 . 4 ± 0 . 1% for the 3’-integration junction and 0 . 056 ± 0 . 004% , for the 5' junction ) . The lower frequency of 5’- than 3’-integration junctions for retrohoming of the Ll . LtrB intron into genomic and plasmid target sites may reflect that a high proportion of the retrohoming events result in the integration of 5’-truncated introns , similar to the situation for human LINE-1 elements where retrotransposition frequently results in 5’ truncations due to abortive reverse transcription [52] . Surprisingly , retrohoming efficiencies with both plasmid and genomic target sites were similar regardless of whether or not the expressed T7 RNAP contained an NLS ( S1 Fig ) . This finding presumably reflects that RNPs resulting from transcription of pLl . LtrB that remains in the cytosol after transfection can still gain access to the genomic target site ( S1 Fig; see Discussion ) . For retrohoming into the plasmid target site , full-length intron integrations requiring all steps in retrohoming were confirmed by conventional PCR and sequencing of the integration junction in the above experiments ( S2 Fig ) and more extensively in genetic assays described below . For the genomic target site , the very low frequency of full-length intron integrations ( 0 . 033% ) made it difficult to recover them from whole-cell DNA by conventional PCR . However , both the 5’- and 3’-integration junctions expected for full-length integrations were detected by Taqman qPCR assays of RNase-treated genomic DNA at levels well above background and with the same excess of 3’ junctions as found in the plasmid assay ( Fig 5B and 5C ) . Additionally , unlike splicing of the Ll . LtrB intron in human cells , which is not dependent upon LtrA protein ( see above ) , retrohoming of the Ll . LtrB intron into both plasmid and genomic DNA target site and the detection of both the 5’- and 3’-DNA integration junctions required the LtrA protein , which is needed for DNA target site recognition as well as reverse transcription ( Fig 5B and 5C ) . Finally , in an important control , no significant retrohoming into the wild-type plasmid or genomic site was detected under any condition for an Ll . LtrB intron retargeted to insert into the CCR5 gene ( Fig 5B ) . We confirmed that this CCR5 targetron retrohomes into a plasmid-borne CCR5 target in HEK-293 cells at frequencies of 0 . 24–0 . 27% for the 3’-integration junction , but could not detect integrations into the genomic CCR5 gene . Although retrohoming frequencies in HEK-293 cells with added Mg2+ were relatively high , we observed that the addition of 80 mM MgCl2 to the culture medium to promote retrohoming resulted in cellular blebbing , a hallmark of apoptosis [53] , with about half of the cells becoming non-adherent and unable to divide in fresh media . Inviable non-adherent cells could potentially have higher targeting rates due to enhanced Mg2+ influx due to more permeable cell membranes . Consistent with this possibility , we found that retrohoming frequencies at 80 mM MgCl2 were substantially higher in non-adherent cells ( 3’-integration junctions 0 . 4–1 . 0% and 0 . 4–1 . 5% for genomic and plasmid target sites , respectively ) than in adherent cells ( 3’-integration junctions 0 . 01–0 . 08% and 0 . 13–0 . 50% for genomic and plasmid target sites , respectively ) ( Fig 5D ) . We tested whether lower MgCl2 concentrations , shorter targeting times , or different Mg2+ salts could alleviate the deleterious effects of added Mg2+ , but found that all treatments that improved cell viability decreased retrohoming frequencies to unattractively low levels ( S3 Fig ) . Cell populations in which the Ll . LtrB intron had integrated into the genomic site at 80 mM Mg2+ were viable and remained adherent in high MgCl2 growth medium indefinitely . Thus , these experiments indicate that the Ll . LtrB intron can retrohome into both plasmid and genomic target sites in viable human cells , the latter at frequencies as high as ~0 . 1% as measured by 3’-integration junctions , so long as extra Mg2+ is added to the culture medium . The finding that retrohoming of the Ll . LtrB intron in human cells is limited by low Mg2+ concentrations led us to test whether we could select Ll . LtrB intron variants that could retrohome more efficiently at low Mg2+ concentrations in human cells . We previously selected Ll . LtrB variants with mutations in the distal stem of domain V ( DV ) that had 10- to 20-fold higher retrohoming efficiencies in a Mg2+-deficient E . coli mutant , as well as decreased Mg2+-dependence for RNA splicing and reverse splicing in vitro [36] . However , neither of the two best such variants had increased retrohoming efficiency into genomic or plasmid target sites in HEK-293 cells with or without 80 mM MgCl2 added to the culture medium ( S4 Fig ) . We also tested an intron variant that was selected for enhanced retrohoming in Xenopus laevis oocyte nuclei , another environment in which low Mg2+ concentrations are stringently limiting for retrohoming [10 , 54] . Although this variant had ~4-fold higher retrohoming efficiency in X . laevis oocyte nuclei , it did not show higher retrohoming frequencies than wild-type Ll . LtrB in human cells in our assays ( S5 Fig ) . A possible explanation is that these Ll . LtrB variants selected in E . coli or X . laevis are optimized for different intracellular environments and Mg2+ concentrations than those in human cells . Thus , we attempted to select Ll . LtrB variants with enhanced retrohoming directly in human cells . For directed evolution in human cells , we adapted an E . coli plasmid-based genetic assay for retrohoming that avoids pitfalls of PCR amplification of low frequency intron-integration events [6] . In this assay , a group II intron carrying a phage T7 promoter retrohomes into a target site cloned on a recipient plasmid upstream of a promoterless tetracycline-resistance gene ( tetR ) resulting in a TetR plasmid that can be selected by transformation of human cell DNA preparations into E . coli ( Fig 6A ) . The retrohoming efficiency of the Ll . LtrB variant containing the phage T7 promoter in HEK-293 cells supplemented with 80 mM MgCl2 was ~70% that of the wild-type intron as measured by Taqman qPCR in plasmid targeting assays ( Fig 6B ) . For in vivo selections , the three Ll . LtrB intron expression plasmids were co-transfected with the recipient plasmid into HEK-293 cells , which were then incubated in culture medium with 80 mM MgCl2 . After 24 h , plasmids were extracted from the HEK-293 cells by an alkaline-lysis procedure and electroporated into E . coli HMS174 ( λDE3 ) to select for TetR colonies , which were screened by colony PCR and sequencing of both 5’- and 3’-integration junctions to confirm retrohoming of the full-length Ll . LtrB intron into the DNA target site ( S6 Fig ) . In controls , no retrohoming was detected by this assay in HEK-293 cells transfected with the same plasmids , but incubated in culture medium without 80 mM MgCl2 . This control confirms that retrohoming detected in the assay occurred in human cells and not after transformation of the donor and recipient plasmids into E . coli , and it provides further evidence that addition of MgCl2 to the culture medium is needed to stimulate Ll . LtrB retrohoming in human cells . We used the HEK-293 cell plasmid selection system to perform eight rounds of in vivo directed evolution in culture medium supplemented with 80 mM MgCl2 via an adaptive walk in which introns that retrohomed into the plasmid target site in each round were amplified by PCR at a relatively high mutagenesis frequency of 3 mutations per intron per round prior to re-cloning into the expression vector for the next round ( Fig 6C ) . After eight rounds , we increased the stringency of the selection by reducing the MgCl2 concentration to 40 mM and performed four additional selection cycles without the addition of new mutations between cycles ( rounds 9–12 ) . The retrohoming efficiency of the selected pools relative to that of the wild-type intron assayed in parallel increased slowly from rounds 6 to 9 and somewhat more rapidly during rounds 10 to 12 . After the 12 rounds of selection shown in the Figure , an additional three rounds of selection with and without mutagenesis gave no further improvement in retrohoming efficiency of the pools relative to the wild-type intron at 40 mM MgCl2 . As described below , high-throughput sequencing indicated that this plateau in retrohoming efficiency reflected that a small number of mutations that moderately enhance retrohoming had overtaken the pool at round 12 and could not be substantially improved by other mutations that were positively selected at either 40 or 80 mM Mg2+ . Although the mutant pools were not increasing in activity at a rapid pace , the possibility remained that individual mutations or combinations of mutations in the pool had enhanced retrohoming . To investigate the mutational diversity of the evolution cycles , we used Pacific Biosciences single-molecule sequencing ( PacBio RS ) , which provides long read lengths ( 1 , 000–15 , 000 nt ) , combined with circular consensus sequencing ( CCS ) , which compensates for sequencing errors by using rolling-circle amplification to generate concatameric-sequencing reads of the same molecule [55] . An advantage of PacBio RS is that it reads single-molecules directly and thus alleviates problems stemming from formation of molecular hybrids during PCR , which can over-estimate the number of unique sequences in molecular diversity experiments [56 , 57] . We further avoided formation of PCR hybrids by preparing the sequencing libraries directly from TetR-positive recipient plasmids that contained integrated introns without PCR . We first sequenced retrohomed introns from round 8 ( NCBI SRA database , accession number SAMN03342363 ) and generated a fitness map that displays the degree of conservation of each nucleotide as a heat map on a secondary structure diagram of the Ll . LtrB intron ( Fig 7 ) . The degree of conservation of different nucleotides displayed a wide range and is shown with a scale ranging from dark to light blue for conserved sites ( 0–0 . 3% mutations ) and from pink to red for mutable sites ( >0 . 3–51% mutations ) ( Fig 7 ) . On average , the round 8 mutant pool contained 4 . 4 mutations per intron . The majority of nucleotides ( 551 of 776 ) in the intron were conserved ( dark or light blue ) over eight cycles of directed evolution . Regions required for ribozyme activity ( e . g . , the catalytic triad in DV , J2/3 , which interacts with DV to form the active site , the branch-point A residue in DVI , and the 5’ and 3’ ends of the intron ) were invariant , with the exception of a few nucleotides previously shown to be less constrained within those regions ( e . g . , the dinucleotide bulge in DV ) . The most variable regions were DIVb , which lies outside the catalytic core , and the two terminal loops of DII . DIVa , which contains a high-affinity LtrA-binding site , showed strong conservation of most nucleotides found to be critical for LtrA binding ( positions 557 , 559 , 561–564 ) , but not position 556 [58 , 59] . A mutation at position 548 in an internal loop in DIVa was positively selected ( green triangle ) and could affect LtrA binding . Although many of the nucleotide changes after 8 cycles of selection appear to be neutral , as they do not bias towards any specific nucleotide , mutations at 25 sites were positively selected ( nucleotides within green triangles in Fig 7 ) , meaning that >2% of the population had a mutation at that position of which >80% had the indicated base . Two of the positively selected mutations were within sequence elements involved in long-range tertiary interactions within the catalytic core ( ζ and θ’ ) , while six of the positively selected mutations disrupted or weakened base-pairing interactions . Mutations at two sites became highly prevalent in the population ( >27% ) . The first was a G282A mutation in EBS1 , which changes a UG to a UA base pair at position -4 of the EBS1/IBS1 interaction between the intron and 5’ exon and had been shown previously to result in an ~50% increase in the efficiency of reverse splicing into a DNA target site in vitro [60] . The second was intron position 642 , which was mutated in 51% of the population at round 8 and 99% at round 12 ( black arrow ) . At round 8 , 63% of the mutations at position 642 were U to A and the other 37% were U to C . Position 642 is located two nucleotides upstream of the transcription start site of the T7 promoter inserted for selection purposes within DIVb . Although mutations at this position could in principle simply attenuate the T7 promoter [61] , leading to less T7-induced toxicity in our E . coli assay , experiments below show that the selected mutations increase retrohoming efficiency in human cells in Taqman qPCR assays . The T7 promoter "TATA-box" region has been shown to interact with TFIID and Pol II in HeLa cell extracts [62 , 63] , and mutations in this canonical “TATA box” could potentially decrease TFIID- and Pol II-binding , leading to increased production of full-length intron transcripts , or could affect retrohoming by some other mechanism . Finally , while the distal stem of DV was mutable , as previously shown in E . coli selections [36] , it was not the site of mutations undergoing positive selection for retrohoming in human cells . This finding is in agreement with the results of S4 and S5 Figs , which show that mutations in the distal stem of DV that increased retrohoming efficiency in E . coli or X . laevis oocyte nuclei , did not increase retrohoming frequency in human cells . To determine whether the mutations that were positively selected in HEK-293 cells at 80 mM Mg2+ ( rounds 1–8 ) were enriched further after more stringent selection without mutagenesis at 40 mM Mg2+ ( rounds 9–12 ( Fig 6C ) , we sequenced retrohoming products from round 12 ( NCBI SRA database , accession number SAMN03342364 ) . In Fig 7 , positions at which the mutation frequency increased or decreased by >2-fold from cycle 8 at 80 mM Mg2+ to cycle 12 at 40 mM Mg2+ are indicted by large green or red arrows , respectively . Surprisingly , over half ( 9 of 16 ) of the positively selected nucleotides that comprised >5% of the population in cycle 8 decreased to less than 0 . 3% of the population in cycle 12 ( red arrows ) . Conversely , six of the eight positively selected mutations that comprised >4% of the population in cycle 12 ( green arrows with indicated nucleotide ) were not prevalent in the population at cycle 8 ( <2% ) . Four of the eight mutations that were positively selected in round 12 weakened or disrupted base pairs in the intron secondary structure . Two positions that were under positive selection at both 80 and 40 mM Mg2+ , the EBS1 mutation G282A and the DIVb mutations U642C and U642A , were present in 64 and 99% of the population , respectively , in cycle 12 . Finally , we identified the top sequencing reads present at highest frequency in cycles 8 and 12 ( S1 Table ) . Many of these contained similar mutations that are candidates for increasing retrohoming activity in human cells . Combinations of these prevalent mutations were tested for linkage disequilibrium ( S2 Table ) to assess covariation between mutations . The majority of mutation pairs had D' values close to 0 , indicating equilibrium , but three mutations in DIVb ( U642A , G651A , and U652C ) compared in pairwise combinations had D' values between 1 and 2 . 3 , suggesting strong covariation . A number of Ll . LtrB variants that were most prevalent in the population and/or contained positively selected nucleotides were assayed for retrohoming in HEK-293 cells with 80 mM MgCl2 added to the culture medium . Ll . LtrB variants having only the mutations G282A ( EBS1 ) or any of the DIVb mutations ( U642A , G651A , U652C ) had retrohoming efficiencies similar to or no greater than 50% better than wild type ( Fig 8 ) . However , the combinations of G282A ( EBS1 ) and either U642C or U642A-G651A-U652C in DIVb had two- to three-fold higher retrohoming frequencies than the wild-type intron ( Fig 8 ) . These findings confirm that selections yielded beneficial mutations that increase retrohoming efficiency with added Mg2+ in human cells . However , all of the beneficial mutations identified lie outside the group II intron catalytic core , the most critical positions of which were invariant in the human cell selections . While the PacBio deep sequencing identified some combinations of Ll . LtrB mutations that increase retrohoming frequency in human cells , separately testing every conceivable combination of mutations is an inefficient means of identifying the best variants for human cells . Instead , we turned to synthetic shuffling [64] of high frequency mutations identified from the fitness maps to screen many mutation combinations at once . Based on the sequencing of variants from rounds 8 and 12 ( Fig 7 ) , we generated a rationally designed synthetic shuffling mutagenesis library by assembly PCR [65] . The library was constructed to test combinations of mutations that showed positive selection and high penetrance during the initial directed evolution ( >80% one nucleotide type present in >5% of the population; subsets of the nucleotides indicated by green triangles or green or black arrows in Fig 7 ) . The library consisted of Ll . LtrB introns in which eighteen such positively selected nucleotides were doped at a 1:1 ratio of the selected to the wild-type nucleotide and position 642 in DIVb was randomized . The library was selected for four cycles of retrohoming in HEK-293 cells at either 80 or 40 mM MgCl2 and tested for retrohoming efficiency compared to the wild-type intron at both Mg2+-concentrations after each cycle . Both selections gave pools of Ll . LtrB variants with increased activity relative to the wild-type intron ( S7 Fig ) , and we then performed PacBio sequencing of the fourth cycle pool for each of the selections ( NCBI SRA database , accession numbers SAMN03342365 and SAMN03342366 ) . The sequencing showed that specific mutations were selected at a number of positions , but these positively selected mutations differed for the selections done at the two different Mg2+-concentrations ( Fig 9A ) . To identify those variations associated with the highest retrohoming activity , we generated separate sequence logos for variants that appeared at least three times in the deep sequencing ( Fig 9B ) . While the positions that were shifting towards the mutant nucleotide were shared between the total sequence reads versus just the highest prevalence sequence reads , the shifts were more pronounced in the latter . Both the EBS1 position 289 and DIVb position 642 mutations were present in 100% of the highest frequency variants . We assayed a number of these high prevalence variants for retrohoming in HEK-293 cells ( Fig 9C–9E ) . All of the variants had 3–4 fold higher frequencies for retrohoming into the plasmid target site than did the wild-type intron . When we tested the best of these variants for retrohoming into the genomic target site , we found that variants 80–4 and 40–1 had about three-fold increased retrohoming frequencies . Although these variants were the best we found , they were only marginally better than the EBS1/DIVb mutation combinations tested in Fig 8 . These findings suggest that the additional positively selected mutations outside EBS1 or DIVb contribute small fitness effects that together lead to increased retrohoming frequencies . The small contributions to enhanced retrohoming by these mutations is consistent with their relatively slow accumulation during the selections compared to the driving mutations in EBS1 and DIVb .
Here we show that a mobile group II intron , the L . lactis Ll . LtrB intron , can retrohome into a chromosomal DNA site in human cells . To do so , we developed a mobile group II intron expression system that overcomes barriers to group II intron proliferation in eukaryotic nuclear genomes , including suboptimal codon usage and translational repression of the intron-encoded RT , NMD of group II intron-containing RNAs , and suboptimal Mg2+ concentrations . NMD was overcome by using phage T7 RNAP rather than Pol II to express the group II intron RNA , while suboptimal codon usage and translational repression were overcome by separately expressing a human codon-optimized group II intron RT from a separate Pol II-transcript . The remaining barrier , suboptimal intracellular Mg2+ concentrations in eukaryotic cells , was overcome simply by adding 80 mM MgCl2 to the cell culture medium . Retrohoming in human cells was demonstrated by sensitive Taqman qPCR assays of both the 5’- and 3’-integration junctions for both plasmid and chromosomal DNA target sites and by conventional PCR and sequencing of recipient plasmids containing fully integrated intron with both of the expected integration junctions . The expression system workarounds enabled the Ll . LtrB intron to splice and retrohome into both plasmid and chromosomal target sites in viable human cells at frequencies up to ~0 . 5% and ~0 . 1% , respectively . However , in vivo selections and synthetic shuffling of positively selected mutations gave only modest further improvements in retrohoming efficiency that still required added Mg2+ in the cell culture medium . The latter findings suggest that low Mg2+ concentrations constitute an effective natural barrier to group II intron proliferation in human cells that is not readily overcome by selecting group II intron variants and may be a major factor in why mobile group II introns failed to persist as such in eukaryotic nuclear genes . The finding that Pol II transcripts containing the Ll . LtrB intron are selectively degraded by NMD in human cells ( Fig 3 ) extends previous findings for S . cerevisiae and suggests that this defense mechanism against mobile group II introns is used generally in eukaryotes [11] . The Ll . LtrB-intron contains multiple stop codons in all three reading frames and could be degraded either by the exon-junction complex ( EJC ) -dependent NMD pathway , if the Ll . LtrB-containing transcript contains cryptic spliceosomal splice sites , or by non-EJC-dependent NMD mechanisms , which are known to operate in mammalian cells [66] . By contrast , a T7 RNAP transcript containing the intron is not subject to NMD and accumulates to the same levels as a parallel control transcript lacking the intron ( Fig 4 ) . Although the T7 RNAP-synthesized Ll . LtrB transcript accumulates to levels sufficient to support retrohoming in human cells , it has a 5’-triphosphate and up-regulates interferon-response genes , such as RIG-I and IFIT1 , which may lead to its sequestration or degradation [45] . Suppression of these innate immune responses could lead to higher levels of T7 RNAP transcripts and retrohoming in human cells than observed here . The finding that supplementation of the cell culture medium with 80 mM Mg2+ was by itself sufficient to enable splicing and retrohoming of T7 transcripts containing the Ll . LtrB intron indicates that intracellular Mg2+ concentrations are limiting for these processes in human cells [67] . This finding extends previous work showing that group II intron RNPs microinjected into Xenopus laevis oocyte nuclei and Drosophila and zebrafish embryos could retrohome efficiently into plasmid target sites only when Mg2+ was injected in addition to the group II intron RNPs [10] . In contrast to yeast , where transcripts containing the Ll . LtrB group II intron RNA are spliced but not translated [11 , 35] , we observed no detectable splicing of Ll . LtrB-transcripts in human cells without Mg2+ supplementation , even when intron RNA degradation by NMD was suppressed . The Pylaiella littoralis Pl . LSU/2 group II intron could also splice in yeast but not in a human cell line ( HCT116 cells; [12] ) . Thus , the intracellular environment in human cells under normal growth conditions appears to be less amenable to group II intron splicing than it is in yeast . Surprisingly , the Mg2+-stimulated splicing of the Ll . LtrB intron in human cells neither required nor was enhanced by the LtrA protein , which is needed for group II intron splicing in bacteria or in vitro [21 , 49] . This IEP-independent splicing could reflect either self-splicing of the Ll . LtrB intron or that human cellular proteins can replace LtrA to stabilize the active intron RNA structure . An intriguing possibility is that the Ll . LtrB intron can be spliced in human cells by a protein evolutionary related to LtrA , such as a LINE-1 or telomerase RT , or the spliceosomal protein Prp8 , which evolved from a group II intron-like RT [32] . Although dispensable for splicing in human cells , the group II intron RT remains essential for retrohoming , where it contributes to DNA target-recognition and is required for target DNA-primed reverse transcription [22 , 68] . The expressed LtrA protein could in principle bind to the group II intron RNA either before or after splicing , the latter being analogous to the reconstitution of active group II intron RNPs in vitro by binding of purified LtrA to self-spliced intron lariat RNA [21] . The similar retrohoming efficiencies when T7 Pol was expressed with or without an NLS ( S1 Fig ) indicate that nuclear transcription and splicing of Ll . LtrB RNA to produce functional RNPs is not required for retrohoming and can also occur from transfected plasmids that remain in the cytosol . Free Mg2+ concentrations may be higher in the cytoplasm than the nucleus , where Mg2+ is sequestered by chelation to chromosomal DNA [69] , thereby favoring group II intron RNA splicing and RNP assembly in that compartment rather than the nucleus . If so , group II intron RNPs may gain access to chromosomal DNA either passively during mitosis or by using a pre-existing RNP transport system . Both mechanisms have been suggested for LINE-1 and other non-LTR-retrotransposon RNPs , which are assembled in the cytoplasm but must gain access to the nucleus for retrotransposition [70–72] . Unlike retrohoming of the Ll . LtrB intron in bacteria , we found that retrohoming of the Ll . LtrB intron into both genomic and plasmid target sites in human cells yields an excess of 3’- over 5’-integration junctions detected by Taqman qPCR assays ( 7–49 fold; Figs 5B–5D and S1 ) . This excess of 3’-integration junctions could reflect the integration of 5’-truncated introns similar to human LINE-1 elements , whose retrotransposition frequently results in the integration of 5’-truncated elements due to abortive reverse transcription [52] . For both group II introns and LINEs , a high frequency of 5’ truncations during retrotransposition could reflect a combination of barriers to reverse transcription , such as RNA-binding proteins , RNase cleavage of the intron or LINE RNA during or prior to cDNA synthesis , and the ability to ligate truncated cDNAs to upstream chromosomal DNA by non-homologous end-joining ( NHEJ ) mechanisms , which are not active in E . coli [73–75] . The excess of 3’-integration junctions for the Ll . LtrB intron could also reflect retrohoming of excised linear intron RNAs , which can carry out only the first step of reverse splicing , resulting in the attachment of the 3’ end of the intron RNA to the 3’ exon; TPRT would then yield a cDNA copy of all or part of the linear intron RNA that is ligated to the 5’ exon by NHEJ but could also potentially remain unattached [73 , 74] . Linear intron RNAs may be generated either by hydrolytic splicing induced by Mg2+ supplementation in the absence of LtrA protein or by debranching of lariat RNAs , possibly via the same enzyme ( Dbr1 ) that functions in the debranching and turnover of spliceosomal intron lariats [76] . The latter could be yet another eukaryotic defense against the proliferation of mobile group II introns . The newly developed mobile group II intron expression system enabled us to select directly for Ll . LtrB intron variants that could retrohome more efficiently in human cells . To do so , we used a plasmid-based mobility assay that enabled selection for low frequency retrohoming events via E . coli transformation and combined it with the long reads of the PacBio RS circular consensus sequencing to identify mutations under positive selection in the evolving populations . Selections at 80 and 40 mM Mg2+ showed that the majority of intron nucleotides were conserved and nucleotides that form the intron RNA’s active site were highly conserved or invariant . Variations were found mainly in terminal loops and at a few scattered positions within the intron . Two mutations , one strengthening the EBS1/IBS1 interaction between the intron and 5’ exon , and the other near the T7 promoter sequence inserted in DIVb , saturated the pool but gave only ~2-fold higher retrohoming efficiency , and other positively selected mutations did not confer substantial additional benefit , even in synthetic shuffling experiments to select for optimal combinations of mutations . Further , mutations selected at 80 mM Mg2+ differed from those selected at 40 mM Mg2 , and Ll . LtrB intron variants selected for enhanced retrohoming in Mg2+-deficient E . coli [36] or X . laevis oocyte nuclei [54] did not show increased retrohoming frequencies in HEK-293 cells . The latter findings may reflect competing effects of altering Mg2+-binding at different sites on intron RNA folding , so that variants selected at one low Mg2+ concentration are not well suited to function at other low Mg2+ concentrations . Previous studies in which variants of the Azoarcus group I intron ribozyme were selected under different conditions showed that different combinations of mutations confer fitness for different environments [77 , 78] . It is possible that very rare mutations not sampled in our selections , different selections , selections with another group II intron , or rational redesign of the group II intron catalytic core based on X-ray crystal structures could yield group II intron variants that retrohome at high frequencies in eukaryotic cells . Until such time , our findings for the Ll . LtrB intron suggest that barriers to group II intron retrohoming in human cells are not readily overcome by mutational variation and selection , possibly reflecting that the group II intron catalytic core cannot be modified readily to function efficiently at lower Mg2+ concentrations . The latter could explain why group II introns failed to evolve into a form that could function in eukaryotes without fragmentation into spliceosomal introns and the spliceosome . Although the Ll . LtrB intron works very efficiently for gene targeting in bacteria [9] , its targeting efficiency via retrohoming in human cells is substantially lower than those for current methods using CRISPR/Cas9 , zinc-finger nucleases or TALEN-based systems [79] . Additionally , retrohoming of the Ll . LtrB intron in human cells requires the addition of Mg2+ to the culture medium , which stresses the cells . Nevertheless , gene targeting efficiencies for the Ll . LtrB intron of near 0 . 1% might be sufficient for gene targeting applications and could potentially be increased substantially by stable rather than transient expression of the group II intron expression plasmids and/or by suppression of innate immune responses and lariat debranching enzyme . It also remains possible that other group II introns can be found that function more efficiently in human cells than does Ll . LtrB . Finally , as DNA target site recognition by mobile group II introns is not dependent upon ribozyme activity , the ability of group II intron RNPs to recognize a DNA target site in the human genome at appreciable frequency as found here suggests they could be used analogously to CRISPR/Cas9 nuclease-null mutants to localize group II intron RT fusion proteins or modified group II intron RNAs with different functionalities to desired chromosomal locations [80] . Mobile group II introns are thought to have evolved in bacteria where the intracellular Mg2+ concentrations are higher than in eukaryotes [1 , 36 , 81 , 82] . They are hypothesized to have entered an ancestral pre-eukaryote , likely an archaeon , with eubacterial endosymbionts that gave rise to mitochondria and chloroplasts , invaded the nucleus , proliferated as mobile elements , and then degenerated with group II intron domains evolving into snRNAs that reconstitute to form the catalytic core of the spliceosome [4 , 34] . Based on their discovery that Pol II transcripts containing the Ll . LtrB group II intron are subject to NMD and translational repression , Belfort and coworkers hypothesized that translational repression resulting from group II intron insertion into protein-coding genes contributed to group II intron loss from eukaryotic nuclear genomes and their evolution into spliceosomal introns [11 , 35] . Considered in the context of the above hypotheses , our results suggest that the ancestral eukaryote must have had relatively high intracellular Mg2+ concentrations that could support proliferation of group II introns in protein-coding genes by retrohoming and that lowering of intracellular Mg2+ concentration in eukaryotes may have been an evolutionary response to selective pressure to restrict group II intron proliferation . Mammals use an analogous defense mechanism based on iron limitation as part of an innate immune response to bacterial infections [83] . In this scenario , a decrease in intracellular Mg2+ concentrations in ancestral eukaryotes would have strongly inhibited group II intron splicing , thereby increasing selective pressure against retaining group II introns as such in protein-coding genes . The evolution of the nuclear membrane , itself hypothesized to be an evolutionary response to group II intron invasion [3] , had the additional advantage of sequestering group II introns into a separate compartment where free Mg2+ concentrations are further decreased by chelation to DNA and chromatin , while enabling the cytosol to maintain higher Mg2+ concentrations for other cellular processes [36 , 67] . A lower free Mg2+ concentration in the eukaryotic nucleus would confer immunity from group II introns that are sporadically acquired by the integration of organellar DNA fragments into nuclear genomes [84] and could resolve the conundrum of why group II introns did not persist in non-coding regions of eukaryotic genomes , where they are not subject to selective pressures caused by translational repression and NMD [13] . Given the inability of multiple group II introns that had inserted into protein-coding genes in an ancestral eukaryote to be cleanly excised simultaneously or to mutate readily into a form that could splice efficiently at low Mg2+ concentration , the evolutionary response was their degeneration into relatively unstructured spliceosomal introns that maintain conserved splice site and branch-point sequences . Reflecting their evolutionary origin , these conserved sequences are recognized by a common splicing apparatus consisting of snRNAs derived from group II intron domains that can now with the aid of proteins promote splicing in the low Mg2+ environment of the eukaryotic nucleus . More generally , our results suggest that differences in intracellular environment had a profound impact on the evolution of introns and gene expression mechanisms in bacteria and eukarya .
Mammalian cells were grown in culture media supplemented with 10% fetal bovine serum ( Gemini Biosystems ) , penicillin , and streptomycin at 37°C with 5% CO2 unless otherwise stated . HEK-293 ( ATCC ) and HEK-293 Flp-In cells ( Invitrogen; Flp-In 293 ) were maintained in Dulbecco's Modified Eagle Medium ( DMEM; Invitrogen ) supplemented with glutaMAX ( Invitrogen ) , and hygromycin B . HeLa cells were maintained in Eagle’s Minimum Essential Medium ( EMEM; Invitrogen ) . COS-7 cells were maintained in DMEM . Antibiotics were added at the following concentrations: ampicillin ( 100 μg/ml ) , carbenicillin ( 150 μg/ml ) , hygromycin B ( 50–100 μg/ml ) , penicillin ( 1 , 000 U/ml ) , streptomycin ( 1 , 000 μg/ml ) , and tetracycline ( 15 μg/ml ) . Transfection reagents were: Fugene 6 ( Roche ) , Lipofectamine 2000 ( Life Technologies ) , Polyfect ( Qiagen ) , and polyethylenimine ( PEI; 40 , 000 linear molecular weight; Polysciences Inc ) . E . coli HMS174 ( λDE3 ) ( Novagen ) was used for the selection of recipient plasmids after retrohoming of the Ll . LtrB intron into the plasmid target site in human cells . Electrocompetent HMS174 ( λDE3 ) were generated as described [10 , 85] and had a transformation efficiency of >2 x 1010 colony-forming units measured using pUC19 plasmid . E . coli strain DH5α was used for cloning . Plasmid phLtrA is a derivative of pAAV ( Stratagene ) that expresses a human codon-optimized LtrA ORF ( hLtrA; see below ) with a 3X myc tag and SV40-NLS fused to its C-terminus . The hLtrA ORF is cloned behind a CMV promoter and followed by a human growth hormone polyadenylation signal . Plasmid phLtrA1 is an earlier hLtrA expression plasmid in which the human codon-optimized LtrA ORF with an SV40-NLS fused to its C-terminus is cloned behind a CMV promoter in a pIRES vector ( Clontech ) . The LtrA ORF contains a small artificial spliceosomal intron , subsequently found to be unnecessary for hLtrA expression , inserted after the start codon and is followed by an SV40 polyadenylation signal . pLtrA is the same except with the native non-codon optimized LtrA ORF . Plasmid pLl . LtrB contains an Ll . LtrB-ΔORF intron RNA ( Ll . LtrB-ΔD4 ( B1-B3 ) [86] ) cloned downstream of a T7 promoter in a TOPO2 . 1 vector ( Invitrogen ) . Variants of this plasmid include pLl . LtrB-GFP in which Ll . LtrB intron and flanking exons interrupts the GFP ORF at position 386; pGFP , which contains a T7-driven GFP ORF with the 35-nt ligated exon sequence that would result from Ll . LtrB intron splicing inserted at position 386; and pLl . LtrB-HPRT and pLl . LtrB-CCR5 in which the wild-type Ll . LtrB-ΔD4 ( B1-B3 ) intron is replaced by one that has been retargeted to insert in the mouse hprt gene ( position 115; [45] ) or human CCR5 gene ( position 332; [6] ) , respectively; pLl . LtrB-T7 is a derivative of Ll . LtrB-ΔD4 ( B1-B3 ) that contains a minimal T7 promoter in DIVb ( positions 627–646 ) ; and pLl . LtrB-stuffer is a derivative that lacks the Ll . LtrB intron and was used for library construction . Plasmid pCMV-BFP , pCMV-BFP-E1E2 , and pCMV-BFP-Ll . LtrB contain the blue fluorescent protein ( BFP ) ORF without or with the ltrB exons 1 and 2 or the Ll . LtrB-ΔD4 ( B1-B3 ) intron flanked by ltrB exons 1 and 2 interrupting the ORF after the start codon cloned in pcDNA5FRT ( Invitrogen ) . Plasmid pT7-NLS contains the T7 RNA polymerase ( T7 RNAP ) ORF with an N-terminal SV40-NLS cloned behind a CMV promoter in pAAV vector ( Agilent ) , and pT7 is the same plasmid containing the T7 RNAP ORF without a NLS . Recipient plasmid pFRT contains a wild-type Ll . LtrB target site ( positions -30 to + 15 from the intron-insertion site ) inserted into the Flp-In recombinase site of pcDNA5/FRT ( Life Technologies ) . The target site region is identical to that inserted into the HEK-293 Flp-In genome . Recipient plasmid pBRRQ is a derivative of pBRR-Tet [6] and contains a wild-type Ll . LtrB target site ( positions -30 to +15 from the intron-insertion site ) flanked by sequences with Tm values optimized for qPCR ( S1 Table ) cloned upstream of a promoter-less tetR gene . Recipient plasmid pBRR-CCR5 is identical to pBBRQ except for containing the CCR5 targetron insertion site ( positions -30 to +15 from the intron insertion site ) . All recipient plasmids carry an ampR marker . The human codon optimized LtrA sequence was generated from overlapping oligonucledotides by assembly PCR [65] . Oligonucleotides containing hLtrA sequence were synthesized by HHMI/Keck Oligonucleotide Synthesis Facility ( Yale ) and PCR reactions were carried out by using Vent DNA polymerase ( New England Biolabs ) , high annealing temperatures ( 58–60°C ) , and manual hot start–i . e . , adding Vent DNA polymerase after sample temperature reached 94°C ) . PCR products were gel-purified and digested with EcoRI and XbaI or HindIII and XbaI , then cloned into pKSBluescript ( Agilent ) to form pKS-hLtrA and confirmed by sequencing . The assembled ORF was re-cloned into a pIRES vector ( Clontech ) to generate phLtrA1 . HEK-293 cells were seeded at equal density into 96-well white plates ( Corning ) , allowed to grow out , and transfected using Fugene 6 ( Roche ) according to manufacturer’s recommendations . After 48 h in culture , cytotoxicity analysis was carried out using the CellTiter-Glo direct lysis kit ( Promega ) according to manufacturer instructions . Luciferase activity was measured on a Mithras Multimode Platereader ( Berthold ) . Trypan blue staining was performed by mixing 10 μl of cells with 10 μl of trypan blue solution ( 0 . 4%; Invitrogen ) and then counting stained and unstained cells on a hemacytometer . For immunoblotting , cells were collected and boiled in 1x Laemmli gel buffer for 5 min . After pelleting insoluble material by centrifugation in a microfuge for 2 min at top speed , the protein samples prepared from the same number of cells were run in 8% polyacrylamide/0 . 1% SDS gel , which was then blotted to a nitrocellulose membrane using a Hoefer SemiPhor blotter ( Amersham ) . Anti-LtrA antibody [49] was used at 1:1 , 000 dilution , and goat anti-rabbit secondary antibody ( Pierce ) was used at 1:60 , 000 dilution , both at room temperature . After developing the immunoblot , the membrane was stained with AuroDye to confirm even loading . For immunofluorescence , cells were washed twice with phosphate buffered saline ( PBS ) and then fixed in 2% paraformaldehyde for 30 min at room temperature . After three more washes with PBS , cells were permeabilized by incubating in 0 . 5% Triton X-100 in PBS for 15 min , followed by three washes with PBS containing 0 . 2% Tween 20 ( PBST ) . Blocking was achieved by incubating the permeabilized cells with 10% normal goat serum and 1% BSA in PBST for 1 h . Primary antibody was pre-incubated with untransfected cell lysate ( prepared by sonication ) to deplete nonspecific antibodies and then incubated with cells at 1:5 , 000 dilution in blocking buffer for 1 h at 4°C . After four 5-min washes in PBST containing 0 . 1 M NaCl , cells were incubated with 1:100 dilution of goat anti-rabbit antibody conjugated with fluorescein in blocking buffer for 1 h , washed with PBST containing 0 . 1 M NaCl five times for 5 min each time , incubated with 2 μg/ml Hoechst dye for 10 min , and washed twice with PBS . Cells were mounted and observed under a fluorescence microscope ( Olympus CKX41 ) . HEK-293 cells were grown to confluence , washed with PBS , blown off the dishes with ice-cold hypotonic buffer ( 10 mM HEPES , 10 mM KCl , 1 ml/100 mm dish ) , and incubated on ice for 15 min . Cells were broken by 15 strokes of a Dounce homogenizer . Nuclei were collected by centrifugation at 800 x g for 5 min at 4°C and then resuspended in the residual buffer in the same tube . After 3 cycles of freezing and thawing , chromosomal DNA was sheared by repeated pipetting , and 5 μl of the solution was used for each reaction . RT assays with Ll . LtrB/E2+10 substrate were carried out as described [47 , 49] in 10 μl of reaction medium containing 5 μl lysate , 40 nM Ll . LtrB template , 400 nM E2+10 primer , 450 mM NaCl , 5 mM MgCl2 , 40 mM Tris-HCl , pH 7 . 5 plus 10 μCi [α-32P]dTTP ( 3 , 000 Ci/mmol; New England Nuclear ) and 0 . 2 mM of each dNTP . The Ll . LtrB/E2+10 substrate consists of Ll . LtrB RNA ( an in vitro transcript containing the Ll . LtrB-∆ORF intron and flanking exons ) with a 20-mer DNA primer ( E2+10 ) annealed to a position in the 3’ exon that corresponds to that of the cleaved bottom strand normally used as the primer for target DNA-primed reverse transcription of the intron RNA during retrohoming . Reactions were initiated by adding dNTPs and incubated at 30°C for 30 min . Incorporation of [α-32P]dTTP was measured by spotting onto DE81 paper ( Whatman ) and counting Cherenkov radiation in a scintillation counter ( LS6500 , Beckman ) . UPF1 and scramble siRNAs ( Dharmacon ) were transfected into ~60% confluent HeLa or HEK-293 cells 24 h prior to transfection of BFP- or GFP-containing plasmids . UPF1 levels were measured in equivalent amount of proteins from crude cell lysates via SDS-PAGE ( 4–12% polyacrylamide gradient gel ) and immunoblotting using a Trans-Blot Turbo system ( Bio-Rad ) to blot the gel to a nitrocellulose membrane , which was then probed with an anti-UPF1 antibody ( ab10510; Abcam ) . Plasmid and siRNA transfections were carried out using Dharmafect as described [87] . For analysis of transcript levels and splicing via RT-qPCR and RT-PCR , respectively , RNA was purified from transfected cells using the ZR RNA Miniprep Kit ( Zymo ) . 1 μg of each RNA sample was treated with DNase I ( Invitrogen ) at 37°C for 1 h to remove DNA and then converted to cDNA with a SuperScript III reverse transcriptase kit ( Invitrogen ) according to manufacturer’s recommendations . RT-PCR was carried out with GC-rich Phusion polymerase mastermix ( New England Biolabs ) under standard conditions , unless otherwise indicated . RT-qPCR was carried out using Power SYBR Green Master Mix ( ABI ) on an Applied Biosystems Viia7 system in 96-well format under standard conditions . For the CMV-BFP cassettes , the primers were pAAV MCSfw 5’ TCTTATCTTCCTCCCACAGCTCCT and GFP-L qPCRrev 5’ TCGTCCTTGAAGAAGATGGTG , and for the T7-GFP cassette , the primers were pTOPOsplicinginfw 5’ TGTCTTCTTGACGAGCATTCC and pTOPOsplicinginrev 5’ TAGGTCAGGGTGGTCACGA . Retrohoming of the Ll . LtrB intron in mammalian cells was assayed by Taqman qPCR using an Applied Biosystems Viia7 system in 384-well format using Taqman probes ( Life Technologies ) . Reactions were performed in technical triplicate in 10-μl volumes for 35 ( plasmid ) or 40 ( genomic ) cycles using Taqman PCR universal mastermix ( Applied Biosystems ) under standard conditions . Standard curves for quantitation used four 10-fold dilutions of either pBRRQ or pFRT plasmid containing an integrated Ll . LtrB intron and had >90% efficiency across the range of concentrations used . Standard curve plasmids were quantified using a Qubit system ( Life Technologies ) . Standard curve dilutions were buffered with 10 ng/μl phage lambda DNA carrier . The primer/probe sets are shown in S3 Table . HEK-293 Flp-In cells ( Invitrogen ) contain a FRT recombinase site in a decondensed region of the genome . A single copy of the wild-type Ll . LtrB insertion site ( position -30 to +15 from the intron-insertion site ) was recombined into the FRT site genomic locus according to manufacturer's recommendations . For retrohoming experiments , HEK-293 Flp-In cells containing the Ll . LtrB target site were seeded in multi-well culture plates ( Corning ) 24 h prior to transfection to reach a confluency of 60–80% on the day of transfection . Cells were dissociated using Stem Pro Accutase ( Invitrogen ) , and cell counting was performed with a hemocytometer or using the Scepter system ( Millipore ) . For genomic targeting experiments , the Ll . LtrB intron expression plasmids , pLl . LtrB , pT7-NLS , and phLtrA were transfected at 276 ng each with 2 . 76 μg branched polyethyleneimine ( PEI ) ( Polysciences , Inc ) per well in a 12-well culture plate for 24 h . For plasmid targeting experiments , recipient plasmid pFRT or pBRRQ was included at 276 ng per well in addition to the above three plasmids . After 24 h , the media was removed and replaced with growth medium supplemented with MgCl2 or other Mg2+ salts for an additional 24 h unless otherwise specified . The next day , when the cells were typically 80–90% confluent , non-adherent cells were removed by vigorously rinsing with PBS three times , and adherent cells were collected into a 1 . 5-ml snap-tube unless otherwise specified . Total DNA was extracted from cell pellets with a Qiagen Blood and Tissue kit with an RNase step or the ZR-genomic miniprep kit ( Zymo research ) according to manufacturer's recommendations . In plasmid targeting experiments , plasmids were extracted from cells using alkaline lysis with the Wizard SV-miniprep system ( Promega ) or total DNA using the ZR-genomic miniprep kit ( Zymo Research ) . Experiments typically used three wells that had been independently seeded and transfected in parallel for determination of SEMs . Biological replicates were performed on separate days and reported with SDs . pLl . LtrB-T7 mutant libraries for each selection cycle were generated by PCR with Mutazyme II ( Stratagene ) according to the manufacturer’s recommendations for 3 mutations per kb . Approximately 200 ng Ll . LtrB DNA template was mutagenized in a 50-μl PCR with primers 309S 5’- CACATCCATAACGTGCGCC and 308A 5’- TAATTGCTAGCCGGCCGCATTAAAAATGATATG for 30 cycles , and then re-amplified to obtain a higher yield using Phusion polymerase ( New England Biolabs ) . The PCR product was purified from an agarose gel stained with Sybr gold ( Invitrogen ) under blue-light illumination and then digested overnight with AatII and NheI-HF ( New England Biolabs ) . After purification , 750 ng of the insert was ligated to 1 μg of linearized and dephosphorylated pLl . LtrB-stuffer for 2 h at room temperature in a volume of 400 μl using T4 DNA ligase ( 4 , 000 units; New England Biolabs ) . The ligation mix was purified and concentrated to a volume of 6 μl using a Zymo clean and concentrator column and then electroporated into 100 μl E . coli MegaXDH10B cells ( Invitrogen ) with total transformants typically reaching >2 x 108 . The resulting library was purified by using an Endotoxin-free MiniKit II ( Omega Biosciences ) and transfected into HEK-293 Flp-In cells for both targeting and selection experiments . In vivo selections in HEK-293 cells were done using a modification of a previously described E . coli plasmid-based retrohoming assay in which a group II intron with a phage T7 promoter inserted in DIVb integrates into a target site cloned in a recipient plasmid upstream of a promoterless tetR gene , thereby activating that gene [10 , 36] . HEK-293 cells were transfected with plasmids for the hybrid Pol II/T7 expression system ( Fig 1 ) , with pLl . LtrB replaced with pLl . LtrB-T7 , which contains a minimal T7 promoter in DIVb , and pBRRQ , which contains an Ll . LtrB target site cloned upstream of a promoter-less tetR gene . After 24 h , plasmids were isolated from transfected cells by alkaline lysis using the Wizard SV plasmid miniprep kit ( Promega ) . An aliquot was diluted and used for Taqman qPCR and the rest was concentrated to 6 μl using a Zymo clean and concentrator column . The concentrated plasmid was electroporated into 100 μl of electrocompetent E . coli HMS174 ( λDE3 ) cells , which were then plated onto LB-agar plates containing tetracycline ( 15 μg/ml ) and grown for 2 days . The resulting colonies were pooled , and the TetR plasmids were isolated by alkaline lysis using a Wizard SV miniprep kit ( Promega ) . Ll . LtrB introns that had successfully retrohomed into the TetR-recipient plasmids were PCR amplified by 21 cycles of PCR with or without mutagenesis as described above using primers that flank the integration site ( primers 200S and 269A; S3 Table ) , and the PCR product was isolated from an agarose gel and used to generate a library for the next round of selection . Assembly PCR was used to generate the synthetically shuffled library [65] . Briefly , multiple 80-120-mer oligonucleotides spanning the length of the intron and containing the randomized or doped positions of interest and complementary overlaps with a Tm of ~55°C were synthesized at the Center for Systems and Synthetic Biology at UT-Austin . For each intron library , the assembly PCR was done with a 500-ng equimolar mix of oligonucleotides for 25 cycles under standard conditions in 50 μl of Phusion PCR mastermix . A 5-μl aliquot was placed in 300 μl of Phusion PCR mix with forward and reverse primers that synthesize the full-length intron and run for an additional 25 cycles . The full-length product was purified by electrophoresis in an agarose gel and used to construct libraries in pLl . LtrB , as described above . Libraries for Pacific Biosciences RS circular consensus sequencing ( CCS ) were generated according to manufacture's recommendations for A-tailed inserts , and sequencing was performed at the Johns Hopkins University Medical School deep sequencing and microarray core facility . Inserts for PacBio sequencing were generated directly from pooled TetR-positive plasmids isolated after directed evolution cycles by digesting >50 μg of plasmid DNA with AatII and EcoRI-HF ( New England Biolabs ) at sites 37-nt upstream and 16-nt downstream of the Ll . LtrB-integration site , respectively , and then purifying the resulting restriction fragment in a 1% agarose gel under blue light using Sybr Gold staining . To assess the sequencing error-rate for the PacBio CCS , we sequenced the wild-type intron and determined the number of substitutions , insertion , and deletion errors . With three rolling-circle sequencing passes of the intron , the substitution error rate was <0 . 01% . The insertion and deletion ( indel ) rates were 0 . 21 and 0 . 07% respectively , and these occurred predominantly at homopolymeric regions . Sequence reads were filtered to remove reads that did not reach at least three circular passes . Raw sequence reads in the FastQ file format were aligned to the wild-type Ll . LtrB reference sequence using Mosaik Aligner 1 . 0 ( https://code . google . com/p/mosaik-aligner/ ) and text files were extracted using the Tablet browser [88] . Insertion gaps were removed using a Perl script , Gapstreeze , available online at ( http://www . hiv . lanl . gov/content/sequence/GAPSTREEZE/gap . html ) , and reads containing deletion-errors were removed . Aligned sequences were then analyzed for nucleotide variation using a Perl script courtesy of Dr . Scott Hunicke-Smith ( UT-Austin ) . All other data analysis , including calculation of nucleotide frequencies and analysis of co-variations was performed using Unix shell scripts , including grep , cut , uniq , sort , and awk . Standard linkage disequilibrium was calculated as D = ( PAB x Pab ) - ( PAb x PaB ) , where PAB is the frequency at which the mutations occur together , PAb and PaB are the mutations occurring independently , and Pab the frequency at which neither occurred . The normalized linkage disequilibrium ( D' ) was calculated by dividing positive D values by the theoretical maximum co-occurrence and negative D values by a theoretical minimum co-occurrence based on the observed individual frequencies in the population . The significance of these values was measured with the r2 value ( the square of the correlation coefficient ) calculated as r2 = D2/PaPbPAPB , and χ2 which is r2 multiplied by the number of sequences analyzed [77] . The Pacific Biosciences sequencing data are available at the NCBI SRA database ( Biosample accession numbers: SAMN03342363 , SAMN03342364 , SAMN03342365 and SAMN03342366 ) . The hLtrA sequence is available from NCBI Genbank ( accession number KP851976 ) . The primary data underlying the Figures are available in S1 Data . | Mobile group II introns are bacterial retrotransposons that are evolutionary ancestors of spliceosomal introns and retroelements in eukaryotes . They consist of an autocatalytic intron RNA ( a ribozyme ) and an intron-encoded reverse transcriptase , which together promote intron mobility to new DNA sites by a mechanism called retrohoming . Although found in bacteria , archaea and eukaryotic organelles , group II introns are absent from eukaryotic nuclear genomes , where host defenses impede their expression and lower intracellular Mg2+ concentrations limit their ribozyme activity . Here , we developed a mobile group II intron expression system that bypasses expression barriers and show that simply adding Mg2+ to culture medium enables group II intron retrohoming into plasmid and chromosomal target sites in human cells at appreciable frequencies . Genetic selections and deep sequencing identified intron RNA mutations that moderately enhance retrohoming in human cells , but not without added Mg2+ . Thus , low Mg2+ concentrations in human cells are a natural barrier to efficient retrohoming that is not readily overcome by mutational variation and selection . Our results have implications for group II intron use for gene targeting in higher organisms and highlight the impact of different intracellular environments on intron evolution and gene expression mechanisms in bacteria and eukarya . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Retrohoming of a Mobile Group II Intron in Human Cells Suggests How Eukaryotes Limit Group II Intron Proliferation |
Establishment of sister chromatid cohesion is coupled to DNA replication , but the underlying molecular mechanisms are incompletely understood . DDX11 ( also named ChlR1 ) is a super-family 2 Fe-S cluster-containing DNA helicase implicated in Warsaw breakage syndrome ( WABS ) . Herein , we examined the role of DDX11 in cohesion establishment in human cells . We demonstrated that DDX11 interacts with Timeless , a component of the replication fork-protection complex , through a conserved peptide motif . The DDX11-Timeless interaction is critical for sister chromatid cohesion in interphase and mitosis . Immunofluorescence studies further revealed that cohesin association with chromatin requires DDX11 . Finally , we demonstrated that DDX11 localises at nascent DNA by SIRF analysis . Moreover , we found that DDX11 promotes cohesin binding to the DNA replication forks in concert with Timeless and that recombinant purified cohesin interacts with DDX11 in vitro . Collectively , our results establish a critical role for the DDX11-Timeless interaction in coordinating DNA replication with sister chromatid cohesion , and have important implications for understanding the molecular basis of WABS .
Cohesion is the process that ensures tethering of newly replicated sister chromatids until they separate in metaphase [1] . This process is mediated by cohesin , an evolutionarily conserved hetero-tetrameric complex ( made of Smc1 , Smc3 , Scc1 and either SA1 or SA2 subunits ) , which has a ring-like structure and is believed to encircle DNA [2–4] . Several proteins interact with cohesin during different phases of the cell cycle and regulate its association with chromatin . In mammalian cells , cohesin is loaded onto DNA in telophase by the action of the loader complex ( Scc2-Scc4 ) [5] . During G1 phase , cohesin association with chromatin is dynamic and cohesin can be unloaded by the activity of the Wapl-Pds5 complex . In S phase ( and subsequent G2 ) binding of cohesin to chromatin becomes stable , as a consequence of acetylation of the Smc3 subunit by two dedicated acetyltransferases ( Esco1 and Esco2 ) [6–7] . This process , known as cohesion establishment , renders cohesin resistant to the action of Wapl-Pds5 and is believed to take place at the replication fork [8–10] . Genetic inactivation of a number of DNA replication factors results in cohesion defects in yeast and in mammalian cells [11–12] . An important role in the chromosomal cohesion process is played by Timeless ( Tof1/Swi1 in yeast ) , which , together with Tipin ( Csm3/Swi3 in yeast ) and Claspin ( Mrc1 in yeast ) , forms the replication fork-protection complex ( FPC ) . The FPC has multiple important functions for maintaining genome stability during DNA replication [13] . First , the FPC is a mediator of the S phase checkpoint promoting ATR-mediated Chk1 phosphorylation . Second , the FPC plays roles that are independent of the S phase checkpoint . It associates with the advancing replisomes and prevents uncoupling of replicative DNA polymerases from the DNA helicase , when DNA synthesis is halted at sites of DNA damage or at natural replication fork barriers . Lastly , FPC components promote chromosomal cohesion in various systems , including yeasts [14–15] , Caenorhabditis elegans [16] , Xenopus laevis egg extracts [17–18] and human cells [19–20] . Genetic studies in yeast have revealed a functional link between the FPC and the cohesion establishment factor Chl1 ( Chromosome loss 1 protein ) [14 , 21–22] . Chl1 , also known as ChlR1 or DDX11 in metazoans , is a super-family 2 ( SF2 ) ATP-dependent DEAH-box DNA helicase that unwinds DNA with a 5’-to3’ directionality [23] . Human DDX11 shares sequence similarity with the Fe-S cluster-containing DNA helicases FANCJ , XPD and RTEL1 . All of these helicases play important roles in genome stability maintenance and are implicated in rare genetic syndromes and cancer development [24–25] . DDX11 is genetically linked to the Warsaw breakage syndrome ( WABS ) , a rare hereditary disease . WABS-affected individuals display a complex pattern of clinical manifestations , including reduced growth , skin rash , heart defects , deafness , and intellectual disability . At the cytological level , WABS patient cells exhibit increased drug-induced chromosomal breakage and sister chromatid cohesion defects [26–27] . We have previously demonstrated that DDX11 and Timeless physically and functionally interact and operate in concert to preserve replication fork progression in stressful conditions in HeLa cells [28] . Nonetheless , the precise molecular mechanism , by which DDX11 and Timeless cooperate with other components of the replication machinery and/or the cohesin complex to promote genomic stability and sister chromatid cohesion , has not yet been elucidated . Herein , we identify a Timeless-binding motif in DDX11 and show that mutations of this sequence compromise the DDX11-Timeless interaction . We demonstrate that DDX11 mutants defective in Timeless binding are unable to rescue sister chromatid cohesion defects of DDX11-depleted HeLa cells . Conversely , DDX11 helicase-dead mutants partially revert the loss-of-cohesion phenotype of these cells . These results suggest that the interaction of DDX11 with Timeless is critical for sister chromatid cohesion . Besides , we demonstrate that DDX11 and cohesin associate with replication forks in HeLa cells and this association is reduced when DDX11 is down-regulated . In addition , we show that DDX11 interacts with the cohesin complex in cell extracts and in vitro . Overall , our data suggest that DDX11 has a scaffolding function in sister chromatid cohesion by anchoring the cohesin complex to the replication machinery and underscore the importance of the DDX11-Timeless interaction for linking replication fork progression to chromosomal cohesion in human cells .
We previously demonstrated that human DDX11 and Timeless directly interact and collaborate to preserve replication fork stability [28] . To identify DDX11 residues responsible for Timeless binding , we carried out an analysis based on tiling peptide microarrays that covered the entire length of the DDX11 sequence . These arrays consisted of 454 15-residue long peptides that were "printed" in duplicate on a glass slide . They were probed with purified recombinant Flag-Timeless and subsequently detected with a fluorescently labelled anti-Flag antibody . As the negative control , an identical peptide array was subjected to mock incubation with the anti-Flag antibody , but without Flag-Timeless . As shown in Fig 1A , two main interaction spots were identified , which were not present in the negative control . These spots were centred around Peptide # 32 and # 44 , which map to the N-terminal portion of DDX11 between helicase motifs I and Ia ( Fig 1B ) . A multiple sequence alignment revealed that this region of human DDX11 ( here named Region T , residues 65–225 ) forms an insertion that is shared only by FANCJ , but not other SF2 Fe-S DNA helicases ( see S1 Fig ) . According to a DDX11 three-dimensional model based on the Thermoplasma acidophilum XPD crystal structure [29] , Region T is predicted to reside on the protein surface in the RecA-homology domain 1 ( HD1; see Fig 1C ) . To identify amino acid residues critical for Timeless binding , we used microarrays containing a full substitution scan of DDX11 Peptide # 32 . In these arrays , each residue of Peptide # 32 was substituted with all 20 natural amino acids . We found that substitution of the two C-terminal residues of Peptide # 32 ( corresponding to Glu201 and Tyr202 of full-length DDX11 ) with lysine completely abolished the interaction with Timeless ( S2 Fig ) . Other changes of the same residues had a less drastic effect on Timeless binding . Then , we carried out site-directed mutagenesis studies of full-length DDX11 to validate the importance of the above residues for Timeless binding ( Fig 1D and 1E ) . We noticed that DDX11 Glu201 and Tyr202 belong to a short highly conserved sequence that we named "EYE" motif . A multiple sequence alignment revealed that this motif is invariant in DDX11 orthologs from vertebrates , whereas it is only partially conserved in DDX11 proteins from fruit fly , worm , budding yeast and fission yeast ( S3B Fig ) . Residues of human DDX11 "EYE" motif were substituted to produce the mutants that were named DDX11 KAE and KAK . We observed an almost complete loss of interaction between Timeless and the DDX11 KAK mutant , when co-pull down experiments were performed in vitro on mixtures of these proteins produced in the recombinant form ( Fig 1D ) . Moreover , interaction of the DDX11 KAE and KAK mutants with the endogenous Timeless was examined by co-immuno-precipitation experiments performed on whole extracts of HEK 293T cells ectopically expressing these DDX11 mutant forms . These analyses revealed that the above DDX11 amino acid changes strongly reduced Timeless binding in human cells ( Fig 1E ) . Therefore , the conserved "EYE" motif of DDX11 is critical for Timeless binding , although we cannot completely exclude that other contact sites could exist between the two proteins . Besides , as the association between the DDX11 KAK and KAE mutants and Timeless is not completely abolished in whole cell extracts , additional protein factors could mediate DDX11:Timeless interaction in vivo . We then examined the relevance of the DDX11-Timeless interaction in sister chromatid cohesion in interphase and mitosis . These analyses were carried out in a HeLa cell line where DDX11 was stably knocked-down . This cell line ( named HeLa 5–5 ) was established by infection with a pantropic retrovirus ( pSuper-Retro-Puro ) that expresses a shRNA targeting the DDX11 coding sequence . At the same time , a control cell line ( named HeLa C1 ) was obtained by infection with an empty retrovirus , as previously described [30–31] . Centromeric cohesion was examined in metaphase chromosome spreads by indirect immuno-fluorescence with the human CREST antibody that specifically recognizes inner centromere/kinetochore proteins ( Fig 2A and 2B ) . As expected , the majority ( about 82% ) of control cells displayed sister chromatid pairs with a typical tight primary constriction . By contrast , a high proportion ( about 73% ) of DDX11-depleted HeLa cells had metaphase chromosomes with a loosened centromere constriction . A small fraction of cells gave rise to metaphase chromosome spreads with a total premature chromatid separation . These findings are consistent with previous reports showing that DDX11 is required for proper chromosomal cohesion [30 , 32] . The DDX11-depleted HeLa cells were transiently transfected with vectors expressing wild type DDX11 and the KAE and KAK mutants to assess the ability of these proteins to rescue the observed chromosomal cohesion defect ( Fig 2B ) . As shown in Fig 2C , Flag-tagged DDX11 mutants were expressed at a level that was comparable with that of the wild type protein . The complementation assays indicated that ectopically expressed wild type DDX11 was able to efficiently rescue the loss-of-cohesion phenotype of the DDX11-depleted cells , with about 72% of spreads having cohered chromatid pairs . In contrast , over-expressed DDX11 KAE and KAK mutants did not revert the chromosomal cohesion defect and about 71–73% of the examined spreads displayed chromosomal cohesion anomalies . These results indicate that the interaction of DDX11 with Timeless is needed for proper sister chromatid cohesion . Previous biochemical studies revealed that substitution of Lys50 with Arg in the helicase motif I ( Walker A ) of human DDX11 ( DDX11 K50R ) abolished ATP binding/hydrolysis and DNA unwinding , but not DNA-binding activity [33] . Besides , human DDX11 with substitution of Gln23 with Ala in the so-called conserved Q motif ( DDX11 Q23A ) was reported to be completely unable to bind/hydrolyze ATP and bind/unwind DNA [34] . We analyzed the ability of these two DDX11 helicase-dead mutants to correct the chromosomal cohesion defects observed in the DDX11-depleted HeLa cells . Our complementation studies revealed that either DDX11 K50R or Q23A was able to correct the centromeric cohesion defect of DDX11-depleted HeLa cells , although not as efficiently as the wild type protein: about 55% and 53% of chromosome spreads showed normal chromosomal pairing , respectively ( Fig 2B ) . We then examined the Timeless-binding capability of DDX11 K50R and Q23A helicase-dead mutants by co-pull down experiments in whole cell extracts , and found that it was not reduced as compared to wild type DDX11 ( Fig 2D ) . These results revealed that DDX11 has a role in sister chromatid cohesion that is not strictly dependent on its catalytic functions ( ATP-binding/hydrolysis and DNA-binding/unwinding ) . We also analysed the effect of Wapl down-regulation in DDX11-depleted HeLa cells and found that the percentage of mitotic cells with premature chromatid separation was reverted to a normal level in cells where expression of the cohesin releasing factor was knocked-down ( see S4A and S4B Fig ) . Thus , these results suggest that at least a function of DDX11 in cohesion establishment is to counteract Wapl . We note , however , that the depletion of DDX11 by siRNA is incomplete . It remains to be tested whether Wapl depletion can rescue the cohesion defects caused by the complete loss of DDX11 . Besides , we found that down-regulation of DDX11 did not substantially affect the level of acetylated Smc3 in HeLa cells ( see S4C Fig ) . Next , we evaluated the ability of the above DDX11 mutants to revert the chromosomal cohesion defects of DDX11-depleted cells in G2 phase by a fluorescence in situ hybridization ( FISH ) assay , using a probe specific for a chromosome 3 locus ( Fig 3A–3C ) . After an overnight thymidine block , HeLa cells were released in fresh medium for 4 hr to enrich them in late S/G2 phase . Distance between FISH signals in G2 phase nuclei was measured to examine the cohesion status in single cells . We found that cohesion defects caused by DDX11 loss were rescued by the wild type protein and the helicase-dead mutants ( K50R and Q23A ) . In contrast , the Timeless-binding defective mutants ( KAE and KAK ) were unable to restore a normal distance between paired FISH dots ( Fig 3B ) . These results revealed that the DDX11-Timeless interaction is critical for chromosomal cohesion even in interphase nuclei . We have recently shown that Mcm2–7-dependent cohesin loading in early S phase is critical for cohesion establishment in human cells [35] . We checked if DDX11 is also required for cohesin loading in S phase . In HeLa cells stably expressing Scc1-Myc we depleted DDX11 using siRNA and examined the level of chromatin-bound Scc1-Myc . In cells arrested in early S phase with thymidine treatment , the intensity of Scc1-Myc was consistently reduced with four individual DDX11 siRNAs and of all four siRNAs ( Fig 4A and 4B ) . Depletion of DDX11 with each DDX11 siRNA was efficient ( Fig 4C ) . Then , we extended this analysis to the cohesin SA2 subunit and found that its association to chromatin was also reduced in DDX11-depleted HeLa cells synchronised in S phase ( S5 Fig ) . Moreover , rescue experiments indicated that Scc1-Myc loading onto chromatin was partially restored by the DDX11 helicase-dead mutants ( K50R and Q23A ) , but not by the Timeless-binding defective mutants ( KAE and KAK ) ( S6 Fig ) . Collectively , these results suggest that DDX11 promotes stable association of cohesin to chromatin during S phase in a way that is dependent on its direct interaction with Timeless . Then , we examined if DDX11 was able to interact with cohesin . To this end , we carried out co-pull down experiments using purified recombinant proteins . Flag-tagged DDX11 wild type and KAK mutant were purified from transiently transfected mammalian cells . Human cohesin was produced in insect cells infected with a single multi-gene baculovirus co-expressing the four core complex subunits ( Smc1 , Smc3-Flag , Scc1 , 10xHis-SA1; S7 Fig ) . Co-pull down experiments with an anti-DDX11 antibody bound to Protein A Sepharose beads revealed that either wild type DDX11 or the KAK mutant was able to bind the cohesin complex ( Fig 5A ) . Besides , we carried out co-immunoprecipitation experiments with anti-Flag beads on extracts of HEK 293T cells transiently transfected with vectors expressing a Flag-tagged DDX11 wild type and its mutants , and found that endogenous cohesin was co-pulled down with each of these proteins ( Fig 5B ) . These findings suggest that DDX11 physically associates with cohesin , and this interaction does not require the "EYE" motif or the helicase activity of DDX11 . Chl1 , the budding yeast counterpart of human DDX11 , was reported to engage with cohesin during S phase in the context of the replication fork [36] . To examine whether human DDX11 and cohesin associate with the advancing replisomes , we carried out co-immuno-precipitation experiments with an anti-Cdc45 antibody bound to Protein A Sepharose beads on the nuclear fraction of control ( HeLa C1 ) and DDX11-depleted HeLa cells ( HeLa 5–5 ) . Cdc45 is an accessory subunit of the replicative DNA helicase ( the Cdc45/Mcm2-7/GINS , CMG , complex ) [37] . As shown in Fig 5C , Western blot analysis of the pulled down samples revealed the association to Cdc45 of DDX11 , cohesin ( Smc3 subunit; binding of Smc1 is shown in S8A Fig ) , Timeless and the Mcm4 protein . Depletion of DDX11 reduced the amount of cohesin that was co-immunoprecipitated with Cdc45 , but had no effect on the pull-down of Mcm4 and Timeless ( Figs 4C and S8A ) . Since Cdc45 associates with chromatin only in S phase as a stable component of the CMG complex together with GINS [37] , our results suggest that DDX11 is bound to the advancing replisomes and plays a critical role in anchoring cohesin to the replication forks . While evidences of an association of cohesin and many cohesin regulators ( such as Esco2 , the Scc2-Scc4 loader and the Wapl-Pds5 releasing complex ) to the replication machinery were reported by various experimental approaches in mammalian cells [35 , 38–40] , localisation of DDX11 at sites of DNA synthesis has remained elusive so far . To further investigate this issue , we used the in situ analysis of protein interactions at DNA replication forks ( SIRF ) technique [41] . As schematically depicted in Fig 6A , this novel methodology is based on a proximity ligation assay ( PLA ) coupled to 5'-ethylene-2'-deoxyuridine ( EdU ) click-it chemistry to identify co-localisation of proteins of interest to nascent DNA in single cells . This analysis revealed that the number of DDX11-EdU PLA spots was significantly higher in EdU-treated cells than in control cells ( see Fig 6B ) , indicating that DDX11 localises at sites of DNA synthesis . Finally , we carried out an additional set of co-immunoprecipitation experiments on extracts of HEK 293T transiently expressing Flag-tagged DDX11 wild type and mutant proteins and found that association of DDX11 KAE and KAK mutants to various components of the DNA replication machinery ( such as Timeless , WDHD1 and Cdc45 ) was noticeably reduced relatively to the wild type protein ( S8B Fig ) . In contrast , binding of the DDX11 helicase-dead mutants ( K50R and Q23A ) to the above replication factors was similar to that of the wild type protein ( see S8C Fig ) . These results suggest that DDX11 acts in concert with Timeless to ensure stable binding of cohesin to the advancing replisomes .
In this study we identified a sequence motif of the human DDX11 DNA helicase that is responsible for the direct interaction with Timeless , a component of the replication fork-protection complex . Using Timeless-binding defective and helicase-dead mutants of DDX11 in complementation studies , we were able to demonstrate that the interaction of DDX11 with Timeless is more critical for sister chromatid cohesion than its DNA helicase activity . Besides , we found that DDX11 and cohesin associate with the ongoing replication forks in S phase synchronized HeLa cells , and binding of cohesin to the replisomes is reduced when DDX11 is down-regulated . These results , together with our finding that DDX11 physically interacts with the cohesin complex in vitro , suggests that DDX11 might play a role in recruiting cohesin to the ongoing replication forks . The short sequence motif of human DDX11 involved in Timeless-binding ( the so-called "EYE" motif ) maps to a 160-residue insertion between helicase boxes I and Ia , which we named Region T . A multiple sequence alignment of human SF2 Fe-S cluster DNA helicases revealed that XPD and RTEL1 do not have a corresponding N-terminal extra-domain . Human FANCJ has a shorter N-terminal insertion between helicase boxes I and Ia ( S1 Fig ) , which was reported to be critical for binding either to the DNA mismatch repair protein MLH1 [42] or to G-quadruplex DNA structures [43] . The "EYE" motif is invariably present in DDX11 orthologs from vertebrates , but it is only partially conserved in the Saccharomyces cerevisiae Chl1 DNA helicase ( see S3A and S3B Fig ) . Consistently , Chl1 was reported to be recruited to the replication forks by the homo-trimeric Ctf4 replication factor . This interaction is mediated by a specific Ctf4-interacting protein ( CIP ) motif that maps in the C-terminal portion of the Chl1 protein between helicase boxes IV and V [36] , as shown in S3A Fig . Similar CIP motifs were demonstrated to mediate the interaction of DNA polymerase α ( via the p180 subunit ) and the GINS complex ( via the Sld5 subunit ) with Ctf4 factor [44] . Multiple sequence alignments revealed that DDX11 vertebrate orthologs lack a conserved CIP motif between helicase boxes IV and V ( S3C Fig ) . Overall , these findings suggest that the network of interactions among protein factors operating at the DNA replication fork is not completely conserved during evolution from yeasts to vertebrates . Our analysis further revealed that two DDX11 site-specific mutants devoid of ATPase and DNA helicase activities ( DDX11 K50R and Q23A ) were able to partially rescue sister chromatid cohesion defects of DDX11-depleted HeLa cells . In contrast , in avian DDX11-knocked-out DT40 cells DDX11 helicase-dead mutant ( K87A , chicken counterpart of human DDX11 K50R mutant ) was unable to rescue the defective sister chromatid cohesion phenotype in complementation assays [45] . However , our finding , that the enzymatic activities of human DDX11 are not strictly required for sister chromatid cohesion , is in agreement with a published study on the budding yeast Chl1 [36] . An important implication of our results is that , even if Chl1/DDX11 DNA helicase activity is involved in Okazaki fragment processing in yeast/human cells , as previously hypothesized [10 , 12] , this function is not required for efficient chromosomal cohesion . This hypothesis is consistent with a recent biochemical study revealing that DNA-bound fission yeast cohesin can capture a second DNA molecule only if that is single-stranded . Thus , second-DNA capture by the cohesin ring is likely to occur at the replication fork prior to Okazaki fragment maturation by a process that is ATP-dependent and strictly requires the Scc2-Scc4 loader [46] . It was proposed that the budding yeast Chl1 protein could also promote cohesion establishment at the replication fork by orienting the cohesin complex in a way that facilitates acetylation of the Smc3 subunit by the specific Eco1 acetyltransferase [36] . Indeed , a remarkable reduction of Smc3 acetylation was observed in yeast cells lacking the Chl1 gene [15] . We found that down-regulation of DDX11 did not substantially affect the level of acetylated Smc3 in HeLa cells ( see S4C Fig ) , in agreement with an analysis carried out in chicken DDX11-knockout DT40 cells [45] . These results suggest that other protein factors assist Esco1/Esco2 in modifying cohesin during cohesion establishment at the replication fork in mammalian cells . Indeed , a recent study by Peters and coworkers has shown that Esco2 directly interacts with the MCM2–7 complex in human cells and this interaction is critical for cohesin acetylation and cohesion establishment [40] . In this study we provide evidence that DDX11 is located at sites of DNA synthesis in single cells by using the SIRF technique [41] . Moreover , our analysis suggests that DDX11 is involved in anchoring the cohesin complex to the replication machinery during fork progression . This hypothesis is supported by our finding that DDX11 physically interacts with cohesin either in cell extracts or in vitro using purified recombinant proteins . The association between cohesin and DDX11 takes place in the context of the ongoing DNA replication forks , as suggested by their co-immunoprecipitation from the nucleoplasm/chromatin fraction of HeLa cell extracts with an antibody specific for Cdc45 , a component of the CMG complex that is bound to replisomes only during S phase [37] . Moreover , we found that depletion of DDX11 in HeLa cells reduces the amount of cohesin that is co-pulled down with Cdc45 . Besides , in cells over-expressing DDX11 mutants with impaired Timeless-binding capability we found a reduced association of either DDX11 or cohesin to the DNA replication machinery ( Figs 5 and S8 ) . Our results and reports by others [19] suggest that Timeless may act upstream of DDX11 to enable a stable association of cohesin rings to the DNA replication forks . However , we cannot exclude that additional replication factors help to stabilize association of cohesin to the replication machinery . In fact , a comprehensive parallel study suggests that cohesin is loaded at the pre-replication complex at the G1/S boundary in a process that requires Scc2-Scc4 , the Mcm2-7 complex and the Dbf4-dependent kinase ( DDK ) ; after replication origin firing and Mcm2-7 DNA helicase activation , cohesin rings are mobilized and held at ongoing replication forks with the help of other replication factors , in addition to Timeless and DDX11 , including WDHD1 and RPA [35] . In a previous work we provided evidence that DDX11 and Timeless directly interact and operate in the same pathway that preserves replication fork progression in stressful conditions ( such as dNTP depletion ) [28] . In the present work , we found that the DDX11-Timeless interaction is critical for establishing chromosomal cohesion , whereas the enzymatic activities of DDX11 are not as essential for this process . This finding suggests that restarting stalled replication forks in stressful conditions and pairing duplicated chromatids are separable functions of DDX11 , even if the interaction with Timeless is needed to efficiently execute both . Our results have important implications for understanding the molecular basis of WABS . The few pathogenic missense mutations described so far were demonstrated to abolish the DDX11 DNA helicase activity in vitro by biochemical studies [26–27] . However , in fibroblasts derived from WABS patients , an almost complete disappearance of DDX11 was observed by immuno-blot analysis [26] , quite likely due to an intrinsic instability of the mutated protein . Thus , in line with our findings , it is likely that the chromosomal cohesion defects observed in WABS patient-derived cell lines are caused by DDX11 protein loss rather than by abrogation of its catalytic activities .
Recombinant human Flag-Timeless was produced in insect cells infected with a recombinant baculovirus , as previously described [47] . DDX11-3xFlag ( wild type and KAK mutant ) were produced in HEK 293T cells transiently transfected with pcDNA 3 . 0 plasmid constructions and purified as previously described [33] . The cohesin complex ( consisting of Smc1 , Smc3-Flag , Scc1 , 10xHis-SA1 ) was produced in Sf9 insect cells infected with a single multi-gene recombinant baculovirus ( a gift from Jan-Michael Peters , Wien , Austria ) , as previously described [48] . Customized PEPperCHIP peptide microarrays were provided by PEPperPRINT GmbH ( Heidelberg , Germany ) on conventional glass slides ( 75 . 4 mm x 25 mm x 1 mm ) . In the tiling peptide microarray DDX11 protein was translated into 454 different 15-amino acid peptides with a peptide-peptide overlap of 13 residues and spotted in duplicate ( 908 peptide spots for each array ) . In the microarray containing a full substitution scan of DDX11 peptide # 32 ( NH2-EQLESGEEELVLAEY-COOH ) , all residues of this peptide were substituted with all 20 natural amino acids ( 300 spots containing 15-amino acid peptides ) . Moreover , each microarray was framed by Flag ( NH2-DYKDDDDKAS-COOH ) and HA ( NH2-YPYDVPDYAG-COOH ) tags as control peptides . Each peptide microarray was incubated using a PEPperCHIP incubation tray ( PEPperPRINT , Heidelberg , Germany ) with shaking at 140 rpm in the following solutions in sequence: blocking buffer ( PBS containing 0 . 05% [v:v] Tween 20 , 1% [w:v] bovine serum albumin ) 1 hr at room temperature; staining buffer ( PBS containing 0 . 05% [v:v] Tween 20 , 0 . 1% [w:v] bovine serum albumin ) for 10 min; blocking buffer containing Flag-Timeless at 50 μg/mL for 16 hr; washing buffer ( PBS containing 0 . 05% [v:v] Tween 20 ) for 5 s at room temperature for three times; blocking buffer containing Cy3-labelled anti-Flag and Cy5-labelled anti-HA antibodies for 30 min at room temperature; washing buffer for 5 s at room temperature for three times . The chip was dipped into de-ionized water , dried in a stream of air and analyzed in a high-resolution fluorescence microarray scanner ( Agilent , model G2565 ) . The following primary antibodies were used for immunoblotting , immunoprecipitations and immunofluorescence to detect human proteins of interest: Timeless ( Abcam , ab72458 ) ; DDX11 ( Santacruz Biotechnology , sc-271711; for the SIRF assay , a rabbit polyclonal antibody donated by Joanna Parish , Birmingham , United Kingdom ) ; CREST ( donated by Florence Larminat , Toulouse , France ) ; Tubulin ( Sigma , T9026 ) ; Smc3 ( Bethyl , A300-060A ) ; GAPDH ( Cell Signalling , 2118S ) ; Smc1 ( Abcam , ab117610 ) ; Scc1 ( Cell Signaling Technology , 4321S ) ; Mcm4 ( Abcam , ab4459 ) ; Cdc45 ( a rat monoclonal antibody donated by Hans-Peter Nasheuer , Galway , Ireland; our own rabbit polyclonal against full length Cdc45 used in the immuno-precipitation experiments of Figs 5C and S8A ) . In addition , the following antibodies specific for protein tags were used: Flag ( Abcam , ab49763; Pepperprint , Cy3-labelled antibody 110802 ) ; HA ( Pepperprint , Cy5-labelled antibody 110801 ) ; Myc ( Roche , 11667203001 ) . A mouse monoclonal anti-biotin ( Invitrogen , 03–3700 ) was used for the SIRF assay . The following horseradish peroxidase-conjugated secondary antibodies were used: anti-rabbit ( Abcam , ab6721 ) ; anti-mouse ( Santacruz Biotechnology , sc-2005 ) ; anti-rat ( Sigma , A5795 ) . A DDX11-defective HeLa cell line ( named HeLa 5–5 ) , was kindly provided by Dr Akira Inoue ( Memphis , TN , USA ) . It was established by infection of HeLa cells with a pantropic retrovirus ( pSuper-Retro-Puro ) expressing a shRNA ( # 5 , targeting the DDX11 coding sequence , U33833 , from nucleotide 2398 ) , together with an HeLa control clone ( named HeLa C1 ) , obtained by infection with an empty retrovirus construction , as previously described [30–31] . HeLa and human embryonic kidney ( HEK ) 293T cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovin serum ( FBS ) and Pen/Strep in a humidified 5% CO2 atmosphere at 37°C . pcDNA 3 . 0 plasmids were constructed that direct over-expression of the DDX11 protein fused to a 3x Flag-tag at the C-terminal end ( wild type and site-specific mutants ) . These plasmid vectors were transfected into HEK 293T cells using PEI at a mass ratio of 3:1 PEI:DNA . At 48 hr after transfection , cells ( about 1 x 108 cells/experiment ) were detached , washed twice in cold PBS . Cell pellets were re-suspended in lysis buffer ( 50 mM Tris-HCl pH 8 . 0 , 150 mM NaCl , 0 . 25% [v:v] Triton X-100 , 10% [v:v] glycerol ) supplemented with a protease inhibitor cocktail . The samples were subjected to sonication on ice using a Branson digital sonifier model SSE-1 ( 8 cycles consisting of 2-s impulses at an output 10% followed by 5-s intervals ) and centrifuged for 10 min at 13 , 000 g at 4°C . Then , 30 μL of Flag-M2 ( Sigma ) beads were added to 2 mg of cell extract total protein . Samples were incubated at 4°C for 2 hr in a rotating wheel . Then , beads were washed four times with lysis buffer and proteins and protein complexes specifically bound were eluted with lysis buffer containing Flag peptide at 0 . 4 mg/mL . Samples were subjected to Western blot analysis using the indicated antibodies . Immuno-precipitations were carried out on nuclear extracts prepared from the indicated HeLa cells ( about 4 x 107 cells/experiment ) . Cell cultures were synchronized in S phase with a single block in thymidine ( at 2 mM ) followed by release in fresh medium for 2 . 5 hr . Cells were collected by centrifugation . Preparation of cell nuclear fraction was according to a published protocol with modifications [49] . Cell pellets were re-suspended in 1 mL of osmotic buffer ( 10 mM Hepes-NaOH pH 7 . 9 , 0 . 2 M potassium acetate , 0 . 34 M sucrose , 10% [v:v] glycerol , 1 mM dithiotreitol , 0 . 1% [v:v] Triton X-100 ) and incubated for 5 min on ice . After centrifugation ( 800 g for 5 min ) , the nucleus/chromatin fraction present in the pellet was re-suspended in 1 mL of hypotonic buffer ( 10 mM Hepes-NaOH pH 7 . 9 , 50 mM NaCl , 1 mM dithiotreitol , 0 . 1% [v:v] Triton X-100 ) containing a protease inhibitor cocktail ( Roche ) . Samples were subjected to sonication on ice using a Branson digital sonifier model SSE-1 ( 10 cycles consisting of 10-s impulses at an output 10% followed by 20-s intervals ) followed by incubation for 20 min at 37°C in the presence of micrococcal nuclease ( 2 units/sample; Sigma , cat . N3755 ) and CaCl2 ( at 10 mM ) . Insoluble material was removed by centrifugation at 16 , 000 g for 30 min . Samples ( containing 0 . 3–0 . 5 mg of protein ) were used in immuno-precipitation experiments with the indicated rabbit antibodies and control rabbit IgG bound to Protein A Sepharose beads ( GE Healthcare ) . They were incubated for a minimum of 3 hr ( or overnight ) at 4°C in a rotating wheel . Beads were washed 4 times with the following buffer: 10 mM Hepes-NaOH pH 7 . 9 , 50 mM NaCl , 1 mM dithiotreitol , 0 . 1% ( v:v ) Triton X-100 . Proteins bound to the beads were re-suspended in SDS-PAGE loading buffer ( 50 mM Tris-HCl , pH 6 . 8 , 10% [v:v] glycerol , 200 mM β-mercaptoethanol , 0 . 5% [w:v] SDS , 0 . 01% [w:v] blue bromophenol ) and analyzed by Western blot using the indicated antibodies . Direct interaction between recombinant purified Timeless ( or cohesin ) and DDX11 proteins was analyzed by co-immuno-precipitation experiments . Mixtures ( 200 μL ) contained purified Timeless ( 0 . 8 μg ) or the cohesin complex , ( 1 μg ) , recombinant DDX11 ( 0 . 5 μg ) and 30 μL of Protein A-beads bound to anti-DDX11 antibody in binding buffer ( 25 mM Tris-HCl , pH 7 . 5 , 150 mM NaCl , 2 mM MgCl2 , 1 mM dithiotreitol , 5% [v:v] glycerol ) . Samples were incubated for 2 hr at 4°C on a rotating wheel . Then , the beads were washed 4 times with washing buffer ( 25 mM Tris-HCl , pH 7 . 5 , 300 mM NaCl , 2 mM MgCl2 , 1 mM dithiotreitol , 5% [v:v] glycerol , 0 . 25% [v:v] Triton X-100 ) and bound proteins re-suspended in 30 μL of SDS-PAGE loading buffer . Samples were subjected to electrophoresis through 7% polyacrylamide-bis ( 29:1 ) gel and analyzed by immuno-blot with the indicated antibodies . Plasmid constructs expressing wild type DDX11 and its indicated site-specific mutants were transfected into DDX11-depleted HeLa cells ( clone 5–5; [30–31] ) . pcDNA 3 . 0 vector constructs ( named pcDNA-DDX11-Flag_WT , _K50R , _Q23A , _KAK and _KAE ) were mutated to make the DDX11 coding sequence resistant to the short hairpin RNA # 5 that is stably produced in the above HeLa cell line to down-regulate the endogenous DDX11 expression . At 24 hr post-transfection , cells were blocked in S phase by adding thymidine at 2 mM into the medium . After 16 hr , cells were released into fresh medium without thymidine . After 9 hr , colchicine at 5 μM was added to the medium and cultures incubated for additional 2 hr . Then , mitotic cells were collected by shake-off , washed once with PBS , treated with 55 mM KCl hypotonic solution at 37°C for 15 min and spun onto microscope slides with a Shandon Cytospin centrifuge . Cells on the slides were first treated with the PHEM buffer ( 25 mM HEPES , pH 7 . 5 , 10 mM EGTA , pH 8 . 0 , 60 mM PIPES , pH 7 . 0 , and 2 mM MgCl2 ) containing 0 . 3% ( v:v ) Triton X-100 for 5 min and then fixed in 4% ( v:v ) paraformaldehyde for 10 min . Fixed cells were washed three times with PBS containing 0 . 1% ( v:v ) Triton X-100 for 2 min each time , and incubated with the human CREST antiserum in PBS containing 3% ( w:v ) BSA and 0 . 1% ( v:v ) Triton X-100 at 4°C overnight . Cells were then washed three times with PBS containing 0 . 1% ( v:v ) Triton X-100 for 2 min each time , and incubated with a fluorescent secondary antibody in PBS containing 3% ( w:v ) BSA and 0 . 1% ( v:v ) Triton X-100 for 1 hr at room temperature . Cells were again washed three times with PBS containing 0 . 1% ( v:v ) Triton X-100 and then stained with 1 μg/mL DAPI for 2 min . Slides were viewed with a 100x objective on a Nikon A1 confocal microscope using a NIS-Elements imaging software . FISH probes that specifically recognize a locus on human chromosome 3 were made as described previously [50] . HeLa cells expressing shCtrl ( HeLa C1 ) or shDDX11 ( HeLa 5–5 ) were transfected with plasmid vectors that express wild type or mutant DDX11 . After a thymidine-block for 16–18 hr , cultures were released to fresh medium and incubated for 4 hr . Then , cells were harvested by treatment with Trypsin , treated with 75 mM KCl hypotonic solution for 25 min at 37°C and fixed with ice-cold methanol and acetic acid ( 3:1 , v:v ) . Fixed cells were dropped onto pre-warmed slides , in situ hybridized at 80°C with DNA probes and incubated at 37°C overnight . Slides were sequentially washed with 0 . 1% ( w:v ) SDS in 0 . 5x SSC at 70°C for 5 min , PBS at room temperature for 10 min and 0 . 1% ( v:v ) Tween 20 in PBS at room temperature for 10 min . Slides were then mounted with ProLong Gold ( Life Technologies ) and viewed with a 100 x objective on a DeltaVision fluorescence microscope ( GE Healthcare ) . Image processing and quantification were performed with the ImageJ software . For siRNA transfection , HeLa Tet-On cells at 20–40% confluency were transfected with Lipofectamine RNAiMAX ( Invitrogen ) according to the manufacturer’s protocols , and analyzed at 24–48 hr after transfection . The siRNAs were transfected at a final concentration of 5 nM . The siRNAs used to downregulate DDX11 expression were as follows: # 1 ( 5'-GCAGAGCUGUACCGGGUUU-3' ) , # 2: ( 5'CGGCAGAACCUUUGUGUAA-3' ) , # 3: ( 5'-GAGGAAGAACACAUAACUA-3' ) , # 4: ( 5'-UGUUCAAGGUGCAGCGAUA-3' ) . Cells were cultured and treated in the Nunc Lab-Tek II CC2 Chamber Slides . They were first treated with the PHEM buffer containing 0 . 5% ( v:v ) Triton X-100 for 5 min and then fixed in 2% ( v:v ) paraformaldehyde for 15 min . Fixed cells were blocked in PBS containing 2% ( w:v ) BSA for 30 min and then incubated with desired antibodies in PBS containing 0 . 1% ( v:v ) Triton X-100 ( PBST ) and 3% ( w:v ) BSA and at 4°C overnight . Cells were then washed three times with PBST for 5 min each time , and incubated with fluorescent secondary antibodies ( Molecular Probes ) in PBST containing 3% ( w:v ) BSA for 1 hr at room temperature . Cells were again washed three times with PBST and stained with 1 μg/mL DAPI in PBS for 5 min . After the final wash with PBS , the slides were mounted with VECTASHIELD anti-fade mounting medium ( Vector Laboratories ) , sealed with nail polish , and viewed with a 100x objective on a DeltaVision fluorescence microscope ( GE Healthcare ) . Image processing and quantification were performed with Image J . Exponential growing MRC5SV40 cells ( described in [51] ) were seeded onto a microscope chamber slide . The day of experiment , cells were incubated with 100 μM EdU for 15 min and treated as indicated . After treatments cells were pre-extracted in CSK-100 buffer ( 100mM NaCl , 300mM sucrose , 3 mM MgCl2 , 10 mM PIPES pH 6 . 8 , 1 mM EGTA , 0 . 2% Triton X-100 , protease inhibitor cocktail at 1x ) for 5 min on ice under gentle agitation and fixed with 4% ( v:v ) paraformaldehyde in PBS for 20 min at RT . Cells were treated with ice-cold methanol at -20°C for 10 s and then blocked in 3% ( w:v ) BSA in PBS for 15 min . The primary antibodies used were diluted as follows: rabbit anti-DDX11 at 1:150 and mouse anti-biotin at 1:500 . The negative control consisted of cells that were not pulsed with EdU . Samples were incubated with secondary antibodies ( OLINK Bioscience ) conjugated with PLA probes MINUS ( anti-rabbit ) and PLUS ( anti-mouse ) . The incubation with all antibodies was accomplished in a humidified chamber for 1 h at 37°C . Next , the PLA probes MINUS and PLUS were hybridized to two connecting oligonucleotides to produce a template for rolling-circle amplification . After amplification , the products were hybridized with a red fluorescence-labelled oligonucleotide . Samples were mounted in Prolong Gold anti-fade reagent with DAPI . Images were acquired randomly using an Eclipse 80i Nikon fluorescence microscope , equipped with a Video Confocal ( ViCo ) system . | Chromosomes are DNA molecules that contain the genetic information . During replication , the two sister DNA molecules covered by proteins ( sister chromatids ) are held together by many copies of a ring-like protein complex named cohesin , in a process called sister-chromatid cohesion . Before a cell divides , the cohesin rings are removed from the two sister chromatids to allow their migration towards the opposite poles of the dividing mother cell . At the end of this process , the two daughter cells have inherited a complete set of chromosomes . Before the next cell division , chromosomes are duplicated with high speed and fidelity . This important task is performed by the DNA replication machinery , a sophisticated apparatus made of several enzymes and proteins . In the present study , we have demonstrated that DDX11 and Timeless , two subunits of the DNA replication machinery , recruit the cohesin rings to promote their stable binding to the newly duplicated chromosomes , that is the establishment of sister-chromatid cohesion . In human cells that were genetically engineered to reduce the level of DDX11 , we observed that sister-chromatid cohesion was loosened and association of cohesin to chromosomes was reduced . Our experimental results contribute to our understanding of the molecular mechanisms underlying the functional coupling between DNA replication and sister-chromatid cohesion in human cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"hela",
"cells",
"enzymes",
"cell",
"cycle",
"and",
"cell",
"division",
"gene",
"regulation",
"biological",
"cultures",
"cell",
"processes",
"enzymology",
"dna",
"helicases",
"chromatids",
"dna",
"replication",
"cell",
"cultures",
"sequence",
"motif",
"analysis",
"dna",
"synthesis",
"phase",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"small",
"interfering",
"rnas",
"chromosome",
"biology",
"bioinformatics",
"proteins",
"gene",
"expression",
"cell",
"lines",
"biochemistry",
"rna",
"helicases",
"cell",
"biology",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"cultured",
"tumor",
"cells",
"non-coding",
"rna",
"chromosomes"
] | 2018 | Interaction of the Warsaw breakage syndrome DNA helicase DDX11 with the replication fork-protection factor Timeless promotes sister chromatid cohesion |
The neuronal system underlying learning , generation and recognition of song in birds is one of the best-studied systems in the neurosciences . Here , we use these experimental findings to derive a neurobiologically plausible , dynamic , hierarchical model of birdsong generation and transform it into a functional model of birdsong recognition . The generation model consists of neuronal rate models and includes critical anatomical components like the premotor song-control nucleus HVC ( proper name ) , the premotor nucleus RA ( robust nucleus of the arcopallium ) , and a model of the syringeal and respiratory organs . We use Bayesian inference of this dynamical system to derive a possible mechanism for how birds can efficiently and robustly recognize the songs of their conspecifics in an online fashion . Our results indicate that the specific way birdsong is generated enables a listening bird to robustly and rapidly perceive embedded information at multiple time scales of a song . The resulting mechanism can be useful for investigating the functional roles of auditory recognition areas and providing predictions for future birdsong experiments .
Songbirds are able to repeat the same , often complex songs with amazing precision . When male birds sing to a female repeatedly , there is on average a 1% temporal deviation across the whole song [1] , [2] . This combination of complexity and precision is remarkable . Studying the neuronal basis of birdsong generation may lead to an understanding of the mechanism underlying how sequences of song syllables are expressed as complex and temporally precise sound wave modulations . More generally , such a mechanism may also be useful for understanding how action sequences at a relatively slow time-scale ( e . g . the words in a sentence ) can be generated by a neuronal system while a high degree of precision is maintained in the output at a fast time-scale ( e . g . the sound wave modulations necessary to form speech sounds ) . Recent findings [1]–[3] have shown that the song generation mechanism in birds is hierarchical where neurons in one particular high-level structure , HVC , fire in a specific sequence with high temporal precision and drive neurons in the lower level structure RA ( robust nucleus of the arcopallium ) . Female birds , at which the songs are typically directed , are expert in registering variables like the speed of the song and the precision and the repertoire of the singer [4]–[8] . Unfortunately , the study of song recognition is more challenging than song generation because experimental indicators for recognition , such as the subsequent behavior of a female bird , are more difficult to measure than indicators for song generation . This has led to a long list of experimental and theoretical findings on song generation and learning while the mechanisms of song recognition remain relatively elusive . Here , we propose that the functional mechanism of song recognition can be obtained from the song generation mechanism . The basic idea underlying this novel modeling approach is that female birds are optimal in song recognition because their mating choice critically depends on the optimal recognition of valuable features of the male which are revealed by subtle indicators in his song . Similarly , male birds should be able to distinguish the songs of their neighbors from the songs of strangers to protect their territories [9] , [10] . Using a recently established Bayesian inference technique for nonlinear dynamical systems [11] , we can emulate this optimal recognition: the key ingredient is a generative model ( a nonlinear dynamical system ) which can generate a specific song . Usually , generative models for complex sensory dynamics , such as the sound wave or spectrum of birdsong , are difficult to derive because it is hard to describe a complex multi-scale structure like birdsong using only differential equations . Fortunately , since the hierarchical birdsong generating system is so well-studied , parts of such a model already exist , in particular at the level of the HVC , RA and vocal tract dynamics [12]–[18] . We have combined these parts into a coherent whole , guided by key experimental results , to form a generative model that can play complex songs . In particular , we combined sequence-generating dynamics , attractor dynamics and a model of vocal tract dynamics [17] in a three-level , hierarchical nonlinear dynamical system . This dynamic model is based on neuronal rate models , thereby describing the biological system at a mesoscopic level . We then used Bayesian inference to derive another set of hierarchical , nonlinear differential equations ( recognition system ) which is , by way of construction , Bayes-optimal in recognizing this song and can be compared to the real birdsong recognition system . To do this , we exposed the agent to several tasks and found that the agent's dynamics and performance were reminiscent of song recognition in real bird brains in aspects such as sensitivity to speed changes [19] and song perturbations [20] , [21] . Thus , by harnessing rich experimental and theoretical results in birdsong generation , we were able to derive a novel , functional model of birdsong recognition . We discuss the experimental evidence that the identified mechanism is indeed used for song recognition by birds . We suggest that the present model may be useful for understanding the functional and computational roles of auditory recognition areas . In addition , the identified recognition mechanism can be used as a novel machine learning tool to recognize sequential behavior from fast sensory input , e . g . in artificial speech recognition .
Lotka-Volterra equations are well known in population biology to describe the competition between species [32] . Rabinovich et al . ( see [33] for a review ) applied this idea more generally to neuronal dynamics under the name of winnerless competition , see [28] and [34] for applications . In the following , we will describe how one can apply this idea to model sequential HVC activity by a nonlinear dynamical system . In the winnerless competition setting , there are equilibrium points which are saddles of a nonlinear dynamical system . Each of these equilibrium points has a single unstable direction and all other directions are stable . One can think of these saddle points as the beads on a string where the unstable manifold of one saddle point is the stable manifold of the next saddle point and this sequence continues in a circular fashion forming a heteroclinic chain . Under some conditions [27] , this sequence is stable , i . e . a solution of the system that starts from a neighborhood of the chain , stays in this neighborhood at all times while traveling through all saddle points . This stable sequential behavior is what we exploit to model the experimentally established sequential activities of HVC ( RA ) ensembles at the highest level . As the solution of the system moves along the string , it visits all saddle points , i . e . each HVC ( RA ) ensemble , one by one thereby activating each ensemble for a brief period until it is deactivated as the next ensemble becomes active . These dynamics can be obtained from a neural mass model of mean membrane potential and action firing potential [35] , reviewed in [33] . We use the equations: ( 1 ) where is the hidden-state vector ( e . g . , mean membrane potentials ) at the third ( HVC ) level , and are scalars , is the sigmoid function applied component-wise and is the connectivity matrix with entries giving the strength of inhibition from state to . The second equation describes the output vector ( or causal-state vector; e . g . , neural firing rates ) where , , is a normalizing function . We also add normally distributed noise vectors and to render the model stochastic . With an appropriately chosen connectivity matrix , one can obtain a system with saddle points forming a stable heteroclinic chain [27] . For the entries of the connectivity matrix , one chooses high inhibition from the previously active neuron to the currently active neuron and low inhibition from the current active neuron to the next neuron which will become active: ( Here when and when ) . Note that , theoretically , one can generate arbitrarily long sequences of HVC activation using the above connectivity matrix . The stability region around the heteroclinic chain will persist for much longer sequences than the one modeled here . For our illustrative simulations described below , we use , i . e . , there are 8 HVC ( RA ) neuronal ensembles but the model works robustly with more HVC ( RA ) ensembles as well ( see Figure S1 ) . A real bird brain has many more HVC ( RA ) ensembles but here we are interested in presenting a general mechanism for which a small selection of HVC and RA ensembles is sufficient . See the third level dynamics in Figure 5A for typical dynamics generated by this system . We control the dynamics of RA ensembles by letting the kth HVC ( RA ) ensemble send a signal to the lower level during its activation time . See the next subsection for details of how this signal vector is computed . The total signal sent to the lower level by all HVC ( RA ) ensembles at any time is a linear combination of the 's: where is the output vector in Eq . ( 1 ) . Note that for typical sequential dynamics at the HVC level , except for the transition times , only one entry in is active ( i . e . , only one entry is close to ) , see Figure 5A . Experimental findings suggest that activation of different HVC ( RA ) ensembles drives the activation of different combinations of RA ensembles [25] . In the present model , we capture this by forming a network of RA neuronal ensembles whose dynamics converge to one of several attractors depending on the input from the HVC level ( see Figure 2 ) . This means that the RA level receives input from the HVC level and produces output which encodes the level of activity of each RA ensemble at a given time . Since we are working with continuous systems , the notion of attractors comes up naturally as the RA ensemble activity flows from one activity pattern to another one . To achieve this smooth flow between RA attractors , we have to use a nonlinear network because otherwise the RA level would simply copy the dynamics of the HVC level . Note that , similar to the HVC level , the intrinsic neuronal dynamics of the RA are not established well experimentally . In this situation , we aim at describing underlying population dynamics which give rise to the experimentally observed key features of RA dynamics [25] . To implement these dynamics , we use a well-established type of an attractor-based network described by Hopfield [29] . Hopfield networks have been mostly used as a model of associative memory where each memory item is encoded by an attractor . When such a system receives noisy sensory input , i . e . it is started at some nearby initial state , it evolves to an attractor ( the memory to be retrieved ) [29] , [36] . Here , we use this idea to encode the activities of RA ensembles by attractors . As the attractor of the network changes continuously due to driving HVC input , the activities of RA ensembles also changes such that some RA ensembles activate and some others deactivate . This gives us the spatiotemporal coding that drives the syrinx dynamics described in the next subsection . We use a Hopfield network with asymmetric connectivity matrices [37]–[39] given by the following equation: ( 2 ) where is the ensemble state vector with ensembles , is a diagonal positive matrix which governs the rate of change of each ensemble's state , is a synaptic connectivity matrix with entries denoting the strength of connection from ensemble to ensemble , is the activation function which we take as tanh function applied component-wise and is the direct input from the HVC level . This equation is similar to Eq . ( 1 ) , i . e . both are continuous-time recurrent neural networks , but in Eq . ( 2 ) we have an additional input vector and different conditions on the connectivity matrix as described below . In addition , the use of the nonlinear activation function brings more plausibility to the network , as compared to linear dynamics , since the effect of one RA ensemble to another one does not increase linearly but saturates . The input vector should be chosen such that RA ensembles get quickly attracted to a desired attractor . An attractor means that a subset of the RA ensembles are ‘active’ ( taking the value ) while all other RA ensembles are inactive ( taking the value ) . The goal is to establish conditions for the network in Eq . ( 2 ) to have a globally asymptotically stable equilibrium point ( a vector that makes the right hand side of Eq . ( 2 ) zero and attracts all the solutions regardless of the initial state ) . These conditions and the proper choice of for the desired attractor have been described in [38] and [40] ( see Theorem 1 in Text S1 ) . Using this technique , we can employ a small number of RA ensembles to encode a larger number of desired attractors to control the lowest level , the motor output . Each HVC ( RA ) ensemble provides a different -vector to the RA level thereby driving the RA ensembles into a unique attractor . The application of this is that each RA level attractor will drive the motor output in a specific way thereby producing a different part of the song . We obtain the equations for the second level by combining the Hopfield network , Eq . ( 2 ) , with the two output equations ( state vectors ) and where superscripts denote the specific level of a variable: ( 3 ) where the exact form of the connectivity matrices , and the HVC input vector are described in Text S1 , is a scalar and are normally distributed noise vectors . is the normalizing function as in Eq . ( 1 ) . Note that squeezes the entries of into the interval but may return values smaller than 1 since more than one entry of can be active ( ) at a given time . The vectors and carry the output of the second level to the first level ( oscillator level ) as described in the next subsection . In the present model , we use ( i . e . , five RA ensembles , Figure 5B ) . Note that there are different ways to activate RA ensembles to produce motor output . We use 7 of these 31 combinations ( one occurring twice ) in Figure 5 for generation of an example song ( with 8 HVC ( RA ) ensembles at the higher level ) . In the figures , we used arbitrary units for both time ( x-axis ) and neuronal activation ( y-axis ) because we consider neuronal ensembles . The avian vocal organ , the syrinx , is located at the base of the trachea ( windpipe ) where the trachea divides into the bronchi . A set of soft tissues within the syrinx , the labia , which are similar to human vocal folds , oscillate with the airstream propelled from the air sacs . Sound waves generated from these oscillations propagate through the trachea and beak . Therefore , these sound waves are modeled as the oscillations of the labia which are produced by the vocal control signals: the air sac pressure , , and the stiffness of the labia , . Such a mathematical model of the vocal fold oscillations was first given by Titze [41] , and similar oscillations were experimentally observed in the bird syrinx [42] . A simplified version of this model ( using a polynomial approximation for the nonlinear dissipation ) can be given as follows [17]: ( 4 ) where is the position of the labia from the midpoint of the syrinx , denotes the air sac pressure , is the linear dissipation constant , is the stiffness of the labia and is a dissipation term to prevent the big amplitude oscillations when the labia meet each other or the walls of the syrinx [43] . The fundamental frequency of the sound wave increases or decreases proportional to . Note that there is a critical value for the pressure such that if , no phonation is produced . This region in the parameter space corresponds to the mini breaths between syllables [44] . Using this simple model , one can obtain accurate copies of some birdsongs such as canary [30] , chingolo sparrow [17] , white-crowned sparrow [31] and cardinal [45] by choosing appropriate vocal control signals for the syrinx ( and ) as described next . Oscillators as in the present model have been widely used to model movement patterns in animals and humans . Central pattern generators are a well-known example of neural networks that are used to generate periodic motor commands such as locomotion [46] . We use the same principle here , and use five oscillators with different frequencies ( one for each RA ensemble ) to let the RA dynamics drive the vocal output ( syrinx ) mechanism , see Eq . ( 4 ) . Note that it is experimentally not well established how the RA level controls the syrinx muscles; our approach is a natural extension of the phenomenological syrinx model described above [17] . The main point here is that the oscillator level ( first level ) is assumed to generate mixtures of oscillations ( hidden states ) where the RA level activity at the supraordinate level controls which oscillations should be produced at a given time . Each RA ensemble is assumed to control the activity of a single oscillator at the level below Therefore , the spatiotemporal coding of the RA level is transformed into the oscillatory activity of the first level which generates the final p ( t ) and k ( t ) dynamics necessary to control the syrinx . As oscillators , we choose simple sine wave equations where the lowest frequency oscillator corresponds to the slowest-changing dynamics of the birdsong . We choose the remaining four oscillators such that their frequencies are integer multiples of this first oscillator's frequency ( ) : , , and . Each one of these sine waves represents faster changing dynamics of the song; being the fastest . In this way , we can model effects in the birdsong which express themselves on different time-scales . We include these five oscillators in the present model at the first level , where each of the five ensembles at the RA level controls the amplitude of one of the oscillators ( through , Eq . ( 5 ) in Text S1 ) . The observable output is obtained by taking a linear combination of these amplitude-modulated sine waves . To drive the vocal model appropriately , we produce two outputs and ( the second output is simply a time-shifted copy of the first one ) , which are involved in producing air sac pressure p ( t ) and the stiffness of the labia k ( t ) . and are described in detail in Text S1 . Laje et al . [31] chose and to form several ellipses in the parameter space where each ellipse corresponds to a different syllable . However , this parameterization may not support complicated syllables which have more fluctuations on the sonogram . Here , we extend their model to increase the complexity of the generated songs by using the linear combination of different frequency sine waves ( and described above ) to parameterize these two functions and obtain a variety of ellipse-like curves in the parameter space ( see Figure 3 ) :where and are the outputs of the first level and the scalars are given in Table 1 . These ellipse-like curves can be plugged into Eq . ( 4 ) to obtain synthetic birdsongs . See Figure 6 for the sonogram obtained using the first level output of the generation process shown in Figure 5C . The sonogram can be played and is reminiscent of a birdsong ( Audio S1 ) . Note that in the real system , longer HVC ( RA ) sequences would be required to produce a song with 6 . 5 seconds duration since HVC ( RA ) bursts last only about 6–10 ms [25] . Here , we assume that each HVC ( RA ) ensemble in the model is a collection of at least 80 HVC ( RA ) neurons that fires sequentially and controls the timing of the song for about 800 ms . In this subsection , we will briefly describe the present recognition scheme for the generated songs . This scheme is a model of vocal communication between conspecific birds but may also serve as a functional model to explain experimental findings along the auditory processing pathway which is less understood than the song pathway . Here , we describe a potential mapping of this Bayesian inference framework to neuronal dynamics at a population level , see [47] , [48] . The inference is based on hierarchical message passing and implements a predictive coding scheme for dynamics . As summarized below , all the update equations of the recognition system ( to reconstruct the hidden states ) consist of differential equations ( as in the generation model ) and therefore may be implemented by neuronal populations and their network interactions via forward , backward and lateral connections [47] , [48] . How can a bird recognize a conspecific's song and decode the information contained in the song ? This decoding is important as it is known that female birds select their mates according to criteria such as the complexity of the male's repertoire [7] or the precision of the vocal performance [8] and they show preference for the songs of their mates or fathers compared to the songs of strangers [4] , [49] , [50] . In general , this suggests that listening birds may have certain expectations ( priors ) about the type of the song they expect to hear . In general , we assume that listening birds have internal models for the songs they have learned before and the generative model of the heard songs should fit to this internal model . Using this concept , we model optimal recognition using Bayesian inference for hierarchical , nonlinear dynamical systems [47] . For the sensory input , we assume that the vocal control signal , given the sound wave , can be readily extracted by the listening bird ( agent ) from the spectrotemporal dynamics , see Figure 3 . Here , we consider the p ( t ) and k ( t ) dynamics , in the recognition step , as an abstract representation of the song spectrum and therefore a phenomenological approximation to the highly nonlinear features of the singing bird's syrinx . This means that we assume that the listening bird has access to these dynamics via some low-level recognition process . For the present implementation of the inference framework , the full inference from the soundwave ( Figure 7 ) would currently be computationally too expensive because this would require a high temporal resolution , e . g . at 12 kHz , and long time-series . However , once an optimized ( parallel ) implementation of the present framework becomes available , the present model can be extended in a straightforward fashion to model recognition that receives a soundwave as sensory input by adding another level that transformed the p ( t ) and k ( t ) dynamics to soundwaves . Given this vocal control signal , we infer the spatiotemporal RA dynamics and the sequential HVC ( RA ) dynamics . The proposed Bayesian inference scheme provides , under some assumptions , optimal inference to decode the RA and HVC ( RA ) dynamics , i . e . to recognize the hidden messages embedded into the vocal control signal . The mathematical description is provided below and can be conceptualized as follows: At each time step t , the recognition system receives sensory input , here the current amplitudes of the p ( t ) and k ( t ) dynamics . Like the generative model , the recognition system has three levels as well . Each of these three levels consists of interacting neuronal populations , which encode predictions , i . e . expectations , about how their internal dynamics will evolve during a song . At the same time , each level receives input from the subordinate level . For the first level , this is the sensory input , which is compared with the internal prediction . The prediction error is forwarded to the second level , where again predictions are used to generate prediction errors , which are forwarded to the third level . Critically , each level adjusts its internal predictions to minimize its prediction error weighted by the prior precision of the internal prediction . At each level , the updated predictions are sent to the subordinate levels to guide their internal predictions by higher level predictions . In summary , each level minimizes its prediction error by a fusion of internal dynamics with top-down ( predictions ) and bottom-up ( prediction error ) messages . The overall result is that a listening bird fuses its dynamic and hierarchically arranged expectations about a song with the actual sensory input . Importantly , due to this dynamic fusion , the recognition is robust against deviations from its expectations by explaining away errors of the singing bird by internal precision-weighted prediction error . The derivation of the update equations to achieve Bayes-optimal online recognition solutions is non-trivial , see Friston et al . [11] . Note that this modeling approach implies that generation and recognition models are fundamentally different from each other in the sense that generation is a top-down process where recognition consists of both top-down and bottom-up processes . Although some of the computations in the generation and recognition model are the same and may provide a computational explanation for mirror neuron accounts [51] , this is not a central issue in the present paper and we assume here that recognition is performed by neuronal populations different from those that generated the song . Clearly , this remains an open question that can only be settled experimentally . For sensory input and a given model , the probability is called the model evidence or marginal likelihood of and is an important quantity for model comparison among different models . In our case , is the vocal control signal for the syrinx which we take as the input and the model ( Figure 4 ) includes all the parameters and equations together with causal and hidden states at all levels . We take to be the set of all hidden states and causal states at all levels of hierarchy . The task for the agent is to infer the states from the sensory input under model m . We assume that the parameters ( such as , and , see Figure 4 ) have been learned previously by the listening bird and are fixed ( Table 1 ) . Our goal is to approximate the posterior density which will give us both the posterior mean of the dynamical states and the uncertainty about this mean . To get a good approximation for the posterior density , we follow a rather indirect way using the marginal likelihood . The marginal likelihood of can be written as . Here , is defined in terms of the likelihood and the prior . Except for a few analytical cases , this integral is usually intractable and needs to be approximated . One way for this approximation is to introduce a free-energy term which is a lower bound for the marginal likelihood . It is not hard to show that:where is the free-energy , is the Kullback-Leibler divergence and is the recognition density . Note that is an auxiliary function that we will use to approximate the posterior density . It is easy to show that , and if and only if . This means is a lower bound for , and if we can maximize , this will minimize giving an approximation for the posterior density . To maximize with respect to , we make the assumption of normally distributed error terms and write where consists of the mode and the variance . Then the problem turns to a maximization problem of the free energy with respect to :which gives the approximation for the posterior density . For the details of this variational process and its extension to time-dependent states , see [11] . Since we apply the variational scheme in a hierarchical setting , we write the equations in our model ( see Figure 4 ) in a generic hierarchical form [11] . We use the same set of equations as in the generative model since we assumed the singing and listening birds have the same internal models . We denote all hidden and causal states at level by and , respectively . In particular , stands for all the and outputs of the th level . We also write and to describe the dynamics of the hidden and causal states in the th level:where denotes the normally distributed fluctuations at the th level . The present model shown in Figure 4 follows this generic form . The causal states ( ) provide input to the subordinate level while the hidden states ( ) are intrinsic to each level . Note that the Gaussian fluctuations in the above hierarchical form quantify different amounts of noise at each level of the singing bird . We list the covariance matrices used in the “Ideal Communication” simulation in Table 1 . Note that sensory input enters the recognition system at the first level: . The optimization process of ( i . e . the estimated mode of causal and hidden states ) can be implemented in a message passing scheme [11] which involves passing predictions down and passing prediction errors up from one level to another . Prediction errors can be written aswhere and denote the predictions from level above for and , respectively . In this scheme , is optimized through gradient descent on prediction errors at each level of the hierarchy . Importantly , the computations required for this gradient descent could be implemented by interacting neuronal populations at each level: Each population comprises causal and hidden state-units that encode the expected states and the error-units , with one matching error-unit for each state-unit , which encode the prediction errors . The estimated mode of the states , i . e . , is described by the activity of the state-units . The error units compare the estimated modes with predictions sent via backward and lateral connections and compute prediction errors , which are passed on via forward and lateral connections . This message passing has been shown to minimize precision-weighted prediction errors and optimize predictions at all levels efficiently ( see [47] , [52] for further details ) . Software Note: The routines ( including commented Matlab source code ) implementing this dynamic inversion , which were also used for the simulations in this paper , are available as academic freeware ( Statistical Parametric Mapping package ( SPM8 ) from http://www . fil . ion . ucl . ac . uk/spm/; Dynamic Expectation Maximization ( DEM ) Toolbox ) .
Here , we simulate the ideal situation in which both the ‘singing bird’ and the ‘listening bird’ have learned how exactly a song should sound . As before , we use eight third level ensembles that are each activated sequentially and , during this time , they control the activities of five second level ensembles ( Figure 2 ) . The third level imposes a sequence of attractors on the second level which in turn produce linear combinations of appropriate sine waves to produce the air sac pressure and labia stiffness , see Figure 4 . To introduce noise ( both internal state noise for the singing bird , and also transmission noise to the listening bird ) , we used normally distributed zero-mean noise with standard deviation of and at all levels . To show that recognition is robust against starting condition ( i . e . the state of the ongoing neuronal activity within the bird brain at song onset ) , the initial states of the recognition are chosen differently from the true initial values used in the generation . As expected , we find that the listening bird starts tracking the sensory input very quickly and follows it robustly during the remainder of the song , see Figure 7 . Next , we show what happens if the listening bird has a different expectation than the singing bird about how a song should sound . In the generative model ( singing bird ) , we use the same third level ensembles and the corresponding second level combinations that we used in the ‘Ideal Communication’ case ( Figure 2 ) . However , the recognition system ( listening bird ) knows a slightly different song where there is a deviation in a single syllable . We model this by changing the effect of the third ensemble at the third level such that it activates only the first ensemble at the second level ( instead of the first and fourth as in the singing bird ) . This means that the motor output and the sonogram look different from the prior expectation of the listening bird but only for the third syllable , see Figure 8 . The internal recognition dynamics of the listening bird register this deviation and show two effects during the third syllable , between time points and : ( i ) Prediction errors in the recognition are distributed throughout all three levels and are not only explained by changes at a single level ( Figure 9 ) . This makes sense since the observed deviation at the first level cannot be explained by the simple oscillatory first level dynamics . Rather , the recognition attempts to explain away the deviation at the first level by using prediction error at the second and third level as well . At the first level , this is quite successful because the recognized dynamics look very similar to the generated dynamics ( see Figure 8 , bottom row ) . However , at higher levels , there are obvious differences between the generated and recognized dynamics , i . e . the listening bird can infer a deviation via the prediction error at the second and third levels . ( ii ) When the deviation has finished , the recognition quickly locks back onto the ongoing song dynamics at all three levels and decodes the song veridically . In summary , this simulation shows that the dynamic recognition hierarchy uses all its levels to compensate for unexpected deviations in the song . This means that all levels of the hierarchy work together in concert to minimize the effects of deviations throughout the hierarchy . In other words , the activity of high-level auditory processing levels in songbirds in response to small deviations in the expected song may be most revealing for their function . This mechanism may be important in social context since the listening bird can recognize subtle variations in the singing bird by its activity in high-level areas and grade the singing bird's overall performance [54] . Considering the anatomical complexity of the brain , genetic and developmental variability is expected in the brains of individuals of the same species . At the macro scale , the general connectivity structure of distinct brain regions may be shared , but at the micro scale , variability is found in size , location and connections between individual neurons or neuronal ensembles [55]–[58] . Here , we simulate a difference in the connectivity structures by using different second-level connectivity matrices W ( Figure 4 and Eq . ( 3 ) ) in the generative model of the singing bird and the recognition system of the listening bird . In other words , the listening bird has a different internal model at the second level as would be prescribed by the generative model of the singing bird at the RA level . How can birds with individual variability in their internal models still extract the same information from a song ? The answer is that differences in the second-level connectivity matrix W can be compensated by a different driving activity I from the third level since I depends on W ( see Theorem 1 in Text S1 ) . In our simulation , we assume that these driving activities have already been learned in the corresponding birds , e . g . during juvenility . As shown in Figure 10 , the states at all three levels can be recognized successfully even though the second levels in the two birds are wired differently . This means that the internal models of generation and recognition do not have to be the same but can cope with structural variations due to anatomical variability at the micro-scale . Critically , this compensation of anatomical variability at the second level relies on the hierarchical configuration and learning of the connectivity from the third level to second level . In song generation , a critical question is which regions of the brain are involved in the timing of syllables or sub-syllable structures . A recent study tackled this question by manipulating the temperature of the HVC and RA regions in the singing bird [53] . Importantly , it was shown that song speed at all time scales slowed down but the acoustic structure stayed the same as the temperature of HVC dropped . In the sonogram , this corresponds to a temporal stretching of the song . Conversely , cooling of RA did not have any effect on the timing of the song . This suggests that HVC is involved in the control of the timing of the song [53] . We observed similar behavior in our model where we modeled the cooling by manipulating the rate ( i . e . speed ) constants and at the three levels . Importantly , changing the rate constant for HVC slows down the song but changing the rate constant for RA does not . In the first simulation ( Figure 11 , left ) , we ‘cooled’ HVC by changing from to . This slows down the dynamics of the HVC level and immediately slows down the RA level as well since the control signals coming from HVC now last twice as long . In other words , we find as in the cooling experiment that HVC , due to its position at the top of the hierarchy , controls directly the timing of the song . To reflect this slowing down in the output we also changed from to ( is kept constant in all simulations ) to adjust the frequencies which were chosen independently from the RA level for simplicity ( where ) . In the second simulation ( Figure 11 , right ) , we changed the rate constant of RA , , from to . This has no observable effect , as in the experiment [53] , on the dynamics of RA ensembles since the timing of attractor activations is controlled by the timing of HVC . A change in only slows down the transition times which has no detectable effect in the output . Speed changes may not only have an experimentally observable effect in the generated song but also in the listening bird . Interestingly , speech changes in song also occur under natural conditions , e . g . in a social context: Male birds sing slightly faster when addressing a female bird ( directed song ) compared to singing towards other males or when alone ( undirected song ) [6] , [59] . Using the present model , we tested whether the listening bird can detect such small changes in the singing bird during directed song . We slowed down the song by 3% , thereby modeling an undirected song , and analyzed the prediction errors in the listening bird which expected the slightly faster , directed version . The listening bird was able to recognize the song successfully but it also reliably distinguished the subtle change in the tempo , as can be seen from the sustained prediction errors at all three levels ( Figure 12 ) .
We have derived a recognition scheme using Bayesian inference . However , bird brains may have established their recognition capabilities by evolutionary processes [4] , [49] , [62] . What are the similarities between the proposed recognition scheme and the biological one ? Note that the present modeling does not suggest that the areas involved in generation and recognition are the same . Many computations during recognition are different from those in generation . The present recognition scheme consists of three hierarchical levels , thereby mirroring the hierarchical generation system . We found that three hierarchical levels are also appropriate for the recognition of a song . Interestingly , experimental findings point to a hierarchical arrangement of the auditory system in songbirds as three major functional levels of processing [63] , [64] where it is partially unclear yet how this hierarchy maps exactly onto the auditory system . Moreover , note that these areas are mostly investigated for male ( zebra finch ) birds and it is quite possible that there could be different areas involved in females or in other bird species . Experimental evidence suggests that HVC may be located at the highest level of this recognition system . In particular , HVC ( X ) neurons ( HVC neurons that project to Area X , see Figure 1 ) are selectively responsive to the bird's own or a conspecific's song [65] , [66] . The firing of HVC ( X ) neurons at temporally precise times during an auditory stimulus [65] is similar to the temporally precise activation of HVC ( RA ) neurons during singing . This suggests that HVC ( X ) neurons may be involved in the representation of the expected sequence of song dynamics . In the present model , the third level encodes both the sequence prediction but also the perceived deviation from this sequence . The circuitry of areas subordinate to HVC during song recognition is not particularly well understood . The caudal mesopallium ( CM ) and caudomedial nidopallium ( NCM ) have been shown to be selective for particular familiar songs or sounds and are involved in auditory memory [23] , [63] , [64] . Similar functions are implemented by the second level of the present recognition model: The second level encodes the expectation of specific spatiotemporal patterns , i . e . it encodes auditory memory by attractors that correspond to specific vocal tract dynamics ( sounds ) . Note that there is a clear distinction between the third and second level in the model: While the third level encodes the expected sequence of sound dynamics , the second level encodes the repertoire of song sounds ( transcribed to sound waves by the vocal tract dynamics ) . This functional separation is also assumed to be implemented in the real bird brain [26] . In the primary auditory area , Field L , spectral-temporal receptive fields ( STRF ) have been proposed to explain the selective responses of neurons [67] . These selective responses may correspond to the recognition dynamics at the first level in the model which decodes the detailed spatiotemporal structure of the auditory stimulus guided by higher level predictions . It is interesting to note that we could use the present recognition model to derive , as done experimentally [67] , the spectral-temporal receptive fields at the first level . Alternatively , one could use experimentally acquired STRFs to adapt the first level of the present model to establish exact equivalence of the model and the real system at the level of primary auditory areas . There are several models that focus on the sequential activation of HVC ( RA ) neurons using single neuron models . Inhibition is believed to be a key element to generate rhythmic ( sequential ) activity in HVC [15] , [16] , [68] . We used winnerless competition which relies on inhibition to sequentially activate HVC ( RA ) ensembles . A similar generation mechanism as described here can be obtained using the synaptic chain scheme: Li and Greenside [12] proposed a conductance-based model for HVC ( RA ) neurons from which they obtained sequential multi-spike bursts . Later , Jin et al . [13] used an intrinsic bursting mechanism to obtain higher firing rates more consistent with the experimental data . This scheme was extended in [14] and was shown to produce robust and highly stereotyped sequential bursts . A learning mechanism was proposed in [69] showing how a sparse temporal code can emerge from a recurrent network . The models mentioned above focus on describing possible ways for the sequential activity of HVC where the downstream areas can be regarded as driven in a feed-forward fashion by HVC . A comprehensive generative model that includes HVC , RA and motor control areas was described in [18] . This study showed that the intrinsic connectivity at the RA level can substantially influence the acoustic features of syllables . This approach is similar to the present where the common research question is which parameterization ( connectivity ) of a recurrent neural network will generate motor control signals that result in realistic acoustic features of birdsong . However , we additionally incorporated recent findings [26] which point to a specific role of RA ensembles in encoding sound wave modulations . Furthermore , we provide evidence that the hierarchical setting of HVC and RA ensembles is the basis for robust and rapid song recognition . Theunissen et al . [70] estimated spectral-temporal receptive fields ( STRF ) of nonlinear auditory neurons using natural sounds as sensory input . The STRFs describe which temporal succession of acoustical features would elicit the maximal neural response and provide useful information for modeling perception of acoustic features , e . g . in the primary auditory area , Field L [67] . A two-level model was introduced [71] where the first level encoded frequency responses identified by an STRF analysis and the second level used these features to model song selective responses of HVC neurons . In another approach , Larson et al . [72] proposed a model for auditory object recognition where the first level uses a distance metric to distinguish between different spike trains and the second level acts as a decision network . However , both of these models propagate auditory signals in a feed-forward fashion from the low to the high level while the present scheme uses dynamical and recurrent bottom-up and top-down message passing thereby providing a more comprehensive model of the neuronal dynamics observed during song recognition . Learning models such as [73] and [74] were proposed which also include birdsong production and evaluation . These models mainly focus on the neural mechanisms of learning but they also provide mechanisms for song evaluation . There have been also attempts for the automated recognition of birdsongs using machine learning methods , e . g . [75] , [76] . However , these models are not concerned with neurobiological plausibility but rather use ad-hoc techniques as used in automated speech recognition , i . e . hidden Markov models and template-based matching of song syllables . There are several implications for future experiments which one can derive from the present model . The first is that we observe prediction errors at all levels when there is an unexpected piece of song ( Figure 9 ) or a song which is slower than expected ( Figure 12 ) . This suggests that there may not be a single area in the auditory pathway ( such as HVC ( X ) or LMAN in the anterior forebrain pathway ) that acts as a comparator between the stimulus and previously memorized tutor song [77] but several levels of the auditory pathway may be involved in this comparison . Comparing the neuronal recordings from a bird that listens to a normal speed song and a slower version of the same song might reveal the locations where these prediction errors are computed . Similar experiments have been done in auditory areas Field L and caudal lateral mesopallium ( CLM ) where some neurons responded robustly to perturbations in vocalization or playback of the bird's own song [21] . A functional model like the one presented here could predict what amount of activity should be expected in experiments given defined deviations , at different levels of the recognition hierarchy . Parallel to this idea , a recent experiment explained the activity in CLM by the level of surprise in the stimulus [20] . Our model could be used to predict the amount of surprise or prediction error at different hierarchical levels . As the present model covers much of the auditory pathway , this prediction technique may be best suited for using functional MRI on birds [78] , [79] where one would model increased activation , relative to some baseline condition , as an increase in prediction error . As noted by several authors , human speech and birdsong have in common that both are complex , hierarchical , sequenced vocalizations which are repetitions and combinations of simple units such as phonemes and syllables [2] , [80] , [81] . Although human speech is far more complex than birdsong , the underlying anatomical and functional features show striking similarities such as the pathways for vocal production , auditory processing and learning [22] , [81] . Songbirds , similar to humans , gain their vocal abilities early in life by listening to adults , memorizing , and practicing their songs [22] . These similarities suggest that one may derive insight about human speech recognition and learning from findings in birdsong research [82] . The present results clearly point to the usefulness of a hierarchical recognition structure to decode sequences of syllables . Such hierarchical models are rarely used in automated speech recognition [83] presumably because the standard model , the hidden Markov model , is mathematically best understood only in a non-hierarchical setting . The present scheme shows that complex spectral dynamics such as birdsong may be modeled as a sequence of nonlinear dynamics , where , in the generative model , each level drives the subordinate level in a highly non-linear fashion . To invert such a hierarchical , nonlinear , dynamical system , one requires sophisticated Bayesian inference machinery [11] , [84] . We described such a mechanism previously for a simple auditory sequence of sounds [85] . The novelty of the current approach is that we use a neurobiologically plausible generative model to derive a functional recognition model that has also the potential to recognize real and complex birdsong . In addition , we hypothesize that the specific arrangement of HVC and RA level ( dynamic sequences driving attractor dynamics at a lower level ) and its Bayesian online inversion will not only play a role in birdsong recognition models but may be successfully used for automated speech recognition as well . The mathematical model that we used to generate birdsongs was previously shown to produce accurate copies of songs such as canary [30] , chingolo sparrow [17] , white-crowned sparrow [31] and cardinal [45] songs . The vocal organs of other birds , e . g . of the zebra finch , can generate highly nonlinear , more complex , acoustic dynamics than the one considered here . For modeling such songs , one would have to replace the syrinx model of Eq . ( 4 ) by a more involved syrinx model such as the one reported in [86] . For our purposes , we focused on one particular song to describe the generation and recognition framework . The recognition of different songs either by the same or different conspecifics could be modeled by using multiple sequences encoded at the third level , where we assume that the recognition will converge to the best fitting sequence . In addition , one could adapt the nonlinear syrinx model to endow a singing bird with its own low-level acoustic characteristics . In the present model , we used rather small numbers of ensembles for visualization and computational purposes . The generative model applies to an arbitrary number of ensembles and similar type of dynamics can be obtained with larger number of ensembles at each level ( see Figure S1 for generation with 100 HVC ensembles ) . For recognition , we performed similar experiments with larger numbers of HVC ensembles ( 32 ) and RA patterns ( 24 ) where the recognition results were as robust as with the reported smaller size models ( see Figure S2 for the simulation ) . This indicates that the model scales to larger model sizes . However , there are two main issues that one will need to address to enable recognition using hundreds of units: ( i ) The computational power required for the recognition quickly increases with the number of ensembles used ( with complexity due to computing a matrix exponential , see [11] ) . This can be resolved by parallelizing the ensemble-specific computations which would be a further step towards biological reality . Currently , we emulate these parallel computations using a single-process Matlab implementation . ( ii ) The complexity of the syrinx model must be matched by the ‘descriptive power’ of the RA level . In other words , if one wanted to increase the number of RA ensembles significantly , one also had to render the model at the syrinx level more complex so that the recognition can infer more RA ensembles from more complex sensory data . However , this increase in model complexity at the syrinx and RA levels would require a more sophisticated syrinx model and is beyond the scope of the present work , in which we provide a proof of concept and introduce the computational framework . Furthermore , we tested the sensitivity of the Bayesian recognition in response to changing specific details of the generative model: ( i ) We used higher noise levels ( standard deviation of and ) as compared to the simulations above , the recognition still robustly inferred the hidden states and causes at all levels ( see Figure S3 ) ( ii ) We found that the recognition is robust against varying the initial conditions of the states in both the generative model and recognition . We tested a wide range of random initial conditions in both generation and recognition and observed that in all simulations the recognition quickly locks into the necessary dynamics . This implies that the listening bird can recognize a song reliably whatever the initial state of itself or the singing bird at the beginning of the song . ( iii ) We also changed the connectivity matrices at the third level ( with the constraint of high inhibition from the previous neuron and low inhibition to the next neuron ) and at the second level ( with the constraint that global stability conditions are satisfied , see Theorem 1 in Text S1 ) of the generative and recognition models . The recognition was still robust with these different connectivity matrices ( see Text S1 and Figure S4 ) . We described a model to generate artificial birdsongs and a scheme for their online recognition . We constructed a model based on key experimental findings in birdsong generation . Our results show that the specific , hierarchical mechanism how birdsong is generated enables robust and rapid decoding by a hierarchical and dynamic Bayesian inference scheme . We have interpreted this as evidence that the birdsong generation mechanism is geared toward making the song robustly decodable by conspecifics and discussed the experimental evidence that songbirds use a recognition mechanism similar to the present Bayesian inference scheme . | How do birds communicate via their songs ? Investigating this question may not only lead to a better understanding of communication via birdsong , but many believe that the answer will also give us hints about how humans decode speech from complex sound wave modulations . In birds , the output and neuronal responses of the song generation system can be measured precisely and this has resulted in a considerable body of experimental findings . We used these findings to assemble a complete model of birdsong generation and use it as the basis for constructing a potentially neurobiologically plausible , artificial recognition system based on state-of-the-art Bayesian inference techniques . Our artificial system resembles the real birdsong system when performing recognition tasks and may be used as a functional model to explain and predict experimental findings in song recognition . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
] | [
"circuit",
"models",
"motor",
"systems",
"nonlinear",
"dynamics",
"mathematics",
"neural",
"networks",
"computational",
"neuroscience",
"biology",
"computational",
"biology",
"sensory",
"systems",
"neuroscience",
"neurophysiology",
"coding",
"mechanisms"
] | 2011 | A Hierarchical Neuronal Model for Generation and Online Recognition of Birdsongs |
CpG islands ( CGIs ) are dense clusters of CpG sequences that punctuate the CpG-deficient human genome and associate with many gene promoters . As CGIs also differ from bulk chromosomal DNA by their frequent lack of cytosine methylation , we devised a CGI enrichment method based on nonmethylated CpG affinity chromatography . The resulting library was sequenced to define a novel human blood CGI set that includes many that are not detected by current algorithms . Approximately half of CGIs were associated with annotated gene transcription start sites , the remainder being intra- or intergenic . Using an array representing over 17 , 000 CGIs , we established that 6%–8% of CGIs are methylated in genomic DNA of human blood , brain , muscle , and spleen . Inter- and intragenic CGIs are preferentially susceptible to methylation . CGIs showing tissue-specific methylation were overrepresented at numerous genetic loci that are essential for development , including HOX and PAX family members . The findings enable a comprehensive analysis of the roles played by CGI methylation in normal and diseased human tissues .
DNA methylation in the mammalian genome arises due to covalent addition of a methyl group to the 5′ position of cytosine in the context of the palindromic dinucleotide , CpG . This modification is established and maintained by a family of DNA methyltransferases that are essential for development and viability [1 , 2] . The pattern of CpG methylation in the human genome distinguishes two fractions with distinct properties: a major fraction ( ∼98% ) , in which CpGs are relatively infrequent ( on average 1 per 100 bp ) but highly methylated ( approximately 80% of all CpG sites ) , and a minor fraction ( <2% ) that comprises short stretches of DNA ( ∼1 , 000 bp ) in which CpG is frequent ( ∼1 per 10 bp ) and methylation-free . The latter are known as CpG islands ( CGIs ) and they frequently colocalise with the transcription start sites ( TSSs ) of genes [3 , 4] . Although CGIs are often free of methylation , there are circumstances in which they become heavily methylated , and this invariably correlates with silencing of any promoter within the CGI . Artificial methylation of CGI promoters has long been known to extinguish transcription when the constructs are introduced into living cells [5] . Moreover , demethylation of endogenous methylated CGIs using DNA methytransferase inhibitors can restore expression of the gene [6] . These findings demonstrate that dense CpG methylation prevents expression of CGI promoters . Because of this biological consequence , it is important to know the extent of CGI methylation in both normal and diseased tissue states . The classical example is X chromosome inactivation in placental mammals , during which hundreds of CGI promoters become methylated and contribute to the stability of gene inactivation on this chromosome [7 , 8] . Genomic imprinting can also depend upon differential CGI methylation between maternal and paternal alleles [9] . Certain “testis-specific antigen” genes possess CGIs that are methylated in all somatic tissues , but not in testis , where the genes are expressed [10] . Several additional candidates for CGI methylation in normal tissues have been reported [11 , 12] , and the number of cases has recently grown due to large-scale bisulfite sequencing [13] and analysis of promoter methylation using microarrays [14] . In the cases of X chromosome inactivation and genomic imprinting , the biological processes were described initially , and CpG methylation was subsequently implicated through mechanistic studies . To uncover new biological roles for CGI methylation in hitherto undiscovered biological processes , it would be advantageous to comprehensively screen genomic DNA for methylated CGIs in normal or diseased cell types . A persistent limitation affecting this kind of approach has been uncertainty concerning CGI identification [15] . The criteria for designating a sequence as CGI-like are currently exclusively bioinformatic in nature , relying on the differences in the base composition and CpG frequencies ( observed/expected ) between bulk genomic DNA and CGIs [16 , 17] . In an attempt to address this limitation and create a resource for future analysis , we developed a method for CGI identification and purification based on their lack of CpG methylation in an otherwise highly methylated genome . Our method utilised a protein domain with a specific affinity for clustered nonmethylated CpG sites [18 , 19] . Using this reagent we physically purified DNA sequences that contain clusters of nonmethylated CpG-rich DNA from human blood DNA . Large-scale sequencing of the fraction identified a CGI set that was annotated on the ENSEMBL database . We found that many CGIs in the set were not associated with promoters of annotated genes , but were either within transcription units or between genes . By arraying the intact CGI sequences , we were able to interrogate genomic DNA fractions from several human tissues in order to identify methylated CGIs . The results revealed large numbers of CGIs that are methylated in normal human tissues , many of which showed tissue-specific methylation .
To enrich for nonmethylated CpG-rich DNA ( CpG islands ) , we developed the technique of CXXC affinity purification ( CAP ) . This uses the cysteine-rich CXXC3 domain that has a high affinity for nonmethylated CpG sites [18 , 19] . A recombinant CXXC domain from mouse Mbd1[19] was expressed in bacteria , and its binding specificity for nonmethylated CpG sites was confirmed ( Figure 1A ) . The CXXC domain had no detectable affinity for DNA containing only methylated CpGs or for DNA lacking CpGs altogether . We linked the CXXC domain to a sepharose matrix and confirmed that this fractionated DNA fragments according to CpG density and methylation status ( unpublished data ) . All DNA bound to the column at 0 . 1 M salt . Methylated DNA and CpG-poor DNA eluted at ∼0 . 4 M NaCl , whereas elution of nonmethylated CpG-rich DNA required 0 . 6–1 . 0 M NaCl . To test the behaviour of CGIs on the column , human genomic DNA was digested with MseI ( TTAA ) [20] and fractionated over the CXXC column ( Figure 1B ) . The reasoning behind use of Mse1 [20] was to cut AT-rich bulk genomic DNA into small fragments ( predicted average = 123 bp ) , but to leave CGIs relatively intact ( predicted average = 625 bp ) . As bulk genomic DNA has a CpG on average every 100 bp , most Mse1 fragments will have too few CpGs to be retained by the CXXC matrix . CGIs on the other hand , with 1 CpG per ∼10 bp , will give rise to long fragments with many CpGs . Eluted fractions were interrogated by PCR using primers specific for a range of known CGIs and non-CGI sequences ( Figure 1C ) . For example , the nonmethylated CGI of the P48 gene eluted at high salt . The X-linked monoamine oxidase ( MAO ) gene eluted as a single high salt peak from male genomic DNA ( where it is nonmethylated ) , but as two separate peaks at high and low salt when female DNA ( with one methylated and one nonmethylated allele ) was fractionated . The CGI associated with the NYESO testis-specific antigen gene ( methylated in somatic tissues ) eluted from the CXXC column by low salt as predicted . The data confirm that CAP may be used to purify a CGI fraction from human genomic DNA . Most or all CGIs are in a nonmethylated state in sperm , but in addition repetitive elements [21] and telomere-proximal sequences [22] , both of which are moderately CpG-rich , are hypomethylated in sperm DNA . To avoid contamination of the CGI fraction with sequences that are nonmethylated , specifically in germ cells , whole human blood was used as a source of CGI fragments . Pooled whole blood DNA from three males was fractionated using the CXXC column . High salt fractions were pooled , diluted , and re-chromatographed before cloning in plasmids . The resulting blood CGI library was analysed by 221 , 860 sequence reads representing 119 , 487 genomic templates . These compiled to give 28 , 013 unique MseI fragments . Plots of DNA insert length versus either G+C content or observed/expected CpG frequency ( CpG[o/e] ) showed that the great majority of clones exhibited a higher G+C content ( average = 62% ) and CpG[o/e] ( average = 0 . 71 ) than bulk genomic DNA ( G+C = 41% and CpG[o/e] = 0 . 2 ) ( Figure 2A and 2B ) . A fraction of small fragments with sequence characteristics resembling bulk genomic DNA was detected by these plots . As these probably represent contamination , we filtered out fragments shorter than 512 bp that had a GC content less than 50% and/or a CpG[o/e] less than 0 . 6 ( see grey dots in Figure 2A and 2B ) . The resulting final sequenced set corresponds to 17 , 387 CGIs and is annotated on the ENSEMBL genome browser ( http://www . ensembl . org/index . html . DAS sources: “CPG island clones” ) . The great majority have classical CGI properties ( Figure 2C ) . Due to their high average GC content , the sequence pass rate was 69% . Assuming that the unsequenced clones reflect the same proportion of CGIs as those that were sequenced , we estimate the total number of CGIs in the library as 25 , 200 . It is likely that a higher proportion of sequence failures affect bona fide CGIs , as GC-richness is known to interfere with sequencing . If so , we estimate that the number of human genomic CGIs may be closer to 30 , 000 . CGIs are identified bioinformatically as DNA sequences with a base composition greater than 50% G+C and a CpG[o/e] of more than 0 . 6 [23] . The DNA length over which this condition applies is critical . Initially the threshold most often used was 200 bp , whereas 500 bp is now more commonly applied [17] . These two criteria are formalised as “NCBI-relaxed” and “NCBI-strict , ” respectively ( http://www . ncbi . nlm . nih . gov/mapview/static/humansearch . html#cpg ) . The relaxed algorithm predicts 307 , 193 CGIs in the human genome , which includes many repeated sequences and gene exons . Over 90% of NCBI-relaxed CGIs are not represented in either our library or the set predicted by the NCBI-strict . This and other arguments suggest that the great majority ( >90% ) are false positives . On the other hand , 77% of clones in the CGI library match CGIs predicted by the “NCBI-strict” algorithm ( Table 1 ) . Examples of the coincidence of NCBI-strict predicted CGIs and sequenced CGI clones are illustrated for the three typical regions of the human genome ( Figure 2D ) . Altogether , NCBI-strict identifies 24 , 163 CGIs in the human genome , which accords with the adjusted CGI library estimate of 25 , 200 . The coincidence of these numbers masks significant differences , however , as 23% of CGIs in the library are not detected by the NCBI-strict algorithm ( 4 , 082 out of 17 , 387; Table 1 ) . Four randomly selected examples of library CGIs not detected by NCBI-strict ( Figure 2D and 2E , numbered ) gave CpG maps resembling CGIs; three of these coincided with the promoters of annotated protein-coding genes ( Figure 2D and 2E ) . The presence of bioinformatically predicted CGIs that are missing from the CGI library is most probably due to sequence failure of ∼31% of library inserts . Analysis of the CGIs missed by the NCBI-strict algorithm shows them to be , as expected , significantly weaker with respect to CpG and G+C content than the total set ( Figure S1 ) . It was not obvious , however , that the algorithm could be easily improved based on this information . Relaxation of the sequence parameters reduces the number of false negatives , but leads to increased numbers of false positives . We suggest that CAP identifies islands that fail the NCBI criteria , but reduces the false discovery rate by excluding spurious methylated CpG-rich sequences . Like the majority of CGIs , most NCBI-missed islands are gene-associated , although with an increased incidence of intragenic islands ( Table S1 ) . The CGI library therefore includes a significant fraction of bona fide CGIs that are missed by one of the best available algorithms . CAP defines a set of CGIs that is coherent with respect to clustering of nonmethylated CpG sites . The genomic distribution of these CGI sequences correlates strongly with gene density ( Figure 2F ) . For example , gene-rich Chromosome 19 is also CGI-rich , whereas gene-poor Chromosome 18 is correspondingly CGI-poor . With respect to annotated protein-coding genes , we found that 76% of CGIs are within 1 . 5 kb of a transcription unit , but only 49% overlap with the TSS ( Table 2 ) . It follows that half of CGIs are not TSS-associated , but are either within downstream regions of transcription units ( 22% ) or located in intergenic DNA . Previous studies have detected CGIs at the TSS of 56% of human protein-coding genes [24] . As 43 . 5% of TSSs overlap sequenced CGIs , we calculate that the sequenced set of 17 , 387 CGIs represents 78% of the CGI complement . According to this calculation , the total CGI number would be 22 , 400 , somewhat less than the figure of 25 , 200 deduced from the fraction of sequenced inserts . CAP selects CGIs from blood DNA based on their lack of methylation and therefore excludes the small fraction of CGIs ( <3% ) that are fully methylated in somatic cells from the set [14] . Indeed , CGIs associated with the human testis-specific antigen genes [10] , which are methylated in somatic tissues , were not enriched by CAP ( Figure 1C ) or present in the library ( unpublished data ) . Despite the absence of these fully methylated CGIs , we reasoned that the blood CGI library provides an opportunity to screen for methylation that affects a fraction of all copies of a specific CGI in whole blood DNA . Also , it permits a screen for differential methylation of CGIs in tissues and cell types other than blood . To investigate CGI methylation in normal human tissues , we constructed an array of sequenced CGIs from the library by immobilising single-stranded PCR-amplified inserts on glass slides using 5′-aminolink chemistry as described ( http://www . sanger . ac . uk/Projects/Microarrays/arraylab/methods . shtml ) . As probes for the array , methylated CGIs were enriched from genomic DNA using MBD affinity purification ( MAP ) , which was shown previously to efficiently bind methylated CGIs [20] ( Figure 3A and 3B ) . Human male and female blood DNA was MseI-digested and ligated to universal catch linkers . We verified by PCR that affinity fractionation using MAP effectively separated known methylated CGIs ( XIST on the active X chromosome and NYESO ) from bulk genomic DNA and nonmethylated CGIs ( P48 and XIST on the inactive X chromosome; see Figure 3B ) . Male and female DNA fractions were pooled after two rounds of MAP , amplified by linker-mediated PCR , cyanine labeled , and hybridized to the CGI microarray . Quadruplicate hybridisations ( inclusive of cyanine dye swaps ) gave mean enrichment values ( MAP/Input ) that allowed a comparison between male and female methylated CGI complements . As expected , these were positively correlated ( R = 0 . 865 Pearson correlation ) suggesting similar overall patterns . As the library comprises MseI fragments that sometimes overlap minimally with the cognate CpG-rich region , we chose to disregard data from spots that contained DNA with an average CpG frequency ( observed/expected ) of less than 0 . 5 . Although the omitted fragments often denote CGIs , they include too little of the CpG-rich domain to be reliable for detection of MAP probes . This refinement reduced the number of analysable CGIs on the array to 14 , 318 . To assess the relationship between hybridization signal relative to input and degree of enrichment by MAP , we measured a selection of CGIs in the probe by quantitative PCR and compared this data with the M values ( log2 [MAP signal]/[Input signal] ) for those sequences ( Figure 3C ) . The results established that M values greater than 1 . 5 denote CGIs that are significantly enriched by MAP and therefore methylated . CGIs of the BEST1 and R4RL1 genes were predicted to be nonmethylated ( M = 0 . 2–0 . 4 ) and methylated ( M = 2 . 2–2 . 8 ) , respectively , based on the array data . Bisulfite genomic sequencing confirmed this expectation ( Figure 3G and 3H ) . The major difference in CGI methylation between male and female DNA was expected to be due to X chromosome inactivation ( see also [25] ) . We therefore compared the methylation status of CGIs on Chr 16 and Chr X in male versus female DNA . Chr 16 CGIs did not vary between males and females , whereas Chr X CGIs were significantly enriched in female DNA as predicted ( Figure 3D–3F; Table S2 ) . Studies of human X chromosome inactivation have indicated that a proportion of genes escape inactivation and are therefore expressed from both chromosomes [26 , 27] . By comparing the microarray data for a set of inactivated and escaping CGIs , we found that inactivated genes had significantly higher M values ( p-value = 1 . 213 ×10−5 ) ( Figure 3I ) . This finding affirms the long-standing link between CGI methylation and gene silencing and validates the present experimental system as a means of detecting genes that are shut down in this way . Methylation of CGIs on the inactive X chromosome and at imprinted genes is well known , but CGI methylation at other chromosomal loci in normal cells and tissues is incompletely characterized [12 , 13 , 28 , 29 , 30] . To investigate this issue on a large scale , we probed CGI arrays with MAP fractions from genomic DNA ( three individuals per pool ) of brain , muscle , spleen , and sperm in addition to blood ( Figure 4A ) . MAP enrichment of methylated CGIs in sperm DNA consistently failed to generate enough DNA for labeling using our standard PCR amplification conditions and was therefore not analysed further . We conclude that the level of CGI methylation in sperm is far lower than in any of the somatic tissues . Taking M values greater than 1 . 5 to signify methylation , we observed between 5 . 7% and 8 . 3% of CGIs methylated in the somatic tissues that were tested ( Figure 4B; Table 3; Dataset S1 ) . Some CGIs were methylated in common between all the tested somatic tissues , whereas others were methylated in only one or a subset of the tissues . We noted that methylated CGIs disproportionately involved those that are remote from the TSS of an annotated gene . In the dataset as a whole , only 8% of TSS CGIs showed evidence of methylation in at least one tissue , whereas 22% of 3′ CGIs were methylated ( Table 4 ) . Do the methylated CGIs differ in sequence characteristics from CGIs that remain methylation-free ? We plotted the CpG[o/e] frequencies of 1 , 657 CGIs that acquired methylation in one or more tissues and found a mean CpG[o/e] of 0 . 77 compared with 0 . 75 for methylated CGIs ( Figure 4C ) . Though statistically significant ( p-value = 1 . 413e-10 ) the biological significance of this small difference is unclear . We checked by bisulfite sequence analysis a panel of seven CGIs with M values suggestive of tissue-specific methylation ( M values differing between tissues by >0 . 75 ) . In each case , bisulfite data confirmed the microarray predictions . CGI I1878 is not associated with an annotated gene ( ±1 . 5 kb ) and is methylated exclusively in muscle and brain ( Figure 4D ) . CGI I2985 spans the transcription start site of the SEC31B gene , whose product is implicated in vesicular trafficking , and is compositely methylated only in blood and spleen ( Figure 4E ) . CGIs I13406 ( Figure 4F ) and I12175 ( Figure 5A ) are methylated specifically in muscle . These overlap the predicted gene 67313 and the 3′ end of OSR1 . CGI I3654 , which is associated with the promoter region of an annotated PAX6 transcript ( Q59GD2 ) , previously shown to contain methylated CpG sites [31] , is specifically methylated in brain ( Figure 5B ) . I11878 is a 3′ CGI of ZN649 and is only methylated in spleen ( Figure 4G ) . Many methylated CGIs were associated with genes that are essential for development ( Figure 5 ) . This was confirmed by analysis of gene ontology , which showed significant overrepresentation of genes whose products are involved in developmental processes , including ectoderm and mesoderm development , neurogenesis , and segment specification ( Table S3 ) . Transcription factors , including homeobox family members and other DNA binding proteins , were twice as abundant as expected by chance . Other gene categories did not show significant enrichment . Among the CGIs whose methylation status was confirmed by bisulfite sequencing , PAX6 is involved in eye development and neurogenesis [32] , the HOXC cluster lays down the embryonic body plan , and OSR1 is related to a gene involved in Drosophila gut development . We examined the extended HOXC and PAX6 loci for CGI methylation status using the MAP-CGI array data . Our library identified 19 CGIs within the 150-kb HOXC gene cluster of which eight were methylated differentially in blood , muscle , and spleen ( Figure 5C ) . Brain was the only tissue that lacked obvious HOXC CGI methylation . Of nine CGIs near PAX6 , two showed differential methylation . In addition to brain-specific methylation of the PAX6-Q59GD2 CGI ( see Figure 5B ) , we observed methylation of a CGI upstream of the major PAX6 promoter in muscle and brain ( Figure 5D ) . The majority of CGIs identified as methylated by MAP-CGI array hybridization display composite methylation ( Figures 3 , 4 , and 5 ) , whereby DNA strands at a specific locus were either heavily methylated or essentially nonmethylated . This can explain why CGIs that were initially selected by being nonmethylated in blood DNA ( by CAP ) nevertheless register as methylated by MAP-CGI array analysis . One potential explanation for composite CGI methylation is that different individuals within the tissue pools exhibit different CGI methylation . To look for such “polymorphism , ” we examined CGI I5134 , which is within the HOXC cluster and shows composite methylation by bisulfite genomic sequencing . Analysis of individuals by MAP-CGI arrays showed highly significant differences between individual C and individuals A and B ( Figure 5E ) . This strikingly confirms individual variability in methylation at this CGI . Another potential explanation for composite CGI methylation is that cell types within the tissue sample possess different CGI methylation profiles . Blood , for example , consists of monocytes and granulocytes , each of which is subdivided into other cell types . As CGI I2985 was methylated at about half of DNA strands in blood , we tested the level of CGI methylation in DNA from monocytes and granulocytes separately . The results showed that monocytes had high methylation levels at this CGI , whereas granulocytes had very low methylation ( Figure 5F ) . These findings indicate a developmental origin for cell type–specific methylation at this genomic CGI .
We describe the characterisation of a comprehensive , verified CGI set derived from human blood genomic DNA that will be beneficial for studies of CGIs in normal human tissues and in disease settings . By focusing on CGIs alone , we excluded ∼98% of the genome from our analysis . While it will ultimately be important to know in detail the methylation status of whole genomes , this currently represents a technical challenge that has been addressed only for the small-genomed plant Arabidopsis [33 , 34] . These studies used indirect microarray-based methods for mapping DNA methylation that depend upon probes enriched in methylated domains . Current enrichment methods require clusters of CpG methylation , which are notably absent from the CpG-deficient majority of the mammalian genome . As a result , much bulk genomic DNA is beyond the resolution limit of this approach . Whole genome bisulfite sequencing , the most direct and reliable method for mapping methylated sites , has not yet been attempted in any organism . We therefore decided to study a discrete genomic fraction with evident biological relevance whose methylation status can be interrogated using microarray-based methods . To isolate nonmethylated duplex CGIs from total genomic DNA , we harnessed the binding specificity of the CXXC protein domain . Extensive sequencing of the resulting library confirmed that CGIs represent a discrete fraction of the human genome with shared DNA sequence characteristics . The present CGI set supercedes a previous human CGI library that was prepared in our laboratory using an indirect affinity purification procedure [20] . The initial library was not comprehensive and appears to have acquired significant levels of non-CGI contamination following amplifications . We estimate that the new library represents ∼25 , 000 CGIs , of which ∼60% have been arrayed as full-length single strands on glass slides . Additional analysis of inserts that initially failed conventional sequencing strategies will generate an array that covers the great majority of CGIs that are nonmethylated in human blood . The choice of blood DNA as a starting material necessarily excludes from the set any CGIs that are nonmethylated in germ cells , but densely methylated in the soma [14] . In the future , it will be instructive to compare an exhaustive sequence analysis of this set with comparable sequences isolated by CAP from sperm DNA . The library prepared using CAP defines CGIs based on the empirical criterion of clustered nonmethylated CpGs , whereas criteria based purely on base sequence and composition necessarily ignore methylation status . Comparing our set with predicted CGIs on the NCBI database shows good overlap with predictions based on the “strict” algorithm . The CGI library did , however , identify 23% of CGIs that were negative by this criterion . This suggests that the software for DNA sequence-based CGI identification misses almost one in four CGIs that the more biological criterion of CAP is able to include . Recent CGI analyses identified large numbers of human CGI promoters that are enriched in methylation at lysine 4 of histone H3 , a mark of transcriptional activity [14 , 35 , 36] . Since it has been proposed that hypomethylation is dependent on germ line and early embryonic transcription [3] , we determined the overlap between our CGI set and the H3K4 sites in human embryonic stem cells [37] . We calculate that 90% of CGIs in the filtered set ( 14 , 318 ) coincide with H3K4 methylated promoters that were reported in the chromatin study . A better test of the relationship between CGIs and H3K4 methylation islands in ES cells is to exclude promoters of annotated genes and focus on intra- and intergenic CGIs . Here again , a high proportion ( 75% ) of CGIs overlap with H3K4 methylation islands . These findings are compatible with the notion that the presence of CGIs is connected with specialised chromatin configurations in early embryonic cells . An intriguing proposal is that H3K4 methylation may be incompatible with docking of de novo methyltransferases [38] . This could in theory insure that these regions remain free of CpG methylation at a time when the rest of the embryonic genome is subject to global methylation . We found that 49% of CGIs overlap the TSS of an annotated gene . In considering the function of the half of CGIs that are remote from an annotated TSS , it is noteworthy that several intragenic CGIs have been shown to coincide with previously unforeseen promoters that initiate bona fide transcripts [39 , 40] . This raises the possibility that all CGIs function as promoters and are therefore TSS-associated [40] . In this connection , it is of interest that genome-wide analysis by tiling arrays detected over 10 , 000 unanticipated human transcripts , many of which may represent noncoding RNAs [41] . It is conceivable that many inter- and intragenic CGIs mark promoters that drive the synthesis of these novel transcripts . The noncoding transcripts XIST and AIR , for example , whose RNA products play regulatory roles [42–44] , both initiate within CGI promoters . The proximity of many methylated CGIs to developmentally important genes raises the possibility that putative CGI transcripts play regulatory roles during development . Recent analyses of the human HOX gene cluster highlight the functional importance of noncoding RNAs [45] . Large numbers of potential CGI promoters within HOX gene loci may therefore contribute to the regulation of these complex loci . CGI methylation has been extensively studied in cancers and their derivative cell lines , but relatively less attention has been paid to the phenomenon in normal tissues . Several studies have reported somatic CGI methylation , but in early examples the bioinformatics procedure used to identify these sequences was often equivalent to the NCBI-relaxed algorithm , which generates a large excess of questionable CGI candidates . The MASPIN gene , for example , scores as a methylated CGI promoter by the relaxed criterion [28] , but it is not detected as such either by the NCBI-strict algorithm or by CAP ( unpublished data ) . A recent report addressing the methylation status of 16 , 000 human promoters identified that 3% of TSS-associated CGIs are normally methylated in somatic tissue [14] , which is somewhat below the levels observed in our study ( 7 . 8%; Table 4 ) . We detect a much higher frequency of methylation at nonpromoter CGIs ( average = 16% ) , which are obviously absent from promoter arrays . In particular , 22% of CGIs near the 3′ ends of genes are methylated . Extensive bisulfite sequence analysis [13] surveyed 512 CGIs on Chrs 6 , 20 , and 21 and reported 9 . 2% to be methylated in somatic tissues . This is similar to the overall level of 11 . 6% methylation among 14 , 318 CGIs detected by our study ( Table 4 ) . Our findings raise important questions about the relationship of CGI methylation to gene expression . On the X chromosome , it is clear that methylated CGIs correlate with inactivated genes whereas unmethylated CGIs correlate with genes known to escape inactivation . The generalisation that CGI methylation silences promoters is therefore supported ( see also [25] ) . The relevance to gene expression of the autosomal methylated CGIs identified here is complicated by the frequent presence of both methylated and nonmethylated alleles in a specific tissue ( see below ) . This means that even if CGI methylation silences a promoter completely , large changes in gene expression are not to be expected . Also , many CGIs are not at promoters of annotated genes , but are within or between transcription units . Their function with respect to transcription , if any , may be positive or negative . Finally , any transcripts originating from these “orphan” CGIs have yet to be identified and cannot be tested . For these reasons , it is difficult to make predictions about the effect of CGI methylation on global transcription levels . We nevertheless mined published expression microarray data to determine whether tissues in which a specific set of promoter CGIs was methylated expressed the associated genes at a different level from tissues where the same CGI was unmethylated . The results showed no obvious correlation between CGI methylation and expression . This , therefore , remains an open question that demands detailed analysis of specific cases . Genes that play an important role in development were prominent among the set of methylated CGIs identified by MAP-CGI array hybridization . Out of 109 CGI-associated genes that contain homeobox-like domains , 27 ( ∼25% ) were unmethylated in at least one tissue compared with ∼11% of all CGI-associated genes ( see Table 4 ) . Specifically , we identified 79 CGIs in the four human HOX gene clusters A–D , of which 22 were methylated in at least one of the tissues that we tested . Given the relatively small selection of tissues analysed in the study , the actual frequency of HOX CGI methylation in all human tissues is likely to be higher than one in four . Interestingly , methylation of HOX gene CGIs is also reported in cancers [46] , raising the possibility that cancer CGI methylation patterns mimic patterns that arise during development . A potential link between normal development and cancer is suggested by the finding that CGIs methylated in cancer preferentially include promoters that are marked by association with polycomb group proteins in embryonic stem cells [47–49] . In contrast , we found little difference between the fractions of all CGIs ( 5 . 9% = 845/14 , 318 ) and of methylated CGIs ( 7 . 7% = 127/1 , 657 ) that were polycomb-associated in embryonic cells [37] . The origins of CGI methylation in cancer may be distinct from the mechanisms that lead to CGI methylation in normal tissues . It was reported that the most CpG-rich CGIs among 512 analysed on Chr 6 , Chr 20 , and Chr 22 were never methylated , suggesting that the CpG-richness may protect from methylation [13] . In a larger CGI set , we detected a very small , but statistically significant , difference in sequence properties between CGIs that become methylated and those that remain immune in the tested cell types . The mean CpG[o/e] was 0 . 75 for methylated CGIs compared with 0 . 77 for bulk CGIs ( Figure 4C ) . Bock and colleagues [50] identified sequence features that were predictive for CGI methylation , including specific repeats , sequence patterns , and DNA structure . Contrary to predictions of this method , methylated CGIs were significantly depleted in repetitive elements and showed no difference in predicted base twist . We did , however , observe small , but statistically significant , increases in simple sequence elements ( TGTG/CACA ) and base-stacking energy ( see Figure S2 ) . The biological relevance of these minimal differences is uncertain . Weber and coworkers [14] identified ∼2 , 000 promoters out of 16 , 000 that were more susceptible to methylation than CGIs themselves . These so-called “weak CpG islands” had an average CpG[o/e] ratio intermediate between CGIs and bulk genomic DNA . We have determined that 75% of weak CpG islands reported by Weber et al . are absent from the CGI library . Weak CGIs may be depleted because they are heavily methylated and therefore not enriched by CAP . Indeed , 22 methylated weak CpG islands [14] were not detected in our library . Alternatively , their relatively low CpG density and somewhat elevated frequency of Mse1 sites may result in too few CpGs per fragment for efficient retention by the CXXC matrix . Those CGIs that were methylated often showed a mixture of heavily methylated and nonmethylated strands by bisulfite analysis . There are several possible explanations for composite methylation patterns . Firstly , at the highest level , it is possible that different individuals contributing to the DNA pool are polymorphic with respect to this epigenetic mark . We analysed specific CGIs in muscle DNA from three individuals and found evidence of individual variation of this kind . A large-scale survey would be required to determine the extent of inter-individual variability . A second possibility is that cells within the analysed tissue are heterogeneous with respect to CGI methylation . Each of the analysed tissues consists of multiple differentiated cell types that should be analysed separately to address this possibility . Analysis of three compositely methylated CGIs in blood showed one that was highly methylated in monocytes , but weakly methylated in granulocytes , indicating that cell type–specific CGI methylation underlay heterogeneous DNA methylation . A third possible explanation for composite methylation is monoallelic methylation . A previous study of 149 CGIs on Chr 21q detected three that were mono-allelically methylated , indicating that this explanation also accounts for some cases of composite CGI methylation [12] .
Cloning of the His-CXXC construct from murine Mbd1a was described previously [19] . The MBD construct was subcloned from pET30bhMeCP2 [51] . A fragment of human MeCP2 corresponding to amino acids 76–167 was PCR-amplified and ligated into the Nde1 and EcoR1 sites of pet30b ( Novagen ) to generate a C terminally His-tagged pet30bMeCP2_76–167 . Primers: pet30bMeCP2_76–167Nde1 CGG TTC ATA ACC ATA TGG CTT CTG CCT CCC CCA AAC AGC GG and pet30bMeCP2_76–167EcoR1 CGG AAG TCA AAG AAT TCT CAT CAG TGG TGG TGG TGG TGC CGG GA . Recombinant peptides were purified from 10 l of induced BL21 ( DE3 ) pLysS ( Stratagene ) culture on Nickel Charged Fast Flow Chelating Sepharose ( GE Healthcare ) . The CXXC construct was further purified by cation exchange using Sp-Sepharose ( GE Healthcare ) cation exchange as previously described [51] . Recombinant protein was bound to Nickel sepharose prior to longer term storage . CXXC-EMSA was carried out essentially as described in [19] . Briefly , binding reactions including 0 , 250 , 500 , 1 , 000 , or 2 , 000 ng of purified recombinant His-CXXC were preincubated in 1× binding buffer ( 5 × binding buffer: 30 mM Tris-HCl [pH8] , 750 mM NaCl , 5 mM DTT , 30 mM MgCl2 , 15% Glycerol , 50 ng/μl BSA , and 0 . 05 μg/μl of poly ( dAdT ) ( Amersham ) . End-labeled CG11[52] probe ( 1 ng ) was added to each reaction and incubated for a further 25 min . Complexes between probe DNA and the CXXC domain were resolved on a 1 . 3% agarose Tris-borate-EDTA gel and imaged by Phosphor Imager ( Molecular Devices ) . Whole blood was collected from voluntary donors and used in anonymized pools . Donors were aware of , and consented to , its use for preparation of DNA . Monocyte and granulocyte cells were prepared from whole human blood using Ficoll gradient centrifugation . Whole blood ( 3 ml ) was layered onto an equivalent volume of Histopaque-1077 Ficoll ( Sigma-Aldrich ) and sedimented according to the manufacturer's instructions . Mononucleocytes were recovered from the plasma-ficoll interphase and granulocytes from the cell pellet . Whole human blood , monocyte , and granulocyte DNA was extracted using the Genomic-tip 500/G ( Qiagen 10262 ) genomic DNA purification kit . Sperm DNA was prepared as described [53] . Human skeletal muscle , spleen , and brain genomic DNAs were purchased from Ambion . 50–60 mg of recombinant CXXC was dialysed into W1 buffer ( 50 mM sodium phosphate buffer [pH8] , 300 mM NaCl , 10% glycerol , 15 mM ß-mercaptoethanol , 0 . 5 mM PMSF ) , bound to nickel-charged sepharose , and then washed with 10 column volumes ( CVs ) of W1 , 10 CVs of W2 ( W1 + 10 mM Imadazole ) , and 10 CVs of W1 . Beads were packed onto a 1-ml Tricorn chromatography column ( GE Healthcare ) . Mse1 digested male DNA ( 100 μg ) pooled from three individuals was bound to the CXXC column in 90% CA buffer ( 20 mM Hepes [pH7 . 9] , 0 . 1% Triton X-100 , 10% glycerol , 0 . 5 mM PMSF , 10 mM 2-Mercaptoethanol ) and 10% CB buffer ( CA + 1 M NaCl ) . Equilibrated DNA was then eluted over an increasing NaCl gradient of 10%–100% CB buffer ( Figure 1B ) . Fractions ( 3 ml ) were collected and 200 μl of each was precipitated and resuspended in 40 μl 1 × TE buffer . Aliquots were PCR- interrogated using Redhot taq DNA polymerase ( Abgene ) for XIST ( for CACGTGACAAAAGCCATG , rev GGTTAGCATGGTGGTGGAC ) , NYESO ( for CCCAGCGTCTGGTAACCATC , revCCACGGGACAGGTACCTC ) , MAO ( for CGGGTATCAGATTGAAACAT , rev CTCTAAGCATGGCTACACTACA ) , P48 ( for cagaaggtcatcatctgcca , rev tgagttgtttttcatcagtcca ) under the following conditions: 2 min at 94 °C; followed by 30 cycles of 94 °C for 50 s , Tann °C for 50 s , 72 °C for 1 min; and a final extension of 72 °C for 7 min . PCR products were resolved on a 1 . 5% TAE-agarose gel ( Figure 1B ) . Fractions retaining nonmethylated CpG-rich Mse1 fragments ( Figure 1B ) were pooled , diluted with CA buffer , and re-chromatographed . The relevant fractions were precipitated and ligated into the NdeI site of pGEM5zf- ( Promega ) . The clone set was arrayed into 384-well plates in glycerol for long-term storage . Copies were taken and DNA prepared for sequencing using a modified alkaline lysis method . Cells were lysed in glucose , Tris , EDTA ( pH 8 ) buffer plus NaOH and SDS and spun through Millipore Montage filter plates directly into propan-2-ol to precipitate , followed by elution in water . In all , 172 , 800 clones were sequenced forward and reverse using T7 and SP6 primers and BigDye V3 . 1 chemistry , under the following conditions: 30 s at 96 °C , followed by 44 cycles of 92 °C 8 s , 55 °C 8 s , 60 °C 2 min . Samples were separated using 3730 XL sequencers ( Applied Biosystems ) . Extraction was performed using sequence analysis v3 . 1 , and base-called using Phred [54] . DNA sequences were identified using NCBI36 and mapped using the ENSMBL Genome Browser ( http://www . ensembl . org/index . html ) . CGIs that mapped within 1 . 5 kb of annotated genes were considered to be gene-associated in order to take into account mis-annotation of transcription start sites within poorly defined 5′ UTRs . Amino-linked clone insert amplicons were generated by vector-specific PCR in 50 mM KCl , 5 mM Tris ( pH 8 . 5 ) , and 2 . 5 mM MgCl2 including 1 M Betaine ( 10 min at 95 °C; followed by 35 cycles of 95 °C for 1 min , 60 °C for 15 s , 72 °C for 7 min; and a final extension of 92 °C for 10 min; 5′ aminolink forward primer 5′ - CTC ACT ATA ggg CgA ATTg g −3′ reverse primer 5′ -CgC CAA gCT ATT TAg gTg AC-3′ ) . PCR products were ethanol-precipitated and resuspended in 1 × microarray spotting buffer ( 250 mM sodium phosphate [pH 8 . 5] , 0 . 01% sarkosyl , 0 . 1% sodium azide ) . Arrays were spotted onto amine-binding slides at 20–25 °C , 40%–50% relative humidity . After an overnight incubation in a humid chamber , the slides were blocked ( 1% ammonium hydroxide for 5 min , followed by 0 . 1% SDS for 5 min ) and denatured ( 95 °C ddH2O for 2 min ) , rinsed in ddH2O , and dried by centrifugation for 5 min at 250 ×g . Human tissue DNA pooled from three individuals was digested with Mse1 , phosphatase-treated , and ligated to 5 μmol of phosphorylated catch-linkers ( upper_GGT CCA TCC AAC CGA TCT and lower_CCA GGT AGG TTG GCT AGA AT phosphate ) that had been annealed in 1× TE for 5 h . DNA was bound to an MBD chromatography column and affinity-purified essentially as described [20] . Fractions containing methylated CpG-rich Mse1 fragments were pooled and re-chromatographed before precipitation ( Figure 3B ) . Purified DNA was resuspended in 1×TE and amplified in parallel with input DNA , using the GC Rich PCR system ( Roche; 2 min at 95 °C; followed by 18 cycles of 95 °C for 1 min , Tann °C for 1 min , 72 °C for 4 min; and a final extension of 72 °C for 7 min; universal primer_GGT CCA TCC AAC CGA TCT TA ) . MAP and Input DNAs ( 200 ng ) were fluorescently labeled by random priming using the Bioprime labeling kit ( Invitrogen ) , 1 × dNTS ( 10× dNTPS; 2 mM of each dATP , dGTP , dTTP , 1 mM dCTP ) and 1 . 5 nmol of Cy3 or Cy5-dCTP ( GE Healthcare ) . The labeled Input and MAP probes were purified ( Invitrogen “Purelink” ) , pooled , and precipitated with 100 μg of human Cot-1 DNA ( Invitrogen ) . Labeled DNA was resuspended in 400 μl of hybridisation buffer ( 2XSSC , 50% deionised formamide , 10 mM Tris-HCl [pH7 . 5] , 10% dextran sulphate , 0 . 1% Tween 20 ) , denatured at 100 °C for 10 min , snap-chilled on ice , and incubated for 1 h at 37 °C . The CGI microarrays were prehybridised with Cot-1 and herring sperm DNA ( Sigma ) before being hybridised for 48 h at 37 °C . Arrays were washed four times at 37 °C in 1× phosphate buffered saline/0 . 05% Tween 20 , three times at 52 °C in 1 × saline sodium citrate , twice at RT in 1 × phosphate buffered saline/0 . 05% Tween 20 , and finally rinsed in water , before being dried by centrifugation ( 500g ) . Arrays were scanned with a GenePix Autoloader 4200AL ( Axon ) and then processed using the GenePix Pro 6 . 0 ( Axon ) software package . All subsequent analysis was carried out with the LIMMA package in the R statistical environment . Features with poor signal-to-noise ratios were stabilised using a base value of 1 , 000 for background-subtracted intensities . Cy3 and Cy5 signals were transformed into M values ( log2[red/green] ) and normalised by print-tip loess . Each tissue analysis is represented by four microarrays comprising two independent replicates with respective dye swaps . Processed values were averaged through linear modeling and used to determine the relative enrichment of MAP DNA relative to Input . An M value of >1 . 5 was designated as the threshold for hypermethylation as determined by quantitative PCR ( Figure 3C ) and bisulfite genomic sequencing ( Figures 3G and 3H , 4D–4G , and 5A and 5B ) . This threshold was confirmed as significant by calculation of a t-statistic by eBayes modeling and BH multiple testing correction . Differential methylation was deduced when features displayed an M value >1 . 5 in one or more tissues and a differential of 0 . 75 between tissues ( upper boundary capped at M = 2 . 5 ) . To avoid complications due to X chromosome inactivation , CGIs on sex chromosomes were not included in the analysis . In addition , spots that gave no signal on the microarray ( NA values ) and spots containing DNA in which CpG[o/e] values were <0 . 5 were excluded . Real-time PCR was carried out on MAP and Input material with iQ SYBR Green Supermix ( Bio-Rad ) on an iCycler ( Bio-Rad ) according to manufacturer's instructions . For primer sequences see Table S4 . Bisulfite treatment of genomic DNA was carried out as described by Feil et al . [55] , and prepared for sequencing as outlined by Suzuki et al . [56] . Genomic DNA ( 5 μg ) was digested by EcoRI prior to bisulfite treatment , and precipitated after the desulfonation step . Samples were resuspended in 1 ×Tris-EDTA buffer for subsequent PCR and sequencing reactions . Bisulfite specific primers were designed both manually and with the aid of the MethPrimer software [3] ( sequences are available on request ) . PCR was carried out on the bisulfite-treated DNA using RedHot Taq DNA polymerase ( Abgene ) under the following conditions: 2 min at 94 °C; followed by 40 cycles of 94 °C for 50 s , Tann °C for 50 s , 72 °C for 1 min; and a final extension of 72 °C for 5 min . PCR fragments were cloned using the Strataclone PCR cloning system ( Stratagene ) and at least ten products amplified ( as above ) and sequenced ( BigDye Terminator v3 . 1 Cycle Sequencing Kit; Applied Biosystems ) . Methylation status and experimental quality control was carried out with the aid of BiqAnalyzer [57] . | The human genome contains about 22 , 000 genes , each encoding one of the proteins required for human life . A particular cell type ( e . g . , blood , skin , etc . ) expresses a specific subset of protein genes and silences the remainder . To shed light on the mechanisms that cause genes to be activated or shut down , we studied DNA sequences called “CpG islands” ( CGIs ) . These sequences are found at over half of all human genes and can exist in either the active or silent state depending on the presence or absence of methyl groups on the DNA . We devised a method for purifying all CGIs and showed that , unexpectedly , only half occur at the beginning of genes near the promoter , the rest occurring within or between genes . Notably , methylation of CGIs causes stable gene silencing . We tested 17 , 000 CGIs in four human tissues and found that 6%–8% were methylated in each . Genes whose protein products play an essential role during embryonic development were preferentially methylated , suggesting that gene expression during development could be regulated by CGI methylation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics"
] | 2008 | A Novel CpG Island Set Identifies Tissue-Specific Methylation at Developmental Gene Loci |
Triatoma dimidiata , currently the major Central American vector of Trypanosoma cruzi , the parasite that causes Chagas disease , inhabits caves throughout the region . This research investigates the possibility that cave dwelling T . dimidiata might transmit the parasite to humans and links the blood meal sources of cave vectors to cultural practices that differ among locations . We determined the blood meal sources of twenty-four T . dimidiata collected from two locations in Guatemala and one in Belize where human interactions with the caves differ . Blood meal sources were determined by cloning and sequencing PCR products amplified from DNA extracted from the vector abdomen using primers specific for the vertebrate 12S mitochondrial gene . The blood meal sources were inferred by ≥99% identity with published sequences . We found 70% of cave-collected T . dimidiata positive for human DNA . The vectors had fed on 10 additional vertebrates with a variety of relationships to humans , including companion animal ( dog ) , food animals ( pig , sheep/goat ) , wild animals ( duck , two bat , two opossum species ) and commensal animals ( mouse , rat ) . Vectors from all locations fed on humans and commensal animals . The blood meal sources differ among locations , as well as the likelihood of feeding on dog and food animals . Vectors from one location were tested for T . cruzi infection , and 30% ( 3/10 ) tested positive , including two positive for human blood meals . Cave dwelling Chagas disease vectors feed on humans and commensal animals as well as dog , food animals and wild animals . Blood meal sources were related to human uses of the caves . We caution that just as T . dimidiata in caves may pose an epidemiological risk , there may be other situations where risk is thought to be minimal , but is not .
Humans have been attracted to caves for much of our history for purposes as varied as religious ceremonies to simply shelter . In Guatemala people use caves for several cultural practices including religious ceremonies , tourism , and shelter [1] . In the department of Petén in northern Guatemala , caves are often used by hunters for sleeping and shelter from rain , especially in the rainy season . Also in northern Guatemala , caves in the department of Alta Verapaz are sacred places and sometimes the site of Mayan religious ceremonies; tourists also visit some caves . The religious ceremonies occur seasonally and are associated with periodic pilgrimages . In Belize , the Rio Frio caves in the Cayo District are a popular tourist attraction and workers are known to sleep in the caves to guard against vandalism . Although bat guano is harvested in many Latin American countries and the United States , our study did not include locations where this occurs . The parasite that causes Chagas disease , Trypanosoma cruzi , is transmitted by blood feeding triatomine insects [2] . These vectors are generally wide-ranging feeders and what they feed on is an important aspect of epidemiological risk; for example , human blood meals may mean increased risk of transmission to humans [3] . The species Triatoma dimidiata is currently the major vector of Chagas disease in Central America [4] . Some vector populations are found entirely in sylvan ecotopes , including caves; whereas others are nearly entirely domesticated and some move between ecotopes [5]–[9] . It is generally thought that sylvan vectors are not important for human transmission; however , few data are available to support this assumption . It is important to know if they are feeding on humans to help prevent transmission and to target control . Because Chagas disease is the most economically important parasitic disease in Latin America with 8–10 million persons infected [10] , we decided to investigate the potential role of cave dwelling T . dimidiata in transmission of T . cruzi to humans . To investigate the possibility that cave-dwelling , sylvatic , T . dimidiata contribute to the human transmission cycle , we determined the blood meal sources of T . dimidiata collected from caves in three regions of Central America where T . dimidiata is the main vector . We tested the hypothesis that T . dimidiata collected from caves would not contain human DNA , and thus are not likely to be involved in transmission of T . cruzi to humans . In addition , we discuss the relationship between human cultural interactions with the cave and the blood meal sources found at each location .
Adult males and females were collected from inside caves in the departments of Petén and Alta Verapaz in Guatemala and Cayo district in Belize ( Table 1 ) . All the vectors had characters typical of the cave dwelling morph of T . dimidiata and are easily distinguished from those found in nearby houses and peridomestic areas by their longer , larger heads and reduced eyes and ocelli , as well as darker color for the adults and nymphs that are lighter in color [5] , [11] . Wearing protective gear , five or six professionals from Laboratorio de Entomología Aplicada y Parasitología ( LENAP ) , Universidad de San Carlos de Guatemala ( USAC ) collected the vectors . These personnel are trained in the safe handling of infectious agents and were experienced in searching for vectors . Personnel searched the walls , ceiling ( when possible ) , and floor of the cave for 25 min , four times at each geographic location , with a 5–10 min . break between searches . Repeated searches are important because it takes vectors some time to emerge from hiding spots , presumably attracted by CO exhaled by the searchers . Triatomine nymphs in these caves are pale colored compared to nearby non-cave vectors , and have been observed to emerge from hiding in cracks and crevices of the cave walls and ceiling within a half hour following humans entering the cave . Adults are darker in color and easy to see in the caves because of the contrast with the limestone walls . Vectors were identified as T . dimidiata sensu lato based on published taxonomic keys [11] . The human cultural practices in the caves differ among locations . Near the municipality of San Luis , Petén , insect vectors were collected from a small cave located on a privately owned cattle farm . This particular cave was selected because the owner of the farm requested a survey to find the origin of vectors found in their home . Access to the cave is restricted and only local hunters and farm workers know of the cave; however , feral dogs from the nearby village may enter the cave in the rainy season . In Petén , hunting is a traditional weekend entertainment and caves are frequently used as overnight shelters . In this region , caves are not used for religious purposes and although hunters and their dogs may sleep and seek shelter in the caves , domesticated animals other than dogs are not brought to caves . There are no houses near the cave , it is surrounded by pasture areas for cattle . In Alta Verapaz , caves near Lanquín and Cahabón were examined . Grutas de Lanquín , a complex of caves known to be infested with T . dimidiata [12] was chosen because the main cave is a tourist attraction and is used for Mayan religious ceremonies . The Lanquín River runs through the center of the cave . Thousands of bats inhabiting the cave provide a popular daily tourist attraction as they leave en masse at twilight . Ceremonial stone altars within the cave are still used by local residents for Mayan religious ceremonies that are performed at night . Candles and offerings to the gods are found inside the cave . There are several cave entrances but municipality guards control access . Human houses are located nearby . Santa Maria cave , in the municipality of Cahabón , maintains the tradition of seasonal pilgrimages organized by descendants of the Maya . Travelers leave offerings to the gods in the cave , including sacrificial animals , pottery , candles , liquor , plant resins , and tamales . There are houses near the one main entrance to the cave , but these houses do not have domestic animals in captivity , and access to the cave is not controlled . The caves at Rio Frio , Belize are a popular tourist attraction with several entrances to a wide open sandy area with the river flowing through the cave being a popular place for bathing . The ceiling is too high to search in this location , thus the walls are the main source of the vectors collected . Guards often sleep in these caves to protect against vandalism but there are no houses near the cave . DNA was extracted from T . dimidiata abdomens as described previously [13] , except we used the E . Z . N . A . Genomic DNA Isolation kit ( Omega Bio-Tek , Norcross , GA ) . The T . cruzi infection status was determined for the vectors from Lanquin , Alta Verapaz using the TCZ primers following previously described methods [14] , prior to starting the cloning procedure . Unfortunately , all of the extracted DNA from Rio Frio and San Luis Peten was used in the cloning assay and therefore we were unable to assess the T . cruzi infection status for the vectors from these locations . However a previous study found 25% T . cruzi prevalence in Peten [15] . The assay to identify blood meal sources was developed and previously reported by us [16] . Because the extractions of DNA from the insect abdomens potentially contain a mixture of blood meal DNA from multiple vertebrates , as well as insect DNA and parasite DNA ( if infected ) , the initial PCR used primers specific for the mitochondrial 12S ribosomal RNA gene of vertebrates [17] . After confirming the PCR products were the expected size ( ∼215 bp ) , they were cloned using the pGEM-T Easy Vector System ( Promega , Madison , WI , USA ) . Using the same 12S primers , cloning was verified by PCR , followed by sequencing in one direction with BigDye v3 . 1 ( Applied Biosystems , Foster City , CA , USA ) and analysis using an ABI PRISM 3730xl ( Beckman Coulter , Fullerton , CA , USA ) . Sequence files were trimmed to 140 bp , edited using Sequencher v4 . 10 ( Gene Codes Corporation , Ann Arbor , MI , USA ) and taxonomic identification of the sequences was determined with a BLAST search using ≥99% identity as the criterion for a match . To determine if feeding on the different blood meal sources varied among the three locations or between sexes , we used one way analysis of variance ( ANOVA ) for each taxa ( humans , dogs , food animals , animals commensal with humans and wild animals ) . All statistical analyses were done using the JMP statistical package ( JMP Version 10 [SAS Institute Inc . , Cary , NC] ) .
Human was the most common blood meal source in all three locations ( Table 2 , Figure 1 ) , with over half the vectors from each cave feeding on human ( Table 2: Petén = 8 of 9 vectors , Alta Verapaz = 7 of 10 , and Cayo = 4 of 7 ) . In Alta Verapaz , pig was equally abundant to human ( both at 70% ) . Data on T . cruzi infection show that in Alta Verapaz , two of the three vectors that had human blood meals sources also had the parasite ( Table 3 ) . Overall we detected blood meals from eleven different taxa in the 24 vectors from the three locations , including duck and two species each of opossum and bat ( Table 2 ) . We examined a total of 218 sequences with 162 ( 74% ) providing information on a blood meal source . On average , 9 . 04 clones were sequenced per vector , 7 . 11 sequences per vector were interpretable as a blood meal source , and individual vectors contained 2–3 blood meal sources ( = 2 . 63 , range 1–5 ) . Using this small region of the 12S gene and the criteria of ≥99% match to a published vertebrate sequence , were able to determine the species for nine of the eleven taxa ( human , dog , pig , mouse , rat , duck and the two bat species ) . We could not distinguish between goat and/or sheep , these two Bovids in the Subfamily Caprinae have identical 12S sequences for the regions examined . Similarly , we were only able to determine the subfamily for the two different opossum sequences ( Didelphinae: Didelphis sp . or Philander opossum ) The non-human hosts fall into four groups that relate to human cultural interactions with caves: companion animal ( dog ) , food animals ( goat/sheep and pig ) , human commensal species ( mouse and rat ) and wild species ( mallard duck and the two species each of bat and opossum ) . The mouse and rat species , Mus musculus and Rattus norvegicus , are human commensals and likely occur at these locations because of human activities . Humans and commensal animals were blood meal sources at all three locations ( Figure 1 ) . Food animals were not a blood meal source in Petén , likewise wild animals were not a blood meal source in Alta Verapaz and there was no evidence of feeding on dogs in Cayo . Feeding on dog ( F = 3 . 57; d . f . 2 , 23; P<0 . 05 ) , food animals ( F = 5 . 56; d . f . 2 , 23; P<0 . 02 ) and sylvatic animals ( F = 4 . 68; d . f . 2 , 23; P<0 . 03 ) were significantly different among locations . Feeding on humans ( F = 0 . 64; d . f . 2 , 23; P>0 . 05 ) and commensal animals ( F = 0 . 23; d . f . 2 , 23; P>0 . 05 ) did not vary among locations . There were no significant differences between males and females in the detection of human ( F = 1 . 14; d . f . = 1 , 22; P>0 . 05 ) , dog ( F = 0 . 05; d . f . = 1 , 22; P>0 . 05 ) , food animal ( F = 0 . 22; d . f . = 1 , 22; P>0 . 05 ) , commensal animal ( F = 0 . 001; d . f . = 1 , 22; P>0 . 05 ) , or sylvatic animal ( F = 0 . 03; d . f . = 1 , 22; P>0 . 05 ) blood meal sources . The number of vectors feeding on bats was small , one each from Petén and Cayo and the bats were different genera . In contrast opossum are a more common blood meal source ( 43% , 3 of 7 vectors examined from both Cayo and Petén , including one vector from Cayo that had feed on both types of opossums ) . None of the vectors from Alta Verapaz had fed on opossum .
Our results indicate that cave dwelling , sylvatic T . dimidiata can contribute to the human transmission cycle . The evidence is that human was the most commonly found blood meal in T . dimidiata collected from caves . At the one location where we have T . cruzi infection data , we found 3 of 10 ( 30% ) of vectors infected , and 2 of these 3 ( 2 of the 10 or 20% of vectors from that location ) also had evidence of human blood . Human blood was found in >50% of all vectors in seven caves from three different locations . However , humans were not more common than the other blood meal sources combined . Of the total blood meals identified 27% at San Luis were from human ( 6 of 14 blood meal sources ( 43% ) , at Alta Verapaz 7 of 30 ( 23% ) , and at Rio Frio 4 of 19 ( 21% ) . Further evidence of association with humans included substantial numbers of blood sources from the companion animal dog , domesticated food animals , and animals commensal with humans . Evidence of frequent human-bug contact was observed in the three different types of caves each with different ecological conditions and uses by humans . These results suggest a risk of transmission of T . cruzi to humans entering caves , however further study measuring the rates of T . cruzi infection in insects from all locations would clarify the transmission risk . The high proportion of bugs that had evidence of human blood meals is surprising , given that caves are considered a sylvatic ecotope , and sylvatic vectors are generally thought not to be important in human transmission . Our results add weight to recent studies challenging this assumption . For example , in sylvan environments in Arizona and California , 38% of Triatoma rubida and Triatoma protracta had human blood sources [16] , in Triatoma infestans in Bolivia 27% of blood meals detected were human [18] . However , a study of Triatoma gerstaeckeri , Triatoma indictiva , Triatoma sanguisuga and T . protracta in Texas found little evidence of feeding on humans [19] . Ecological factors including human encroachment into natural areas have been associated with emerging diseases worldwide [20] . The bugs from Santa Isabel Cave in Péten , a privately owned cattle farm with remnant forest patches , showed the highest incidence of human blood consumption ( 43% , Figure 1 ) . We have anecdotal evidence from the owner of the farm that hunters and farm workers use the relatively cool cave for overnight sleeping or daytime resting , especially during the heat of the dry season . In addition , at the entrance of the cave we have found a clean wild animal skull , a common artifact at contemporary Maya hunting shrines [21] , and machetes and plastic water bottles have been found inside the cave ( M . C . Monroy , personal observation ) . Wild and human commensal animals were the next most preferred blood sources ( 29% and 21% , respectively ) . It is surprising that dog blood meals are rare ( only 7% ) , since dogs usually accompany the hunters , but perhaps the dogs are eating bugs that come close [22] . Vectors from caves on this cattle farm had fed on the common vampire bat but interestingly had not fed on cattle , suggesting they do not leave the caves . Humans and food animals are the main blood meal sources for T . dimidiata ( both 30% ) in the caves of Lanquín and Cahabón in Alta Verapaz , located within heavily deforested areas and surrounded by agricultural fields and houses ( Figure 1 ) . Frequent entry of human into the caves for tourism and religious ceremonies could explain the high proportion of blood meals from humans as these caves are popular destinations for tourists and Mayan religious ceremonies . The people performing the rituals may stay overnight in the caves; however , tourists are not likely to stay overnight in the caves , but rather camp near the caves ( M . C . Monroy , personal communication or personal observation ) . Chickens , pigs , goats , and other food animals are free ranging in this area , which could explain high food animal blood sources ( 30% ) in the bugs . The human companion animal dog ( 22% ) and human commensal animals ( 17% ) are the next most frequently found sources , and no wild animal blood sources were found in the bugs from these caves , demonstrating the strong association of humans and bugs in this area . Deforestation removes sylvan animal habitat and therefore sylvan blood sources . Anthropogenic change is known to alter blood meal sources [23] . Bugs found in Rio Frio , Cayo , Belize cave had fed mostly on human and human commensal animals ( 29% each , Figure 1 ) . The cave is relatively small , has wide openings at both ends and is surrounded by forest . The high association with humans can be explained as the cave is a popular tourist attraction , frequented by swimmers , with guards often sleeping in the cave at night to protect against vandalism . Food animals and wild animals are the next most frequent blood source ( 21% each ) which suggests it is a transit and resting place for wild and feral animals . For example , wild pigs have been reported in this area ( http://www . belizeanway . com/destination-cayo ) . The bat fed upon in Cayo was most likely hairy-legged myotis ( 100% match to Myotis keaysi ) but possibly elegant myotis ( 99% match to Myotis elegans ) [24] . There was no evidence of dog blood meals in vectors collected in Rio Frio . The only nearby settlement is a British military base , and since the Belize military supervises the entrance to the cave it may be more restricted to hunters and their dogs . Of the eleven taxa detected in the blood meals , about one-third were wild ( five species total , duck , two opossum and two bat ) and about one-third ( 7 of 24 ) of the vectors had fed on sylvatic animals . Due to land use practices , none of the vectors from Alta Verapaz had evidence of feeding on wild animals . The other locations showed similar wild animal blood sources: opossum was the most common ( 43% in both ) . In contrast , the number of vectors feeding on bats was small , one each from each locality ( 14% ) . Although the vectors included in this study were all collected within caves , it is not known which of the blood meals we identified were actually obtained in the caves as it is possible the vectors could leave the caves to feed . However , our results are consistent with these vectors spending most or all of their lives in the caves . In Petén , although there is abundant cattle surrounding the caves , cattle blood was not found in the bugs . In addition several reports establish that T . dimidiata in caves in Central America differs morphologically from those found outside the caves suggesting they are different populations that do not intermingle [5] , [25] . The bugs included in this study are the cave morphology . Therefore our results suggest the human-bug contact occurs when humans enter the cave . This may not be true everywhere as some T . dimidiata collected in houses in Colombia may be migrants from caves [5] , [26] and in some localities T . dimidiata blood sources show evidence of migration between ecotopes [27] . Although a large number of different taxa ( 11 ) were detected in the abdomens of the 24 T . dimidiata this is likely a minimal estimate of blood sources as older blood meals may have been too degraded for detection [28] . Caves at all three locations had evidence of Chagas disease vectors feeding on humans and human commensal animals; as well as some combination of dogs , domesticated food animals and wild animals . The variation of blood meal sources among caves is understandable in the context of human uses of the caves at the three locations . We caution that human encroachment into natural areas may expose individuals to unanticipated risks of disease transmission . | Caves have enduring appeal to humans , and their lure in Central America includes tourism , religious ceremonies and shelter . The major Chagas disease vector in this region , Triatoma dimidiata , inhabits caves throughout its range . We challenge the assumption that cave-dwelling vectors are not important for human transmission by determining blood meal sources of vectors collected in caves from three locations that differ in the activities of humans at the caves , and link the results to cultural practices that differ among locations . Seventy percent of cave-collected vectors were positive for human DNA , and fed on 10 additional vertebrates with relationships to humans varying from companion animal ( dog ) , food animals ( pig , sheep/goat ) , wild animals ( duck , bat , opossum ) to commensal animals ( mouse , rat ) . Feeding sources relate to human activities that vary among locations , for example , human and food animals were the main sources in Cahabón caves , located within deforested areas that are now agricultural fields and houses . We tested vectors from one location for infection with the Chagas disease parasite and found 30% ( 3 of 10 ) infected , including two of the three vectors from this location that had evidence of human blood . Humans should be aware of potential consequences of visiting and sleeping in caves . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"parasitic",
"protozoans",
"plant",
"science",
"protozoans",
"plant",
"pathology",
"population",
"biology",
"genetic",
"epidemiology",
"infectious",
"diseases",
"environmental",
"epidemiology",
"epidemiology",
"disease",
"vectors",
"behavioral",
"ecology",
"community",
"ecology",
"ecology",
"vector",
"biology",
"trophic",
"interactions",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"organisms",
"terrestrial",
"environments"
] | 2014 | Hunting, Swimming, and Worshiping: Human Cultural Practices Illuminate the Blood Meal Sources of Cave Dwelling Chagas Vectors (Triatoma dimidiata) in Guatemala and Belize |
Plants defend themselves against pathogens by activating an array of immune responses . Unfortunately , immunity programs may also cause unintended collateral damage to the plant itself . The quantitative disease resistance gene ACCELERATED CELL DEATH 6 ( ACD6 ) serves to balance growth and pathogen resistance in natural populations of Arabidopsis thaliana . An autoimmune allele , ACD6-Est , which strongly reduces growth under specific laboratory conditions , is found in over 10% of wild strains . There is , however , extensive variation in the strength of the autoimmune phenotype expressed by strains with an ACD6-Est allele , indicative of genetic modifiers . Quantitative genetic analysis suggests that ACD6 activity can be modulated in diverse ways , with different strains often carrying different large-effect modifiers . One modifier is SUPPRESSOR OF NPR1-1 , CONSTITUTIVE 1 ( SNC1 ) , located in a highly polymorphic cluster of nucleotide-binding domain and leucine-rich repeat ( NLR ) immune receptor genes , which are prototypes for qualitative disease resistance genes . Allelic variation at SNC1 correlates with ACD6-Est activity in multiple accessions , and a common structural variant affecting the NL linker sequence can explain differences in SNC1 activity . Taken together , we find that an NLR gene can mask the activity of an ACD6 autoimmune allele in natural A . thaliana populations , thereby linking different arms of the plant immune system .
Plants rely on a sophisticated immune system to defend themselves against pathogens . A central challenge for plants is how to achieve a fast , effective response upon pathogen attack , while at the same time preventing spontaneous firing of the signaling machinery in the absence of danger [1] . Inappropriate activation of immune signaling can reduce growth or even damage the plant’s own cells , while an inefficient immune response makes plants more likely to succumb to pathogen attack [2–6] . Although highly effective immune alleles may have background activity , they will nevertheless be favored when pathogen pressure is high . In contrast , in locations and years with low pathogen pressure , such alleles tend to be selected against [7] . If differently active alleles exist at the same locus , such temporal or spatial variation in pathogen pressure will maintain both types of alleles at ratios that reflect the prevalence of the different environments; this is one example of the phenomenon of balancing selection [8–10] . Very different alleles have been described for several disease resistance ( R ) genes of the nucleotide binding site leucine-rich repeat ( NLR ) class , although formal evidence for balancing selection is still rare [11–15] . NLR immune receptors detect the presence of so-called pathogen effector molecules , leading to effector-triggered immunity ( ETI ) . NLRs do so in several different ways , and the biochemical and structural basis for effector detection , including direct interaction with effectors or with effector-modified host proteins , is increasingly well understood [16 , 17] . Variation in NLR immune receptors is linked to the fact that effectors are rarely essential for pathogen survival , and that even closely related pathogens can greatly vary in effector content [16 , 18] . In mutant screens , gain-of-function alleles have been identified for several NLR genes , based on their autoimmune phenotypes . The mutant proteins are active regardless of effector presence , and thus can increase resistance to a range of unrelated pathogens [19–26] . Whether highly active autoimmune alleles of NLR genes are being deployed to enhance broad-spectrum pathogen resistance in nature is unknown , but the presence of NLR alleles with modest autoimmune activity in breeding programs is well documented , and phenotypic signs of autoimmunity have been exploited to follow resistance genes in segregating crop populations ( e . g . , [27] ) . In addition to specialized NLRs , which often engender very strong , qualitative disease resistance , plants employ various genes for quantitative resistance [28–31] . An example in the wild plant A . thaliana is ACCELERATED CELL DEATH 6 ( ACD6 ) [32–35] . Although ACD6 does not encode an NLR immune receptor , the locus features extensive copy number and sequence variation in wild populations , reminiscent of what is found at many NLR loci . Importantly , natural populations segregate for ACD6 allele classes with clear functional differences [36–38] . One class , first identified by genetic analysis of the natural Est-1 accession , protects in the greenhouse against a wide range of unrelated pathogens , from microbes to insects [36] . This unusually large benefit of the hyperactive ACD6 allele appears to be due to enhanced and partially constitutive defense responses . On the other hand , the autoimmunity seen in plants with the ACD6-Est allele substantially compromises growth under specific laboratory conditions , reducing both plant size and the tempo with which new leaves are being produced [36 , 39 , 40] . That the allele is found at a frequency of over 10% in natural populations is compatible with the idea that selection maintains this allele despite a fitness trade-off [36–38] . The frequency of functionally distinct ACD6 alleles differs between local populations , suggestive of a role of ACD6 in local adaptation [37 , 38] . Alternatively , since the expression of immune and growth traits often differ between laboratory and field conditions [41–43] , the effects the ACD6-Est allele in nature might be modified by the environment , and there is evidence from other laboratories that the expression of ACD6-Est associated lesions is environment-dependent [39 , 40 , 44 , 45] . ACD6 encodes a transmembrane protein with intracellular ankyrin repeats , which positively regulates cell death and defense , acting in part via the immune hormone salicylic acid ( SA ) and the SA transducer NONEXPRESSER OF PR GENES 1 ( NPR1 ) [32 , 33] . The precise mechanism of ACD6 biochemical action remains enigmatic , but ACD6 is found in a complex with several immune-related proteins , including pathogen-associated-molecular-pattern ( PAMP ) receptors , indicating a role for ACD6 in PAMP triggered immunity ( PTI ) [34 , 35] . While PTI and ETI have often been considered as different arms of the plant immune system , the distinction between ETI and PTI is becoming more and more blurred , not only because of substantial overlap in downstream responses , but also because not all effectors are highly variable , and not all PAMPs are highly conserved [46] . Here , we report that natural alleles at several independent loci can modulate the activity of the ACD6-Est autoimmune allele . We have characterized one locus , encoding the NLR protein SUPPRESSOR OF NPR1-1 , CONSTITUTIVE 1 ( SNC1 ) in detail and show that a common structural variant affecting the NL linker explains differences in ACD6-Est dependent SNC1 activity . We propose that allelic diversity at SNC1 and other loci contributes to the maintenance of the ACD6-Est autoimmune allele in natural A . thaliana populations .
Sanger sequencing analysis of a limited collection of 96 natural accessions had shown that 19 , or 20% , had the ACD6-Est allele , which triggers autoimmunity and reduced growth under specific laboratory conditions [36] . To extend our knowledge of the global distribution of the ACD6-Est allele , we attempted to use Illumina short read data of 1 , 135 accessions [47 , 48] , with the goal of ascertaining the presence of two codons that are causal for ACD6-Est autoimmune activity [36] . Likely because of linked sequence diversity and copy number variation , the first codon , encoding amino acid 566 in the reference sequence , GCA ( Ala ) in Col-0 and AAC ( Asn ) in Est-1 , could not be confidently typed with short reads . At the second codon , encoding amino acid 634 in the reference , 823 accessions could be assigned to have either Col-0- or Est-1-like codons . Of these , 721 , or 88% , had CTT ( Leu ) , diagnostic for the Col-0 reference , and 102 , or 12% , had TTT ( Phe ) , diagnostic for Est-1 , confirming that ACD6-Est alleles are not uncommon in the global A . thaliana population . Of the 19 accessions with an ACD6-Est allele examined by Todesco and colleagues [36] , two did not show any necrotic lesions , an obvious sign of autoimmunity . The remaining 17 accessions differed in their phenotypic severity as well , with three classified as expressing mild , five intermediate , and nine strong necrosis . We extended this analysis to a total of 54 accessions for which we had confirmed the presence of an ACD6-Est allele by Sanger sequencing ( S1 Table ) . Seven had no or only very mild lesions , 22 had intermediate lesions , and 25 were similarly affected as the Est-1 strain , in which this allele had been originally identified ( Fig 1A ) . While we could confirm that there is substantial variation in the extent of autoimmune phenotypes exhibited by ACD6-Est carriers , this observation did not inform about the genetic cause of this variation . To determine whether this was due to sequence differences at the ACD6 locus itself , we directly tested ACD6 activity in the first two ACD6-Est accessions identified as lacking necrotic lesions , Pro-0 and Rmx-A180 [36] . ACD6 was expressed well in both accessions , indicating that suppression of the ACD6-Est phenotype was not due to reduced RNA accumulation ( Fig 1B ) . We introduced full-length genomic ACD6 fragments from both accessions into the Col-0 reference strain , which carries a standard ACD6 allele , and into acd6-2 , a Col-0 derivative with a T-DNA insertion in ACD6 . All four classes of transformants from 10 T1 lines had small rosettes and necrotic lesions , similar to what has been reported for Est-1 [36] ( Fig 1C ) . Together , these results demonstrated that the ACD6 alleles from Pro-0 and Rmx-A180 have similar activity as the original ACD6-Est allele outside their native genetic backgrounds , suggesting that accessions such as Pro-0 and Rmx-A-180 carry extragenic suppressors of ACD6-Est activity . ACD6 acts in a feed-forward loop that regulates the accumulation of SA , a key hormone in both disease signaling and autoimmunity [49] . We therefore asked whether the suppressed autoimmunity of the ACD6-Est allele in Pro-0 was accompanied by reduced SA levels . Indeed , Pro-0 had much less SA than the ACD6-Est reference strain Est-1 or acd6-1 , which carries an EMS-induced hyperactive ACD6 allele in the Col-0 background [33] . SA levels in Pro-0 were even lower than in Col-0 , which carries the standard ACD6 allele ( Fig 2A ) . Similarly , the expression of SA responsive marker gene PR1 was reduced in Pro-0 compared to Est-1 ( Fig 2B ) . As expected , knocking down ACD6 with an artificial miRNA [36 , 50] resulted in greatly reduced SA levels in both acd6-1 and Est-1 , while the effect in Pro-0 was much smaller ( Fig 2A ) . In agreement , knocking-down ACD6 in Pro-0 did not substantially affect PR1 expression either ( Fig 2B ) . Finally , to assess not only autoimmunity but also true immune responses , we measured H2O2 accumulation after stimulation with the PAMP flg22 [51] . Est-1 responded strongly to flg22 , with the H2O2 production being greatly reduced by knocking down ACD6 . Pro-0 responded much more weakly , and knocking down ACD6 was of little consequence ( Fig 2C ) . In conclusion , these results confirmed attenuation of the effects of the hyperactive ACD6-Est allele in Pro-0 . To understand the population genetic architecture underlying extragenic suppression of the activity of ACD6-Est-type alleles , we crossed three accessions that carried such alleles , but did not show necrosis , to the necrotic Est-1 accession . The ratios of necrotic and non-necrotic F2 individuals suggested more than one locus in all mapping populations ( S2 Table ) . We used RAD-seq [52] followed by QTL mapping to identify causal regions of the genome in these accessions . As expected , we identified multiple significant QTL in all mapping populations ( Fig 3A ) . The QTLs explained 4% to 23% of the observed phenotypic variation , further supporting the quantitative nature of phenotypic suppression of ACD6-Est alleles ( S3 Table ) . In most , but not all , cases did the Est-1 allele increase necrosis ( S1 Fig ) . We decided to focus on the QTL in the center of chromosome 4 , near ACD6 , which explained 18% of phenotypic variance in the Pro-0 x Est-1 cross , because the confidence interval ( Chr 4: 7 . 5–9 . 7 Mb , LOD = 13 . 1 , p<0 . 0001 ) contained a well-studied locus that had been linked to suppression of autoimmunity . Laboratory-induced loss-of-function alleles of this locus , SNC1 , can fully or partially suppress the phenotypic defects of different autoimmune mutants [53 , 54] . In addition , SNC1 gain-of-function alleles reduce plant size , similar to ACD6 hyperactivity , while down-regulation of SNC1 homologs increases plant size [19 , 55] . SNC1 is located in a complex NLR gene cluster , RECOGNITION OF PERONOSPORA PARASITICA 4/5 ( RPP4/5 ) , which has alleles that mediate resistance to different strains of the oomycete pathogen Hyaloperonospora arabidopsidis [56 , 57] . SNC1 and different RPP4 and RPP5 alleles are closely related in sequence [19 , 55 , 56 , 58] . To test whether SNC1 contributes to the suppressed leaf necrosis in Pro-0 , and whether SNC1 can modulate ACD6 activity , we took a transgene approach . We first introduced a SNC1-Pro-0 genomic fragment into Pro-0 . This transgene caused autoimmunity , consistent with additional NLR copies often leading to spontaneous activation of immunity , but the degree of autoimmunity was very modest [59–61] . In contrast , when we introduced a chimeric transgene containing the SNC1 promoter from Pro-0 , but coding sequences from Est-1 , into Pro-0 , we observed strong necrosis and dwarfism , both in primary transformants and in the T2 generation ( Fig 3B and 3C , S2 Fig ) . In agreement , Pro-0 plants transformed with the SNC1-Est allele were more resistant to infection by P . syringae pv . tomato strain DC3000 ( Pto DC3000 ) than untransformed Pro-0 plants or Pro-0 plants carrying an extra SNC1-Pro copy ( p<0 . 0001 ) ( Fig 3D ) . This suggests that differences in the transcribed portion of SNC1 , most likely differences between the SNC1-Est and SNC1-Pro proteins , are causal for attenuated effects of ACD6-Est alleles . However , since we did not have Near Isogenic Lines ( NILs ) at our disposal , with which we could directly compare the phenotypic effects of the QTL region with the SNC1 transgenes , we cannot exclude that the QTL on chromosome 4 includes further genes that contribute to attenuated ACD6-Est activity in Pro-0 . To gauge whether attenuation of ACD6-Est activity by variant SNC1 alleles is likely to be common , we wanted to assess the population frequency of different SNC1 alleles . We compared the SNC1 alleles from Est-1 and Pro-0 with RPP4/RPP5/SNC1-like sequences from a REN-seq dataset of 65 A . thaliana accessions [62] . A protein-phylogeny of 142 RPP4/RPP5/SNC1 homologs indicated that 34 genes , from 29 accessions , form a single clade of SNC1 orthologs , which are predicted to encode highly similar Toll-Interleukin-1 receptor ( TIR ) domains . Five accessions , Ws-2 , Liri-1 , Marce-1 , DraIV 6–22 and MNF-Che-2 , have two SNC1 homologs each ( Fig 4A , S3 Fig , S5 Table ) . The most noticeable sequence differences among SNC1 proteins affect the NL linker between the nucleotide binding ( NB ) and leucine-rich repeat ( LRR ) domains [19] . The NL linker ( amino acids 544–671 in SNC1-Col-0 ) is duplicated in 15 of the 34 SNC1 proteins ( Fig 4A and 4B , S3 and S4 Figs ) . A SNC1 allele that differs functionally from the Col-0 reference allele has been identified in Ws-2 , SNW [59] . The SNC1 homolog that we found in addition to SNW in Ws-2 ( Ws-2-T078 ) has a duplicated linker and is highly similar to the Pro-0 variant ( Fig 4B , S5 Fig ) . Similarly , the MNF-Che-2 and DraIV 6–22 accessions have two SNC1 genes , one each encoding a single and a duplicated linker ( Fig 4B ) . Overall , our analysis revealed a complex history of SNC1 diversification within and between accessions . To address whether SNC1 linker variation is associated with suppression of ACD6 activity across accessions , we determined the ACD6 allele type in all 15 accessions with the duplicated SNC1 linker . Three accessions , Ty-1 , Ct-1 and MNF-Che-2 , had ACD6-Est alleles ( S6 Fig ) , but only MNF-Che-2 had symptoms of necrosis . This is in agreement with our hypothesis that SNC1-Pro allele can function as genetic suppressors of ACD6 hyperactivity ( Fig 4C ) . That ACD6-Est induced necrosis was not suppressed in MNF-Che-2 may be due to the additional SNC1 homolog without a duplicated linker in this accession ( Fig 4C ) . A gain-of-function allele of SNC1 induced by EMS mutagenesis , snc1-1 , has a single amino acid substitution in the NL linker domain , pointing to the functional importance of the NL linker [19] . Since the NL linkers of SNC1 from Pro-0 and Est-1 differ notably , we suspected that they might be responsible for differential ACD6-Est effects in different backgrounds . To test this hypothesis , we introduced three different SNC1 constructs into Col-0: a genomic SNC1-Pro fragment , a genomic SNC1-Est fragment , and a chimeric SNC1-Pro fragment with the dNL linker of Pro-0 replaced with the sNL linker from Est-1 . Sixteen of 53 SNC1-Est T1 transgenic plants had a severe autoimmune phenotype ( Fig 5A ) . However , none of 70 SNC1-Pro T1 transgenic plants showed severe , and only 10% showed mild phenotypes . The ability of SNC1-Pro in triggering leaf necrosis was fully restored when introducing the SNC1-Est linker in an otherwise Pro-0 protein ( SNC1-Pro-NL-Est ) ( Fig 5A ) . To further assess the contribution of the NL linker on SNC1-induced cell-death signaling , we transiently expressed both SNC1-Est and SNC1-Pro in N . benthamiana and monitored chlorosis indicative of cell death 7 days after infiltration . Expression of SNC1-Pro resulted in a much weaker cell death and significantly less ion leakage than either SNC1-Est or SNC1-Pro-NL-Est ( Fig 5B , S7 Fig ) . Taken together , we conclude that polymorphisms in the NL linker region explain functional differences between SNC1-Est and SNC1-Pro .
The naturally hyperactive ACD6-Est allele in A . thaliana can induce an autoimmune syndrome associated with necrosis and reduced growth in many , but not all , accessions with this allele [36] . We have shown that variation in the effects of ACD6-Est alleles does not map to the ACD6 locus itself , but rather is caused by extragenic modifiers of the ACD6-Est phenotype . Notably , alleles at several different loci can have such an effect , including a common allele at the NLR locus SNC1 . Genetic mapping experiments with three different accessions carrying an ACD6-Est allele identified a minimum of six genomic regions that can modulate ACD6-Est effects . The identified QTL only explain a minority of observed phenotypic variation , suggesting that several additional minor-effect loci contribute to modulation of ACD6 activity . An attractive hypothesis is that suppressors of ACD6-Est-like activity have arisen as a consequence of excessive deleterious effects of ACD6-Est and perhaps hyperimmunity alleles at other loci . Alternatively , the ACD6-Est allele itself may first have evolved on a background that masked its effects . We found that about a quarter of A . thaliana accessions share the SNC1 allele with a duplicated NL linker , which can attenuate ACD6-Est effects . The fraction of accessions with an ACD6-Est allele among those with this SNC1 allele is similar to the frequency of ACD6-Est alleles in the global A . thaliana population . While there is no obvious indication for linkage between specific SNC1 and ACD6 alleles , the numbers examined so far are very small , and the situation could be different in local A . thaliana populations . Another explanation for why we did not see an enrichment of specific SNC1 alleles among ACD6-Est carriers might be that other modifiers are more important in natural populations . It has been argued that tandem duplication of NLR genes within clusters can provide an effective mechanism to acquire novel NLR specificities without sacrificing old ones [63] . SNC1 features both relevant structural and copy number variation , and functionally distinct variants can occur in the same genome , offering an opportunity to test such hypotheses about NLR evolution . Since SNC1 has a dosage effect on plant innate immunity [59] , it will be interesting to ask whether there are situations in which it is advantageous to have two functionally different SNC1 copies , as we have seen in MNF-Che-2 , Ws-2 and DraIV 6–22 . Mutations in other NLRs can either trigger autoimmunity [19 , 21–26 , 64–69] or suppress autoimmunity in various mutants with a range of biochemical defects [67 , 70–77] . As for SNC1 itself , the lethality of bon1 ( bonzai1 ) bon3 double mutants is caused by inappropriate activation of at least five NLRs , including SNC1 [77] . Some of the other QTL regions that modify ACD6-Est effects include NLR genes as well ( S8 Fig ) , and it will be interesting to learn whether the situation for ACD6-Est is similar as for bon1 bon3 mutants , and whether ACD6-Est activity can be attenuated by alleles at NLR loci other than SNC1 . In summary , our genetic analyses of accessions with a hyperactive ACD6 allele have shown that natural A . thaliana populations frequently harbor extragenic modulators of ACD6 activity , underscoring the importance of epistatic interactions for natural phenotypes [29 , 78–83] . In addition , our work has revealed a new connection between ACD6 , which has before been primarily linked to PAMP recognition and PTI [34 , 35 , 84] , and NLR proteins , which are central to ETI . There is already genetic evidence for cell death triggered by PAMP receptor complexes being suppressed by the NLR SNC1 [59 , 85] . An arbitrary selection of evidence for the distinction between PTI and ETI being fluid [46] includes the overlap in the signaling network downstream of AvrRpt2-triggered ETI and the network activated by the well-known PAMP flg22 [86] , or MAPK3 and MAPK6 as widely shared downstream signaling components for both PTI and ETI [86–88] . Together , our findings illustrate the potential of studies of naturally occurring autoimmunity in either A . thaliana inbred strains or hybrids [89] to contribute to our understanding of the immune system in healthy and diseased plants .
Seeds of Arabidopsis thaliana accessions ( S1 and S5 Tables ) were from stocks available in the lab . Seeds were germinated and cultivated in growth rooms at a constant temperature of 23°C ( temperature variability about ± 0 . 1°C ) , air humidity of 65% and long days ( LD , 16 hr day length ) or short days ( SD , 8 hr day length ) , with light ( 125 to 175 μmol m-2 s-1 ) provided by a 1:1 mixture of Cool White and Gro-Lux Wide Spectrum fluorescent lights ( Luxline plus F36W/840 , Sylvania , Erlangen , Germany ) . Leaf necrosis was scored in SD at 23°C . RNA was extracted from three biological replicates ( 4- to 5-week-old entire rosettes ) using TRIzol Reagent ( Thermo Scientific , Waltham , MA ) and treated with DNase I ( Thermo Scientific , Waltham , MA , USA ) . 2 μg total RNA was used as a template for reverse transcription ( M-MLV reverse transcriptase kit; Thermo Scientific , Waltham , MA , USA ) . Quantitative real-time PCR reactions were performed using Maxima SYBR Green Master Mix ( Thermo Scientific , Waltham , MA , USA ) according to the manufacturer’s directions on a CFX384 real time instrument ( Bio-Rad , Hercules , CA ) . Transcript abundance was normalized to the TUBULIN BETA CHAIN 2 ( At5g62690 ) transcript . Relative expression compared to the control ( usually Col-0 ) was quantified based on [90] . Primers used for qRT-PCR are listed in S7 Table . Leaf discs ( 5 mm diameter ) were punched from 5-week-old 23°C SD-grown plants and immediately floated with the adaxial side up in individual wells of a 96-well plate ( Greiner Bio-One , Frickenhausen , Germany ) containing 200 μL of water . They were incubated overnight ( 12–16 hr ) covered with a transparent lid , to recover . The water was replaced with the elicitation solution containing 10 μM Luminol ( Sigma-Aldrich , MO ) , 10 μg mL-1 horseradish peroxidase and 100 nM flg22 ( QRLSTGSRINSAKDDAAGLQIA , >85% purity; EZBiolab , Westfield , IN [91] ) . H2O2 production was monitored by luminescence immediately after the elicitation solution was added , for at least 90 minutes ( Tecan Infinite PRO multimode reader , Tecan , Männedorf , Switzerland ) . Controls were mock treated with the same solution without flg22 . Each genotype was assayed at least three independent times , with leaves coming from 4 biological replicates each time . Each sample consisted of 10 ( ±1 ) mg freeze-dried , ground plant material of 4-week-old rosettes , with 8 biological replicates per genotype , of plants that had been grown in randomized complete block design at 23°C SD . SA was extracted twice with 400 μl 20% methanol ( LCMS-grade ) / 0 . 1% hydrofluoroalkane by 5 min ultrasonic extraction , followed by 20 min incubation on ice , and removal of solids by centrifugation for 10 min at 13 , 500 g . 320 μl supernatant were removed after each extraction step and combined in a new vial . A third extraction step with 400 μl of 100% methanol ( conditions see above ) was performed and the supernatant was combined with the previous ones . The total volume was split in half before drying in a speed vac . For analysis of the conjugated and free salicylic acid the pellets were redissolved in 30 μl 50% methanol / 0 , 1% formic acid . Ultra Performance Liquid Chromatography Mass Spectrometry ( UPLC-MS ) analysis was performed on a Waters Acquity UPLC system coupled to a SYNAPT G2 QTOF mass spectrometer equipped with an ESI-Source ( Waters Corporation , Milford , MA ) at the University of Tübingen—ZMBP Analytics Laboratory . MassLynx v4 . 1 was used to control the LCMS system and TargetLynx ( Waters Corporation ) to perform data integration . RAD-seq library preparation using a set of 192 adaptor sequences containing PstI and Mse1 recognition sequences were prepared according to [52 , 92] . Ninety six or 192 individual samples were pooled per library and sequenced on one Illumina HiSeq 2000 lane ( 100 bp single-end reads ) . Raw reads were pre-processed with the SHORE ( https://sourceforge . net/projects/shore ) pipeline . Default parameters were used for de-multiplexing , read trimming and mapping . The BWA [93] option , also with default parameters , was used to map the reads to the TAIR9 Col-0 reference genome . The consensus sequences were computed , which were used to create a matrix containing the genotypes at specific marker positions for all F2 individuals ( S4 Table ) . We only considered bi-allelic SNPs that agreed with SNPs from both parents and the F1 hybrid of a given cross . Scripts are available at https://github . com/mzaidem/ACD6_suppressors_mapping_RADseq_scripts/tree/master . Pro-0 x Est-1 , Rmx-A180 x Est-1 and Bs-5 x Est-1 F2 individuals were phenotyped for late-onset necrosis in a binary manner . QTL mapping and testing for QTL effects and interactions were performed using the R/qtl package [94] . Lod scores were calculated under a single-QTL model using the function ‘‘scanone” . Lod score significance thresholds were established using 1 , 000 permutations . Bayesian credible intervals were estimated for individual QTL . The standard expectation-maximization algorithm was always used as ‘‘method . ” Using the markers closest to the two highest LOD scores , designated as Q1 and Q2 , we conducted direct linear modeling . The “Full” model has the equation: y ~ Q1 + Q2 + Q1:Q2 , where the additive effect as well as the interaction effect of the selected QTL’s are used to calculate the amount of necrotic phenotype variation explained by the model . The “Additive” model ( y ~ Q1 + Q2 ) , calculates how much of the necrotic phenotype variation observed is only due to the additive effect of Q1 and Q2 . While the “Interaction” model ( y ~ Q1:Q2 ) , accounts for how much of the necrotic phenotype variation observed is due to the interaction of Q1 and Q2 . ACD6 fragments were amplified from genomic DNA with PCR primers designed based on ACD6 from Est-1 ( [36] , S7 Table ) . The Pro-0 ( Rmx-A160 ) genomic DNA fragment was 9 . 1 ( 10 . 3 ) kb long , including 2 . 5 ( 1 . 8 ) kb sequences upstream of the initiation codon and 2 . 8 ( 3 . 2 ) kb downstream of the stop codon . Restriction enzyme sites of Gibson Assembly ( NEB , Ipswich , the USA ) were used to generate chimeric constructs [95] . An amiRNA against ACD6 ( TTAATGGTGACTAAAGGCCGT ) [36] was used to knock down ACD6 . All constructs were individually cloned into the Gateway entry vector pCR8/GW/TOPO ( Invitrogen ) and moved into binary vector pFK206 by LR reaction ( Thermo Scientific , Waltham , MA , USA ) . Constructs in pFK206 were transferred to Agrobacterium tumefaciens strain ASE and plants were transformed by floral dipping [96] ( S6 Table ) . SNC1 ( AT4G16890 ) , RPP4 ( AT4G16860 ) , and RPP5 ( AT4G16950 ) like protein sequences from 63 accessions were extracted from an in-house NLR database of Araport II . These sequences and the single RPP4/5 homolog of Arabidopsis lyrata were used as template to query the database with exonerate v2 . 2 . 0 allowing a minimal mapping score of 95% [97] . We used MEGA6 [98] to realign these sequences with SNC1-Est-1 , SNC1-Pro-0 , SNC1-Ws-2 ( SNW , GenBank accession AY510018 ) , RPP5-Ler-0 ( AF180942 ) and the RPP4/RPP5/SNC1 homolog At4g16900 from Col-0 ( Araport II ) . One-hundred thirty-six RPP4/5 protein sequences were used to infer phylogenetic relationships with both Neighbor-Joining ( NJ ) and Maximum likelihood ( ML ) approaches in MEGA6 , either over their entire lengths or only the TIR domain ( positions 1–226 in SNC1-Col-0 ) . Node confidence was estimated with 1 , 000 bootstrap replications . All sequences that we could confidently assign to the SNC1 clade were aligned separately to identify functional polymorphisms . A similar approach was used to generate a phylogenetic tree of ACD6 protein sequences extracted from the in-house NLR database , using ACD6 ( AT4G14400 ) and the adjacent ACD6 homolog AT4G14390 from Araport II as references . Pseudomonas syringae pv tomato strain DC3000 was grown to OD600 around 3 . 0 , collected and resuspended in 10 mM MgCl2 at 5 × 105 colony-forming units ( cfu ) /ml . The bacterial suspension was infiltrated into 4-week-old leaves with a needle-less syringe . Bacterial growth was determined 3 days post inoculation ( dpi ) . Agrobacterium tumefaciens with binary plasmids was grown to OD600 of 1 . 6 , and incubated in induction medium ( 10 mM MES pH 5 . 6 , 10 mM MgCl2 and 150 μM acetosyringone ) for three hours . Cell suspensions were normalized to OD600 of 0 . 8 for co-infiltration into the abaxial side of leaves of N . benthamiana leaves that had been grown under SD at 23°C . Images were taken 7 days after infiltration . Conductivity assays to measure ion leakage were performed according to [58] . Six independent leaves were infiltrated with A . tumefaciens with binary constructs encompassing different SNC1 alleles and two discs ( 0 . 6 mm diameter ) were harvested per leaf 5 days after infiltration . The fresh leaves were briefly rinsed with ddH2O and placed in a tube with 8 ml ddH2O . Samples were gently shaken for 24 hr before measuring ion content ( C1 ) using an Orion Conductivity ( Thermo Scientific , Waltham , MA , USA ) instrument . Total ion content ( C2 ) was determined from leaf samples that had been boiled for 15 min . C1/C2 ratios were calculated as indicators of ion leakage . Nicotiana benthamiana leaves expressing GFP only were used as control . Statistic differences were calculated by one-way-ANOVA . | Plants defend themselves against pathogens by activating immune responses . Unfortunately , these can cause unintended collateral damage to the plant itself . Nevertheless , some wild plants have genetic variants that confer a low threshold for the activation of immunity . While these enable a plant to respond particularly quickly to pathogen attack , such variants might be potentially dangerous . We are investigating one such variant of the immune gene ACCELERATED CELL DEATH 6 ( ACD6 ) in the plant Arabidopsis thaliana . We discovered that there are variants at other genetic loci that can mask the effects of an overly active ACD6 gene . One of these genes , SUPPRESSOR OF NPR1-1 , CONSTITUTIVE 1 ( SNC1 ) , codes for a known immune receptor . The SNC1 variant that attenuates ACD6 activity is rather common in A . thaliana populations , suggesting that new combinations of the hyperactive ACD6 variant and this antagonistic SNC1 variant will often arise by natural crosses . Similarly , because the two genes are unlinked , outcrossing will often lead to the hyperactive ACD6 variants being unmasked again . We propose that allelic diversity at SNC1 contributes to the maintenance of the hyperactive ACD6 variant in natural A . thaliana populations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"plant",
"anatomy",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"brassica",
"alleles",
"plant",
"science",
"model",
"organisms",
"signs",
"and",
"symptoms",
"experimental",
"organism",
"systems",
"plants",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"bioinformatics",
"biological",
"databases",
"leaves",
"genetic",
"loci",
"necrosis",
"eukaryota",
"plant",
"and",
"algal",
"models",
"diagnostic",
"medicine",
"sequence",
"databases",
"phenotypes",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"autoimmunity",
"organisms"
] | 2018 | Modulation of ACD6 dependent hyperimmunity by natural alleles of an Arabidopsis thaliana NLR resistance gene |
The HLA ( Human Leukocyte Antigens ) genes are well-documented targets of balancing selection , and variation at these loci is associated with many disease phenotypes . Variation in expression levels also influences disease susceptibility and resistance , but little information exists about the regulation and population-level patterns of expression . This results from the difficulty in mapping short reads originated from these highly polymorphic loci , and in accounting for the existence of several paralogues . We developed a computational pipeline to accurately estimate expression for HLA genes based on RNA-seq , improving both locus-level and allele-level estimates . First , reads are aligned to all known HLA sequences in order to infer HLA genotypes , then quantification of expression is carried out using a personalized index . We use simulations to show that expression estimates obtained in this way are not biased due to divergence from the reference genome . We applied our pipeline to the GEUVADIS dataset , and compared the quantifications to those obtained with reference transcriptome . Although the personalized pipeline recovers more reads , we found that using the reference transcriptome produces estimates similar to the personalized pipeline ( r ≥ 0 . 87 ) with the exception of HLA-DQA1 . We describe the impact of the HLA-personalized approach on downstream analyses for nine classical HLA loci ( HLA-A , HLA-C , HLA-B , HLA-DRA , HLA-DRB1 , HLA-DQA1 , HLA-DQB1 , HLA-DPA1 , HLA-DPB1 ) . Although the influence of the HLA-personalized approach is modest for eQTL mapping , the p-values and the causality of the eQTLs obtained are better than when the reference transcriptome is used . We investigate how the eQTLs we identified explain variation in expression among lineages of HLA alleles . Finally , we discuss possible causes underlying differences between expression estimates obtained using RNA-seq , antibody-based approaches and qPCR .
The HLA region is the most polymorphic in the genome , and also shows the greatest number of disease associations , which has made it very well characterized at the genomic , population and functional levels [1 , 2] . Decades of research have also shown that the HLA genes are targets of natural selection , likely a consequence of their role in responding to pathogens [2 , 3] . This combination of evolutionary and biomedical interest has resulted in an extensive catalogue of HLA variation in human populations , with the frequency of HLA alleles defined for various populations [4–7] . The most intensely studied genes in the region are the classical HLA loci . These include Class I genes ( HLA-A , HLA-C , and HLA-B ) , whose proteins present endogenous antigens to CD8+ T cells , and also regulate innate immune responses by interacting with killer cell immunoglobulin-like receptors ( KIR ) expressed on natural killer ( NK ) cells , and Class II genes ( HLA-DRA , HLA-DRB1 , HLA-DQA1 , HLA-DQB1 , HLA-DPA1 , and HLA-DPB1 ) whose proteins present exogenous antigens to CD4+ T cells [1 , 8] . Variation within or near HLA genes has been convincingly linked to resistance and susceptibility to both autoimmune and infectious diseases [2 , 9 , 10] . Although in many cases the mechanistic basis of the associations remains poorly understood , there has been an effort to identify whether associations can be linked to features such as variation at the level of specific amino-acids , HLA alleles , HLA haplotypes , non-coding variants near HLA genes , or HLA expression levels ( reviewed in [2 , 10 , 11] ) . In some instances , the association of one feature results from the fact that it tags another feature . Several studies have shown that HLA alleles may have a protective effect because they mark overall gene expression [12–15] . Other studies showed that a combination of two features may be important . For example , mortality of transplant recipients was associated with increased expression of HLA-C alleles which harbor specific amino-acids [16] . An understanding of how HLA expression varies among individuals , and the identification of genetic variants involved in the regulation of expression , will play a central role in understanding the contribution of HLA genes to normal and disease phenotypes . However , until recently , little information existed about the regulatory variation and population-level expression patterns of HLA genes , a result of the difficulty in quantifying expression for genes which show an unusually high polymorphism and are members of a multi-gene family [9 , 17] . Efforts have been made to develop antibody-based methods to quantify HLA protein on the cell surface [12–14 , 18–20] , or hybridization-based approaches to quantify mRNA , such as qPCR [13 , 21 , 22] and microarray methods [23] . However , the design of PCR primers , microarray probes , or antibodies that span the diversity of possible variants represents a technically challenging and labor-intensive undertaking . In addition , qPCR technologies are not appropriate for comparison of expression levels among different loci , an important concern when seeking to understand how expression of HLA genes responds to environmental challenges . The results obtained to date using qPCR [13 , 21 , 22] , antibody-based approaches [12–14 , 18–20] and customized arrays [23] have contributed to the understanding of HLA expression and its underlying genetic regulation . However , they do not take advantage of the large amount of RNA-seq data generated by studies of whole transcriptomes in large samples [24–26] , often involving populations from various regions of the world , which represents an attractive resource to investigate HLA expression . Although such whole-transcriptome RNA-seq studies do provide expression estimates for HLA genes , they bring new challenges . RNA-seq pipelines may provide biased expression estimates for two reasons: 1 ) many short reads originating from genes with extreme polymorphism fail to map to the reference genome , due to high degree of variation ( which results in a large number of mismatches between the reference genome and that of most individuals being analyzed ) , and 2 ) the presence of paralogues makes it difficult to map a read uniquely to a specific gene , leading to the exclusion of many reads . This raises concerns about the reliability of RNA-seq approaches to quantify HLA expression , given that these loci represent both the extreme of polymorphism in the human genome and are part of a multi-gene family [27–29] . A strategy to overcome these challenges is the mapping of reads to an HLA-personalized reference ( an index containing sequences of the individual being quantified ) , rather than to a single reference genome . For example , seq2HLA is a tool developed by Boegel et al . [30] to provide in-silico HLA types and expression estimates , and later applied to demonstrate that different tumor types are associated with different HLA expression levels [31] , and also to provide a large catalog of HLA expression in 56 human tissues and cell types [32] . AltHapAlignR [33] is another software which infers the HLA references which are the closest to the individual’s HLA haplotypes , and maps reads to them . The authors reanalyzed a large RNA-seq dataset [24] , and provided comparisons with conventional read mapping , showing an improvement in accuracy with the HLA-tailored pipeline . In this article we present HLApers , a personalized pipeline which we have developed to reliably quantify HLA expression from RNA-seq data . We compare HLApers to conventional pipelines , and discuss for the first time the impact of accurate estimation of HLA expression on downstream analyses such as eQTL ( expression Quantitative Trait Loci ) mapping and allele-specific expression . We show that it is possible to adapt different computationally efficient methods to work under the personalized reference framework , providing reliable quantification of HLA expression from RNA-seq data . We find that implementations with either a conventional read mapper [34] or a pseudoaligner [35] show similar expression estimates . We use simulations to assess accuracy , showing that the HLA-personalized pipeline is more accurate than conventional mapping , and apply the tool to reanalyze RNA-seq data from the GEUVADIS Consortium [24] , which made available whole-transcriptome RNA-seq data for Lymphoblastoid Cell Lines ( LCLs ) from 462 European and African individuals . We evaluate the impact of more accurate expression estimates obtained with HLApers on downstream analyses by carrying out a detailed survey for allele-specific expression and eQTL mapping at the classical HLA loci . Surprisingly , we find that conventional RNA-seq pipelines provide gene-level expression estimates and identify eQTLs which are highly correlated with those obtained under the HLA-personalized approach . However , we identify gains of using a pipeline tailored for HLA expression: higher accuracy , expression estimates at the HLA allele-level , and eQTLs with higher probabilities of being causal .
We developed the HLApers pipeline ( for HLA expression with personalized genotypes ) to measure HLA expression from whole-transcriptome RNA-seq data . The pipeline can use either ( 1 ) a suffix array-based read mapper ( STAR [34] ) followed by quantification with Salmon [36] ( henceforth called STAR-Salmon ) , or ( 2 ) a pseudoaligner with built-in quantification protocol ( kallisto [35] ) . The key feature of our implementation is the use of an index supplemented with a set of sequences covering the breadth of known HLA sequences ( see Materials and methods: Index supplemented with the HLA diversity ) . We implemented a two-step quantification approach , where ( 1 ) we align reads to reference sequences corresponding to all known HLA alleles , and identify those which maximize the read counts at each locus to infer the genotype which is present ( in-silico genotyping ) , and ( 2 ) we use this inferred HLA genotype to create a personalized index which we use to quantify expression ( Fig 1 ) . A problem for the analysis of HLA is that of multimaps: reads originating from a particular allele which map to other alleles of the same locus , or to other loci . To deal with multimaps different approaches have been used . One possibility is to discard them ( as in [33] ) , and another is to split them evenly among the compatible references ( as in [30] ) . In HLApers , we use maximum likelihood estimates of expression obtained by an expectation-maximization ( EM ) algorithm ( implemented within Salmon [36] and kallisto [35] ) , which infers the quantities of each HLA allele that maximize the probability of observing the set of sequenced reads . Because the in-silico typing is an important step for accurate expression estimates , we assessed the concordance between our RNA-seq based HLA typing and the HLA allele calls experimentally determined for 5 HLA loci using Sanger sequencing [6] . The concordance was higher than 97% for all of the HLA genes compared ( S1 Table ) . This is consistent with previous results showing that RNA-seq provides reliable HLA alleles calls [30 , 37–40] . Our HLApers pipeline integrates widely used quantification tools such as Salmon and kallisto , and provides both allele and locus level estimates of HLA expression . The pipeline is available at https://github . com/genevol-usp/HLApers . We next investigated how using the HLA-personalized index affects the quantifications of HLA expression . To this end we simulated an RNA-seq experiment using the Polyester package [41] , with a dataset of 50 individuals , with the read length and counts matching those of the observed data ( see Materials and methods: Simulation ) . We analyzed the simulated dataset with three different methodologies: ( 1 ) the two-step approach in HLApers , where we first inferred the personalized HLA genotype and then aligned reads to it; ( 2 ) alignment to the reference transcriptome ( Gencode release 25; primary assembly ) ; and ( 3 ) alignment to the reference genome ( GRCh38 ) . In approaches ( 1 ) and ( 2 ) the alignment is followed by expression quantification using Maximum Likelihood ( ML ) , which provides a statistical framework for dealing with multimap reads , whereas in approach ( 3 ) quantification is performed using only uniquely mapped reads . For each HLA locus and methodology , we assessed the proportion of simulated reads which successfully aligned . Because previous studies for genome sequencing identified a correlation between mapping success and the number of mismatches between the HLA allele an individual carries and the reference genome [28] , we analyzed how alignment success behaves as a function of the number of mismatches between each HLA allele and the reference genome ( Fig 2 ) . The use of an HLA-personalized index results in the largest proportion of successfully aligned reads , no matter how different the allele carried by the individual is from the allele in the reference genome . This is expected , since the personalized HLA component guarantees that a sequence close or identical to that originating the read will be present . When the alignment was performed using the reference transcriptome , there was a marked reduction in the proportion of successfully aligned reads for HLA-DRB1 , HLA-DQA1 , HLA-DQB1 , driven by decreased alignment success for alleles with a greater proportion of mismatches with respect to the reference genome . When using uniquely mapped reads there was a massive read loss for HLA-A , HLA-B and HLA-DPB1 , regardless of the divergence to the reference genome , as well as a lower proportion of successfully aligned reads across other loci . This shows that both discarding multipmaps , as well as not including a personalized index , have a negative impact on mapping success . Finally , for the least polymorphic HLA loci , mapping should not be sensitive to the specific reference used . This is precisely what we find , with all pipelines performing similarly for HLA-DRA and HLA-DPA1 . Having demonstrated that including an individual’s HLA alleles in the index improves the success of read alignment in the simulated data ( Fig 2 ) , we set out to address two questions with real data by applying HLApers to the GEUVADIS dataset [24] . First , we examined how expression varies among HLA loci , when the personalized index is used ( Fig 3 ) . Secondly , we compared expression estimates with and without the use of the personalized index , so as to evaluate the impact of its usage on real data ( Fig 4 ) . By summing the estimates for the 2 alleles at each HLA locus , we obtain gene-level expression estimates ( Fig 3 ) . We observe that HLA-B is the highest expressed gene overall . Among the Class I genes , HLA-B is followed by HLA-A with similar levels , and by HLA-C which has about 50% of the expression levels of HLA-B . For Class II genes , HLA-DRA is the most highly expressed . Although we observe a general concordance with the original GEUVADIS quantifications [24] , there are some notable differences: in the original quantifications , HLA-B is twice as expressed as HLA-A , and HLA-DPA1 is more expressed than HLA-DRB1 ( S1 Fig ) . We found that the correlation between results using the reference transcriptome or the personalized index was greater than 0 . 87 for every locus except for HLA-DQA1 ( Fig 4 ) . The loci with the lowest correlations between indices ( HLA-DRB1 , HLA-DQA1 and HLA-DQB1 ) , are also those with the greatest read loss when divergence from the reference allele is high in the simulation ( Fig 2 ) . We then investigated if the quantification tool used in HLApers ( STAR-Salmon or kallisto ) influences expression estimates . Correlations between STAR-Salmon and kallisto approaches were on average r > 0 . 99 for read counts , dropping to r = 0 . 8 for TPM estimates for Class I genes , likely due to different bias correction models ( S2 Fig ) . These results show that the key features influencing alignment success are the use of a personalized index and the statistical treatment of multimaps ( as opposed to discarding them ) , with the specific alignment tool being less influential . Finally , we evaluated the reproducibility of the HLA expression estimates obtained with HLApers . To this end , we analyzed replicates for 97 European individuals in GEUVADIS . We observed an average correlation of expression estimates between replicates of 0 . 92 over the 9 classical HLA loci . This shows that HLA quantifications from RNA-seq with HLApers are highly reproducible ( S3 Fig ) . The key role played by HLA loci in the immune response and their strong and abundant associations with infectious and autoimmune diseases have motivated studies to uncover their regulatory architecture [15 , 20 , 21 , 23 , 42–45] . Here we use accurate expression estimates obtained with HLApers , together with genotype data from the 1000 Genomes Project [46] , to identify SNPs which are associated with variation in expression levels ( eQTLs ) . Because multiple SNPs can affect expression , it is interesting to identify independent contributions made by distinct SNPs . This has previously been done by using a best eQTL ( i . e . , the one with the most extreme p-value ) as a covariate in subsequent searches for an additional variant . Here we use a related approach implemented in QTLtools , a collection of tools to perform eQTL analysis [47] . The module QTLtools cis allows for the identification of groups ( or “ranks” ) of SNPs associated with independent signals . For each rank , we identified the site with the most extreme association ( Fig 5 ) . Transcriptome studies can quantify expression for various biological features: individual SNPs , exons , isoforms , genes . In the case of HLA loci , a natural unit of interest is the HLA allele . HLApers provides expression estimates for individual alleles , since it is the allelic sequences which are included in the index . The immunogenetics literature has shown that many HLA alleles are associated with specific phenotypes of evolutionary and medical importance ( reviewed in [2 , 10 , 11] ) . Gauging information about the expression levels of alleles can therefore provide an additional layer of information . In order to explore allele-level estimates , we first grouped alleles in “lineages” , which comprise groups of alleles which are evolutionarily and functionally related , since the large number of alleles would make sample sizes per allele too sparse . Although the expression of individual allelic lineages is highly variable among individuals , there is an overall significant difference in expression among lineages ( Welch’s ANOVA p-values ranging from 3 . 7 × 10−8 for HLA-DPB1 to 6 × 10−51 for HLA-DQA1 ) . Enhanced coordination of gene expression has been proposed as an advantage for gene clustering , as seen at the HLA region [1 , 2] . We found a high correlation of expression both within the group of Class I and Class II genes , and lower levels between Class I and Class II genes , which are more than 1Mb apart ( Fig 8A ) ( but see [22] for a result of no co-expression among Class I genes ) . A possible cause for co-expression of genes which are physically close to each other is that regulatory activity is structured in domains ( CRDs , for Cis Regulatory Domains ) [52] . Such domains comprise contiguous regions along a chromosome , and their existence predicts co-expression along haplotypes for genes associated with the same CRD . In fact , previous studies did find evidence that expression of some genes in the HLA region was a feature associated with haplotype membership [23 , 62] . We used our inferences of HLA haplotype structure to investigate if there is an haplotypic effect on coordination of expression among nearby HLA genes . Specifically , we tested the hypothesis that co-expression is stronger between alleles located within the same haplotype than between those on different haplotypes . We did not find a consistently higher correlation of expression for alleles on the same haplotype ( correlation within haplotypes being higher in only 7 out of 18 locus pairs surveyed within Class I or Class II , Fig 8A ) . This result suggests that correlation of expression among HLA loci in LCLs is a result of factors acting at the gene level , and is not driven predominantly by properties of the haplotype . For example , we observe correlation of expression between Class II genes and CIITA , the Class II Major Histocompatibility Complex Transactivator . This correlation is not driven by proximity ( CIITA is on chromosome 16 ) , but rather by a trans regulatory mechanism ( Fig 8B ) .
The contribution of HLA variation to normal and disease phenotypes goes beyond peptide specificity , and includes other factors which influence the strength of immune responses , such as HLA expression levels . However , we are still only starting to understand the regulation of expression of these genes , a consequence of their extreme polymorphism which imposes challenges to the methodologies available to measure mRNA and protein levels . There has been an increasing effort to use RNA-seq data for in-silico HLA typing ( [63] and other methods reviewed in [64] ) and to estimate expression levels for HLA genes [30 , 33] . In this paper , we present HLApers , an HLA-personalized pipeline that quantifies HLA expression based on RNA-seq . We used these expression estimates to identify eQTLs for HLA genes , to estimate the degree of allelic imbalance in expression , and to investigate how the alleles at eQTLs explain variation in expression among HLA alleles . When new technologies/methods are developed , it is a good practice to evaluate if the results are in agreement with the current knowledge , which in the case of HLA was mostly derived from qPCR or antibody-based studies . However , in our survey of this literature , we found it difficult to compare our newly obtained expression estimates based on RNA-seq with those from previous studies . There are several reasons underlying this difficulty , which we now discuss . First , expression estimates based on qPCR and antibody-based approaches are in general not comparable among different HLA loci , since different primers/antibodies are often developed for each locus . Thus , the information for expression differences which we observe among HLA loci are usually not available from studies using these techniques . A few studies developed antibodies with comparable affinities for HLA-A/B/C , allowing comparison of expression between loci . Using such a method , Apps et al . [19] found an at least 12-fold reduction in HLA-C expression with respect to HLA-A and HLA-B as measured by flow cytometry , whereas the difference we documented based on RNA-seq was only twofold . However , the Apps et al . [19] study was restricted to a specific haplotype , and because our results show substantial within locus variation in expression levels , it is difficult to compare studies . Second , many additional layers of biological factors make a direct comparison across studies challenging . For example , the abundance of total protein , of proteins on cell surface or of mRNA in the cell , represent different molecular phenotypes associated with gene expression , given the existence of various post-transcriptional and post-translational regulatory processes [18 , 20 , 42 , 43 , 65 , 66] . As a consequence , there are good biological reasons to expect differences when expression is estimated at each of these levels . In addition , it is well known that gene expression varies among tissues and cell types [26 , 32] , posing an additional challenge in comparisons between our findings and those previously reported in the literature , which are often based on analyses of different cell types or tissues . Third , regarding mRNA quantification , qPCR and RNA-seq each have their own sources of bias , and the finding that there are differences between studies using these approaches should not , on its own , provide definitive evidence that one or the other is inaccurate . Future studies will need to evaluate the degree to which these two technologies can be compared for quantification of HLA expression . Nevertheless , given the availability of qPCR based quantifications for HLA expression , as well as the interest in obtaining allele-level estimates of expression , we compared the expression for individual HLA lineages obtained by RNA-seq with those from qPCR from previous studies . We found that RNA-seq and qPCR show high concordance for the relative ordering of lineage-level expression for HLA-B and HLA-C ( 66% and 100% , respectively ) , whereas for HLA-A concordance was low , at 28% . Therefore results suggest that for HLA-B and HLA-C , when we examine lineages that are sufficiently different in their expression relative to one another ( see S1 Text ) , the two methods are in high agreement , but further research is required to identify the sources of discordance between RNA-seq and qPCR for HLA-A . It is important to recall that these analyses rely on lineage-level expression estimates obtained by qPCR which are indirect , since they estimate locus-level expression ( whereas RNA-seq data analysis with HLApers distinguishes among alleles ) , adding another source of differences among methods . Our perspective is that differences across studies call for further investigation , and highlight the potential biological variation that needs to be accounted for when comparing cell types , infection status , methodological approach , and molecule being quantified . In this study we present HLApers , and use simulations and empirical data analyses to show that it produces estimates for HLA expression from RNA-seq with high precision and accuracy . HLApers incorporates a Maximum-Likelihood estimator to deal with instances of reads mapping to multiple alleles or loci , following a strategy that has been widely used in RNA-seq studies [26 , 67] . Our pipeline can use different alignment strategies ( e . g . STAR-Salmon [34 , 36] or kallisto [35] ) , but we show that the accuracy of expression depends on the sequences contained in the index , and is less dependent on the program used . The impact of using an HLA-personalized index on expression estimates varies markedly among loci . We found that for HLA-DQA1 there are large differences between gene-level expression estimates obtained using the reference transcriptome and those obtained using the HLA-personalized index . However , this difference is quite low for Class I genes , and intermediate for Class II loci other than HLA-DQA1 ( Figs 2 and 4 ) . However , even for the Class I genes , using the personalized index does result in changes in expression estimates with respect to the reference transcriptome . Therefore , we asked whether these changes have an impact on downstream analyses , such as eQTL mapping . When we compare the eQTLs identified using either HLApers or the reference transcriptome , 6 out of 8 loci had non-overlapping eQTLs . However , in most cases the same biological signal was being captured ( S2 Table ) , and the p-values ( Fig 5 ) and the causality ( S5 Fig ) of the eQTLs obtained with HLApers are only modestly better than when the reference transcriptome is used . We show that most of the eQTLs we identified ( 14/17 ) share a signal with previously reported regulatory SNPs ( either experimentally validated SNPs and eQTLs ( S3 Table ) or CRD-QTLs ( S5 Table ) ) . This indicates that , despite of the improvement in expression estimates that personalized pipelines can generate , larger sample sizes are necessary in order to identify eQTLs with greater probability of being causal , and to identify novel eQTLs with smaller effects . We note that the use of transformed cell lines ( such as LCLs ) may lead to expression profiles specific to this environment , and as a consequence some of the regulatory architecture underlying the expression of the HLA genes may not be shared with other tissues , cell types or treatments . The HLA-personalized approach provides expression estimates at the HLA allele level , which is not a product of standard RNA-seq pipelines . We integrate allele-level information with the eQTLs mapped for the genes , showing that the HLA allele is a relevant layer of information to understand the regulation of gene expression , because in some instances the regulatory architecture is linked to specific HLA alleles . This joint mapping of regulatory variants and assessment of expression of HLA alleles can illuminate the understanding of the HLA regulation , and contribute to disentangle specific contributions to disease phenotypes .
All data were obtained from third party sources and no additional ethical approval was required . In order to create the index , we downloaded 16 , 187 nucleotide sequences for 22 HLA loci ( HLA-A , HLA-B , HLA-C , HLA-E , HLA-F , HLA-G , HLA-DMA , HLA-DMB , HLA-DOA , HLA-DOB , HLA-DPA1 , HLA-DPB1 , HLA-DQA1 , HLA-DQB1 , HLA-DRA , HLA-DRB1 , HLA-DRB2 , HLA-DRB3 , HLA-DRB4 , HLA-DRB5 , HLA-DRB7 , HLA-DRB8 ) from the International Immunogenetics/HLA database ( IMGT ) ( release 3 . 31 . 0 available at https://github . com/ANHIG/IMGTHLA ) . For many alleles , sequence data is not available for the entire coding region ( e . g . , only for exons 2 and 3 for class I , and exon 2 for class II genes , which are called ARS exons ) . Because the lack of sequences for much of the coding region would cause the exclusion of many reads , including those mapping to the boundaries of the available exons , for each allele with partial sequence we used the available sequence to find the closest allele which has the complete sequence , and attributed the sequence from this allele . This is expected to introduce little bias in either the genotyping step ( because ARS exons are the most polymorphic and sufficient to distinguish specific alleles ) or the expression estimation ( because the non-ARS sequence attributed is likely very similar to the real one ) . For the final index file , we replaced the HLA transcripts in the reference transcriptome ( Gencode v25 , primary assembly ) with the HLA diversity described above . STAR’s module genomeGenerate , Salmon’s index , and kallisto’s index compile an index from these sequences . In order to select the alleles to be used in the HLA-personalized index , we observed that a simple procedure of selecting the 2 alleles with the largest number of estimated read counts after applying a zygosity threshold is sufficient to produce calls with accuracy of ≥ 95% . However , in order to avoid false homozygotes and false heterozygotes , we implemented additional steps . First , we selected the top 5 alleles and applied an intra-lineage threshold of 0 . 25 , meaning that only alleles which had at least 25% of the total expression in their lineage were considered for further steps . For each individual , we compiled an index containing only these ( up to ) 5 alleles and estimated their expression . We then determined if the individual was heterozygote at the lineage level by applying a threshold of 0 . 15 on lineage expression levels . The lead allele from each lineage was selected to compose the genotype . A zygosity threshold of 0 . 15 was applied to decide whether the genotype was heterozygous at the allele-level . For each locus , the reads mapped to the lead allele were removed , and another step of alignment and quantification was performed in order to determine if the second allele was real , or just noise due to extensive similarity to the lead allele . If the second allele had at least 1% of the locus read counts , it was kept , otherwise the genotype was considered to be homozygous for the lead allele . The thresholds described above were chosen because they maximized the concordance with the Sanger sequencing typings [6] , while also minimizing the rate of false homozygotes and heterozygotes . We implemented two versions of the HLApers pipeline: ( 1 ) one using STAR ( v2 . 5 . 3a ) [34] to map reads followed by Salmon ( v0 . 8 . 2 ) [36] to quantify the expression , and ( 2 ) using kallisto ( v0 . 43 . 1 ) [35] , which performs pseudoalignment and quantification . The quantification pipeline is structured in a two-stage process , first identifying the most expressed allele ( s ) at each HLA locus in order to infer the genotype which is present , and next quantifying expression for these inferred genotypes as well as for the rest of the transcriptome ( Fig 1 ) . Reads were aligned directly to the transcriptome . STAR alignments were passed to Salmon for quantification ( module quant under alignment mode ) , whereas kallisto directly produces quantifications with the quant module . In both the HLA typing step ( for which the index contains all HLA sequences in the IMGT database ) and the quantification step ( for which an HLA-personalized index is used ) , short reads can map to more than one locus , or more commonly to multiple alleles of the same locus ( multimaps ) . The quantification methods we are using deal with multimaps by inferring maximum likelihood estimates optimized by an expectation-maximization algorithm to probabilistically assign reads to each reference in the index , and also include models to account for sequencing bias . For the mapping with STAR , we tuned parameters in order to avoid discarding multimaps and to accommodate mismatches . For quantification , we used all bias correction options available ( –seqBias and –gcBias in Salmon , and –bias in kallisto ) . Polyester is an R package designed to simulate RNA-seq datasets [41] . We used the function simulate_experiment_countmat to simulate transcriptome data for 50 randomly chosen GEUVADIS individuals . Simulations were based on the read lengths and counts in the original data , with library sizes of 30 million reads , sampling without bias from a normal distribution of read start sites ( with average fragment length of 250bp and sd of 25bp , and error rate of 0 . 005 ) . The code for the generation of the simulated datasets is available at https://github . com/genevol-usp/hlaexpression/tree/master/simulation/data . We processed the simulated reads with STAR-Salmon to perform the quantifications using different indices: To investigate the relationship between quantifications and sequence divergence with respect to the reference , we used the R function adist to calculate the proportion of mismatches between the HLA alleles carried by the individuals and the alleles in the reference genome . We quantified HLA expression based on RNA-seq data for 358 European individuals , in samples of LCLs ( Lymphoblastoid Cell Lines ) , made available by the GEUVADIS Consortium [24] ( we excluded samples from the original dataset which are not in the 1000 Genomes phase 3 ) . We performed expression quantification using the HLApers and reference transcriptome pipelines . We evaluated the reproducibility of the HLA quantifications by analyzing a subset of 97 individuals for which a replicate was available ( https://github . com/genevol-usp/hlaexpression/blob/master/geuvadis_reanalysis/replicates/data/write_sample_info . R ) . For the eQTL analysis , we used only autosomal genes which are expressed in a large proportion of samples , exploring the thresholds of TPM > 0 in at least 25% , 50% , or 75% of samples . In order to correct the expression data for technical effects , we sequentially removed the effect of the first 0 to 100 PCs and ran an eQTL analysis for each condition ( S6 Fig ) . The configuration of thresholds and number of PCs which maximized the eQTL discovery ( at FDR = 5% ) was considered . This resulted in the use of genes expressed in ≥ 50% of samples ( 19 , 613 genes ) , and 60 PCs . The PCA analysis and data correction were performed with QTLtools v1 . 1 [47] , using the modules pca and correct respectively . For the genetic variant data , we used the 1000 Genomes Phase 3 biallelic variants , lifted to GRCh38 coordinates , after filtering for MAF ≥ 0 . 05 in the individuals included in this study ( 6 , 837 , 505 variants in total ) . In order to control for population structure in the eQTL analysis , we ran a PCA on the variant genotype data and assessed the PCs which captured the structure . We used QTLtools pca requiring that variants should be at least 6kb apart . After visual inspection of the plots in S7 Fig , PCs 1–3 were used as covariates in the eQTL analysis . We used QTLtools cis to conduct the cis-eQTL analysis using the following model: PCA - corrected and standard normal expression ∼ SNPs + covariates ( PCs for population stratification ) The permutation pass was performed with 1000 permutations and a cis-window of 1Mb . P-values were computed by beta approximation and significance was determined by running the script runFDR_cis . R provided by QTLtools with FDR of 5% . Multiple eQTLs with independent effects on a particular gene were mapped with a conditional analysis based on step-wise linear regression ( see Supplementary method 8 in [47] ) . The method automatically learns the number of independent signals per gene and provides sets of candidate eQTLs per signal . In order to investigate the putative function of the eQTLs we mapped , we investigated whether these eQTLs were present in ENCODE [53] regulatory elements annotated for LCLs . We used three types of functional annotations: open chromatin regions given by DNAse footprinting , transcription factor binding sites ( TFBS ) assayed by ChIP-seq , and histone modifications . We performed an RTC [48] analysis as described in [47] to investigate whether our eQTLs tagged the same causal variant as a GWAS variant or previously reported eQTL . We downloaded the GWAS catalog data ( v1 . 0 . 1 ) from https://www . ebi . ac . uk/gwas/api/search/downloads/alternative and selected associations with p-value < 10−8 . We obtained the coordinates of recombination hotspots from http://jungle . unige . ch/QTLtools_examples/hotspots_b37_hg19 . bed . We applied the RTC module implemented in QTLtools ( QTLtools rtc ) , selecting the HLA region only , using a D’ threshold of 0 . 5 , and turning on the conditional flag ( –conditional ) to test all independent eQTLs for a gene . Following the recommendation of Delaneau et al . ( see Supplementary Note 7 in [47] ) , we considered that two SNPs tagged the same functional signal if the RTC score was >0 . 9 . To investigate whether there is a haplotypic coordination of expression at HLA , we used phased HLA genotype data to verify if there was more correlation of expression between alleles on the same haplotype than on different haplotypes ( Fig 8 ) . To also assess the phasing between HLA alleles and the eQTLs ( Fig 6 ) , we included the eQTLs mapped for each HLA gene in the phasing procedure , accounting for the fact that their phasing was already known from 1000 Genomes Project . In order to be conservative in the phase estimation , we used only the haplotype calls which were concordant between two approaches for phasing . First , we used PHASE [68] to determine the haplotype of each allele in the genotype , providing HLA allele designations and phased eQTL genotypes as input . Second , we checked the compatibility of the individual HLA genotypes with the phased SNP haplotypes from 1000 Genomes . Given all possible haplotypes for each individual at HLA-G∼A∼E∼C∼B∼DRA∼DRB1∼DQA1∼DQB1∼DPA1∼DPB1 , we checked which combination of 2 haplotypes minimized the number of differences to the 1000 Genomes data to infer the haplotypes present in the individual . This resulted in 417 haplotypes completely concordant between the two approaches . The HLApers pipeline is available at https://github . com/genevol-usp/HLApers . The entire analysis , including simulations , index compilation , quantification of expression , eQTL mapping , etc is available at https://github . com/genevol-usp/hlaexpression . | The level at which a gene is expressed can have important influence on the phenotype of an organism , including its predisposition to develop diseases . One way to estimate gene expression is by quantifying the abundance of RNA . RNA-seq has become the method of choice to provide such estimates at the genomewide scale . However , the application of RNA-seq to HLA genes —key players in the immune adaptive response— has remained a rarely explored approach . This is due to the problem of mapping bias , which causes deficient read alignment at genes which are very polymorphic and different from the reference genome . This has motivated approaches that replace the single reference genome with personalized sequences , comprised of the individual’s specific HLA genotype . Here we explore the use of computational frameworks to obtain reliable expression levels for HLA genes from RNA-seq datasets . We present a pipeline in which the quantification of HLA expression is carried out using methods which account for HLA diversity , avoiding the biases of standard approaches . We then evaluate the impact of this form of quantifying HLA expression on downstream analyses . The pipeline also allows us to integrate information on eQTLs with expression levels at the HLA allele-level , which can help disentangle different contributions to disease phenotypes and help understand the regulatory architecture at the HLA region . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"genome-wide",
"association",
"studies",
"variant",
"genotypes",
"alleles",
"genetic",
"mapping",
"genome",
"analysis",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"sequence",
"alignment",
"bioinformatics",
"gene",
"expression",
"genetic",
"loci",
"haplotypes",
"heredity",
"database",
"and",
"informatics",
"methods",
"genetics",
"transcriptome",
"analysis",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology",
"human",
"genetics"
] | 2019 | Expression estimation and eQTL mapping for HLA genes with a personalized pipeline |
Human infection with Bwamba virus ( BWAV ) and the closely related Pongola virus ( PGAV ) , as well as Nyando virus ( NDV ) , are important causes of febrile illness in Africa . However , despite seroprevalence studies that indicate high rates of infection in many countries , these viruses remain relatively unknown and unstudied . In addition , a number of unclassified bunyaviruses have been isolated over the years often with uncertain relationships to human disease . In order to better understand the genetic and evolutionary relationships among orthobunyaviruses associated with human disease , we have sequenced the complete genomes for all 3 segments of multiple strains of BWAV ( n = 2 ) , PGAV ( n = 2 ) and NDV ( n = 4 ) , as well as the previously unclassified Mojuí dos Campos ( MDCV ) and Kaeng Khoi viruses ( KKV ) . Based on phylogenetic analysis , we show that these viruses populate 2 distinct branches , one made up of BWAV and PGAV and the other composed of NDV , MDCV and KKV . Interestingly , the NDV strains analyzed form two distinct clades which differed by >10% on the amino acid level across all protein products . In addition , the assignment of two bat-associated bunyaviruses into the NDV group , which is clearly associated with mosquito-borne infection , led us to analyze the ability of these different viruses to grow in bat ( RE05 and Tb 1 Lu ) and mosquito ( C6/36 ) cell lines , and indeed all the viruses tested were capable of efficient growth in these cell types . On the basis of our analyses , it is proposed to reclassify the NDV strains ERET147 and YM176-66 as a new virus species . Further , our analysis definitively identifies the previously unclassified bunyaviruses MDCV and KKV as distinct species within the NDV group and suggests that these viruses may have a broader host range than is currently appreciated .
The Bunyaviridae are a large , diverse group of more than 350 viruses divided into 5 genera , of which more than 150 belong to the genus Orthobunyavirus [1] . Importantly , these viruses represent a significant cause of arthropod-borne human disease worldwide , with infection often associated with a febrile and/or encephalitic illness , and in rare cases also hemorrhagic manifestations [2] . However , despite their importance for public health , from both a genetic and an evolutionary standpoint the family has been only poorly characterized . In addition to the well-known agents of human bunyavirus disease in Africa , such as Rift Valley Fever virus ( RVFV; genus Phlebovirus ) and Crimean-Congo Hemorrhagic fever virus ( CCHFV; genus Nairovirus ) , there are a number of viruses in the genus Orthobunyavirus that are also agents of human disease , with many of them being highly prevalent within their endemic areas . Among these viruses , by far the most prevalent appears to be Bwamba virus ( BWAV ) , which has been reported to be among the most common arthropod-borne diseases in Africa [3] . Infection is associated with a relatively non-specific febrile illness that , while usually self-limiting , is frequently associated with exanthema and can include meningeal involvement [4] . However , recently a group of 14 fatal cases of BWAV infection with hemorrhagic complications , particularly bleeding from the oral mucosa and into the gastrointestinal tract , were reported during an outbreak among Rwandan refugees [4] , indicating that BWAV infection can also be associated with very severe disease manifestations . To date a total of only 21 human cases of BWAV infection have been reported from various countries ( including Uganda [4] , [5] , Central African Republic [6] , Kenya [7] and Tanzania [4] ) . However , virus isolation and/or serological studies suggest that this virus circulates in several additional countries ( i . e . Mozambique , South Africa and Nigeria ) ( Figure 1A ) and show that seropositivity exceeds 90% in some populations [8]–[11] . Thus , these data clearly suggest that a lack of concerted surveillance efforts , together with the frequency of infections with other pathogens causing febrile illness in the affected areas , have contributed to under-reporting , and as a result an under-appreciation of the disease burden imposed by BWAV infection . Interestingly , a closely related virus , Pongola virus ( PGAV ) , has been isolated from mosquitoes ( Aedes circumluteolus ) in South Africa [9] and appears to have been responsible for a single reported case of human infection associated with febrile illness in Uganda [12] . While serological studies also indicate a high prevalence ( 9–26% ) of PGAV infection in several countries ( i . e . South Africa , Mozambique , Namibia , Botswana , Angola [8] , [13]–[15] ) , it must be noted this virus is highly cross-reactive with BWAV in many serological tests [5] , [9] , [16] , and that it has historically been difficult to distinguish the distributions and relative abundance of these two viruses; particularly since the affected geographical regions appear to substantially overlap ( Figure 1A and B ) . As a consequence of this high degree of serological cross-reactivity , these viruses are classified together into a single serogroup . Similar to BWAV and PGAV , Nyando virus ( NDV ) is also capable of causing moderate-severe febrile disease in humans . Although to date only a single human case with multiphasic fever , myalgia and vomiting has been reported from the Central African Republic [6] , [17] , again serological studies indicate a high level of seroprevalence in many countries , including Kenya , Uganda and Senegal [18] , [19] ( Figure 1C ) . Together with repeated isolations from mosquito pools [17] , [18] , [20] these findings indicate that , as with many of the African orthobunyaviruses , NDV virus might also be much more prevalent than is currently appreciated . Intriguingly , the limited sequencing data previously available for this virus suggests that Nyando virus ( strain ArB16055; GenBank Accession AM709781 ) is closely related to Bunyamwera virus [21] , a finding that appears to be at odds with the lack of serological cross-reactivity between these viruses . Indeed Nyando virus is classified into a distinct serogroup ( Nyando serogroup ) , of which it is presently the only member [22] . Until now , bunyavirus classification has relied almost exclusively on serological testing , which may include complement fixation , hemagglutination inhibition , immunofluorescence assay and/or viral neutralization assays , and on the basis of these assays , the genus Orthobunyavirus is presently divided into 18 serogroups [22] . However , cross-reactivity between viruses is common and often limits our ability to make definitive identifications based on these methods [23] , [24] . Further , many other bunyaviruses remain uncharacterized as a result of an inability to assign them a position within this serological classification [25] . Examples of such unclassified bunyaviruses include Mojuí dos Campos virus ( MDCV ) , which was isolated from an unknown bat species in Brazil [26] ( Figure 1D ) , and Kaeng Khoi virus ( KKV ) , which has been isolated from bats ( Tadarida plicata and Taphozous theobaldi ) in Thailand and Cambodia [27]–[29] as well as from bedbugs ( Stricticimex parvus and Cimex insuetus ) [27] ( Figure 1E ) . While no informative serological relationships could be established for KKV [29] , MDCV was originally observed to show some cross-reactivity with Nyando virus as well as San Angelo virus ( California encephalitis serogroup ) [25] , raising the possibility that it may be related to one or both of these viruses . While genome sequencing is known to be both a rapid and accurate means of identifying viruses , and is indeed the standard for the identification of most other virus families , it is dependent on the availability of sufficient pre-existing sequence data , something that is presently lacking for bunyaviruses . Indeed , where genetic analysis of these viruses has been performed it is often extremely limited and focuses almost exclusively on the S-segment . However , this issue has been increasingly recognized within the field and is beginning to be addressed , particularly as a result of large-scale de novo sequencing efforts [30]–[32] . In light of the tri-segmented genome structure of these viruses , which allows them to evolve by both antigenic drift and antigenic shift ( i . e . reassortment ) , this leaves us with an incomplete understanding of the exact identity of many bunyaviruses , and the relationships and diversity that exists among them . However , without substantial improvements in the availability of genetic information for these viruses , such determinations will be difficult , if not impossible . In order to improve our understanding of the evolutionary relationships and genetic diversity among orthobunyaviruses causing human disease in Africa , and particularly the viruses of the Bwamba/Pongola virus and Nyando virus serogroups , we have undertaken the complete genome sequencing of multiple strains of each of these viruses . In addition , we have determined the first complete sequences of MDCV and KKV , which has allowed their definitive identification , and based on the genetic relationships identified in our analyses , we have begun to explore the possibility that additional host and vector species may be involved in the ecology of these viruses .
Vero E6 ( African green monkey kidney , ATCC CRL-1586 ) , Tb 1 Lu ( Tadarida brasiliensis lung , ATCC CCL-88 ) and RE05 ( Rousettus aegyptiacus fetus; kindly provided by Ingo Jordan , ProBioGen AG [33] ) cells were grown in Dulbecco's modified Eagle's medium ( DMEM; Sigma-Aldrich ) , supplemented with 10% fetal calf serum ( FCS; Life Technologies ) , 2 mM L-glutamine , 50 U/ml penicillin and 50 µg/ml streptomycin ( Life Technologies ) at 37°C in the presence of 5% CO2 . C6/36 ( Aedes albopictus larva , ATCC CRL-1660 ) were grown in Eagle's Minimum Essential Medium ( EMEM; Life Technologies ) supplemented with 10% fetal calf serum ( FCS; Life Technologies ) , 2 mM L-glutamine , 50 U/ml penicillin and 50 µg/ml streptomycin ( Life Technologies ) at 28°C in the presence of 5% CO2 . The virus strains used in this study were kindly provided by the Centers for Disease Control and Prevention ( CDC ) , Division of Vector-Borne Diseases ( DVBD ) arbovirus reference collection and the World Reference Center for Emerging Viruses and Arboviruses ( WRCEVA ) arthropod-borne virus reference collection . Information regarding their origin is summarized in Table 1 . Virus stocks were grown in Vero E6 cells in DMEM supplemented with 2% FCS , 2 mM L-glutamine , 50 U/ml penicillin and 50 µg/ml streptomycin and 10 µg/ml Mycokill AB ( GE Healthcare ) . Virus growth was monitored based on the appearance and progression of cytopathic effect ( CPE ) in cells . When advanced CPE was observed ( 50–75% of cells detached ) , the culture supernatants were harvested for RNA isolation . Cell culture supernatants from infected cells were spun twice at 1 , 000× g for 5 min at 4°C to pellet cell debris . For Sanger sequencing , RNA was then extracted using the QIAamp viral RNA extraction kit ( Qiagen ) according to the manufacturer's directions . Alternatively , for Next Generation sequencing ( NGS ) , samples were further purified and concentrated through a centrifugal filtration device ( Millipore ) prior to RNA extraction , as previously described [34] . cDNA for NGS was synthesized using a previously described modification [34] of the protocol described by Palacios et al . [35] . Briefly , first strand cDNA was synthesized using the Superscript III Reverse Transcriptase system ( Life Technologies ) with 100–1 , 000 ng of total RNA using a random octamer linked to a defined 17-mer primer ( 5′-GTT TCC CAG TAG GTC TCN NNN NNN N-3′ ) . RNA was then hydrolyzed in NaOH and the single-stranded cDNA ( ss-cDNA ) products purified using the QIAquick system ( Qiagen ) . The resulting ss-cDNAs were randomly amplified using a 1∶9 mixture of the arbitrary 17-octamer primer and a primer targeting a specific 17-mer sequence ( 5′- CGC CGT TTC CCA GTA GGT CTC-3′ ) . The resulting ss-cDNA templates were used as template for PCR using Platinum Taq polymerase . PCR products were purified using the QIAquick kit following the manufacturer's protocol ( Qiagen ) and used as template for sequencing on the 454 Titanium FLX sequencer ( 454/Roche Life Sciences ) . cDNA samples were quantitated using Picogreen reagent ( Life Technologies ) and prepared according to the Rapid Library Preparation Method Manual – GS FLX Titanium Series October 2009 ( 454 Life Sciences ) . A multiplex was titrated in medium volume emulsion ( MVE ) format to determine the optimal copy-per-bead ratio ( CPB ) which produced the best sequencing quality . A 454 Titanium sequencing run was then performed using 1 CPB . Genomic viral sequences were produced on the 454 FLX sequencer and de novo assembled using GS De Novo Assembler v2 . 6 ( 454 Life Sciences ) and CLC Genomics Workbench 4 . 0 ( CLC Bio ) . Translated BLAST ( blastx ) was performed to remove non-viral contaminants and the initial assembly was performed using Sequencher v5 . 0 ( Gene Codes ) . Assembled contigs were then verified , refined or corrected by mapping the 454 reads using GS Reference Mapper v2 . 6 ( 454 Life Sciences ) . Where needed , several rounds of manual assembly and trimming were performed in Sequencher with verification done using GS Reference Mapper to eliminate discrepancies or errors discovered during the prior reference mapping procedure . Based on the assembled data obtained from NGS , primers for reverse transcription PCR ( RT-PCR ) and Sanger sequencing were designed ( primer sequences available upon request ) . RT-PCR was performed with Superscript III reverse transcriptase ( Life Technologies ) and iProof DNA polymerase ( Bio-Rad ) . The 3′ and 5′ termini of each genome RNA segment were amplified using both a 3′ and 5′ RACE approach based on ligation-anchored PCR , as previously described [36]–[38] , with some sequences additionally confirmed using a commercially available 5′ RACE System ( Life Technologies ) according to the manufacturer's instructions . The nucleotide sequences obtained for each genome segment , or the deduced amino acid sequences of each of the open reading frames , were aligned with the representative sequences of other known members of the genus Orthobunyavirus from GenBank ( Table S1 ) . Sequences were aligned using the MUSCLE algorithm and the evolutionary history for each tree construction was inferred using the neighbor-joining ( NJ; [39] ) and maximum likelihood methods ( ML; [40] ) , as implemented in MEGA 5 [41] . For the NJ analyses , the evolutionary distances were computed using the Maximum Composite Likelihood method [42] . Statistical support for the tree topology obtained with all methods was evaluated based on bootstrap re-sampling [43] with values calculated based on 1 , 000 replicates . RE05 , Tb 1 Lu and C6/36 cells were grown for 80–90% confluence in 6 well plates , and the various virus strains indicated were used to infect these cells at an MOI of 0 . 1 . The formation of CPE was monitored daily from 24–72 h and supernatants were harvested at 72 h post-infection for titration via plaque assay . Briefly , a 10-fold dilution series of supernatants were prepared in DMEM without FCS or supplements and 500 ul per well was applied to 12 well plates . Following incubation for 1 h at 37°C virus dilutions were removed and wells were overlaid with 0 . 9% agarose in 1× MEM containing 2% FCS , 2 mM L-glutamine , 50 U/ml penicillin and 50 µg/ml streptomycin . Once solidified plates were incubated at 37°C for 3 d ( NDV ( MP401 ) , NDV ( ERET147 ) , MDCV and KKV ) or 5 d ( BWAV and PGAV ) prior to fixation overnight in 10% formalin containing 0 . 1% crystal violet ( Sigma-Aldrich ) . The genome sequences determined in this study were deposited in GenBank under the following accession numbers ( S segment , M segment , and L segment ) : KJ867176 , KJ867177 and KJ867178 ( PGAV , strain SA Ar1 ) ; KJ867179 , KJ867180 and KJ867181 ( PGAV , strain 191B-07 ) ; KJ867182 , KJ867183 and KJ867184 ( BWAV , strain M459 ) ; KJ867185 , KJ867186 and KJ867187 ( BWAV , strain UgAr1888 ) ; KJ867188 , KJ867189 and KJ867190 ( NDV , strain MP401 ) ; KJ867191 , KJ867192 and KJ867193 ( NDV , strain UgAr1712 ) ; KJ867194 , KJ867195 and KJ867196 ( NDV , strain ERET147 ) ; KJ867197 , KJ867198 and KJ867199 ( NDV , strain YM176-66 ) ; KJ867200 , KJ867201 , and KJ867202 ( MDCV , strain BeAn 276121 ) ; KJ867203 , KJ867204 and KJ867205 ( KKV , strain PSC-19 ) .
In order to better understand orthobunyavirus evolution as it relates to the relationships between and degree of diversity among BWAVs , PGAVs and NDVs , we have determined the complete genome sequences for all 3 viral RNA segments from multiple strains of each of these viruses . In addition , we have determined the complete sequences of a single strain of each of the “uncharacterized” bunyaviruses MDCV and KKV ( listed in Table 1 ) . Based on phylogenetic analysis of these viruses in relation to previously published data obtained from GenBank for other members of the genus Orthobunyavirus ( Table S1 ) , it is apparent that BWAV and PGAV form a distinct virus clade , separate from that formed by the NDV strains ( Figure 2 ) , consistent with their assignment to distinct serogroups . For BWAV and PGAV , all three segments form a lineage that diverged from a common ancestor shared with the California encephalitis virus ( CEV ) group , indicating that these two groups are closely related at the genetic level . The NDV clade is less closely related to any currently recognized orthobunyavirus group forming a very distinct genetic grouping branching ancestrally to both the CEV and BWAV/PGAV lineages . However , closer analysis of the NDV group also shows considerable variation between NDV strains , such that the NDV group actually forms two distinct clades: one composed of strains MP401 and UgAr1712 , while the other is composed of strains ERET147 and YM176-66 . The groupings and relative positions of these viruses were well-preserved regardless of whether a criterion-based method ( ML; Figure 2 ) or a clustering method ( NJ; Figure S1 ) was used for construction of the phylogenetic trees . Further , the same relationships are observed regardless of whether complete genome segment nucleotide sequence , coding region nucleotide sequence or amino acid sequence datasets are used for the analysis ( data not shown ) . Early serological data for MDCV had indicated a possible distant relationship to both CEV and/or bunyamwera group viruses . Further , a preliminary phylogenetic analysis of very short sequence data fragments available for KKV in GenBank ( accession numbers JN010801 and AY843028–AY843038 ) indicated a possible , but weakly supported , evolutionary relationship to the NDV group ( data not shown ) . On this basis we additionally undertook full-length sequencing of a single strain each of MDCV and KKV using de novo sequencing . Based on these complete genome data we found that both viruses clearly grouped together with the NDV viruses we had sequenced ( Figure 2 ) . However , despite NDV being the closest known relative of both of these viruses , at both the nucleotide and amino acid levels these viruses were considerably divergent from all NDV strains examined , demonstrating only 53–73% nucleotide identity and 39–70% amino acid identity for MDCV , and 53–75% nucleotide identity and 39–72% amino acid identity for KKV , clearly indicating that these should be considered as distinct virus species ( Tables S2–S4 ) . Analysis of the genome structures for all of these viruses indicates that they are consistent with what is known for previously analyzed orthobunyavirus genomes ( Figure 3 ) . The S segments for these viruses ranged from 902 [NDV ( MP401 ) ] to 1061 ( MDCV ) nucleotides and encoded both a nucleoprotein of 233 ( NDV , MDCV , KKV ) to 235 ( PGAV ) amino acids in length and an NS protein of 92 ( BWAV , PGAV , NDV , MDCV ) to 106 ( KKV ) amino acids derived from an alternate downstream ATG . The M segment and L segment were found to be between 4395 ( MDCV ) and 4568 ( KKV ) nucleotides , and 6866 ( KKV ) and 6994 ( MDCV ) nucleotides in length , respectively , and encoded only one long open reading frame ( ORF ) each , corresponding to the glycoprotein precursor ( GPC ) and the RNA-dependent RNA polymerase ( L ) . The GPC protein of the various viruses were found to be between 1415 ( MDCV ) and 1447 ( BWAV ) nucleotides in length , while the L proteins were between 2249 ( KKV ) and 2268 ( NDV ) amino acids in length . In order to apply rational criteria to the taxonomic assignments within these virus groups , we next analyzed the levels of sequence divergence between the different strains of BWAV , PGAV and NDV ( Tables S2 , S3 , S4 , S5 , S6 , S7 ) . Based on these analyses we found 95–99% nucleotide identity among BWAV strains , while values for amino acid sequence conservation were between 98–99% . Among PGAV strains even higher levels of sequence identity were observed with 98%–99 . 6% sequence conservation at the nucleotide level and 99%–100% at the amino acid level . Despite the relationship between strains of either virus , sequence conservation between these two groups decreased to 67%–89% at the nucleotide level and 63%–86% at the amino acid level across all three segments . This clearly supports the classification of BWAV and PGAV as distinct virus species despite previous reports of strong serological cross-reactivity , which renders them nearly indistinguishable in some tests [5] , [9] , [16] . Among the NDVs , the situation observed was rather different . While the NDV clade is clearly highly divergent from all other recognized virus groups , it also demonstrated much more divergence between strains . Analysis showed that the MP401 and UgAr1712 strains exhibit 92%–99% identity at the nucleotide level and 96%–100% identity at the amino acid level , consistent with the levels of conservation observed between BWAV and PGAV strains . Similarly , the ERET147 and YM176-66 strains showed the expected high levels of sequence conservation , with 80%–98% identity at the nucleotide level and 90%–98% at the amino acid level . However , between these groups much more substantial differences in sequence were observed and identity levels dropped to 61–89% at the nucleotide level and 57%–86% at the amino acid level , respectively . These levels are then very similar to those seen when comparing BWAV and PGAV , members of two different virus species , and suggest that these two NDV clades should also be recognized as distinct orthobunyavirus species . Given the paucity of full-length genome sequences available for bunyaviruses , including the orthobunyaviruses , little is known about the sequence and/or arrangement of their terminal untranslated regions ( UTRs ) ( Figure S2 ) . Based on a comparison of our full-length BWAV/PGAV and NDV/MDCV/KKV group sequences we noted that all sequences determined for these virus groups contained the well-conserved terminal sequences believed to be characteristic of all orthobunyavirus genome sequences ( 3′-AGTAGTGTAC…GCACACTACT-5′ ) . In addition we found that the downstream 5 nt fit well to the sequences determined for Jamestown Canyon virus ( JTCV ) , a member of the closely related California Encephalitis virus group . Further , where deviations from this prototype sequence were observed , compensatory mutations are found in the other UTR , which would then maintain base pairing at these positions . Such deviations from the established sequences of CEV group members were seen for the MDCV and KKV S segment UTRs and the BWAV and PGAV M segment UTRs ( Figure S2 ) . This observation supports previously proposed base pairing models , based on in vitro work , which have suggested direct interactions between the 3′ and 5′ UTR sequences that need to be maintained for functionality [44] . Beyond these well conserved terminal sequences we found that the UTR sequences exhibit a general lack of conservation in both sequence and length between different virus species . The 3′ UTRs varied in length from 24–85 nt ( S segment: 38–85 nt , M segment: 24–49 nt , L segment: 27–45 nt ) , with the PGAV M segment 3′UTR being uncommonly long in comparison to other 3′ UTR sequences . Compared to the 3′ UTRs , the 5′ UTR sequences were generally much longer , ranging from 85–310 nt , and with sequences in excess of 200 nt determined for the BWAV , MDCV and KKV S segment 5′ UTRs and the KKV M segment 5′ UTR . Interestingly , despite the marked variability of these sequences between virus species , within a single species these sequences are in fact highly conserved . This can be clearly seen when examining the NDV ( MP401 ) and NDV ( UgAr1712 ) UTR sequences in comparison to those of NDV ( ERET147 ) and NDV ( YM176-66 ) ( Figure S2 ) . As such , both UTR sequence and length may provide useful additional criteria/markers for species delineation . The genetic assignment of MDCV and KKV to the NDV clade was surprising , since neither of these viruses has been previously associated with transmission from mosquitos , which are the sole established vector for the transmission of BWAV , PGAV and NDV . In contrast , MDCV and KKV were both originally isolated from bats , which is rather unusual for orthobunyaviruses and has not been reported for BWAV , PGAV or NDV . However , it raises the possibility that these viruses might have a broader host/vector range than is currently appreciated . In order to establish the feasibility of a role for these additional vector and host species in nature , we assessed the ability of representative viruses from the BWAV , PGAV , NDV , MDCV and KKV groups to productively infect cells from African ( Rousettus aegyptiacus ) and South American ( Tadarida brasiliensis ) bat species , as well as an Aedes albopictus mosquito cell line . Our data clearly indicate that all of these viruses have the ability to grow in both bat and mosquito cell types with titre increases in infected cells of between 1 . 5–4 logs between 0 and 72 hours post-infection ( Figure 4 ) . During the same time frame all of these viruses showed ∼3 logs of growth in VeroE6 cells , which are highly permissive for infection with a broad range of orthobunyavirus . There were no identifiable trends observed regarding which viruses ( i . e . mosquito associated African orthobunyaviruses [BWAV , PGAV , NDV] or bat associated orthobunyaviruses from other regions [MDCV , KKV] ) showed more efficient growth in any of these cell types . During infection , all of the viruses tested showed prominent cytopathic changes ( CPE ) in each of the mammalian cell types examined ( i . e . VeroE6 , RE05 and Tb 1 Lu cells; data not shown ) . In contrast , none of the viruses produced clear CPE in C6/36 cells . This lack of CPE in C6/36 cells occurred despite all viruses showing 3–4 logs of virus growth in these cells , comparable to what is seen , for instance , with VeroE6 cells where strong CPE formation is observed . Thus , it appears that this lack of CPE formation is a feature of infection in C6/36 cells rather than being influenced by the different viruses tested . Overall , based on these data it appears at least possible for KKV and MDCV to productively infect mosquito cells . Similarly infection of bat cells with BWAV , PGAV and NDV is also possible and leads to productive infection associated with cytopathological changes .
In this study we have determined the first full-length sequences for all three segments of multiple strains of BWAV and PGAV , as well as NDV , and for single strains of the related MDCV and KKV . As a result we have been able to definitively establish the relationships among these viruses , as well as their relationship to other orthobunyavirus groups . This work has not only clarified previously uncertain assessments about their relationships based on serology but has also led to the identification of two previously unclassified bunyaviruses , MDCV and KKV , as close relatives of NDV , and the identification of existing NDV strains as highly diverse , warranting classification into distinct virus species . Our complete genome-based analysis of all three segments of BWAV and PGAV confirmed their placement within the orthobunyavirus genus , consistent with what has been previously reported based on S-segment analysis alone [45] . In particular , the position of these groups , branching immediately ancestral to the CEV group viruses for all three segments , explains previous reports of cross-reactivity of BWAV to members of the CEV group . Interestingly , despite their high degree of serological cross-reactivity in many assays , BWAV and PGAV display amino acid divergence values for all viral proteins that clearly support their classification as separate virus species . Indeed , the high degree of serological cross-reactivity between these viruses , including in neutralization assays [5] , [9] , [16] , is surprising given that they exhibit only 64% amino acid identity in GPC . This suggests that there may be an unusually high degree of conservation among neutralizing epitopes in these viruses , in relation to the overall levels of sequence conservation , and/or that there may be a marked immunodominance of a few well conserved epitopes . Unlike for BWAV and PGAV , the relative position identified by our analysis for NDV was not consistent with a previous report examining the S segment sequence of NDV ( strain ArB16055 ) [21] . This previous study had suggested a close genetic relationship to Bunyamwera virus , an observation that was apparently at odds with serological evidence assigning NDV to a distinct serogroup [19] . In contrast , we clearly observed that all four of our Nyando virus isolates fell into a distinct clade ancestral to those formed by BWAV/PGAV and CEV and that this position was consistent for all three segments . We did not see any evidence for Nyando virus strains that were genetically related to bunyamwera virus on the S segment , suggesting that the identity of NDV ( strain ArB16055 ) needs to be closely re-examined , as it may either have been misidentified or may be reassortant in nature . However , based on the currently available data , there does not appear to be any evidence for reassortment involving either the BWAV/PGAV or NDV groups . In our study , very little genetic divergence was noted among different strains of BWAV and PGAV , despite substantial differences in the location and/or time at which they were collected , indicating that molecular detection methods based on information from one or a few virus strains may be adequate to detect all of the genetic diversity present for BWAV and PGAV . However , given the small number of isolates available for each of these virus species , we cannot exclude the possibility that other strains showing much more substantial sequence divergence also exists . In contrast to the situation for BWAV and PGAV , for NDV a large degree of genetic divergence was noted among the strains analyzed . Indeed , this genetic divergence was to such an extent that , based on a cut-off value of 10% amino acid divergence , a division of the existing NDV strains into two distinct virus species would be warranted . Given the prevailing naming conventions among bunyaviruses ( i . e . naming based on the location of initial isolation ) we would , therefore , propose that the prototype NDV ( MP401 ) , isolated from the Nyando river valley [18] , and the closely related NDV ( UgAr1712 ) continue to be referred to as “Nyando virus” , while NDV ( ERET147 ) , could be reclassified as “Manéra virus” , reflecting its original isolation from the Manéra forest in Ethiopia , along with the closely related NDV ( YM176-66 ) . Based on their close serological relatedness , it is also likely that other ERET and YM series viruses ( e . g . ERET124 , YM120-68 and YM259-68 ) isolated as part of the same studies [46] , [47] will also fall into this genetic group . Our observation that NDV forms two highly distinct genetic groups supports early findings indicating that these viruses are serologically quite distinct [19] . Further , the closer genetic grouping of the ERET147 and YM176-66 strains is also supported by serological findings obtained during the initial isolation of YM176-66 [48] . The placement of MDCV within a larger NDV clade is also supported by early serological data which indicated reactivity by immunofluorescence assay ( IFA ) and complement fixation ( CF ) , but not neutralization ( NT ) , to both NDV and CEV group viruses [25] . Thus this is consistent with its genetic placement close to , but distinct from , both of these groups , as shown in this study . Initially , the existence of such a close relationship of these bat-associated bunyaviruses to what have been , until now , strictly mosquito-borne viruses was surprising . However , this finding appears to only contribute to the increasing number of bunyaviruses shown to be associated with bats . Interestingly , while for MDCV the virus has only been isolated from a single live bat with unreported health status [26] , for KKV , infection in bats appears to be detrimental in a significant proportion of infected animals , as shown by frequent and consistent virus isolations from dead bats [28] , [29] . Indeed , our in vitro data also indicate that bats should be more closely considered in future ecological and epidemiological investigations looking at BWAV , PGAV and NDV , as all of these viruses display at least a fundamental ability to infect cells derived from various bat species . Further , we cannot currently exclude that productive infection of bat cells may in fact be a feature of a wide range of other orthobunyavirus species as well , opening up the possibility that a broader host range than is currently appreciated might generally exist for orthobunyaviruses , and specifically that consideration of bats as a potential host for other orthobunyaviruses may also be warranted . While for MDCV no arthropod vector has yet been established , for KKV , until now only bedbugs have been identified as a potential vector species [27] . On this basis , the classification of KKV in the NDV group was particularly surprising , since orthobunyavirus transmission appears to be almost exclusively mosquito-borne , with the exception of the Tete virus group , which is mainly transmitted by ticks , and a few specific instances of culicoid fly vectored viruses [1] . No other example of an orthobunyavirus associated with infection of and transmission via bedbugs has been reported . Interestingly , we also observed that both KKV and MDCV were able to infect mosquito cells in vitro , and while these data only show that infection of mosquitos is fundamentally possible at a cellular level , when taken together with the close genetic relationship of these viruses to a number of well-documented mosquito-borne viruses , this clearly indicates that the possibility of mosquito-borne transmission should be considered in future field studies . Such studies will be particularly important given that evidence exists supporting the relevance of KKV for human infection . In particular , high levels of seroprevalence among Guano miners working in caves known to have infected bats and/or bedbugs have been reported [28] , [29] , and anecdotal reports suggest that working in such caves is associated with a mild generalized febrile illness in new workers [27] , [28] . The inclusion of MDCV and KKV in the NDV clade was also surprising given their geographical distribution , having been isolated exclusively from South America and Southeast Asia , respectively . The existence of such a highly genetically diverse clade , which spans three continents , suggests on the one hand that additional related but unrecognized virus species likely exist , and also that transmission between these non-contiguous geographical locations has most likely been facilitated by an unknown host species common to all of these virus groups . In particular , transmission of KKV and MDCV via migratory bird routes through various flyways would appear possible ( i . e . via the American Flyways<>Black Sea and Mediterranean flyways<>Asia Flyways<>American Flyways ) , and indeed infection of bird species has been shown to be possible for several other orthobunyaviruses , including Turlock virus and Mermet virus [49] , [50] . Alternatively , we must also consider that transmission between Africa and Asia may also have been facilitated by bats species endemic to these regions , some of which also populate broad geographical areas . In future the availability of comprehensive genome sequencing datasets , such as that determined in this study , will be important not only for molecular-based detection of virus infection ( i . e . in infected mosquito samples , acutely infected humans , etc . ) but will facilitate the development of recombinant antigen-based detection systems , which will be necessary for undertaking broader serological surveillance/screening efforts aimed at defining the geographic areas affected by these viruses , as well as estimating seroprevalence in animal species and/or larger segments of the human population to better define the public health impact of these viruses in the endemic areas . In addition , our study , which was restricted to only a small number of available isolates , also highlights the need for increased sample collection for these and other neglected tropical disease agents , and particularly the collection of human isolates , in order to develop a clearer picture of the actual extent of virus genetic diversity . Overall the data contained in this study have not only led to the genetic identification of two previously uncharacterized viruses , but in doing so , has considerably expanded our knowledge of virus diversity along the BWAV/PGAV and NDV genetic lineages . On this basis we have also presented arguments for a more refined and evidence based approach to the taxonomic classification of the viruses in these groups , something that is increasingly appreciated as being sorely needed within the Orthobunyavirus genus . Further , our in vitro data , informed by the genetic relationships established as part of our sequencing efforts , have identified the possibility of infection with these viruses in an expanded range of host and vector species . The availability of complete genetic information for these viruses , as well as a better understanding of their genetic relationships , will be instrumental in assisting future surveillance efforts aimed at determining the distribution and public health impact of these viruses , as well as efforts in identifying the contributions of various potential host and vector species . | Bunyavirus infections cause febrile illnesses of varying severity worldwide; however , despite their public health importance most remain relatively unstudied . In order to clarify the genetic relationships among African orthobunyaviruses associated with human infection , we have sequenced multiple strains of Bwamba ( BWAV ) , Pongola ( PGAV ) and Nyando virus ( NDV ) . Based on genetic analysis we showed that , while different BWAV and PGAV virus strains are closely related , NDV strains were highly variable and warrant classification as two distinct virus species . In addition , sequencing of the previously unclassified Mojuí dos Campos ( MDCV ) and Kaeng Khoi ( KKV ) viruses showed that both are closely related to NDV . This was unexpected considering that these viruses were isolated in South America and Southeast Asia , respectively , and are mainly associated with bats . Further , our experiments also indicated that BWAV and PGAV , as well as NDV , MDCV and KKV , are able to infect both bat and mosquito cell lines , suggesting that ecological studies focusing on these potential host and vector species are warranted . In the future , the availability of complete genetic information for these viruses , together with an understanding of their genetic relationships , will aid in better defining the distribution and public health impact of these viruses . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"organismal",
"evolution",
"comparative",
"sequence",
"analysis",
"medicine",
"and",
"health",
"sciences",
"viral",
"classification",
"emerging",
"viral",
"diseases",
"vector-borne",
"diseases",
"microbiology",
"viruses",
"microbial",
"evolution",
"molecular",
"biology",
"techniques",
"bunyaviruses",
"sequence",
"analysis",
"infectious",
"diseases",
"medical",
"microbiology",
"microbial",
"pathogens",
"arboviral",
"infections",
"molecular",
"biology",
"viral",
"evolution",
"arboviruses",
"high",
"throughput",
"sequencing",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"evolutionary",
"biology",
"organisms"
] | 2014 | Molecular Characterization of Human Pathogenic Bunyaviruses of the Nyando and Bwamba/Pongola Virus Groups Leads to the Genetic Identification of Mojuí dos Campos and Kaeng Khoi Virus |
Treponema pallidum subsp . pertenue ( TPE ) is the causative agent of yaws , a multistage disease endemic in tropical regions in Africa , Asia , Oceania , and South America . To date , seven TPE strains have been completely sequenced and analyzed including five TPE strains of human origin ( CDC-2 , CDC 2575 , Gauthier , Ghana-051 , and Samoa D ) and two TPE strains isolated from the baboons ( Fribourg-Blanc and LMNP-1 ) . This study revealed the complete genome sequences of two TPE strains , Kampung Dalan K363 and Sei Geringging K403 , isolated in 1990 from villages in the Pariaman region of Sumatra , Indonesia and compared these genome sequences with other known TPE genomes . The genomes were determined using the pooled segment genome sequencing method combined with the Illumina sequencing platform resulting in an average coverage depth of 1 , 021x and 644x for the TPE Kampung Dalan K363 and TPE Sei Geringging K403 genomes , respectively . Both Indonesian TPE strains were genetically related to each other and were more distantly related to other , previously characterized TPE strains . The modular character of several genes , including TP0136 and TP0858 gene orthologs , was identified by analysis of the corresponding sequences . To systematically detect genes potentially having a modular genetic structure , we performed a whole genome analysis-of-occurrence of direct or inverted repeats of 17 or more nucleotides in length . Besides in tpr genes , a frequent presence of repeats was found in the genetic regions spanning TP0126–TP0136 , TP0856–TP0858 , and TP0896 genes . Comparisons of genome sequences of TPE Kampung Dalan K363 and Sei Geringging K403 with other TPE strains revealed a modular structure of several genomic loci including the TP0136 , TP0856 , and TP0858 genes . Diversification of TPE genomes appears to be facilitated by intra-strain genome recombination events .
The infectious agent of yaws , Treponema pallidum subsp . pertenue ( TPE ) , causes chronic infections in children and young adults , which is characterized by skin lesions including nodules and ulcerations of the skin , which is later accompanied by joint , soft tissue , and bone manifestations ( reviewed in [1] ) . Unlike the syphilis treponemes , Treponema pallidum subsp . pallidum ( TPA ) , TPE and Treponema pallidum subsp . endemicum ( TEN , the causative agent of endemic syphilis ) are transmitted between individuals mostly through direct skin contact . However , possible sexual transmission has been reported for TEN , a treponeme highly related to TPE [2–4] . Only a limited number of TPE strains/isolates have been characterized to date , mainly as a result of the uncultivable character of TPE , low number of available laboratory strains , and a limited number of clinical isolates with sufficient numbers of treponemal DNA copies per sample . However , the recent study by Edmonson et al . [5] showed a successful long-time in vitro cultivation of syphilis treponemes that could be also potentially applied to TPE strains . This could result in an increase in the number of characterized TPE strains . So far , seven TPE strains have been completely sequenced , including five strains of human origin ( CDC-2 , CDC 2575 , Gauthier , Ghana-051 , and Samoa D ) [6 , 7] and two TPE strains ( Fribourg-Blanc and LMNP-1 ) isolated from a Guinea baboon ( Papio papio ) in West Africa [8] and an olive baboon ( Papio anubis ) from Tanzania [9] , respectively . In addition to these complete genomes , genomes of 6 other TPE isolates of human origin [10] and 7 from nonhuman primates have been sequenced to draft genome quality [9] . TPE strains have been shown to be highly similar to syphilis-causing strains of T . pallidum subsp . pallidum ( TPA ) [6] and to the TEN strain Bosnia A [11] . While there is an increasing understanding of genome structure and plasticity in TPA and TEN [12 , 13] , the genome characteristics of TPE remain largely unexplored . As a result , little is known about intra-strain recombinations that occur in TPE strains and their role in genome evolution and diversification . In this communication , we compared the complete genome sequences of two strains of TPE isolated in Indonesia to other available TPE whole genome sequences and identified regions resulting from intra-strain genome recombinations . While TPE genomes appear to be relatively conserved compared to the genomes of other uncultivable pathogenic treponemes , including TPA and TEN strains , genetic diversification of TPE genomes appears to be facilitated by intra-strain genome rearrangements .
TPE strains Kampung Dalan K363 and Sei Geringging K403 originated from the study of Noordhoek et al . [14] , where involved persons or parents of involved children gave informed consent for sample collection . No vertebrate animals were used in the study . Two TPE strains , Kampung Dalan K363 and Sei Geringging K403 , were used in this study . TPE Kampung Dalan K363 was isolated on January 5 , 1990 and TPE Sei Geringging K403 on May 14 , 1990 in villages in the Pariaman region of Sumatra , Indonesia . Both TPE strains were isolated from patients having skin lesions . Skin biopsies were first homogenized in PBS and intra-dermally inoculated into the shaved inguinal areas of Syrian Golden hamsters , which were later transported to the Netherlands [14] . Hamsters that developed skin lesions were sacrificed and the inguinal lymph nodes were homogenized in PBS and inoculated into the testes of New Zealand White rabbits . Several serial passages in rabbits were performed before samples were taken for isolation of treponemal DNA . TPE strains Kampung Dalan K363 and Sei Geringging K403 were provided as DNA samples by Dr . S . Bruisten ( Public Health Laboratory , Department of Infectious Diseases GGD Amsterdam ) who derived the DNA samples from crude treponemal lysates which were kindly donated by Dr . G . Noordhoek who collected these samples in Indonesia and processed them in the Netherlands [14] . The determination of the number of treponemal DNA copies per μl of samples was not performed . Total DNA of both samples was first amplified using multiple displacement amplification ( REPLI-g kit , QIAGEN , Valencia , CA , USA ) according to the manufacturer’s instructions . The amplified DNA was then diluted 50-times and used as a template for TPE whole genome amplification , which was performed with treponemal specific primers as described previously [6–8 , 11 , 15] . For amplification of individual amplicons ( n = 278 , S1 Table ) , PrimeSTAR GXL DNA Polymerase ( Takara Bio Inc . , Otsu , Japan ) was used . PCR products were generated using touchdown PCR under the following cycling conditions: initial denaturation at 94°C for 1 min; 8 cycles: 98°C for 10 s , 68°C for 15 s ( annealing temperature gradually reduced by 1°C/every cycle ) , and 68°C for 6 min; 35 cycles: 98°C for 10 s , 61°C for 15 s , and 68°C for 6 min ( 43 cycles in total ) ; followed by final extension at 68°C for 7 min . All overlapping PCR products were purified using a QIAquick PCR Purification Kit ( QIAGEN , Valencia , CA , USA ) and mixed in equimolar amounts . Each pool of PCR products ( n = 4 for each of TPE strains , S1 Table ) was then used for whole genome sequencing . Individual pools of both TPE samples were used for Illumina Nextera XT library preparation and subsequently sequenced on a MiSeq platform ( 2x300 bp ) at the sequencing facility of CEITEC ( Brno , Czech Republic ) . Resultant sequencing data were quality pre-processed using Trimmomatic ( v0 . 32 ) [16] with a sliding window length of 4 bp and a Phred quality threshold value equal to 17 . After pre-processing , sequencing reads shorter than 50 bp were removed . Sequencing results for individual pools are summarized in S2 Table . The Illumina sequencing reads corresponding to individual pools of both TPE samples were handled separately and de novo assembled using SeqMan NGen v4 . 1 . 0 software ( DNASTAR , Madison , WI , USA ) using default parameters . A total of 222 , 249 , 158 , and 265 contigs from the TPE Kampung Dalan K363 strain and 124 , 92 , 93 , and 87 contigs from the TPE Sei Geringging K403 strain were obtained for Pools 1–4 , respectively ( S2 Table ) . The resulting contigs obtained from both strains were then separately aligned to the TPE Samoa D genome ( CP002374 . 1 [6] ) using Lasergene software ( DNASTAR , Madison , WI , USA ) . In parallel , the Illumina sequencing reads corresponding to individual pools were mapped to the corresponding pool sequences ( S2 Table ) of the TPE Samoa D genome ( CP002374 . 1 [6] ) and both the de novo and reference-guided approaches were compared . All genome gaps and discrepancies were resolved using Sanger sequencing . Altogether , 11 and 6 genomic regions of the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains were Sanger sequenced , respectively . The consensual sequences for individual pools were then used to compile the complete genome sequences of both TPE strains . To determine the number of repetitions within the arp ( TP0433 ) gene , the repetitive sequences ( between coordinates 462430–463157 in the TPE Samoa D genome [6] ) were amplified and Sanger sequenced using the primers 32BrepF1 ( 5′-CGTTTGGTTTCCCCTTTGTC-3′ ) and 32BrepR1 ( 5′-GTGGGATGGCTGCTTCGTATG-3′ ) as described elsewhere [17] . Similarly , the repetitive sequences within the TP0470 gene ( Samoa D coordinates 498895–499200 ) were amplified and sequenced using the primers TPI34F4 ( 5′-GTCTTGTGCACATTATTCAAG-3′ ) and TPI34R5 ( 5′-CTTCGTGCAACATCGCTACG-3′ ) . The intra-strain variability , relative to the length of several G/C-homopolymeric tracts , was identified in both genomes and the prevailing length of G/C regions was used in the final genome sequences . Genes were annotated using Geneious software ( v5 . 6 . 5 ) [18] as described previously [11] and were tagged with TPEKDK363_ and TPESGK403_ prefixes . Locus tag numbering corresponded to tag numbering for the orthologous genes annotated in the TPE Samoa D genome ( CP002374 . 1 [6] ) . The TPE Samoa D genome contains one ( TPESAMD_0005a ) additionally annotated gene compared to the TPE Kampung Dalan K363 and TPE Sei Geringging K403 genomes . This gene is fused to TP0006 and is designated as TPEKDK363_0006 and TPESGK403_0006 , respectively . Four genes ( TPEKDK363_0146 , TPEKDK363_0520 , TPEKDK363_0812 , and TPEKDK363_0856a ) and eight genes ( TPESGK403_0126 , TPESGK403_0146 , TPESGK403_0312a , TPESGK403_0435a , TPESGK403_0520 , TPESGK403_0812 , TPESGK403_0865 , and TPESGK403_0924a ) were annotated as pseudogenes in the TPE Kampung Dalan K363 and TPE Sei Geringging K403 genomes , respectively . Since the tprK gene showed intra-strain variability , the corresponding nucleotide positions were denoted with an “N” in the complete genome sequences . For proteins with unpredicted functions , a 150 bp-gene size limit was applied . The identification of genetic heterogeneity was carried out as described by Strouhal et al . [7] . Briefly , individual Illumina reads were mapped to the final version of the genome sequence using SeqMan NGen ( v4 . 1 . 0 ) software with default parameters and requiring at least a 93% read identity relative to the reference genome . For the determination of the frequency of each nucleotide in every single genome position , the haploid Bayesian method was used for SNP calculation using the same software . Individual reads supporting a less frequent allele located at the 3’-terminus ( i . e . , five or less nucleotides ) were omitted . At least thirty independent reads from both directions were required . Nucleotide positions located within homopolymeric tracts ( defined as a stretch of six or more identical nucleotides ) were excluded from analysis . Chromosomal loci showing genetic heterogeneity within TPE genomes were defined as those containing more than 8% alternative reads in regions having a coverage depth greater than 100x . Candidate sites were then visually inspected using SeqMan NGen ( v4 . 1 . 0 ) software; the tprK ( TP0897 ) gene , showing intra-strain variability , was excluded from the analysis . Phylogenetic trees of the TPE strains were constructed from available whole genome sequences ( S3 Table ) including the Samoa D ( CP002374 . 1 [6] ) , CDC-2 ( CP002375 . 1 [6] ) , Gauthier ( CP002376 . 1 [6] ) , Fribourg-Blanc ( CP003902 . 1 [8] ) , Ghana-051 ( CP020365 . 1 [7] ) , CDC 2575 ( CP020366 . 1 [7] ) , and LMNP-1 ( CP021113 . 1 [9] ) strains . Moreover , 6 additional draft genomes of TPE strains isolated on Solomon Islands [10] were used to determine the phylogenetic relatedness of TPE Kampung Dalan K363 and Sei Geringging K403 strains . The genome of TEN Bosnia A strain ( CP007548 . 1 [11] ) was used as an outgroup . Whole genome alignment was constructed using SeqMan software ( DNASTAR , Madison , WI , USA ) and phylogenetic trees were constructed using the Maximum Likelihood method based on Tamura-Nei model [19] and with MEGA software [20] . Since there were chromosomal regions that included: ( 1 ) the tprD and tprK genes , ( 2 ) intergenic regions within both rrn operons , and ( 3 ) sequences in the arp and in the TP0470 genes , which are recombinant or repetitive in TPE strains ( S3 Table ) , these regions were excluded from the phylogenetic analyses . For analysis of the modular structure of the TP0136 , TP0856 , and TP0858 genes , additional available treponemal whole genome sequences were used including: TPA strains Nichols ( CP004010 . 2 [21] ) , SS14 ( CP004011 . 1 [21] ) , DAL-1 ( CP003115 . 1 [22] , Mexico A ( CP003064 . 1 [23] ) , Chicago ( CP001752 . 1 [24] ) , and Sea81-4 ( CP003679 . 1 [25] ) and T . paraluisleporidarum ecovar Cuniculus strain Cuniculi A ( CP002103 . 1 [26] ) . For each TPE genome sequence ( n = 9; S3 Table ) , the number of canonical k-mers of length 9–33 nt were determined using Jellyfish software ( v2 . 0 . 0 ) [27] . The number of unique k-mers saturated at a length of 17 nts and the 17-mers and longer k-mers were used for further evaluation . In order to determine their exact locations and their exact numbers , the detected k-mers were mapped to the TPE genomes using EMBOSS fuzznuc ( v6 . 6 . 0 ) [28] . Subsequently , k-mers were divided into two groups with the first group comprised of k-mers with exactly the same numbers in all tested TPE genomes and the second group comprised of k-mers with different numbers in at least one TPE genome . Localization of k-mers in the annotated genes of each TPE genome was carried out using BEDTools intersect ( v2 . 26 . 0 ) [29] . For each gene , the number and type of overlapping k-mers was determined using R ( v3 . 4 . 1 , packages rio v0 . 5 . 5 , dplyr v0 . 7 . 3 ) [30] . Similarly , for each k-mer , the number and type of overlapping genes was determined and k-mers with more than a single localization in at least one genome were extracted . The complete genome sequences from TPE Kampung Dalan K363 and TPE Sei Geringging K403 were deposited in the GenBank under accession number CP024088 . 1 and CP024089 . 1 , respectively .
Both TPE strains were sequenced using the pooled segment genomic sequencing ( PSGS ) protocol as previously described [6–8 , 11 , 15] . Illumina sequencing resulted in 7 , 545 , 122 paired reads and 1 , 241 , 564 , 236 total bases , with an average coverage depth of 1 , 021x for the TPE Kampung Dalan K363 genome , and 3 , 784 , 916 paired reads and 784 , 165 , 636 total bases , with an average coverage depth of 644x for the TPE Sei Geringging K403 genome . A total of 222 , 249 , 158 , and 265 contigs for each of the pools 1–4 of the TPE Kampung Dalan K363 strain and 124 , 92 , 93 , and 87 contigs for the 4 pools of the TPE Sei Geringging K403 strain were obtained by de novo assembly . Detailed characteristics of Illumina sequencing and de novo assembly are shown in S2 Table . The genome structures of both TPE strains were similar to other previously characterized TPE strains with no major chromosomal rearrangements . The summarized genomic features of TPE Kampung Dalan K363 and TPE Sei Geringging K403 were compared to the most closely related TPE Samoa D genome ( CP002374 . 1 [6] ) . Details are shown in Table 1 . The genomes of TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains differed in the number of repetitions within the arp ( TP0433 ) and TP0470 genes ( S3 Table ) . The TPE Kampung Dalan K363 strain contained 4 and 35 repetitions in the arp and TP0470 genes , respectively , while the TPE Sei Geringging K403 strain contained 2 and 28 repetitions within these genes , respectively . In both strains , the same repeat motif ( Type II ) within the arp gene was identified as previously shown in other TPE strains [31] . Both TPE genomes showed the same constitution of intergenic spacer regions within the rrn operons , i . e . , tRNA-Ile/tRNA-Ala pattern [13 , 32] ( S3 Table ) , and both contained the tprD2 allele in the tprD locus [33] ( S3 Table ) . Moreover , both TPE genomes differed in the sequences of the tprK gene variable regions [34–37] . The whole genome sequences of the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains were analyzed with respect to the occurrence of nucleotide diversity between both strains . As a result , the genomes differed in 38 single nucleotide positions ( S4 Table ) . In addition , both genomes differed in 18 nucleotide positions within the TP0858 gene . Moreover , there were differences in both genomes in the number of 2 nt-long ( TG ) and 9 nt-long ( TCCTCCCCC ) repetitive sequences between coordinates 390964–390969 and 1051995–1052003 ( according to the TPE Samoa D genome [6] ) , respectively . The genome of the TPE Kampung Dalan K363 strain contained 2 and 2 of these repetitive sequences , while the TPE Sei Geringging K403 genome contained 3 and 1 of these repetitions ( S4 Table ) , respectively . Since both TPE strains Kampung Dalan K363 and Sei Geringging K403 underwent serial passages in hamsters and rabbits after their isolation from human patients , the identified genetic differences resulted either from the cultivation experiments in animals or were already present during infection of humans . As revealed by our phylogenetic analyses , the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains clustered together and were more distantly related to other complete genomes of TPE strains of human or baboon origin ( Fig 1 ) . In contrast to the Indonesian strains used in this study , all other TPE strains originated from Africa with the exception of the TPE Samoa D strain , which was isolated in Western Samoa in 1953 ( S3 Table ) . Nevertheless , additional phylogenetic analysis including the recently published TPE draft genome sequences from 6 individuals from Solomon Islands [10] did not show clustering of TPE Kampung Dalan K363 and TPE Sei Geringging K403 with these strains ( S1 File ) . While all Solomon Island isolates [10] clustered together and were more closely related to TPE Samoa D , TPE Kampung Dalan K363 and TPE Sei Geringging K403 belonged to distinct cluster . The TPE Kampung Dalan K363 and TPE Sei Geringging K403 genomes were inspected for the presence of genetic intra-strain heterogeneity [38] . While the genome of the TPE Kampung Dalan K363 strain contained 3 intra-strain heterogeneous sites , the genome of the TPE Sei Geringging K403 strain harbored only a single such site ( S5 Table ) . The TPE Kampung Dalan K363 strain contained heterogeneous sites in genes TP0448 ( encoding uracil phosphoribosyltransferase ) , TP0488 ( coding for methyl-accepting chemotaxis protein ) , and TP1032 ( encoding hypothetical protein ) . In the TPE Sei Geringging K403 strain , a single heterogeneous site was found in the TP0363 gene , encoding chemotaxis protein CheA , which is a sensor histidine kinase . All four heterogeneous sites resulted in amino acid replacements in the corresponding proteins ( S5 Table ) . In both TPE genomes , intra-strain variability in the length of G/C-homopolymeric tracts was identified as previously shown in other treponemal genomes [7 , 39 , 40] . Based on the prevailing length of G/C regions in the final genome sequences , 16 out of 44 such regions were found to be different when comparing the TPE Kampung Dalan K363 and TPE Sei Geringging K403 genomes ( S6 Table ) . Therefore , four genes ( TPESGK403_0126 , TPESGK403_0312a , TPESGK403_0865 , and TPESGK403_0924a ) were annotated as pseudogenes in the TPE Sei Geringging K403 genome ( S6 Table ) . Although both Indonesian TPE strains were highly related to each other , a relatively long stretch of nucleotide differences in the TP0858 gene sequence of both analyzed strains suggests a potential recombination event . Compared to the TP0858 sequence of the TPE Sei Geringging K403 strain ( which was similar to the other TPE strains ) , the TP0858 sequence in the TPE Kampung Dalan K363 strain differed in 18 nucleotide positions ( coordinates 819–853 in the TP0858 gene of the TPE Samoa D [6] ) . Moreover , the nucleotide sequence present in the TP0858 gene of the TPE Kampung Dalan K363 strain ( i . e . , r5 sequence; see Fig 2 ) was found to be identical with the one found between coordinates 798–832 in the TP0856 gene ( TPE Samoa D gene coordinates ) . Interestingly , the same sequence ( i . e . , r5 sequence; see Fig 2 ) was detected also in the TP0858 gene of the TPA Sea 81–4 ( coordinates 819–853 according to the TPE Samoa D TP0858 gene ) . Analysis of additional treponemal genomes revealed that the TEN Bosnia A strain contained identical sequences in the above described regions of both the TP0856 and TP0858 genes ( i . e . , r6 sequence; see Fig 2 ) , even though these sequences in the TEN Bosnia A strain and the TPE Kampung Dalan K363 strain differed ( Fig 2 ) . Upstream of this sequence , between coordinates 768–809 in the TP0858 gene ( TPE Samoa D gene coordinates ) , there was a 42 nt-long DNA region ( i . e . , r8 and r4 sequences; see Fig 2 ) showing identical sequences within TPE and TPA/TEN strains , respectively , but different in the TPE and TPA/TEN comparison ( Fig 2 ) . In addition , the TPA Sea81-4 strain contained identical sequences ( i . e . , r3 sequence; see Fig 2 ) in both the TP0856 and TP0858 genes between coordinates 538–573 and 579–612 ( TPE Samoa D gene coordinates ) , respectively . The modular structure of the TP0856 and TP0858 genes comprising all completely sequenced treponemal strains is depicted in more detail in Fig 2 . Protein sequence analyses revealed that the repetitive modules found within both TP0856 and TP0858 genes ( e . g . , r3 in TPA Sea81-4 , r4 in TPA and TEN strains , r5 in TPE Kampung Dalan K363 and TPA Sea81-4 , and r6 in TEN Bosnia A ) used the same reading frame and therefore yielded the same amino acid sequence in both TP0856 and TP0858 proteins . Protein function analysis of TP0856 and TP0858 revealed presence of UPF0164 , an uncharacterized protein family found only among T . pallidum strains . Members of this protein family belong to the membrane beta barrel superfamily . No motifs were found within these genes using the Motif search ( https://www . genome . jp/tools/motif ) and Pfam , NCBI-CDD and PROSITE Profile databases . However , as described recently [41] , TP0856 and TP0858 proteins showed structural similarity to FadL , a long fatty acid transporter . In addition , the sequences of TP0856 and TP0858 were analyzed by I-TASSER server [42] to predict the protein structure . The analyses revealed that most of the variable sites ( i . e . , r1 , r2 , r4 , r5 , r6 , r8 and r9 ) of TP0856 and TP0858 represent coil sequences at the outer surface of β-barrels suggesting that these protein loci are exposed to the external milieu . The detailed overview of predicted structure for module sequences in TP0856 and TP0858 genes are shown in S7 Table . An alignment of both whole genome sequences of the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains revealed striking sequence differences in the TP0136 gene compared to other TPE strains . A detailed analysis of treponemal TP0136 gene orthologs identified a modular structure in the region between coordinates 158103–158250 ( coordinates according to the TPE Samoa D genome [6]; see Fig 3 ) . While in the TPE Samoa D and TPE Gauthier strains , the DNA region between coordinates 158196–158228 is represented by a 33-nt long sequence ( i . e . , r6 and r4 sequences; see Fig 3 ) , in TPE strains CDC-2 , CDC 2575 , Fribourg-Blanc , and LMNP-1 , the same region contains an additional copy of this 33-nt long sequence ( Fig 3 ) . In contrast , the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains contain within this region a duplicated segment from coordinates 158229–158250 ( i . e . , r5 and r2 sequences; see Fig 3 ) , followed by another duplicated sequence from positions 158128–158195 ( i . e . , r3 and r4 sequences; see Fig 3 ) . The sequence of the TP0136 gene from the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains thus resembles the TP0136 gene sequences found in TPA strains ( Fig 3 ) , where two 96 nt-long repetitions are present ( i . e . , comprising r1-r2-r3-r4 sequences ) . The modular structure of the TP0136 gene comprising all completely sequenced treponemal strains is depicted in more detail in Fig 3 . To systematically detect genes showing a modular genetic structure , a whole genome analysis of the presence of direct or inverted repeats of 17 or more nucleotides in length was performed . The length of 17 or more nucleotides was based on an analysis of identified sequentially unique k-mers present in TPE genomes . Starting with k-mers 9 nts in length , the number of different k-mers increases with the length of k-mers until it reaches a maximum at 11 nts and then decreasing ( Fig 4 ) . In k-mers 17 nt in length , the number of detected different k-mers remains stable and therefore this length of k-mers was selected for identification of positions and multiplicity of these k-mers in the TPE strain genomes . The results of position- and multiplicity-mapping of k-mers are summarized in Table 2 . Besides tpr genes ( tprCDEFGIJK ) , a frequent presence of repeats was found in the region spanning the TP0126–TP0136 genes , and in TP0856 , TP0858 , and TP0896 genes . Examples of other treponemal genes showing a different modular structure are presented in Fig 5 .
Two TPE strains isolated in 1990 from villages in the Pariaman region of Sumatra , Indonesia , were completely sequenced in this study using the pooled segment genomic sequencing ( PSGS ) approach . This approach allowed assembly and compilation of complete genome sequences without gaps or ambiguous nucleotide positions . The only exception was the variable regions within the tprK gene where consensus sequences were not determined due to intra-strain nucleotide sequence variability [34–37] . Both Indonesian TPE strains , i . e . , Kampung Dalan K363 and Sei Geringging K403 , were genetically related to each other and both strains were more distantly related to other , previously characterized TPE strains . Most of the available complete genome sequences of TPE strains originated in Africa except for the TPE Kampung Dalan K363 and Sei Geringging K403 strains that were isolated in Indonesia [14] and the TPE Samoa D strain that was isolated from the Samoan Islands in the central South Pacific , forming part of Polynesia [44] . This opens the question of whether TPE strains differ with respect to their geographical origin as shown by molecular typing studies of TPA [13] . A recent paper on TPE isolates sequenced from the Solomon Islands revealed 8 draft genome TPE sequences from 6 patients [10] showing that the Solomon Islands genome sequences represented a discrete TPE clade that was distinct from all previously sequenced TPE strains . Nevertheless , the phylogenetic analysis including also TPE strains from Solomon Islands [10] did not show clustering of TPE Kampung Dalan K363 and TPE Sei Geringging K403 with these strains ( S1 File ) despite their close geographical origin . However , the draft genome status of TPE Solomon Islands strains needs to be taken into account in the interpretation of the phylogeny shown in S1 File . Interestingly , the genetic features within the TP0858 gene of the TPE Kampung Dalan K363 strain , presented in this study , was similar to those found in all of the Solomon Islands isolates . This comprised a short sequence within the TP0858 gene that was conserved in all the Solomon Islands isolates suggesting that this sequence is representative of isolates from the South Pacific region [10] . Moreover , this sequence is a part of the reverse primer binding site within the TP0858 gene , which was the target of a PCR assay designed by Chi and colleagues [45] , which leads to false-negative PCR results on samples with this recombination [10 , 46] . A limited amount of genetic diversity within individual TPE strains was found in this study . Although it has been shown that the number of identified intra-strain heterogeneous sites correlates positively with the average depth of sequencing coverage , these genomes revealed just one and three such sites , although the average depth of sequencing coverage was well above 600x . Čejková et al . [38] proposed that the number of heterogeneous sites also reflects T . pallidum subspecies classification , where the majority of heterogeneous sites were found among TPA strains and not among TPE strains . This work appears to be consistent with this prediction same as the recently sequenced genomes of TPE strains Ghana-051 and CDC 2575 , which showed a relatively limited number of heterogeneous sites ( n = 13 , n = 5; respectively ) [7] . As shown in previous studies , all the alternative alleles identified in this study encoded non-synonymous amino acid replacements , suggesting an adaptive character for this genetic variability [38] . In general , pathogenic treponemes comprising TPA , TPE , and TEN strains or isolates , lack mobile genetic elements including pathogenicity islands , prophages , and plasmids [12 , 13] . It was long believed that the lack of mobile genetic elements is related to the absence of genetic recombination both within and between treponemal strains . Yet , due to recent accumulation of genetic data , treponemes appear to recombine genetic material both within and between genomes . One of the first observations describing intra-strain genetic recombinations ( recombinations within genomes ) came from studies on the tprK gene , which shows increasing variability during the course of human infection [35 , 36] . The underlying mechanism here is gene conversion using sequences from the flanking regions of tprD [34] . Later , Gray et al . [47] demonstrated that intra-genomic recombination has played a significant role in the evolution of tpr genes ( tprCDIGJK ) , which have evolved through gene duplication and gene conversion . As an example , the occurrence of tprD and tprD2 , both found within TPA clusters ( Nichols-like and SS14-like ) and within TPE strains [33] , suggests a gene conversion mechanism in copying the tprC allele ( that is identical with the tprD allele ) to the tprD locus [33 , 47] . Similarly , the TP0136 locus of the Treponema paraluisleporidarum ecovar Cuniculus strain Cuniculi A contains an almost identical copy of the TP0133 gene sequence , suggesting the same mechanism as for tprD/tprD2 allele alternation [13 , 26] . A similar situation was recently found in TPE samples isolated on Lihir Island , Papua New Guinea [48] , where the TP0136 allele also had an intriguing sequence identity to the TP0133 gene . The authors proposed a possible interstrain recombination between treponemal species , however , intra-strain recombination by copying the TP0133 allele to the TP0136 locus appears to be more plausible . As shown in Čejková et al . [32] , two rRNA ( rrn ) operons occurred in two different rrn spacer patterns ( i . e . , tRNA-Ala/tRNA-Ile and tRNA-Ile/tRNA-Ala patterns ) and these variants were found independently of species/subspecies classification , time , and geographical source of the treponemal strains , suggesting the existence of reciprocal recombination in treponemes . Besides intra-genomic recombination events , traces of interstrain ( intergenomic ) recombination between TPA , TPE , and TEN strains have been proposed for several genetic loci [3 , 11 , 23] . Comparisons of genome sequences of the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains as well as analysis of other TPE strains revealed a modular structure for at least three gene loci including TP0136 , TP0856 , and TP0858 , suggesting that the recombination within treponemal genomes can result in substantial changes in gene and protein sequences . Further systematic analyses revealed additional gene loci with a modular genetic structure that differ in certain strain ( s ) compared to others , these genes included TP0126 , TP0126b , TP0126c , TP0127b , TP0128 , TP0130 , TP0898 , and tprCDFI ( TP0117 , TP0131 , TP0316 , and TP0620 ) , indicating that this mechanism , which enables genetic diversification , is quite common in treponemal genomes . Moreover , there were additional genes identified that have direct or inverted repeats ( summarized in Table 2 ) and thus have the potential for genetic reshuffling . The analysis of these genes revealed that these loci were limited to specific and relatively short genomic regions . In addition , these regions were often found in paralogous gene families including tpr genes ( tprCDEFGIJK ) , the paralogous family of TP0133 , TP0134 , TP0136 , and TP0462 genes , and the paralogous family of TP0548 , TP0856 , TP0858 , TP0859 , and TP0865 genes . As a consequence of the inherent variability of these paralogous families , restriction fragment length polymorphism ( RFLP ) analysis of the tprE ( TP0313 ) , tprG ( TP0317 ) , and tprJ ( TP0621 ) genes became a part of the CDC-typing scheme that determines , in addition to the tprEGJ RFLP pattern , a number of 60-bp tandem repeats within arp ( TP0433 ) [49] . Interestingly , members of two additional paralogous families including TP0136 and TP0548 are targets of sequencing-based molecular typing [50–55] . Similar to the TP0136 protein , TP0856 and TP0858 are predicted lipoproteins [43] . Structure similarity of TP0856 and TP0858 proteins to FadL , a long chain fatty acid transporter , was recently published [41] and both proteins are members of a FadL-like family ( TP0548 , TP0856 , TP0858 , TP0859 , TP0865 ) found in T . pallidum . Moreover , most of the variable sites ( i . e . , r1 , r2 , r4 , r5 , r6 , r8 and r9 ) of TP0856 and TP0858 were located in loops suggesting that these protein loci are exposed to the external milieu . The TP0136 gene has been shown to have heterogeneous sequences among T . pallidum strains [56 , 57] . Moreover , the TP0136 lipoprotein was demonstrated to be exposed on the surface of the bacterial outer membrane and was shown to bind to the extracellular matrix glycoproteins fibronectin and laminin [56] . Immunization with recombinant TP0136 delayed ulceration in experimentally infected rabbits but did not prevent infection or the formation of skin lesions [56] . The NH2-terminus of the TP0136 protein comprises a region with a modular structure overlapping the major fibronectin binding activity domain [57] . The modular structure was identified within the two 96 nt-long repetitions that are present in TPA strains and in the TEN Bosnia A strain . Interestingly , the sequence of the TP0136 gene in the TPE Kampung Dalan K363 and TPE Sei Geringging K403 strains resembled the TP0136 gene sequences found in TPA strains representing a new molecular type in the yaws MLST typing scheme [48] . Moreover , the specific insertion in TP0136 in the DAL-1 genome [22] was predicted to contain donor sequences for the tprK gene of T . pallidum [58] . In the case of TprC protein , one of the predicted antigenic epitopes on the 3D predicted structure , E3 ( residues 575–583 ) , partially overlaps with a recombinant region [59 , 60] . Our findings are therefore consistent with relatively frequent genetic recombinations operating at certain treponemal loci and these recombinations likely result in novel amino acid sequences exposed to the external milieu . In summary , although TPE genomes appear to be the most conserved genomes of the uncultivable pathogenic treponemes [12 , 13] , diversification of TPE genomes appears to be facilitated by intra-strain genome recombination events and rearrangements . Analysis of additional genomes will likely reveal more potential recombinations in the future . | Treponema pallidum subsp . pertenue ( TPE ) is the causative agent of yaws , a multi-stage disease that is endemic in tropical regions of Africa , Asia , Oceania , and South America . TPE belongs to the pathogenic treponemes and causes several human and animal infections . Whole genome sequences of two TPE strains isolated from patients in Indonesia were determined in this study . While both strains were highly related to other TPE strains isolated from humans and baboons , detailed genetic analyses revealed a modular character of several genes and genomic regions . While TPE genomes appear to be the most conserved genomes of uncultivable pathogenic treponemes , diversification of TPE genomes appears to be facilitated by intra-strain genome recombination events . In addition to genes with an identified modular structure , we identified additional genes that have direct or inverted repeats and thus have the potential for genetic reshuffling . | [
"Abstract",
"Introduction",
"Material",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"sequence",
"assembly",
"tools",
"membrane",
"proteins",
"outer",
"membrane",
"proteins",
"genome",
"analysis",
"sequence",
"motif",
"analysis",
"cellular",
"structures",
"and",
"organelles",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"bioinformatics",
"medical",
"microbiology",
"microbial",
"pathogens",
"comparative",
"genomics",
"cell",
"membranes",
"genetic",
"loci",
"treponema",
"pallidum",
"dna",
"sequence",
"analysis",
"cell",
"biology",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology"
] | 2018 | Complete genome sequences of two strains of Treponema pallidum subsp. pertenue from Indonesia: Modular structure of several treponemal genes |
Mycobacterium tuberculosis ( Mtb ) impairs dendritic cell ( DC ) functions and induces suboptimal antigen-specific CD4 T cell immune responses that are poorly protective . Mucosal T-helper cells producing IFN-γ ( Th1 ) and IL-17 ( Th17 ) are important for protecting against tuberculosis ( TB ) , but the mechanisms by which DCs generate antigen-specific T-helper responses during Mtb infection are not well defined . We previously reported that Mtb impairs CD40 expression on DCs and restricts Th1 and Th17 responses . We now demonstrate that CD40-dependent costimulation is required to generate IL-17 responses to Mtb . CD40-deficient DCs were unable to induce antigen-specific IL-17 responses after Mtb infection despite the production of Th17-polarizing innate cytokines . Disrupting the interaction between CD40 on DCs and its ligand CD40L on antigen-specific CD4 T cells , genetically or via antibody blockade , significantly reduced antigen-specific IL-17 responses . Importantly , engaging CD40 on DCs with a multimeric CD40 agonist ( CD40LT ) enhanced antigen-specific IL-17 generation in ex vivo DC-T cell co-culture assays . Further , intratracheal instillation of Mtb-infected DCs treated with CD40LT significantly augmented antigen-specific Th17 responses in vivo in the lungs and lung-draining lymph nodes of mice . Finally , we show that boosting CD40-CD40L interactions promoted balanced Th1/Th17 responses in a setting of mucosal DC transfer , and conferred enhanced control of lung bacterial burdens following aerosol challenge with Mtb . Our results demonstrate that CD40 costimulation by DCs plays an important role in generating antigen-specific Th17 cells and targeting the CD40-CD40L pathway represents a novel strategy to improve adaptive immunity to TB .
Critical to the success of Mycobacterium tuberculosis ( Mtb ) as a pathogen is its ability to manipulate host innate and adaptive immune responses to its benefit . Despite the development of antigen-specific T cell responses following infection , Mtb is able to persist within the host , indicating that Mtb-specific T cell immunity is suboptimal and ineffective at eliminating the pathogen [1 , 2] . Indeed , several studies have shown that mice infected with Mtb exhibit delayed initiation of antigen-specific CD4 T cell responses , which is preceded by delayed migration of Mtb-containing dendritic cells ( DCs ) from the lung to draining lymph nodes [3–5] . Moreover , although IFN-γ and T-helper 1 ( Th1 ) responses are important for controlling infection , they are not sufficient to eradicate bacteria and do not protect against developing tuberculosis ( TB ) [6–8] . Recently , IL-17 and Th17 responses have emerged as important for protective immunity to TB [9–16] . Studies in mice suggest that early induction of IL-17 in the lung promotes control of mycobacterial growth , and balanced Th1/Th17 responses in the lung have been reported to be more effective [17–19] . We previously reported that an avirulent hip1 ( Hydrolase important for pathogenesis 1; Rv2224c ) mutant Mtb strain induced significantly higher IL-17 and IFN-γ responses compared to infection with wild type Mtb due to enhanced functions of infected DCs [20 , 21] . Together , these studies suggest that wild type Mtb subverts DCs to prevent optimal T-helper responses and that augmenting DC functions during infection may be beneficial for improving protective immunity . While several studies have reported that Mtb manipulation of DC functions leads to suboptimal Th1 responses [21–24] , we lack insights into Th17 generation during Mtb infection . To gain insight into host pathways involved in generating Th17 responses during Mtb infection , we sought to define the molecular mechanisms underlying Th17 responses following Mtb infection of DCs . As the primary antigen-presenting cells in the immune system , DCs are instrumental in shaping adaptive immunity and determining the types of antigen-specific T-helper subsets that are generated in response to infection . Upon phagocytosis of the pathogen , DCs present pathogen-derived antigens to naïve CD4 T cells , provide critical costimulatory signals and produce cytokines; these signals initiate antigen-specific T-helper cell activation and polarization towards specific subsets [25–27] . However , beyond the role of cytokines such as IL-1β , IL-6 , and IL-23 in polarizing and committing antigen-specific CD4 T cells towards a Th17 phenotype , the DC-T cell interactions underlying Th17 polarization during Mtb infection are poorly defined . We previously showed that eliminating Hip1-dependent immune evasion mechanisms in Mtb enhanced the capacity of DCs to induce Th17 responses and was accompanied by significantly higher expression of the costimulatory molecule , CD40 , on infected DCs [21] . Because costimulation of naïve T cells in the context of cognate interactions between DCs and T cells is critical for optimal activation and differentiation of antigen-specific T cells , these data suggested that impaired CD40-dependent costimulation during wild type Mtb infection may lead to suboptimal Th17 responses in TB . CD40 has previously been implicated in generating Th1 responses during Mtb infection [28] , but its role in the polarization of the Th17 subset during infection is not defined . We therefore sought to investigate the contribution of the CD40 costimulatory pathway in Th17 generation during Mtb infection and determine the effects of augmenting CD40 costimulation on bacterial control . In this study , we show that CD40 expression on DCs is required for the generation of IL-17 responses to Mtb infection and that interaction between CD40 on DCs and CD40L on CD4 T cells is critical for generating antigen-specific IL-17 responses . Importantly , we found that engaging CD40 on DCs via crosslinking with a multimeric CD40 agonist reagent ( CD40LT ) significantly enhanced antigen-specific IL-17 responses to Mtb . Further , intratracheal instillation of Mtb-infected DCs treated with CD40LT led to significant enhancement of antigen-specific Th17 responses in the lungs and mediastinal lymph nodes ( MLN ) of mice , showing that engaging the CD40-CD40L pathway can overcome suboptimal Th17 responses to Mtb in vivo . Finally , we show that CD40 engagement in the setting of a DC transfer model enhances control of Mtb following aerosol challenge . Our results demonstrate that the CD40-CD40L pathway is critical for generating IL-17 responses and that targeting this costimulatory pathway represents a novel strategy to potentially improve protection against TB .
To test whether CD40 expression is required for differentiation of naïve CD4 T cells into IL-17-producing cells in response to Mtb infection , we used DC-T cell co-culture assays as previously described [21] . We infected bone marrow derived DCs from wild type C57BL/6 mice ( B6 ) or from CD40-/- mice for 24 hours , followed by co-culture with purified naïve TCR-transgenic ( Tg ) CD4 T cells isolated from OT-II mice in the presence of OVA323–339 peptide ( Fig 1A ) . Supernatants were harvested 72 hours after co-culture and assayed for IL-17 , IL-2 and IFN-γ by ELISA . Mtb-infected DCs from B6 mice induced increasing levels of IL-17 , IL-2 and IFN-γ cytokines with increasing concentrations of peptide . In contrast , CD40-/- DCs were significantly impaired in their ability to induce IL-17-producing cells in response to Mtb , but retained the capacity to induce IFN-γ and IL-2 ( Fig 1A ) . These data indicate that CD40 is specifically required to generate antigen-specific IL-17 responses . To assess whether the defect in IL-17 production was specific for CD40 deficiency , we examined the contribution of the costimulatory molecules CD80 and CD86 , which are known to be essential for IL-2 production and are required for optimal T cell proliferation [29 , 30] . While DCs that were doubly-deficient in CD80 and CD86 were severely impaired in IL-2 production , their ability to induce antigen-specific IL-17 responses were comparable to DCs from B6 mice ( Fig 1B ) and did not exhibit the defective IL-17 responses observed in CD40-/- DCs . These data indicate that CD40-dependent costimulation plays an essential and specific role in the generation of IL-17 responses to Mtb . Cytokines produced by infected DCs are known to be critical for polarizing antigen-specific CD4 T cell subsets [17 , 31] . Since IL-6 , IL-1β , and TGF-β have been shown to induce Th17 polarization , we sought to assess whether defective IL-17 responses seen in Mtb-infected CD40-/- DCs was due to defects in their ability to produce innate cytokines following Mtb infection . However , levels of IL-6 , IL-1β , and IL-12 produced by DCs from CD40-/- mice were comparable to the levels seen in DCs from B6 mice ( Fig 1C ) and bioactive TGF-β was undetectable in all culture conditions . Thus , the inability of CD40-/- DCs to induce IL-17 responses is not due to impaired innate cytokine responses , suggesting that interaction of CD40 expressed on DCs with its ligand , CD40L ( CD154 ) , may be necessary for production of IL-17 by CD4 T cells following Mtb infection . CD40L is expressed on antigen-activated T cells and binding of CD40 with CD40L provides accessory costimulatory signals that are necessary for optimal activation and differentiation of antigen-specific T cells . In order to determine whether interaction of CD40 with CD40L was required for IL-17 generation , we carried out co-culture assays using antigen-specific CD4 T cells isolated from OT-II mice crossed to mice lacking CD40L ( CD40lg-/- x OT-II ) . This allowed us to test whether CD40L on T cells was required for IL-17 generation in a setting where CD40 expression on DCs remained intact . We found that CD40lg-/- CD4 T cells were attenuated in their ability to generate IL-17 responses after co-culture with Mtb-infected DCs ( Fig 2A ) , concordant with the defective IL-17 response seen with CD40-/- DCs ( Fig 1 ) . Interestingly , CD40lg-/- T cells also displayed attenuated IFN-γ and IL-2 responses ( S1 Fig ) , which suggests that lack of CD40L leads to broader defects in T cell responses compared to the absence of CD40 . These results show that both CD40 and CD40L are required for optimal IL-17 generation . To further extend our genetic knockouts studies , we carried out co-culture assays in which we blocked CD40-CD40L interactions using saturating doses of a non-agonistic , anti-CD40L monoclonal antibody ( clone MR1 ) . This antibody has been shown to successfully block CD40-CD40L interactions in vitro ( S2 Fig ) and in vivo [32] . Blockade of CD40-CD40L interaction between Mtb-infected DCs and CD40L-replete antigen-specific CD4 T cells significantly reduced antigen-specific IL-17 responses ( Fig 2B ) . Together , these data show that interaction between CD40 and CD40L is critical for production of IL-17 by CD4 T cells during Mtb infection . The requirement for the CD40-CD40L pathway in IL-17 generation suggested that boosting interactions between CD40 and CD40L could serve as a tool to augment IL-17 responses . To exogenously engage CD40 on Mtb-infected DCs , we utilized a multimeric CD40 agonist in which two trimeric CD40L constructs are artificially linked ( CD40L trimers; CD40LT ) . The CD40LT reagent effectively aggregates and activates CD40 independently of T cells . Addition of CD40LT to Mtb-infected B6 DCs produced enhanced levels of IL-12 ( Fig 3A ) , consistent with previous reports showing IL-12 induction after CD40 engagement [33] . Importantly , treatment with CD40LT significantly enhanced production of IL-6 and IL-23 , which are key cytokines for Th17 polarization and expansion ( Fig 3A ) . IL-1β , which also promotes Th17 differentiation in combination with IL-6 and IL-23 , was not altered by treatment with CD40LT ( Fig 3A ) . Moreover , CD40LT-treated Mtb-infected DCs induced significantly higher levels of IL-17 from co-cultured ESAT-6 TCR-Tg CD4 T cells compared to Mtb-infected DCs lacking CD40 engagement ( Fig 3B ) . In contrast , CD40LT-treatment did not alter production of IFN-γ from antigen-specific CD4 T cells in vitro ( Fig 3B ) . These data show that CD40 engagement augments antigen-specific IL-17 generation . Costimulatory signals synergize with antigen-specific signals downstream of T cell receptor ( TCR ) ligation to promote full activation of T cells . Absence of signaling through the CD80/86-CD28 costimulatory pathway , for example , results in suboptimal T cell activation and anergic responses [29 , 30 , 34 , 35] . CD28 signaling is thought to function by lowering the T cell activation threshold , thus facilitating optimal T cell activation and IL-2 production . To investigate whether CD40 engagement on DCs similarly impacts the activation threshold of Mtb-specific T cells and whether this , in turn , influences IL-17 production , Mtb-infected DCs were either treated with CD40LT or left untreated , pulsed with increasing concentrations of ESAT-61–20 peptide , and co-cultured with ESAT-6 TCR-Tg CD4 T cells . We found that CD40LT-treated DCs displayed an enhanced capacity to induce IL-17 responses at all antigen doses compared to untreated conditions ( Fig 3C ) . The ability of Mtb-infected DCs to induce IL-17 at lower concentrations of peptide after CD40LT treatment suggests that signals induced by CD40 engagement lowers the threshold for antigen-specific production of IL-17 . Thus , CD40-dependent costimulation may serve to overcome suboptimal generation of IL-17 responses elicited early in Mtb infection when antigen levels are low . In order to dissect the relative contributions of Th17-polarizing cytokines and CD40-CD40L interaction on IL-17 responses , Mtb-infected DCs were treated with or without CD40LT and then co-cultured with ESAT-6 TCR-Tg CD4 T cells in the presence of the MR1 CD40L-blocking antibody as described in Fig 2B . Interestingly , antibody blockade of CD40-CD40L interaction significantly decreased antigen-specific IL-17 responses even in the presence of CD40LT ( Fig 3D ) . These data suggest that exogenous engagement of CD40 on DCs that results in enhanced production of Th17-polarizing cytokines is not sufficient for generating antigen-specific IL-17 responses in a setting where CD40 cannot interact with CD40L on antigen-specific CD4 T cells . Induction of early IL-17 responses on mucosal surfaces of the lung is thought to be important for immunity to Mtb and inducing balanced Mtb-specific Th1/Th17 responses may enhance protective immunity . To determine whether CD40 engagement on DCs can enhance the induction of Mtb-specific lung Th17 responses in vivo , we utilized a mucosal transfer approach via intratracheal instillation of DCs . This approach allows for targeted manipulation of Mtb-infected DCs without potential confounding from off-target effects such as CD40 engagement of alveolar macrophages . Transferred DCs have been shown to prime adoptively transferred Mtb-antigen-specific T cells in lymph nodes and lungs of mice by 3 days post-intratracheal instillation [23] . We adoptively transferred naïve CD45 . 2+ ESAT-6 TCR-Tg CD4 T cells into CD45 . 1+ congenic hosts . The next day , we transferred DCs infected with Mtb in the presence or absence of CD40LT by intratracheal instillation ( Fig 4A ) . At 6 and 12 days after DC transfer , we assessed Th17 responses in the lungs and MLN by determining the expression of IL-17 and RORγt in CD45 . 2+ ESAT-6-specific CD4 T cells by intracellular cytokine staining ( ICS ) and flow cytometry . Engaging CD40 on Mtb-infected DCs using CD40LT enhanced the frequency of ESAT-6-specific RORγt+IL-17+ T cells in the lungs ( Fig 4B ) and MLN ( Fig 4C ) . Notably , the majority of IL-17+ cells expressed RORγt , the transcription factor that determines Th17 lineage commitment [36] , indicating that CD40LT-treated Mtb-infected DCs polarized CD4 T cells into Th17 cells . To determine Th1 and Th17 responses in the lungs and MLN of mice at 6 and 12 days after intratracheal instillation of DCs , we assessed IFN-γ and IL-17 production in CD45 . 2+ ESAT-6 TCR-Tg T cells by flow cytometry ( Fig 5A ) . Transfer of Mtb-infected DCs that were treated with CD40LT resulted in a greater expansion of ESAT-6 TCR-Tg CD4 T cells compared to Mtb DCs that did not receive exogenous CD40LT , and was comparable to the expansion of ESAT-6 TCR-Tg CD4 T cells in response to hip1 mutant Mtb-infected DCs ( Fig 5B ) . Moreover , transfer of CD40LT-treated , Mtb-infected DCs significantly enhanced the frequencies of antigen-specific Th17 cells in lungs and MLN compared to Mtb-infected DCs alone and was comparable to the Th17 frequencies elicited by hip1 mutant Mtb-infected DCs ( Fig 5C ) . We also observed higher frequencies of antigen-specific IFN-γ+ CD4 T cells in the lung , but not MLN , on day 6 post-intratracheal transfer of either Mtb-infected CD40LT-treated DCs or hip1 mutant Mtb-infected DCs compared to their untreated counterpart ( Fig 5D ) . 12 days after intratracheal instillation of DCs , CD45 . 2+ ESAT-6-specific IFN-γ responses in the lungs were comparable , suggesting that DCs that did not receive CD40LT were delayed in inducing Th1 responses relative to Mtb-infected CD40LT-treated and hip1 mutant Mtb-infected DCs . Interestingly , antigen-specific CD4 T cells producing IL-17 and IFN-γ were mutually exclusive populations and double producing cells were not detected . These data demonstrate that engagement of the CD40 pathway can overcome deficits in Th17 generation during Mtb infection and leads to enhanced antigen-specific Th1 and Th17 responses in vivo . DCs loaded with Mtb antigens have been previously shown to confer better anti-mycobacterial immunity than BCG ( Bacillus Calmette-Guérin ) vaccination in mouse models [37 , 38] . Therefore , DC-based vaccination provides a useful model to study the impact of boosting CD40-engagement on priming of antigen-specific T cell pools and on the control of Mtb infection in vivo . We exposed DCs to heat-killed ( HK ) Mtb followed by treatment with CD40LT . DCs stimulated with HK Mtb and CD40LT were equivalent to ex vivo assays using live Mtb ( S3 Fig ) . Comparison groups included transfer of uninfected DCs , DCs stimulated with HK wild type Mtb or with HK hip1 mutant Mtb . Upon transfer of antigen-loaded DCs into mouse lungs by intratracheal instillation , we assessed immune responses generated by transferred DCs by measuring the activation of endogenous CD4 T cell responses and frequencies of Th17 and Th1 cells in the lungs of mice 6 and 12 days after DC transfer . 15 days after DC transfer , we challenged mice with low-dose aerosolized Mtb . At 5 weeks post-challenge ( day 50 ) , we determined lung bacterial burden and Mtb-specific Th1 and Th17 responses ( Fig 6A ) . CD40LT treatment induced significantly higher frequencies of activated CD44+ CD4 T cells ( Fig 6B ) and higher frequencies of lung IL-17+ CD4 T cells 6 and 12 days after DC transfer ( Fig 6C ) . IFN-γ+ CD4 T cell frequencies were higher on day 6 in mice receiving CD40LT-treated DCs compared to untreated Mtb-DCs , but were comparable by day 12 . As expected , transfer of hip1 mutant Mtb-stimulated DCs induced robust Th17 and Th1 responses in the lungs of mice on day 6 and day 12 post-DC transfer . Following aerosol challenge with low dose Mtb 15 days after intratracheal transfer of DCs , we assessed bacterial burden in the lungs of mice 5 weeks after challenge by plating for CFU . As shown in Fig 6D , transfer of DCs stimulated with HK Mtb resulted in significant reductions in lung bacterial burden at day 50 compared to transfer of DCs alone . Interestingly , CD40LT treatment reduced bacterial burden even further , showing that boosting CD40-CD40L interactions could overcome pathogen-mediated impairment of CD40 costimulation and promote enhanced anti-mycobacterial immunity . Notably , transfer of DCs exposed to hip1 mutant Mtb also showed comparable reductions in bacterial burden . These results are consistent with our previous report showing that hip1 mutant Mtb inherently induces superior DC responses compared to wild type Mtb , i . e . , significantly higher induction of Th1- and Th17-polarizing cytokines , higher expression of CD40 , enhanced antigen presentation and balanced Th1/Th17 responses [21] . To assess Mtb-specific Th1 and Th17 responses in the lungs of mice post-challenge , we stimulated lung cells ex vivo with Mtb whole cell lysate ( WCL ) and determined IFN-γ+ and IL-17+ CD4 T cell frequencies by flow cytometry . We found significantly enhanced Th17 responses in mice that intratracheally received CD40LT-treated DCs or hip1 mutant Mtb-stimulated DCs compared to those that received Mtb-stimulated DCs . However , there was no discernible difference between the groups in terms of lung CD4 IFN-γ responses ( Fig 6E ) . Importantly , lung IL-17 responses inversely correlated with bacterial burden , while there was no significant correlation between IFN-γ responses and lung bacterial burden ( Fig 6F ) . Our data show that we can improve protection against Mtb challenge by overcoming Mtb-mediated impairments in CD40 costimulation .
The findings in this study identify the CD40-CD40L pathway as a critical mechanism for the generation of antigen-specific Th17 responses and highlight the importance of DC-T cell crosstalk in immunity to Mtb infection . Importantly , we provide insights into improving adaptive immunity to TB by augmenting the functions of DCs and show that exogenously engaging CD40 on DCs significantly enhances control of Mtb burden in the lungs of infected mice . Costimulatory signals provided by antigen presenting cells such as macrophages and DCs are critical for full activation of naïve antigen-specific CD4 T cells and promote their rapid expansion into cytokine-producing effector cells , which exert their antimicrobial functions at the site of infection . While differentiation of activated CD4 T cells into IFN-γ+ Th1 subsets is relatively well understood , the molecular mechanisms underlying the generation of Th17 cells , particularly during Mtb infection , are less clear . Moreover , the mechanisms by which Mtb induces delayed , suboptimal T cell immunity , which enables the pathogen to successfully evade adaptive immunity and persist within the host , remain poorly understood . Several studies , including our own , have shown that Mtb impairs antigen presentation in infected DCs and dampens production of Th17-polarizing cytokines , such as IL-6 , IL-23 and IL-1β [21 , 22 , 24 , 39–41] . However , very little was known about the role of costimulatory pathways in driving Th17 development in TB prior to our study . Our work shows that innate cytokines are important for the generation of IL-17 responses ( Figs 1 and 3 ) and is consistent with other studies showing a critical role for CD40-dependent IL-6 and IL-23 in the induction and expansion of Th17 cells [42–44] . Interestingly , our results show that blocking CD40-CD40L interaction with MR1 attenuates IL-17 responses to Mtb-infected DCs despite treatment of DCs with CD40LT , which suggests that optimal induction of IL-17 to Mtb infection requires CD40-CD40L interaction ( Fig 3D ) . However , studies have shown that exogenous addition of supraphysiological levels of Th17-polarizing cytokines can drive CD40-/- DCs to induce IL-17 [42] . Our data suggest that costimulatory interactions between Mtb-infected DCs and T cells are required for optimal generation of IL-17 responses . In addition , localization of CD4 T cells in close proximity to infected DCs is likely to be an important determinant of the type of antigen-specific CD4 T cells mobilized after infection . Recent work has demonstrated that uninfected MLN-resident DCs acquire antigen from infected lung DCs and can prime Mtb-specific CD4 T cells to produce IFN-γ [23] . It is possible that while MLN-resident DCs acquire antigen , their maturation status and costimulatory capacity may be suboptimal without Mtb infection and , thus , not amenable to generating CD4 T cell responses beyond IFN-γ . Moreover , within the Th1 subset , studies have shown distinct IFN-γ-producing CD4 T cells in the vasculature and parenchyma of Mtb-infected mice [45] . However , localization of Th17 cells within lung compartments and the role of lung-specific DC subsets in driving the polarization of Th1 and Th17 during Mtb infection are poorly understood . Our study uses bone marrow derived DCs and therefore the extent to which our experiments model in vivo-generated lung DCs needs to be investigated further . Overall , our data showing that CD40-CD40L interaction is required for optimal Th17 generation in response to Mtb and that boosting CD40-CD40L interactions augments Th1 and Th17 responses suggests that restriction of costimulatory pathways is an important virulence mechanism used by Mtb for inducing suboptimal T-helper responses that benefits the pathogen . Our finding that exogenous induction of CD40-mediated costimulation , via CD40LT treatment , is able to elicit IL-17 production at lower concentrations of peptide stimulation ( Fig 3 ) than by Mtb-infected DCs alone leads to an interesting speculation . In early stages of Mtb infection , low levels of antigen in the lung , combined with impaired CD40 induction on Mtb-infected DCs , likely results in suboptimal costimulation of naïve CD4 T cells and therefore suboptimal and delayed induction of Th17 responses . However , engaging the CD40-CD40L pathway and promoting interactions between these two molecules likely facilitates better Th17 generation , even when lung antigen levels are low during early stages of infection . It has been reported that higher peptide concentrations are required for inducing Th17 polarization compared to Th1 in a study that examined activation of Smarta-2 TCR-Tg T cells [42] . Efficient CD40-mediated costimulation may serve to lower the threshold for T cell activation and Th17 polarization , and overcome the need for high antigen loads . Interestingly , hip1 mutant Mtb-loaded DCs induced higher Th17 responses compared to wild type Mtb , even without CD40LT treatment ( Fig 5 ) , and enhanced protection ( Fig 6 ) . We have previously shown that hip1 mutant Mtb induces high levels of CD40 and Th17 responses [21] . Therefore , it is likely that elimination of Hip1 results in efficient CD40-dependent costimulation , and bypasses the need for exogenous engagement of the CD40-CD40L pathway . However , we do not rule out the possibility that hip1 mutant Mtb activates alternate DC pathways that promote robust T cell immunity and further investigation into the common and exclusive immune pathways activated by CD40LT and hip1 mutant Mtb is of interest . Previous work by Demangel et al demonstrated that lung Th1 responses can be augmented by transferring BCG-infected DCs in conjunction with agonistic anti-CD40 mAb [33] . However , this approach did not significantly restrict Mtb lung burdens following challenge compared to BCG-infected DCs alone . We speculate that the use of heat killed Mtb in our study as well as the timing of the aerosol challenge at 2 weeks after intratracheal instillation of DCs ( in contrast to 2 days post-DC transfer in the Demangel et al study ) likely established higher frequencies of antigen-specific Th17 and Th1 precursors , leading to better control of Mtb . Additionally , recent work by Griffiths et al showed that mice vaccinated with BCG followed by intratracheal delivery of Ag85B peptide loaded DCs , one day before and four days after challenge with Mtb HN878 , had enhanced bacterial control [46] . Interestingly , they achieved similar reductions in bacterial burden after administration of TLR-9 and CD40 agonists together with Ag85B peptide and also showed higher levels of lung IFN-γ and IL-17 responses . The study by Griffiths et al complements our results , which provide mechanistic evidence that the CD40-CD40L pathway is critical for the generation of Mtb-specific lung Th17 responses . While IL-17 responses appear to be required for resistance against infection with the hypervirulent Mtb HN878 strain , IL-17 may also be important for generating efficacious vaccine-induced immunity . Our data show an association between enhanced IL-17 responses and lower bacterial burden after aerosol Mtb challenge ( Fig 6F ) , but do not directly link Th17 responses with increased protection . While we have demonstrated that engaging CD40 on DCs confers enhanced Th17 responses in the lungs in a setting of mucosal DC transfer , we also observed augmented Th1 responses in vivo prior to challenge ( Figs 5 and 6 ) . Therefore , our data demonstrate that CD40 engagement on DCs improves adaptive immunity to TB , likely due to induction of a balanced Th1/Th17 response . Although we have not shown that the Th17 cells generated in the lung following transfer of DCs stimulated with HK Mtb + CD40LT or HK hip1 mutant Mtb are directly responsible for the increased protection seen in Fig 6 , our studies provide a platform to further investigate the potential of designing vaccination strategies that overcome Mtb immune evasion , either by augmenting CD40 costimulation and/or deletion of immunomodulatory factors such as hip1 ( in BCG or other live attenuated Mtb vaccine strains ) that impair DC functions . Our studies on understanding the role of CD40 costimulation in Th17 responses significantly extend our understanding of the CD40-CD40L pathway during infection , as previous investigations studying this pathway in TB as well as in other infections have focused on Th1 responses . CD40 has been shown to promote Th1 responses by synergizing with TLR signaling to induce high levels of IL-12 production from antigen presenting cells in several infections [33 , 47 , 48] . While our own data show that CD40-/- DCs and CD40LT-treated DCs infected with Mtb do not affect IFN-γ responses in a closed system in vitro ( Figs 1 and 3 ) , treatment of Mtb-infected DCs with CD40LT does augment IFN-γ responses in the lungs 6 days after intratracheal instillation of DCs ( Figs 5 and 6 ) , suggesting that engaging the CD40-CD40L pathway enhances both Th1 and Th17 responses in vivo and may lead to a more balanced Th1/Th17 immunity to TB . Engagement of CD40 is not uniquely important for Th17 generation , as previous investigations on the role of CD40 in mycobacterial diseases have supported the importance of CD40 in the amplification of Th1 responses . CD40-/- mice were shown to be susceptible to aerosol infection with Mtb due to a defective Th1 response [28] , but CD40lg-/- mice were reported to be resistant to Mtb infection and capable of establishing Th1 immunity [28 , 49] . Together with our data showing that Mtb poorly induces CD40 expression on infected DCs [21] , these studies suggest that , while CD40L may be dispensable for generating Th1 responses that control bacterial burden , engaging the CD40-CD40L pathway may be important for generating balanced Th1/Th17 responses that may better control Mtb infection . Moreover , while IL-17 responses were not examined in those studies , mucosal Th17 cells are also likely to contribute to controlling Mtb in CD40-/- mice in vivo; this may be dependent on antigen load as the reported susceptibility of CD40-/- mice disappeared after high dose aerosol challenge [28] . Our work showing that promoting CD40-CD40L interaction augments early Th17 responses in the lung ( Figs 5 and 6 ) is consistent with several previous reports showing an important role for Th17 cells in protection at mucosal surfaces such as in the lung and intestine [18 , 19 , 50 , 51] . In TB , it has been suggested that Th17 cells in the lung may act directly on infected cells or by recruiting additional immune cells , such as IFN-γ+ Th1 , to combat infection . Notably , in Figs 5 and 6 , we show that intratracheal instillation of Mtb-infected DCs treated with CD40LT is associated with an earlier IFN-γ response in the lungs compared to Mtb-DC , which supports the idea that induction of early antigen-specific Th17 can serve to recruit antigen-specific Th1 cells . Our work highlights the importance of augmenting DC costimulation in order to improve adaptive immunity to TB and provides evidence that specifically augmenting DCs through CD40 can enhance antigen-specific mucosal immunity . The generation of robust antigen-specific immunity that goes beyond IFN-γ-producing Th1 responses is an important consideration for vaccines and host-directed therapeutics for TB . The IL-12/STAT-1/IFN-γ axis is important for the control of Mtb , but robust induction of IFN-γ alone does not correlate with enhanced protection against developing TB disease in a variety of vaccine trials [6 , 7] , and there is mounting evidence for IFN-γ independent and Th17-mediated mechanisms of Mtb control [9 , 52 , 53] . In fact , recent work in mice has demonstrated that IFN-γ plays a more important role in control of bacterial burden at extra-pulmonary sites such as the spleen and must be restrained to prevent lung pathology [54] . In humans , bi-allelic mutations in RORC , leading to abrogated IL-17 responses , is associated with susceptibility to mycobacteria , suggesting a role for IL-17 responses in human TB [55] . In addition , the emerging importance of mucosal Th17 responses in protective and vaccine-induced immunity to Mtb [18 , 19 , 51] highlights the need to design and evaluate candidate vaccines that induce robust early Th17 responses . It is important to keep in mind , however , that unbalanced production of IL-17 can be pathogenic [56] . Over-exuberant induction of IL-17 at non-mucosal sites via repeated subcutaneous BCG exposure can lead to worsening of disease [57] and damaging neutrophilia , while IFN-γ receptor signaling limits excessive Th17-mediated neutrophilia [58] . In this context , future studies aimed at augmenting CD40 costimulation would benefit from studying how augmenting this pathway impacts neutrophil responses . In summary , our studies demonstrate a novel role for CD40 costimulation in generating Th17 responses in TB and show that augmenting the CD40-CD40L pathway , either through DC-targeted strategies or deletion of immune-evasion genes in the pathogen , can bolster adaptive immunity in TB . Our results indicate that targeting DC costimulatory pathways in the context of subunit vaccines or live attenuated vaccines represents a novel strategy to induce balanced Th1/Th17 immunity and improve control of Mtb infection .
All experiments using animals or tissue derived from animals were approved by the Institutional Animal Care and Use Committee ( IACUC ) at Emory University ( Protocol number YER-2003476-060919GN ) . Experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . Mtb H37Rv was grown as described previously [21 , 40] . Briefly , bacteria were grown at 37°C in Middlebrook 7H9 broth or 7H10 agar supplemented with 10% oleic acid-albumin-dextrose-catalase ( OADC ) ( Becton Dickinson , Franklin Lakes , NJ ) , 0 . 5% glycerol , and 0 . 05% Tween 80 ( for broth ) , with the addition of 25 μg/ml kanamycin ( Sigma-Aldrich , St . Louis , MO ) for hip1 mutant Mtb . For heat inactivation , bacterial stocks in 7H9 were grown to midlog phase , sonicated , washed twice with PBS and inactivated at 80°C for 2 hours . All mice were housed under specific pathogen-free conditions in filter-top cages within the vivarium at the Yerkes National Primate Center , Emory University , and provided with sterile water and food ad libitum . C57BL/6 , and C57BL/6 CD45 . 1+ congenic mice , CD80-/-CD86-/- , and CD40-/- mice were purchased from The Jackson Laboratory . OT-II TCR Tg mice specific for OVA323–339 peptide were obtained from Dr . Bali Pulendran ( originally generated in the laboratory of Dr . F . Carbone , University of Melbourne ) , and TCR-Tg mice specific for early secreted antigenic target 6 ( ESAT-6 ) 1–20/I-Ab epitope were obtained from Dr . Andrea Cooper ( Trudeau Institute ) and were bred at the Yerkes animal facility . CD40lg-/- x OT-II Tg mice were obtained from Dr . Mandy Ford ( Emory University ) and bred at the Yerkes animal facility . For generating murine bone marrow derived DCs , bone marrow cells from indicated strains of mice were flushed from excised femurs and tibias and grown in RPMI 1640 medium ( Lonza , Walkersville , MD ) supplemented with 10% heat-inactivated FBS ( HyClone , Logan , UT ) , 2 mM glutamine , 1x β-mercaptoethanol , 10 mM HEPES , 1 mM sodium pyruvate , 1x nonessential amino acids , and 20 ng/ml murine recombinant GM-CSF ( R&D Systems , Minneapolis , MN ) . Incubations were carried out at 37°C with 5% CO2 . Fresh medium with GM-CSF ( 20ng/ml ) was added on days 3 and 6 , and cells were used on day 8 for all experiments . We routinely obtained >85% CD11c+ CD11b+ MHCII+ cell purity by flow cytometry . DCs were further purified using CD11c microbead kits as per the manufacturer’s instructions ( Miltenyi Biotec , Auburn , CA ) . 3x105 CD11c-purified bone marrow derived DCs were plated onto 24-well plates overnight prior to infection . For live infections , bacteria were filtered through 5 μm filters , resuspended in complete medium , and sonicated twice for 5 seconds before addition to the adherent monolayers . Bacteria were used for infection ( in triplicate ) at a multiplicity of infection ( MOI ) of 10 or as indicated . Infection of DCs was carried out for 4 hours , after which monolayers were incubated with amikacin ( 200 μg/ml; Sigma-Aldrich ) for 45 minutes to kill extracellular bacteria and then washed four times with PBS before incubating in complete medium . To determine bacterial input , a set of wells was lysed in PBS containing 0 . 5% Triton X-100 and plated onto 7H10 agar plates for CFU enumeration after 21 days . For stimulation of DCs with heat killed Mtb , DC were exposed to heat-killed Mtb at an MOI of 10 in complete medium as determined by CFU enumerated from bacterial stocks prior to heat killing . Uninfected DCs were used as controls for each experiment . For some experiments , DCs were treated with multimeric CD40LT reagent ( Adipogen ) concurrent with infection or stimulation . Cell free supernatants were collected after 24 hours to assay for cytokines: IL-12p40 , IL-12p70 , IL-6 , IL-1β ( BD OptEIA , San Jose , CA ) and IL-23 ( Biolegend , San Diego , CA ) by ELISA according to manufacturers’ instructions . CD4 T cells were purified from single-cell suspensions of spleen and lymph nodes from 6–8 week old transgenic mice ( Naïve CD4 negative selection kit , Stemcell Technologies ) of the indicated strain . Purified CD4 T cells show ≥ 99% purity by FACS analysis . DCs were incubated with 10 μg/ml ( or as indicated ) of OVA323–339 or ESAT-61−20 peptide for 6 hours , washed with PBS , and infected with Mtb with or without CD40LT for 24 hours . DCs were then washed twice with PBS and co-cultured with antigen-specific CD4 T cells to achieve a 1:4 DC:T cell ratio for 72 hours . Cell free supernatants collected from co-cultured cells were analyzed for IFN-γ ( Mabtech , Cincinnati , OH ) , IL-17 ( ELISA Ready-Set-Go , eBioscience ) , and IL-2 ( BD Biosciences ) by ELISA according to the manufacturers’ instructions . 2x104 CD11c-purified DCs were seeded in 96-well plates overnight , pulsed with relevant peptide and treated with the indicated conditions for 24 hours . Afterwards , purified antigen-specific CD4 T cells were incubated with 20 μg/ml anti-CD40L ( clone MR1 ) blocking antibody and co-cultured with DCs to achieve a 1:10 DC:T cell ratio . Co-cultured cells were incubated at 37°C with 5% CO2 for 72 hours prior to harvest of supernatants for ELISAs . CD11c-purified DCs were stimulated with indicated conditions for 24 hours . DCs were then washed twice and resuspended in sterile PBS at 1x106/50 ul and injected intratracheally into isoflurane-anesthetized mice . For some experiments , recipient mice ( CD45 . 1+ ) received purified 1x106 ESAT-61−20 TCR-Tg CD4 T cells ( CD45 . 2+ ) one day prior to DC instillation by tail-vein injection . 6 and 12 days post-intratracheal instillation , lungs and mediastinal lymph nodes were harvested . Lungs were digested with 1 mg/ml collagenase D ( Worthington ) at 37°C for 30 min . For some experiments , the upper right lobe of the lung was used for determining CFU and the rest of the lung was used for cellular assays . Homogenized single-cell lung suspensions were obtained through mechanical disruption and filtered through a 70-μm cell strainer ( BD Biosciences ) , treated with RBC lysis buffer for 3–5 min , and washed twice with cell culture media . Cells were counted and used to set up stimulations for intracellular cytokine staining and flow cytometry . Single cell suspensions were stimulated with media , ESAT-61−20 ( 10 μg/ml ) , Mtb whole cell lysate ( 10 μg/ml ) , or PMA ( 80 ng/ml ) and ionomycin ( 500 ng/ml ) as indicated . BFA ( 5 μg/ml ) and monensin ( 1:1500 ) were added to the stimulated cells after 1 . 5 hours and cells were cultured for an additional 4 . 5 hours , or 16 hours for whole cell lysate stimulations , and then stained for flow cytometry . Live cells were discriminated by a live/dead fixable aqua dead cell stain ( Molecular Probes ) . For staining DCs , murine anti-CD11c PE-Cy7 ( clone N418 , eBioscience ) , anti-CD11b APC-Cy7 ( clone M1/70 , Biolegend ) , anti-CD40 PE-Cy5 ( clone 1C10 , eBioscience ) , anti-CD86 APC ( clone GL1 , eBioscience ) , and anti–MHC II PE ( clone M5/114 . 15 . 2 , BD ) were utilized . For staining T cells , murine anti-CD3 V450 ( clone 500A2 , BD ) , anti-CD4 Alexa700 ( clone RM4-5 , BD ) , anti-CD8 PerCP ( clone 53–6 . 7 , BD ) , anti-TCR γδ BV605 ( clone GL3 , Biolegend ) , anti-CD44 APC-Cy7 ( clone IM7 , BD ) , anti-CD45 . 1 BV785 ( clone A20 , BioLegend ) , and anti-CD45 . 2 BV650 ( clone 104 , BioLegend ) were utilized to stain for surface markers . Murine anti-RORγt PE ( clone B2D , eBioscience ) , anti-TNFa PE-Cy7 ( clone MP6-XT22 , BD ) , anti-IFN-γ APC ( clone XMG1 . 2 , eBioscience ) , anti-IL-2 FITC ( clone JES6-5H4 , BD ) , and anti-IL-17 PE-CF594 ( clone TC11-18H10 , BD ) were stained intracellularly with the BD Cytofix/Cytoperm or BD Transcription Factor kit as per manufacturer’s instructions . Staining for cell-surface markers was done by resuspending ∼1-2x106 cells in 100 ml PBS with 2% FBS containing the antibody mixture at 4°C for 30 min and then washing with PBS containing 2% FBS . Data were immediately acquired using an LSRII flow cytometer ( BD Biosciences ) . Data were analyzed with FlowJo software ( FlowJo LLC , Ashland , OR ) . Mtb H37Rv was grown to OD600 of ∼0 . 6–0 . 8 , washed two times in 1× PBS . 1-ml aliquots were frozen at −80°C and used for infection after thawing . Single-cell suspensions of these aliquots were used to deliver ∼100 CFU into 8–10 week old C57BL/6J mice using an aerosol apparatus manufactured by In-Tox Products ( Moriarty , NM ) . Bacterial burden was estimated by plating serial dilutions of the lung homogenates on 7H10 agar plates on day 1 ( for entry ) or as indicated . CFU was enumerated after 21 days . The statistical significance of data was analyzed using the Student’s unpaired t-test for comparisons between two groups or one-way analysis of variance ( ANOVA ) with a Tukey posttest correction for multiple comparisons for analysis of two or more groups ( GraphPad Prism 6 . 0h ) . In order to calculate correlation , a linear regression was utilized to generate a best-fit line and Spearman’s correlation coefficient calculated ( GraphPad Prism 6 . 0h ) . Data are shown as mean ±S . D . of one representative experiment from multiple independent experiments . | Tuberculosis ( TB ) remains a serious global health problem and understanding how to induce protective immunity to M . tuberculosis ( Mtb ) remains a major challenge . While antigen-specific CD4 T cells and IFN-γ are important for controlling Mtb infection , they are not sufficient for protecting against TB . We need insights into host pathways that can be targeted to overcome suboptimal antigen-specific immunity induced by Mtb . Dendritic cells ( DCs ) are antigen presenting cells that orchestrate the adaptive immune response to infection , but Mtb subverts DC-T cell interactions . Therefore , improving the crosstalk between DCs and T cells during Mtb infection has the potential to enhance anti-mycobacterial immunity . Here we identify interaction between CD40 on DCs and CD40L on T cells as a critical mechanism for generating lung Th17 cells . By engaging CD40 on DCs using a multimeric reagent , we significantly augmented early Mtb-specific Th17 responses in lungs . Intratracheal DC instillation in conjunction with CD40 engagement provided a balanced Th1/Th17 response and improved control of bacterial burden after aerosol challenge with Mtb . Our studies show that the CD40-CD40L pathway is important for the generation of Mtb-specific Th17 responses and targeting CD40-CD40L interactions is a promising avenue for improving adaptive immunity to TB . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] | [
"blood",
"cells",
"t",
"helper",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"cytokines",
"immunology",
"tropical",
"diseases",
"cell-mediated",
"immunity",
"cloning",
"bacterial",
"diseases",
"developmental",
"biology",
"molecular",
"development",
"molecular",
"biology",
"techniques",
"bacteria",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"tuberculosis",
"t",
"cells",
"actinobacteria",
"molecular",
"biology",
"immune",
"response",
"immune",
"system",
"cell",
"biology",
"mycobacterium",
"tuberculosis",
"immunity",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] | 2017 | Engaging the CD40-CD40L pathway augments T-helper cell responses and improves control of Mycobacterium tuberculosis infection |
Uterine smooth muscle cells remain quiescent throughout most of gestation , only generating spontaneous action potentials immediately prior to , and during , labor . This study presents a method that combines transcriptomics with biophysical recordings to characterise the conductance repertoire of these cells , the ‘conductance repertoire’ being the total complement of ion channels and transporters expressed by an electrically active cell . Transcriptomic analysis provides a set of potential electrogenic entities , of which the conductance repertoire is a subset . Each entity within the conductance repertoire was modeled independently and its gating parameter values were fixed using the available biophysical data . The only remaining free parameters were the surface densities for each entity . We characterise the space of combinations of surface densities ( density vectors ) consistent with experimentally observed membrane potential and calcium waveforms . This yields insights on the functional redundancy of the system as well as its behavioral versatility . Our approach couples high-throughput transcriptomic data with physiological behaviors in health and disease , and provides a formal method to link genotype to phenotype in excitable systems . We accurately predict current densities and chart functional redundancy . For example , we find that to evoke the observed voltage waveform , the BK channel is functionally redundant whereas hERG is essential . Furthermore , our analysis suggests that activation of calcium-activated chloride conductances by intracellular calcium release is the key factor underlying spontaneous depolarisations .
The human genome contains a large number of distinct ion channels , each of which may be assembled into multimeric complexes that modulate the electrical activity of the cell in a subtly different way [1] . Determination of the total ion channel repertoire of a given cell ( its ‘conductance repertoire’ ) is cumbersome via conventional biophysical techniques . The latter rely on the specificity and availability of suitable pharmalogical agents or protocols , which may or may not be able to differentiate between combinations of conductances that can give rise to similar behaviors at the electrophysiological level . By contrast , transcriptomic analysis accurately surveys the complete complement of mRNA coding for all potential conductances . However , such data sets are semi-quantitative at best , absent a straightforward relationship between mRNA levels and surface expression/functionality of the electrogenic proteins . The classic approach in electrophysiology has been to fit a suitable mathematical model to the data , where the parameters to be estimated represent not only the densities of the electrogenic entities but also their biophysical properties [2–4] . This approach is hampered by the large number of distinct electrogenic species . One option is to combine the contributions from several electrogenic species ( which are often oligomeric complexes ) into single entities , resulting in a ‘macroscopic current’ model [4–10] . Such models may not be sufficiently detailed for pharmacological purposes and accurate assessment of currents within native cells can be technically challenging . The gold standard , therefore , is to determine the conductance repertoire at the level of individual molecular species . The present paper integrates transcriptomics with the information encoded by the action potential ( AP ) waveform . Electrophysiological data do not necessarily fix a unique solution; there are infinitely many alternative conductance repertoires , that are all equally compatible with the available electrophysiological data . We should consider all conductance repertoires that are consistent with the data , and focus on the properties that they share as a class . This motivates a formal characterisation of the functional redundancy of the system . Our approach is based on two key ideas . First , we use two kinds of data to constrain the space of possibilities , namely mRNA sequencing to determine which species the cell is capable of expressing , combined with parameter estimation to determine the conductance repertoire from electrophysiological data . Second , we obtain the parameters associated with the gating kinetics from the analysis of published data , which means that the only remaining free parameters are the surface densities of the species on the list compiled from transcriptomics . We focus on a cell type of considerable clinical interest , namely the myometrial smooth muscle cell ( MSMC ) , which is the principal unit of electrical activity in the uterus , normally quiescent except in labor , when spontaneous action potentials are generated by the MSMC [11] . A detailed , comprehensive characterisation of these cells would aid the development of therapeutics to manage both preterm birth and perinatal complications associated with uterine contractility , such as postpartum haemorrhage [12] . From a pharmacological point of view , the most promising targets are electrogenic entities whose contribution to the overall electrophysiological behavior is essential in the sense that the role of such an entity cannot be fulfilled by some combination of other ion channels . We show that our analysis is capable of identifying such entities and verify these predictions by means of additional experiments .
Physiological predictions were made , using the simulation model in its free-running mode , on the basis of the conductance repertoire as per the minimal ℓ1-norm criterion . The predictions were validated against experimental data obtained from voltage-clamp experiments for currents isolated by antagonist application . Sets of currents were measured for control and antagonist-inhibited current , and specific current was obtained by subtracting the current recorded in the presence of antagonist from the total current; measurements were reported as mean current at 450 ms ± SEM . The simulations were entirely independent of these additional data . To inhibit and thereby isolate the Kv2 . 1 currents , we applied stromatoxin ( ScTx ) , an inhibitor of Kv2 . 1 homomeric channels , Kv2 . 1/Kv6 . 1 , Kv2 . 1/Kv9 . 3 , as well as the A-type channel Kv4 . 2 . The late outward current was measured to separate Kv2 . 1 from A-type currents . The measured ScTx-sensitive late outward current ( at t = 450 ms ) , resulting from voltage-clamp experiment at 40 mV after application of 100 nM ScTx , was 3 . 71±0 . 41 pA/pF ( n = 19 ) . The total peak Kv2 . 1 current elicited by the simulation for the same voltage step ( 40 mV ) was 3 . 77 pA/pF ( the channels Kv2 . 1 , Kv6 . 1 , and Kv9 . 3 contribute to this current ) . Dofetilide is a selective inhibitor of hERG channels . The isolated dofetilide-sensitive current at 450 ms observed in voltage-clamp experiments at 40 mV after the application of 1 μM dofetilide was 1 . 24±0 . 61 pA/pF ( n = 7 ) , as compared to 1 . 544 pA/pF for the simulated current elicited by hERG at the same voltage . Fig 3 shows the experimental current densities under the voltage-clamp conditions . The simulated values of peak currents at selected voltage steps agree , to within the margins of experimental error , with the currents measured for Kv2 . 1 channels and hERG channel . The functional redundancy is encoded by the kernel of a linear transformation . An informative graphical representation of this functional redundancy can be obtained by transforming the basis of the kernel to a suitable echelon form , which is then visualised as a heat map . These redundancy maps , shown in Figs 4 and 5 , are to be read one row at a time . A unit change in the density of the channel appearing in the leading ( left-most ) position of each row can be functionally compensated , by upward or downward shifts represented by the trailing elements ( i . e . , those to the right ) . The redundancy map thus represents a compensation signature , charting the cell’s ability of exhibiting the same electrophysiological behavior despite altered channel densities . This compensation is computed on the basis of the voltage waveform that was used to derive the basis vectors , but it is not unreasonable to expect that the compensatory adjustments prescribed by the redundancy map are applicable for a wider range of behaviors . An additional caveat is that these adjustments are exact only for infinitesimal changes in the leading channel density; for finite perturbations in the leading variable , the corresponding shifts dictated by the trailing elements are only a first-order approximation . This means that there is a finite operating range within which the approximation can be expected to be reasonably accurate . Moreover , the operating range is further constrained by considerations of physiologic feasibility: channel densities cannot be negative and , similarly , there will be a upper bound to the current density that can be physically sustained by the cell membrane ( there will also be a physiological upper bound to the quantities of channel proteins that can be synthesised and inserted into the membrane ) . To estimate such upper bounds we used the procedure detailed in Materials and Methods , “Physiological upper bounds on channel density values . ” Uterine contraction during labor is characterised by its automaticity [43] . The myogenic mechanisms underlying uterine activity are not fully understood , although it is generally accepted that a major contribution is made by the increase in the concentration of intracellular calcium of MSMCs via L-type calcium channels [43] . Here we explore three mechanisms that could result in the initiation of periodic spontaneous contractions . The first mechanism acts through the activation of the purinergic membrane receptor ( P2RX4 ) , which is an adenosine triphosphate ( ATP ) -gated ion channel . The second mechanism depends on the periodic depletion of phosphatidylinositol-4 , 5 bisphosphate ( PIP2 ) , which modulates the activation of a number of plasma membrane potassium ion channels . Closure of these channels would promote depolarisation and subsequent contractions . The third mechanism is based on periodic intracellular Ca2+ release from sarcoplasmic reticular stores , leading to activation of the Ca2+-activated Cl− conductance ANO1 . Using free-running mode simulations in which the cell was activated through ATP variation ( as outlined above ) , we evaluated the effect of increasing or decreasing various channel densities . The results are shown in Fig 13 . When the BK channel density was set to zero , no significant effect was observed on the waveform , in accordance with our earlier characterisation of BK as functionally redundant . On the other hand , variation of Kv2 . 1 density , from around 30 channels/pF ( double the estimated channel density of Kv2 . 1 ) to 0 channels/pF ( complete block of Kv2 . 1 ) , had a substantial effect on the waveform ( Fig 13 ) . When Kv2 . 1 was completely blocked , there was a more marked overshoot of the AP that rapidly depolarised to a plateau level with increased AP frequency . The slight change in the resting membrane potential ( RMP ) might imply a minor role for Kv2 . 1 in maintaining the RMP . In order to determine what role Kv2 . 1 currents might play in the whole tissue and to test the predictions of our model , we examined ∼3 × 10 mm muscle strips from the longitudinal layer of murine myometrium under current-clamp conditions using a sharp microelectrode to record fluctuations in membrane potential over time . Typical examples of the activity observed in the myometrial muscle strips are shown in Fig 14A and 14B , where we observe a spontaneous AP generated from a RMPs of −45 mV to −50 mV under control conditions for both Day 15 ( D15 ) and Day-18 ( D18 ) , respectively . The application of 100 nM ScTx elicits an increase in AP spike amplitude in the D15 sample and also in the D18 sample ( albeit to a lesser extent ) . The summary data ( n = 4 and n = 6 for D15 and D18 , respectively , p ≤ 0 . 05 ) for these experiments are presented in the graphs in Fig 14C , 14D , 14E and 14F , where mean AP spike amplitude was significantly increased from 42 . 87±5 . 07 mV to 53 . 25±3 . 78 mV in D15 and from 45 . 5±4 . 96 mV to 56 . 21±3 . 37 mV ( p ≤ 0 . 05 ) in D18 tissue . Application of 100 nM ScTx also significantly increased AP frequency in D15 samples from a mean of 1 . 69±0 . 18 AP min−1 to 2 . 31±0 . 34 AP min−1 and 1 . 11±0 . 22 AP min−1 to 1 . 38±0 . 19 AP min−1 . A modest yet consistent reduction in the duration of AP was apparent in APs recorded from D15 tissue . The number of experiments carried out in the current study was insufficient to detect a significant change in RMP with ScTx application in either D15 or D18 cells ( Fig 14F ) . To assess how Kv2 . 1 regulates spontaneous contractions , we recorded isometric tension from ∼3 × 10 mm muscle strips of the longitudinal layer of the mouse myometrium , examining the effects of 100 nM ScTx on the activity . Fig 15A and 15B show representative traces of the mechanical activity of D15 and D18 myometrium . In the presence of 100 nM ScTx , contraction amplitude was increased ( 13 . 32±1 . 79 to 14 . 24±1 . 87 mN , D15; 9 . 69±1 . 79 to 11 . 02±1 . 87 mN , D18; Fig 15C ) , as was frequency ( 0 . 74±0 . 05 to 0 . 89±0 . 07 contractions per minute , D15; 0 . 82±0 . 06 to 0 . 99±0 . 07 contractions per minute , D18; Fig 15E ) , whereas contraction half-width was reduced ( 22 . 97±1 . 38 to 17 . 65±1 . 86 s , D15; not significant in D18; Fig 15D ) .
Laser-capture microdissection [67] and mRNA-sequencing have led to significant advances in gene expression analysis in both single cells and tissues . Application of these techniques provided a complete repertoire of the mRNA population in MSMCs , which we used to construct a list of possible electrogenic entities by combining all possible combinations of ion channel subunits that are potentially expressed in the MSMC . Since many channels exist as various combinations of subunits , we obtain a repertoire of over thirty electrogenic entities that could be present in the plasma membrane of the MSMC . A mathematical model for every individual potential oligomeric channel complex was formulated on the basis of biophysical data in the literature ( typically obtained via heterologous expression ) . We used native cell behavior to drive the individual conductance currents ( viz . voltage and calcium time series as detailed in Materials and Methods ) . The remaining unknowns were the densities in the plasma membrane of the various species . The observed membrane potential time series is linear in these densities , allowing us to calculate the space of all possible solutions of ion channel densities that are consistent with the observed data; the functional redundancy of the system is represented by the basis vectors of this space . To fix the channel densities we considered the most parsimonious solution defined as having the smallest ℓ1-norm subject to non-negativity . Alternative criteria are possible , such as optimising for agreement with mRNA expression levels for the conductances . However , this approach is conditional on accurate quantitative measurements of expression levels , and on the availability of a reliable mapping from transcriptomics to functional proteomics; neither condition is satisfactorily met at the present state of the art ( life times and translation efficiencies of mRNA species vary; moreover , the protein may be stored or degraded instead of being translocated to the membrane ) . Predictions were made using our model based on the conductance repertoire imposed by the criterion of the smallest ℓ1-norm . These predictions were tested against experimental data , obtained in voltage-clamp experiments for isolated currents that were sensitive to inhibition by antagonist application . The simulated values of peak currents in response to voltage steps closely matched the values that were observed experimentally . In the case of certain channels , for example Kir7 . 1 , the simulated data failed to predict the observed data accurately , which can be attributed to the fact the particular voltage waveform that was used to drive the model does not distinguish well between currents ( e . g . between the Kir7 . 1 current and the background potassium current ) . The myometrium has been modeled at the levels of both tissue and the whole organ [5–7] . Bursztyn et al developed an excitation-contraction model of the uterine muscle cells , comprising voltage-gated calcium channels , calcium transporters , and Na+/Ca2+ exchangers [8] . This model explicitly accounts for the processes of myosin light chain phosphorylation and stress production as per the cross-bridge model of Hai and Murphy [9] , but did not include the inactivation process of the calcium current and several other ionic currents . By contrast , Rihana et al [10] modeled the macroscopic currents for each of the relevant ionic species based on voltage-clamp experiments; the dimension of the state space of this model was an order of magnitude lower than ours . Nevertheless , they were able to reproduce various phenomena , such as a single AP , or a train of APs with an RMP of −35 mV to −40 mV and a duration of about 133–200 ms . Another macroscopic ionic current model , accounting for fourteen types of current , was presented by Tong et al [4] . Gating kinetics in the present study were largely derived from equations and parameter estimates taken from the literature and as such representative of the state of the art in the field; accordingly , not all models may have been evaluated to the same standard of rigour as advocated by Fink and Noble [68] . Huys et al [69] proposed an approach similar to ours . However , they illustrated their method on simulated data generated by a macroscopic current model , as opposed to the present work in which real-life data have been confronted with a microscopically detailed account of all relevant currents . One distinguishing feature of our approach is that we can chart the functional redundancy of the system at the level of individual species of electrogenic entities . The redundancy maps show which combinations of channels can be substituted for one another . We investigated the potential scope for redundancy of the system of some of the most-studied potassium channels in MSMCs ( e . g . BK , Kv2 . 1 , SK3 , hERG ) . The use of singular-value decomposition ( SVD ) to determine the directions of substantial variation in parameter space is well-established [70] . A key advantage of SVD over a naive one-factor-at-a-time approach [71] , is that the spectrum of singular values expresses directly which linear combinations of parameters strongly affect the read-out of interest . Sher et al . [72] applied the technique to ionic channel modelling and used the SVD of the sensitivity matrix collecting the partial derivatives of the read-out quantities of interest with respect to the model parameters to probe identifiability and sensitivity , whereas the present method takes a single quantity of interest , the voltage , but treats its entire time course as the output and is accordingly time-global . A regression-based method to perform parameter sensitivity analysis in electrophysiological models was proposed by Sobie [73] , who derived quasi-equivalent linear models by treating simulation outputs obtained under randomised parameter variation as statistical input data . Saltelli and Annoni [71] emphasize the essential locality of “one-at-a-time exploration” given the lack of information that would be provided by the higher-order derivatives ( including cross terms ) , but this caveat is not applicable here since the model is linear in the vector of parameters of interest ( our model is non-linear in the parameters that were fixed from extraneous data but this is not germane to the point made by these latter authors ) . The role of BK channels in modulating uterine excitability remains controversial: whilst the presence of the protein and an active conductance are not disputed [74–76] , the impact of the BK current is not fully understood [40] . Our functional redundancy analysis indicates that the effect of BK can be compensated by small changes in the channel densities of the other channels in the model . This suggests that the effect which BK exerts on the voltage waveform ( as observed in the present study ) is readily replaceable by the currents carried by other electrogenic entities . This observation is consistent with the findings of Aaronson et al , who demonstrated that application of BK inhibitors has no effect on spontaneously contracting rodent myometrial strips [40] . If BK can be readily substituted by suitable combinations of other channels , one might ask why it should be part of the conductance repertoire at all . It may be relevant that the unitary conductance of the BK channel is unusually large [77] , which means that the cell can be silenced by expressing even a minor number of these molecules . Perhaps BK serves the function of an “emergency shut-down . ” We also investigated the potential involvement of Kv2 . 1 in shaping the action potential ( AP ) . We investigated the effect of changes in Kv2 . 1 channel density on the AP waveform by running free-running simulations . Decreasing the Kv2 . 1 channel density from the estimated value to zero increased the amplitude of the initial AP as well as the rate at which steady state is reached during the plateau phase of the AP . The predicted effects on AP amplitude and frequency were confirmed by experiments although the effect on the resting membrane potential ( RMP ) could not be verified with certitude . It appears that Kv2 . 1 channels do play a role in modulating the excitability of the myometrium by inhibiting both contraction frequency and amplitude , while contributing little to the RMP . As regards the hERG channel , our functional redundancy analysis indicated that small changes in the hERG channel density could not be compensated by physiologically realistic adjustments in the densities of the other channels . In addition , the simulations carried out free-running mode suggest that variations in the hERG channel density substantially affect the AP waveform ( Fig 13 ) . Taken together , these findings indicate that hERG is a major suppressor of bursting activity in MSMCs , in keeping with Greenwood et al [41] , who observed that the ERG-specific blocker dofetilide-induced contractions in quiescent tissues , while ERG-channel blockers were able to induce contractions in tissue strips that failed to develop spontaneous rhythmic activity . In contrast to hERG’s essential role , the SK channels were found to be readily susceptible to compensation . The redundancy analysis indicates that SK2 can be functionally replaced by adjustments in SK3 and SK4 densities . This is in accordance with the close physiological similarities between the members of the SK family . We also demonstrated the utility of the model in exploring various mechanistic hypotheses , addressing three mechanisms that could trigger spontaneous contractions . Simulations provided supporting evidence for the involvement of extracellular ATP and/or calcium ‘sparks’ affecting CaCCs , whereas we found no support for a role of PIP2 acting on potassium channels . Although far from conclusive in their own right , such simulations can help experimentalists to pinpoint promising research avenues . The model can be improved in several ways . In its present form , the dynamics of intracellular calcium are represented in a minimalistic way and a more physiologically faithful model would be desirable . However , this is subject to the quality of the temporospatial resolution of the cytosolic and sarcoplastic-reticular levels of Ca2+ . Another mechanism we have not considered here is the feedback from force-generation on the dynamics of membrane potential , which acts via stretch-sensitive ion channels . Furthermore , our model describes a single cell , whereas the in vivo activity of MSMCs is critically dependent on their interconnectivity . Thus , a multi-cell model with realistic network properties would constitute another future improvement . In the present study a spontaneous voltage signal was imposed . It is also possible to impose artificial voltage waveforms ( AVWs ) , for instance composed as a linear combination of B-splines , and compare the current response to such an AVW to the current response of the model . The advantage of this approach would be that the AVW can be optimised to discriminate maximally between two given conductances , for instance by minimising the inner product between the current responses of the respective conductances , where the two are considered as functions of time . A tighter upper bound on the functional redundancy of the system can be obtained in this manner . The functional redundancy we map is strictly relative to the waveform that is imposed . As more behaviors are explored , more “essential roles” are revealed and the dimensionality of the kernel can generically be expected to diminish; in fact , it could come down all the way to zero , which would imply that there is no intrinsic redundancy in the conductance repertoire . Alternatively , the kernel dimensionality settles on a non-zero value “in the limit of arbitrarily many explored behaviors”—this value would then denote the irreducible functional redundancy of the system . It stands to reason that sundry physiological limitations will impose further constraints on the system , but the considerations in the present paper concern only the functional redundancy vis-à-vis the membrane conductance repertoire . Two key advantages of our approach are that it is modular and expandable . The model can readily accommodate new information regarding additional conductances or novel kinetic properties . Moreover , being expressed in terms of microscopic conductances that correspond to particular molecular entities , the model can in principle be related directly to transcriptomics . This has important practical advantages , because transcriptomics data can be acquired in different tissues , and at different time points , accurately surveying all potential molecular entities . Another facet of the expandability of our approach is that the characterisation of functional redundancy , becomes more accurate , that is , comes closer to the irreducible functional redundancy , as a wider range of physiological behavior of a given cell ( both spontaneous and evoked ) is observed . Our analysis of functional redundancy provides a means to address the well-known difficulty in electrophysiological modelling that electrophysiology at the whole-cell level is insufficient to constrain the densities of all channels . Traditionally , this problem has been solved by postulating macroscopic currents that represent the combined action of several similar conductances , but this approach has the drawback that a direct correspondence to genomic or transcriptomic data is lost . We would even argue that the approach based on macroscopic currents is a false economy since , from an experimental point of view , the acquisition of the macroscopic currents is itself underdetermined , whereas microscopic currents can be unequivocally isolated and thereby fully characterised .
All procedures were conducted within the guidelines of The Declaration of Helsinki and were subject to local ethical approval ( REC-05/Q2802/107 ) . Prior to surgery , informed written consent for sample collection was obtained . The complete repertoire of the electrogenic proteins that are potentially expressed in the MSMC was determined on the basis of mRNA expression data , which have been made available as entire raw data via the GEO database ( GEO series accession number GSE50599 ) and , in processed form , as the supplementary material associated with the original publication by Chan et al [13] . We constructed a complete list of every individual oligomeric channel complex that has been attested in the literature and that is consistent with the subunits in the mRNA expression set; the mathematical model , shown in Fig 1A , incorporates 31 time-dependent ionic currents . These currents include outward currents such as the voltage-gated potassium current , the voltage- and calcium-gated potassium current , as well as the calcium-gated potassium current . In addition , the model comprises two inward , depolarising currents attributed to the two voltage-gated calcium channels ( L-type and T-type ) . Chloride fluxes are represented as a calcium-activated chloride current , as well as a background current . A background potassium current was also accounted for . Finally , the model includes pumps and exchangers: the Na+-Ca2+ exchanger ( NCX ) , the plasma membrane calcium ATPase ( PMCA ) , and the Na+-K+ pump ( NaK ) . A separate mathematical model for each oligomeric channel was formulated , either based on models that have already been proposed in the literature or on the basis of the available data . The biophysical and kinetic parameters of several of these entities could in many cases be taken directly in the literature , or else have been obtained by means of least-squares fitting to the experimental data on heterologous expression systems , taken from the literature , as detailed below . The mathematical expressions that describe the individual channel kinetics , together with the associated parameters descriptions , units , and values were previously listed in McCloskey et al [39] and are reproduced here for the sake of clarity in the Supplementary Materials , while a summary of the potential conductance species included in our model is shown in Table 1 . | A well-known problem in electrophysiologal modeling is that the parameters of the gating kinetics of the ion channels cannot be uniquely determined from observed behavior at the cellular level . One solution is to employ simplified “macroscopic” currents that mimic the behavior of aggregates of distinct entities at the protein level . The gating parameters of each channel or pump can be determined by studying it in isolation , leaving the general problem of finding the densities at which the channels occur in the plasma membrane . We propose an approach , which we apply to uterine smooth muscle cells , whereby we constrain the list of possible entities by means of transcriptomics and chart the indeterminacy of the problem in terms of the kernel of the corresponding linear transformation . A graphical representation of this kernel visualises the functional redundancy of the system . We show that the role of certain conductances can be fulfilled , or compensated for , by suitable combinations of other conductances; this is not always the case , and such “non-substitutable” conductances can be regarded as functionally non-redundant . Electrogenic entities belonging to the latter category are suitable putative clinical targets . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"voltage-gated",
"ion",
"channels",
"uterus",
"calcium-activated",
"potassium",
"channels",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"membrane",
"potential",
"electrophysiology",
"neuroscience",
"ion",
"channels",
"cellular",
"structures",
"and",
"organelles",
"proteins",
"calcium",
"channels",
"biophysics",
"cell",
"membranes",
"physics",
"biochemistry",
"myometrium",
"anatomy",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"potassium",
"channels",
"physical",
"sciences",
"voltage-gated",
"calcium",
"channels",
"neurophysiology"
] | 2016 | Reconstruction of Cell Surface Densities of Ion Pumps, Exchangers, and Channels from mRNA Expression, Conductance Kinetics, Whole-Cell Calcium, and Current-Clamp Voltage Recordings, with an Application to Human Uterine Smooth Muscle Cells |
In the peripheral nervous system ( PNS ) myelinating Schwann cells synthesize large amounts of myelin protein zero ( P0 ) glycoprotein , an abundant component of peripheral nerve myelin . In humans , mutations in P0 cause the demyelinating Charcot-Marie-Tooth 1B ( CMT1B ) neuropathy , one of the most diffused genetic disorders of the PNS . We previously showed that several mutations , such as the deletion of serine 63 ( P0-S63del ) , result in misfolding and accumulation of P0 in the endoplasmic reticulum ( ER ) , with activation of the unfolded protein response ( UPR ) . In addition , we observed that S63del mouse nerves display the upregulation of many ER-associated degradation ( ERAD ) genes , suggesting a possible involvement of this pathway in the clearance of the mutant P0 . In ERAD in fact , misfolded proteins are dislocated from the ER and targeted for proteasomal degradation . Taking advantage of inducible cells that express the ER retained P0 , here we show that the P0-S63del glycoprotein is degraded via ERAD . Moreover , we provide strong evidence that the Schwann cell-specific ablation of the ERAD factor Derlin-2 in S63del nerves exacerbates both the myelin defects and the UPR in vivo , unveiling a protective role for ERAD in CMT1B neuropathy . We also found that lack of Derlin-2 affects adult myelin maintenance in normal nerves , without compromising their development , pinpointing ERAD as a previously unrecognized player in preserving Schwann cells homeostasis in adulthood . Finally , we provide evidence that treatment of S63del peripheral nerve cultures with N-Acetyl-D-Glucosamine ( GlcNAc ) , known to enhance protein quality control pathways in C . elegans , ameliorates S63del nerve myelination ex vivo . Overall , our study suggests that potentiating adaptive ER quality control pathways might represent an appealing strategy to treat both conformational and age-related PNS disorders .
In the peripheral nervous system ( PNS ) , Schwann cells discontinuously myelinate axons to promote the fast saltatory conduction of the nerve impulses . During myelin biogenesis , large amounts of both lipids and proteins are produced . After the completion of myelination , the maintenance of proper myelin homeostasis is essential for preserving axon functionality [1 , 2 , 3] . Mutations in myelin proteins , such as myelin protein zero ( P0 ) , cause different forms of Charcot-Marie-Tooth ( CMT ) neuropathy with mild-to-severe peripheral nerves dysfunctions [4 , 5] . In particular , the deletion of serine 63 in the extracellular domain of P0 ( P0-S63del ) results in a CMT1B neuropathy characterized by developmental hypomyelination followed by demyelination , with formation of onion bulbs-like structures [6 , 7] . Wild type ( WT ) P0 , a type I transmembrane protein , is normally synthesized on ER-bound ribosomes , N-glycosylated at asparagine 122 and delivered to myelin via the secretory pathway [8 , 9 , 10] . The mutant P0-S63del is instead misfolded and retained in the ER , where it elicits a chronic ER stress that activates a dose-dependent unfolded protein response ( UPR ) [7 , 11 , 12] . The UPR , a cellular response aimed at rebalancing ER homeostasis , relies on the activation of signaling pathways which attenuate global protein synthesis and stimulate ER quality control ( ERQC ) systems , such as folding capacities and ER-associated degradation ( ERAD ) [13] . In ERAD of glycoproteins , the misfolded substrates are demannosylated , recognized by mannose-specific lectins and directed to multi-protein channels , the dislocons , that retrotranslocate them to the cytosol for proteasomal degradation [14 , 15] . An involvement of the ERAD pathway in the clearance of the misfolded P0-S63del protein was suggested by the observation that genes codifying for factors of the ERAD/proteasome system are globally upregulated in S63del mouse nerves [11 , 16] . Here we show that P0-S63del glycoprotein is degraded via ERAD in vitro and that the Schwann cell-specific ablation of the dislocon component Derlin-2 causes ERAD impairment and worsens the CMT1B phenotype in S63del mice . Furthermore , we show that ablation of Derlin-2 from WT Schwann cells results in a late-onset motor-predominant peripheral neuropathy . Thus , an efficient ERAD is not only protective in the S63del-CMT1B disorder , but is also important to preserve adult myelin integrity by maintaining ER homeostasis in Schwann cells . Compounds that enhance protein quality control pathways such as ERAD might therefore represent a new therapeutic avenue for ER stress- and age-related neuropathies .
Transcriptomic analysis of S63del mouse nerves , which are characterized by a chronic UPR [7 , 12] , revealed the upregulation of several ERAD genes ( S1A Fig ) [11] . Accordingly , qRT-PCR and Western blot experiments performed on post-natal day 28 ( P28 ) sciatic nerve samples confirmed the induction of a selection of ERAD factors at both the mRNA and protein levels ( S1B and S1C Fig ) [11] . This increase appeared to occur specifically in S63del Schwann cells , as suggested by immunofluorescence staining for the ERAD dislocation factor Derlin-2 on teased nerve fibers ( S1D Fig ) . Co-immunoprecipitation experiments performed on P28 sciatic nerve lysates showed that Derlin-2 interacted with P0 in WT nerves and that this interaction was increased in S63del nerves , where also Derlin-1 appeared to interact with P0 ( S1E , S1F and S1G Fig ) . These data and the observation that P0-S63del glycoprotein , but not wild-type P0 , is polyubiquitinated [16] suggested that the ERAD/proteasome system might be involved in the degradation of the mutant P0-S63del in vivo . Several limitations , however , hamper the possibility to directly verify this hypothesis in nerves . In S63del Schwann cells both the wild type P0 and the mutant P0-S63del proteins are abundantly co-expressed [7] and , currently , there are no specific antibodies that discriminate between the two forms [16] . Thus , to test our hypothesis , we generated stable , tetracycline inducible HEK293 cells that express HA-tagged versions of either wild type P0 ( P0-wt ) , P0-S63del or P0-S63C proteins . The latter was used as a control since , despite carrying a mutation in the same amino acid , the P0-S63C glycoprotein exits the ER-Golgi stacks and reaches the myelin sheath [7 , 17] . Immunofluorescence analysis confirmed that also in this in vitro system P0-wt and P0-S63C reached the cell surface , whereas P0-S63del protein is retained in the ER , as shown by its co-localization with the ER marker Calnexin ( CNX ) ( Fig 1A ) . Accordingly , endoglycosidase H ( EndoH ) assay showed that the N-glycan portion of P0-S63del protein was largely sensitive to EndoH cleavage , indicative of a folding defect that blocks the protein into the ER . The control proteins were instead mostly EndoH resistant , since they assume a mature and transport-competent conformation ( Fig 1B ) . As expected , all these proteins were sensitive to treatment with peptide-N-glycosidase F ( PNGaseF ) , an enzyme that removes almost all types of N-glycan independently of the intracellular localization of the proteins ( Fig 1B ) [18] . Western blot analysis performed 17 and 48 hrs after induction with tetracycline showed that the steady state levels of both P0-S63C and , to greater extent , P0-S63del proteins were reduced as compared to P0-wt ( Fig 1C ) . However , pulse-chase experiments did not detect any gross difference in their rate of synthesis ( S2A and S2B Fig ) . These observations suggest that the mutant P0s , and particularly the P0-S63del protein , may have reduced intracellular stability and faster degradation rate as compared to P0-wt . To test whether the ERAD/proteasome system degrades the two P0 mutants , we treated the induced cells with the proteasome inhibitor PS341 and performed pulse-chase analysis followed by immunoprecipitation and SDS-PAGE ( Fig 1D and 1E ) . Western blot against ubiquitin confirmed that PS341 blocked proteasome activity , resulting in a global accumulation of polyubiqitinated proteins ( S2C Fig . ) Differently from P0-wt and P0-S63C , P0-S63del protein displayed higher electrophoretic mobility in pulse-chase experiments ( Fig 1D ) , typical of mannose trimming that occurs when misfolded proteins are retained in the ER before degradation [19] . Importantly , P0-S63del protein showed faster degradation rates as compared to controls and accumulated following proteasome inhibition ( Fig 1D and 1E ) . We also noted that two additional polypeptides with MW around 90 and 70 KDa appeared to co-immunoprecipitate with the P0-S63del protein , but not with the other two P0 variants ( see arrowheads in Fig 1D and S2A Fig ) . Based on their crucial role as ER retention factors for unstructured glycoproteins , we hypothesized that these two proteins could be CNX and binding immunoglobulin protein ( BiP ) . To verify this , induced cells were treated with PS341 and subjected to immunoprecipitation against P0 , followed by Western blot analysis to specifically reveal BiP and CNX ( Fig 1F ) . As expected , these chaperones were substantially detectable only in P0-S63del immunocomplexes ( Fig 1F ) and , in addition , pulse-chase followed by double co-immunoprecipitation experiments , which reveal only stable interactions , confirmed these results ( S2D and S2E Fig ) . Altogether these data corroborate our hypothesis that the ERAD pathway participates in the clearance of the ER-retained P0-S63del glycoprotein , limiting its toxicity . Based on these observations , we hypothesized a protective role for ERAD in the S63del CMT1B neuropathy . To test this , we impaired ERAD in peripheral nerves via the Schwann cell-specific ablation of Derlin-2 , a well characterized ERAD factor and UPR target gene [20 , 21 , 22] , whose deletion was previously shown to cause defective ER dislocation in vivo [23] . Importantly , Derlin-2 protein is upregulated in S63del Schwann cells and it appears to interact with P0 ( S1 Fig ) . We crossed Der2fl/fl mice [23] with P0Cre mice [24] to obtain P0Cre//Der2fl/fl or fl/+ mice . These mice were then bred with S63del//Der2fl/+ animals to generate S63del//P0Cre//Der2fl/fl mice ( S63del//Der2SCKO ) , P0Cre//Der2fl/fl mice ( Der2SCKO ) and the respective P0Cre-negative controls . To test the efficiency and specificity of P0Cre-mediated recombination , we performed PCR reactions on genomic DNA extracted from sciatic nerves and several other tissues ( S3A and S3B Fig ) . In P5 sciatic nerves the recombined Der2KO ( 600bp ) band was specifically detected in all P0Cre-positive samples , but not in P0Cre negative control nerves , as expected ( S3A Fig ) . Moreover , in P21 Der2SCKO mice , the recombination band appeared only in sciatic nerves and not in other tissues , with the exception for a faint Der2KO band visible in skeletal muscles , possibly because of the presence of nerve terminals ( S3B Fig ) . Accordingly , Derlin-2 mRNA and protein levels were consistently decreased in Der2SCKO and S63del//Der2SCKO nerves as compared to the respective controls ( S3C–S3E Fig ) . To assess ERAD impairment in Derlin-2 knockout Schwann cells , we measured the levels of the lectin chaperone osteosarcoma amplified 9 ( OS9 ) and of inositol-requiring enzyme 1α ( IRE1α ) , two endogenous ERAD substrates that get stabilized in several ERAD deficient tissues [23 , 25 , 26 , 27 , 28] . In both Der2SCKO and S63del//Der2SCKO nerves we found remarkably high amounts of OS9 protein , but minimally changed mRNA levels ( S3F–S3H Fig ) , indicating that , also in peripheral nerves , the lack of Derlin-2 prevalently leads to OS9 protein stabilization because of a less efficient degradation . Similarly , IRE1α protein was strongly increased in Der2SCKO and S63del//Der2SCKO nerves ( S3I and S3J Fig ) . A mild stabilization of endogenous ERAD substrates was also observed in S63del controls ( S3F–S3H , S3I and S3J Fig ) , possibly because of the diminished proteasomal efficiency in these nerves [16] . Given the highly secretory nature of myelinating Schwann cells , we reasoned that the impairment of ERAD might have detrimental effects on myelination . To test this hypothesis , we collected Der2SCKO sciatic nerves at P5 , P15 and P28 and evaluated their morphology . At all these time points , Der2SCKO nerves did not display any gross myelin abnormality ( Fig 2A and S4A Fig ) . At P28 , myelin thickness was normal as measured by morphometric g-ratio analysis on semithin sections ( mean g-ratio: Der2SCKO 0 . 69±0 . 001; WT 0 . 68±0 . 003 , n . s . ; Fig 2B ) , and the axon size distribution was unchanged as compared to WT controls ( Fig 2D ) . To assess whether Derlin-2 is instead required for remyelination , we performed sciatic nerve crush experiments on Der2SCKO mice and looked at the extent of remyelination 45 days after injury ( T45 ) . Morphological analysis performed on semithin sections did not reveal any significant difference in the extent of remyelination , with comparable numbers of remyelinated and degenerated fibers in Der2SCKO and WT injured controls ( S4B–S4D Fig ) . Moreover , g-ratio analysis performed on T45 EM sections confirmed that myelin reached a similar thickness in both Der2SCKO and WT remyelinated nerves ( S4E and S4F Fig ) , suggesting that Derlin-2 is not required for correct remyelination after injury . Notably however , ablation of Derlin-2 in S63del nerves worsened the developmental hypomyelination ( Fig 2A ) , with a significant increase in the mean g-ratio at P28 ( S63del//Der2SCKO 0 . 73±0 . 003; S63del 0 . 71±0 . 001 , P = 0 . 015; Fig 2C ) , but no effects on the axon size distribution ( Fig 2D ) . Detailed EM analysis confirmed equal myelination in Der2SCKO and WT developing nerves , and increased hypomyelination in S63del//Der2SCKO nerves as compared to S63del control , without significant alterations in Schwann cell morphology ( S5A and S5B Fig ) . Finally , Derlin-2 deletion did not impact on Schwann cell number in P21 sciatic nerve sections ( S5C and S5D Fig ) . In fact , whereas S63del nerves showed supernumerary Schwann cells as compared to WT , as previously reported [12] , the number of cells nuclei was basically unaltered in both Der2SCKO and S63del//Der2SCKO as compared to the respective WT and S63del controls ( S5C and S5D Fig ) . Altogether these data indicate that Derlin-2 is neither required for normal developmental myelination nor for remyelination after injury , but it appears to be protective in early stages of the CMT1B neuropathy . We have previously shown that an increase in P0-S63del protein expression determines higher UPR induction and a more severe neuropathy in transgenic S63del mice [7 , 12] . We reasoned that , since ERAD appears to be involved in the degradation of the P0-S63del glycoprotein , its impairment should stabilize the misfolded P0 in the ER and increase the levels of stress , leading to the worsening of the phenotype observed in S63del//Der2SCKO mice . To test this hypothesis , we first checked the effects of siRNA-driven Derlin-2 depletion on the stabilization of P0-S63del protein in HEK293 cells . We confirmed the efficient silencing of Derlin-2 by Western blot ( Fig 3A ) and then performed pulse-chase experiments , followed by immunoprecipitation and SDS-PAGE , to monitor the levels of radiolabeled P0-S63del protein ( Fig 3B and 3C ) . After 150 min of chase , Der2 knocked-down cells displayed 20% higher levels of radiolabeled P0 as compared to control cells , confirming that the depletion of Derlin-2 leads to P0-S63del protein accumulation ( Fig 3B and 3C ) . In addition , we also found that the silencing of Derlin-2 , even in the absence of P0-S63del expression , activates a stress response , as shown by increased homocysteine-induced endoplasmic reticulum protein ( HERP ) protein levels [29 , 30] , whose induction is further intensified following the accumulation of P0-S63del protein ( Fig 3D ) , in agreement with increased expression of HERP in P28 S63del nerves ( S1A and S1B Fig ) . Next , as a measure of ER-retained P0-S63del levels in vivo , we checked the expression of ER stress/UPR markers such as BiP , C/EBP-Homologous Protein ( CHOP ) , Glucose Regulated Protein 94 ( GRP94 ) , phosphorylated eukaryotic Initiation Factor 2 α ( P-eIF2α ) and spliced X-box Binding Protein 1 ( Xbp1s ) in sciatic nerves at P28 . As hypothesized , S63del//Der2SCKO nerves showed a consistent trend towards higher levels for all the markers analyzed as compared to S63del nerves , suggesting higher P0-S63del levels in the ER ( Fig 4A–4I ) . Despite overall normal Schwann cells development and myelination ( Fig 2 ) , and in line with what seen in the HEK293 cells following Derlin-2 silencing , also Der2SCKO sciatic nerves showed signs of modest ER stress and UPR induction as compared to WT controls ( Fig 4A–4I ) . These observations resemble what was previously reported in Derlin-2 deficient hepatocytes and B-cells , in which both the secretory function and cell development appeared normal despite mild UPR activation elicited by lack of Derlin-2 [23] . In S63del mice demyelination occurs predominantly in adulthood [7 , 12] . We thus analyzed the effects of ERAD impairment also in 6–12 months old ( mo ) sciatic nerves . In S63del//Der2SCKO nerves , the myelin sheath appeared thinner as compared to S63del controls ( mean g-ratio: S63del//Der2SCKO 0 . 73±0 . 007; S63del 0 . 71±0 . 004 , P = 0 . 043; Fig 5A and 5B ) , consistently with what already observed at P28 . Rather unexpectedly , adult Der2SCKO sciatic nerves showed reduced myelin thickness as well ( Der2SCKO 0 . 68±0 . 002; WT 0 . 65±0 . 009 , P = 0 . 032 ) , suggestive of mild nerve pathology ( Fig 5A and 5B ) . In addition , at both 6 and 12 mo , S63del//Der2SCKO sciatic nerves showed a strong increase in the number of demyelinated fibers as compared to S63del nerves ( Fig 5A and 5C ) , and features of progressive demyelination appeared also in Der2SCKO mice ( Fig 5A and 5C ) . Of note , in both S63del//Der2SCKO and Der2SCKO mice , large naked axons and onion bulbs predominantly appeared within the motor areas of sciatic nerves , composed of fascicles of medium-large caliber fibers ( Fig 5A and 5D ) . Quantitative EM analysis confirmed the hypomyelination in S63del//Der2SCKO nerves as compared to S63del nerves , and in Der2SCKO nerves as compared to WT ( Fig 6A and 6B ) . Ultrastructural analysis showed that myelin periodicity was grossly unaltered in Der2SCKO as compared to WT , indicating that the reduction in myelin thickness was likely due to a reduced number of myelin wraps rather than a defect in myelin compaction ( Fig 6C–6C’ ) . Next , we evaluated nerve functionality by measuring nerve conduction velocity ( NCV ) , F-wave latency ( FWL ) and compound muscle action potential ( CMAP ) amplitude . In line with the observed phenotypes , NCVs were significantly reduced in both Der2SCKO and S63del//Der2SCKO nerves as compared to the respective controls ( WT 44 . 24±0 . 678; Der2SCKO 39 . 01±1 . 243; S63del = 35 . 98±1 . 384; S63del//Der2SCKO31 . 45±0 . 947; Fig 6D ) . FWL , which was higher in S63del nerves as compared to WT as expected , was slightly increased by the ablation of Derlin-2 and , remarkably , 30% of the FWLs in Der2SCKO and 7% in S63del//Der2SCKO nerves were absent ( Fig 6E ) . Derlin-2 ablation in both WT and S63del backgrounds , instead , did not significantly alter CMAP amplitudes ( Fig 6F ) . Altogether these data indicate that a functional ERAD contributes to myelin maintenance in normal adult nerves and protects Schwann cells from demyelination in CMT1B neuropathy with activated UPR . The morphological analysis outlined above suggested that onion bulbs were more frequent in the motor fascicles rather than in the sensory ones of sciatic nerves in both Der2SCKO and S63del//Der2SCKO mice ( Fig 5A and 5D ) . Thus , we reasoned that the motor component of nerves could be more susceptible to ERAD perturbation as compared to the sensory component . To test this hypothesis , we took advantage of femoral nerves , in which a motor-predominant branch ( quadriceps nerves ) and a sensory branch ( saphenous nerve ) can be analyzed separately . Der2SCKO quadriceps nerves showed more pronounced onion bulbs formation as compared to sciatic nerves ( compare Fig 7A with Fig 5A ) , whereas Der2SCKO saphenous nerves appeared grossly normal ( Fig 7A ) . In S63del//Der2SCKO mice , instead , both quadriceps and saphenous nerves were severely compromised , although with some remarkable differences . In fact S63del//Der2SCKO quadriceps nerves , but not saphenous nerves , showed extensive onion bulbs formation and signs of ongoing demyelination ( Fig 7A–7C and 7E ) as compared to the control nerve ( Fig 7A and 7D ) , similar to what observed in high P0-S63del overexpressor ( S63del-H ) [7] . Despite this difference however , both nerves displayed signs of axonal degeneration ( Fig 7F and 7G ) . Overall these data indicate that ablation of Derlin-2 in WT mice causes an age-related , motor-predominant , demyelinating neuropathy , suggesting that motor nerves are more sensitive to ERAD impairment as compared to sensory nerves . In S63del mice instead , ERAD impairment aggravates the CMT1B disease phenotype in all nerves analyzed , with motor fibers presenting the most severe demyelination . The progressive and severe worsening of the neuropathy in S63del//Der2SCKO mice and the appearance of a phenotype in adult Der2SCKO nerves prompted us to analyze stress responses also at this stage . In 6 mo nerves ERAD impairment due to Derlin-2 ablation was still evident , as shown by the increased levels of OS9 and IRE1α protein in both Der2SCKO and S63del//Der2SCKO sciatic ( S6A–S6C Fig ) and quadriceps ( S6D–S6F Fig ) nerves . Western blot and qRT-PCR experiments showed a strong increase in both ER stress levels and UPR activation in adult S63del//Der2SCKO nerves as compared to S63del controls ( Fig 8A–8I and S7A–S7E Fig ) . Conversely , Der2SCKO sciatic nerves , which showed significant increase of BiP and GRP94 protein levels ( Fig 8A , 8B , 8D and 8E ) , did not display canonical activation of the UPR branches . In fact , BiP and Xbp1s mRNAs ( targets of ATF6 and IRE1 respectively ) were not induced ( Fig 8G–8I ) and only P-eIF2α levels appeared to be increased as compared to WT ( Fig 8C and 8F ) , although the phosphorylation of eIF2α was uncoupled from CHOP induction ( Fig 8H ) . In Der2SCKO quadriceps nerves , where demyelination was more pronounced ( Fig 7 ) , we instead detected a trend towards an increase of CHOP ( S7E Fig ) , correlating with previous observations suggesting that CHOP activation underlies demyelination in peripheral nerves [12] . Taken together these results suggest a complex scenario in which in S63del//Der2SCKO nerves ERAD impairment determines an increase in the ER retention of P0-S63del protein augmenting ER stress/UPR , progressively worsening the neuropathic phenotype . In Der2SCKO sciatic nerves instead , the perturbation of ERAD transiently activates a mild UPR during development , most likely favouring cell fitness [31] . The UPR however fails to be maintained into adulthood , possibly determining a proteostatic failure that may underlie myelin degeneration during aging . Overall our data suggest that in CMT1B neuropathy ERAD is an adaptive ERQC pathway that limits the toxic effects of the misfolded P0 . Potentiating the adaptive ERQC systems might therefore represent an appealing approach for the treatment of misfolded-protein diseases . In this respect , it has been shown that the stimulation of the hexosamine biosynthetic pathway ( HBP ) , which generates intermediates for N- and O-glycosylation of proteins [32] , promotes stress resistance , relief from proteotoxicity and lifespan extension in C . elegans by globally enhancing protein degradation systems , including ERAD [33] . Thus , we hypothesized that treatments with the HBP intermediate N-acetyl-glucosamine ( GlcNAc ) could improve S63del myelination . To test this hypothesis , both WT and S63del dorsal-root-ganglia ( DRG ) explants were allowed undergoing myelination ex vivo for 2 weeks in presence or absence of GlcNAc . These treatments appeared well tolerated , as suggested by normal myelination in treated WT DRG explants ( Fig 9A–9C ) . Treatment with GlcNAc significantly improved myelination in S63del DRG explants as measured by the increased number and length of MBP+ internodes ( Fig 9A–9C ) and increased P0 protein , as measured by WB ( Fig 9D and 9E ) . This improvement was accompanied by a significant reduction in stress levels as measured by qRT-PCR for CHOP ( Fig 9F ) , even though we could detect only a small , not significant increase in the mRNA for the ERAD component Sel1L ( Fig 9G ) , that in C . elegans appears the main ERAD target of the HBP pathway [33] . Further studies will be required to test whether and how the stimulation of the HBP pathway can alleviate nerve pathology in vivo .
The S63del-CMT1B neuropathy is characterized by developmental hypomyelination , followed by demyelination with onion bulbs formation and compromised nerve functionality [6 , 7] . Deletion of Derlin-2 from S63del Schwann cells exacerbates both the early and late features of the neuropathy , indicating that ERAD is protective , most likely because of its role in degrading the misfolded P0-S63del protein , as our in vitro data suggest . In S63del//Der2SCKO nerves , in fact , ER stress and UPR levels are higher as compared to S63del controls . This is in line with the idea of increased P0-S63del accumulation in the ER , and accordingly , the phenotype of S63del//Der2SCKO nerves closely recapitulates what happens in S63del-H mice [7 , 12] . In S63del//Der2SCKO mice , quadriceps and saphenous nerves manifest a dramatic exacerbation of the disease , whereas sciatic nerves appear only moderately worsened . The reason for this difference is currently unknown . It would be intriguing to assess whether these nerves differ in their inflammatory response or susceptibility to inflammation , since a similar disease feature is often observed in sporadic inflammatory neuropathies [39 , 40] . Finally , the observation that deletion of Derlin-2 causes per se a late onset demyelinating neuropathy ( see below ) and that its deletion synergizes with the P0-S63del mutation worsening the CMT1B phenotype , may suggest that also ERAD genes , alongside integral nodal component such as Nrcam and Scn8a [41 , 42 , 43] , could be among those genetic modifiers that influence the large phenotypic variability of many forms of CMTs . The number of CMT1B-causing mutations that activate a UPR is constantly increasing [44 , 45] and , in addition , other mutant myelin proteins of both PNS and CNS are characterized by ER retention and/or proteasomal degradation , such as PMP22 [46 , 47 , 48] , connexin-32 [49] , MAG [50] and PLP mutants [51 , 52] . This enlarges the spectrum of myelinating disorders for which treatments based on pharmacological modulation of ERQC systems and/or UPR might be a promising therapy [36] . Here we show that absence of Derlin-2 dramatically exacerbates the S63del-CMT1B neuropathy , pointing at the enhancement of ERAD as a potential strategy to treat the disease . In line with this , overexpression of Derlin-2 was shown to protect renal podocytes from apoptosis caused by ER dysfunction [53] and upregulation of EDEM was shown to preserve against ER proteotoxicity and age-related physiological decline in Drosophila [54] . In addition , pharmacological inhibition of USP14 , a deubiquitinating enzyme , was shown to enhance the proteasomal degradation of some disease-associated misfolded proteins in cultured cells [55 , 56] and to increase the rate of general proteolysis in S63del sciatic nerves ex vivo [16] . However , USP14 inhibitors proved extremely toxic when used in myelinating DRG explants , hampering the possibility to test their effects on S63del myelination . Still , here we show that treatments of S63del DRG explants with GlcNAc , a HBP metabolite known to enhance ERAD , the proteasome and autophagy in C . elegans [33] , ameliorated the extent of S63del myelination and reduced the levels of ER-stress . It should be noted that GlcNAc administration might also increase lipid synthesis and accumulation [57] , which are known to be downregulated in S63del mice [11] . Further studies are needed to assess whether and how administration of GlcNAc can alleviate the disease in mouse models of conformational neuropathies . Although the proteasome regulates the levels of wild type PMP22 [58 , 59] and of some PMP22 mutants [48] , the involvement of ERAD in myelination has been poorly investigated and its role in the degradation of ER-retained myelin proteins is only currently emerging [50 , 60] . Of note , ablation of BiP and CNX , ER retention factors upstream of ERAD , has been shown to cause myelin abnormalities in both the PNS and the CNS [61 , 62 , 63] . Here we show that Schwann cell-specific deletion of Derlin-2 impairs ERAD in WT nerves , but does not affect developmental myelination or remyelination after nerve injury . Despite this , P28 Der2SCKO sciatic nerves display moderate ER stress and transient UPR induction . This suggests that low levels of stress can be tolerated by the developing Schwann cells , in which ERQC pathways and stress responses might be sufficiently tonic to cope with the reduced protein degradation efficiency , similarly to B-cell and hepatocytes [23] . Aged Der2SCKO mice instead develop a demyelinating neuropathy , indicating that an efficient ERAD becomes important in adulthood to preserve myelin integrity . In humans , peripheral neuropathies associated to aging are highly diffuse but their pathogenetic mechanisms are still largely unknown [64 , 65] . Our data , together with the observation that aged tissues , including peripheral nerves , encounter a natural decline in expression and performance of general proteostatic networks [11 , 66 , 67 , 68 , 69] , make it tempting to speculate that ERQC failure could be one of the factors that contribute to the onset of age-related neuropathies . Indeed , in line with the idea of a decline in the efficiency of stress responses , in adult Der2SCKO sciatic nerves the UPR fails to be activated . Only the phosphorylation of eIF2α remains sustained although its downstream target CHOP is not significantly induced . This could suggest that PERK , upstream of the P-eIF2α/CHOP arm of the UPR , might not be the only kinase responsible for eIF2α phosphorylation in Der2SCKO nerves . Several works indicated that the GCN2/P-eIF2α branch of the integrated stress response ( ISR ) [70] , mainly activated upon amino acid starvation and UV irradiation , can be , in some cases , a less potent inducer of CHOP as compared to PERK [71 , 72] . Moreover , proteasome inhibition was shown to activate the ISR because of reduced amino acid recycling [73] . Thus , it is reasonable to imagine that impaired ER dislocation may limit the amount of proteins destined for proteasomal degradation , rendering the recycling of amino acids less efficient and favoring the activation of the GCN2/P-eIF2α branch in Derlin-2 deficient Schwann cells . Our data also suggest that motor fibers appear to rely on an efficient ERAD to greater extent as compared to sensory fibers , although the reason for this difference is unknown . A possibility is that different subpopulations of myelinating Schwann cells might exist , capable of re-adjusting their phenotypes depending on the type of fiber they myelinate and , thus , showing different susceptibility to ERQC failure . This would be in agreement with what was suggested by Höke et al . and Brushart et al . , who showed that myelinating Schwann cells express distinct phenotypes at least for what concerns the pattern of growth factors production [74] and central-peripheral location [75] . Alternatively , these differences might be intrinsic to each nerve and depend on its anatomy , origin , function and/or usage . Currently , effective therapies for CMT and age-related neuropathies are missing . Our data allow envisaging two , not mutually exclusive , approaches to target ERQC for therapeutic intervention in ER-stress related CMTs: the attenuation of known maladaptive arms of the UPR [11 , 76] and the enhancement of adaptive pathways , such as ERAD . Similarly , interventions aimed at modulating the proteostasis network may prove beneficial also for age-related neuropathies . Further efforts in the discovery of drugs able to potentiate ERQC pathways are therefore highly desirable .
All experiments involving animals were performed in accordance with Italian national regulations and covered by experimental protocols reviewed by local Institutional Animal Care and Use Committees . Approval number 359/2015-PR . S63del ( S63del-L line ) , P0Cre and Derlin-2fl/fl mice have been previously described [7 , 23 , 24] . S63del mice were maintained on FVB/N genetic background , whereas P0Cre and Derlin-2fl/fl lines were maintained on C57/BL6N background ( Charles River , Calco , Italy ) . Der2SCKO ( P0Cre//Der2fl/fl ) and S63del//Der2SCKO ( S63del//P0Cre//Der2fl/fl ) mice ( FVB//C57BL6—F2 generation ) were obtained by crossing S63del//Der2fl/+ ( FVB//C57BL6—F1 generation ) and P0Cre//Der2fl/fl or fl/+ mice ( C57/BL6N ) . Age matched S63del ( S63del//Der2fl/fl or +/+ ) and WT ( Der2fl/fl or +/+ ) littermates were used as controls . For genotyping and evaluation of P0Cre-mediated recombination , genomic DNA was extracted from tail , sciatic nerve , skeletal muscle , brain , heart , spleen and kidney . All PCR products were stained with SYBR Safe DNA Gel Stain ( Invitrogen ) , run in 2% agarose gels and detected with UVP GelDOC-It Imaging System . PCR protocols for genotyping of S63del and P0Cre mice were previously described [7 , 24]; Derlin-2 PCR primer sequences were: 5’-GGTTCATGCAGACAAACCATGATCGC-3’; 5’AGAGTGAAATGGCAGTTGGGTGTG-3’; 5’-GCTTTCACAAACCTGCAAGCTCCT-3’ DRG explants were isolated from E13 . 5 embryos , seeded on rat collagen I-coated coverslips and maintained in culture as previously described [77] . Myelination was induced with 50μg/ml ascorbic acid ( Sigma-Aldrich ) added to culture medium . Treatments with N-acetyl-D-Glucosamine ( GlcNAc; Sigma-Aldrich ) , dissolved in culture medium , were performed for 2 or 3 weeks in parallel to myelination induction . Culture medium was refreshed every two days . Samples were fixed and prepared for immunofluorescence and the average number and length of MBP+ internodes per field were measured with NIH Image-J software . 8 non-overlapping images per DRG were acquired with a Leica DM5000 microscope ( 10x and 20x objectives ) equipped with a Leica DFC480 digital camera . At least 3 independent dissections were performed . Inducible cell lines where generated using the Flp‐In T‐REx system ( Invitrogen ) . HA-tagged P0-wt , P0-S63C and P0-S63del cDNAs were cloned in Hind III ‐ EcoR V sites in the pCDNA5/FRT/T0 vector . The constructs were subjected to DNA sequencing to confirm the DNA preparations . To obtain the inducible cell lines , Flp‐In T‐REx HEK293 cells were co‐transfected with pCDNA5/FRT/T0 plasmids encoding the P0 variants and with a pOG44 construct that constitutively expresses the Flp recombinase . Cells were selected using medium supplemented with 150μg/ml hygromycin and 15μg/ml blasticidin . Flp‐In T‐REx HEK293 cells expressing the P0 proteins were cultured in DMEM supplemented with 10% FBS , 150μg/ml hygromycin and 15μg/ml blasticidin . Induction of P0s transgenes expression was obtained by adding 100ng/ml tetracycline to cell culture medium . Nerves were freshly dissected , fixed in 2% glutaraldehyde in phosphate buffer , osmicated in 1% OsO4 , alcohol dehydrated , infiltrated with propylene oxide and embedded in Epon . Transverse semithin sections and ultrathin sections were cut with an Ultracut microtome [11 , 78] . Semithin sections were stained with toluidine blue and acquired with a Leica DM5000 microscope equipped with a DFC480 digital camera , whereas ultrathin sections were stained with lead citrate and photographed with a Zeiss ( Oberkochen , Germany ) EM10 electron microscope . g-ratio ( axon diameter/fiber diameter ) was measured on semithin sections with semi-automated computer based morphometric analysis using Leica QWin V3 software [11]; four-six images per nerve were acquired with a 100x objective; ~800–2000 fibers per condition were measured . On EM images , g-ratio was measured using ImageJ software . 50–70 myelinated fibers from 10–12 images per animal were analyzed , from three mice per genotype . The number of demyelinated ( naked ) axons was counted blind to genotype on images acquired with a 100x objective from sciatic nerve semithin sections . The number of onion bulbs was measured on entire quadriceps nerves: single images were acquired with a 40x objective and nerves were reconstructed with Adobe-Photoshop CS4 ( Adobe Systems , San Jose , CA ) . Three-five animals per genotype were used . Electrophysiological tests were performed using an EMG system ( NeuroMep Micro , Neurosoft , Russia ) . Mice were anesthetized and placed under a heating lamp to maintain a constant body temperature . Monopolar needle electrodes were inserted subcutaneously to stimulate the tibial nerve at the ankle and , subsequently , the sciatic nerve at the sciatic notch; the cathode was placed close to the nerve and the anode was inserted proximally to the cathode . The stimulation consisted of single 100μs , 1Hz supramaximal pulses . The muscular response was recorded by inserting the active electrode into muscles in the middle of the paw and the reference electrode in the skin between the first and second digit . NCV ( m/s ) , peak-to-peak CMAP amplitude ( mV ) and FWL ( ms ) were measured . FWL measurement was obtained by stimulating the tibial nerve at the ankle and recording the responses in the paw muscles , using the same pair of needle electrodes used for the nerve conduction study [79] . Nerves were freshly dissected , desheated in PBS and teased after 20 min of fixation with 4% paraformaldehyde ( PFA ) . Nerve fibers were gently separated , let adhere onto slides and stored at -80°C until immunofluorescence . Images were acquired with Volocity Software at Perkin Elmer Ultraview ERS Confocal microscope with a 63x objective and processed with Adobe Photoshop CS4 ( Adobe Systems , San Jose , CA ) . For cryosections preparation , sciatic nerves were immediately embedded in Killik cryostat embedding medium ( Bio-Optica ) , frozen in liquid nitrogen and stored at -80°C until analysis . The following rabbit antibodies recognized PMP22 ( 1:10000; Abcam ) , GRP78/BiP ( 1:1000; Stressgene ) , Calnexin ( 1:2000; Sigma Aldrich; 1:3000 WB—1:500 IP -1:100 IF; kind gift by A . Helenius ) , OS9 ( 1:10000 , Abcam ) , eIF2α and P-eIF2α ( 1:2000; Cell Signalling XP-Technology ) , Ubiquitin ( 1:2000 , Dako ) . Rabbit anti-Derlin-1 and anti-Derlin-2 antibodies were received from H . L . Ploegh [23] . Rabbit antibody against HERP was a generous gift of K . Kokame . Rat antibodies recognized GRP94 ( 1:2000; Abcam ) and MBP ( 1:5 ) . Chicken monoclonal antibody recognized Neurofilament-M ( 1:1000; BioLegend ) . Mouse antibodies recognized ß-Tubulin ( 1:5000/1:10000; Sigma Aldrich ) , HA tag ( 1:500–1:100; Hybridoma 12CA5; Santa Cruz ) and KDEL ( 1:200 IF , ENZO Life Sciences; 1:700–500 WB-IP , Stressgene ) . HEK293 cells were plated on glass poly‐lysine coated coverslips . After 17 hr of induction , cells were fixed using 3 , 7% formaldehyde and blocked with 10% goat serum . Primary and secondary antibodies , diluted in 10% goat serum , were incubated for 2 hr and 30 min respectively . Microscopy images were collected using a laser scanning confocal microscope ( Leica DI6000 microscope stand , SP5 scan head ) equipped with a HCX PL APO CS 63X oil UV objective . DRG explants , sciatic nerve cryosections and teased nerve fibers were fixed for 15 min with 4% PFA and permeabilized with ice-cold methanol or 0 . 1% Triton-X100 ( Sigma ) in blocking solution . Samples were blocked in normal goat serum ( NGS; Dako ) /1% bovine serum albumine ( BSA; Sigma ) /PBS for 1 hr at room temperature ( RT ) . Primary and secondary antibodies were diluted in 1% BSA . Primary antibodies were incubated 1 hr RT or overnight at 4°C , whereas secondary antibodies 45 min at RT in dark condition . Nuclei were marked using Hoechst or DAPI . Samples were mounted onto slides with Vectashield mounting medium ( Vector Laboratories ) . For siRNA-based interference , Flp‐In T‐REx HEK293 cells expressing the P0-S63del protein were grown in DMEM supplemented with 10% FBS . Cells at 50% confluence were transfected with 50 pmol/dish siRNA duplex ( hs_DERL2 FlexiTube siRNA Qiagen ) using Lipofectamine 2000 according to the manufacturer’s instructions . 30 h after siRNA transfection , the expression of P0-S63del was induced by adding 100ng/ml tetracycline to cell culture medium . 48 h after siRNA transfection , cells were lysed or subjected to pulse-chase analysis as described below . Induced cells were washed with PBS and incubated with starving medium ( DMEM , 50mM Hepes , 1% Glutamax ) for 10 min at 37°C . [35S]‐methionine/cysteine mix ( SIGMA‐Aldrich ) was directly added to a final concentration of 0 . 2mCi/ml and cells were pulsed for 10 min . Label medium was removed and cells were chased in DMEM supplemented with 5mM non‐labeled methionine/cysteine . Cells were washed with PBS containing 20mM N-ethyl-maleimide ( NEM ) for 1 min and then lysed with 2% CHAPS ( Anatrace ) in HEPES‐buffered saline ( HBS ) , pH 6 . 8 , supplemented with 20mM NEM and protease inhibitors for 20 min on ice . Supernatants were collected by centrifugation at 4°C/10000xg for 10 min , immunoprecipitated and subjected to SDS‐PAGE as described below . After exposure of the gels to autoradiography films ( GE Healthcare , Fuji ) , films were scanned with the Typhoon FLA 9500 ( Software Version 1 . 0 ) . For immunoprecipitation , cell lysates were incubated with protein-A beads ( SIGMA , 1:10 , w/v swollen in PBS ) and the specific antibody . After 90 min , the immunocomplexes were washed with HBS , 0 . 5% CHAPS , pH 6 . 8 . Beads were resuspended in sample buffer and denatured for 10 min at 65°C . Samples were subjected to SDS‐PAGE ( see below ) . Peripheral nerves were dissected and frozen in liquid nitrogen . Frozen nerves were pulverized on dry ice and proteins were extracted in denaturing lysis buffer ( Tris/HCl 50mM PH7 . 5 , NaCl 150mM , EDTA 10mM , 2% SDS ) containing protease inhibitor cocktail ( PIC 100X roche ) , Na3VO4 and NaF . Total protein concentration was determined by BCA assay ( Pierce ) following manufacturer’s instructions . Equal amounts of proteins were separated by SDS-PAGE ( Biorad ) and gels were transferred onto nitrocellulose membrane ( GE Healthcare ) . Membranes were blocked with 5% milk ( milk powder/1x PBS-Tween 0 . 05% ) and incubated with primary antibodies diluted in 5% Milk or 5% BSA/1x PBS-Tween 0 . 05% at 4°C overnight . HRP-conjugated antibodies were diluted in 5% Milk/1x PBS-Tween 0 . 05% and incubated 1 hr RT . Signals were detected by ECL method and autoradiography film ( GE Healthcare ) with Classic E . O . S . AGFA Developer Machine . Densitometric analysis was performed with NIH-Image-J software . For inducible cells , protein samples were prepared as described above and separated in SDS‐PAGE under reducing conditions after boiling in DTT‐containing sample buffer for 10 min at 65°C . Membranes were developed using the Luminata Forte ECL detection system ( Millipore ) and signals were detected with the ImageQuant LAS 4000 system in the standard acquisition mode ( GE Healthcare Life Science ) . Bands were quantified using Multi Gauge Analysis tool ( Fujifilm ) . The linearity of the detected signal range was ensured with appropriate loading controls . Total RNA was extracted with Trizol ( Roche Diagnostic GmbH , Germany ) and retrotranscribed as previously described [7] . TaqMan assays were performed following manufacturer’s instructions ( TaqMan , PE Applied Biosystems Instruments ) on an ABI PRISM 7700 sequence detection system ( Applied Biosystems Instruments ) [11 , 12] . Normalization was performed using 18S rRNA as reference gene . Target and reference genes PCR amplification were performed in separate tubes with Assay on Demand ( Applied Biosystems Instruments ) : 18S assay , Hs99999901_s1; Ddit3/Chop assay , Mm00492097_m1; Xbp-1s assay , Mm03464496_m1; Hspa5/BiP assay , Mm00517691_m1; Derl3 assay , Mm00508292_m1; Derl2 assay , Mm01245788_m1; Derl1 assay , Mm00470296_g1; Sel1L assay , Mm01326442_m1; HRD1/SYVN1 assay , Mm00511995_m1; EDEM1 assay , Mm00551797_m1; OS9 assay , Mm00617153_m1; Herpud1 assay , Mm00445600_m1 . Sample size was not predetermined with any statistical method , but our sample size is similar to that generally used in the field . Graphs and data were analyzed using GraphPad Prism Software and/or Microsoft Excel . Data show the mean ± Standard Error of Mean ( SEM ) . Unpaired , 2 tails , Student’s t test or One-way ANOVA with Tukey’s post hoc test were used as specified in the figure legends; significance levels ( P values ) were marked on figures as follows: *P ≤ 0 . 05 , **P ≤ 0 . 01 , ***P ≤ 0 . 001; only comparisons between WT vs Der2SCKO , WT vs S63del and S63del vs S63del//Der2SCKO groups are illustrated in all figures . | Charcot-Marie-Tooth neuropathies are a large family of peripheral nerve disorders , showing extensive clinical and genetic heterogeneity . Although strong advances have been made in the identification of genes and mutations involved , effective therapies are still lacking . Intracellular retention of abnormal proteins has been recently suggested as one of the pathogenetic events that might underlie several conformational neuropathies . To limit the toxic effects of accumulated mutant proteins , cells have developed efficient protein quality control systems aimed at optimizing both protein folding and degradation . Here we show that ER-associated degradation limits Schwann cells stress and myelin defects caused by the accumulation of a mutant myelin protein into the ER . In addition , we also describe for the first time the importance of Schwann cells ERAD in preserving myelin integrity in adult nerves , showing that genetic ERAD impairment leads to a late onset , motor-predominant , peripheral neuropathy in vivo . Effort in the design of strategies that potentiate ERAD and ER quality controls is therefore highly desirable . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [] | 2019 | Schwann cells ER-associated degradation contributes to myelin maintenance in adult nerves and limits demyelination in CMT1B mice |
Treatment and morbidity control of schistosomiasis relies on a single drug , praziquantel . Hence , there is a pressing need to develop additional therapeutics against schistosomiasis . The antimalarial drug mefloquine shows antischistosomal activity in animal models and clinical trials , which calls for further investigations . We comparatively assessed the efficacy and tolerability of the following treatments against Schistosoma haematobium in school-aged children in Côte d'Ivoire: ( i ) praziquantel ( 40 mg/kg; standard treatment ) ; ( ii ) mefloquine ( 25 mg/kg ) combined with praziquantel ( 40 mg/kg ) ; and ( iii ) mefloquine-artesunate ( 3× ( 100 mg artesunate +250 mg mefloquine ) ) combined with praziquantel ( 40 mg/kg ) ( treatments administered on subsequent days ) . Two urine samples were collected before , and on days 21–22 and 78–79 after the first dosing . Sixty-one children were present on all examination time points and had complete datasets . No difference in efficacy was observed between the three treatment groups on either follow-up . On the 21–22 day posttreatment follow-up , based on available case analysis , cure rates of 33% ( 95% confidence interval ( CI ) 11–55% ) , 29% ( 95% CI 8–50% ) , and 26% ( 95% CI 5–48% ) were observed for praziquantel , mefloquine-artesunate-praziquantel , and mefloquine-praziquantel , respectively . The corresponding egg reduction rates were 94% and above . On the second follow-up , observed cure rates ranged from 19% ( praziquantel ) to 33% ( mefloquine-artesunate-praziquantel ) , and egg reduction rates were above 90% . Praziquantel monotherapy was the best tolerated treatment . In the mefloquine-artesunate-praziquantel group , adverse events were reported by 91% of the participants , and in the mefloquine-praziquantel group , 95% experienced adverse events . With the exception of abdominal pain at moderate severity , adverse events were mild . The addition of mefloquine or mefloquine-artesunate does not increase the efficacy of praziquantel against chronic S . haematobium infection . Additional studies are necessary to elucidate the effect of the combinations against acute schistosomiasis .
Schistosomiasis is a neglected tropical disease caused by a chronic infection with blood-dwelling parasitic flatworms of the genus Schistosoma [1] . More than 230 million people in the tropics and subtropics are infected and the global burden of schistosomiasis is estimated at 3 . 3 million disability-adjusted life years ( DALYs ) [2] , [3] . The main strategy to control schistosomiasis is preventive chemotherapy that is the periodic administration of the antischistosomal drug praziquantel to populations at-risk of morbidity , most importantly school-aged children [4]–[6] . Although no clinically relevant resistance to praziquantel has been documented thus far , reliance on a single drug is a risky endeavor [7] , [8] . Moreover , praziquantel is quite ineffective against the young developing stages of schistosomes [9] , [10] . To address this inherent shortcoming of praziquantel , treatment of acute infections will have to be repeated or postponed until worms will have matured [1] . It is thus essential to find additional therapeutics against schistosomiasis , ideally compounds that are active against all stages of the parasite . There is presently no other broad-spectrum antischistosomal drug available and the drug development pipeline is empty . Against this background and taking into consideration scarce resources for research and development of neglected tropical diseases , repurposing of drugs that are already approved for human use is a promising strategy [11] . Indeed , such a strategy is more rapid , less risky , and less costly than developing new drugs [12] , [13] . Since 2008 , the antimalarial drug mefloquine is undergoing detailed in vitro , in vivo , and clinical investigation for its trematocidal properties . For example , in the Schistosoma mansoni-mouse model , mefloquine exhibited high worm burden reductions following single-dose regimen against juvenile and adult schistosomes [14]–[16] . Mefloquine also revealed a high activity against the other two major human schistosome species , S . haematobium [17] and S . japonicum [14] , [18] . It was therefore concluded that mefloquine has a similarly broad spectrum of activity than praziquantel . In an exploratory trial in Côte d'Ivoire , a mefloquine-artesunate combination showed a moderate cure rate ( 61% ) and high egg reduction rate ( 96% ) in school-aged children infected with S . haematobium [19] . Recently , mefloquine , used as intermittent preventive therapy against malaria in pregnancy ( IPTp ) , showed high egg reduction rates in women with a concomitant S . haematobium infection [20] . However , combination therapy with mefloquine and praziquantel , which showed high worm burden reductions in laboratory animals [21] , has not yet been studied in Schistosoma-infected patients . The aim of the current study was to assess the efficacy and tolerability of mefloquine and mefloquine-artesunate combined with praziquantel against S . haematobium in school-aged children . Since prior in vivo studies revealed synergistic effects when mefloquine and praziquantel were administered on subsequent days , and drug interaction between mefloquine and praziquantel have not been studied before , drug administration of the antimalarials and praziquantel was spaced by a day . For comparison , one group of children was treated with praziquantel only , using the current standard dose of 40 mg/kg . Treatment outcomes were assessed twice , on days 21–22 and 78–79 after the first dosing to determine the effect against pre-patent and patent S . haematobium infection .
Ethical clearance was obtained by the ethics committee in Basel ( EKBB; reference no . 70/08 ) and the Ministère de la Santé et de l'Hygiène Publique en Côte d'Ivoire . Parents/guardians of participating children signed a written informed consent for their children , and children assented orally . Participation was voluntary and the children were informed that they could withdraw anytime without further obligation . The trial is registered with Current Controlled Trials ( ISRCTN00393859 ) . The study was carried out in Sahoua , a village in the Taabo district , located about 170 km north-west of Abidjan , the economic capital of Côte d'Ivoire . Sahoua is situated in the V-Baoulé , the transition zone between the rainforest in the South and the Savannah in the North , at the north-western edge of the Taabo health and demographic surveillance system ( HDSS ) [22] . The climate is tropical with the main rains occurring between April and July and in September/October . People are primarily engaged in subsistence farming ( e . g . , cassava , plantains , and yams ) , whilst cacao is the predominant cash crop . The village of Sahoua is close to the Bandama River . The inhabitants coming from neighboring countries Mali and Burkina Faso are fishers . Women perform household chores ( e . g . , washing dishes or clothes ) at the water's edge . School-aged children are in frequent contact with the water during recreational activities ( e . g . , bathing and swimming ) . The field and laboratory work was carried out between November 2011 and February 2012 . The study aim was explained and approved by the local health authorities , including the director of Taabo-Cité hospital , the district health officer of Tiassalé , the village chief , and the school director . Parents/guardians of the children provided written informed consent , while children assented orally . All school children from grade 3 ( CE1 ) to 6 ( CM2 ) were invited to participate in the prescreening . A total of 130 school children provided a urine sample . Urine samples were collected between 10:00 and 14:00 hours and labeled with unique identifiers . Samples were transferred to the laboratory of the Taabo-Cité hospital for macroscopic and microscopic examination of S . haematobium . 77 children were identified as positive and invited to participate in the study . These children were asked to provide an additional urine sample and a stool sample the next day . Those children who had complete parasitological datasets were invited for a clinical examination , which included a physical examination , weight measurement ( using an electronic balance recording to the nearest 0 . 1 kg ) , assessing temperature ( using battery-powered ear thermometers to the nearest 0 . 01°C ) , and a finger-prick blood sample . From the blood sample , hemoglobin concentration was determined using a portable Hemocue 301 ( HemoCue AB; Ängelholm , Sweden ) . Additionally , thick and thin blood films were prepared on microscope slides , labeled with unique identification numbers , and air-dried . Children were excluded if any of the following criteria were met: ( i ) fever ( temperature ≥37 . 5°C ) ; ( ii ) pregnancy first trimester assessed verbally; ( iii ) presence of any abnormal medical condition , judged by the study physician; ( iv ) history of acute or severe chronic disease; ( v ) psychiatric disorders such as epilepsy; ( vi ) recent use of anthelmintic or antimalarial drugs ( within the past month ) ; and ( vii ) weight below 20 kg . S . haematobium-infected children who were excluded from the study were offered praziquantel ( 40 mg/kg ) free of charge . Mefloquine ( 250 mg lactabs ) , and mefloquine-artesunate blisters containing 3×100 mg artesunate and 3×250 mg mefloquine were purchased from Viktoria Apotheke ( Zurich , Switzerland ) . Praziquantel ( 600 mg tablets ) was purchased from Inresa ( Bartenheim , France ) . Children included in the study received one of three treatments under direct medical observation , following a computer-generated randomization code: ( i ) mefloquine 25 mg/kg single dose ( body weight <30 kg ) or a split dose spaced by 6 hours ( body weight ≥30 kg ) plus a single dose of praziquantel ( 40 mg/kg ) on the next day; ( ii ) mefloquine-artesunate ( 1×100 mg artesunate and 1×250 mg mefloquine ) once daily for 3 consecutive days plus a single dose of praziquantel ( 40 mg/kg ) on treatment day 4; and ( iii ) praziquantel , standard single dose ( 40 mg/kg ) . Mefloquine and praziquantel were administered to the nearest half tablet according to the calculated dose per kg of body weight . All participating children received a snack shortly after drug administration . Children were kept for observation and interviewed for the presence of acute adverse events 3 hours posttreatment . In addition , adverse events were assessed 24 hours after each treatment dose ( prior to the next dose for the mefloquine-artesunate-praziquantel and mefloquine-praziquantel treatment groups ) . Adverse events were graded ( i . e . , mild , moderate , severe , and life-threatening ) , and symptomatic relief provided if necessary . Urine samples were examined visually for macroscopic blood and then analyzed for microhematuria using reagent strips ( Hemastix , Siemens Healthcare; Zurich , Switzerland ) . For detection of S . haematobium eggs , urine samples were subjected to a filtration method [19] . In brief , samples were carefully homogenized and 10 ml of urine pressed through a 13-mm diameter filter with 25 µm pores ( Sefar AG; Heiden , Switzerland ) . The filters were placed on microscope slides , a drop of Lugol's solution added before examination under a microscope at a magnification of ×100 by two experienced technicians . Slides were re-examined by a senior technician in case of differing results among the two technicians . Stool specimens were subjected to duplicate Kato-Katz thick smears , using standard 41 . 7 mg templates [23] and examined under a microscope . The number of S . mansoni , Ascaris lumbricoides , Trichuris trichiura , hookworm , and other helminth eggs were counted and recorded for each species separately . Additionally , approximately 2 g of stool was preserved in sodium acetate-acetic acid-formalin ( SAF ) , and processed with an ether-concentration method [24] . Samples were examined microscopically at a magnification of ×100 for helminths , and at a magnification of ×400 for intestinal protozoa ( e . g . , Blastocystis hominis , Chilomastix mesnili , Endolimax nana , Entamoeba coli , Entamoeba hartmanni , Entamoeba histolytica/E . dispar , Giardia intestinalis , and Jodamoeba bütschlii ) . Thick and thin blood films were stained with Giemsa , and prepared and read as described elsewhere [19] . Parasite counts were documented as the number of Plasmodium per µl of blood , assuming a standard count of 8 , 000 white blood cells per µl of blood . By definition , pilot studies are conducted to serve as a starting point for further studies and are primarily intended to yield information about the feasibility and implementation possibilities of novel treatments . Thus , the choice of an adequate sample size for a pilot study is mainly based on practical considerations of the pilot trial rather than on statistical sample size calculations [25] . Allowing for up to 50% drop-outs , we aimed for 23–25 children per treatment arm . Data were double entered into Excel and Access ( Microsoft 2010 ) , cross-checked , and analyzed using Stata version 10 . 1 ( StataCorp . ; College Station , United States of America ) and Statsdirect version 2 . 7 . 9 ( Statsdirect , Chesire , United Kingdom ) . All children with primary endpoint data were included in the analysis ( available case analysis ) . S . haematobium egg counts from the two urine samples were averaged for each child ( arithmetic mean ( AM ) ) and the AM and geometric mean ( GM ) egg count for each treatment group calculated . Cure rate ( percentage of children excreting no S . haematobium eggs at the posttreatment follow-ups ( i . e . , 21–22 and 78–79 days after drug administration ) among parasitological-confirmed children at baseline ) and egg reduction rate ( reduction of AM and GM egg count among S . haematobium-positive children posttreatment compared to the respective AM or GM pretreatment ) were calculated . Bootstrap resampling method with 10 , 000 replicates was used to calculate 95% confidence intervals ( CIs ) for egg reduction rates of GM [26] . Differences in egg reduction rates were determined under the assumption that non-overlapping CIs indicate statistical significance . To test whether there was an association between cure rates and dose , the actual doses administered were determined and analyzed using logistic regression . To compare baseline and follow-up parameters , Mann-Whitney U test and Wilcoxon matched pairs test were used , as appropriate . Pearson's χ2 was used to compare the proportion of reported adverse events between treatment arms .
Overall , 71 S . haematobium-infected children were randomized to the three treatment arms ( Figure 1 ) . Ten children were lost at the first or second follow-up , mainly because of travels at the time of the surveys . Demographic and clinical baseline characteristics of the 61 children included in the available case analysis are summarized in Table 1 . Treatment groups were well balanced in terms of age ( mean age: 10 . 4–10 . 9 years ) , weight ( mean weight: 28 . 0–29 . 1 kg ) , and height ( mean height: 138 . 8–139 . 9 cm ) . However , more boys ( n = 47 ) than girls ( n = 14 ) participated in the trial . Hemoglobin values were in the normal range . Most of the included children suffered from heavy S . haematobium infection , as defined by ≥50 S . haematobium eggs per 10 ml of urine ( n = 36 , 59% ) . No difference was observed in infection intensity between the three treatment groups; the GM of S . haematobium eggs per 10 ml of urine ranged from 45 . 2 to 61 . 8 . Macroscopic examination of urine revealed visible blood in 33 samples ( 46% ) . Indirect screening approaches based on reagent strips to detect microhematuria and proteinuria , revealed prevalences of 82% and 85% , respectively . Most children were coinfected with Plasmodium falciparum . No infection with S . mansoni was diagnosed . Coinfections with hookworm were observed in three children . Intestinal protozoa infections were common; Endolimax nana and Entamoeba coli were the predominant species in all treatment groups . Our results showed no difference in cure rates against S . haematobium among the three treatment arms , at neither treatment follow-up . At the first follow-up 21–22 days posttreatment , we observed low cure rates; namely , 26% ( 95% CI 5–48% ) for mefloquine plus praziquantel , 29% ( 95% CI 8–50% ) for mefloquine-artesunate plus praziquantel , and 33% ( 95% CI 11–55% ) for praziquantel monotherapy ( Table 2 ) . Cure rates were higher among children with light-intensity S . haematobium infections ( 38–67% ) . Cure rates determined with indirect screening approaches by visual inspection of urine for macrohematuria and reagent strip testing for microhematuria were as high as 71% . No significant association between cure rates and exact dose was observed . In the praziquantel treatment arm , the exact dose of praziquantel administered ranged from 33 to 45 mg/kg; in the mefloquine-praziquantel treatment arm , the exact dose of mefloquine ranged from 22 to 25 mg/kg and that of praziquantel from 33 to 45 mg/kg; in the mefloquine-artesunate-praziquantel treatment arm , the exact dose of mefloquine ranged from 13 to 36 mg/kg , that of artesunate from 5 . 4 to 14 . 2 mg/kg , and the administered dose of praziquantel was 33–43 mg/kg . High egg reduction rates ( 94–96% based on the GM eggs per 1 g of stool ( EPG ) ) were observed for the three treatments against S . haematobium at the first follow-up . Additionally , all children treated with mefloquine-artesunate plus praziquantel were cured from Plasmodium infections . Seventeen out of 18 children treated with mefloquine plus praziquantel had a negative laboratory diagnosis of Plasmodium . The two groups treated with antimalarials had significantly higher hemoglobin values ( p<0 . 05; mean increase of 0 . 57 g/dl hemoglobin in mefloquine-praziquantel and 1 . 15 g/dl in mefloquine-artesunate plus praziquantel treated children ) in contrast to children treated with praziquantel alone ( mean increase of hemoglobin 0 . 36 g/dl ) . As expected , no effect on concomitant Plasmodium infection was observed in children who were treated with praziquantel singly . The prevalence of intestinal protozoa infection was similar at baseline and the first treatment follow-up . At the second follow-up 78–79 days posttreatment , cure rates ranged from 19% ( 95% CI 1–37% ) ( praziquantel ) to 33% ( 95% CI 11–55% ) ( mefloquine-artesunate plus praziquantel ) ( Table 3 ) . No difference was observed between the three treatment arms and cure rates . Additionally , no difference was observed between cure rates at the first and second follow-up . Egg reduction rates were high ( 92–94% based on the GM ) . Visual inspection and reagent strip analysis resulted in cure rates of 53–67% and 47–62% , respectively . Eleven children in the antimalarial treatment groups had re-acquired a malaria infection . Hemoglobin levels in children treated with mefloquine-artesunate plus praziquantel remained significantly higher compared to baseline values ( 12 . 9 versus 11 . 6 g/dl; p<0 . 001 ) . In addition , infection with E . coli were most commonly observed ( 37 children ) , followed by E . nana and hookworm infections ( 24 and 17 children , respectively ) . During the clinical examination at baseline 20 of the 61 participating children reported symptoms , mainly headache and abdominal pain . The number of children experiencing adverse events , stratified by treatment arm at each of the two treatment follow-ups , is summarized in Table 4 . Table 5 presents the number of specific mild and moderate adverse events , observed in each treatment arm assessed at different examination time points . We did not observe any life-threatening adverse events following treatment and , with the exception of abdominal pain at moderate severity ( 12 children ) , adverse events were mild . Praziquantel and mefloquine-artesunate were significantly better tolerated than mefloquine ( p<0 . 05 ) , as assessed 24 hours posttreatment . Nearly all children treated with mefloquine-artesunate plus praziquantel ( 91% ) and mefloquine plus praziquantel ( 95% ) experienced adverse events over the four respectively two treatment days . More than half of the children ( 56% ) stated adverse events following a single dose of praziquantel . Abdominal pain ( mild and moderate episodes ) was the most commonly observed adverse event in all treatment groups . Children treated with mefloquine-artesunate plus praziquantel and mefloquine plus praziquantel reported significantly more mild abdominal pain than children treated with praziquantel ( p<0 . 05 ) . Vomiting was also commonly reported by children treated with mefloquine-artesunate plus praziquantel ( p<0 . 05 ) and mefloquine plus praziquantel . Other common adverse events in all treatment groups included vertigo , headache , and diarrhea .
Reliance on a single drug for individual treatment and community-based morbidity control of schistosomiasis – one of the most important parasitic diseases in sub-Saharan Africa – bears the risk of parasites developing resistance . No alternative antischistosomal drugs are in the development pipeline . Oxamniquine and metrifonate – two drugs that have been widely used against S . mansoni and S . haematobium , respectively – are ( with the exception of oxamniquine in Brazil ) no longer commercially available [27]– . A promising approach for identifying new drugs against schistosomiasis is to repurpose existing drugs that are already on the market for the treatment of other diseases . This strategy is popular in many medical fields , including tuberculosis [30] , cancer [31] , and malaria [32] . In fact , a recent analysis of the research and development landscape of drugs and vaccines for neglected diseases from 2000 to 2011 showed that most new drugs in this therapeutic area are repurposed versions of existing products [11] . In the present exploratory trial , we assessed whether antimalarials ( mefloquine and mefloquine-artesunate ) plus praziquantel have a higher efficacy than standard single-dose praziquantel . Mefloquine and mefloquine-artesunate combination were selected as combination partner for praziquantel since laboratory studies have shown synergistic effect for mefloquine-praziquantel combinations in vitro and in vivo [21] . Furthermore , stage-specific susceptibility studies have shown that , in contrast to the biphasic activity of praziquantel , juvenile worms are particularly vulnerable to mefloquine and the artemisinins [14] , [33] . Hence , we hypothesized that a mefloquine-praziquantel combination has an increased spectrum of activity compared to praziquantel alone . Note that drugs were administered on consecutive days as drug interactions have not been studied to date and the treatment schedule administering the antimalarials prior to praziquantel had achieved the highest activity in vivo [21] . At the first posttreatment follow-up 21–22 days after drug administration , a marked reduction in the intensity of infection with high egg reduction rates ( 94–96% ) but low cure rates ( 26–33% ) were observed in the three treatment groups . We were surprised about the low cure rates achieved by praziquantel , although previous studies also reported low cure rates when administering praziquantel against S . haematobium ( e . g . , 40% in Cameroon [34] and 37% in Mali [35] ) . As described before [6] , [34] , these low cure rates most likely reflect that children treated with praziquantel had high infection intensities prior to drug administration . However , most prior studies have reported higher cure rates . For example , Stothard and colleagues recently reviewed the literature and meta-analyzed the data , which revealed an overall cure rate of 70% in response to a single dose of praziquantel against S . haematobium [6] . Unexpectedly , the co-administration of either mefloquine or mefloquine-artesunate with praziquantel showed similarly low cure rates than the paziquantel single treatment group . Our findings therefore contrast with previous studies . In Nigeria , a combination of praziquantel and artesunate ( using a similar treatment schedule than in the current investigation ) achieved higher cure rates and egg reduction rates compared to single praziquantel or single artesunate [36] . In addition , a previous study conducted in a nearby village , revealed a cure rate of 61% in S . haematobium-infected children treated with a mefloquine-artesunate combination [19] . Hence , since mefloquine and artesunate exhibit antischistosomal properties [19] , we expected to observe higher cure rates combining these antimalarials with praziquantel compared to praziquantel singly . A limitation of our study is that the viability of excreted eggs [37] was not determined , and hence counts of dead eggs might have been included in the analysis , and hence our reported cure rates might underestimate the true situation . At the second follow-up examination 78–79 days posttreatment , cure and egg reduction rates were comparably low as in the first follow-up . A slight ( not significant ) decrease in the estimated cure rate of praziquantel was noted ( from 33% to 19% ) . Given the small sample size and low cure rates observed already at the first follow-up , a conclusion whether the addition of mefloquine and/or artesunate would expand the activity profile of praziquantel targeting juvenile schistosomes cannot be drawn . As expected , praziquantel was the best tolerated treatment , perhaps explained by only one type of drug administered . Mefloquine-praziquantel and mefloquine-artesunate-praziquantel on the other hand were administered over 2 and 4 days , respectively . The adverse event rate calculated as the number of adverse events per group , divided by the person-time at risk in each group was similar among the treatment groups ( data not shown ) . Whether adverse events following praziquantel administration in the mefloquine-artesunate-praziquantel and mefloquine-praziquantel treated children are due to praziquantel or due to the long systemic exposure of the antimalarials is not known . Similar to our previous study [19] , most children treated with mefloquine and mefloquine-artesunate reported mild or moderate adverse events , mainly gastrointestinal complaints , including abdominal pain , nausea , and vomiting . In conclusion , our results suggest that a drug combination containing mefloquine-artesunate or mefloquine has no benefit over standard praziquantel against chronic S . haematobium infection regarding efficacy ( cure and egg reduction rate ) and safety ( frequency and severity of adverse events ) . Further studies are required to elucidate the effect of these combinations on acute schistosomiasis . There is a pressing need to develop additional antischistosomal drugs , as long as praziquantel remains efficacious against different Schistosoma species parasitizing man . | The antimalarial drug mefloquine shows activity against blood flukes that cause the disease schistosomiasis . In animal studies it has been found that a mefloquine-praziquantel combination kills blood flukes more effectively than praziquantel alone . Combining praziquantel with another drug might therefore increase efficacy , broaden the spectrum of activity , and delay the development of drug resistance . We designed a study in Ivorian school children to assess the efficacy and tolerability of mefloquine and mefloquine-artesunate combined with praziquantel against the blood fluke Schistosoma haematobium . The administration of the antimalarials and praziquantel was spaced by a day . Treatment outcomes were assessed twice , on days 21–22 and 78–79 after the first dosing to determine the effect against adult and juvenile S . haematobium , respectively . At both follow-ups , high reduction in the intensity of infection ( egg reduction rates of 94–96% ) , but low cure rates ( 26–33% ) were observed in the three treatment groups . Adverse events were common , particularly in children treated with mefloquine-praziquantel and mefloquine-artesunate-praziquantel . Our study suggests that the addition of mefloquine and mefloquine-artesunate to praziquantel has no benefit in the treatment of chronic S . haematobium infection . However , further investigations are warranted to evaluate the effect of combination therapy on juvenile flukes and longer-term morbidity profiles . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"helminth",
"infections",
"infectious",
"diseases",
"schistosomiasis",
"medicine",
"and",
"health",
"sciences",
"neglected",
"tropical",
"diseases",
"tropical",
"diseases",
"parasitic",
"diseases"
] | 2014 | Praziquantel, Mefloquine-Praziquantel, and Mefloquine-Artesunate-Praziquantel against Schistosoma haematobium: A Randomized, Exploratory, Open-Label Trial |
Hepatitis C virus ( HCV ) infection is a leading cause of liver disease worldwide . The HCV RNA genome is translated into a single polyprotein . Most of the cleavage sites in the non-structural ( NS ) polyprotein region are processed by the NS3/NS4A serine protease . The vital NS2-NS3 cleavage is catalyzed by the NS2 autoprotease . For efficient processing at the NS2/NS3 site , the NS2 cysteine protease depends on the NS3 serine protease domain . Despite its importance for the viral life cycle , the molecular details of the NS2 autoprotease activation by NS3 are poorly understood . Here , we report the identification of a conserved hydrophobic NS3 surface patch that is essential for NS2 protease activation . One residue within this surface region is also critical for RNA replication and NS5A hyperphosphorylation , two processes known to depend on functional replicase assembly . This dual function of the NS3 surface patch prompted us to reinvestigate the impact of the NS2-NS3 cleavage on NS5A hyperphosphorylation . Interestingly , NS2-NS3 cleavage turned out to be a prerequisite for NS5A hyperphosphorylation , indicating that this cleavage has to occur prior to replicase assembly . Based on our data , we propose a sequential cascade of molecular events: in uncleaved NS2-NS3 , the hydrophobic NS3 surface patch promotes NS2 protease stimulation; upon NS2-NS3 cleavage , this surface region becomes available for functional replicase assembly . This model explains why efficient NS2-3 cleavage is pivotal for HCV RNA replication . According to our model , the hydrophobic surface patch on NS3 represents a module critically involved in the temporal coordination of HCV replicase assembly .
Hepatitis C virus ( HCV ) is a single-stranded positive-sense RNA virus belonging to the family Flaviviridae and is with 170 million infected individuals worldwide an important cause of chronic liver disease [1] . The positive sense RNA genome contains a 5’-untranslated region ( UTR ) , a single open reading frame ( ORF ) that encodes both structural as well as non-structural ( NS ) viral proteins and a 3’ UTR . Cap-independent translation of the viral genome yields a single polyprotein that is co- and posttranslationally processed into the individual proteins by host signal peptidases and two viral proteases NS2-NS3 and NS3-4A . The host signal peptidases cleave at the junctions of Core/E1 , E1/E2 , E2/p7 and p7/NS2 [2–4] . The NS3/NS4A serine protease complex mediates the cleavages of the non-structural proteins NS3-NS5B [5 , 6] . The chymotrypsin-like serine protease domain residing in the N- terminal 180 amino acids of NS3 requires NS4A as a cofactor for full activity [5 , 7 , 8] . NS3 harbors downstream of the protease domain ATPase and helicase activities [9] . The NS2 protein ( 217 amino acids , aa ) is membrane-associated via its N-terminal domain that consists of three putative transmembrane segments with a perinuclear ER localization [10 , 11] . The C-terminal protease domain ( aa 94–217 ) resides on the cytoplasmic face of the ER membrane [10 , 12] and , with the N-terminal domain of NS3 , forms the NS2-NS3 autoprotease that catalyzes the cleavage at the NS2/NS3 site [7 , 13 , 14] . The putative catalytic triad of the NS2-NS3 protease resides entirely in NS2 and autocleavage at the NS2/NS3 junction is independent of the NS3 serine protease activity [7 , 15 , 16] . The NS2 protease domain is highly conserved among HCV genotypes and its crystal structure indicates that a dimer forms a composite active site with a catalytic triad analogous to those of cysteine proteases [16] . Recently , NS2 , followed by only two residues of NS3 , has been shown to be a bona fide protease exhibiting low-level intrinsic protease activity . This NS2 protease activity is stimulated by the NS3 serine protease domain ( residues 1–180 ) defining this domain as stimulatory cofactor for NS2 [17] . Cleavage between NS2 and NS3 is essential for RNA genome replication of the full-length virus and subgenomic NS2-NS5B replicons but not for NS3/4A serine protease activity [18–20] . Furthermore , the NS2 protein , but not its proteolytic activity , is required for the production of infectious virus [21–29] . The crystal structure of the NS2 protease domain represents the post-cleavage conformation , which likely differs from the one of the NS2-NS3 precursor [16] . Due to the lack of the NS2-NS3 structure , little is known about the NS2-NS3 cleavage mechanism and its regulation by NS3 . Mechanistically , it was proposed that conserved surface areas of NS2 and NS3 may interact to contribute to a functional catalytic NS2-NS3 environment and correct positioning of the scissile bond [16] . To identify molecular features that are critical for NS2 protease stimulation by NS3 , we conducted a comprehensive mutagenesis screen of the entire NS3 protease domain . We identified a conserved hydrophobic NS3 surface patch that is essential for efficient NS2 protease activation in the context of uncleaved NS2-NS3 and demonstrate that NS2 protease stimulation mainly depends on hydrophobic protein-protein interactions . Furthermore , mutational analysis of this NS3 surface patch revealed that this area is also pivotal for viral RNA replication and NS5A hyperphosphorylation . Based on these findings we propose a sequential cascade of molecular events where the hydrophobic NS3 surface patch orchestrates NS2 protease stimulation , NS5A hyperphosphorylation and viral RNA replication . Moreover , this study defines an unexpected effect of the NS2-NS3 cleavage and offers a molecular understanding why efficient NS2-NS3 cleavage and the generation of free NS3 are essential for viral RNA replication .
The NS2 protease has a low-level intrinsic ability to cleave the NS2-3 precursor protein and this function is strongly enhanced by the NS3 protease domain ( aa 1–180 ) [17 , 30] . The NS2-NS3 cleavage can be analyzed in a cell-based assay by expressing HCV NS2-NS3 polyprotein fragments from T7 promoter of pcite-FLAG-NS2-NS3 ( 1–172 ) GST plasmid in T7 polymerase producing Huh-7/T7 cells additionally boosted by the MVA-T7pol vaccinia virus system followed by Western blot analysis ( Fig . 1A and 1B ) . To investigate the mechanism of NS2-NS3 cleavage , we aimed at the identification of residues in the NS3 protease domain that are critical for NS2 stimulation . Accordingly , we established an experimental set-up that retains robust NS2-NS3 cleavage but yet is sensitive to perturbations of NS2 activation when mutating residues in NS3 . To this end , analysis of C-terminal truncations of the NS3 protease domain revealed that NS3 residues 1–169 were able to detectably activate NS2-NS3 cleavage . Furthermore , a NS3 protease domain consisting of aa 1–172 was found to be sufficient to stimulate NS2-NS3 cleavage to a level comparable with NS2-NS3 ( 1–180 ) ( Fig . 1C and 1D ) . Therefore , we used pcite-FLAG-NS2-NS3 ( 1–172 ) GST as the basis for our di-alanine scanning mutagenesis of the entire ( genotype 1b ) -derived NS3 protease domain and determined the impact of these di-alanine NS3 mutations on the NS2 activation by measuring the NS2-NS3 cleavage by Western blotting ( Fig . 1A and 1B ) . Among all tested NS3 di-alanine mutants we identified 8 candidates that either strongly interfere with ( IP114/115AA and GI152/153AA ) or inhibit ( IT3/4AA , LY104/105AA , LV106/107AA , LL126/127AA , PL142/143AA and DF168/169AA ) the NS2-NS3 cleavage ( Fig . 1E and 1F ) . To rule out that these NS3 mutations affect NS3 protein function ( s ) not related to the NS2-activation , we introduced the inhibitory mutations into pcite-NS3 ( 1–172 ) GST/BK and determined their NS3 serine protease activity in a trans-cleavage assay in the presence of the NS4A cofactor with a NS4B/NS5A serine protease substrate ( S1A Fig . ) . While the NS3 mutations IT3/4AA and IP114/115AA exhibit serine protease activity comparable to wild type NS3 in this trans-cleavage assay , mutations LY104/105AA and LL126/127AA reduced NS3 serine protease activity but still allowed for detectable cleavage of the NS4B/NS5A substrate ( S1B and S1C Fig . ) . In contrast , the NS3 double mutations LV106/107AA , PL142/143AA , GI152/153AA and DF168/169AA did not display detectable NS3 serine protease activity in this assay ( S1B and S1C Fig . ) . Mapping of these NS3 residues on a NS3 structure revealed that residues LV106/107 , PL142/143 , GI152/153 and DF168/169 are directed towards the center of the NS3 protein suggesting that changing these amino acids pairwise to alanine might affect proper NS3 protein folding ( S2 Fig . ) . Based on these observations only the NS3 mutations IT3/4AA , LY104/105AA , IP114/115AA and LL126/127AA block the NS2 protease activating function of NS3 but still allow for NS3/NS4A serine protease activity . Accordingly , NS3 mutations LV106/107AA , PI142/143AA , GI152/153AA and DF168/169AA were not further analyzed . Since activation of NS2 by NS3 should require interaction of NS2 with NS3 , we hypothesized that the amino acids involved , should be localized on the NS3 surface . Mapping of NS3 mutations that selectively blocked NS2 protease activation onto the NS3 crystal structure [31] revealed that I3 , Y105 , P115 and L127 constitute a continuous hydrophobic surface patch , while the residues T4 , L104 , I114 and L126 are directed more towards the protein core ( Fig . 2A ) . To investigate if this surface area is critical for the NS2 protease stimulation by NS3 , we mutated these residues individually as well as simultaneously and analyzed their impact on NS2-NS3 cleavage . While the individual NS3 mutations Y105A , P115A and L127A allowed for NS2-NS3 cleavage to different degrees , their combinations inhibited NS2-NS3 processing , indicating that the identified hydrophobic patch is indeed pivotal for the NS3-mediated NS2 protease stimulation ( Fig . 2B ) . Since the I3A mutation is located in the proximity of the NS2/NS3 site and thus might act as a cleavage site mutant we did not focus on this residue . In order to evaluate if the hydrophobic character or the amino acid identity are critical for NS2 activation , we mutated the surface amino acids Y105 , P115 and L127 individually and simultaneously to either phenylalanine or arginine . While all single or double phenylalanine exchanges were functionally tolerated to various degrees ( Fig . 2D ) , the introduction of a single charged amino acid at position 105 ( Y105R ) or 127 ( L127R ) strongly inhibited NS2 protease activation ( Fig . 2D and S3 Fig . ) . In contrast , a single proline-to-arginine exchange at position 115 ( P115R ) had only a moderate effect on NS2-NS3 cleavage indicating more flexibility at this position provided that the surrounding area remains hydrophobic ( Fig . 2D and S3 Fig . ) . These results revealed that the hydrophobicity of the amino acid side chains rather than their identities determine the function of this NS3 surface area for NS2 protease stimulation . The fact that the NS3 residues Y105 , P115 and L127 are conserved among all HCV genotypes suggested that this hydrophobic patch and its role in the NS3-mediated NS2 activation represents a conserved feature in the HCV life cycle . To confirm this assumption , we introduced the mutations into pcite-FLAG-NS2-NS3 ( 1–172 ) GST/JFH1 that is based on genotype 2a ( JFH1 ) NS2-NS3 sequence and determined the extent of the NS2-NS3 cleavage . Overall , the efficiency of the NS2-NS3 cleavage in the presence of the minimal NS3-cofactor domain ( aa 1–172 ) for JFH1 appears to be increased when compared to the genotype 1b ( compare Fig . 2B and 2D ) . As observed for genotype 1b , individual exchanges of P115A and L127A had a minor influence on the NS2-NS3 cleavage efficiency in this genotype 2a context with Y105A displaying a moderate effect ( Fig . 2F ) . In contrast , mutating two ( YP105/115AA , YL105/127AA and to a lesser extent PL115/127AA ) or all three ( YPL105/115/127AAA ) of the conserved surface residues simultaneously to alanine , strongly ( YP105/115AA , YL105/127AA and YPL105/115/127AAA ) or moderately ( PL115/127AA ) reduced the NS2 activation by NS3 ( Fig . 2D ) . To further confirm this observation , the Flag-NS2-NS3 ( 1–172 ) GST derivatives with either single ( Y105A and P115A ) or double alanine substitutions ( YP105/115AA ) in the hydrophobic patch of two different HCV genotypes , genotype 1b ( BK ) and genotype 2a ( JFH1 ) , were analyzed by a RIP assay . The WT and the NS2/C184A derivatives of both genotypes served as positive and negative controls , respectively . As shown in Fig . 3 , the impact of the NS3 mutations on NS2-NS3 cleavage is in agreement with our Western blot results ( compare Fig . 3B with Fig . 2C and 2F ) . Furthermore , the corresponding mutations in both genotypes have a similar impact on the NS2-NS3 cleavage when compared to their respective WT derivate ( Fig . 3B ) . This demonstrates that the mechanism of the NS3-mediated NS2 protease stimulation is conserved among different genotypes . Assuming that the hydrophobic patch on the NS3 surface functions only in NS3-cofactor mediated NS2-stimulation , surface mutations in the context of a NS3-5B/JFH1 replicon should not affect polyprotein processing or RNA replication . To test this hypothesis , we introduced the NS3 mutations into the luciferase-expressing genotype 2a reporter replicon pFKI389-Luc/NS3-3’ and determined the RNA replication kinetics of the mutant RNAs relative to the wild type ( WT ) and non-replicative ( GND ) RNAs . Besides the four single mutations ( I3A , Y105A , P115A and L127A ) , two simultaneous mutations ( YP105/115AA and YPL105/115/127AAA ) were analyzed . The majority of the single alanine mutations ( I3A , Y105A or P115A ) as well as the combination of Y105 and P115 mutations ( YP105/115AA ) had no detectable effect on RNA replication , indicating that the I3 , Y105 and P115 surface exchanges do not inhibit any of the NS3 functions required for RNA replication in the NS3-NS5B/JFH1 replicon context ( Fig . 4A ) . In contrast , the L127A exchange resulted in a strongly reduced RNA replication level when present either alone or in combination with Y105A or P115A mutations in the NS3-5B/JFH1 replicon , suggesting that this residue plays a critical role in HCV genome replication ( Fig . 4A ) . To characterize the replication phenotype of L127A in more detail , we first determined if this mutation is affecting the NS3-4A serine protease activity . Accordingly , we introduced the entire set of NS3 mutations into a pcite-NS3-3’/JFH1 plasmid which allows the replication-independent expression of the NS3-5B polyprotein by using the MVA-T7pol vaccinia virus system . Western blot analysis confirmed that all NS3 surface mutations ( including L127A ) had no detectable effects on NS3 stability or NS3-NS5B polyprotein processing in this system ( Fig . 4B ) . In contrast , the analysis of the NS5A phospho-form distribution led to an intriguing observation regarding levels of hyperphosphorylated NS5A . While the NS3 mutations I3A , Y105A , P115A and YP105/115AA exhibited NS5A hyperphosphorylation levels comparable to wild type , the L127A mutation individually or in combination with Y105A , P115A or YP105/115AA showed a strong reduction in NS5A hyperphosphorylation ( Fig . 4B ) . Strikingly , the L127A-replication phenotype correlated with the significant reduction in the ratio of hyper- to basally phosphorylated NS5A ( Fig . 4A and 4C ) . However , we cannot rule out that L127A also inhibits other NS3 function ( s ) , not linked to its serine protease activity , which may also contribute to the negative replication phenotype . NS2-NS3 cleavage is essential to liberate NS3 , a step that has been shown to be critical for viral RNA replication [32] . Accordingly , NS3 mutations that interfere with the NS2 protease activation in the context of the NS2-5B/JFH1 polyprotein should not only reduce the NS2-NS3 cleavage but also inhibit RNA replication similar to what has been demonstrated with NS2 mutations inactivating the NS2 protease [32] . To test this assumption , we introduced the NS3 mutations into the genotype 2a replicon pFKI389-Luc/NS2-3’ and determined the RNA replication kinetics of the mutant RNAs relative to the wild type ( WT ) and the non-replicative ( GND ) replicon RNAs . A replicon carrying a NS2-active site mutation ( NS2/C-184-A ) was used as a further non-replicative control which is defective in NS2-NS3 cleavage but retains its serine protease activity . The single mutations I3A , Y105A and P115A in the NS2-5B replicon RNA allowed for RNA replication to different degrees: the P115A mutant replicon RNA replicated to wild type levels whereas the mutant replicon RNAs I3A and Y105A showed a greater than 100-fold reduction in RNA replication ( Fig . 5A ) . The L127A mutation in the NS2-5B/JFH1 replicon inhibited RNA replication to levels similar to the ones observed in the NS3-5B/JFH1 replicon . Most interestingly , the NS2-5B/JFH1 replicon with the YP105/115AA mutation that interferes with the NS2 activation by NS3 did not detectably replicate indicating that inefficient NS2-NS3 cleavage triggered by NS3 surface mutations blocks viral RNA replication ( Fig . 5A ) . Several determinants for NS5A hyperphosphorylation have been mapped to NS3 , NS4A and NS4B and expression of the NS3-5A polyprotein is required for this process . As demonstrated above , the NS3 mutation L127A strongly reduced NS5A hyperphosphorylation as well as RNA replication ( Fig . 4 ) . At the same time , L127 , together with Y105 and P115 , constitutes the NS3 surface patch that is required for activation of the NS2 protease in the context of uncleaved NS2-NS3 ( Fig . 2 ) . Therefore , we asked whether NS2-NS3 cleavage in a NS2-5B polyprotein is mechanistically linked to NS5A hyperphosphorylation . This hypothesis was based on the assumption that the accessibility of L127 on the NS3 surface should be compromised in the context of uncleaved NS2-NS3 with respect to its function in NS5A hyperphosphorylation . To assess the role of NS2-NS3 cleavage efficiency in regulating NS5A hyperphosphorylation experimentally , a panel of pcite-FLAG-NS2-3’/JFH1 plasmids was used for NS2-NS5B polyprotein expression in Huh-7/T7cells . For a rigorous test , a pcite-FLAG-NS2-3’/JFH1 variant carrying an NS2 protease active-site mutation ( NS2/C-184-A ) was used . As expected , the analysis of the NS2/C-184-A mutant showed that NS2-NS3 cleavage was abrogated without affecting polyprotein processing downstream of NS3 by the serine protease activity ( Fig . 5B ) . In addition , the abrogation of NS2-NS3 cleavage correlated with a complete loss of NS5A hyperphosphorylation ( Fig . 5B ) . Therefore , the NS3 release from uncleaved NS2-NS3 appears to be a prerequisite for NS5A hyperphosphorylation . Further analyses revealed that single NS3 mutations had small but detectable effects on NS2-NS3 cleavage efficiency compared to wild type , while combinations of NS3 mutations ( YP105/115AA , PL115/127AA and YL105/127AA ) further decreased NS2-NS3 cleavage ( Fig . 5B and 5C ) and the triple mutation ( YPL105/115/127AAA ) almost abolished cleavage . Importantly , for double and triple mutants with decreased NS2-NS3 cleavage efficiency the analysis of NS5A revealed strongly reduced levels of hyperphosphorylated NS5A ( Fig . 5B and 5C ) . In addition , analysis of NS5A hyperphosphorylation revealed that the small but detectable effects on NS2-NS3 cleavage efficiency for the single NS3 mutants I3A and Y105A also correlated with a reduced NS5A hyperphosphorylation compared to either wild type or P115A . In contrast , L127A decreased the production of NS5A hyperphosphorylation to levels comparable with the double or triple mutations ( Fig . 5B ) . Thus , the effect of L127A is similar to what has been observed in the NS3-5B polyprotein . The quantification of NS2-NS3 cleavage and NS5A hyperphosphorylation indicates that an efficient NS2-NS3 cleavage increases the ratio of hyper- to basal-phosphorylation of NS5A ( Fig . 5B and 5C ) . Together , these results strengthen our observation that NS5A hyperphosphorylation is functionally linked to NS2-NS3 cleavage in a NS2-5B polyprotein context . Recent work characterizing NS5A hyperphosphorylation in replicons of HCV genotype 1b ( Con1 ) and genotype 2a ( JFH1 ) isolates revealed significant differences concerning the degree of correlation between NS5A hyperphosphorylation and viral RNA replication [33–36] . Therefore , we determined if L127 is also critical for NS5A hyperphosphorylation in the context of the NS3-5B/Con1 replicase . Accordingly , the NS3 mutations were introduced into a NS3-5B/Con1 luciferase reporter replicon and analyzed for their effect on RNA replication ( Fig . 6A ) . NS3-5B/Con1 polyprotein processing as well as the NS5A phospho-form distribution were determined by Western blotting using Huh-7/T7 cells and the MVA-T7pol vaccinia virus system ( Fig . 6B ) . In agreement with the results obtained for the NS3-5B/JFH1 replicon , the NS3-5B/Con1 replicons encoding for the NS3 mutants Y105A , P115A and YP105/115AA showed robust RNA replication , comparable to wild type , while a replicon encoding the NS3 L127A exchange was replication-deficient ( Fig . 6A ) . Transient protein expression in cells transfected with the corresponding set of mutant pcite-NS3-5B constructs followed by immunoblotting assays did not reveal a substantial inhibition of the NS3-5B/Con1 polyprotein processing indicating that no apparent defect in the NS3-NS4A serine protease was induced by the NS3 mutations ( Fig . 6B ) . In agreement with our results for JFH1 , only the substitution L127A significantly reduced the ratio of hyper- to basal phosphorylated NS5A relative to wild type in the context of the NS3-5B/Con1 polyprotein ( Fig . 6B ) . Next , we also tested in the Con1 replicon system if a correlation between the NS2-NS3 cleavage efficiency and NS5A hyperphosphorylation exists . To this end we first determined the replication capacities of mutant NS2-5B/Con1 replicons in a transient replication assay . Among the mutant NS2-5B/Con1 replicons that were examined only the one encoding the NS3 P115A exchange replicated to wild type level ( Fig . 7A ) . All other NS2-5B/Con1 replicons examined did not replicate to detectable levels . The observation that the NS2-5B/Con1 replicons with the NS2/C184A , the YP105/115AA and the L127A mutations showed no replication corroborated our earlier observations in the NS2-5B/JFH1 replicon system . The fact that the NS2-5B/Con1 Y105A replicon derivative did also fail to replicate was somewhat unexpected and differs from the Y105A NS2-5B/JFH1 mutant replicon that still replicates at low level compared to WT ( Fig . 5A ) . Processing of the respective Y105A NS2-5B/Con1 polyprotein upon expression by the MVA-T7 pol vaccinia virus system revealed that the NS2-NS3 cleavage was already strongly reduced compared to either wild type or the NS3 P115A mutant NS2-5B ( Fig . 7B and 7C ) . Moreover , in cells expressing the NS2-5B/Con1 Y105A polyprotein , NS5A hyperphosphorylation was almost undetectable in our Western blot system when compared to the wild type or NS2-5B/Con1 ( P115A ) polyprotein ( Fig . 7B and 7C ) . Together , these findings again emphasize the functional link between NS2-NS3 cleavage efficiency and NS5A hyperphosphorylation . Collectively , these results point to a conserved and , so far , unappreciated role of efficient NS2-NS3 cleavage as a prerequisite for the NS5A hyperphosphorylation during the biogenesis of the HCV replication complex in different HCV genotypes in a process that is critical for viral genome replication . NS2 is required for virion assembly and this function has been assigned to both the N-terminal membrane association- and the C-terminal protease domain [19 , 21 , 22 , 37 , 38] . Accordingly , the NS2 protease domain not only catalyzes the NS2-NS3 cleavage but also provides determinants for virion morphogenesis . This observation together with the finding that mature NS2 is interacting with NS3 in a way that has been implicated to coordinate virion assembly [24] , prompted us to test whether mutations in the hydrophobic NS3 surface patch have effects on infectious virus production . To this end we introduced the NS3 mutations Y105A , P115A , L127A and YP105/115AA into the JFH1ad-R2a_NS2EI3 genome ( S4 and S5 Figs . ) . In this bicistronic context , virus assembly can be analyzed independently from NS2-NS3 cleavage and thus uncouples NS2-NS3 processing from replication and virus assembly . Both single mutations replicated to wild type level , whereas the double mutation ( YP105/115AA ) exhibited slightly reduced replication ( S4B Fig . ) . The L127A mutation did not replicate in this bicistronic context ( S5 Fig . ) confirming our earlier observation that this mutation inhibited RNA replication ( Figs . 4A and 5A ) . The replication-competent NS3 mutants ( Y105A , P115A and YP105/115AA ) have no detectable effect on NS5A hyperphosphorylation in the context of the bicistronic JFH1ad-R2a_NS2EI3 genome ( S6 Fig . ) as expected from our transient expression experiments with the NS3-5B polyprotein ( Fig . 4B ) . Infectious virus production was similar to wild type for both single mutants , while the double mutant showed a reduction of infectious titers of about 10-fold ( S4C and S4D Fig . ) . When we calculated the infectivity release efficiency as a ratio of infectivity release and replication the double mutation exhibited a 5-fold reduction compared to wild type , whereas the single mutations showed efficiencies of about 60% of wild type level ( S4E Fig . ) . The double mutant YP105/115AA strongly reduces NS2 protease stimulation by NS3 ( Fig . 5B ) most likely by disturbing surface interactions between NS2 and NS3 in the NS2-NS3 precursor protein . The same mutations mildly reduce the infectivity release efficiency to 20% of wild type in a bicistronic background ( S4E Fig . ) . We also examined whether the NS3 mutations influence binding between NS2 and NS3 , which could , in case of the double mutant YP105/115AA , explain the observed moderate reduction in virus production . To this end , Huh7 . 5 cells were transfected with wild type JFH1ad_HAF-NS2EI3 ( carrying a HA-Flag epitope sequence , HAF , in NS2; S7A Fig . ) or mutant RNAs encoding individual NS3 substitutions . After 72 hours cells were harvested , lysed and lysates were used for immunoprecipitation . Western blot analysis of immunoprecipitated HAF-NS2 protein as well as the co-immunoprecipitated NS3 indicated that the mutations in NS3 did not detectably influence NS2:NS3 interaction ( S7B Fig . ) . Together , these results showed that virus mutants with substitutions in the hydrophobic NS3 surface patch are still capable of virus production albeit at slightly ( Y105A and P115A ) or moderately ( YP105/115AA ) reduced levels compared to wild type . This indicates that the hydrophobic NS3 surface residues required for NS2 protease stimulation are not critical for virus assembly in the context of a bicistronic genome .
Polyprotein processing is a key step in the HCV life cycle and the NS2-NS3 autocleavage has been demonstrated to be essential for RNA replication of full-length virus and subgenomic NS2-5B replicons [18 , 32] . NS2 in its cleaved form is also required for virion morphogenesis [19 , 22 , 37–39] . Previous studies have shown that the NS2 protease activity critically depends on the presence of the NS3 cofactor [17] and that the formation of an NS2 dimer is required for NS2-NS3 cleavage due to the unique composite nature of the NS2 protease active site [12 , 16 , 20] . Since the structure of the NS2-NS3 protein in its pre-cleavage conformation has not been determined , the molecular mechanism of the NS2-activation and its regulation remained enigmatic . Our finding that a conserved hydrophobic NS3 surface patch composed of Y105 , P115 and L127 is essentially involved in NS2 protease stimulation reveals that NS2 protease stimulation is mainly directed by specific hydrophobic NS3 surface residues ( Fig . 2 ) . This observation is supported by our mutational characterization of this surface area . While the introduction of single charged amino acids in this hydrophobic core strongly reduces ( P115R ) or abolishes ( Y105R , L127R ) NS2 protease activation , more conservative phenylalanine exchanges of Y105 , P115 and L127 were tolerated with regard to NS2-NS3 cleavage ( Fig . 2B ) . Accordingly , the hydrophobic character of this surface patch is pivotal for the NS3 cofactor function in NS2 protease activation . The finding that only combinations of these amino acids inhibit NS2-NS3 cleavage suggests that multiple hydrophobic NS3 residues are required to create a continuous hydrophobic NS3 surface area that interacts with NS2 to stimulate NS2 protease activity ( Fig . 2A ) . This NS2/NS3 interaction in uncleaved NS2-NS3 likely promotes the efficient formation of a functional conformation required for productive NS2-NS3 autoprocessing . Interestingly , the NS3 surface area implicated in the formation of a proteolytically active conformation of the NS2-NS3 complex is situated in the previously defined NS3 Zn2+-binding domain [17] . This domain was shown to represent the functional NS2-protease activating domain and can also activate NS2-NS3 cleavage in trans [17] . Our results indicate that intrinsic structural preferences in the NS2-NS3 precursor have important regulatory roles in the NS2-NS3 autocleavage reaction . Mechanistically , it has been proposed that the intramolecular NS2-NS3 cleavage is performed by active NS2-NS3 dimers in the course of co- and posttranslational polyprotein processing [40] . Although critical residues for NS2 dimerization are not known [16] , the detection of NS2 interactions in mammalian cells by co-immunoprecipitation in the absence of NS3 suggests that NS3 is not essential for NS2 dimer formation [40] . In theory , the formation of a proteolytically active NS2-NS3 precleavage complex could be promoted by either intermolecular or intramolecular interactions between NS2 and NS3 ( i . e . , between the NS2 moiety of one molecule of NS2-NS3 and the NS3 moiety of a second NS2-NS3 molecule or intramolecular of NS2-NS3 molecules which form dimers ) . Future experiments will aim at the identification of corresponding interaction sites on NS2 and to investigate if the stimulation of NS2 by NS3 in uncleaved NS2-NS3 is based on intra- or intermolecular protein interactions . Such regulation of polyprotein processing is also seen in other polyprotein-encoding viruses . For instance , the regulation of P23 polyprotein processing in alphaviruses is achieved by extensive intramolecular contacts between nsP2 and nsP3 which may function in positioning and recognition of the P2/3 cleavage site [41 , 42] . In contrast , poliovirus 3CD has no intramolecular contacts and cleavage of the solvent exposed cleavage site is regulated by intermolecular contacts between 3CD molecules [43] . A major finding of this study is that the hydrophobic NS3 surface involved in activating the NS2 protease has a second important function in the regulation of viral RNA replication . The L127A mutation in NS3 massively reduced NS5A hyperphosphorylation and this defect strongly correlated with an inhibition of viral RNA replication of NS3-5B replicons of different HCV genotypes without affecting the NS3-4A serine protease activity ( Figs . 4 and 6 ) . Determinants for NS5A hyperphosphorylation have been mapped to NS3 , NS4A and NS4B and expression of the NS3-5A polyprotein is required for this process [44–47] . These findings suggest that the molecular basis for the multifactorial nature of NS5A hyperphosphorylation is its dependency on the assembly of the viral replicase . These results point to a surprising functional overlap of NS3 surface determinants that are involved in NS2 protease activation and NS5A hyperphosphorylation , a process depending on viral replicase assembly , with L127 being a structural constituent of both NS3 functions . The dual role of L127 in promoting NS2-NS3 cleavage and regulating NS5A hyperphosphorylation prompted us to test whether NS2-NS3 cleavage is a prerequisite for NS5A hyperphosphorylation . The detailed mechanism underlying the NS5A hyperphosphorylation is still poorly understood [34 , 35 , 48 , 49] . It was hypothesized that the generation of authentic NS2 by NS2-NS3 cleavage is important for NS5A hyperphosphorylation [50] . However , hyperphosphorylated NS5A was also detected in the absence of NS2 suggesting that the authentic N-terminus of NS3 , generated by NS2-NS3 cleavage , represents the critical determinant for NS5A hyperphosphorylation [44 , 51] . We could demonstrate that mutations either inactivating the NS2 protease ( NS2 C184A ) or blocking the NS2 protease cofactor function of NS3 ( YP105/115AA ) were blocking or strongly reducing NS5A hyperphosphorylation in NS2-5B polyproteins of two different HCV genotypes . Reductions in the NS2-NS3 cleavage efficiency correlated not only with a decrease in NS5A hyperphosphorylation but also with a block in viral RNA replication of the respective NS2-5B replicons ( Figs . 5 and 7 ) . This correlation became apparent by using the NS2-5B replicons of genotype 1b and 2a . When we compared the impact of the NS3 mutations on RNA replication in these systems , we observed that the Y105A mutation inhibited NS2-5B/Con1 replication but was still allowing for RNA replication in the NS2-5B JFH1 background ( compare Figs . 5A and 7A ) . This difference is most likely due to the significantly lower replication capacity of the genotype 1b replicon RNA and is also reflected by the detectable NS5A hyperphosphorylation only in the case of the NS2-5B/JFH1 ( Figs . 5B and 7B ) . These results are in line with the recent observation that , although NS2-NS3 cleavage is not limiting Con1 replication , the formation of the membranous HCV replication complexes ( RC ) might be less efficient in Con1 compared to the biogenesis of JFH1 RCs [52] . Based on our data , we propose the following order of events: NS2 in uncleaved NS2-NS3 interacts with the hydrophobic NS3 surface area that includes L127 resulting in NS2 protease stimulation . As a consequence , in uncleaved NS2-NS3 the NS3 surface surrounding L127 is most likely not accessible for protein-protein interactions since this region is engaged in interactions with the NS2 protease domain . Upon NS2 release the NS3 surface area around L127 becomes available for novel protein-protein interactions that finally allow NS5A hyperphosphorylation to occur . We propose that this process is functionally linked to the assembly of the viral replicase complex and likely involves interactions of NS3 with NS4A [32 , 47] . Moreover , after NS2 release NS3 can undergo a structural change to adopt a conformation required for its function within the viral replicase . These changes are suggested to promote inter-domain co-operations between NS3 helicase and serine protease domain [53] and lead to efficient NS4A binding [32] . Interestingly , such a scenario is supported by the observation that uncleaved NS2-NS3 exhibits a lower affinity to NS4A when compared to NS3 [32] . In this context it is remarkable that mutations in the NS4A acidic domain could be rescued by second site mutations in the NS3 protease domain [47] . One of these residues is in close proximity to L127 on the NS3 surface supporting our finding that this region is critical for NS5A hyperphosphorylation and likely for the assembly of the viral replication complex . Together , our data indicate that NS2-NS3 cleavage is mechanistically linked to NS5A hyperphosphorylation as well as the assembly of the viral replicase . These intriguing observations reveal an unexpected function for the NS2-NS3 cleavage and might explain why an efficient liberation of functional NS3 is a prerequisite for viral genome replication . The identification of a hydrophobic surface segment on NS3 as an essential module of the NS2 protease cofactor in NS3 which also plays a critical role in replicase assembly is an important step towards a better understanding of the sequential processes involved in the functional assembly of the HCV RNA replication complex at the molecular level which are most certainly accompanied by conformational transitions within the different macromolecular complexes .
Huh7 Lunet [44] and Huh7/T7 [54] cells were maintained in Dulbecco's modified minimal essential medium supplemented with 10% FCS , 100 U penicillin/100 μg/ml streptomycin , and 2 mM L-glutamine . Huh-7/T7cells were cultured in the presence of 400 μg/ml G418 . Genomes Con1 [44 , 55] , BK [56] , JFH1 [57] have been described . All mutations were introduced via QuikChange ( QC ) mutagenesis . Details concerning the generation of constructs and their properties can be found in the supplementary material . The experimental procedures used to generate in vitro transcripts from cloned HCV sequences and transfection of Huh-7 cells by electroporation have been described [37] . After electroporation , cells were immediately transferred to complete DMEM and seeded as required for the assay . At each time point ( 4 , 24 , 48 , and 72 h ) , cells were washed with PBS , scraped into 1 ml of PBS and collected by centrifugation . The cells were lysed in 40 μl of lysis buffer ( PJK-GmbH ) . 20 μl of the lysate was analyzed using the Beetle Juice luciferase assay system ( PJK-GmbH ) and measured in a luminometer ( Junior LB9509 , Berthold ) . The applied procedures have been described [58] . HCV nonstructural proteins were expressed from pcite plasmids . Briefly , 2 x 106 Huh-7/T7 cells were infected with MVA-T7pol vaccinia virus [59] and subsequently transfected with 4 or 8 μg of plasmid DNA by using Superfect reagent ( QIAGEN ) . Proteins were separated in polyacrylamide-Tricine gels . After SDS-PAGE , proteins were transferred onto a nitrocellulose membrane ( Pall , USA ) . The membrane was blocked with 5% ( w/v ) dried skim milk in phosphate-buffered saline with 0 . 05% ( v/v ) Tween 20 ( Invitrogen ) . For antigen detection , anti-NS5A 9E10 [60] , mouse monoclonal antibody against NS3 of the JFH-1 isolate ( 4D11 ) ( generated in a cooperation between Harish Ramanathan , Michael Engle , Michael S . Diamond and Brett D . Lindenbach ) or anti-NS3 ( 2E3 ) [61] , anti-NS2 ( YAL-4-70-8 , Cell Essentials ) , anti-NS4B [62] , anti-FLAG ( Sigma ) , anti-V5 ( Invitrogen ) , anti-HA ( HA . 11 clone 16B12 , Covance ) and anti-GST ( GE Healthcare ) , antibodies were used in 2% ( w/v ) dried skim milk in phosphate-buffered saline with 0 . 05% ( v/v ) Tween 20 . For primary antibody detection , horseradish peroxidase-conjugated species-specific secondary antibodies ( Dianova ) were used at a 1:3000 dilution and Western Lightning Chemiluminescence Reagent Plus ( Perkin Elmer ) was applied prior to imaging using a LAS 4000 imaging system ( Biorad , Munich ) . Quantifications of Western blots were carried out using ImageJ 1 . 47t software ( NIH , Bethesda ) . HCV nonstructural proteins were expressed from pcite plasmids . For transient protein expression , 2 x 106 Huh-7/T7 cells were infected with MVA-T7pol vaccinia virus [59] and subsequently transfected with 4 μg of plasmid DNA by using Superfect reagent ( QIAGEN ) . Cells were kept in DMEM culturing medium for 2 h . Medium was changed to DMEM ( lacking L-Methionine , L-Cysteine and L-Glutamine ) supplemented with 1% Glutamax ( Gibco ) . After 30 min the medium was changed again for DMEM ( lacking L-Methionine , L-Cysteine and L-Glutamine ) supplemented with 1% Glutamax ( Gibco ) , containing 70 μCi ( 1Ci = 37GBq ) [35S]-labeled methionine/ cysteine ( Hartmann Analytics ) . After 6 h cells were lysed in 250 μl RIPA ( G ) [150 mM NaCl , 1% ( vol/vol ) Nonidet NP40 , 0 . 5% ( wt/vol ) deoxycholate , 0 . 1% ( wt/vol ) SDS , 50 mM Tris ( pH 8 ) ] , containing 1 mM PefablocSC ( Roth ) . The following steps were performed at 4°C , mixing was ensured by placing samples on a spinning wheel . Lysates were incubated for 30 min and then centrifuged for 30 min at 16 . 000 g . The supernatants were incubated with anti-GST antibody ( GE Healthcare ) 1:400 in RIPA ( G ) for 1 h , then 50 μl of a 20% ( vol/vol ) Protein-A-Sepharose suspension were added and incubated for another hour . The Protein-A-Sepharose was pelleted at 16 . 000 g and washed 3 x with RIPA ( G ) . Proteins were denatured in sample buffer containing 5% β-Mercaptoethanol at 95°C for 10 min and then separated in polyacrylamide-tricine gels with 8% polyacrylamide . SDS-Gels were fixed in a solution containing 40% methanol and 10% acetic acid , dried and exposed to Imaging screens ( Fuji ) for 1–3 days . Readouts were performed using Phosphorimager ( Fuji BAS ) . Quantifications were carried out with AIDA Image Analyzer ( Version 3 . 52 ) software using a background substraction method . | Hepatitis C virus ( HCV ) replicates its genome in close association to cellular membranes which serve as assembly site of multi-subunit replication complexes . The process of replication complex maturation must be properly controlled to prevent the non-functional maturation/assembly of these complexes . In this process , the temporal regulation of viral polyprotein processing often plays a pivotal role as exemplified by the strict requirement for NS2-NS3 cleavage for HCV genome replication . We demonstrate here that a conserved hydrophobic NS3 surface patch activates the NS2 protease to stimulate NS2-NS3 cleavage . By dissecting the role of these NS3 surface residues in viral RNA replication , we show that one of these NS3 residues is also a critical determinant for HCV genome replication by negatively regulating NS5A hyperphosphorylation . Surprisingly , further experiments revealed that the NS2-NS3 cleavage is a prerequisite for NS5A hyperphosphorylation . To fulfill the requirements for gradual assembly into functional replication complexes , an ordered cascade of molecular events takes place: in uncleaved NS2-NS3 , the hydrophobic NS3 surface patch promotes NS2 protease stimulation; upon NS2-NS3 cleavage , this surface region becomes available for functional replicase assembly . As a consequence , the hydrophobic surface patch on free NS3 can promote NS5A hyperphosphorylation as an indication of functional replicase assembly . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | A Conserved NS3 Surface Patch Orchestrates NS2 Protease Stimulation, NS5A Hyperphosphorylation and HCV Genome Replication |
Identifying biomarkers for tuberculosis ( TB ) is an ongoing challenge in developing immunological correlates of infection outcome and protection . Biomarker discovery is also necessary for aiding design and testing of new treatments and vaccines . To effectively predict biomarkers for infection progression in any disease , including TB , large amounts of experimental data are required to reach statistical power and make accurate predictions . We took a two-pronged approach using both experimental and computational modeling to address this problem . We first collected 200 blood samples over a 2- year period from 28 non-human primates ( NHP ) infected with a low dose of Mycobacterium tuberculosis . We identified T cells and the cytokines that they were producing ( single and multiple ) from each sample along with monkey status and infection progression data . Machine learning techniques were used to interrogate the experimental NHP datasets without identifying any potential TB biomarker . In parallel , we used our extensive novel NHP datasets to build and calibrate a multi-organ computational model that combines what is occurring at the site of infection ( e . g . , lung ) at a single granuloma scale with blood level readouts that can be tracked in monkeys and humans . We then generated a large in silico repository of in silico granulomas coupled to lymph node and blood dynamics and developed an in silico tool to scale granuloma level results to a full host scale to identify what best predicts Mycobacterium tuberculosis ( Mtb ) infection outcomes . The analysis of in silico blood measures identifies Mtb-specific frequencies of effector T cell phenotypes at various time points post infection as promising indicators of infection outcome . We emphasize that pairing wetlab and computational approaches holds great promise to accelerate TB biomarker discovery .
Mycobacterium tuberculosis ( Mtb ) continues to be a global public health threat , responsible for 1 . 3 million deaths due to tuberculosis ( TB ) and 8 . 6 million new infections in 2013 [1] . While only 10% of infected individuals develop clinically active TB , the other 90% harbor bacteria and are considered to be clinically latent [2] ( latent TB , or LTBI ) . Clinically latent individuals can undergo reactivation to active TB , and thus serve as a large reservoir for disease transmission . A major hurdle in controlling TB is the lack of accurate biomarkers that correlate prognosis and progression to infection [3 , 4 , 5 , 6] . Identification of biomarkers for both infectious and non-infectious diseases is a focus of much current biomedical research . While blood or urine can be obtained from patients to measure biomarkers , events occurring in these physiological compartments may not accurately reflect dynamics at sites of infection , such as lungs [7] . We recently showed that during Mtb infection , T cell responses in blood do not consistently reflect T cell responses within granulomas , sites of Mtb infection in lungs [8] . Biomarkers associated with the observed spectrum of infection , ranging from control of infection ( LTBI ) to clinically active TB [9] , are unknown . This lack of understanding is present in settings of both natural and vaccine-induced immunity [6] . Here we focus our study in a natural immunity setting . Non-specific markers of inflammation , when considered alone , do not have sufficient predictive value for clinical use in TB [3 , 4] . For decades , the Tuberculin Skin Test ( TST ) has been the most common diagnostic tool for Mtb exposure . However , lack of specificity for detection of active TB disease , inability to distinguish between BCG vaccination and Mtb infection , and inability to provide insight into disease progression limit the predictive power of TST [3 , 4 , 5] . IFN-γ Release Assays ( IGRA ) , which measure Mtb-specific release of IFN-γ from blood cells , have higher specificity for detection an ongoing TB infection ( ~80% ) [10] but fail both as a useful correlate of vaccine-induced protection and a reliable predictor of disease progression ( i . e . , due to low sensitivity or true positive rate ) [3 , 4] . Recent association studies suggest that ratios of various T cell subpopulations in blood ( e . g . , CD4+ vs CD8+ T cells ) may help distinguish among stages of TB progression [2 , 11 , 12 , 13] . Time course data in humans are typically only available from blood , and not from sites of Mtb infection ( lungs or lymph nodes ( LNs ) . In addition , it is challenging to determine the time of exposure and infection progression status in humans , limiting the ability to use these samples for predictive studies . Gene expression profiling [2 , 14 , 15] , microRNA and metabolism-based discovery [16 , 17] , and plasma antibody profiling in Mtb infected humans [18] and animal models [3 , 19] have been paired with data mining and machine learning tools to uncover potential TB biomarkers . Imaging has been recently used successfully as a diagnostic and prognostic tool in TB studies ( i . e . , PET/CT scan ) , both in the context of natural infection [20 , 21] and drug-treatment [22 , 23] . While informative , these studies have thus far been unable to determine either a single or suite of practical biomarkers , or predictive correlates of protection that would be useful in clinical practice , particularly in developing countries where TB is most prevalent . This is likely due to the complexity of TB disease , that blood may not reflect lung infection dynamics [8] , intrinsic limitations in ex vivo studies and limited availability of longitudinal in vivo data . Moreover , a spectrum of infection outcomes overlying binary classifications of active TB and latent infection has been identified , making biomarker discovery for infection status even more challenging [9] . Here we present an integrated experimental and computational modeling approach toward the discovery of TB biomarkers with the goal of predicting Mtb infection outcomes . Fig 1 shows a methodology roadmap of the different datasets generated and analyses performed in this study . First , using machine learning , immunologic data from blood of Mtb-infected cynomolgus macaques was interrogated for biomarkers that would predict infection outcome . The analysis of the NHP blood datasets did not identify any potential TB biomarker . A separate dataset was then used to help build and calibrate a computational model of the immune response during Mtb infection . Our unique multi-scale and multi-physiological compartmental computational model generates in silico data on dynamics of infection in both blood and lung , capturing formation of independent granulomas in lungs and at the same time T cell profiles in blood . We then constructed virtual non-human primate hosts based on granuloma data and infection status from macaques , and used these models to identify potential biomarkers that can predict infection outcome . We found that Mtb-specific frequencies of effector T cell phenotypes ( i . e . , both CD4+ and CD8+ ) in the blood can be targeted to distinguish infection outcomes , with a clear separation between median trajectories of active versus latent TB late during infection progression ( ~300 days ) .
We assessed levels of CD4+ and CD8+ T cells ( and memory subsets based on the expression of the two markers CD45RA and CD27 , see Table 1 for details ) and their cytokine production in the blood of 28 non-human primates ( NHPs ) ( cynomologus macaques ) at 10 time points over the course of experimental Mtb infection ( up to 6 months post infection ) ( Fig 2A , Table 1 , S1–S6 Tables ) . Such an extensive time course of sampling in the NHP model of TB , which recapitulates human Mtb infection , has not been previously available for biomarker studies . These NHPs were clinically classified as having either latent infection or active TB disease as previously described [24 , 25] , but exhibit a range of outcomes encompassing these clinical classifications [26] . S1 Fig shows representative flow cytometry plots outlining gating strategies employed in assessing T cell levels . To interrogate these unique and extensive NHP infection datasets for biomarkers ( step 1 in Fig 1 ) , we applied four supervised classification algorithms: classical and penalized discriminant analysis ( LDA and PLDA ) , quadratic discriminant analysis ( QDA ) and logistic regression ( full details are given in the Methods and Supplemental materials ) . Sensitivity , specificity and misclassification error rates ( MER ) are shown in Table 2 for both training and test sets . Results of the training dataset show high accuracy , even if overfitting is likely driving it ( since we have 28 samples for more than 150 features ) . However , similar results were not observed in the test data set , where none of the four methods used was able to discriminate disease outcomes ( MERs for the test set ~50% , Table 2; S2 Fig ) . The stark differences in predictive accuracy between the training and test datasets indicate over-fitting , which could potentially be mitigated by increasing sample size . We next re-analyzed the blood datasets using unsupervised clustering algorithms ( multidimensional scaling , Ward’s method and other hierarchical methods , see Materials and Methods for full details ) . These algorithms suggested that the 28 macaques can be subdivided into 3–25 optimal clusters ( S7 Table ) , but there was no agreement among the methods in terms of what the clustering should be . The clusters contain mixtures of NHPs with active disease and latent infection , with no clear separation between subjects and no distinct delineator of infection stage . This is likely due to the overlapping spectrum of pathology present in animals that are clinically classified as having active disease or latent infection [8 , 9 , 21 , 22 , 25] , as also described for humans [27] . Overall , our multiple attempts to use longitudinal blood T cell data obtained from NHPs to predict the clinical binary classification ( i . e . latent infection or active TB ) were not successful . Our NHP blood dataset may not be predictive of infection outcomes if blood does not reflect lung infection dynamics ( as suggested by our previous work [8] ) and/or if the sample size does not have the necessary statistical power to discriminate binary outcomes . In addition , TB exists on a spectrum and identifying binary outcomes may be an artificially imposed constraint and not practical [9] . These shortfalls could be mitigated using a computational model that describes the immune response to Mtb infection in three physiological compartments capturing blood , LNs and lung dynamics , calibrated with our NHP datasets ( step 2 in Fig 1 and Fig 2B ) . Building on previous work from our group , we developed a multiscale and multicompartment model by linking our computational model GranSim that captures individual granuloma formation and function in the lungs using an agent-based model framework [28 , 29 , 30 , 31 , 32 , 33] with 2 additional non-linear ODE models capturing dynamics in both LN and blood [28 , 29] ( Fig 2 , S1 Text ) . Figs 3 and 4 illustrate the calibration of our in silico model to our NHP dataset , with both spatial and temporal fits to distinct data from individual granulomas in lungs ( Fig 3 ) and blood ( Fig 4 ) ( ranges for parameter values given in S8 and S9 Tables ) . In addition , bacterial burden per granuloma ( in colony forming units , CFU ) in the lung are calibrated to recently published [8 , 22] and unpublished NHP datasets ( Fig 3A , S1 Table ) . Time courses for the number of Mtb bacteria from a representative contained granuloma are shown in Fig 3B . We also show a comparison between granuloma spatial distribution ( e . g . , location of immune cells , bacteria , necrotic center ) from two NHP granuloma images and two in silico granuloma snapshots with similar Mtb bacteria levels and lesion size ( Fig 3C ) . NHP T cell dynamics ( S5 Table ) from 9 animals were used to calibrate our in silico blood dynamics , and in silico trajectories fall within the levels of the CD4+ and CD8+ T cells from NHP blood ( Fig 4 ) . Currently , Mtb-specificity of T cells in the NHP model cannot be directly measured ( for example , tetramers are not available ) . As a surrogate , we used our NHP blood dataset to calculate frequency of T cells that produce any of 6 cytokines measured ( IFN-γ IL-2 , IL-6 , IL-10 , IL-17 and TNF ) ( S6 Table ) following stimulation with two Mtb-specific immunodominant antigens ESAT-6 and CFP-10 ( S6 Table ) . These T cell frequencies are then used as proxies for the frequencies of Mtb-specific T cells ( at least from the perspective of a functional response to these antigens ) . This does not provide a direct measure of cells that are truly Mtb-infection specific . The NHP used in this study are outbred monkeys that could have been exposed to non-tuberculous mycobacteria ( NTMs , or environmental exposure ) and would nonetheless respond to stimulation with these antigens prior to our infecting them with Mtb . This may explain the non-zero frequencies in the NHP dataset present in many of the memory T cell phenotypes observed early during infection ( red dots , Fig 5 ) . The computational model does not account for these potential effects . We next used our calibrated in silico model to predict the spatio-temporal dynamics of Mtb-specific T cells ( see Materials and Methods section for details ) . To parallel our computational model of virtual NHPs that would also previously be exposed to mycobacterial antigens , our simulations start with non-zero initial conditions in the same frequencies as the NHP data . Our in silico model predictions for frequencies of Mtb-specific Central Memory ( Fig 5A and 5C ) and Effector T cells ( Fig 5B and 5D ) are within the variation of the NHP data that derived as proxies for Mtb-specific T cell responses . We next apply data mining techniques and correlation methods on these in silico datasets for discovery of potential biomarkers ( step 4 in Fig 1 ) . Since our computational model can predict dynamics of Mtb-specific cells in blood , lung and LNs , we use our model to generate large in silico datasets ( 10 , 000 virtual granuloma ) , pairing lung outputs ( i . e . , CFU/granuloma dynamics ) with blood measures ( immune cell dynamics ) over the time course of infection ( step 3 in Fig 1 ) . We first apply principal component analyses ( PCA ) to the in silico blood readouts only ( S3 Fig ) , and then we extend the analysis to the 3 compartment readouts combined together ( e . g . , blood , lymph node and single granuloma in the lung , see S11 Table and S4 and S5 Figs ) . If the analysis is performed only on Mtb-specific T cell variables in the blood , the top 2 principal components alone explain ~70% of dataset variability ( S3A Fig ) . Even if no cluster emerges ( S3B Fig ) , the top 2 principal components are dominated by Mtb-specific effector CD4+ and effector CD8+ T cells , as early as 42 days post infection , as well as Mtb-specific Central Memory CD4+ and CD8+ T cells later during infection ( S3C Fig ) . To explore this finding further , we narrowed the analysis of the in silico datasets by correlating virtual Mtb-specific frequencies of T-cells in blood to virtual CFU/granuloma at the site of infection ( lungs ) . In silico frequencies of Mtb-specific Effector CD4+ and CD8+ T cells in blood predict well the granuloma-scale bacterial burden , with a clear separation between low vs high CFU/granuloma ( all results shown in supplement at the granuloma scale- S6 and S7 Figs , S10 Table ) . However , since infection usually results in multiple granulomas within a single host , a host scale-readout , not a single granuloma-scale readout , would be more useful . We provide such an analysis below . We next generate host-scale predictions by combining NHP lung necropsy data and the in silico repository described in the previous section ( Fig 1 , step 5 ) . Our main assumption is that host-scale clinical outcomes can be determined by the combined effect of a host’s heterogeneous granuloma bacterial burden [21] . In fact , we previously demonstrated substantial heterogeneity and variability in NHP granulomas at the bacterial and immune cell levels , both among animals and within a single animal [8 , 21 , 22] . Total bacterial burden in NHP lungs at necropsy correlates with infection outcome ( i . e . NHP with active disease have higher bacterial burden than those with latent infection ) [25] . The contribution of individual ( and total ) granulomas in a single host to factors that can be tracked in blood is not clear , and likely is responsible for the lack of correlation of T cell responses in blood with those in granulomas or with clinical status in our previous study [8] . To scale our computational model results from single lung granuloma to whole host , we first generate a large sample of virtual NHP hosts by using our in silico dataset to simulate NHP infections ( Fig 6 ) that recapitulate known CFU/granuloma data ( S1 Table ) . We generated virtual hosts that contained the number of granulomas and CFU/granuloma to match our NHP dataset ( step 5 in Fig 1 ) . For each virtual NHP host , we sample from our in silico repository of 10 , 000 granulomas ( see Materials and Methods section for detail ) . We then combine all the sampled in silico granulomas and examine their corresponding in silico blood data on immune cell levels as generated from the computational model . This technique allows us to predict time courses of many different T cell phenotypes in the blood using our computational model ( step 6 in Fig 1 ) . Our goal here is to predict the collective behavior of latent and active TB groups rather than in silico blood trajectories of single NHP . Thus the scaling to host analysis does not include the computation of specificity , sensitivity and misclassification rates . Predicted numbers of CD4+ and CD8+ T cells ( including the subsets of Mtb-specific T cells ) in blood were generated for 43 virtual NHP ( 20 latent infection , 23 active TB as in the experimental NHP dataset ) ( Fig 7 and S8 and S9 Figs ) . These predictions match the NHP blood dataset , suggesting that one cannot distinguish between active and latent TB outcomes by monitoring total T cell levels ( Fig 7A–7C and S8A and S8C–S8E Fig ) . In contrast , predicted Mtb-specific T cell levels in blood do allow us to distinguish infection outcomes ( active vs latent ) by ~300 days post infection ( both CD4+ and CD8+ , Fig 7C and 7D ) . The predicted frequencies of Mtb-specific Effector CD4+ and CD8+ T cells in virtual active vs latent hosts separate after 300 days post infection ( Fig 7E and 7G ) . The predicted frequencies of Mtb-specific Effector Memory CD4+ and CD8+ T cells are also indicative of outcome but at later time points ( Fig 7F and 7H ) . Overall , we predict that frequencies of Mtb-specific effector CD4+ and CD8+ T cells in blood are significantly higher ( from 2- to 4- fold ) in an active versus a latent Mtb-infected NHP and thus a combination of these cells and various time points post-infection should be targeted as potential biomarkers of Mtb infection progression .
One of the greatest tools in disease diagnosis and treatment is a robust biomarker . In TB , there been has much debate regarding whether biomarkers exist and , if so , what could serve as appropriate biomarkers [3 , 4 , 5 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 34 , 35] . To date , no biomarker ( or set of biomarkers ) has been shown to be useful in discriminating the extent of infection and disease in humans . One of the complications in predicting Mtb infection status is the spectrum of disease outcomes encompassed within the binary classifications of active TB and latent infection [9 , 36 , 37] . This variability in disease outcome is also paralleled by heterogeneity in granuloma outcomes both between , and within , individual NHPs . We previously reported that a spectrum of granulomas , in terms of numbers , types , immune responses and bacterial burden are found in individual animals and among animals with active or latent infection [8 , 22 , 25] . Recent studies from our group support that progressive and healed granulomas can coexist within the same animal , with nearly all animals capable of sterilizing at least a subset of individual granulomas [21] . However , animals with active TB show a subset of lesions that do not control infection , which presumably results in dissemination [22] . Not surprisingly , given this heterogeneity within hosts , systemic T cell responses in the blood do not accurately reflect granuloma T cell responses [8] . Here we collected a unique and extensive set of phenotypic and functional T cell data in blood from 28 NHPs experimentally infected with a low dose of Mtb and with a known clinical outcome ( active or latent TB ) . We present an approach that integrates these experimental NHP data into a computational model capturing immune dynamics in 3 physiological compartments . We apply classical data mining techniques to temporal datasets from blood that are derived from both NHP studies and from our computational model . A major benefit of using data mining techniques on computational data is that in silico datasets are exhaustive in both time and density , allowing for greater sample sizes and statistical power . This can rule out small sample size as a potential error in the corresponding animal study and point to other factors at play . Standard supervised and unsupervised classification algorithms returned no clear decomposition or optimal cluster distribution . This suggested that the resolution of NHP blood measures that were collected was not able to identify potential biomarkers of disease progression , possibly due to the overlap between the binary classifications of infection outcome . Moreover , if we speculate that clinical outcome of hosts can ultimately be driven by the combined activity of an individual’s granuloma burden , sampling the blood will likely reflect the average dynamics of variable local immune responses to infection . Ultimately , we were unable to identify biomarkers from NHP experimental data collected in a systemic compartment ( blood ) , indicating the complex nature of localized lung disease in TB . To offer complementary analyses , we built a novel and unique multi-organ computational model that tracks infection dynamics in lungs , blood and LNs , using the extensive NHP datasets distinctly for model building , calibration and validation . This allows generation of a large ( on the order of thousands ) parallel in silico dataset to mine for potential biomarkers . One aspect of our in silico system is the ability to accurately simulate Mtb-specific T cell frequencies over time . These data indicate that Mtb-specific effector CD4+ and CD8+ T cell frequencies in the blood could serve as potential biomarkers to predict infection outcome . Although full ranges of Mtb-specific T cell frequencies are not currently measurable in primates , many studies are now available on Mtb protein epitopes ( e . g . , ESAT-6 , Ag85B , 16kDa , 19kDa , Hsp65 , Rv1490 ) for the most common human HLAs ( reviewed in [38] ) . For each epitope , ranges for frequencies of antigenic-specific T cells span from 0 . 05% up to 1–2% . If we assume that an entire Mtb-specific frequency repertoire can be represented by the sum of the responses to many of these epitopes , our new scaling-to-host methodology offers predictions that are quantitatively comparable to these ranges ( Fig 7 and S8 Fig ) . Because of the variability in human HLA molecules and the large number of potential Mtb antigens , distinguishing total Mtb-specific T cell frequencies in blood is not yet feasible , especially for CD8+ T cells . However , recent studies have identified novel epitope sequences from Mtb that may soon be targets to discriminate and quantify Mtb-specific T cell responses [38 , 39] . This will allow prediction of infection status or treatment outcome using peripheral blood samples from infected hosts [40 , 41 , 42 , 43 , 44 , 45] . For example , a recent study by Sette et al [39] provides an extensive list of Mtb-derived epitopes recognized by CD4+ T cells from healthy and LTBI individuals . The authors emphasize how CD4+ T cells from both groups recognize non-tuberculous mycobacteria ( NTMs , or environmental exposure ) epitopes ( likely from previous exposure ) . This might also explain the large variability identified in the NHP T cell data derived herein ( Fig 5 ) . The list of peptides that reflect true Mtb-specific versus environmental responses is by no means complete , but we can reasonably speculate that more Mtb-specific epitopes will be identified in the near future . By measuring these pathogen-specific responses , we will have a more adequate estimate of Mtb-specific immunity generated upon infection to correlate with protection or outcomes . Moreover , by including these combinations of Mtb-specific epitopes into our computational model , we could eventually predict infection outcome earlier or discriminate among a spectrum of infection outcomes . A recent study on MDR-TB patients shows how our approach can be viable in a clinical setting . Riou et al [40] identified a subset of activated and proliferating Mtb-specific CD4+ T cells ( i . e . , Ki67+ HLA-DR+ ) as a potential marker in peripheral blood that predicts the time to sputum culture conversion in TB patients at the start of treatment . Another recent prospective proof-of-concept study uses a novel T-cell activation marker-tuberculosis ( i . e . , TAM-TB ) assay on cryopreserved peripheral blood mononuclear cell samples to diagnose active TB in children based on ratios of CD27 phenotype of CD4 T cells producing IFNγ in response to Mtb antigens ( i . e . , ESAT6 and CFP10 ) [42] . It has been also suggested that a shift of Mtb-specific Central Memory to Mtb-specific Effector Memory CD4+ T cells precedes the clinical diagnosis of active TB by many months [44] . If we restrict our analysis to ratios of Mtb-specific Effector Memory to Mtb-specific Central Memory T cell frequencies ( in silico data ) , no significant difference can be shown between latent and active TB infection groups ( see S9A Fig ) . However , our in silico predictions for the time courses of the ratios of Mtb-specific Effector to Mtb-specific Central Memory CD4+ T cells , and higher ratios in the active TB group ( statistically significant only after day 300 post infection ) confirm the finding of Schuetz et al ( S9B Fig and [44] ) . In complex diseases such as heart disease and cancer , a suite of biomarkers best predicts disease outcomes and treatment intervention points . Our studies support that for TB , a complex lifelong infection that is compartmentalized primarily in the lung , single biomarkers in blood may not be a feasible goal . Coupled to the idea that infection outcome in hosts likely occurs over a spectrum , identification of a single biomarker may be a misguided goal . Our novel systems biology approach combined with the ongoing progress to elucidate and winnow down Mtb-specific epitopes can significantly influence hypothesis generation for biomarker prediction for TB . Any in silico prediction can then be validated on human clinical data and animal model studies , adding important knowledge to human immunity to TB and TB intervention studies .
The goal of this study was to develop a methodology to identify biomarkers for TB infection outcome . Our systems biology approach integrates non-human primate ( NHP ) datasets together with computational modeling , allowing the generation of virtual hosts that we use for biomarker discovery . The next sections illustrate: i ) the experimental datasets , ii ) classification algorithms , iii ) computational modeling framework , iv ) in silico datasets analysis , v ) model calibration and vi ) virtual hosts generation . Refer to Fig 1 for a general roadmap of all the different datasets that have been generated and all the many analysis that have been performed in this study . All experimental manipulations , protocols , and care of the animals were approved by the University of Pittsburgh School of Medicine Institutional Animal Care and Use Committee ( IACUC ) . The protocol assurance number for our IACUC is A3187-01 . Our specific protocol approval numbers for this project are 11090030 and 12060181 . The IACUC adheres to national guidelines established in the Animal Welfare Act ( 7 U . S . C . Sections 2131–2159 ) and the Guide for the Care and Use of Laboratory Animals ( 8th Edition ) as mandated by the U . S . Public Health Service Policy . All macaques used in this study were housed at the University of Pittsburgh in rooms with autonomously controlled temperature , humidity , and lighting . Animals were singly housed in caging at least 2 square meters that allowed visual and tactile contact with neighboring conspecifics . The macaques were fed twice daily with biscuits formulated for non human primates , supplemented at least 4 days/week with large pieces of fresh fruits or vegetables . Animals had access to water ad libitem . Because our macaques were singly housed due to the infectious nature of these studies , an enhanced enrichment plan was designed and overseen by our nonhuman primate enrichment specialist . This plan has three components . First , species-specific behaviors are encouraged . All animals have access to toys and other manipulata , some of which will be filled with food treats ( e . g . frozen fruit , peanut butter , etc . ) . These are rotated on a regular basis . Puzzle feeders foraging boards , and cardboard tubes containing small food items also are placed in the cage to stimulate foraging behaviors . Adjustable mirrors accessible to the animals stimulate interaction between animals . Second , routine interaction between humans and macaques are encouraged . These interactions occur daily and consist mainly of small food objects offered as enrichment and adhere to established safety protocols . Animal caretakers are encouraged to interact with the animals ( by talking or with facial expressions ) while performing tasks in the housing area . Routine procedures ( e . g . feeding , cage cleaning , etc ) are done on a strict schedule to allow the animals to acclimate to a routine daily schedule . Third , all macaques are provided with a variety of visual and auditory stimulation . Housing areas contain either radios or TV/video equipment that play cartoons or other formats designed for children for at least 3 hours each day . The videos and radios are rotated between animal rooms so that the same enrichment is not played repetitively for the same group of animals . All animals are checked at least twice daily to assess appetite , attitude , activity level , hydration status , etc . Following M . tuberculosis infection , the animals are monitored closely for evidence of disease ( e . g . , anorexia , weight loss , tachypnea , dyspnea , coughing ) . Physical exams , including weights , are performed on a regular basis . Animals are sedated prior to all veterinary procedures ( e . g . blood draws , etc . ) using ketamine or other approved drugs . Regular PET/CT imaging is conducted on most of our macaques following infection and has proved very useful for monitoring disease progression . Our veterinary technicians monitor animals especially closely for any signs of pain or distress . If any are noted , appropriate supportive care ( e . g . dietary supplementation , rehydration ) and clinical treatments ( analgesics ) are given . Any animal considered to have advanced disease or intractable pain or distress from any cause is sedated with ketamine and then humanely euthanatized using sodium pentobarbital . Twenty-eight cynomolgus macaques ( Macacca fasicularis ) were infected with a low dose of Mtb ( Erdman strain , ~25–50 CFU ) and monitored clinically for signs and symptoms of TB up to six months post infection . Infection was confirmed by tuberculin skin test conversion and/or lymphocyte proliferation assay six weeks post-infection . The macaques were classified ( as previously described in [22 , 23 , 25] ) to have either clinically latent TB infection or active TB ( S1 Table , third column ) based on clinical , radiological and microbiologic criteria . Peripheral blood was collected at time-points: pre Mtb infection ( only for 9 NHPs ) and at days 10 , 20 , 30 , 42 , 56 , 90 ( or M3 , 3 months ) , 120 ( or M4 ) , 150 ( or M5 ) and 180 ( or M6 ) post infection . PBMC were isolated , stimulated with peptide pools of Mtb specific antigens ESAT-6 and CFP-10 ( 10μg/ml of each peptide ) or phorbol dibutyrate ( PDBu ) and ionomycin as positive control and media as negative control , in the presence of Brefaldin A ( BD biosciences ) for 6 hours at 37°C with 5%CO2 . Multi-parametric intracellular flow cytometry was performed on fresh PBMC to assess CD4 and CD8 T cell cytokine profiles ( GMCSF ( clone: M5D12 ) , IFN-γ ( B27 ) , IL-2 ( MQ1-17H12 ) , IL-4 ( 8D4-8 ) IL-6 ( MQ2-6A3 ) , IL-10 ( JES3-9D7 ) , IL-17 ( 64CAP17 ) and TNF ( MA611 ) ) during the course of Mtb infection . S1 Fig shows representative flow cytometry plots where we outline our gating strategies employed in the analysis of T cells in PBMC . We generated four different datasets: a single cytokine dataset ( where each cell was labeled for a single cytokine , S2 Table ) , a multiple cytokine dataset ( where each cell was labeled for the simultaneous presence of more than 1 cytokine , S3 Table ) , a memory single cytokine dataset ( S4 Table , where the same cytokines profile of the single cytokine dataset is stratified by CD4+ and CD8+ memory sub-populations based on the expression of CD45RA and CD27 , namely Naïve-N [CD45RA+ CD27+] , Central Memory-CM [CD45RA-CD27+] , Effector Memory-EM [CD45RA-CD27-] and Terminally Differentiated-TD or Effector-E [CD45RA+CD27-] and a T cell dataset ( where only CD4+ and CD8+ memory phenotypes levels are measured , S5 Table ) . S6 Table summarizes the frequencies of ESAT6 or CFP10-stimulated memory T cells that produced any cytokine ( i . e . , IFN-γ , IL-2 , IL-6 , IL-10 , IL-17 and TNF ) upon stimulation . All these datasets are displayed in Fig 1 under the grey box labeled NHP data , with a detailed description given in Table 1 . The complete datasets are available as supplementary material ( S1–S6 Tables ) . The experimental design is described in Fig 2A . The time points measured in the T cell dataset ( S5 Table ) differ slightly from the other three datasets ( S2–S4 Tables , see Table 1 for details on the four datasets ) . The first three datasets were used for machine learning ( step 1 in Fig 1 ) , while the T cell dataset ( S5 Table ) was used to build ( Fig 2B ) and calibrate the in silico model for the blood compartment of the computational model ( ( step 2 in Fig 1 and Fig 4 ) . Moreover , the ESAT6-CFP10 Memory T cell dataset ( S6 Table ) was used for a tentative model validation of the Mtb-specific frequency predictions of our in silico model ( Fig 5 ) . S1 Table describes on all the NHPs enrolled in the study . A total of 58 NHPs have been monitored ( all have been classified as either latent [green] or active [red] TB ) , 28 of which have immunologic data from blood ( single , multiple and memory cytokine datasets , as well as T cell dataset ) . As for the number of granulomas and CFU per granuloma data , S1 Table has all the details on the 43 NHPs that have been necropsied ( with time of necropsy included ) , 12 of which are included in the 28 NHPs of the blood data . S1 Table was used to calibrate the lung compartment of the computational model ( step 2 in Fig 1 and Fig 3 ) . The data for some cytokine profiles was imputed ( ascribed ) using the techniques described in [46] , which relies on a “nearest neighbor” or local averaging approach to fill in missing values . The imputation approach is stochastic in nature [46] , and hence is sensitive to an initial random seed . The results presented below were obtained by averaging over 1000 trials of imputing and classifying . The results were qualitatively unchanged regardless of whether cytokine profiles were scaled to have unit variance . The analysis was performed in R ( version R 2 . 14 . 2 ) and in Matlab ( ( R2011b v7 . 13 ) ) . In order to identify possible correlates of protection on the experimental data in S2–S4 Tables , we applied two types of classification techniques , discriminant analysis and logistic regression [47] . The most popular versions of these techniques rely on a linear relationship ( i . e . , linear discriminant analysis-LDA ) between the correlates of protection and TB infection outcome of the macaques . Due to the relatively poor performance of these classical approaches , we also applied three other techniques that feature additional flexibility . The first is quadratic discriminant analysis ( QDA ) , which can be appropriate if quadratic relationships exist between possible correlates and the TB infection outcome [47] . The last two techniques utilize a statistical technique , termed l1 regularization , that relies on constrained optimization to identify important variables in the model and hence minimize the misclassification error rate of the different models[47] . We refer to the last two models as penalized linear discriminant analysis ( PLDA ) and logistic regression . The cross-validation functions within the R package are used to select the regularization parameters for PLDA and logistic regression [48 , 49] . In addition to performing all techniques on the data sets , we also applied each technique to the data projected onto the first 3 principal components , which captured over 50% of variability in the data . Our goal in doing this was to try to improve performance by reducing the number of cytokine covariates . Our goal is to predict TB infection outcome , a binary variable ( 1 for active; 0 for latent ) and performances for the different methods are measured with a set of special metrics called sensitivity , specificity and misclassification error . Sensitivity is the rate at which active TB infections are correctly predicted ( true positive rate ) , specificity is the rate at which latent TB infections are correctly predicted ( true negative rate ) , and misclassification error rate is the overall prediction error rate [48] . To avoid reporting overly optimistic ( biased due to over-fitting ) results , we withhold 50% of the data when estimating the model parameters and test the fitted model on the withheld data . Henceforth , we refer to the withheld data as test data and the rest of the data used to estimate the model as training data . Results are qualitatively unchanged for a different number of macaques in the test dataset vs training set . Note also that data are considered by each of the four techniques to be cross-sectional , i . e . , direct application of each method ignores temporal information . Since we found that six macaques in the study were consistently incorrectly classified by our classification algorithms , we re-analyzed the blood data using unsupervised clustering algorithms ( i . e . , multidimensional scaling , Ward’s method and other hierarchical methods , and k-means ) [47] . These algorithms attempt to find inherent groupings in the cytokine profiles , and hence could potentially identify a possible spectrum of latency rather than delineating between active and latent cases . Multidimensional scaling ( MDS ) is a visualization technique that forms a two-dimensional representation of the macaques , where the location of each macaque in two-dimensional space is optimized so that macaques close together are most similar . Complete and Average link hierarchical clustering , as well as Ward’s method are called “agglomerative” or bottom-up approaches , because they start with each macaque as its own cluster and then cluster them together until a stopping criteria is met . In contrast to agglomerative approaches , K-means is a top-down partitioning method , where initially all macaques belong to a single cluster . The clusters are then progressively split into smaller clusters ( reviewed in [47] ) . One would expect to find consistent results between these different techniques , even though they take varying approaches to discovering macaque groupings . Wildly varying results across methods would indicate that the data feature high levels of noise that mask a possible structure between macaques . Our hybrid agent-based model ( ABM , labeled GranSim ) tracks Mtb infection in 3 physiological compartments: lung ( site of infection ) , draining lymph node ( LN , site of generation of adaptive immunity ) and blood ( a measurable compartment , see Fig 2B for details ) . Our computational model captures single granuloma formation and function in the lung [30 , 31 , 32 , 33] , while LN and blood compartments[28 , 29] represent dynamics of the whole body in response to infection . We simulate infection over a time span of 600 days , with rules and interactions solved on a 10 min time step which can be found at http://malthus . micro . med . umich . edu/GranSim/ . The lung environment represents a 2x2 mm section of lung parenchyma tissue . Details on lung initialization can be found in [30 , 31 , 32] . Cell types captured in the model are macrophages and T cells ( or T lymphocytes ) . Macrophages transition between the following states: resting , activated , infected and chronically infected . T cells are represented by their functions: Tγ ( i . e . , IFNγ-producing T cells , necessary for macrophage activation ) , Tcyt ( i . e . , cytotoxic T cells ) and Treg ( regulatory T cells ) . Two effector molecules are also tracked , namely tumor necrosis factor-α ( TNF ) and interleukin 10 ( IL-10 ) . LN and blood dynamics are each captured by compartmentalized system of ordinary differential equations ( ODEs ) ( see S1 Text for the equations and derivation ) . We updated our published multi-compartment computational models34 , 35 , in which we now track CD4+ and CD8+ T lymphocytes with different memory phenotypes ( i . e . , Naïve , Effector , Central and Effector Memory ) . Two mechanisms new to GranSim are used herein: T cell recruitment and proliferation at the site of infection ( described in detail in S2 Text—T cell recruitment and T cell proliferation in the lung ) . A great advantage of the in silico representation is that we can track both Mtb-specific and non Mtb-specific T cells ( the T cell dataset does not discriminate between Mtb-specific and non-Mtb-specific cell types ) . The lymph node-blood dynamics are captured by 31 ODE equations with 21 independent parameters . Measure units are cell counts in the LN and cell/mm3 in blood . Details on the initial conditions of the model as well as parameter ranges are given in S8 and S9 Tables . Fig 2B illustrates the different T lymphocyte phenotypes tracked in the model and some assumptions regarding cell priming , trafficking and differentiation . Our experimental data show that a single LN drains multiple granulomas forming in lungs [20 , 21 , 22] . We coarse-grain cell migration from the site of infection ( GranSim ) to a draining lymph node . Antigen presentation and priming occurring in LNs is driven by a proxy for antigen presenting cell dynamics ( APCs , Eqn ( 1 . 1 ) in S1 Text ) , which is the count of macrophages in the lung that interacted with Mtb ( MMtb ) at any time during infection ( see [28] for details on the implementation ) . However , only a small fraction of cells ( e . g . , ~5–10% ) that encounter Mtb in lungs is thought to be able to traffic and migrate to LNs [50 , 51 , 52 , 53 , 54] . Due to computational limitation , the lung compartment tracks formation and progression of a single granuloma and not multiple granulomas developing simultaneously . Thus , the MMtb count is an overestimate for the APC trafficking to the LN from a single granuloma . However , we use the MMtb count as a reasonable proxy for APC dynamics for the whole host . This assumption implies that we are calibrating the computational model as if the host is developing between 10 and 20 similar granulomas in the lungs ( matching the ~5–10% cells that likely migrate to the LN upon infection ) [25] . This allows us to use the T cell dataset to calibrate the computational model in the blood ( since the blood data measured in that dataset reflect TB infection of a whole NHP ) , as well as in the lung by using our data on CFU per granuloma in NHPs [21 , 22] . Based on the above assumptions , when multiple granulomas are combined together to generate virtual NHPs and capture heterogenous granuloma outcomes within a single host , the in silico blood readouts for the whole host are computed as averages and medians ( instead of sums ) across all the granulomas sampled ( see Figs 6 and 7 and S8 and S9 Figs ) . We are currently working on a computational platform where multiple granulomas can be simulated simultaneously and have potential to interact with each other , mimicking a whole lung infection dynamics . Two T-cell phenotypes traffic between LN and blood , namely naïve and central memory cells ( separate equations are described for CD4+ and CD8+ T cells , as well as for Mtb-specific and non-Mtb-specific , see S1 Text for details ) . Mtb-specific naïve T cells are primed by APCs in LNs to become precursor T cells , which proliferate and ultimately differentiate into Mtb-specific Central Memory and/or Mtb-specific Effector , depending on the strength of APC stimulation [55 , 56 , 57 , 58] . Mtb-specific Central Memory cells can be re-stimulated in LNs ( similarly to Naïve ) and become precursor again [59 , 60] . Mtb-specific Effector cells can differentiate into Mtb-specific Effector Memory cells [61] . All cells in the LN compartment ( except APC and precursor ) migrate into blood ( as shown in Fig 2B ) . We modeled Mtb-specific CD4+ T cell processes in the LN and blood compartments identical to how CD8+ T cells are modeled , with the exception of Mtb-specific Naïve CD8+ priming which is dependent of Mtb-specific Naïve CD4+ priming . We modeled non-Mtb-specific T cells ( grey circles in Fig 2B ) similarly to their respective Mtb-specific cell types ( colored circles in Fig 2B ) counter parts . However , non-Mtb-specific cells do not respond to antigen , therefore , no priming occurs in any of these cell populations in LNs and no precursor cells are generated . We assume that neither effector nor effector memory cells re-enter the LN after migrating into blood: they recirculate through blood and eventually migrate to lung ( as shown in Fig 2B ) . The production of non-Mtb-specific effector cells was captured as a source term in blood and was calibrated to the T cell dataset prior to infection ( S5 Table ) . GranSim is an agent-based model implemented using the C++ programming language in conjunction with Boost libraries ( distributed under the Boost Software License–available at www . boost . org ) . The graphical user interface ( GUI ) was built using the Qt framework ( open-source , distributed under GPL–available at qt . digia . com ) , which allows us to display , track and plot different readouts of the in silico granuloma simulation in real-time . The lymph node and blood compartments are modeled together as an ordinary differential equation ( ODE ) system ( as shown in Fig 2B ) . They are interfaced with GranSim by numerical ODE solvers implemented as part of the C++ platform all within our own code . The three-compartmental model can be used with or without GUI visualization and is cross-platform ( Mac , Linux , Windows ) . Computational model simulations were performed on XSEDE’s OSG Condor pool and NERSC’s Edison Cray XC30 . Initial ODE model calibration and analysis of the results were performed in Matlab , as well as all post-processing analysis of data generated by our in silico model . We use the min-max range NHP pre-infection data from the T cell dataset to establish ranges for the initial conditions for the blood ODE model ( homeostasis , see S8 Table for details on the initial conditions ) . We assume that these levels represent a flow of cells constantly trafficking through LNs . Therefore , initial numbers of cells in the lung draining LN were calculated by dividing the number of those in the blood , by the number of LNs in the host ( i . e . , parameter host_Ln in S9 Table ) . We also assumed a frequency of Mtb-specific Naïve T cells ( parameter λ ) between 0 . 001 and 0 . 00001 ( i . e . , [1e-3 , 1e-5] ) [62] . Most model parameter values used herein are taken from our previously published studies [28 , 29 , 30 , 31 , 53 , 63 , 64] and are listed in S9 Table . Additionally , we rely on uncertainty analysis techniques to efficiently explore the parameter space and inform on baseline behaviors of the system ( uncertainty analysis—UA ) . Here we used Latin Hypercube Sampling ( LHS , reviewed in [65] ) for UA . The LHS algorithm is a stratified Monte Carlo sampling method without replacement [65] . It was used to generate 1 , 000 unique parameter sets , which are simulated in replication 10 times ( a total of 10 , 000 in silico simulations ) . We varied 21 parameters/mechanisms in blood and LN compartments , as well as 8 initial conditions for Naïve , Effector , Effector Memory and Effector Memory CD4+ and CD8+ ( see S8 and S9 Tables for the ranges used to generate the in silico granulomas and blood/LN dynamics ) . Parameters/mechanisms in GranSim have been updated to reflect latest knowledge and data ( see [30] for details and http://malthus . micro . med . umich . edu/GranSim/ ) and were held fixed in our LHS experiments as blood and LN parameters were varied . The in silico dataset comprises sets of 10 , 000 model simulations ( i . e . , 1 , 000 x 10 replications ) of single granulomas coupled to LN and blood dynamics over a time span of 600 days post infection ( step 3 in Fig 1 ) . We analyzed the following readouts across the three compartments: blood compartment ( 16 variables ) , lymph node ( 15 variables ) , macrophage and T cell counts in the lung , as well as measures in the lung such as lesion size , TNF and IL-10 total levels , cytotoxic killing , and infected macrophage bursting rate ( total of 48 , see S11 Table for the complete list ) . We used the same time points from the experimental data ( e . g . , 9 time points from the T cell dataset , days 10 , 20 , 30 , 42 , 56 , 90 , 120 , 150 and 180 post infection ) and computational model simulations for direct comparison between the in vivo and in silico model outputs for the blood measures ( step 2 in Fig 1 ) . We extended the analysis up to 600 days to perform the scaling-to-host step , since we matched to NHPs that have been necropsied between 200 and 600 days post infection . Principal component analyses ( PCA ) were performed on the in silico dataset ( step 4 in Fig 1 ) , after the computational model was calibrated to the experimental data . We defined Mtb-specific frequencies for each T cell phenotype as the ratio between the number of Mtb-specific T cells over the total number of T cells . The computational model was calibrated with respect to i ) CFU/granuloma dynamics , based on our recent NHP experimental data [21 , 22] ( see S1 Table and step 2 in Fig 1 ) , and on ii ) blood T cell dynamics as measured in the T cell dataset ( see S5 Table and step 2 in Fig 1 ) . S8 and S9 Tables shows the ranges used to generate our in silico dataset of 10 , 000 granulomas ( step 3 in Fig 1 ) . The 10 , 000 model simulations returned CFU dynamics as well as T cell dynamics in the blood that we compare to the NHP experimental data ( as shown in Figs 3 and 4 ) . Since the T cell dataset measured total T cell populations rather than Mtb-specific subsets , we summed Mtb-specific and non Mtb-specific in silico cell predictions to match our experimental data . Due to the large variability in the T cell dataset , we did not superimpose strict criteria for model calibration . The computational model was considered to be adequately calibrated if minimum , mean , median and maximum in silico trajectories were within the longitudinal data measured in the experimental settings . Only 12 NHPs ( of the 28 in the datasets used for classification ) have both blood and lung data available , so we used these 12 NHPs for the initial model calibration ( Fig 4 ) . We then merged both the lung data on CFU of the 43 NHPs with the blood data of the 12 NHPs to see if all the virtual NHPs have blood dynamics within the experimental ranges ( Fig 4 ) . The 12 NHPs are biased towards latent ( 9 ) vs active TB ( 3 ) and limited to 6 months ( while the lung experimental data are available up to 600 days ) . Fig 5 has the T cell Memory phenotype data shown as frequencies of T cell producing any cytokine in response to ESAT6/CFP10 stimulation . For the reasons outlined above and since time courses of NHP blood experimental data are not available to match our Mtb-specific T cell frequencies predictions ( step 6 in Fig 1 ) , Figs 7 , S8 and S9 only show in silico time courses . To allow our model to be directly comparable to NHP and human data that track infection progression , we developed a method to scale single granuloma outcomes in lungs to host level readouts in blood . The virtual NHP generation approach uses a different set of NHP data compared to the machine learning approach ( step 5 in Fig 1 ) . We guide the generation of virtual NHPs by replicating granuloma heterogeneity and variability in the lung up to 600 days post infection ( i . e . , # of granuloams and CFU/granuloma in S1 Table ) . We then analyze the in silico blood dynamics to make predictions on potential biomarkers for infection outcome . The complete protocol , illustrated in Fig 6 , comprises the following steps: i ) select the number of granulomas to replicate upon infection over time in each known NHP from our NHP experimental data ( see S1 Table ) , ii ) mimic CFU/granuloma values by sampling only granulomas that have similar CFU burden from our repository of 10 , 000 in silico granulomas ( based on Figs 3 and 4 , the computational model readouts are within the range of variability measured by experimental data in the lung and in the blood , thus the sampling is warranted ) , iii ) collect in silico blood readouts predicted for each granuloma , and then average them across all granulomas , iv ) repeat steps ii ) and iii ) K times ( to capture within host variability , we used K = 1 , 000 ) and generate statistics ( e . g . , means , medians , standard deviations ) on the K replications to recapitulate in silico blood data on the single virtual host , v ) repeat steps i ) -iv ) for all the NHPs in the dataset . For example , NHP 21710 ( latent TB , row 14 in S1 Table ) has 12 granulomas; 8 of them are sterile and 4 have the following CFU burdens: 38 , 142 , 38 and 1133 . Virtual NHP 21710 was built by sampling 12 in silico granulomas from our repository at the day NHP 21710 was necropsied ( i . e . , 370 days post infection ) . For sterile granulomas , we use the criteria of CFU<1 per granuloma , and for the non-sterile granulomas we sample from subsets of in silico granulomas that fell within a ±α range of the NHP experimental data ( Fig 6 ) . We used α = 10% . Numbers of in silico granulomas that satisfy our condition for sterile granulomas ( i . e . CFU<1 ) at day 370 post infection is 2257 out of 10 , 000 . For the first non-sterile granuloma of NHP 21710 ( CFU = 38 ) , we first select in silico granulomas with CFU in the ±10% range , namely [30 , 46] ( i . e . , 66 granulomas ) , and then randomly choose one without replacement ( thus the third granuloma , which has the same CFU , will not be assigned to the in silico granuloma selected for the first one ) . The second granuloma will be selected from the range [113 , 170] ( i . e . , 12 granulomas ) , and so on . Once all the in silico granulomas have been sampled , we average the blood readouts across all the simulated granulomas ( 12 granulomas for NHP 21710 ) and then compute the values for total T cell levels as well as for the Mtb-specific T cell frequencies in the blood . We repeat the same procedure K times ( K = 1000 ) for the same virtual NHP to account for granuloma outcome variability and generate statistics ( mean , median , standard deviations ) for the variables used to correlate in silico blood time courses to infection outcome ( e . g . , total T cell levels as well as Mtb-specific T cell frequencies , see Fig 6 for details ) . We applied the scaling to host method to 43 NHPs that have been classified as either latent ( 20 NHPs ) or active ( 23 NHPs ) TB , and that have data on numbers of granulomas in lung and CFU/granuloma for each granuloma ( highlighted in yellow in S1 Table ) . Once all the blood readout statistics have been computed on the 43 virtual NHPs , we generated trajectories ( see Figs 7 , S8 and S9 ) by grouping active vs latent virtual NHPs and we test if any of them are significantly different ( t-test ) over time and likely predictive of infection outcome ( step 6 in Fig 1 ) . Figs 7 , S8 and S9 show how we used in silico predictions for the time courses of total CD4+ and CD8+ T cell levels , as well as Mtb-specific T cell frequencies to predict the virtual NHPs infection outcome , using the clinical known classification between latent and active TB groups from S1 Table . | Tuberculosis ( TB ) is a disease that is caused by infection after inhaling the bacterium Mycobacterium tuberculosis . Not everyone infected with TB bacteria becomes sick . As a result , two TB-related conditions have been categorized: latent TB infection ( not sick but still harboring the bacteria ) and active TB disease . If not treated properly , active TB disease can be fatal . Almost 1 . 3 million die of TB worldwide each year , with ~8 , 6 million new infections in 2013 . No effective vaccine is available to protect against TB and treatment of infection with multiple antibiotics is lengthy ( 6–9 months ) , with non-compliance being a major factor for the emergence of drug-resistant strains . A key step in developing effective vaccines and possibly shorter treatment regimens is the ability to identify biomarkers that correlate prognosis and progression to infection ( similar to how cholesterol levels are a measure of heart health ) . In this study we show how pairing computer modeling , statistics and mathematics with datasets derived from non-human primate studies can accelerate biomarker discovery , and offer a new approach to identifying correlates of protection that will be useful in clinical practice , particularly in developing countries where TB is most prevalent . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"body",
"fluids",
"granulomas",
"immunology",
"tropical",
"diseases",
"biomarkers",
"bacterial",
"diseases",
"cytotoxic",
"t",
"cells",
"bacteria",
"infectious",
"diseases",
"white",
"blood",
"cells",
"memory",
"t",
"cells",
"animal",
"cells",
"tuberculosis",
"t",
"cells",
"actinobacteria",
"hematology",
"biochemistry",
"blood",
"cell",
"biology",
"anatomy",
"mycobacterium",
"tuberculosis",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] | 2016 | Computational and Empirical Studies Predict Mycobacterium tuberculosis-Specific T Cells as a Biomarker for Infection Outcome |
Collaboration between heterogeneous pattern recognition receptors ( PRRs ) leading to synergistic coordination of immune response is important for the host to fight against invading pathogens . Although complement receptor 3 ( CR3 ) and Dectin-1 are major PRRs to detect fungi , crosstalk between these two receptors in antifungal immunity is largely undefined . Here we took advantage of Histoplasma capsulatum which is known to interact with both CR3 and Dectin-1 and specific particulate ligands to study the collaboration of CR3 and Dectin-1 in macrophage cytokine response . By employing Micro-Western Array ( MWA ) , genetic approach , and pharmacological inhibitors , we demonstrated that CR3 and Dectin-1 act collaboratively to trigger macrophage TNF and IL-6 response through signaling integration at Syk kinase , allowing subsequent enhanced activation of Syk-JNK-AP-1 pathway . Upon engagement , CR3 and Dectin-1 colocalize and form clusters on lipid raft microdomains which serve as a platform facilitating their cooperation in signaling activation and cytokine production . Furthermore , in vivo studies showed that CR3 and Dectin-1 cooperatively participate in host defense against disseminated histoplasmosis and instruct adaptive immune response . Taken together , our findings define the mechanism of receptor crosstalk between CR3 and Dectin-1 and demonstrate the importance of their collaboration in host defense against fungal infection .
Diseases caused by fungal pathogens have become an important cause of morbidity and mortality over the last decades due to the increasing number of immunocompromised patients [1] . To reveal the cellular and molecular mechanisms of the interaction between host and fungal pathogens will be helpful for the development of new therapeutic strategies . Innate immune cells recognize pathogen-associated molecular patterns ( PAMPs ) on fungi through pattern recognition receptors ( PRRs ) [2 , 3] . The fungal cell wall is composed predominantly of glucans , chitin , mannose and other covalently-linked proteins with the composition varies between species and even between the different strains and morphological forms of the same species [4 , 5] . Since a single pathogen is composed of multiple PAMPs , innate immune cells are likely to simultaneously or sequentially utilize a complex set of PRRs to interact with a specific pathogen . The coordination between PRRs leading to activation or inhibition of downstream signaling is referred to as “receptor crosstalk” [6] . PRR collaboration is known to be important for the host to control invading pathogens . It has been reported that Dectin-1 functions synergistically with TLR2 in amplifying innate cell cytokine response when stimulated with their respective ligands [7–10] . Collaboration between Dectin-1 and TLR2 is mediated by the activation of both Dectin-1/Syk and TLR/Myd88 pathways which results in increased NF-κB activity [7 , 9] . Receptor crosstalk between C-type lectin receptors ( CLRs ) has also been reported . Dectin-1 collaboration with SIGNR1 or with Dectin-2 is largely dependent on the activation of Syk kinase to enhance immune responses against Candida albicans [11 , 12] . Stimulation with C . albicans induces colocolization and physical association of Dectin-1 and SIGNR1 whose collaboration results in macrophage oxidative response [11] . Aside from collaborating with TLRs and CLRs , Dectin-1 is also known to collaborate with CR3 in macrophage cytokine response to H . capsulatum in a Syk kinase-dependent manner [13] . However , the coordinated mechanisms in CR3 and Dectin-1 collaboration leading to cytokine production are undefined . CR3 ( Mac-1 , αMβ2 , or CD11b/CD18 ) is the principal β2 integrin known to contribute to fungal recognition in innate immune cells [3] . CR3 contains two ligand binding sites , I domain and lectin-like domain , which bind to protein ligands ( such as iC3b , ICAM-1 , and fibronectin ) and β-glucan , respectively [14] . CR3 is an enigmatic receptor which transduces diverse and distinct signaling upon engagement with different ligands [15 , 16] . Activation of CR3 is mediated by conformation change and regulated by inside-out and outside-in signals [17] . The inside-out signaling to activate CR3 can be initiated from other receptors , such as TLRs and Dectin-1 [18 , 19] . Engagement of CR3 also elicits outside-in signaling which activates innate immune effector functions , such as phagocytosis , cytotoxic killing , and cytokine production [17] . Despite many studies revealing the molecular mechanisms regulating CR3 phagocytosis [20] , the signaling pathway ( s ) responsible for its cytokine response is yet to be addressed . In contrast to CR3 , the understanding of Dectin-1 signaling is growing during the last decade . Dectin-1 engagement induces the phosphorylation of its intracellular ITAM-like motif leading to the recruitment and activation of Syk [21] . Syk facilitates the phosphorylation of PLCγ2 , allowing subsequent activation of MAPKs , AP-1 and NFAT and the assembly of Card9-Bcl10-Malt1 signalsome which mediates the canonical NF-κB activation [22–24] . A recent study also showed that Card9 bridges the interaction between Ras-GRF-1 and H-Ras , leading to downstream ERK activation upon Dectin-1 ligation [25] . In addition , Dectin-1 triggers Syk-independent Raf-1 activation through which antagonizes Syk-induced noncanonical NF-κB activation [26] . The requirement for Syk and the use of Card9 in Dectin-1 signaling differs in different macrophage and dendritic cell ( DC ) populations [27 , 28] . Thus , signaling pathway downstream of Dectin-1 and signaling crosstalk between Dectin-1 and heterologous PRRs can very well be cell-type specific . Here we used H . capsulatum and particulate ligands to study the molecular mechanism of collaboration between CR3 and Dectin-1 in macrophages . Our findings clearly showed that collaboration between CR3 and Dectin-1 in induction of TNF and IL-6 production was through synergistic activation of their downstream Syk-JNK-AP-1 signaling axis . In addition , while both CR3 and Dectin-1 were recruited and clustered on lipid raft microdomains upon encountering H . capsulatum , disruption of lipid raft hampered their collaboration in signaling activation and the subsequent cytokine response . Interestingly , CR3 and Dectin-1 cooperatively participated in host defense against disseminated histoplasmosis . Taken together , our results revealed the molecular mechanism underlying crosstalk between CR3 and Dectin-1 in enhancing cytokine response and demonstrated that they orchestrate adaptive antifungal immune response .
By use of blocking antibodies we previously showed that macrophage utilizes CR3 to phagocytose and both CR3 and Dectin-1 to mediate cytokine response to H . capsulatum [13] . Employing Itgam-/- , Clec7a-/- , and Itgam-/-Clec7a-/- macrophages here we investigated the mechanism of receptor crosstalk between CR3 and Dectin-1 ( S1A Fig ) . While losing either or both CR3 and Dectin-1 did not affect the propagation of intracellular H . capsulatum ( S2A Fig ) , TNF and IL-6 responses to either heat-killed or viable H . capsulatum were reduced in Itgam-/- and Clec7a-/- macrophages and the reduction was further enhanced in macrophages deficient in both receptors ( Figs 1A and S3A , and also refer to S2B Fig ) . Blocking Dectin-1 in Itgam-/- macrophages or blocking CR3 in Clec7a-/- macrophages also revealed that signals from these two receptors had additive effect in TNF and IL-6 production ( S4A Fig ) . It is worth noting that heat treatment did not change H . capsulatum recognition by macrophages . Both CR3 and Dectin-1 are known to signal through activation of Syk kinase [21 , 29] . We examined whether Syk is involved in the collaborative cytokine response induced by CR3 and Dectin-1 . Stimulation with heat-killed or viable H . capsulatum activated Syk kinase and Syk activation was dampened in Itgam-/- and Clec7a-/- macrophages as well as in WT macrophages blocked by anti-CR3 or anti-Dectin-1 antibody ( Figs 1B , S3B , S5A and S4B ) . The level of phosphorylated Syk was further reduced in Itgam-/-Clec7a-/- macrophages and in WT macrophages blocked by both anti-CR3 and anti-Dectin-1 antibodies ( Figs 1B , S3B and S4B ) . These results suggest that signals from CR3 and Dectin-1 cooperatively activate Syk kinase . Treatment with Syk inhibitors significantly reduced H . capsulatum-induced TNF and IL-6 production and the reduction was not due to the cytotoxic effect of inhibitors ( Figs 1C and S6 ) . While Syk deficiency did not affect the expression of CR3 and Dectin-1 , the deficiency almost completely abolished the production of TNF ( Figs 1D and S1B ) . These results clearly demonstrate that CR3 and Dectin-1 act in concert in macrophage cytokine response to viable as well as heat-killed H . capsulatum by intensifying Syk activation . To verify the nature of collaboration between receptors , specific particulate ligands for CR3 ( iC3b-coated bead ) and Dectin-1 ( depleted zymosan ) were used to stimulate cells . iC3b-coated beads , but not uncoated beads , and depleted zymosan additively induced TNF and IL-6 production in WT macrophages while Itgam-/- macrophages did not respond to stimulation by iC3b-coated beads neither did Clec7a-/- macrophages to depleted zymosan ( Figs 1E and S7A ) . Co-stimulation of macrophages with iC3b-coated beads and depleted zymosan enhanced the level of phosphorylated Syk compared to stimulation by either ligand alone ( Figs 1F and S7B ) and treatment with Syk inhibitors abolished TNF and IL-6 response induced by these two ligands ( Fig 1G ) . These data collectively reveal that Syk is the point where signals from CR3 and Dectin-1 converge to mediate collaborative cytokine response . It has been demonstrated that localization of Dectin-1 and CR3 to lipid raft microdomains is critical for signaling activation of each respective receptor [30 , 31] . While both CR3 and Dectin-1 diffusely distributed in the cytosol and on cell membrane in unstimulated macrophages , they were recruited and colocalized on lipid raft microdomains at the interface between macrophage and H . capsulatum yeasts upon stimulation ( Fig 2A ) . Isolation of lipid rafts by sucrose gradient confirmed that CR3 and Dectin-1 translocated to flotillin-1-enriched membrane microdomains ( Fig 2B ) . Stimulating macrophages with H . capsulatum induced the phosphorylation of Syk which was associated with lipid raft microdomains ( Fig 2C ) . It is noted that Syk was present in the lipid raft fractions on H . capsulatum-stimulated as well as on unstimulated cells , suggesting that rather than recruitment , H . capsulatum stimulation triggers phosphorylation of Syk that was already present on lipid rafts ( S8A and S8B Fig ) . Disruption of lipid rafts by methyl-β-cyclodextrin ( MβCD ) significantly reduced TNF and IL-6 production and cholesterol replenishment rescued the ability to produce TNF but not IL-6 ( Fig 2D ) . H . capsulatum-induced Syk phosphorylation was diminished in cells treated with MβCD and it was partially restored by cholesterol replenishment ( Fig 2E ) . These results demonstrate that activated Syk is associated with lipid raft microdomains where CR3 and Dectin-1 cluster and the integrity of lipid rafts is important to macrophage cytokine response and signaling activation induced by H . capsulatum . We used Micro-Western Arrays ( MWAs ) by employing 92 antibodies to probe for phosphorylated ( 84 antibodies ) and non-phosphorylated ( 8 antibodies ) signaling proteins known to be involved in the pathways downstream of PRRs for phagocytosis and cytokine production to screen for signaling molecules activated by H . capsulatum ( S9 Fig and S1 Table ) . The heat map shows that H . capsulatum induced phosphorylation of Syk , Raf-1 , PLCγ2 , PKC and molecules in the PI3K/Akt , NF-κB , and MAPK pathways at as early as 15 min and c-Jun and c-Fos ( two components of AP-1 ) at 60 min after stimulation ( Figs 3A and S9 ) . Activated signaling molecules were validated by conventional Western blot analysis . Consistent with the MWA data , H . capsulatum stimulation caused the increase of phosphorylated Syk , Akt , Raf-1 , JNK , ERK , p38 , IKKα/IKKβ , IκBα and NF-κBp65 at 15 min , while phosphorylation of c-Jun and c-Fos occurred at 60 min after stimulation ( Fig 3B ) . However , the amount of phosphorylated PLCγ2 and PKC isoforms ( PKCε , PKCη , PKCθ ) were below the limit of detection . Taken together , our data show that H . capsulatum stimulation leads to activation of Syk , Raf-1 , AP-1 , as well as molecules involved in the Akt/PI3K , NF-κB and MAPK pathways . We used pharmacological kinase inhibitors to identify the signaling molecule ( s ) that participates in macrophage cytokine response to H . capsulatum . Treatment with PI3K , JNK and ERK inhibitors greatly diminished TNF and IL-6 production , yet inhibiting Raf-1 enhanced TNF and IL-6 responses ( Fig 4A ) . Interestingly , p38 inhibitor had disparate effects on TNF and IL-6 production ( Fig 4A ) . LDH assay showed that pharmacological inhibitors at the concentrations we used did not have cytotoxic effect on the cells ( S6 Fig ) . Interestingly , Syk deficiency abrogated phosphorylation of Raf-1 and JNK but not that of Akt , ERK or p38 ( Fig 4B ) although inhibiting their activation diminished TNF and IL-6 production ( Fig 4A ) . Results in Fig 4C show that inhibition of JNK activation in cells stimulated with iC3b-coated beads and depleted zymosan significantly reduced TNF and IL-6 production yet inhibition of Raf-1 did not affect the production of either cytokine . These results indicate that JNK , a signaling molecule downstream of Syk , plays an important role in the coordinated CR3 and Dectin-1cytokine response . We next investigated the link between CR3 and Dectin-1 engagement and JNK activation . Results showed that JNK phosphorylation was reduced to about 50% of WT control in macrophages lacking either receptor across all time points after stimulation with heat-killed or viable H . capsulatum and JNK phosphorylation was almost completely abolished in Itgam-/-Clec7a-/- macrophages ( Figs 4D , S3C and S5B ) . Similarly , blocking either receptor reduced JNK phosphorylation and blocking both receptors reduced JNK phosphorylation even further ( Fig 4E ) . Stimulation of WT macrophages with both iC3b-coated beads and depleted zymosan increased the level of JNK phosphorylation compared to stimulation with either ligand alone ( Fig 4F ) and the additive effect of these two ligands was diminished in macrophages deficient in either receptor ( Fig 4G ) . Taken together , our results show that engagement of CR3 and Dectin-1 separately induces JNK phosphorylation and engagement of both have an additive effect on JNK activation which is downstream of Syk . Results in Fig 3 show that stimulation of macrophages with H . capsulatum triggered the activation of both NF-κB and AP-1 . To identify whether NF-κB or AP-1 mediates the collaboration between CR3 and Dectin-1 for cytokine production , we first clarified whether they were activated in Syk-/- macrophages . Fig 5A shows that H . capsulatum-induced c-Fos and c-Jun phosphorylation was greatly diminished in Syk-/- macrophages while IκBα phosphorylation and degradation was not affected , indicating that AP-1 , but not NF-κB , is the transcription factor downstream of Syk . In addition , while IκBα and NF-κBp65 phosphorylation was not affected in single and double knockout macrophages , activation of c-Fos and c-Jun was dampened in Itgam-/- and Clec7a-/- macrophages and almost completely abolished in Itgam-/-Clec7a-/- macrophages ( Figs 5B , 5C , S3D and S5C ) . Stimulation of WT macrophages with iC3b-coated beads and depleted zymosan together had an additive effect over stimulation with either one alone on c-Jun and c-Fos but not NF-κB activation ( Fig 5D and 5E ) . The additive effect was abolished in single knockout macrophages ( Fig 5F ) . When c-Jun and c-Fos were silenced by respective siRNA separately , TNF and IL-6 production was greatly reduced ( Fig 5G and 5H ) . Together these data show that AP-1 , but not NF-κB , is the transcription factor that mediates the collaborative cytokine response induced by CR3 and Dectin-1 . We employed disseminated histoplasmosis model , which is characterized by splenomegaly with large numbers of macrophages infiltrating the spleen , to investigate the contribution of CR3 and Dectin-1 in host defense against H . capsulatum infection [32] . WT , Itgam-/- , Clec7a-/- and Itgam-/-Clec7a-/- mice were intravenously infected with a low sublethal dose of H . capsulatum ( 2 . 5 × 105 ) . While the fungal loads in Itgam-/- and Clec7a-/- single knockout mice were comparable to ( Itgam-/- ) or slightly higher than ( Clec7a-/- ) WT mice , that in double knockout mice were significantly higher than either of the single knockout mice on 7 days after infection ( Fig 6A ) . Accompanying higher fungal loads , TNF , IFN-γ and IL-17 levels in splenocyte cultures of double knockout mice were lower than in either of the single knockout mice and IL-6 levels were lower than in Clec7a-/- mice ( Fig 6B ) . Interestingly , the percentages of IFN-γ-producing CD4+ and CD8+ cells were significantly reduced in Itgam-/- and Clec7a-/- mice compared to WT mice , and they were reduced even further in double knockout mice ( Fig 6C ) . Consistent with in vivo IFN-γ response , Itgam-/- , Clec7a-/- and Itgam-/-Clec7a-/- dendritic cells stimulated with H . capsulatum produced significantly less IL-12p35 transcripts and IL-12p40 transcripts was reduced in Itgam-/-Clec7a-/- dendritic cells compared to WT cells ( S10A and S10B Fig ) . These results together reveal that CR3 and Dectin-1 act in concert in instruction host adaptive immunity against H . capsulatum . Infection with a higher dose of H . capsulatum ( 5 × 106 ) , Itgam-/-Clec7a-/- mice had significantly greater fungal burden , higher mortality and succumbed at an earlier time point ( 10 to14 days ) compared to Itgam-/- , Clec7a-/- or WT mice ( Fig 6D and 6E ) . The number of total splenocytes in infected mice was significantly lower in Itgam-/-Clec7a-/- compared to either WT or single knockout mice ( Fig 6F ) . Impressively , infection greatly reduced the proportion and number of F4/80+ cell population in Itgam-/-Clec7a-/- mice , which was significantly lower than in either of the single knockout mice and WT mice ( Fig 6G and also refer to S11A and S11B Fig ) . These results suggest that macrophages in Itgam-/-Clec7a-/- mice fail to respond to H . capsulatum infection and/or rapidly undergo apoptosis . Macrophages being the major player in the interaction between the host and the fungal pathogen H . capsulatum , deficiency in innate receptors that recognize the pathogen for phagocytosis ( CR3 ) and proinflammatory cytokine production ( CR3 and Dectin-1 ) [13] greatly affects host defense against this pathogen . Furthermore , these two innate receptors CR3 and Dectin-1 , collaboratively function not only in innate immune response but also orchestrate adaptive immune response against disseminated histoplasmosis .
Recognition of invading pathogens by PRRs triggers innate immune responses and shapes the adaptive immunity . Fungal cell wall is complex in its composition which varies between species , strains , and morphological forms . Innate immune cells use a unique set of PRRs to recognize and respond to a given fungus . Therefore , there is a lot of interest to dissect the molecular mechanism of coordination between heterogeneous PRRs . In the present study , we showed that CR3 and Dectin-1 collaborate in regulating macrophage pro-inflammatory cytokine response . Engagement of CR3 and Dectin-1 induces their clustering on lipid raft microdomains which function as a platform for downstream signaling . Utilizing MWA , genetic approach and pharmacological inhibitors , we demonstrated that receptor crosstalk between CR3 and Dectin-1 enhances the activation of their downstream Syk-JNK-AP-1 signaling pathway . Furthermore , our results showed that CR3 and Dectin-1 both participate and function cooperatively in host defense against H . capsulatum by facilitating adaptive antifungal immunity . The pathogenic fungus H . capsulatum is assorted to chemotype I and II based on the absence and presence of α- ( 1 , 3 ) -glucan in the yeast cell wall [4] . H . capsulatum strain 505 we used in this study having its β- ( 1 , 3 ) -glucan readily exposed does not express α- ( 1 , 3 ) -glucan on the yeast cell surface which makes it likely to be assorted to chemotype I ( S12A Fig ) . Rappleye et al . showed that H . capsulatum yeast strain G186A AGS ( + ) expressing α- ( 1 , 3 ) -glucan on the outer cell wall layer ( S12B Fig ) is not recognized by Dectin-1 [33] . The isogenic strain ags1-null mutant lacking α- ( 1 , 3 ) -glucan is recognized by Dectin-1-expressing cells and induces TNF production in phagocytic cells [33] . We showed in a previous study and here that strain 505 induces macrophage TNF and IL-6 production through both CR3 and Dectin-1 , a phenomenon possibly unique to H . capsulatum that classified as chemotype I [13] . Interestingly , unlike what is reported in C . albicans [34] , heat treatment does not alter β-glucan expression on H . capsulatum strain 505 ( S12A and S12C Fig ) . We also discovered that Syk-JNK-AP-1 signaling and TNF and IL-6 production triggered by heat-killed H . capsulatum are comparable to that induced by viable organism . In the case of PRR crosstalk , previous studies showed that Syk is required for the synergy between TLR and Dectin-1 and acts as the convergence point of Dectin-1 and Dectin-2 signaling , both of which situations lead to NF-κB activation [7 , 9 , 12 , 35] . We demonstrated here that signaling crosstalk between CR3 and Dectin-1 through stimulation with H . capsulatum initiates at and converges on Syk . However , unlike receptor crosstalk between TLR and Dectin-1 and that between Dectin-1 and Dectin-2 , activation of Syk through CR3 and Dectin-1 does not connect to NF-κB pathway . Instead , Syk leads to activation of JNK and AP-1 . Consistent with H . capsulatum stimulation , co-stimulation with iC3b-coated beads and depleted zymosan enhances activation of Syk , JNK and AP-1 , but not NF-κB . Our findings clearly define Syk-JNK-AP-1 axis as the signaling pathway downstream of the collaborative interaction between CR3 and Dectin-1 . CR3 is unique among the members of β2 integrin family that it contains two ligand binding sites , I domain and lectin-like domain . Single or dual ligation of I domain and lectin-like domain in CR3 triggers disparate signaling pathways and cellular responses [15 , 17] . We have previously shown that CR3 recognizes H . capsulatum through both I domain and lectin-like domain within CD11b , demonstrating that H . capsulatum mediates dual CR3 ligation [13] . Here we observed that stimulation of macrophages with iC3b-coated beads ( single CR3 ligation of I domain ) does not induce Syk phosphorylation , while co-stimulation with iC3b-coated beads and depleted zymosan ( dual CR3 ligation and Dectin-1 ligation ) triggers Syk phosphorylation and the level of phosphorylation is greater than stimulation by depleted zymosan alone ( single CR3 ligation of lectin-like domain and Dectin-1ligation ) . Interestingly , stimulating Clec7a-/- macrophages with both iC3b-coated beads and depleted zymosan only minimally activates Syk . The lack of inside-out signaling provided by Dectin-1 may account for the failure for iC3b-coated beads and depleted zymosan to activate CR3 [19] . Together , our results show that dual ligation of CR3 optimizes the signaling synergy induced by simultaneous engagement of CR3 and Dectin-1 . An early study showed that H . capsulatum yeasts activate the alternative complement pathway which may lead to iC3b deposition [36] . Although CR3 can directly recognize and respond to H . capsulatum , whether complement opsonization would enhance the collaboration between CR3 and Dectin-1 or change PRR usage in macrophage interaction with H . capsulatum needs to be investigated . PRR clustering in membrane lipid microdomains is crucial for host cells to optimize detection of fungal pathogens by formation of phagocytic synapse and serving as a platform for signaling synergy [31 , 37 , 38] . In this study , we show that the spatial nearness of CR3 and Dectin-1 and that both of them being mobilized to and associated with lipid rafts are important to achieve their collaboration upon engagement with H . capsulatum . However , little is known about how PRRs are recruited to lipid rafts and how signaling crosstalk is induced by heterogeneous PRRs . It has been reported that intracellular osteopontin ( iOPN ) is essential for clustering of heterologous PRRs , including Dectin-1 , TLR2 and mannose receptor , that recognize Pneumocystis [38] . In addition , iOPN is involved in Dectin-1 and TLR2 downstream signaling by acting as a scaffold protein which associates with their respective downstream molecule Syk and IRAK1 [38] . It remains to be answered whether the steric assembly and the signaling synergy of CR3 and Dectin-1 induced by H . capsulatum is also mediated by iOPN , or by other adaptor molecule ( s ) . Macrophage plays multiple roles in host defense against H . capsulatum . Besides being a host cell , it also functions as cytokine/chemokine-producing cell and antigen donor cell , and it serves as effector cell when stimulated with IFN-γ , IL-17 and GM-CSF [39–42] . However , little is known about the signaling pathways downstream of functional receptors after macrophage encountering H . capsulatum . By MWA , this study is the first to uncover signaling pathways activated by H . capsulatum in macrophages . Our results reveal that PI3K , NF-κB , MAPK and AP-1 signaling pathways are activated by H . capsulatum . It is interesting to note that although PI3K , ERK and p38 activities modulate TNF and IL-6 production , they do not function downstream of Syk . In addition , inhibition of NF-κB does not affect H . capsulatum-induced TNF and IL-6 responses ( S13 Fig ) . Thus , we postulate that other pathways may act in parallel with Syk-JNK-AP-1 pathway to regulate H . capsulatum-induced TNF and IL-6 response in macrophages . We show that H . capsulatum induces Raf-1 activation in a Syk-dependent manner . Inhibition of Raf-1 increases H . capsulatum-induced TNF and IL-6 production . Our results define Raf-1 as a negative regulator in macrophage cytokine response to H . capsulatum . A previous study showed that Dectin-1 induces Raf-1 activation in a Syk-independent manner [26] . Activated Raf-1 antagonizes Syk-dependent non-canonical NF-κB activation by promoting inactive p65/RelB dimer formation [26] . Ligation of DC-SIGN activates Raf-1 which downregulates Borrelia burgdorferi- and TLR-induced TNF and IL-6 response by destabilizing their mRNAs and suppresses IL-12 response by impairing nucleosome remodeling at IL-12p35 promoter [43] . Together , these data suggest that activation of Raf-1 by ligation of PRRs negatively regulates and fine-tunes innate immune response . More studies are needed to demonstrate how Raf-1 negatively regulates H . capsulatum-induced TNF and IL-6 production . There are only limited studies on the role of AP-1 in PRR crosstalk and in host defense against fungal infections . It has been shown that TLR2 and Dectin-1 cooperatively regulate zymosan-induced IL-10 production in DCs through an ERK-dependent , but AP-1-independent mechanism [44] . Other reports showed that Dectin-1 engagement in DCs triggers AP-1 activation , while curdlan stimulates a PLCγ2-dependent pathway , β-glucan on Aspergillus fumigatus activates a Syk-dependent pathway [22 , 45 , 46] . Our results demonstrate that CR3 acts in concert with Dectin-1 to activate both c-Jun and c-Fos . In addition , knocking down c-Jun and c-Fos in macrophages decreases H . capsulatum-induced TNF and IL-6 response , highlighting the role of AP-1 in host defense against fungal infections . Our MWA data show that PLCγ2 is activated by H . capsulatum . However , whether PLCγ2 is involved in CR3/Dectin-1-mediated AP-1 activation still remains to be determined . It is interesting to note that AP-1 activation is known to be associated with several disorders ( ex . cancer and autoimmune diseases ) by regulating genes involved in cell proliferation , angiogenesis and inflammation and inhibition of AP-1 activation is identified as a promising therapeutic strategy [47–49] . Our findings raise the possibility that administration of AP-1 inhibitor may increase susceptibility to fungal infections by suppressing proinflammatory cytokine production . Thus , AP-1inhibitor should be used with caution as a treatment modality for cancer and inflammatory disorder . In addition to acting as a PRR to interact with pathogens , CR3 plays multiple roles in cellular processes including leukocyte extravasation , adhesion and chemotaxis [17] . This adds to the complexity of experimental design in addressing the role of CR3 in vivo . Intranasal or intraperitoneal infection of Itgam-/- mice with S . pneumonia results in infiltration of a large number of neutrophils to the infection sites [50 , 51] . Neutrophil recruitment to the peritoneum is reduced in mice lacking CR3 after intraperitoneally challenge with C . albicans , yet accumulation of neutrophils in the kidney is comparable between WT and CR3-deficient mice intravenously infected with C . albicans [19 , 52] . These studies demonstrated that abnormal neutrophil infiltration to the infection site should be considered as a factor influencing susceptibility in in vivo studies employing Itgam-/- mice . Indeed , we observed that unlike in WT mice , there is massive neutrophil infiltration to the lungs of Itgam-/- mice after intratracheal infection with H . capsulatum , a condition which may interfere the interaction between macrophages and the yeasts ( S14A and S14B Fig ) . By contrast , the percentage and the number of neutrophils recruited to the spleen of Itgam-/- mice were commensurate with that in WT mice after intravenous inoculation of H . capsulatum . To focus on the roles of CR3 and Dectin-1 in macrophage interaction with H . capsulatum , we resorted to employ the disseminated histoplasmosis model by intravenous inoculation of the organism instead of pulmonary infection . PRR signaling is known to act as a bridge that links innate and adaptive immunity . Signaling transduced by Dectin-1 can induce Th1 and Th17 responses [53] as well as priming cytotoxic T cells [54] . In addition , it has been reported that blockade of CR3 significantly reduces the Th1 and Th17 responses induced by A . fumigatus [55] . H . capsulatum infection triggers IFN-γ and IL-17 production by both CD4+ and CD8+ T cells , and both IFN-γ and IL-17 activate macrophages for inhibition of the replication of intracellular H . capsulatum [39 , 41 , 56 , 57] . Mice deficient in IFN-γ or treated with neutralizing antibody against IFN-γ or IL-17 are more susceptible to H . capsulatum , showing the importance of IFN-γ and IL-17 in clearing this intracellular fungal pathogen [57–59] . However , whether and how PRRs participate in the development of H . capsulatum-induced IFN-γ and IL-17 response remains unclear . In this study , we show that both CR3 and Dectin-1 contribute to and function collaboratively in regulating TNF , IL-6 , IL-17 and IFN-γ responses induced by H . capsulatum . TNF is a critical factor for the host to control H . capsulatum , and acts together with IFN-γ and IL-17 to provide protection [57 , 60] . Although the role of IL-6 in histoplasomsis has not been well addressed , previous studies showed that the generation of H . capsulatum-induced IL-17 response is dependent on IL-6 and IL-6 deficiency leads to impairment of Th1 response in mice infected with C . albicans or A . fumigatus [57 , 61 , 62] . Our results also showed that , in addition to regulate macrophage TNF and IL-6 production , both CR3 and Dectin-1 are involved in IL-12 response in dendritic cells , suggesting the role of these receptors in regulating dendritic cells cannot be ignored . Moreover , our in vitro study showed that deficiency in either CR3 or Dectin-1 or both did not affect the intracellular growth of H . capsulatum in macrophages , strengthening the possibility that CR3 and Dectin-1 deficiency resulting in susceptibility to disseminated H . capsulatum infection is due to their roles in regulating IFN-γ and IL-17 responses . It is interesting to note that macrophages in the spleens of Itgam-/-Clec7a-/- mice were greatly diminished ( almost exhausted ) after infection with a lethal dose of H . capsulatum , presenting a picture that the macrophages are losing the battle to the fungal pathogen . While dendritic cells are major antigen-presenting cells and macrophage are the host , cytokine-producing cell and effector cell in infection by H . capsulatum , our study reveals that CR3 and Dectin-1 are of vital importance not only in their collaborative roles in macrophage cytokine production but also in instructing adaptive immune response against disseminated histoplasmosis . In summary , we demonstrate for the first time the mechanism of receptor crosstalk between a member of the integrin family and CLR resulting in enhanced cytokine response . The collaboration between CR3 and Dectin-1 is through activation of Syk-JNK-AP-1 signaling pathway and dependent on formation of PRR clusters on lipid rafts . Our results also highlight the importance of CR3 and Dectin-1 in innate recognition that instructs antifungal adaptive immune response . Collectively , our findings provide a better understanding of the molecular mechanisms underlying the collaboration between CR3 and Dectin-1 and offer a valuable model for disentangling the intricacies of host-pathogen interactions .
All animal experiments were undertaken in accordance with the Guidebook for the Care and Use of Laboratory Animals , 3rd Ed . , 2007 , published by The Chinese-Taipei Society of Laboratory Animal Sciences , approved by the Institutional Animal Care and Use Committee ( IACUC , Permit number: 20130231 ) of National Taiwan University College of Medicine . Histoplasma capsulatum ( Hc ) strain 505 yeast cells were used in the whole study . Yeast cells were cultured at 37°C on brain heart infusion ( BHI ) agar supplemented with 1 mg/ml cysteine and 20 mg/ml glucose . Heat-killed yeast cells were prepared by treatment at 65°C for 2 h . To examine the surface expression of α-glucan and β-glucan , viable or heat-killed yeasts were fixed with 4% paraformaldehyde and stained with antibodies against α- ( 1 , 3 ) -glucan ( Clone MOPC-104E ) ( Biolegend , San Diego , CA , USA ) and β- ( 1 , 3 ) -glucan ( Biosupplies , Parkville , Australia ) followed by analysis with flow cytometry ( BD FACSCanto II , BD Biosciences ) . Itgam-/- ( Stock number: 003991 ) and wild-type C57BL/6 ( Stock number: 000664 ) mice were originally purchased from the Jackson Laboratories ( Bar Harbor , ME , USA ) and Clec7a-/- mice were generated by Dr . Gordon D . Brown [63] . Itgam-/-Clec7a-/- mice were generated by crossing Itgam-/- and Clec7a-/- mice . Mice heterozygous for a deletion in the Syk locus ( Syk+/- ) were obtained from Dr . Clifford Lowell ( University of California , San Francisco , CA , USA ) [64] . All strains used in this study were on C57BL/6 background . They were maintained and bred in the National Taiwan University College of Medicine Laboratory Animal Center ( NTU CMLAC ) or in the National Laboratory Animal Center ( NLAC , Taiwan ) under specific pathogen-free ( SPF ) conditions . In vivo infection experiments were performed following biosafety level 2 ( BSL-2 ) guidelines . Peritoneal macrophages were collected by lavage from mice at 4 days after peritoneal injection of 1 ml of 3% thioglycollate medium ( Sigma-Aldrich , St Louis , MO , USA ) . Macrophages deficient in Syk were derived from fetal liver cells obtained from Syk-/- mouse embryos ( E13 . 5-E15 . 5 ) . Syk-/- embryos were separated from Syk+/+ and Syk+/- embryos by their exhibition of severe petechiae and confirmed by genotyping [65] . Single-cell suspensions from fetal liver tissues were cultured in 20% L929-cell conditioned medium for 7 days . Over 95% of the adherent cells were F4/80+ which were identified as fetal liver-derived macrophages ( FLDMs ) . Syk inhibitors SykI and BAY 61-3606 , PI3K inhibitor LY294002 , ERK inhibitor U0126 , JNK inhibitor SP600125 , and p38 inhibitor SB203580 were purchased from Calbiochem-Merck ( Darmstadt , Germany ) . Raf-1 inhibitor GW5074 , methyl-β-cyclodextrin ( MβCD ) , and water-soluble cholesterol were obtained from Sigma-Aldrich . Antibodies against phospho ( p ) -Zap-70 ( Tyr319 ) /Syk ( Tyr352 ) , p-Akt ( Tyr308 ) , p-c-Raf ( Ser338 ) , p-ERK1/2 ( Thr202/Tyr204 ) , p-JNK ( Thr183/Tyr185 ) , JNK , p-p38 ( Thr180/Tyr182 ) , p-IKKα/β ( Ser176/180 ) , p-NF-κBp65 ( Ser536 ) , p-IκBα ( Ser32 ) , IκBα , p-c-Jun ( Ser63 ) , c-Jun , p-c-Fos ( Ser32 ) , and c-Fos were purchased from Cell Signaling ( Beverly , MA , USA ) . Anti-Syk , anti-β-actin , HRP-conjugated anti-rabbit IgG , and HRP-conjugated anti-mouse IgG antibodies were purchased from GeneTex Inc . ( Irvine , CA , USA ) . Blocking antibodies against CR3 ( clone 5C6 ) and Dectin-1 ( clone 2A11 ) were purchased from Serotec ( Oxford , UK ) . Antibodies for cell surface staining , anti-CD11b ( clone M1/70 ) , anti-CD18 ( clone GAME-46 ) , and APC-conjugated anti-F4/80 antibody were obtained from eBioscience ( San Diego , CA , USA ) , and anti-Dectin-1 ( clone 218820 ) was purchased from R&D Systems ( Minneapolis , MN , USA ) . Macrophages were stimulated with viable or heat-killed H . capsulatum yeasts ( at a yeast-to-macrophage ratio of 20/1 ) or iC3b-caoted beads ( at a bead-to-cell ration of 10/1 ) and 50 μg/ml depleted zymosan ( InvivoGen , San Diego , CA , USA ) . The iC3b-coated beads were prepared as described previously [66] . Briefly , 2 × 108 of 3 μm Latex beads ( Sigma-Aldrich ) were incubated with PBS containing 20 μg/ml human IgM ( Sigma-Aldrich ) at 37°C for 60 min . Beads were washed with PBS then resuspended in freshly isolated mouse serum ( diluted 1:1 in PBS ) and incubated at 37°C for another 20 min . Beads were washed with Hank's balanced salt solution then resupspended in RPMI 1640 medium supplemented with 10% heat-inactivated FBS . During the process , the classical pathway of complement cascade was activated , resulting in C3b deposition on the surface of beads where it was rapidly and completely converted to iC3b [67] . Macrophages were stimulated with or without H . capsulatum or particulate ligands . Culture supernatants were collected after incubation at 37°C for different periods of time . The concentrations of TNF and IL-6 in the supernatants were quantified by enzyme-linked immunosorbent assay ( ELISA ) kit ( eBioscience ) following the manufacturer’s instructions . Cells were lysed by PhosphoSafe Extraction Reagent cell lysis buffer ( Novagen , Madison , WI , USA ) . Whole cell lysates were subjected to electrophoresis at 10% sodium dodecyl sulfate polyacrylamide gel ( SDS-PAGE ) , and transferred to PVDF membrane . The membrane was incubated in buffer containing primary antibody against molecule of interest , followed by HRP-conjugated secondary antibody . The blot was developed by chemiluminescence using ECL solution ( Millipore , Billerica , MA , USA ) . For normalization , the intensity of blots was quantified by ImageJ software ( NIH , Bethesda , MD , USA ) . Macrophages ( 3 × 107 ) were lysed with 0 . 5% Brij in TNE buffer [25 mM Tris ( pH 7 . 5 ) , 150 mM NaCl , 5 mM EDTA , protease inhibitors , 1 mM Na3VO4 , and 1 mM NaF] and let stand on ice for 1 h . Lysates were then mixed with equal volume of 80% sucrose in TNE buffer and overlaid with 30% and 5% sucrose in the same buffer . The gradients were centrifuged at 40 , 000 × g in a SW55Ti rotor ( Beckman Coulter , Fullerton , CA , USA ) at 4°C for 18 h . Twelve fractions were collected and the proteins in the fractions were subjected to electrophoresis at 10% SDS-PAGE and Western blot analysis by using antibodies against CD11b ( GeneTex ) , Dectin-1 ( Santa Cruz Biotechnology , Santa Cruz , CA , USA ) , Syk , p-Syk , and flotillin-1 ( Cell Signaling ) . Macrophages were allowed to adhere on cover slide overnight and stimulated with heat-killed H . capsulatum yeasts . After stimulation , cells were fixed with 3% paraformaldehyde followed by permeabilization with 0 . 5% Triton X-100 . Cells were blocked with PBS containing 5% heat-inactivated FBS and stained with Alexa Fluor 647-conjugated cholera toxin B ( Invitrogen , Carlsbad , CA , USA ) , anti-p-Syk ( Cell Signaling ) , anti-Dectin-1 ( Serotec ) , and/or PE-conjugated anti-CD11b ( eBioscience ) antibodies . Cells were then stained with secondary Alexa Fluor-conjugated secondary antibodies ( Jackson ImmunoResearch , West Grove , PA , USA ) . Cell nuclei were stained with Hoechst 33258 . The images were acquired with a Zeiss Axiovert 100TV confocal microscope ( Carl Zeiss Inc . , Jena , Germany ) and analyzed by Zen software ( Carl Zeiss Inc . ) and ImageJ software ( NIH ) . Peritoneal macrophages were stimulated with or without heat-killed H . capsulatum . Cells were lysed at different time points , and Micro-Western Arrays ( MWAs ) were performed to measure protein expression as previously described [68] . Blots were analyzed by Odyssey analysis software ( Li-Cor Biosciences , USA ) . Heat maps were created by using PermutMatrix software ( LIRMM ) . Peritoneal macrophages ( 1 × 106 ) were transfected with 30 pmol of control siRNA or siRNAs targeting c-Fos or c-Jun ( Santa Cruz Biotechnology ) using the Amaxa Nucleofector kit for mouse macrophages ( Lonza , Basel , Switzerland ) with a Nucleofector II electroporation device ( Lonza ) . After transfection , cells were gently resuspended in RPMI 1640 medium supplemented with 20% heat-inactivated FBS and plated in 12-well tissue culture plate . Adherent cells were collected for cytokine assay and Western blot analysis 48 h later . Mice were injected intravenously with low ( 2 . 5 × 105 ) or high ( 5 × 106 ) dose of H . capsulatum yeasts suspended in PRMI 1640 medium . For survival studies , mice were monitored for up to 30 days . For immunological studies , mice were killed on day 7 ( low dose ) or day 9 ( high dose ) after infection . To determine the fungal burden , spleens were weighted and homogenized in sterile RPMI 1640 medium . The homogenates were serially diluted and plated on glucose-pepton agar plates . Mycelial colonies were counted 10 days after incubation as described elsewhere [40] . To determine the percentage of leukocyte populations in spleen , splenocytes were stained for surface CD4 , CD8 , B220 , F4/80 , Ly6G and CD11c and analyzed by flow cytometry . All antibodies were purchased from eBioscience . To study cytokine production by splenocytes , single cell suspensions were prepared from the spleen . Five million splenocytes were cultured in RPMI 1640 complete medium containing 400 pg/ml of rIL-2 for 48 h . The concentrations of TNF , IL-6 , IL-17A and IFN-γ in the culture supernatants were quantified by ELISA . To analyze intracellular IFN-γ , 1 × 106 splenocytes were cultured in RPMI 1640 complete medium for 24 h and monensin ( 2 μM , Sigma-Aldrich ) was added at 6 h before harvest . Cells were stained with anti-CD4 , anti-CD8 and anti-IFN-γ antibodies as described previously [40] . The percentage of IFN-γ-producing cells in the total CD4+ or CD8+ T cell populations was calculated by dividing the % of IFN-γ+CD4+ cells or IFN-γ+CD8+ by the % of CD4+ or CD8+ cells . All antibodies were purchased from eBioscience . The comparisons between multiple groups were analyzed with one-way ANOVA followed by Tukey post-hoc test or by Duncan post-hoc analysis using SPSS 22 . 0 statistical software ( IBM , Armonk , NY , USA ) . The differences between two groups were tested by two-tailed t-test . Generalized Wilcoxon test was used to analyze mouse survival . Differences were considered significant at a P value of < 0 . 05 . The accession numbers in the UniPortKB/SwissProt database of the proteins mentioned in this study are followed: CD11b , P05555; CD18 , P11835; Dectin-1 , Q6QLQ4; Syk , P43404; JNK , Q91Y86 ( JNK1 ) and Q9WTU6 ( JNK2 ) ; c-Fos , P01101; c-Jun , P05627; Raf-1 , Q99N57; PLCγ2 , Q8CIH5; Akt , P31750; ERK , Q63844 ( ERK1 ) and P63085 ( ERK2 ) ; p38 , P47811; IKKα , Q60680; IKKβ , O88351; IκBα , Q9Z1E3; NF-κBp65 , Q04207; PKCε , P16054; PKCη , P23298; PKCθ , Q02111; β-actin , P60710; GAPDH , P16858; flotillin-1 , O08917; TNF , P06804; IL-6 , P08505; IL-12p35 , P43431; IL-12p40 , P43432; IL-17A , Q62386; IFN-γ , P01580 . | The incidence of life-threatening fungal infections is increasing during the last decades . A better understanding of the interactions between fungal pathogen and its host cell is important to the development of new therapeutic strategies against fungal infections . Dimorphic fungus Histoplasma capsulatum becomes disseminated and threatens life in immunocompromised individuals . This fungal pathogen utilizes complement receptor 3 ( CR3 ) and Dectin-1 , two pattern recognition receptors on the surface of innate immune cells , to induce macrophage cytokine response . In this study , we demonstrated that CR3 and Dectin-1 act collaboratively to induce macrophage TNF and IL-6 response through a mechanism dependent on activation of the Syk-JNK-AP-1 signaling axis . CR3 and Dectin-1 are recruited and form clusters on lipid raft microdomains upon stimulation by H . capsulatum , leading to activation of their signaling convergence at Syk kinase and induction of subsequent cytokine response . In addition , we showed that CR3 and Dectin-1 cooperatively instruct the adaptive antifungal immunity to defense against H . capsulatum infection . Our findings define the molecular mechanisms underlying receptor crosstalk between CR3 and Dectin-1 and provide a valuable model for receptor collaboration in the context of host-fungus interactions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | CR3 and Dectin-1 Collaborate in Macrophage Cytokine Response through Association on Lipid Rafts and Activation of Syk-JNK-AP-1 Pathway |
Dyskeratosis congenita ( DC ) is a heterogeneous inherited bone marrow failure and cancer predisposition syndrome in which germline mutations in telomere biology genes account for approximately one-half of known families . Hoyeraal Hreidarsson syndrome ( HH ) is a clinically severe variant of DC in which patients also have cerebellar hypoplasia and may present with severe immunodeficiency and enteropathy . We discovered a germline autosomal recessive mutation in RTEL1 , a helicase with critical telomeric functions , in two unrelated families of Ashkenazi Jewish ( AJ ) ancestry . The affected individuals in these families are homozygous for the same mutation , R1264H , which affects three isoforms of RTEL1 . Each parent was a heterozygous carrier of one mutant allele . Patient-derived cell lines revealed evidence of telomere dysfunction , including significantly decreased telomere length , telomere length heterogeneity , and the presence of extra-chromosomal circular telomeric DNA . In addition , RTEL1 mutant cells exhibited enhanced sensitivity to the interstrand cross-linking agent mitomycin C . The molecular data and the patterns of inheritance are consistent with a hypomorphic mutation in RTEL1 as the underlying basis of the clinical and cellular phenotypes . This study further implicates RTEL1 in the etiology of DC/HH and immunodeficiency , and identifies the first known homozygous autosomal recessive disease-associated mutation in RTEL1 .
Hoyeraal Hreidarsson syndrome ( HH ) is a clinically severe variant of the telomere biology disorder dyskeratosis congenita ( DC ) [1] . DC is a heterogeneous inherited bone marrow failure syndrome ( IBMFS ) diagnosed by the presence of the classic triad of dysplastic nails , abnormal skin pigmentation , and oral leukoplakia . However , substantial clinical heterogeneity has been observed and the phenotype may include pulmonary fibrosis , liver disease , esophageal , urethral , or lacrimal duct stenosis , developmental delay , and/or other complications . Individuals with DC are at very high risk of bone marrow failure ( BMF ) , myelodysplastic syndrome , and cancer [2] . The clinical consequences of DC manifest at variable ages and in different patterns , even within the same family . Independent of the classic triad , lymphocyte telomere lengths less than the first percentile for age are diagnostic of DC [3] . Depending on the affected gene , DC can be inherited in X-linked recessive ( XLR ) , autosomal dominant ( AD ) , or autosomal recessive ( AR ) patterns . Germline mutations in DKC1 result in XLR inheritance , mutations in TERC , TERT , RTEL1 , or TINF2 result in AD inheritance , and mutations in TERT , RTEL1 , CTC1 , NOP10 , NHP2 , or WRAP53 result in AR inheritance [4]–[7] [8]; mutations in these genes account for approximately one-half of classic DC cases . Patients with HH have many of the DC features listed above; however , severe immunodeficiency [9] , non-specific enteropathy , intrauterine growth retardation ( IUGR ) , and developmental delay may be the presenting features . In addition to features of DC , the presence of cerebellar hypoplasia is often the basis for a diagnosis of HH [1] . Patients with HH have extremely short telomeres , even when compared with other DC patients [3] . Germline mutations in DKC1 ( XLR ) , TINF2 ( AD ) , or TERT ( AR ) have been shown to cause HH . The causative mutation in HH is known in less than one-half of cases . We clinically characterized individuals with HH from two different families . The affected individuals had IUGR , immunodeficiency , enteropathy , and extremely short telomeres . In both families , we discovered homozygous recessive germline mutations in Regulator of Telomere Elongation Helicase 1 ( RTEL1 ) and characterized the telomere defect that resulted from these mutations . While RTEL1 mutations have been previously implicated in AD and AR compound heterozygous cases of DC , HH , and DC-like cases [6] , [7] , this report is the first instance of a homozygous DC-causative mutation in this gene .
This study demonstrates the clinical and molecular consequences of homozygous autosomal recessive mutations in RTEL1 . We identified two families with children who had HH , were of AJ ancestry , and had the same homozygous RTEL1R1264H mutations . These data provide further evidence that defects in RTEL1 function can lead to clinical phenotypes consistent with the HH variant of DC [6] . Our molecular analyses indicate that the homozygous RTEL1R1264H mutation results in short , heterogeneous telomeres . Additionally , cell lines bearing this mutation produce excess extrachromosomal T-circles , but only in the presence of functioning DNA replication machinery . RTEL1 is proposed to resolve T-circles to enable proper telomeric replication; in the absence of this activity , T-loops are inappropriately resolved as a circle when encountered by the replication machinery , resulting in a shortened telomere [18] . T-circle formation in the presence of RTEL1R1264H is SLX4-dependent , similar to T-circle formation in RTEL1-deficient cells [14] . RTEL1 also aids in suppressing inappropriate recombination throughout the genome . We have shown that the RTEL1R1264H mutation results in a modest enhancement in sensitivity to DNA damage , as well as an increase in SCE , indicating that the RTEL1R1264H mutation impairs both telomeric and non-telomeric aspects of RTEL1 function . The fact that both the probands were homozygous for the identical risk haplotype suggests that there is an ancestral haplotype that is shared by parents in both families ( Figure 1A and 1B ) . We were able to reconstruct the haplotype based on the genotypes obtained using Sanger sequencing . This haplotype was also seen without the mutation in 14/378 ( TSI/GBR/FIN ) samples of EUR ethnicity in the 1000 Genomes data . Together with the occurrence of the risk haplotype in the two families with AJ ethnicity , the evidence supports the interpretation that this mutation is confined to EUR populations and is most likely an AJ founder mutation . We have not extended the 34 kb haplotype further since the number of individuals with this rare recessive disorder in our study is too small to investigate the age of the mutation based on haplotypes and population history . We and others recently reported that AD nonsense RTEL1 mutations are present in HH and that an additional missense mutation in the helicase domain further exacerbates the clinical and telomere length phenotype , while the presence of only a single missense mutation in the helicase domain resulted in a less clinically severe phenotype [6] , [7] . [8] The current study provides important insight into the function of the C-terminal end of the human RTEL1 protein . RTEL1 deficiency confers embryonic lethality in mice [19] , suggesting that the R1264H allele is hypomorphic . As is the case for the two families described here , hypomorphs are usually recessive; for example , AR partial loss-of-function mutations in FANCD2 cause Fanconi anemia and AR LIG4 mutations result in Ligase IV syndrome [20] , [21] . Furthermore , this mutation is distal to the RTEL1 helicase domain , and is thus unlikely to directly affect enzymatic activity . Nevertheless , the phenotypic impact of RTEL1R1264H at the cellular level was pronounced . The RTEL1R1264H mutation falls within exon 34 , which encodes a predicted C4C4 RING domain of RTEL1 , lying downstream of a putative PIP box . Many RING domain-containing proteins are E3 ubiquitin ligases that interact with E2 ubiquitin-conjugating enzymes through their RING domains . BRCA1 , MDM2 , and Parkin are all examples of RING domain-containing proteins that are involved in human disease [22] . The putative RTEL1 RING domain is distant from the helicase domain , suggesting that the RTEL1R1264H mutation may affect the RING domain while leaving the helicase activity intact . Given the severity of the clinical and cellular phenotypes of this mutation , the data suggest that this domain exerts a significant influence on the biological function of RTEL1 . Further analysis of this domain to define the mechanism ( s ) of its influence is ongoing . These findings , together with the recent report that non-coding SNPs in RTEL1 have been found to be associated with susceptibility to high-grade glioma [23]–[25] , broadly implicate the RTEL1 locus in human cancer susceptibility . Given the cellular phenotypes of DC/HH and those reported here , the clinical features of DC are likely sequelae of defects in maintenance and functions of the telomere . We have demonstrated that the RTEL1R1264H mutation affects both the telomeric and non-telomeric functions of RTEL1 . Individually , proteins involved in either telomere maintenance or DNA repair can result in immunodeficiency when perturbed: DC is an example of the former , and Bloom syndrome of the latter . The patients described here exhibit severe immunodeficiency , which may be the result of a mutation affecting both of these pathways . However , future studies are required to better understand this observation .
This research was approved by the Institutional Review Boards ( IRB ) of the National Cancer Institute and Memorial Sloan Kettering Cancer Center . All participants or their parents signed IRB-approved informed consent forms . Patient NCI-318 and her family were participants in an IRB-approved longitudinal cohort study at the National Cancer Institute ( NCI ) entitled “Etiologic Investigation of Cancer Susceptibility in Inherited Bone Marrow Failure Syndromes” ( NCI 02-C-0052 , ClinicalTrials . gov Identifier: NCT00027274 ) . In this study , patients and their family members complete questionnaires and undergo thorough clinical evaluations at the NIH Clinical Center [2] . Telomere length was measured by flow cytometry with fluorescent in situ hybridization ( flow FISH ) in leukocytes [26] . THE MSKCC proband was ascertained on IRB-approved protocol 95-091 entitled “Collection of Hematopoietic Progenitor Cell and/or Blood Samples From Patients For Research Studies . ” Other family members consented to germline testing in the Clinical genetics Service , as well as MSKCC 93-102 “Ascertainment of Peripheral Blood or Saliva Samples for Genetic Epidemiology Studies of Familial Cancers , ” as well as a specific consent for the novel homologous recombination gene described in this report . Whole exome sequencing for family NCI-318 was performed at the NCI's Cancer Genomics Research Laboratory as previously described [6] . Reads were aligned to the hg19 reference genome using Novoalign software version 2 . 07 . 14 ( http://www . novocraft . com ) , Picard software version 1 . 67 ( http://picard . sourceforge . net/ ) and the Genome Analysis Toolkit ( GATK , http://www . broadinstitute . org/gatk/ ) [27] . Variant discovery , genotype calling , and annotation were performed as described [6] using data from the UCSC GoldenPath database ( http://hgdownload . cse . ucsc . edu/goldenPath/hg19/database/ ) , the ESP6500 dataset from the Exome Variant Server , NHLBI Exome Sequencing Project ( ESP ) , Seattle , WA ( http://evs . gs . washington . edu/EVS/ ) ( accessed August 2012 ) , the Institute of Systems Biology KAVIAR ( Known VARiants ) database ( http://db . systemsbiology . net/kaviar/ ) [28] , the National Center for Biotechnology Information dbSNP database ( http://www . ncbi . nlm . nih . gov/projects/SNP/ ) [29] build 137 , and the 1000 Genomes ( http://www . 1000genomes . org/ ) [12] . Variants were also annotated for their presence in an in-house database consisting of over 700 whole exomes that were sequenced in parallel with our DC families . Variants within each family were filtered and categorized as indicated in Table S1 . Validation of exome sequencing findings in the NCI-318 trio was performed by sequencing coding exons of RTEL1 . Primer sequences are shown in Table S4 . All samples were amplified using KAPA2 RobustHotstart Readymix ( 2× ) ( Kapa Biosystems , Johannesburg , South Africa ) and the following cycling conditions: 3 min at 95° , followed by 30 cycles of 15 sec at 95° , 15 sec at 60° , 15 sec at 72° , followed by 10 min at 72° . Amplicons were purified using Agencourt's Ampure XP beads , then libraries were constructed and barcoded using the Ion Xpress Plus Fragment Library Kit ( Life Technologies , Carlsbad , CA , USA ) . DNA tagged beads were generated for sequencing using Life Technologies' OneTouch and run on an Ion 316 chip on the Ion PGM Sequencer ( Life Technologies ) . The default TMAP aligner and variant caller was used to generate a variant list per sample . Targeted resequencing of DNA damage response genes was instrumental in the discovery of the RTEL1 mutation at MSKCC . Genomic enrichment via microfluidic PCR was conducted using the primer pool from Raindance Technologies [30] . Resulting libraries were prepared for sequencing using the SOLiD 4 sequencer ( Life Technologies , Carlsbad ) . Read alignment and base-calling was done using the ABI Bioscope software with parameters optimal for targeted resequencing . Reads were filtered for mapping quality . RTEL1 contained the most biologically relevant non-synonymous exonic variant . MSK-41 was included in a panel of 24 cell lines in which targeted DNA sequencing of approximately 300 DNA damage response genes ( including RTEL1 ) was carried out ( see methods [13] ) . PolyPhen-2 [31] ( http://genetics . bwh . harvard . edu/pph2 ) , SIFT [32] ( http://sift . jcvi . org ) , and Condel [33] ( http://bg . upf . edu/condel/home ) were used to predict the severity of RTEL1 amino acid substitutions . Multiple sequence alignments were generated for homologous RTEL1 protein sequences using T-Coffee [34] ( www . tcoffee . org ) to evaluate conservation . Alignments were generated with NCBI Reference Sequence , GenBank or Ensembl proteins ENSP00000353332 ( Homo sapiens ) , NP_001124929 . 1 ( Pongo abelii ) , NP_001091044 . 1 ( Bos taurus ) , and EDL07405 . 1 ( Mus musculus ) . Telomere FISH was performed as described [35] . Images were captured at 100× magnification , with precisely the same exposure time for each genotype ( MSK-41 hTERT and BJ hTERT ) . Sensitivity ( gain ) is increased to saturation , and chromosome ends for which there still appears no signal are scored as signal-free ends . The heterogeneity observed in Figure 2C was reproducible over several experiments , and with different probes ( data not shown ) . Cells were collected from 2 to 3 10 cm plates at 70% confluence for each condition . Genomic DNA extraction was performed as described [36] . DNA was double digested by AluI/HinfI restriction enzymes overnight before starting TCA assay and then Southern Blot as described [37] with minor modifications to Phi29 DNA polymerization ( MBI Fermentas ) with a mammalian telomeric primer and a mammalian telomeric probe for hybridization . Blot images were captured and quantified with Storm 840 scanner and ImageQuant TL software ( Amersham Biosciences ) . Sister chromatid exchange analysis and telomere FISH were carried out as described previously [35] . Mitomycin C sensitivity assays were as described [38] . To detect SLX4 levels in the various knockdown conditions , we immunoprecipitated SLX4 ( 1 . 5 mg protein lysate , 10 µg of antibody ) with rabbit polyclonal antibody ( A302-269A ) followed by western blotting with polyclonal rabbit antibody A302-270A . Both antibodies were from Bethyl . T-circles were detected and quantified as previously described [14] . Immortalized conditional RTEL1F/- MEFs were as previously described [14] and were cultured in DMEM containing 10% fetal bovine serum . Cre recombination was carried out with Ad5-CMV-Cre adenovirus ( Vector Biolabs ) for 96 hr as described [39] . Cells were either not treated or treated with aphidicolin ( 5 µM ) for 24 hrs . | Patients with dyskeratosis congenita ( DC ) , a rare inherited disease , are at very high risk of developing cancer and bone marrow failure . The clinical features of DC include nail abnormalities , skin discoloration , and white spots in the mouth . Patients with Hoyeraal-Hreidarsson syndrome ( HH ) have symptoms of DC plus cerebellar hypoplasia , immunodeficiency , and poor prenatal growth . DC and HH are caused by defects in telomere biology; improperly maintained telomeres are thought to be a major contributor to carcinogenesis . In half the cases of DC , the causative mutation is unknown . By studying families affected by DC for whom a causative mutation has not yet been identified , we have discovered a homozygous germline mutation in RTEL1 , a telomere maintenance gene that , if mutated , can result in HH . The mutations result in the inability of the RTEL1 protein to function properly at the telomere , and underscore its important role in telomere biology . | [
"Abstract",
"Introduction",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | A Recessive Founder Mutation in Regulator of Telomere Elongation Helicase 1, RTEL1, Underlies Severe Immunodeficiency and Features of Hoyeraal Hreidarsson Syndrome |
It was recently reported that the recBC mutants of Escherichia coli , deficient for DNA double-strand break ( DSB ) repair , have a decreased copy number of their terminus region . We previously showed that this deficit resulted from DNA loss after post-replicative breakage of one of the two sister-chromosome termini at cell division . A viable cell and a dead cell devoid of terminus region were thus produced and , intriguingly , the reaction was transmitted to the following generations . Using genome marker frequency profiling and observation by microscopy of specific DNA loci within the terminus , we reveal here the origin of this phenomenon . We observed that terminus DNA loss was reduced in a recA mutant by the double-strand DNA degradation activity of RecBCD . The terminus-less cell produced at the first cell division was less prone to divide than the one produced at the next generation . DNA loss was not heritable if the chromosome was linearized in the terminus and occurred at chromosome termini that were unable to segregate after replication . We propose that in a recB mutant replication fork breakage results in the persistence of a linear DNA tail attached to a circular chromosome . Segregation of the linear and circular parts of this “σ-replicating chromosome” causes terminus DNA breakage during cell division . One daughter cell inherits a truncated linear chromosome and is not viable . The other inherits a circular chromosome attached to a linear tail ending in the chromosome terminus . Replication extends this tail , while degradation of its extremity results in terminus DNA loss . Repeated generation and segregation of new σ-replicating chromosomes explains the heritability of post-replicative breakage . Our results allow us to determine that in E . coli at each generation , 18% of cells are subject to replication fork breakage at dispersed , potentially random , chromosomal locations .
The bidirectional replication of the Escherichia coli circular chromosome starts at the replication origin oriC and ends when forks meet in the opposite region , the chromosome terminus . Replication forks are arrested in the terminus region by specific sites called ter where binding of the Tus protein blocks replication forks in an orientation-specific manner ( reviewed in [1 , 2] ) . ter sites are oriented to form a replication fork trap , replication forks can enter the trap but their exit is delayed by pauses at several successive ter sites ( Fig 1A and 1B ) . As chromosome segregation is concurrent with replication in bacteria , the origin and terminus regions are also the first and the last DNA sequences to be segregated during chromosome partitioning [3–5] . Following replication initiation , the two origins first remain associated at mid-cell for about 20 min and then move to the ¼ and ¾ positions of the cell . Then , the chromosome arms segregate from mid-cell to these positions as they are replicated . Finally , the terminus regions are also replicated at mid-cell and only separate shortly before cell division [3–5] . The chromosome terminus is organized in a large Ter macrodomain of about 780 kilobases ( kb ) by binding of the MatP protein to specific DNA motifs , the matS sites [6] . MatP also interacts with the septum protein ZapB , and thus maintains the Ter macrodomain at midcell during septum formation [7–9] . The terminus region is centred on a specific site called dif , the target of recombinases XerC and XerD for chromosome dimer resolution ( reviewed in [10 , 11] ) . dif is positioned opposite oriC on the circular chromosome ( Fig 1A ) , and is the inversion point of the GC strand skew . Specific motifs , KOPS ( FtsK oriented polar sequences ) , which provide directionality of chromosome segregation , converge at the dif site ( reviewed in [12] ) . They are recognized by the C-terminal domain of a septum-protein , the FtsK translocase which acts as an oriented DNA pump . KOPS motifs point from the origin of replication towards dif , allowing FtsK to bring newly replicated dif sites together at mid-cell and to remove DNA from the constricting septum [13 , 14] . As a result dif sites are the last region to be segregated away from mid-cell [5 , 15] . Recently a new phenomenon was described in the terminus region . Sequencing of the entire genome and analysis of DNA sequence coverage as a function of position on the chromosome ( Marker Frequency Analysis , MFA ) has revealed a deficit of sequences in the chromosome terminus region in the recB mutant [16–18] DNA double strand break ( DSB ) repair in E . coli is entirely dependent on homologous recombination , first steps of which are catalysed by RecBCD and RecA ( reviewed in [20–22] ) . RecBCD is a heterotrimeric complex that binds to double-stranded DNA ( dsDNA ) ends . RecB and RecD are helicases , and RecB also acts as a nuclease . RecBCD degrades dsDNA ends until it encounters specific DNA motifs called chi sites , after which it continues to degrade the 5’ end . It then loads RecA on the protruding 3’ tail for homology search , strand invasion and strand exchange . The resulting Holiday junctions are resolved by RuvABC resolvase to generate recombination products . In the absence of RecA , DSBs lead to chromosome degradation because of the potent exonuclease activity of RecBCD . Indeed the complex was originally characterised as the major E . coli exonuclease , Exo V . recB and recC null mutants are deficient for DSB repair , but because the RecBC complex can still catalyze strand opening and RecA loading , recD mutants are Rec+ . However , Exo V activity is abolished in all three null mutants , recB , recC and recD , even though the recD mutants still degrade linear DNA in vivo at 50% of the wild-type rate [23] Finally , RecBCD-dependent homologous recombination is coupled with replication restart , which allows chromosome replication to resume after the repair by homologous recombination of broken replication forks ( reviewed in [24] ) . In a previous study we showed that the deficit of terminus DNA sequences observed in the chromosome of recB mutant cells , which we call terminus DNA loss ( Fig 1C , S1 Fig ) , was independent of all known DNA processing events to take place in the terminus: replication fork merging , dimer resolution and decatenation of the two circular replicated chromosomes [19] . It also occurred in cells lacking FtsK-mediated chromosome segregation , but in an ftsK mutant , terminus DNA loss became less centred at dif , indicating a role for FtsK in the positioning of the peak of DNA loss around the site of convergence of KOPS sequences [19] . Our study led to the following key observations: ( i ) terminus DNA loss occurred during septum closure and required cell division , ( ii ) a first cell division generated one daughter cell that lacked the terminus sequence , and one that retained it ( the initial event ) , ( iii ) the daughter cell that carried the terminus sequence generated again a non-proliferating terminus-less cell and a viable terminus-containing cell , at each following generation ( heritable , transmitted events; [19]; Fig 1D ) . Furthermore , our analysis by RecA ChIP suggested that these terminus DSBs did not occur in wild-type cells , and were thus caused by the absence of RecBCD [19] . Here we have taken forward our previous study and used MFA and cell biology techniques to understand these mysterious observations . We propose and test a model in which , in a recB mutant , replication fork breakage triggers a terminus DSB during cell division in a heritable manner . Our results allow us to conclude that in wild-type , untreated E . coli cells , chromosome DSBs occur mainly at replication forks , and to determine the frequency of spontaneous replication fork breakage to be ~18% per cell per generation .
We studied terminus DNA loss by a combination of MFA and microscopy analyses . For microscopy , we used strains that constitutively express the yGFP-ParBpMT1 fusion protein from a chromosome-inserted gene and carry a parSpMT1 site at one of three different loci ( Fig 1B ) . Binding of yGFP-ParBpMT1 to its cognate recognition site allows the visualization of each parS sequence as a fluorescent focus [25] . Three different strains were used , which carry ydeV::parSpMT1 between dif and terC , 10 kb from each , or yoaC::parSpMT1 about 300 kb away from dif on the left replichore , or ycdN::parSpMT1 about 500 kb away from dif on the right replichore [19] ( Fig 1B , S1 Table ) . All experiments were carried out in M9 glucose medium ( called M9 henceforth ) . Exponentially growing wild-type cells showed one or two foci . Cells with two foci depended on whether the parSpMT1 site was replicated and segregated and therefore decreased with distance of the site from the origin [25] ( S2 Table ) . In a recB mutant ~30% of cells showed no dif-proximal focus ( ydeV::parSpMT1 ) , and ~7–8% showed no dif-distal focus ( yoaC::parSpMT1 , ycdN::parSpMT1 ) [19] ( Table 1 , S2 Table ) . Time-lapse microscopy experiments allowed the real time visualization of focus loss in recB mutant cells: ~18% of the divisions produced a focus-less cell and a daughter cell with a focus [19] ( “% initial events” in Table 1; S1 Video ) and focus loss was heritable in ~75% of the cases [19] ( Fig 1D; “% transmitted” in Table 1; S1 Video; these inherited events are not counted in the 18% initial events ) . The molecular model depicted in Fig 2 explains these observations and has been tested in the present work . The model is as follows: a dsDNA end formed by breakage of one replication fork , at a dispersed and potentially random chromosomal location , results in a structure called a σ-replicating chromosome . This consists of an entire circular chromosome covalently linked to a linear partial chromosome arm by one intact replication fork ( Fig 2 , step A ) . The linear arm is repaired by homologous recombination in wild-type cells , but remains unrepaired in a recB mutant , in which σ-replicating chromosomes have been proposed to prevent cell growth [26 , 27] . We propose that in a recB mutant the linear and circular parts of this σ-replicating chromosome segregate to the two halves of the cell , while the intact replication fork progresses toward the terminus , and pauses at the ter sites ( Fig 2 , step B ) . However , the linear arm of the σ-replicating structure necessarily passes through mid-cell and is processed by FtsK , which precisely positions dif in the constricting septum ( Fig 2 , step C ) . The trapped DNA is broken during cell division , producing one daughter cell containing a linear , partial chromosome ( focus-less cell ) and the other one containing a σ-replicating chromosome with a shortened tail ( Fig 2 , step C ) . The DNA ends made during septum closure are located near dif and are slowly degraded by exonucleases . A second round of replication is initiated at oriC ( Fig 2 , step D ) and the tail of the σ-replicating chromosome is enlarged by the entire newly replicated sequence when the intact replication fork of the σ-replicating chromosome merges with the fork of the second replication round ( Fig 2 , step E ) . This new σ-replicating chromosome contains a complete linear chromosome attached to the terminus of a circular chromosome . The circular and linear parts segregate to daughter cells , and the region around the dif site , maintained in the path of the septum by the FtsK translocase , is cleaved again during cell division ( Fig 2 step F ) . This accounts for the efficient transmission of the phenomenon to the progeny in recBC mutants , as terminus breakage creates again a circular chromosome with a short tail and therefore the cycle of events can resume ( Fig 2 , step G ) . Importantly , we propose here that the initial DSB occurs at a replication fork , because a DSB elsewhere in the replicated region would leave both forks intact ( Fig 3A ) . Replication would produce a circular chromosome with no scar and a linear chromosome interrupted at a random sequence , which cannot account for our observations of heritable terminus DNA loss during division and DNA degradation centred on dif . In a recA mutant , dsDNA ends are acted upon by RecBCD and linear DNA is very efficiently degraded . We predicted that both the first linear tail created by fork breakage and the second , smaller linear tail created by division-induced breakage should be degraded by RecBCD in recA cells , reducing initial events and transmission of the phenomenon , respectively ( Fig 3B ) . We observed that the percentage of focus-less cells was three-fold lower in the recA mutant ( 9% ) than in the recB mutant ( ~32% , Table 1 , S2 Table ) . Time-lapse experiments showed that focus loss occurred in recA cells with some of the characteristics of recB cells: it occurred most frequently at the septum , always at the time of cell division and in one daughter cell only ( Fig 4A left panel; complete movie is shown in S2 Video ) . However , the proportion of initial events in the recA mutant was 7% of total divisions , nearly three-fold less than in the recB mutant ( 17 . 7% , Table 1 , Fig 4A left panel ) . Furthermore , transmission of the phenomenon to progeny was less efficient in the recA than in the recB mutant , since ( i ) ~37% of events were transmitted to progeny instead of ~75% in recB cells , and ( ii ) the number of successive generations undergoing terminus DNA loss was reduced compared to the recB mutant: for example , among the events that could be followed for more than 3 generations , 19 out of 27 continued focus loss in the recB mutant versus only 2 out of 12 in the recA mutant , the other ones mostly returning to normal growth . Note that the percentage of heritable events decreased from 13 . 3% of all divisions in the recB mutant ( 75% of 17 . 7% of the divisions ) to 2 . 6% in the recA mutant ( 37% of 7% of the divisions ) . Furthermore , 5–10% of divisions in the recA mutant were preceded by cell elongation , and some elongated cells produced focus-less cells ( S3 Video ) . This cell elongation could result from a partial degradation of the long DNA tail , which might prevent a correct DNA segregation and , in turn , block septum formation until the following replication round . In addition , in recA mutant cells we observed a similar percentage of cells lacking the dif-proximal ydeV::parSpMT1 locus and the yoaC::parSpMT1 locus further from dif ( ~9%; Table 1; S2 Table ) , and no terminus DNA loss could be detected by MFA ( [17]; Fig 4B left panel , S2 Fig ) . The recA mutants are known to lose entire nucleoids , and ~10% loss of terminus corresponds to such recA mutant cells without chromosomes [28] . We propose that DNA degradation by RecBCD extends further around DSBs , degrading the entire chromosome in the 9% focus-less recA cells and thus preventing detection of DNA loss by MFA . To test whether the lower efficiency of focus loss in the recA mutant results from the DNA degradation activity of RecBCD in the absence of RecA ( Fig 3B ) , we used a recA recB mutant . The percentage of focus-less cells was similar in recA recB and recB mutants for the dif proximal site ydeV::parSpMT1 and for the distal sites yoaC::parSpMT1 and ycdN::parSpMT1 ( Table 1 , S2 Table ) . Furthermore , time-lapse experiments showed that focus loss occurred at the time of division , in one cell only , and was transmitted to progeny ( Fig 4A right panel ) . The frequency of initial events ( 21% , Table 1 , Fig 4A right panel ) and the high rate of transmission to progeny ( 83 . 7% ) were similar in recA recB to the RecA+ recB strain . Furthermore , the MFA profiles were similar in recA recB and recB mutants ( Fig 1C , Fig 4B right panel , S1 and S2 Figs ) . This result shows that in a recA single mutant the frequency of terminus DNA loss is reduced due to the presence of RecBCD . In a recA recD mutant , DSBs are not repaired because homologous recombination is inactivated by the recA mutation , and dsDNA ends are slowly degraded because the recD mutation inactivates the Exo V activity of the RecBCD complex ( the RecB nuclease is not active in the RecBC complex lacking RecD , reviewed in [20–22] ) . recA recD mutant chromosomes were analysed by MFA ( Fig 4B middle panel , S2 Fig ) . Terminus chromosome degradation covered a much larger region and was less steep than in recB cells , but was still centred on dif , the region of GC skew inversion . We propose that terminus DSBs occur in recA recD cells and that the very broad zone of DNA degradation around the terminus is due to the processive and potent helicase activity of RecBC , which in the absence of RecD produces ssDNA from dsDNA ends efficiently , and thus facilitates the action of ssDNA exonucleases [23 , 29] . Microscopy experiments confirmed DNA loss of a larger terminus region in the recA recD compared to recB mutant cells , since 27 . 3% of them lacked the dif-proximal ydeV-parSpMT1 focus , 23% lacked the dif-distal yoaC-parSpMT1 focus and only 11% lacked the ycdN::parSpMT1 locus , the furthest from dif ( Table 1 , S2 Table ) . Time-lapse microscopy analysis of ydeV-parSpMT1 foci in recA recD cells showed that focus loss occurred as in the recB mutant: most often at the septum , always at the time of cell division and in one daughter cell only , and it was transmitted to the progeny ( Fig 4A middle panel , another example is shown in S4 Video ) . The frequency of initial events was 16 . 1% and these events were transmitted to progeny in 65% of the cases , without cell elongation ( Table 1 , Fig 4A middle panel ) . We conclude that terminus DNA loss is limited in recA cells by the Exo V activity of RecBCD . Recently , terminus DNA loss was also observed in a recA sbcB sbcD mutant [30] . In this mutant RecBCD is present but does not degrade DNA efficiently because DNA degradation requires dsDNA ends to be made blunt by SbcB and SbcCD exonucleases [31 , 32] . In agreement with a lack of DNA degradation by RecBCD in the recA sbcB sbcD mutant , microscopy results in the recA sbcB sbcD mutant were similar to the recA recB mutant ( Table 1 , Fig 5A ) , while inactivation of only sbcB or sbcCD in the recA mutant had a partial effect ( Table 1 ) . Finally , our model predicts that heritable terminus DNA loss should occur at a low efficiency in a recB sbcB sbcD mutant , which lacks RecBCD but where DSBs are repaired by the RecFOR pathway of recombination ( reviewed in Michel and Leach , 2012 ) . Actually in this mutant initial events were decreased nearly two-fold ( to around 10% , Table 1 ) and focus loss was less frequently transmitted to progeny ( 27 , 3% heritable events , Table 1 ) . These results are in agreement with the repair of dsDNA ends by the RecFOR recombination pathway , even though MFA analysis suggested that recB sbcB sbcD mutants initiate unscheduled replication in the terminus , and an unexplained high level of focus-less cells in growing cultures suggested that additional phenomena occur in the terminus region of the recB sbcB sbcD mutant ( [30]; Fig 5B; S3C and S3D Fig; Table 1 ) . Altogether , these results demonstrate that both homologous recombination and RecBCD-mediated DNA degradation should be inactivated to observe heritable terminus DNA loss , as predicted from our model ( Figs 2 and 3 ) . To date , only one particular replication fork breakage event is specific for recB and recA recD mutants , and those breaks result from RuvABC-catalysed resolution of a Holliday junction made by replication fork reversal [22 , 33] . Replication fork reversal is a reaction that involves the annealing of leading- and lagging-strand ends at a blocked fork , resulting in a dsDNA end adjacent to a Holliday junction [22 , 33] . In recBC and in recA recD mutants , the dsDNA end is neither recombined nor degraded , and resolution of the Holliday junction by RuvABC produces fork breakage [22 , 33] . Fork breakage by RuvABC in a recB mutant is a hallmark of replication fork reversal , and we tested a putative role of RuvABC in the production of the DSBs that lead to terminus DNA loss . The inactivation of ruvAB did not reduce the percentage of focus-less cells in recB ruvAB ( 37% , Table 1 ) or in recA recB ruvAB cells ( 38% , Table 1 ) . Focus loss in the recB ruvAB and recA recB ruvAB mutants occurred at the time and most often at the site of cell division , in one daughter cell , and was transmitted to progeny ( Fig 5C ) . Focus loss was quantified by time-lapse experiments in recA recB ruvAB cells , where only recombination-independent Holliday junctions can form . The frequency of initial events was unchanged by RuvAB inactivation ( about 21% ) , and transmission of focus loss to progeny was slightly lower than in the Ruv+ recA recB mutant but remained high ( 60% ) . Furthermore , DNA loss in the dif region was still observed by MFA in the recB ruvAB mutant ( Fig 5C , S3A and S3B Fig ) . We conclude that RuvAB is not required for terminus DNA loss in the recB mutant , which implies that replication fork reversal is not the main source of fork breakage in this mutant . The model predicts that the focus-less cell generated by the first cell division carries a truncated linear chromosome lacking all sequences between the original random DSB and the terminus , therefore potentially lacks essential genes . In contrast , focus-less cells generated in the following generations , which are delimited by two DSB events in the terminus region , contain a complete linear chromosome . This prediction could be tested by comparing the ability to propagate of these two types of focus-less cells . For this experiment we had to use a hipA hipB deleted strain since this toxin-antitoxin locus is adjacent to dif and its degradation in ydeV-parSpMT1 focus-less cells prevents proliferation [34 , 19 , 35] . In a hipA recB mutant 30% of the first focus-less cells did not divide while all the second focus-less cells divided ( <3% did not divide , Table 2 ) . This indicates that 30% of the first focus-less cells lacked some essential proteins that were expressed by the second focus-less cells . This is in agreement with the proposal that the first focus-less cells originally carry a truncated linear chromosome and thus differ from the subsequent focus-less cells that are born with a full linear chromosome . According to the model presented in Fig 2 , transmission of the phenomenon to progeny requires the production of a σ-replicating chromosome , in which a linear and a circular chromosome are attached by a replication fork in their terminus ( Fig 2 Step E ) . Therefore transmission should be prevented by using cells in which the naturally circular E . coli chromosome has been converted to a linear chromosome , artificially interrupted in the dif region . We used a strain that carries the terminus sequence tos of the linear phage N15 , 3 kb from dif on the right replichore , and that expresses the N15 telomerase TelN , which processes the tos sequence ( Fig 6A ) [36] . This strain propagates with a linear chromosome , interrupted 3 kb from dif [36] . As a control for these experiments , we used an isogenic strain with a circular chromosome , which carries the tos site but lacks the gene encoding the TelN protein ( S1 Table ) . Cells with linear chromosomes were studied by fluorescence microscopy , using ydeV::parSpMT1 or gusC::parSpMT1 markers on the left replichore ( 13 kb or 105 kb from the chromosome end , respectively ) , and yddW::parSpMT1 or pspE:: parS pMT1 markers on the right replichore , ( 19 kb or 217 kb from the chromosome end , respectively ) ( Fig 6A , Table 3 ) . It should be noted that the hipA hipB locus is adjacent to dif , therefore it will be degraded together with the ydeV::parSpMT1 or gusC::parSpMT1 markers , inhibiting growth of these focus-less cells . In contrast , because it is separated from the other chromosome arm by the tos site , it will remain intact in cells that lose the yddW::parSpMT1 or pspE:: parSpMT1 markers , allowing the multiplication of the cells that lose these loci . The proportion of cells lacking the end-proximal ydeV::parSpMT1 focus increased from 4 . 8% in the RecB+ strain to 20 . 7% in the recB mutant , while the proportion of cells lacking the end-distal gusC::parSpMT1 focus reached 10 . 5% in the recB mutant ( Table 3 ) . In contrast , the proportion of the cells devoid of the end-proximal yddW::parSpMT1 focus increased from 4 . 1% in RecB+ to nearly 60% in the recB mutant , while the proportion of cells lacking the end-distal pspE::parSpMT1 focus reached 56% in the recB mutant . As expected , in control isogenic strains with a circular chromosome , the proportion of cells lacking the dif-proximal loci ( ydeV::parSpMT1 or yddW::parSpMT1 ) was increased from about 1% in RecB+ to around 30% in the recB mutant , and was higher than the loss of a dif-distal locus ( pspE::parSpMT1 , 15% focus-less cells in a recB mutant , Table 3 ) . The difference between right and left replichores was specific for linear chromosomes , suggesting that the proportion of focus-less cells could be largely influenced by the position of the hipA hipB locus . To precisely quantify terminus DNA loss , ydeV::parSpMT1 and yddW::parSpMT1 foci were analysed in recB by time-lapse microscopy experiments . Results in the control recB mutant that carries tos but harbours a circular chromosome owing to the absence of TelN protein were similar to those observed in MG1655 , with a loss of ydeV::parSpMT1 or yddW::parSpMT1 foci occurring at the time of cell division , in one of the two daughter cells , and transmitted to progeny ( S5 Video ) . We counted 15 . 9% initial events for the yddW::parSpMT1 locus and more than 80% of the events were transmitted to progeny ( Table 4 ) . In cells with a linear chromosome , a similar percentage of initial events was observed with the terminus-proximal markers on the left and right replichores ( 14–17% ) but , importantly , the phenomenon was generally not transmitted to progeny , as only 11 to 19% of the events were heritable ( Table 4 , note that this level corresponds to the percentage of initial events and could therefore correspond to independent events occurring by chance after a first one ) . This result indicates that the transmission of focus loss to the progeny requires circularity of the chromosome . In addition , time-lapse experiments allowed us to observe that ydeV::parSpMT1 focus-less cells did not multiply , as expected from the concomitant degradation of the hipA hipB locus ( Fig 6B right panels , another example of ydeV::parSpMT1 focus loss from a linear chromosome is shown in S6 Video ) . In contrast , cells that lose the yddW::parSpMT1 locus on the right replichore could multiply for at least three generations ( Fig 6B left panels , complete movie is shown in S7 Video ) . Therefore , the high level of yddW::parSpMT1 and pspE::parSpMT1 focus-less cells can be simply explained by the propagation of focus-less cells carrying an intact hipA hipB locus . Genomes of the RecB+ and recB mutant linear strains were analysed by MFA ( Fig 6C , S4 Fig ) . A depletion of DNA sequences was observed on one chromosome arm , while nearly no DNA loss was observed on the chromosome arm carrying hipA hipB , possibly because the MFA technique is not sensitive enough to detect the weak level of recB-dependent DNA loss on this arm ( 16% , Table 3 ) . Although the MFA profile was therefore not informative regarding terminus DNA loss , it was in full agreement with the microscopy results . We conclude from these experiments that focus-less cells , which reflect a lack of terminus DNA , could be observed at either of the two ends of a recB mutant chromosome linearized at position 1585 kb . The phenomenon shares some common features with terminus DNA loss observed in circular chromosomes ( focus loss in one daughter cell , at the time of division ) , but , importantly , the capacity to lose terminus DNA in one daughter cell was not heritable . These results indicate that chromosome circularity , and thus DNA continuity of the terminus region is required for the heredity of the phenomenon , although it is not required for the formation of a first focus-less cell ( initial events ) . According to our model , transmission of terminus DNA loss to progeny depends on the persistence of the short DNA tail formed at each generation by septum closure until the arrival of the following replication round ( Fig 2 step D ) . In a recB mutant , DNA degradation is mediated by the action of helicases and exonucleases and is expected to be much slower than RecBCD-catalysed DNA degradation [37 , 38] . In a recA mutant , this short tail is the target of the potent RecBCD Exo V activity and should be efficiently degraded , which explains why only 37% of the initial events , instead of 80% in the recA recB mutant , were transmitted to progeny at least for one generation . The length of this tail is defined by the distance between the site of breakage ( the dif region ) and the position at the time of division of the intact replication fork that is slowed down by ter sites ( Fig 2 step C ) . Therefore , the duration of replication blockage at ter is expected to control heredity of terminus DNA loss in a recA mutant . We measured terminus DNA loss in a tus recA mutant , in which replication forks do not arrest at ter . tus inactivation increased the percentage of initial events from 7% to 11% , and increased the percentage of heritable events in the recA mutant from 37% to 64% , similar to the recA recD level ( Table 1 , S2 Table ) . This result shows that in a recA mutant replication arrest at ter limits terminus DNA loss and particularly the transmission of terminus DNA loss to the progeny . The model presented in Fig 2 implies that the two terminus sequences remain covalently attached . In wild-type cells , this covalent attachment cannot be directly visualized , as the two newly-synthesized terminus regions are anyway co-localized at the septum position when MatP is present . In contrast to wild-type cells , in a matP mutant terminus sequences readily separate after replication [6 , 7] . We used a matP mutant to test the attachment of the newly synthesized terminus sequences in the recB mutant . As previously described , all matP cells exhibited an early segregation of the ydeV::parSpMT1 loci to the ¼ and ¾ positions in the cell , owing to the lack of attachment of the terminus macrodomain to the septum ( [6] arrows in Fig 7A ) . MFA and microscopy experiments showed that terminus DNA loss occurred in matP recB as in the recB single mutant ( Table 1 , Fig 7B and 7C , S6 Fig ) . In time-lapse experiments , focus loss occurred at the septum , at the time of division , in one of the two daughter cells , and in a heritable manner ( Fig 7B ) . Measures of initial events and heredity showed that DNA loss was unaffected by matP inactivation ( 15 . 5% initial events , 86 . 4% heredity; Table 1 ) . However , although in most recB matP cells ydeV-parSpMT1 foci segregated prematurely to the ¼ and ¾ positions ( arrows in Fig 7B ) , in ~15–16% of cells foci remained together at the site of septum formation until division ( cells circled with a full white line in Fig 7B ) . Interestingly , focus loss occurred specifically in those cells where the two replicated ydeV::parSpMT1 foci remained nearby in the division plane , or , in other words , the lost focus was always one of the two foci that remained at the septum position after replication , in spite of the absence of MatP ( Fig 7B , focus-less cells are circled with a dashed white line ) . The specific loss of one of the two non-segregated loci in the matP recB mutant supports the idea that the two replicated chromosomes are linked at a position close to the ydeV locus ( Fig 2 ) . FtsK also contributes to the positioning of the chromosome terminus at the septum via binding of its C-terminal domain to KOPS sequences and chromosome translocation [5 , 39] . Nevertheless , in the matP ftsKΔCter recB mutant , which lacks the two functions known to position the terminus at the septum , ~40% focus-less cells were observed ( Table 1 ) . The MFA experiment showed an enlarged degraded region confirming that FtsK is not required for terminus DNA loss . Furthermore , DNA degradation was no longer centred on dif and spanned the entire fork trap , delimited by oppositely-oriented ter sites ( Fig 7D ) , confirming that FtsK translocation activity is responsible for the localization of the peak of DNA degradation around dif . Importantly , terminus DNA loss is observed in the absence of the functions that position the chromosome terminus at the septum , which supports the idea that the two terminus sequences are attached covalently . To confirm the post-replication attachment of two terminus regions in a MatP+ strain , we analysed chromosome segregation using cells where division is blocked by cephalexin , an inhibitor of the late septum protein FtsI [40] . As expected , cephalexin treatment caused the formation of elongated cells , and most wild-type cells showed regularly spaced ydeV::parSpMT1 foci , while 15–25% showed non-segregated foci ( Fig 8A and 8D ) . The proportion of cells with non-segregated ydeV::parSpMT1 foci was similar in all recombination proficient cells: between 11% and 25% non-segregated ydeV::parSpMT1 loci ( dif proximal ) and between <0 . 5% and 6 . 3% non-segregated yoaC::parSpMT1 loci ( 300 kb away from dif ) ( Fig 8D , see wild-type , recD , sbcB sbcD , recB sbcB sbcD and the circular chromosome control cell ) . Septum assembly is essential for dimer resolution owing to the role of the FtsK C-terminal domain in XerCD activation [41 , 42] , and about 15% of cells contain a chromosome dimer [43] . Consequently , the percentage of recombination proficient cells showing non-segregated dif-proximal loci can be accounted for by the lack of dimer resolution . In support of this idea , because dimers only form in circular chromosomes , nearly all cells harbouring a linear chromosome showed proper segregation of ydeV::parSpMT1 loci upon cephalexin treatment ( 0 . 7% non-segregated , Fig 8D ) . In a recA mutant , 10–17% cells showed non-segregated ydeV::parSpMT1 foci . Since chromosome dimers do not form in the absence of homologous recombination ( recA mutant ) , these 10–17% cells suffer a dimer-independent segregation defect ( Fig 8D ) . The marker further from dif ( yoaC::parSpMT1 ) was less affected and showed only 4% non-segregated cells . Inactivation of recD , recB or sbcB sbcD in the recA mutant increased the proportion of cells showing non-segregated ydeV::parSpMT1 foci to 17–29% , therefore , in recA mutants the lack of segregation of the ydeV::parSpMT1 foci after cephalexin treatment ( Fig 8D ) is correlated with the frequency of initial loss events ( Table 1 ) . This result supports the idea that terminus DNA loss occurs in cells in which the two termini remain covalently linked after replication . Interestingly , in the recB and recC mutants the percentage of cells presenting a segregation defect was as high as 22–44% for the dif proximal locus and increased to 9–14% for the dif-distal locus ( Fig 8B and 8D ) . Since dimer formation is half as frequent in the recB mutant as in wild-type cells [43] , the proportion of cells in which ydeV::parSpMT1 foci did not segregate independently of dimer formation could be as high as 15–30% , as in recA recD and recA sbcB sbcD mutants . This percentage correlates with the level of terminus DNA loss observed in dividing cells ( nearly 20% of initial events ) . Note that in cephalexin-treated cells focus segregation was similar to wild-type in recB sbcB sbcD ( Fig 8D ) , although this mutant showed an intermediate level of initial events between wild-type and recB mutant , ( 10% , Table 1 ) . To account for this observation , we propose that dsDNA end repair is slower when catalysed by RecFOR and RecA ( recB sbcB sbcD cells ) than when catalysed by RecBCD and RecA ( wild-type ) . Consequently , in recB sbcB sbcD cells that do not divide ( cephalexin treated ) , initial DSBs are repaired , although slowly , which allows segregation of sister chromosomes , while in dividing cells σ-replicating chromosomes are not always repaired prior to division and are sometimes cleaved . Finally , as expected from its high level of initial events , the recB mutant with a linear chromosome showed a high level of cells with an abnormal pattern of ydeV::parSpMT1 foci after cephalexin treatment ( 24–40% ) . However , in the linear chromosome recB mutant the abnormal cephalexin-induced filaments presented a deficit of ydeV::parSpMT1 foci ( Fig 8C ) instead of non-segregated foci , as observed in recB cells and in other mutants with a circular chromosome ( Fig 8B ) . As described below this is expected from the random breakage of one replication fork in a linear chromosome ( S6 Fig , see Discussion ) . In conclusion , a defect in segregation of the two replicated dif regions is observed in cells that lack homologous recombination and Exo V mediated DNA degradation both in the presence ( in a matP mutant ) and the absence ( in cephalexin-treated cells ) of cell division . This finding supports the idea that terminus DNA loss results from septum closure on non-separated chromosome termini .
Initial events rely on the persistence of a σ-replicating chromosome tail after fork breakage , which can lead to a focus-less cell only if the linear tail is neither degraded nor recombined , and segregates to the future daughter cell ( Fig 2B and 2C ) . The observation that initial events are three-fold less frequent in recA than in a recA recB mutant suggests that in two thirds of cases the potent Exo V activity of RecBCD ( variable but up to 800–900 bp per sec , [44 , 45] ) catches up with the progressing fork ( 500–600 bp per sec , [46 , 47] ) and fully degrades this first long tail , which prevents initial events ( Fig 3B pathway B ) . In a recA mutant the frequency of both initial and secondary events is increased by tus inactivation . The increase of initial events could be explained by two ways . Firstly , complete DNA degradation of the first tail is expected to be delayed by the progression of the active replication fork across the terminus . Secondly , in a subpopulation of cells , the progression of one of the two intact replication forks beyond the terminus , in the direction opposite to the main transcription direction , might increase replication fork blockage , as previously proposed , and in turn replication fork breakage and σ-replicating chromosome formation [48 , 49] . Increased heredity in the recA tus compared to the recA mutant supports the idea that heredity relies on the persistence of the truncated tail after terminus DNA breakage , hence on the length of this tail ( Fig 2D–2F ) . Growing cultures of recA mutants were reported to contain 5 to 10% anucleate cells ( see for example [32 , 50] ) , which corresponds to the percentage of focus-less cells observed in this work . Interestingly , in the recA mutant we did not observed loss of parSpMT1 foci at any time other than cell division . This observation suggests that most anucleate cells in MM cultures of a recA mutant result from the degradation of a linear chromosome formed by two successive DSBs: one at a random position during replication and one close to dif during septum closure ( Fig 3B pathway C ) . The formation of a focus-less cell is not transmitted to progeny when the chromosome is linearized by tos/TelN , in agreement with the idea that heredity requires circularity of the chromosome for the merging of the intact replication fork with the following replication round ( Fig 2 ) . A model showing the events expected to occur in the recB mutant harbouring a linear chromosome , according to the model shown in Fig 2 , is presented in S6 Fig . In the recB mutant with a linear chromosome , accidental breakage of one replication fork , while the other replication fork progresses to the chromosome end , leads to a linear head-to-head dimer composed of one entire chromosome and one truncated chromosome , linked by the telomerase TelN recognition site ( S6 Fig , 3 first steps ) . The two halves of this dimer segregate to the two future daughter cells , with the TelN recognition site at mid-cell . TelN action at this site produces an intact linear chromosome , which segregates to form the focus-carrying cell , and a truncated chromosome ( focus-less cell ) . Cells that harbour a truncated chromosome lacking the ydeV site do not multiply while those that lack the yddW locus multiply . Note that the reaction starts by fork breakage as on a circular chromosome , but the missing terminus , which fails to be copied by the broken replication fork , is not copied by the other fork ( and then degraded ) , since the chromosome is linear ( S6 Fig , progression of the intact fork to the end ) . Accordingly , in time-lapse experiments we did not observe a duplication of the ydeV::parSpMT1 or yddW::parSpMT1 focus prior to focus loss ( Fig 6 ) , and after cephalexin treatment abnormal elongated cells showed regions devoid of focus ( Fig 8 ) . Linearization in the terminus by TelN separates the intact from the truncated linear chromosomes after replication completion ( S6 Fig , last step ) , and no DSB occurs during cell division . Our results account for the long-standing observation of three types of cells in a recB mutant culture: non-dividing cells ( our focus-less cells ) , residually dividing cells ( the cells that produce a focus-less cell ) , and normally dividing cells [51] , Furthermore , the viability of recB cells is lower than that of recA mutant cells although , in addition to DSB repair , the latter also lack single-strand gap recombinational repair and induction of all DNA repair genes under the control of the SOS response [51–53] . It was proposed that the tail of a σ-replicating chromosome is a lethal form of damage in a recBC mutant , and that σ-replicating chromosomes are less deleterious in a recA mutant where the linear tail can be degraded by RecBCD [26 , 27 , 52] . Our study strongly supports the idea that σ-replicating chromosomes are the major cause of the low viability of the recB mutant but they do not simply cause lethality . Instead , one cell remains alive while most of the tail is segregated and cleaved off into a doomed daughter cell at each generation . Several kinds of replication impairments render RecBC , and sometimes also RecA , essential for viability [54 , 55] . The reverse assumption , that the viability defect of recBC and recA mutants directly reflects a correspondingly high level of spontaneous replication impairment , was often postulated . However , in contrast with this assumption , flow cytometry and MFA analyses showed that chromosome replication proceeds with a rate similar to wild-type in recB and recA mutants [19 , 56] . Replication fork blockage or breakage was not observed , although it should have been detected if it were responsible for the low viability of these mutants . Our model provides an explanation for this paradox . Our data allow us to determine for the first time that the level of spontaneous replication fork breakage is ~18% per cell per generation ( 9% per fork ) , which is too low to be directly detected in population studies . Finally , our findings raise future questions to be addressed: how does spontaneous replication fork breakage occur , and how are terminus DSBs catalysed ? We have previously shown that the periplasmic endonuclease Endo 1 is not involved [19] and no nuclease has been reported to be specifically associated with the septum .
Strains are described in S1 Table . Most strains were constructed by P1 transduction . New mutations were constructed as described in [57] , using DY330 [58] . Oligonucleotides used for constructions and mutation checking are shown in S3 Table . recA and recB mutations were checked by measuring UV sensitivity . recD mutations were checked by comparing the plating efficiencies of wild-type T4 and T4gpIIam phages ( the unprotected T4gpIIam only multiplies on recBC and recD mutants [59] ) . sbcCD mutations were checked by comparing the plating efficiencies of wild-type λ a λ carrying a long palindrome ( the λDRL154 phage that carries a long palindrome only multiplies on sbcCD mutants , [60] ) . In the course of this work , we fortuitously discovered that our microscopy strains are Phi80 lysogens . In contrast with the reported effects of Phi80 lysogeny in AB1157 [61 , 62] , Phi80 lysogens in MG1655 are only very weakly UV sensitive ( around 10% survival at 40 J/m2 ) , do not affect T4 or λ phages plating , and do not show a recD or sbcCD mutant phenotype . These background differences presumably result from the high divergence of the AB1157 and MG1655 genomes . All strains used for MFA are Phi80-free and experiments with Phi80-free recA and recB mutants confirmed that the cryptic phage did not affect the microscopy results ( S4 Table ) . MFA were performed and analysed as described in [19] , with cells grown in M9 glucose at 37°C . The MFA data have been submitted to the ArrayExpress repository . The access number for these data is E-MTAB-6122 . Microscopy experiments were performed and analysed as described in [19] . For snapshot analysis cells were grown in M9 glucose at 37°C . Time-lapse experiments were realized on M9 glucose at 30°C . | The Escherichia coli recBC mutant , deficient for DNA double-strand break ( DSB ) repair , shows a viability defect and a specific deficit in the level of chromosome terminus DNA sequences . We previously showed that this deficit results from heritable terminus DNA loss , owing to cell-division dependent DSBs in the chromosome terminus . Here , we used whole genome sequencing and microscopy to analyse the phenomenon . Our results allow us to conclude that in E . coli most spontaneous DSBs occur at replication forks , and that such breaks occur in 18% of cells at each generation . In a recBC mutant the linear chromosome arm made by replication fork breakage is neither repaired nor degraded . Thus it remains attached to the circular chromosome part , which triggers a DSB in the chromosome terminus during cell division in a heritable reaction . In wild-type cells , broken replication forks are repaired and these terminus DSBs do not occur . Our study reconciles the idea that replication fork impairment is a major source of chromosome breakage with the observation that most DSBs in a recBC mutant occur in the chromosome terminus during cell-division and reveals the links between these two phenomena . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"chromosome",
"structure",
"and",
"function",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"dna",
"replication",
"forms",
"of",
"dna",
"circular",
"dna",
"dna",
"homologous",
"recombination",
"chromosome",
"biology",
"genetic",
"loci",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"recombination",
"chromosomes"
] | 2018 | Broken replication forks trigger heritable DNA breaks in the terminus of a circular chromosome |
Diphthamide is a highly modified histidine residue in eukaryal translation elongation factor 2 ( eEF2 ) that is the target for irreversible ADP ribosylation by diphtheria toxin ( DT ) . In Saccharomyces cerevisiae , the initial steps of diphthamide biosynthesis are well characterized and require the DPH1-DPH5 genes . However , the last pathway step—amidation of the intermediate diphthine to diphthamide—is ill-defined . Here we mine the genetic interaction landscapes of DPH1-DPH5 to identify a candidate gene for the elusive amidase ( YLR143w/DPH6 ) and confirm involvement of a second gene ( YBR246w/DPH7 ) in the amidation step . Like dph1-dph5 , dph6 and dph7 mutants maintain eEF2 forms that evade inhibition by DT and sordarin , a diphthamide-dependent antifungal . Moreover , mass spectrometry shows that dph6 and dph7 mutants specifically accumulate diphthine-modified eEF2 , demonstrating failure to complete the final amidation step . Consistent with an expected requirement for ATP in diphthine amidation , Dph6 contains an essential adenine nucleotide hydrolase domain and binds to eEF2 . Dph6 is therefore a candidate for the elusive amidase , while Dph7 apparently couples diphthine synthase ( Dph5 ) to diphthine amidation . The latter conclusion is based on our observation that dph7 mutants show drastically upregulated interaction between Dph5 and eEF2 , indicating that their association is kept in check by Dph7 . Physiologically , completion of diphthamide synthesis is required for optimal translational accuracy and cell growth , as indicated by shared traits among the dph mutants including increased ribosomal −1 frameshifting and altered responses to translation inhibitors . Through identification of Dph6 and Dph7 as components required for the amidation step of the diphthamide pathway , our work paves the way for a detailed mechanistic understanding of diphthamide formation .
Regulation of biological processes by posttranslational modification can involve the function , distribution and interaction capabilities of the modified protein [1]–[3] . Though most modification pathways such as phosphorylation and ubiquitin conjugation target many different proteins , some exceptional ones uniquely target just a single polypeptide [4] . One prominent example is diphthamide formation on eukaryal translation elongation factor 2 ( eEF2 ) [5] . Strikingly , this modification is pathobiologically important because it is hijacked for eEF2 inhibition by sordarin fungicides and by diphtheria toxin ( DT ) produced by pathovarieties of Corynebacterium diphtheriae that cause the severe human disease syndrome diphtheria [6]–[8] . Both agents efficiently block protein synthesis by inactivating the essential function of the modified translation factor , via stalling the diphthamide modified form of eEF2 on ribosomes and irreversible ADP ribosylation of eEF2 , respectively [9]–[12] . Diphthamide itself is a highly modified histidine residue on eEF2 – 2-[3-carboxyamido-3- ( trimethylamino ) -propyl]-histidine – which is conserved from yeast ( H699 ) to man ( H715 ) ( Figure 1 ) [5] , [8] , [13] . Intriguingly , it is absent from the bacterial eEF2 analog , EF-G , thus conferring immunity on the DT producer . Among the archaea and eukarya , diphthamide formation involves a conserved biosynthetic pathway , which has been extensively dissected in Saccharomyces cerevisiae via isolation of mutant strains that confer resistance to growth inhibition by DT and sordarin . This has led to the identification of the diphthamide synthesis genes DPH1-DPH5 [7] , [12] , [14]–[16] ( Figure 1 ) . The first step in diphthamide synthesis involves transfer of a 3-amino-3-carboxypropyl ( ACP ) radical from S-adenosyl-methionine ( SAM ) to the histidine imidazole ring , generating the ACP modified intermediate of eEF2 [17]–[19] . ACP radical transfer requires the proteins Dph1-Dph4 [16] , where Dph1 and Dph2 are paralogous iron-sulfur cluster containing partner proteins that copurify and interact with Dph3 , potentially as part of a multimeric complex [6] , [20]–[22] . Dph3 and Dph4 are thought to chaperone Dph1-Dph2 by maintaining their iron-sulfur clusters in redox states required for proper ACP radical generation . In line with this , Dph3 and Dph4 have electron carrier activities [23] , [24] , while Dph3 ( also known as Kti11 [25] ) additionally partners with Elongator subunit Elp3 [6] , [20] , an iron-sulfur cluster and radical SAM enzyme with roles in protein and tRNA modifications [26]–[28] . Formation of diphthine , the second pathway intermediate ( Figure 1 ) , requires trimethylation of the amino group in ACP and is catalyzed in yeast by diphthine synthase Dph5 , using SAM as methyl donor [29]–[31] . Intriguingly , reconstitution of archaeal Dph5 activity has shown that the trimethylamino group formed in diphthine is prone to elimination in vitro [32] . Finally , the carboxyl group of diphthine is amidated by an elusive ATP dependent diphthamide synthetase ( Figure 1 ) . Once fully modified , diphthamide can be efficiently targeted by NAD+-dependent ADP ribosylase toxins including DT , Pseudomonas exotoxin A [33] and Vibrio cholix toxin [34] . However , the intermediate diphthine is also a very weak substrate for inhibitory ADP ribosylation [29] , [31] . Together with data showing that growth inhibition by sordarin also depends on DPH1-DPH5 [6] , [7] , translation factor eEF2 constitutes an ‘Achilles heel’ for yeast , study of which has provided important insight into the pathobiological relevance of posttranslational protein modification [35] . Physiologically , the function of the diphthamide modification is enigmatic . Yeast mutants unable to synthesize diphthamide confer elevated frequency of ribosomal frameshifting [6] , [36] but are viable and grow normally [14] , although substitution of the modified histidine in eEF2 by other amino acids confers growth defects in some instances [37] . However , loss of diphthamide synthesis leads to delayed development and is embryonic lethal in homozygous DPH3 knockout mice [38]–[40] . Together with the association of mammalian DPH1 with tumorigenesis [16] , [38] as well as neuronal and embryonic development , this indicates that diphthamide modification plays an important biological role . Whether or not this implies structural or regulatory roles for diphthamide modified eEF2 remains to be seen , but the latter notion is intriguing given the possibility of endogenous cellular ADP ribosylases that target eEF2 [4] . Interestingly , no DT resistant yeast mutants have been identified to date that affect the final amidation step in the pathway , probably because diphthine is targetable , albeit inefficiently , by ADP ribosylation [29] , [31] . Thus amidase-deficient mutants may display DT sensitivity in vivo and thereby escape identification in screens for DT resistant yeast mutants . Indication that additional proteins are involved in diphthamide biosynthesis has come from recent work on WDR85 and its potential yeast ortholog YBR246w [41] , [42] , while our preliminary investigation of the yeast DPH1 genetic interaction network [13] implicated both YBR246w and YLR143w as novel proteins potentially involved in the diphthamide pathway . Here we further exploit yeast genome-wide genetic interaction and chemical genomics databases [43] , [44] to demonstrate that YLR143w ( DPH6 ) and YBR246W ( DPH7 ) cluster tightly with all known members of the diphthamide gene network . We find that dph6 and dph7 mutants phenocopy sordarin and DT traits typical of the bona fide dph1-dph5 mutants , which are defective in the first two steps of diphthamide synthesis . Importantly , we show that DPH6 and DPH7 deletions block the final amidation step of the diphthamide pathway , cause diphthine modified forms of eEF2 to accumulate and consequently abolish ADP ribose acceptor activity upon DT treatment . Thus conversion of diphthine to diphthamide depends on Dph6 and Dph7 .
To identify factors involved in the terminal amidation step of the diphthamide modification pathway ( Figure 1 ) , we took advantage of synthetic genetic array ( SGA ) screens , which previously enabled systematic mapping of genetic interactions among yeast deletion collections using high-density arrays of double mutants [45] , [46] . SGA analysis provides the set of genetic interactions for a given gene – the genetic interaction profile – and thereby the phenotypic signatures indicative of functions of both known genes and unassigned ORFs [47] . For example , genes with similar interaction profiles are often functionally related in shared biochemical pathways and/or protein complexes [48] , [49] . Consistent with this , SGA analysis revealed that the diphthamide gene network members have highly correlated interaction profiles and tightly cluster in the global genetic interaction landscape from yeast [45] . Since our preliminary examination of the yeast genetic interaction landscape placed two uncharacterized yeast ORFs , YLR143w and YBR246w , within the diphthamide gene network [13] , we next examined this network in more detail by mining the SGA DRYGIN database for quantitative S . cerevisiae genetic interactions [44] , [50] . We compared DPH1 , DPH2 , DPH4 , DPH5 , YLR143w and YBR246w gene interactions with every array ORF represented in the SGA network and deposited at DRYGIN , ranking the similarity between all possible pairwise profiles according to their Pearson correlation coefficient ( PCC; see Table S1 for full details ) . As expected , the other known DPH genes scored significantly highly among the correlation profiles generated for each specific DPH query gene , consistently being ranked among the top ten genetic interactors ( Figure 2A ) . Strikingly , YLR143w and YBR246w were among the top interactors of DPH1 , DPH2 , DPH4 and DPH5 , while the most correlated interactors for YLR143w and YBR246w included each other and several bone fide DPH genes ( Figure 2A ) . Such highly correlated interaction patterns suggest that YLR143w and YBR246w are both functionally interrelated and qualify as candidate ORFs of the pathway for eEF2 modification by diphthamide . In line with this notion , the two eEF2 encoding gene copies , EFT1 and EFT2 , also ranked among the top ten interactors of DPH1 , DPH2 and DPH5 ( Figure 2A ) . For independent validation of these correlations , we searched the FitDB yeast fitness database [51] , which contains genome-scale phenotypic profiles for diploid yeast deletion collections in response to more than 1100 different growth conditions [43] , [52] . Here , scoring gene interaction profiles by homozygous co-sensitivity revealed that among the top loci to phenocluster with YBR246w are DPH2 , DPH4 and DPH5 , while top interactors of YLR143w include DPH4 , DPH5 , YBR246w and DPH2 ( Figure 2B ) . A similar pattern of interaction is shown by DPH5 ( Figure 2B ) , DPH2 and DPH4 ( data not shown ) . Based on correlated interaction profiles , FitDB ascribes GO terms enriched for processes concerning peptidyl-diphthamide biosynthesis from peptidyl-histidine to YLR143w and YBR246w with p-values of 2×10−3 and 9×10−4 respectively ( Figure 2C ) . Collectively , the FitDB and DRYGIN profiles thus provide robust phenotypic signatures suggesting novel roles in the diphthamide pathway for YBR246w and YLR143w , which are tightly clustered within the DPH gene network ( Figure 2C ) . This notion is consistent with a recent report that YBR246w and its mammalian homolog , WDR85 , have a diphthamide related function [41] , [42] . Since YLR143w is as yet unassigned in the Saccharomyces genome database ( SGD ) , based on the evidence below that YLR143w and YBR246w are indeed diphthamide synthesis genes we have named them DPH6 ( YLR143w ) and DPH7 ( YBR246w ) . To verify the predicted roles for DPH6 and DPH7 in the diphthamide pathway , we next examined strains deleted for these ORFs for phenotypes specifically linked to defects in diphthamide formation on eEF2 , namely sordarin resistance and response to DT [6] , [7] . Sordarin traps eEF2 on the 80S ribosome [53] , blocking mRNA translation elongation and yeast cell growth [54] in a fashion that depends on diphthamide synthesis [6] , [7] . As a result , diphthamide mutants dph1-dph5 efficiently protect against sordarin inhibition [6] , [7] . Like dph1-dph5 , dph6 and dph7 mutants showed robust resistance towards sordarin at 10 µg/ml , a concentration inhibitory to the wild-type ( Figure 3A ) . This resistance was comparable to that shown by eEF2 substitution mutants eft2H699I and eft2H699N ( Figure 3A ) , which lack the His699 acceptor residue for diphthamide modification [37] . Thus DPH6 and DPH7 are novel sordarin effectors , a feature they share with the diphthamide synthesis genes DPH1-DPH5 [6] , [7] . Diphthamide modification plays a key effector role for inhibitory ADP ribosylation of eEF2 by DT , hence dph1-dph5 mutants in both yeast and mammalian cells confer resistance towards DT [14] , [16] . We therefore compared DT-dependent ADP ribosylation of eEF2 in vitro between wild-type cells and dph1 , dph5 , dph6 and dph7 mutants . While the translation factor from wild-type cells was efficiently modified by the toxic ADP ribosylase ( Figure 3B ) , eEF2 extracted from dph1 , dph5 , dph6 and dph7 mutants could not be ADP ribosylated by exogenously added DT under the conditions used ( Figure 3B ) . Such lack of ADP ribose acceptor activity in vitro strongly suggests defects in the diphthamide pathway and that DPH6 and DPH7 encode novel functions required for diphthamide formation . To further address this experimentally , we assayed the response of dph6 and dph7 mutants to intracellular expression of the ADP ribosylase domain of DT ( DTA ) using GALS , a truncated variant of the GAL1 promoter [55] . When DTA expression was induced by 0 . 1% galactose , dph6 and dph7 mutants were indeed found to show some protection against DTA in contrast to wild-type cells ( Figure 3C ) , consistent with defects in diphthamide formation . However , at a higher level of expression on 2% galactose , they showed wild-type like sensitivity to DTA whereas dph1 and dph5 mutants remained fully resistant ( Figure 3C ) . This suggests that eEF2 forms from dph6 or dph7 mutants , although not substrates in vitro ( Figure 3B ) , can nonetheless be ADP ribosylated in vivo if DTA is expressed at a sufficiently high level [30] . While our work was in progress , eEF2 from a ybr246w/dph7 mutant was shown to be a very weak substrate for ADP ribosylation when treated with 10 mM DT [42] , a 500-fold increase in concentration over that used in our in vitro ADP ribosylation assays ( Figure 3B ) . Thus eEF2 from the dph6 or dph7 mutants is resistant to sordarin and shows a vastly reduced ability to be ADP-ribosylated by DT , strongly suggesting that the diphthamide pathway is defective . Since the intermediate diphthine can serve as a sub-optimal substrate for ADP ribosylation using excess levels of DT or upon overexpressing its toxic ADP ribosylase domain from inside cells [29] , [31] , the properties of eEF2 from dph6 and dph7 mutants are consistent with a defect in the final step of the pathway that converts diphthine to diphthamide . Our analysis is therefore entirely consistent with the above database predictions and indicates DPH6 and DPH7 constitute novel candidate loci for diphthamide biosynthesis . Given the above evidence , we next examined whether eEF2 prepared from cells deleted for either DPH6 or DPH7 carried any modification on His699 , the eEF2 residue that is modified to generate diphthamide . eEF2 preparations made from wild-type and gene deletion strains expressing His6-tagged eEF2 were digested with trypsin and examined by mass spectrometry . The His6-tagged form was chosen as the source of eEF2 since expression rescued the inviability of an eft1 eft2 double mutant lacking eEF2 function , and it is thus considered to be biologically active [56] . Strains lacking either DPH1 , in which the first step of diphthamide biosynthesis is blocked , or lacking DPH5 ( encoding diphthine synthase ) , were used respectively as controls for complete lack of modification and presence of ACP , the first intermediate in the diphthamide pathway [14] , [16] , [30] , [32] . All strains expressed similar levels of His6-tagged eEF2 ( data not shown ) . The modified histidine in eEF2 ( His699 ) is located in the tryptic peptide 686-VNILDVTLHADAIHR-700 and , as expected , unmodified versions of this peptide were readily detected in eEF2 prepared from the dph1 mutant ( Figure S1C ) . Unmodified peptide was also found in eEF2 prepared from dph5 , dph6 and dph7 deletion strains as well as from wild-type cells ( Figures S1 and S2 ) . Thus even in wild-type cells not all of the eEF2 is modified by diphthamide . In addition to the unmodified peptide , we readily detected diphthamide-modified peptide in eEF2 prepared from the wild-type strain ( Figure 4A ) , but failed to detect this in any of the mutants . Instead , ACP-modified peptide was found in eEF2 prepared from the dph5 mutant ( Figure 4B ) , as expected given its known role in generating diphthine [32] from the ACP intermediate in the pathway . In contrast , eEF2 from the dph7 mutant generated spectra consistent with the presence of diphthine on His699 , in which the m/z values for both the parent ions and the daughter ions in the MS/MS spectra were higher in a manner consistent with the 0 . 984 Da extra mass associated with presence of a carboxyl group in diphthine rather than the amide group in diphthamide ( Figure 4C ) . Thus each of the doubly-charged daughter ions in Figure 4C is larger by an m/z of ∼0 . 5 than the corresponding ion in the wild-type spectrum ( Figure 4A ) . Furthermore , the quite different elution times of the diphthine-modified and diphthamide-modified peptide that are evident from the extracted ion chromatograms ( Figure S3 ) are consistent with differently modified forms of eEF2 . As noted in previous studies [32] , [33] , [36] , some of the ions in our MS/MS spectra had undergone neutral loss of the trimethylamino group during MS/MS , as indicated by loss of 59 . 110 mass units . Two types of spectra corresponding to the peptide with modified His699 were seen when eEF2 from the dph6 mutant was analyzed . In some spectra ( Figure 4D ) , the parent ion m/z and MS/MS data indicated the presence of diphthine as in the dph7 mutant , with some daughter ions again showing neutral loss of the trimethylamino group during MS/MS as noted above . However , we also detected peptide forms in which elimination of the trimethylamino group had occurred prior to analysis , as indicated by the lower parent ion m/z ( Figure 4E ) and an MS/MS spectrum in which all assignable peaks corresponded to ions lacking the trimethylamino group . Such trimethylamino elimination prior to mass spectrometry was observed previously when diphthine-modified Pyrococcus horikoshii EF2 was generated in an in vitro reaction [32] , indicating that this modification might be unstable . However , we failed to detect any pre-mass spectrometry loss of the trimethylamino group when eEF2 from the dph7 mutant was analyzed . Thus while eEF2 from both mutants carries diphthine , the modification appears to be more labile in the dph6 mutant and may be protected from trimethylamino elimination by the absence of Dph7 . Figure S3 shows extracted ion chromatograms for ions with m/z values corresponding to the His699 containing peptide modified with diphthamide , diphthine or with ACP , indicating that the ACP modified peptide was only present in the dph5 mutant , the diphthine modified peptide was only present in dph6 and dph7 mutants , and diphthamide-modified peptide was only seen in wild-type cells . Our mass spectrometry analysis therefore shows that in yeast strains lacking either DPH6 or DPH7 , modification of His699 progresses only as far as diphthine . Thus both loci indeed qualify as novel diphthamide synthesis genes with likely roles in conversion of diphthine to diphthamide . Although Dph6 and Dph7 appear to function within the same step of the diphthamide synthesis pathway , using co-immune precipitation they were not found to interact either with one another or with Dph2 and Dph5 , players involved in the two earlier pathway steps ( Figure S4; Figure S5 and data not shown ) . However , in support of our evidence that Dph6 is a diphthamide biosynthetic factor , we observed by co-immune precipitation that Dph6-HA bound to a fraction of ( His ) 6-tagged eEF2 ( Figure 5A ) . Intriguingly , this interaction was independent of Dph7 ( Figure 5A ) , suggesting Dph7 may not mediate interaction between Dph6 and the translation factor . Dph7 is also unlikely to play an indirect role through regulation of DPH6 gene expression because Dph6 protein levels were unaltered in the DPH7 deletion strain ( Figure 5A ) . Inactivation of WDR85 , the mammalian homolog of Dph7 , was recently shown to dramatically enhance association of diphthine synthase Dph5 with eEF2 [41] . We therefore examined whether Dph7 impacts on the interaction between Dph5 and eEF2 in budding yeast . We found that a much higher level of affinity tagged eEF2 could be co-immune precipitated with HA-tagged Dph5 from extracts of the dph7 mutant in comparison to wild-type extracts ( Figure 5B ) . A smaller increase was also seen with the dph6 mutant ( Figure 5B ) . This strongly suggests a conserved role for Dph7/WDR85 as a regulator of the Dph5•eEF2 interaction . Remarkably , we also found similarly enhanced binding of Dph5 to eEF2 in the dph1 mutant , which has a defect in the first step of the diphthamide pathway and therefore lacks the ACP modification that is the immediate substrate of diphthine synthase ( Figure 5B ) . Strikingly , DPH5 overproduction from a galactose-inducible promoter was found to be highly detrimental to cells deleted for DPH7 and to all mutants blocked at the first step of the pathway , but had little effect on the dph6 mutant and no effect on wild-type or dph5 cells ( Figure 6A ) . Intriguingly , this cytotoxicity goes hand in hand with the enhanced Dph5•eEF2 interaction profiles we observed in dph1 , dph6 and dph7 cells under conditions of wild-type DPH5 copy number and normal Dph5 expression levels ( Figure 5B ) . Taken together , our results suggest that binding of Dph5 to incompletely modified eEF2 may be inhibitory to the function of the translation factor . Our data also indicate that both unmodified eEF2 , and diphthine-modified eEF2 in the absence of Dph7 , show strongly enhanced binding to Dph5 . Furthermore , since we failed to detect association between Dph5 and Dph6 despite demonstrating interaction of each with eEF2 , it is likely that Dph5 and Dph6 do not bind concurrently to eEF2 and that their binding may therefore be mutually exclusive . Although the precise biological function of diphthamide is unclear , its location at the tip of the eEF2 anticodon mimicry domain IV predicts a potentially important role in translation . Consistent with this , structure-function studies have shown that domain IV is sufficiently proximal for interaction with tRNA in the decoding P-site of the ribosome [57] and alterations of invariant tip residues , including H699 substitutions that cannot be diphthamide modified , confer biologically significant traits including thermosensitive growth defects [37] , [58] . Nonetheless , when compared to their wild-type parental strain , we found no significant changes in the growth performance of dph1-dph7 mutants in either liquid or on solid media and at standard cultivation temperatures of 30°C ( Figure S6 ) . Even increasing the cultivation temperatures to 39°C had no discernable effect on dph cell growth except for the dph3/kti11 mutant ( Figure S6 ) , which is known to be thermosensitive due to additional functions unrelated to diphthamide [6] . However , intrigued by previous reports that diphthamide defects can induce ribosomal frame-shifts [6] , [36] , we next studied whether DPH6 and DPH7 deletions affect the accuracy of eEF2 in the translation process ( Figure 6B ) . Using lacZ-based reporters to monitor programmed +1 and −1 frameshift signals derived from Ty elements [36] , [59] , dph1-dph7 mutants failed to induce significant ribosomal +1 frameshifts ( data not shown ) . However , dph1 , dph2 , dph3 , dph5 and dph6 mutants significantly enhanced lacZ expression dependent on a −1 frameshift , with dph6 and dph3 cells scoring as the top −1 frameshifters followed by lower but statistically significant effects in dph1 , dph2 and dph5 mutants ( Figure 6B ) . This confirms increased −1 frameshifting in dph2 and dph5 mutants seen previously [36] and demonstrates an even larger defect in dph3 and dph6 strains . Ribosomal −1 frameshift induction by dph7 and dph4 , though slightly increased in relation to wild-type controls , was considered statistically insignificant ( Figure 6B ) . The −1 frameshifting phenotype shared between dph6 and bona fide dph mutants is consistent with a role for diphthamide in promoting translational accuracy of eEF2 . In line with a role for diphthamide in the fine tuning of translation elongation , growth assays performed under thermal and/or chemical stress conditions showed that certain dph mutants including DPH6 and DPH7 deletion strains displayed altered responses to translation elongation indicator drugs such as hygromycin , anisomycin or paromomycin ( Figure S7 ) . In conclusion , our data indicate that diphthamide mutant strains such as dph6 increase ribosomal errors typical of −1 translational frameshifts and that the diphthamide modification function of Dph6 , which is required for completion of diphthamide synthesis , is likely to assist eEF2 in reading frame maintenance during translation .
Dph6 contains three conserved domains consistent with it functioning as an enzyme ( Figure S8 ) . The amino-terminal 225 residues constitute an Alpha_ANH_like_IV domain ( cd1994 in the NCBI Conserved Domain Database [62] , also known as DUF71 ) , a member of the adenine nucleotide alpha hydrolase superfamily that is predicted to bind ATP . Many DUF71 proteins from archaea to mammals contain the highly conserved motif –E215GG ( D/E ) XE220– ( Dph6 numbering ) , which has been proposed to be involved in substrate binding and catalysis and which is replaced by –ENGE ( F/Y ) H– in a group of related DUF71 proteins implicated in biotin synthesis [60] . Based on this we generated a dph6 allele encoding two substitutions in this region ( G216N , E220A ) and tested its functionality by monitoring complementation of sordarin resistance in a yeast dph6 knockout strain . Figure 7 clearly shows that this small change completely inactivates the function of Dph6 , demonstrating that the Alpha_ANH_like_IV domain is critical for the conversion of diphthine to diphthamide . The C-terminal portion of Dph6 contains two domains related to the YjgF-YER057c-UK114 protein family ( eu_AANH_C1: cd06155 and eu_AANH_C2: cd06166 ) that may promote homotrimerisation and formation of an inter-subunit cleft that has been proposed to bind small molecule ligands [63]–[65] . Several key residues in human UK114 required for homotrimerisation and ligand binding [66] are present in Dph6 ( Figure S8 ) including arg-107 , which in E . coli TdcF forms a bidentate salt bridge with the carboxylic acid group of bound ligands [63] . Deletion of residues 335–415 encompassing much of the YjgF-YER057c-UK114 region abolished the function of Dph6 as monitored by sordarin resistance ( Figure 7 ) , while truncation of Dph6 at the first of the two conserved domains by insertion of a myc tag also eliminated Dph6 function ( Figure 7 ) despite detectable expression of the truncated polypeptide ( data not shown ) , indicating that the YjgF-YER057c-UK114 domains are also important for Dph6 function and that the Alpha_ANH_like_IV domain is nonfunctional on its own . Since Salmonella enterica YjgF has an enamine/imine deaminase activity that is conserved in human UK114 [67] it is possible that the YjgF-YER057c-UK114 domains in Dph6 are used to generate ammonia for diphthamide formation . Taken together , these properties suggest a direct , ATP-dependent role for Dph6 in diphthine amidation proceeding via an adenylated intermediate and with ammonia acting as the source of the amide group . Such a direct role has now been demonstrated by Su et al . , who have used an in vitro assay system to show that Dph6 has diphthamide synthetase activity [61] . Although proteins showing Dph6-like domain organization are readily identified in fungi , plants , amphibians and insects ( Figure S8 ) , they are largely absent from archaeal and mammalian proteomes . However , mammals and archaea have separate proteins showing strong similarity to either the adenine nucleotide alpha hydrolase domain or to the YjgF-YER057c-UK114 related regions ( Figure S8 and data not shown ) , suggesting Dph6 functionality may be split between different polypeptides in these cases . It is therefore surprising that expression of the human DPH6 ortholog in a yeast dph6 mutant can restore diphthamide biosynthesis [61] despite lacking the YjgF-YER057c-UK114 domains that are essential in the yeast protein ( Figure 7; [61] ) . Thus while the core function of the enzyme must therefore reside in the Alpha_ANH_like_IV domain , it will be interesting to determine the role of the YjgF-YER057c-UK114 domains in Dph6 from lower eukaryotes . Dph7 has four well-defined WD40 repeats ( Figure S9 ) and its predicted structure consists exclusively of β-sheet elements [41] , [68] . Although its human homolog WDR85 has been implicated in the first step of diphthamide biosynthesis [41] , our work and that of Su et al . [42] show that the pathway can proceed as far as diphthine in the absence of DPH7 and that the block is therefore in conversion of diphthine to diphthamide . Furthermore , this block cannot be bypassed simply by introducing DPH6 on a multicopy plasmid to increase the level of diphthamide synthetase ( data not shown ) . How then might Dph7 contribute to diphthine amidation ? Its domain structure suggests it could act as an adaptor molecule for diphthine amidation [42] , but this notion is at odds with our failure to detect interaction between Dph7 and Dph6 ( see above ) . Our intriguing finding that eEF2 binds much more Dph5 in the absence of Dph7 suggests an alternative role , namely that Dph7 is needed to displace Dph5 from diphthine-modified eEF2 to allow the amidation reaction to occur . Similar findings in mammalian cells upon inactivation of WDR85 support this notion [41] . Together with our data showing that viability of dph1-dph4 and dph7 cells is extremely sensitive to excess Dph5 in comparison to wild-type or dph6 cells , it appears that binding of Dph5 to eEF2 is inhibitory to the function of the translation factor and negatively interferes with cell growth unless eEF2 carries the completed diphthamide modification . Perhaps in addition to catalyzing methylation of ACP-modified eEF2 , Dph5 binds to newly-synthesised eEF2 to exclude it from functioning in translation until the diphthine amidation step takes place ( Figure 8 ) . Consistent with this proposal is our observation that the level of Dph5 associated with eEF2 in the dph1 mutant , in which modification of His699 cannot be initiated , is drastically increased and virtually indistinguishable from the enhanced Dph5-eEF2 interaction seen when Dph7 is absent . Dph7 may be needed to displace Dph5 once diphthine has been generated so that Dph6 can carry out the diphthine to diphthamide conversion ( Figure 8 ) , a notion consistent with the sensitivity of the dph7 mutant to DPH5 overexpression . In contrast , the dph6 mutant may tolerate Dph5 overexpression because Dph7 is present to displace it . Two other seemingly unrelated functions have been previously proposed for DPH7 . Firstly , it emerged from a genetic screen as a potential negative regulator of RNA polymerase I ( Rrt2 ) , although no other DPH genes were similarly identified [69] . Secondly , DPH7 has been implicated in retromer mediated endosomal recycling and named ERE1 [68] . The connection between endosomal recycling and diphthamide biosynthesis is currently unclear and it remains to be determined whether Dph7 is multifunctional or if these other roles are linked to its eEF2 modification function . Diphthamide on eEF2 is the target for bacterial ADP-ribosylase toxins and also affects toxicity of sordarin and ricin , a ribosome inhibiting protein toxin from plants [70] . Although this emphasizes its pathological relevance , the physiological significance of diphthamide remains enigmatic and elusive . Nonetheless , the evolutionary conservation of the diphthamide pathway among eukaryotes and the embryonic lethality of mice that cannot synthesize diphthamide [38] strongly suggest that it is important in translation related processes . In support of this notion , evidence presented here and by others shows that diphthamide mutants cause increased translational frameshifting , a defect also observed in mammalian cells [6] , [36] , [71] . Diphthamide modification may have particular importance in multicellular organisms or when cells are stressed [4] . Mutation of mammalian diphthamide synthesis genes affects cell proliferation and development: inactivation of DPH3/KTI11 is associated with tRNA modification defects and neurodegeneration and mutations in DPH1/OVCA1 revealed a tumor suppressor role for this diphthamide synthesis gene in ovarian cancer [27] , [38]–[40] , [72] . Regardless of its physiological functions , our data indicate that the diphthamide pathway is more complex than originally anticipated and comprises , in addition to Dph1-Dph5 , two further components , Dph6 and Dph7 , which operate in the terminal amidation step ( Figure 8 ) . While it is now clear that Dph6 is diphthamide synthetase [61] , in the future it will be important to understand why the archaeal and mammalian orthologs can dispense with the otherwise conserved YjgF-YER057c-UK114 domains and to define the precise role of Dph7 . It will also be crucial to explore the potential role of diphthine synthase ( Dph5 ) as a potential regulator of the entire pathway and the reasons for apparent lability of diphthine in the dph6 mutant that is suggested by our data ( Figure 8 ) .
Yeast strains used in this study are listed in Table S2 and plasmids in Table S3 . Cultures were grown in complete ( YPD ) or minimal ( SD ) media [73] at 30°C unless otherwise stated . For phenotypic assays , YPD was supplemented with 10 µg/ml sordarin sodium salt from Sordaria araneosa ( Sigma-Aldrich ) . Yeast transformations with plasmid DNAs were performed following the lithium acetate protocol [74] . Diphtheria toxin ( DT ) growth assays in vivo involved expression of the toxin's cytotoxic ADP ribosylase fragment ( DTA ) from vector pSU8 ( p415-GALS-DTA ) , essentially as previously described for dph1-dph5 mutants [6] . pSU8 was made by cloning the BamHI fragment encoding DTA from pLMY101 [30] into plasmid p415-GALS , a single-copy E . coli-yeast shuttle vector with a truncated GAL promoter [55] , which allows for conditional DTA induction on galactose-containing media . [55] . The translational frameshift reporter assay essentially involved previously published protocols together with the described lacZ reporter plasmids pJD204 . 0 ( wild-type control ) , pJD204 . −1 ( −1 frame ) and pJD204 . +1 ( +1 frame ) [36] , [59]; the pJD204 plasmid series was kindly provided by T . Kinzy ( UMDNJ , USA ) . The relative values for +1 and −1 frameshifting were statistically analyzed using one-way ANOVA followed by Dunnett's multiple comparison post test and was performed with Graphpad Prism 5 . 0 software essentially as previously described [75] . Details of all primers used in numerous PCR dependent genomic manipulation experiments can be found in Table S4 . Gene deletions were performed using in vivo PCR-based one step-gene disruption protocols in combination with marker plasmids YDpKl-L , YDpKl-U or YDpSp-H [76] and knockout primers ( Table S4 ) including those previously described [6] , [25] , [77] . Gene deletions were confirmed via diagnostic PCR on genomic DNA preparations using target ORF-specific primer pairs ( Table S4 ) as well as sordarin response assays . C-terminal tagging of DPH1 , DPH2 , DPH5 , DPH6/YLR143w and DPH7/YBR246 was performed according to previously published in vivo PCR-based epitope tagging protocols [78] using appropriate S3/S2 primer pairs ( Table S4 ) . Tagged genes were confirmed by Western blot detection with anti-HA or anti-c-Myc antibodies ( Santa Cruz Biotechnology A-14 and F7 , respectively ) . Detection of HA- or c-Myc-tagged Dph1 , Dph2 , Dph5 , Dph6 and Dph7 as well as Dph3 and Elp2 in co-immune precipitation ( Co-IP ) assays were performed as previously described [6] , [77] , [79] . pSU6 was generated by insertion into YCplac111 [80] of a genomic PCR fragment including DPH6 together with 829 bp of upstream and 59 bp of downstream sequence flanked by EcoRI and BamHI sites incorporated using PCR primers ( Table S4 ) . The insert was verified by sequencing and shown to complement a dph6 knockout strain . pSU7 was made by cloning the DPH6 insert from pSU6 into YEplac181 [80] . To generate a G216N E220A dph6 mutant , pSU6 was digested with AgeI and BsmBI and the small DPH6 fragment replaced by an identical synthetic fragment ( Integrated DNA Technologies ) carrying the G216N E220A mutations , generating independent clones pMS61 and pMS62 . The replaced region was verified by DNA sequencing . pMS67 and pMS68 were generated from pSU6 by replacing the BsmBI-SalI fragment carrying the C-terminal region of DPH6 and downstream sequence with a synthetic BsmBI-SalI fragment in which codons 335–685 were replaced by sequence encoding the linker and triple myc tag from pYM23 [81] . To generate pMS72 , the smaller NheI-SpeI fragment of pSU7 was excised and the large fragment ligated to generate an in-frame fusion that removed DPH6 codons 347–471 , checking the resulting fusion by DNA sequencing . Yeast cell extracts were prepared as described previously [15] . ADP ribosylation reactions were performed at 37°C for 1 hour in a volume of 40 µl ADP ribosylation buffer ( 20 mM Tris-HCl , pH 7 . 4 , 1 mM EDTA , 50 mM DTT ) containing 50 µg of yeast extract , 50 ng fully-nicked DT and 10 µM 6-biotin-17-NAD ( Trevigen ) . Samples were then mixed with SDS sample buffer , boiled for 5 min and run on 4–25% SDS-PAGE gradient gels ( Invitrogen ) . The proteins were transferred to nitrocellulose membranes and Western blotting was performed using streptavidin-IR conjugate ( Rockland Immunochemicals , Gilbertsville , PA ) and scanned on an Odyssey Infrared Imager ( LICOR Biosciences , Lincoln , NE ) . BY4741 wild-type yeast cells as well as dph1 , dph5 , ylr142w/dph6 and ybr246w/dph7 mutants thereof carrying an eft2 null-allele were transformed with plasmid pTKB612 ( a kind gift from A . R . Merrill , University of Guelph , Ontario , Canada ) , which expresses a ( His ) 6-tagged form of translation elongation factor eEF2 ( Table S3 ) that is fully functional and able to complement an eft1 eft2 double mutant [56] . In order to express and purify ( His ) 6-tagged eEF2 for MS/MS analysis , 750 ml of yeast culture were grown in YPD to an OD600 2 . 0 and harvested by centrifugation . The pellet was resuspended in 3 ml B60 buffer ( 50 mM HEPES-KOH pH 7 . 3 , 60 mM KOAc , 5 mM Mg ( OAc ) 2 , 0 . 1% Triton X100 , 10% ( v/v ) glycerol , 1 mM NaF , 20 mM glycerophosphate , complete protease inhibitor [Roche] ) without DTT and cells were lysed in a bead beater . The lysate was centrifuged twice at 13 , 500 rpm for 30 min . and the protein concentration measured with a NanoDrop spectrophotometer . Five mg total protein was applied to 2 mg anti- ( His ) 6-tag Dynabeads ( Invitrogen , #101-03D ) and purified according to manufacturer's instructions . The identity of purified eEF2 fraction was confirmed by SDS-PAGE and Western blot analysis using an anti- ( His ) 6 antibody ( Abcam , #ab18184 ) . Crude yeast eEF2 preparations from wild-type and dph mutants strains were separated by SDS-PAGE using 4–12% Bis-Tris precast gels ( Invitrogen , Carlsbad , USA ) and the area of the gel containing eEF2 was excised after staining with Instant Blue Coomassie ( Expedeon , Cambridge , UK ) . In-gel digests were performed using trypsin , subsequent to reduction and alkylation with dithiothreitol and iodoacetamide , with the resulting peptides cleaned over C18 columns . Peptides were then analyzed via HPLC-MS/MS using a Dionex U300 HPLC ( Dionex California ) with a 15 cm PepMap C18 column coupled to a Thermo Orbitrap Velos mass spectrometer ( Thermo Fisher Scientific , Bremen , Germany ) . The peptides were eluted from the C18 column at 300 nL/min over 120 min using a linear 5–90% ( v/v ) acetonitrile gradient . The Orbitrap Velos was operated in positive ion mode , with an ion source voltage of 1 . 2 kV and capillary temperature 200°C , using a lock mass of 445 . 120024 . The initial survey scan was performed at 60000 resolution , FTMS scanning from 335–1800 Da . The top 15 most intense ions were selected for MS/MS sequencing , using collision-induced dissociation ( CID; MS/MS charge state 1+ rejected , >2+ accepted ) . Protein identification was performed using MaxQuant 1 . 2 . 2 . 5 [82] against a proteome database generated from the Saccharomyces Genome database [83] . Manual annotation of the modified peptide spectra corresponding to the modified EF2 peptide and generation of extracted ion chromatograms were done using the Thermo Xcalibur software for spectra visualization . | Diphthamide is an unusual modified amino acid found uniquely in a single protein , eEF2 , which is required for cells to synthesize new proteins . The name refers to its target function for eEF2 inactivation by diphtheria toxin , the disease-inducing agent produced by the pathogen Corynebacterium diphtheriae . Why cells require eEF2 to contain diphthamide is unclear , although mice unable to make it fail to complete embryogenesis . Cells generate diphthamide by modifying a specific histidine residue in eEF2 using a three-step biosynthetic pathway , the first two steps of which are well defined . However , the enzyme ( s ) involved in the final amidation step are unknown . Here we integrate genomic and molecular approaches to identify a candidate for the elusive amidase ( Dph6 ) and confirm involvement of a second protein ( Dph7 ) in the amidation step , showing that failure to synthesize diphthamide affects the accuracy of protein synthesis . In contrast to Dph6 , however , Dph7 may be regulatory . Our data strongly suggest that it promotes dissociation of eEF2 from diphthine synthase ( Dph5 ) , which carries out the second step of diphthamide synthesis , and that Dph5 has a novel role as an eEF2 inhibitor when diphthamide synthesis is incomplete . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"systems",
"biology",
"biochemistry",
"computer",
"science",
"model",
"organisms",
"molecular",
"cell",
"biology",
"genetics",
"chemistry",
"biology",
"genomics",
"microbiology",
"genetics",
"and",
"genomics",
"toxicology"
] | 2013 | The Amidation Step of Diphthamide Biosynthesis in Yeast Requires DPH6, a Gene Identified through Mining the DPH1-DPH5 Interaction Network |
Barrier epithelia that are persistently exposed to microbes have evolved potent immune tools to eliminate such pathogens . If mechanisms that control Drosophila systemic responses are well-characterized , the epithelial immune responses remain poorly understood . Here , we performed a genetic dissection of the cascades activated during the immune response of the Drosophila airway epithelium i . e . trachea . We present evidence that bacteria induced-antimicrobial peptide ( AMP ) production in the trachea is controlled by two signalling cascades . AMP gene transcription is activated by the inducible IMD pathway that acts non-cell autonomously in trachea . This IMD-dependent AMP activation is antagonized by a constitutively active signalling module involving the receptor Toll-8/Tollo , the ligand Spätzle2/DNT1 and Ect-4 , the Drosophila ortholog of the human Sterile alpha and HEAT/ARMadillo motif ( SARM ) . Our data show that , in addition to Toll-1 whose function is essential during the systemic immune response , Drosophila relies on another Toll family member to control the immune response in the respiratory epithelium .
Although the innate immune system is a primitive host defense mechanism , it involves a sophisticated repertoire of humoral and cellular responses both acting systemically and locally [1] . In recent years , the Drosophila model organism has proven to be an invaluable system in dissecting in great details the genetics and cellular mechanisms regulating the innate immunity [2]–[3] . One fundamental mechanism common to humans and Drosophila immunity involves signaling by receptors of the Toll family . Upon microbial infection , human TLRs activate the synthesis of cytokines and other regulatory molecules that stimulate the adaptive immune system [4] . In Drosophila , Toll signalling leads to the activation of the systemic immune response , which is characterized by the synthesis of AMPs by the fat body cells [5] . Upon secretion into the circulating blood , these AMPs provide systemic protection against fungi and bacteria . In the mammalian innate immune response , bacteria are directly sensed by TLRs [6] . In contrast , these microorganisms are detected in Drosophila by another class of proteins , the Peptidoglycan Recognition Proteins ( PGRPs ) , also present in the human proteome [7]–[9] . Recognition of Lys-type peptidoglycan ( PGN ) ( mainly found in Gram-positive bacteria cell wall ) by the circulating PGRP-SA protein triggers a protease cascade involving successively Mod-Sp , Grass , and the Spätzle-Processing-Enzyme ( SPE ) [8] , [10]–[11] . Upon activation , SPE becomes competent to transform the zymogen pro-Spätzle into an active ligand for the Toll receptor , inducing its dimerization and intracellular signalling [12] . Production of AMPs after infection by Gram-negative bacteria , is however largely independent of the Toll pathway but rather relies on another NF-κB signalling cascade named IMD [13] . Sensing of Gram-negative bacteria upstream of the IMD pathway takes place at the plasma membrane , through PGN recognition by the transmembrane PGRP-LC receptor . Binding of DAP-type PGN ( present in Gram-negative bacterial cell wall ) to PGRP-LC induces its dimerization , which , in turn , triggers IMD-dependent intracellular events enabling the nuclear translocation of the NF-κB transcription factor Relish [14]–[17] . Although penetration of infectious microbes into the body cavity , and consequently , activation of a systemic immune response , is a rare event , interactions between microbes and epithelia take place constantly throughout the life of all metazoan . This implies that these barrier epithelia must be armed with efficient systems for microbial detection and elimination . However , these epithelia that act as interfaces with the external environment , share some characteristics that could be seen as detrimental for the needs of an effective immune system . Indeed , they usually have large surface areas and consist of thin structural layers , thus representing ideal entry points for pathogens . In this respect , the airway epithelium is unique among all epithelia , since it has a very delicate structure and is constantly exposed to a plethora of airborne pathogens . This could explain the occurrence of a great variety of inflammatory lung diseases , including asthma , chronic obstructive pulmonary disease or cystic fibrosis [18]–[19] . Elucidation of the primary steps that lead to chronic inflammation of the mammalian lung is obstructed by the complexity of inflammatory responses in this organ . Animals with a much simpler organization , such as the fruit fly , might help us clarify the basic architecture of this epithelial immune response , thereby helping to unravel the mechanisms that lead to chronic inflammation of the airways . Although of much simpler organization , the fly's airway system shows striking similarities with the mammalian lung regarding both its architecture and its physiology [20]–[25] . In this report , we present a detailed description of the mechanisms that regulate AMP production in the Drosophila respiratory epithelium . We show that , in contrast to systemic fat body immune response , the IMD pathway can be activated non-cell autonomously in the tracheal network . We present evidence that IMD pathway activation is tightly regulated in the cells of the respiratory epithelium . We demonstrate that the molecular mechanisms underlying IMD down-regulation following infection , are different from those previously reported in the gut and in the fat body , and rely on a dialog between two antagonist pathways . The production of AMPs in the trachea is positively regulated by the IMD pathway , which is counterbalanced by a negative regulation from a signalling cassette , whose upstream receptor is a member of the Toll family , Toll-8/Tollo . Our data suggest that the Spz2/DNT1 cytokine is a putative Tollo ligand in this process , and that Ect4 , the Drosophila ortholog of the human Toll/Interleukin-1 Receptor ( TIR ) domain-containing protein SARM mediates Tollo signalling during tracheal immune response .
The tracheal system is a relatively simple model system that has provided an important insight into the biology of branching morphology [22] . It is a tubular structure covered by a lumenal cuticular lining that forms a physical barrier against dehydration and invading microorganisms [20] . This network consists of a monolayer epithelium made up of two dorsal trunks ( DT ) connected to several visceral branches ( VB ) bringing oxygen to internal tissues . Upon infection by the entomopathogenic bacterium Erwinia carotovora carotovora ( Ecc ) , these epithelial cells produce a cocktail of AMPs including Drosomycin , Drosocin and Attacin [26]–[27] . In order to perform a detailed spatiotemporal analysis of this epithelial response , we used a Drosomycin-GFP reporter transgene ( Drs-GFP ) and monitored the response of this tissue after infection . When reared on conventional medium , a few larvae showed sporadic tracheal Drs-GFP expression mainly in VB , and rarely in the posterior part of the DT , namely the posterior spiracles ( PS ) ( Figure 1A and 1B ) . This basal Drs-GFP tracheal expression was qualitatively and quantitatively similar in larvae reared conventionally or under axenic conditions ( data not shown ) . Upon infection with Ecc , Drs-GFP activation followed a somewhat stereotypical pattern . Responding larvae were categorized into three classes according to their Drs-GFP expression pattern ( Figure 1A and 1C ) , namely , larvae expressing GFP in PS and in posterior VB only ( class I ) , in PS and all VB ( class II ) and in VB and DT ( class III ) . Kinetic experiments showed that GFP signal was first detected in PS and then spread into VB ( Figure 1A , C ) . The reporter was only later activated in the DT and , unexpectedly , first in the anterior half and then in the entire trunk ( Figure 1A ) . Although we appreciate that the expression patterns of reporter transgenes can slightly deviate from those of the actual AMP , the Drs-GFP expression patterns observed cannot easily be attributed to a progressive diffusion of bacterial elicitors ( such as PGN ) from the spiracles into the tracheal network , but rather speak for more complicated mechanisms in tracheal AMP activation . Next , we compared the mode of activation of AMP following forced immune pathway activation in the trachea and in another immune tissue , the fat body . Although activation of Drosomycin transcription is mainly controlled by the Toll pathway in the fat body ( but can be activated by ectopic IMD pathway triggering , see later ) , it is strictly IMD-dependent in the trachea [5] , [26] , [28] . Indeed , over-activation of the IMD ( UAS-PGRP-LCa , UAS-IMD ) , but not of the Toll pathway ( UAS-spz act , Tl3 ) , is sufficient to induce tracheal expression of Drs-GFP in non-infected larvae ( Figure S1A and S1B ) . Concomitantly , loss-of-function mutations in IMD pathway components ( Relish , PGRP-LC , IMD ) prevent Drs-GFP tracheal activation in infected larvae , whereas Toll signalling mutants such as spz or Dif do show a wild-type tracheal response upon infection . To analyze whether all tracheal cells were competent to trigger AMP production upon IMD pathway activation , we induced UAS-IMD expressing clones in tracheal cells , using fat body clones as controls . Overexpression of IMD led to a strictly cell-autonomous and fully penetrant activation of both Drs-GFP and Dipt-Cherry in the fat body ( Figure 2B and 2C ) . In the trachea , although most IMD-expressing cells showed Drs-GFP expression , a fraction did not ( Figure 2A ) . In addition , Drs-GFP activation was not always associated with the expression of the UAS-IMD transgene ( Figure 2A ) , suggesting that IMD pathway activation in trachea is not strictly cell autonomous . These results were confirmed by using a UAS-PGRP-LCa transgene that activated Drs-GFP both autonomously and non-autonomously in trachea cells ( Figure 2E ) but strictly cell-autonomously in the fat body ( Figure 2F ) . We next addressed whether PGRP-LC function was required cell-autonomously for IMD pathway activation in the trachea upon infection . Analysis of MARCM loss-of-function clones for PGRP-LC indicates that tracheal cells mutant for PGRP-LC were totally impaired in their ability to trigger Drs-GFP expression , following Ecc infection ( Figure 2D and H ) . These results indicate that , although PGRP-LC is essential in tracheal cells for IMD pathway triggering , IMD pathway activation in one tracheal cell can spread to neighboring cells . This contrasts with the strictly cell-autonomous IMD-dependent immune response observed in fat body cells . In order to get a further insight into the mechanisms that control AMP induction in trachea , we looked for putative immune genes expressed in this tissue . A recent report identified the repertoire of all immune genes expressed in the trachea [29] . One of the striking data of this study , [confirmed by FlyAtlas ( http://flyatlas . org/ ) ] was that , in addition to Toll itself , two other Toll family members , 18-Wheeler ( Toll-2 ) and Tollo ( Toll-8 ) , are strongly expressed in trachea . 18-wheeler being implicated in developmental processes with indirect impacts the immune response [30] , we focused our study on the putative function of the Tollo transmembrane protein in the tracheal immune response . Using Lac-Z reporter lines ( data not shown ) and q-RT-PCR ( Figure 3A ) , we confirmed that Tollo mRNA is highly enriched in the tracheal epithelium , and expressed at lower levels in other tissues . To investigate the subcellular localization of the Tollo protein , we genetically associated a UAS-Tollo::Myc construct with the trachea-specific Breathless-Gal4 driver ( Btl-Gal4 ) . Anti-Myc antibody staining suggested that Tollo was localized apically at the cell membrane facing the airway lumen ( Figure 3B ) . Double staining experiments showed that Tollo::Myc partially co-localized with the apical marker Cadherin::GFP , but was mutually exclusive with Viking::GFP , a basal membrane-associated protein ( Figure 3B ) . These results indicate that Tollo is a protein enriched in the tracheal epithelium with an apical subcellular localization . The rather restricted expression pattern of Tollo mRNA in the trachea and the apical subcellular distribution of Tollo protein prompted us to investigate its putative function in the immune response . For that purpose , we used two previously characterized hypomorphic alleles ( Tollo145 and TolloR5A ) together with a complete loss-of-function allele ( TolloC5 ) that we generated by P-element mediated homologous recombination [31]–[32] ( Figure S2A ) . All Tollo mutants were viable with no obvious developmental defects and gave rise to phenotypically normal pharate adults indicating that Tollo has no essential role in Drosophila development . We tested the ability of these Tollo mutant larvae to mount an immune response . In the absence of infection , approximately 5% of wild-type larvae showed Drs-GFP expression in VB and/or PS ( Figure S3 ) . Similar Figures were obtained with Tollo mutants suggesting that Tollo is not required to set up the basal level of AMP production in the absence of infection ( Figure S3 ) . After bacterial infection , however , the immune response was much stronger in Tollo mutants than in wild-type sibling larvae ( Figure 4A and 4B and Figure S3 ) . While we could identify the three previously described classes of Drs-GFP positive larvae in both control and Tollo mutants , the relative proportion of these was significantly different between genotypes ( Figure S3 ) . The percentage of larvae showing no GFP expression was reduced to 5–10% in Tollo mutant ( compared to 40% in controls ) , whereas Class III larvae , which represented 15% of controls , reached up to 50% in the Tollo mutant larvae ( Figure S3 ) . Similar results were obtained with three independent Tollo alleles and in trans-heterozygous allelic combination demonstrating that this phenotype was , indeed , due to Tollo inactivation and not to other mutations on the chromosome ( data not shown ) . The effects were not only qualitative but also quantitative . In most infected Tollo mutant larvae , Drs-GFP expression was intense , whereas it was rarely the case in controls ( Figure 4A ) . q-RT-PCR experiments indicate that Drosomycin , Drosocin and Attacin mRNA levels were respectively increased by 6 , 7 and 2 . 6 fold after infection in Tollo mutants compared to wild-type trachea ( Figure 4B ) . To ensure that this effect was indeed a consequence of Tollo inactivation in the tracheal network itself , we combined the Btl-Gal4 driver with a UAS-TolloIR RNA interference construct . As shown in Figure 4A and Figure S3 , larvae , in which Tollo was eliminated specifically in trachea , also showed increased Drs-GFP expression in this tissue after infection . In addition to regulating Drosomycin expression in the trachea , the IMD pathway also controls Diptericin transcription locally in the gut and systemically in the fat body [13] , [26] . To test whether Tollo acts as a general negative regulator of IMD-dependent mechanisms in other immune tissues , we analyzed the effects of inactivating Tollo on IMD pathway activation in the gut and the fat body . Using the Dipt-Cherry reporter construct for the larval stage ( Figure 4D ) and q-RT-PCR for both larval and adult stages ( Figure 4E and 4F ) , we showed that Tollo was not implicated in IMD negative regulation in either tissue . This was the case for both immune responses induced by septic injury or by oral ingestion . In addition , we showed that Tollo mutants were unaffected in their ability to activate the Toll pathway during a Gram-positive bacteria-mediated systemic immune response ( Figure 4F ) . Altogether these results demonstrate that the Tollo receptor is specifically required to dampen IMD pathway-dependent responses in the tracheal network after infection . Since AMPs are induced upon cellular stress , we tested whether Drosomycin expression in Tollo mutants was a secondary consequence of a possible implication of Tollo in tracheal formation . The following reasons led us to believe that it was not the case . 1 ) Tollo mutant embryos gave rise to viable adults , suggesting that Tollo mutant trachea are fully functional in larvae and adults . 2 ) Tracheal cell morphology of Tollo and control larvae appeared similar when observed under transmission electron microscopy ( Figure 5A ) . 3 ) No constitutive AMP transcription was detected in non-infected Tollo mutant larvae ( Figure 4B and Figure S3 ) . We then wondered whether Drosomycin over-activation could be linked to the presence of higher levels of potential immune elicitors in Tollo mutant and RNAi trachea . This could be due to the presence of higher bacterial load in the trachea . However , as shown in Figure 5B and Figure S4B , the number of Ecc-GFP in Tollo mutant trachea was identical in Tollo mutants and in controls . Alternatively , Tollo mutant trachea could be more permeable to contaminated external fluid . To test this hypothesis , wild type and Tollo mutant larvae were incubated in the presence of a fluorescent dye , bromophenol blue . External fluid penetration inside the trachea lumen was not different in wild-type and Tollo mutant larvae ( Figure 5C and 5D ) . This indicates that over-activation of Drosomycin in both Tollo mutant and RNAi trachea cannot be attributed to an increase in fluid penetration , and therefore putative immune elicitor load within the tracheal lumen . Altogether , these results demonstrate that the infection-dependent Drosomycin over-activation observed in Tollo mutant , is not secondary to defective trachea but rather suggests a direct implication of the Tollo protein in the regulation of IMD-dependent Drosomycin expression . We then tried to identify the intra- and extra-cellular components that may mediate Tollo signalling in trachea . The Drosophila Toll-1 receptor and vertebrate TLR functions have all been shown to be mediated by TIR domain-containing proteins , respectively , the Drosophila dMyd88 and the mammalian Myd88 , TRIF , SARM , TRAM and MAL [33]–[35] . However , the Drosophila proteome contains two TIR domain proteins , Ect4/SARM the Drosophila ortholog of vertebrate SARM and dMyd88 , the latter mediating the Toll signalling during dorso-ventral axis specification and immune response [36]–[38] . In order to test whether Tollo acts through a TIR domain-containing protein , we analyzed the tracheal immune response of dMyd88 and Ect4/SARM mutants . We showed that larvae carrying Ect4/SARM mutations ( Figure S2B ) display a strong over-activation of Drosomycin expression ( visualized with the Drs-GFP reporter transgene and quantified by q-RT-PCR ) upon infection , phenotype that was not observed in dMyd88 mutant larvae ( Figure 6A and 6B and Figure S5A ) . This suggests that Ect4/SARM is the bone fide TIR domain adaptor transducing Tollo signalling in the tracheal immune response . Whereas TLRs function as Pattern Recognition Receptors by directly binding to microbial motifs , previous work has shown than the Drosophila Toll-1 receptor is activated during both embryonic dorso-ventral axis specification and immune response by its ligand Spätzle [39]–[40] . Since spz mutant larvae did not present higher activation of Drs-GFP in the respiratory tract after infection , we believe that Spz is not a functional ligand for Tollo ( Figure 6C ) . Since Spz-like genes are present in the fly genome , we screened them for a Tollo-like phenotype . We observed that removing Spz2 ( known as DNT1 ) function in trachea phenocopies Tollo mutant as far as Drosomycin over-activation is concerned ( Figure 6C and 6D and Figure S5B ) . This suggests that DNT1 could be the , or one of the ligand ( s ) , responsible for Tollo activity in the tracheal immune response . Similarly to Tollo , q-RT-PCR data indicate that Ect4 and DNT1 are specifically acting in the tracheal epithelium and do not contribute to IMD pathway regulation in the gut and in the fat body ( Figure 6B and 6D ) . Taking into account the above results , it appears that the function of Tollo is specifically to down-regulate the IMD pathway in the tracheal cells following infection . In order to genetically place Tollo with respect to known IMD pathway components , we performed epistatic experiments . We showed that the Tollo mutant phenotype requires functional PGRP-LC receptor and intracytoplasmic adaptor IMD , since double mutant Tollo- , PGRP-LC- and imd -; Tollo- trachea did not show any signs of Drs-GFP activation after infection ( Figure 4A ) . This epistatic relationship was confirmed by q-RT-PCR on Drosomycin mRNA ( Figure 4C ) . In genetic terms , Tollo is hypostatic , or acts in parallel to PGRP-LC and imd . Consistently , Relish nuclear translocation monitored with an anti-Relish antibody was higher in Tollo RNAi-infected tracheal cells than in controls ( Figure S6 ) . These results suggest that Tollo is not directly involved in IMD pathway activation per se but that , in its absence , IMD pathway activation is more efficient upon infection .
Epithelial responses are first and foremost local responses to prevent the epithelium from unnecessary immune reactions . Since the recognition steps in Drosophila respiratory epithelia involve the transmembrane receptor PGRP-LC and occur within the extracellular space , it is expected that molecular mechanisms must be at work to prevent constitutive or excessive immune response in this tissue , particularly essential for animal growth and viability . In this report , we present data demonstrating that the transmembrane receptor Tollo is part of a signalling network , whose function is to specifically down-regulate AMP production in the trachea . We show that Tollo antagonizes IMD pathway activation in the respiratory epithelium , and that DNT1/Spz2 and Ect4/SARM are putative Tollo ligand and transducer , respectively , in this process . Our data demonstrate that , in addition to the family founder Toll-1 , another member of the Leucine-Rich-Repeats family of Toll proteins , is regulating the Drosophila innate immune response . Although it has been abundantly documented that every single mammalian TLR has an immune function [4] , the putative implication of Toll family members , other than Toll-1 itself , in the Drosophila immune response has been a subject of controversy [41] . Data showing that Drosophila Toll-9 over-expression was sufficient to induce AMPs expression in vivo has prompted the idea that Toll-9 could maintain significant levels of anti-microbial molecules , thus providing basal protection against microbes [42] . However , our recent analysis of a complete Toll-9 loss-of-function allele has shown that this receptor is neither implicated in basal anti-microbial response nor required to mount an immune response to bacterial infection [43] . The present data are also fully consistent with a recent report showing that Toll-6 , Toll-7 and Toll-8 are not implicated in systemic AMP production in flies [44] , and demonstrate that a Toll family member , Tollo , is a negative regulator of local airway epithelial immune response upon bacterial infection . In contrast to Toll-1 , whose activation is inducible in the fat body , Tollo pathway activation seems to be constitutive in the trachea . Despite these differences , both receptors use a member of the Spz family as ligand . Interestingly , sequence similarities , intron's size and conservation of key structural residues , indicate that Spz2/DNT1 is phylogenetically the closest family member to the Toll ligand Spz [45] . Furthermore , both Spz and Spz2/DNT1 have been shown to have neurotrophic functions in flies [46] . It would be of great interest to test whether Tollo also mediates Spz2 function in the nervous system . Both during embryonic development and immune response , Spz is activated by proteolytic cleavage [10] , [47]-[48] . This step depends upon the Easter protease that is implicated in D/V axis specification and on SPE for Toll pathway activation by microbes . Since Spz orthologs are also produced as longer precursors , they are likely to be activated by proteolysis . The fact that Tollo and Spz2 loss-of-function phenotypes correspond to excessive AMP production , suggests that in wild-type conditions , the Tollo pathway is constitutively activated by an active form of the Spz2 ligand . This situation is reminiscent to that observed in the embryonic ventral follicle cells , in which a Pipe-mediated signal induces a constitutive activation of the Easter cascade leading to Spz cleavage , Toll activation and , in turn , ventral fate acquisition [49] . It should be noted that Easter and one Pipe isoform are very strongly expressed in the trachea cells ( Flyatlas ) , and are candidate proteins in mediating Tollo activity in the respiratory epithelia . The fact that Ect4 , but not dMyd88 mutant , loss-of-function mutant phenocopies Tollo mutant suggest that Ect4 could be the TIR domain adaptor transducing Tollo signal in the tracheal cells . Alternatively , Ect4/SARM could mediate Tollo function by interfering with IMD pathway signalling . In mammals , SARM is under the transcriptional control of TLR and negatively regulates TLR3 signalling by directly interfering with the association between the RHIM domain-containing proteins TRIF and RIP [50] . Since PGRP-LC contains a RHIM domain as TRIF , and IMD is the Drosophila counterpart of RIP , one can envisage that Drosophila SARM could act by interfering with the PGRP-LC/IMD association required for IMD pathway signalling . Similarly to its function as a negative regulator in fly immunity , SARM is the only TIR domain-containing adaptor that acts as a suppressor of TLR signalling [36] , [50] . One obvious question relates to the mode of action of Tollo on IMD pathway downregulation . Two mechanisms have been recently described that result in the down-regulation of the IMD pathway . The first one regulates PGRP-LC membrane localization , and is dependent on the PIRK protein [51]–[53] . Upon infection , the intracellular PIRK protein is up-regulated and , in turn , represses PGRP-LC plasma membrane localization leading to the shutdown of the IMD signalling [53] . In infected pirk mutants , IMD-dependent AMPs are overproduced in both the gut and the fat body . In our conditions , however , inactivation of PIRK specifically in the trachea did not influence Drosomycin activation in trachea ( Figure S7A ) . To verify whether Tollo is acting via a mechanism similar to PIRK , we looked at PGRP-LC membrane localization using a UAS-PGRP-LC::GFP construct . PGRP-LC membrane localization was identical in wild-type and Tollo mutant tracheal cells ( Figure S7B ) . The second mechanism that modulates IMD activation , acts directly on the promoters of IMD target genes . Ha et al . ( 2005 ) have shown that the Caudal transcription factor sits on some of the IMD target promoters preventing their activation by Relish [54] . We thus tested the putative implication of Caudal in Tollo signalling by using Drs-GFP reporter transgenes containing either wild-type Caudal Responsive Elements ( CDREs ) or mutated versions unresponsive to Caudal activity [55] . Upon infection , Drs-GFP with mutated CDREs was activated in fat body but not in gut or trachea ( Figure S7C ) . In conclusion , Caudal acts as a transcriptional activator , rather than a repressor , for the Drs-GFP reporter in trachea . These results indicate that Tollo does not regulate the IMD pathway via PGRP-LC membrane localization or through promoter targeting of Caudal . One challenging task for the future will be to identify the mechanism used by Tollo to counter-balance tracheal PGRP-LC activation . It has been reported that the loss of Tollo function in ectodermal cells during embryogenesis alters glycosylation in nearby differentiating neurons [31] , [56]–[57] . Since the pattern of oligosaccharides expressed in a cell can influence its interactions with others and with pathogens , Tollo could function by modifying glycosylation pattern in response to microbes . It could be envisaged that Tollo mediates PGRP-LC glycosylation , and thereby reduces its ability to respond to bacterial elicitors . Further work will be required to address the above hypothesis , whereby Tollo activity and glycosylation modification could be linked in order to regulate the IMD pathway activation in trachea .
The following microorganisms were used: Erwinia carotovora carotovora 15 2141 ( Ecc ) , Erwinia carotovora carotovora 15 pOM1-GFP spectinomycinR ( Ecc-GFP ) , Escherichia coli 1106 ( E . coli ) and Micrococcus luteus CIPA270 ( M . luteus ) . Bacterial load of surface sterilized individuals was quantified by plating appropriate serial dilutions of lysates obtained from 6 dissected guts or trachea ( from larvae ) on nutrient agar plates ( Luria Bertani + spectinomycin 100 µg/ml ) . Biological triplicates were collected for each experimental condition at 4h and 24h after Ecc-GFP infection . Homogenization of tissues was performed using the Precellys 24 tissue homogenizer ( Bertin technologies , France ) and 0 , 75-1mm glass beads in 500 µL of LB + spectinomycin . PGRP-LCDE12 is a complete deletion of the PGRP-LC locus [16] . Flies carrying this mutation are unable to activate the IMD pathway . spzrm7 is a null allele which prevents Toll pathway activation [5] . yw , Drs-GFP [27] , Dpt-Cherry [58] , TolloC5 ( this work ) , TolloR5A [32] , Tollo145 [59] , UAS-TolloIR ( VDRC #9431 ) , UAS-Tollo::Myc [31] , DNT141 [46] , UAS-spz2IR ( VDRC #26115 ) , Ect4EY04273 BL#15733 , Df ( 3L ) ED4408 BL#8065 , Tl3 BL#3238 ( a dominant gain-of-function allele of Tl , Btl-Gal4 BL#8807 , UAS-myrRFP BL#7118 , act>CD2>Gal4 BL#4780 , cad-EGFP BL#30875 , Vkg-GFP ( a gift from Michel Sémériva ) , hs-Gal4 BL#2077 , RelishE20 [60] , imd1 [5] , UAS-spz act [30] , dMyd88c03881 [34] , UAS-PGRP-LC::GFP ( a gift from François Leulier ) and Dif1 [61] . Generation of the TolloC5 allele was performed as described in [62] using the two following inserted elements: d01565 and PBacf05248 [63] . Complete deletion of the Tollo gene was confirmed by sequencing genomic DNA extracted from TolloC5 mutants ( molecular details upon request ) . Fly stocks were raised on standard cornmeal-agar medium at 25°C . Cells from overnight bacterial cultures were recovered by centrifugation at 4 , 000 g for 10 min at 4°C . The supernatant was discarded and the pellet was resuspended in fresh LB media . Cell suspensions were serially diluted in PBS , and the concentration of cells was determined by optical-density ( OD ) measurement . 200 µl of an overnight bacterial culture of Ecc ( OD = 200 ) were directly added on top of feeding third instar larvae into a standard cornmeal-agar medium at 25°C . A similar volume of LB broth was used in control experiments . Larvae were monitored for Drosomycin and Diptericin transcription by fluorescence analysis using Drs-GFP and Dpt-cherry reporters respectively , and by qRT-PCR , 24h after infection . Septic injuries were performed by pricking adult males with a thin needle contaminated with M . luteus or E . coli . 200 µl of Bromophenol Blue ( SIGMA # B8026 ) at 10 g/l were directly added on top of feeding third instar larvae . For Drs-GFP study , Drs-GFP;UAS-myrRFP;act>CD2>Gal4 females were crossed to either ywhsflp;;UAS-IMD or to ywhsflp;; UAS-PGRP-LCa males . For Dpt-Cherry study , ywhsflp; UAS-GFP; act>CD2>Gal4 females were crossed to Dpt-Cherry; UAS-IMD or to Dpt-Cherry , UAS-PGRP-LCa males . In both cases , larvae of the progeny were heat shocked at early-mid L3 stage ( 72h-96h after egg deposition , AED ) and observed 24 h later . Generation of MARCM clones in trachea was performed by crossing MARCM virgin females of genotype ywhsflp;; Tub-Gal80 FRT2A en masse to the Drs-GFP; Blt-Gal4 , UAS-myrRFP; PGRP-LCDE12 FRT2A line . Resulting embryos were submitted to a heat shock 4–6 hr AED for 1 hr at 38°C in a circulating water bath , and kept at 25°C until larvae reached early-mid third instar ( 72h-96h AED ) , when they were infected by Ecc and observed 24 h later . Larval tissue were dissected in PBS and fixed for 20 min in 4% paraformaldehyde on ice . After several rinses in PBT ( PBS + 0 . 1% Triton X-100 ) , they were blocked for 1 hr in PBT-3% BSA at 4°C and then incubated with antibody at the appropriate dilution in PBT-BSA 3% overnight at 4°C . Primary antibodies were: rabbit Anti-Relish ( 1∶500 ) or Mouse Anti-Myc ( 9E10 Santa Cruz at 1∶ 500 ) . Several washes in PBT were followed by a 2 hr incubation with secondary antibody at RT ( Alexa Fluor 546 goat anti-rabbit IgG and Alexa Fluor 555 goat anti-mouse IgG diluted 1∶500 , Molecular Probes ) , then 5 washes in PBT . The tissues were finally mounted in Vectashield ( Vector Laboratories ) fluorescent mounting medium , with DAPI . Images were captured with a LSM 510 Zeiss confocal microscope . Quantitative real-time PCR and SYBR Green analysis were performed as previously described [58] . Primer information can be obtained upon request . The amount of mRNA detected was normalized to control rp49 mRNA values . Normalized data was used to quantify the relative levels of a given mRNA according to cycling threshold analysis ( ΔCt ) . For electron microscopic sections , third instar larvae trachea were dissected and fixed at RT in 4% PFA and 2% glutaraldehyde in 0 . 12 M sodium cacodylate buffer at pH 7 . 4 for 1 h . The trachea were then washed 3×10 min in 0 . 12 M sodium cacodylate buffer , post-fixed in 2% OsO4 in 0 . 12 M sodium cacodylate buffer for 1 h and washed again 3×10 min . Samples were subsequently dehydrated through series of ethanol gradients and infiltrated with propylene oxide , embedded in epoxy resin ( Fluka , Sigma ) and polymerized at 80°C . Ultrathin ( 80 nm ) plastic sections were cut using a Leica UltraCut microtome with a diamond Diatome knife and post-stained with 2% uranyl acetate , followed by treatment with Reynolds'lead citrate , and stabilized for transmission EM by carbon coating . Examination was performed with a Zeiss Leo 912 microscope at 100 kV . Images were captured using a Gatan 792 Bioscan camera using Digital Micrograph software . | Invertebrates solely rely on innate immune responses for defense against microbial infections . Taking advantage of its powerful genetics , the fly Drosophila melanogaster has been extensively used as a model system to dissect the molecular mechanisms that control innate immunity . This work led to the discovery of the essential role of the Toll-1 receptor in triggering the systemic immune response in flies , and paved the way for the discovery of the function of members of the Toll-like receptor ( TLR ) family in mammalian immunity . Whereas all TLRs are implicated in the mammalian immune response , Toll-1 was , so far , the only Drosophila Toll family member to be involved in the regulation of the immune response . In the present study , we show that another Toll family member , Toll-8 ( Tollo ) , plays an important role in controlling the respiratory epithelium immune response . Our data indicate that , by antagonizing the IMD pathway , Tollo is preventing over-activation of the antibacterial response in the airway epithelium . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] | [
"immunity",
"innate",
"immunity",
"immunology",
"biology"
] | 2011 | Toll-8/Tollo Negatively Regulates Antimicrobial Response in the Drosophila Respiratory Epithelium |
MicroRNAs ( miRNAs ) are small non-coding RNAs that regulate gene expression by binding to partially complementary regions within the 3’UTR of their target genes . Computational methods play an important role in target prediction and assume that the miRNA “seed region” ( nt 2 to 8 ) is required for functional targeting , but typically only identify ∼80% of known bindings . Recent studies have highlighted a role for the entire miRNA , suggesting that a more flexible methodology is needed . We present a novel approach for miRNA target prediction based on Deep Learning ( DL ) which , rather than incorporating any knowledge ( such as seed regions ) , investigates the entire miRNA and 3’TR mRNA nucleotides to learn a uninhibited set of feature descriptors related to the targeting process . We collected more than 150 , 000 experimentally validated homo sapiens miRNA:gene targets and cross referenced them with different CLIP-Seq , CLASH and iPAR-CLIP datasets to obtain ∼20 , 000 validated miRNA:gene exact target sites . Using this data , we implemented and trained a deep neural network—composed of autoencoders and a feed-forward network—able to automatically learn features describing miRNA-mRNA interactions and assess functionality . Predictions were then refined using information such as site location or site accessibility energy . In a comparison using independent datasets , our DL approach consistently outperformed existing prediction methods , recognizing the seed region as a common feature in the targeting process , but also identifying the role of pairings outside this region . Thermodynamic analysis also suggests that site accessibility plays a role in targeting but that it cannot be used as a sole indicator for functionality . Data and source code available at: https://bitbucket . org/account/user/bipous/projects/MIRAW .
MicroRNAs ( miRNAs ) are a family of ∼22-nucleotide ( nt ) small RNAs that regulate gene expression at the post-transcriptional level . They act by binding to partially complementary sites on target genes to induce cleavage or repression of productive translation , preventing the target gene from producing functional peptides and proteins . Despite advances in understanding miRNA:mRNA interactions , the rules that govern their targeting process are not fully understood [1–4] . While many miRNA targets have been computationally predicted only a limited number have been experimentally validated . Moreover , although a variety of miRNA target prediction algorithms are implemented , results amongst them are generally inconsistent and correctly identifying functional miRNA targets remains a challenging task . The majority of prediction tools are based on the assumption that it is the miRNA seed region—generally defined as a 6 to 8 nucleotide sequence starting at the first or second nucleotide—that contains almost all the important interactions between a miRNA and its target and their focus is on these canonical sites . This seed-centric view has been supported by structural studies [5] and a widely cited report [6] that investigated the importance of other ( non-canonical ) regions within a miRNA and concluded their contributions had relatively low relevance compared to the ( canonical ) seed region . However , more recent studies have revealed that many relevant targets are implemented via non-canonical binding and involve nucleotides outside the seed region , indicating that the entire miRNA should be considered in target prediction algorithms [3 , 7 , 8] . This is also supported by the performance of target prediction tools which typically identify approximately 80% of known miRNA targets , indicating the mechanisms associated with the remaining 20% of non-canonical targets remain poorly understood . Thus , there is an opportunity for novel approaches to improve knowledge of miRNA-regulated processes . In turn , this can lead to better understanding the effects of mutations in the non-coding region of the genome in terms of function and disease . To this end , in this work , we apply deep learning techniques to investigate the role of non-canonical sites and pairing beyond the canonical seed region in microRNA targets . Almost all target prediction methods are rule-based or adopt machine learning ( ML ) methodology with varying success . Rule-based systems incorporate various human-crafted descriptors to represent miRNA:gene target binding ( e . g . type of pairs in the site , binding stability , or conservation of the target site among species ) . Machine learning techniques also use human crafted descriptors , but as input features to machine learning models . The limitation of both these approaches is the process of feature selection and representation , which is constrained by the use of handcrafted descriptors to model a process that is not fully understood . Recent increases in computational power have permitted the rise of methods that can dispense with human-crafted features; making it possible to deal directly with raw data and autonomously learn and identify patterns to appropriately represent data . In particular , deep learning ( DL ) [9] has been shown to be an effective method for classification tasks in domains with complex feature representation . Generally , DL methods represent raw data by incorporating multiple hierarchical levels of abstraction . While this approach is typically applied to standard ML problems such as image classification [10] , natural language processing [11] or speech recognition [12] , it is now finding use in the life sciences for applications such as RNA splicing prediction [13] and gene expression inference [14 , 15] . DL has also been applied to the miRNA target prediction problem . Cheng et al . [16] used convolutional neural networks to analyze matrices of miRNA:site features , but the selected features were still human-crafted descriptors and thus the method faces similar problems as rule-based and ML approaches . A more recent work , DeepTarget [17] , relied on recurrent neural networks to identify potential binding sites and assess their functionality . However this work is still oriented to the identification of canonical sites and relies on a limited small data set for the training phase . In this paper we present miRAW , a novel miRNA target prediction tool that works with raw input data and which makes no assumptions about suitable input descriptors . miRAW scans the 3’UTR of the gene to identify potential target sites . It then uses DL to identify relevant patterns by directly analyzing the whole mature miRNA transcript , rather than focusing on the seed region and analyzing precomputed descriptors . It is trained and tested against experimentally verified positive and negative datasets . The resulting predictions can then be refined by incorporating exogenous information . When compared to other state-of-the-art miRNA target prediction tools , miRAW demonstrates a significant improvement in performance , highlighting the importance of considering pairing beyond the seed region . In order to gain a deeper understanding of the characteristics of non-canonical targets , we also investigated the prediction results in terms of model design ( i . e . , how different configurations affect the type of predictions obtained ) and from a biological perspective ( i . e . , how different classes of predicted target sites varied in terms thermodynamic stability and binding structures ) . In particular , results reveal ( i ) many potential functional non-canonical binding structures that are supported by experimentally verified miRNA:mRNA target data and ( ii ) commonly prioritized features such as site accessibility energy and seed region structure are relevant but not sufficient for discerning between functional and non-functional target sites .
A key factor for successful application of any ML classification technique is access to a sufficiently variable and representative dataset that will generalize a trained model to new and unseen data . For the miRNA target prediction problem , this requires a comprehensive dataset of verified positive and negative targets that encompass both canonical and non-canonical examples . While there are multiple data repositories providing information regarding experimentally validated positive miRNA targets [20–22] , there are significantly fewer experimentally verified negative targets . This is not an issue for methods that use rule-based approaches to describe positive matches [6] , but it represents a major concern for ML-based approaches that require similar numbers of labeled examples for both classes . Here , we focused on human data and used ( i ) Diana TarBase [21]—the most comprehensive publicly available dataset , which contains information for both positive ( 121 , 090 ) and negative ( 2 , 940 ) experimentally verified human miRNA:mRNA interactions—and ( ii ) MirTarBase [20]—containing 410 , 000 experimentally verified positive targets—as the knowledge core for our study . Annotation related to transcripts and miRNA binding site locations were obtained by cross-referencing Diana TarBase identifiers with miRBase release 21 [23] and Ensembl release 87 [24] entries . As a preliminary step , the Diana and MirTarBase data were parsed to ( i ) remove inconsistent entries that were marked both as positive and negative targets—due to contradictory results in different experimental validations—and ( ii ) combine entries that were validated more than once by different verification methods . This produced a final dataset of 303 , 912 positive ( + ) and 1 , 096 negative ( - ) miRNA:mRNA interactions . The data was then split into two parts ( each consisting of 151 , 956+ and 548- interactions ) for the training and testing stages ( see Fig A in S1 and S2 Files ) . Selection of candidate MBSs in a mRNA is another key step for a miRNA target prediction algorithm as it identifies which regions within a mRNA have the potential to be a target binding site . Most target prediction methods follow a similar approach for candidate selection: they scan the 3’UTR of the gene looking for sites that are partially complementary to the miRNA transcript; if a site fulfills certain criteria , it is considered to be a candidate site and is subjected to further analysis . Candidate site selection methods ( CSSMs ) that focus on the retrieval of canonical targets only consider those sites that have perfect complementary within the miRNA seed region ( nucleotides 2 to 8 , see Fig 2a ) and will return the smallest number of predicted targets . Methods willing to accept non-canonical sites have looser restrictions: some accept a limited number of bulges , mismatches or wobble pairs in the seed region whilst others accept such mismatches only if there are compensatory nucleotide pairs outside the seed region ( Fig 2b and 2c ) . In an ideal scenario where the training dataset contained sufficient examples of all the possible forms of positive and negative targets , the CSSM would not be required as , theoretically , an ANN would be able to estimate the function acting as CSSM . In reality , there are limited numbers of reliable experimentally verified miRNA:targets ( especially for negatively validated sites ) and the CSSM step effectively narrows the search space to simplify the ANN classification task . The CSSM used by miRAW ( CSS miRAW ) for searching the 3’UTR follows a similar approach to other prediction tools –investigating successive 30-mer segments– but employs a more relaxed set of restrictions that reflect recent experimental studies that relax the requirement of perfect pairing in the seed region and acknowledge a possible role for the other nucleotides . For example , Kim et al [8] report the role of nucleotide 9 in several miRNA binding sites and Grosswendt et al [2] found that a significant number of miRNAs do not require perfect complementarity within the seed region and compensate for this in non-seed nucleotides . Finally , a recent structural study by Klum et al [31] clarify a role for the 3’ end of the miRNA in the targeting process . Based on the findings from these and other related studies , we investigated three different approaches that expand the analysis beyond the typical 7mer seed region and relax the broadly adopted requirement for perfect pairing within the seed region . In particular , we consider a site to be a candidate MBS if there is a minimum number of base pairs—considering both Watson-Crick ( WC ) pairing and wobbles—within an extended seed region and investigated three different configurations: In each case , base pairs do not need to be consecutive in order to accommodate the presence of gaps and bulges . Thus , these models can accommodate both standard canonical MBSs as well as a broader range of non-canonical target site structures ( see Fig 2 ) , including the vast majority ( up to 97 . 63% ) of experimentally validated sites from Diana TarBase and CLIP/CLASH binding site datasets . Moreover , while these relaxed conditions for the seed region generate a much larger number of candidate sites , the decision of whether a site represents a functional target is delegated to the ANN ( which considers the entire miRNA & mRNA sequence ) . In this way , we ensure that minimal assumptions , and hence bias , are incorporated into the analysis . To further evaluate the impact of choice of CSSM , we also implemented the CSSMs used in two of the most commonly used miRNA target prediction tools: Both these CSSMs are subsets of CSSM-miRAW-6-1:10 and CSSM-miRAW-7-1:10 ( Fig 2 ) . Implementation of different CSSMs served the primary purpose of fine-tuning miRAW but also allowed us to investigate the targeting process from a biological perspective . The 5 proposed methods encapsulate different target ranges . At one extreme , CSS-miRAW-TS and CSS-miRAW-P adopt conservative approaches oriented towards canonical sites but they also consider a limited number of non-canonical sites with small irregularities in the seed region; at the other extreme , the other non-canonical CSSMs follow a greedier approach that allows the consideration of several non-canonical sites with broader irregularities in the seed region . These differences produce variations in both the canonical and non-canonical predicted targets . As an ANN requires numerical data for input , we transformed the miRNA and candidate mRNA site transcripts to binary values using one hot encoding . Each of the mRNA and miRNA nucleotides was translated to a binary vector of dimension 4 , corresponding to the four possible nucleotide values ( see Table 1 ) . Thus , each miRNA target is represented by two concatenated binary vectors: one composed of dimension 120 ( 4x30nt , where 30nt accommodates the longest known miRNA ) corresponding to the mature miRNA transcript , and a second composed of dimension 160 ( 4x40nt ) corresponding to the mRNA site ( 30 nt ) and 5 additional upstream and downstream nucleotides . These additional nucleotides seek to capture any influence that the flanking sequence may exert on the target [32 , 33] . The optimal number of additional upstream/downstream nucleotides was determined by evaluating how it affected the predictive power of the neural network ( see Fig B in S1 File ) . The number of additional nucleotides also conditions the window step size used when scanning the 3’UTR—a smaller window would result in a redundant analysis of potential sites by the neural network whilst a larger step would result in unscanned regions within the 3’UTR . Classification of candidate miRNA:MBSs was performed using a feed forward deep ANN . As we rely on the network to identify the relevant relationships between a sequence and the features that describe the miRNA:mRNA interaction , the input of the network consisted of the binarized transcripts of the miRNA and the MBS . The network was configured so that the number of inputs in the input layer was equal to the dimensionality of the binarized representation of the miRNA:mRNA transcripts , and the output layer consisted of two outputs ( positive and negative class classification ) . In addition , transcripts were aligned so the start of the seed region corresponded always to the same input node . The deep ANN was composed of eight dense hidden layers ( comprising rectifier activation function –RelU– nodes ) whilst the output layer comprised two softmax output nodes . The shape of the network was consistent with its intended functionality: ( i ) the first hidden sparse layer increases the dimensionality of the problem allowing the representation of data in a more complex dimension ( over-completion ) ; this layer does not necessarily improve the efficiency of the network autoencoder but it gives it “room” to explore the search space . ( ii ) Hidden layers one to five aim to identify the relevant features representing the data; they correspond to the first half of a stacked autoencoder . These layers were pre-trained as an isolated autoencoder in order to learn the features that are most representative of miRNA:MBS duplexes . ( iii ) the last three layers are responsible for classifying the features learned by the autoencoder; and follow the typical shape of a feedforward classification network . The number of nodes per layer was chosen experimentally using the guidelines in [34] as a starting point and resulted in the structure shown in Fig C and Fig D in S1 File . The size of the autoencoder was determined by minmizing the number of layers required to compress the data without losing important information ( relative error < 0 . 05 ) ; i . e . , a smaller network may struggle to capture important information whereas a larger one may require additional training time and would have higher overfitting risk . To ensure the network’s capacity to deal with newly observed data and to avoid overfitting , training was performed with a dropout rate of 0 . 2 . The maximum number of epochs was set to 500 in order to prevent excessive training time and overfitting . We tested two different loss functions for the network: negative log likelihood ( NLL ) and cross entropy ( XENT ) . After performing cross-validation , and determining the best model configuration ( see “Results:Neural Network Evaluation” ) we generated miRAW’s ANN model by retraining the network using the complete training dataset ( with the same proportion of positive and negative class instances ) . The two neurons of the output layer correspond to the negative ( output 0 , o0 ) and positive ( output 1 , o1 ) classes . Therefore , the class of the site is determined by the values of the two output neurons: c l a s s = { 1 if o 1 - o 0 > 0 - 1 if o 1 - o 0 ≤ 0 ( 1 ) This method will assign a positive or negative classification even if there is only a small difference between the positive and negative output neurons . This scenario corresponds to situations where the network is not confident about the classification of the input data . To deal with such uncertainty a constant parameter K was used to define a ‘grey area’ in which the network is not able to provide a reliable prediction: c l a s s = { 1 if o 1 - o 0 ≥ K - 1 if o 1 - o 0 ≤ - K u n k n o w n if - K < o 1 - o 0 < K ( 2 ) According to [35] , we consider that a miRNA targets an mRNA if any of the potential MBSs of the mRNA are functional . In the representation of the targeting process implemented within miRAW , we require the neural network classify at least one candidate site as positive to consider a miRNA:mRNA pair as a positive targeting event . In our model , given a miRNA m and a gene g , a candidate site selection method sm ( m , g ) determines a set of potential MBSs for that pair , i . e . s m ( m , g ) = C S ( 3 ) C S = { c s 0 … c s i } ( 4 ) To determine if the miRNA is targeting the gene , each candidate site within the miRNA:mRNA segment is binarized and input to miRAW’s deep ANN . The result of the targeting prediction T ( m , g ) corresponds to the disjunction of the neural network outputs ( ann ( m , csi ) ) for all the candidate sites csi ∈ CS in the gene g . T ( m , g ) = ⋁ i = 0 | C S | a n n ( m , c s i ) ( 5 ) The fact that it only requires a single candidate site to be classified as positive for the miRNA:mRNA prediction to be positive implies that miRAW is particularly sensitive to false positives . A false negative for a single candidate site can be abrogated by a positive classification for any of the remaining candidate sites but a single false positive cannot be corrected by any number of negative candidate sites . Hence , the more potential sites a CSSM identifies , the higher the probability of obtaining a false-positive prediction , reducing the performance of the classification due to a lower precision . P ( F P C S S M ) = 1 - P ( ! F P C S S M ) = 1 - ( 1 - F D R ) | C S | P ( F P C S S M ) = 1 - p r e c i s i o n | s i t e s | ( 6 ) where FDR corresponds to the false discovery rate of the neural network ( that can also be definied as 1—precision ) . This also implies that CSSMs that adopt a greedier approach will end up obtaining more false positives by chance . The presence of false positives in miRAW’s ANN can be partially attributed to the fact that not all the information concerning miRNA targets can be obtained from the miRNA:MBS duplex and , therefore , cannot be inferred by the neural network . For instance , aspects such as site accessibility [36] require accessing additional external data sources . This external information can be used to refine ANN outcomes by removing sites unlikely to be functional . In an attempt to reduce the likelihood of false positives , we included an a posteriori filtering step based on accessibility energy . It is known that miRNA binding sites that are more easily accessible tend to have higher chances of being functional targets [36]; for this reason , several tools usch as PITA , miRMAP [37] or PACMIT [38] combine this information with the binding site minimum free energy ( ΔGduplex ) to produce a refined target prediction . The site accessibility energy ( ΔGopen ) of a MBS can be defined as the energy required to unfold the secondary structure of the mRNA in order to accommodate the miRNA [36 , 38] . As the calculation of ΔGopen requires information that extends beyond the MBS and which involves the whole mRNA sequence , it is particularly well suited for use as a posteriori filter in miRAW . Following the site accessibility energy definition of [38] , we implemented an ΔGopen filter that removed all predicted sites presenting a ΔGopen higher than a threshold thsa . Based on results from previous studies [36 , 38] , we set thsa = −10kcal/mol . For accuracy and robustness , we computed local site accessibility following the guidelines defined in [39] and [38] . Specifically , we used the ViennaRNA package [26] and considered the 200nt surrounding the target rather than folding the whole mRNA sequence as this may result in less accurate and more complex secondary structures [38] . miRAW was implemented using Java . RNACoFold from the ViennaPackage [26] was used for computing the candidate sites . Implementation of the deep neural network was done using the DeepLearning4Java ( DL4J ) library [40] . DL4J allows the use of both CPU and GPUs for neural network training and classification . All the analyses presented in this paper were performed using GPUs due to its improved performance; however , a CPU based version of miRAW is also available . To assess miRAW’s performance , we compared it against the following commonly used target prediction tools: TargetScan release 7 . 1 [6] , Diana microT-CDS v . 4 [41] , PITA v . 6 [36] , miRanda ( built upon the mirSVR predictor ) [42] , mirzaG [43] , Paccmit [44] and mirDB [45] . These represent the current gold standards ( i . e . , most commonly referenced ) for microRNA target prediction software . These software periodically release datasets containing all available predicted target databases ( as of January 2017 ) with the test datasets defined in “Methods:Dataset Preparation” . We performed 10-fold validation for the TarBase test , and repeated the analysis for 100-fold ( Fig F in S1 File ) and the full TarBase test ( Fig G , Fig H and Table D in S1 File ) . For the 10-fold analysis we also characterized the predictions in terms of the structure and site accessibility energy distributions to try and to gain a better understanding of the identified targets in these terms . If a miRNA:mRNA was present in the test dataset , but missing in a release dataset , then that interaction was assigned the lowest possible score for that algorithm . For genes with multiple annotated 3’ UTRs we selected the isoform used to build the test dataset . In prediction datasets where such an isoform was not available or not specified , we tested the longest isoform available . The comparison was performed using the optimal reported configuration for each tool . For the release datasets that provided a prediction score , we used a variable threshold and the pRoc R package [46] to build receiver operating characteristic ( ROC ) curves . To assess the significance of the results , we performed a Wilcoxon signed rank test for each of the evaluated metrics; Results were considered significant for p < 0 . 05 unless otherwise stated ( see S3 File for specific p-values ) . To evaluate the correlation between the target predictions for a miRNA and the corresponding gene expression profiles in the Transfection Test Dataset we computed the coefficient of determination ( r2 ) between the reported predictions for each different algorithm and the changes of gene expression levels present within the corresponding transfection dataset . As some methods such as miRAW or microT adopt binary approaches that only evaluate whether or not a miRNA is targeting a gene , rather than quantifying the repression efficacy of the miRNA target , we also examined the expression changes for the top 1% of predicted targets for each of the algorithms , counting how many of the predicted targets presented significant changes in their expression levels . In contrast to the r2 test , this approach determines if higher confidence predictions are reliable and does not penalize methods that follow less restrictive approaches in order to favour sensitivity . As before , we considered the interactions not present in the release datasets as negatives and we assigned them the lowest score for that algorithm . mirSVR , mirDB and mirza-G have been excluded from this test due to absence of predictions for the transfected miRNAs we evaluated here . All genes within the testing datasets correspond to the GRCh38 release of the reference human genome whilst all the miRNAs appear in miRBase version 19 and later . All tested target prediction datasets with the exception of PITA and miRanda were built using these , or more recent , releases of miRBase and the reference human genome . This might impair the performance of PITA and miRanda as some of their negative predictions may not have been tested . TargetScan , Paccmit and MIRZA-G offer two different databases in each release , one providing target sites highly conserved among species ( TS_Conserved , Paccmit_Cons MG_Conserved ) and one providing sites not-necessarily conserved among species ( TS_NonConserved , Paccmit_NonCons , MG_NonConserved ) . In both cases , the two versions of the databases were considered .
Cross validation of miRAW’s ANN presented good results in terms of predicting both positive and negative sites . This was independent of the loss function used during training , with all evaluated metrics resulting in scores higher than 0 . 90 ( Fig 3 and Tables A and B in S1 File ) . Nonetheless , accuracy and area under the curve ( AUC ) metrics show that the XENT ( accuracy = 0 . 92 , AUC = 0 . 96 ) loss function resulted in a statistically significantly ( Wilcoxon signed sank test ) better network compared to the NLL function ( accuracy = 0 . 91 , AUC = 0 . 93 ) . This was reflected in both prediction of positive targets , where the XENT network achieved higher precision , sensitivity and F1-score compared to the NLL network . For negative target prediction , the XENT network returned a larger number of predictions than the NLL but nevertheless achieved a similar negative precision . It is worth noting that , across the different folds , the XENT network was less consistent in terms of negative precision than in positive precision and that for most of the folders it presented more FN than FP . This , combined with the difference in sensitivity and specificity values ( 0 . 92 vs 0 . 94 ) , suggests that the XENT network is slightly biased towards negative predictions as it predicted more negative than positive sites for each fold . Despite this fact , the statistically significant higher accuracy , AUC and F1-scores ( both positive and negative ) indicate that the XENT network is more appropriate than the NLL network for miRNA target prediction . Fig 4 shows the receiver operating characteristic ( ROC ) curves for the NLL and XENT networks . The XENT network has a larger AUC , indicating superior performance . Moreover , there is a clear difference in shape of the curves and distribution of data points . The XENT network exhibits a smooth curve with relativity evenly spaced points , the NLL curve is more discontinuous and the data points are concentrated within a smaller region . This indicates a stronger polarization of the NLL network , where all the predictions are strongly classified as a positive or a negative target ( i . e . class value is very close to 1 or -1 ) . Conversely , the smoothness of the XENT network represents a more progressive classification , allowing the presence of less polarized predictions , resulting in a more generalized predictive ability . The shape of the NLL curve also suggests that the NLL network might be overfitted and that it might struggle to classify new observations that significantly differ from the training data—this is also supported by the average epoch numbers used by each network to reach its optimal set of weights , 7 . 32 for the XENT network versus 11 . 21 for the NLL network . The general consistency of calculated parameters and ROC curves across the different folds in the two networks ( Tables A and B in S1 File ) indicates that the model performance is not dependent on the training and test datasets used . Fold 7 of the XENT network achieved the highest performance in terms of all the considered evaluation metrics and so this ANN model was selected for testing in the gene prediction stage . To investigate the impact of the site selection method , we compared the performance of five different CSSMs ( miRAW-6-1:10 , miRAW-7-1:10 , miRAW-7-2:10 , miRAW-TS and miRAW-Pita ) in the presence and absence of a site-accessibility energy ( AE ) filter of -10kcal/mol , summarized in Fig 5 and Table C in S1 File . All the methods achieve accuracies between 0 . 64 and 0 . 74 with significant differences depending on if the site-accessibility filtering is present ( AE ) or absent ( NF ) . This effect can be seen when the different CSSMs are ordered by accuracy . miRAW-Pita-NF , miRAW-TS-NF , miRAW-7-2:10-AE and miRAW-7-1:10-AE obtain similar accuracies ( ≈ 0 . 72 ) with no statistically significant differences in the metric , although miRAW-6-1:10-AE has a slightly poorer performance ( 0 . 71 ) . However , miRAW-7-2:10-NF , miRAW-7-1:10-NF , miRAW-Pita-AE and miRAW-TS-NF ranked in the bottom of the table in terms of accuracy . Thus , while the non-canonical CSSMs specifically derived for miRAW obtain better results in the presence of filtering , the canonical derived CSSMs ( miRAW-Pita and miRAW-TS ) exhibit improved performance in the absence of filtering . The F1-scores summarize how well a particular class is classified by a particular CSSM , an optimal CSSM will perform well for both positive and negative targets . Fig 5 and Table C in S1 File show that CSSMs with reported low accuracy underperform in at least one of the F1-scores: miRAW-7-2:10-NF , miRAW-7-1:10-NF and miRAW-6-1:10 have high positive F1-scores but a low negative F1-score caused by an excess of false positives ( causing a low specificity ) whilst miRAW-Pita-AE and miRAW-TS-AE have a low positive F1-score cause by the excess of false negatives ( causing a low sensitivity ) . However , miRAW-Pita-NF , miRAW-TS-NF , miRAW-7-2:10-AE and miRAW-7-1:10-AE all obtain balanced F1-scores ranging between 0 . 71 and 0 . 74 , indicating an ability to effectively predict both positive and negative targets . Fig 6 summarizes the composition of site types by each CSSM . Fig 6a shows the average number of canonical ( blue ) and non-canonical ( green ) sites identified for each miRNA:gene pair in the test dataset whilst Fig 6b shows the relative proportions of each identified type . As expected , CSSMs methods following conservative approaches ( miRAW-Pita and miRAW-TS ) identified more canonical than non-canonical potential sites , whereas the miRAW CSSMs identified larger total numbers of potential sites , of which many more were non-canonical . The figure also shows that the number of predicted canonical sites varies according to the selected CSSM , with the conservative approaches obtaining more canonical sites than the greedy approaches . While this seems to contradict the expectation that all the CSSMs should identify similar canonical sites , the difference can be understood when the higher number of accepted binding structures recognized by the non-canonical oriented CSSMs are taken into account . Several of the sites identified by miRAW-Pita or miRAW-TS overlap with non-canonical binding sites predicted by the miRAW specific CSSMs that present greater stability and are therefore preferentially selected ( Fig 7 ) . Fig 6b also shows that the application of the site accessibility filter does not significantly alter the ratio of canonical and non-canonical sites for any CSSM , suggesting that site accessibility filters do not act as a discriminator between canonical and non-canonical sites . Fig 5 shows that site accessibility filtering has very different effects in the canonical and non-canonical CSSMs . This difference can be understood by considering the different approaches taken by the canonical ( conservative ) and non-canonical ( greedy ) models . The conservative models only consider canonical sites containing close to perfect complementarity ( ≥7mers ) in the seed region and a restricted number of non-canonical sites , resulting in a limited amount of candidate sites . Conversely , the greedy models not only recognize canonical sites but also screen a wide range of non-canonical sites that follow unconventional target structures , obtaining a much higher number of potential target sites . This is illustrated in Fig 8a which shows the average and median number of potential MBSs identified for a miRNA:mRNA pair for each CSSM . The boxplot shows that , in the absence of filtering , more potential MBSs are identified for the more relaxed non-canonical CSSMs . For example , on average , miRAW-Pita_NF and miRAW-TS_NF identify between 3 and 4 sites respectively per miRNA:mRNA pair , whereas miRAW-7-1:10_NF and miRAW-6-1:10_NF identify ∼22 and ∼13 sites respectively . However , as more potential MBSs are identified , the chance of incorrectly obtaining a false positive increases . From the training results for the XENT network ( Table A in S1 File ) , we can estimate the overall probability of the network obtaining a false positive prediction as 0 . 068 ( P ( FP ) = 1 − precision ) . However , there is greater variation when we independently consider the various CSSMs ( Eq 6 ) . Fig 8b shows the relationship between the probability of obtaining a false positive and the average number of sites obtained by each CSSM . The non-canonical CSSMs with filtering are at the bottom left of the curve , indicating the efficacy of the filtering step , a-posteriori filtering for these CSSMs reduces the number of identified potential MBS which , in turn , lowers the probability of returning a false-positive error in the network and obtaining a false positive miRNA:mRNA classification– . Conversely , application of a posteriori filtering in conservative CSSMs significantly reduces the number of candidate sites , leading to the exclusion of true binding sites and increasing the probability of classifying a miRNA:mRNA pair as a false negative . We next investigated the relationship between the ΔGopen threshold applied in miRAW and the accuracy and the negative and positive F1-score metrics . The results are summarized in Fig 9a–9c . Again , the canonical and non-canonical CSSMs curves exhibit distinct characteristics . All the metrics for the canonical CSSMs improve with increasing ΔGopen , i . e . , these CSSMs are most effective in the absence of any threshold filtering . Conversely , the non-canonical CSSMs improve with low ΔGopen thresholds . While the positive F1-score progressively increases with high ΔGopen thresholds , the negative F1-score achieves a peak value between -8 and -10kcal/mol . This reflects the fact that , as the ΔGopen threshold is increased , more sites are considered by miRAW: increasing sensitivity ( more true positives ) but in turn reducing specificity ( less true negatives and more false positives ) . The observation that site accessibility filtering increased the performance of non-canonical CSSMs but decreased performance in canonical CSSMs suggests a potential bias towards low site-accessibility energy in non-canonical sites . To explore this possibility , we examined the site accessibility energy distribution of the sites predicted by the different CSSMs and separated them into canonical and non-canonical sites found in true positive ( TP ) and false positive ( FP ) miRNA:mRNA pairs . The results are shown in Fig 10a–10d . Fig 10a groups data according to type of site and classification outcome ( canonical TP , canonical FP , non-canonical TP and non-canonical FP ) . For the canonical TP sites , the different CSSM energy distributions do not present statistically significant differences between them ( pairwise comparison using Kolmogorov-Smirnov test; we considered differences as statistically significant when p < 0 . 05 ) ( Fig 10b ) ; this is explained by the fact that all the site selection methods identify similar canonical sites . For the FP canonical sites there are no significant differences among the three non-canonical CSSMs , however the miRAW-Pita CSSM does have significant differences with these CSSMs . From the corresponding graph in Fig 10a ( top right ) the peak in the miRAW-Pita distribution occurs around a threshold of -10kcal/mol whereas for the non-canonical CSSMs the peaks occur between -14 and -15 kcal/mol . The reason for this difference is unclear as we would anticipate a similar set of false positive canonical sites for all the CSSMs –as all consider sites with perfect seed region complementarity– . However , one possibly for this divergence may be a consequence of the fact that , in contrast to the other CSSMs , miRAW-Pita is the most conservative and does not consider pairing beyond the seed region as a factor for determining the binding site . This is also consistent with the situation shown in Fig 7 where some canonical sites identified by miRAW-Pita can accommodate more stable non-canonical bindings . This explanation is also consistent with the lower peak in the miRAW-TS CSSM curve , miRAW-TS only considers orthodox non-canonical sites ( involving several consecutive WC pairs outside the seed region ) compared to miRAW CSSMs . For the non-canonical TPs , there are significant differences between the canonical and non-canonical CSSMs . Although the profiles of the curves for the miRAW-Pita and non-canonical distributions appear similar , there are significantly more sites predicted at the peak energy ( ≈ -12kcal/mol ) for miRAW-Pita than for the non-canonical CSSMs , although it is unclear whether this is the only feature responsible for the estimated the significant differences between these distributions . Finally , for the non-canonical FP sites , there are significant differences between all the CSSMs , with the non-canonical CSSMs generally presenting distinct distributions compared to the TS and Pita CSSMs . As the exact differences between normalized energy distribution curves were unclear even between statistically significantly different situations , we also performed pairwise comparisons between the mean ( Mann-Whitney U test; p = 0 . 05 ) and median values ( Mood’s median test; p = 0 . 05 ) for each of the distributions . We found that the miRAW-Pita CSSM tended to have higher ΔGopen in non-canonical sites ( both TP and FP ) , which can be attributed to the fact that miRAW-Pita only considers non-canonical sites based on the seed region whereas both TS and CSSM-miRAW accommodate sites beyond this . Therefore MBSs in Pita non-canonical sites have less dependency on accessibility as they only need to accommodate a ( smaller ) seed region compared to the broader accessibility by the non-canonical sites in the other CSSMs . This is also observed to some degree between miRAW-TS and the miRAW specific CSSMS; miRAW-TS requires several consecutive binding sites outside of the seed region but the miRAW-CSSMs accommodate even more flexible structures . Fig 10c shows the same site energy distribution information , but with each plot grouped by CSSM . In all cases , the TP curves have smoother distributions than the FP curves and Kolmogorov-Smirnov tests Fig 10d ) report statistically significant differences ( p < 0 . 05 ) , between the energy distributions of the predicted canonical and non-canonical sites for almost all the CSSMs . False positives present more irregular distributions in all the CSSMs . Despite these observed differences , there is no clear explanation for either the distinction between canonical or non-canonical sites or true and false positives . In summary , the results in Fig 10 indicate that: there are differences in the energy distributions of the sites obtained using different CSSMs; there are differences between canonical and non-canonical sites; and there are differences between the energy distributions for the true and false positives . Nevertheless , these differences are not sufficient to identify any clear discriminatory features between MBSs , i . e . , ΔGopen application of a ΔGopen filter improved performance of CSSMs by reducing the amount of potential MBS ( thus reducing the probability of a false positive ) , rather than by identifying a relationship between ΔGopen and true or false positives . Fig 11 and Fig G and Table C in S1 File compare the performance of the best miRAW configurations—miRAW-Pita-NF , miRAW-TS-NF , miRAW-6-1:10-AE , miRAW-7-1:10-AE and miRAW-7-2:10-AE– with state-of-the art target prediction software tools—TargetScan release 7 . 1 [6] , Diana microT-CDS v . 4 [41] , PITA v . 6 [36] , miRanda ( built upon the mirSVR predictor ) [42] , Paccmit [44] , MirzaG [43] , and mirDB [45]—using the dataset defined in “Methods:Dataset Preparation” . All the miRAW implementations generally obtained ( statistically significant—See S3 File - ) better results than each of the other prediction tools for all the evaluated metrics , except for specificity , where most of the methods obtain a similar score , and for precision where mirDB also obtained similar results . Generally , the other methods presented low accuracies: TargetScan , PITA , miRanda and mirDB values were around 0 . 50; Micro-TDS achieved a value of 0 . 61 , but this was still much less than that reported for miRAW ( 0 . 78 ) . For mirDB , miRanda , TS_Conserved and TS_NonConserved , the low accuracies seem a consequence of their tendency to misclassify true targets as negative; i . e . , despite reporting high specificities ( > 0 . 70 ) , their negative precision , and hence their F1-score , was low . Conversely , PITA reported better sensitivity than specificity , but obtained similar positive and negative precision . Finally , microT-CDS did not show a particular bias towards any of the classes . It presented balanced specificity ( 0 . 63 ) and sensitivity ( 0 . 59 ) and similar precision ( 0 . 62 ) and negative precision ( 0 . 61 ) . Nevertheless , it was still outperformed by all the tested miRAW configurations . The balance behaviour observed for microT-CDS and miRAW is also reflected in the shape of their ROC curves when using the full dataset ( Fig H in S1 File ) , which presented higher AUCs ( miRAW > 0 . 75 , microT-CDS ∼ 0 . 72 ) . These results also highlight how the consideration of interspecies site preservation influences the prediction results . This is particularly apparent in the performance of the two different TargetScan releases . TargetScan achieved low accuracy for both conserved ( 0 . 53 ) and non-conserved datasets ( 0 . 56 ) ; in both cases the reason for such low accuracy are a consequence of the high number of false negatives . The TS_Conserved model presented a high specificity ( 0 . 94 ) , meaning that it classified almost all negative targets correctly , but a large number of positive sites were also misclassified as negative , which caused the negative precision to drop to 0 . 52 . Despite filtering positive targets using interspecies conservation information , the TS_Conserved precision ( 0 . 68 ) was still lower than the values achieved with any miRAW configuration . For the non-conserved sites dataset ( TS_NonConserved ) the increase in the number of positive predicted sites augmented the number of true positives and sensitivity ( 0 . 40 ) but this , in turn , increased the number of false positives , which reduced precision to 0 . 59 and specificity to 0 . 72 . Similarly , Mirza-G and Paccmit achieved better performance in the releases that do not consider interspecies conservation as a prediction factor . Comparison of TargetScan performance with miRAW reveals that both releases obtained lower performances than any miRAW configuration in all the evaluated metrics with the exception of specificity , which is a consequence of the conservative approach used by TargetScan regarding positive classification . Two of the tested methods ( PITA and mirSVR ) rely on thermodynamic features such as site accessibility ΔGopen or duplex stability ΔGduplex for target refinement . However , neither method achieved good performance . PITA obtained a relatively high sensitivity compared to other methods ( 0 . 62 ) meaning that it retrieved most of the positive sites , however it has a low precision ( 0 . 48 ) meaning that several negative targets were misclassified . Considering that PITA identifies mostly canonical sites and that it bases its classification on the combination of ΔGduplex and ΔGopen , this indicates that thermodynamic features alone are not sufficient for differentiating a positive and a negative target , consistent with our results in Figs 8–10 . As a consequence several negative targets with low accessibility energy are wrongly classified as positive . Conversely , the miRAW-Pita-NF results , which share the same site selection method , presents better scores in all the evaluated metrics . Considering that miRAW-Pita-NF uses the same canonical target-oriented CSSM as PITA but does not use ΔGopen for determining the functionality of the miRNA:mRNA pair , this also indicates that ΔGopen does not appear to be the most important ( i . e . most effective ) feature for evaluating canonical sites . This is consistent with the fact that miRAW-Pita-NF and miRAW-TS-NF outperformed miRAW-Pita-AE and miRAW-TS-AE , which only considered sites with low ΔGopen . However , our observation that the methods based on the non-canonical CSSMS had improved performance in the presence of a ΔGopen filter , suggests that this feature does have a role in target functionality . This apparent contradiction can be understood by recognizing that such a role is primarily linked to the broader set of non-canonical sites which correspondingly have a larger range of ΔGopen values , many of which possess higher secondary structure stability , therefore making binding site access difficult . Fig 12B shows the coefficient of determination ( r2 ) between the target predictions made by the different algorithms and the expression level changes present in the the transfection test datasets . Four non canonical-oriented variants of miRAW ( miRAW-7-1:10-NF , miRAW-7-2:10-NF and miRAW-6-1:10-NF , miRAW-6-1:10-AE ) obtained the best correlation , with r2 ≈ 0 . 033 , followed by microT-CDS ( r2 = 0 . 027 ) . Focusing on miRAW , non-canonical oriented methods clearly outperformed the canonical oriented ones , which ranked at the bottom of the table . In this test , site accessibility energy filtering did not improve the r2 score of the miRAW methods , although miRAW-6-1:10 performed similarly with and without filtering . Regarding the rest of the methods , microT outperformed TargetScan , Paccmit and Pita; this ranking is consistent with the one obtained by the different miRAW configurations as methods considering more non-canonical binding structures outperform the ones focused on canonical ones . The r2 values obtained by all the methods were low , this can be explained by the fact that methods tend to adopt a binary classification approach to predict if a miRNA is targeting a gene and that prediction scores reflect confidence on the predictions rather than the efficiency of the target itself; quantifying the effects of such targeting would require a regression approach . In addition error in the microarray measurements , different miRNA transfection efficiencies or secondary effects of introducing the miRNAs may caused unexpected variability in the expression profiles that can affect r2 . The low scores obtained in this analysis by even the best methods might also be consequence of the simplicity of the test: the datasets contain the gene expression changes provoked by an isolated miRNA in a specific tissue sample , a situation that is rarely observed in in vivo scenarios , where miRNAs and genes have many-to-many relations . Fig 12C shows the proportion of the most 161 confident predictions ( top 1% ) of the different algorithms that corresponded to genes differentially expressed in the microarrays of the transfection test dataset . The functional predictions provided by the different miRAW configurations ranged between 66 . 87% and a 16 . 62% depending on the tested microarray , being miRAW-6-1:10-NF the configuration that presented the best results ( μ = 29 . 25% ) . In the top ranked methods , configurations that did not rely on site-accessibility filtering obtained slightly better results ( μ = 29 . 25% , μ = 28 . 00% ) than those ones that filtered unaccessible sites ( μ = 29 . 15% , μ = 27 . 75% ) ; nevertheless , those differences cannot be considered significant . Similarly , non-canonical oriented methods obtained slightly better performances than canonical-oriented ones although the difference was not significant . Regarding the rest of tested methods , the TargetScan release that considered interspecies site conservation performed similar to miRAW ( 29 . 06% ) whilst the other methods were outperformed . The obtained results are fairly consistent with the results of the previous test as , in general , methods that provided higher precisions reported a higher proportion of differentially expressed genes in their predictions , being Paccmit the exception to that trend .
The imprecise nature of miRNA targeting allows the generation of complex regulatory networks and understanding the mechanisms and functions of these networks requires systematic experimental investigation . In the ideal world it would be possible to experimentally verify the target set of all miRNAs , but both the cost and limited throughput of current methods means that miRNA studies depend on computational predictions to complement experimental data . The requirement for complimentary base pairing within the seed region for miRNA targeting has been established through numerous experimental studies and forms the basis of all current prediction tools . Initially , it was assumed that seed region binding was a core requirement for all targets but , as more non-canonical targets were experimentally identified , prediction tools evolved to try to accommodate this divergence . The differences in how the various prediction tools recognize the relevance of specific deviations from canonical binding highlights the complexity surrounding the targeting process . The most conservative tools only consider targets that achieve full complementary pairing in the seed regions , whereas other tools allow compensatory binding to accommodate seed mismatches . Moreover , to accommodate non-canonical binding sites , current target prediction tools rely on the use of human crafted descriptors in an attempt to summarize current knowledge regarding miRNA:mRNA interactions , maintaining a bias towards properties associated with the miRNA seed region . Also , as knowledge has increased , so has the complexity of feature descriptors and consequently there is limited consistency amongst the different tools . Thus , researchers tend to use multiple prediction tools to adopt a “carpet bombing” approach to investigate target space , retaining only those targets that are common among a certain fraction of tools . This further biases predictions back towards the most conservative ( i . e . canonical ) targets . In this study , we adopted a neutral approach towards the prediction process , avoiding incorporating any knowledge related to the targeting process . The performance gap between miRAW and the descriptor-based approaches suggests that current knowledge is still not sufficient to accurately capture all aspects of the miRNA targeting process . This is consistent with recent studies , e . g . , [3] , [7] and [31] , which demonstrate that the whole miRNA can play a relevant role in many functional miRNA targets . Based on these findings , we took advantage of deep learning methodology to incorporate the whole miRNA sequence for target prediction . As deep learning has the capacity to automatically extract its own data feature descriptors , miRAW is not limited by the assumptions incorporated into current target prediction tools . Our experiments showed that miRAW consistently outperforms current techniques , suggesting that the descriptors learned by the deep neural network are able to encode current knowledge and include additional yet to be understood information . Moreover , we attempted to remove any preconceptions from the learning stage by including all miRNA and mRNA nucleotides as input to our model . The only knowledge we apply is during the selection of candidate targets where we implement a selection step to retain miRNA:mRNA pairs that have established binding within a relaxed seed region that spans nucleotides 1 through 10 . Despite the application of this selection step , the entire miRNA:mRNA sequence is used as input to the deep learning model . This has the benefit of narrowing the search space while retaining a larger number of candidate targets including non-canonical target types . In an ideal scenario , with enough representative positive and negative data samples , the selection step could be skipped as a deep enough neural network should be able to map such information into its weights . Relaxation of the seed region allows the consideration of both canonical and non-canonical targets , including the ones defined in recent studies that stated the importance of considering nucleotides beyond the 7th position [3 , 8] . This also aligns with recent studies which investigated potential binding sites based on microarray expression data that indicate a significant role for miRNA nucleotide 9 [8] and structure studies [31] that demonstrate off-site targeting in the 3’ region of the miRNA is achieved by a pivoting structural element , α helix-7 , within the Ago2 protein that permits rapid making and breaking of miRNA:target base pairs in the 3’ end of the seed region . This allows Ago2 to rapidly screen potential targets to dynamically search for non-canonical sites . The impact of the candidate site selection model can be seen from the results for different CSSMs within the miRAW model . Conservative approaches ( miRAW-Pita and miRAW-TS ) presented slightly better accuracies than more relaxed approaches ( miRAW-6-1:10 , miRAW-7-1:10 and miRAW-7-2:10 ) but their predictions were heavily biased towards canonical sites . On the other hand , more relaxed models identified a higher number of potential MBSs following both canonical and non-canonical structures . For the latter , the higher number of identified sites generated higher numbers of false positives that decreased precision and specificity . This problem was addressed by post filtering sites with a high ΔGopen value , increasing accuracy , precision and specificity to levels analogous to the ones obtained by miRAW-Pita and miRAW-TS but with a broader spectrum of binding structures . The contrasting performance among miRAW-Pita , miRAW-TS and the different miRAW specific CSSMs can be understood by the way in which the methods filter the candidate targets . miRAW-Pita has a conservative approach that discards any site containing more than one mismatch within the seed region without considering further the remainder of the mature miRNA transcript . This enhances the reliability of the positive prediction , but at the cost of increasing the number of false negatives as non-canonical sites are discarded . At the other extreme , miRAW adopts a more open strategy to maximize the types of sites ( both canonical and non-canonical ) that are considered . This more accommodating approach allows the detection of more non-canonical sites , but with the consequence of an increased number of false positives . This argument is also consistent with the results observed for miRAW-TS , which has the most restrictive CSSM—small irregularities are permitted in the seed region but this requires compensatory pairing in the 3’ end of the miRNA . The extended seed region permitted by miRAW leads to selection of positive sites with irregular bindings in the canonical seed region , supporting the argument that pairing beyond the seed region has a more important role . This is observed even with the miRAW-TS and miRAW-Pita CSSMs ( which still feed the whole transcript sequences into the machine learning model ) which obtain better results than their Pita and TargetScan counterparts . As deep learning has the capacity to automatically extract its own data feature descriptors , by incorporating the entire miRNA and 3’ UTR target region , miRAW is not limited by the assumptions incorporated into current target prediction tools . Our experiments showed that miRAW consistently outperforms current techniques , suggesting that the descriptors learned by the deep neural network are able to encode current knowledge and include additional yet to be understood information . Furthermore , as the amount of available target data increases , CSSM constraints can be relaxed which , in turn , will facilitate the discovery or disposal of additional non-canonical miRNA binding structures . Another important task within this work was the processing of different data sources to transform them into suitable training , testing and evaluation datasets . For a ML classifier to learn the patterns needed to distinguish different classes it is necessary not only to have good quality training data but also to have a balanced number of instances for each class . We selected Diana TarBase and mirTarBase as our core data sources as they represented the most comprehensive set of evidence for miRNA:mRNA functional interactions , spanning a range of different experimental methods and providing multiple evidence for many interactions . However , for most of the validated experiments the databases do not provide exact details of the target site for the supported interactions . To obtain reliable binding site information we processed and integrated PAR-CLIP and CLASH datasets -which reveal information regarding binding sites and binding structures but not regarding functionality- and cross-referenced them with TarBase and mirTarBase . Generally , there are many resources for experimentally verified positive data but access to experimentally verified negative data is more scarce . Some approaches solve this problem by generating synthetic negative examples , but these may not accurately represent real negative targets and are particularly inappropriate for the DL approach we implemented here . Thus , we generated our negative data by carefully selecting sites that had both the potential of providing stable miRNA binding and were associated with an experimentally verified negative target . Despite the enhanced performance demonstrated by miRAW , it is prudent to consider some of the potential limitations of automatic feature learning approaches such as DL . The hierarchical and opaque internal data representation learned by a neural network can be difficult to interpret and map into human interpretable knowledge , hence it is not possible to directly identify the features that determine the classification . To address this issue , studies on neural networks knowledge transferability [47] and weight constraining [48] may aid the interpretation process , which is the next logical step in our work . Another issue is incorporating knowledge that is external to the miRNA:MBS duplex transcripts . For example , some of the broadly incorporated features in current prediction tools , such as interspecies conservation or site accessibility energy , cannot be inferred by deep learning as these features are built upon external information not contained in the miRNA:MBS duplex transcript . In miRAW this is addressed by applying a posteriori filtering that refines the outcomes of the neural network using external information . As a first test of such a posteriori filtering , the ΔGopen filter proved to be effective at narrowing the search space in CSSMs possessing an elevated false positive probability ( caused by the high number of identified MBSs ) . An analysis of the ΔGopen energy distribution showed significant differences between canonical and non-canonical sites , and between true positive and false positive predictions . Nevertheless these differences were not enough to discriminate between these categories , indicating that ΔGopen energy is relevant in the targeting process but is not a sufficient indicator to identify target classes . Regarding interspecies conservation , the fact that miRAW , regardless of the CSSM , outperformed current methods without considering interspecies conservation information suggests that this has limited applicability as a descriptor for miRNA sites . This is also supported by the accuracy and F1-Score results obtained by TargetScan NC , which outperformed the ones obtained by TargetScan C . This is also consistent with research from a recent study which found that interspecies preservation filtering can be disregarded for functionally important non-canonical target sites [2] . Another consideration is the combinatory effect of multiple but weak binding sites which , acting in concert , can have significant functional roles [49] . Consistent with other target prediction tools , miRAW’s binding site centric approach cannot evaluate the joint regulatory effect of multiple potential weak target sites in a mRNA as sites are analyzed independently . Nevertheless , we observe that many miRNA:mRNA duplexes where the CSSM detected a high number of potential MBSs tended to be classified as functional . Although in many cases the classification might simply be a reflection of the increased probability of the ANN producing a false positive , it could also be a consequence of the ANN recognizing features associated with cumulatively strong targeting as proposed by [49] . Similarly , some of the false positive predictions obtained by the non-canonical CSSMs without Δopen filtering might correspond to weak targets without clear regulatory effects . This hypothesis could be explored in miRAW by updating the ANN in order to add a third class ( negative , positive and weak ) and by transforming the prediction aggregation Eq ( 5 ) into a cumulative function . Nevertheless , this approach would require the construction of a training dataset containing enough weak target examples , a challenging task considering that most existing miRNA target resources address the targeting from a binary perspective . The work presented in this paper focused on the prediction of human miRNA targets , nonetheless the methodology can be readily applied to build target prediction models in other species . Bearing in mind that data availability is crucial for building reliable machine learning classifiers , a logical next step is to implement a target prediction model for mouse . Beyond this , the presented approach will benefit from further experimental studies that will serve to validate new predictions obtained by miRAW but also to generate new experimental data to reliably expand the training of the model . Additionally , considering the a posteri filtering step can be applied in retrospective way , it can be used to re-investigate the relevance of some miRNA target descriptors , such as interspecies conservation . Finally , as miRAW considers the whole miRNA:mRNA transcript for its predictions , this also allows the use of miRAW to assess the impact of target site mutations and miRNA isoform variations on the targeting process , which have been shown to have functional roles and characteristic populations that can vary amongst different conditions . | microRNAs are small RNA molecules that regulate biological processes by binding to the 3’UTR of a gene and their dysregulation is associated with several diseases . Computationally predicting these targets remains a challenge as they only partially match their target and so there can be hundreds of targets for a single microRNA . Current tools assume that most of the knowledge defining a microRNA-gene interaction can be captured by analysing the binding produced in the seed region ( ∼ the first 8nt in the miRNA ) . However , recent studies show that the whole microRNA can be important and form non-canonical targets . Here , we use a target prediction methodology that relies on deep neural networks to automatically learn the relevant features describing microRNA-gene interactions for predicting microRNA targets . This means we make no assumptions about what is important , leaving the task to the deep neural network . A key part of the work is obtaining a suitable dataset . Thus , we collected and curated more than 150 , 000 experimentally verified microRNA targets and used them to train the network . Using this approach , we are able to gain a better understanding of non-canonical targets and to improve the accuracy of state-of-the-art prediction tools . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"transfection",
"learning",
"neural",
"networks",
"engineering",
"and",
"technology",
"gene",
"regulation",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"artificial",
"neural",
"networks",
"micrornas",
"cognitive",
"psychology",
"mathematics",
"nuclear",
"engineering",
"forecasting",
"statistics",
"(mathematics)",
"artificial",
"intelligence",
"computational",
"neuroscience",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"mathematical",
"and",
"statistical",
"techniques",
"gene",
"expression",
"molecular",
"biology",
"biochemistry",
"rna",
"site",
"selection",
"psychology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"structural",
"engineering",
"physical",
"sciences",
"computational",
"biology",
"non-coding",
"rna",
"cognitive",
"science",
"statistical",
"methods"
] | 2018 | miRAW: A deep learning-based approach to predict microRNA targets by analyzing whole microRNA transcripts |
In West Africa , envenoming by saw-scaled or carpet vipers ( Echis ocellatus ) causes great morbidity and mortality , but there is a crisis in supply of effective and affordable antivenom ( ISRCTN01257358 ) . In a randomised , double-blind , controlled , non-inferiority trial , “EchiTAb Plus-ICP” ( ET-Plus ) equine antivenom made by Instituto Clodomiro Picado was compared to “EchiTAb G” ( ET-G ) ovine antivenom made by MicroPharm , which is the standard of care in Nigeria and was developed from the original EchiTAb-Fab introduced in 1998 . Both are caprylic acid purified whole IgG antivenoms . ET-G is monospecific for Echis ocellatus antivenom ( initial dose 1 vial ) and ET-Plus is polyspecific for E . ocellatus , Naja nigricollis and Bitis arietans ( initial dose 3 vials ) . Both had been screened by pre-clinical and preliminary clinical dose-finding and safety studies . Patients who presented with incoagulable blood , indicative of systemic envenoming by E . ocellatus , were recruited in Kaltungo , north-eastern Nigeria . Those eligible and consenting were randomly allocated with equal probability to receive ET-Plus or ET-G . The primary outcome was permanent restoration of blood coagulability 6 hours after the start of treatment , assessed by a simple whole blood clotting test repeated 6 , 12 , 18 , 24 and 48 hr after treatment . Secondary ( safety ) outcomes were the incidences of anaphylactic , pyrogenic and late serum sickness-type antivenom reactions . Initial doses permanently restored blood coagulability at 6 hours in 161/194 ( 83 . 0% ) of ET-Plus and 156/206 ( 75 . 7% ) of ET-G treated patients ( Relative Risk [RR] 1 . 10 one-sided 95% CI lower limit 1 . 01; P = 0 . 05 ) . ET-Plus caused early reactions on more occasions than did ET-G [50/194 ( 25 . 8% ) and 39/206 ( 18 . 9% ) respectively RR ( 1 . 36 one-sided 95% CI 1 . 86 upper limit; P = 0 . 06 ) . These reactions were classified as severe in 21 ( 10 . 8% ) and 11 ( 5 . 3% ) of patients , respectively . At these doses , ET-Plus was slightly more effective but ET-G was slightly safer . Both are recommended for treating E . ocellatus envenoming in Nigeria . Current Controlled Trials ISRCTN01257358
Bites by saw-scaled or carpet vipers ( Echis ocellatus ) are frequent in the savanna region of West Africa , where agricultural workers and their children are at greatest risk [1]–[6] . Systemic haemorrhage , consumption coagulopathy , shock and debilitating local tissue necrosis may ensue [6] . Fatalities result from haemostatic failure susceptibility to which is conveniently detected and monitored by the 20 minute whole blood clotting test [6] which correlates with a plasma fibrinogen concentration of about 0 . 5 g/l [7] and has been used in several previous antivenom trials [3]–[6] , [8]–[17] . In Nigeria , where untreated case fatality exceeds 10–20% , E . ocellatus causes hundreds of deaths each year [1] , [18] , [19] . In recent years , antivenom has become scarce , costly and inaccessible to most patients [18] , [20]–[23] . This provides an entrée for unscrupulous marketing of geographically-inappropriate products that can prove clinically disastrous [3] , [24] , [25] . Improving the treatment of snake bite victims in Nigeria demands solutions to economic , logistical , marketing , distribution and storage problems associated with antivenom supply and provision of better training for medical personnel to optimize antivenom use [26]–[28] . The development of safe , effective and affordable antivenoms is a priority [28] . In the 1990s , the Federal Ministry of Health in Nigeria ( FMHN ) supported the development of a new ovine Fab monospecific antivenom raised against Nigerian E . ocellatus venom ( EchiTAb-Fab antivenom ) by MicroPharm , UK . This antivenom was tested [8] , [29] , registered by the Nigerian National Agency for Food and Drug Administration and Control ( NAFDAC ) and used in Nigeria from 1998–2000 . However , its use , like that of CroFab® in the United States [30] , was complicated by recurrent envenoming [31] attributable to rapid clearance of the Fab fragments [8] . To overcome this problem , it was replaced by a caprylic acid-refined , whole IgG antivenom ( EchiTAb-G ) ( ET-G ) with the same specificity . This proved clinically effective during compassionate clinical release in Nigeria , becoming the standard of care . During pre-trial use at Kaltungo , in early 2005 , 146 of 182 ( 80%; 95% confidence interval 74%–85% ) patients envenomed by E . ocellatus showed permanent restoration of blood coagulability 6 hours after an initial dose of 1 vial of ET-G ( unpublished data ) . This antivenom was registered by NAFDAC ( registration number A6-0078 ) . Recently , a new equine whole IgG antivenom ( EchiTAb-Plus-ICP ) ( ET-Plus ) was prepared by Instituto Clodomiro Picado , Costa Rica , also refined using caprylic acid [32]–[34] . It was raised against venoms of E . ocellatus , puff adder ( Bitis arietans ) and black-necked spitting cobra ( Naja nigricollis ) , the taxa of greatest medical importance in Nigeria . Pre-clinical ( rodent ED50 assay ) and preliminary clinical dose-finding and safety studies suggested that initial doses of 1 vial of ET-G and 3 vials of ET-Plus might cure the coagulopathy in at least two thirds of patients [35] . A natural progression of this work was a larger Phase III head-to-head comparison of both antivenoms in a non-inferiority design , since ET-G was the established standard of care . Since the untreated case fatality rate for E . ocellatus envenoming , inferred from results of treatment with inappropriate non-specific antivenoms , has been reported as 12 . 1% ( 95% CI: 6 . 3–22 . 1% ) [3] and 15 . 8% ( 95% CI: 10 . 4–23 . 4% ) [1] , it was considered unethical to include a placebo comparator arm . In this paper , we compare the effectiveness in correcting coagulopathy and safety of ET-Plus with those of ET-G ( standard treatment ) for envenoming by E . ocellatus , in a randomised controlled double-blind non-inferiority trial .
In the face of the current crisis in antivenom availability in Nigeria , our main aim was to evaluate effective and acceptably safe treatments for E . ocellatus envenoming in a randomised controlled double-blind non-inferiority trial , comparing ET-Plus , a new antivenom , with ET-G , an antivenom of established effectiveness which has been the standard for care in Nigeria since 2005 . The trial was designed to demonstrate non-inferiority of ET-Plus compared to ET-G . A 10% non-inferiority margin was deemed acceptable ( i . e . we stipulated that at least 70% of ET-Plus patients must have permanent restoration of blood coagulability at 6 hours for it to be deemed non-inferior to ET-G ) . A sample size of 198 in each group provides 80% power to detect this non-inferiority margin difference of 10% , at a 5% one-sided significance level . It was anticipated that there would be no losses to follow up . Patients were screened for eligibility and enrolled by one of the study clinicians who then contacted the hospital pharmacist directly for randomisation and provision of the allotted antivenom . Patients were allocated to receive either ET-Plus or ET-G with equal probability ( allocation ratio 1∶1 ) using simple randomisation . The random number sequence was generated using a table of random numbers . Treatment allocations were concealed by using sequentially numbered opaque sealed envelopes held by the hospital pharmacist who was otherwise independent of the study . He provided the masked antivenoms after reconstituting them to total volumes of 40 mls with sterile water for injection in an unmarked syringe . Diluted in this way , ET-Plus 3∶4 , ET-G 1∶4 , both antivenoms appeared as identical colourless solutions . The patient , ward staff and treating clinicians , who also assessed the outcome , were thus masked to the identity of the particular antivenom used . Patients were analysed in the groups to which they were assigned regardless of deviation from the protocol or treatment received ( intention-to-treat population ) . Baseline demographic factors and clinical characteristics were summarised using counts ( percentages ) for categorical variables , mean ( standard deviation [SD] ) for normally distributed continuous variables , or median ( interquartile [IQR] or entire range ) for other continuous variables . To determine the magnitude and direction of the treatment effects for dichotomous outcomes , relative risks and one-sided 95% confidence intervals were calculated; the lower limit was provided when comparing antivenom effectiveness ( where an increase in positive events is desirable ) whereas the upper limit was provided when comparing safety ( where a decrease in negative events is desirable ) . Continuous outcomes were checked for normality and the treatment groups were compared using either the t test ( for normally distributed data ) or the Mann Whitney-U test ( for non-normally distributed data ) . Treatment effects were presented as a corresponding difference in means or medians ( plus one-sided 95% confidence intervals ) . ET-G and ET-Plus were given limited registration for clinical trials by NAFDAC . The trial was sanctioned by NAFDAC and the Gombe State Medical Research Ethics Committee . Written informed consent ( in English or Hausa , the common language of this region ) was obtained from the patients after they had read the information sheet and discussed it with medical staff . Oral explanations in Tangale or Fulani were also available .
The trial was conducted during the period 2005–2007 , but recruitment was possible for only 9 months , because of staffing problems and delayed importation and authorisation by NAFDAC of the second consignment of ET-Plus antivenom . During this time , 1102 patients presented to the hospital with a history of snake bite and were assessed for eligibility ( Figure 1 ) . Among 646 patients who were ineligible because their blood was coagulable , 74 had been bitten by snake species other than E . ocellatus [night adders ( Causus maculatus ) , burrowing asps ( Atractaspis dahomeyensis and A . watsoni ) and cobras ( Naja nigricollis ) identified by examination of the dead snakes brought by these patients] and 572 showed only local or no envenoming . Among 456 with incoagulable blood ( 20WBCT ) , indicating systemic envenoming by E . ocellatus , 26 refused to join the study , 4 were excluded because of difficult venous access and 26 were ineligible for the following reasons as per protocol: pregnancy ( 10 ) , antivenom treatment for their present bite ( 2 ) , evidence of coma/cerebral haemorrhage on admission ( 4 ) , severe underlying illness ( HIV/AIDS ) ( 2 ) , bitten more than 72 hr previously ( 8 ) . Of 400 recruited to the study , 68 were children under the age of 14 . One hundred and ninety four patients allocated ET-Plus proved to be broadly similar to 206 allocated ET-G in their demographic and clinical characteristics , including age , height and weight , proportion who brought the dead snake ( Echis ocellatus ) that had bitten them , sex , proportion of children , delay between bite and admission , site of bite , severity of local envenoming and incidence of spontaneous systemic bleeding ( Table 1 ) .
In more than three quarters of the patients in this trial , blood coagulability was restored within 6 hr of the initial doses of ET-Plus and ET-G antivenoms , with no recurrence of incoagulability 12 , 18 , 24 or 48 hr later . The initial doses had been derived from preclinical test potency and confirmed by preliminary open dose-finding and safety studies [35] . Restoration of blood coagulability has been used as a surrogate marker of antivenom effectiveness in many clinical studies of viper-bite-induced consumption coagulopathy [3]-[6] , [8]-[17] . In this respect , ET-Plus not only proved non-inferior to ET-G , but showed weak evidence of superiority ( RR 1 . 10 one-sided 95% CI lower limit 1 . 01; P = 0 . 05 ) . The low incidence of recurrent coagulopathy ( 7 after ET-Plus , 17 after ET-G see Table 3 ) contrasted with the frequency of this phenomenon when Fab fragment antivenoms are used . This is partly attributable to the slower elimination of whole IgG antivenoms [8] , [30] , [36] . The effectiveness of ET-Plus and ET-G compares favourably with the results published for a new candidate equine F ( ab' ) 2 antivenom with specificity for E . ocellatus venom , African Antivipmyn® ( Laboratorios Silanes , Mexico ) [12] , . Among 289 patients recruited to an open multi-centre trial in Benin , West Africa , 79% had incoagulable blood suggesting systemic envenoming by E . ocellatus and 3% died . After treatment with an initial dose of 2 vials of African Antivipmyn® ( based on results of pre-clinical rodent assays ) that was repeated according to clinical criteria , 18% of the patients still had incoagulable blood 24 hr after starting treatment and restoration of blood coagulability was delayed beyond 60 hours in 6% of them [5] . Recurrent coagulopathy occurred in 25% of their patients compared to 6% in ours . Bleeding was arrested within 2 hours in 60% and within 24 hours in 80% , compared to an upper limit of 32 minutes in our patients . Surprisingly , results of a subsequent preliminary dose finding study of African Antivipmyn® in 129 patients suggested that reducing the average total dosage of antivenom from 3 . 81 to 2 . 21 vials per patient did not reduce the efficacy of the treatment , implying a rather flat dose-response curve [38] . ET-Plus caused one or more reactions in 50/194 ( 25 . 8% ) patients compared to 39/206 ( 18 . 9% ) for ET-G . The incidence of these early antivenom reactions should be compared to 15 . 2% and 57% with 10 ml and 20 ml of the original lyophilised EchiTAb-Fab [8] and 17% with SAVP Echis antivenom [9] . In the trial of African Antivipmyn® , “unexpected events” were observed in 19% of patients , including shock , dyspnoea , cough and angioedema [5] . In our trial , bronchospasm or gastrointestinal symptoms , classified as severe reactions , were more frequent with ET-Plus ( 10 . 8% compared to 5 . 3% ) probably reflecting the three times greater dose of IgG protein administered as the initial and subsequent doses of this antivenom ( 1 . 20 g ) compared to ET-G ( 0 . 37 g ) [35] and also possibly the greater dilution of ET-G ( 1∶4 ) compared to ET-Plus ( 3∶4 ) antivenom in the 40 ml injectate administered to all the patients . Only one quarter of the patients reported for follow up 2 weeks after the bite and among these , late serum sickness reactions were reported by 10 . 2% who had received ET-Plus and 5 . 2% ET-G ( Table 2 ) . However , these incidences may have been exaggerated if those with symptoms were more likely to return to the hospital . There were no fatalities among the study patients but without a placebo control group the prognosis of untreated E . ocellatus victims fulfilling our entry criteria is uncertain . However , the danger of E . ocellatus envenoming is shown by the fact that , at Kaltungo Hospital , 9 E . ocellatus victims died after the supply of antivenom faltered in June 2009 . Throughout the African savanna north of the equator , the overwhelming need is for an antivenom to treat envenoming by Echis spp . ( E . leucogaster , E . jogeri , E . pyramidum and especially E . ocellatus ) , the predominant cause of fatal and debilitating snake bite [1] , [3] , [4] , [6] , [8] , [18] . Systemic envenoming by Echis spp . is conveniently detected by simple tests of whole blood coagulability , such as 20WBCT , allowing the use of monospecific antivenoms such as ET-G or SAVP Echis antivenom even when the causative snake cannot be identified directly . However , ET-Plus is raised against venoms of two other medically important species complexes in the same region , puff adders ( Bitis arietans ) and spitting cobras ( Naja nigricollis ) , broadening its potential usefulness . Cross-neutralisation of venoms of several Echis , Bitis and Naja spp . by ET-Plus has been demonstrated in rodents [34] but its efficacy against envenoming by these species must be addressed by future clinical studies . Traditionally , the clinical use of antivenoms , a neglected class of biological drugs , has been based almost entirely on results of laboratory tests in animals . In future , however , WHO [28] will strongly encourage the procedure carried out in the present programme: preclinical tests and preliminary dose-finding and safety studies [35] followed by formal Phase III clinical trials . In the present study , conventional pre-clinical rodent ED50 assays reliably predicted clinical effectiveness of ET-Plus and ET-G . Surely such tests should become a minimal requirement before an antivenom is selected for clinical use in a particular country and to prevent the unscrupulous marketing of geographically inappropriate antivenoms [24] . This is now the policy of the Nigerian regulatory agency NAFDAC . In the case of E . ocellatus bites , in which coagulopathy is a clinically relevant and objectively measurable effect of envenoming , results of our preliminary Phase I dose-finding and safety tests , using a novel 3+3 dose escalation protocol , were confirmed by the outcome of the Phase III RCT . This protocol should be considered in similar situations as a substitute for conventional Phase I studies that many consider unethical for antivenom testing [39] . Comparison of initial doses of 3 vials of ET-Plus ( a new antivenom ) with 1 vial of ET-G ( whose clinical efficacy had been established through pre-trial use and has been the standard of care since 2005 ) demonstrated that , at this dose , ET-Plus was slightly more effective in correcting haemostatic effects of E . ocellatus envenoming in Nigeria . ET-Plus has the potential advantage of a broader spectrum of activity . However , 1 vial of ET-G was slightly safer . Both antivenoms are recommended for treating E . ocellatus envenoming in Nigeria . | Snake bite threatens millions of poor rural folk throughout Africa . In Nigeria , as in many countries of sub-Saharan Africa , it takes a terrible toll on human life and limb . Over the years , the news for those exposed to snake bite has been generally bad: withdrawal of antivenom manufacturers , increasing cost and , most recently , the marketing of ineffective or fake antivenoms in the region . Our paper reports encouraging results achieved by two antivenoms created as a direct consequence of the present crisis in antivenom supply for Africa . They have been assessed in the most powerful trial ever attempted in this field . The trial showed that in people with non-clotting blood following carpet viper bite , the commonest cause of snake bite morbidity and mortality in the West African savannah , administration of the antivenoms- EchiTAb G and EchiTAb Plus-ICP led to permanent restoration of blood clotting in 76% and 83% of the patients within 6 hours , respectively . Generally mild early adverse reactions were recorded in 19% and 26% , respectively . Both antivenoms proved effective and acceptably safe and can be recommended for treating carpet viper envenoming in Nigeria . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"public",
"health",
"and",
"epidemiology/global",
"health",
"public",
"health",
"and",
"epidemiology/occupational",
"and",
"industrial",
"medicine"
] | 2010 | Randomised Controlled Double-Blind Non-Inferiority Trial of Two Antivenoms for Saw-Scaled or Carpet Viper (Echis ocellatus) Envenoming in Nigeria |
The Actinomycetales bacteria Rhodococcus opacus PD630 and Rhodococcus jostii RHA1 bioconvert a diverse range of organic substrates through lipid biosynthesis into large quantities of energy-rich triacylglycerols ( TAGs ) . To describe the genetic basis of the Rhodococcus oleaginous metabolism , we sequenced and performed comparative analysis of the 9 . 27 Mb R . opacus PD630 genome . Metabolic-reconstruction assigned 2017 enzymatic reactions to the 8632 R . opacus PD630 genes we identified . Of these , 261 genes were implicated in the R . opacus PD630 TAGs cycle by metabolic reconstruction and gene family analysis . Rhodococcus synthesizes uncommon straight-chain odd-carbon fatty acids in high abundance and stores them as TAGs . We have identified these to be pentadecanoic , heptadecanoic , and cis-heptadecenoic acids . To identify bioconversion pathways , we screened R . opacus PD630 , R . jostii RHA1 , Ralstonia eutropha H16 , and C . glutamicum 13032 for growth on 190 compounds . The results of the catabolic screen , phylogenetic analysis of the TAGs cycle enzymes , and metabolic product characterizations were integrated into a working model of prokaryotic oleaginy .
Bio-Diesel is an energy-rich portable fuel derived mainly from triacylglycerols ( TAGs ) . Biodiesel and related fuels are extracted from oleaginous organisms , both photosynthetic and non-photosynthetic , that use available energy sources to fix carbon into high levels of stored lipids . In chemoheterotrophic organisms TAGs are synthesized by bioconversion of organic compounds such as the sugars and organic acids derived from globally-abundant cellulosic biomass . A genetic understanding of the oleaginous metabolism of chemoheterotrophic species like Rhodococcus provides critical insight for biofuels development . The high GC content Gram-positive Actinomycetales bacteria Rhodococcus opacus PD630 and Rhodococcus jostii RHA1 , a close relative that has a completely sequenced genome [1] , were previously shown to accumulate large amounts of TAGs and wax esters ( WEs ) [2] , [3] , [4] . Rhodococcus species present an attractive target for industrial processes due to high substrate tolerances and high density culturing on a rapid time scale as compared to many photosynthetic organisms [5] , [6] . The oleaginous metabolism of Rhodococcus goes beyond abundant lipid biosynthesis to include diverse hydrocarbon catabolism . R . jostii RHA1 was isolated from soil containing 1 , 2 , 3 , 4 , 5 , 6-hexachlorocyclohexane ( Lindane ) [7] , while R . opacus PD630 was enriched on phenyldecane as a sole carbon source after isolation from soil sampled at a gas works plant [2] . Rhodococcus can catabolize and detoxify several aromatic hydrocarbons that contaminate soil from industrial waste products . These toxic substrates include polychlorinated biphenyls ( PCBs ) [8] , [9] , [10] , [11] and other halogenated compounds such as Lindane that was used in large quantity for agricultural practices . Limitation of an essential nutrient stimulates enzymatic conversion of the non-limiting essential nutrients into stored polymers such as phosphorous conversion to poly-phosphate [4] , acetyl- and other short acyl-CoAs conversion to polyhydroxyalkanoates ( PHAs ) [12] , [13] , [14] , or the production of TAGs and WEs from these same short chain acyl-CoA primers [2] , [15] . Most prokaryotes store carbon as polyhydroxyalkanoic acids ( PHAs ) when other essential nutrients such as reduced nitrogen are limiting . By contrast bacteria in the order Actinomycetales have uniquely developed a storage lipid cycle that leads to accumulation of TAGs and WEs [16] . Abundant TAGs accumulation in Rhodococcus provides a pool of fatty acids for β-oxidation as cellular fuel , components of the plasma membrane , and substrates for the enzymatic production of the very-long and highly-modified extracellular lipids characteristic of Actinomycetales . Lipid metabolism in the genus Mycobacterium has been a major focus of scientific research due to the effect pharmacological inhibitors of lipid biosynthesis such as isoniazid [17] , [18] , [19] , thiolactamycin , and pyrazinamide [20] have on killing pathogenic mycobacteria . Whole-genome views of lipid metabolism in mycobacteria reveal these bacteria have developed several lipid biosynthesis systems and a large number of genes to support diverse and abundant lipid biosynthesis . In Mycobacterium an order to the enzyme activity of the lipid synthases has been established through genetic , biochemical , and pharmacological evidence; wherein lipids are biosynthesized de novo by the multifunctional FAS type 1a enzyme followed by further elongation via the FAS II system and the multifunctional MAS-family type 1b synthases . Collectively these 3 fatty acid synthase systems produce 2 classes of fatty acyl-CoAs that differ in chain length . One class contains lipids <20 carbons ( C20 ) that are components of the plasma membrane and the storage lipids TAGs and WEs . Another class contains lipids >C26 that are built by multiple synthases and can reach lengths as long as C90 in some Mycobacterium species [19] , [21] but only C60 in Rhodococcus opacus [22] and C54 molecules have been observed in R . equi [23] . The longer chain length lipids are used in Actinomycetales to build a protective extracellular coat that helps these bacteria survive in harsh environments whether it be the phagosome of a macrophage or contaminated soil enriched in toxic organic compounds . The interplay between multiple lipid biosynthesis systems in Actinomycetales [19] requires genetic understanding for engineering the flow of carbon to desired lipid types .
To establish a genetic model of Rhodococcus metabolism , we generated a high quality draft sequence of the Rhodococcus opacus PD630 genome . DNA sequencing with 454 shotgun and 3 kb paired-end reads resulted in 16 large scaffolds containing 9 . 27 Mb of assembled DNA sequence . We stitched the gapped genome scaffolds based on extensive chromosomal-synteny with the complete genome sequences of related Rhodococcus species R . jostii RHA1 and R . opacus B4 ( Figure S1 ) . The R . opacus PD630 genome contained 8632 genes that underwent metabolic reconstruction using pathway tools software [24] resulting in a model containing 1735 metabolic reactions . Enzymes were connected to metabolic reactions based on enzyme commission numbers ( EC# ) that were assigned by the EFICAz2 algorithm [25] and by gene-name recognition within pathway tools software . This automated EC# assignment allowed for multiple genomes to undergo metabolic reconstruction in parallel . Comparisons between metabolic reconstructions for a set of 8 phylogenetically related and one outlier species Ralstonia eutropha H16 that were assembled in this way can be browsed at ( http://tinyurl . com/opacuscyc14-5-comparative ) . The resulting initial metabolic reconstruction of R . opacus PD630 Opacuscyc14 . 5_comparative contains 400 metabolic pathways and 135 transport reactions . A more complete metabolic reconstruction of R . opacus PD630 by pathway hole filling using the pathway tools 14 . 5 software , additional EC # assignments made with the database at Kyoto Encyclopedia of Gene and Genomes KEGG ( http://www . genome . jp/kegg/ ) , and limited manual curation ( outlined in Figure S2 ) resulted in a metabolic reconstruction containing an additional 282 metabolic reactions , 44 metabolic pathways , and eight transport reactions . Opacuscyc_14 . 5 was improved by refining the metabolic model of TAGs biosynthesis and degradation by metabolic product characterization of uncommon fatty acids that accumulate to high levels in Rhodococcus . The results of a screen for bacterial growth on 190 metabolic compounds were used as a multi-genic test of the reconstruction , described in more detail below . Comparison of the results of the catabolic screen with the metabolic pathway predictions revealed that precision was 65% before refinement and 71% after refinement ( Table S1 ) . The current working model of Opacuscyc14 . 5 contains 2017 metabolic reactions , 444 metabolic pathways , and 143 transport reactions that can be browsed at ( http://tinyurl . com/4dv5m32 ) . Several eubacteria of the order of Actinobacteria including Rhodococcus , Corynebacterium , Mycobacterium , and Bifidobacterium are distinguished for having both the type 1a fatty acid synthase ( FAS ) ( Figure 1a ) and a FAS II lipid biosynthesis system . FAS is a polyketide synthase related protein containing all of the necessary enzymatic activities for de novo lipid biosynthesis ( Figure 1b ) . In mycobacteria FAS has been shown to elongate C2 and C3 carbon acyl-CoAs into C16 , C18 , C24 , and C26 fatty acyl-CoAs [26] , [27] . In the suborder of bacteria Corynebacterineae , the FAS II system of lipid biosynthesis elongates FAS product fatty acyl-CoAs through the enzymatic activity of 6 families of enzymes ( Figure 1c ) for biosynthesis of mycolic acids and related cell wall associated lipids . The FAS II system in these Actinomycetales operates on C16–C26 length fatty acyl-CoAs not the usual short chain lipid biosynthesis substrates characteristic of the related enzymes in bacteria , chloroplasts , and mitochondria . The number of unique FAS genes is shown in Figure 1a in the context of a phylogenetic tree of genera built by AMPHORA [28] . Unlike the widespread taxonomic representation of the FAS II genes , the FAS type 1a gene in Rhodococcus has genus representation in prokaryotes is limited to only within Actinomycetales . The FAS gene is likely to have emerged in Actinobacteria ( Figure 1a ) and was horizontally transferred to only those eukaryotic branches containing fungi and stramenopiles ( Figure 2a ) . Our genome-based metabolic reconstruction revealed close metabolic pathway relationships when we compared the lipid metabolism enzymes of the phylogenetically related genera Mycobacterium , Nocardia , and Rhodococcus . These Actinomycetales contain multiple pathways for lipid biosynthesis utilizing a unique combination of multifunctional fatty acid synthases ( FAS ) , a type 1a synthase and a related MAS-family type 1b synthase that generate linear- and branched-fatty acids respectively ( Figure 1D ) . Diacyl glycerol acyl transferase ( DGAT ) enzymes convert fatty and other carboxylic acids into TAGs [29] . DGAT genes are also of limited genus representation in eubacteria and interestingly are present in six out of seven genera that also contain at least one gene of the MAS-family type 1b synthases ( Figure 1a ) . Closely related genera to Rhodococcus share expanded gene-families of DGATs [29] , [30] . A phylogenetic tree based on 701 shared genes present in a single copy and general genomic features of representative species used for comparisons with Rhodococcus are presented in Figure S3 . The phylogenetic distance between the bacterial taxa that contain the FAS type 1a suggest this gene has likely been horizontally transferred within Actinomycetales . The FAS gene was duplicated in Corynebacterium and horizontally transferred between Bifidobacterium and Corynebacterium . The FAS type 1a protein from Rhodococcus is highly related , ranging between 60–65% amino acid identity , to the enzyme in mycobacteria . The FAS protein is comprised of 3128 amino acids with 7 domains ( Figure 1b ) , six of which catalyze distinct biochemical reactions ( Figure 1c ) . Multiple sequence alignment of Actinomycetales FAS type 1a proteins reveals conservation across all domains ( Dataset S1 and Dataset S2 ) . The order of type 1a synthase protein domains , high sequence similarity , and presence of all three key lipid biosynthesis systems ( FAS type 1a , MAS-family type 1b , and FAS II ) suggest that the substrates and products are similar for the shared enzymatic network of lipid synthases within the suborder Nocardiacae and Mycobacterium . A working model for carbon flow in Rhodococcus lipid biosynthesis is presented in Figure 1b . Acetyl-CoA is the product of many catabolic reactions and a key substrate in lipid biosynthesis ( http://tinyurl . com/4fgl6zo ) . Acetate and longer chain carboxylic acids captured from the environment can be converted to -CoA derivatives by CoA synthetases ( EC 6 . 2 . 1 . 1 ) thus feeding these organic acids into lipid metabolism ( Figure 1b ) . The first committed step in lipid biosynthesis is catalyzed by acetyl-CoA carboxylases that function as α/β complexes in Rhodococcus and related bacteria of the suborder Corynebacterineae to generate malonyl-CoA that is utilized by FAS type 1a , MAS-family type 1b , and FAS II enzymes for fatty acid biosynthesis ( Figure 1b and 1d ) . Malonyl-CoA is generated by ATP-hydrolysis dependent carboxylation of acetyl-CoA in reaction EC 6 . 4 . 1 . 2 by an expanded family of AccA ( α ) and AccD ( β ) enzymes in Rhodococcus ( Figure 3a ) . Homologous Rhodococcus AccA and AccD proteins were analyzed phylogenetically with those from related species ( Figure S4 ) . Some paralogous enzymes from these gene families were reported to recognize the –CoA derivatives of distinct carbon chain length organic acids [31] , [32] , [33] , [34] suggesting that a diverse pool of organic acids could be carboxylated and incorporated into cellular lipids by Rhodococcus . The pool of propionoyl-CoA in Rhodococcus can be converted into methylmalonyl-CoA via ATP dependent carboxylation by propionoyl-CoA carboxylase in reaction EC 6 . 4 . 1 . 3 ( Figure 1d ) . In Mycobacterium and Streptomyces methylmalonyl-CoA is a lipid biosynthetic substrate that is incorporated into methyl-branched lipids by the type 1b fatty acid synthases , a member of the mycocerosic acid synthase MAS-family of proteins [35] . OPAG_06239 is homologous ( 41% identity over 3527 amino acids ) to a protein within the MAS-family with all of the same twelve functional domains and common domain architecture as PKS12 ( Figure 1d ) , suggesting a related role in branched-lipid biosynthesis that was shown in M . tuberculosis to be production of C30–34 length methyl-branched phospholipids containing a 4 , 8 , 12 , 16 , 20-pentamethylpentacosyl lipid subunit [36] . Only one MAS-family type 1b gene is encoded in R . opacus PD630 and related R . opacus B4 , and R . jostii RHA1 genomes; whereas in the genera Mycobacterium , Frankia , and Streptomyces expansion of this gene-family has occurred ( Figure 2b ) . In Mycobacterium the MAS-family synthases generate methyl-branched lipids adding to the diversity observed in their cell wall lipids that have been shown to be key in persistence , pathogenicity , and immune recognition [37] , [38] , [39] . By contrast , Streptomyces also contain several MAS-family proteins but do not encode for the FAS type 1a protein ( Figure 1a ) . Streptomyces are rich in methyl-branched lipids [40] , [41] that are shorter than the methyl-branched lipids observed in Mycobacterium indicating a difference in the biosynthetic workload between MAS-family and FAS II genes that could result from the absence of the type 1a synthase gene in Streptomyces . The TAGs cycle includes twenty distinct enzymatic reactions starting from acetyl-CoA . The biochemical details of this cycle are presented with the corresponding EC number ( Figure S5 ) ( http://tinyurl . com/TAGs-cycle ) . The large expansions in homologous genes implicated in the TAGs cycle we identified in our initial metabolic reconstruction led us to further analyze the gene-families . We grouped the implicated TAGs cycle genes and families of genes based on protein similarity using the TribeMCL algorithm [42] on a small set of related bacterial species . We found that the genus Rhodococcus was deeply enriched in TAGs cycle genes including gene-families of very different sizes and there were no metabolic deficiencies in the multi-step metabolic-cycle ( Figure 3a ) . 261 candidate R . opacus PD630 TAGs cycle genes were identified for the TAGs cycle reactions ( Figure 3b ) . The largest gene family in the Rhodococcus TAGs cycle corresponds to the FAD dependant acyl-CoA dehydrogenases that operate in the β-oxidation of fatty acyl-CoAs ( EC 1 . 3 . 99 . 3 ) . R . opacus PD630 contains 71 of these acyl-CoA dehydrogenase genes whereas Corynebacterium glutamicum 13032 only contains two genes predicted for this reaction ( Figure 3a ) . Large gene families identified in the Rhodococcus TAGs cycle also include: DGATs ( EC 2 . 3 . 1 . 20 ) resolved in a phylogenetic tree ( Figure S6 ) , TAG lipases ( EC 3 . 1 . 1 . 3 ) , acyl-CoA synthetases ( EC 6 . 2 . 1 . 3 ) , enoyl-CoA hydratases ( EC 4 . 2 . 1 . 17 ) , and acetyl-CoA C-acyltransferases ( EC 2 . 3 . 1 . 16 ) . Rhodococcus species contained at least 261 genes that could contribute to this metabolic cycle without consideration of membrane transport genes as they have yet to be defined for each catabolic substrate . The TAGs cycle -CoA ligases/synthetases that ligate –CoA with carboxylic acids play a role in initiation as well a β-oxidation of lipids have been grouped into 1 category represented as ( EC 6 . 2 . 1 . 1/2/3 ) because we could not resolve the homologous enzymes that vary in their substrate chain length specificity ( acetyl- EC 6 . 2 . 1 . 1 , propionoyl- EC 6 . 2 . 1 . 2 , acyl- EC 6 . 2 . 1 . 3 ) . The largest number of genes in R . opacus PD630 dedicated to the TAGs cycle were mostly attributed to the acetate/acyl-CoA synthetases ( EC 6 . 2 . 1 . 1/2/3 ) ( 18 genes ) , acyl-CoA dehydrogenases ( EC 1 . 3 . 99 . 3 ) ( 71 genes ) , and enoyl-CoA hydratases ( EC 4 . 2 . 1 . 17 ) ( 64 genes ) . These large gene families are central to the β-oxidation pathway suggesting that Rhodococcus can catabolize and extract energy through aerobic respiration from a diverse range of carboxylic acids as well as biosynthesize from these compounds a diverse array of lipid products . The related Actinomycetale C . glutamicum 13032 provided stark contrast to the large number of lipid metabolism genes we observed in Rhodococcus . C . glutamicum 13032 had four pathway holes ( Figure 3a ) with a total of 19 genes implicated in the TAGs cycle ( Figure 3b ) . A substrate-permissive glycerol-3 phosphate acyl transferase could bypass the EC 2 . 3 . 1 . 15 pathway hole allowing for production of essential phospholipids; however the 3-hydroxyacyl-CoA dehydrogenase in the β-oxidation of fatty acids ( EC 1 . 1 . 1 . 35 ) is missing in C . glutamicum 13032 , a result that is consistent with the observed deficiency in catabolism of fatty acids described below . C . glutamicum 13032 is also missing the DGAT enzyme responsible for the final step in TAGs and WE biosynthesis . A complete table of genes for the twelve species presented in Figure 3 can be found in Table S4 . During fermentation of glucose , R . opacus PD630 and R . jostii RHA1 produced abundant lipids that were likely odd in carbon number based on their elution profile on gas chromatograph-flame ionizing detector analysis ( GC-FID ) ( Figure 4a and 4b ) . The odd-carbon lipid species were found in Rhodococcus TAGs that were purified by thin layer chromatography ( TLC ) prior to conversion into fatty acid methyl esters ( FAMEs ) for assay by GC-FID ( Figure 4c and 4d ) . During growth on glucose these odd-carbon fatty acids increase in relative abundance accumulating to as much as 30% in R . opacus PD630 and 40% in R . jostii RHA1 of the total lipids detected in the GC-FID assay ( Figure 4e ) . Fermentation analytics of media concentrations of ammonium and glucose as well as cellular total fatty acids , and residual dry weight ( Figure S7 ) indicated glucose depletion at 96 hours further stimulated R . jostii RHA1 to produce higher levels of odd-carbon lipids than R . opacus PD630 . The FAS type 1a enzyme purified from Mycobacterium phlei was shown to convert malonyl-CoA and the C3 substrate propionoyl-CoA into undefined fatty acids without propionoyl-CoA carboxylase in vitro [27] , suggesting that straight-chain odd-carbon lipids could be made by the type 1a FAS . We analyzed the chemical identity and structure of the Rhodococcus stored odd-carbon fatty acids to evaluate whether these lipids are methyl-branched or straight-alkyl chains in order to provide insight into the enzyme ( s ) responsible for this uncommon lipid biosynthesis . The known substrate preferences of the FAS type 1a enzymes are straight-chain substrates ( acetyl-CoA , propionyl-CoA , and malonyl-CoA [26] , [27] ) while the MAS-family type 1b synthases incorporate the C3 methyl-branched lipid substrate methylmalonyl-CoA [35] , [37] , [39] , [43] , [44] . To determine the identity and chemical structure of the putative Rhodococcus odd-carbon lipids we purified them as FAMEs by taking advantage of their enrichment when R . opacus PD630 and R . jostii RHA1 were grown on propionate as the sole carbon source ( Figure 4f ) . Propionate can be converted intracellularly to propionoyl-CoA through the activity of propionoyl-CoA ligase ( EC 6 . 2 . 1 . 2 ) ( Figure 1b ) allowing for degradation via the methylcitrate cycle [45] or incorporation into lipids via two routes . In the straight-chain lipid biosynthesis pathway diagramed in Figure 1b , propionoyl-CoA is a substrate in the initial condensation reaction with malonyl-ACP . By contrast , in the branched-lipid biosynthesis pathway diagrammed in Figure 1d , propionoyl-CoA is first converted into methylmalonyl-CoA prior to incorporation into lipids by the MAS-family type 1b synthases . FAMEs from propionate grown cells were purified via reverse phase HPLC then analyzed on a coupled gas chromatograph/electron ionization-mass spectrophotometer GC-EI-MS . We identifed ions with masses that corresponded to the methyl esters of pentadecanoic acid m/z 256 Da C15∶0 ( Figure S8 ) , heptadecanoic acid m/z 284 Da C17∶0 ( Figure S9 ) , and heptadecenoic acid m/z 282 Da C17∶1 ( Figure S10 ) . Fragmentation of these full length FAMEs resulted in ions that also matched previously reported spectra for electron ionized ions with these odd-carbon chain length FAMEs [46] . To discriminate between structural isomers with the same molecular weight ( straight-chain and methyl-branched lipids ) , we performed 1H-NMR on the HPLC-purified FAMEs from Rhodococcus ( Figure S11 ) . We saw evidence of the methyl esters as expected from the transesterification of cellular lipids with methanol at 3 . 68 ppm; however , the FAMEs purified from Rhodococcus showed no evidence of methyl-branching at 0 . 86 ppm along the aliphatic chain of the C15∶0 , C17∶0 , and C17∶1 fatty acids as compared to the branched control 16-methylheptadecanoate . The NMR spectra in addition to the mass spectra demonstrate that these odd-carbon lipids contained straight-chain alkanes and a cis-alkene , not the methyl-branched forms expected from a methylmalonyl-CoA intermediate generated by propionoyl carboxylase in reaction EC 6 . 4 . 1 . 3 . The absence of methyl-branched lipids in the shorter odd-carbon storage lipids indicated that de novo biosynthesis of methyl-branched lipids in Rhodococcus if present is restricted to the longer chain length lipids as previously described for the cell wall associated lipids in Mycobacterium . The identification of the stored pentadecanoic acid , heptadecanoic acid , and heptadecenoic acid odd-carbon straight-chain fatty acids could result from the FAS type 1a enzyme that is known to incorporate the three carbon molecule propionoyl-CoA and malonyl-CoA [27] . The odd-carbon fatty acids isolated from Rhodococcus grown on glucose contained predominantly seventeen carbons , that is in the range where most FAS type 1a products are released from the type 1a FAS ( C16–18 ) [26] . Rhodococcus fermentation of low-cost organic substrates into oil requires a more complete understanding of the catabolic capabilities of these species . We tested 190 organic compounds as the sole carbon source in time course growth assays with four soil-derived bacterial species including the Actinomycetales R . opacus PD630 , R . jostii RHA1 , Corynebacterium glutamicum 13032 , and the Gram-negative β-proteobacterium Ralstonia eutropha H16 . Four chemical categories of compounds including carboxylic acids , nitrogen containing , carbohydrates and alcohols , and oligosaccharides were tested in these bacterial time course growth assays . Compounds capable of supporting growth yielded growth values that were clustered hierarchically to show catabolic-relationships between growth substrates and the species being compared . Carbohydrates and alcohols are an important class of compounds to evaluate for converting natural organic streams such as cellulose and cellulose derived sugars , sugarcane , and beet sugars into biofuels . Seventeen oligosaccharides were screened for growth resulting in the identification of 13 , 8 , 3 , and 0 growth substrates for R . opacus PD630 , R . jostii RHA1 , C . glutamicum 13032 , and R . eutropha H16 respectively ( Figure 5a ) . R . eutropha H16 showed no ability to degrade the oligosaccharides tested and relatively few monosaccharides were catabolized indicating that this species has many pathways to catabolize organic streams such as the rich pool of carboxylic acids defined below or fix CO2 by using the energy derived from splitting H2 [47] , [48] , [49] . Corynebacterium and Rhodococcus catabolized the disaccharides sucrose and maltose as well as the trisaccharide maltotriose . Rhodococcus species grew poorly on disaccaharide maltose but well on the trisaccharide maltotriose suggesting a possible membrane-transport preference . R . opacus PD630 gene OPAG_05551 is a glycogen hydrolase that contains the predicted activity to account for maltotriose growth that are part of an operon containing OPAG_05551 that breaks down glycogen ( glucose α1–4 ) with Pi to yield glucose-1-P . Glucose 1-P isomerase is also utilized in galactose catabolism to generate the glucose 6-P that is common to glycolysis , pentose phosphate , and Entner Duoderoff catabolic pathways that are all complete pathways in our sequence based metabolic reconstructions of Rhodococcus ( http://tinyurl . com/GLC-6-P ) . The compound-specific growth assays indicated that galactose and oligogalactoside metabolisms differ between R . jostii RHA1 and R . opacus PD630 ( Figure 5b ) . The ability of R . opacus PD630 to metabolize galactose enables efficient growth on the oligogalactosides lactose , lactitol , melibionic acid , melibiose , lactulose , raffinose , and stachyose ( Figure 5a ) . A search of our metabolic reconstruction for the genetic basis of the galactose phenotypic differences between R . opacus PD630 and R . jostii RHA1 led to identification of a galactose-catabolic region of Rhodococcus genomes that is shared between the two R . opacus species B4 and PD630 , but not in the closely related R . jostii RHA1 species . Two divergent polycistrons in the galactose catabolic region are fully syntenic between the R . opacus PD630 and R . opacus B4 species that contain the hydrolytic α- and β-galactosidases as well as solute binding protein ( SBP ) , solute binding protein transporters ( SBPT ) , and transcription regulatory protein of the DeoR family ( Figure 5c ) that could collectively hydrolyze and transport mono- and oligo-galactosides supporting growth on these compounds . Carboxylic acids appear to be an important carbon source for Rhodococcus and R . eutropha H16 . To determine the utilization of carboxylic acids as carbons source , we tested 64 carboxylic acids for growth using the previously described assay ( Figure 6 ) . These analyses led to the identification of 39 growth substrates for R . opacus PD630 and R . jostii RHA1 ( 61% of carboxylates tested ) , 15 growth substrates for C . glutamicum 13032 ( 21% of carboxylates tested ) , and 34 growth substrates for R . eutropha H16 ( 53% of carboxylates tested ) ( Figure 6 ) . Acetic and propionic acids support growth in all species tested as did other common intermediates in central metabolism such as pyruvic , succinic , and citric acids . All of the species tested grew on gluconic acid that is degraded by the pentose phosphate pathway . Similar to the observed galacto-saccharide phenotypes , R . opacus PD630 uniquely degrades D-galactonic acid-g-lactone ( Figure 6 ) . By contrast , growth on the short chain hydroxy acid glycolic acid is specific to R . jostii RHA1 and R . eutropha H16 ( Figure 6 ) . Growth was observed for the longer chain α and β hydroxybutyric acids by both Rhodococcus species and R . eutropha H16 , whereas weak growth on γ hydroxybutyric acid was again only seen for R . jostii RHA1 ( Figure 6 ) suggesting that this species has a more diverse hydroxy-acid metabolism . R . opacus PD630 , R . jostii RHA1 , and R . eutropha H16 grew well on the C10 dicarboxylate sebacic acid suggesting that β-oxidation of these longer chain dicarboxylates provides rapid growth . β-oxidation could also explain the growth on Tween 40 and Tween 80 compounds by Rhodococcus and Ralstonia ( Figure 6 ) . The Tween compounds are converted to fatty acids upon ester-hydrolysis by cutinase proteins encoded for in the genomes of Actinomycetales [50] . Tween 20 supported growth of only R . eutropha H16 likely due to toxicity for the other species resulting from the hydrolytic release of a C12 fatty acid . C . glutamicum 13032 is limited in carboxylates catabolism as seen by the inability to catabolize sebacic acid and the longer chain Tween 40 and Tween 80 ( Figure 6 ) . A limited carboxylate catabolic-profile observed for C . glutamicum 13032 in our growth assay is consistent with genes missing for 3-hydroxyacyl-CoA dehydrogenase within the fatty acids β-oxidation pathway ( Figure 3a ) . The identification of the sialic acid or N-acetylneuraminic acid , an abundant component of extracellular glycoproteins , as a growth substrate for the soil bacterium C . glutamicum 13032 ( Figure 6 ) was unexpected because this pathway has only been described for bacteria that colonise animals [51] , [52] . Examination of the C . glutamicum 13032 genome reveals a likely sialic acid catabolic operon containing genes for a secreted sialidase ( cg2935 ) , a sialic acid ABC transporter ( cg2937–2940 ) of the satABCD type seen in Haemophilus ducreyi [53] and a full set of catabolic genes ( cg2928–9 and cg2931–3 ) genes that would allow C . glutamicum to degrade sialic acid to fructose-6-phosphate , pyruvate , and ammonia [51] . A related but incomplete catabolic operon is seen in C . diptheriae NCTC 13129 . The sialic acid catabolic genes are not conserved in the other species tested nor were related transporters shown to facilitate uptake of sialic acid [54] , consistent with these phenotypic results . Rhodococcus displayed the most diverse catabolism of nitrogenous compounds in our species-comparative time course growth assays . Clustered heats maps relating the time course growth of 38 nitrogenous compounds tested . 26 nitrogenous compounds supported growth in at least one of the species tested ( Figure S12 ) . A notable feature of the Rhodococcus species nitrogenous metabolism is the ability to catabolize the branched amino acids that result in the formation of propionoyl-CoA thereby establishing a cycle of propionoyl-CoA in carbon storage and amino acid catabolism ( http://tinyurl . com/propionyl-CoA ) . A ranked order list of growth values on 190 compounds for each species in the screen is presented in Table S2 . Comparison of the protein domains ( Pfam database , Sanger Institute UK ) within a set of related Actinomycetales and the Gram-negative outlier R . eutropha H16 revealed that the R . opacus PD630 genome contained 3 membrane transport protein families in the top 9 expanded families of total encoded protein domains ( Figure S13 ) . The membrane transport major facilitator super family ( MFS ) is the most prevalent Pfam domain in R . opacus PD630 . There are 229 MFS genes in R . opacus PD630 as compared to 176 in R . jostii RHA1 , 104 in R . eutropha H16 , and 47 in C . glutamicum 13032 ( Figure S14 ) . The high relative number of membrane transporters likely enables the broad catabolism we observed in R . opacus PD630 and R . jostii RHA1 . The ability to degrade sterols is shared between Rhodococcus and Mycobacterium . Following intravenous injection , M . tuberculosis has been observed to colonize lung tissue that is rich in lipid bodies and cholesterol crystals [55] . Genetic analysis of M . tuberculosis cholesterol catabolic pathways showed this sterol to be an essential carbon source during M . tuberculosis infections in mouse lung models [56] , [57] , [58] . Complete sterols degradation requires catabolism of both of the aliphatic branched side chain as well as the terpene polycyclic rings; however catabolism studies of M . tuberculosis showed that 14C labeling at the fourth position of the steroid A ring was released as CO2; whereas the label at position 26 within the sterol branched side chain was converted into phthiocerol dimycocerosate ( PDIM ) [57] . This study indicated that part of the sterol was being degraded for energy and the other was being used for assembly of cell wall associated branched-lipids . Sterol A and B ring degradation results in propionoyl-CoA and pyruvate while side chain degradation results in a 2∶1 formation of propionoyl-CoA: acetyl-CoA . In M . tuberculosis the the propionoyl-CoA from cholesterol degradation is converted into methylmalonyl-CoA then incorporated into branched-fatty acids such as PDIM by the PKS12 MAS-family type 1b synthase [35] , [37] , [43] , [44] , [59] . A large region ( ∼0 . 28 Mb ) of Rhodococcus chromosomal DNA has been identified through transcriptomic and genetic analysis of R . jostii RHA1 grown on cholesterol as a sole source of carbon [60] . Within this large chromosomal region , there are six clusters of genes that encode for the multiple enzymes dedicated to sterol degradation including: membrane transport , side chain degradation , sterol A and B ring degradation , and sterol C and D ring degradation , for recent review [61] . We performed whole genome alignments between R . opacus PD630 , R . jostii RHA1 , and R . opacus B4 , M . sp JLS , M . vanbaalenii PYR-1 , M . smegmatis MC2 155 , M . tuberculosis H37Rv , C . glutamicum 13032 , and S . avermitilis MA 4680 to evaluate gene conservation within the six gene clusters implicated in cholesterol degradation ( Figure S14 ) . We found extensive conservation in the cholesterol degradation genes between Mycobacterium and Rhodococcus as has been reported previously [60] . Our analysis indicates Streptomyces contains many of the key genes for cholesterol degradation but is lacking homologues to the Mce sterol-transport genes [57] . C . glutamicum 13022 did not contain the genes implicated for cholesterol degradation . The only major difference found between the previously described sterol degradation chromosomal region in R . jostii RHA1 and R . opacus PD630 is the presence of a transposase gene in gene cluster 2 ( OPAG_09155 ) .
The Rhodococcus genomes encode multiple biosynthetic pathways for making lipids and expanded gene families within those pathways that contribute to the diversity and abundance of lipid products seen with some Actinomycetales . We demonstrated that Rhodococcus TAGs contain the uncommon straight-chain odd-carbon lipids pentadecanoate , heptadecanoate , and heptadecenoate by mass and structural determination . The high abundance of the straight-chain odd-carbon lipids are yield-controllable through feeding of the three carbon organic salt propionate . Propionate is converted to propionoyl-CoA , a metabolite that is generated during the catabolism of the branched amino acids isoleucine , valine , and threonine as well as sterols . Propionoyl-CoA levels likely increase intracellularly during nitrogen and glucose starvation resulting from elevated protein and amino acid degradation as the physiology of these cells adapt to these nutrient limitations . Elevated propionoyl-CoA has been proposed to explain increased production of the branched-lipid PDIM when mycobacteria is grown on cholesterol as a sole source of carbon [59] . The FAS type 1a protein in the genus Rhodococcus is highly related to the mycobacterial protein that has been demonstrated to incorporate propionoyl-CoA during de novo lipid biosynthesis and produce lipids with the same lengths observed in the stored lipids TAGs and WEs . Our genomic analysis indicates conservation of all three interconnected lipid biosynthesis systems in Nocardiacae and Mycobacterium namely FAS type 1a , MAS-family type 1b , and FAS II . This group of bacteria also shares the ability to store lipids in high abundance . This conserved lipid biosynthetic network provides means for extrapolating functional information derived from decades of lipid metabolism research in Mycobacterium to the elaborate lipid metabolism in Rhodococcus . Pharmacological studies in combination with enzyme activities performed with purified systems demonstrated that Mycobacterium uses FAS type 1a to convert two and three carbon organic acids into straight-chain fatty acids containing 16 , 18 , 24 , and 26 carbons [20] , [26] , [27] . FAS type 1a products are released as the –CoA derivative of a fatty acid due to the dual activity of the FAS type 1a malonyl palmitoyl transferase ( MPT ) domain [21] . The FAS acyl-CoA products can then proceed through multiple lipid biosynthetic pathways . Studies of mycobacterial lipid metabolism have established a genetic model for this related oleaginous bacterium that connects multiple biosynthetic systems for the building of storage , plasma membrane , as well as very long chain cell wall-associated lipids . The close phylogenetic relationship between the genus Mycobacterium and Rhodococcus resulted from common ancestry wherein these genera share many features of their elaborate lipid metabolisms . Much of what is known about Actinomycetales odd-carbon lipid metabolism comes from work that described longer chain extracellular lipids that are methyl-branched . Prior to this study , it was unclear whether the <C20 odd-carbon lipids in Rhodococcus are methyl-branched as a result of incorporating methylmalonyl-CoA [16] . Using 1H-NMR to analyze purified lipids we could distinguish between the structural isomers of branched and straight-chain fatty acids demonstrating that the stored odd-carbon lipids in Rhodococcus are straight-chain . The odd-carbon storage lipids in Rhodococcus could result from propionoyl-CoA and malonyl-ACP condensing in the initiation phase of FAS type 1a biosynthesis ( Figure 1b ) . The substrate requirements for production of straight-chain odd-carbon lipids matches the reported substrate specificity of the multifunctional FAS type 1a enzyme and its product chain lengths . The combination of metabolic product identification and metabolic pathway mapping through genomic sequence supports a serial order to the elaborate network of lipid biosynthesis in Rhodococcus , similar to Mycobacterium , wherein FAS type 1a initiates and elongates fatty acids releasing –CoA products that are in the range of C16–C18 and C24–C26 . FAS produced fatty acyl-CoAs can be further elongated in the type II system or by the MAS-family type 1b synthase . The Rhodococcus type 1b synthase OPAG_06239 , like the homologous PKS12 gene from Mycobacterium , contains two modules consisting of 6 protein domains in a direct repeat ( Figure 1d ) . The PKS12 enzyme displays dual specificity for malonyl-CoA and methylmalonyl-CoA within the same polypeptide and oligomerizes in a tail-to-head fashion to perform multiple elongation cycles on C16- and C18-CoA molecules resulting in C30–34 multiply methyl-branched fatty acids . Enzyme assays with single modules ( 6 domains ) of PKS12 and site-specific mutations within the PKS12 acetyl transferase ( AT ) domains demonstrated that the N-terminal module incorporates methylmalonyl-CoA whereas the C-terminal module incorporates malonyl-CoA [43] . This mode of biosynthesis results in branched fatty acids that contain a methyl-branch at every fourth carbon from the point of initial condensation by the type 1b enzyme . This alternating of elgonation substrates mechanism explains the curious methylation at every fourth carbon in the mannosyl-β-1-phosphomycoketide ( MPM ) molecules isolated from Mycobacterium [36] . In Mycobacterium , the enzymes of the type 1b MAS-family and FAS II system share a common preference for longer chain length fatty acyl-CoAs resulting in the observed order in biosynthesis enzyme activity . The order of conserved domains and sequence homology of the Rhodococcus type 1b synthase OPAG_06239 suggests that this enzyme will function similarly in some aspects of branched lipid biosynthesis that is characteristic of the MAS-family enzymes in the related Mycobacterium species; however no methyl-branched lipids have been identified in Rhodococcus to date . In species such as Mycobacterium that have both FAS type 1a and 1b synthases , there is a distribution in labor amongst the synthases that is dictated by chain length; wherein FAS type 1a functions as the initiating synthase and the type 1b synthase incorporates methyl branch substrate methylmalonyl-CoA during lipid biosynthesis of longer chain length lipids . In more distantly related Actinomycetales , the genera Saccharopolyspora , Salinspora , Frankia , and Streptomyces we observed a type 1b but no type 1a synthase ( Figure 1a ) ; thus preventing the distribution of labor observed in Nocardiacae and Mycobacterium . Streptomyces have been reported to store abundant methyl-branched lipids [41] that are shorter in chain length than those isolated from Mycobacterium . The observation of odd-carbon lipids in the TAGs from Rhodococcus intrigued us as a possibility that type 1b synthase activity was contributing to the accumulated TAGs . The chemical analysis we performed indicated the stored TAGs were straight-chain lipids of the appropriate length to have been built by type 1a FAS through iterative biosynthesis with two of this enzymes known substrates ( propionyl-CoA and malonyl-CoA ) . How is the strict distribution of labor observed between type 1a and the other type II and 1b synthases maintained in Nocadiacae and Mycobacterium ? The crystal structure of the related FAS from S . cerevisae [62] revealed this synthase is a hexamer complex of apoenzymes . A hexamer complex also explains the behavior in ultracentrifugation studies of the FAS type 1a complex from mycobacteria [26] . The structural studies of fungal FAS provided a structure-based model wherein fatty acids are biosynthesized within the cavity of a 2 . 6 MDa β barrel structure . The enzyme FAS-ACP domain with growing acyl chain accesses individual catalytic domains shuttling biochemical intermediates from one active site to the next within the FAS type 1a hexamer complex [62] . Fatty acid products of the appropriate length are released as acyl-CoAs by the MPT domain of FAS type 1a . The FAS hexamer in Mycobacterium display a bimodal distribution of product chain lengths that are ( C16–18 and C24–26 ) [26] . We propose the FAS 1a acyl-CoA products become accessible to the other synthases once released from within the β barrel structure of the FAS hexamer by the palmitoyl transferase activity of the MPT domain . Acetyl/acyl transferase ( AT ) domains in type 1 fatty acid synthases display remarkable substrate diversity . In the type 1a synthase , the AT domain loads acetyl–CoA on the FAS-ACP domain then is subsequently transferred to a cysteine in the keto synthase ( KS ) domain . The second substrate for condensation is the elongation substrate malonyl-CoA that is transfered by the FAS-MPT domain to the FAS-ACP domain then delivered to the KS domain for decarboxylating condensation with acetyl-CoA resulting in acetoacetyl-ACP . The type 1b synthases that inititiate fatty acid biosynthesis , like the mammalian FAS , lack the MPT domain thus use their AT domain to load both acetyl-CoA and malonyl-CoA . MAS-family type 1b enzymes display significant substrate diversity by varying both the acyl substrates ( acetyl-CoA , acyl-CoAs ) as well as elongation substrates ( malonyl-CoA , methylmalonyl-CoA ) . We found seven Actinomycetales genera that contained the MAS-family type 1b synthases compared to 4 genera that encode the type 1a synthase . Six out of seven Actinomycetales genera that encode type 1b synthases also encode at least one storage enzyme of the DGAT/WE family . We conclude the relatively widespread taxonomic representation of the type 1b synthases and the diversity of substrates these enzymes react on suggests that there are lipid variants yet to be identified . The large number of genes dedicated to lipid metabolism in Actinomycetales is the result of gene duplications , multifunctional FAS type 1a and MAS-family type 1b gene emergence , horizontal gene transfer , and emergence of the DGAT/WE enzymes that catalyze the transesterification of fatty acyl-CoAs with diacylglycerol . We performed phylogenetic analysis on genes identified by metabolic reconstruction in an attempt to predict the most likely activities for each enzyme encoded in the R . opacus PD630 genome . We used the TribeMCL algorithm to describe the size of the gene family and all of the family members for a set of related Actinomycetales . Phylogenetic tree formation with these related species was used to resolve the orthologous from paralogous enzymes , thus facilitating the extrapolation of functional data for related proteins to our in silico model of Rhodococcus Opacuscyc14 . 5 . The broad catabolism and elaborate lipid biosynthesis described in Rhodococcus indicates that there is substantial enzymatic activity on chemically-related compounds that could be explained by expansion in gene-family sizes and genetic drift to encode enzymes with slightly variant substrate recognitions from ancestrally-related enzymes but still catalyze similar chemical reactions . The 8632 genes in R . opacus PD630 provide a large arsenal of metabolic enzymes for an oleaginous lifestyle within soil . We found protein domains contained in transport proteins to be the most abundant class of domains encoded in Rhodococcus as well as a distinctive profile of transporter types that are distinguishing for Rhodococcus ( Figure S13 ) . Identification of structural genes and operons that play a role in the lipid body assembly process have already begun to benefit from genomic sequence [63] . Chemoheterotrophic organisms capable of fermenting sugars and a broad spectrum of organic compounds derived from cellulosic and other natural resource biomasses through bioconversion provides an industrial process to convert agricultural side- , natural resource- , and industrial waste-streams into fungible fuels . R . opacus PD630 differs from R . jostii RHA1 in its ability to catabolize the cellulosic sugar galactose and oligogalactosides . Both Rhodococcus species degraded the cellulosic sugars glucose and rhamnose; however , none of the species tested were able to degrade the disaccharide cellobiose containing β1–4 linked glucose . Cellulose catabolism requires hydrolysis of β1–4 linked glucose , indicating that these species do not encode for these hydrolases . The most abundant component of hemicellulose is xylan that is broken down to xylose . Consistent with the cellobiose deficiency these species were unable to catabolize xylose . Genetic complementation of cellulosic degradation pathways in Rhodococcus provides a streamlined approach for cellulosic biomass conversion into oil-based fuels . Rhodococcus species catabolized a diverse array of carboxylic acids that corresponded to an expansion in the acyl-CoA ligase gene family that link organic acids with biosynthesis and catabolism . Organic acids are produced in the mixed-acid fermentations of cellulose degrading organisms thus indicating a potential next-generation strategy for 2-phase fermentations of cellulosic biomass to TAGs that could be done without genetic modification . Our working model of the R . opacus PD630 metabolism began with genome sequence that allowed phylogenetic comparisons to be made with related species that have been studied in far greater molecular detail . Phenotypic information about catabolism in Rhodococcus provided a powerful multigenic test that guides the metabolic reconstruction towards completion through phenotype-directed pathway curation . Literature-based pathway curation united the reported biochemical reactions of M . phylei FAS type 1a protein with metabolic products that we purified and characterized during metabolic model refinement of R . opacus PD630 . The improvements to our genetic model of R . opacus PD630 metabolism provides a template for further refinement with the integration of data from genetics , biochemistry , metabolomics , lipidomics , and transcriptomics that will be the focus of future work .
R . opacus PD630 was obtained from the DSMZ strain 44193 . R . eutropha H16 ATCC17699 and C . glutamicum ATCC13032 were from ATCC . R . jostii RHA1 was a gift from Lindsay Eltis at the University of British Columbia . Genomic DNA was extracted from R . opacus PD630 as in [64] without the addition of mutanolysin . Genomic DNA was sequenced at 454 Life Sciences ( A Roche Company ) according to manufacturer protocol for shotgun ( Rapid Library ) and 3 kb paired-end reads using GS FLX Titanium sequencing chemistry . The resulting DNA sequence was assembled using the GS De Novo Assembler software version 2 . 0 . Open Reading Frames ( ORFs ) were predicted from assembled genome sequence using a combination of in silico ORF predictions and gene mapping from the annotated R . jostii RHA1 genome . In silico ORFs were predicted using GeneMark [65] and Glimmer3 [66] . For synteny based gene-mapping Nucmer [67] was used to find local alignments between every contiguous sequence read ( contig ) in the R . jostii RHA1 and every contig in the R . opacus PD630 . A whole-genome synteny map was then built from these local alignments by chaining together collinear hits , allowing up to 10 , 000 bases of undetectable similarity between anchors , then filtering out chains that overlap a larger chain on either sequence by more than 90% of their length . The remaining chains , corresponding to syntenic regions of DNA , were globally aligned using LAGAN [68] . Each transcript in the R . jostii RHA1 genome was mapped onto the R . opacus PD630 by attempting two mapping techniques and then selecting the transcript with the higher-scoring global alignment to the reference transcript's coding sequence . The first method projects the boundaries of each gene onto the target genome using the coordinates in the raw whole-genome alignment . The second method uses the whole-genome alignment to define a target region containing the reference transcript , and then uses GeneWise to build a gene model by aligning the reference protein to the target region . Final ORFs were defined by comparison of in silico ORFs and mapped ORFs with hits to Pfam [69] and the top blast hits against the non-redundant protein database . ORFs with overlap to non-coding RNA features ( see below ) were reviewed and removed when appropriate . Discrepancies in the final ORFs were resolved via manual review . Ribosomal RNAs ( rRNAs ) were identified with RNAmmer [70] . The tRNA features were identified using tRNAScan [71] . Other non-coding features were identified with RFAM [72] . Every annotated gene in the Rhodococcus opacus PD630 genome is assigned a locus number of the form OPAG_##### both at the Broad Institute web site and in GenBank with accession ABRH01000000 . The R . opacus PD630 V3 genome assembly contained 293 ( contigs ) , containing 16 scaffolds longer than 2 , 000 residues . The 16 large scaffolds of the R . opacus PD630 genome , 10 of which were largely syntenic with R . jostii RHA1 and R . opacus B4 linear chromosomes , were ordered and oriented to reflect the observed synteny followed by the smaller contigs . Nucmer and mummerplot , components of the mummer 3 . 0 software package [67] , were used to align the genomes and create graphical representations of those alignments . The mummerplot graph files were annotated to show boundaries between each of the scaffolds in the mummerplot graphs . Through manual inspection of these graphs , the syntenic order and orientation of these contigs were assembled into a stitched FASTA file with the addition of 500 N nucleotides at scaffold termini . Enzymes for nine comparative species in Opacuscyc14 . 5_compare were computationally predicted using the EFICAz2 algorithm , enzyme prediction using gene-name matching , pathway prediction [24] , transcription unit prediction [73] , transporter prediction [74] , and pathway hole filling [75] was performed with the Pathway Tools 14 . 5 software [24] . These databases can be accessed at ( http://rhodocyc . broadinstitute . org ) . In addition , Opacuscyc14 . 5_working also predicted enzymes using homology to proteins with an EC# assignment in the database at Kyoto Encyclopedia of Gene and Genomes KEGG ( http://www . genome . jp/kegg/ ) . Subsequent literature-based manual curation was used to refine Opacuscyc14 . 5 TAGs biosynthesis and degradation cycle . A set of directed gene pairs was generated by performing an all-against-all BLASTP search ( min % aligned = 10 and e-value<1e−5 ) between a comparative set of genomes ( all accession numbers available in Table S3 ) . Genes were clustered using OrthoMCL [76] with a Markov inflation index of 1 . 5 and a maximum e-value of 1e−5 . Gene clusters were identified to which the M . tuberculosis FAS , MAS , PKS12 , and DGAT genes belonged to . These clusters had all members plotted on the AMPHORA phylogenetic tree of the Actinobacteria [28] . In the phylogenetic analysis of FAS genes , we included three fungal FAS genes and two stramenopile FAS genes , while in the phylogenetic analysis of MAS-family genes we added four animal FAS genes . Amino acid sequences were aligned using MAFFT [77] using the E-INS-i method . A maximum likelihood phylogeny was estimated using the PROTGAMMABLOSUM62 model in RAxML [78] with 1000 bootstrap replicates . FAS type 1a protein alignments were visualized using Jalview [79] and domains were annotated based on predictions from the Conserved Domain Database [80] . A custom blast database containing seven Rhodococcus & Mycobacterium genomes , including R . opacus B4 , R . opacus PD630 , R . jostii RHA1 , M . tuberculosis H37Rv , M . vanbaalenii PYR-1 , M . sp JLS & M . leprae Br4923 was generated . Protein sequences annotated as either AccA or AccD within Mycobacterium tuberculosis H37Rv were identified and accessions extracted into separate lists , whereas the WS/DGAT protein accessions were identified from the R . jostii RHA1 genome annotation . Each set of amino acid sequences were blasted against the database using BLASTP with an expect threshold of 1E−30 . Unique matches were identified and whole sequences extracted for alignment with MUSCLE v3 . 7 [81] , using a maximum of 24 iterations . Alignments were manually checked with ClustalX v2 . 0 [82] , at which point WS/DGAT sequences without any residues aligning to the proposed active site motif ( H[L/S/P]xxxDG ) [4] were rejected . The multiple sequence alignments were converted to phylip format for passing to ProtTest v2 . 4 [83] , which determines the best-fit substitution model and produces a phylogenetic tree with maximum likelihood estimation , using PhyML v3 . 0 [84] . Newick formatted trees were represented with iTOL [85] . GC-MS analyses were conducted on a TraceGC Ultra DSQ mass spectrometer ( Thermo Scientific ) equipped with an AT-5 ms column from Alltech ( 60 m×0 . 25 mm i . d . ×0 . 25 mm df ) . The injector and transfer line were maintained at 280°C while the ion source was set at 180°C in electron ionization ( EI ) mode . High purity helium was used as carrier gas at a flow rate of 1 mL/minute . The sample was injected onto the column in split mode and heated at 40°C for a minute . The GC oven temperature was increased to 280°C at 20°C per minute . The NMR spectra were recorded on a Varian Inova instrument , operating at 500 MHz for 1H and 125 MHz for 13C , equipped with a three channel , 5 mm , indirect detection probe , with z-axis gradients . The solvent was chloroform-d , and the temperature was 25°C . The chemical shifts for 1H were referenced to the residual solvent signal , 7 . 27 ppm on the tetramethylsilane scale . Proton spectra were acquired in 4 transients , with a 30° pulse , an acquisition time of 5 s and no relaxation delay . The intensity of the signals was referenced to the signal of the terminal methyl in the alkyl chain , at 0 . 88 ppm , 3H . 100 ml cultures of R . opacus PD630 and R . jostii RHA1 were grown in defined media [5] with substitution of 1% propionate/0 . 056% NH4SO4 for 1 week at 30°C with agitation followed by addition of 1 gram of propionate from sterile 20% stock solution . 50 mls of culture was collected after 2 weeks by pelleting and freeze dried prior to lipid extraction . 50 mg of dried cells were transesterified in a 2 ml volume of 50% CHCl3/42 . 5% methanol/7 . 5% H2SO4 for 2 . 5 hours at 100°C in sealed 16×125 ml glass tubes ( Kimble Glass Co . ) . Rhodococcus fatty acid methyl esters ( FAMEs ) preps were then concentrated under a stream of N2 , resuspended in dichloromethane ( DCM ) , and analyzed ( 10-mL injections ) using an Agilent 1050 High Performance Liquid Chromatography ( HPLC; Agilent Technologies , Inc , Wilmington , DE USA ) coupled to a Wakosil II RS-Prep C18 column ( 5 mm , 20 mm×250 mm; Wako Chemicals USA , Inc . , Richmond , VA USA ) and a Sedere Sedex 75 Evaporative Light Scattering Detector ( ELSD; SEDERE , Alfortville Cedex , France ) for determination of relevant fractionation range . The ELSD drift tube was set at 50°C , and the air nebulization pressure at 3 . 5 bar . The dual solvent system used is a linear gradient program based on the method published by Mansour [86] , with an extended run time , and substitutes ( DCM ) for chloroform . Starting with 98% acetonitrile ( MeCN ) and 2% DCM , the program ramps to 60% MeCN linearly by 100 minutes , holds at that ratio until 105 minutes , then ramps to 100% DCM by 110 minutes holding until 115 min , and returns to the starting conditions by 120 minutes . Methyl 12-methyltetradecanoate ( C15∶0 ) and Methyl 16-methylheptadecanoate ( C18∶0 ) branched FAMEs standards ( Sigma ) were used to determine the HPLC retention times of analytes of interest . Fractions were collected for each FAME species: 17∶1 eluted at 12 . 8 minutes , C 15∶0 eluted at 13 . 5 , C17∶0 eluted at 16 . 6 minutes for subsequent GC-EI-MS and 1H-NMR analysis . TLC purification of TAGs was performed as previously described for single solvent system [5] followed by scraping of the TAG species that were detected by water staining . Scraped TAGs were extracted with 1∶1 ( vol/vol ) chloroform methanol for 1 Hr prior to filtration with 0 . 2 µM PTFE membrane ( VWR International ) . Extracts were dried under nitrogen then subjected to transesterification and analyzed by ( GC-FID ) [5] . R . opacus PD630 TAGs cycle genes were identified through metabolic reconstruction and literature-based curation . The genes assigned to biochemical reactions within the TAGs cycle were analyzed to identify all significant gene pairs ( BLASTP , E< = 1e-5 ) followed by Markov clustering as implemented by TribeMCL [42] . Candidate gene families were then aligned using MAFFT [77] and manually curated in Jalview 2 . 6 . 1 [79] using neighbor joining trees and curation of alignments . Because we were classifying by EC number which is a broad classification of function and not necessarily homology , a strict cutoff for percent overlap or percent similarity was not used to retain as many members of particular EC category as annotated in KEGG ( http://www . genome . jp/kegg/ ) . A full list of curated EC # related Gene families implicated at each biochemical step of the TAGs cycle are represented with 11 other bacterial species in Table S4 . Multiexperiment viewer [87] was used to hierarchically cluster the species in the TAGs cycle and visualize in a heat map . Clustering was pearson correlated with average linkage . The threshold for color-saturation was set to 15 genes per reaction category . Chemical compounds capable of supporting growth as a sole source of carbon were identified using a tetrazolium-based growth assay developed by Biolog Incorporated; wherein growth of cells and aerobic respiration were measured by Dye D ( Biolog , Inc ) reduction resulting in purple color and turbidity read at 590 nm absorbance in Fluorostar plate reader . Innoculation cultures were grown for 1 day on LB agar plates at 30°C from which 3 . 8E−3 ODu/well were transferred from 0 . 5 ml water resuspensions after measurement in nanodrop ( Thermo Scientific ) at 600 nm . During growth , plates were wrapped in aluminum foil and measured at 44 , 72 , and 96 hours . The measured A590 values were normalized by first subtracting values from uninoculated plates . The average of normalized values were background subtracted with average values of inoculated wells that lacked a carbon source ( negative control ) . The background subtracted growth values were separated into chemical categories by filtering compound classifications in Excel . The normalized growth values for each compound were clustered using Cluster 3 . 0 [88] with centered correlation . Heat maps were generated in Java TreeView 1 . 1 . 3 [88] . Growth substrates were identified at 96 hours incubation for compounds that had A590 values >0 . 2 . | Biofuels research is focused on understanding the energy-related metabolic capabilities of a broad range of biological species . To this end we sequenced the genome of Rhodococcus opacus PD630 , a bacterium that accumulates close to 80% of its cellular dry weight in oil , a rare trait in the prokaryotic and eukaryotic kingdoms . R . opacus PD630 has a large 9 . 27 Mb genome that contains many homologous genes dedicated to lipid metabolism . The number and novelty of these predicted genes presents a challenge to the complete and accurate metabolic reconstruction of this species' metabolism based only on genome sequence . To refine our sequence-based metabolic reconstruction , we developed a multidisciplinary approach that included integrating the identification of abundant yet uncommon straight-chain odd-carbon lipid biosynthesis and the results of a catabolic screen for growth substrates . Comparative analysis of the R . opacus PD630 genome sequence with those of a group of related species provided a view into how this bacterium became such a remarkable TAGs producer and led to the identification of a set of biofuels target genes for this group of bacteria . Our synthesis of genome sequence and phenotypic information supports a model for the genetic basis for prokaryotic oleaginy and provides key insights for the engineering of next-generation biofuels with genes that are conserved in both prokaryotic and eukaryotic kingdoms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biotechnology",
"biochemistry",
"lipids",
"enzymes",
"genetics",
"biology",
"genomics",
"microbiology",
"metabolism",
"genetics",
"and",
"genomics"
] | 2011 | Comparative and Functional Genomics of Rhodococcus opacus PD630 for Biofuels Development |
Amyloid-like inclusions have been associated with Huntington's disease ( HD ) , which is caused by expanded polyglutamine repeats in the Huntingtin protein . HD patients exhibit a high incidence of cardiovascular events , presumably as a result of accumulation of toxic amyloid-like inclusions . We have generated a Drosophila model of cardiac amyloidosis that exhibits accumulation of PolyQ aggregates and oxidative stress in myocardial cells , upon heart-specific expression of Huntingtin protein fragments ( Htt-PolyQ ) with disease-causing poly-glutamine repeats ( PolyQ-46 , PolyQ-72 , and PolyQ-102 ) . Cardiac expression of GFP-tagged Htt-PolyQs resulted in PolyQ length-dependent functional defects that included increased incidence of arrhythmias and extreme cardiac dilation , accompanied by a significant decrease in contractility . Structural and ultrastructural analysis of the myocardial cells revealed reduced myofibrillar content , myofibrillar disorganization , mitochondrial defects and the presence of PolyQ-GFP positive aggregates . Cardiac-specific expression of disease causing Poly-Q also shortens lifespan of flies dramatically . To further confirm the involvement of oxidative stress or protein unfolding and to understand the mechanism of PolyQ induced cardiomyopathy , we co-expressed expanded PolyQ-72 with the antioxidant superoxide dismutase ( SOD ) or the myosin chaperone UNC-45 . Co-expression of SOD suppressed PolyQ-72 induced mitochondrial defects and partially suppressed aggregation as well as myofibrillar disorganization . However , co-expression of UNC-45 dramatically suppressed PolyQ-72 induced aggregation and partially suppressed myofibrillar disorganization . Moreover , co-expression of both UNC-45 and SOD more efficiently suppressed GFP-positive aggregates , myofibrillar disorganization and physiological cardiac defects induced by PolyQ-72 than did either treatment alone . Our results demonstrate that mutant-PolyQ induces aggregates , disrupts the sarcomeric organization of contractile proteins , leads to mitochondrial dysfunction and increases oxidative stress in cardiomyocytes leading to abnormal cardiac function . We conclude that modulation of both protein unfolding and oxidative stress pathways in the Drosophila heart model can ameliorate the detrimental PolyQ effects , thus providing unique insights into the genetic mechanisms underlying amyloid-induced cardiac failure in HD patients .
Amyloidosis constitutes a large group of diseases characterized by the misfolding of proteins and the accumulation of protein aggregates in different tissues [1]–[3] . Huntington's disease ( HD ) is an inherited neurodegenerative disorder caused by mutations in the Huntingtin ( HTT ) protein which result in expanded Poly-glutamine ( PolyQ , CAGn ) repeats that cause aggregation-prone amyloidosis [4]–[7] . The molecular mechanism that leads to HD is not fully understood and presently no effective treatment exists [4] , [8]–[10] . It has been well established that the length of the PolyQ repeat is important in the progression of disease [4] , [5] . HTT with 6–35 PolyQ repeats does not cause HD . However , HTT with more than 40 PolyQ ( CAG40 ) repeats results in HD [4] , [5] , [11] . In general HD is primarily considered as an aggregation-based disease; however , some studies have shown that disease-causing PolyQ repeats in HTT make it prone to misfolding and aggregation [4]–[7] , [12]–[15] . HTT is expressed in several tissues in addition to the brain , including heart and skeletal muscles [11] , [16]–[18] and is known to be involved in protein trafficking , vesicle transport and transcriptional events [4] , [11] . HD is also associated with skeletal muscle atrophy [11] , [19] and multiple epidemiological studies have shown that cardiovascular diseases and cardiac failure are the second leading cause of mortality in HD patients [8] , [18] , [20] . Cardiac failure is implicated as the cause of death in over 30% of HD patients , compared to 2% of the age-matched non–HD patients [8]–[11] , [18] , [20] . Although , the mechanism whereby mutant HTT causes muscle atrophy and cardiac defects is not known , it is possible that an increase in protein misfolding and the consequent high energy burden in cardiac cells play roles [8] , [10] , [11] , [16] , [18] . In support of this , recent evidence demonstrates nuclear and cytoplasmic PolyQ aggregates in non-CNS tissue [11] , [16] , [18] . Furthermore , neuronal expression of mutant HTT protein with expanded PolyQ or cardiac-specific expression of only the PolyQ pre-amyloid oligomers in mice leads to cardiac defects [10] , [21] , [22] . Furthermore , expression of mutant PolyQ-81 in mice and in rat neonatal cardiomyocytes results in amyloid as well as PolyQ-positive aggregates in the cytoplasm and over-expression of a chaperone αB-crystallin reduces PolyQ-induced aggresomes [21] , [23] . Moreover , reduction of aggresomes upon over-expression αB-crystallin results in higher levels of amyloid oligomer and enhances toxicity [23] . Despite the availability of cell and mouse models to examine/investigate PolyQ expression in the heart , little is known about the mechanism that leads to cardiac dysfunction . We have previously established the Drosophila heart as a useful model system to generate insights into the genetic basis of heart development and to elucidate the genetic interactions underlying heart physiology and age-dependent deterioration [24]–[27] . Recently , we showed in this genetic model that knock-down of the chaperone UNC-45 significantly reduced myosin expression and led to severe cardiac dilation [28] . Protein folding and oxidative stress pathways have previously been implicated in the development/pathology of HD [5] , [6] , [10] , [18] , [19] , however , their involvement with cardiac phenotypes has not been explored . In the current study , we manipulate these two pathways to attempt to suppress the cardiac defects induced by mutant HD-associated PolyQ repeat lengths . These defects include the accumulation of GFP-positive aggregates , mitochondrial defects , oxidative stress and both functional and morphological cardiac abnormalities . We also found that cardiac over-expression ( OE ) of the chaperone UNC-45 suppressed amyloid deposition and ameliorated the heart function defects to some extent . In addition , OE of SOD as well as feeding with the dietary antioxidant resveratrol also partially suppressed the amyloid-induced cardiac dysfunction , whereas hydrogen peroxide feeding aggravated the heart defects . Our data suggest that the protein folding and ROS pathways interact in mediating the effects of mutant HTT as a near complete reversal of the cardiac defects was achieved when both pathways were modulated simultaneously in flies expressing disease-causing PolyQ repeats . Thus , our data show a deleterious effect of mutant PolyQ aggregates on cardiac function and indicate that these effects are the result of protein misfolding and/or concomitant oxidative stress .
To evaluate cardiac function following PolyQ-induced cardiomyopathy , we obtained Drosophila transgenic lines [5] expressing enhanced-GFP-tagged control or mutant Htt fragments ( UAS-Httex1-QneGFP ) with different PolyQ repeat lengths ( Q25 , Q46 , Q72 , and Q103 ) . For simplicity , Httex1-Q25-eGFP , Httex1-Q46-eGFP , Httex1-Q72-eGFP and Httex1-Q102-eGFP are referred to as PolyQ-25 , PolyQ-46 , PolyQ-72 and PolyQ-102 , respectively . Using the heart-specific driver Hand , we observed severe cardiac defects and/or extreme dilation upon expression of disease-causing PolyQ-46 , PolyQ-72 and PolyQ-102 , whereas the shorter PolyQ-25 had no measurable effect . Figure 1A shows contracted hearts from 1-week old flies with cardiac-specific expression of PolyQ-72 and an age-matched control expressing PolyQ-25 . The heart tube expressing PolyQ-72 is clearly much less contracted during systole than is the PolyQ-25 control . Heart wall diameters during both systole and diastole are indicated by double headed arrows in the M-mode records produced from high speed movies; the line of pixels used to produce these records is indicated by the blue line in the still image . The ability of the PolyQ-72 heart to contract during systole is much reduced compared to the control heart . A comparison of representative M-mode records for 3-week old flies from each of the transgenic lines and wild type controls is shown in Figure 1B and demonstrates a progressive increase in cardiac arrhythmia with increased length of PolyQ . A comparison of PolyQ-25 , PolyQ-46 and PolyQ-72 heart contractility , dilation and arrhythmia is also shown in supplementary movie S1 . Hearts expressing the longer repeats , PolyQ-72 and PolyQ-102 , also showed significant cardiac dilation in addition to arrhythmia ( Figure 1B ) . Hearts from flies with cardiac-specific expression of mutant PolyQ also exhibited additional functional and morphological defects including floppy , non-contractile ostia and one or more non-contractile myocardial cells , primarily in the conical chamber ( CC ) and the adjacent chamber ( Figure 1C ) . In PolyQ-72 and PolyQ-102 expressing hearts there were frequent asystolic periods as well as hearts that were completely unable to beat . The incidence of these qualitative defects is quantified in Figure 1C . In addition to these heart function defects , cardiac specific expression of the two longer PolyQ proteins ( PolyQ-72 and PolyQ-102 ) significantly shortens the lifespan of flies ( Fig . 1 D and Table S1 ) . Quantification of functional parameters from the high-speed movies demonstrated a significant increase in diameters during both systole ( Figure 2A ) and diastole ( Figure 2B ) compared to hearts expressing PolyQ-25 and wild-type controls; this dilation was more severe for the longer PolyQ repeats ( PolyQ-72 and PolyQ-102 ) . The observed cardiac dilation was accompanied by a significant reduction in heart contractility , measured as a decreased fractional shortening ( % FS ) in the PolyQ-46 , PolyQ-72 and PolyQ-102 expressing hearts ( Figure 2C ) . Long PolyQ repeats induced significant increases in both the systolic and diastolic intervals and this effect again appeared to be dependent on the “dose” of PolyQ ( Figure 2D , 2E ) . The incidence in cardiac arrhythmias was also quantified ( arrhythmia index ) [27]–[29] , and showed a PolyQ dose-dependent increase ( Figure 2F , see also Figure 1B ) . Cardiac specific expression of PolyQ-25 did not significantly alter any of the measured cardiac function parameters when compared to hearts from control flies that lacked any PolyQ expression ( Hand-Gal4/+ ) . Similar alterations in cardiac physiological parameters were observed in response to disease-causing PolyQ expression ( PolyQ-46 , PolyQ-72 and PolyQ-102 ) in younger , 1 week old flies ( Figure S1A to S1F ) . Taken together our data indicate that all the cardiac defects we observe are PolyQ-length dependent and suggest that the PolyQ-72 repeat length is sufficient to exert the maximal deleterious effect on these hearts . To explore whether the cardiac physiological dysfunction in response to cardiac-specific expression of PolyQ is the result of amyloid accumulation we used a Green Fluorescent Protein ( GFP ) tag to visualize PolyQ proteins and amyloid deposits and phalloidin to detect F-actin , revealing the myofibrillar organization within myocardial cells . Hearts expressing non-disease causing PolyQ-25 ( control ) show densely packed actin-containing myofibrils arranged in a circumferential pattern within the cardiomyocytes ( Figure 3A ) and GFP-tagged PolyQ was found to be distributed homogeneously throughout the cytoplasm ( Figure 3B ) . In contrast , expression of disease-causing PolyQ-72 resulted in noticeably reduced myofibrillar content and in severe myofibrillar disorganization ( Figure 3C , dashed box ) . Interestingly , we also observed the presence of many GFP-positive aggregates of various sizes throughout the cardiomyocytes ( Figure 3D ) . We used a well-documented filter trap assay [30]–[32] to confirm this increase in aggregate formation upon expression of mutant PolyQ-72 . Although expression of both control PolyQ-25 and mutant PolyQ-72 protein was virtually the same relative to histone H2B ( Fig . 3G , top and middle panels ) , the heart-specific expression of mutant PolyQ-72 resulted in a significant increase in GFP-positive aggregates compared to control hearts ( Figure 3G , bottom panel ) . We also used antibody against muscle myosin to explore the effect of mutant PolyQ on the myosin organization within the myofibrils . Myosin organization in myofibrils from control hearts exhibits a similar circumferential arrangement as for F-actin ( thick arrow in Figure 3E ) ; however , the myosin pattern appears significantly aberrant upon expression of PolyQ-72 ( Figure 3F ) . In fact the majority of the staining visible in Fig . 3F is due to myosin in the non-cardiac longitudinal muscle fibers that run ventrally along the cardiac tube ( thin arrows in Figure 3E and 3F ) . Some disorganized myosin-containing myofibrils are still seen upon expression of mutant PolyQ ( Figure 3F , dashed boxes ) . This phenotype is reminiscent of knock-down of the myosin-specific chaperone UNC-45 [28] , suggesting that long PolyQ aggregates might interfere with chaperone function . These data indicate that the presence of toxic aggregates leads to a reduction in cardiac myosin-actin content with disorganized myofibrils . The ultrastructure of Drosophila cardiac muscle has been described previously in detail by Lehmacher et al . [33] . Transmission electron micrographs of transverse sections of 4 week-old hearts from PolyQ-25 control flies reveal a layer of contractile cardiomyocytes and a supporting layer of non-cardiac ventral-longitudinal fibers ( VL , Figure 4A ) . Myocardial cells from PQ-25 controls contain mitochondria with densely packed cristae ( 4A , MT ) that can be seen adjacent to the myofibrils ( Figure 4A , MF ) . In contrast , micrographs from PolyQ-46 hearts show evidence of myofibrillar degeneration ( 4B , arrow ) and severe mitochondrial fragmentation ( 4B , B' , asterisks ) . Such mitochondrial fragmentation and alterations in cristae structure have previously been linked to increased apoptotic activity in primary striatal cultures from YAC128 HD transgenic mice as well as in neuronal expression of mutant HTT protein with expanded PolyQ in a mouse heart model [10] , [34] . These defects are even more severe in hearts expressing longer forms of PolyQ-72 with near complete loss of myofibrillar architecture . We also looked for autophagosome/lysosome structures and observed a significant amount of LysoTracker positive punctae upon expression of mutant PolyQ-72 ( Figure S2 E ) . Significantly , most of the staining is co-localized with PolyQ-GFP punctae ( Figure S2 F ) . In contrast , almost no GFP- or LysoTracker-positive punctae are seen upon expression of non-disease causing PolyQ-25 ( Figure S2 A–C ) suggesting a direct link between expression of mutant PolyQ and activation of the autophagy pathway . The mitochondrial defects , myofibrillar disorganization and cardiac function abnormalities observed upon expression of mutant PolyQ could arise from aggregate-induced oxidative stress . We used dihydroethidium ( DHE ) to evaluate the role of reactive oxygen species ( ROS ) production in mediating the effects of disease-causing PolyQ . Cardiac expression of PolyQ-46 and PolyQ-72 resulted in 2- and 5-fold increases in DHE staining respectively ( Figure 5E and 5H ) compared to age-matched PolyQ-25 control hearts ( Figure 5B ) . Furthermore , a number of the mutant-PolyQ induced GFP-aggregates colocalize with areas of strong DHE staining ( arrows in Figure 5D and E; and Figure 5G and H ) while expression of PolyQ-25 shows almost no GFP-positive punctae or DHE staining . These results confirm an association between PolyQ-induced aggregates and oxidative stress . To explore whether induction of oxidative-stress could aggravate the PolyQ phenotype , flies expressing PolyQ-25 and PolyQ-46 in cardiac tissue were fed H2O2 for 3-weeks during adulthood . PolyQ-46 expressing hearts in the presence of oxidant showed significantly increased cardiac dilation . However , no such enlargements of cardiac diameters were seen in PolyQ-25 expressing hearts in the presence of oxidant ( Figure 6A–6B ) . Although fractional shortening was decreased and cardiac arrhythmias were increased in PolyQ-46 expressing hearts by H2O2 feeding , both were affected to a similar extent as were control PolyQ-25 expressing hearts ( Fig . 6C–F ) . Feeding oxidant to non-PolyQ expressing wild-type flies ( Hand-Gal4/+ ) had similar minimal effects on cardiac parameters as for the PolyQ-25 controls ( Figure S3A–S3F ) . The increased incidence of arrhythmia and reduced contractility of hearts expressing PolyQ-72 was also aggravated in the presence of oxidant ( Figure S4A , S4B ) . While PolyQ-46 hearts in the absence of oxidant do exhibit sparsely distributed amyloid-aggregates , the presence of H2O2 results in an increase ( 30% ) in the density of aggregates ( green GFP-punctae , Figure 6G , 6I ) . Furthermore , treatment with oxidant also resulted in more myofibrillar disorganization and loss compared to age-matched PolyQ-46 without oxidant ( compare Figures 6G and 6H ) . Interestingly , muscle fiber organization remained virtually unchanged in hearts from wild-type controls ( Hand/+ ) and PolyQ-25 expressing controls when fed H2O2 ( compare Figure S3G with S3H and S3I with S3J ) . These data support the idea that oxidative stress enhances the accumulation of mutant PolyQ aggregates ( GFP-punctae ) , myofibrillar disorganization and loss of actin-containing myofibrils and that these aggregates contribute to the cardiac dilation and heart function defects we observe . Additionally , our results show that oxidative stress exerts differential effects on heart function and structure depending upon the presence or absence of PolyQ aggregates . It is also possible that treatments with H2O2 may lead to secondary stress ( such as initiation of the heat shock program ) . Assuming secondary stresses in response to H2O2 were similar between these two groups of flies , the only explanation for these more severe defects in the PolyQ-46 expressing group is an interaction between ROS signaling and mutant PolyQ expression . The H2O2 treatments may also affect other tissues such as neuronal tissue; therefore , it will be interesting to determine if control and mutant PolyQ affect cardiac and neuronal tissue under oxidative stress in a similar manner . Because mutant PolyQ expressing hearts exhibited oxidative stress , mitochondrial defects as well as aggravated cardiac defects in response to hydrogen peroxide feeding , we tested whether over-expression of superoxide dismutase ( SOD ) could rescue the PolyQ-induced cardiomyopathy . We over-expressed SOD-1 or SOD-2 along with PolyQ-72 in fly hearts and examined the effect on cardiac function ( Figure 7A–7F ) . In hearts from 3-week old flies over-expressing SOD-1 along with PolyQ-72 the cardiac dilation was significantly reduced compared to hearts expressing PolyQ-72 alone and was nearly the same as for wild-type Hand/+ hearts ( Figure 7A , 7B ) . Cardiac contractility was also improved in SOD overexpressing hearts ( Figure 7C ) . Both diastolic and systolic intervals were significantly lower ( Figure 7D , 7E ) and the incidence of arrhythmias was significantly reduced to nearly wild-type levels ( Figure 7F ) in PolyQ-72 hearts overexpressing SOD compared to hearts expressing PolyQ-72 alone . Analysis of the myofibrillar organization also showed a partial rescue of PolyQ-72–induced myofibril disarray upon SOD-1 expression and both the size and the density of mutant PolyQ aggregates were markedly reduced ( Figure 7G , 7H , 7K , 7L ) . Co-expression of SOD-2 and PolyQ-72 produced a similar suppression of the PolyQ-72-induced cardiomyopathy ( Figure S5A to S5F ) . No cardiac defects were seen when SOD-1 or SOD-2 was co-overexpressed with the non-disease causing PolyQ-25 ( not shown ) . Over-expression of SOD has been shown to suppress the cardiac defects associated with knock-down of mitochondrial assembly regulatory factor ( MARF ) [35] . We attempted to rescue Poly-Q associated cardiac abnormalities with transgenic expression of MARF . However , over-expression of MARF does not result in any significant suppression of cardiac defects associated with mutant PolyQ-72 ( Figure S5A to S5F ) . In fact , it has been shown that over-expression of mitofusion 2 promotes cardiomyocyte apoptosis via a mitochondrial death pathway in cultured mammalian cardiomyocytes [36] . We also tested the effects of feeding flies the antioxidant resveratrol . As for SOD-over-expression in PolyQ-72 hearts , resveratrol treatment reduced the dilated systolic and diastolic diameters , the diastolic and systolic intervals , and the arrhythmias ( Figure 7A , 7B , 7F ) . It increased contractility and significantly reduced the PolyQ-72-induced increase in aggregate size and density ( Figure 7C , 7I , 7K , 7L ) . These data indicate that the cardiac defects seen in response to expression of disease-causing PolyQ can be partially suppressed upon over-expression of antioxidant agents , such as SOD or resveratrol . Since over-expression of SOD or feeding with the antioxidant resveratrol rescued mutant PolyQ-induced cardiac defects , we examined whether over-expression of SOD could rescue the mitochondrial defects associated with expression of mutant PolyQ-46 . In contrast to 4 week-old PolyQ-46 hearts , which contained areas of myofibrillar degeneration ( Figure 8A , arrow ) , the majority of myofibrils in PolyQ-46 hearts overexpressing SOD were intact ( Figure 8B , arrow ) . Hearts expressing PolyQ-46 contained fragmented mitochondria ( Figure 8A , A' asterisks ) , while PolyQ-46 hearts overexpressing SOD contained normally shaped mitochondria with densely packed cristae ( Figure 8 B , B' ) , similar to PolyQ-25 controls ( Figure 4 ) . It is known that accumulation of mutant PolyQ interferes with the protein folding machinery in neurons and it has been predicted that PolyQ has the same effect in the heart [37] , [38] . Therefore , we reasoned that over-expression of the chaperone UNC-45 , which may enhance proper protein folding , might improve cardiac function in hearts compromised by disease-causing PolyQ expression . To address this we over-expressed UNC-45 along with PolyQ-72 in the fly heart . Indeed , transgenic over-expression of UNC-45 completely suppressed mutant PolyQ-72 induced cardiac dilation ( compare to Hand/+ wild-type controls , Figure 7A , 7B ) . UNC-45 over-expression improved contractility ( Figure 7C ) and the regularity of the heart rhythm ( Figure 7F ) . Most interestingly , over-expression of UNC-45 in the presence of PolyQ-72 significantly reduced the density of mutant PolyQ aggregates compared to hearts expressing PolyQ-72 alone ( Figure 7G , 7J , 7L , 7M , compare Figure 7G with 7J and compare lane 1 and 4 in Figure 7M ) . However , the mean aggregate size was not altered ( Figure 7K ) . Additionally , hearts overexpressing UNC-45 showed a slightly more wild-type organization of actin-containing myofibrils ( Figure 7J ) . Over-expression of UNC-45 in the presence of PolyQ-72 also resulted in a restoration some of the normal structure and content of the myosin-containing myofibrillar network ( Figure S6B ) , which were nearly absent in the cardiomyocytes upon expression of PolyQ-72 alone ( Figure S6A ) . These results suggest that disease-causing PolyQ may act by interfering with chaperone function , which is required for proper myosin folding/accumulation [28] , [39] . In contrast , over-expression of UNC-45 alone or with PolyQ-25 resulted in only minor changes in functional cardiac parameters ( Figure S7A to S7F ) . These results suggest that protein unfolding may play a role in mediating PolyQ-induced cardiomyopathy . We examined whether improving protein folding and oxidative stress pathways might interact to suppress PolyQ-induced cardiac defects . To test this , we co-expressed UNC-45 and SOD-1 in conjunction with PolyQ-72 . Co-expression of UNC-45 and SOD-1 restored cardiac contractility ( Figure 9C ) and suppressed cardiac dilation ( Figure 9A , 9B ) , as well as cardiac arrhythmias ( Figure 9F , and Movie S2 ) . Over-expression of UNC-45 and SOD-1 also nearly completely suppressed the formation of GFP-positive aggregates that were dramatically induced by expression of PolyQ-72 ( Figure 9H , 9K , 9P , 9Q ) . Furthermore , PolyQ-72 hearts expressing both UNC-45 and SOD-1 exhibited more organized actin-containing myofibrillar structures ( Figure 9G vs . 9J ) . Together this suggests that mutant PolyQ aggregates induced by abnormal protein folding and increased oxidative stress are linked to cardiac physiological and structural defects . To further confirm an association of both protein-unfolding and oxidative stress pathways with PolyQ-induced cardiomyopathy , we co-expressed PolyQ-72 and UNC-45 in the presence of the antioxidant resveratrol ( Figure 9 ) . Over-expression of UNC-45 in the presence of resveratrol almost completely suppressed PolyQ-induced cardiac dilation , cardiac arrhythmia and amyloid aggregation ( Figure 9A , 9B , 9F , 9N , 9P , 9Q ) . Furthermore , the cardiac contractility was improved as was myofibrillar organization compared to PolyQ-72 expression alone ( Figure 9C , 9M ) . Over-expression of UNC-45 and SOD or over-expression of UNC-45 in the presence of resveratrol also reduced the diastolic and systolic intervals to wild-type levels compared to age-matched PolyQ-72 ( Figure 9D , 9E ) . A summary of cardiac parameters for wild-type controls as well as hearts expressing PolyQ-25 and PolyQ-72 ( with or without antioxidant treatment ) is shown in Figure S8A . Finally , mutant PolyQ-induced lifespan reduction was rescued by transgenic over-expression of SOD or co-expression of SOD and UNC-45 but not with UNC-45 over-expression ( Figure S8B and Table S1 ) . Overall , the genetic interactions that we have identified in this study demonstrate that the protein-misfolding and oxidative stress pathways induced by accumulation of HD-causing PolyQ aggregates are linked and associated with cardiac dysfunction .
Huntingtin protein is expressed in many tissues including the heart and epidemiological studies suggest that HD patients have a higher susceptibility to cardiac failure compared to age-matched controls without HD [8] , [9] , [11] , [16] , [17] . However , the cellular mechanisms underlying the cardiac dysfunction in HD have yet to be studied in the heart . Using the genetically tractable model system Drosophila , we now show a direct correlation between the levels of amyloid accumulation , overall ROS production and the severity of cardiac dysfunction . Cardiac-specific expression of disease-causing Htt-PolyQ ( PolyQ-46 , PolyQ-72 and PolyQ-102 ) all elicited cardiac dysfunction compared to hearts expressing the non-disease-causing PolyQ-25 . In addition the qualitative as well as quantitative defects that we observed in response to PolyQ expression were dose-dependent . Since mutant Htt-PolyQ protein was expressed specifically in the heart , it is unlikely that our observations reflect a neuronal contribution to these cardiac defects . Our data suggest that the increased risk of cardiac disease in HD patients is possibly due to cardiac amyloid accumulation , mitochondrial defects as well as oxidative stress and that the severity of disease depends upon the length of the PolyQ repeat ( Figures 3–5 ) . Our data also demonstrate that the likely cause of the observed functional defects is the severe myofibrillar disorganization and reduced myosin and actin content in myocardial cells resulting from cardiac-specific expression of disease causing PolyQ ( Figure 3C and 3F ) . Recently we showed that the chaperone UNC-45 is required for preserving myosin accumulation/folding in Drosophila cardiomyocytes , as its reduction leads to severe disorganization of myosin-actin containing myofibrils and thus sarcomeres [28] . The current results extend this observation and are the first demonstration of a role for UNC-45 in amyloidosis-induced cardiac defects . In support of this hypothesis , it has previously been shown that nuclear or cytoplasmic aggregates ( inclusion bodies ) of polyglutamine proteins contain chaperones involved in protein folding [11] , [37] . Furthermore , and consistent with our results , over-expression of the chaperone αB-crystallin reduces PolyQ-induced aggregation in rat neonatal cardiomyocytes; however , over-expression of αB-crystallin enhances amyloid oligomer formation and toxicity [23] . In the present study co-over-expression of UNC-45 with disease-causing PolyQ-72 dramatically reduced amyloid aggregate density ( Figure 7L and 7M ) and ameliorated cardiac dysfunction by decreasing the incidence of cardiac arrhythmia , suppressing the mutant-Htt-induced cardiac dilation ( Figure 7A , 7B ) and improving cardiac contractility to a dramatic extent ( Figure 7C ) . Importantly , over-expression of UNC-45 in the presence of PolyQ-72 restored myosin-containing myofibrils ( Figure S6 ) , suggesting that one effect of amyloid aggregation is to interfere with proper folding of muscle myosin in cardiomyocytes . The fact that UNC-45 over-expression did not completely suppress the mutant Htt-PolyQ-induced cardiac physiological defects and lifespan reduction is consistent with the idea that amyloid accumulation affects additional cellular pathways that result in cardiac abnormalities . As reported for αB-crystallin , suppression of aggregates is not sufficient to reduce toxicity [23] and this possibility may also exist in the case of UNC-45 . It is also possible that the overall high level of oxidative stress produced by mutant PolyQ is the main determinant for lethality . Expression of mutant PolyQ leads to mitochondrial defects due to increased oxidative stress [4]–[10] , [40]–[42] . Several neuronal studies have shown that expression of mutant polyQ affects SOD expression [43]–[46] . Manipulation of SOD seems to be directly correlated with levels of oxidative stress in several neurodegenerative diseases [4]–[10] , [40]–[42] . Additionally , SOD over-expression reduces diabetic cardiomyopathy and some forms of neurodegeneration by reducing oxidative stress [47] , [48] . However , neither UNC-45 nor SOD has been shown previously to suppress the PolyQ-induced phenotypes in either neuronal or cardiac animal disease models . Our data also support a role for oxidative stress pathways in amyloid-induced cardiac dysfunction and lethality . Treatment with oxidants aggravated the moderate effects of PolyQ-46 on heart function , causing an increase in amyloid aggregate density and more severe cardiac defects ( Figure 6 ) . This suggests a possibly causal relationship between oxidative stress , the formation of aggregates and cardiac dysfunction . Furthermore , our ultrastructural analysis clearly shows mutant PolyQ-induced mitochondrial defects , while DHE staining indicates that excess ROS production occurs upon expression of mutant PolyQ ( Figures 4 and 5 ) . Interestingly , some of the PolyQ aggregates co-localize with concentrated DHE staining ( Figure 5 ) . Significantly , we were able to reduce the size and density of mutant PolyQ-aggregates as well as the severity of the PolyQ-72-induced cardiac defects by over-expression of SOD or by feeding the anti-oxidant resveratrol ( Figure 7 ) . This is consistent with findings that resveratrol provides protection in neuronal models of Huntington's disease [49]–[54] . Interestingly , the anti-oxidant resveratrol has been shown to affect expression of anti-oxidative enzymes , including enhanced expression of SOD-1 [49]–[54] . Expression of mutant PolyQ may both induce oxidative stress and interfere with protein folding pathways [4] , [21] , [40]–[42] . A study using cultured mouse neurons showed that oxidative stress increases PolyQ aggregation and that over-expression of SOD1 in conjunction with the chaperone HSP-70/HSP-40 could suppress Htt-polyQ-induced aggregation and toxicity [42] . However , simultaneous manipulation of both of these genetic pathways has not previously been attempted in vivo . In addition to neurons , expression of the mutated Htt protein or expression of pre-amyloid oligomers cause cardiac defects by affecting several pathways including oxidative stress , mitochondrial abnormalities , presence of protein aggregates and increased autophagosomal content [10] , [21] , [30] , [31] , [34] . However , no attempt had thus far been made to suppress PolyQ-induced cardiac defects , a crucial step for understanding the mechanistic basis of disease progression and amelioration . Indeed , in our in vivo cardiac model , co-expression of UNC-45 and SOD-1 or expression of UNC-45 in the presence of resveratrol had a tendency to suppress the PolyQ-72-induced amyloid aggregation and concomitant cardiac dilation more efficiently than either treatment alone ( Figures 7 and 9 ) . Thus , our results suggest that suppression of both protein aggregates and ROS may be required for the amelioration of PolyQ-induced cardiomyopathy . As HD is primarily a neurological disease , the effect of such suppression is worth exploring in neural tissues . In addition to interfering with protein folding pathways , expression of mutant PolyQ may lead to myofibril loss by directly interacting with muscle proteins . Previous studies have suggested that mutant PolyQ may bind directly to contractile proteins and disturb their function [55] , [56] . Integrity of contractile proteins is also required for maintaining mitochondrial organization and cardiomyocyte function [35] , [57]–[60] . Additionally , expression of aggregation-prone mutant PolyQ may induce oxidative stress due to mitochondrial damage in the cardiomyocytes , which are heavily dependent on mitochondrial function and are vulnerable to oxidative stress [60]–[62] . For example , knockdown of SOD results in mitochondrial defects and severe dilated cardiomyopathy phenotype in a mouse model [63] . Our results do show a dramatic increase in overall ROS levels in mutant PolyQ expressing hearts . Moreover , the GFP-positive PolyQ aggregates co-localize with areas of strong DHE staining and the observation that antioxidant treatments partially rescue the cardiac defects further support this hypothesis . Overall , accumulation of amyloid in the cardiomyocytes can induce mechanical deficits by affecting the integrity of contractile proteins as well as mitochondria and lead to cardiomyocyte death , possibly through activation of autophagy . Consistent with our finding , a similar mechanism has been proposed for cardiomyopathy associated with amyloid producing mutant αB-crystallin [58]–[60] , [64] . Both mutant αB-crystallin and mutant PolyQ caused aggregate formation in cardiomyocytes suggesting a common mechanism for underlying cardiomyocyte degeneration [21] , [23] , [58]–[60] , [64] . It is unclear at this point whether the presence of toxic aggregates in cardiomyocytes is directly interfering with mitochondrial organization leading to cardiac defects or whether oxidative stress produced by mutant PolyQ leads to mitochondrial dysfunction that triggers cardiomyocyte dysfunction . A full understanding of all the molecular details involved in mutant PolyQ induced cardiomyopathy will require additional study but we have now identified some of the key players and interactions in vivo . Although dense granular deposits , immunoreactive to an anti-Huntingtin antibody , have been found in muscle tissue from an HD patient , no such study has been performed on HD heart biopsy samples [65] . Thus , our study suggests that it would be useful to look for accumulation of amyloid protein in the hearts of HD patients , especially those with heart disease . Delineating how these aggregates might be toxic to cells will be critical not only for an understanding of PolyQ-induced cardiomyopathy but also for gaining insights into aggregation-based neural degeneration . The Drosophila heart model provides a genetically tractable system whereby these interactions can be examined in the context of a functioning organ . Indeed , elucidating the genetics underlying PolyQ-induced cardiomyopathy should also have an impact on our understanding of other cardiac diseases associated with oxidative stress , mitochondrial dysfunction , the unfolded protein response and proteostasis in general .
To evaluate cardiac function following PolyQ-induced cardiomyopathy , we obtained Drosophila transgenic lines expressing enhanced-GFP-tagged mutant Htt fragments ( UAS-Httex1-QneGFP ) with different PolyQ lengths ( Q25 , Q46 , Q72 , and Q102 ) [5] . Heart-specific expression was achieved with a UAS-Gal4 system using the Hand driver [66] , [67] , crossed to the different UAS-Httex-GFP lines ( Httex1-Q25-eGFP , Httex1-Q46-eGFP , Httex1-Q72-eGFP and Httex1-Q102-eGFP ) . For simplicity Httex1-Q25-eGFP , Httex1-Q46-eGFP , Httex1-Q72-eGFP and Httex1-Q102-eGFP are referred to as PolyQ-25 , PolyQ-46 , PolyQ-72 and PolyQ-102 , respectively in this study . F-1 progeny ( i . e . Hand-Gal4/+ , Hand-Gal4>UAS-PolyQ-25 , Hand-Gal4>UAS-PolyQ-46 , Hand-Gal4>UAS-PolyQ-72 or Hand-Gal4>UAS-PolyQ-102 ) of each transgenic cross were collected , separated by sex and cultured at 25°C . Lifespan of female progeny was determined with survivorship being monitored every third day with a food change as previously described [28] . Adult flies were analyzed at 1 and 3 weeks of age . The cardiac tissue-specific Hand-Gal4 driver was gift from Eric Olsen [68] . Transgenic unc-45 , SOD and MARF lines were generated as previously described [69]–[71] . Semi-intact hearts were prepared as described previously [29] , [72] . Direct immersion optics were used in conjunction with a digital high-speed camera ( up to 200 frame/sec , Hamamatsu EM-CCD ) to record 30 s movies of beating hearts; images were captured using HC Image ( Hamamatsu Corp . ) . Cardiac function was analyzed from the high speed movies using semi-automatic optical heartbeat analysis software ( a MatLab-based image analysis software ) which quantifies heart period , diastolic and systolic diameters , diastolic and systolic intervals , cardiac rhythmicity , fractional shortening and produced the M-mode records [29] , [72] . Dissected hearts ( from 1 and 3 week old flies ) were briefly exposed to 10 mM EGTA and then fixed with 4% paraformaldehyde in PBS as previously described [73] . Fixed hearts were probed with myosin antibody followed by goat-anti-rabbit IgG-Cy5 ( Chemicon , Temecula , CA ) and Alexa555-phalloidiin ( Invitrogen , Carlsbad , CA ) to stain F-actin . Fluorescence imaging of Drosophila heart tubes was carried out using an Apotome Imager Z1 ( Zeiss ) and an AxioCam MRm ( Zeiss ) as previously described [73] , [74] . To detect extended PolyQ-induced aggregates in the fly heart , we used Htt-PolyQ-GFP [5] in conjunction with anti-myosin or phalloidin . Dihydroethidium ( DHE ) and LysoTracker were employed for the detection of oxidative stress and autophagosomes/lysosomes respectively using a modified protocol previously described for use in other tissue [71] , [75] . Briefly , semi-intact hearts were prepared as described above and stained with DHE ( Molecular Probes , Carlsbad , CA ) at 2 µM final concentrations in artificial hemolymph for 30 min , followed by three washes with artificial hemolymph . Hearts were relaxed with 10 mM EGTA and mounted in Vectashield . A similar staining procedure was applied for the staining with LysoTracker red ( Molecular Probes , Carlsbad , CA ) : 30 min , 1 µM final concentration in artificial hemolymph . Fluorescence imaging was carried out using an Apotome Imager Z1 ( Zeiss ) and an AxioCam MRm ( Zeiss ) as previously described [73] , [74] . DHE intensity was quantified using ImageJ software . Semi-intact heart preparations were prepared for transmission electron microscopy using a modified protocol described previously [76] . Briefly , hearts were relaxed with 10 mM EGTA followed by a primary fixation protocol ( 3% formaldehyde , 3% glutaraldehyde in 0 . 1 M cacodylate buffer , pH 7 . 4 ) and secondary fixation ( 1% OsO4 , 100 mM phosphate buffer , and 10 mM MgCl2 , pH 7 . 4 ) . The samples were block stained in 2% uranyl acetate and dehydrated with an acetone series , followed by orientation and embedding in Epon-filled BEEM capsules . Polymerization was performed at 60°C under vacuum . Thin sections ( 50 nm ) were cut using a Diatome diamond knife on a Leica ultramicrotome and picked up on formvar-coated grids . Slices were stained with 2% uranyl acetate for 10 min and Sato's lead stain [77] for 2 min . Images were obtained at 120 kV on a FEI Tecnai 12 transmission electron microscope . We employed standard genetic and transgenic techniques [69] to co-express chaperones or SOD in flies expressing UAS-PolyQ in the heart using the Gal4 driver . Genetic crosses using multiple balancers were carried out with transgenic flies expressing unc-45 , SOD-1 or SOD-2 . Briefly , the wild-type genomic dunc-45 was used as previously described [69] . To ameliorate PolyQ-72 induced cardiac defects , adult males homozygous for the PolyQ-72 ( w1118/Y; +/+; PolyQ-72/PolyQ-72 , flies were crossed with female flies homozygous for the wild-type dunc-45 transgene on the 1st chromosome P[w+ , dunc-45/P[w+ , dunc-45]; Hand-Gal4/Cyo; +/+ [69] . The following progeny ( P[w+ , dunc-45]/w1118; Hand-Gal4/+; PolyQ-72/+ ) were analyzed to determine ability of UNC-45 over-expression to suppress Poly-Q induced cardiac defects . A similar genetic suppression approach was used with the wild-type dunc-45 transgene on the 2nd chromosome after crossing w1118/w1118; P[w+ , dunc-45]/P[w+ , dunc-45]; +/+ to w1118/Y; Cyo/Hand-Gal4; PolyQ-72/PolyQ-72 to obtain w1118/w1118; P[w+ , dunc-45]/Hand; +/PolyQ-72 . For expression of SOD-1 ( 2nd chromosome ) , w1118/w1118; P[w+ , SOD-1]/P[w+ , SOD-1]; +/+ flies were crossed to w1118/Y; Cyo/Hand-Gal4; PolyQ-72/PolyQ-72 to obtain w1118/w1118; P[w+ , SOD-1]/Hand; +/PolyQ-72 . Finally , for co-expression of unc-45 , SOD-1 in PolyQ-72 expressing flies , females homozygous for the wild-type dunc-45 transgene on the 1st chromosome P[w+ , dunc-45]/P[w+ , dunc-45]; Hand-Gal4/Cyo; +/+ were crossed to w1118/Y; P[w+ , SOD-1]/Cyo; PolyQ-72/PolyQ-72 to obtain P[w+ , dunc-45]/w1118; Hand-Gal4/SOD-1; PolyQ-72/+ . Similar genetic approaches were used to co-express UNC-45 or SOD in flies expressing UAS-Poly-25 in the heart using the Gal4 driver . To study the effects of MARF over-expression on mutant PolyQ-induce cardiac defects , flies with MARF overexpressing transgenes on their second or third chromosomes were crossed with extant stocks ( w1118/Y; Cyo/Hand-Gal4; PolyQ-72/PolyQ-72 ) to obtain w1118/w1118; P[w+ , MARF]/Hand; +/PolyQ-72 or w1118/w1118; +/Hand; P[w+ , MARF]/PolyQ-72 . We did not see any difference in the time to eclosion or the number of progeny between the mutant PolyQ and cardiac-specific SOD1/UNC-45/MARF overexpression lines . For treatment with resveratrol , flies expressing Hand-Gal4>UAS-PolyQ ( control or disease causing ) , Hand-Gal4; w1118 or ( P[w+ , dunc-45]/w1118; Hand-Gal4/+; PolyQ-72/+ ) were raised separately in standard media in the presence or absence of resveratrol ( final concentration , 1 mg/ml , a dose previously used in the Drosophila model [48] ) . Flies were collected on eclosion , separated by sex and cultured at 25°C; food was changed every 3 days . Adult flies were analyzed at 3 weeks of age . Additionally , we did not see any difference in the time to eclosion or the number of progeny between the mutant PolyQ and resveratrol fed organisms ( concentration 1 mg/ml ) . For treatment with H2O2 , flies were collected on eclosion and separated by sex . One group was cultured on food containing 1% H2O2 and a second group received standard food . Food was changed every 3 days and adult flies were analyzed at 3 weeks of age . Semi-intact hearts from flies expressing various length Htt-PolyQ-GFP [5] were prepared and fixation was carried out at described above . The numbers of aggregates in micrographs taken using the GFP wavelength ( 488 nm ) were quantified using ImageJ software ( Particle Count and Analysis function ) . Briefly , each immunofluorescence micrograph was divided into 50×50 µM square boxes ( 2500 µm2 unit area ) and the total number of aggregates within each box was quantified . Between three and four boxes per heart were analyzed and the number of aggregates per box was averaged for each heart . We also quantified the 2D surface area of the aggregates within the box using ImageJ software . Thus , we quantified mutant Htt-polyQ deposits in terms of the total number of aggregates per unit area as well as in terms of the size of aggregates from at least four hearts per genotype . For each heart aggregate density/size was determined for each of three 50×50 micron regions and averaged and four to six hearts were examined for each genotype/treatment . We used a filter trap assay for the quantification of Htt-polyQ GFP aggregates in the Drosophila heart as previously described [30]–[32] . Briefly , hearts were dissected ( 30–50/genotype ) and harvested in SDS lysis buffer ( 20 mM Tris-HCl , pH 7 . 5 , 200 mM NaCl , 2% SDS ) ; samples were diluted to 1 µg/100 µl with TBS ( 20 mM Tris-HCl , pH 7 . 5 , 500 mM NaCl ) and 500 ng were loaded onto a cellulose acetate membrane ( 0 . 2 µm pore size , Whatman , Piscataway , NJ ) , using the 96-well BioDot Apparatus ( Bio-Rad , Hercules , CA ) . Protein content of homogenized heart samples was determined using the DC Protein Assay Kit II ( Bio-Rad , Hercules , CA ) . Non-specific binding sites were blocked using 5% nonfat milk in TBA buffer for 2 hours . Immunodetection was performed after incubation with mouse anti-GFP ( Covance Research Products , Dedham , MA ) and secondary antibodies ( KPL , Inc . , Gaithersburg , MD ) using Thermo Scientific SuperSignal West Dura Substrate on a Bio-Rad ChemiDoc XRS System . A parallel series of 500 ng samples was blotted onto nitrocellulose ( Bio-Rad , Hercules , CA ) to confirm GFP expression in non-aggregating PolyQ-GFP , using H2B antibody ( Cell Signaling , Danvers , MA ) as a loading control . Quantification of the immunoblot's density was carried out using ImageJ software . For all quantitation except lifespan analysis , statistical significance was determined using one-way analysis of variance ( ANOVA ) followed by Dunnett's post-hoc test to determine significance between groups with Prism 6 . 0 ( Graph Pad ) software . Significant differences were assumed for p<0 . 05 . For lifespan studies , data were analyzed using the Gehan-Breslow-Wicoxon test followed by multiple comparisons between control and experimental groups . Significance was taken at p values less than the Bonferroni-corrected threshold of p<0 . 0125 . | Huntington's disease ( HD ) is associated with amyloid-like inclusions in the brain and heart , and accumulation of amyloid protein is associated with neurodegeneration and cardiomyopathy . Recent studies suggest that HD patients show increased susceptibility to cardiac failure . However , the mechanisms by which disease-causing poly-glutamine repeats ( PolyQ ) cause heart dysfunction in these patients are unclear . We have developed a novel Drosophila heart model that exhibits significant GFP-positive aggregates upon HD-causing PolyQ expression in myocardial cells resulting in PolyQ length-dependent physiological defects . Modulation of protein folding and oxidative stress pathways in this system reduced the number of aggregates and reversed the cardiac dysfunction in response to expression of disease-causing PolyQ . The ability to explore PolyQ-associated mechanisms of cardiomyopathy in a genetically tractable whole organism , Drosophila melanogaster , promises to provide novel insights into the relationship between amyloid accumulation and heart dysfunction . Our findings not only impact the understanding of PolyQ-induced cardiomyopathy but also other human cardiac diseases associated with oxidative stress , mitochondrial defects and protein homeostasis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2013 | Huntington's Disease Induced Cardiac Amyloidosis Is Reversed by Modulating Protein Folding and Oxidative Stress Pathways in the Drosophila Heart |
The trafficking of primordial germ cells ( PGCs ) across multiple embryonic structures to the nascent gonads ensures the transmission of genetic information to the next generation through the gametes , yet our understanding of the mechanisms underlying PGC migration remains incomplete . Here we identify a role for the receptor tyrosine kinase-like protein Ror2 in PGC development . In a Ror2 mouse mutant we isolated in a genetic screen , PGC migration and survival are dysregulated , resulting in a diminished number of PGCs in the embryonic gonad . A similar phenotype in Wnt5a mutants suggests that Wnt5a acts as a ligand to Ror2 in PGCs , although we do not find evidence that WNT5A functions as a PGC chemoattractant . We show that cultured PGCs undergo polarization , elongation , and reorientation in response to the chemotactic factor SCF ( secreted KitL ) , whereas Ror2 PGCs are deficient in these SCF-induced responses . In the embryo , migratory PGCs exhibit a similar elongated geometry , whereas their counterparts in Ror2 mutants are round . The protein distribution of ROR2 within PGCs is asymmetric , both in vitro and in vivo; however , this asymmetry is lost in Ror2 mutants . Together these results indicate that Ror2 acts autonomously to permit the polarized response of PGCs to KitL . We propose a model by which Wnt5a potentiates PGC chemotaxis toward secreted KitL by redistribution of Ror2 within the cell .
Primordial germ cells ( PGCs ) are embryonic precursors of the gametes that arise before other major cell lineages in most multicellular animals [1] . This early specification necessitates a lengthy migration through the developing embryo in order to reach the nascent ovaries or testes . In mice , epiblast-derived cells seal their germline commitment at the embryo periphery ∼e7 . 25 , then enter the forming endoderm and travel through the elongating hindgut epithelium . PGCs make a coordinated exodus into the surrounding mesentery at e9 . 5 and then converge on the gonadal ridges between e10 . 5 and e11 . 5 . Though exquisitely coordinated , this process is also imperfect; by e12 when migration is over , stragglers consistently remain outside the gonad in midline tissues , and are eliminated by apoptosis [2] . The importance of balanced regulation of PGC survival and migration is evident by the consequences of dysregulation: failure to survive or reach the gonad can lead to sterility , whereas inappropriate survival can lead to germ cell tumors [3] , [4] . The molecular mechanisms underlying the migration of these evolutionarily essential but relatively inaccessible cells remain largely unknown in the mammalian germline . Here we conducted a forward genetic screen for germ cell defects in mouse embryos and identified an allele of Ror2 . Ror2 is a highly conserved receptor tyrosine kinase with homologs in many metazoans from Aplysia to Drosophila to humans [5] . Widely expressed during development , Ror2 has been implicated in chondrocyte differentiation , cochlear , craniofacial , heart , limb and gut morphogenesis in mice and humans [6]–[9] . Work in a number of different organisms suggests that Ror2 signaling affects cell polarity . In the developing mouse gut epithelium , the protein exhibits apicobasal polarity in its distribution [10] . Polarity is requisite for cells undergoing directed migration , cell division in a particular orientation , as in asymmetric divisions , and for the organization or shape of cells with respect to their neighbors , for example in convergent extension . Defects in cell shape and convergent extension have been reported in the mouse gut , organ of Corti , and Xenopus gastrula as a result of Ror2 signaling loss [9] , [11]–[13] . Ror2-mediated polarized cell division has been reported in C . elegans [14] . A role for Ror2 signaling in directional migration has been reported in the mammalian palate [15] and in several cell lines , via c-Jun N-terminal Kinase and the actin-binding protein FilaminA [16]–[18] . Phenotypic resemblance between mouse embryos with targeted deletions of Ror2 and those deficient for Wnt5a first suggested that these genes share a common pathway [6] , [8] , [17] , [19] . Biochemical approaches later confirmed ligand-receptor interactions between Wnt5a and Ror2 via the cysteine-rich ( frizzled-like ) extracellular domain of Ror2 [17] . Indeed , the expression patterns of Wnt5a and Ror2 virtually overlap in the primitive streak , tail mesoderm and limb buds of midgestation mouse embryos [19]–[21] . Wnt5a was similarly invoked in aspects of cell polarity , including orienftation of cell division in the limb [22] , convergent extension movements and cell shape in the Xenopus gastrula [23] , [24] , and polarized migration in a melanoma cell line [25] , [26] . Many of these different Wnt5a-Ror2 pathway mutants exhibit similarly altered distribution of polarity mediators , such as Disheveled [23] , [27] , [28] , the Dlg-Lgl complex [24] , [29] , Van Gogh [29] , or adhesion receptor complexes [26] . The identification of the Ror2Y324C mutant in an unbiased screen for PGC defects brings to light a previously unrecognized function of Ror2 in germ cell development . We show here that Ror2 and its putative ligand Wnt5a promote efficient migration of PGCs to the embryonic gonads . These studies demonstrate a cell intrinsic function for Ror2 in potentiating the polarized response to secreted KitL , drawing a new link between Ror2 and Kit signaling in PGC migration .
As an unbiased approach to identifying new genes involved in mouse germ cell development , we conducted a genome-wide recessive ethylnitrosourea ( ENU ) mutagenesis screen for PGC defects in e9 . 5 embryos [30] . One of the mutations identified based on the presence of ectopic PGCs mapped to the region of Ror2 . An A to G transition in exon 7 at nucleotide 1203 causes a tyrosine to cysteine substitution at position 324 ( Y324C ) of the ROR2 predicted protein ( Figure 1A ) . This missense mutation falls in the kringle domain , a conserved structural motif in the ROR2 extracellular domain . Ror2Y324C homozygous embryos exhibit defects in tail elongation ( Figure 1B , 1C ) and somite segmentation , similar to the Ror2 targeted deletion allele ( Figure 1D ) [8] , [17]; like the knockout , Ror2Y324C mutants die perinatally . Ror2 immunoblotting on e10 . 5 embryo lysates revealed a double band at approximately 200 kD; both bands were present in similar amounts between WT and Ror2Y324C mutants ( Figure 1E ) . In humans , missense mutations in the hRor2 cysteine rich , kringle and tyrosine kinase domains that are associated with Robinow syndrome cause the protein to be retained in the endoplasmic reticulum [31] . We examined the expression of ROR2 at e11 . 5 by intracellular staining with an antibody directed against the cytoplasmic tail of the receptor; by flow cytometry signal was present at similar levels in WT and Ror2Y324C ( Figure 1F , right ) . These experiments suggest that the mutation does not affect protein stability but do not discriminate between its normal or abnormal subcellular localization . To determine whether Ror2 is expressed in PGCs , we employed a transgenic mouse strain , Oct4ΔPE-EGFP , which expresses Enhanced Green Fluorescent Protein ( GFP ) under a modified Oct4 reporter that is specific to PGCs during mid-gestation [10] , [32] ( Figure 1F ) . By flow cytometry , ROR2 intracellular staining was present within the GFP+ population at e11 . 5 ( Figure 1F ) . Furthermore , when Oct4ΔPE-EGFP+ PGCs were purified flow cytometrically , Ror2 transcript could be detected by semi-quantitative RT-PCR; more transcript appeared to be present in GFPnegative cells from embryo tails ( denoted ‘soma’; Figure 1G ) , where high levels of Ror2 have been previously detected by in situ hybridization [8] . The purity of sorted PGCs was confirmed by RT-PCR for Oct4 , which was absent in somatic cells , and KitL , which was confined to soma ( Figure 1G ) . ROR2 protein was similarly detected in histologic sections with two different antibodies; signal appeared to be concentrated at the apical surface of the hindgut and somites [33] and in the ventral neural tube in wild type embryos ( Figure 1H–1H′ ) , as previously reported [34] . ROR2 was also present throughout the e10 . 5 dorsal mesentery and enriched at the membrane of wild type PGCs ( Figure 1I , 1I′ ) . These studies confirm the expression of Ror2 mRNA and protein in migratory and postmigratory PGCs , as suggested by previous microarray data [35] , and demonstrate the stable expression of Ror2Y324C mutant protein . A major ligand for Ror2 is believed to be Wnt5a . Wnt5a mRNA expression in the tail and hindgut of the embryo overlaps that of Ror2 , although precisely which cells secrete Wnt5a remains unclear [19]–[21] . By RT-PCR we determined that Wnt5a transcript is present in sorted Oct4ΔPE-EGFP+ PGCs , although it is more abundant in GFPnegative somatic cells of the tail and hindgut ( Figure 1G ) . In histological sections stained with a WNT5A antibody , we observed bright foci as well as intercellular signal in the intestine and gonadal ridges ( Figure 1J–1J′ ) , which both lie on the PGC migratory route . Upon closer examination , WNT5A could be detected at variable levels at or near the surface of PGCs ( Figure 1J , inset ) . These results collectively identify a role for Ror2 in PGC development and raise the possibility that PGCs perceive paracrine or autocrine WNT5A signals via the Ror2 receptor . We next characterized the phenotypes of PGCs in Ror2Y324C mutants . In e10 . 5 embryos stained with SSEA1 antibody [36] , [37] , PGCs can be visualized migrating through the dorsal mesentery ( Figure 2A ) . In Ror2Y324C mutants , PGCs do not migrate rostrally , but remain in the mesentery surrounding the caudal hindgut ( Figure 2B ) , as well as on the surface of the tail and in the allantois ( Figure S3B , S3C ) . At e11 . 5 , immunostaining with GCNA ( a marker of postmigratory PGCs [38] ) revealed a reduction in the number of PGCs within Ror2Y324C gonad primordia compared to wild type; furthermore , the distribution of Ror2Y324C PGCs was skewed toward the caudal end of the gonad and extragonadal PGCs were increased in midline tissues ( Figure 2E , 2F ) . At e12 . 5 , male and female Ror2Y324C gonads appeared less densely populated with PGCs ( Figure 2I , 2J , female not shown ) . We developed techniques for the quantification of PGCs in the entire embryo or embryonic gonad with confocal imaging and 3D analysis ( see Methods ) . The mean number ± standard deviation of PGCs in mutants at e10 . 25 ( 443±73 ) was similar to wild type ( 551±157; Figure 2D ) , in spite of their abnormal distribution . However , at e11 . 5 , the number of Ror2Y324C PGCs in gonads was diminished ( 1243±369 ) compared to wild type ( 2598±265 , p = 0 . 0002; Figure 2H ) . At e12 . 5 this difference persisted ( p = 0 . 009 ) , as 7058±2282 PGCs were counted in wild type gonads and 3825±1144 in Ror2Y324C ( Figure 2L ) ; male and female were combined here , as their numbers were similar . The PGC estimates and corresponding doubling time found in wild type embryos ( 13 . 4–16 . 7 hours ) are similar to those reported previously [39] . The doubling time for Ror2Y324C PGCs falls within this range for postmigratory PGCs , but was more protracted from e10 . 25–11 . 5 ( 20 hours ) , predicting an earlier decline in proliferation or rise in apoptosis . We compared the phenotype of Ror2Y324C PGCs to that of a targeted Ror2 knockout allele [8] . At e11 and e12 , we observed a similar PGC decrease compared to age-matched wild type C57Bl/6 littermates ( Figure S1 ) . Despite genetic background differences , the PGC deficit in Ror2−/− embryos was indistinguishable from that resulting from our point mutation . This similarity suggests that Ror2Y324C is a strong loss of function allele . Previous work has shown that Ror2 lies downstream of Wnt5a both biochemically and genetically [17] , [19] , [40] . Therefore , we examined the PGCs of Wnt5a null mutants and found a more pronounced and earlier deficit compared to Ror2 . At e10 . 5 , Wnt5a−/− PGCs were similarly caudally distributed ( Figure 2C ) but were already depleted in number ( 242±121 ) compared to wild type ( Figure 2D ) . We noted significant reductions at e11 . 5 , when 310±148 PGCs were present in Wnt5a gonads ( Figure 2G , 2H ) , and by e12 . 5 this number increased to 1587±985 ( Figure 2K , 2L ) . Consistent with biochemical data [17] , [40] , the greater severity of the Wnt5a germ cell phenotype suggests that this ligand operates through other receptors besides Ror2 . Together , these studies demonstrate that Wnt5a and Ror2 mutants are phenocopies in the PGC compartment , which corroborates their function there as ligand and receptor . To investigate the cellular mechanism underlying the PGC deficit in Ror2 and Wnt5a mutants , we extended our quantitative imaging in the embryonic gonad to include markers of proliferation and death . We performed triple immunofluorescence for GCNA , as well as phospho-histone H3 ( PHH3 ) , and cleaved PARP to quantify subsets of mitotic and apoptotic PGCs , respectively ( Figure S2 ) . No differences were observed in cPARP expression among postmigratory PGCs in wild type , Ror2 , or Wnt5a gonads ( Figure 3 ) . However , analysis of e10 . 5 embryo sections revealed an increase in cPARP expression among still migratory PGCs in Ror2Y324C ( 11 . 9±1 . 6% ) and Wnt5a ( 11 . 4±4 . 8% ) relative to wild type ( 4 . 3±1 . 3%; Figure 3A , staining shown in Figure S3 ) . This discrete wave of apoptosis preceded any observed loss in cell number in Ror2 mutants . We next compared the frequencies of programmed cell death between PGCs within and outside the e11 . 5 gonad . Caspase3 staining in histologic sections revealed similar frequencies in properly localized PGCs , but increased apoptosis among extragonadal Ror2Y324C germ cells ( 14 . 0±1% ) compared to wild type ( 2 . 4±2 . 4% ) and Wnt5a ( 4 . 9±1 . 2%; Figure 3B ) . Examination of PGC proliferation by PHH3 staining did not reveal significant differences in the frequency of proliferating PGCs between wild-type , Wnt5a , or Ror2Y324C embryos at e10 . 25 or e11 . 5 ( Figure 3C ) ; despite several hundred PGCs counted in each genotype at each stage , the variation was large . Together these results demonstrate that increased apoptosis rather than reduced proliferation contributes to the PGC deficit in Ror2 and Wnt5a mutants . Restriction of the observed burst of programmed cell death to migratory PGCs , together with its absence in gonadal PGCs , suggested that the location of mutant germ cells could be a factor in their elimination . On one hand , migrating mutant PGCs could be more sensitive to the reduced levels of survival factors such as KITL and SDF1 in the dorsal mesentery as compared to the gonad [2] , [41] , [42] , [43] , where they are more protected from death . On the other hand , inefficient migration may lead to an accumulation of ectopic Ror2 PGCs , which die in an environment lacking survival factors [2] . To distinguish between these possibilities , we rescued PGC apoptosis in Ror2 mutants by generating double mutants with a targeted knockout of the pro-death gene Bax . Previous work established an increase in ectopic PGCs in e11 . 5 Bax single mutants due to the lack of apoptosis of mis-migrated PGCs , although the total number of PGCs remained unchanged [2] . Genetic ablation of Bax in Ror2Y324C mutants increased the number of midline and ectopic PGCs , but did not restore the number of PGCs in the gonads . At e11 . 5 , 1815±362 PGCs were counted in Ror2; Bax double mutant gonads , which did not differ from 1275±359 in stage matched Ror2 littermates ( p = 0 . 07; Figure 3D ) . Although Bax does not rescue PGCs in Ror2 gonads , a significant increase in the total number of PGCs in the entire aorta-gonad-mesonephros region of double mutants compared to Ror2 single mutants ( p = 0 . 036; data not shown ) reflects rescue of ectopic PGC death throughout the midline in Ror2; Bax ( Figure 3G , compared to Figure 3E , 3F ) . This result suggests that defects in migration are primary to the defects in PGC survival in Ror2 mutants . We next directly compared the efficiency of PGC migration in mutants . When quantified in histological sections at e11 . 5 ( Figure 3H , 3I ) , ectopic ( extragonadal ) PGCs comprised over 70% of the total PGCs in Wnt5a mutants , and 30% in Ror 2Y324C , compared to less than 5% in wild type ( Figure 3J ) . Poor cell trafficking could therefore account for the loss of gonadal PGCs of both mutants at e11 . 5 . However , it remained unclear whether morphologic differences in the caudal hindgut of both mutants cause the observed migration defects . Indeed , morphological and molecular analysis revealed a shortening and widening of the Ror2Y324C caudal hindgut at e9 . 5 ( Figure 1E , Figure S4 ) , which corresponds to the PGC exodus from the hindgut . Upon examining embryos before hindgut formation , we confirmed that the location and number of early PGCs were indistinguishable from wild type in Ror 2Y324C as well as Wnt5a at e7 . 5–8 . 0 ( Figure S5 ) . By e9 . 0 , we observed ectopic PGCs accumulated in the allantois , throughout the tail mesoderm , and caudal hindgut of Ror2Y324C mutants ( Figure S6A–S6C ) . However , this phenotype does not distinguish between the possibilities of an intrinsic PGC migration defect versus a structural abnormality that hindered the passage of PGCs from the allantois into the hindgut pocket . [44] . Given the previously demonstrated expression of Wnt5a throughout the allantois and primitive streak [19] , we wondered whether it could act chemotactically to draw PGCs from the allantois into the hindgut . To address this possibility , we implanted beads coated with WNT5A into the caudal region of e8 . 0 embryos . Control BSA-coated beads delivered to the hindgut pocket did not disrupt embryo or PGC development over 24 h culture ( Figure S6D , S6G ) . Beads soaked in recombinant WNT5A or concentrated conditioned medium did not alter the course of PGCs , whether placed directly in or near their path ( Figure S6E , S6H ) . Consistent with previous reports [43] , beads similarly impregnated with the known chemoattractants SDF1and Stem Cell Factor ( SCF , or secreted KitL ) affected migration of PGCs at close range , inducing occasional deviation from their normal route ( Figure S6F , S6I ) . Although we did not assess biological activity of Wnt5a-soaked beads , when PGCs were explanted and cultured over 24 hours , we did measure a modest increase in their number in the presence of recombinant WNT5A , suggesting that WNT5A is biologically active ( Figure S6J ) . Collectively these results could indicate that WNT5A may not act as a direct chemotactic cue for PGCs; rather , they suggest that Wnt5a and Ror2 could have a permissive role to allow the response of PGCs to other navigation signals . Reduced PGC colonization of the gonads in Ror2 and Wnt5a mutants could result from disruptions in hindgut architecture or from intrinsic defects in PGC migration . The expression of Ror2 in both PGCs and their surrounding tissues does not provide any insight . In fibroblasts , previous work showed that WNT5A induces motility , cell shape change , and chemotaxis via Ror2 [16] , [45] . We did not observe PGC chemotaxis toward a WNT5A source in cultured embryos . Other work shows that WNT5A polarizes melanoma cells when a chemotactic gradient is present [26] . We sought a direct test of migratory capacity of isolated Ror2 PGCs . However , when sorted from e9 . 5–10 . 5 embryos using the Oct4ΔPE-EGFP reporter , we did not observe any migration of wild type PGCs toward SDF1 or SCF in a transwell assay , as previously reported [46] . However , Farini et al . also showed that SCF elicited cytoskeletal changes and membrane protrusions in isolated PGCs over a short period [46] . We replicated this result using flow cytometrically purified Oct4ΔPE-EGFP+ cells from e9 . 5 embryo posteriors and maintained on Matrigel in serum-free media . Without the support of feeder cells , which provide growth factors , survival was poor and PGCs appeared round and devoid of filopodia ( Figure 4A ) . As reported [46] , the addition of SCF induced morphological changes in PGCs , including the acquisition of membrane protrusions and ellipsoid shape ( Figure 4B ) . We noted that the shape assumed by Ror2 PGCs cultured in these conditions differed from wild type , and therefore endeavored to quantify this morphology . Using phalloidin to define the F-actin cytoskeleton , we measured the longest cellular axis and the orthogonal short axis of the cell body; we then computed an Elongation Index ( ALong−AShort ) / ( ALong+AShort ) , which approaches zero for round cells , such as the example in Figure 4C . Elongated cells often extended filopodia or lamellopodia , which were not included in the measurement , but which usually aligned with the long axis ( Figure 4D–4D′ ) . Following 7 hours of culture without SCF , a mean Elongation Index ( EI ) of 0 . 044 was observed in wild type PGCs , which increased to 0 . 088 in the presence of SCF ( p = 0 . 0005; Figure 4E ) . PGC elongation continued to increase in culture up to 20 hours in SCF , to a mean EI of 0 . 169 , ( Figure 4F ) . By contrast , Ror2 PGCs mutants cultured in parallel exhibited a mean EI of 0 . 114 in SCF , which is significantly lower than mixed wild type and heterozygous PGCs ( p = 0 . 005 ) . When SCF was excluded from the media , but a strip of Matrigel was introduced along one side of the culture well to produce a gradient , the elongation response of WT PGCs was similar to that in static SCF , with a mean EI of 0 . 20; the graded source of SCF did not increase the EI of Ror2 PGCs: mean EI of 0 . 11 , p = 0 . 0004 ( Figure 4G ) . Short axis dimensions did not differ between wild-type and mutant PGCs ( data not shown ) , but as we did not assess the z-axis length , these results do not exclude the possibility that Ror2 PGCs occupy less volume instead of remaining more spherical than wild type following SCF treatment . We also examined the capacity of PGCs to align with a chemotactic gradient . Using the long cellular axis explained above , we measured the angle between this axis and a line orthogonal to the source of SCF ( schematized in Figure 4H ) . When SCF was uniformly present in the media ( here termed static ) , the orientation of WT PGCs was randomly distributed between 0 and 90° from an arbitrary line , as would be expected . However , following 20 hours in an SCF gradient , wild-type PGC orientations were biased toward lower angles; that is , they showed greater alignment parallel to the gradient ( Figure 4I , p = 0 . 0018 ) . Ror2 PGCs did not preferentially orient toward the SCF source , but were randomly distributed in their orientations ( Figure 4I , p = 0 . 0004 ) . Taken together , these in vitro studies reveal a compromised ability of Ror2 PGCs to respond to SCF , either by elongating or orienting toward a chemotactic gradient . Because these assays were carried out in the absence of feeder cells using Oct4-ΔPE-EGFP+ cells sorted to >95% purity , the observed defects must be cell-intrinsic . Polarized cell migration depends upon the perception of an extracellular chemotactic gradient , the acquisition of polarized molecular or membrane components , and ensuing changes in cellular organization , including cytoskeletal elements and organelles [47] . We observed an overall reduction in Ror2 PGC shape change and alignment in the presence of an SCF gradient compared to wild type . This phenotype could result from the impaired perception of a chemotactic cue or diminished capacity to respond . As little is known about PGC polarity , we first examined the localization of two subcellular structures involved in polarized responses of migratory cells , the Golgi apparatus and the centrosome; identified here by GM130 ( Golgi ) and Pericentrin ( centrosome ) immunofluorescence , these organelles are positioned by microtubules in response to polarity cues [48] . Following SCF exposure , cultured wild type and Ror2 PGCs both elaborated F-actin-rich extensions ( Figure 5A′ , 5B′ , 5C′ ) . GM130 and Pericentrin staining was observed colocalized in three discrete cellular geometries . Asymmetric localization of GM130 and Pericentrin to one extreme of the nucleus in elongated cells was denoted Class I ( Figure 5A″ ) . Central positioning of GM130 and Pericentrin adjacent to or above the nucleus was denoted Class II ( Figure 5B″ and 5C″ ) . Class III included geometrically rounded cells with eccentric GM130 and Pericentrin ( Figure 5D″ ) . Finally , GM130 was occasionally observed as dispersed foci ( not shown ) , Class IV , which is most likely the configuration in mitotic cells [48] . The tabulated results of several experiments are shown ( Figure 5E ) . A similar frequency of wild-type Class I PGCs was observed in static and graded SCF ( 69% and 73% , respectively ) . This distribution of the Golgi and centrosome appears to be nonrandom given the relatively large cellular area occupied by the PGC nucleus . Strikingly , a significant overall reduction of Class I Golgi position was observed in Ror2: 45% in static SCF and 48% in a gradient , both of which differ from wild type ( p = 0 . 005 ) . This result suggests that the coordination of centrosome/Golgi position and cell shape is affected in Ror2 PGCs . However , if we consider the Golgi position apart from cell shape—since the rounded Class III cells could retain molecular and organelle polarization– it becomes apparent that the cells in Class III also exhibit Golgi/centrosome asymmetry . In this line of reasoning , we find that the frequency of combined Class I and III PGCs does not differ between wild type and Ror2 in graded SCF ( p = 0 . 23 ) and is barely significant in static SCF ( p = 0 . 04 ) . This analysis could suggest that Ror2 PGCs are defective in cell elongation , but not polarized positioning of the centrosome and Golgi . Conversely , if we compare only geometrically elongated cells , or those in Classes I and II , we find a decreased incidence of polarized Golgi position ( Class I ) of Ror2 PGCs cultured in static SCF ( p = 0 . 029 ) , but not in gradient SCF ( p = 0 . 18 ) compared to wild type . Given that the majority of Ror2 PGCs elongate to some degree in SCF , this discrepancy in Golgi position could reveal a more subtle defect in their polarized response . Finally , although a rare class , the incidence of Class IV or dispersed GM130 appears elevated in Ror2 PGCs ( Figure 5E ) . This uptick could reflect a slight increase in proliferation of the mutant PGCs that we have observed in vitro ( Figure S7 ) . Taken together these experiments demonstrate a decoupling between cell elongation and polarized position of the Golgi and centrosome in Ror2 mutant PGCs; however with the alternate interpretations of Class III cells as either randomly positioned or polarized Golgi/centrosome within a rounded cell , it remains possible that Ror2 acts as a cell polarity effector or else in the associated cell shape changes . Several previous studies have implicated Ror2 in cell polarity , including polarized cell division in C . elegans [14] , directional migration in the limb and several mammalian cell lines [18] , [26] , [49] , and apicobasal polarity in the mouse gut [9] . A polarized distribution of ROR2 within the developing gut epithelium [9] prompted us to examine ROR2 localization in PGCs following culture in SCF . Immunofluorescence revealed asymmetry of ROR2 within the cytoplasm as well as on the surface membrane of PGCs . The distribution of ROR2 at one extreme of the cell coincided with GM130 ( Figure 5F , 5F′ ) . ROR2 was also observed prominently on the membrane protrusions of cultured PGCs . Returning to the embryo , we examined the subcellular distribution of ROR2 and GM130 in PGCs in vivo . In e10 . 25 histologic sections , immunostaining revealed an apical enrichment of ROR2 in the hindgut and dorsal neural tube , colocalized with GM130 ( Figure 6A–6A′ ) . At this stage , PGCs identified by the expression of Stella are migrating through the dorsal mesentery toward the gonadal ridges ( Figure 6A ) . Within these PGCs , ROR2 appeared to be enriched on one side in most instances ( Figure 6B , 6C , 6D ) . This enrichment was coincident with GM130 ( Figure 6B″ , 6C″ ) and , unexpectedly , Stella ( Figure 6B′ , 6C′ ) . As a polarized Stella distribution has not been previously reported , we wondered whether this pattern could reflect the plane of section . When PGCs were instead immunostained with the SSEA-1 antibody , the asymmetric distribution of ROR2 persisted ( Figure 6D ) , but SSEA-1 appeared to be localized consistently around the PGC border ( Figure 6D′ ) , as did β-catenin ( Figure 6D″ ) . Together , these data demonstrate a polarized distribution of ROR2 in PGCs that are responding to chemotactic cues in vitro and migrating in vivo . Its localization on filopodia and segregation on the same side of the cell as the Golgi suggests Ror2 could be important in the polarization response of the cell in response to SCF . Upon examining histologic sections from Ror2Y324C embryos , we did not observe a comparable degree of asymmetric ROR2 distribution or Stella distribution in PGCs , and GM130 staining was present but dimmer ( Figure 6E , 6E′ , 6E″ ) . This result suggests that a functional Ror2 receptor could localize asymmetrically on a PGC responding to chemotactic cues in order to amplify or enhance the polarized response . We next asked whether the elongation phenotype of Ror2 PGCs in culture could be observed in vivo . As the expression of KitL changes dynamically after 9 . 5 to become restricted to the genital ridges [2] , PGCs migrating within the dorsal mesentery toward the gonadal ridges probably experience a gradient of secreted KITL analogous to the graded SCF in vitro . We examined PGCs migrating through the dorsal mesentery of sectioned e9 . 75–e10 . 75 embryos using DAPI and SSEA-1 staining to ensure that the entire cell was captured . Visually , many PGCs in Ror2Y324C embryos at this stage appeared rounded ( Figure 7B , 7C ) . In confocal stacks , we located the largest cellular cross section for measuring the longest cellular axis and the short axis orthogonal to this one ( Figure 7A , 7D ) . In migratory wild-type PGCs , we measured a mean EI of 0 . 18±0 . 08 ( Figure 7E ) . This is in line with the mean EI of 0 . 20±0 . 12 that was determined in two dimensions from SCF-gradient cultured PGCs . In Ror2Y324C embryos , migratory PGCs exhibited a mean EI of 0 . 09±0 . 07 , which is significantly less than in wild type ( p<0 . 0001 ) . For comparison , we examined PGCs located within the gonads of e10 . 75–11 . 5 embryos , since previous studies reported that following their arrival in the gonad , PGCs acquire a rounded morphology [50] . Wild-type postmigratory PGCs measured 0 . 08±0 . 05 mean EI . Together these in vivo observations suggest that Ror2 enhances the polarized response that leads to cell elongation of migratory PGCs .
Our analysis shows a surge in apoptosis , an increase of ectopic PGCs in Ror2Y324C and Wnt5a , and disrupted hindgut architecture in Ror2Y324C embryos; consistent with this , PGC diminution in e11 . 5 gonads and concomitant increase in ectopic PGCs was reported in Wnt5a mutants as this manuscript was under revision [53] . However it remains unclear which of these three defects– apoptotic elimination , mismigration or anatomic barriers– primarily causes the PGC phenotype . On one hand , PGCs that migrate inefficiently could be increasingly subject to death , or on the other hand , dying or unhealthy PGCs could migrate poorly . This dilemma is resolved by rescuing cell death with genetic ablation of the proapoptotic gene Bax , which does not restore the loss of PGCs in Ror2Y324C gonads . By contrast , in Steel mutants , which lack both membrane and secreted KitL , one or two Bax null alleles is sufficient to rescue KitLSteel/Steel gonadal PGCs [2] . Together these results argue that migration is the primary defect in Ror2 PGCs , and apoptosis in the periphery arises as a consequence of reduced survival factor exposure . The possibility remains that ectopic Ror2 PGCs are increasingly sensitive to the withdrawal of survival factors . KITL and SDF1 , known to be the most important PGC survival factors [42] , [43] , [54] , are both concentrated in the e11 . 5 gonads and their absence in peripheral tissues likely leads to Ror2 PGC death . In the developing mammalian palate , a series of bead and cell implantations suggest that WNT5A is sufficient for directional movement of cells via Ror2 [15] . However , similarly implanted beads coated with WNT5A did not divert the migration of PGCs in our studies ( Figure S6 ) , and we could not detect any role for Wnt5a as a chemoattractant . Instead , we suggest that Wnt5a may act permissively in PGC migration . This is not unprecedented , as in a melanoma cell line , Witze et al . showed that WNT5A acts permissively to regulate the polarized distribution of adhesion receptors in response to a chemokine gradient [26] . In PGCs , the known chemoattractive factors include SDF1 [42] , [43] and KitL [2] , [41] , [55] . Recognized as the most critical growth and survival factor for PGC , KitL was first postulated as a guidance cue for PGCs from the analysis of the SteelDickie mutant [56]; SCF ( secreted KitL ) was later shown to induce PGC migration ex vivo , as well as inducing cell shape changes [46] , [57] . Aberrant cell shape was previously noted in PGCs from Steel mutants [58] . When the survival of Steel PGCs was restored in compound mutants with Bax , functions for KitL in motility , adhesion and colonization of the gonad were identified [2] , [41] . The resemblance of these cellular phenotypes to what we observed in Ror2Y324C PGCs prompted us to ask whether Ror2 could enhance the response to KitL chemotactic cues . Our ex vivo experimental approach addresses three separate aspects of cell migration: polarity , cell shape change and orientation toward a chemotactic cue . We find that a gradient of SCF induces geometric elongation as well as a nonrandom alignment of wild type PGCs within the field . This chemotropic function of SCF was previously recognized in mast cells but not PGCs [59] , [60] . Similar to a recent report [57] , our results also reveal a chemokinetic , or non-directional function of SCF in migration , as wild-type PGCs assumed an elongated , polarized morphology when SCF was uniformly present . This morphology was accompanied by a polarization of the Golgi apparatus , which is typical of migratory cells [61] . In all of these cellular behaviors , Ror2 PGCs exhibit a mitigated response to SCF; their orientation is randomized instead of aligned with respect to the gradient , they elongate less , and the frequency of polarized Golgi distribution reduced . We find , strikingly , a similar difference in the shape of migratory PGCs in Ror2 mutant embryos . This result confirms in vivo a function for Ror2 in the polarized migration of PGCs . Taken together , the in vivo and in vitro experiments suggest that Ror2 signaling enhances the chemotactic response of PGCs to KitL emanating from the gonadal ridges . Although the precise function of Ror2 in this polarized migration remains to be determined , its nonrandom protein distribution throughout migratory PGCs may provide a clue . The observed pattern of ROR2 within the cytoplasm and near the cell membrane is reminiscent of the asymmetry within the hindgut epithelium [9] . The potential colocalization with the Golgi apparatus is intriguing and warrants further investigation . On the other hand , the distribution of ROR2 on membrane protrusions is reminiscent of the reported expression on the dendrites of hippocampal neurons [62] . The altered distribution of ROR2 in Ror2Y324C PGCs argues for its specificity and functional significance , and leads us to propose that ROR2 distribution becomes polarized in response to directional KitL cues and thus reinforces the polarization of the cell . In other words , Ror2 might potentiate asymmetry in Kit signaling , or even transform it from a general signal to a polarized signal . It is also possible that the localization of ROR2 on protrusions promotes the growth and selection of filopodia into a clear lamellopodia or a leading edge , similar to the axonal path finding function of the homolog in C . elegans [7] . Elucidating the dynamic distribution of ROR2 in PGCs undergoing polarized responses will be an important future pursuit . The connection established in these studies between Ror2 and SCF-induced cell polarization is new , and the molecular nature of the relationship is unclear . The robust evidence for the specificity of SCF for cKit rules out the possibility of biochemical interaction between SCF and Ror2 [63] . However , the absence of feeder cells or serum and purity of sorted PGCs in our culture system demonstrates that both proteins are acting in the same cell . Based on the detection of Wnt5a transcript and protein in PGCs , a plausible scenario would involve autocrine WNT5A secreted from PGCs in the cultures . In this model , PGC polarization is initiated by KitL and amplified by localized extracellular concentrations of WNT5A secreted from either PGCs or surrounding mesentery . In regions of high Wnt5a expression , which correspond to successive targets of PGCs such as the hindgut and the gonadal ridges , the sensitivity to KitL may be amplified to help guide PGCs toward these targets . Within PGCs , Ror2 engagement by Wnt5a could lead to the redistribution of ROR2 at the cell surface , for example through ligand-receptor endocytosis [64]; this could rapidly lead to asymmetric ROR2 distribution between leading and trailing edges of the cell . Ror2 could ultimately reinforce the polarization response initiated by KitL-cKit in a number of different ways , such as by augmenting common downstream signaling components , such as cytoskeletal reorganization machinery , or perhaps by inhibiting the responsiveness to KitL in some regions of the cell . Understanding the relationship between Wnt5a-Ror2 and KitL-cKit at a mechanistic level is an important next step . Both tyrosine kinase receptors function in the development of multiple tissues as well as cancer; cKit is a proto-oncogene implicated in melanoma [65] and gastrointestinal stromal tumors [66] , and Wnt5a expression has been correlated with tumor invasiveness [25] . Therefore the relationship between these two pathways is likely to extend beyond PGCs .
All animal work was carried out in compliance with care and use standards at each institution . Ror2Y324C was identified in a recessive ENU screen at e9 . 5 for mouse mutants with PGC defects [23] . Other mouse strains used included: Bax ( MGI:1857429 ) , Wnt5a ( MGI:1857617 ) , and Oct4-ΔPE-GFP [67] with genotyping performed as described elsewhere [19] , [50] . Mice were maintained on C3H or mixed C3H/FvB genetic backgrounds . Embryos were generated in timed matings by monitoring for copulatory plugs . Pregnant females were sacrificed and embryos staged by the following anatomic landmarks: 27–33 somite pairs was designated e10 . 0 , 34–39 somite pairs as late e10 , 45–48 somite pairs and the appearance of the otic vesicle as e11 . 5 , and the presence of embryonic kidneys designated e12 . 5; gonad sex was determined by the appearance of tubules and the coelomic vessel in e12 . 5 males , and SRY genotyping [68] . MIT SSLP markers were used to map Ror2Y324C to chromosome 13 to a ∼10 Mb interval between D13 MIT 176 and D13 MIT 13 . Sequencing of the Ror2 ORF revealed an A to G transition at position 1203 , which creates a restriction site . Y324C genotyping was carried out by PCR amplification of a 238 bp fragment in 25 uL reactions heated to 95°C for 3 min , followed by 45 cycles of 94°C for 30 sec , 57°C for 1 min , 72°C for 30 sec , and a 7 min hold at 72°C using the following primers: 5′-ACC AGT GCT ACA ACG GCT CT-3′ and 5′-AGT TCC ACG CGT ACG TTT TT-3′ ) . Subsequent digestion 5 h with 3 U HpyCh4 V ( NEB ) produced fragment sizes 152 and 86 bp for the wild type allele , 152 , ∼50 and ∼30 bp for mutant allele . Embryos were dissected at e9 . 5–11 . 5 in cold PBS/0 . 2% BSA and the posterior fragment or gonads dissociated in 0 . 25% trypsin/EDTA for 3–5 minutes at 37°C followed by 1 mg/mL DNaseI for 5 min . For Ror2 intracellular flow cytometry , cells were prepared using the Cytofix/cytoperm Kit ( Beckton Dickson ) and stained at 1∶50 ( Santa Cruz Biotech A17 ) . Live cell staining was carried out in phenol red-free DMEM/2% fetal bovine serum/10 mM EDTA . Dead cells were excluded on the basis of Sytox Blue ( Invitrogen ) signal . PGCs , delineated as Oct4 ( ΔPE ) -GFP+ were sorted directly into lysis buffer and extracted with RNeasy Kit ( Qiagen ) , DNAse I treated , and reverse-transcribed with qScript ( Quanta Biosciences ) or Superscript III ( Invitrogen ) . PCR primers were designed with Primer Express software ( Applied Biosystems ) . Amplification was carried out with 50 or 100 cell equivalents of cDNA on a Mastercycler EP ( Eppendorf ) using the following primer sets: 5′-GACTTCAACAGCAACTCCCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′ for Gapdh; 5′- AATGCACAACTGCCATCTCC-3′ and 5′-AGGAATGCCTAGACTACTGGAAAA-3′ for KitL [69]; 5′-AGTCTGGAGACCATGTTTCTGAAG T-3′ and 5′-TACTCTTCTCGTTGGGAATACTCAATA-3′ for Oct4; 5′- GAGATCAGCTTGTCCAC-3′ and 5′- AGCATCGCCTCTTGCCGG-3′ for Ror2 , 5′- GCAGACCGAACGCTGTCATT-3′ and 5′- CCACAATCTCCGTGCACTTCT-3′ for Wnt5a . For Western blotting , day 9 . 5 embryos were lysed in RIPA buffer containing 1% Nonident P-40 , 0 . 25% Deoxycholate acid , 150 mM NaCl , 0 . 1% SDS , 50 mM HEPES ( pH 7 . 4 ) and proteinase inhibitor cocktail ( Roche ) . 40 ug protein was separated by SDS-PAGE gel and probed with anti-ROR-2 antibody ( Santa Cruz Biotech , H-76 ) . For immunofluorescence histology , embryos fixed in 4% paraformaldehyde were embedded in OCT and cryosectioned at10 um . Slides were blocked 1 h in 10% calf serum + 0 . 1% Tween in PBS and stained overnight @ 4°C in the blocking buffer followed by 3×15 minute washes in PBS . Primary antibodies used included SSEA1 ( Developmental Studies Hybridoma Bank , 1∶200 ) , Wnt5a ( R&D Systems AF645; 1∶20 ) , activated Caspase 3 ( Promega , G7481 , 1∶250 ) , phospho-histone H3 ( Sigma , clone HTA28 , 1∶200 ) , E-cadherin ( Invitrogen , 13–1900 , 1∶200 ) , Ror2 ( Santa Cruz Biotech A-17 1∶50 ) , Pericentrin ( Covance , PRB-432C , 1∶50 ) , GM130 ( Beckton Dickinson , monoclonal , 1∶100 ) , the latter was preceded by 2 min treatment in Ficin ( Invitrogen ) . Bromodeoxyuridine ( Abcam ab6326 , 1∶40 ) was preceded by treatment in 4N HCl for 10 minutes and 5 min in 0 . 1 M Borate buffer pH 8 . 6 . Secondary antibodies and fluorescent Phalloidin were purchased from Invitrogen were incubated for 1 hour at room temperature and used at 1∶200–1∶500 . Sections were mounted in Vectashield ( Vector Labs ) . For whole mount immunofluorescence , embryos or gonads were fixed in methanol∶dimethylsulfoxide ( 4∶1 ) at −20°C overnight , rehydrated and rocked @ 4°C overnight in PBSMT ( PBS/2% nonfat dry milk/0 . 5% Tween ) with antibodies to SSEA1 ( 1∶200 ) , cleaved PARP ( Cell Signaling #9544 , 1∶50 ) , phospho histone H3 ( 1∶50 ) , GCNA ( a kind gift of George Enders , undiluted supernatant ) ; triple washes in PBSMT were followed by overnight rocking with secondary antibodies diluted 1∶200 in PBSMT , followed again by washing 3× . Gonads were mounted on slides in Vectashield ( Vector Labs ) , whereas whole embryos were serially dehydrated and cleared in Methyl Salicylate for viewing . Oct4-ΔPE-GFP+ PGCs sorted from e9–9 . 75 embryos were seeded in chambered slides ( Lab-Tek II ) coated with 1 mg/ml Matrigel then incubated @ 37°C in 5% CO2 with DMEM/15% Knockout Serum Replacement ( Invitrogen ) , 1000 U/mL LIF ( Millipore ) , 5 uM Forskolin , and added 250 ng/mL WNT5a ( RnD ) , 50 ng/ml SCF ( Invitrogen ) . Gradients were produced by placing 100 ng/mL SCF in Matrigel along one edge of the chamber . Following culture , cells were fixed for 10 min in 4% paraformaldehyde and immunostained . Embryos were dissected in RPMI/10 mM Hepes/10% FBS at e8 . 0 with membranes intact . Heparin coated glass beads ( Sigma ) or Affygel Blue 100–200 mesh beads ( BioRad ) were washed3× in PBS , soaked 1 hr in PBS-BSA , 25 ug/mL SCF ( RnD ) , 25–50 ug/mL SDF1 ( RnD ) , 50 ug/mL WNT5a ( RnD ) or 20–50-fold concentrated WNT5A conditioned media , washed , and implanted into the proximal allantois , hindgut pocket or axial mesoderm with microforceps . Embryos were cultured in organ culture dishes ( Falcon ) containing 50% DMEM HG with Pen-Strep/50% Heat inactivated Rat Serum ( Taconic ) at 37°C in 5% C02 for 24 hours . Fast red staining was carried out as detailed elsewhere [70] . Brightfield imaging was performed on an Olympus MVX10 stereomicroscope . Confocal imaging was carried out with a 10× , 20× or 63× objective on a Leica SP5 TCS microscope equipped with 405 , 488 , 543 , 594 , and 633 nm lasers . Stacks were analyzed using Volocity ( Improvision ) . The number of PGCs in wholemount e10 . 5 immunostained embryos or e11 . 5–12 . 5 gonads was estimated using a measurement protocol created in Volocity 5 . 0 acquisition software . Objects were identified by in the SSEA1 or GCNA channel using the “Find Objects Using Standard Deviation ( SD ) Intensity” task , with a lower limit of 3 . 2–3 . 7 SDs above the mean . Holes were filled in objects , and those under 20 mm3 were excluded , and touching objects separated using a size guide of 200 mm3 in the GCNA channel or 750 mm3 in the SSEA1 channel . “Exclude Objects by Size” task was repeated to eliminate objects less than 20 mm3 created by the previous command . Objects were visually inspected to determine the approximate size cutoff for single objects . For gonads colabeled with antibodies against PHH3 or cleaved PARP , subsequent selection of colocalized objects was carried out using the intensity and colocalization functions . All measurement results were exported to Excel ( Microsoft ) for calculations . Clustered objects exceeding the defined threshold of single PGCs were summed and divided by the average PGC size . For instances in which over 20% of the total measured GCNA volume remained clustered , the quantity of PGCs was estimated by dividing this total volume over the average object size in well scattered specimens ( 300 um3 used here for GCNA ) . Cell axes lengths were quantified in confocal stacks visualized in Volocity using the measurement function and exported to Excel for analysis . Image reconstructions with the long axis marked were exported to ImageJ ( NIH ) for angle measurements . | Egg and sperm derive from precursors in the early embryo called primordial germ cells ( PGCs ) . The mechanisms underlying the migration of PGCs through the embryo to the forming gonads remain unclear . In a genetic screen , we identified a role for the receptor Ror2 and its ligand Wnt5a in promoting PGC colonization of the embryonic gonads . By ex vivo culture , we show that Ror2 acts autonomously in PGCs to enhance their polarized response to the chemotactic factor SCF . Asymmetric distribution of ROR2 within PGCs in vitro and in vivo suggests that signaling via Ror2 locally amplifies cell polarity in response to other directional cues . These studies identify a novel relationship between Ror2 and cKit signaling in polarized migration . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"embryology",
"organism",
"development",
"stem",
"cells",
"embryonic",
"stem",
"cells",
"genetic",
"screens",
"genetics",
"organogenesis",
"biology",
"genetics",
"and",
"genomics"
] | 2011 | Ror2 Enhances Polarity and Directional Migration of Primordial Germ Cells |
A century after its discovery , Chagas disease still represents a major neglected tropical threat . Accurate diagnostics tools as well as surrogate markers of parasitological response to treatment are research priorities in the field . The purpose of this study was to evaluate the performance of PCR methods in detection of Trypanosoma cruzi DNA by an external quality evaluation . An international collaborative study was launched by expert PCR laboratories from 16 countries . Currently used strategies were challenged against serial dilutions of purified DNA from stocks representing T . cruzi discrete typing units ( DTU ) I , IV and VI ( set A ) , human blood spiked with parasite cells ( set B ) and Guanidine Hidrochloride-EDTA blood samples from 32 seropositive and 10 seronegative patients from Southern Cone countries ( set C ) . Forty eight PCR tests were reported for set A and 44 for sets B and C; 28 targeted minicircle DNA ( kDNA ) , 13 satellite DNA ( Sat-DNA ) and the remainder low copy number sequences . In set A , commercial master mixes and Sat-DNA Real Time PCR showed better specificity , but kDNA-PCR was more sensitive to detect DTU I DNA . In set B , commercial DNA extraction kits presented better specificity than solvent extraction protocols . Sat-DNA PCR tests had higher specificity , with sensitivities of 0 . 05–0 . 5 parasites/mL whereas specific kDNA tests detected 5 . 10−3 par/mL . Sixteen specific and coherent methods had a Good Performance in both sets A and B ( 10 fg/µl of DNA from all stocks , 5 par/mL spiked blood ) . The median values of sensitivities , specificities and accuracies obtained in testing the Set C samples with the 16 tests determined to be good performing by analyzing Sets A and B samples varied considerably . Out of them , four methods depicted the best performing parameters in all three sets of samples , detecting at least 10 fg/µl for each DNA stock , 0 . 5 par/mL and a sensitivity between 83 . 3–94 . 4% , specificity of 85–95% , accuracy of 86 . 8–89 . 5% and kappa index of 0 . 7–0 . 8 compared to consensus PCR reports of the 16 good performing tests and 63–69% , 100% , 71 . 4–76 . 2% and 0 . 4–0 . 5 , respectively compared to serodiagnosis . Method LbD2 used solvent extraction followed by Sybr-Green based Real time PCR targeted to Sat-DNA; method LbD3 used solvent DNA extraction followed by conventional PCR targeted to Sat-DNA . The third method ( LbF1 ) used glass fiber column based DNA extraction followed by TaqMan Real Time PCR targeted to Sat-DNA ( cruzi 1/cruzi 2 and cruzi 3 TaqMan probe ) and the fourth method ( LbQ ) used solvent DNA extraction followed by conventional hot-start PCR targeted to kDNA ( primer pairs 121/122 ) . These four methods were further evaluated at the coordinating laboratory in a subset of human blood samples , confirming the performance obtained by the participating laboratories . This study represents a first crucial step towards international validation of PCR procedures for detection of T . cruzi in human blood samples .
A century after its discovery [1] Chagas disease still represents a health threat to an estimated 28 million people in the Americas , being the second highest illness burden among neglected tropical diseases [2]–[3] . The infection by the protozoan Trypanosoma cruzi can be acquired from blood-sucking triatomine bugs , blood transfusion , transplacental transmission or by the oral contamination foodstuffs by infected triatomine faeces [2]–[3] . Since 1990 , a series of international initiatives based on vector control , systematic screening of blood donors in all endemic countries , and detection and treatment of congenital transmission have been launched for control and elimination of Chagas disease . These strategies have led to significant reduction in the number of infected people worldwide . According to information from 21 countries where the disease is endemic , the number of infected persons today is estimated to be 7 , 694 , 500 , most of them at the chronic stage of disease [2]–[3] . Traditional parasitological procedures , such as xenodiagnosis and haemoculture are laborious and time-consuming and show poor sensitivities in cases of low-level parasitaemias , limiting their usefulness in diagnosis and monitoring of drug efficacy [4]–[6] . Since the past decade , the application of polymerase chain reaction ( PCR ) to detect T . cruzi directly in blood samples has opened new possibilities for the diagnosis of infection and evaluation of trypanocidal chemotherapy in different clinical and epidemiological settings [7]–[22] . These PCR procedures have revealed highly variable levels of sensitivity and specificity , depending on a number of technical factors such as , the volume of sample collected , the conditions of conservation of the sample , the methods used to isolate DNA , the parasite sequences and primers selected , the reagents used as well as the thermo-cycling conditions . Variability in PCR sensitivity could also be in part explained by the intermittent presence and quantity of circulating parasites at the time of blood collection . In addition , molecular targets from strains belonging to six different T . cruzi discrete typing units ( DTUs , [23] ) with dissimilar DNA content and gene dosage [22] , [24]–[25] have been used for molecular diagnosis by different laboratories . In addition , sequence polymorphisms within amplified fragments among strains from different DTUs may influence the efficiency of amplification [26]–[27] . Moreover , false negative findings due to interference of PCR inhibitory substances co-purified during lysis and DNA extraction of blood samples and false positive results mostly due to carry over DNA contamination [22] , [28] may arise . In this context , the assessment of the performances of currently available PCR tests for detection of T . cruzi infection in blood samples and DNA control sets was launched by expert laboratories in PCR detection of T . cruzi infection from different countries of America and Europe . We aimed to compare the performance of currently used PCR strategies for detection of T . cruzi DNA in sets of blind samples , including purified DNA from reference culture stocks from different T . cruzi discrete typing units , human blood samples spiked with cultured parasite cells and clinical samples from seropositive and seronegative patients from different endemic countries , in order to select the best performing tests for validation .
An access database form was distributed to the participants to standardize reporting of results . Those laboratories performing more than one PCR test per sample sent a separate report for each test . The results were analyzed by using SAS Software and Microsoft Excel . Due to the exploratory nature of the study , a descriptive analysis of results is provided . For set A , the following parameters were evaluated: 1 ) specificity ( Sp ) : the proportion of negative PCR results in the three negative samples , 2 ) coherence: ( Co ) the ability of reporting positive PCR findings in a consecutive way , from the highest to the lowest detected DNA concentration for each series of DNA dilutions of parasite stocks and 3 ) the detection limits ( DL ) for each stock . A test was defined as Good Performing Method ( GPM ) if it was 100% specific and coherent and capable of detecting 10 fg/ul or less DNA for all parasite DTU stocks . For set B the same parameters were evaluated: Sp , Co and DL . A test was defined as GPM if it was 100% specific and coherent and capable of detecting 5 parasite equivalents/mL of Guanidine Hidrochloride-EDTA treated blood or less . For each sample of set C , a consensus PCR result was obtained on the basis of the reports by GPM tests in sets A and B , as done in other PCR interlaboratory studies [34] . A sample was considered PCR positive by consensus if more than 50% of the GPM gave positive results and PCR negative if more than 50% of GPM tests gave negative results . Those samples for which 50% of the GPM methods gave positive reports and 50% gave negative ones were considered indeterminate . The sensitivity , specificity , accuracy and kappa index of the different PCR tests were calculated by using 1 ) the above mentioned consensus PCR results and 2 ) the serological diagnosis as the reference methods . Inter-observer kappa coefficients were calculated using GraphPad Software on-line statistical calculators ( http://www . graphpad . com/quickcalcs/kappa1 . cfm ) . Kappa values<0 . 01 indicate no concordance , those between 0 . 1 and 0 . 4 indicate weak concordance , those between 0 . 41 and 0 . 60 indicate clear concordance , those between 0 . 61 and 0 . 80 indicate strong concordance , and those between 0 . 81 and 1 . 00 indicate nearly complete concordance . Accuracy was calculated as reported [35] .
Table 2 shows the data obtained by the 48 PCR tests on DNA dilutions from the 3 parasite stocks representing DTUs I , IV and VI . The seven PCR tests targeting sequences other than Sat-DNA or kDNA failed to detect the most concentrated DNA sample ( 10 fg/ul ) of one , two or all three parasite stocks ( Tests C3 , C4 , C6 , S1 and S3; Detection Limit = ND ) , or reported false positive findings in the negative controls without DNA ( Tests C5 and U2 ) and thus were not included in the following analysis . Out of the 41 tests based on kDNA ( 28 tests ) or Sat-DNA sequences ( 13 tests ) , 25 ( 51 . 2% ) provided specific and coherent results for all three parasite stocks ( Sp = Y , Co = Y , Table 2 ) . Fourteen of them targeted kDNA , representing 50% of the reported kDNA-PCR tests and 11 targeted Sat-DNA , representing 84 . 6% of Sat-DNA PCR tests . These data indicated that PCR tests based on Sat-DNA sequences were more specific than those based on kDNA . Figures 1A and 1B show the distribution of the detection limits ( DL ) of the above mentioned 11 Sat-DNA PCR and 14 kDNA PCR tests , respectively , for each T . cruzi stock . Analysis of T . cruzi I DNA series: Nine out of 11 Sat-DNA- and all 14 kDNA-PCR tests were capable of detecting at least the most concentrated T . cruzi I DNA sample ( Figure 1 A and B , grey bars ) and 2 Sat-DNA and 8 kDNA-PCR tests could detect 0 . 1 fg/µl of T . cruzi I DNA . The lowest detection limit for T . cruzi I DNA was 0 . 01 fg/ul obtained by laboratory W using conventional kDNA-PCR ( Table 2 and Figure 1B ) . Thus , kDNA- PCR tests were more sensitive than Sat-DNA PCR tests to detect T . cruzi I DNA . Analysis of T . cruzi IV DNA series: 8 out of 11 Sat-DNA- and all 14 kDNA-PCR tests were capable of detecting the most concentrated T . cruzi IV DNA sample ( Figure 1 A and B , black bars ) . The lowest detection limit ( 1 fg/µl ) was reached by three Sat-DNA- and six kDNA-PCR tests , suggesting similar analytical sensitivities of methods based on both molecular targets to detect T . cruzi IV DNA . Analysis of T . cruzi VI DNA series: All 11 Sat-DNA- and 13 out of 14 kDNA-PCR tests were capable of detecting the most concentrated T . cruzi VI DNA sample ( Figure 1 A and B , white bars ) . The only test that did not detect Cl-Brener DNA amplified the constant kDNA region ( G1 , kDNAc , Table 2 ) . The lowest detection limit ( 0 . 01 fg/µl ) was obtained by 2 Sat-DNA PCR tests ( Z and F1 , Table 2 ) followed by 0 . 1 fg/µl obtained by 2 conventional kDNA-PCR tests ( R and W , Table 2 ) . Overall , the reported PCR tests were less sensitive for detecting DNA from the T . cruzi IV reference stock . Twenty PCR tests showing specific and coherent results and detecting at least the most concentrated DNA samples from each of the parasite stocks were considered Good Performing Methods for Set A ( bold fonts , Table 2 ) . They comprised 53 . 8% of 13 Sat-DNA-PCR and 46 . 42% of 28 kDNA-PCR tests . Ten GPM tests used in-house ( IH ) PCR mixtures and 10 used commercial master mixes ( Kt ) , representing 35 . 7% of the 28 IH and 76 . 9% of the 13 Kt PCR reagent mixes . In addition , 12 GPM used conventional amplification and eight used real time PCR ( C and RT in Table 2 ) , representing 38 . 7% of 31 C and 72 . 7% of 11 RT tests . These data showed that commercial master mixes and real time PCR offered better PCR performance in purified DNA samples . Out of the 44 PCR tests reported for spiked Guanidine Hidrochloride-EDTA blood samples , the three tests targeting sequences other than Sat-DNA or kDNA were not further analyzed , because they failed to detect the most concentrated sample ( S1 and S3 tests ) or showed false positive findings in the non-spiked control ( U2 test ) ( Set B , Table 2 ) . Twenty five out of 41 PCR tests based on kDNA and Sat-DNA sequences showed specific , coherent results and detection limits of at least 5 par/ml ( GPM , bold fonts , Table 2 , Set B ) . They included 14 kDNA and 11 Sat-DNA PCR tests , representing 50% of 28 kDNA and 84 . 6% of 13 Sat-DNA based tests . Ten GPM used in-house extraction methods and 15 used DNA extraction kits , representing 41 . 6% of 24 IH and 62 . 5% of 24 Kt tests . Thus , methods using commercial DNA extraction and Sat-DNA as amplification target resulted in better performance . Procedures based on kDNA presented more variation in sensitivity than Sat-DNA tests ( Figure 1C , white bars ) . The smallest detected concentration was 5×10−3 par/ml , recorded by three laboratories using conventional kDNA-PCR after DNA extractions with Chelex resine , a blood extraction kit or solvent extraction with Phenol ( LbE , LbL1 and LbW , respectively , Table 1 ) . Tests based on Sat-DNA presented sensitivities between 0 . 05 and 0 . 5 par/ml ( 10/11 tests , Figure 1C , black bars ) with the only exception of one test based on solvent DNA extraction and IH conventional Sat-DNA PCR ( C2 , Table 1 ) that reached a detection limit of 5×10−5 par/mL ( Table 2 , and Figure 1C ) . Out of the 44 PCR tests performed on clinical samples of Set C , a 18s-rDNA PCR ( S3 ) and a SL-DNA PCR ( C6 ) tests did not detect any positive sample and the 24s α rDNA-PCR test ( U2 ) had only 40% of specificity . Consequently , they were not included for subsequent analysis . The levels of agreement among the 41 remaining PCR tests on the reports for each clinical sample are presented in Table 3 . For each sample , 3 series of consensus PCR results were calculated: 1 ) consensus based on the 28 kDNA-PCR tests , 2 ) consensus based on the 13 Sat-DNA PCR tests and 3 ) consensus based on the 16 tests defined as GPM in both sets A and B samples . The sensitivity of consensus kDNA-PCR was 65 . 62% ( 21 PCR positive/32 seropositive samples ) , that of consensus Sat-DNA was 62 . 5% ( 20 PCR positive/32 seropositive samples ) and that of consensus GPM was 56 . 25% ( B ( 18 PCR positive/32 seropositive samples ) being 4 samples indeterminate ( 13 , 18 , 20 , 21 , Table 3 ) because the levels of agreement among GPM tests was 50% . The individual performance of the 41 PCR tests was evaluated in comparison with the consensus PCR results reached by the 16 GPM in sets A and B ( 18 PCR positive , 20 PCR negative samples ) and in comparison with serologic diagnosis ( 10 seronegative , 32 seropositive samples ) ( Table 4 ) . There was a high variability among the performances of the different methods ( Table 4 ) . The median values of sensitivity , specificity and accuracy of the 41 tests were 72 , 77 . 5 and 68 . 4% , respectively in comparison to consensus GPM PCR reports , and 59 . 4 , 70 and 59 . 5% , respectively in comparison to serological diagnosis ( Table 4 ) . Four GPM showed the best operational parameters in set C ( Table 4 ) . Tests LbD2 and LbD3 used solvent DNA extraction followed by conventional hot-start and Real time PCR targeted to Sat-DNA , respectively ( primer pairs TCZ-F/TCZ-R ) . Test LbF1 used a commercial kit for DNA extraction based on glass fiber columns and Real Time PCR targeted to Sat-DNA ( primer pairs cruzi 1/cruzi 2 and TaqMan probe cruzi 3 ) and test LbQ used solvent DNA extraction and conventional hot-start PCR for kDNA ( primer pairs 121/122 ) . The performance of these four tests was further evaluated at the coordinating laboratory on a subset of samples from seropositive and seronegative patients , analysed in four independent experiments ( Table 5 ) . Examples of the outputs of each method are shown in Figure 2 . The degree of concordance among the reported results by the BPM was between 87 . 5% and 90 . 62% . This intralaboratory evaluation showed that the selected methods depicted similar operational parameters than when performed by the corresponding laboratories in the international study ( Tables 4 and 6 ) .
In set A , GPM included kDNA and sat-DNA PCR tests in similar proportions . However , Sat-DNA PCR tests were less sensitive than kDNA-PCR tests to detect T . cruzi I DNA . This is most likely due to the fact that T . cruzi DTU I harbors approximately four to ten-fold less number of satellite repeats than DTUs II , V and VI , which has been demonstrated by different molecular approaches [22] , [45] . Regarding T . cruzi IV that also harbors a lower dosage for satellite sequences [22] , similar analytical sensitivities of kDNA and Sat-DNA PCR tests were observed , being lower than that obtained for the other two tested DTU representative stocks . The genome size and relative DNA contents of Can III cl1 ( 116 . 44 Mb , 95% CI 110 . 4–122 . 63 and 1 . 090 , respectively ) and CL Brener ( 108 . 55 Mb , 95% CI 101 . 41–115 . 89 and 1 . 017 ) are similar [25] , although Can III cl1 harbors about 5 fold less satellite repeats than CL-Brener [22] . The relative contribution of the nucleus and kinetoplast has not been measured but normally , kDNA represents 20–25% of the total DNA content [46] . There are no available data regarding the number of minicircles in the kinetoplast of Can III cl1 , so it could be speculated that the lower analytical sensitivity of most PCR tests to detect DNA from this clone respect to the other ones , could be due to a lower minicircle copy dosage . Set B allowed evaluation of the influence of DNA extraction procedures in the PCR performance . A 72 . 2% of DNA extraction methods based on commercial kits led to GPM in set B , whereas 57 . 8% of phenol-chlorophorm extracted DNA led to GPM reports . These findings indicated that Guanidine Hidrochloride-EDTA blood was suitable for DNA extraction using kits based on lysis buffers containing Guanidine salts . Out of the 25 GPM in set B , 14 had a sensitivity of 0 . 05 par/ml , which should be adequate for diagnosis of infection in chronic patients [20] . Indeed , the necessary detection limit in chronic Chagas disease has been stated as one parasite cell in 10 mL of blood [20] . Analysis of PCR performance in set C clinical samples showed that the four best performing tests presented strong concordance with respect to consensus PCR results obtained by the 16 tests defined as GPM in sets A plus B ( kappa index between 0 . 7 and 0 . 8 ) . Out of them , three tests targeted sat-DNA sequences and only one targeted kDNA . These data are in agreement with previous works showing that PCRs targeting Sat-DNA performed better than PCRs targeting kDNA sequences [47]–[49] , although kDNA based PCR has been more widely used [20] . Moreover , two of the sat-DNA best performing tests used Real Time PCR , one with a Sybr Green fluorescent dye ( LbD2 ) and the other one with a TaqMan probe ( LbF1 ) . It must be pointed out that LbD2 and LbD3 tests were performed by the same laboratory . Out of the 16 GPM performed by 11 different laboratories , 3 laboratories performed two methods ( LbD , LbK and LbP ) and one lab developed 3 tests ( LbG ) . These data point to laboratory dependence concerning PCR performance , which may be due to multiple factors including technical expertise , correct use of quality controls , instrumentation and reagents . For example , tests LbF1 , LbS2 ( GPM ) and LbZ ( not GPM ) were all based on sat-DNA Real Time PCR using the same primer pair ( cruzi 1 – cruzi 2 ) , differing in the trade marks of the DNA extraction and Master Mix kits . Some tests shown as GPM in sets A+B had very low sensitivities in set C ( LbK2 , LbP2 , LbV1 , Table 4 ) , suggesting that quality controls might have failed to distinguish false negative clinical samples . A major drawback of most PCR tests is that they do not contain an internal amplification control ( IAC ) . An IAC is a non target DNA sequence present in the same sample reaction tube , which is co-amplified simultaneously with the target sequence [50] . In a PCR without an IAC , a negative result can indicate that the reaction was inhibited , as a result of the presence of inhibitory substances in the sample matrix . The presence of PCR inhibition in Guanidine Hidrochloride-EDTA treated blood samples has been described [22] . The European Standardization Committee ( CEN ) , in collaboration with International Standard Organization ( ISO ) has proposed a general guideline for PCR testing that requires the presence of IAC in the reaction mixture [51] . Therefore , only IAC-containing PCRs should undergo multicentre collaborative trials , which is a prerequisite for validation . Some other tests shown as GPM in sets A+B had very low specificities ( LbI2 , LbW , LbG2 , LbG3 , Table 4 ) . Amplicon carry-over contamination is one of the most probable causes . PCR master mixes with dUTP and Uracil-DNA N-glycosylase ( UNG ) intended to abolish amplicon carry-over contamination were used in some tests ( LbF , LbG , LbL , LbS , Table 1 ) . Nevertheless , some of them did not show good specificity in set C ( LbG2 , LbG3 , LbG4 , LbL1 , LbL2 , Table 4 ) , suggesting that problems during sample processing , such as sample to sample contamination could have arisen . The median values of the sensitivities obtained in testing the Set C samples with the 16 tests determined to be GPMs by testing the Set A and Set B samples varied considerably depending on the clinical characteristics of the persons from whom the Set C samples were drawn . Indeed , sensitivity was 100% ( 25-75p = 100−100 ) for immunosuppressed heart transplanted pts , 56 . 5% ( 25-75p = 39 . 1–66 . 3% ) for asymptomatic and 57 . 1% ( 25-75p = 14–75% ) for symptomatic chronic Chagas disease patients . These data point to the limitations of PCR strategies for diagnosis of patients at the chronic phase of disease . In addition , some of these samples had been stored at 4°C for at least two years before this PCR study; thus higher PCR positivity might be obtained in prospective clinical studies but it is unlikely that the current PCR methods will have a sensitivity comparable to serological assays for diagnosis of chronic Chagas disease . The four BPM methods were transferred to the Coordinating laboratory , where they were evaluated in a subset of clinical samples , each one tested in four independent assays , obtaining good concordance and confirming the performance reported by the participating laboratories in the previous international study ( Tables 5 and 6 ) . Further work is still needed to validate them through prospective studies in different settings . In this regard , this collaborative evaluation constitutes a starting point towards technical improvement and development of an international standard operating procedure ( SOP ) for T . cruzi PCR . In this context , the BPMs could be recommended for alternative diagnostic support , such as in the following settings: a ) post-treatment follow-up of patients to look for failure of therapy to achieve parasitologic response [12] , [14]–[16] , [20] , [22]; b ) diagnosis of congenital Chagas disease in newborns in whom the presence of maternal anti-T . cruzi antibodies make serological studies useless [11] , [15] , [48]; c ) early diagnosis of reactivation after organ transplantation of T . cruzi infected recipients under immunosuppressive therapy [18] , [41] , d ) differential diagnosis of Chagas reactivation in patients with AIDS [39] , and e ) suspicion of oral transmission [52] . Moreover it can be useful for post-treatment follow-up of experimental animals to look for failure of therapy to achieve parasitologic cure [53]; in diagnosis in naturally infected triatomines or triatomines used for xenodiagnosis , since it has been shown that PCR tests are much more sensitive than microscopic examination of intestinal contents [37] , [54]; and diagnosis of T . cruzi infection in mammalian reservoirs for which serologic tools have not been developed [38] . | A century after its discovery , Chagas disease , caused by the parasite Trypanosoma cruzi , still represents a major neglected tropical threat . Accurate diagnostics tools as well as surrogate markers of parasitological response to treatment are research priorities in the field . The polymerase chain reaction ( PCR ) has been proposed as a sensitive laboratory tool for detection of T . cruzi infection and monitoring of parasitological treatment outcome . However , high variation in accuracy and lack of international quality controls has precluded reliable applications in the clinical practice and comparisons of data among cohorts and geographical regions . In an effort towards harmonization of PCR strategies , 26 expert laboratories from 16 countries evaluated their current PCR procedures against sets of control samples , composed by serial dilutions of T . cruzi DNA from culture stocks belonging to different lineages , human blood spiked with parasite cells and blood samples from Chagas disease patients . A high variability in sensitivities and specificities was found among the 48 reported PCR tests . Out of them , four tests with best performance were selected and further evaluated . This study represents a crucial first step towards device of a standardized operative procedure for T . cruzi PCR . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"molecular",
"biology",
"microbiology/parasitology"
] | 2011 | International Study to Evaluate PCR Methods for Detection of Trypanosoma cruzi DNA in Blood Samples from Chagas Disease Patients |
Neuronal activity is mediated through changes in the probability of stochastic transitions between open and closed states of ion channels . While differences in morphology define neuronal cell types and may underlie neurological disorders , very little is known about influences of stochastic ion channel gating in neurons with complex morphology . We introduce and validate new computational tools that enable efficient generation and simulation of models containing stochastic ion channels distributed across dendritic and axonal membranes . Comparison of five morphologically distinct neuronal cell types reveals that when all simulated neurons contain identical densities of stochastic ion channels , the amplitude of stochastic membrane potential fluctuations differs between cell types and depends on sub-cellular location . For typical neurons , the amplitude of membrane potential fluctuations depends on channel kinetics as well as open probability . Using a detailed model of a hippocampal CA1 pyramidal neuron , we show that when intrinsic ion channels gate stochastically , the probability of initiation of dendritic or somatic spikes by dendritic synaptic input varies continuously between zero and one , whereas when ion channels gate deterministically , the probability is either zero or one . At physiological firing rates , stochastic gating of dendritic ion channels almost completely accounts for probabilistic somatic and dendritic spikes generated by the fully stochastic model . These results suggest that the consequences of stochastic ion channel gating differ globally between neuronal cell-types and locally between neuronal compartments . Whereas dendritic neurons are often assumed to behave deterministically , our simulations suggest that a direct consequence of stochastic gating of intrinsic ion channels is that spike output may instead be a probabilistic function of patterns of synaptic input to dendrites .
The appropriate level of physical detail required to understand how complex processes such as cognition and behavior emerge from more simple biological structures is unclear [1] , [2] . For example , while it is possible to account for certain aspects of nervous system function using models that represent each neuron as a simple integrate and fire device , it is increasingly clear that this approach does not capture the full range of computations that many real neurons carry out [3] , [4] . Dendritic and axonal morphology are defining features of neuronal cell types and have important influences on the computations that a neuron performs [5] . Differences in morphology determine how neurons respond to synaptic input and are sufficient to produce distinct patterns of spontaneous activity [6] and degrees of action potential back-propagation from the soma into the dendrites [7] . Cable theory and compartmental modeling provide a foundation for predicting the propagation of electrical signals in the dendrites and axons of neurons [8] , [9] . However , while the assumption that transitions between open and closed states of ion channels can be treated as a deterministic process may be sufficient for some purposes , recent evidence suggests that stochastic transitions between the states of individual ion channels could influence computations carried out by neurons [10]–[17] . Stochastic opening and closing of ion channels causes ‘noisy’ fluctuations in the current or voltage recorded from a neuron [18]–[20] . While cable theory suggests that fluctuations of this kind might be particularly important in fine structures such as axons and dendrites [21] , we nevertheless know very little about how neuronal morphology and stochastic gating of ion channels interact to determine how neurons respond to synaptic input . Given the difficulty of reducing detailed morphological models to simple analytical forms that could also incorporate stochastic gating of individual ion channels [22] , experimentally constrained numerical simulations will be important to enable these issues to be explored systematically . Investigation of stochastic ion channel gating using numerical simulations has been limited by trades-offs between simulation accuracy and computation time [22] . A simple approach is to add noise sources to deterministic models . However , as ion channels have multiple functional states with transitions that often depend on the membrane voltage [11] , [15] , [22] , [23] , this may not accurately account for the noise introduced by ion channel currents . A more accurate alternative is to explicitly model transitions between different functional states for each ion channel on a neuron's membrane . However , for neurons with complex axonal or dendritic architectures there are two substantial obstacles to this approach . First , typical central neurons express large numbers of ion channels and simulations must be repeated many times to obtain statistically valid descriptions [24] . This is a formidable computational task and even relatively straightforward simulations of the consequences of stochastic channel gating can require substantial computing time ( see e . g . [12] , [13] ) . Second , each neuronal ion channel occupies a specific location on the extra-cellular membrane , whereas most neuronal models represent the distribution of ion channels as the density of a deterministic conductance across an area of membrane . Although this formalism has been successful for simulating many aspects of neuronal activity , it is of less use for models that explore the consequences of the localization of individual ion channels , for example to evaluate the macroscopic effects of short range interactions between ion channels and other signaling molecules [25] , or the consequences of spatially heterogeneous distributions of ion channels within relatively small sub-cellular structures such as dendritic spines and axon terminals [26] , [27] . To address the functional consequences of stochastic ion channel gating in neurons with extensive dendritic or axonal arborizations we developed a parallel stochastic ion channel simulator ( PSICS ) , which enables efficient simulation of the electrical activity of neurons with complex morphologies and arbitrary localization of stochastic ion channels on the extracellular membrane , while also addressing limitations of previous approaches . We have also developed an interactive tool ( ICING ) for visualization and development of models of neurons containing uniquely located ion channels . Here , we illustrate the use of PSICS and ICING , outline the computational strategies used and provide benchmark data for evaluation . We then identify previously unappreciated differences between the effects of stochastic ion channel gating on somatic and dendritic membrane potential activity in several different morphological classes of neuron . We show that the consequences of stochastic gating depend on dendritic morphology and suggest novel functional roles for the kinetics of ion channel gating . Using a previously well-validated realistic model of a CA1 pyramidal neuron we demonstrate that stochastic ion channel gating influences spike output in response to dendritic synaptic input . We show that stochastic gating of axonal or dendritic ion channels substantially modifies synaptically driven dendritic and axonal spike output , with stochastic gating of voltage-dependent sodium and potassium channels having the greatest impact and hyperpolarization-activated channels the least . By demonstrating that neuronal responses to dendritic synaptic input can be intrinsically probabilistic , these results offer a new and general perspective on synaptic integration by central neurons . Full documentation for PSICS/ICING as well as the software , source code and examples are available from the project website ( http://www . psics . org ) .
To investigate the functional consequences of stochastic ion channel gating for neurons with complex dendritic or axonal morphologies , we first developed new software tools that enable accurate , fast simulation ( PSICS ) and visualization ( ICING ) of neuronal models that contain stochastically gating ion channels . The organization and development of the new software tools are described in Text S1 , Figure S1 and in more detail on the project website ( http://www . psics . org ) . Here , we briefly outline novel features of model specification and visualization , before describing key benchmark data and simulation experiments that evaluate the functional impact of stochastic ion channel gating in different neuronal cell types . The new software uses a simple XML file structure that enables components of a model either to be constructed manually , to be configured using a graphical interface ( Figure 1A ) , or in the case of ion channels and morphologies to be imported from other programs and databases that allow saving of models in the NeuroML format . For example the morphology of the model CA1 pyramidal neuron shown in Figure 1 was downloaded as a NEURON simulation from the modelDB website ( http://senselab . med . yale . edu/modeldb ) and exported from NEURON as a . xml file . Similar methods can be used to import models developed with Neuroconstruct ( http://www . neuroconstruct . org/ ) [28] . To specify the membrane conductance we adopted a new approach in which the location of each individual ion channel is first uniquely determined ( Figure 1 ) . This approach is complementary to that of the program MCell [29] , which simulates movement and reactions of molecules within and around cells . In contrast , other neuronal modeling software approximates ion channel location as an average conductance density across a region of membrane . Before simulations are run in PSICS the neuron is discretized into sections that are then treated as isopotential compartments . As neurons are rarely at steady-state and have conductance that varies with membrane voltage , we implemented a discretization procedure that balances the capacitive charging rates for adjacent compartments ( see Methods ) . The granularity of the discretization process is set by the user and determines the number of channels in a particular compartment . After discretization the PSICS simulation engine will by default compute the activity of the population of channels in an isopotential compartment , rather than the activity of each individual ion channel . Modifying the granularity of the discretization process changes the number of channels per compartment , but not the actual distribution or location of channels in the model . Since the presently available tools for visualization and development of neuronal models are aimed primarily at deterministic simulations , we developed a graphical tool ( ICING ) to allow display and manipulation of neuronal models with complex three-dimensional architectures and many discrete membrane ion channels ( Figure 1 ) . ICING reads neuron morphologies specified either in NeuroML or as . swc files generated by the Neurolucida reconstruction program and used by the NeuroMorpho . org database ( http://neuromorpho . org/ ) . This enables components of a PSICS model to be visualized and edited . For example , to: 1 ) specify the size of compartments to use for the simulation , 2 ) select ion channels to be included in the model neuron , 3 ) select sections of the model for insertion of a particular ion channel class , 4 ) set rules that dictate the distribution of ion channels in their designated sections . The model neuron and its associated ion channels can be displayed in a variety of formats , for example to emphasize labeled sub-regions of the model ( Figure 1A ) , to illustrate the compartmental boundaries in a model ( Figure 1B ) , or to provide a detailed 3-dimensional exploration of the neuron morphology and ion channel distribution ( Figure 1C–D ) . We represent ion channels using Markov models , in which each ion channel may be in one of a number of discrete states with the probability of transition to any other state determined independently of the channel's previous history [24] , [30] , [31] . To efficiently simulate stochastic transitions between states of a channel we developed a modified version of the tau leap method ( see Methods ) [32] , [33] . The algorithm we use is equivalent to sampling an exact realization of the number of channels in a particular state at the end of each time step . In principle this results in shorter simulation times than algorithms that track the exact times of transitions between states [34] , [35] , or methods that permit a maximum of one transition per ion channel during each step [11] , [23] . To further reduce the simulation time the algorithm considers only channels with a non-negligible probability of making a transition during a particular step ( see Methods ) . At any particular sample time point and membrane potential , the tau leap algorithm should not produce any systematic error in the mean or variance of the current . However , the modified tau leap algorithm will not explicitly represent transitions that take place between time-points . We show below how this algorithm is particularly advantageous for current-clamp simulations in which high frequency current fluctuations are filtered by the neuronal membrane . We will also address how the choice of simulation parameters determines the accuracy and computation time . To first evaluate the modified tau leap algorithm for stochastic simulations we consider a simple three-state Na+ channel model recorded with an ideal voltage-clamp ( Figure 2 and Figure S2 ) . At a fixed membrane potential the simulated current through deterministic Na+ channels is constant , whereas the equivalent stochastic simulation reveals large fluctuations in the Na+ channel current ( Figure 2A ) . With sufficiently long periods of simulated channel activity , the mean amplitude of the stochastic current converges to the amplitude of the deterministic current ( Figure 2B ) and the estimated variance of the stochastic current converges to the value predicted from the number of channels and their single channel current amplitude ( Figure 2C ) . In deterministic simulations positive voltage steps from a negative holding potential elicit smoothly varying inward currents that activate rapidly , inactivate and are followed by a resurgent component after repolarization to the negative holding potential ( Figure S2 ) . In corresponding stochastic simulations the current response contains step-like fluctuations and differs from trial to trial , with the average waveform over many trials converging on the equivalent deterministic waveform ( Figure 2D ) . To determine whether the expected number of single channels and their single channel conductance could be retrieved from the simulated macroscopic currents , we carried out variance-mean analysis [36] , [37] . Both parameters could be estimated from either the activating ( Figure 2E ) or the inactivating phase of the current ( Figure 2F ) . Estimates of the number of single channels ( Figure 2G ) and their single channel conductance ( Figure 2H ) converged over many simulations onto their predicted values . These data demonstrate that our modified tau leap algorithm accurately simulates stochastic voltage-gated ion channel activity . This is further illustrated by the convergence towards zero of the error for the fit of the variance-mean function ( Figure 2I ) . On the order of 104 simulations were required for the fits to reliably converge to within 1% of the actual values , highlighting the importance of obtaining large numbers of repeated observations for estimating single channel properties using variance-mean analysis . Estimates obtained from 104 or fewer simulations varied around the actual values depending on the number of simulations used ( Figure 2G–I ) , suggesting an additional use of PSICS to quickly simulate the range of errors likely for estimates obtained from variance-mean analysis of macroscopic currents generated by channels with different gating schemes . Before comparing simulations of neurons with different morphologies , we first established the accuracy of simulation of current and voltage propagation using standard compartmental models for which there are analytical descriptions of the equivalent cable structures [38] . We examined the membrane potential of a simple cable containing stochastic leak Na+ and K+ channels ( Figure 3A ) . While fluctuations in the membrane potential are very small when both leak channels have a very small single channel conductance ( 0 . 01 pS ) , when either single channel conductance is increased to physiological values ( >1 pS ) , there is a substantial increase in the membrane noise ( Figure 3B ) . Although of physiological relevance , this noise makes comparison with analytical results problematic and therefore for validation of stochastic simulations we used a single channel conductance of 0 . 01 pS . Membrane potential responses to injection of a current step at one end of the cable , simulated with PSICS using either stochastic or deterministic ion channels , were effectively identical to the analytical result ( Figure 3C ) . Moreover , using a model of a branching dendrite ( Figure 3D ) , stochastic or deterministic ion channel simulations with PSICS also accurately reproduce the predicted voltage change in response to current injection ( Figure 3E ) . Thus , PSICS accurately simulates passive propagation of signals in compartmental models of cable structures , and when stochastic ion channels have very small single channel conductance the electrical behavior of a multi-compartment model is similar to models containing deterministic ion channels . We next assessed simulation of excitable neurons . In a model of a cylinder containing active membrane conductances [38] , simulations with PSICS that used deterministic ion channel models or stochastic implementations of channels with very small single channel conductance , produced essentially identical results to well established deterministic simulation software ( Figure 3F ) . By contrast , when we increased the single channel conductance to more physiological values , we found that while the action potential waveform was similar , the stochastic ion channel gating introduced jitter into the timing of the action potentials , such that reproducible timing of spike firing was not maintained between trials ( Figure 3F ) . We also compared simulation of an excitable model of a CA1 pyramidal cell shown in Figure 1 , with published data obtained with the same model [39] . The initiation and back propagation of action potentials were reproduced by simulation of this model with PSICS using either deterministic or stochastic ion channels ( data not shown ) . While open probability and single channel conductance influence the amplitude of current fluctuations generated by stochastic ion channel gating , little attention has been given to the functional impact of channel kinetics or of interactions between channel properties and the membrane capacitance . The simplified models we used to evaluate PSICS also allowed us to begin to investigate these issues . To avoid non-linearities from voltage-dependent gating , we simulated single-compartment models that contain only passive leak Na+ and K+ channels . Each channel has one open and one closed state , with an open probability of 0 . 7 , and the relative density of the channels was adjusted to produce a resting membrane potential of −60 mV . We compared a version of the model in which the forward and reverse rate constants for entering the open state were 0 . 07 ms−1 and 0 . 03 ms−1 ( slow gating ) with a version in which the corresponding rate constants were 7 ms−1 and 3 ms−1 ( fast gating ) . The model containing the slow gating channels produced membrane currents in voltage-clamp , or membrane potentials in current-clamp , that fluctuated at frequencies below approximately 15 Hz ( Figure 4A ) . By contrast , channels with faster gating kinetics produced current fluctuations with similar total power , but smaller amplitude at low frequencies ( <∼15 Hz ) and larger amplitude at higher frequencies ( >∼15 Hz ) ( Figure 4B ) . In current-clamp simulations , the corresponding high-frequency membrane potential fluctuations were filtered by the membrane capacitance . As a result , the membrane potential fluctuations span a similar range of frequencies to the slow gating model , but have substantially smaller amplitude ( Figure 4B ) . Thus , the functional impact of stochastic channel gating is determined by gating kinetics in conjunction with the membrane capacitance , as well as by open probability and single channel conductance . To examine how the choice of a suitable simulation time step is constrained by these properties , we initially used the simple passive models described above . For the model containing slow leak channels , simulations with time-steps as large as 1 ms reproduced the dominant components of the power spectra of current and voltage fluctuations ( Figure 4A ) . By contrast , for the model containing fast gating channels , simulation time-steps of 0 . 01 ms were required to satisfactorily simulate fluctuation of voltage-clamped currents , whereas time steps of duration up to 0 . 1 ms were sufficient to simulate membrane potential fluctuations in current-clamp conditions ( Figure 4B ) . Thus , selection of the simulation time-step requires evaluation of the recording configuration , the power spectra of channel activity , the membrane time constant and the kinetics of the simulated channels . Since simulation of complex neuronal morphologies can take considerable time , even using optimized computational algorithms , before simulating neuronal morphologies we first investigated strategies to minimize the time required for simulations without affecting accuracy of the results . We evaluated a stochastic implementation of the Hodgkin-Huxley Na+ channel model in a single compartment voltage-clamped at a fixed potential ( Figure 4C–D ) . With simulation time-steps in the range of 1–1000 µs , currents simulated using PSICS had mean and variance that correspond well to the predicted values ( Figure 4C ) . However , as the duration of the time-step is increased , the power spectra of the simulated currents reveal aliasing-like effects and failure to accurately simulate high frequency fluctuations ( Figure 4D ) . Thus , as we expect from the properties of the tau leap algorithm , longer time steps will produce currents with the correct variance and mean amplitude , but will not accurately simulate high frequency components of current fluctuations . To determine if improvements in simulation efficiency expected from use of the modified tau leap algorithm and an optimized computational core translate into practical reductions in simulation time , we compared the time required for simulations using PSICS to simulations run in the widely used NEURON simulation environment [40] . Stochastic ion channel gating can be simulated in NEURON using the next-transition algorithm which tracks the exact times at which each channel changes state [34] , [35] . For simulations of voltage-clamped currents using short time-steps and relatively few ion channels , the simulation time with PSICS was approximately three-fold faster than using NEURON ( Figure 4E–F ) . This difference increased to a more than 10 fold reduction in simulation time when PSICS is used for simulations with larger time steps and more ion channels . We next evaluated performance using the spiking single cable model also used for the simulations in Figure 3F . This model contains several types of ion channel distributed across multiple compartments and has a rapidly fluctuating membrane potential . The simulation time per unit biological time was constant for simulations run with NEURON and was independent of the compartment size . By contrast , the times for simulations run with PSICS were faster at all time steps . This difference was >100 fold with relatively large numbers of channels per compartment and long time-steps ( Figure 4G ) . For simulation parameters likely to be appropriate for many detailed neuronal models we estimate that PSICS obtains approximately an order of magnitude or greater reduction in simulation time . Together , these data establish that the new tools we have developed enable accurate and efficient stochastic simulations of neurons with complex morphologies , with performance that is superior to other general-purpose software . Does morphology influence the functional consequences of stochastic ion channel gating ? To address this possibility , we first compared the membrane potential noise resulting from stochastic ion channel gating in a hypothetical dendritic tree that obeys Rall's 2/3 power law , with membrane potential noise resulting from stochastic ion channel gating in the corresponding equivalent cable structure [41] ( Figure 5 ) . In both structures spontaneous opening and closing of fast leak K+ and Na+ channels causes noisy fluctuations in the membrane potential . These fluctuations increase in amplitude by more than ten fold between the proximal and the distal ends of the branched dendrite model ( Figure 5D–E ) ( p<<1e-9 ) , but have relatively small amplitude throughout the equivalent cylinder ( Figure 5A–B ) ( p = 0 . 35 ) . Similar differences in amplitude and location-dependence were present for models that instead contained the slow gating leak channels , but were otherwise identical ( Figure 5B , E ) . Under what conditions do channel kinetics determine the impact of stochastic ion channel gating ? Whereas our earlier simulations indicated an important role for channel kinetics ( Figure 4A–B ) , in our initial simulations of the hypothetical dendritic tree and equivalent cable the kinetics of the leak current did not affect the amplitude of the membrane potential fluctuations ( p = 0 . 63 ) ( Figure 5G ) . However , since the branching tree and the equivalent cable have a membrane time constant on the order of 0 . 1 ms , whereas many central neurons have membrane time constants on the order of 10 ms , we re-evaluated these models after increasing the membrane capacitance to bring the time constant into this range ( Figure 5C , F ) . In this case , the amplitude of the membrane potential fluctuations was also dependent on location in the branched dendrite ( p<<1e-9 ) ( Figure 5F ) , but not in the equivalent cable ( p = 0 . 474 ) ( Figure 5C ) . Moreover , in contrast to the models with the fast membrane time constant , for models with a more physiological membrane time constant the kinetics of the leak current profoundly influenced the amplitude of membrane potential fluctuations ( Figure 5F ) . In models containing the fast leak channels the amplitude of membrane potential fluctuations was reduced ( p<<1e-9 ) , but in models containing the slow leak channels their amplitude was similar ( p = 0 . 54 ) ( cf . Figure 5G , H ) . Thus , the cellular effects of stochastic ion channel gating depend on morphology , while the specific effects of morphology depend on the kinetics of the ion channels found in the membrane . Since the consequences of stochastic ion channel gating are sensitive to their specific cellular context , establishing the impact of stochastic gating in particular central neurons will require simulations account for details of their morphology . Do realistic neuronal morphologies influence the functional impact of stochastic ion channel gating ? While the simulations described above suggest this may be the case , they also suggest that the consequences of stochastic gating depend on the specific details of neuronal morphology and ion channel kinetics . To address this question directly , we therefore reasoned that if neuronal morphology is an important determinant of the impact of stochastic ion channel gating , then simulations using identical rules to introduce identical stochastic ion channels into neurons with distinct dendritic morphologies , should predict differences between neurons ( Figure 6 ) . We simulated resting membrane potential activity in 29 reconstructed neurons downloaded from neuroMorpho . org [42] . The neuronal models spanned 6 distinct morphological classes: neocortical layer V pyramidal neurons ( n = 5 ) , cerebellar Purkinje neurons ( n = 5 ) , dopaminergic substantia nigra neurons ( n = 4 ) , parvalbumin-positive interneurons ( n = 5 ) , hippocampal CA1 pyramidal neurons ( n = 5 ) and hippocampal dentate gyrus granule cells ( n = 5 ) . Fluctuations in the membrane potential were apparent in all neurons simulated using stochastically gating ion channels ( Figure 6A ) . However , the amplitude of these fluctuations varied significantly both between neurons of the same morphological class ( p<0 . 01 for all classes ) , between neurons of different morphological classes ( p<<1e-9 ( Figure 6B ) , and as a function of dendritic location within neurons ( p<<1e-9 ) . For example , pyramidal neurons from the neocortex demonstrate relatively small amplitude membrane potential fluctuations ( Figure 6 ) . This is consistent with a previous modeling study of stochastic ion channel activity in a single layer V pyramidal neuron [43] . By contrast , membrane potential fluctuations could be substantially larger in hippocampal dentate gyrus granule cells ( Figure 6 ) . Thus , the impact of stochastic gating of dendritic ion channels on neuronal electrical properties is determined by neuronal morphology and can vary according to dendritic location . As the impact of stochastic gating in the abstract models described above depended on channel kinetics ( Figure 4A–B and Figure 5 ) , we asked if this is also the case in the models based on reconstructed neurons . We focused on models of cortical layer V pyramidal neurons and on models of granule cells from the dentate gyrus of the hippocampus . When the fast gating leak channels used for the simulations in Figure 6 were replaced with an equivalent deterministic conductance , we found almost no difference in the amplitude of membrane potential fluctuations recorded from somatic or dendritic locations ( DG neurons average 1 . 11 fold difference , p = 0 . 02; Layer V neurons , average 1 . 14 fold difference , p = 1 . 5e-6 ) ( Figure 7 ) . Thus , in the configuration used for simulations in Figure 6 , the membrane potential fluctuations are primarily driven by stochastic gating of voltage-gated ion channels , but not by the leak channels . By contrast , when we replaced the fast gating leak channels with otherwise identical slow gating leak channels , the membrane potential fluctuations were approximately three-fold larger than fluctuations recorded from models containing deterministic or fast-gating stochastic leak channels ( DG neurons average 3 . 13 fold difference , p<<1e-9; Layer V neurons , average 3 . 08 fold difference , p<<1e-9 ) ( Figure 7 ) . Thus , slow gating leak channels can increase the amplitude of spontaneous membrane potential fluctuations . This suggests a novel mechanism for modulation of neuronal activity , whereby modulation of channel gating , without affecting open probability or single channel conductance , could profoundly influence fluctuations in a neuron's somatic or dendritic membrane potential . The previous simulations establish that in principle stochastic gating of intrinsic ion channels might be important for neuronal function , but the impact of stochastic ion channel gating on neuronal responses to physiological patterns of synaptic input is not known . We therefore asked if stochastic gating of post-synaptic ion channels in dendritic neurons influence the transformation of synaptic input into spike output obtained with realistic neuronal morphologies and ion channel properties ? In the models described so far , ion channel distributions were chosen to facilitate comparisons between morphologies . To address more realistic ion channel distributions we adopt a model of a CA1 pyramidal neuron that has previously been shown to account well for somatic and dendritically initiated action potentials [44] ( Figure 8A ) . To further increase the approximation of the model to a real CA1 pyramidal neuron we introduced HCN channels with distribution following previous experimental descriptions [45]–[47] . We then examined responses of the model to ongoing activation of 1502 synaptic inputs distributed throughout the basal and apical dendrites , each activated independently according to a Poisson process with an average frequency of 5 . 5 Hz . We focus here on results of simulations in which the neuron was driven by synaptic input to fire at frequencies of approximately 20 Hz , which is towards the upper end of firing frequency of active CA1 neuron in vivo [48] . Similar results are obtained for synaptic input that drives firing at lower frequencies and when synaptic input is distributed so that spikes are triggered primarily by depolarization of the distal apical dendrites ( not shown ) . We consider here only asynchronous and distributed synaptic input , which is likely to correspond to activity during the theta state in awake animals [49] . As in experimental studies [49] , [50] , forward propagating apical dendritic spikes were only evoked in the model using highly coincident and spatially localized stimuli , but were not observed in response to the patterns of input that we investigate here . Compared to the deterministic model , the stochastic version generated “extra” spikes at times when the equivalent deterministic neuron was silent and “dropped” spikes at times when the equivalent deterministic neuron fired action potentials ( Figure 8B ) . Importantly , the “extra” spikes observed during simulation of the stochastic model occurred at similar time-points from trial to trial ( Figure 8C and Figure S3 ) . Thus , stochastic channels allow probabilistic detection of features in the stimulus waveform that would not produce responses in a deterministic neuron . Likewise , not all spikes observed in the deterministic simulation were “dropped” in the stochastic simulation , but rather “dropped” spikes were more likely at particular time points ( Figure 8C and Figure S3 ) . Comparison of spike times from repeated trials demonstrated that stochastic ion channel gating also introduced considerable jitter into the timing of action potentials . Therefore , whereas deterministic neurons encode information using a fixed response to particular patterns of synaptic input , these results suggest that stochastic gating of intrinsic ion channels enables pyramidal neurons to generate probabilistic responses . Thus , while for both stochastic and deterministic neurons certain combinations of synaptic input evoke spikes with high reliability and other combinations fail to elicit spikes , in stochastic neurons some patterns of synaptic input have an intermediate probability of evoking spikes , which is observed as trial-to-trial variability ( Figure 8C ) . While this intermediate probability might not be decoded in a single trial from a single neuron , if each trial is instead considered as the response of a different neuron within a large population , then the probabilistic responses could quite easily be decoded from the population activity ( Figure 8C ) . To evaluate the mechanism for probabilistic initiation of action potentials , we recorded membrane potential from the soma and from proximal parts of each primary dendrite . While some somatic action potentials were preceded by dendritic depolarizations that resemble fully propagating dendritic spikes ( Figure 8D ) , most were preceded by smaller amplitude dendritic depolarizations ( Figure 8E–G ) . The all-or-nothing nature of these smaller events suggests that they reflect dendritic action potentials that propagate passively to the soma ( Figure 8E–G and Figure S3 ) . This is consistent with experimental recordings from basal dendrites of cortical pyramidal neurons [51] . In the stochastic model “extra” somatic spikes could result from additional actively propagating dendritic spikes ( Figure 8D ) or additional smaller all-or-nothing dendritic depolarizations of sufficient amplitude to elicit somatic action potentials ( Figure 8G ) , while “dropped” somatic spikes resulted from failures to initiate all-or-nothing dendritic depolarizations ( Figure 8E–F ) . These observations point to the importance of local dendritic signaling for the functional consequences of stochastic ion channel gating in pyramidal neurons and suggest that synaptic initiation and active propagation of dendritic spikes may be particularly sensitive to stochastic membrane potential fluctuations . To evaluate the relative roles of stochastic axonal compared with stochastic dendritic ion channels we implemented versions of the model in which one population of ion channels was deterministic and the other stochastic . Both axonal and dendritic stochastic channels caused “dropped” and “extra” dendritic spikes ( Figure 9A–B , D ) . When only axonal channels were stochastic , the number of “dropped” and “extra” somatic spikes ( p<1e-6 in both cases ) and dendritic spikes model ( p<1e-6 in both cases ) were less than in the fully stochastic model ( Figure 9A ) . In contrast , when only the dendritic channels gated stochastically , we found that the number of “dropped” and “extra” somatic spikes was not significantly different compared to the fully stochastic model ( p = 0 . 99 and p = 0 . 85 respectively ) , while the number of “extra” ( p<1e-6 ) , but not the number of “dropped” ( p = 0 . 81 ) dendritic spikes differed from the fully stochastic model ( Figure 9B ) . The number of “dropped” and “extra” spikes was much smaller in the stochastic axon model , compared with the stochastic dendrite model ( p<1e-6 in all cases ) . Stochastic gating of axonal channels also caused very little additional jitter in the timing of action potentials , whereas stochastically gating dendritic ion channels could account for almost all of the spike jitter ( Figure 9E ) . These data suggest that while stochastic gating of axonal channels can modify spike patterns , stochastic dendritic channels account for most of the impact of stochastic gating on synaptically driven spike output . This is consistent with the substantial effects of stochastic gating on initiation of dendritic spikes ( Figure 8D–G and Figure S3 ) . Since the model CA1 pyramidal neuron contains several types of ion channel that differ in their kinetics , voltage-dependence and single channel conductance [44] , we asked if any particular channel type mediates the consequences of stochastic gating ? We compared versions of the model in which only one type of ion channel was simulated stochastically and the others were simulated deterministically . These simulations demonstrated that stochastic gating of any single type of ion channel is insufficient to fully account for all of the “dropped” or “extra” spikes observed in the fully stochastic model ( Figure 9D–E ) . The greatest impact on spike output came from stochastically gating voltage-dependent Na+ channels , followed by A-type and delayed rectifier potassium channels ( Figure 9D–E and Figure S4 ) . Thus , the number of “dropped” somatic and dendritic spikes did not differ between the model in which only Na+ channels gated stochastically compared with the fully stochastic model ( p = 0 . 98 and 0 . 3 ) , whereas there were fewer “extra” somatic and dendritic spikes ( p<1e-4 in both cases ) . Models in which only one of the other ion channel types gated stochastically differed significantly from the fully stochastic model in all measures of “extra” and “dropped” spikes ( p<1e-3 ) . Nevertheless , models containing stochastic gating voltage-dependent K+ channels generated considerably more than 50% of the number of “extra” and “dropped” spikes observed in the fully stochastic model . Interestingly , stochastic gating of Ih channels alone had very little impact on axonal spikes or spike jitter , but nevertheless increased the number of “extra” dendritic spikes . The relative lack of effect of Ih can be explained by the small single channel conductance , while the primary effect on additional dendritic spikes may reflect slow gating kinetics and dendritic localization of these channels ( Figure 9C ) . Together , these data suggest that in a well-validated , realistic model of a CA1 pyramidal neuron experiencing distributed synaptic input sufficient to drive action potentials at physiologically relevant rates , stochastic gating of dendritic ion channels substantially modifies spike output . While no single ion channel is sufficient to fully account for modified spike output , stochastic gating of dendritic voltage-gated Na+ and K+ channels may be particularly important .
Gating of single ion channels is one of the better-understood stochastic processes in biology [24] , [30] , [52] . Nevertheless , the functional consequences of discrete transitions between open and closed states of ion channels found in the membranes of morphologically complex neurons are not well understood and for technical reasons have received relatively little attention . The reductions in computation time obtained with PSICS enable this issue to be addressed systematically for the first time using detailed simulations of large numbers of reconstructed neurons ( Figures 4–9 ) . Modification by stochastic ion channel gating of the pattern and timing of spikes generated in response to synaptic input to a previously well validated model CA1 pyramidal neuron ( Figures 8 and 9 ) , suggests that stochastic ion channel gating substantially influences synaptic integration by dendritic neurons . However , our simulations also suggest reasons for caution in extrapolating between cell types as the consequences of stochastic gating depended on neuronal morphology ( Figures 5 and 6 ) . Thus , the impact of stochastic opening and closing of ion channels varies as a function of sub-cellular location within a neuron and may differ both between neuronal cell types and across neurons of the same cell type . Moreover , our finding that ion channels with identical open probability , but distinct gating kinetics produce different membrane potential fluctuations ( Figure 7 ) , while suggesting a previously unexplored mechanism for control of neuronal activity , also indicates that single-channel recordings of ion channels in dendrites and axons ( e . g . [53]–[55] ) will be important to constrain stochastic models of excitable neurons . Given debates over the accuracy of aspects of reconstructed neuronal morphologies [56] , [57] , our comparison of neuronal cell types should be considered as a proof of principle rather than a definitive description of a particular neurons activity . Since experimental tests that replace a neurons stochastic with equivalent deterministically gating dendritic ion channels are not currently possible , accumulation of accurate morphological and biophysical data will be particularly important for further investigator of the roles of stochastic gating in particular cell types . Irrespective of the details of any particular model , our results suggest that neuronal morphology and ion channel properties interact to determine the functional consequences of ion channel gating . Comparison of model neurons with different morphology , but containing identical ion channels , indicates that dendritic morphology plays a key role in determining the functional consequences of stochastic ion channel gating ( Figures 6–7 ) . Diversity between neurons in vivo in their expression of particular ion channels [58] , could accentuate or attenuate distinctions between neurons predicted on the basis of their morphology . Simulations of the detailed CA1 pyramidal neuron model in which only one ion channel type was implemented stochastically suggest several further insights into the roles of particular types of ion channel . First , stochastic gating of any single ion channel type was insufficient to fully account for probabilistic behavior of the fully stochastic neuron . Second , quantitative differences between the probabilistic behavior of the fully deterministic and stochastic neurons were considerably less than the sum of the differences between the fully deterministic neuron and each version of the model in which only one ion channel type gated stochastically . This suggests considerable redundancy in the functional consequences of gating by any particular type of ion channel . Third , stochastic gating of only a single ion channel type , for example voltage-gated Na+ channels in Figure 9 , can nevertheless substantially modify spike output . The latter two conclusions suggest that the results of our simulations will be robust to different assumptions about single channel properties of particular ion channels and at worst may under-estimate the influence of stochastic ion channel gating ( see Methods ) . Fourth , the impact of stochastic gating differs between ion channels types . For example , compare the model CA1 pyramidal neuron in which only Ih channels gate stochastically , with equivalent models in which other ion channels gate stochastically ( Figure 9 ) . The relatively small impact of Ih is perhaps not surprising given the relatively small underlying single channel conductance , which is estimated to be on the order of 1 pS [13] . Given that Ih is a major contribution to the resting dendritic membrane conductance of pyramidal neurons [13] , [46] , it might at first appear surprising that stochastic gating of other ion channel types can have such large effects . However , we have previously shown that as HCN channels are deactivated by depolarization , at potentials closer to spike threshold their impact is minimal compared to stochastic gating of other types of voltage-dependent ion channel [11] , [59] . The substantial influence on synaptically driven spike output of stochastic gating of voltage-gated Na+ and K+ channels is consistent with this explanation ( Figure 9 ) . Since it is rarely possible to reduce electrical activity within morphologically complex excitable neurons to tractable analytical models , mechanistic simulation of axons and dendrites relies on well-constrained compartmental models . Compartmental simulations necessarily involve trade-offs between accuracy and simulation time . This is a particularly important problem for simulations that aim to account for the stochastic transitions between the states of each individual ion channel in a realistic model neuron . To enable efficient and accurate simulation , we adopted an algorithm that generates a correct distribution of ion channel states at each simulation time point , while sacrificing explicit representation of ion channel states during the interval between time points . Relatively short time steps are required for accurate simulation of voltage-clamped conductances with rapid kinetics ( Figure 4 ) . In contrast , during simulation of membrane voltage recorded in the current-clamp configuration , high frequency current fluctuations are filtered by the membrane capacitance and therefore have little impact on neuronal activity . Therefore , high frequency component of the conductance fluctuations do not have to be explicitly simulated ( Figure 4A–B ) and so simulations using the modified tau leap algorithm can take advantage of longer time steps . Our new approach has several advantages over previous methods for simulation of stochastic ion channel gating . While approaches that add noise terms to the equations used to calculate the membrane currents are computationally efficient [14] , [60] , the noise term is at best only indirectly related to the biophysical properties of the simulated channels . Thus , these methods may be of use for efficiently simulating some of the phenomenological consequences of stochastic channel gating , but are of less utility for addressing the relationship between properties of single ion channels and macroscopic neuronal activity [22] . One approach to account for single channel properties is to explicitly track the probability that each channel makes a transition between states during each time step [11] , [23] . However , this method is accurate only if sufficiently small time steps are used [22] and is computationally very expensive . An alternative method is to explicitly track the exact time of transitions of channels between states , while counting only the number of channels in each particular state [34] , [35] . This demands less computation time than explicitly tracking each channel , but nevertheless requires generation of random numbers between time steps and therefore becomes inefficient for longer time steps or large numbers of channels . In contrast , PSICS uses a simulation algorithm that accurately simulates the distribution of ion channel states at each time step without having to track the exact time of each transition , while also counting only the number of ion channels in each state without having to track the states of individual channels . We show that this can lead to an order of magnitude or greater improvement in simulation time without loss of accuracy ( Figure 4 ) . For all approaches , including those introduced here , parallel computing produces a further linear reduction in computation time with additional processors simply by enabling the multiple simulations required for statistical evaluation of models to be carried out simultaneously . By implementing a previously well-validated model of a CA1 pyramidal neuron using stochastically gating ion channels , our simulation results provide evidence that synaptic integration by dendritic neurons is probabilistic . While the instantaneous output of a single neuron functioning in this way is relatively unreliable , instantaneous representations distributed across a population of stochastic neurons could be read out by summation of their outputs . The impact of such probabilistic integration on information processing and computation by populations of pyramidal neurons remains to be determined . For CA1 pyramidal neurons in the hippocampus , one possibility is that this probabilistic integration is important for encoding of location by the timing of action potentials relative to ongoing network rhythms [61] . Indeed , our results are consistent with relatively unreliable encoding of location by the timing of individual action potentials , but suggest that coding mechanism might be considerably more reliable when the activity of large ensembles of neurons is considered . Challenges for future experimental and theoretical studies include determining the conditions , additional cell types and sub-cellular locations in which stochastic gating of ion channels affects spike output , and to establish the consequences for computations carried out by neural circuits . At some sub-cellular locations noise introduced by stochastic gating of single ion channels might impose physical constraints on the computational properties of neurons [12] and may limit the efficiency of neural coding [62] . Alternatively , neuronal noise sources may promote detection of signals [63]–[66] , enable multiplication of synaptic responses [67] , or control the pattern of action potential firing [11] . The tools we introduce here will enable these and other possibilities to be addressed systematically . In addition to exploring physiological mechanisms , systematic simulations may be important for understanding the functional consequences of changes in morphology or ion channel localization that accompany nervous system disorders . For example , changes in dendritic morphology are reported in several forms of mental retardation [68] , but the functional implications of interactions between stochastic gating of dendritic ion channels and disease or behaviorally related changes in dendrite morphology are yet to be addressed .
Channel positions are allocated according to user-specified probability densities over the structure such that each channel has a position in three dimensions . The input morphology is sub-sampled at 1 µm for computing local number densities . Axial positions for channels are generated either by sampling a Poisson distribution for the distance to the next channel , or by taking uniform increments to give the desired average density . The angles at which channels occur around a section are allocated randomly from a uniform distribution . At present , these angles only affect the visualization since the structure is later discretized into elements with end faces perpendicular to the axis . The seeds used for the random number generator are stored with the model so that exact allocations can be reproduced . So that allocations are portable across platforms , the generator used is a Mersenne Twister [70] , which is included as part of PSICS rather than relying on a system library . For computing the propagation of membrane potential changes , the structure is compartmentalized into elements such that all elements have approximately the same value of: where and are the positions of the end points of the compartment along the structure , r is the local radius and was used throughout this study . This is a purely geometrical property that provides a discretization that balances conductance between compartments against charging rates for membrane capacitance and is independent of the membrane time constant . PSICS allows post-hoc adjustment of the discretization in view of the conductances arising from allocated channels , but the facility was not used for the present study . In general , the resulting elements are neither straight nor of uniform radius so conductances between compartments are also computed as integrals along the structure . Ion channels are represented by kinetic schemes . Each scheme has one or more complexes , and each complex is an inter-connected set of states with expressions for the transition rates between them . Models using the Hodgkin-Huxley gating particle description are mapped to the corresponding scheme where each gate corresponds to a two-state complex [24] . For deterministic calculations , multi-complex schemes are used directly . For stochastic calculations , multi-complex schemes are mapped to the equivalent single-complex scheme by “multiplying-out” [24] , [71] . For a scheme with n states , the probabilities of the channel being in each of the states can be expressed by a probability density vector of length , where is the probability of being in state . A channel in state can only transition into one of the states directly connected to that state . If the classical transition rate between state and state is then when channels are in state , the number of transitions per unit time between from state to state would be for large ( if there is no direct connection from states to state ) . In this case , the single channel probabilities obey a similar relation giving the rate of change of as the difference between the total fluxes into and out of that state: By gathering up the terms , this can be written in matrix form as a master equation [72]:where , and In general , will depend on the membrane potential , and could incorporate other dependencies such as calcium or other second messengers . Note that , as with the transition rates themselves , only has non-zero elements along the diagonal and where direct transitions are possible in the channel scheme . Using directly to update state occupancy probabilities and sample the resulting distribution to generate a stochastic simulation of the channel corresponds to the explicit method ( e . g . [11] ) . If it is not necessary to follow every transition , then the simulation can be made more efficient by integrating the master equation over a time interval yielding: As this integral is exact for constant , the state occupancy probabilities can be computed for any future time . In practice is often voltage dependent , so should be kept small enough that there are no significant changes in over a single time step . In calculations , significant efficiency improvements over explicit methods can be achieved by using the fixed step transition matrix , , which allows channel to move through many briefly occupied states in a single calculation step without having to follow the dynamics of each transition . Unlike , is typically non-zero everywhere , since there is a non-zero probability of the channel making the multiple transitions necessary to get from one state to any other state within the time step . For a deterministic calculation , the state occupancy probability vector is replaced by an occupancy density vector , , while the rest of the derivation remains the same , so the update step for is: For a stochastic calculation , the element is the probability that a channel currently in state will be in state at the end of the time-step . The update step for a single channel involves generating a uniform random number , r , between 0 and 1 and then comparing it to the elements of the 'th column of , where is the current state of the channel . The selected state , , is such that: That is , the elements act as the widths of bins and the random number is used to pick one of the destination bins according to their relative sizes . For small populations , the update step is applied to each channel individually . For larger populations , significant computation time can be saved by updating only the channels that have a non-negligible probability of changing state during the step and rescaling the bins accordingly . Thus , if the total number of channels in a given state is and the probability of any channel leaving the state in the next time step is then the number of channels leaving the state in one step has a binomial distribution:and the probability that or more channels will leave is: For a statistically reliable result , we require a bound on the number of missed events , here set at 2 in 108 . This requires such that: In general , however , finding the smallest that meets this condition is a computationally costly process and , since it needs to be done for each state of each channel population at each time step , these costs could wipe out any benefits of not advancing the channels individually . To avoid this problem , the following fitting formulae are used to produce a safe estimate of in terms of the probability of leaving the current state , and the number of channels , , in that state: The first applicable formula is used . If none of the conditions is met , all the channels are advanced individually . These formulae were arrived at by a combination of experimentation and numerical optimization to find expressions that approximate the form of the exactly computed values of as a function of and over a logarithmically sampled grid . They ensure that the chance of more channels making a transition out of the current state than are considered at one step is less than 2 in 108 . For ion channels the modified tau leap method removes the major cost of the next transition method [34] , [35] , which lies in re-computing next transition times whenever the membrane potential changes significantly , even when no transition has occurred . In the tau-leap method , a different membrane potential simply requires the use of different transition probabilities for the next step . Since all the required tables can be pre-computed , then if the steps are small enough that no significant potential change occurs within a step ( a condition that is easily met ) , the membrane potential change only adds very slightly to the computational cost . The other efficiency saving of a tau-leap style algorithm over channel-by channel calculation comes from treating a group of molecules ( channels ) as a single stochastic unit . On a neuron the unit size is determined by the electrical compartmentalization of the membrane . Therefore , depending on the geometry and channel distribution , clusters of up to 5000 channels may be possible . The channel update step yields conductances ( assumed Ohmic ) and reversal potentials for membrane currents on each compartment , which are used in the Crank-Nicolson or implicit Euler methods to compute new membrane potentials . The process is then repeated for the next step . The allocation of channels , discretization of the structure and tabulation of transition matrices is performed in a pre-processing stage written in Java . The core calculations are written in Fortran . For simulations illustrated in Figures 6–8 neuronal morphologies were downloaded from the Neuromorpho database ( www . neuromorpho . org ) . Neurons were identified in the database as follows . Layer V pyramidal cells: p18 and p22 from Dendritica; g0692P1 , g0699P1 and gR002P1 from Svoboda lab . Dentate gyrus granule cells: n271 , n272 and n518 from Turner lab; 428883 , B106885 from Claibourne lab . Purkinje cells: alxP , e4cb3a1 and e1cb4a1 from Martone lab; p19 and p20 from Dendritica . CA1 pyramidal cells: n409 from Turner lab; NM1 from Ascoli lab; ri04 and ri06 from Spruston lab; pc4c from Gulyas lab . Substantia nigra dopaminergic neurons: Nigra2a955 , Nigra11h941-1 , Nigra24a953 , Nigra12h945 from Dendritica . Parvalbumin interneurons: pv08e , pv22b , pv22e , pv22j and pv22m from Gulyas lab . In these simulations the densities of voltage-gated channels were based on previously published studies [6] . The leak conductance was modeled as voltage-independent Na+ and K+ channels with open probabilities of 0 . 7 . The following channels were included: fast Na+ channels ( 1/µm2 ) ; non-inactivating K+ channels ( 0 . 05/µm2 ) ; high-voltage Ca2+ channels ( 0 . 15/µm2 ) ; Na+ and K+ leak channels ( 0 . 016/µm2 ) . The resting membrane potential was set by modifying the ratio of Na+ to K+ leak channels . In all simulations reported here this was −60 mV . We chose single-channel conductances of 20 pS for all ion channels , as this is similar to values reported for single channel recordings made from neuronal dendrites [54] , [55] . This value is intermediate for cloned mammalian ion channels , which can have single channel conductances from <1 pS up to >150 pS [30] . Membrane capacitance was set to 0 . 75 µF/cm2 and axial resistivity to 150 Ω cm [6] . For models of neurons that are known to have dendritic spines ( all models except the parvalbumin-expressing interneurons ) , the dendritic membrane capacitance and the number of dendritic ion channels were doubled . For each model neuron , membrane potential was recorded at the soma and at all dendritic locations 100 µm , 200 µm , 300 µm , etc . , from the soma . All reported results were obtained from at least 3 s of simulated biological time . The simulation time step was 10 µs . The simulations of a detailed model of CA1 pyramidal neuron ( Figures 8 and 9 ) used previously published ion channels , morphology and channel distributions [44] . In this model voltage-dependent ion channels are distributed in the soma , axon and dendrites according to previous experimental measurements . The only modification to the model was the addition of Ih conductance and channel distribution taken from a different publication from the same group [45] and consistent with data from other groups [46] . The densities of Na+ and K+ leak channels were automatically adjusted to achieve a resting potential of approximately −70 mV throughout the cell , while maintaining a total leak conductance consistent with the original model . The single channel conductance of the delayed rectifier K+ channels and voltage-dependent Na+ channels were set to 20 pS , which is similar to estimates from single channel recordings [54] , [55] . For simplicity , in the reported experiments the single channel conductance of A/D type K+ channels and leak channels were also set to 20 pS , which is similar to experimental measurements for D-type channels [55] , somewhat larger than estimates for dendritic A-type channels [55] and towards the low end of the range of single-channel conductance reported for leak channels [30] . Thus , our simulations of models with only stochastically gating voltage-dependent Na+ and delayed rectifier K+ channels can be considered as fully constrained predictions given currently available data , while our simulations of the fully stochastic model likely estimate a lower limit for the consequences of stochastic ion channel gating . This is because our results from simulations when A/D type K+ channels are deterministic , but voltage-dependent Na+ or delayed rectifier K+ channels are stochastic , nevertheless demonstrate highly probabilistic spike firing , indicating that a smaller single channel conductance for A/D type K+ channels would have little impact on the results , while a possible larger single channel conductance for the leak channels would be expected to increase the impact of stochastic gating . Our simulations of A/D type stochastic gating alone should be considered as setting an upper limit for stochastic effects based on known properties of these channels , whereas the simulations of leak channels alone are less well constrained and serve as an illustrative example . Unlike other ion channels , the single channel conductance of Ih channels is set at 1 pS , which is consistent with noise-analysis of dendritic Ih recorded from cortical neurons [13] and the absence of step-like single channel waveforms from measurements of Ih obtained with cell-attached recordings from CA1 pyramidal neurons [46] . Synapses were modeled as bi-exponential conductance changes of rise time 0 . 2 ms , decay time 2 ms and peak conductance 0 . 18 nS . Synapses were distributed randomly across all dendrites >30 µm from the soma at an average density of 0 . 1/µm2 ( 1502 in total ) . Each synapse was activated independently according to a Poisson process with a mean frequency of 5 . 5 Hz . For analysis dendritic spike times were calculated as upward voltage crossings above a −60 mV threshold . Visual inspection of traces confirmed that this threshold successfully isolated all-or-nothing dendritic events . Additional analysis of simulation data was carried out using IGORpro ( Wavemetrics ) . Statistical analysis used R ( http://www . r-project . org/ ) . Comparisons of group data use ANOVAs . For analysis of data in Figures 8 and 9 this was followed by Tukey Honest Significant Difference test for comparisons between individual groups . Significance values referred to in the text refer to probabilities indicated by the latter test after adjustment for multiple comparisons . Simulations with NEURON used version 7 . 0 ( www . neuron . yale . edu ) . For comparisons between NEURON and PSICS , simulations were run on the same hardware with minimal competing activity . Although other factors such as cache size and output options may contribute to performance differences in some cases , the simulation algorithm was the main factor determining simulation time . Parallel simulations were run on multi-processor PCs , or on a cluster of servers ( maximum 1456 processors ) at the Edinburgh Compute and Data Facility ( ECDF ) . | The activity of neurons in the brain is mediated through changes in the probability of random transitions between open and closed states of ion channels . Since differences in morphology define distinct types of neuron and may underlie neurological disorders , it is important to understand how morphology influences the functional consequences of these random transitions . However , the complexities of neuronal morphology , together with the large number of ion channels expressed by a single neuron , have made this issue difficult to explore systematically . We introduce and validate new computational tools that enable efficient generation and simulation of models containing ion channels distributed across complex neuronal morphologies . Using these tools we demonstrate that the impact of random ion channel opening depends on neuronal morphology and ion channel kinetics . We show that in a realistic model of a neuron important for navigation and memory random gating of ion channels substantially modifies responses to synaptic input . Our results suggest a new and general perspective , whereby output from a neuron is a probabilistic rather than a fixed function of synaptic input to its dendrites . These results and new tools will contribute to the understanding of how intrinsic properties of neurons influence computations carried out within the brain . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"neuroscience/neuronal",
"signaling",
"mechanisms",
"biophysics/theory",
"and",
"simulation",
"neuroscience/theoretical",
"neuroscience",
"computational",
"biology/computational",
"neuroscience"
] | 2010 | Stochastic Ion Channel Gating in Dendritic Neurons: Morphology Dependence and Probabilistic Synaptic Activation of Dendritic Spikes |
Defining the mechanisms of Mycobacterium tuberculosis ( Mtb ) persistence in the host macrophage and identifying mycobacterial factors responsible for it are keys to better understand tuberculosis pathogenesis . The emerging picture from ongoing studies of macrophage deactivation by Mtb suggests that ingested bacilli secrete various virulence determinants that alter phagosome biogenesis , leading to arrest of Mtb vacuole interaction with late endosomes and lysosomes . While most studies focused on Mtb interference with various regulators of the endosomal compartment , little attention was paid to mechanisms by which Mtb neutralizes early macrophage responses such as the NADPH oxidase ( NOX2 ) dependent oxidative burst . Here we applied an antisense strategy to knock down Mtb nucleoside diphosphate kinase ( Ndk ) and obtained a stable mutant ( Mtb Ndk-AS ) that displayed attenuated intracellular survival along with reduced persistence in the lungs of infected mice . At the molecular level , pull-down experiments showed that Ndk binds to and inactivates the small GTPase Rac1 in the macrophage . This resulted in the exclusion of the Rac1 binding partner p67phox from phagosomes containing Mtb or Ndk-coated latex beads . Exclusion of p67phox was associated with a defect of both NOX2 assembly and production of reactive oxygen species ( ROS ) in response to wild type Mtb . In contrast , Mtb Ndk-AS , which lost the capacity to disrupt Rac1-p67phox interaction , induced a strong ROS production . Given the established link between NOX2 activation and apoptosis , the proportion of Annexin V positive cells and levels of intracellular active caspase 3 were significantly higher in cells infected with Mtb Ndk-AS compared to wild type Mtb . Thus , knock down of Ndk converted Mtb into a pro-apoptotic mutant strain that has a phenotype of increased susceptibility to intracellular killing and reduced virulence in vivo . Taken together , our in vitro and in vivo data revealed that Ndk contributes significantly to Mtb virulence via attenuation of NADPH oxidase-mediated host innate immunity .
The ability of Mycobacterium tuberculosis ( Mtb ) to adapt and thrive intracellularly relies on a variety of strategies to alter mechanisms of the host innate immunity . In particular , interference with phagosome biogenesis was highlighted as a significant aspect of Mtb persistence and replication within the macrophage [1] , [2] . How Mtb circumvents phagosomal acidity , bactericidal enzymes , and reactive oxygen species ( ROS ) remains a central question for many cellular microbiologists . ROS are produced by the phagocyte NADPH oxidase ( NOX2 ) complex and were classified 30 years ago as powerful microbicidal agents against many intracellular pathogens [3] . In vivo evidence for the contribution of NOX2 to the innate immunity arsenal was deduced from field observations of high susceptibility of chronic granulomatous disease patients ( CGD ) to opportunistic pathogens [4] , [5] . Such observations were experimentally confirmed in mouse models of CGD [6] , [7] . Recent years have seen a growing body of evidence to suggest a crucial role for ROS in the control of mycobacterial infections [7] . In particular , one group has recently identified Mtb nuoG as a potential virulence factor operating at the level of NOX2 by mechanisms yet to be defined [8] . The NOX2 complex consists of two constitutively associated transmembrane proteins , gp91phox and gp22phox and four cytosolic subunits: p40phox , p47phox , p67phox , and Rac1 , a small GTPase [9] . Fully functional NOX2 requires membrane translocation of p40phox , p47phox , active Rac1 ( GTP-bound form ) and p67phox , and their assembly around gp91phox and gp22phox subunits [10] . NOX2 assembly leads to gp91phox activation to generate superoxide through a redox chain by transferring electrons from cytosolic NADPH to phagosomal oxygen [9] . The production of superoxide is in turn converted into several other microbicidal molecules , such as hydrogen peroxide and hydroxyl radicals , along with peroxynitrite when combined with nitric oxide radicals [9] . While the role of NOX2 in innate immunity is well established , several reports suggested that it might act beyond the control of intracellular infections to trigger macrophage apoptosis [11] , [12] , a central event that paves the road to adaptive immunity [13]–[15] . Previous results from our laboratory identified Mtb nucleoside diphosphate kinase ( Ndk ) as a GTPase Activating Protein ( GAP ) acting on Rab5 and Rab7 GTPases , leading ultimately to reduced phagolysosome fusion [16] , [17] . In the present study , we examined whether Ndk GAP activity extends to other GTPases , with a particular focus on Rho GTPases . We found that Mtb Ndk interacts specifically with Rac1 and inactivates it leading to inhibition of NOX2 assembly and activation in the macrophage . We also established a link between Ndk-dependent NOX2 attenuation and inhibition of apoptosis response to Mtb . Consistent with these findings , Ndk knock down significantly reduced Mtb survival in vitro and in vivo .
We recently showed that mycobacterial Ndk plays an essential role in intracellular survival of the attenuated M . bovis BCG strain by a mechanism dependent on phagosome maturation arrest [17] . To examine whether Ndk also contributes to survival of virulent Mtb , we first attempted to generate an Ndk mutant in the Mtb strain H37Rv using various methods , including a gene disruption approach utilizing transducing mycobacteriophages [18] . Unfortunately , ndkA gene disruption affected severely the growth of bacteria . We therefore opted for protein knock down with mRNA antisense , the only approach developed so far to study essential genes in Mtb [19] . To do so , we transformed Mtb with the integrative vector pJAK1 . A , previously designed by us [20] , to express a stable full length antisense ( or sense , control ) mRNA sequence to ndkA . Thus , we generated a strain ( Mtb Ndk-AS ) in which Ndk protein expression was undetectable by western blot , even after many passages in the absence of the selection marker kanamycin , indicating a stable knock-down with the pJAK1 . A vector ( Fig . 1A ) . Fortunately , Mtb Ndk-AS displayed a similar growth profile to that of wild type Mtb and the control sense strain ( Mtb Ndk-S ) in standard culture media ( Fig . 1B ) , as well as in the presence of oxidative stress ( H2O2 , Fig . S1 ) . Thus potential fitness disadvantage that could be associated with genetic manipulation of Mtb are unlikely . Knock down of Ndk significantly affected Mtb survival in RAW 264 . 7 macrophages to the extent that at 72 h post infection , numbers of Mtb Ndk-AS dropped by 2 log colony-forming units ( CFUs ) , relative to the wild type or Ndk-S strains ( Fig . 1C ) . These findings suggested that the Ndk protein might contribute to Mtb virulence in vivo . Virulence during the early acute phase of infection is essentially controlled by innate immune responses and can be rapidly assessed in the SCID mouse model where innate immune responses are intact [21] , [22] . In this regard , experiments of SCID mice infection by aerosol showed significant reduction ( ∼70% ) of bacterial load in the lungs of Ndk-AS infected animals compared to those infected with wild type and Ndk-S strains ( P = 0 . 002 , unpaired t-test , Fig . 1D ) . Indicators of morbidity were apparent in the mice within 6 weeks with no significant difference between the three infection test groups ( Fig . S2 ) . However , when infected subcutaneously , time to death was extended to 12–15 weeks . Under these conditions , Kaplan Meier survival analysis clearly demonstrated that animals inoculated with Mtb Ndk-AS survived significantly longer ( ∼20 days , P<0 . 0001 ) than the control strain expressing Ndk sense mRNA , which caused the death of mice at similar rates seen in mice infected with the wild type strain ( Fig . 1E ) . Taken together , these data demonstrated clearly that Ndk contributes to Mtb survival in the host through mechanisms that we have attempted to elucidate . Our recent findings that Ndk expresses GAP activity towards Rab5 and Rab7 [17] suggested that this activity might extend to other host GTPases . Therefore , we examined whether Ndk targets macrophage Rho GTPases , known to play essential roles in early events of innate immunity against intracellular pathogens [23] , [24] . To do so , macrophages were allowed to ingest Ndk-coated latex beads and then cell lysates were subjected to immunoprecipitation with Ndk antibody . Proteins associated with Ndk were analyzed by western blot with Rac1 , Rho , or Cdc42 antibodies . The results obtained showed that only Rac1 was interacting with Ndk within the macrophage ( Fig . 2A ) . Rac1 binding to Ndk was further confirmed with reverse pull down experiments using Rac1 antibody and western blotting with Ndk antibody , which showed clearly a physical association between Mtb Ndk and Rac1 ( Fig . 2B ) . Results obtained with coated latex beads clearly demonstrated the specificity of Ndk-Rac1 interactions within the macrophage . Accordingly , we then examined Ndk-Rac1 interaction in cells infected with the bacterium instead of beads and showed that Rac1 antibodies are able to pull-down Ndk-Rac1 complexes from cells infected with wild type and Ndk-S but not Ndk-AS Mtb ( Fig . 2C ) The results shown above ( Fig . 2 ) suggested that Mtb Ndk must cross the phagosomal membrane towards the cytosol to bind to and inactivate Rac1 . We first confirmed the hypothesis of cytosolic translocation of Ndk using i ) confocal microscopy analysis , which showed diffused staining of Ndk distant from phagosomes containing wild type and Mtb Ndk-S but not from those containing Mtb Ndk-AS ( Fig . S3A ) and ii ) immunogold staining and EM analysis , which clearly demonstrated that Mtb Ndk effectively crosses the phagosomal membrane toward the cytosol ( Fig . S3B ) . We next examined the level of Rac1 activation in infected macrophages with pull down experiments using binding domain derived from Rac1 interacting protein ( PAK-1 PBD ) , which interacts with Cdc42 as well . We also examined levels of Rho activation using Rho interacting protein ( Rhotekin RBD ) . These binding domains interact only with GTP-bound forms of Rho GTPases [25] . Mtb infected RAW cells were exposed to LPS in order to activate the Rho GTPases , then cell lysates were examined for the amount of active Rac1 , Rho , or Cdc42 . Western blot analyses with Rac1 , Cdc42 and Rho antibodies demonstrated that Mtb significantly inhibits the level of LPS-induced Rac1 activation ( Fig . 3A , top panel ) . In contrast , Mtb had no apparent effect on Cdc42 and Rho activation . This GAP activity was also observed in macrophages ingesting Ndk-coated beads , as opposed to Mtb bacilli , demonstrating a specific Ndk GAP activity on Rac1 ( Fig . 3A , lower panel ) . To further examine Mtb effects on Rac1 and the phagosomal events it regulates , we performed a time-course Rac1 activation assay with macrophages infected by Mtb Ndk-AS and wild type Mtb . The results obtained showed a dramatic reduction of active Rac1 levels 15 min post infection and undetectable levels 1 h later in macrophages infected with wild type Mtb ( Fig . 3B , top panel ) . In contrast , levels of active Rac1 remain unchanged in macrophages infected with Mtb Ndk-AS ( Fig . 3B , bottom panel ) . Taken together , these data clearly demonstrated that Mtb Ndk expresses GAP activity on both basal and induced Rac1-GTP levels in the macrophage . Active Rac1 ( GTP bound form ) has been shown to translocate to early phagosomes in order to facilitate recruitment of its binding partner , the NOX2 subunit p67phox [26] , [27] . Given that Ndk expresses GAP activity towards Rac1 ( GTP into GDP switch ) , we examined whether Mtb interferes with phagosomal recruitment of p67phox . We first applied intracellular staining and confocal microscopy to estimate the proportion of Rac1 and p67phox positive phagosomes in Mtb-infected RAW cells . Results obtained ( Fig . 4A and 4B ) showed a substantial reduction of Rac1 and p67phox positive phagosomes ( 13% and 29% respectively ) in cells infected with live Mtb relative to those infected with killed Mtb ( 86% Rac1 and 88% p67phox positive phagosomes , respectively ) . In contrast , recruitment of p47phox to live Mtb phagosomes was comparable to that of phagosomes containing killed Mtb . To demonstrate that the NOX2 assembly defect is related to Ndk GAP activity , we applied similar confocal analyses to cells infected with Mtb Ndk-AS . The images ( Fig . 4C and 4D ) clearly showed a restoration of Rac1 and p67phox recruitment to Mtb Ndk-AS containing phagosomes to a level similar to those observed in cells infected with killed Mtb ( ∼% and 72% positive phagosomes , respectively ) . As expected , much lower numbers of Rac1 and p67phox positive phagosomes ( 13% and 30% respectively ) were observed in cells infected with control strain Mtb Ndk-S . As an alternative approach , a previously developed quantitative FACS analysis method [28] was used to assess the level of NOX2 components on individual phagosomes . To adapt this method to mycobacterial phagosomes , macrophage plasma membrane was stained with CellMask Deep Red ( detectable by FL4 channel ) , and then cells were infected with Mtb strains expressing fluorescent DsRed protein ( FL2 ) . Following cell disruption , mycobacteria included in cell membrane-derived vacuoles ( double FL2/FL4 positive events ) were readily identified by flow cytometry analysis ( Fig . S4 ) . Phagosome preparations were then stained with Rac1 or p67phox antibodies and FITC-conjugated secondary antibodies ( FL1 ) . Samples were subjected to flow cytometry analysis and mean fluorescence intensities ( MFI ) were deduced from fluorescence histograms . Results obtained ( Fig . 4E ) showed higher recruitment of Rac1 and p67phox to phagosomes containing Mtb Ndk-AS ( MFI: 49 . 6 and 42 . 6 respectively ) relative to phagosomes containing Mtb Ndk-S ( MFI: 25 . 3 and 21 . 8 respectively ) or Mtb wild type ( MFI: 23 . 9 and 15 . 7 respectively ) . To establish a direct link between Ndk GAP activity and defective NOX2 assembly , additional flow cytometry analyses were applied to phagosomes containing coated beads ( Fig . S5 ) and showed a marked decrease of Rac1 and p67phox recruitment to the Ndk bead phagosomes ( MFI: 8 . 1 and 3 . 8 respectively ) relative to control phagosomes containing BSA-beads ( MFI: 14 . 6 and 7 . 3 respectively ) . Taken together , these findings showed for the first time that Mtb uses Ndk GAP activity to disrupt phagosomal assembly of NOX2 via interference with Rac1-dependent recruitment of p67phox . Previous findings that Rac1 and p67phox subunits are essential for NOX2 assembly and activation of gp91phox to generate superoxide [10] suggested that disruption of Rac1/p67phox translocation to the phagosome by Ndk would affect NOX2-dependent ROS production . To verify this hypothesis , we applied a luminol-dependent chemiluminescence assay to assess ROS production in Mtb infected cells . Luminol is a membrane diffusible reagent commonly used for quantitative detection of superoxide anion radicals and hydrogen peroxide molecules . Bone marrow derived macrophages ( BMDM ) from C57BL/6 mice were infected with Mtb strains and assayed for kinetics of chemiluminescence production over a period of 60 min . Relative luminescence profiles obtained ( Fig . 5A ) revealed that cells infected with Mtb Ndk-AS induced significantly higher amounts of ROS production ( peak value = 256 RLU ) compared to those infected with wild type Mtb or Mtb Ndk-S ( peak value ∼120 RLU ) . Thereafter , we confirmed the apparent inhibitory effect of Ndk with experiments using coated beads ( Fig . 5B ) , which showed minor ROS response to Ndk beads ( peak value = 32 RLU ) relative to ROS production induced by BSA beads ( peak value = 76 RLU ) . Additionally we applied confocal microscopy to visualize intracellular accumulation of ROS using CM-DCFDA , a cell-permeable probe that is non-fluorescent until oxidized within the cell . Thus , in RAW cells infected with Mtb Ndk-AS , the confocal images showed a strong colocalization of oxidized CM-DCFDA ( green fluorescence ) with bacterial phagosomes ( red fluorescence ) indicating accumulation of large amounts of ROS around Mtb Ndk-AS ( Fig . 5C and 5D ) . Conversely , green signal was totally absent in cells infected with either wild type Mtb or Mtb Ndk-S . This effect of Ndk on ROS production was also reproduced when BMDM were used instead ( Fig . S6 ) . Previous studies reported that mitogen-activated protein kinases ( MAPKs ) play an important role in the signaling pathway of NOX2 activation [29] , [30] . To verify whether Ndk also interferes with MAPK activation , macrophages were allowed to ingest Ndk-beads or BSA-beads ( control ) , and then stimulated with PMA or LPS to activate ERK1/2 , and p38MAPK respectively . Cell lysates were then examined for the level of phospho-ERK1/2 and phospho-p38MAPK , which reflects kinase activation . The western blot results obtained ( Fig . S7 ) did not reveal any changes in the levels of kinase phosphorylation in cells infected with Ndk-beads relative to those infected with BSA-beads . Therefore , Ndk effect on NOX2 is clearly independent of MAPK inhibition . Collectively , these experiments demonstrated that the macrophage oxidative response to Mtb is marginal and that knock down of Ndk converts the bacterium into a potent inducer of the ROS response . Mtb is known to inhibit macrophage apoptosis [14] , [15] by mechanisms yet to be clarified . Based upon previous findings that NOX2 activity might extend beyond intracellular killing to induce apoptosis [8] , [14] , we examined whether Mtb uses Ndk to disrupt the NOX2-apoptosis link . First , we applied Annexin V cell surface staining , a popular approach for detection of phosphatidylserine ( PS ) translocation to the extracellular membrane leaflet , which reflects early stages of apoptosis events [31] . Adherent RAW cells on coverslips were infected with Mtb strains for 48 h then stained with Alexa Fluor 488 conjugated Annexin V and examined by confocal microscopy . The images showed very low numbers of Annexin V positive cells in samples infected with wild type Mtb and Mtb Ndk-S ( 5% and 6% positive , respectively ) . In contrast , a higher number of Annexin V positive cells ( 44% ) was observed in samples infected with Mtb Ndk-AS ( Fig . 6A , top panel ) . To establish a direct link between ROS and apoptosis in infected cells , Annexin V staining was repeated on macrophages exposed to Mtb Ndk-AS in the presence of a specific gp91phox peptide inhibitor ( gp91 INH ) or its control scrambled version ( gp91 SCR ) [32] . The results obtained showed clearly that gp91 INH , but not gp91 SCR , reversed completely Mtb Ndk-AS-induced PS translocation to the cell surface ( 5% Annexin V positive , Fig . 6A , bottom panel ) . The effect of the gp91phox inhibitor was confirmed with experiments showing that gp91 INH completely inhibited ROS production in cells infected with Mtb Ndk-AS , which was normally elicited in the presence of gp91 SCR ( Fig . S8 ) . In a complementary series of experiments , we analyzed caspase 3 activation , which occurs during the final stages of apoptosis [33] . Thus , infected macrophages were subjected to intracellular staining with antibody to cleaved ( i . e . active ) caspase 3 and Alexa Fluor 647 conjugated secondary antibody , then analyzed by FACS . Results obtained ( Fig . 6B ) showed higher numbers of apoptotic macrophages in sample tests infected by Mtb Ndk-AS ( 11 . 8% positive events ) relative to those infected with wild type Mtb or Mtb Ndk-S ( ∼4 . 6% ) . Not surprisingly , the wild type and Ndk-S strains inhibited the spontaneous apoptosis observed in control non-infected cells ( 7 . 3% positive cells ) . As expected , Mtb Ndk-AS-induced caspase 3 cleavage was abolished in the presence of the gp91phox inhibitor . It is well known that apoptosis is also induced by nitric oxide ( NO ) in mouse macrophages [34] , [35] . Therefore , the effect of Ndk on macrophage apoptosis might be the result of simultaneous inhibition of ROS and NO production . To verify this possibility we examined IFN-γ-induced NO production in cells infected with Ndk-beads and BSA-beads ( control ) and the results deduced from the Griess assay ( Fig . S9 ) demonstrated that Ndk has no effect on NO production . Taken together , our data demonstrated that Mtb blocks macrophage apoptosis by a mechanism dependent , at least in part , on Ndk-mediated attenuation of NOX2 activity . Results presented above ( Fig . 6 ) together with initial experiments showing attenuated Mtb Ndk-AS survival in RAW macrophages ( Fig . 1 ) suggested that Ndk-mediated inhibition of ROS reduces the macrophage killing capability . To verify this hypothesis , we repeated the survival assay using primary murine macrophages in which ROS production was blocked with gp91 INH . At 72 h post-infection , control experiments showed a significant reduction ( ∼1 . 5 Log10 ) in CFU counts when infecting with Mtb Ndk-AS relative to wild type or Ndk-S ( Fig . 7A ) . However , in the presence of gp91 INH , Ndk-AS survival was restored to a level comparable to that of wild type Mtb at every time point measured ( Fig . 7B ) . These observations were validated with assays in the presence of control scrambled peptide , which did not affect Ndk-AS survival . Taken together , these experiments clearly demonstrated that down modulation of ROS production by Ndk contributes significantly to Mtb persistence in the macrophage .
Previous studies from this laboratory showed that Mtb Ndk exhibits GAP activity towards Rab5 and Rab7 leading ultimately to diminished phagosomal recruitment of their respective effectors EEA1 and RILP [16] , [17] . Defective recruitment of EEA1 and RILP correlated with reduced maturation of phagosomes containing Ndk mutant M . bovis BCG or Ndk-coated beads [17] . In the current study , we demonstrate that Ndk further enhances Mtb virulence by additional GAP activity towards the Rho GTPase Rac1 . We provided direct evidence that Ndk blocks phagosomal recruitment for both Rac1 and its partner molecule p67phox leading ultimately to inhibition of NOX2-mediated ROS production and ROS-mediated apoptosis . A link between Mtb GAP activities and virulence was established with the observation of reduced survival of Ndk mutant Mtb in vitro and in vivo . Ndk is a ubiquitous small protein ( ∼15 kDa ) found in virtually all organisms , from eukaryotes to prokaryotes . In Mtb , Ndk catalyzes the production of nucleoside triphosphates as precursors for RNA , DNA and polysaccharide synthesis , which are critical for normal bacterial physiology [36] . This possibly explains why our attempts to knock out the ndkA gene in Mtb were unsuccessful , suggesting that Ndk is probably essential for Mtb growth . Contrasting with this hypothesis , an effort to comprehensively identify all genes required for Mtb growth using the transposon site hybridization ( TraSH ) technique suggested that the Ndk gene is not essential for Mtb growth [37] . However , as cautioned by the authors of that study , TraSH is simply a screening tool and therefore cannot provide a definitive conclusion about gene essentiality . Indeed , several genes known to be essential for Mtb growth , such as ideR [38] , rmlD [39] and whiB2 [40] have not been identified as essential by the TraSH approach . Thus , whether or not Ndk is essential is a research question that is still open for further investigation and remains beyond the scope of our current study , which focused instead on deciphering the mechanisms by which Ndk promotes Mtb survival in the host . Mycobacterial Ndk has been shown to interact with and inactivate recombinant Rho , Cdc42 and Rac1 proteins [41] . Here we found that within the macrophage , both Mtb and recombinant Ndk ( delivered on the surface of latex beads ) interact with and inactivate native Rac1 , but not Rho or Cdc42 . This suggests that results obtained from cell free systems do not always reflect host-pathogen interactions in the whole cell system . Not surprisingly , this type of discrepancy has been observed with other pathogens that use GAP activities as a mechanism of virulence . For instance , secreted YopE from Yersinia , and SptP from Salmonella were shown to have GAP activity towards all three Rho GTPases extracellularly . However , YopE acts only on Rac1 and RhoA , [42] whereas SptP inactivates Rac1 and Cdc42 , but not RhoA , [43] in cultured cells . In the case of Yersinia , a recent study established a direct link between YopE-mediated inactivation of Rac1 and inhibition of ROS production [44] consistent with our findings that selective Ndk GAP activity towards Rac1 is sufficient to block ROS production in the macrophage . Inhibition of ROS production in nascent phagosomes has also been reported in macrophages infected with the protozoan Leishmania , an intracellular pathogen that is structurally and metabolically distinct from Mtb , which interferes with NOX2 by a mechanism independent of GAP activities [45] . Indeed , Leishmania was shown to use its abundant surface lipophosphoglycan to restrict phagosomal recruitment of both p47phox and p67phox but not Rac1 . Conversely , our study showed that Ndk disrupts the recruitment of Rac1 and its binding partner p67phox but not p47phox . The p47phox subunit contains a PRR ( Prolin Rich Region ) at its C-terminus that binds with high affinity to the C-terminal SH3 domain of p67phox in the cytosol [46] , [47] . It also contains a PH ( Pleckstrin Homology ) domain that interacts specifically with membrane PI[3 , 4]P2 and phosphatidic acid [48] . While the tail-to-tail association of p47phox and p67phox plays a crucial role in NOX2 assembly [27] , recent studies showed that it is rapidly disrupted after membrane translocation [49] . Therefore dissociated p47phox and p67phox would remain separately attached to the phagosome via membrane lipids and Rac1 respectively . This phagosomal configuration of NOX2 subunits is consistent with the specific dissociation of p67phox from Mtb phagosomes as a result of Ndk-mediated Rac1 inactivation . The overall emerging picture from ongoing studies of phagosome remodelling by Mtb suggests that more than one virulence determinant might act in concert to modulate a single event of phagosome biogenesis . For instance , the cell wall glycolipid lipoarabinomannan , which blocks the Ca2+ signaling pathway [50] , synergizes with the acid phosphatase SapM , which hydrolyzes PI[3]P [51] , to abolish phagosome maturation processes that are dependent on recruitment of EEA1 . Such a synergism appears to also be the case for mycobacterial interference with NOX2 activity on the phagosomal membrane . Indeed , a recent study showed that the NuoG subunit of the type I NADH dehydrogenase also promotes Mtb interference with NOX2 activity , as evidenced by increased levels of ROS on Mtb ΔnuoG phagosomes [8] . However , the finding that NuoG is not secreted raises a question about the mechanistic connection between distant NuoG , contained within the bacterial cytosol , and NOX2 components on the cytosolic face of the phagosome membrane . Conversely several different groups have shown that Mtb Ndk is secreted [52]–[54] , suggesting that Ndk could translocate to the cytosolic surface of the vacuole where it interacts with Rac1 . In fact , our EM and confocal data revealed that secreted Ndk crosses the phagosomal membrane towards the cytosol . Consistent with these findings , previous studies showed that live Mtb exports a variety of proteins and glycolipids intracellularly [55]–[57] , and that many of them cross the phagosomal membrane towards the host cell cytosol to interact with and inhibit critical regulators of phagosome biogenesis [51] , [56] , [58] . A possible mechanism for the cytosolic translocation of mycobacterial products is the generation of a semi-porous phagosome membrane by the Mtb ESX-1 secretion system [59] , which was also shown to play an essential role in Mtb escape from the phagosome in later stages of infection [60] , [61] . Therefore it is possible that the ESX-1 secretion system also mediates translocation of Ndk to the cytosol . A highly relevant finding from the present study is that Ndk knock down converted virulent Mtb into an attenuated strain that lost resistance to the hostile environment of the host cell . Indeed , Mtb Ndk-AS infected cells were able to generate NOX2-dependent ROS production and also to undergo apoptosis thus ensuring maximal restriction of bacterial proliferation . In contrast , virulent Mtb strains were shown to down-modulate apoptosis in favor of necrosis [62] , [63] , which releases viable intracellular bacilli for further spreading of the infection and tissue damage during active tuberculosis disease . The link between ROS production , apoptosis and intracellular killing demonstrated in our study is consistent with earlier findings that intracellular oxidative stress induces phosphatidylserine externalization and increased caspase 3 activity [64] , [65] , and that apoptosis induced by the Fas ligand attenuates Mtb survival within the macrophage [66] . In addition to restricting the niche for mycobacterial replication , macrophage apoptosis contributes indirectly to the initiation of adaptive immunity mediated by dendritic cells . Indeed , infected macrophages undergoing apoptosis shed vesicles loaded with bacterial material ( or apoptotic blebs ) that prime dendritic cells for enhanced presentation of mycobacterial antigen to T cells [13]–[15] . In summary , while the role of Ndk in physiological processes has been intensively investigated , its contribution to Mtb pathogenesis has not been previously addressed . Our recent findings and current investigations have extended the knowledge of the biological effects of Ndk , to include inactivation of GTPase effector functions in the macrophage , therefore highlighting a novel strategy used by Mtb to circumvent host innate immunity
DMEM , Fetal calf serum ( FCS ) , and HBSS were purchased from Gibco Laboratories ( Burlington , ON , Canada ) . Luminol , CM-DCFDA , Annexin V-488 , and CellMask Deep Red , were purchased from Invitrogen ( Burlington , ON , Canada ) . Endotoxin-free culture reagents were from StemCell Technologies ( Vancouver , BC , Canada ) . Protease inhibitor mixture , PMSF , and trypsin-EDTA were purchased from Sigma-Aldrich ( St . Louis , MO ) . Protein A-agarose beads were from Bio-Rad laboratories ( Hercules , CA ) . Aldehyde/sulfate latex beads ( diameter , 4 µm ) were obtained from Interfacial Dynamics ( Portland , OR ) . gp91phox inhibitor peptide and its scrambled version [34] were synthesized by GenScript ( Piscataway , NJ ) . Rac1 , RhoA , Cdc42 antibodies were purchased from Millipore ( Temecula , CA ) . p67phox antibody was purchased from BD Transduction Laboratories ( Mississauga , ON , Canada ) and 47phox antibody was purchased from Santa Cruz Biotechnology ( Santa Cruz , CA ) . Cleaved caspase-3 ( Asp175 ) antibody was purchased from Cell Signaling ( Danvers , MA ) . Alexa Fluor 647-conjugated anti-rabbit IgG was purchased from Invitrogen . FITC-conjugated anti-rabbit and anti-mouse IgG were purchased from Sigma Aldrich . M . tuberculosis H37Rv and its derivative strains were grown in Middlebrook 7H9 broth ( BD Diagnostic Systems , Mississauga , ON , Canada ) supplemented with 10% ( v/v ) OADC ( oleic acid , albumin and dextrose solution; BD Diagnostic Systems ) and 0 . 05% ( v/v ) Tween 80 ( Sigma-Aldrich ) at 37°C on standing culture . Mtb Ndk-AS and Ndk-S were generated using our integrative pJAK1 . A plasmid ( selection marker , kanamycin [17] encoding the full length ndkA gene in sense and anti-sense orientations as described [17] . To generate red-fluorescent bacteria , Ndk-S , Ndk-AS , and wild type Mtb strains were transformed with pSMT3 vector ( selection marker , hygromycin ) encoding the DsRed protein as described [16] . RAW 264 . 7 macrophages ( ATCC , Manassas , VA ) were maintained in 10 cm diameter culture dishes ( Corning Inc . , Corning , NY ) at a density of ∼105 per cm2 in Endotoxin-free DMEM containing 5% FCS and 1% each of L-glutamine , penicillin-streptomycin mixture , HEPES , non-essential amino acids ( 100× solution , StemCell ) . Bone marrow derived macrophages ( BMDM ) were obtained by flushing out femurs and tibias of 6–8 week old female C57BL/6 ( Jackson Laboratory , Sacramento , CA ) according to protocols approved by the University of British Columbia Animal Care and Use Committees . Cells were then maintained in complete DMEM containing 10 ng/mL M-CSF for 6 days . For macrophage infection , bacteria in mid-log phase were harvested by 5 min centrifugation at 8 , 000× g . They were subsequently washed twice with 7H9 plus 0 . 05% tween and passed several times through 25 gauge needles to break bacterial clumps . Thereafter , numbers of bacteria were normalized by optical density ( OD600 1 . 0 = 3×108 bacteria/ml ) and adjusted for the desired MOI . Macrophages were then exposed to Mtb strains in complete media without antibiotic for 2 h at 37°C and then washed thrice to remove extracellular bacteria . Cells were reincubated in complete media plus gentamicin ( 50 ug/ml ) for the desired time periods . Groups of 4- to 6-week old female Fox Chase SCID mice ( CB17/Icr-Prkdcscid/IcrCrl ) were infected with ∼150 bacteria by inhalation using a Glas-Col inhalation exposure system ( Terre Haute , IN ) . Two mice from each group were processed on day 1 following infection to confirm bacterial deposition in the lung . Remaining animals were monitored for signs of morbidity . Mice were then euthanized and the bacterial load ( CFUs ) in the lung was determined . Organs were homogenized and serial dilutions plated in duplicate on nutrient 7H10 agar . In other experiments , SCID mice were injected subcutaneously in the scruff of the neck with 106 Mtb strains and then monitored for morbidity over a period of ∼15 weeks . All animals were maintained in accordance with protocols approved by the Animal Care and Use Committees at the University of British Columbia . Experiments were approved by the Animal Care and Usage Committees and performed according to the Canadian Council on Animal Care Guidelines . The animal assurance welfare number is A11-0247 . Ndk was expressed as a C-terminal 6×His tagged fusion protein in E . coli strain BL21 and purified using Ni-NTA purification resin ( Qiagen ) as described [17] . The purity of eluted Ndk was confirmed by SDS-PAGE and Coomassie Blue staining ( Fig . S10A ) . Rabbit Ndk antibody was generated by GenScript , using KLH conjugated ELASQHYAEHEGK peptide fragment corresponding to amino acids 44 to 56 of Mtb Ndk . The specificity of Ndk antibody is shown in Fig . S10B . Ndk or the control BSA were non-covalently linked to latex beads as described [17] . Coverslips were mounted on microscope slides and examined by digital confocal microscopy as described [17] . Immunogold staining was conducted at the EM Facility of the James Hogg Research Centre ( Saint Paul Hospital , Vancouver , BC , Canada ) . In brief , Mtb-infected macrophages were fixed with 4% paraformaldehyde , embedded in 4% low melting point agarose and dehydrated in ethanol . Samples were then transferred to LR White resin . After polymerization at 50°C , 60 nm sections were cut with a Leica EM UC6 microtome ( Leica Microsystems , Switzerland ) and collected on nickel grids . Samples were labelled with Ndk antibodies then F ( ab′ ) 2 of ultra-small goat-anti-rabbit IgG ( Electron Microscopy Sciences , Hatfield , PA ) . Sections were then post-fixed in 2% glutaraldehyde and subjected to silver enhancement for gold labeling with Silver R-Gent SE-EM ( Aurion , Wageningen , Netherlands ) . Samples were then washed in distilled water , stained in 2% uranyl acetate , washed again , air dried and examined with a Tecnai 12 electron microscope ( FEI Company , Hillsboro , OR ) . Confluent RAW cells seeded on 6 cm plates were infected by Mtb strains for 1 h at a MOI of 20∶1 . Thereafter , cells were incubated in the presence of 200 ng/ml LPS for 15 min to induce Rho GTPase activation . Subsequently , cells were lysed in cold buffer containing 30 mM HEPES ( pH 7 . 2 ) , 100 mM NaCl , 10% glycerol , 1% Triton X100 , 1 mM EDTA , 10 mM MgCl2 , and 1 mM PMSF . Soluble protein fractions were analyzed for levels active Rho-GTPases by using a Rac/Rho/Cdc42 activation assay kit ( Millipore ) . Macrophages were cultured in complete DMEM in 96 well white plates ( Corning ) at 105 per well . Prior to ROS assay , cell media was replaced with DMEM without phenol red and luminol was added to a final concentration of 50 µM . Wells were then infected with Mtb strains ( MOI 10∶1 ) or coated beads ( MOI 5∶1 ) . Thereafter , plates were loaded into a Tropix TR717 microplate luminometer ( Applied Biosystems , Bedford , MA ) adjusted to 37°C and relative luminescence was then measured at 60 sec intervals over 60 min . Intracellular detection of ROS was achieved by incubating adherent macrophages to cover slips with 10 µM CM-DCFDA for 30 min at 37°C prior to infection with Mtb strains expressing DsRed . Cells were then analyzed by confocal microscopy . BMDMs , but not RAW cells , were primed with LPS ( 100 ng/ml , 48 h ) prior to ROS assays because expression of fully functional NOX2 complex in BMDMs requires priming with LPS or TNFα [67] , [68] . Adherent RAW cells on coverslips were infected with Mtb strains . At 48 h post phagocytosis , cells were washed twice with cold PBS and then incubated in Alexa Fluor 488 Annexin V ( 1∶20 , Invitrogen ) in staining buffer containing 10 mM HEPES ( pH 7 . 4 ) , 140 mM NaCl , and 2 . 5 mM CaCl2 , for 20 min at room temperature . Coverslips were then analyzed by confocal microscopy . Alternatively , infected cells were scraped and fixed in PBS plus 2% paraformaldehyde for 15 min at room temperature . Cells were then washed with PBS and incubated with anti-cleaved caspase-3 ( 1∶250 ) in permeabilization buffer ( 0 . 1% Triton X100 and 1% BSA in PBS ) for 20 min at room temperature . Thereafter , cells were washed and stained with Alexa Fluor 647-conjugated goat anti-rabbit IgG ( 1∶200 ) for 20 min at room temperature , washed again and analyzed by FACS . | Mycobacterium tuberculosis ( Mtb ) is a very successful intracellular pathogen that infects lung macrophages . Its resistance to intracellular killing has been linked to the development of pulmonary tuberculosis ( TB ) in humans . Thus , understanding the mechanism by which Mycobacterium tuberculosis ( Mtb ) persists in the host is a prerequisite for development of efficient strategies to control TB disease . We have previously shown that Mtb nucleoside diphosphate kinase ( Ndk ) contributes to phagosome maturation arrest via inactivation of Rab5 and Rab7 . In this study , we show that Ndk also targets and inactivates the small GTPase Rac1 , an essential component of the macrophage NADPH oxidase ( NOX2 ) complex . Ndk-dependent inactivation of Rac1 was associated with reduced NOX2-mediated production of reactive oxygen species ( ROS ) and ROS-dependent apoptosis . Conversely , disruption of Ndk expression converted Mtb into a mutant strain that induces strong ROS and apoptosis responses . This phenotype was associated with reduced survival of Ndk mutant in vitro and in vivo . Altogether , our findings demonstrate that Ndk contributes significantly to mycobacterial virulence . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"host-pathogen",
"interaction",
"biology",
"microbiology"
] | 2013 | Mycobacterium tuberculosis Nucleoside Diphosphate Kinase Inactivates Small GTPases Leading to Evasion of Innate Immunity |
Many animal organs are composed largely or entirely of polarized epithelial tubes , and the formation of complex organ systems , such as the digestive or vascular systems , requires that separate tubes link with a common polarity . The Caenorhabditis elegans digestive tract consists primarily of three interconnected tubes—the pharynx , valve , and intestine—and provides a simple model for understanding the cellular and molecular mechanisms used to form and connect epithelial tubes . Here , we use live imaging and 3D reconstructions of developing cells to examine tube formation . The three tubes develop from a pharynx/valve primordium and a separate intestine primordium . Cells in the pharynx/valve primordium polarize and become wedge-shaped , transforming the primordium into a cylindrical cyst centered on the future lumenal axis . For continuity of the digestive tract , valve cells must have the same , radial axis of apicobasal polarity as adjacent intestinal cells . We show that intestinal cells contribute to valve cell polarity by restricting the distribution of a polarizing cue , laminin . After developing apicobasal polarity , many pharyngeal and valve cells appear to explore their neighborhoods through lateral , actin-rich lamellipodia . For a subset of cells , these lamellipodia precede more extensive intercalations that create the valve . Formation of the valve tube begins when two valve cells become embedded at the left-right boundary of the intestinal primordium . Other valve cells organize symmetrically around these two cells , and wrap partially or completely around the orthogonal , lumenal axis , thus extruding a small valve tube from the larger cyst . We show that the transcription factors DIE-1 and EGL-43/EVI1 regulate cell intercalations and cell fates during valve formation , and that the Notch pathway is required to establish the proper boundary between the pharyngeal and valve tubes .
Epithelial tubes are fundamental components of animal organs , and perform many functions such as the transport of liquids , gases or food ( reviewed in [1] ) . Epithelial tubes range in shape from simple cylinders to the branched , convoluted structures of the Drosophila trachea or mammalian kidney ( reviewed in [2] , [3] ) , and tube formation can involve extensive remodeling of the constituent cells [4]–[8] . The C . elegans digestive tract provides a genetic model system for understanding how epithelial tubes form and remodel . The tract consists primarily of three linked tubes , the pharynx , valve , and intestine ( Figure 1 ) . The intestine is a simple , cylindrical tube composed of 20 similar cells that derive from a single early blastomere [9] . The pharyngeal tube is similar in size to the intestine , about 50 microns at hatching , and has only a slightly more complex shape . However , the pharynx is derived from multiple early blastomeres , and contains 80 cells that differentiate into five cell types [9] , [10] . The different pharyngeal cell types have diverse shapes , and even cells of the same type have distinct , position-specific morphologies associated with pharyngeal structure and function [10] . The intestine and pharynx form from separate , but adjacent , primordia that polarize at different times in development , apparently using different polarization cues [11] , [12] . The valve tube , which contains only six cells , links these larger tubes to form a continuous digestive tract [9] , [10] . The pharyngeal primordium includes the future valve cells , and begins as an aggregate of precursor cells that each express the transcription factor PHA-4/FoxA , a key regulator of pharyngeal development ( Figure 1 ) [13]–[15] . The precursors organize into a bilaterally symmetrical array we call the double plate , which resembles two adjacent plates of cells , each plate one cell in thickness ( Figure 1 ) [12] , [16] . After cell ingression and division complete the double plate , PAR polarity proteins localize near the junction of the left and right plates; cell membranes at the junction are the future apical surfaces . This localization appears to result from a laminin-dependent cue at the opposite , future basal , surfaces at the perimeter of the double plate . Apical constriction reshapes most cells into wedges and transforms the double plate into a rounded cyst ( Figure 1 ) [12] . How cyst cells remodel into the valve has not been analyzed . However , previous studies suggest that multiple mechanisms contribute to individual pharyngeal cell shapes . Pharyngeal gland cells have long , thin processes that connect to the lumen and secrete mucin-like proteins [10] , [17] . The basic gland cell shape appears to result from the cell body migrating away from the site of lumen attachment , spooling out the secretory process in its wake , and gland cell shape is further modified by interactions with surrounding pharyngeal muscles [9] , [18] . The longitudinal processes of some pharyngeal neurons may develop from a similar spooling mechanism , as they develop independently of several genes required for axon guidance [19] . In the anterior of the cyst , where pharyngeal cells must ultimately link with epidermal cells to complete the digestive tract , remodeling does not appear to involve either cell migration or cell intercalation [20] . Instead , the anterior cyst cells reorient their apicobasal axes and apical membranes to align with adjacent epithelial cells , thus forming a topologically continuous apical surface [20] , [21] . At the posterior of the pharynx , a single donut-shaped cell called pm8 ( pharyngeal muscle 8 ) has the critical role of forming a boundary with the valve . pm8 begins as a dorsal cell in the cyst , but extends a process through ventral cells along a transient tract of laminin . pm8 then autofuses into a toroid , possibly by wrapping around finger-like projections from neighboring cells [22] . Here , we examine how the valve tube originates from the cyst and links with the pharyngeal and intestinal tubes . The terminal pharyngeal cell , pm8 , and all the cells that form the valve are located at the posterior of the cyst , initially within a disc of polarized , wedge-shaped cells . The adjacent intestinal cells appear to coordinate valve cell polarity by preventing a polarizing cue , laminin , from accumulating at the posterior surface of the disc . Multiple pharyngeal and valve cells extend lateral , actin-rich lamellipodia between neighboring cells that precede several examples of cell repositioning . The developing valve tube must align with the intestinal tube , and we show that linking cells called v3D and v3V dock at the left-right boundary of the intestinal cells . Other cells organize symmetrically around v3D and v3V , and begin bilaterally symmetrical , highly patterned cell intercalations that eventually reshape the disc into three smaller rings of valve cells . These morphogenetic events involve the transcription factors DIE-1 and EGL-43 . The pharynx links to the valve at the boundary between pm8 and the most anterior valve cell , v1 . pm8 and v1 have similar origins , occupy symmetrical positions in the cyst , and initiate bilaterally symmetrical wrapping around the lumenal axis such that both cells become toroidal or donut-shaped . During wrapping , however , pm8 intercalates asymmetrically , anterior to v1 , and thus joins with other pharyngeal cells . Notch signaling occurs in pm8 , but not v1 , prior to morphogenesis , and our results suggest that Notch signaling regulates multiple , pharyngeal-specific properties of pm8 . Thus , the asymmetry in intercalation order is important because only pm8 has the potential to become a pharyngeal cell .
The valve is a short tube of six cells that are organized into three rings , referred to here as rings v1–v3 . The first , most anterior ring consists of the single v1 cell , the second contains three v2 cells , and the third contains two v3 cells ( Figure 2A ) . The posterior end of the valve connects with the intestine through a ring of four intestinal cells ( the int1 ring ) ; other intestinal cells are organized in sequential rings of two cells each ( int2–int9 rings ) [9] . All 20 intestinal cells are clonal descendants of an early embryonic blastomere called E; this paper focuses on the E16 and E20 stages of the intestinal primordium , which contain 16 and 20 E descendants , respectively . The anterior end of the valve connects with the pharynx , which consists of myoepithelial cells called pharyngeal muscles ( pm ) , structural cells called marginal cells ( mc ) , epithelial cells , neurons , and gland cells ( Figures 2A and S1A ) [10] . Most pharyngeal cell types can be subdivided into groups based on region-specific morphological and molecular differences along the anterior-posterior axis ( Figures 2A and S1A ) . Each group contains one to six cells , and cells within a group typically are organized with threefold rotational symmetry around the midline or lumenal axis . Cells can be referenced by a group name such as pm3 ( pharyngeal muscle group 3 ) , or by a specific anatomical name such as pm3DL ( pm3 Dorsal Left; Figure S1A ) . For simplicity , we use group names whenever possible , and provide specific names in Figure S1 for reference . Because the C . elegans lineage is essentially invariant , most cells at birth have predictable , specific fates within differentiated tissues [9] . In the present study , we found no examples of wild-type fate variability by either direct lineage analysis or by lineage-specific transgene expression . Thus , we refer to each cell at birth by its future , differentiated fate rather than its lineage name . For example , we refer to the cell MSaaapapp simply as pm8 . Tube formation and remodeling were examined by confocal microscopy of live embryos expressing fluorescent reporters . A typical experiment is shown in Figure 2B , and combines reporters for either pharynx-specific or general plasma membranes [mig-13::MIG-13::GFP , pha-4::GFP::CAAX , or pie-1::mCherry::PH ( PLC1δ1 ) ] [20] , [23] , [24] , and a marginal cell reporter used for spatial reference [pax-1::HIS-GFP] [25] . In addition , we included reporters for intestinal cell membranes and/or nuclei [end-1::GFP::CAAX , F22B7 . 9::GFP] [26] , [27] . Optical planes shown are sagittal ( S ) , horizontal ( H ) or transverse ( T ) as illustrated in Figure S1A . Cells were identified according to their descent from specific early embryonic blastomeres , and/or by reporter expression and later position within the differentiated pharynx , valve , or intestine . We examined cell contacts as the pharyngeal and valve precursors aggregate into the double plate primordium , and as the double plate transforms into the cyst ( Figures 1 and S1B ) . Many pharyngeal cells maintained contact with their immediate neighbors during the double plate to cyst transformation , consistent with previous observations on nuclear positions [16] . By contrast , cells at the posterior end of the double plate , which include pm8 and all of the valve cells , appeared to change many of their neighbors ( Figure S1B ) . To better visualize these events , we identified and traced contours of all cells at the interface between the pharyngeal/valve primordium and the intestinal primordium , then used the tracings to generate 3D models of the interface cells ( Figure 2C–2E , Video S1 ) . At about 307 minutes in embryogenesis ( double plate-E16 stage ) , the interface contains 19 cells that include the valve cells and diverse pharyngeal cell types ( Figure 2C ) . All of the valve cells and most pharyngeal cells have completed their terminal divisions , and most are cuboidal in shape . The valve cells are split between separate dorsal ( v1 , v3D ) and ventral clusters ( v2R , v2L , v2V , v3V ) . The anterior end of the E16 intestinal primordium contains four cells at this time , but these are not the same four cells that comprise the anterior , int1 ring of the intestine ( compare Figure 2C with 2A ) : The dorsal two intestinal cells at the interface are the parents of the int1 cells that form the int1 ring . We refer to these dorsal cells throughout this report , and for convenience call them the int1p cells ( for int1 parents ) . The ventral two intestinal cells at the interface ( int2D and int2V ) eventually form the second , or int2 , ring of the intestine . By 355 minutes ( cyst-E16 stage ) , the interface has been reduced to 12 cells ( Figure 2D ) . Most pharyngeal and valve cells have undergone apical constriction , resulting in wedge-shaped cells that surround the midline of the cyst ( compare transverse views in Figure 2C and Figure 2D ) [12] . The previously separate clusters of valve cells have moved closer together , and this movement occurs as the intervening pharyngeal cells split away from the intestinal cells; the only pharyngeal cells that remain at the interface are muscles ( pm8 , pm7L , and pm7R ) . As described further below , the dorsal valve cell v3D becomes positioned diametrically opposite v3V ( red cells in Figure 2D ) , and a line drawn between these cells corresponds to a new , dorsal-ventral axis that prefigures bilaterally symmetrical movements of the flanking cells . Only three intestinal cells remain at the interface because one of the ventral cells ( int2D ) intercalates behind one of the dorsal cells ( the right int1p cell ) , and both dorsal cells spread ventrally . By 370 minutes ( cyst-E20 stage ) , the pm7 muscles ( pm7R , pm7L ) have spread to nearly cover the anterior surfaces of the ventral valve cells , thus separating those valve cells from other pharyngeal cells ( Figure 2E ) . The pm7 muscles have lost most of their direct contact with the intestinal primordium . The two int1p cells have divided , and their four daughters close together ventrally to complete the final int1 ring . The int2 cells ( int2D and int2V ) have now intercalated behind the int1 cells , essentially wrapping clockwise around the lumenal axis ( as viewed from the anterior ) to form the second , int2 ring of the intestine . We conclude that the valve cells initially are intermingled with pharyngeal cells , and that both cell types make extensive contacts with intestinal cells . As the cyst forms and begins to remodel , however , only the valve cells and a single pharyngeal cell , pm8 , retain contact with the intestinal primordium . The posterior end of the differentiated valve lumen is centered between the v3D and v3V valve cells ( Figure 2A ) . These cells connect with the anterior end of the intestinal lumen , which is centered on the int1 ring at the intersection of the left-right and dorsal-ventral boundaries of the int1 cells ( Figure 2A ) . Because the pharyngeal/valve and intestinal primordia polarize at different times in development , their separate midline axes must be brought into alignment to form a continuous lumenal surface ( Figure S2A , S2B ) [11] , [12] . We found that both v3 cells dock at the left-right boundary of the int1p intestinal cells , and remain there until the dorsal-ventral division of the int1p cells completes the int1 ring . The parent of v3D creates or occupies a deep pocket between the left and right int1p cells as the latter cells become polarized; after division , the sister of v3D undergoes apoptosis [9] while v3D continues to occupy the pocket ( Figure 3A ) . Reporters for the adhesive protein HMR-1/E-cadherin and the polarity protein PAR-6 become highly enriched where the intestinal cells contact v3D ( closed arrowheads , Figure 3B ) , but not at adjacent contacts with other cyst cells ( open arrowheads , Figure 3B; see also S2C ) . The v3D cell body spreads dorsally during the next 40 minutes , however , a thin extension remains embedded in a small cleft along the left-right boundary between the int1p cells ( Figures 3C and S2C ) . Thus , in transverse view v3D has a wedge shape ( Figure 3D ) that resembles the shape of later cyst cells , but that contrasts with the shape of its contemporary neighbors in the double plate primordium ( Figure 3E ) . This change in v3D shape appears to be initiated by a dorsal-directed lamellipodium ( arrow , Figure 3F ) rather than by apical constriction , and begins before other pharyngeal or valve cells show a polarized localization of PAR proteins ( Figure S2B ) . The v3V cell , which is the ventral counterpart of v3D , initially shows an analogous localization at the left-right boundary between ventral E16 cells ( int2D and int2V; white arrow in Figure 4A; see also Figure 2C ) . However , v3V does not , and cannot , remain at this boundary because int2D and int2V do not contribute to the int1 ring ( Figure 4A , Video S1 ) . Instead , the int1p cells spread ventrally until they reach , and stop at , the left and right margins of v3V ( Figure 4A; see also Figure 2E ) . Thus , both v3D and v3V align at the left-right boundary of the int1p cells , near the final position of the intestinal lumen; v3D moves to the intestinal cell boundary , whereas the boundary extends to flank v3V . Previous studies showed that the zinc-finger transcription factor DIE-1 ( dorsal intercalation and elongation defective-1 ) is broadly expressed in pharyngeal and intestinal precursors at the late double plate stage , and that die-1 mutants can have variable gaps in the apical junctions of the digestive tract [28] . We fixed and immunostained die-1 embryos with an antibody that recognizes a component of apical junctions and found that the gaps in the apical junctions of the digestive tract occurred anterior to the intestine , near the normal positions of valve cells ( 4/11 die-1 embryos , Figure 5A ) . These gaps contained multiple cells and were associated with a misalignment of the pharyngeal and intestinal midlines . In the wild type digestive tract , pharyngeal , valve , and intestinal cells all share a common , radial axis of apicobasal polarity [10] , [11] . By contrast , some presumptive ventral valve cells in die-1 mutants had axes of apicobasal polarity that were not aligned with the adjacent intestinal cells ( Figure 5A ) . We examined the development of die-1 mutant embryos to determine the origin of the apparent valve defects . In each of four die-1 embryos , v3D was born at its normal position in the double plate , but v3D did not extend dorsally along the left-right boundary between the int1p cells ( Figure 3G ) . In one of these embryos , v3D extended an abnormal process between the int1p and int2 cells , but the process later retracted ( data not shown ) . On the ventral side of the cyst , v3V was born at the normal time and place in the die-1 embryos . However , v3V remained abnormally rounded , and showed little or no contact with the int1p cells ( Figure 4B and data not shown ) . Thus , both v3D and v3V require DIE-1 activity for proper association with the left-right intestinal boundary . In the course of analyzing die-1 embryos , we noticed penetrant and abnormal cell contacts between the germ cell precursors and one or more of the ventral valve cells ( compare Figure 4B with 4A at 360 minutes ) . During the wild-type double plate stage , processes from the int2 intestinal cells begin to intercalate between the germ cell precursors and the parents of the valve cells v2L and v2R ( each valve cell parent is labeled p in Figure 5B , 5C ) . By the time v2L and v2R are born ( asterisks in Figure 5B , 5C ) , their posterior surfaces are covered completely by the int2 cells . Shortly thereafter , the additional ventral valve cells ( v2V and v3V ) intercalate between v2L and v2R , and similarly are separated from the germ cell precursors ( Figure 4A and data not shown ) . The int2 processes cover the posterior surfaces of the ventral valve cells until these contacts are replaced by processes from the int1p , and finally int1 , cells as described above ( Figure 5B ) . Thus , intestinal cells normally cover the posterior surfaces of the ventral valve cells through all stages of cyst formation and remodeling ( brackets , Figure 5B ) . By contrast , we found that the intestinal cells of die-1 mutants either failed to spread across the posterior surfaces of the ventral valve cells , or the intestinal cells extended only small processes that later retracted ( Figure 5D ) . The close association of intestinal cells with the posterior surfaces of valve cells was intriguing for two reasons . First , laminin at the periphery of the double plate appears to orient the apicobasal polarity of many or all pharyngeal cells [12] . Second , despite accumulating at other peripheral surfaces of the double plate and cyst , laminin appears to be excluded from the posterior surface ( arrowhead , Figure 5E ) . We identified the valve cells v2L and v2R in fixed and immunostained wild-type embryos at the double plate and cyst stages ( asterisks in Figure 5G ) , and confirmed that laminin is not detectable at the posterior surfaces of these cells ( arrowhead , Figure 5G ) . To determine whether contact with the intestinal primordium prevents laminin accumulation at the posterior of the double plate and cyst , we used a laser microbeam to kill the intestinal precursor in 12-cell , wild-type embryos . Each of 18 embryos that developed without an intestinal primordium showed ectopic laminin accumulation on part or all of the posterior surface of the double plate and cyst ( arrowhead , Figure 5F ) . Similarly , we found that unablated die-1 mutants embryos had ectopic laminin on the posterior surfaces of v2L , v2R , and neighboring cells that fail to contact intestinal cells ( arrowhead , Figure 5H ) . Ablated wild-type embryos that were allowed to develop to terminal stages all developed a relatively normal pharynx , with a basement membrane delimiting the pharynx from the valve ( n = 20; Figure 5I , 5J ) . Pharyngeal cells in the ablated embryos appeared to have the normal , radial axis of apicobasal polarity , as did the most anterior valve cells ( Figure 5I–5L ) ; indeed , two of the embryos hatched and ingested material ( data not shown ) . However , the posterior valve cells formed an abnormal , ball-like structure surrounding a truncated lumen ( Figure 5J , 5L ) . The apical surfaces of the most posterior valve cells were oriented anteriorly , indicating that they have a longitudinal or oblique axis of apicobasal polarity instead of the normal , radial axis ( Figure 5J , 5L ) . These results suggest that contact with the intestinal primordium normally contributes to valve cell polarity by preventing a polarizing cue , laminin , from accumulating on the posterior surfaces of valve cells . Moreover , these results suggest that the apical junction and polarity defects in die-1 mutant embryos stem from earlier defects in intestinal cell intercalation . A comparison of interface cells at the late cyst stage ( Figure 2E ) with their final positions and shapes ( Figure 2A ) indicates that extensive remodeling must occur after the v3 cells dock with the intestine . To visualize the behaviors of individual cells in detail , we wanted to develop a reporter that was expressed in a subset of the interface cells . Previous studies showed that LIN-12/Notch is expressed in a large number of cells throughout embryogenesis , including pm8 and its immediate relatives near the time of their birth in the double plate primordium ( Figure 6A ) [22] . We identified an enhancer element ( lin-12pm8 ) within a lin-12 intron that can drive transgene expression in pm8 and its relatives , but not in other cells that normally express LIN-12 ( Figure 6A; Materials and Methods ) . The pm8 relatives include two muscles ( pm3DL and pm4L ) and one marginal cell ( mc3L ) ; for convenience we refer to pm8 and these three cells as the pm8 family . We used the lin-12pm8 enhancer to construct membrane-localized and nuclear-localized reporters , and used these in conjunction with non-specific membrane reporters to examine the development of the pm8 family . We found that all cells in the pm8 family developed dynamic lamellipodia that appeared to probe neighboring cells during the cyst stage ( Figure 6B , Video S2 ) . We constructed a strain where the pm8 family expresses a reporter for plasma membranes plus a second reporter with the actin-binding domain of moesin fused to GFP [29] ( Figure 6C ) . In time-lapse movies , moesin-GFP was strongly enriched at leading edges of the lamellipodia , suggesting that they are rich in filamentous actin ( arrow , Figure 6C ) . Simultaneous imaging from horizontal and transverse planes showed that the lamellipodia emerge from sub-basal , lateral surfaces of the wedge-shaped cyst cells , and that the lamellipodia usually extend across one , or occasionally two , neighboring cells ( Figure 6B , Video S2 ) . While most of the lamellipodia appear transient and variable , several of the longer-lived lamellipodia are highly reproducible and precede cell repositioning . For example , in normal development pm4L is a lateral muscle , but its sister is a dorsal muscle separated from pm4L by a longitudinal row of marginal cells ( Figure S1A ) . In live imaging , we found that these sisters initially are on the same , dorsal side of the marginal cell row . However , pm4L develops a lamellipodium that breaches the marginal cell row , and the bulk pm4L cytoplasm follows the lamellipodium into the lateral muscle group ( Figure 6B ) . Importantly , pm4L maintains its basic wedge-shaped appearance and association with the midline throughout the repositioning , such that the apparent circumferential migration is essentially a rotation of the pm4L cell body around the midline ( Videos S2 and S3 ) . Another prominent lamellipodium extends from the right side of pm8 , beginning at about 350 minutes ( arrow , Figure 6D ) . As the lamellipodium crosses the anterior face of v3D , it meets a mirror-image lamellipodium extending from the left side of v1 ( Figures 6D and S1C ) . The pm8 and v1 processes remain , without crossing , on the anterior face of v3D for at least 45 minutes , during most of the remodeling events described below . We found that HMR-1/E-cadherin::GFP becomes highly enriched in or near the tips of the pm8 and v1 lamellipodia as they converge , suggesting that E-cadherin contributes to the stability of these cell contacts ( Figure 6E ) . The above results show that after v3D docks at the left-right boundary between the intestinal cells , the flanking pm8 and v1 cells come together , and possibly anchor , on the face of v3D . At this stage , pm8 and the valve cells are discontinuous , wedge-shaped cells at various positions around the midline of the cyst ( Figure 6F ) . For example , the valve cells that later form the v2 ring ( v2R , v2V , and v2L ) are all on the ventral side of the midline . At about 350 minutes , the int1p cells divide to generate the final int1 ring , thus defining the position of the intestinal lumen . Shortly after the int1p division , pm8 and the valve cells begin large-scale intercalations that eventually encircle the midline . The general pattern of the intercalations is illustrated in Figure 6F , with specific intercalation paths diagrammed in Figure 6G and described below . Between 400 and 410 minutes , pm8 extends a ventral-directed lamellipodium that is followed by the pm8 nucleus ( Figure 7A ) . In transverse view , the lamellipodium sweeps down the left side of the cyst , across the ventral side , and up the right side to close back on itself ( Figures 7A , 7B and S3A , S3B , Video S3 ) . This live imaging provides direct confirmation for our previous model that pm8 becomes a toroid or donut-shaped cell by circumferential wrapping [22] . The wrapping requires that pm8 intercalates between pairs of neighboring cells ( Figure 6G ) , and we found that intercalation was defective in die-1 mutants: pm8 extends a ventral process and shows limited nuclear migration into the process in die-1 mutants , but pm8 fails to complete wrapping ( Figure 7C ) . Thus , die-1 mutants appear defective in several different examples of cell intercalation during cyst remodeling . Because the remodeling of cells in the anterior of the cyst is associated with a re-orientation of the apicobasal axis ( see Introduction ) , we wanted to examine whether remodeling in the posterior cyst might involve changes in cell polarity . At and after the cyst stage , PAR-6 appears to be distributed continuously along the midline of the cyst , suggesting that PAR-6 remains at the apical surfaces of at least some posterior cells during remodeling ( Figure S2A and data not shown ) . We then used the lin-12pm8 enhancer to examine PAR-6::GFP expression specifically within the pm8 family . The pm8 family members localized PAR-6::GFP to their midline-facing , apical surfaces during the double plate to cyst transformation , similar to other pharyngeal cells . PAR-6::GFP remained apical throughout the morphogenesis of the pm8 family members , including during the rotational repositioning of pm4L away from the other family members ( double-headed arrow in Figure 7E ) and the ventral wrapping of pm8 ( arrowhead in Figure 7E ) . Thus , cyst cells appear to maintain their radial axis of apicobasal polarity while remodeling their orthogonal , lateral surfaces . Previous studies showed that a novel , radially oriented tract of laminin develops within the posterior cyst just before pm8 begins to change shape ( Figure 7F ) : The ventral extension of pm8 occurs in close proximity to this tract , and laminin function is required for normal pm8 morphogenesis ( Figure 7F ) [22] . With transverse optical sectioning , we found that the laminin tract has a well-defined wedge shape that closely approximates the shape of the adjacent marginal cell , mc3V ( Figure 7F and data not shown ) . This suggests that the pm8 lamellipodium encounters the tract only after it begins to probe between ventral cells , and raises the possibility that the different behaviors of the various pm8 lamellipodia result from the different environments they encounter . About 10–15 minutes after pm8 begins to extend ventrally on the left side of the cyst , v1 develops a symmetrical , ventral-directed lamellipodium on the right side of the cyst ( Figures 6F , 6G and S3C ) . The pm8 and v1 lamellipodia meet and pass dorsal-directed lamellipodia extending from v2L and v2R , respectively ( Figure S3C , S3D , Video S4 ) . The v2L and v2R lamellipodia precede their nuclei and bulk cytoplasm , shifting both cells bodies to the dorsal side . Similar to the pm8 and v1 intercalations , v2L and v2R appear to follow invariant trajectories through the cyst . For example , the v2R lamellipodium initially spreads dorsally between an int1p daughter ( one of the ventral int1 cells ) and the v1 lamellipodium or cell body ( n = 9 embryos; Figure 6G and S3C ) . The timing of the v2L and v2R intercalations are highly reproducible with respect to the timing of the int1p divisions: Shortly after the left ( or right ) int1p cell divides into a dorsal-ventral pair of int1 cells , v2L ( or v2R ) begins spreading on the adjacent , ventral int1 cell . The migrating v2L ( or v2R ) nucleus reaches the dorsal-ventral boundary between the int1 cells at 48+/−6 minutes ( n = 5 embryos; Figure S3C , S3D ) . This reproducibility is intriguing because the absolute division times of the int1p cells can vary by up to 20 minutes in different embryos , and the division times of the left and right int1p cells can differ by up to 15 minutes within the same embryo ( n = 18 embryos ) . Thus , we speculate that the int1p divisions might trigger valve cell intercalation , for example by modulating adhesion to the intestinal cell surface . Because E-cadherin has been shown to regulate intercalations in other systems [30] , [31] , we examined HMR-1/E-cadherin localization during valve cell intercalation . We found that embryos expressing a rescuing HMR-1::GFP transgene [32] showed a striking enrichment of HMR-1 along lateral membranes at the posterior of the cyst during valve formation ( arrow , Figure 8A ) . This localization contrasts with that in intestinal cells and most other cyst cells , where HMR-1/E-cadherin is predominantly at the midline or apical surface in association with apical junction proteins such as AJM-1 ( Figure 8A; see also Figure S2B ) . We identified the dorsal lines of HMR-1 as the boundaries between v3D and either v1 or pm8 , as described above for earlier embryos ( Figure 6E ) . This reproducible , and apparently persistent , zone of expression contrasted with variable localization of HMR-1/E-cadherin in the ventral cyst . In live embryos , we found that HMR-1::GFP localization in the ventral cyst is dynamic , and concentrated particularly around pm8 and the subset of valve cells that undergo nuclear migration ( v2L and v2R ) . For example , HMR-1::GFP is enriched in or by the ventral lamellipodium from pm8 , initially appearing as a single line of fluorescence ( 406 minutes , arrow in Figure 8B ) . The line of HMR-1::GFP then bifurcates as the pm8 nucleus and bulk cytoplasm shift ventrally ( 422 minutes , Figure 8B ) . HMR-1::GFP persists at high levels as the pm8 nucleus continues moving to the ventral interior of the cyst , where it is enriched between pm8 and mc3V near the same location as the laminin tract ( 430 minutes , Figure 8C ) . Similarly , high levels of HMR-1::GFP flank v2L as the v2L nucleus and cytoplasm move dorsally , but little or no HMR-1::GFP is associated with v2V , which stays largely in place ( Figure S3E ) . HMR-1/E-cadherin is expressed both maternally and embryonically; previous studies showed that embryos lacking embryonic expression appear to have relatively normal assembly of most tissues and organs , but fail in hypodermis ( skin ) -mediated processes such as ventral closure and body elongation [33] , [34] . We found that hmr-1 mutant embryos appeared to initiate all of the above intercalation movements , but had defects in nuclear migration and transfer of bulk cytoplasm . For example , the migrating v2R nucleus reached the dorsal-ventral boundary between int1 cells about 48 minutes after the int1p division , similar to wild-type embryos described above . In wild-type embryos , the v2R nucleus and bulk cell body continue to travel dorsally , and pass beyond the dorsal-ventral boundary in an additional 8+/−1 minutes ( n = 5 embryos; Figure 8D ) . However , in hmr-1 embryos , the v2R nucleus and bulk cell body appeared stalled at the dorsal-ventral boundary for at least 1 hour , although the v2R lamellipodium could extend further dorsal between the valve cells v1 and v3D ( arrowheads , Figure 8E; n = 3 embryos ) . At terminal stages , hmr-1 embryos showed defects in the apical junctions in the valve , including apparent gaps within , or anterior to , the v2 cells ( Figure 8F , 8G ) . In summary , valve cells create a tube by circumferential intercalation , or wrapping , around the midline . These movements condense the radius of the posterior cyst symmetrically around the midline while extending the longitudinal axis , effectively extruding the small , valve tube from the larger cyst . Larvae that are defective in the LIN-12/Notch signaling pathway have severe defects in the structure of the pm8/valve boundary , although the Notch pathway appears to be activated only in pm8 [22] . At least part of the role of the Notch pathway is to prevent pm8 from fusing with the adjacent v1 cell: In normal development , the toroidal or donut shapes of pm8 and v1 require that each cell undergoes autofusion while avoiding cross-fusion . Notch signaling in pm8 accomplishes this by activating expression of the pm8-specific fusogen , AFF-1 , while repressing the expression of the v1-specific fusogen , EFF-1 . LAG-1 is the transcriptional effector of the LIN-12/Notch pathway , and in lag-1 mutants pm8 can express EFF-1 and cross-fuse with v1 [22] . If preventing the expression of EFF-1 were the sole function of Notch signaling , lag-1; eff-1 double mutants might be expected to have a normal boundary , similar to eff-1 single mutants . Instead , we found that most lag-1; eff-1 larvae had boundary defects , with posterior mispositioning of one or more pharyngeal nuclei between valve or intestinal cells ( n = 20/32 embryos; open arrows in Figure 9A ) . A similar mispositioning of pharyngeal nuclei has been described for mutations affecting INA-1/alpha-integrin [22] , but is never observed in wild-type larvae ( n = 0/100 ) , and only rarely observed in eff-1 ( ok1021 ) mutant larvae ( n = 2/74 ) or eff-1 ( ok1021 ) aff-1 ( tm2214 ) double mutant larvae ( n = 2/20 ) . To better understand the role of Notch signaling in pm8 and the valve , we made live recordings of lag-1 embryogenesis using membrane and nuclear reporters for the pm8 family . We found that the timing and pattern of pm8 morphogenesis appeared essentially identical to that of wild-type embryos , with pm8 wrapping around the midline ( Figure 7D , compare with wild-type Video S3 ) . We next used a computational approach to search for target genes that might be regulated by Notch signaling in pm8 . Most , if not all , of the numerous Notch interactions that occur in embryogenesis regulate expression of the ref-1 family of transcriptional repressors [35] , [36] . However , Notch signaling activates expression of the transcription factor CEH-24 specifically in pm8 , suggesting that additional pm8-specific targets might exist [22] . We identified 256 genes that have ( 1 ) predicted LAG-1 binding sites ( RTGGGAA ) in their upstream sequences , and ( 2 ) conservation of the LAG-1 sites in at least 3/5 sequenced Caenorhabditis genomes ( see Materials and Methods; Table S1 ) . Thirteen of these genes have been described in the literature and/or public database annotations as having pharyngeal or pharyngeal muscle expression , and pm8 expression has been noted for five genes in addition to ref-1 and ceh-24 [25] , [37]–[39] . We created reporters for three of these genes ( pax-1 , inx-11 , and inx-20; Figure 9B , 9D , 9E ) and found that mutation of the conserved LAG-1 site to RAGGCAA in each reporter abrogated pm8 expression ( Figure S4A–S4C ) . We further showed that a DNA sequence including the inx-11 LAG-1 site , but not a mutated version , was able to compete for LAG-1 binding in vitro ( Figure S4E ) . We noticed that inx-11 and pax-1 both contained predicted and conserved PHA-4/FoxA binding sites ( TRTTKRY ) [40] near the LAG-1 site , and found that mutations in the PHA-4/FoxA sites also abrogated pm8 expression ( Figure S4A , S4D ) . 60 of the genes with predicted LAG-1 binding sites had one or more predicted and conserved PHA-4/FoxA sites within 110 bp of the LAG-1 site . Of seven of these genes tested , three showed pm8 expression ( tpra-1 , F52D10 . 2 , F52E4 . 5; Figure 9F , 9G and data not shown ) , and for each of two genes tested the expression was Notch dependent ( Figure S4A ) . These results indicate that the presence of combined , conserved LAG-1 and PHA-4/FoxA sites is a predictor of Notch-activated expression in pm8 . Previous studies showed that Notch-regulated gene expression can be detected within 25 minutes after a Notch-expressing cell contacts a ligand-expressing cell [35] . Because Notch signaling is activated shortly after the birth of pm8 and before morphogenesis , we anticipated that Notch targets would be expressed during pm8 morphogenesis [22] . However , most of the Notch targets examined were expressed only late in embryogenesis , long after pm8 completes wrapping and the LIN-12/Notch protein is no longer detectable by immunostaining ( Figure 9D–9G and data not shown ) . Although we do not understand the basis for the delayed onset of Notch target gene expression , we found that the LAG-1 protein persists at high levels in the pm8 nucleus throughout embryogenesis , several hours after LAG-1 is no longer detectable in any other nucleus ( Figure 9H ) . We did not observe pharynx/valve defects in mutants defective in the Notch targets ref-1 , ceh-24 , pax-1 , inx-11 , or inx-20 ( data not shown ) , but observed pharyngeal pumping defects in inx-20 mutants . In wild-type larvae , the coordinated contraction of pharyngeal muscles propels food particles posteriorly into the intestine [41] , and we found that pm8 normally contracts ( opens ) only after pm7 contracts ( Figure 9I , Video S5 ) . In inx-20 larvae , we found that the contractions of pm8 and pm7 were not coordinated; for example , pm8 could open before pm7 , apparently causing the anterior regurgitation of food particles from the valve or intestine into the pharynx ( Figure 9J , Video S5 ) . In summary , Notch has a role in establishing or maintaining the pharynx/valve boundary beyond the regulation of fusogen expression , but does not appear critical for pm8 to complete its basic morphogenetic program of wrapping around the midline . Our identification and analysis of Notch-regulated targets in pm8 suggests that most of these targets do not function in morphogenesis , and instead function in the late , pharyngeal or muscle-specific differentiation of pm8 . In parallel with the above studies on Notch targets , we used a genetic screen to isolate mutants with posterior mispositioning of pharyngeal nuclei , similar to the phenotypes of lag-1; eff-1 double mutants and ina-1/alpha-integrin mutants ( Figure 9A ) . Although most mutants examined had severe and general defects in development , two recessive , embryonic and larval lethal mutants , zu470 and zu471 , had relatively normal morphogenesis of non-pharyngeal tissues , and these were selected for molecular analysis . DNA sequencing showed that zu470 contains a nonsense mutation in ina-1/alpha-integrin , and was not analyzed further in this study . Mapping and complementation showed that zu471 is a new allele of egl-43 ( see Materials and Methods ) . EGL-43 is a zinc-finger transcription factor orthologous to the mammalian proto-oncogene EVI1 , and egl-43 mutants have defects in motor neuron migration and vulval morphogenesis [42]–[44] . DNA sequencing showed that zu471 causes a R489H substitution in the C-terminal zinc finger domain; R489 is conserved in EVI1 proteins from nematodes to mammals , is predicted to contact DNA , and is essential for mammalian EVI1 to bind DNA in vitro ( Figure 10A ) [42] , [45] . Embryos and larvae homozygous for mnDf24 , a deficiency that removes egl-43 [46] , [47] , closely resemble egl-43 ( zu471 ) homozygotes , suggesting that zu471 is a strong loss-of-function allele ( Table 1 ) . Most egl-43 ( zu471 ) mutants appeared to have the correct numbers of several specific cells including marginal cells ( mc2 , mc3 ) , muscles ( pm5–pm8 ) , and some valve cells ( v1 ) ( Table 1 ) . As predicted from the posterior mispositioning of pharyngeal nuclei , we found that the pharynx-valve boundary did not form properly in egl-43 ( zu471 ) animals , with the pharyngeal basement membrane failing to extend between pm8 and v1 ( double-headed arrows in Figure 10B , compare with Figure 5K ) . pm8 and v1 had relatively normal positions in the mutant larvae , and appeared to surround the midline . However , both cells usually appeared perforated , and had variable and abnormal processes extending anterior and/or posterior from the cell body ( 15/16 larvae , Figure 10C ) . egl-43 ( zu471 ) larvae expressing a reporter for both the v2 and v3 valve cells showed cell number and/or positioning defects for these cells ( 25/25 animals ) , with the defect often appearing to result from the mispositioning or absence specifically of the dorsal valve cell , v3D ( 11/25 animals; Figure 10C , 10D ) . In egl-43 ( zu471 ) embryos , pm8 and the valve cells appeared to be in their normal positions at the double plate stage . However , v3D did not spread dorsally at the left-right boundary of intestinal cells , and instead remained in the interior of the cyst ( n = 2 embryos; Figure 3H ) similar to die-1 mutant embryos ( Figure 3G ) . At later stages , when v2L and v2R normally spread dorsally along the int1 surfaces before intercalating anterior to v3D ( Figures 6G and S3C , S3D ) , the v2L and v2R cells in egl-43 mutant embryos continued dorsally along the intestinal cells ( Figure S3F ) . In examining egl-43 mutants with additional reporters for pm8 , we noticed that pm8 lacked or showed variable expression of several genes ( ref-1 , aff-1 , pax-1 , ceh-24; Figures 9C and 10F , Table 1 ) . All of these genes are LIN-12/Notch targets . For example , ref-1 is expressed in the cyst cells e2V , mc3V , and pm8 ( Figure 10E ) , but only pm8 expression is dependent on LIN-12/Notch [22] . In egl-43 mutants , ref-1 was expressed in e2V and mc3V , but not in pm8 ( Figure 10F ) . We found that reporters driven by the lin-12pm8 enhancer were not expressed in egl-43 mutant embryos , suggesting a defect in LIN-12/Notch expression ( Figure 10G , 10H , Table 1 ) . The lin-12pm8 enhancer contains a predicted EVI1 binding site ( TCCGGT ) [48] , [49] that is conserved in several related Caenorhabditis species , however , mutation of this site to CGCTGT did not abrogate reporter expression ( data not shown ) . We next scored LIN-12 expression directly by immunostaining the progeny of heterozygous egl-43 parents , using a balancer marked with a fluorescent reporter ( see Materials and Methods ) . Only 8 . 8% of embryos with LIN-12 expression in the pm8 family were homozygous for the egl-43 mutation , indicating that EGL-43 contributes to LIN-12 expression in these cells ( n = 23/262; p<0 . 01 , chi-squared test ) . We constructed a rescuing egl-43::EGL-43::GFP fosmid reporter ( Figure 10A ) to determine the expression pattern of EGL-43 . In addition to neuronal expression described previously [47] , EGL-43::GFP was expressed at about 150 minutes in four descendants of a blastomere called MS; these descendants are pharyngeal/valve precursors ( Figure 10I , 10M ) . Surprisingly , PHA-4/FoxA was not required for EGL-43::GFP expression in these precursors ( Figure 10J ) , in contrast to most other examples of pharyngeal gene expression [40] , [50] . EGL-43::GFP was not detectable in most embryos lacking either of the transcription factors TBX-35 ( 22/36 negative , Figure 10K ) or CEH-51 ( 32/34 negative , Figure 10L ) , which have partially redundant functions in regulating PHA-4/FoxA [51] , [52] . We found that ceh-51 ( tm2123 ) mutants had apical junction and basement membrane defects in the valve that closely resembled egl-43 mutants , while tbx-35 ( tm1789 ) mutants had junctional defects throughout the pharynx ( Figure 10B and data not shown ) . In wild-type development , the anterior two MS descendants that express EGL-43::GFP do not contribute cells to the posterior pharynx or valve . However , the posterior MS descendants produce pm8 , v1 , and v3D , each of which are defective in egl-43 mutants ( Figure 10M ) [9] . These posterior blastomeres also produce the dorsal mc3 marginal cells ( mc3L and mc3R ) , but not the ventral mc3 cell or any of the mc1 or mc2 marginal cells . We noticed that the dorsal mc3 nuclei in egl-43 mutant embryos typically had less expression of pax-1::GFP than wild-type mc3 nuclei , while the mc1 and mc2 nuclei appeared to have normal levels . For example , a wild-type , dorsal mc3 nucleus has a level of pax-1::GFP that is often higher ( 47% ) and never less than the adjacent mc2 nucleus ( upper #3 and #2 nuclei in Figures 9B and 7F; n = 48 embryos ) . By contrast , the dorsal mc3 nuclei in egl-43 mutants often expressed less ( 43% ) and never more than mc2 nuclei ( upper #3 and #2 nuclei in Figure 9C; n = 38 embryos ) . Thus , egl-43 mutants have pharynx-valve defects that stem from defects in LIN-12/Notch expression in MS descendants , and additional defects in closely related cells such as v3D , mc3L and mc3R that do not undergo Notch signaling .
Our results suggest that the valve cell v3D has an early and special role in forming the valve and aligning the valve with the intestine . We showed that v3D undergoes unique changes in cell shape at the double plate stage , when neighboring pharyngeal and valve cells remain cuboidal . The v3D cell body becomes embedded in the left-right boundary of the int1p cells , whose later dorsal-ventral division produces the int1 ring of intestinal cells . Thus , v3D is pre-positioned at one of the two intestinal cell boundaries that define the crosshair positioning of the intestinal lumen . We showed that proper v3D development requires the transcription factors DIE-1 and EGL-43; in mutants defective in either factor , v3D does not extend along the left-right boundary of the intestinal cells . After the wild-type v3D docks at the left-right boundary , pm8 and v1 move into bilaterally symmetrical positions , flanking v3D . Symmetry with respect to the intestinal boundary per se does not appear essential for pm8 and v1 morphogenesis , as both cells appear to form normally in embryos lacking the intestine . However , symmetry might facilitate intercalation of the ventral valve cells that move dorsally along the surfaces of pm8 and v1 , and that must do so in the short period before the embryo elongates and begins muscle contraction . For example , we showed that mutants lacking HMR-1/E-cadherin have defects in the dorsal intercalation of the v2L and v2R cells , and these mutants subsequently develop gaps or breaks in apical junctions within the valve . Cells throughout the digestive tract must have a common , radial axis of apicobasal polarity to form a continuous lumen; laminin is required to coordinate polarization of the pharyngeal primordium , but not the intestinal primordium [12] . Laminin localizes to the perimeter of the double plate primordium , except at the posterior surface which contacts the intestinal primordium . We showed here that intestinal cells prevent the posterior accumulation of laminin . The intestinal cells might simply block laminin access by adhering to the pharyngeal cells , or influence laminin accumulation through a more specific signaling pathway . Previous studies of laminin mutants showed that expression of laminin only in intestinal cells was sufficient to rescue their defects in apicobasal polarity , but it is not known which intestinal surfaces secrete laminin in those experiments [12] . In die-1 mutants , laminin accumulates inappropriately on the posterior surfaces of ventral double plate cells , including the v2 valve cells ( v2L and v2R ) . We propose that this defect results from the failure of the int2 intestinal cells to intercalate ventrally , between the parents of the v2 cells and the germ cell precursors . At later stages , some of the pharyngeal or valve cells in die-1 mutants develop an axis of apicobasal polarity that is oblique to the radial axes of other pharyngeal or intestinal cells . Together , these results suggest that ectopic , posterior laminin has the potential to cue the inappropriate , anterior localization of apical proteins such as PAR-6 . By preventing laminin accumulation , the intestinal cells ensure that posterior valve cells have the same , radial axis of apicobasal polarity as other cells in the digestive tract . As lamellipodia track circumferentially through the cyst they appear to follow invariant trajectories , intercalating between certain pairs of cells while bypassing others . For example , pm8 and v1 occupy bilaterally symmetrical positions in the cyst and extend symmetrical dorsal , and then ventral , lamellipodia . The ventral lamellipodia intercalate between bilaterally symmetrical pairs of cells ( pm7/int1 then pm7/v2 ) before reaching the ventral marginal cell , mc3V . At or near mc3V , pm8 always crosses anterior to v1 . The posterior face of mc3V is associated with a similarly shaped tract of laminin that forms shortly before , and disappears after , pm8 wrapping , and both laminin and a candidate laminin receptor , INA-1/integrin , are essential to form the pharynx-valve boundary ( this report and [22] ) . Moreover , HMR-1/E-cadherin appears enriched at the boundary between pm8 and mc3V , apparently coincident with the laminin tract . Thus , adhesion to either the tract or mc3V might allow the ventral lamellipodium from pm8 to cross asymmetrically , anterior to the lamellipodium from v1 . Shortly before forming their ventral lamellipodia , pm8 and v1 extend dorsal lamellipodia that meet and stop , without crossing , on the anterior face of v3D . The zone of contact with v3D is highly enriched for HMR-1/E-cadherin , although HMR-1 is localized primarily to the apical surfaces of other cyst cells at this stage . HMR-1 enrichment between v3D and pm8 or v1 does not mediate permanent adhesion , as v2L and v2R subsequently intercalate between v3D and pm8 or v1 . Instead , we speculate that adhesion to v3D might prevent the dorsal lamellipodia of pm8 and v1 from crossing randomly before the proper intercalation order is established by events on the ventral side of the cyst . In addition or alternatively , HMR-1/E-cadherin might participate in adhesive complexes that allow the subsequent intercalation of the v2L and v2R cell bodies . We showed that v2L and v2R can extend thin lamellipodia between v3D and either v1 or pm8 in mutants lacking embryonic expression of HMR-1/E-cadherin , but v2L and v2R fail to shift their nuclei and bulk cytoplasm . We do not yet know whether mutants lacking both embryonic and maternally supplied HMR-1 have more severe valve defects; such embryos have general defects in body morphogenesis that result in the rupture and disintegration of internal tissues , and were not analyzed in this study [33] . We propose that the asymmetry of Notch signaling dictates that the ventral lamellipodium from pm8 must cross anterior to v1 . pm8 and v1 are lineal homologs , which are non-sister cells born from similar sublineages , and many such cells in C . elegans have equivalent development potential . However , only pm8 expresses the LIN-12/Notch receptor , and only pm8 expresses Notch target genes . We found no evidence that Notch signaling regulates the basic morphogenetic events of pm8 intercalation and wrapping , which mirror the Notch-independent morphogenesis of v1 . Instead , Notch signaling likely regulates multiple pharyngeal-specific features of pm8 . Previous results showed that pm8 fails to express a pharyngeal myosin-specific reporter in Notch mutants [22] , and pharyngeal muscle expression has been described for several genes that we identified here as candidate Notch targets . Moreover , we showed that mutant larvae defective in inx-20 , a validated Notch target in pm8 , fail to coordinate pm8 muscle contraction with that of other muscles . Since only pm8 has the potential to become a muscle , it must cross anterior to v1 in order to join other pharyngeal muscles . Notch signaling generates a nuclear localized , ternary complex of proteins that includes CSL proteins such as LAG-1 in C . elegans ( reviewed in [61] ) . Our results suggest that several genes with predicted , conserved binding sites for both LAG-1/CSL and PHA-4/FoxA binding sites are likely to be direct Notch targets in pm8 . The subset of genes that contained both predicted sites but that did not show expression in pm8 might instead be targets in other Notch interactions . For example , PHA-4/FoxA is expressed in the intestine and rectal epithelium during some stages of development , and both of these tissues undergo Notch signaling [15] , [62] . Transcription factors other than PHA-4/FoxA might also collaborate with LAG-1 in driving Notch expression in pm8 . Previous studies identified an enhancer in the ceh-24 gene that can drive Notch-dependent expression in pm8 [22] , [63]; this enhancer lacks an obvious binding site for PHA-4/FoxA , but mutational analysis identified other conserved sites that are critical for pm8 expression [63] . Moreover , EGL-43 provides an example of a transcription factor that is expressed in the pm8 lineage independent of PHA-4 . Thus , different target genes could be regulated combinatorially in pm8 by LAG-1 plus PHA-4/FoxA , or by LAG-1 plus at least one other transcription factor . Notch is activated in pm8 shortly before morphogenesis , and studies in both C . elegans and Drosophila indicate that direct Notch targets can be expressed within minutes after signaling [35] , [64] . However , several direct Notch targets are first detectable in pm8 as late as five hours after signaling , and similar , late expression of some Notch targets has been described recently in Drosophila [64] . We showed that the LAG-1 protein remains in the nucleus of pm8 long after signaling , raising the possibility that LAG-1 might remain at target gene promoters for long periods of time . Previous studies noted a high density of LAG-1 binding sites within the lag-1 gene itself [65] , and we speculate that these sites might contribute to persistent LAG-1 expression We showed that LIN-12/Notch expression depends , at least in part , on the transcription factor EGL-43 . EGL-43::GFP expression in pharyngeal and valve precursors does not require PHA-4/FoxA , but requires two other transcription factors , TBX-35 and CEH-51 . These data indicate that a PHA-4-independent pathway influences pharyngeal cell differentiation through EGL-43 , and one likely candidate is the POP-1/TCF pathway [66] . Sister cells born in anterior-posterior divisions throughout the embryo have unequal levels of nuclear POP-1/TCF , including the pharyngeal cells that express EGL-43 ( Figure 10M ) , their daughters , and at least some later descendants of these cells [66] , [67] . POP-1/TCF is thought to collaborate with a variety of transcription factors in regulating cell fate , and POP-1 function in the pharynx has been shown to regulate pharyngeal muscle specification [68] . We showed that neighboring cells can undergo complex and very different programs of morphogenesis during the development of the valve tube , and this complexity could depend on systems like POP-1 that can generate transcriptional diversity between closely related , or even sister , cells .
See Text S1 . Embryos were mounted as described [9] . Time-lapse movies were acquired with a Hamamatsu C9100 camera on a Nikon TE-2000 inverted microscope equipped with a Yokogawa CSU-10 spinning disk and running Volocity 5 . 3 . 3 ( Improvision ) . The absolute times of the image sequences were scaled relative to either the division of int1p to int1 at 350 minutes or the pm8 nucleus reaching the ventral perimeter of cyst at 430 minutes [9] . For 3D modeling of the double plate to cyst transition , movies were taken of developing embryos and cell contours were generated from Z stacks of each time point using TrakEM2 software [69] . The mesh model was reduced using quadratic edge collapse decimation ( MeshLab open source software; http://meshlab . sourceforge . net/ ) , and viewed with Photoshop 3D animation ( Photoshop CS5 ) . Surface images of cell shapes as in Video S2 were rendered from Z stacks using Volocity 3D Opacity software . The following antibodies/antisera were used: anti-LAG-1 ( gift from Judith Kimble ) , anti-LIN-12 ( gift from Stuart Kim ) , anti-UNC-52 [70] ( gift from Don Moerman ) , MH27 [71] , anti-laminin ( mAbGJ2 ) [22] and anti-GFP ( ab6556 , Abcam ) . Worm and embryo fixation procedures were performed essentially as described [11] . We used a set of custom perl scripts to identify conserved predicted LAG-1/CSL sequences in WormBase release WS230 of the C . elegans , C . brenneri , C . briggsae , C . japonica , and C . remanei genomes [38] , [72] . We limited our analysis to the 15 , 752 C . elegans genes with single orthologs in at least two other Caenorhabditis genomes . We searched 5 kb upstream of these genes and their orthologs for occurrences of the sequence RTGGGAA , which has previously been shown to be bound by LAG-1 [65] . We conducted pairwise alignments of these upstream regions using PhyME [73] to identify conserved sequence blocks ( regions of at least 14 nt with at least 70% identity ) . We considered a predicted LAG-1/CSL sequence conserved if it aligned to blocks containing RTGGGAA sequences from at least two other Caenorhabditis genomes . Multiple sequence alignments were performed using MLAGAN [74] and edited with Jalview ( http://www . jalview . org ) . A 440-nm laser microbeam ( Photonics Instruments ) was used to ablate the E blastomere . Immediately following ablation , embryos were removed from the microscope slide and allowed to develop without compression on an agar filled petri dish . | Tubes composed of epithelial cells are universal building blocks of animal organs , and complex organs typically contain multiple interconnected tubes , such as in the digestive tract or vascular system . The nematode Caenorhabditis elegans provides a simple genetic system to study how tubes form and link . Understanding these events provides insight into basic biology , and can inform engineering strategies for building or repairing cellular tubes . A small tube called the valve connects the two major tubular organs of the nematode digestive tract , the pharynx and intestine . The pharynx and valve form from the same primordium , while the intestine forms from a separate primordium . Cells in each primordium polarize around a central axis , and valve formation involves connecting these axes . Using live imaging , we show that valve cells initially resemble other pharyngeal cells , but undergo additional and extensive intercalations around the lumenal axis , effectively squeezing a small tube from the larger primordium . Valve cells develop the same polarity axis as intestinal cells , and we show that this depends on interactions with the intestinal cells . We show that valve formation involves dynamic changes in the localization of adhesive proteins , and identify transcription factors that play a role in valve cell specification and intercalation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Cell Interactions and Patterned Intercalations Shape and Link Epithelial Tubes in C. elegans |
Plague , a zoonosis caused by Yersinia pestis , is found in Asia and the Americas , but predominantly in Africa , with the island of Madagascar reporting almost one third of human cases worldwide . Plague's occurrence is affected by local climate factors which in turn are influenced by large-scale climate phenomena such as the El Niño Southern Oscillation ( ENSO ) . The effects of ENSO on regional climate are often enhanced or reduced by a second large-scale climate phenomenon , the Indian Ocean Dipole ( IOD ) . It is known that ENSO and the IOD interact as drivers of disease . Yet the impacts of these phenomena in driving plague dynamics via their effect on regional climate , and specifically contributing to the foci of transmission on Madagascar , are unknown . Here we present the first analysis of the effects of ENSO and IOD on plague in Madagascar . We use a forty-eight year monthly time-series of reported human plague cases from 1960 to 2008 . Using wavelet analysis , we show that over the last fifty years there have been complex non-stationary associations between ENSO/IOD and the dynamics of plague in Madagascar . We demonstrate that ENSO and IOD influence temperature in Madagascar and that temperature and plague cycles are associated . The effects on plague appear to be mediated more by temperature , but precipitation also undoubtedly influences plague in Madagascar . Our results confirm a relationship between plague anomalies and an increase in the intensity of ENSO events and precipitation . This work widens the understanding of how climate factors acting over different temporal scales can combine to drive local disease dynamics . Given the association of increasing ENSO strength and plague anomalies in Madagascar it may in future be possible to forecast plague outbreaks in Madagascar . The study gives insight into the complex and changing relationship between climate factors and plague in Madagascar .
Plague is a vector-borne , highly virulent zoonotic disease present today in the Americas , Asia and Africa . It is caused by infection with the bacterium Yersinia pestis , which triggers serious illness with up to seventy percent case fatality in human populations if left untreated . The infection is easily treated with antibiotics , yet these are often difficult to access in time in low income settings . The increasing frequency of the disease in many parts of the world [1] , [2] , [3] has been partly attributed to changes in climate [4] . Presently Africa accounts for more than ninety percent of all human plague cases reported worldwide . Within African countries , the majority of cases are reported from Madagascar and the Democratic Republic of Congo [5] , with Madagascar reporting almost one third of human cases worldwide . Plague is endemic in the highland region of Madagascar and more than one hundred human cases are reported every year , though the true number of cases is likely to be higher . The reasons for such pronounced foci in these areas include extreme poverty and lack of health infrastructure , as well as unique climate features . Yersinia pestis bacteria are transmitted between rodent hosts via their fleas; humans are accidentally infected when in contact with rodent fleas or infected animal tissue . Like many vector-borne diseases , plague's occurrence varies temporally and spatially on a variety of scales . The primary mechanisms directing this heterogeneity are thought to be driven by local variation in factors such as temperature and rainfall . For example temperature influences the rate of development of the bacterium in the flea , and the survival and development times of the fleas themselves , while precipitation affects food availability and thus fecundity of the rodent host [6] , as well as having impacts on habitat characteristics and human-rodent contact ( via rodent responses to flooding ) . Seasonal climate in many parts of the world is affected by the El Niño Southern Oscillation ( ENSO ) , a periodic fluctuation in sea surface temperature and air pressure in the Pacific Ocean which modifies the general flow of the atmosphere in the tropics . Warm ( cold ) phases of the ENSO , called El Niño ( La Niña ) , are associated with a warming ( cooling ) of the tropical troposphere [7] . ENSO impacts on the climate of Madagascar in a manner similar to that observed over southern Africa by Nicholson and Selato [8] . In general , an austral spring-summer El Niño starting from September causes warmer and drier conditions than usual during the austral summer-autumn four to seven months later , and cooler and wetter conditions eight to twelve months later . La Niña has the opposite effect , leading to wetter and cooler , followed by drier and warmer , conditions than average . The Indian Ocean Dipole ( IOD ) [9] is a periodic fluctuation in the relative sea surface temperatures of the western and eastern parts of the Indian Ocean . As the western pole of the IOD is located near to Madagascar , IOD events affect the convection locally and in turn influence the climate of the island . Thus , a positive IOD event is associated with warmer and wetter conditions over the island , while the opposite is true for a negative event [9] . We present here the first analysis of the effects of ENSO and IOD on the temporal distribution of confirmed human plague cases in Madagascar . We use a forty-eight year time-series from 1960–2008 . First we identified periods of anomalous incidence and evaluated their occurrence against any coinciding trends of various climate variables . Second we use wavelets to investigate the strength and direction of association between the incidence of plague and ENSO , IOD , temperature and precipitation . Wavelet analysis allows detection of relationships between two time series in time-frequency space . Our approach uses analytical methods which allow a more thorough analysis than the usual identification of long-term unidirectional trends .
Data on all confirmed human plague cases reported in Madagascar from 1956 to 2008 were made available by the World Health Organisation Plague Reference Laboratory of the Institut Pasteur de Madagascar . A time-series of monthly incidence from 1960 to 2008 was created using the date of onset of symptoms for each confirmed bubonic , pneumonic or septicaemic case , 5-yearly human population growth estimates from the United Nations and the last population census from 1993 [10] . Incidence was calculated for each month using the number of cases , multiplying it by 100 , 000 for scale and dividing it by the relevant time-specific population estimate . Data from before 1960 was available but was omitted due to vaccination campaigns which ceased in 1959 . For effective immunisation , people have to be re-vaccinated yearly . Malaria prevention programs , which occur in some areas of Madagascar and use indoor residual spraying of insecticide , have the potential to impact flea populations . However , this could not be quantified or corrected for . Re-calculated climate variables were downloaded via the climate explorer website [11] based on the National Center for Environmental Prediction ( NCEP ) and the National Center for Atmospheric Research ( NCAR ) reanalysis data [12] or the Centre of Environmental Data Archival website ( precipitation ) [13] . For the El Niño Southern Oscillation variable the index of the Japan Meteorological Agency ( JMA ) was retrieved from the Centre for Ocean-Atmospheric Prediction Studies website [14] . Monthly surface temperature and precipitation anomalies for the geographical area of Madagascar ( 42 . 18°E–49 . 68°E; 24 . 76°S-11 . 42°S ) for the period 1960–2008 were used to describe seasonal cycles and in wavelet analyses . A plague incidence anomaly of >0 . 1 or <−0 . 1 was used to define months of anomalously high/low plague incidence respectively . When months of anomalously high or low plague incidence were sequential or less than 3 months apart , they were considered to be part of the same plague event; otherwise , they were treated as separate plague events . A plague event could therefore be for a single month or for several months . For plague events of >1 month duration , the single month which had the maximum or minimum incidence anomaly ( depending on whether it was a high or low incidence event ) was identified . We tested for three types of association between plague incidence anomaly and the explanatory variables: ( i ) using Analysis of Variance , to test for associations between positive or negative plague anomalies and the value of the explanatory variables ( JMA , IOD , temperature , precipitation ) in the month or peak month of a plague event; ( ii ) using Analysis of Variance , to test for associations between positive or negative plague anomalies and the mean value of the explanatory variable for the year centred on the month or peak month of a plague event; ( iii ) using Fisher Exact Test , to test for associations between positive or negative plague anomalies and positive or negative trends in the explanatory variables for the year centred on the month or peak month of a plague event . For the third analysis , we obtained the sign of the linear regression coefficient for the trend of the 12 monthly values ( with the 7th month as the peak plague anomaly month ) and use this to determine if the explanatory variable was tending to increase ( positive coefficient ) , or decrease ( negative coefficient ) , at the time of a plague event . Time-series analysis using wavelets was undertaken on ENSO/IOD/temperature/precipitation and human plague incidence anomaly datasets spanning the 48-year study period . The objective was to establish any associations in time between plague and the climate variables , and , if present , their direction and periodicity . Wavelet analysis [15] , [16] is a powerful means to identify statistical relationships between signals , and is especially useful when there is non-stationarity; i . e . the periodicity changes with time [15] , [17] . Wavelet analysis is a widely recognised tool to investigate temporal dynamics of infectious diseases [18] , [19] , [20] , [21] , [22] , but has never been used to study the relationship between climate variables and plague in Madagascar . To detect temporal patterns , their variations and coherence , we applied wavelet analysis according to the methods of Grinsted [17] using the software R v . 2 . 15 [23] and Matlab 8 . 0 v . R2012a with Wavelet Toolbox [24] . The following procedure was used . First , the five variables of interest - precipitation , temperature , ENSO , IOD and plague incidence - were tested for normality , and where necessary , normalised using a Johnson transformation [25] . A low-pass Gaussian filter was used to remove the intra-seasonal variability in the time-series . Second , the stationarity or non-stationarity of each variable was determined using continuous wavelet decomposition . Each time-series was decomposed and the continuous wavelet transform plot was examined to confirm the presence of high significant variance and to establish its periodicity and any changes within the time-series . Third , the strength of any relationship between certain pairs of variables was investigated using cross-wavelet analysis , which identifies high common power between two signals ( time frames where both signals vary together ) . The direction of the vectors reveals information about the phase relationship ( i . e . time-lags ) between two time-series . A vector pointing to the right indicates the time-series cycle in-phase and a vector to the left indicates cycling in anti-phase . Thus , any red/yellow areas within figures show periods during which the signals cycle with high common power with the vectors signifying the phase and the time lag . Lastly , the presence and direction of any relationship ( positive/negative ) between variables was established using wavelet coherency analysis , with vectors again indicating the direction of association and time lag . Here , a vector pointing to the right indicates positive association and a vector to the left indicates negative association . Downward or upward pointing vectors reveal information about which time-series leads . Wavelet coherence can be understood as an association between signals in the power spectrum space . It identifies regions of the power-space spectrum where vectors point in one direction . For both cross-wavelet and wavelet coherence analysis , information on time-lags between the time-series is revealed by the direction of the vectors . If two signals cycle with significant common power in cross-wavelet analysis or show association in wavelet coherence analysis with a 2 year periodicity , a right-pointing vector means they are cycling in phase; a left pointing vector means they are cycling in anti-phase ( one lags the other by half the periodicity , i . e . 1 year ) ; a downward pointing vector means the first signal leads the second signal by one quarter of the periodicity ( half the difference between in-phase and anti-phase , i . e . 6 months; and so forth ) while an upward pointing vector means the second signal leads the first signal by one quarter periodicity . A cone of influence ( COI ) was applied to the cross-wavelet and wavelet coherence transforms to mark the limits of the time scale within which signal behaviour can be discussed with confidence . The statistical significance level of the wavelet coherence is estimated using Monte Carlo methods [16] . Further information about these methods is provided by Grinsted et al . [17] .
Plague cases occur year-round but there is strong seasonality , with most cases occurring from September to March ( the austral summer ) and reaching a peak from November to January ( Figure 1A ) . The largest inter-annual variability in plague cases occurs from October to December . The seasonal cycles of temperature and rainfall show highest values from November to April , and December to March respectively ( Figure 1B and 1C ) . December to March are also the months with the largest inter-annual range in rainfall . The inter-annual range in temperature values is approximately equal in all months . Thus , the warm , wet season is the time of the highest plague incidence and greatest variation in incidence .
It is widely accepted that ENSO and IOD are dominant drivers of earth's year-to-year climate variability , and have wide-ranging implications for public health , including influencing the periodicity of many infectious diseases [26] , [27] , [28] , [29] , [30] , [31] . This study is the first , however , to demonstrate an influence of global climate drivers on plague incidence in Madagascar . The results demonstrate a strong association between ENSO and plague in Madagascar , a country where about one third of the world's human cases occur and there is still significant mortality from the disease . For much of the time series ENSO cycles on a 2–5 year time scale ( later 2–7 years ) . This appears to lead to a similar periodicity ( also changing with time ) in the occurrence of plague . However , the association is non-stationary: early in the time series , plague leads ENSO by a few months; later they are anti-phase; and at the end of the time series they are in phase . An analysis of the climate conditions around the time of high and low plague anomalies found evidence for larger plague outbreaks being associated with increasing ENSO signal ( i . e . increasing El Niño conditions ) . Our analyses find some association , also non-stationary , between IOD and plague . This association is less strong , however , than the association of ENSO and plague . ENSO and IOD are unlikely to directly affect plague in Madagascar . Instead they influence local climate which in turn affects the disease . In this study we demonstrate that ENSO and IOD influence temperature in Madagascar and that temperature cycles are associated with plague cycles . Furthermore , the pattern of association between ENSO and temperature in terms of changes in lag and periodicity is similar to the pattern of association between the cycles of plague and temperature . We did not find a clear effect of ENSO or IOD on precipitation; and we found only limited evidence for association between cycles of precipitation and plague . However , while ENSO and IOD may have limited influence on precipitation , we nevertheless found that larger than usual plague outbreaks were significantly associated with rainier than usual years . Hence , the effects of ENSO and IOD on plague appear to be mediated more by temperature , but precipitation also undoubtedly influences plague in Madagascar . During the late 1990s , ENSO events became progressively stronger , increasing in both frequency and amplitude . From 1976 , a shift towards warmer and wetter conditions in the tropical Pacific was detected , with widespread climatic and ecological consequences [32] . Thus , changes in the association between ENSO , sea surface temperature and several diseases may be explained by their modulation by the decadal background and changes in variability . The analysis of ENSO and its association with plague incidence revealed a relationship between plague anomalies and an increase in the intensity of ENSO events and precipitation . In our study the correlation between ENSO and plague turned from weakly positive to strongly negative and finally to positive again while the association between incidence and the IOD changed from negative to positive and became stronger with time . This non-stationarity is likely caused by changes to the lags in the response of Madagascar's climate to ENSO/IOD , or changes to the lags in the response of plague to local climatic conditions . The epidemiology of plague is sufficiently complex that the lag between climate and plague incidence may be long; as , for example , rodent and flea populations respond to favourable conditions . Thus , cycling of ENSO and plague in phase during the later part of the time-series does not necessarily imply that plague was responding immediately to ENSO . Instead , it could be that its response is an entire cycle behind ( 1–2 years ) . Equally , at first sight it seems hard to reconcile that ENSO affects plague , and yet in the early part of the time series plague leads ( comes before ) ENSO . It is possible , however , that plague is actually most of a cycle behind ENSO . The plague system is highly seasonal in the plague focus area which suggests that the dry and cold months from May to September are not favourable for transmission . However , there is a changing relationship between ENSO and its effects on temperature and in turn on plague . Until the 1980s ENSO affects temperature positively on the island 12 to 18 months later , while temperature is associated negatively with plague . From then onwards until the end of the time-series , the temperature response to ENSO accelerates to 8–9 months . This shift in response is most likely related to changes in the frequency and magnitude of ENSO events and a warmer decadal background as shown by the overall increase in temperature . At times ENSO and IOD may interact in their effects on plague . During the first phase of positive association between ENSO and incidence from 1960 to 1975 , a period dominated by La Niña events , notable plague outbreaks are related to El Niño causing warmer than usual temperatures during a time of cooler and drier conditions usually brought by La Niña . In the second phase during the 1980s , plague is negatively associated with ENSO during a time when the temperature response to ENSO accelerates , changing the timing of an increase in temperature . Finally , from 1995 onwards , the intensity and magnitude of ENSO events increases drastically and plague shows positive correlation with ENSO again . At the same time the IOD starts impacting on plague probably via its effects on temperature 6 months later and almost immediate effects on rainfall . Together , positive ENSO and IOD events are creating warmer and wetter conditions . This is exemplified by 1997 which saw one of the largest positive plague incidence anomalies , and was the year with the strongest El Niño and positive IOD in the time-series . The intensity of ENSO and IOD events and corresponding increases in temperature , as well as increases in rainfall , show strong associations with an increase in plague incidence in Madagascar . The seasonality of plague confirms the strong link between disease transmission and optimal environmental conditions , via their effects on vector and host . Higher temperature and increased precipitation during the cold and dry season in Madagascar are likely to increase flea survival and shorten flea development time [33] , [34] , [35] , [36] . An ENSO event influencing the temperature around this time of year would therefore create more favourable conditions for plague transmission . Changes in the climate–disease relationship over time , as detected here , have been found for certain other diseases like cholera [37] , most likely modulated by long-term climate change effects and their influence on epidemiological systems . Plague incidence has also been linked to global climate in other parts of the world , and several studies have hypothesized about the potential mechanisms . Temperature thresholds exist for pathogen survival and transmission , and temperature and precipitation affect the environment of both vector and rodent host . In China , an increased rate of human plague at the province level was associated with the Southern Oscillation Index and Sea Surface Temperature of the tropic Pacific east of the equator [38] . Similarly , inter-annual variability of disease incidence in the Americas correlates with global climate phenomena , with the link hypothesised to be caused by effects on the rodent host community [39] , [40] , [41] . In the US the association between global climate and plague was shown to depend on time-lagged precipitation events , presumably increasing rodent populations via food availability , and on relatively cool summer temperatures during the plague transmission season , potentially increasing the abundance of infectious fleas [21] . Therefore a coherency of ENSO and plague incidence is most likely due to the influence of ENSO on temperature which in turn affects host and vector ecology and transmission potential [2] , [41] , [42] . There are of course other drivers for plague , which have not been considered here . These might also be related to climate effects and act on plague indirectly through anthropogenic and socioeconomic factors such as poverty , migration and cultural practises , all of which can influence disease transmission risks [43] , [44] , [45] . Global climate influences an array of factors which affect the epidemiology of plague , many of which will depend on both ecological and anthropogenic characteristics . The implications of a link between human plague and ENSO and IOD in Madagascar are exceedingly complex but this study leads the way to understanding the relationship between large scale climate and plague in a country where one third of the world's cases occur . | Plague is a vector-borne bacterial infection with rodents and their fleas as its principal hosts . Transmission to humans occurs via the bite of an infected flea . In the highlands of Madagascar , plague is endemic and more than one hundred human cases are reported every year . Global climate is known to affect many infectious diseases and has been shown to affect plague incidence in other areas of the world . The ENSO and the IOD are global climate drivers affecting rainfall and temperature in Madagascar . Our study investigates the effect of global climate drivers on human plague incidence on the island . We found a link between ENSO , IOD , temperature and precipitation and plague incidence throughout the 48-year time-series although it was not constant over time . The correlation between ENSO and plague turned from weakly positive to strongly negative and then to positive , and the association with the IOD became stronger with time . We demonstrate that during periods of high ENSO intensity , plague incidence is likely to increase via ENSO's impact on temperature and precipitation . This shows that climate indices can be a tool to help predict human plague incidence . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"tropical",
"diseases",
"microbiology",
"bacterial",
"diseases",
"plant",
"science",
"yersinia",
"global",
"health",
"plant",
"pathology",
"neglected",
"tropical",
"diseases",
"bacterial",
"pathogens",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"yersinia",
"pestis",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"earth",
"sciences",
"biology",
"and",
"life",
"sciences"
] | 2014 | A Non-Stationary Relationship between Global Climate Phenomena and Human Plague Incidence in Madagascar |
Functional MRI ( fMRI ) studies have traditionally relied on intersubject normalization based on global brain morphology , which cannot establish proper functional correspondence between subjects due to substantial intersubject variability in functional organization . Here , we reliably identified a set of discrete , homologous functional regions in individuals to improve intersubject alignment of fMRI data . These functional regions demonstrated marked intersubject variability in size , position , and connectivity . We found that previously reported intersubject variability in functional connectivity maps could be partially explained by variability in size and position of the functional regions . Importantly , individual differences in network topography are associated with individual differences in task-evoked activations , suggesting that these individually specified regions may serve as the “localizer” to improve the alignment of task-fMRI data . We demonstrated that aligning task-fMRI data using the regions derived from resting state fMRI may lead to increased statistical power of task-fMRI analyses . In addition , resting state functional connectivity among these homologous regions is able to capture the idiosyncrasies of subjects and better predict fluid intelligence ( gF ) than connectivity measures derived from group-level brain atlases . Critically , we showed that not only the connectivity but also the size and position of functional regions are related to human behavior . Collectively , these findings suggest that identifying homologous functional regions across individuals can benefit a wide range of studies in the investigation of connectivity , task activation , and brain-behavior associations .
In functional MRI ( fMRI ) studies , comparing functional characteristics between subjects or groups requires aligning the individual’s data to an “average brain” based on global brain morphology [1] . It is becoming increasingly recognized that interindividual variability exists not only in macroscopic and microscopic brain anatomy [2–4] but also in the organization of functional systems; i . e . , the size , shape , position , and connectivity profile of the functional regions may vary drastically across individuals [5 , 6] . Standard procedures for cross-subject alignment according to macroscopic anatomy may not establish the proper functional correspondence between subjects and can obscure biologically important signals both at the subject level and the group level , especially in the heteromodal association networks that are not strongly tied to anatomical structures [7 , 8] . For example , one of the best-studied function-anatomy dissociations is in the language network , which may be dominated either by the left hemisphere or by the right hemisphere in different subjects [9–11] . More generally , the high level of intersubject variability in the association functions may be a fundamental principle of brain organization and a critical outcome of human brain evolution [5 6 , 12] . Recognizing the significance of intersubject variability in functional organization [13 , 14] , the field of neuroimaging has been making rapid progress towards mapping functional regions at the level of individual subjects [15–20] , especially using connectivity measured by resting state fMRI . For example , Hacker and colleagues proposed an artificial neural network to localize the motor cortex in individual subjects [21]; integrating information derived from functional and anatomical imaging , Glasser and colleagues proposed a multimodal approach to parcellate the cerebral cortex into hundreds of areas [22] . We recently developed an iterative parcellation procedure to map the individual subject’s cortical functional networks and demonstrated that the results were comparable to the current gold standard , invasive cortical stimulation mapping , in patients undergoing brain surgery [17] . Focusing on subject-level analyses , Gordon and colleagues [19] and Braga and colleagues [23] carefully examined a few subjects who were densely sampled and discovered important features of brain networks that were missed in group-based templates but are evident within the individuals . These technical advances in subject-level functional mapping will not only facilitate the investigation of within-subject functional dynamics [24] that are necessary for personalized medicine [25–27] but will also benefit traditional group-level functional studies by providing more meaningful landmarks for between-subject comparison . Specifically , aligning subjects based on homologous functional regions is expected to improve the specificity of functional signals in the networks being studied , and will lead to increased statistical power in group-level analyses [22 , 28] . Here , we tested these hypotheses using large-scale resting state and task-fMRI data provided by the Human Connectome Project ( HCP ) [29–31] , and systematically examined how the individually specified functional regions may benefit group-level studies of functional connectivity and task-evoked activations , and in turn facilitate the discovery of brain-behavior associations .
Using a subject-specific , iterative functional network parcellation strategy that was guided by a group-level functional atlas ( Yeo’s atlas ) [17 , 32] , we mapped 18 cortical networks for each of the 677 subjects provided by the HCP ( selected from the HCP S900 release; see Materials and methods for subject inclusion criteria ) . The resulting cortical networks were then compared with the initial group-level atlas , which consisted of 116 discrete regions of interest ( ROIs ) across the 18 cortical networks . Using a template matching approach , we localized the homologous ROIs in each individual subject’s cortical networks ( see Materials and methods and S1 Fig ) . A majority of the 116 ROIs could be identified in each individual . Across the 677 subjects , 92 homologous ROIs were found in all subjects and these ROIs covered 85 . 6% ± 1 . 5% of the cortical area ( Fig 1A ) . A few small ROIs were not detected in all individuals , either due to technical limitations or because some functional regions may be truly absent in some subjects [15] . Functional ROIs extracted from the individuals’ networks also demonstrated good reproducibility across different days . Using the data of two scan sessions from the same individual , test-retest reliability was quantified for each ROI . The mean Dice’s coefficient across all ROIs was 69 . 8% ± 12 . 9% ( Fig 1B ) . Test-retest reliability was relatively low for the ROIs in the basal frontal and temporal regions , which are more prone to MRI susceptibility artifacts . Importantly , the individually specified ROIs exhibited substantial intersubject variability in size , shape , and position ( Fig 1C ) . These homologous ROIs established a one-to-one correspondence between subjects and thus could be used as the basis for between-subject comparisons in functional studies . Previous studies have repeatedly demonstrated that functional connectivity is highly variable across individuals , especially in the higher-order association areas [5 , 33] , and that the connectivity variability could be related to individual differences in cognitive ability or behavior . However , intersubject variability in connectivity was estimated after aligning data based on brain anatomy; thus , it was influenced both by variability in network topography and variability in connectivity strength among functional regions . Localizing homologous regions in individual subjects allows one to directly evaluate intersubject variability in the size , position , and connectivity strength of the functional regions and investigate their potential relationship to behavior , respectively . We first quantified intersubject variability in vertex-based functional connectivity maps across the 677 subjects using the approach established in previous studies [5] and successfully replicated the earlier findings ( Fig 2A ) . Intersubject variability in size and position of the 116 ROIs , as well as connectivity among the individually specified ROIs , was then evaluated ( Fig 2B–2D , see Materials and methods ) . We found that intersubject variability in vertex-based functional connectivity maps was not only associated with the variability in connectivity strength among the ROIs ( r = 0 . 55 , p < 0 . 001 , S2 Fig ) but was also associated with intersubject variability in position and size of the ROIs ( r = 0 . 49 , p < 0 . 001 and r = 0 . 26 , p < 0 . 005 , respectively ) . These results indicated that previous findings of individual differences in connectivity were likely influenced by variability in network topography . Intriguingly , when variability in functional anatomy was controlled , functional ROIs in the visual and auditory cortices demonstrated strong intersubject variability in their connectivity strength with other ROIs ( Fig 2D ) , although , traditionally , these primary functions were considered less variable . Thus , while visual and auditory ROIs showed relatively low intersubject variability in position ( Fig 2B and 2C ) , which is consistent with previous knowledge that visual and auditory functions are strongly tied to anatomical structures , their connectivity profiles may be significantly more variable across individuals than previously thought ( Fig 2D ) . Taken together , our findings indicated that the size , position , and connectivity of functional regions are dissociable and may be related to different aspects of individual variability in brain functions . Spontaneous brain activity and task-evoked activity are bound by the same anatomical connectivity infrastructure; however , the exact relation between resting state connectivity and task-evoked activations remains unclear . Here , we examined whether individual differences in the cortical functional anatomy are related to individual differences in task activation patterns . Intersubject variability in cortical functional anatomy was computed according to the Dice’s overlap between two subjects’ functional ROI distributions ( i . e . , 1—Dice’s coefficient , “rest distance” ) . Task-evoked activity was first estimated on the cortical surface of each individual ( i . e . , beta values derived from the general linear model ) ; intersubject variability was then measured as the “distance” between two subjects’ cortical activation patterns ( i . e . , 1—spatial correlation between the two task-fMRI maps , “task distance” ) . We found that intersubject variability in cortical functional anatomy was significantly associated with intersubject variability in task-evoked activations ( Fig 3 , r = 0 . 28 , p < 0 . 001 for seven tasks combined and p < 0 . 05 for all individual tasks except the motor task ) . In other words , subjects who showed similar cortical distribution of the functional regions tended to show similar task activation maps . These findings provided a theoretical ground for using the ROIs derived from resting state functional connectivity for task-fMRI studies . We examined whether task-evoked activity could be better aligned across individuals using the individually specified functional ROIs than using the atlas-based ROIs . Task activation maps were first estimated on the cortical surface of each individual using the general linear model , and then activation values ( beta values ) were averaged within each ROI . A subject’s whole-brain activation pattern was thus represented by the activation values in these ROIs . Similarity between two subjects was estimated by correlating their activation patterns ( Fig 4A ) . We found that task-evoked activation patterns were significantly more similar between two subjects when activations were estimated using our individually specified ROIs , as opposed to using the ROIs from group-level atlases including Yeo’s atlas [32] and Glasser’s atlas [22] . This remained true when the data were processed using MSMAll , an improved cross-subject registration approach based on a multimodal surface matching ( MSM ) algorithm and some useful imaging modalities released by the HCP [28 , 34] ( p < 0 . 001 for 5 of the 6 tasks in the HCP data , block bootstrap test accounted for the family structure , 1 , 000 iterations . See Fig 4B for the Language task and Working Memory task as examples . See S3 Fig for the results of the other tasks ) . Improved cross-subject alignment may result in an increased statistical power in group-level task-fMRI analyses . To test this hypothesis , a group-level one-sample t test was performed using the individuals’ mean activation values ( beta values ) within each ROI . The task-relevant regions were first identified using the full sample of 677 subjects . Group-level statistical analyses were then carried out using subsets ( 20 , 30 , 40 , 50 subjects ) of the cohort , and the regions showing significant activations were plotted to the brain surface ( see Fig 4C for the results of the Language and Working Memory tasks as examples ) . In these subsets of subjects , activations were more significant when the analyses were performed using the individually specified ROIs than the atlas-based ROIs . Task-activated regions were also mapped using a series of significance thresholds and compared with the maps derived from the full dataset of 677 subjects . In these subsets of subjects , a higher percentage of task-relevant regions could be detected when the group-level statistical analyses were carried out using the individually specified ROIs than the atlas-based ROIs , across different selections of significance thresholds ( Fig 4D , see S3 Fig for results of other tasks ) . Finally , we found MSMAll processing could improve cross-subject alignment but did not outperform our approach based on individually specified ROIs ( Fig 4D ) . Functional regions identified in individuals may capture the idiosyncrasies of subjects and lead to the discovery of meaningful imaging biomarkers for cognitive functions and behavior . Here , we explored the possibility of predicting individual subjects’ levels of fluid intelligence ( gF ) based on connectivity among the individually specified ROIs . A support vector regression ( SVR ) algorithm combining the leave-one-family-out cross validation ( LOFOCV ) was employed for the prediction ( see Materials and methods ) . A variety of potential confounds , including sex , age , age2 , sex*age , sex*age2 , head size , overall head motion , and acquisition date , were regressed from both the imaging measures and the gF scores before the prediction . The prediction analysis was first carried out using functional connectivity values among the individually specified ROIs derived from our method . The predicted gF scores showed a significant correlation with the observed gF ( Fig 5A , r = 0 . 303 , p < 0 . 001 , permutation test accounted for the family structure , 1 , 000 permutations ) . Connections that were most predictive of gF involved ROIs in the frontoparietal network ( FPN ) , salience network ( SAL ) , default mode network ( DMN ) , and motor-sensory network ( MOT ) ( Fig 5B ) . Specifically , higher gF appeared to be associated with stronger connectivity strength between FPN and some other networks , including the DMN , SAL , and MOT . In contrast , the correlation between the predicted and observed gF was reduced ( p = 0 . 002 , z = 2 . 849 , Steiger’s z test ) when the model was trained using connectivity among the 92 corresponding ROIs defined in Yeo’s atlas [32] ( Fig 5C , r = 0 . 207 , p = 0 . 028 , permutation test accounted for the family structure , 1 , 000 permutations ) . More importantly , the predictive connections identified using these two approaches were largely different ( Fig 5B and Fig 5D ) and only showed small overlap ( Dice’s coefficient = 0 . 25 , see S4A Fig ) . To better understand why atlas-based connectivity became less predictive of gF compared with individually specified connections , we directly examined the correlation between gF and connectivity values among ROIs . Correlations between gF and these predictive connections derived from individual ROIs were significantly stronger ( p < 0 . 001 , t = 4 . 49 , paired t test , see S4B Fig ) than correlations between gF and the same connections defined using group-level ROIs . This result indicated that brain-behavior correlation was already obscured by the group-level atlas before the prediction model was applied , thus impairing the prediction of gF . Repeated analysis using another group-level atlas provided by Glasser and colleagues [22] , which consisted of 360 ROIs , yielded similar results ( correlation between predicted and observed gF: r = 0 . 215 , p = 0 . 041 , permutation test accounted for the family structure , 1 , 000 permutations , see S5 Fig ) . Our analysis above indicated that intersubject variability in the size , position , and connectivity of the functional regions can be dissociated using the individually specified ROIs ( Fig 2 ) . Here , we examined whether the topographic features ( size and position ) of the functional regions are behaviorally relevant . We found that the size and position of the individually specified ROIs could also predict gF scores ( Fig 6 , r = 0 . 266 , p < 0 . 001 for size; r = 0 . 274 , p < 0 . 001 for position; r = 0 . 298 , p < 0 . 001 for size and position combined; permutation test accounted for the family structure , 1 , 000 permutations ) . Specifically , we observed a mild negative correlation ( r = −0 . 125 , p = 0 . 001 ) between gF and size of the DMN regions ( network 15 and 16 , see S7 Fig for the network labels ) but a positive correlation between gF and the size of the motor-sensory regions ( r = 0 . 095 , p = 0 . 013 ) . Additionally , higher gF may be related to a more anterior position of a functional region ( network 17 ) in the inferior frontal gyrus , which is likely related to language ( r = 0 . 099 , p = 0 . 010 ) . Using functional connectivity and topography features together , the correlation between the predicted and observed gF increased to r = 0 . 347 ( p < 0 . 001 , permutation test accounted for the family structure , 1 , 000 permutations ) . These results indicated that the size , position , and connectivity of the functional ROIs provide nonredundant information for the prediction of behavior . For comparison , we repeated the prediction analysis using ROIs defined in group-level atlases and data aligned by MSMAll [34] . After this multimodal alignment , functional connectivity strength among the atlas-based ROIs could better predict gF compared with connectivity derived from the unaligned data ( r = 0 . 255 for Yeo’s atlas with MSMAll versus r = 0 . 207 for Yeo’s atlas without MSMAll; r = 0 . 300 for Glasser’s atlas with MSMAll versus r = 0 . 215 for Glasser’s atlas without MSMAll; see S5A Fig ) . Specifically , Glasser’s atlas combined with data aligned using MSMAll could predict gF with an accuracy comparable to that based on connectivity strength among the individualized ROIs . However , because the size and position of an individual subject’s brain regions cannot be specifically examined after MSMAll alignment , this atlas-based strategy ( r = 0 . 300 ) did not outperform our individualized approach ( r = 0 . 347 ) , which takes advantage of individual differences in both connectivity strength and network topography . Interestingly , we also found a negative correlation between gF and head motion ( r = −0 . 122 , p = 0 . 001 in the 677 subjects , see S6 Fig for analyses on the effect of head motion ) , consistent with previous reports [35] . The correlation increased when we included the subjects with greater head motion . In the 815 subjects who had completed resting state fMRI runs , the correlation between gF and head motion was r = −0 . 176 ( p = 4 . 4 × 10−7 ) . To investigate whether the prediction of gF was influenced by head motion , we calculated the partial correlation between predicted and observed gF , while controlling for head motion . We found that controlling for head motion had little effect on the correlations , indicating the correlations between observed and predicted behaviors were not driven by motion ( S5B Fig ) . Finally , we repeated the prediction analysis based on the individual-specified features , using 10-fold cross validation , and found that our conclusions remained unchanged ( correlation between predicted and observed gF was as follows: r = 0 . 295 for connectivity , r = 0 . 249 for ROI size , r = 0 . 270 for ROI position , see Materials and methods ) . When inspecting the functional connections that were predictive of gF ( as shown in Fig 5B ) , we found that the majority of them were connections between different functional networks rather than connections within the same network . In these predictive connections , many between-network connections appeared to be positively correlated with gF ( red lines in Fig 5B ) , although some connections showed a negative correlation ( blue lines in Fig 5B ) . To examine how between-network connectivity is related to gF , we averaged the connectivity values of all predictive between-network connections and found that mean between-network connectivity showed a mild positive correlation with gF ( r = 0 . 131 , p = 6 . 09 × 10−4 ) , indicating that subjects with higher gF tend to have stronger between-network connectivity , especially between the FPN and several networks , including the DMN , SAL , and MOT . To understand why between-network connectivity derived from individually specified ROIs could better predict gF than that derived from atlas-based ROIs , we investigated how between-network connectivity was changed by the functional alignment . We found that the strength of between-network connectivity showed an average decrease of 12 . 07% when the ROIs were individually specified compared with group-level ROIs based on the Yeo atlas ( S7 Fig , p < 0 . 001 for 16 of 18 networks , paired t test , Bonferroni correction for 18 comparisons ) . These findings indicated that between-network connectivity values were significantly overestimated in studies that directly applied the group-level atlas to individual subjects; thus , it must be interpreted with caution because the inflated connectivity values are more prone to type I errors [36] . Intriguingly , although the absolute values of between-network connectivity were significantly reduced after the functional alignment , they became more predictive of gF , suggesting that our individualized ROIs improved the specificity and accuracy of functional connectivity estimates .
Standard imaging processing procedures use volume-based [37 , 38] or surface-based [39 , 40] registration to align an individual subject’s functional data to a population-level brain template . These registration methods are based on anatomical features such as brain shape , curvature , sulcal depth , or their derivatives ( e . g . , spectral features of cortical anatomy ) [41] . While they can align the macroanatomy of subjects to some extent , these approaches are not capable of aligning functional regions that are often dissociated from macroscopic anatomical landmarks . Recent progress in resting state functional connectivity research has made it possible to align data across subjects based on resting state networks and shows great promise in improving the estimate of connectivity and task-fMRI activations [28] . The functional alignment procedure aims to control for the “nuisance variance” introduced by the topography of networks; thus , one can accurately measure the connectivity strength or task-evoked activation across a group of subjects ( but see Discussion below ) . In the present study , we showed that data alignment using subject-specific functional regions could significantly improve the group-level estimates of task activations ( Fig 4 ) and functional connectivity , especially for the connections between different networks that tend to be overestimated by traditional methods ( S7 Fig ) . The improved connectivity measures in turn can benefit the discovery of imaging biomarkers for cognitive abilities ( Fig 5 and Fig 6 ) . Identifying functional regions in individual subjects not only improves task-fMRI and connectivity estimates , but also enables the investigation of intersubject variability in functional network topography ( Fig 2 ) . For example , we found that ROIs in association areas are highly variable in terms of their spatial distribution . In contrast , positions of the functional regions in the visual and auditory cortices are less variable across individuals , which is consistent with our previous knowledge that visual and auditory functions are more strongly tied to anatomical structures than association functions . However , after aligning the data based on homologous functional regions , intersubject variability in connectivity strength demonstrated an unexpected distribution ( Fig 2D ) and showed a high degree of intersubject variability in these primary functional areas . This implies that intersubject variability of visual and auditory functions may be mostly reflected in their connectivity strength with other brain regions . Further work is required to investigate how the connectivity strength variability in these areas may relate to individual differences in auditory and visual functions . This unexpected observation may lead to new testable hypotheses about individual differences in auditory and visual processing . An important question in the field of neuroimaging that has yet to be answered is whether resting state functional connectivity could serve as the “functional localizer” for task-fMRI analyses . Some previous studies have used simple fMRI tasks to localize functional ROIs in individual subjects prior to quantitative analyses of functional signals at the population level and have shown great potential in improving statistical power [42–45] . Nevertheless , functional mapping using task-based MRI at the single subject level generally suffers from poor signal-to-noise ratio ( SNR ) , limited test-retest reliability [46–48] , and inconsistency with respect to the current gold standard of functional mapping in individuals , i . e . , invasive electrical cortical stimulation ( ECS ) [49 , 50] . Intrinsic functional connectivity may be an alternative , as it has demonstrated great strengths in individual-level functional mapping; however , understanding the exact relationship between intrinsic connectivity and task-evoked activation remains one of the key questions in brain imaging . At the population level , regions with strong intrinsic functional connectivity at rest tend to co-activate during tasks [51] , indicating that spontaneous and task-evoked activity were bound by common functional configurations . In addition , the network architecture revealed by resting state connectivity is present across a wide variety of task states [52] . In line with these findings , we have previously demonstrated that at the single subject level , the whole-brain functional connectivity network architecture derived from task-fMRI data largely resembles that derived from resting state data [17] . Using a machine learning strategy , Tavor and colleagues recently showed the possibility of predicting individual subjects’ task-evoked activity based on combinations of resting state functional connectivity maps [53] . In the present study , we directly quantified the correlations between individual differences in cortical functional anatomy and individual differences in task-evoked activation patterns . The results indicate that spontaneous and task-evoked activity are tightly related to each other ( Fig 3 ) , supporting the possibility of using resting state connectivity as the functional localizer for task-fMRI analyses . We further showed that task-evoked activations were more robustly detected in the individually specified functional ROIs than in the atlas-based ROIs ( Fig 4C and 4D ) . It was recently hypothesized that fMRI analyses based on the signals averaged within functional parcels might benefit from a “neurobiologically constrained” smoothing , which could improve the SNR and statistical power by avoiding the deleterious effects of spatial smoothing [22] . We found that task activations averaged within the individually specified ROIs were significantly more similar between individuals than activations averaged in the atlas-based ROIs ( Fig 4B ) , thus suggesting that the improved statistical power can not only result from the neurobiologically constrained smoothing within subjects but also from the more accurate alignment of functional regions between subjects . Taken together , these data demonstrate the feasibility and advantages of using connectivity as the functional localizer for task-fMRI studies . Recent evidence suggests that individual differences in human behavior and cognition , such as intelligence quotient , musical skills , and reading ability , may be related to variability in brain connectivity [54–60] . In a previous study , we carried out a meta-analysis and demonstrated that loci predicting individual differences in the behavioral and cognitive domains are predominantly located in the association cortex , including the language , executive control , and attention networks that are known to be wired more differently between individuals than the unimodal regions [5] . This observation implies that associations between functional connectivity estimates and behavioral measures may be underestimated or undetected if functional regions are not tailored to individual subjects . Previous studies mostly quantified functional connectivity within regions of population-based atlases , and then correlated these estimates with the individual subject’s behavioral and cognitive measures [59 , 61 , 62] . Here , we showed that performing analyses based on individually specified functional regions will improve the correspondence of functional connectivity and cognitive as well as behavioral measures , thereby facilitating the discovery of new imaging biomarkers for cognition and behavior . Importantly , we found that between-network connectivity measurement will greatly benefit from the subject-specific ROIs . Although the absolute values of between-network connectivity were significantly reduced after functional alignment , they became more predictive of gF . We observed that individuals with higher gF tend to show stronger between-network connectivity , especially between the FPN and several networks , including DMN , SAL , and MOT . Moreover , accurate qualification of between-network connectivity based on individualized ROIs will have particularly strong implications for clinical research , as recent studies have suggested that changes in between-network connectivity may signify normal brain development [63] as well as pathological changes [64] . The analytical framework developed in this study can be conveniently extended to the investigations of brain-behavior associations in clinical populations [65 , 66] . A particularly important finding of this study is that not only the functional connectivity but also the size and position of the functional regions are related to gF ( Fig 6 ) , which is also known to be substantially heritable [67] . Our observations are in line with two recent studies that both stressed the importance of network topography . Bijsterbosch and colleagues [68] examined how individual differences in topographic features may influence the modelling of brain connectivity and demonstrated that the spatial arrangement of functional regions could predict nonimaging measures of behavior and lifestyle . By comparing the spatial topography of 17 networks across subjects , Kong and colleagues [69] also showed evidence that individual differences in large-scale network topography could predict individual differences in multiple behavioral phenotypes across cognition , personality , and emotion . While our results are consistent with these recent findings , the present study has proposed a framework that allows one to investigate the brain-behavior association for each discrete functional region and specifically examine imaging features that are largely dissociable , including the size , position , and connectivity of each region . The significance of network topography in human brain function and behavior has been indicated in numerous studies ( e . g . , [70] ) but has not been systematically investigated at the whole-brain level until recently . In our previous study , we observed that the sizes of some functional networks demonstrated strong hemispheric lateralization , which was also related to handedness and task-fMRI activation [17] . Here , we not only showed that individual differences in network topography are associated with individual differences in task activation patterns ( Fig 3 ) but also showed that they are behaviorally relevant ( Fig 6 ) . These observations strongly suggest that variance in size and position should not be treated as nuisance variance and simply removed by the alignment procedure . Finally , given that gF is a heritable trait , future research can take advantage of our parcellation approach and specifically investigate whether the size , position , and connectivity strength of functional regions are influenced by different genetic factors . Remarkable progress and exciting discoveries have been made in the field of functional imaging research over the past two decades; however , only few of them have been directly connected to clinical interventions . A critical bottleneck for expanding the clinical use of fMRI is the ability to robustly localize functional circuits relevant to disorders in individual patients . For example , previous work using functional connectivity to identify potential biomarkers of neurological [71–74] and psychiatric [75–77] illnesses has repeatedly found evidence for altered network architecture in patients , as compared with healthy control participants . And yet , such group-level observations have failed to yield any biomarkers that can predict treatment response or provide confirmatory evidence of a patient's current symptoms and diagnosis . To meet clinical demands , a marker must reliably reflect a patient's current or future symptom load in a manner that can be applied to the management of individual patients [78] . Here , we demonstrated that individually specified functional regions can improve the detection of associations between imaging measures and cognitive abilities at the group level ( Fig 5 and Fig 6 ) , implying that this approach may facilitate the identification of neural circuits associated with symptom severity in patients . Critically , this subject-specific strategy may not only help to identify the symptom-related circuits but , at the same time , can map these circuits onto the individual patient’s brain , thus providing personalized targets for intervention . There are several limitations related to the methodology employed in the present study . First , the functional ROIs were derived from a connectivity network parcellation , which uses a “winner-takes-all” approach . However , the resting state of the human brain is not a single static state , but consists of multiple states that dynamically emerge and dissolve . Functional network parcellation based on connectivity should thus be seen as a statistical estimate of the co-activation probability among brain regions , as opposed to a collection of fundamental functional units separated by sharp boundaries . Second , the functional regions identified in individual subjects depend on the validity of the network parcellation . The present study is based on the 18-network parcellation that has been widely used in the literature [32] . However , the optimal number of networks is yet to be investigated and will most likely remain equivocal . Moreover , the 18-network parcellation cannot reveal fine-grained subdivisions of important areas such as the auditory and visual cortices . Taking this crude parcellation as the basis for constructing individual-level parcels will inevitably limit its usage in the investigation of highly specialized functions within some areas . Third , performing group-level analyses using the subject-specific ROIs relies on the identification of homologous regions across subjects . Because the functional localization problem is inherently ambiguous , the procedure of matching homologous functional regions across individuals may introduce error or bias . For example , although the individualized parcellation approach may be able to segment the hemispheres of people with left-lateralized and right-lateralized language dominance differently , in extreme cases when the language area is missing in one hemisphere , it may not improve “true” functional alignment relative to a group atlas because it would not match one subject’s right-lateralized language region to another subject’s left-lateralized one . Fourth , the present study only focused on cortical regions; it did not include functional regions in subcortical structures . The involvement of cortico-subcortical circuits in various cognitive processes and brain disorders is well recognized . Future work on functional network parcellation in individual subjects’ subcortical structures will greatly advance our ability to characterize functional brain architecture . Finally , the reliability of functional ROIs is also dependent on the scan length . Recent studies , including our own [16 , 79] , have provided evidence that sufficient scan length is crucial for the individual-level test-retest reliability of functional connectivity measurements . The relatively low test-retest reliability of the individualized parcellation boundaries in several brain regions may introduce noise relative to group atlases , which are defined with many more data than the data collected for a single subject . Whether functional ROIs derived from short resting state scans can benefit group-level functional analyses must be further explored .
The present study used data made publicly available by the HCP , supported by the WU-Minn Consortium . Written informed consent was obtained from each participant in accordance with relevant guidelines and regulations approved by the local institutional review board at Washington University in St . Louis ( IRB #201204036 ) . The present study used data from the HCP S900 data release , which consisted of 955 young healthy subjects . After quality control , 677 subjects ( 372 female , age range 22–35 years , except for one subject who was over 36 years ) were selected for subsequent analyses . Each participant underwent two fMRI sessions on two different days . Each fMRI session consisted of two 15-minute resting state runs and about 30 minutes of task-fMRI . A battery of behavioral tests was performed by each participant . The present study examined the association between gF and neuroimaging measures . gF was selected because its association with functional connectivity has been reported in previous studies [54 , 59 , 80 , 81] . gF is one’s capacity to solve problems in novel situations . More details about the participants and data acquisition can be found in S1 Text . The “ICA-FIX” denoised fMRI data of the HCP subjects , represented as time series of grayordinates [28] , were used . The data were already preprocessed in the HCP pipeline using FSL ( FMRIB Software Library ) , FreeSurfer , and Connectome Workbench’s command line functions [28 , 31 , 82 , 83] . Each subject's preprocessed resting state fMRI data were resampled to a common standard cortical surface mesh representation ( fs_LR 32k mesh ) . Studies have reported that global physiological noise and motion-related artifacts were not fully removed by ICA-FIX method [28 , 84] . We took the following additional processing procedures for resting state fMRI analysis: ( 1 ) normalizing the resting state fMRI time series at each vertex to zero mean and unit variance; ( 2 ) linear detrending and band-pass filtering ( 0 . 01–0 . 08 Hz ) ; ( 3 ) regressing out 12 head-motion parameters and whole-brain signal; and ( 4 ) smoothing on the 32k fs_LR surface using a Gaussian smoothing kernel ( sigma = 2 . 55 mm ) . Task-fMRI data were already preprocessed and analyzed by the HCP on the fs_LR 32k surface [31 , 85] . Task activation maps with 4-mm Gaussian smoothing were downloaded from the HCP , and we did not perform any additional processing on the task-fMRI data . For the group-level analyses on task activations shown in Fig 4 , we computed the mean beta values in our individually specified ROIs , as well as in the ROIs defined by the atlas . Here , we used beta values ( task effect size ) instead of Z values ( a ratio between beta and unexplained variance ) ; thus , we could estimate the BOLD signal changes induced by tasks within a parcel [22] . The significance levels were estimated for each ROI using a one-sample t test . For comparison purposes , we included resting state and task data that were processed using MSMAll , which is the improved intersubject registration based on a MSM algorithm and features from multiple imaging modalities released by the HCP [28 , 34] . Except where noted , the description of analysis applies to data aligned using the traditional cortical folding-based registration method . A population-level functional atlas including 18 cortical networks was obtained using data from 1 , 000 healthy subjects [17 , 32] . The original atlas consisted of 17 networks and was further divided into 114 discontinuous ROIs . The hand sensorimotor areas were then defined using a hand motor task and separated from other regions [86] , resulting in 116 ROIs in total . Vertices at the boundaries of the ROIs were excluded because of the indefinite network affiliations . These population-level cortical ROIs were used as the functional template , and the homologous ROIs were identified in each individual subject . The procedure to localize functional ROIs in individual subjects consisted of the following steps: The individual subject’s connectivity profile was represented by the connectivity strength among ROIs . The mean signal of an ROI was computed by averaging the preprocessed BOLD signal across all vertices within the ROI . Connectivity between two ROIs was then estimated using Pearson’s correlation and converted into Z values using Fisher’s Z transformation . When ROIs were individually specified , we used the 92 ROIs that were consistently detected in all 677 HCP subjects; as a result , each individual subject’s connectome was represented by a 92 × 92 matrix . For comparison purposes , functional connectivity was also estimated using the corresponding 92 ROIs in the population-level functional atlas [32] ( “Yeo’s atlas” ) and Yeo’s atlas with fMRI data aligned using MSMAll [34] ( “Yeo’s atlas with MSMAll” ) . We applied the ROIs from Yeo’s atlas in the main analysis to ensure the number of ROIs was consistent with our individualized ROIs . In another comparison , 360 ROIs in a fine-grained atlas derived from the multimodal parcellation of the cortex proposed by Glasser and colleagues [22] ( “Glasser’s atlas” ) and the corresponding ROIs with fMRI data aligned using MSMAll ( “Glasser’s atlas with MSMAll” ) [22] were applied . Functional connections were separated into within-network and between-network connections according to whether they connected two ROIs in the same network or different networks ( see S7 Fig ) . Within-network and between-network connectivity values were estimated for each subject . To compute the within-network connectivity of a specific network , we averaged the connectivity values of all ROI pairs within the network . To compute the between-network connectivity of a specific network , we averaged the connectivity values of all ROI pairs that involved an ROI within the network and an ROI outside the network . The size of a functional region was calculated as the number of vertices that fell within that region . For each ROI , intersubject variability in size was calculated as the standard deviation of size across subjects . Intrasubject variability was estimated as the difference in ROI size between two scan sessions . To control for the impact of noise and other technical confounds on intersubject variability estimates , intersubject variability in size was corrected by regressing out the mean intrasubject variability using the similar strategy described in Mueller and colleagues [5] . The position of a functional region was represented by the coordinates of its center of mass . For each ROI , intersubject variability in ROI position was estimated as the average geodesic distance among the ROI centers across subjects . Intrasubject variability in ROI position was estimated as the geodesic distance between the ROI centers localized in two sessions . Intersubject variability in position was also corrected by regressing out the mean intrasubject variability . For a given ROI , intersubject variability in size and position was evaluated using only the subjects in whom the homologous ROIs could be detected . Intersubject variability in vertex-wise and ROI-based functional connectivity was estimated using an approach similar to Mueller and colleagues [5] . For each vertex in the fsLR_32k surface mesh ( 59 , 412 vertices ) , the functional connectivity profile was represented by its connectivity with other vertices on the fsLR_32k surface . The functional connectivity profile for each individually specified ROI on the surface was computed as its connectivity with all other ROIs . Intersubject variability in ROI-based functional connectivity was also corrected by regressing out the mean intrasubject variability . A SVR algorithm ( L2-regularized L2-loss SVR model ) implemented in the LIBLINEAR package ( https://www . csie . ntu . edu . tw/~cjlin/liblinear/ ) was used to predict gF based on functional connectivity . Prediction performance was evaluated using LOFOCV . Family structure was kept intact in the prediction; i . e . , subjects from the same family were not split into the training and testing datasets . During the LOFOCV , model parameters were trained using the data of left-in subjects , and then the trained model was applied to the left-out subjects ( i . e . , one family ) to predict the subjects’ gF scores; the procedure was repeated for each family to predict the gF score of all subjects . The performance of the prediction model was evaluated by the correlation between predicted and observed gF scores . Specifically , each LOFOCV procedure included feature selection , model learning , and testing . Before selecting effective features , covariates including sex , age , age2 , sex*age , sex*age2 , brain size , head motion , and acquisition quarter were regressed from the features and the observed gF scores . The regressing weights were applied to the left-out dataset . To reduce redundant information and prevent possible over-fitting , functional connectivity that showed significant correlations with the gF scores in the training dataset were selected to train the model . We applied different significance thresholds ( p < 0 . 001 and p < 0 . 0005 ) in feature selection , corresponding to large and small numbers of features . The de-confounded features from the testing data were inputted into the trained model to derive the predicted gF scores . To compare prediction performance based on different methods for defining functional ROIs or cross-subject alignment , the maximum correlation between predicted and observed gF was obtained from a different number of selected features . To predict gF based on the size or position of the ROIs , we also used the SVR model described above . The position of a functional region on the cortical surface was evaluated as the coordinates of the region’s center of mass in the right–left axis , anterior–posterior axis , superior–inferior axis ( RAS ) coordinate system . In each LOFOCV , imaging features ( i . e . , size and/or position of ROIs ) in the training data were applied to train the model . To test if topographic features from individually specified ROIs and functional connectivity among them can provide nonredundant information for the prediction of gF , we also trained the prediction models using different features and then averaged the outputs from different models . The above prediction analyses based on connectivity , ROI size , and ROI position were also repeated using 10-fold cross validation . Specifically , we trained the model using 90% of the families and tested the model in the remaining 10% of the families . We ensured that family members were not split between folds . The 10-fold cross validation was repeated 50 times , and the mean prediction accuracy was reported . To ensure that the prediction was not affected by head motion , we also investigated the relation between motion and gF scores , ROI size , and ROI reliability ( See S6 Fig ) . A nonparametric permutation test was performed to determine whether the prediction of gF scores exceeded the chance level . The observed gF values were randomly reshuffled among the subjects ( 1 , 000 permutations ) , and the prediction procedures were repeated . To account for family structure , members in one family were not split during the permutation [87] . The permutation p-value was estimated by calculating the percentage of permutations that yielded a prediction-observation correlation value higher than the prediction-observation correlation based on the real data . Contributions of the functional connections ( connection weight ) were averaged across all LOFOCV folds . If one feature was not selected in one fold , its contribution was set to zero in this fold . The contribution of a given ROI was calculated by summing up the contributions of all connections involving that ROI . If one ROI was not associated with any of the selected features , its contribution to the prediction was set to zero . For the purpose of visualization , all imaging results were visualized using the Connectome Workbench display tool provided by the HCP ( https://www . humanconnectome . org/ ) [83 , 88] . The connectograms in Fig 5B and Fig 5D were created using Circos ( http://circos . ca/ ) . | No two individuals are alike . The size , shape , position , and connectivity patterns of brain functional regions can vary drastically between individuals . While interindividual differences in functional organization are well recognized , to date , standard procedures for functional neuroimaging research still rely on aligning different subjects’ data to a nominal “average” brain based on global brain morphology . We developed an approach to reliably identify homologous functional regions in each individual and demonstrated that aligning data based on these homologous functional regions can significantly improve the study of resting state functional connectivity , task-fMRI activations , and brain-behavior associations . Moreover , we showed that individual differences in size , position , and connectivity of brain functional regions are dissociable , and each can provide nonredundant information in explaining human behavior . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"methods",
"and",
"resources",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"neural",
"networks",
"social",
"sciences",
"problem",
"solving",
"neuroscience",
"magnetic",
"resonance",
"imaging",
"cognitive",
"psychology",
"mathematics",
"brain",
"mapping",
"intelligence",
"discrete",
"mathematics",
"combinatorics",
"neuroimaging",
"research",
"and",
"analysis",
"methods",
"language",
"computer",
"and",
"information",
"sciences",
"imaging",
"techniques",
"behavior",
"psychology",
"radiology",
"and",
"imaging",
"diagnostic",
"medicine",
"permutation",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"cognitive",
"science"
] | 2019 | Performing group-level functional image analyses based on homologous functional regions mapped in individuals |
Buruli ulcer is a chronic painless skin disease caused by Mycobacterium ulcerans . The local nerve damage induced by M . ulcerans invasion is similar to the nerve damage evoked by the injection of mycolactone in a Buruli ulcer mouse model . In order to elucidate the mechanism of this nerve damage , we tested and compared the cytotoxic effect of synthetic mycolactone A/B on cultured Schwann cells , fibroblasts and macrophages . Mycolactone induced much higher cell death and apoptosis in Schwann cell line SW10 than in fibroblast line L929 . These results suggest that mycolactone is a key substance in the production of nerve damage of Buruli ulcer .
Buruli ulcer is a disease characterized by the painless nature of its lesion . The disease is basically characterized by the ulcer without pain [1] , but some pain is noted at the wound care dressing service [2] . These studies suggest that Buruli ulcer lesions are initially painless , but the patients experience pain after chemotherapy , probably due to nerve regeneration . Studies of the pathological mechanism have revealed that local nerves are invaded and damaged by the causative agent , M . ulcerans [3] , and that similar nerve damage is evoked by the injection of mycolactone in a mouse model [4] . In both instances , Schwann cells , which play the major role in maintaining nerve function , showed vacuolar degeneration . Also , nerve damage was histopathologically confirmed in human Buruli ulcer lesions [5] . To further elucidate the mechanism of nerve damage in Buruli ulcer , we tested the cytotoxic effect of mycolactone on a cultured Schwann cell line ( SW10 ) . Because mycolactone evokes cell death and apoptosis in fibroblasts [6 , 7] , macrophages [7] , adipocytes [8] keratinocytes [9] , vascular endothelial cells [10] and skeletal muscle satellite cells [11] , it is necessary to compare the cytopathic pattern produced by mycolactone on Schwann cells to that on other cells . Therefore , mouse fibroblast cell line L929 and macrophage cell line J774 were used for comparison studies . Synthetic mycolactone A/B [12] was used for the evaluation of mycolactone alone . In addition , the cytotoxic effect of synthetic mycolactone A/B remote diastereomer ( stereocenter present outside a self-contained box ) [12] was compared with that of synthetic mycolactone A/B .
L929 mouse fibroblast cells ( ATCC CCL-1 ) were purchased from the American Type Culture Collection and passaged in Dulbecco‘s Eagle's Minimum Essential Medium supplemented with 10% heat-inactivated horse serum at 37°C with 5% CO2 . Mouse macrophage cells J774A . 1 ( ATCC TIB-67 ) , C2C12 mouse myoblast ( ATCC CRL-1772 ) , Neuro-2a mouse neuroblast ( ATCC CCL-131 ) , sNF96 . 2 human Schwann cells ( ATCC CRL-2884 ) were purchased from the American Type Culture Collection and passaged in Dulbecco‘s Modified Eagle’s Medium supplemented with 10% heat-inactivated fetal calf serum at 37°C with 5% CO2 . HUVEC human endothelial cells ( Lonza CC-2519 ) were purchased from Lonza and passaged in Endothelial Cell Growth Medium 2 Kit ( Lonza C-22111 ) at 37°C with 5% CO2 . SW10 mouse Schwann cells ( ATCC CRL-2766 ) were purchased from the American Type Culture Collection and passaged in Dulbecco’s Modified Eagle’s Medium supplemented with 10% heat-inactivated fetal calf serum at 33°C with 5% CO2 . Synthetic mycolactone A/B and synthetic mycolactone A/B remote diastereomer were supplied by one of the coauthors ( Yoshito Kishi ) , as ethanol-diluted solutions ( 1 mg/ml ) . The purity of synthetic mycolactone A/B and mycolactone A/B remote diastereomer was confirmed by 1H- and 13C-nuclear magnetic resonance and also by high performance liquid chromatography . The 0 . 20 mg/ml stock solution was prepared as follows: Firstly , 10 . 30 mg mycolactone A/B ( the weight was determined by a Mettler ultra-micro balance ) was dissolved in 10 . 3 ml ethanol to prepare a 1 . 0 mg/ml solution . Secondly , 2 . 0 ml of the above solution was diluted with 8 . 0 ml ethanol to prepare a 0 . 2 mg/ml solution . 0 . 50 ml each of the solution was transferred to a brown ampoule and sealed under argon . Ampoules containing the 0 . 20 mg/ml stock solution were kept in dark at -20°C . Thirdly , the concentration of the stock solution was further confirmed by the optical density at 362 nm ( UV λmax 362 nm ( log ε 4 . 35 ) ) . The same procedure was used for preparation of the 0 . 20 mg/ml stock solution of mycolactone A/B remote diastereomer; in this series , 10 . 20 mg of synthetic mycolactone A/B remote diastereomer was used . They were diluted using culture medium to 30 μg/ml , 3 μg/ml , 300 ng/ml , 30 ng/ml , 3 ng/ml and 300 pg/ml . Ethanol similarly diluted with culture medium to a final ethanol concentration of 3 μg/ml was used as the negative control . L929 fibroblasts and J774 macrophages were cultured in 24-well plates ( 2 . 0x104 cells/well ) for 24 hrs . Synthetic mycolactone A/B similarly diluted ten-fold from 30 μg/ml to 300 pg/ml was added and morphological evaluations were performed under a phase-contrast microscope ( Nikon TMS ) every 24 hrs for up to 60 hrs . L929 fibroblasts and SW10 Schwann cells were cultured in 24-well plates ( each well containing 2 . 0x104 cells ) for 24 hrs . Synthetic mycolactone A/B with a final concentration of 300 ng/ml , 30 ng/ml , or 3 ng/ml was added to the culture wells and incubated . The counting of dead cells after trypan blue staining and the TUNEL assay were performed at 24 hrs and 48 hrs as described below . SW10 mouse Schwann cells and L929 mouse fibroblast cells were cultured for 24 hrs in the same manner as described above . Synthetic mycolactone A/B diluted to a final concentration of 300 ng/ml , 30 ng/ml , or 3 ng/ml was added to the cells and incubated for 12 , 24 , 48 and 72 hrs . Photomicrographs were taken using the phase-contrast microscope . SW10 mouse Schwann cells , L929 mouse fibroblasts , J774A . 1 mouse macrophage , C2C12 mouse myoblast , Neuro-2a mouse neuroblast , sNF96 . 2 human Schwann cells and HUVEC human endothelial cells were cultured for 24 hrs . Synthetic mycolactone A/B with a final concentration of 300 ng/ml , 30 ng/ml , or 3 ng/ml was then added and the cells were incubated further . Floating and adhered cells were collected at 12 , 24 and 48 hrs time points . Ethanol similarly diluted with culture media to a final ethanol concentration of 300 ng/ml was used as the negative control . In addition , actinomycin-D ( Sigma ) diluted with culture medium for 24 hrs was used as the positive control . Following the bicinchoninic acid assay ( BCA assay ) , Western blot analysis was performed using rabbit anti-cleaved caspase-3 ( Cell Signaling #9661 ) , rabbit anti-caspase-3 antibody ( Cell Signaling #9662 ) , mouse monoclonal anti-histone H2A . XS139ph ( phospho Ser139 ) antibody ( GENETEX , Inc . GT2311 ) , and mouse monoclonal anti-α-tubulin ( Sigma T-9026 ) as an internal control . Horseradish peroxidase ( HRP ) -labeled goat anti-mouse IgG ( 7076 ) and goat anti-rabbit IgG ( 7074 ) , purchased from Cell Signaling , were used as secondary antibodies . Immunoreactive bands were visualized using a chemiluminescence reagent Immuno Star LD ( Wako ) . Fibroblasts and Schwann cells were cultured in chamber slides for 24 hrs . Synthetic mycolactone A/B with a final concentration of 300 ng/ml , 30 ng/ml , or 3 ng/ml was added before further culturing for 12 and 24 hrs . Ethanol similarly diluted with culture media to a final ethanol concentration of 300 ng/ml was used as the negative control . In addition , actinomycin-D diluted with culture medium for 24 hrs was used as the positive control . Following fixation with paraformaldehyde and Triton-X treatment , the cultures were fluorescently stained with the following reagents: cleaved caspase-3 was stained fluorescent red ( Rabbit anti-Cleaved Caspase-3 ( 1:1000 ) /Alexa Fluor 594 Goat Anti-Rabbit IgG ) , nuclear DNA was stained blue ( Hoechst 33342 ) , and intracellular actin was stained green ( Alexa Fluor 488 Phalloidin ) . Cells were examined under a confocal laser scanning microscope ( Olympus: FV10i-DOC Laser Scanning Microscope ) . L929 fibroblasts and SW10 Schwann cells were cultured , and treated with the same concentration of mycolactone A/B and mycolactone A/B remote diastereomer . Trypan blue staining and the TUNEL assay were performed to count dead cells and evaluate apoptosis , respectively . The Mann-Whitney U-test was applied for the comparison of data obtained from the two groups .
Synthetic mycolactone A/B exerted cytotoxicity against the tested cell lines . Both fibroblasts and macrophages exhibited detachment of most adhered cells 24 hrs after the addition of 30 μg/ml and 3 μg/ml of synthetic mycolactone A/B . Treatment with 300 ng/ml and 30 ng/ml of synthetic mycolactone A/B resulted in partial detachment , while no floating cells were found in the 3 ng/ml , 300 pg/ml and negative control cultures . At 60 hrs , the number of floating cells increased , but 3 ng/ml and 300 pg/ml cultures contained no floating cells . Fibroblasts underwent shrinkage and detachment 24 hrs after the addition of 300 ng/ml and 30 ng/ml of mycolactone . However , 3 ng/ml and 300 pg/ml produced no morphological changes . Schwann cells also displayed partial detachment 24 hrs after the addition of 300 ng/ml and 30 ng/ml of mycolactone . However , no morphological changes were detected in the 3 ng/ml and 300 pg/ml cultures . At 48 hrs , floating cells were observed in the culture containing 30 ng/ml of mycolactone . Cytotoxicity levels of mycolactone were compared between Schwann cells and fibroblasts . As shown in Fig 1 , cell death as evaluated by trypan blue staining showed that Schwann cells are more sensitive to mycolactone than fibroblasts at the concentrations of 30 ng/ml ( 24 hrs , p < 0 . 01; 48 hrs , p < 0 . 001; 72 hrs , p < 0 . 001 ) and 300 ng/ml ( 24 hrs , p < 0 . 002; 48 hrs , p < 0 . 001; 72 hrs , p < 0 . 001 ) . Apoptosis induced by mycolactone was evaluated for the Schwann cells and fibroblasts . As shown in Fig 2 , the TUNEL assay also showed that Schwann cells are more sensitive to mycolactone than fibroblasts . As shown in Fig 3 , control cells and cells treated with 3 ng/ml of mycolactone contained no floating cells at 24 , 48 , or 72 hrs . Fibroblasts exhibited no changes until 48 hrs , but partial detachment began at 72 hrs with 30 ng/ml of mycolactone A/B . In contrast , Schwann cells displayed round shrinkage and floating at 24 hrs with 30 ng/ml of mycolactone A/B . Some adherent cells remained at 48 hrs , but all had detached at 72 hrs . As shown in Fig 4A , SW10 Schwann cells showed induction of cleaved caspase 3 and phosphorylated histone H2A . X at 12 and 24 hrs with 30 and 300 ng/ml . At 48 hrs , cleaved caspase 3 became negative , but p-histone H2A . X was expressed . In J774 macrophages , mycolactone slightly induced cleaved caspase 3 , but it did not induce p-histone H2A . X . Mycolactone slightly induced cleaved caspase 3 and p-histone H2A . X at 48 hrs in the fibroblasts . Fig 4B shows comparison of seven cell lines at 24 hrs with 30 ng/ml of mycolactone . Mouse Schwann cells ( SW10 ) and human Schwann cells ( sNF96 . 2 ) showed strong induction of both cleaved caspase 3 and p-histone H2A . X by mycolactone . Mouse neuroblasts also showed induction of p-histone H2A . X . However , fibroblasts ( L929 ) , endothelial cells ( HUVEC ) , myoblasts ( C2C12 ) and macrophages ( J774 ) did not show clear induction . Expression of cleaved caspase-3 was compared at 12 and 24 hrs after the administration of mycolactone . In the four conditions ( 12 and 24 hrs , 30 and 300 ng/ml mycolactone ) , the expression of cleaved caspase 3 was observed in the cytoplasm of SW10 Schwann cells ( 10–21% ) and in some of fibroblasts ( 2–3% ) ( Fig 5 ) . As shown in Fig 6 , synthetic mycolactone A/B and its remote diastereomer exerted identical cytotoxicity to both fibroblasts and Schwann cells at the same concentration .
Patients with Buruli ulcers are thought to have no pain unless secondary bacterial infection occur [1] . However , Alferink et al . [2] showed the presence of severe pain during wound care in about 30% of patients . Addison et al . [13] also reported that 54 . 8% of patients experienced wound pain associated with wound dressing at a primary health care center , where secondary bacterial infection was 9% . These studies suggest that Buruli ulcer lesions are initially painless , but the patients experience pain after chemotherapy , probably due to nerve regeneration . Antibiotic therapy decreases mycolactone concentration in Buruli ulcer lesions [14] . In our previous study with a single injection of mycolactone to mouse footpads [4] , we had already demonstrated thin myelins ( remyelination ) 28 days after the injection and normal myelins ( completion of remyelination ) on day 42 , indicating that mycolactone-induced nerve damages are reversible . These findings lead us to further study cytopathic mechanism of mycolactone to Schwann cells . To determine whether synthetic and purified mycolactone have similar biological activities and the optimal mycolactone concentration for the cytotoxic studies , various amounts of mycolactone were added to L929 fibroblasts and J774 macrophages and cultured for 24 , 48 and 72 hrs . At 24 hrs , no detachment was observed with mycolactone concentrations of 3 ng/ml and 300 pg/ml , but many floating cells were found following treatment with concentrations of 30 μg/ml , 3 μg/ml , 300 ng/ml and 30 ng/ml . These results were compatible with the study of George et al . [4] using mycolactone purified from cultured M . ulcerans . The biological activity of synthetic mycolactone in cultured cells was confirmed in this study . The threshold concentration of cytotoxicity was between 30 ng/ml and 3 ng/ml . Sarfo et al . [14] measured the concentration of mycolactone in 80 untreated human Buruli ulcer lesions by mass spectrometry , and the median ( range ) concentration was 26 ng/ml ( 0–1970 ) , which also supports the adequacy of our present study . Trypan blue vital staining reflects cell membrane damage , which might be caused either by the necrotic pathway or late-stage apoptotic process . The counting of trypan blue-positive cells showed that a much higher percentage of SW10 Schwann cells ( 76 . 7% ) than L929 fibroblasts ( 7 . 1% ) were dead 24 hrs after exposure to 30 ng/ml of mycolactone . No significant increase of trypan blue-positive cells was observed at 48 hrs . The TUNEL reaction stains cellular nuclei undergoing the apoptotic process , but not those of the necrotic pathway . In the present study , TUNEL-positive cells were much more frequent in SW10 Schwann cells than in L929 fibroblasts 24 hrs following exposure to 30 ng/ml of mycolactone . The amounts linearly increased to 84 . 1% ( SW10 ) and 25 . 4% ( L929 ) at 48 hrs . Chronological differences between the cell death study and apoptosis study may be reflective of the cytotoxic mechanism of mycolactone , but further study is required . As well as caspase-7 , Caspase-3 is a key protease involved in cellular apoptotic processes [15] . That mycolactone induced cleaved caspase-3 , a late apoptosis marker , as well as phospho-histone H2A . X , a marker for early apoptosis , in Schwann cells but not in fibroblasts or macrophages was confirmed by Western-blotting and immunocytochemistry in the present study ( Fig 4A ) . Moreover , in the comparison of various cell lines ( Fig 4B ) , mouse and human Schwann cells showed strong induction of both cleaved caspase 3 and p-histone H2A . X by mycolactone . Mouse neuroblasts also showed induction of p-histone H2A . X . In contrast , clear induction was not observed in fibroblasts , myoblasts , endothelial cells or macrophages . These overall results suggest that Schwann cells and neurons are more sensitive to mycolactone than other cell types . Given the previous study demonstrating that mycolactone diffuses passively across the cell membrane [16] , it is conceivable that susceptibility to mycolactone could be defined by several factors ( e . g . , distribution , expression level , or affinity ) associated with cellular target molecules of mycolactone rather than particular membrane receptors . In this context , mycolactone is known to directly bind to WASP/NWASP , resulting in cytoskeletal rearrangements and detachment of target cells [17] . Considering the physiological importance of WASP in the development of Schwann cells [18] , it is possible that cellular susceptibility to mycolactone may be affected by cell type specific WASP/NWASP expression and distribution , or their affinity to mycolactone . Clearly , further studies will be needed to clarify which factors define cellular susceptibility to mycolactone . Recently , Marion et al . [19] demonstrated that mycolactone induces hypoesthesia by eliciting signaling through type 2 angiotensin II receptors , leading to potassium-dependent hyperpolarization of neurons . In their study , murine sciatic nerves innervating the infected footpads did not show morphological changes . In our previous studies [3] [4] , murine nerve bundles at the inoculation sites showed degeneration , but sciatic nerves were not examined as they are distant from the lesion and pathology was not expected in our animal model . In our experimental model , we have demonstrated apoptosis at 12 hrs ( Western blot ) by mycolactone at a concentration similar to that found in human Buruli ulcer lesions ( 30ng/ml ) . On the other hand , Marion et al . applied 350ng/ml of mycolactone to primary hippocampal neuronal culture and demonstrated voltage change but no cell death within 20 min . As the temporal axes of two studies are different , direct comparison is impossible , but approximately 30 to 300 ng/ml of mycolactone seem to have both short-term effect and long-term effect to cultured cells . Guenin-Macé et al . [17] showed that mycolactone modifies actin assembly and distribution in the cytoplasm by hijacking the Wiskott-Aldrich syndrome protein ( WASP ) family , resulting in defective cell adhesion and directional migration of epithelial cells . Detachment of cultured cells in our study may be evoked by this WASP-mediated process [17] . They also found that a distinct truncated version of mycolactone minimal structure inhibits the inflammatory cytokine responses of human primary cells at noncytotoxic doses in vitro [20] . Such studies could reveal new therapeutic uses for modified mycolactone substances . Recent studies revealed that mycolactone inhibits the production of inflammatory mediator by blocking the translocation into endoplasmic reticulum [21] . This immunosuppressive process is evoked by selective blockage of Sec-61 translocon by mycolactone [22] . Lack of inflammatory cells in the Buruli ulcer lesion is reasonably explained by these studies . In summary , we demonstrated the cytotoxicity of synthetic mycolactone A/B and its remote diastereomer in mouse Schwann cells as well as fibroblasts and macrophages . A quantitative study showed that Schwann cells are relatively more sensitive than fibroblasts to mycolactone 24 and 48 hrs after exposure to mycolactone A/B diastereomer . The painless nature of Buruli ulcer may be caused by the cytotoxicity of mycolactone A/B in cultured Schwann cells . | Buruli ulcer is a chronic skin disease caused by Mycobacterium ulcerans , and the disease is characterized by the painless nature of its lesion . Similar to leprosy , loss of pain often hinders the patients from taking proper medical care , resulting in gross deformities . A toxic lipid mycolactone produced from Mycobacterium ulcerans was thought to block the sensory system of the lesion , either by direct cellular damage ( cytotoxicity ) to the regional nerve tissue , or by a more sophisticated , non-toxic paralyzing mechanism . In the peripheral nerve , Schwann cells nourish axons and accelerate nerve conduction . In this study , we have compared the cytotoxic potential of mycolactone on cultured Schwann cells and that on fibroblasts , and found that mycolactone A/B induced much higher cell death and apoptosis in Schwann cell line SW10 than in fibroblast line L929 . These results support the cytotoxic theory and suggest that mycolactone is a key substance in the production of nerve damage of Buruli ulcer . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"blood",
"cells",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"endothelial",
"cells",
"cell",
"processes",
"immunology",
"tropical",
"diseases",
"macroglial",
"cells",
"fibroblasts",
"toxicology",
"epithelial",
"cells",
"animal",
"models",
"bacterial",
"diseases",
"model",
"organisms",
"connective",
"tissue",
"cells",
"experimental",
"organism",
"systems",
"schwann",
"cells",
"neglected",
"tropical",
"diseases",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"buruli",
"ulcer",
"white",
"blood",
"cells",
"animal",
"cells",
"connective",
"tissue",
"biological",
"tissue",
"mouse",
"models",
"glial",
"cells",
"cytotoxicity",
"cell",
"biology",
"anatomy",
"apoptosis",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"macrophages"
] | 2017 | Mycolactone cytotoxicity in Schwann cells could explain nerve damage in Buruli ulcer |
Australia uses a protocol combining human rabies immunoglobulin ( HRIG ) and rabies vaccine for post-exposure prophylaxis ( PEP ) of rabies and Australian bat lyssavirus ( ABLV ) , with the aim of achieving an antibody titre of ≥0 . 5 IU/ml , as per World Health Organization ( WHO ) guidelines , as soon as possible . We present the course of PEP administration and serological testing for four men with complex requirements . Following dog bites in Thailand , two men ( 62 years old , 25 years old ) received no HRIG and had delayed vaccine courses: 23 days between dose two and three , and 18 days between dose one and two , respectively . Both seroconverted following dose four . Another 62-year-old male , who was HIV-positive ( normal CD4 count ) , also suffered a dog bite and had delayed care receiving IM rabies vaccine on days six and nine in Thailand . Back in Australia , he received three single and one double dose IM vaccines followed by another double dose of vaccine , delivered intradermally and subcutaneously , before seroconverting . A 23-year-old male with a history of allergies received simultaneous HRIG and vaccine following potential ABLV exposure , and developed rash , facial oedema and throat tingling , which was treated with a parenteral antihistamine and tapering dose of steroids . Serology showed he seroconverted following dose four . These cases show that PEP can be complicated by exposures in tourist settings where reliable prophylaxis may not be available , where treatment is delayed or deviates from World Health Organization recommendations . Due to the potentially short incubation time of rabies/ABLV , timely prophylaxis after a potential exposure is needed to ensure a prompt and adequate immune response , particularly in patients who are immune-suppressed or who have not received HRIG . Serology should be used to confirm an adequate response to PEP when treatment is delayed or where a concurrent immunosuppressing medical condition or therapy exists .
Without appropriate management , infection with rabies virus or with Australian bat lyssavirus can lead to progressive , fatal neurologic illness . Whilst Australia is free of classical rabies , Australian bat lyssavirus ( ABLV ) is endemic in local bat populations [1] . Further , Australians are taking increasing numbers of short , return international trips annually , including to regional destinations where rabies is endemic . Many – 64 of 65 individuals requiring post exposure prophylaxis ( PEP ) in a recent Australian paper [2] – travel without pre-exposure rabies prophylaxis . National guidelines for PEP of rabies and ABLV , using human rabies immunoglobulin ( HRIG ) and rabies vaccine , are used in Australia [3] , [4] . Reported local exposures to lyssaviruses managed in Queensland are assessed in conjunction with the local Public Health Unit ( PHU ) . The aim of post-exposure vaccination is to achieve an antibody titre of ≥0 . 5 IU/mL , as per World Health Organization ( WHO ) guidelines [5] , as quickly as possible . In line with the United States [6] , Australia moved from a five dose to a four dose standard PEP protocol in November 2010 [7] . Current PEP guidelines for both potential rabies or ABLV contact require that healthy individuals without previous rabies vaccination receive four vaccine doses on days 0 , 3 , 7 , and 14 after exposure , with a fifth dose recommended ( day 28 ) only in the case of immune impairment ( through disease or treatment ) [4] . Patients who have not undergone pre-exposure prophylaxis receive HRIG as part of PEP to provide early protection against migration of the virus to the central nervous system , until a protective vaccine-induced titre is achieved [8]; usually seen by day 14 [9] . For patients who have received previous rabies vaccination , HRIG is not required and only two doses of vaccine are given on days 0 and 3 [3] . Once commenced , every effort should be made to comply with dosing and timing for PEP schedules , including both HRIG and vaccine . Whilst short interruptions of some days in receiving scheduled doses are generally not of concern , the impact of longer delays of weeks is not known [10] . In these situations , serological testing , to monitor the immune response , taken seven to 14 days following the final vaccine dose in the series , has been recommended [11] . The Australian Immunisation Handbook states that confirmatory serology is not routinely necessary , but should be done two to three weeks following pre-exposure prophylaxis in immunosuppressed patients at risk of exposure to ABLV or rabies , and at two to four weeks following PEP in immunosuppressed patients after the recommended fifth dose given at day 28 [3] . ABLV was first identified in the brain of a young black flying fox ( Pteropus alecto ) found in 1996 unable to fly , and subsequently in a 1995 archived specimen from the same species [12] , [13] . ABLV has been identified as the cause in two human deaths , with neither case having received PEP [14] , [15] . Bat ABLV seroprevalance is low ( <1% ) in general surveys , but higher in sick , injured , or rescued bats [1] . Thirty-one ( 5 . 2% ) of 600 bats submitted to Queensland Scientific Services between 1998 and 2006 were positive for ABLV by direct fluorescent antibody testing [16] . There is no direct empiric evidence demonstrating the effectiveness of current PEP regimens for preventing ABLV in humans . An early study showed mice administered a variety of commercial animal and human rabies vaccines were uniformly protected against intracerebral ABLV challenge [13] . A 2005 study showed 15 of 19 mice vaccinated twice via the intraperitoneal route using a human diploid cell vaccine ( HDCV ) survived peripheral ABLV exposure , and 10 of 20 survived intracranial exposure [17] . Whilst somewhat reassuring , there are to date no published effectiveness or serology reports on individuals who have received PEP in Australia for potential ABLV exposure . When potential rabies exposure occurs overseas , the traveller may return home having had an altered or delayed PEP course or not having commenced treatment [2] . Those with a pre-existing medical condition or using medication that results in immunosuppression may also require monitoring of PEP response . Here we present four cases with potential exposure to rabies or ABLV between 2009 and 2011 where treatment deviated from standard recommendations or occurred with concurrent immunosuppression .
One potential ABLV and three rabies exposures , non-standard in nature , were referred to Brisbane North PHU for public health physician advice . In this case series , we report PEP schedules and serological responses ( Table 1 ) . Vaccination date refers to the number of days post exposure , day 0 being the date of injury/exposure . Serology specimens were tested by Queensland Health using an ELISA assay , Platelia Rabies II kit ( Bio-Rad , Hercules , California , USA ) , according to manufacturer's guidelines . This assay has very high agreement for detecting total anti-rabies virus glycoprotein antibodies compared to the WHO-recommended standard [18] . Results are expressed in equivalent units per millilitre ( EU/mL ) , which correlate with international units ( IU/mL ) : a value ≥0 . 50 EU/mL represents seroconversion [18] . In this case series , serum specimens were taken from patients to test for rabies antibodies . Oral informed consent , as opposed to written consent , was provided by patients for these procedures as they were part of routine clinical care . This case series was assembled in retrospect by the physicians who participated in the management of these patients . As such , none of the interventions or management were part of a pre-designed research study which would require prior ethics committee approval . We have written consent from each of the four patients for the inclusion of their de-identified details and clinical story in the case series .
Patient 1 , a 62-year-old male traveller , suffered an unprovoked dog bite ( rabies status unknown ) to his thigh in Thailand in February 2009 . The wound was cleaned with soap and water for at least five minutes . On the same day , he received an intramuscular dose of cell-culture derived inactivated rabies vaccine ( Verorab , Sanofi Pasteur ) , but no HRIG . On day 3 he received a second dose of vaccine in Thailand , but had no further treatment until he sought medical care on day 26 , several days after returning to Australia . At this point he was given a third dose of rabies vaccine ( Rabipur , CSL Biotherapies/Novartis Vaccines ) . Rabies serology titre was collected 24 hours later: this was prior to completion of the standard four dose course , and returned a value of 0 . 31 EU/mL . After consultation with a public health physician , he was offered a dose of rabies vaccine on the day that the serology became available ( day 30 ) , a further dose of rabies vaccine on day 33 , with repeat serology and a final dose of rabies vaccine on day 40 . Serology collected on day 33 was 1 . 99 EU/mL . This gentleman received a sixth dose of vaccine on day 40 , despite having adequate immunity on day 33 and not in keeping with the five-vaccine protocol in use at the time [3] . The patient remains well . Patient 2 , a 25-year-old male , was bitten on the left ring finger by a stray puppy ( rabies status unknown ) in Thailand in January 2010 . The wound was cleaned with soap and water for at least five minutes . He received an intramuscular dose of rabies vaccine in Thailand on the same day , but no HRIG . Shortly afterwards , he returned to Australia but did not seek further treatment until day 18 . His general practitioner contacted the local PHU and was asked to perform rabies serology prior to giving two doses of intramuscular rabies vaccine simultaneously , meaning both testing and treatment were not consistent with standard recommendations . The rabies titre result was 0 . 38 EU/mL . Dose four of rabies vaccine was given and further serology performed ( day 22 ) which gave a titre of 0 . 44 EU/mL . Further doses of rabies vaccine were given on days 29 and 36 . Repeat serology ( day 29 , not available on day 36 ) showed a rabies titre of 2 . 17 EU/mL , and the patient remains well . Patient 3 , a 62-year-old male , suffered an unprovoked dog bite ( rabies status unknown ) on the back of his left calf on a Thai beach in October 2010 . The patient did not report wound cleaning . The patient was HIV-positive and had non-insulin dependent diabetes mellitus . His most recent CD4 count , taken approximately six months earlier , was 560/µL ( within normal limits ) . The patient sought care in Thailand six days after the event , where he was given a dose of rabies vaccine ( Verorab ) and HRIG , and received a second dose of vaccine at the same clinic three days later . On his return to Australia , he presented to his local emergency department ( ED ) on day 14 where he was given an intramuscular dose of vaccine ( Rabipur ) . He had further doses on days 20 and 34 . In view of his HIV status , rabies serology was performed on day 49 , which showed a sub-therapeutic titre of <0 . 12 EU/mL . At this stage , in consultation with the PHU , an infectious diseases physician and a HIV specialist , he was given a double dose of intramuscular Rabipur immediately ( day 63 ) and serology was repeated four weeks later ( titre <0 . 12 EU/mL ) . On day 161 a double dose of vaccine ( HDCV , Sanofi Pasteur ) was given intradermally and subcutaneously ( relative volumes given intradermally and subcutaneously not available from clinical notes ) . Serology performed two weeks later provided a titre of 1 . 39 EU/mL . The patient remains well . Patient 4 , a 23-year-old male with a medical history that included eczema , asthma , and allergies ( eggs , dairy products , some food colourings ) , was scratched on the back of his neck by a bat while standing under a tree in Brisbane , Queensland , in April 2010 . Bat urine also entered his eye . The patient did not report wound cleaning . The bat flew away and was unavailable for identification and ABLV testing . He attended the local ED a few hours after injury , and was advised at that time to return the next morning . HRIG and vaccination ( Rabipur ) were administered simultaneously the following day , and ten minutes later he developed a generalised rash , facial oedema and tingling in his throat . He was treated for urticaria and allergic angio-oedema with prednisolone ( 50 mg p . o ) , promethazine ( 25 mg IM ) , and ranitidine ( 300 mg p . o . ) . He was discharged on the same day with a prescription for two further daily doses of 50 mg prednisolone . After consultation between ED and PHU , the remaining doses of rabies vaccine were given in ED . In light of the patient's egg allergy , the brand of vaccine was changed to an inactivated rabies vaccine which does not contain traces of egg protein ( HDCV , Sanofi Pasteur ) , for doses on days 4 , 8 , 18 , and 32 . Because steroids were co-administered , serology was performed to monitor rabies antibody titres . On days 8 and 18 , his titre was 0 . 16 EU/mL , but had increased to 3 . 03 EU/mL on day 32 . A fifth dose of vaccine was administered to this case in keeping with national guidelines , requiring a five dose PEP regimen following ABLV exposure , at the time [3] . The patient remains well .
Failure of rabies PEP regime has been documented [19] , [20] , but is uncommon if properly administered [21] . As illustrated by our case series , PEP is often complicated by exposures in tourist settings where reliable prophylaxis may not be available , treatment being delayed until return home , or where variations in administration exist [2] , [8] . Delays in starting PEP and not receiving the recommended full course of HRIG and vaccines is common in Australian travellers [2] . These cases also illustrate the importance of timely prophylaxis after a potential rabies/ABLV exposure to ensure an adequate immune response , particularly in the context of medical conditions or treatment that may result in immunosuppression . None of the individuals in this case series had previously received rabies vaccine . Pre-exposure prophylaxis is recommended for Australian travellers spending more than a month , or working with mammals , in a rabies-endemic region; those likely to receive bites or scratches from Australian bats; or are laboratory personnel working with live lyssaviruses [3] . A recent Australian study questioned the adequacy of these recommendations given that , in their case series , most injuries occurred within 30 days of arrival in a rabies-endemic region , most were injured whilst participating in common tourist activities , more than a third did not initiate contact with animals , and the most common injury sites were hands and fingers – high risk sites for rabies transmission due to rich nerve supply [2] . These findings are reinforced by the pattern of exposure in the three subjects in our case series who were bitten by a dog in Thailand . The authors recommended all travellers to rabies endemic regions be counselled about high-risk behaviours , avoiding animal bites , and be offered pre-exposure vaccination [2] . In each of the cases described , there was a delay between injury and seroconversion . This is particularly important given the incubation period for rabies/ABLV may be as short as 2 weeks , or , rarely , several days [22] , [23] . Most rabies deaths occur when there is deviation from WHO PEP guidelines [21] , reinforcing the importance of monitoring the serological response to treatment in certain patients , such as those presented above . Patient 1 and 2 received no HRIG as part of their PEP . This is concerning , as rabies has been documented where HRIG has not been infiltrated into the wounds of exposed patients [24] , [25] . Failure to access HRIG whilst overseas , particularly in developing countries , appears to be a common problem , even following severe hand or facial injuries [2] , [26] , [27] . Given the recent shift to a standard four dose PEP schedule , it is encouraging to note that all but one case seroconverted after four doses of rabies vaccine , despite disruptions to recommended timing . There were a number of deviations from standard testing and treatment in our cases . These testing deviations in this case series have provided serological data at non-standard time points . Both patient 1 and 2 had serology collected before receiving four doses of vaccine , and , as may be anticipated , had not seroconverted . However , these assays showed the immune response in both patients had risen above baseline . Patient 2 received a double dose of vaccine following a prolonged delay after dose 1 , and both patients 1 and 2 received six doses of vaccine in total . These events are likely an outcome of excessive caution and confusion that surrounds prolonged deviations from recommended PEP schedules . The delay in patient 3 between his last intramuscular dose ( day 63 ) and effective intradermal/subcutaneous dosing ( day 161 ) was due to discussions about the most appropriate management plan , in the absence of strong evidence , and communicating this with the patient . Rabies is an almost invariably fatal disease meaning uncertainty about PEP implementation generates much anxiety for both patients and clinicians . Deviating from existing guidelines , particularly with early serology or extra vaccine doses , is likely to be inefficient and potentially confusing for the patient and others involved in their care . Every effort should be made to comply with recommended testing and treatment protocols following potential ABLV- or rabies-prone injury . Patient 4 in our series received short course oral steroids immediately following his first vaccine , and had a protective titre demonstrated on day 32 , 14 days after his fourth dose . Of note , short-course oral corticosteroids equivalent to a prednisolone dose of less than 20 mg/day are not thought to interfere with the immune response to vaccination [28] . This patient took 50 mg/day for three days , but had adequate seroconversion two weeks after dose 4 . Failure of pre-exposure prophylaxis and PEP has been documented in individuals with immune deficiency secondary to HIV infection [29] , [30] . Patient 3 had well controlled HIV , but had not mounted a protective antibody response 161 days after seven IM doses , despite a recent normal CD4 count . Interestingly , he developed a titre of 1 . 39 EU/mL shortly after a double dose of HDCV vaccine , given in part intradermally and subcutaneously . There are very limited data about using intradermal rabies vaccine in HIV-infected patients as PEP [31] , and whether it may be more effective at producing an immune response in HIV-infected patients than IM vaccine remains unclear . Care is needed in assessing HIV patients after potential rabies exposure even when immunosupression is thought to be mild . Further clinical trials are required to identify optimal PEP regimens in HIV-infected patient groups with varying immunosuppression . In conclusion , clinicians should be aware that a delay between injury/exposure and PEP administration , schedule interruption , or complicating immunosuppression may inhibit or delay seroconversion . Due to the potentially short incubation period of rabies/ABLV , particular care must be taken in such patients where HRIG has not been administered . In cases with delayed treatment , complicating medical history , or concurrent steroid or other immunosuppressing therapy or medical condition , serology should be used , where available , to monitor the response to PEP and confirm serconversion has occurred . | In Australia , the administration of rabies post-exposure prophylaxis ( PEP ) occurs for potentially exposed returned travellers from endemic regions or for potential local exposure to Australian Bat Lyssavirus . For Australian tourists , delays in commencing PEP or not receiving HRIG or all recommended doses of vaccine are common . We report a case series where serology provided information in four patients with delayed , incomplete , or complicated PEP treatment . Three of these patients reported a dog bite in Thailand and the fourth was scratched by a bat and had bat urine enter his eye in Australia . Management was complicated by lack of HRIG administration , delays in the recommended timeframe for receipt of vaccine doses , and immunosuppression caused by co-administration of steroids and by HIV infection with a normal CD4 count . All patients seroconverted but this was delayed in some cases , and in the HIV-positive subject required a double dose of vaccine delivered intradermally and subcutaneously . In complex or non-standard PEP delivery , including delayed treatment and immunosuppression due to steroid treatment , HIV or another immunosuppressing medical condition , serology can be used to guide further treatment and should be used to confirm seroconversion . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"vaccines",
"medicine",
"vaccination",
"infectious",
"diseases",
"rabies",
"clinical",
"immunology",
"immunity",
"travel-associated",
"diseases",
"neglected",
"tropical",
"diseases",
"immunology"
] | 2013 | Using Serology to Assist with Complicated Post-Exposure Prophylaxis for Rabies and Australian Bat Lyssavirus |
While acute lung injury ( ALI ) contributes significantly to critical illness , it resolves spontaneously in many instances . The majority of patients experiencing ALI require mechanical ventilation . Therefore , we hypothesized that mechanical ventilation and concomitant stretch-exposure of pulmonary epithelia could activate endogenous pathways important in lung protection . To examine transcriptional responses during ALI , we exposed pulmonary epithelia to cyclic mechanical stretch conditions—an in vitro model resembling mechanical ventilation . A genome-wide screen revealed a transcriptional response similar to hypoxia signaling . Surprisingly , we found that stabilization of hypoxia-inducible factor 1A ( HIF1A ) during stretch conditions in vitro or during ventilator-induced ALI in vivo occurs under normoxic conditions . Extension of these findings identified a functional role for stretch-induced inhibition of succinate dehydrogenase ( SDH ) in mediating normoxic HIF1A stabilization , concomitant increases in glycolytic capacity , and improved tricarboxylic acid ( TCA ) cycle function . Pharmacologic studies with HIF activator or inhibitor treatment implicated HIF1A-stabilization in attenuating pulmonary edema and lung inflammation during ALI in vivo . Systematic deletion of HIF1A in the lungs , endothelia , myeloid cells , or pulmonary epithelia linked these findings to alveolar-epithelial HIF1A . In vivo analysis of 13C-glucose metabolites utilizing liquid-chromatography tandem mass-spectrometry demonstrated that increases in glycolytic capacity , improvement of mitochondrial respiration , and concomitant attenuation of lung inflammation during ALI were specific for alveolar-epithelial expressed HIF1A . These studies reveal a surprising role for HIF1A in lung protection during ALI , where normoxic HIF1A stabilization and HIF-dependent control of alveolar-epithelial glucose metabolism function as an endogenous feedback loop to dampen lung inflammation .
Acute lung injury ( ALI ) is an inflammatory disease of the lungs that is characterized by hypoxemic respiratory failure with bilateral pulmonary infiltrates , not attributable to left heart failure [1]–[4] . Among the hallmarks of ALI are severe arterial hypoxemia and uncontrolled accumulation of inflammatory cells into different compartments of the lungs , in conjunction with cytokine release and inflammatory activation of recruited or resident cells . Particularly alveolar epithelial injury plays a key role in the pathogenesis of ALI , leading to disruptions of the alveolar-capillary barrier function , resulting in extensive pulmonary edema , attenuated gas exchange , and uncontrolled lung inflammation [2] . While there is currently no specific therapy available , management consists of aggressive treatment of the initiating cause , vigilant supportive care , and the prevention of nosocomial infections . From a clinical perspective , it is important to point out that ALI is among the leading causes of morbidity and mortality of patients requiring critical care medicine . For example , a large scale prospective , population-based , cohort study indicates that each year in the United States there are close to 200 , 000 cases of ALI , which are associated with 74 , 500 deaths and 3 . 6 million hospital days [5] . Moreover , long-term disabilities in ALI survivors are significant . A landmark study who followed ALI survivors over 5 years revealed that exercise limitation , physical and psychological sequelae , decreased physical quality of life , as well as increased costs and use of health care services are important legacies of severe lung injury [6] . Taken together , these studies highlight the importance for finding novel ALI treatment approaches with the goal to dampen excessive lung inflammation [2] , [7]–[9] or to support its resolution phase [10]–[16] . While ALI has a major impact on the morbidity and mortality of patients requiring critical care medicine and mechanically ventilation , many episodes of ALI are self-limiting , and resolve spontaneously through molecular pathways that are transcriptionally controlled [10]–[12] , [17] . In spite of profound inflammatory responses to surgery and mechanical ventilation [18] , patients undergoing major thoracic surgery for lung cancer have an overall incidence of ALI of less than 5% [19] , open heart surgery with cardiopulmonary bypass less than 0 . 5% [20] , or kidney transplantation of less than 0 . 2% [21] . Based on these clinical observations , we hypothesized that innate adaptive pathways exist that dampen acute pulmonary edema and lung inflammation during ALI in mechanically ventilated patients . To identify novel endogenous pathways important for lung protection and ALI resolution that are turned on by mechanical ventilation , we utilized stretch exposure of pulmonary epithelial cells as a single-cell-based in vitro model for ventilator-induced ALI [22] , [23] . Very surprisingly , genome-wide microarray screening for alterations in gene transcription of human pulmonary epithelial cells exposed to cyclic mechanical stretch conditions revealed a transcriptional response that shared many similarities with ambient hypoxia exposure . Subsequent studies identified a molecular pathway resulting in normoxic stabilization of the transcription factor hypoxia-inducible factor HIF1A during stretch conditions in vitro or during ALI induced by mechanical ventilation in vivo . HIF is a transcription factor that is critical for tissue adaptation to conditions of hypoxia [4] , [24]–[31] . More recently , HIF has also been implicated in the control of immune responses during inflammatory processes including ischemia and reperfusion , cancer or mucosal inflammation [3] , [4] , [32]–[34] . Sites of acute inflammation are frequently characterized by considerable shifts in the supply and demand of metabolites that result in limited oxygen availability ( inflammation-associated hypoxia ) [7] , [35]–[38] . Consistent with a role for hypoxia-signaling in cellular adaptation to conditions of limited oxygen availability [31] , [36] , [39] , several studies implicate hypoxia-dependent signaling pathways in the attenuation of mucosal inflammation [7] , [33] , [40]–[43] . Surprisingly and in contrast to some of the above studies on inflammatory hypoxia [36] , [37] , [44] , [45] , we observed that pulmonary stabilization of HIF1A during ALI occurs under normoxic conditions . Consistent with previous studies [46]–[48] , our findings implicate a metabolic pathway in normoxic HIF stabilization , including stretch-induced inhibition of succinate dehydrogenase ( SDH ) activity . In fact , previous studies had indicated that normoxic stabilization of HIF in cancer involves inhibition or mutation of the metabolic enzyme SDH [46] , [48] . Extension of these initial findings utilizing pharmacologic and genetic models of hypoxia-signaling during ALI revealed that HIF controls a functional link between alveolar epithelial metabolism and lung inflammation during ALI , thus implicating pharmacologic strategies to activate these HIF-dependent metabolic pathways for ALI treatment . Indeed , compounds such as HIF activators have been used safely in patients [49] .
Previous studies had demonstrated that ALI is associated with significant alterations of gene-expression that frequently resembles endogenous adaptive responses [2] , [7] , [35] . To systematically examine alterations of pulmonary gene expression during ALI , we utilized a simple in vitro model of ALI by exposing human pulmonary epithelia to cyclic mechanical stretch conditions . This in vitro model resembles alveolar injury as occurs during ventilator-induced ALI in humans [22] , [23] , [40] . For this purpose , we examined human alveolar epithelial cells ( Calu-3 ) following 24 h stretch exposure at 30% intensity by performing a genome microarray screen ( http://www . ncbi . nlm . nih . gov/projects/geo/query/acc . cgi ? acc=GSE27128 ) . We subsequently confirmed the array results utilizing real-time RT-PCR and Western blotting for a subset of genes that were regulated ( Table S1 ) . Surprisingly , computerized pathway analysis to examine alterations in gene transcription ( Ingenuity IPA , Version 11631407 ) revealed that hypoxia-signaling resembled the dominant stress response pathway when comparing stretch-exposed pulmonary epithelia to un-stretched controls ( Figure 1A; Figure S1 ) . In fact , subsequent studies of Calu-3 pulmonary epithelia exposed to different time-periods of stretch ( Figure 1B ) or studies utilizing a HIF reporter plasmid transfected into pulmonary epithelia ( A549 ) and exposed to stretch conditions revealed stabilization of HIF1A—the key transcription factor for hypoxia adaptation ( Figure 1C ) [30] . Similarly , exposure to different degrees of cyclic mechanical stretch ( Figure 1D ) showed stretch-dose-dependent stabilization of HIF1A . Moreover , studies in primary human alveolar epithelial cells demonstrated robust HIF1A stabilization following stretch exposure ( Figure 1E ) . To examine if HIF1A stabilization during stretch has functional consequences , we generated pulmonary epithelial cell lines with stable siRNA-mediated repression of HIF1A ( Figure S2A–C ) . As HIF1A functions as an important regulator of glycolysis [50] we first explored the possibility of stretch controlling transcription of glycolytic enzymes . Exposure of control-transduced pulmonary epithelial cells demonstrated very robust induction of the transcript levels for phosphofructokinase m ( PFKM ) , pyruvate dehydrogenase kinase 1 ( PDK1 ) ( Figure S3A ) , and lactate dehydrogenase a ( LDHA , Figure 1F ) with stretch exposure . In contrast , stretch-induced induction of glycolytic enzymes was completely abolished in pulmonary epithelial cells with HIF1A repression ( Figure S3A and Figure 1F ) . Similarly , stretch-induced increases of lactate levels in their supernatant or increases in glycolytic flux as measured by metabolic turnover of 13C-labeled glucose during stretch were HIF1A-dependent ( Figure 1G , H ) . These findings were specific for HIF1A , as cells with pulmonary epithelial cells with siRNA-mediated HIF2A repression behaved similar to controls ( Figures S2 and S3 ) . Together , these studies demonstrate stabilization of HIF1A during cyclic mechanical stretch exposure of human pulmonary epithelial cells in vitro and reveal transcriptional and functional consequences of HIF1A stabilization on carbohydrate metabolism . Based on the above findings that identified stabilization and increased transcriptional and functional activity of HIF1A during cyclic mechanical stretch conditions , we next pursued studies to address the mechanism of stretch-induced HIF1A stabilization . As a first option , we considered the possibility that stretch-associated increases in oxygen consumption could lead to a decrease in oxygen availability , and subsequent hypoxic stabilization of HIF1A . Previous studies had shown that shifts in cellular oxygen levels between the cytosol and the mitochondria—as caused by nitric oxide—can cause HIF stabilization [51] . To address this possibility , we performed measurements of oxygen partial pressures within the supernatant of stretched pulmonary epithelial cells ( Figure 2A ) [52] . However , we observed high oxygen partial pressures in controls or stretched pulmonary epithelial cells , indicating that HIF1A stabilization during stretch occurs under normoxic conditions . Previous studies had indicated that normoxic stabilization of HIF in cancer involves inhibition or mutation of the metabolic enzyme SDH [46] . These studies demonstrate that mutations of SDH are associated with inhibition of prolylhydroxylases ( PHDs ) —a group of enzymes responsible for tagging HIF for proteasomal degradation [24] , [53]–[55]—thereby causing normoxic HIF stabilization [46]–[48] . To examine the possibility that stretch-induced HIF1A stabilization could involve SDH inhibition , we next examined the consequences of pulmonary epithelial stretch exposure on SDH activity . We observed that stretch exposure was associated with a very robust attenuation of SDH activity ( Figure 2B ) . To address if increased succinate levels with following PHD inhibition were the mechanism for HIF1A stabilization associated with low SDH activity , we generated cells with siRNA repression of succinate-CoA ligase ( SUCLG; Figure 2C ) . While SDH inhibition leads to the accumulation of succinate and succinate-elicited PHD inhibition , SUCLG deletion would prevent the conversion of succinate-CoA to succinate and thereby lower succinate levels [46]–[48] . As shown in Figure 2D , stretch-induced HIF1A stabilization was completely abolished in cells with siRNA-mediated SUCLG repression , thereby implicating a “succinate leak”—caused by stretch-induced SDH inhibition—in normoxic HIF1A stabilization during cyclic mechanical stretch of human pulmonary epithelial cells . To further address the possibility of stretch-elicited inhibition of SDH as a mechanism of normoxic HIF stabilization , we next performed studies with cell permeable α-ketoglutarate in studies on succinate-dependent HIF1A stabilization . Earlier studies found that succinate-mediated inhibition of PHD is competitive and is reversed by pharmacologically elevating intracellular α-ketoglutarate [56] . Introduction of α-ketoglutarate derivatives restores normal PHD activity and HIF1A levels and thereby alleviates pseudo-hypoxia . In fact and as shown in Figure 2E , α-ketoglutarate treatment prevented HIF1A stabilization in pulmonary epithelia at 8 and 24 h of stretch . To address possible mechanisms of stretch/stress-induced SDH inhibition , we next pursued studies on the effect of stretch on stress kinases . Recent studies found enhanced mitogen-activated protein kinase ( MAPK ) activation following stretch of lung epithelia [57] . As such we measured SDH activity during stretch with and without inhibitors of JNK , ERK , or p38 . As shown in Figure 2F–H , inhibition of these MAPKs abolished stretch-mediated SDH inhibition . Based on reports of normoxic HIF1A stabilization by pro-inflammatory cytokines [58] , we next tested for stretch-induced HIF1A expression under conditions of neutralization of TNFA or IL-6 using neutralizing antibodies . As shown in Figure 2I , TNFA or IL-6 neutralizing antibodies blunted stretch-mediated HIF1A stabilization in lung epithelia . As TNFA is able to up-regulate stress kinases such as p38 [59] , these findings indicate that TNFA release by stretched pulmonary epithelia might be involved in the up-regulation of stress kinases with concomitant inhibition of SDH activity . To address the functional consequences of HIF on stretch-associated alterations of metabolism beyond anaerobic glycolysis , we next examined tricarboxylic acid cycle ( TCA ) flux and mitochondrial functions of human pulmonary epithelial cells ( Calu-3 ) following siRNA-mediated HIF1A repression . As shown in Figure 3A , studies tracing different cellular metabolites in the presence of labeled glucose ( 13C-glucose ) revealed that TCA flux was significantly increased in control cells , but not following siRNA-mediated HIF1A repression . As HIF1A functions to improve mitochondrial respiration and concomitant ATP production during hypoxia , we next assessed if HIF1A-dependent TCA flux increases during normoxia could be associated with increases of HIF1A-dependent mitochondrial complex IV ( COX4 ) activity [60] . In fact , stretch exposure was associated with increased COX4 activity and ATP levels in controls , but not in pulmonary epithelial cells with HIF1A repression ( Figure 3B , C ) . Together , these findings indicate that HIF1A optimizes metabolic functions during stretch conditions in vitro . We next performed studies to better understand increases in TCA substrate flux despite up-regulation of PDK1 . As shown , stretch stress causes normoxic HIF1A stabilization and HIF1A-dependent glycolytic enzymes expression . Among them pyruvate dehydrogenase kinase isoenzym 1 ( PDK1 ) is up-regulated . This enzyme is not to be confused with phosphoinositide-dependent-kinase-1 ( also abbreviated PDK1 ) . The pyruvate dehydrogenase kinase isoenzym 1 ( PDK1 ) phosphorylates the enzyme of the same name ( specifically PDE1 ) , which is the major component of the pyruvate dehydrogenase complex ( PDC ) . Phosphorylation of pyruvate dehydrogenase at serine residue 1 of 3 possible ones by PDK1 will almost completely inhibit activity of the PDC . As shown in Figure 1G , inhibition of the PDC fits well with increased production of lactate , as pyruvate is expected to accumulate because it is less metabolized in the first step of the TCA controlled by PDC activity . However , how does up-regulation of PDK1 , a master inhibitor of the PDC , fit to a several-fold increase in TCA/C13-Gucose ratios ( Figure 2E ) . To address this question , we next measured mitochondrial PDC activity [61] during stretch conditions . In particular , we measured the activity of the pyruvate dehydrogenase ( PDH ) . As shown in Figure 3D we found increases of PDH activity in a HIF1A-dependent manner upon stretch . These findings further supports our hypothesis of normoxic HIF1A stabilization and HIF1A-dependent increases of TCA cycle activity despite findings of PDK1 mRNA induction ( Figure S3 ) . Consistent with a mitochondrial dysfunction in HIF1A knock-down cells , we next observed increased reactive oxygen species ( H2O2 , Figure 3E ) . Moreover , markers for epithelial cell inflammation were elevated in Calu-3 cells with HIF1A repression ( Figure 3F–G ) . Similarly to the metabolic alterations , these findings were specific for HIF1A , since HIF2A knockdown cells showed similar responses as controls ( Figure S3C ) . Together , these findings indicate a functional role for normoxic HIF1A stabilization during conditions of cyclic mechanical stretch in attenuating pulmonary epithelial inflammation by optimizing cellular metabolism of carbohydrates . After having observed a transcriptional hypoxia-program in pulmonary epithelia exposed to stretch conditions in vitro , we went on to examine HIF1A during in vivo conditions of ALI . For this purpose , we utilized a previously described model of mechanical ventilation-induced ALI [23] , [40] , [62] . We gained initial insight from immune-histochemical studies of pulmonary tissues following exposure to ALI . Consistent with in vitro studies of stretch exposure , we observed increased HIF1A staining following ALI induction ( Figure 4A ) . Similarly , Western blotting indicated that HIF1A stabilization during ALI is time-dependent and occurs more pronounced with increased stretch conditions—such as those encountered during mechanical ventilation at higher inspiratory pressure levels ( 35 mbar versus 45 mbar; Figure 4B , C ) . In addition , we examined pulmonary HIF1A levels utilizing a previously characterized HIF reporter mouse model [63] . Consistent with the above studies utilizing Western blotting for HIF1A , we observed elevated HIF1A levels following ALI induction ( Figure 4D ) . Interestingly , studies utilizing different concentrations of oxygen in the inspired air during ventilator-induced ALI ( 20% oxygen versus 100% oxygen ) demonstrated that HIF1A stabilization is not affected by increasing the level of inspired oxygen concentration—suggesting that HIF1A stabilization occurs under conditions with sufficient oxygen availability ( Figure 4E ) . Similarly to stretch-associated HIF stabilization , we observed robust repression of SDH activity following exposure to ALI induced by mechanical ventilation ( Figure 4F ) , thereby implicating SDH-associated increases in succinate in the normoxic stabilization of HIF during ALI . Together , these findings reveal normoxic stabilization of HIF during ventilator-induced ALI . Based on the above finding that HIF1A is stabilized during ALI , we next performed pharmacologic studies to address the functional role of HIF1A stabilization on ALI outcomes . First , we utilized the HIF1A activator dimethyl-oxaloylglycine ( DMOG ) . DMOG-associated stabilization of HIF1A involves inhibition of PHDs [64] , [65]—a mechanism very similar to succinate-dependent HIF stabilization . For the purpose of these studies , we pretreated mice with DMOG ( 1 mg i . p . ) 4 h prior to the induction of ALI . This pharmacologic approach was associated with robust HIF1A stabilization in murine lungs ( Figure S4A ) . Functional studies revealed that ALI-associated increases in pulmonary edema were dramatically improved following DMOG treatment ( Figure 5A ) . In addition , pre-treatment with DMOG was associated with attenuated lung inflammation , as assessed by measurements of myeloperoxidase levels in the lungs or in the BAL , indicating that pretreatment with DMOG attenuates pulmonary neutrophil accumulation in the lungs during ALI ( Figure 5B , C ) . Similarly , attenuated gas exchange during VILI was improved in mice pre-treated with DMOG ( Figure 5D ) . Finally , pre-treatment with DMOG was also associated with a very dramatic increase of survival time during VILI exposure ( Figure 5E ) . Together , these pharmacologic studies indicate that stabilization of HIF during ALI functions to dampen pulmonary edema and lung inflammation during ALI induced by mechanical ventilation in vivo . After having observed very robust therapeutic effects for the treatment with a pharmacologic HIF activator , we next examined the consequences of a pharmacologic inhibitor of HIF1A activity . For these studies we utilized echinomycin , which prevents binding of HIF1A to the DNA , thereby preventing functional HIF1A activation [66] . Echinomycin pretreatment prevented HIF-dependent induction of pulmonary HIF target genes , such as PFKM , PDK1 , or LDHA transcript levels ( Figure S4B ) . In contrast to studies with the HIF1A activator DMOG , we observed increased pulmonary edema ( Figure 6A ) , in conjunction with increased pulmonary and BAL myeloperoxidase levels ( Figure 6B , C ) , attenuated gas exchange ( Figure 6D ) , and decreased survival time during ventilator-induced ALI in mice that were pre-treated with the HIF inhibitor echinomycin ( Figure 6E ) . Together , these pharmacologic studies indicate that inhibition of the transcriptional functions of HIF during ALI is highly detrimental , implicating normoxic HIF1A stabilization in an endogenous feedback loop to protect the lungs from excessive inflammation and pulmonary edema during ALI . To gain insight into the tissue-specific source of HIF1A-dependent lung protection , we next performed genetic studies for HIF1A during ventilator-induced ALI . Due to the fact that Hif1a−/− mice die during embryogenesis [67]–[69] , we utilized transgenic mice with a “floxed” HIF1A gene [70] to systematically delete HIF1A in different tissue compartments of the lungs . As first step , we generated mice with induced deletion of HIF1A in all tissues , including the lungs ( Hif1af/fActinCre+ , Figure S4C ) . Consistent with our above findings with pharmacologic HIF inhibition , we observed dramatic decreases in survival time , paired with increased pulmonary edema and attenuated gas exchange during ALI exposure of Hif1af/f ActinCre+ mice compared to ActinCre+ controls ( Figure 7A ) . In contrast , mice with deletion of HIF1A in vascular endothelia ( Hif1af/f CadherinCre+; Figure 7B , Figure S5A ) , myeloid cells ( Hif1af/f LysozymCre+; Figure 7C ) [70] , or the conducting airways ( Hif1af/f ClaraCellCre+; Figure 7D , Figure S5B ) [71] showed no difference in survival time , pulmonary edema , or gas-exchange during ALI as compared to corresponding controls . However , mice with induced deletion of HIF1A in their alveolar epithelial cells ( Hif1af/f SurfactantCre+; Figure 7E , Figures S5C and S6 ) [72] showed a similar phenotype as mice with induced deletion of HIF in all tissues . Indeed , Hif1af/f SurfactantCre+ demonstrated attenuated survival time , increased pulmonary edema , and attenuated gas exchange during ventilator-induced ALI . Taken together , these studies provide genetic in vivo evidence for a protective role for alveolar-epithelial-specific HIF1A-signaling during ALI . Based on the above studies implicating alveolar epithelial HIF1A in lung protection during ALI , we performed a more detailed examination of mice with conditional deletion of HIF1A in pulmonary epithelia ( Hif1af/f SurfactantCre+ ) . Isolation of alveolar epithelial cells from Hif1af/f SurfactantCre+ or control mice after exposure to mechanical ventilation confirmed alveolar-epithelial HIF1A stabilization in controls but not in Hif1af/f SurfactantCre+ ( Figure 8A , Figure S6 ) . Additional functional studies of Hif1af/f SurfactantCre+ demonstrated a more profound degree of lung inflammation during ALI , including increased pulmonary neutrophil accumulation ( Figure 8B ) , pulmonary IL-6 ( Figure 8C ) , CXCL1 ( Figure 8D ) , and pulmonary TNF-α levels ( Figure 8E ) . Similarly , pulmonary injury as assessed by a VILI score ( Figure 8F ) and as shown in a representative histologic slides ( Figure 8G ) was dramatically increased in Hif1af/f SurfactantCre+ upon exposure to ALI induced by mechanical ventilation . In addition , the lung protective effects of pharmacologic treatment with the HIF activator DMOG on albumin leakage ( Figure 8H ) , pulmonary gas exchange ( Figure 8I ) , and neutrophil accumulation in the lungs ( Figure 8J ) were abolished in mice with conditional HIF1A deletion in the alveolar epithelia , indicating that DMOG-elicited lung protection requires the presence of pulmonary epithelial HIF1A , as opposed to alternative tissue-specific sources for HIF1A such as endothelia or myeloid cells . Taken together , these findings implicate alveolar-epithelial HIF1A in dampening lung inflammation during ALI . Based on the above in vitro studies on the role of HIF1A in optimizing pulmonary epithelial carbohydrate metabolism during conditions of cyclic mechanical stretch , we next examined metabolic functions of alveolar-epithelial HIF1A during ALI in vivo . For this purpose—and as we have done in previous studies [73]—we employed liquid chromatography–tandem mass spectrometry analysis of 13C-glucose metabolites . In line with the hypothesis that alveolar-epithelial HIF1A optimizes carbohydrate metabolism during ALI in vivo , we found that elevations of pulmonary glucose levels during ALI were completely abolished in Hif1af/f SurfactantCre+ ( Figure 9A ) . Similarly , transcript levels of the glucose transporter Glut1 were dramatically increased following ALI in control mice but not in Hif1af/f SurfactantCre+ ( Figure 9B ) . Moreover , we observed that ALI-associated elevations of [13C6]-fructose 1 , 6 bisphosphate and [13C6]-lactate following exposure to ALI were completely abolished in Hif1af/f SurfactantCre+ ( Figure 9C , D ) . Similarly , ALI-associated elevations of the transcript levels of glycolytic enzymes were attenuated in Hif1af/f SurfactantCre+ mice ( Figure S7A ) . Taken together , these data indicate that ALI induced by mechanical ventilation is associated with a transcriptional program under the control of alveolar-epithelial HIF1A that leads to an increase in glycolytic flux . To address the functional role of HIF-dependent induction of the glycolytic pathway , we next performed in vivo studies with the glycolysis inhibitor 2-deoxy-D-glucose ( 2-DG ) , a well-characterized inhibitor of glycolysis [74] . Here we found that treatment with 2-DG ( 200 mg/kg i . p . prior to the induction of ventilator-induced ALI ) was associated with increased pulmonary edema ( Figure 9E ) , attenuated gas exchange ( Figure 9F ) , enhanced pulmonary neutrophil accumulation as assessed by MPO levels ( Figure 9G ) , and cytokine production ( IL-6; Figure 9H ) . Moreover , survival time during ALI exposure of anesthetized mice was significantly reduced following pretreatment with 2-DG as compared to vehicle treatment in wild-type mice ( Figure 9I ) , while 2-DG treatment of Hif1af/f SurfactantCre+ mice had no effect on survival time ( Figure 9J ) . Similar , while 2-DG treatment increased H2O2 production in wild-type mice—indicating mitochondrial stress—2-DG treatment did not further increase H2O2 levels in Hif1af/f SurfactantCre+ mice ( Figure S7B ) . Taken together , these studies indicate that enhancing the glycolytic carbohydrate flux during ALI is mediated exclusively through alveolar epithelial HIF1A . Moreover , pharmacologic studies using an inhibitor for glycolysis implicate transcriptional increases in lung epithelial glucose metabolism as an endogenous adaptive pathway to dampen inflammation during ALI in vivo . To gain additional insight into the metabolic functions of epithelial HIF1A during ALI in vivo , we next pursued metabolic studies on TCA flux and mitochondrial metabolism in Hif1af/f SurfactantCre+ mice . Consistent with the above studies in stretched pulmonary epithelial cells in vitro , we observed that ALI-associated increases in [13C]-malate and TCA flux rates were completely abolished in mice with gene-targeted deletion of HIF1A in alveolar epithelia ( Figure 10A , B ) , suggesting that alveolar epithelial HIF1A specifically promotes TCA flux during ALI in vivo . Based on our in vitro studies implicating HIF1A in stretch-induction of mitochondrial complex IV ( COX4 ) , and previous studies showing that HIF1A regulates cytochrome oxidase subunits to optimize efficiency of respiration in hypoxic cells [60] , we next performed an analysis of COX4 activity along with the protein expression of the cytochrome oxidase subunit COX4-2 during ALI . Here , we found a very robust increase of COX4 activity and the induction of COX4-2 protein levels in control mice exposed to ALI ( Figure 10C , D ) . In contrast , in vivo studies using Hif1af/f SurfactantCre+ mice during ALI demonstrated that the observed induction was completely abolished in mice with alveolar epithelial deletion of HIF1A . Accordingly , increases in pulmonary ATP levels following ALI were abolished in Hif1af/f SurfactantCre+ mice ( Figure 10E ) . Consistent with a mitochondrial dysfunction in Hif1af/f SurfactantCre+ mice during ALI , we found higher ROS levels ( Figure 10F ) . Similarly , complete inhibition of mitochondrial respiration utilizing rotenone treatment in wild-type animals exposed to ventilator associated lung injury was associated with dramatically increased lung inflammation ( IL-6 levels displayed; Figure 10G ) or alveolar lung leakage ( BAL albumin; Figure 10H ) , thereby mimicking our previous findings in Hif1af/f SurfactantCre+ mice . Taken together , these studies indicate that HIF-dependent optimization of glucose uptake , glycolysis , TCA flux , and respiration of the alveolar epithelium represents an endogenous adaptive process that counter-regulates pathologic lung inflammation and improves survival during ventilator-induced ALI ( Figure 11A ) .
The present study was designed to identify stress-elicited transcriptional pathways that could be targeted as novel treatment for ALI . For this purpose , we performed a genome-wide screen to identify transcriptional responses to cyclic mechanical stretch of pulmonary epithelial cells , such as occurs during mechanical ventilation in the context of ALI . Quite surprising , these studies revealed a transcriptional response that resembled hypoxia signaling . We found that the key transcription factor for mediating hypoxia adaptation—HIF1A—was stabilized during stretch conditions in vitro , or during ALI induced by mechanical ventilation in vivo . Unexpectedly , HIF1A stabilization occurred in the absence of tissue hypoxia , and involved normoxic inhibition of SDH , a molecular pathway that has previously been implicated in the normoxic activation of HIF during cancer [46]–[48] . In order to define the functional role of HIF1A during ALI , we subsequently performed pharmacologic studies with HIF activators or inhibitors . These findings demonstrated a protective role for HIF stabilization during ALI . Systematic deletion of the Hif1a gene in different lung tissues identified a predominant role for alveolar epithelial HIF1A signaling in these responses . Subsequent metabolic studies indicated that stretch exposure of pulmonary epithelial cells in vitro or ventilator-induced ALI in vivo is associated with the normoxic stabilization of HIF1A , thereby leading to increased glycolytic capacity , TCA flux , optimized mitochondrial respiration , and finally increased ATP generation . Indeed , HIF-dependent prevention of mitochondrial dysfunction during ALI resulted in increased alveolar epithelial capacity to produce ATP , while concomitantly preventing ROS accumulation and lung inflammation ( Figure 11 ) . We were surprised to find a functional role for pulmonary hypoxia-signaling during ALI—particularly because the lungs represent one of the best oxygenated environments in the body . However , this is not the first time that studies uncovered a surprising manifestation for hypoxia-signaling under normoxic conditions . As such , lactate and pyruvate have been shown to regulate hypoxia-inducible gene expression independently of hypoxia by stimulating the accumulation of HIF1A in human gliomas and other cancer cell lines [75] . In contrast to those studies , we observed that HIF stabilization occurred during ALI under normoxic conditions . Combinations of in vitro stretch studies of pulmonary epithelial cells and in vivo studies of ALI induced by mechanical ventilation implicate stretch-elicited inhibition of SDH in HIF activation . These findings are consistent with previous studies of SDH-dependent HIF activation [46]–[48] . Studies on mutations in human cancers had implicated SDH in altered cellular metabolism and cellular transformation . Inactivating SDH causes the accumulation of succinate , which inhibits 2-oxoglutarate-dependent enzymes , including PHDs that mark HIF1A for polyubiquitylation and proteasomal degradation [48] . In support of these findings , other studies have redundantly shown the critical role of metabolism in regulating HIF1A-dependent gene expression [76] . Also consistent with the current studies , previous studies have shown molecular pathways for normoxic stabilization of HIF1A . For example , a recent study described hypoxia-independent activation of HIF1A by enterobacteriaceae and their siderophores [77] . Several studies had provided evidence for HIF1A stabilization during bacterial infections [78]–[80] . In extension of these findings , the authors pursued the role of bacterial siderophores in HIF1A activation during infection with Enterobacteriaceae . Infection of mice with Y . enterocolitica led to functional activation of HIF1A in Peyer's patches . Because mice with deletion of HIF1A in the intestinal epithelium showed a significantly higher susceptibility to orogastric Y . enterocolitica infections , activation of HIF1A in host cells during bacterial infection represents a host defense mechanism in this study . Additional studies with Y . enterocolitica , S . enterica subsp enterica , or E . aerogenes , and moreover , application of their siderophores ( yersiniabactin , salmochelin , aerobactin ) , caused a robust , dose-dependent HIF1A response in human epithelia and endothelia , independent of cellular hypoxia [77] . The authors conclude that bacterial siderophores account for the normoxic stabilization of HIF1A during infection with human pathogenic bacteria . In the context of lung injury , previous studies had shown an indirect role for hypoxia-signaling in attenuating ALI induced by bacterial infection [81] . For example , an important study tested the hypothesis that oxygenation weakens a tissue-protecting mechanism triggered by hypoxia and concomitant HIF activation [41] . The authors found that the hypoxia-dependent transcriptional control of extracellular adenosine signaling protects the lungs from the toxic effects of overactive immune cells such as neutrophils [35] . In line with the present studies , these findings demonstrate a protective role for the activation of the HIF-pathway in the lungs by attenuating pathologic lung inflammation during ALI . These studies pointed towards a protective role for HIF-elicited increases of adenosine signaling through the Adora2a adenosine receptor [7] , [8] . During conditions of ischemia or inflammation , different cells release nucleotides that can be enzymatically converted to adenosine [2] , [23] , [38] , [82]–[85] . HIF1A can function in multiple ways to enhance the protective effects of extracellular adenosine signaling during ALI , including enhanced enzymatic adenosine production by transcriptionally inducing the ecto-5′-nucleotidase ( conversion of extracellular AMP to adenosine ) , which is a known HIF target gene [86] , [87] . Similarly , other studies had shown a protective role for the extracellular production of adenosine and signaling events through adenosine receptors in lung protection—for example , by attenuating vascular leakage [88]–[91] or improving alveolar fluid transport [40] . Moreover , other studies also implicate HIF in attenuating adenosine uptake via repression of equilibrative nucleoside transporters , and thereby enhancing extracellular-dependent tissue protection via adenosine signaling events [92]–[95] . Interestingly , CD73 was among the genes regulated during stretch exposure of pulmonary epithelial cells in our present study ( Table S1 ) . Consistent with a role of HIF-dependent increases of CD73-dependent adenosine generation and lung protection during ALI , we also found that the protective effects of the HIF activator DMOG was attenuated in gene-targeted mice for cd73 ( Figure S8 ) . Taken together , these studies indicate the likelihood that HIF-dependent lung protection during ALI may also involve HIF-elicited increases in extracellular adenosine receptor signaling [96] , [97] . However , it remains presently unclear if the findings of HIF1A-elicited improvements in carbohydrate metabolism of pulmonary epithelial cells are related or independent of adenosine metabolism and signaling . Moreover , it will be important to examine the role of this pathway in additional models of lung injury , such as ischemia reperfusion and second organ reflow injury of the lung [98] or in models of inflammatory lung disease [99] , [100] . Other previous studies had implicated a functional role of HIF-dependent induction of glycolysis in inflammatory diseases . For example , previous studies had shown that HIF signaling is required for myeloid cell metabolism , and for their capability to function during conditions of inflammatory hypoxia [70] . These studies demonstrated that stabilization of HIF1A is essential for myeloid cell infiltration and activation in vivo via functioning as a essential regulator for their glycolytic capacity: when HIF1A is absent , the metabolic defect results in profound impairment of myeloid cell aggregation , motility , invasiveness , and bacterial killing [70] . However , during conditions of ALI—such as those that were used in the present studies—the lungs remain normoxic , and PMN-dependent HIF stabilization and HIF-elicited increases of PMN-inflammatory functions appear to be an unlikely scenario . In contrast , stretch-induced stabilization of HIF1A under normoxic conditions in epithelial cells appears to contribute to lung protection during ALI . Indeed , HIF-dependent increases in alveolar epithelial cell glycolysis , TCA cycle flux , and mitochondrial respiration allows for a metabolic adaptation that is critical to dampen pathologic lung inflammation during ventilator-induced lung injury . Consistent with the present studies implicating HIF in mitochondrial respiration during ALI , a previous study had demonstrated a functional role of HIF1A in mitochondrial respiration via altering mitochondrial complex IV ( COX4 ) subunit composition under hypoxic conditions [60] . Moreover , the current metabolic in vivo studies in mice with tissue-specific deletion of HIF1A in alveolar epithelia are consistent with previous studies from the 1970s that examined normoxic or hypoxic metabolism of cultured pulmonary epithelia . During air cultivation , alveolar epithelial cells had a high rate of aerobic and anaerobic glycolysis [101] . However , the authors observed concomitant increases of anaerobic and aerobic glycolysis under hypoxic condition [101] . These findings are consistent with the present studies showing that HIF1A is responsible for increases of glycolysis and lactate production and , at the same time , for oxidative glucose flux rates during ALI in vivo . Our findings utilizing liquid chromatography–tandem mass spectrometry analysis of metabolites following the infusion of 13C-glucose-labeled glucose in mice with tissue-specific deletion of alveolar epithelial HIF1A link these early observations with a functional role for HIF1A in controlling alveolar epithelial metabolism during ALI . Taken together , the present studies demonstrate a critical role for the stabilization of HIF1A during conditions of cyclic mechanical stretch in vitro , or during ALI in vivo , representing an endogenous protective mechanism . We conclude that targeting HIF-dependent metabolism of alveolar epithelia during ALI represents a powerful therapeutic strategy to dampen lung inflammation . These findings are also interesting from a translational perspective: a pharmacologic activator of HIF has recently been studied in the treatment of patients with renal anemia , and there were no safety concerns observed [49] . As such , treatment of ALI with HIF activators could potentially be considered in a clinical trial of ALI . Based on the specificity of our findings for pulmonary epithelial HIF1A , an inhaled treatment approach with less systemic concentrations of the compound may also be feasible .
Experimental protocols were approved by the Institutional Review Board at the University of Colorado Anschutz Medical Campus , Aurora , Colorado . They were in accordance with the U . S . Law on the Protection of Animals and the National Institutes of Health guidelines for use of live animals . Calu-3 human airway epithelial cells or human pulmonary epithelial A549 cells were cultured as described previously [23] , [40] . Primary pulmonary epithelial cells ( HPAEpiC ) ( ScienCell Research Laboratories , California ) were cultured according to the supplier's instructions . To study the consequences of cyclic mechanical stretch , we adopted a previously described in vitro model resembling mechanical ventilation by applying cyclic mechanical stretch . In short , Calu-3 , A549 , or HPAEpiC were plated on BioFlex culture plates-collagen type I ( BF-3001C; FlexCell International ) and allowed to attach and grow to 80% confluence . All cells were cultured in 4 ml media: Calu-3 cells were grown in Advanced MEM ( GIBCO ) , A549 cells were grown in DMEM-F12 ( GIBCO ) , both cell lines with 10% FBS , 0 . 02% L-Glutamine . Plates were then placed on a FlexCell FX-4000T Tension Plus System and stretched at percentage stretch indicated , 30% maximum , 0 . 7% stretch minimum , and sine wave 5 s on , 5 s off . Cells were collected at indicated time points from duplicate wells , flash-frozen , and stored at −80°C for further analysis . For control , cells were cultured under similar conditions at rest ( no cyclic mechanical stretch ) . Array data have been deposited at GEO ( accession number GSE 27128 ) ( http://www . ncbi . nlm . nih . gov/projects/geo/query/acc . cgi ? acc=GSE27128 ) . The transcriptional profile in Calu-3 cells subjected to 30% stretch for 24 h was assessed from total RNA ( isolated using Qiagen RNeasy kit ) using quantitative GeneChip arrays ( Human genome U133 Plus 2 . 0 Array ) . The integrity of RNA was assessed using an Agilent 2100 Bioanalyzer ( Agilent Technologies ) , and RNA concentration was determined using a NanoDrop ND-1000 spectrophotometer ( NanoDrop , Rockland , Delaware ) . Biotinylated cRNAs for hybridization to Affymetrix 3′-arrays were prepared from total RNA using the Affymetrix two-cycle target labeling assay with spike in controls ( Affymetrix Inc . , Santa Clara , California ) . Labeled-cRNA was fragmented and hybridized to Human Genome Arrays following the manufacturer's protocols . After hybridization and staining , the arrays were scanned using a GCS3000 with the latest 7G upgrade . Each array was subjected to visual inspection for gross abnormalities . Several other QC metrics were used to monitor hybridization efficiency and RNA integrity over the entire processing procedure . Raw image files were processed using Affymetrix GCOS 1 . 3 software to calculate individual probe cell intensity data and generate CEL data files . Using GCOS and the MAS 5 . 0 algorithm , intensity data were normalized per chip to a target intensity TGT value of 500 and expression data and present/absent calls for individual probe sets calculated . Quality control was performed by examining raw DAT image files for anomalies , confirming each GeneChip array had a background value of less than 100 , monitoring that the percentage present calls was appropriate for the cell type , and inspecting the poly ( A ) spike in controls , housekeeping genes , and hybridization controls to confirm labeling and hybridization consistency . According to our experimental setup , the arrays were normalized , grouped , and analyzed for differentially expressed transcripts based on different statistical tests . Different clustering algorithms allowed us to identify transcripts that show similar expression profiles . Using the “Ingenuity Pathway Analysis , ” we were able to identify biological mechanisms , pathways , and functions most relevant to our experimental dataset ( Table S1 , Figure S1 ) . All antibodies used were anti-HIF1A mouse monoclonal [H1alpha67] ( Abcam ) , anti-HIF1A mouse monoclonal [H1alpha67] ( Novus ) , anti-HIF2A rabbit polyclonal ( Novus ) , and anti-COX 4|2 mouse monoclonal ( R&D ) . COX4|2 immunoblotting was performed using mitochondrial protein fractions . Stable cell cultures with decreased HIF1A and HIF2A expression were generated by lentiviral-mediated shRNA expression . pLKO . 1 lentiviral vectors targeting HIF1A had shRNA sequence of CCG GCC AGT TAT GAT TGT GAA GTT ACT CGA GTA ACT TCA CAA TCA TAA CTG GTT TTT ( TRCN 0000003809 ) , and HIF2A had a sequence CCG GCC ATG AGG AGA TTC GTG AGA ACT CGA GTT CTC ACG AAT CTC CTC ATG GTT TTT ( TRCN 0000003807 ) . For controls , nontargeting control shRNA ( SHC002;Sigma ) was used . HEK293T cells were co-transfected with pLK0 . 1 vectors and packaging plasmids to produce lentivirus . Filtered supernatants were used for infection of Calu-3 and cells were selected with puromycin ( 30 ug ml−1 ) for at least two passages before initiating stretch experiments . Total RNA was isolated from human lung epithelia or murine lung tissue and transcript levels were determined by real-time RT-PCR ( iCycler; Bio-Rad Laboratories Inc . ) [90] . Primers were Quantitect from Qiagen . To assess partial oxygen pressures ( pO2 ) and lactate from cell supernatants or to determine pulmonary gas exchange ( paO2 ) in mice from arterial blood obtained via cardiac puncture , samples were analyzed immediately after collection with the I-STAT Analyzer ( Abbott ) . All 1H-NMR spectra were obtained at the Bruker 500 MHz DRX NMR spectrometer using an inverse Bruker 5-mm TXI probe . High-resolution 1H- and 13C-NMR experiments were performed with the Bruker 500 MHz DRX spectrometer equipped with an inverse 5-mm TXI probe ( Bruker BioSpin , Fremont , California ) and 31P-NMR experiments with the 300 MHz Bruker Avance system with a 5-mm QNP probe . For proton NMR , a standard water presaturation pulse program was used for water suppression; spectra were obtained at 12 ppm spectral width , 32K data arrays , 64 scans with 90-degree pulses applied every 12 . 8 s . Trimethylsilyl propionic-2 , 2 , 3 , 3 , -d4 acid ( TSP , 0 . 5 mmol/L ) was used as an external standard for metabolite chemical shift assignment ( 0 ppm ) and quantification . 13C-NMR spectra with proton decoupling were recorded using the C3-lactate peak at 21 ppm as chemical shift reference ( spectral width was 150 ppm , 16K data arrays , 20K scans applied every 3 s ) . [3-13C] lactate satellite peak ( at 1 . 23 ppm ) from 1H-NMR spectra served as an internal standard for 13C-NMR spectra ( at 21 ppm ) for calculation of 13C-enrichment of glucose and glucose metabolites . 31P-NMR spectra were obtained using the spectral width of 50 ppm and 16K data arrays , with 6K–10K scans being applied every 3 . 5 s . Before 31P-NMR spectra were recorded , EDTA ( 100 mmol/L ) was added to each PCA extract to complex divalent cations . Methylene diphosphonic acid ( 2 mmol/L ) was used as an external standard for chemical shift references ( 18 . 6 ppm ) and for metabolite quantification in 31P-NMR . For nuclear magnetic resonance ( NMR ) analysis of metabolites , cells were shock frozen immediately after stretch . Rate of glycolysis or TCA cycle flux is given as ratio of incorporated 13C intermediates of glycolysis compared to 13C glucose levels . All data were processed using the Bruker WINNMR program . All NMR experiments were performed at the Metabolomics NMR University of Colorado Cancer Center Core . Succinate Dehydrogenase ( SDH , Complex II Enzyme Activity Microplate Assay Kit , Abcam ) or COX4 ( Complex IV Enzyme Activity Microplate Assay Kit , Abcam ) or Pyruvate Dehydrogenase ( Pyruvate dehydrogenase ( PDH ) Enzyme Activity Microplate Assay Kit ) activities were determined from mitochondrial extracts using enzyme activity assays manufactured by Abcam . A549 cells were either grown on inserts or in 60-mm Petri dishes . SMARTpool siRNA targeting SUCLG was synthesized by Dharmacon ( Lafayette , Colorado ) . SMART pool reagents combine four SMART selection-designed siRNAs into a single pool , resulting in even greater probability that the SMARTpool reagent will reduce target mRNA to low levels . In addition , to further increase transfection efficiency and off-target effects , siRNA was selected by Dharmacon according ON-TARGETplus criteria . As control , siRNA ( Dharmacon , Lafayette , Colorado ) with at least four mismatches to any human , mouse , or rat gene was used . A549 cell loading was accomplished using standard conditions of DharmaFECT ( Dharmacon , Lafayette , Colorado ) , when cells had reached 40–60% confluence . After 48 h of loading , RNA/Protein was isolated as described before [102] . Octyl-α-ketoglutarate as a stable , cell-permeable form of α-ketoglutarate that accumulates rapidly and preferentially in cells with a dysfunctional TCA cycle was used at 1 mM . To inhibit MAP kinases in vitro , AMG 548 [Tocris , potent and selective inhibitor of p38 ( at 10 µM inhibits p38α , p38β , p38γ , and p38δ ) ] , FR 180204 [Tocris , selective ERK inhibitor ( at 1 µM inhibits ERK2 and ERK1 ) ] , and BI 78D3 [Tocris , competitive c-Jun N-terminal kinase ( JNK ) inhibitor ( 500 nM ) ] was used . For TNF-α neutralizing , D1B4 ( Cell Signaling ) Rabbit mAb was used . For IL-6 , anti-IL-6 antibody ( ab6672 ) from abcam was used . The snap-frozen lungs were thawed , weighed , and transferred to different tubes on ice containing 1 ml of Tissue Protein Extraction Reagent ( T-PER; Pierce Biotechnology ) . Cells or lung tissues were homogenized at 4°C . Cell or lung homogenates were centrifuged at 9 , 000 g for 10 min at 4°C . Supernatants were transferred to clean microcentrifuge tubes , frozen on dry ice , and thawed on ice . Total protein concentrations in the lung tissue homogenates were determined using a bicinchoninic acid kit ( Pierce Biotechnology ) . IL-6 ( R&D Systems ) , IL-8 ( R&D Systems ) , KC ( R&D Systems ) , TNF-α ( R&D Systems ) , and MPO ( Hycult Biotech ) levels were evaluated in lung tissue homogenates using a mouse ELISA kit according to the user's manual . ATP from cells or tissue was determined using ATP Bioluminescence Assay Kit CLS II for ATP Measurement from Roche . Intracellular hydrogen peroxide levels were measured using an Amplex Red Hydrogen Peroxide Assay kit ( Molecular Probes/Invitrogen ) according to the manufacturer's instructions . In brief , total cell lysates and supernatants were harvested 24 h after stretch in the FlexCell FX-4000T Tension Plus System . Lung tissue exposed to VILI was homogenized in Tissue Protein Extraction Reagent ( T-PER; Pierce Biotechnology ) and spun down at 13 , 000 g at 4°C for 10 min . Supernatants were collected and used in the assay . Bronchoalveolar lavages from VILI exposed animals were collected in 1 mL of saline . On a 96-well microplate , samples were added and reactions were initiated immediately by adding Amplex Red reaction mixture . Fluorescence was measured on Synergy 2 Multi-Mode Microplate reader ( BioTek ) in excitation range of 540/25 and emission detection of 620/40 . Fluorescence levels were normalized to the protein concentration . Wild-type mice ( BL6C57 ) , Hif1a reporter mice ( FVB . 129S6-Gt ( ROSA ) 26Sortm2 ( HIF1A/luc ) Kael/J ) [63] , Hif1af/f ( B6 . 129-Hif1atm3Rsjo/J ) , ActinCre+ ( B6 . Cg-Tg ( CAG-cre/Esr1* ) 5Amc/J ) [103] , CadherinCre+ ( B6 . Cg-Tg ( Cdh5-cre ) 7Mlia/J ) [104] , SPC-rtA ( B6 . Cg-Tg ( SFTPC-rtTA ) 5Jaw/J ) [105] , and Tet-O-Cre ( B6 . Cg-Tg ( tetO-cre ) 1Jaw/J ) [106] were purchased from Jackson laboratories . Cre exclusively expressed in the conducting airway ( ClaraCellCre+ ) were obtained from Thomas Mariani [107] . To obtain specific Hif1a−/− mice , Hif1af/f mice were crossed with the appropriate Cre mouse . Whole body Hif1a−/− mice knockout were achieved by a 5-d treatment of tamoxifen ( 1 mg/d ) i . p . For Hif1a tissue-specific knockout in the alveolar epithelium , triple transgenic mice ( Hif1af/f SPC-rtA Tet-O-Cre ) were induced by Doxycyclin therapy over 5 d i . p and p . o . as described [106] . All animal protocols were in accordance with the guidelines of the National Institute for Health for the use of laboratory animals and approved by the Institutional Animal Care and Use Committee of the University of Colorado . Ventilator-induced lung injury was induced as described previously [40] . To obtain BAL fluid , the tracheal tube was disconnected from the mechanical ventilator and the lungs were lavaged 3 times with 0 . 5 ml of PBS . All removed fluid was centrifuged immediately , and the supernatant was aliquoted for measurement of the albumin concentration . Mice were given a single i . p . injection of a mixture of luciferin ( 50 mg/kg ) in sterile water . Following normal ventilation ( 15 mbar , 3 h ) or VILI ( 45 mbar , 3 h ) , lungs were immediately excised and placed in a light-tight chamber equipped with a charge-coupled device IVIS imaging camera ( Xenogen , Alameda , California ) . Photons were collected for a period of 5–20 s , and images were obtained by using living image software ( Xenogen ) and igor image analysis software ( WaveMatrics , Lake Oswego , Oregon ) . Expression of the Hif1a reporter genes was assayed using the tissue homogenates in Tper Tissue Protein Extraction Reagent ( Pierce ) . The homogenates were centrifuged for 30 min at 4 , 900×g at 4°C . The luciferase gene expression was measured by using the Dual-Luciferase Reporter Assay System from Promega according to the manufacturer's instructions using a Biotek Synergy 2 Multimode Microplate Reader . Rotenone ( Tocris ) , 2-Deoxy-D-glucose ( 2-DG ) , Dimethyloxalylglycine ( DMOG ) , and Echinomycin were purchased from Sigma-Aldrich , USA and used at the concentrations as indicated . Following 4 h of VILI , mice were euthanized by pentobarbital overdose . Lungs were lavaged with sterile PBS , then perfused with PBS via the right ventricle . As previously described , dispase ( BD , USA ) was instilled followed by a low-melting point agarose plug . Lungs were removed intact and incubated at 37°C for 30 min . Tissue was dissociated manually and cells progressively filtered ( 70 µm/40 µm , Fisher , USA ) . Cells were treated with FcR blocking mAB ( 24G2 ) and labeled with anti-EpCAM biotin ( eBioscience , USA ) and anti-biotin Dylight 633 . EpCAM positive cells were positively selected using streptavidin magnetic microbeads ( Miltenyi , USA ) . Purified cells were analyzed for EpCAM expression using a LSRII flow cytometer ( BD , USA ) and FlowJo v8 . 8 . 4 software ( TreeStar , USA ) [108] . Following ventilation at the settings indicated in the figure legends , the mice were euthanized and lungs were fixed by instillation of 10% formaldehyde solution via the tracheal cannula at a pressure of 20 mbar . Lungs were then embedded in paraffin and stained with hematoxylin and eosin . Two random tissue sections from four different lungs in each group were examined by a pathologist who was blinded to the genetic background and treatment of the mice . For each subject , a 5-point scale was applied: 0 , minimal damage; 1 to >2 , mild damage; 2 to >3 , moderate damage; 3 to >4 , severe damage; and 4+ , maximal damage . Points were added together and are expressed as median ± range ( n = 4 ) . Hif1af/f SPC-rtA Tet-O-Cre mice or littermate controls ( induced SPC-rtA Tet-O-Cre mice ) matched in age , weight , and gender were exposed to 2 h of VILI [45 mbar] . 13C-glucose was administered 30 min prior to the onset of VILI via intraperitoneal injection . Lung tissue was snap-frozen with clamps pre-cooled to the temperature of liquid nitrogen . Seven metabolites ( glucose , lactate , pyruvate , malate , fructose-1 , 6-bisphosphate , glutamate , and creatine ) were measured by ultrahigh performance liquid chromatography–mass spectrometry ( UPLC-MS ) using a Waters Acquity ultrahigh-performance liquid chromatograph coupled to a Waters Synapt HDMS quadrupole time-of-flight mass spectrometer , which was operated with an atmospheric pressure electrospray ionization ( ESI ) source as described previously . [32] Data were compared by two-factor ANOVA with Bonferroni's posttest or by Student's t test where appropriate . Values are expressed as mean ± s . d . from 3–6 animals per condition . For analysis of changes in transcript , a one-way ANOVA was carried out and multiple comparisons between control and treatment groups were made using the Dunnett posttest . The Mantel Cox test was used for analysis of survival curves . Data are expressed as mean ± s . d . ; p<0 . 05 was considered statistically significant . For all statistical analyses , GraphPad Prism 5 . 0 software for Windows XP was used . The authors had full access to and take full responsibility for the integrity of the data . All authors have read and agree to the manuscript as written . | Acute lung injury is a devastating lung disease caused by injuries or acute infections to the lung . In patients it manifests itself as acute respiratory distress syndrome . Severe pulmonary edema and uncontrolled lung inflammation are typical symptoms of acute lung injury , which make it hard for patients to breath efficiently . In the clinical course of the disease , patients require mechanical ventilation to support their lung function and to provide sufficient oxygen levels to vital organs such as the brain , the heart , or the kidneys . We hypothesized that stretch conditions—such as those that occur during mechanical ventilation—result in transcriptional adaptation of alveolar epithelial cells—the innermost lining of the lungs . In this study we identified an unexpected involvement of the transcription factor hypoxia-inducible factor HIF1A in lung protection . We observed that during acute lung injury , stabilization of HIF1A is mediated by increased levels of succinate , an intermediate product of the citrate cycle . Interestingly , we show that HIF1A in alveolar epithelia functions to induce a transcriptional program that improves the efficiency of carbohydrate metabolism by injured lungs , thereby helping to adapt to the injurious conditions of mechanical stretch and to reduce lung inflammation . These preclinical findings highlight the potential for pharmacological HIF1A stabilization in the treatment of acute lung injury . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | HIF1A Reduces Acute Lung Injury by Optimizing Carbohydrate Metabolism in the Alveolar Epithelium |
The Pseudomonas syringae acetyltransferase HopZ1a is delivered into host cells by the type III secretion system to promote bacterial growth . However , in the model plant host Arabidopsis thaliana , HopZ1a activity results in an effector-triggered immune response ( ETI ) that limits bacterial proliferation . HopZ1a-triggered immunity requires the nucleotide-binding , leucine-rich repeat domain ( NLR ) protein , ZAR1 , and the pseudokinase , ZED1 . Here we demonstrate that HopZ1a can acetylate members of a family of ‘receptor-like cytoplasmic kinases’ ( RLCK family VII; also known as PBS1-like kinases , or PBLs ) and promote their interaction with ZED1 and ZAR1 to form a ZAR1-ZED1-PBL ternary complex . Interactions between ZED1 and PBL kinases are determined by the pseudokinase features of ZED1 , and mutants designed to restore ZED1 kinase motifs can ( 1 ) bind to PBLs , ( 2 ) recruit ZAR1 , and ( 3 ) trigger ZAR1-dependent immunity in planta , all independently of HopZ1a . A ZED1 mutant that mimics acetylation by HopZ1a also triggers immunity in planta , providing evidence that effector-induced perturbations of ZED1 also activate ZAR1 . Overall , our results suggest that interactions between these two RLCK families are promoted by perturbations of structural features that distinguish active from inactive kinase domain conformations . We propose that effector-induced interactions between ZED1/ZRK pseudokinases ( RLCK family XII ) and PBL kinases ( RLCK family VII ) provide a sensitive mechanism for detecting perturbations of either kinase family to activate ZAR1-mediated ETI .
Both plants and animals use cell membrane-spanning receptor kinases to sense extracellular molecular patterns produced by invading pathogens . While agonists of these receptors can trigger signal cascades that result in protective host immune responses ( pattern-triggered immunity , or PTI ) , Gram-negative bacteria equipped with a type III secretion system ( T3SS ) can dampen such basal immune responses by delivery of effector proteins directly into host cells through the needle-like T3SS pilus . In plant cells , such T3SS-delivered effectors ( T3Es ) can be recognized by nucleotide-binding , leucine-rich repeat ( NLR ) proteins , resulting in ‘effector-triggered immunity’ ( ETI ) , which is often accompanied by a hypersensitive cell death response ( HR ) [1] . NLRs are members of the STAND ( signal transduction ATPases with numerous domains ) class of P-loop NTPases that also includes the NOD-like receptors of animal cells [2] . These proteins share common central nucleotide-binding and carboxy-terminal LRR domains , but are preceded by distinct classes of amino-terminal domains–plant NLRs have either ‘coiled-coil’ ( CC ) or ‘Toll/Interleukin-1 Receptor’ ( TIR ) domains at their amino-termini [2–5] . Plant NLRs become activated either by direct sensing of effectors , or by indirect sensing of effector-modified substrates [6] . Such effector-modified substrates can be virulence targets whose functions limit bacterial growth , or mimics of these virulence targets referred to as ‘decoys’ [7] . For example , the Arabidopsis kinase PBS1 ( AvrPphB-Susceptible 1 ) , is cleaved by the P . syringae T3E HopAR1 ( previously known as AvrPphB ) , resulting in activation of the Arabidopsis NLR RPS5 [8 , 9] . There are 45 PBS1-like kinases ( PBL kinases , or PBLs ) in Arabidopsis that together with PBS1 comprise receptor-like cytoplasmic kinase ( RLCK ) subfamily VII . At least eight of these additional PBLs are also sensitive to cleavage by HopAR1 , including BIK1 and PBL1 , which are both involved in plant immunity [10] . PBS1 orthologues have been reported in diverse angiosperms such as wheat [11] and barley [12] but are ‘guarded’ by phylogenetically distinct NLRs [12] . This conservation ( and the convergent evolution of distinct NLR guards ) likely reflects important immunity functions provided by PBS1 and other PBLs that make them attractive virulence targets for P . syringae and other plant pathogens . The Arabidopsis NLR ZAR1 ( HopZ-Activated Resistance 1 ) belongs to the coiled-coil class of plant NLRs and is required for ETI against at least three distinct T3Es: HopZ1a and HopF2a from Pseudomonas syringae , and AvrAC from Xanthomonas campestris [13–15] . In addition to ZAR1 , recognition of all three T3Es requires members of RLCK family XII . Members of this RLCK family lack at least one of several well-established and highly conserved kinase motifs , and are therefore considered to encode pseudokinases [16–18] . Recognition of the acetyltransferase HopZ1a requires the RLCK XII pseudokinase ZED1 ( HopZ ETI-Deficient 1 ) , which is encoded in a gene cluster with seven additional ZED1-related pseudokinases ( ZRKs ) . ZRK1/RKS1 and ZRK3 pseudokinases are required for recognition of AvrAC and HopF2a , respectively [14–16] . ZED1 , ZRK1/RKS1 and ZRK3 have all been shown to interact with ZAR1 in planta , although different roles in T3E recognition have been proposed [15 , 17 , 18] . ZED1 is acetylated by HopZ1a , but a ZED1 knockout does not appear to influence the virulence of P . syringae [16] . As such , it has been proposed that ZED1 is a decoy substrate monitored by ZAR1 to guard against the acetylation of other ( kinase ) substrates of HopZ1a [16–18] . In contrast , ZRK1/RKS1 functions as an adaptor for ZAR1 by recruiting PBL2 ( PBS1-like kinase 2 ) proteins uridylated by AvrAC to elicit ETI [14 , 19 , 20] . Recent structural studies indicate that binding of uridylated PBL2 to ZAR1-bound ZRK1/RKS1 promotes ADP/ATP exchange by ZAR1 and results in formation of a pentameric ‘resistosome’ similar to the inflammasomes and apoptosomes formed by mammalian NLRs [19 , 20] . ZRK3 is also thought to act as a ZAR1 adaptor for an as-yet-unidentified kinase that is ADP-ribosylated by the P . syringae T3E HopF2a [15] . Overall , the ZED1/ZRK family of pseudokinases have expanded the T3E recognition capacity of ZAR1 by directly sensing T3E modifications and/or serving as adaptors that bridge interactions between the NLR and T3E-modfied plant kinases . Highly-conserved kinase motifs were initially identified by comparative sequence analysis [21 , 22] , but since then , determination of the molecular structures of an ever-increasing number of diverse kinase domains in the presence or absence of nucleotides , metal ions , substrate peptides , and/or pharmacological inhibitors has provided rational explanations for the functional importance of these motifs [23] . The kinase domain nucleotide-binding pocket is formed by a cleft between a mostly β-stranded amino-terminal lobe and an α-helical carboxy-terminal lobe . In the amino-terminal lobe , a glycine-rich motif in the loop connecting strands β1 and β2 ( ‘G-loop’ ) acts as a lid through backbone interactions with the β- and γ-phosphates of bound ATP [24] . Appropriate positioning of ATP phosphates also requires metal ions ( usually Mg2+ or Mn2+ ) , a lysine from strand β3 that is stabilized by a salt-bridge with a glutamate from helix αC , and an aspartate from the ‘DFG motif’ . The DFG motif also defines the beginning of the ‘activation loop’ , a conformationally flexible ( and often disordered ) region of 20–35 amino acids that typically contains serine and/or threonine residues that can be phosphorylated ( by autophosphorylation or by upstream kinases ) . Phosphorylation of the activation loop can stabilize an open conformation ( i . e . , a ‘DFG-in’ orientation , and a fully-assembled hydrophobic ‘regulatory spine’ ) that is associated with active kinases [25] . This open conformation allows substrate peptides to approach the γ-phosphate ( when appropriately-positioned , as described above ) for phospho-transfer by the aspartate from an ‘HRD motif’ in the catalytic segment . These dynamic structural features play essential roles in regulating kinase activity as well as modulating their interactions with other proteins [23 , 26] . ZED1 and the ZRKs have degeneracies in nearly all of these canonical kinase motifs , and as such they may all represent pseudokinases that perform T3E sensing functions similar to those proposed for ZED1 , ZRK1/RKS1 and ZRK3 [14–16 , 18] . Early surveys of the complete kinase domain complements from organisms including human , mouse , Caenorhabditis elegans , Dictyostelium and Arabidopsis found that pseudokinases like ZED1 account for ~10% of the kinase domains in higher eukaryotes and 20% of Arabidopsis RLCKs [27 , 28] . Plant pseudokinases are emerging as important mediators of development and immunity [29–31] . Rather than mere non-functional kinases resulting from relaxed ( or absent ) selective constraints , there is instead a growing body of evidence indicating that pseudokinases are important signaling molecules in spite of their predicted lack of catalytic activity , often functioning as adaptors or scaffolds that modulate the activity of true kinase partners [27 , 30 , 32 , 33] . In this report we show that the P . syringae T3E HopZ1a induces interactions between ZED1 and PBL kinases and promotes the formation of a ZAR1-ZED1-PBL ternary complex . The pseudokinase features of ZED1 mediate its conditional interactions with PBL kinases , recruitment of ZAR1 , and importantly , its regulation of ZAR1-mediated immunity . Overall , our results support the hypothesis that ZED1-PBL ( and more generally ZRK-PBL ) pseudokinase-kinase interplay provides a sensor for perturbations of host kinases by pathogen-delivered effector proteins .
In order to assess whether PBL kinases are plausible targets of HopZ1a , we tested whether HopZ1a can interact with members of this family . Using a yeast two-hybrid ( Y2H ) screen ( S1 Fig; interaction scheme A ) with all 46 PBL kinases as prey and HopZ1a as bait ( wild-type or a C216A catalytic site mutant ) we found strong binding between nine PBL kinases ( PBL21 , PBL27 , PBL8 , PBL2 , PBL3 , PBL4 , PBL18 , PBL15 and PBL13 ) and HopZ1aC216A , but no interactions were observed with the wild-type effector , HopZ1awt ( Fig 1A ) . We hypothesized that failure to observe binding between HopZ1awt and PBLs may reflect efficient enzymatic activity and rapid substrate turnover ( i . e . , ‘catch and release’ catalysis ) . Such interactions would be stabilized in the case of HopZ1aC216A due to frustrated enzymatic activity . To test whether HopZ1a can acetylate PBL kinases , we co-expressed FLAG-tagged derivatives of both HopZ1a and PBS1 in yeast , prepared anti-FLAG immunoprecipitates from yeast cell extracts and subjected these to liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) analysis . Although PBS1 was a weaker HopZ1a interactor , we chose to test its acetylation by HopZ1a based on its established role in ETI . We observed three PBS1 peptides with mass increases of 42 Da ( corresponding to addition of an acetyl group ) when PBS1 was co-expressed with HopZ1awt but not with HopZ1aC216A , indicating HopZ1a-mediated acetylation . Fragmentation analysis of these three peptides established that the specific sites acetylated were T32 , S244 , and S405 ( Fig 1B; S3 Fig; S1 File ) . These results support our hypothesis that HopZ1a can acetylate PBL kinases . Given the previously reported finding that AvrAC promotes interactions between PBL2 and ZRK1/RKS1 [14] , we sought to develop an assay to investigate whether HopZ1a can promote interactions between ZED1 and PBL kinases . We devised a yeast three-hybrid ( Y3H ) assay wherein interactions between a ZED1/ZRK bait protein and PBL preys were assessed in the absence and presence of single-copy chromosomal integrations of a T3E of interest ( S1 Fig , interaction schemes A , B ) . As proof-of-principle we integrated Xanthomonas AvrAC at the yeast ho locus and investigated its influence on ZRK1/RKS1 interactions with PBL kinases . In the absence of AvrAC expression , ZRK1/RKS1 exhibited high-affinity binding to several PBLs ( PBL21 , PBL17 , PBL8 and PBL15 ) and moderate binding to several others , including PBL2 ( S4 Fig ) . Binding between ZRK1/RKS1 and PBL2 was enhanced by the presence of AvrAC ( as was binding between ZRK1/RKS1 and PBL3 / PBL29; S4 Fig ) , indicating that our Y3H system can be used to examine modulation of ZRK/PBL interactions by T3Es . We then integrated hopZ1a ( wild-type or C216A alleles ) at the yeast ho locus and examined its influence on ZED1-PBL interactions . Only weak ZED1-PBL interactions were observed in the absence of effector ( Fig 2B; Fig 3B , column 1 ) or when HopZ1aC216A was expressed ( Fig 2A; Fig 3B , column 3 ) . Notably , expression of wild-type HopZ1a resulted in enhanced interactions between ZED1 and 11 PBL kinases ( PBL21 , PBL22 , PBL5 , PBS1 , PBL27 , PBL17 , PBL8 , PBL4 , PBL9 , PBL15 , and PBL13 ) , with the strongest interactions occurring between ZED1 and PBL5 / PBL27 / PBL17 / PBL8 / PBL15 ( Fig 2A; Fig 3B , column 2 ) . These results indicate that HopZ1a activity can promote interactions between the ZED1 pseudokinase and PBL kinases . In order to assess whether stimulation of ZED1-PBL binding was a property specific to HopZ1a , we also created yeast strains with chromosomal integrations of genes encoding both closely-related ( HopZ1b; 65% amino acid identity to HopZ1a over 349 non-gapped sites ) and more distantly-related HopZ family members ( HopZ2 and HopZ3; 26% and 23% identical to HopZ1a over 335 and 306 non-gapped sites , respectively ) . Interactions between ZED1 and PBLs were assessed in the presence of each of these HopZ family effectors or with the unrelated effector AvrAC ( Fig 2B; S5 Fig; Fig 3B , columns 2–7 ) . Like HopZ1a , HopZ1b promoted strong interactions between ZED1 and PBL27 / PBL17 / PBL8 , but in contrast , interactions between ZED1 and PBL5 / PBL15 / PBL21 were weaker with HopZ1b than with HopZ1a , and HopZ1b did not promote interactions between ZED1 and PBL22 / PBS1 / PBL4 / PBL9 ( Fig 2 and Fig 3B , columns 2 , 4 ) . Notably , the interactions between ZED1 and PBLs were unchanged by the expression of HopZ2 , HopZ3 , or AvrAC ( Fig 3B , columns 5–7; S5 Fig , panels A , B ) although integrated effectors are expressed to comparable levels ( S5 Fig , panel C ) , demonstrating that promotion of ZED1/PBL interactions is specific to HopZ1a ( and to a lesser extent , HopZ1b ) . The kinase domains of ZED1 and the ZRKs have diverged significantly from those of true kinases , with degeneracies in one or more of the established kinase motifs ( S6 Fig ) [16] . They also generally lack kinase activity under in vitro conditions that are sufficient for canonical kinases to phosphorylate model substrates [16 , 34] . To determine the importance of ZED1 pseudokinase features for ( HopZ1a-induced ) binding to PBLs , we identified conspicuous ZED1 sequence elements based on their divergence from canonical kinase motifs and designed mutations intended to ‘restore’ these degenerate motifs . We made a series of ZED1 variants with various combinations of mutations at five different sites ( Fig 3A ) , described below . We tested the PBL-binding activity of each of these rationally-designed ZED1 mutations ( individually and in various combinations ) both in the absence and presence of HopZ1a expression using Y2H and Y3H assays . We examined the binding of ZED1 mutants to a restricted subset of 15 PBLs that includes ten representatives with robust HopZ1a-induced binding affinity to wild-type ZED1 as well as five with relatively weak HopZ1a-dependent ZED1-binding activity ( PBL24 , PBL36 , PBL18 , PBL12 , PBL11 ) . Our mutational analysis found that ZED1N173D ( HRD restoration ) , ZED1W193G ( DFG restoration ) and ZED1V212T ( activation loop restoration ) all gained affinity for several PBL kinases in the absence of HopZ1a , with ZED1N173D and ZED1V212T showing the strongest interactions ( S8 Fig and S9 Fig; Fig 3B , compare columns 12 , 13 , 18 , 19 with columns 1 , 10 , 17 ) . These mutations induced HopZ1a-independent ZED1 binding to PBL kinases showing both strong ( PBL21 , PBL27 , PBL15 , PBL17 , PBL4 ) and weak ( PBL18 ) HopZ1a-dependent binding . HopZ1a-independent binding between PBS1 and ZED1 was only observed in the context of the ZED1N173D mutation ( i . e . , HRD restoration; ZED1N173D and ZED1D92E N173D; S8 Fig and S9 Fig; Fig 3B , columns 12 , 14 , 18 ) . The triple mutant ZED1‘3xG’ ( G-loop restoration ) disrupted HopZ1a-induced PBL binding , and only moderate HopZ1a-independent binding was observed ( S9 Fig , panel B; Fig 3B , column 21 ) . Combining ZED1 mutations differentially influenced subsets of PBL interactions . For example , combining N173D with W193G ( ZED1N173D W193G; S8 Fig , panel B; Fig 3B , column 15 ) or with D92E and W193G ( ZED1‘EDG’; S8 Fig , panel B; Fig 3B , column 16 ) resulted in a loss of PBS1 binding ( observed with ZED1N173D and ZED1D92E N173D; S8 Fig; Fig 3B , columns 12 , 14 , 18 ) but increased ZED1 binding affinity for PBL21 and PBL24 ( not observed with any of the three individual mutations; S8 Fig , panel A and S9 Fig , panel A; Fig 3B , columns 11 , 12 , 13 , 18 ) . When we combined N173D with the restored G-loop ( ZED1‘3xG’ N173D ) , all binding to PBLs was abolished whether HopZ1a was present or not ( S9 Fig , panel B; Fig 3B , column 22 ) . However , a subset of interactions was restored in the context of a phospho-accepting activation loop in ZED1 ( ZED1‘3xG’ V212T and ZED1‘3xG’ ‘DT’ , which adds both N173D and V212T; S9 Fig , panel B; Fig 3B , columns 23 , 24 ) . These results suggest that the effects of ZED1 mutations differentially affect binding affinity for distinct PBLs , indicating that PBLs differ in their abilities to interact with ZED1 through perturbation of ( pseudo ) kinase features either through mutation or by T3E modification . Based on the results of our ZED1 mutagenesis , we hypothesized that acetylation of one or more sites on ZED1 might influence ZED1-PBL binding by shifting between ‘pseudokinase-like’ and ‘kinase-like’ states . We have previously described acetylation of ZED1 by HopZ1a at T125 and T177 [16] , and subsequent experiments have indicated that S84 and T87 can also be acetylated by HopZ1a ( Fig 1C; S3 Fig; S1 File ) . We therefore made glutamine substitutions as potential mimics of acetylated serine and/or threonine residues based on similarities in sidechain length and branching structure ( Fig 4A ) . We tested the PBL-binding capacity of three ZED1 glutamine substitution mutants: ZED1T177Q , which is proximal to the ( non-catalytic ) active site residue N173 , as well as ZED1S84Q and ZED1T87Q , which are both predicted to be part of a solvent-exposed surface of helix αC ( S7 Fig , panel A; S10 Fig , panel A ) . None of the ZED1 glutamine substitution mutants ( or any of the three possible pairwise combinations ) resulted in HopZ1a-independent binding to any of the PBLs tested ( S10 Fig , panel B ) . In contrast , each of the mutants bearing the T177Q substitution ( ZED1T177Q , ZED1S84Q T177Q , ZED1T87Q T177Q ) lost all capacity for HopZ1a-dependent binding ( S10 Fig , panel B ) , indicating that ZED1T177Q represents a loss-of-function allele . We also tested whether the PBS1 acetylation sites described above ( Fig 1B; S3 Fig; S1 File ) are important for its interaction with ZED1 . Similar to our ZED1 findings , we found that none of the PBS1 glutamine substitutions ( alone or in combination ) enable HopZ1a-independent ZED1-binding activity ( S10 Fig , panel C ) . Furthermore , as for ZED1 T177 , each of the PBS1 mutants bearing a glutamine substitution of S244 lost capacity for HopZ1a-dependent binding to ZED1 ( S10 Fig , panel C ) . The combined insights from our mutagenesis of ZED1 and PBS1 acetylation sites suggest that glutamine does not mimic acetylated serine/threonine residues and may instead act as a loss-of-function mutation by preventing acetylation and/or phosphorylation . We also considered isoleucine substitutions as possible mimics of acetylated serine and/or threonine residues based on similarities in hydrophobicity and sidechain branching ( Fig 4A ) . We therefore tested glutamine and isoleucine ( possible acetyl-mimics ) as well as alanine ( loss-of-function ) substitutions in PBL15 , which was selected as a representative PBL based on its strong HopZ1a-dependent binding to ZED1 ( stronger than PBS1; Fig 2A ) . We mutated PBL15 S260 , an activation loop residue equivalent to PBS1 S244 ( which is acetylated by HopZ1a; Fig 1B ) . Of these three substitutions of PBL15 S260 , only PBL15S260I demonstrated ZED1 binding in the absence of HopZ1a activity ( Fig 4B ) . The strength of this interaction was comparable to the ZED1-PBL15wt binding induced by HopZ1a , and HopZ1a-induced ZED1 binding was abolished by the alanine and glutamine substitutions of S260 . Overall , these results suggest that isoleucine substitutions can mimic acetylation of serine ( and possibly threonine ) residues and that acetylation of the activation loops of PBL kinases by HopZ1a is sufficient to promote their interactions with ZED1 . Having established that HopZ1a is able to stimulate protein-protein interactions between ZED1 and PBLs , we wished to investigate whether HopZ1a can also promote formation of a ternary complex between PBLs , ZED1 , and ZAR1 . For this purpose , we developed a yeast four-hybrid system where ZAR1 baits and PBL kinase preys ( 11 with HopZ1a-dependent ZED1 binding ) were expressed in the presence or absence of chromosomally-integrated hopZ1a ( wild-type or the C216A catalytic site mutant ) and ZED1 ( S1 Fig , interaction scheme C ) . Since modulation of inter-domain contacts is known to be important for the activation of plant NLRs [19 , 20 , 37 , 38] , we created various domain truncations of ZAR1 ( panel A in both Fig 5 and S11 Fig ) in order to investigate domain-specific ZAR1 binding interactions with ZED1/PBLs that might be modulated by HopZ1a activity . Full-length ZAR1 bait ( ZAR1wt ) interacted with prey PBL kinases ( PBL5 , and weaker interactions with PBL17 , 4 , 18 , 15 ) when co-expressed with both ZED1 and HopZ1awt , but not when co-expressed with ZED1 and HopZ1aC216A ( Fig 5B , compare columns 4 and 6 ) . These interactions were dependent on ZED1 expression ( Fig 5B , compare columns 3 and 4 ) and also required the ZAR1 LRR domain , since no interactions were observed with the ZAR1ΔLRR construct ( Fig 5B , compare columns 4 and 12 ) . Consistent with this finding , the isolated LRR domain displayed weak HopZ1awt/ZED1-dependent PBL interactions , but the CC and NB domains did not ( S11 Fig , panel B , columns 4 , 10 , 16 ) . Interestingly , ZAR1ΔCC showed stronger HopZ1a/ZED1-dependent interactions with 9 of the 11 PBLs ( Fig 5B , compare columns 4 and 20 ) , suggesting that the ZAR1 coiled-coil domain negatively regulates ZAR1-ZED1-PBL complex formation . Although we have demonstrated above that ZED1 mutants with restored kinase motifs gain HopZ1a-independent binding affinity for PBL kinases , it is possible that these interactions are formed by a binding interface that differs from that required for HopZ1a-dependent PBL-ZED1-ZAR1 interactions . To address this , we integrated genes encoding ZED1wt or ZED1‘DT’ ( ZEDN173D V212T ) at the yeast ho locus to create a HopZ1a-independent Y3H system in which to test possible PBL-ZED1-ZAR1 interactions . Expression of ZED1‘DT’ allowed HopZ1a-independent interactions between ZAR1wt and PBL17 / PBL9 , whereas ZED1wt did not ( Fig 5B , compare columns 7 and 8 ) . Moreover , in the presence of ZED1‘DT’ ( but not ZED1wt ) , ZAR1ΔCC interacts more robustly and with more PBLs ( eight of the eleven tested; Fig 5B , compare columns 7 and 8 with 23 and 24 ) , indicating that the coiled coil domain negatively regulates formation of the ZAR1-ZED1-PBL ternary complex in HopZ1a-independent contexts as well . Importantly , ZAR1ΔLRR did not interact with PBL kinases in the presence of ZED1‘DT’ indicating that the leucine-rich repeat domain is required for both HopZ1a-dependent and HopZ1a-independent interactions . We hypothesized that ZED1/PBL interactions promote ZAR1-mediated immunity , similar to the ZRK1/RKS1-PBL2-ZAR1-mediated immunity triggered by AvrAC [14] . As such , mutant ZED1 alleles that allow HopZ1a-independent interactions with PBL kinases should also promote immunity in the absence of the effector . To test this , we transformed the Arabidopsis ecotype Col-0 with four dexamethasone-inducible , HA epitope-tagged ZED1 alleles ( ZED1wt , ZED1N173D , ZED1V212T , and ZED1‘DT’ ) , and examined the first generation ( T1 ) for a dexamethasone-induced tissue collapse similar to the HR observed in plants expressing the P . syringae effector AvrRpt2 [39] ( see column 2 in Fig 6A and 6B , S12A and S12B Fig and S13A and S13B Fig ) . We established expression profiles for at least 8 individual T1 transformants from each ZED1 allele ( panel C in Fig 6 , S12 Fig , S13 Fig ) prior to testing for dexamethasone-inducible cell death activity . These same plants ( with one leaf removed ) , along with untransformed Col-0 and zar1-1 mutant [13] plants as controls , were then sprayed with dexamethasone and observed visually for signs of dexamethasone-induced HR . As expected , untransformed wild-type Arabidopsis ( Col-0 ) and zar1-1 mutant plants were both unaffected by dexamethasone treatment ( columns 1 and 3 in Fig 6A and 6B , S12A and S12B Fig and S13A and S13B Fig ) . Arabidopsis plants expressing ZED1V212T and ZED1‘DT’ displayed tissue collapse similar to AvrRpt2-expressing plants 72 hours after dexamethasone treatment ( columns 4 and 5 in S13A and S13B Fig; columns 6 and 7 in Fig 6A and 6B ) , whereas plants expressing ZED1wt and ZED1N173D did not ( columns 4 , 5 in Fig 6A and 6B; columns 4–7 in S12A and S12B Fig ) . These results suggested that PBL-interacting ZED1 mutants can trigger HopZ1a-independent immunity and led us to investigate whether this response is ZAR1-dependent . Nicotiana benthamiana has two ZAR1 homologues ( NbZAR1 and NbZAR2 ) that are ~57% identical to Arabidopsis ZAR1 [40] . Despite lacking closely-related homologues of ZED1 and ZRKs ( S14 Fig ) , ZAR1-dependent immune responses triggered by HopZ1a can be recapitulated in N . benthamiana when complemented with Arabidopsis ZED1 [40] . Cell death resulting from an HR is observed as early as 8 hours following induction of transgene expression in N . benthamiana leaves co-transformed with ZED1 and HopZ1a , but not in leaves transformed with either ZED1 or HopZ1a alone [40] . Since BLASTP searches demonstrate that Nicotiana spp . do have highly-similar homologues of Arabidopsis PBLs ( S14 Fig ) , we hypothesized that our ZED1 mutants with HopZ1a-independent PBL binding may be sufficient to trigger ZAR1-dependent HR in N . benthamiana . We therefore tested the ability of ZED1 mutants described above to cause HR in N . benthamiana using transient transformations . As described previously [40] , macroscopic HR-like cell-death was apparent by 8–24 h post-induction of protein expression in N . benthamiana tissue co-transformed with ZED1wt and HopZ1a , but not in tissue transformed with either ZED1wt or HopZ1a alone ( S15 Fig ) . In contrast , ZED1‘DT’ and ZED1V212T were able to cause HR in N . benthamiana even in the absence of HopZ1a ( S15 Fig , panel A ) . Overexpression of wild-type ZED1 did not induce HR despite similar levels of expression as ZED1‘DT’ ( N173D V212T; S15 Fig , panel B ) . We also tested ZAR1-dependence of these effector-independent cell death phenotypes by subjecting N . benthamiana plants to virus-induced gene silencing ( VIGS ) [40–42] two weeks prior to transformation and induction of ZED1 expression . Although the HR-mediated cell death induced by ZED1‘DT’ , ZED1V212T and HopZ1a/ZED1wt was unaffected in plants experiencing silencing of the negative control GUS gene ( Fig 7A , left ) , the HR induced by all of these transformations was completely abolished in plants that received a ZAR1 silencing construct ( Fig 7A , right ) . To determine if pathogen-induced perturbations of ZED1 are sufficient to activate ETI , we tested whether mimicking HopZ1a acetylation of ZED1 T177 can also activate plant immunity . Given that an isoleucine substitution of a PBL acetylation site acted as a gain-of-function mutant with respect to ZED1 / PBL interactions in yeast ( Fig 4 ) , we tested the ability of ZED1T177I to induce HR in N . benthamiana . Like ZED1V212T , expression of ZED1T177I was also capable of inducing effector-independent HR ( Fig 7B; S15 Fig , panel C ) . Importantly , this HR was also dependent on ZAR1 , supporting our hypothesis that acetylation of ZED1 by the type III effector HopZ1a activates ZAR1-mediated immunity [14] .
One conspicuous pseudokinase feature of ZED1—even among ZRKs—is the degenerate kinase catalytic site ( N173 ) , wherein the catalytic aspartate ( acidic side chain ) typically present in active kinases is replaced by asparagine ( structurally-similar to aspartate , but with a basic side chain ) . We created the ZED1N173D mutant to test the consequences of this substitution both in Y2H assays and in an inducible expression system in planta . Although our yeast assays revealed that ZED1N173D gains effector-independent affinity for PBLs , we did not observe induction of cell death ( immunity ) following expression of ZED1N173D in Arabidopsis ( S12 Fig ) or N . benthamiana ( Fig 7A ) . However , a distinct substitution at this position ( N173S; which introduces a shorter , polar side chain ) was previously shown to display ZAR1-dependent developmental and immunity-related phenotypes that are conditionally observed at elevated temperature ( 25°C ) , but not at a lower temperature ( 18°C ) [43] . Since our in planta experiments were carried out at ambient ( room ) temperatures ( ~20°C ) , it is conceivable that some of our ZED1 mutants ( such as ZED1N173D ) may display temperature-dependent phenotypes . Although both ZED1N173D and ZED1V212T mutants acquired strong HopZ1a-independent PBL binding activity , only the ZED1V212T and ZED1N173D V212T mutants ( with restored activation loops ) were sufficient to induce immunity in planta ( Fig 6; S13 Fig; Fig 7A; S15 Fig , panel A ) . ZED1 is unique among the ZRKs in that it lacks any candidate phospho-accepting residues in the activation loop ( S6 Fig , panel B ) . This feature is not specific to the Col-0 ecotype of Arabidopsis , since examination of 813 unique ZED1 sequences representing more than 1000 distinct and globally-distributed ecotypes identified only four with potential phospho/acetyl-accepting sites between the imperfect ‘DFG’ and ‘APE’ motifs that define the activation loop ( S16 Fig; see Materials and Methods ) . The HopZ1a acetylation target S244 of PBS1 ( Fig 1B ) provides further evidence for the functional importance activation loops in kinase-pseudokinase interactions . Glutamine substitution of this residue blocks HopZ1a-induced ZED1 binding activity ( S10 Fig ) , whereas an isoleucine substitution of the homologous activation loop position in PBL15 ( S260I ) acts as a gain-of-function mutant that mimics HopZ1a-induced ZED1-PBL binding ( Fig 4B ) . These results suggest that kinase-pseudokinase ( PBL-ZRK ) interactions can be promoted by activation loop-dependent structural changes in either of these protein families . We therefore speculated that an interface between ZED1 and PBLs involving the sites implicated by acetylation and/or mutational analyses may be similar to the previously described structure of a mutant IRAK4 kinase domain [44] . This mutant ( bearing a catalytic site aspartate to asparagine substitution ) gained the ability to dimerize in solution and was crystalized as an asymmetric dimer mediated by interactions between activation loops [44] . Remarkably , homology modeling that superimposes ZED1 and PBS1 on the two monomers of the asymmetric IRAK4 homodimer indicates that such a ‘front-to-front’ interface positions the acetylated PBL activation loop residue ( i . e . , PBS1 S244 or PBL15 S260 ) in very close proximity to ZED1 residues implicated by our mutational analysis ( N173 , T177 , and V212; S17 Fig ) . This hypothetical ZED1-PBL interface is corroborated by recent cryo-EM-derived structures of ZAR1-ZRK1/RKS1-PBL2 complexes; both a nucleotide-free intermediate form [19] and the pentameric , ATP-bound , activated resistosome [20] feature pseudokinase-kinase interfaces where the uridylated PBL2 activation loop residues are closely associated with the ZRK1/RKS1 activation loop and occupy the cleft between the amino-terminal and carboxy-terminal pseudokinase subdomains ( S18 Fig , S19 Fig ) . Kinase activation loops are also targeted by other T3Es . Cleavage of PBS1 by the T3E HopAR1 occurs after the lysine ( K243 ) immediately preceding S244 [8] , a modification that is perceived by the NLR RPS5 [9] . AvrAC uridylates serine and threonine residues in the activation loops of PBL2 , BIK1 and RIPK [45 , 46] , similar to the acetylation of PBS1 by HopZ1a . These conserved AvrAC uridylation sites are required for PBL2 interactions with ZRK1/RKS1 [14] , and are just two positions ‘downstream’ of PBS1 S244 ( S6 Fig , panel B ) . In human cells , YopJ ( a HopZ-related effector from the human plague pathogen , Yersinia pestis ) acetylates the activation loops of MAP kinase kinases ( MEK2 , MAPKK6 ) and both the α and β subunits of the IκB kinase ( IKK ) complex , resulting in inhibition of MAP kinase and NF-κB signaling pathways [47 , 48] . It has long been established that the phosphorylation status of activation loops has a role in influencing conformational changes that both position catalytic components and regulate access of substrates to the catalytic site [49 , 50] . As such , kinase activation loops may represent attractive targets for invading pathogens , while also providing the structural influence required to sense and transduce kinase domain perturbations . Our yeast four-hybrid data indicate that ZAR1 forms a ternary complex with ZED1 and PBL kinases that is stimulated by the presence of HopZ1a . Formation of this ternary complex is dependent on ZED1 ( Fig 5 ) and is likely similar to the ZAR1-ZRK1/RKS1-PBL2 complexes formed as a result of uridylation of PBL2 by AvrAC [20] . Notably , an intact ZAR1 LRR domain is required for both ZAR1-ZRK1/RKS1 interactions [14] and also for formation of the ZAR1-ZED1-PBL ternary complex promoted by HopZ1a ( Fig 5 ) , consistent with co-immunoprecipitation experiments in N . benthamiana with Arabidopsis ZED1 and ZAR1 truncation mutants [40] and with the ZAR1-ZRK1/RKS1 interface observed in cryo-EM structures [19 , 20] . Although direct interactions between ZED1 and the ZAR1 coiled-coil ( CC ) domain have been previously reported [16 , 40] , our data suggest indicate that the CC domain of ZAR1 can negatively regulate formation of a ZAR1-ZED1-PBL ternary complex ( Fig 5B ) . Indeed , inactive ADP-bound ZAR1-ZRK1/RKS1 complexes are stabilized by CC-mediated interdomain contacts with the HD1 , WHD , and LRR domains [19]—dramatic repositioning of the NBD and HD1 domains and a fold-switch in the CC domain are required for ADP/ATP exchange and subsequent assembly of the resistosome [19 , 20] . These findings are distinguished from the PBS1-RPS5 model wherein the CC domain of RPS5 is required for its interaction with intact PBS1 [9] . Nevertheless , cleavage of PBS1 by HopAR1 is similarly expected to activate RPS5 through a rearrangement of NLR interdomain contacts that is promoted by cleavage-induced perturbations of the PBS1 kinase domain [51] . Overall , our results support models suggesting that acetylation of ZED1 and/or of PBL kinases by HopZ1a promotes formation of ternary complexes with ZAR1; this process requires the LRR domain and is likely to be negatively-regulated by the ZAR1 CC domain ( Fig 8 , models B , C ) . Ternary complex formation may occur by stabilizing a preformed ZAR1-ZED1 complex , as described for the ZAR1-ZRK1/RKS1-PBL2 interactions [14 , 19 , 20] . Alternatively , HopZ1a-stabilized ZED1/PBL dimers may bind ZAR1 as a pre-formed unit to activate immunity , since we have found that HopZ1a-induced ZED1-PBL interactions are ZAR1-independent ( Fig 2A; Fig 4B; S8 Fig; S9 Fig; S10 Fig ) . According to our original decoy model of ZED1 function , acetylation of ZED1 should be sufficient to activate ZAR1 and trigger immunity [16] ( Fig 8 , model A ) . In this report however , we have described HopZ1a-induced ZED1-PBL ( and ZAR1-ZED1-PBL ) interactions that are consistent with the adaptor model proposed by Wang et al . [14] and recent structures of a ZAR1-ZRK1/RKS1-PBL2UMP resistosome [19 , 20] ( Fig 8 , model B ) , suggesting that the ability to conditionally interact with PBLs may be a general feature of ZRK pseudokinases . Nevertheless , our experiments have not ruled out a role for effector-mediated modification of ZED1 in the activation of ZAR1 . Indeed , HopZ1a can acetylate ZED1 at sites proximal to kinase motifs of known functional importance ( S7 Fig , panel A ) , glutamine substitution of the ZED1 acetylation site T177 abolishes HopZ1a-dependent ZED1-PBL binding ( S10 Fig , panel B ) , and isoleucine substitution of T177 is sufficient to cause ZAR1-dependent induction of immunity in N . benthamiana ( Fig 7B ) . We speculate that , like ZRK1/RKS1 [19 , 20] , ZED1 may act as a nucleotide exchange factor for ZAR1 to activate immunity , although in this case nucleotide exchange-promoting activity may be activated by direct acetylation of ZED1 . T177 is located between the degenerate catalytic site residue ( N173 ) and a trio of residues contributing to the ‘catalytic spine’ ( I179 , F180 , I181 ) , and its modification by HopZ1a may influence kinase structure by repositioning catalytic and regulatory spines and/or the activation loop ( S19 Fig ) [23] . Such structural rearrangements may allosterically regulate the relative orientations of ZAR1 subdomains to promote ADP/ATP exchange . PBL interactions may also be required for ZED1-dependent activation of ZAR1 , since HopZ1a activity can promote ZED1-PBL binding ( Fig 2A ) and formation of ZAR1-ZED1-PBL ternary complexes ( Fig 5 ) . Binding of PBL kinases to ( acetylated ) ZED1 might further stabilize conformational changes required to induce ADP/ATP exchange by ZAR1 , resistosome assembly , and immune activation . Although PBLs are also acetylation targets of HopZ1a ( Fig 1 ) , ZED1 mutations that restore degenerate pseudokinase motifs can also promote HopZ1a-independent ZED1-PBL interactions and ZAR1-ZED1-PBL complex formation ( Fig 3B; S8 Fig; S9 Fig , Fig 5B ) . We speculate that these mutations allow ZED1 to adopt an activated pseudokinase structure similar to that of acetylated ZED1 . We therefore introduce an additional model for activation of ZAR1 , recognizing that effector-mediated modification of either ZRKs or PBL kinases may be sufficient to form kinase-pseudokinase dimers and activate ETI ( Fig 8 , models B and C ) . We hypothesize that T3E-induced modifications introduce structural changes in ZRKs that can recruit unmodified PBL kinases to activate ZAR1 ( Fig 8 , model C ) . Together with the ZRK1/RKS1-PBL2 adaptor model ( Fig 8 , model B ) , this new model suggests that weak basal kinase-pseudokinase interactions are enhanced by structural transitions of either binding partner between complementary active/inactive conformations . The recently-described ZAR1-ZRK1/RKS1-PBL2 complexes [20] may indeed reflect a structural inactivation induced by post-translational modification ( PTM ) , since the structure of PBL2 is only partly defined , and its B-factors are elevated compared to those for ZRK1/RKS1 ( S18 Fig ) and ZAR1 . In contrast , while the ‘catalytic’ and ‘regulatory’ spines of ZRK1/RKS1 do not appear to undergo conspicuous structural changes in response to PBL2 binding or ZAR1 nucleotide occupancy , binding of PBL2UMP stabilizes the activation loop of ZRK1/RKS1 ( S19 Fig ) [19 , 20] . The PTMs catalyzed by HopZ1a ( acetylation; adds 42 Da ) and AvrAC ( uridylation; adds 324 Da ) differ in size and in physicochemical properties , and the diversity of enzymatic functions possessed by these and other effectors likely represents a significant force contributing to the evolutionary pressures that have driven diversification of the ZRKs . We note that by using pseudokinase adaptors that monitor kinase domain conformations rather than specific PTMs , plants would be able to recognize an even greater diversity of potential microbial effectors with diverse enzymatic activities . Further structural characterization of ZRK pseudokinases ( both alone and in complex with PBLs and/or ZAR1 ) will be important for critical assessment of our model presenting ZRK/PBL modules as molecular switches that are sensitive to perturbations of kinase structure . It is likely that the PBL kinases with HopZ1a-dependent ZED1 binding include virulence targets that are manipulated to promote P . syringae pathogenesis . Notably , a number of PBLs have been implicated in plant immunity , including BIK1 [52–55] , RIPK [56 , 57] , PBL13 [58] , and PBL27 [59 , 60] , and we have shown that the latter two participate in HopZ1a-dependent interactions with ZED1 ( Fig 2A ) . PBL kinases were not identified in our earlier forward genetic screen for loss of HopZ1a-induced HR [16] ( in contrast to the susceptibility of Arabidopsis pbl2 mutants to infections with AvrAC-expressing X . campestris [14] ) . If PBLs do in fact play a role in activation of ZAR1-mediated ETI by binding to ( acetylated ) ZED1 , we expect that there are multiple PBLs capable of fulfilling this function . Functionally-redundant PBL kinases would as result be individually dispensable for HopZ1a-induced ETI , or alternatively , one or more of these PBLs may also be essential for viability , precluding their recovery as loss-of-function mutants . HopZ1a is a member of a large and diverse family of bacterial effector proteins with similarity to the YopJ effector from Y . pestis . YopJ-like effectors are produced by a variety of pathogens of both plants and animals , and even among P . syringae strains there are five recognized distinct lineages of HopZ effectors ( HopZ1 through HopZ5 ) [61 , 62] . This inter-allelic sequence variation confers significant functional differences , since only HopZ1a is capable of inducing ETI in Arabidopsis . HopZ1b , the allele most closely-related to HopZ1a ( 64% identity ) , can trigger a weak , HR-like tissue collapse , but this response is ZAR1-independent and is only observed in ~25% of infiltrated leaves [13 , 63] . HopZ1b induced only a subset of the PBL-ZED1 interactions promoted by HopZ1a , consistent with its inability to activate robust ZAR1-dependent immunity . The PBL kinases that interact with ZED1 specifically ( or more strongly ) in the presence of HopZ1a than HopZ1b thus represent promising candidates for key regulators of ZAR1 activation by HopZ1a . HopZ2 ( 26% identical to HopZ1a ) did not promote interactions between ZED1 and PBL kinases ( S5 Fig ) and is able to promote P . syringae virulence in Arabidopsis without activating ZAR1 immunity [63] . Likewise , HopZ3 ( 23% identical to HopZ1a ) does not activate ZAR1 , however it is able to interact with PBLs ( RIPK , PBS1 , BIK1 , and PBL1 ) and can also acetylate the activation loop of RIPK ( S251 and S252—identical to the sites uridylated by AvrAC ) to dampen ETI mediated by the NLR RPM1 [64] . In our assays , however , co-expression with HopZ3 did not confer ZED1 binding activity upon RIPK or any other PBL ( S5 Fig ) . Overall , the promotion of most ZED1-PBL interactions is specific to HopZ1a/HopZ1b , suggesting that other HopZ alleles may have adopted distinct modification strategies to avoid recognition of their modified substrates by ZED1 and ZAR1 in Arabidopsis . Interestingly , a related HopZ-like T3E from Xanthomonas perforans , XopJ4/AvrXv4 ( 25 . 5% identical to HopZ1a across 337 non-gapped sites ) triggers a ZAR1-dependent ETI in N . benthamiana that is dependent on an RLCK XII protein named XOPJ4 IMMUNITY 2 ( JIM2 ) [65] . ZED1 is the Arabidopsis protein most closely-related to JIM2 ( BLASTP E-value of 5e-73; 41% identity spanning 92% of the JIM2 query sequence ) and notably , JIM2 has degenerate kinase motifs ( including a ‘dead’ HRD catalytic motif—‘YRI’ ) , suggesting that NbZAR1 may also use a pseudokinase/kinase module to detect perturbations of N . benthamiana signaling [65] . Kinases and pseudokinases possess switch-like structural features that are often allosterically regulated by post-translational modifications ( PTMs ) to influence their activity ( in the case of active kinases ) and/or alter their protein interaction profiles , resulting in reorganization and redistribution of signaling network activities . The plant immune system appears to have harnessed these switch-like features to detect kinase/pseudokinase perturbations that are induced by pathogen-delivered effector proteins . Modifications of either the ZED1/ZRK pseudokinases ( RLCK XII ) or of PBL kinases ( RLCK VII ) can promote interactions between these two families , as well as subsequent/concomitant interactions with the NLR ZAR1 to activate ETI . ZRK/PBL interactions are likely determined by a structural switch that is flipped by effector-mediated PTMs ( Fig 8 ) , providing an array of discriminating sensors that can detect diverse perturbations of the plant kinome ( including both kinases and pseudokinases ) . Further investigations of these intermolecular interactions , the ways in which they are influenced by bacterial effectors , and the structural features required of these components for robust immunity signaling will together provide valuable insights into the molecular mechanisms underlying pathogen perception by plant immune systems .
PBL coding sequences were obtained from the ABRC ( where available ) as pENTR223 or pUNI51 clones ( S1 Table ) . Additional PBL coding sequences were synthesized by General Biosystems , Inc . ( USA ) as GatewayTM-compatible pDONR207 clones . Recombination reactions using ‘LR clonase’ ( Invitrogen ) were used to shuttle coding sequences into a Gateway-compatible derivative of the prey plasmid pJG4-5 ( which encodes HA-tagged amino-terminal fusions of the B42 activation domain with nuclear localization sequences; NLS-B42-HA-prey ) . Donor plasmids carrying truncations of ZAR1 lacking the LRR ( ΔLRR , nucleotides 1–1545 , amino acids 1–515 ) or coiled-coil ( ΔCC , nucleotides 433–2556 , amino acids 145–852 ) domains , or carrying isolated individual domains ( CC , nucleotides 1–432 , amino acids 1–144; NB , nucleotides 433–1545 , amino acids 145–515; LRR , nucleotides 1546–2556 , amino acids 516–852 ) were prepared by PCR amplification and recombination into pDONR207 with BP clonase ( Invitrogen ) . Subsequent LR reactions were used to shuttle these constructs into the bait plasmid pEG202 ( which encodes an amino-terminal LexA DNA-binding domain; LexA-bait ) . Oligonucleotide primers used for site-directed mutagenesis are indicate in S5 Table . Plasmids were introduced into yeast strains according to standard LiAc/PEG methods [66 , 67] following sequence verification by sequencing both DNA strands of the entire coding sequences and across cloning junctions for each clone of interest , using Sanger sequencing services provided by the Center for the Analysis of Genome Evolution and Function ( CAGEF , University of Toronto ) . PBS1 was shuttled from pDONR207 into the centromere-based yeast plasmid pBA350V [68] using LR clonase to create a galactose-dependent expression vector for production of PBS1-FLAG in yeast . pBA350V::PBS1-FLAG was introduced into a derivative of yeast strain Y7092 with chromosomally-integrated hopZ1a-FLAG at the ho locus , as described previously [16 , 69] . FLAG-tagged proteins were expressed in yeast , and cell lysates were prepared and probed with anti-FLAG-conjugated agarose resin ( Sigma ) as described previously [16 , 69] . FLAG-tagged proteins immunoprecipitated in this manner were eluted by incubating with 100 uL of FLAG peptide solution ( 150 ug/mL FLAG peptide in TBS ) for one hour at 4°C . Eluted material was then dried to a pellet under vacuum and stored at -80°C prior to subsequent mass spectrometry analysis . Dried protein samples were re-solubilized in 50 mM ammonium bicarbonate ( pH 7 . 8 ) and then subjected to reduction with dithiothreitol at 56°C , alkylation with iodoacetamide at room temperature , and overnight digestion with sequencing-grade trypsin ( Promega ) at 37°C . This enzymatic reaction was terminated by addition of formic acid ( to 3% ) , and digestion products were purified and concentrated with Pierce C18 spin columns ( Thermo Fisher Scientific ) , then again dried to a pellet under vacuum . Peptide samples were then solubilized in 0 . 1% formic acid prior to LC-MS/MS analyses . Subsequent analytical separation was performed on a homemade , 75 μm i . d . column ( New Objective , Woburn , MA ) gravity-packed with 10 cm of 100 Å , 5 μm Magic C18AQ particles ( Michrom , Auburn , CA ) . Peptide samples were loaded onto the analytical column using a variable gradient with a flow rate of 300 nL/min . The gradient utilized two mobile phase solutions: A—water/0 . 1% formic acid; and B—80% acetonitrile/0 . 1% formic acid . Samples were analyzed on a linear ion trap-Orbitrap hybrid analyzer outfitted with a nano-spray source and EASY-nLC 1200 nano-LC system . The instrument method consisted of one MS full scan ( 400–1400 m/z ) in the Orbitrap mass analyzer , an automatic gain control target of 500 , 000 with a maximum ion injection of 500 ms , one microscan , and a resolution of 60 , 000 . Six data-dependent MS/MS scans were performed in the linear ion trap using the three most intense ions at 35% normalized collision energy . The MS and MS/MS scans were obtained in parallel fashion . In MS/MS mode automatic gain control targets were 10 , 000 with a maximum ion injection time of 100 ms . A minimum ion intensity of 1000 was required to trigger an MS/MS spectrum . The dynamic exclusion was applied using an exclusion duration of 145 s . Each sample was analyzed in triplicate . A spectral library was created using Proteome Discoverer 2 . 0 ( Thermo Fisher Scientific ) . Proteins were identified by searching all MS/MS spectra against a large database composed of the complete proteome of Saccharomyces cerevisiae strain S288C ( ATCC 204508; UniProt proteome ID UP000002311 ) supplemented with sequences for P . syringae HopZ1a ( WP_011152901 . 1 ) , and the Arabidopsis kinases PBS1 ( NP_196820 . 1 ) or ZED1 ( NP_567053 . 1 ) using SEQUEST [70] . A fragment ion mass tolerance of 0 . 8 Da and a parent ion tolerance of 30 ppm were used . Up to two missed tryptic cleavages were allowed . Methionine oxidation ( +15 . 99492 Da ) , cysteine carbamidomethylation ( +57 . 021465 Da ) , and acetylation ( +42 . 01057 Da; for serines and threonines only ) were set as variable modifications . Additional description of these data is provided in S1 File . Derivatives of plasmid pBA2262 used to generate yeast strains with chromosomally-integrated copies of hopZ1awt , hopZ1aC216A , hopZ1b , hopZ2 and hopZ3 have been described previously [69] . Briefly , genes cloned into pBA2262 are under the control of the GAL promoter ( which is positively regulated by galactose ) , they are linked to the downstream NATR gene ( which provides resistance to nourseothricin , also known as clonNAT ) , and are flanked by 5’ and 3’ fragments of the HO gene ( which encodes the homing endonuclease required for mating-type switching in wild yeast but is dispensable and inactivated—i . e . , hoΔ—in domesticated laboratory strains ) . pBA2262 cannot replicate in yeast so pBA2262 derivatives are linearized by digestion with the restriction enzyme NotI prior to transformation , and chromosomal integrants resulting from double recombination events that replace the endogenous hoΔ locus with the gene of interest are isolated by selection on YPDA plates ( yeast extract , peptone , dextrose , adenine sulfate ) containing clonNAT at 100 μg/mL . The avrAC coding sequence was amplified from pUC19-35S-avrAC-HA ( plasmid generously provided by Jian-Min Zhou; Chinese Academy of Sciences , Beijing ) to first create pDONR207-avrAC with BP clonase ( Invitrogen ) , which in turn allowed subsequent creation of pBA2262-avrAC with LR clonase ( Invitrogen ) . Similarly , ZED1wt and ZED1N173D V212T were shuttled from pDONR207-based plasmids to pBA2262 using LR clonase . The haploid yeast strain EGY48 ( ‘alpha’ mating type; i . e . , MAT α ) and derivative strains bearing chromosomally-integrated additional genes ( encoding bacterial effectors or alleles of the Arabidopsis pseudokinase ZED1 ) were transformed with query bait plasmids ( pEG202 derivatives; HIS+ ) , and transformants were isolated by selecting for prototrophy on Synthetic Defined ( SD ) minimal media containing 2% glucose and lacking histidine ( SD +Glc -His ) . Similarly , haploid yeast strain RFY206 ( ‘A’ mating type; i . e . , MAT A ) carrying the lacZ-bearing reporter plasmid pSH18-34 ( URA+ ) and a derivative strain bearing chromosomally-integrated ZED1wt were both transformed with prey plasmids ( pJG4-5; TRP+ ) , and transformants were isolated by selecting for prototrophy on SD glucose media lacking uracil and tryptophan ( SD +Glc -Ura -Trp ) . Expression of all 46 PBL prey fusions was confirmed in cultures of transformed derivatives of the haploid yeast strain RFY206/pSH18-34 following overnight growth in SD minimal media containing 1% raffinose , 2% galactose and lacking tryptophan and uracil ( SD +Raf +Gal -Ura -Trp ) ( S20 Fig ) . Protein extracts were prepared from cell pellets by precipitation with trichloroacetic acid ( TCA ) as described previously [69] , were resolved by electrophoresis through SDS-PAGE gels ( 10% polyacrylamide ) , and were probed with antibodies against the amino-terminal HA tag ( Roche ) . Expression of bait fusions ( and chromosomally-integrated genes ) was confirmed in cell lysates prepared in a similar fashion , except that overnight cultures were first grown in the absence of induction ( SD +Raf -His ) prior to dilution ( to OD600 = 0 . 2 ) in inducing media ( SD +Raf +Gal -His ) , growth for an additional two generations ( ~6–8 h ) and precipitation of total protein with TCA . Representative blots showing the expression of various ZAR1 bait constructs are shown in panel B of Fig 5 and S11 Fig . Simultaneous expression of bait-fusion , prey-fusion and integrated FLAG-tagged effectors in diploid cells carrying bait , prey and reporter plasmids is demonstrated in similar western blots shown in panel C of S5 Fig . Primary antibodies against HA ( Roche ) , the LexA DNA-binding domain ( Sigma ) and FLAG ( Sigma ) , as well as HRP-conjugated secondary antibodies ( Cell Signalling ) were all applied as 1:10 , 000 dilutions in TBST ( 50 mM Tris-Cl , pH 7 . 5; 150 mM NaCl; 0 . 05% Tween-20 ) with 3% powdered skim milk . Following confirmation of expression of all bait , prey and integrated genes in this way , 12 x 8 arrays of bait ( EGY48; MAT α ) and prey ( RFY206; MAT A ) transformants were first arranged manually on appropriate selective media , and all subsequent array manipulations were then performed using a 96-pin replicator . Diploid strains carrying bait , prey and reporter plasmids were created by co-incubation of bait and prey arrays on YPDA media ( 8–18 h ) , followed by two selections on SD glucose media lacking histidine , uracil and tryptophan ( SD +Glc -His -Ura -Trp ) . Bait-prey interactions were assessed on SD minimal media reporter plates containing 2% raffinose , 1% galactose , sodium phosphate ( 0 . 05 M , pH = 7 . 0 ) , X-gal ( 10 mg/mL ) and lacking histidine , uracil and tryptophan ( SD +Raf +Gal +X-gal -His -Ura -Trp ) . All yeast plates contained 2% agar and were incubated at 30°C . Reporter plates were imaged from the bottom using a flat-bed scanner ( Epson ) against a black felt background for contrast . The graphical summaries of the yeast interaction data shown in Figs 1–3 , Fig 5 , S4 Fig , S5 Fig , S8 Fig , S9 Fig and S11 Fig were prepared by finding the average colour for an 81-pixel square ( i . e . , 9 pixels x 9 pixels ) at the center of each colony . Averaged pixel intensities from yeast interaction plates were extracted by defining array positions with an interactive interface implemented in Python ( www . python . org ) using Processing ( www . processing . org; py . processing . org ) . Colours of text labels ( white or black ) for the prey array interaction summaries accompanying yeast reporter plate images ( and for the leaves on the phylogenetic tree shown in Fig 3B ) were determined according to a threshold based on the relative intensities of the red , green and blue channels for the 9 x 9 averaged pixel: specifically , the value of the ratio of bluered+green was evaluated at each PBL array position , and positions where this ratio was ≥ 0 . 727 ( relative interaction strength ≥0 . 455; i . e . , mostly blue; strong interactions ) were assigned white labels , while all other positions were labeled in black ( S2 Fig ) . This value was used as a working threshold for distinguishing strong interactions from the background since it bisects the path described by the total interaction dataset presented in this study ( S2 Fig ) . Pixel plotting and analysis was implemented in Python using Matplotlib [71] . Relevant scripts are available at https://github . com/DSGlab/Yeast-Array-Analysis . For construction of dexamethasone-inducible plant expression vectors , we cloned coding sequences ( lacking stop codons ) for hopZ1a and for wild-type and mutant ZED1 alleles into pMAC14 , which includes a dexamethasone-responsive promoter upstream of a multiple cloning site , as well as linked genes encoding resistance to kanamycin and glufosinate ( also known as BASTA ) [72] . hopZ1a and ZED1 alleles were cloned between XhoI and StuI restriction sites present in the pMAC14 multiple cloning site , placing the coding sequence in frame with a downstream , HA-tag-encoding sequence ( and stop codon ) . Agrobacterium strain GV3101 was transformed according to standard methods [73] , and transformants were selected by screening for plasmid-encoded kanamycin resistance . We used the floral dip method [74] to generate germ-line transformants of Arabidopsis using Agrobacterium strains transformed with pMAC14 derivatives . Arabidopsis transformants were selected by germinating seeds on soil infused with 0 . 1% BASTA ( v/v in water ) . Transformed ( BASTA-resistant ) seedlings were then transplanted into fresh soil free of herbicide along with control wild-type Arabidopsis ( Col-0 ) , third-generation ( T3 ) AvrRpt2-expressing plants , or zar1-1 mutants . Transgene expression was induced by spraying each flat with ~ 30 mL of 20 μM dexamethasone in ddH2O with 0 . 01% surfactant Silwet-L77 . Leaves of N . benthamiana plants were locally transformed by pressure infiltration with suspensions of pMAC14-transformed Agrobacterium strains using a needleless syringe applied to the underside of the leaf . Transgene expression was induced by spraying whole leaves with 20 μM dexamethasone in water with 0 . 01% surfactant Silwet-L77 ( 8–24h following infiltrations ) . Gene silencing was performed as described previously [40–42] , using Agrobacterium strains generously provided by Maël Baudin and Jennifer D . Lewis ( University of California , Berkeley ) . Expression of transgenes in Arabidopsis was assessed by punching three leaf cores from a single leaf which were then floated on 20 μM dexamethasone solution ( in ddH20 ) for ~18h . Expression of transgenes in N . benthamiana was induced by spraying infiltrated plants with a solution of 20 μM dexamethasone in ddH20 with 0 . 01% Silwet L-77 . For each transformation of interest , tissue was harvested by punching three circular leaf cores 5 mm in diameter . These tissue samples were then flash-frozen in liquid nitrogen before manual grinding with a mini-pestle in a 1 . 5 mL Eppendorf tube . Powdered frozen plant tissue was then suspended in 100 μL of Grinding Buffer ( 40 mM Tris pH = 7 . 5 , 150 mM NaCl , 1 mM EDTA , 5 mM DTT , 1% Triton X-100 , 0 . 1% sodium dodecyl sulfate ) supplemented with a 1:500 dilution of plant protease inhibitor cocktail P9599 ( Sigma ) . 15 uL from each of these extracts was then resolved by electrophoresis through 10% polyacrylamide SDS-PAGE gels prior to protein transfer to nitrocellulose membranes . Membranes were blocked with 5% powdered skim milk solution in TBST ( 50 mM Tris-Cl , pH 7 . 5; 150 mM NaCl; 0 . 05% Tween-20 ) . Primary and secondary antibodies were both diluted 1:10 , 000 in TBST with 3% powdered milk . Sequences for 46 PBL kinases and 12 ZED1-related pseudokinases were retrieved from TAIR ( www . arabidopsis . org ) , aligned with Muscle [75] , and trimmed to remove variable-length amino-terminal and carboxy-terminal sequences flanking the conserved kinase domain . These isolated kinase domain sequences were then supplemented with sequences from structurally-characterized kinase domains of the Arabidopsis receptor kinases BRI1 ( PDB: 5LPV ) [35] and BAK1 ( PDB: 3UIM ) [36] , the tomato Pto kinase ( PDB: 3HGK ) [76] , and the human interleukin-1 receptor-associated kinase , IRAK4 ( PDB: 4U97 ) [44] . These 62 sequences were then realigned with Muscle and any amino-terminal or carboxy-terminal sequences flanking the kinase domain were again manually trimmed . The resulting alignment is provided as a supplementary file ( S2 File ) and was used as input for phylogenetic inference using PhyML [77 , 78] . Confidence in nodes of the resulting phylogenetic tree was assessed using the approximate Likelihood Ratio Test ( aLRT ) [79 , 80] . A representation of the resulting tree ( rooted on IRAK4 and including only those nodes with greater than 70% confidence ) is presented in S6 Fig , panel A , and the input tree is provided as a supplementary file ( S3 File ) . A derivative of the same tree is also shown in Fig 3B and is the result of a pruning step to remove any non-PBL leaves . Kinase domain trees were visualized using iTOL [81] . The ZED1 sequences for available Arabidopsis ecotypes were derived from DNA SNP information curated by the 1001 genomes project ( http://1001genomes . org/ ) [82] and were annotated using Maker ( http://www . yandell-lab . org/software/maker . html ) [83] . 1121 translated ZED1 sequences were collapsed into 813 unique sequences and aligned using Muscle [75] ( S4 File ) . 25 of these unique sequences represent more than one Arabidopsis ecotype , while the remaining 788 unique sequences are ‘singletons’ representing a single Arabidopsis ecotype ( S5 File ) . Although confidence in translated ZED1 sequences depends on DNA sequence quality ( which varies among individual sequenced ecotypes ) , 92 . 5% of the unique ZED1 sequences ( mean un-gapped sequence length of 332 . 7 amino acids ) had fewer than 38 positions translated as an ‘X’ due to ambiguous DNA sequence information ( i . e . , ≤ 11 . 4% of the mean sequence length ) , and 76 . 3% of unique sequences had fewer than 23 positions translated as X ( ≤ 6 . 9% of the mean sequence length; S16 Fig , panels A , B , top left ) . A cropped alignment consisting of only the activation loop sequences ( mean un-gapped sequence length of 29 . 0 amino acids ) had a similar distribution of X position prevalence , with 89 . 0% of activation loop sequences containing two or fewer X positions ( ≤ 6 . 9% of mean activation loop length; S16 Fig , panels A , B , bottom left ) . In contrast , the distributions of serine and threonine positions ( S+T ) were markedly different when comparing the activation loop region with full-length protein sequences . While 88 . 6% of the unique Arabidopsis ecotype ZED1 sequences have at least 36 serine or threonine positions ( ≥ 10 . 8% of the mean sequence length ) across their full length ( S16 Fig , panels A , B , top right ) , the vast majority of ecotypes ( 99 . 5% ) lack any serine or threonine residues when considering activation loop sequences alone ( S16 Fig , panels A , B , bottom right ) . Notably , the four ecotypes with threonines in their activation loop sequences are all outliers with respect to sequence quality , with either three or six ambiguous positions translated as X ( sequence ambiguity that is representative of ≤ 7% of all unique sequences; S16 Fig ) , so it is possible that these rare apparent phospho/acetyl-accepting activations residues are in fact artefacts resulting from poor sequence quality . All three-dimensional protein structural models were prepared using the interface to Modeller [84] provided within UCSF Chimera [85] . Model templates were selected based on high-scoring BLASTP hits in the PDB database . The ZED1 model presented in S7 Fig and S10 Fig was constructed using the multiple template option of Modeller and is based on homology with both Solanum pimpinellifolium ( tomato ) Pto kinase ( PDB: 3HGK; 30% identity spanning 81% of query; E-value: 4e-31 ) [76] and the kinase domain of Arabidopsis receptor kinase BRI1 ( PDB: 5LPV; 30% identity spanning 66% of the query; E-value: 5e-27 ) [35] . The PBS1 model in S10 Fig was created much like the ZED1 model but using structures of BIK1 ( PDB: 5TOS; 48% identity spanning 70% of query; E-value: 1e-97 ) [55] and BRI1 ( PDB: 5LPV; 44% identity spanning 62% of query; E-value: 1e-71 ) [35] as model-building templates . Modeling was based on two templates so as to generate a model reliably representing the entire kinase domain . For homology modeling of the hypothetical asymmetric ZED1-PBS1 heterodimer shown in S17 Fig , multiple candidate templates were considered ( S18 Fig ) . ZRK-PBL dimers have been described as part of ZAR1-ZRK1/RKS1-PBL2UMP complexes observed by cryo-EM [19 , 20] , however in these structures the electron density corresponding to PBL2UMP is not well resolved ( mean per-residue B-factors: ~195 , ~225 ) and represents less than half of the PBL2 sequence ( S18 Fig , panels A , B ) . Although a disordered , partly-unfolded state may be a biologically realistic consequence of PBL2 uridylation by the X . campestris effector AvrAC , the increased disorder of PBL2 relative to ZRK1/RKS1 ( and ZAR1 ) is also likely a result of the averaging of ZAR1-ZRK1/RKS1-PBL2 particles—structural variation within and between particles increases in proportion to the radial distance from the centre of protein complexes [86] . While the structure of the PBL kinase BIK1 has been determined by X-ray crystallography [55] and is accordingly of higher resolution ( mean per-residue B-factor: ~57 ) , in this case the kinase dimer interface is in a ‘back-to-back’ orientation that may be a crystal packing artefact and therefore not biologically-relevant ( S18 Fig , panel C ) . In contrast , the X-ray crystal structure of human IRAK4 [44] is of similarly high resolution ( mean per-residue B-factor: ~64 ) and furthermore presents a ‘front-to-front’ interface in which the activation loops from both protomers are intimately involved in the dimer interface ( S18 Fig , panel D ) . IRAK4 ( PDB: 4U97 ) [44] was therefore used as model template for both ZED1 ( 30% identity across 63% of query sequence; E-value 7e-21 ) and PBS1 ( 42% identical across 62% of query sequence; E-value 8e-56 ) in addition to the top-scoring BLASTP hits for ZED1 ( Pto ) and PBS1 ( BIK1 ) described above . In this case , however , activation loop residues from Pto and BIK1 were deleted from input alignments prior to modeling of ZED1 and PBS1 so as to force modeled activation loops to be constrained only by those from chains A and B from PDB entry 4U97 ( IRAK4 ) . Photographs were acquired with a Nikon D5200 DSLR camera . We have previously described an ImageJ macro ( PIDIQ , for Plant Immunity and Disease Image-based Quantification ) that quantifies pixel areas corresponding to green or yellow-shaded regions of input images [87] . Here we used a similar approach to replicate the behaviour of a factory-installed camera filter that serendipitously produces images with pixels representing dead/dying plant tissue and surrounding soil converted to grayscale , while those pixels representing healthy Arabidopsis leaf tissue remain unchanged . We determined that this behavior can be replicated by applying logic that outputs a full-colour , RGB pixel where the log-transformed ratio between pixel intensities for red and green channels ( log10redgreen ) is ≤−0 . 07 , and a grayscale pixel where this condition is not satisfied ( S21 Fig ) . Pixel analysis and image processing were implemented in Python using Matplotlib [71] and a script is available at https://github . com/DSGlab/Live-Dead-Filter . | All plants must ward off potentially infectious microbes , and those grown in large-scale crop operations are especially vulnerable to the rapid spread of disease by successful pathogens . Although many bacteria and fungi can supress plant immune responses by producing specialized virulence proteins called ‘effectors’ , these effectors can also trigger immune responses that render plants resistant to infection . We studied the molecular mechanisms underlying one such effector-triggered immune response elicited by the bacterial effector HopZ1a in the model plant host Arabidopsis thaliana . We have shown that HopZ1a promotes binding between a ZED1 , a ‘pseudokinase’ required for HopZ1a-triggered immunity , and several ‘true kinases’ ( known as PBLs ) that are likely targets of HopZ1a activity in planta . HopZ1a-induced ZED1-PBL interactions also recruit ZAR1 , an Arabidopsis ‘resistance protein’ previously implicated in HopZ1a-triggered immunity . Importantly , ZED1 mutants that restore degenerate kinase motifs can bridge interactions between PBLs and ZAR1 ( independently of HopZ1a ) and trigger immunity in planta . Our results suggest that equilibria between active and inactive kinase domain conformations regulate ZED1-PBL interactions and formation of ternary complexes with ZAR1 . Improved models describing molecular interactions between immunity determinants , effectors and effector targets will inform efforts to exploit natural diversity for development of crops with enhanced disease resistance . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"protein",
"interactions",
"brassica",
"plant",
"physiology",
"plasmid",
"construction",
"fungi",
"plant",
"science",
"model",
"organisms",
"experimental",
"organism",
"systems",
"plant",
"pathology",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"dna",
"construction",
"plants",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"sequence",
"alignment",
"animal",
"studies",
"proteins",
"bioinformatics",
"chemistry",
"molecular",
"biology",
"yeast",
"biochemistry",
"plant",
"defenses",
"eukaryota",
"plant",
"and",
"algal",
"models",
"post-translational",
"modification",
"plant",
"disease",
"resistance",
"acetylation",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"organisms"
] | 2019 | Perturbations of the ZED1 pseudokinase activate plant immunity |
Abscission is the final step of cytokinesis that involves the cleavage of the intercellular bridge connecting the two daughter cells . Recent studies have given novel insight into the spatiotemporal regulation and molecular mechanisms controlling abscission in cultured yeast and human cells . The mechanisms of abscission in living metazoan tissues are however not well understood . Here we show that ALIX and the ESCRT-III component Shrub are required for completion of abscission during Drosophila female germline stem cell ( fGSC ) division . Loss of ALIX or Shrub function in fGSCs leads to delayed abscission and the consequent formation of stem cysts in which chains of daughter cells remain interconnected to the fGSC via midbody rings and fusome . We demonstrate that ALIX and Shrub interact and that they co-localize at midbody rings and midbodies during cytokinetic abscission in fGSCs . Mechanistically , we show that the direct interaction between ALIX and Shrub is required to ensure cytokinesis completion with normal kinetics in fGSCs . We conclude that ALIX and ESCRT-III coordinately control abscission in Drosophila fGSCs and that their complex formation is required for accurate abscission timing in GSCs in vivo .
Cytokinesis is the final step of cell division that leads to the physical separation of the two daughter cells . It is tightly controlled in space and time and proceeds in multiple steps via sequential specification of the cleavage plane , assembly and constriction of the actomyosin-based contractile ring ( CR ) , formation of a thin intercellular bridge and finally abscission that separates the two daughter cells [1–8] . Studies in a variety of model organisms and systems have elucidated key machineries and signals governing early events of cytokinesis [1–6] . However , the mechanisms of the final abscission step of cytokinesis are less understood , especially in vivo in the context of different cell types in a multi-cellular organism [2 , 4 , 5] . During the recent years key insights into the molecular mechanisms and spatiotemporal control of abscission have been gained using a combination of advanced molecular biological and imaging technologies [4 , 7 , 9–15] . At late stages of cytokinesis the spindle midzone transforms to densely packed anti-parallel microtubules ( MTs ) that make up the midbody ( MB ) and the CR transforms into the midbody ring ( MR , diameter of ~1–2 µm ) [4 , 10 , 16 , 17] . The MR is located at the site of MT overlap and retains several CR components including Anillin , septins ( Septins 1 , 2 and Peanut in Drosophila melanogaster ) , myosin-II , Citron kinase ( Sticky in Drosophila ) and RhoA ( Rho1 in Drosophila ) and eventually also acquires the centralspindlin component MKLP1 ( Pavarotti in Drosophila ) [4 , 16 , 18 , 19] . In C . elegans embryos the MR plays an important role in scaffolding the abscission machinery even in the absence of MB MTs [20] . Studies in human cell lines , predominantly in HeLa and MDCK cells , have shown that components of the endosomal sorting complex required for transport ( ESCRT ) machinery and associated proteins play important roles in mediating abscission [4 , 7 , 9–15] . Abscission occurs at the thin membrane neck that forms at the constriction zone located adjacent to the MR [9 , 10 , 17] . An important signal for initiation of abscission is the degradation of the mitotic kinase PLK1 ( Polo-like kinase 1 ) that triggers the targeting of CEP55 ( centrosomal protein of 55 kDa ) to the MR [21] . CEP55 interacts directly with GPP ( 3x ) Y motifs in the ESCRT-associated protein ALIX ( ALG-2-interacting protein X ) and in the ESCRT-I component TSG101 , thereby recruiting them to the MR [13–15 , 22] . ALIX and TSG101 in turn recruit the ESCRT-III component CHMP4B , which is followed by ESCRT-III polymerization into helical filaments that spiral/slide to the site of abscission [9 , 11 , 13–15 , 23] . The VPS4 ATPase is thought to promote ESCRT-III redistribution toward the abscission site [23] . Prior to abscission ESCRT-III/CHMP1B recruits Spastin that mediates MT depolymerization at the abscission site [9 , 10 , 24] . ESCRT-III then facilitates membrane scission of the thin membrane neck , thereby mediating abscission [9 , 10] . Cytokinesis is tightly controlled by the activation and inactivation of mitotic kinases at several steps to ensure its faithful spatiotemporal progression [7 , 8] . Cytokinesis conventionally proceeds to completion via abscission , but is differentially controlled depending on the cell type during the development of metazoan tissues . For example , germ cells in species ranging from insects to humans undergo incomplete cytokinesis leading to the formation of germline cysts in which cells are interconnected via stable intercellular bridges [25–27] . How cytokinesis is modified to achieve different abscission timing in different cell types is not well understood , but molecular understanding of the regulation of the abscission machinery has started giving some mechanistic insight [25 , 26 , 28–30] . The Drosophila female germline represents a powerful system to address mechanisms controlling cytokinesis and abscission in vivo [29 , 31] . Each Drosophila female germline stem cell ( fGSC ) divides asymmetrically with complete cytokinesis to give rise to another fGSC and a daughter cell cystoblast ( CB ) [31–33] . Cytokinesis during fGSC division is delayed so that abscission takes place during the G2 phase of the following cell cycle ( about 24 hours later ) [31] . The CB in turn undergoes four mitotic divisions with incomplete cytokinesis giving rise to a 16-cell cyst in which the cells remain interconnected by stable intercellular bridges called ring canals ( RCs ) [27 , 32] . One of the 16 cells with four RCs will become specified as the oocyte and the cyst becomes encapsulated by a single layer follicle cell epithelium to form an egg chamber [34 , 35] . Drosophila male GSCs ( mGSCs ) also divide asymmetrically with complete cytokinesis to give rise to another mGSC and a daughter cell gonialblast ( GB ) [33 , 36 , 37] . Anillin , Pavarotti , Cindr , Cyclin B and Orbit are known factors localizing at RCs/MRs and/or MBs during complete cytokinesis in fGSCs and/or mGSCs [29 , 31 , 36 , 38–43] . Mathieu et al . recently reported that Aurora B delays abscission and that Cyclin B promotes abscission in Drosophila germ cells and that mutual inhibitions between Aurora B and Cyclin B/Cdk-1 control the timing of abscission in Drosophila fGSCs and germline cysts [29] . However , little is known about further molecular mechanisms controlling cytokinesis and abscission in Drosophila fGSCs . Here we characterize the roles of ALIX and the ESCRT-III component Shrub during cytokinesis in Drosophila fGSCs . We find that ALIX and Shrub are required for completion of abscission in fGSCs , that they co-localize during this process and that their direct interaction is required for abscission with normal kinetics . We thus show that a complex between ALIX and Shrub is required for abscission in fGSCs and provide evidence of an evolutionarily conserved functional role of the ALIX/ESCRT-III pathway in mediating cytokinetic abscission in the context of a multi-cellular organism .
The ESCRT-associated scaffold protein ALIX promotes cytokinetic abscission in human cultured cells [13–15] . We were interested to characterize the role of ALIX in cytokinesis in vivo using Drosophila melanogaster as a model because of its power for elucidation of mechanisms of cytokinesis and abscission in different cell types in a developing organism [2 , 5 , 6 , 29] . We first raised an antibody against Drosophila ALIX ( CG12876 ) ( Fig . 1A ) and examined its subcellular localization during S2 cell division . During meta- , ana- and early telophase ALIX localized at centrosomes ( Fig . 1B-E ) , where it co-localized with Centrosomin ( S1A-S1C Fig . ) . ALIX localization at centrosomes has been detected in human cultured cells in interphase [15] , but to our knowledge ALIX localization at centrosomes during different phases of mitosis has not previously been shown . Strikingly , at mid telophase a fraction of ALIX re-localized from the spindle poles to two pools within the intercellular bridge on each side of the MR/dark zone ( Figs . 1F and S1D ) . Finally , ALIX localized to the central region of the intercellular bridge during late telophase/cytokinesis ( Figs . 1G and S1E ) . Here it appeared to localize to the MR because it formed a ring-like structure around the MTs of the MB ( Fig . 1G ) at the dark zone ( S1E Fig . ) . The pre-immune serum neither stained centrosomes , nor the intercellular bridge or MR ( S1A-S1E Fig . ) . This spatiotemporal redistribution from centrosomes to the MR suggested a possible role for ALIX in cytokinesis in Drosophila cells . We further addressed the role of Drosophila ALIX in cytokinesis in vivo by analyzing two different alix mutant alleles , alix1 and alix3 ( Fig . 2A ) . ALIX is highly expressed in Drosophila embryos , larvae , pupae , adult females and males , as well as in ovaries and testes ( S2A Fig . ) . Interestingly , homozygous mutant offspring of both the alix1 and alix3 mutants could survive to adulthood ( even though they clearly lack the full-length ALIX protein ) ( S2B Fig . and see below ) and we detected none or only minor bi-nucleation clearly attributed to cytokinesis failure in the somatic cell types we analyzed ( S2C-S2F Fig . and S1–S2 Tables ) . Fertility tests of alix1 mutant flies however revealed that both female and male fertility was reduced ( S3A–S3B Fig . ) . In particular female fertility was severely compromised , manifested by very low egg lay and hatch rates ( S3A–S3B Fig . ) . We therefore asked whether oogenesis of alix mutant flies might be altered . Wild type egg chambers contain 16 germ cells and an oocyte with 4 RCs ( Fig . 2B , 2D , 2G , 2E and 2H ) . Curiously , egg chambers in ovaries of both alix1 and alix3 mutant females lacking full-length ALIX ( Fig . 2C and 2F ) often contained exactly 32 germ cells and an oocyte with 5 RCs ( Fig . 2D and 2G ) . Quantifying the egg chamber phenotypes of alix1 and alix3 mutant ovaries revealed that about 60% of the egg chambers in both alleles contained 32 germ cells ( Fig . 2E and 2H ) . We also detected low percentages of egg chambers with more than 32 germ cells in both alix mutant alleles ( Fig . 2E and 2H ) . We next analyzed whether the increased germ cell number in egg chambers was specifically due to loss of alix gene function . Firstly , alix1 and alix3 alleles combined either with two different deficiencies lacking the alix gene or with each other gave rise to 50–60% of egg chambers with 32 germ cells , similar to homozygous alix1 and alix3 mutants ( S3C-S3E Fig . ) . Secondly , two genomic rescue lines containing the full alix gene locus rescued the 32-germ cell phenotype of both the alix1 and alix3 alleles ( S4A–S4F Fig . ) . Finally , RNAi-mediated gene silencing of alix specifically in female germ cells using the maternal triple MTD-GAL4 driver [44–46] resulted in about 50% of egg chambers with 32 or more germ cells ( S4G-S4I Fig . ) showing that absence of ALIX specifically in germ cells causes the 32-germ cell phenotype . We conclude that loss of ALIX function in the Drosophila female germline causes the formation of a high frequency of egg chambers with 32 or more germ cells . Egg chambers with 32 germ cells may arise via encapsulation of two 16-cell cysts by the follicle cell epithelium , an extra round of mitosis in germline cysts or a delay in abscission in fGSCs [29 , 32 , 35 , 47 , 48] . The fact that the egg chambers with 32 germ cells contained one oocyte with 5 RCs excluded that they arose via defective encapsulation of two 16-cell cysts . We further discriminated between the two latter mechanisms by performing RNAi-mediated gene silencing of alix specifically in the germline using either Nanos-GAL4 ( expresses in all germ cells; fGSCs , CBs and 2–16-cell cysts ) or Bam-GAL4 ( expresses in CBs to 8-cell cysts , but not in fGSCs ) to test whether the phenotype originated from fGSCs or cell autonomously in germline cysts . Interestingly , alix-RNAi using Nanos-GAL4 ( Nanos-GAL4 or UAS-Dicer; Nanos-GAL4 ) gave rise to 40–60% egg chambers with 32 germ cells , whereas alix depletion using Bam-GAL4 resulted in normal egg chambers with 16 germ cells only ( S5A-S5D Fig . ) . These data linked alix depletion in fGSCs to the formation of egg chambers with 32 germ cells and suggested that they did not arise from an extra round of mitosis of germline cysts . This thus indicated a role for ALIX in abscission in fGSCs in agreement with recent work showing that a delay in abscission in fGSCs can give rise to the formation of stem cysts in which the fGSC is connected to several daughter cells [29] . If abscission eventually takes place , a 2-cell cysts may pinch off and subsequently undergo four rounds of mitosis , giving rise to a 32-cell cyst [29] . We thus investigated whether or not we could detect stem cysts following loss of ALIX function . Stem cysts are characterized by their elongated fusomes , their weak Nanos expression as in stem cells , their lack of expression of the cyst differentiation factor Bam and that the cell in direct contact with the stem cell niche is positive for p-Mad [29] . The cells within the stem cysts are moreover found to divide synchronously [29] . Importantly , alix1 and alix3 germaria as well as germaria with alix-RNAi in fGSCs displayed chains of weakly Nanos-positive germ cells interconnected by elongated fusomes in which the most anterior cell was in contact with the cap cells in the stem cell niche of the germarium ( Figs . 3A-C and S5E-S5H ) . The cell in contact with the cap cells in such alix-deficient cysts was moreover p-Mad-positive ( S6A Fig . ) and the cysts were Bam-negative ( S6B-S6D Fig . ) . We also detected synchronously dividing cells in the anterior tip of alix-deficient germaria ( Fig . 3I ) . Taken together , these characteristics defined the alix-deficient cysts as stem cysts and indicated a role for ALIX in abscission in fGSCs . Each fGSC divides with complete cytokinesis giving rise to another stem cell and a daughter cell CB [31 , 32] ( Fig . 2B ) . fGSC cytokinesis progression can be monitored using markers for the fusome and RCs ( hereafter referred to as MRs ) [31 , 32 , 49 , 50] . To determine the nature and frequency of the abscission defects upon loss of ALIX function we quantified fGSC morphologies in wild type , alix1 and alix3 germaria using markers for the fusome ( hts-F ) , MRs/MBs ( Cindr ) [38] and nuclei ( Figs . 3D-G and S7A-S7H ) . We categorized fGSC phenotypes as indicated in Fig . 3H ( and as illustrated in S7E Fig . ) . Wild type fGSCs displayed only normal phenotypes: ~50% fGSCs with a spectrosome , ~40% fGSC-CB pairs with an MR and ~10% fGSC-CB pairs with an MB ( Figs . 3D-E , 3H , S7A-S7B and S7E ) . These frequencies of different fGSC cell cycle stages are consistent with previous reports [49 , 51] . alix3 and alix1 mutant germaria contained smaller fractions of fGSCs with a spectrosome ( ~15% for both mutants ) , fGSC-CB pairs with an MR ( ~25% in alix3 and ~10% in alix1 ) and fGSC-CB pairs with an MB ( 0% in alix3 and ~1% in alix1 ) compared to wild type ( Figs . 3H and S7E ) . Importantly , more than half of the alix mutant fGSCs showed abscission defects: linear chains ( ~20% in alix3 and ~10% in alix1 ) , branched chains ( ~40% in alix3 and ~30% in alix1 ) or polyploidy ( ~2% in alix3 and ~30% in alix1 ) ( Figs . 3F-H and S7C-S7H ) . Abscission defects appeared in the majority of both alix3 and alix1 mutant germaria , and never in wild type ( S3–S4 Tables ) . Consistently , upon alix-RNAi in the germline using Nanos-GAL4 ( Nanos-GAL4 or UAS-Dicer; Nanos-GAL4 ) the majority of germaria contained stem cysts in which the fGSC was interconnected to multiple daughter cells via fusome and MRs ( S7I-S7K and S5H Figs . ) . We occasionally detected MBs in stem cysts in alix1 and alix3 mutant germaria ( even though MRs predominated ) , indicating abscission events , and cysts of exactly two cells in the process of pinching off ( S7L-S7N Fig . ) . This is consistent with the model of how 32-cell cysts appear following delayed abscission in fGSCs as previously described [29] . Collectively , these results showed that loss of ALIX caused a delay in abscission in fGSCs with the consequent formation of a high frequency of stem cysts . The fact that cells in stem cysts were interconnected in chains via MRs ( Figs . 3F-H , S7C-S7E and S7I-S7J ) together with the infrequent observation of fGSC-CB pairs with an MB upon loss of ALIX function ( Figs . 3H and S7E ) suggested that ALIX plays a role in promoting closure of the MR to mediate fGSC abscission . We conclude that ALIX is required for completion of abscission in Drosophila fGSCs . We further asked whether ALIX may also be required for abscission in asymmetrically dividing Drosophila mGSCs . We stained testes tips from wild type , alix1 and alix3 mutants with antibodies to visualize the hub to which the mGSCs are attached , the fusome , MRs and MBs . In alix1 and alix3 mutant testes that lack full-length ALIX ( S8A Fig . ) ~20% of alix1 and ~40% of alix3 mutant mGSCs were found interconnected to chains of daughter cells by MRs and fusome ( S8B-S8E Fig . ) . These results suggest that ALIX promotes abscission in both female and male GSCs . To examine the subcellular localization of ALIX during cytokinesis in fGSCs we generated transgenic flies with GFP-tagged ALIX under the control of the UASp promoter ( UASp-GFP-ALIX ) and expressed it in fGSCs and germline cysts using MTD-GAL4 or Nanos-GAL4 . We then visualized the progressive stages of fGSC cytokinesis using markers for the fusome and MRs/MBs ( Cindr ) or MTs ( α-tubulin ) . In fGSC-CB pairs in which a small fusome plug had formed within the MR ( G1 ) we detected GFP-ALIX overlapping mainly with the fusome plug ( Fig . 4A ) . At this point we detected anti-parallel MT bundles with a dark zone to which the fusome plug started localizing ( S9A Fig . ) . Then , as the fusome adopted bar morphology in G1/S GFP-ALIX localized at the MR and at this point the MTs were largely degraded ( S9B Fig . ) . GFP-ALIX remained at MRs throughout G1/S , S and early G2 ( Figs . 4B and S9B-S9C ) and then localized at MBs during abscission ( G2 ) ( Fig . 4C ) . We thus conclude that GFP-ALIX is recruited to the center of the MR and then moves to the MR during G1/S , is detected at MRs throughout cytokinesis progression and then finally localizes to MBs during abscission in Drosophila fGSCs . This spatiotemporal dynamics of ALIX during late stages of fGSC cytokinesis is consistent with a role for ALIX in abscission in fGSCs . We next asked by which molecular mechanisms ALIX may act during abscission in fGSCs . The ESCRT-III component and CHMP4 orthologue Shrub ( CG8055 ) was an interesting candidate to mediate abscission together with ALIX because of the important role of ESCRT-III in promoting membrane scission during cytokinetic abscission and because ALIX directly interacts with and recruits the ESCRT-III subunit CHMP4B to the MB to promote abscission in human cells [13 , 15 , 52 , 53] . The interaction between ALIX and CHMP4B is mediated via a motif within the Bro1 domain of human ALIX ( MxxxIxxxL , aa 199–216 ) and a motif in the CHMP4 C-terminus ( MxxLxxW , aa 214–220 ) [13 , 15 , 54] . Importantly , these mutual consensus interaction sites are conserved in Drosophila ALIX and Shrub , respectively ( ALIX: LxxxIxxxL , aa 198–215 and Shrub: MxxLxxW , aa 218–224 ) ( Fig . 5A ) [54] . We therefore tested the possible interaction by co-immunoprecipitation analyses of GFP-tagged Shrub and endogenous ALIX from Drosophila Dmel cell lysates . These analyses showed that GFP-Shrub and ALIX indeed were detected in the same complex ( Fig . 5B ) . We next examined the relative localization of ALIX and Shrub during fGSC cytokinesis . For this purpose GFP-Shrub was expressed using Nanos-GAL4 and ALIX was detected with our anti-ALIX antibody . Interestingly , GFP-Shrub localized at MRs and MBs during cytokinesis in fGSCs ( consistent with observations by [55] ) and ALIX co-localized with GFP-Shrub at MRs in G1/S and S phase and then at MBs during abscission in G2 ( Fig . 5C-F ) . Strikingly , GFP-Shrub additionally localized at the fusome ( Fig . 5C-D and [55] ) . Furthermore , ALIX and GFP-Shrub co-localized at bright dot-like structures on the fusome in fGSCs ( Fig . 5E ) . These most likely represented MB remnants that have been reported to be inherited by the fGSC following cytokinesis completion [36] . Consistently , we detected that GFP-ALIX on MB remnants was preferentially retained in fGSCs following abscission ( data not shown ) . We also noted that GFP-Shrub was weakly detected along the membrane at the point that anti-parallel MTs were detected in fGSC-CB pairs in G1 ( S9D Fig . ) and then accumulated at MRs from G1/S ( Figs . 5C-D and S9D ) . Taken together our results suggested that ALIX and GFP-Shrub co-localize at MRs from G1/S and then at MRs and MBs throughout cytokinetic abscission in Drosophila fGSCs . We next analyzed the role of Shrub as well as the possible functional relationship between ALIX and Shrub during fGSC cytokinesis . We first performed RNAi-mediated depletion of shrub using the Nanos-GAL4 driver . Control germaria displayed normal fGSC and egg chamber phenotypes ( Figs . 6A , 6E , S10A and S10E ) . Upon alix-RNAi about 40% of fGSCs were found in stem cysts ( linear or branched ) ( Fig . 6B and 6E ) and ~50% of the egg chambers contained 32 germ cells ( S10B and S10E Fig . ) . Importantly , following shrub-RNAi about 45% of fGSCs formed stem cysts ( Fig . 6C and 6E ) , ~10% of the fGSCs were polyploid ( Fig . 6E ) and ~50% of the egg chambers contained 32 germ cells ( Figure S10C and S10E Fig . ) . Consistently , stem cysts were also present in 70% of heterozygous shrubG5/+ mutant germaria ( S10F Fig . ) suggesting that the stem cysts appeared specifically due to loss of Shrub function . These results showed that loss of Shrub function caused delayed abscission in Drosophila fGSCs and that Shrub is required for completion of abscission in these cells . To test the functional relationship between ALIX and Shrub in fGSCs we performed combined shrub- and alix-RNAi using Nanos-GAL4 . We detected about 55% of the fGSCs in stem cysts ( Fig . 6D-E ) , 15% polyploid fGSCs ( Fig . 6E ) as well as about 40% of egg chambers with 32 germ cells and 15% of egg chambers with more than 32 germ cells ( compared to 3–4% in alix- or shrub-RNAi ) ( S10D-S10E Fig . ) . The increased frequency of egg chambers with more than 32 germ cells suggested an even more delayed abscission rate upon combined ALIX and Shrub depletion than following reduction of either ALIX or Shrub levels alone . Consistently , reducing the Shrub levels in the alix1 mutant background ( shrubG5/+; alix1 ) gave , in addition to stem cysts , rise to the appearance ~10% of germaria with polyploid fGSCs , ~10% of agametic germaria as well as fewer stem cells per germarium than normal suggesting that reduction of the ALIX and Shrub levels in all cell types in the germarium both caused abscission defects and affected germ cell viability ( S10F-S10G Fig . ) . The fact that reducing the levels of both ALIX and Shrub in fGSCs simultaneously caused even more delayed abscission kinetics in fGSCs as compared to decreasing the levels of either of them alone indicated that ALIX and Shrub are required for the same process to promote abscission in Drosophila fGSCs . We next asked whether the complex formation between ALIX and Shrub is important for abscission in fGSCs . In human cells interfering with the interaction between ALIX and CHMP4 causes multi-nucleation and defective midbody morphology [13 , 15] . We introduced point mutations in Drosophila GFP-ALIX ( GFP-ALIX-F198D and GFP-ALIX-I211D ) of residues which have previously been shown to mediate the interaction with CHMP4 in human cells [13 , 15 , 54] . Indeed , we could verify the importance of these residues for the ALIX-Shrub interaction since wild type GFP-ALIX co-precipitated substantially more Shrub than the two mutant proteins in GFP trap analyses ( Fig . 7A ) . To further assess the functional importance of the interaction between ALIX and Shrub in abscission in fGSCs we generated flies expressing GFP-ALIX , GFP-ALIX-F198D or GFP-ALIX-I211D using Nanos-GAL4 either alone or in the alix1 mutant background . Both GFP-ALIX-F198D and GFP-ALIX-I211D localized at MRs and MBs like wild type GFP-ALIX ( Figs . 7B and S11A ) and their expression per se did not induce the formation of stem cysts ( Fig . 7B-C ) . Importantly , wild type GFP-ALIX rescued the fGSC abscission defects in alix1 mutant germaria from 76% of fGSCs in stem cysts to 22% ( Fig . 7C , p < 0 . 05 ) . GFP-ALIX-F198D or GFP-ALIX-I211D could on the other hand not rescue the fGSC abscission defects as 59% and 56% of fGSCs were found in stem cysts following their expression in alix1 mutant germ cells , respectively ( Fig . 7B-C , borderline significant , p = 0 . 05 ) . In agreement , the stem cyst lengths upon expression of GFP-ALIX-F198D or GFP-ALIX-I211D in the alix1 mutant background were similar to the stem cyst lengths in the alix1 mutant , whereas they were shorter upon expression of GFP-ALIX ( S11B Fig . ) . These results suggest that ALIX requires the interaction with Shrub to mediate abscission in fGSCs . Moreover , consistent with the stem cyst phenotypes the expression of wild type GFP-ALIX in the alix1 mutant background rescued the number of egg chambers with 32 cells from 49% to 13% ( p < 0 . 05 ) , whereas neither GFP-ALIX-F198D nor GFP-ALIX-I211D expression in alix1 mutant ovaries could rescue the 32-cell phenotype ( 40% and 39% , respectively , p < 0 . 05 ) ( S11C Fig . ) . Collectively , these results demonstrate that the direct interaction between ALIX and Shrub is required for completion of abscission with normal kinetics in Drosophila fGSCs .
The mechanisms controlling the kinetics of cytokinetic abscission in different cell types in the context of a multi-cellular organism are not well understood . The Drosophila female germline has emerged as a powerful genetically amendable model system to address mechanisms of cytokinetic abscission in vivo [29] . In this study we show that the scaffold protein ALIX and the ESCRT-III component Shrub form a complex to mediate completion of cytokinetic abscission in Drosophila fGSCs with normal kinetics . Loss of ALIX or/and Shrub function or inhibition of their interaction delays abscission in fGSCs leading to the formation of stem cysts in which the fGSC remains interconnected to chains of daughter cells via MRs . As abscission eventually takes place a cyst of e . g . 2 germ cells may pinch off and subsequently undergo four mitotic divisions to give rise to a germline cyst with 32 germ cells [29] . Consistently , loss of ALIX or/and Shrub or interference with their interaction caused a high frequency of egg chambers with 32 germ cells during Drosophila oogenesis . We also found that ALIX controls cytokinetic abscission in both fGSCs and mGSCs and thus that ALIX plays a universal role in cytokinesis during asymmetric GSC division in Drosophila . Taken together we thus provide evidence that the ALIX/ESCRT-III pathway is required for normal abscission timing in a living metazoan tissue . Our results together with findings in other models underline the evolutionary conservation of the ESCRT system and associated proteins in cytokinetic abscission . Specifically , ESCRT-I or ESCRT-III have been implicated in abscission in a subset of Archaea ( ESCRT-III ) [56–58] , in A . thaliana ( elch/tsg101/ESCRT-I ) [59] and in C . elegans ( tsg101/ESCRT-I ) [20] . In S . cerevisiae , Bro1 ( ALIX ) and Snf7 ( CHMP4/ESCRT-III ) have also been suggested to facilitate cytokinesis [60] . In cultured Drosophila cells , Shrub/ESCRT-III mediates abscission and in human cells in culture ALIX , TSG101/ESCRT-I and CHMP4B/ESCRT-III promote abscission [9 , 11 , 13–16] . ALIX and the ESCRT system thus act in an ancient pathway to mediate cytokinetic abscission . Despite the fact that we find an essential role of ALIX in promoting cytokinetic abscission during asymmetric GSC division in the Drosophil a female and male germlines , we did not detect strong bi-nucleation directly attributed to cytokinesis failure in Drosophila alix mutants in the somatic cell types we have examined . This might have multiple explanations . One possibility is that maternally contributed alix mRNA may support normal cytokinesis and development . Whereas ALIX and CHMP4B depletion in cultured mammalian cells causes a high frequency of bi- and multi-nucleation [14 , 15] it is also possible that cells do not readily become bi-nucleate upon failure of the final step of cytokinetic abscission in the context of a multi-cellular organism . Consistent with our observations of a high frequency of stem cysts upon loss of ALIX and Shrub in the germline , Shrub depletion in cultured Drosophila cells resulted in chains of cells interconnected via intercellular bridges/MRs due to multiple rounds of cell division with failed abscission [16] . Moreover , loss of ESCRT-I/tsg101 function in the C . elegans embryo did not cause furrow regression [20] . These and our observations suggest that ALIX- and Shrub/ESCRT-depleted cells can halt and are stable at the MR stage for long periods of time and from which cleavage furrows may not easily regress , at least not in these cell types and in the context of a multi-cellular organism . It is also possible that redundant mechanisms contribute to abscission during symmetric cytokinesis in somatic Drosophila cells . Further studies should address the general involvement of ALIX and ESCRT-III in cytokinetic abscission in somatic cells in vivo . Different cell types display different abscission timing , intercellular bridge morphologies and spatiotemporal control of cytokinesis [10 , 26 , 29] . In fGSCs we found that ALIX and Shrub co-localize throughout late stages of cytokinesis and abscission . In human cells ALIX localizes in the central region of the MB , whereas CHMP4B at first localizes at two cortical ring-like structures adjacent to the central MB region and then progressively distributes also at the constriction zone where it promotes abscission [9–11 , 13 , 15 , 23] . ALIX and CHMP4B are thus found at discrete locations within the intercellular bridge as cells approach abscission in human cultured cells . In contrast , ESCRT-III localizes to a ring-like structure during cytokinesis in Archaea , resembling the Shrub localization at MRs we observed in Drosophila fGSCs [56 , 57] . Moreover , ALIX and Shrub are present at MRs for a much longer time ( from G1/S ) prior to abscission ( in G2 ) in fGSCs than in human cultured cells . Here , ALIX and CHMP4B are increasingly recruited about an hour before abscission and then CHMP4B acutely increases at the constriction zones shortly ( ~30 min ) before the abscission event [9 , 11] . How may ALIX and Shrub be recruited to the MR/MB in Drosophila cells in the absence of CEP55 that is a major recruiter of ALIX and ultimately CHMP4/ESCRT-III in human cells [13 , 15] ? Curiously , we detect a GPP ( 3x ) Y consensus motif within the Drosophila ALIX sequence ( GPPPGHY , aa 808–814 ) resembling the CEP55-interacting motif in human ALIX ( GPPYPTY , aa 800–806 ) . Whether Drosophila ALIX is recruited to the MR/MB via a protein ( s ) interacting with this motif or other domains is presently uncharacterized . Accordingly , alternative pathways of ALIX and ESCRT recruitment have been reported [61–64] , as well as suggested in C . elegans , where CEP55 is also missing [20] . Further studies are needed to elucidate mechanisms of recruitment and spatiotemporal control of ALIX and ESCRT-III during cytokinesis in fGSCs and different cell types in vivo . We found that the direct interaction between ALIX and Shrub is required for completion of abscission with normal kinetics in fGSCs . This is consistent with findings in human cells in which loss of the interaction between ALIX and CHMP4B causes abnormal midbody morphology and multi-nucleation [13 , 15] . Following ALIX-mediated recruitment of CHMP4B/ESCRT-III to cortical rings adjacent to the MR in human cells , ESCRT-III extends in spiral-like filaments to promote membrane scission [9–11 , 13 , 15 , 23] . Due to the discrete localizations of ALIX and CHMP4B during abscission in human cells ALIX has been proposed to contribute to ESCRT-III filament nucleation [15 , 53] . In vitro studies have shown that the interaction between ALIX and CHMP4B may release autoinhibitory intermolecular interactions within both proteins and promote CHMP4B polymerization [54 , 65] . Specifically , ALIX dimers can bundle pairs of CHMP4B filaments in vitro [65] . Moreover , in yeast , the interaction of the ALIX homologue Bro1 with Snf7 ( CHMP4 homologue ) enhances the stability of ESCRT-III polymers [66 , 67] . There is a high degree of evolutionary conservation of ALIX and ESCRT-III proteins [52–54 , 68 , 69] and because ALIX and Shrub co-localize and interact to promote abscission in fGSCs it is possible that ALIX can facilitate Shrub filament nucleation and/or polymerization during this process . Our findings indicate that accurate control of the levels and interaction of ALIX and Shrub ensure proper abscission timing in fGSCs . Their reduced levels or interfering with their complex formation caused delayed abscission kinetics . How cytokinesis is modified to achieve a delay in abscission in Drosophila fGSCs and incomplete cytokinesis in germline cysts is not well understood [25–27] . Aurora B plays an important role in controlling abscission timing both in human cells and the Drosophila female germline [29 , 70 , 71] . During Drosophila germ cell development Aurora B contributes to mediating a delay of abscission in fGSCs and a block in cytokinesis in germline cysts [29] . Bam expression has also been proposed to block abscission in germline cysts [29 , 32 , 72 , 73] . It will be interesting to investigate mechanisms regulating the levels , activity and complex assembly of ALIX and Shrub and other abscission regulators at MRs/MBs to gain insight into how the abscission machinery is modified to control abscission timing in fGSCs . We found that intercellular bridge MTs in fGSC-CB pairs were degraded in G1/S when the fusome adopted bar morphology . Abscission in G2 thus appears to occur independently of intercellular bridge MTs in Drosophila fGSCs . This has also been described in C . elegans embryonic cells where the MR scaffolds the abscission machinery as well as in Archaea that lack the MT cytoskeleton [20 , 56 , 57] . In mammalian and Drosophila S2 cells in culture , on the other hand , intercellular bridge MTs are present until just prior to abscission [9 , 11] . It is interesting to note a resemblance of the stem cysts that appeared upon loss of ALIX and Shrub function to germline cysts in that the MRs remained open for long periods of time similar to RCs . Some modification of ALIX and Shrub levels/recruitment may thus contribute to incomplete cytokinesis in Drosophila germline cysts under normal conditions . Because we detected stem cysts in the case when ALIX weakly interacted with Shrub it is also possible that inhibition of their complex assembly/activity may contribute to incomplete cytokinesis in germline cysts . Abscission factors , such as ALIX and Shrub , may thus be modified and/or inhibited during incomplete cytokinesis in germline cysts . Such a scenario has been shown in the mouse male germline where abscission is blocked by inhibition of CEP55-mediated recruitment of the abscission machinery , including ALIX , to stable intercellular bridges [26 , 30] . Altogether our data thus suggest that ALIX and Shrub are essential components of the abscission machinery in Drosophila GSCs , and we speculate that their absence or inactivation may contribute to incomplete cytokinesis . More insight into molecular mechanisms controlling abscission timing and how the abscission machinery is modified in different cellular contexts will give valuable information about mechanisms controlling complete versus incomplete cytokinesis in vivo . Summarizing , we here report that a complex between ALIX and Shrub is required for completion of cytokinetic abscission with normal kinetics during asymmetric Drosophila GSC division , giving molecular insight into the mechanics of abscission in a developing tissue in vivo .
Fly crosses and experiments were performed at 25°C unless noted otherwise . w1118 was used as a wild type control . w1118 , w1118; Nanos-GAL4 , UAS-Dcr-2 , w1118; Nanos-GAL4 , y v; attP2 , TRiP-alix ( UAS-shRNA-alix , chr 3 , TRiP# HMS00298 ) , TRiP-shrub ( UAS-shRNA-shrb , chr 2 , TRiP# HMS01767 ) [45] , MTD-GAL4 [45] , yw; P{EPgy2}ALiXEY10362 , Df ( 3R ) BSC499/TM6C , Sb , w1118; Df ( 3R ) BSC739/TM6C , Sb and w*; shrbG5 P{neoFRT}42D/CyO , P{GAL4-twi . G}2 . 2 , P{UAS-2xEGFP}AH2 . 2 ( shrubG5/+ ) were from BDSC ( Indiana University ) and PBac{WH}ALiXf03094 from Exelixis at Harvard Medical School ( referred to as alix1 ) . The alix3 allele was generated by imprecise excision of the P-element of the yw; P{EPgy2}ALiXEY10362 line . The breakpoints were determined by sequencing and this allele lacks 860 bp in the 5’ end of the gene ( nucleotides 23534881 to 23525741 on 3R missing ) , thus removing the alix gene start codon , exons 1 , 2 and most of exon 3 . The FRT82B , alix1 and FRT82B , alix3 lines were generated by recombination of to FRT82B chromosomes by standard procedures . The generation of the genomic alix rescue lines is described below . The alix1 and alix3 alleles were kept as stocks balanced over TM6B , Tb and TM6B , dfd gfp chromosomes . UASp-GFP-Shrub and Nanos-GAL4 , UASp-GFP-Shrub were generated as described in [55] . Bam-GAL4 ( chr 3 ) was a kind gift from M . Fuller ( Stanford School of Medicine , CA ) , hsflp , tubulin-GAL4 , UAS-GFP; FRT82 , tubulin-GAL80/TM6B , Tb ( MARCM82 ) was a kind gift from M . Peifer ( University of North Carolina ) . ALIX ( CG12876 ) antibodies were generated by immunizing a guinea pig with two peptides ( CIQSTYNGASEEEKG-/CERLLDEERDSDNQL-amide ) ( BioGenes ) from the Bro1 domain . Primary antibodies and dilutions for immunofluorescence ( IF ) or Western blot ( WB ) were guinea pig anti-ALIX ( IF: 1:1000–3000 , WB: 1:1000 ) , mouse anti-ALIX ( WB: 1:1000 , a kind gift from T . Aigaki , Tokyo Metropolitan University , Japan ) , rabbit anti-Cindr ( IF: 1:1000 ) ( Haglund et al . , 2010 ) , mouse anti-hts-F ( IF: 1:50 , 1B1 , DSHB ) , rabbit anti-Shrub ( WB: 1:1000 , a kind gift from F-G . Bao , University of Massachusetts Medical School , MS [74] , mouse anti-α-spectrin ( IF: 1:25 , 3A9 , DSHB ) , goat anti-Vasa ( IF: 1:100 , dC-13 , Santa Cruz Biotechnology ) , rabbit anti-Nanos ( IF: 1:1000 , a kind gift from A . Nakamura , RIKEN Center for Developmental Biology ) , mouse anti-Bam ( IF: 1:10 , DSHB ) , mouse anti-α-tubulin ( WB: 1:10 , 000 , Sigma ) , sheep anti α-tubulin ( IF: 1:250 , Cytoskeleton ) , guinea pig anti-Cnn ( IF: 1:500 , a kind gift from T . C . Kaufman , Indiana University ) , rabbit anti-phospho-Histone H3 ( IF: 1:500 , Millipore ) , rabbit anti-phosphotyrosine ( IF: 1:500 , Sigma ) , mouse anti γ-tubulin ( IF: 1:500 , Sigma ) , mouse anti-Fasciclin III ( FasIII , IF: 1:50 , 7G1 , DSHB ) , rabbit anti-phospho-Smad1/5 ( Ser463/465 ) ( IF: 1:100 , 41D10 , Cell Signaling ) . GFP–Booster_Atto488 ( IF: 1:200 ) was from Chromotek . To visualize F-actin , Alexa Fluor 647 phalloidin ( 1:50 ) , Alexa Fluor 488 phalloidin ( 1:100 ) or rhodamine phalloidin ( 1:400 ) ( Molecular Probes ) were included in secondary antibody incubations . Secondary antibodies were conjugated to Alexa Fluor 488 , Alexa Fluor 594 ( 1:200 , Molecular Probes ) , Cy3 or Cy5 ( 1:500 , Jackson Immunoresearch ) . DNA was stained using Hoechst 33342 ( 1μg/μl , Invitrogen ) . pOT2-ALIX as well as pAGW and pPGW vectors were from the Drosophila Genomics Resource Center ( DGRC ) ( Bloomington , IN ) . pAc-Shrub-GFP was a kind gift from T . Takeda and D . Glover ( University of Cambridge , Cambridge , UK ) . S2 GFP-α-tubulin cells were a kind gift from E . Griffis ( University of Dundee , UK ) and S2 cells were from ATCC ( CRL-1963 ) ( a kind gift from R . Palmer , Umeå University , Sweden ) . Drosophila D . Mel-2 ( Dmel ) cells ( a kind gift from P . P d’Avino and D . Glover , University of Cambridge , Cambridge , UK ) were grown in Express Five SFM medium ( Gibco ) containing 2 mM L-glutamine , 100 U/ml penicillin and 100µg/ml streptomycin and Drosophila Schneider 2 ( S2 ) cells were cultured in Schneider’s Drosophila Medium ( Gibco ) supplemented with 10% fetal calf serum , 2 mM L-glutamine , 100 U/ml penicillin and 100µg/ml streptomycin . S2 cells were seeded on coverslips for two hours before 12 min fixation at room temperature in 4% formaldehyde ( EM grade , Polysciences ) in PHEM buffer ( 60 mM Pipes pH 6 . 8 , 25 mM Hepes pH 7 . 0 , 10 mM EGTA pH 8 . 0 , 4 mM MgSO4 ) . The cells were then washed three times with PBS and incubated in PBS + 5% BSA + 0 , 1% Triton X-100 ) for at least 1 h . Primary antibodies were diluted in PBS + 1% BSA + 0 , 1% Triton X-100 ( PBT ) and cells incubated with primary antibodies over night at 4 degrees . Cells were then washed twice in PBT for 15 min and then incubated with secondary antibodies diluted in PBT for 2 hrs at room temperature . They were then washed twice in PBT as before followed by incubation with Hoechst 33342 diluted in PBS to 1μg/μl for 5 min . Cells were finally rinsed with PBS and mounted in Mowiol . Ovaries or testes were dissected in PBS and fixed using 4% formaldehyde ( EM grade , Polysciences ) for 30 min either on ice ( all samples including anti-Cindr antibodies ) or at room temperature ( RT ) ( prior to anti-α-tubulin staining ) . Tissues were subjected to permeabilization ( 3 × 15min ) and blocking ( 30 min ) in PBS + 0 . 3% bovine serum albumin ( BSA ) + 0 . 3% Triton X- 100 ( PBT ) at RT and then incubated with primary antibodies diluted in PBT at 4°C over night . Samples were then washed three times 15 min in PBT , incubated with secondary antibodies diluted in PBT for 2 hrs at room temperature followed by three 15 min washes in PBT . For DNA staining , samples were subsequently stained with Hoechst 33342 ( 1 μg/μl ) diluted in PBS for 10 min . Samples were mounted in anti-fading mounting medium ( Prolong Antifade , Molecular Probes or Vectashield , Vector laboratories ) . For anti-ALIX staining , ovaries were fixed in ice-cold methanol for 7 min and subsequently stained as above with the addition of GFP-Booster ( 1:200 ) in the secondary antibody solution . For p-Mad detection the ovaries were fixed for 40 min in 4% formaldehyde with phosphatase inhibitor cocktail ( Sigma , 1:200 ) and stained according to the protocol by Luo et . al . [75] . Images were captured using Zeiss LSM 780 , Zeiss LSM 710 or Zeiss LSM 5 DUO laser scanning confocal microscopes ( Carl Zeiss , Inc . ) equipped with NeoFluar 63×/1 . 4 NA and 100×/1 . 45 NA oil immersion and Plan Apochromat 20×/0 . 8 NA objectives at 20°C . Image processing and analysis were done using the Zeiss LSM 510 ( Version 3 . 2 , Carl Zeiss , Inc . ) and Zen 2009 softwares and Adobe Photoshop CS4 ( Adobe ) . Images are planar projections of sections from z-stacks of germaria unless otherwise noted . Ovaries of 2 to 4-day-old females ( unless otherwise noted ) that had been fed with yeast paste and kept with a couple of males for 2 days were dissected , fixed and stained with antibodies to visualize the fusome ( hts-F ) , MRs/MBs ( Cindr ) , RCs ( pTyr ) /F-actin ( fluorescently labeled phalloidin ) and nuclei ( Hoechst ) as described above . Confocal z-stacks of germaria were acquired at the confocal microscope and fGSC phenotypes were analyzed from z-stacks and three-dimensional reconstructions of z-stacks . fGSC identity was determined based on its anterior localization in the germarium , its fusome morphology and contact with the cap cells . Phenotype scoring was based on the fusome morphology , presence , absence , number and position of Cindr-positive MRs/MBs , cell-cell boundaries and nuclei . We categorized fGSCs into normal morphologies: ( i ) fGSCs with a spherical spectrosome , ( ii ) fGSC-CB pairs with an MR ( includes plug , bar , dumbbell and fusing fusome morphologies ) and ( iii ) fGSC-CB pairs in abscission with an MB between them ( exclamation point fusome ) as well as abnormal abscission-defective morphologies: ( iv ) linear chains of cells interconnected via fusome and MRs , ( v ) branched chains or ( vi ) polyploid , bi- or multinucleate fGSCs . Egg chamber phenotypes were scored at the microscope based on the number of RCs to the oocyte and the number of germ cell nuclei . To examine whether differences between controls and alix1 or alix3 germaria were significant within experiments , each germarium was classified as either normal or non-normal ( the latter being the case if at least one non-normal phenotype was present—linear , branched or polyploid ) . Fisher’s exact test was then used to determine significance . To test whether differences of fGSC phenotypes ( classified as above ) or egg chamber phenotypes between Nos-GAL4/GFP-ALIX and alix1 , Nos-GAL4/GFP-ALIX-F198D; alix1 or Nos-GAL4/GFP-ALIX-I211D; alix1 ovaries were significant we used a mixed factor model with each experiment as random factor . To generate N-terminally GFP-tagged alix , a PCR fragment corresponding to the whole-length alix cDNA ( except the START codon ) was amplified from a cDNA clone from the BDGP Gold cDNA Collection ( DGRC , Bloomington , IN ) using the primers 5’AATGGATCCGGTCGAAGTTTCTGGGCGTGCCG3’ and 5’AATGCGGCCGCTTACCAGCCAGGTGGCTTCTG3’ and the Phusion High-Fidelity PCR Kit ( New England Biolabs ) . The alix cDNA was purified using the QIAquick PCR Purification Kit ( Qiagen ) , and cloned into the pENTR1A Gateway entry vector using the T4 DNA ligase ( Roche ) . The alix gene was then transferred by LR recombination using the Gateway LR clonase II enzyme mix ( Invitrogen ) to the pPGW ( for generation of fly lines ) or pAGW ( for cell lines ) destination vectors ( DGRC , Bloomington , IN ) . Site-directed in vitro mutagenesis was used to introduce point mutations in the pENTR1A-alix vector using primers containing the specific mutations and the Phusion High-Fidelity PCR Kit ( New England Biolabs ) . For generating ALIX-F198D the primers 5’CCAAGCGCAGGAGGTTGACATTCTGAAGGCAATTAAGG3’ and 5’CCTTAATTGCCTTCAGAATGTCAACCTCCTGCGCTTGG3’ were used , and for ALIX-I211D the primers 5’CTTGAAGGACCAGGACATCGCCAAGCTTTGCTGC3’ and 5’GCAGCAAAGCTTGGCGATGTCCTGGTCCTTCAAG3’ were used . The plasmid was then treated with DpnI ( New England Biolabs ) for one hour at 37°C after PCR amplification . The mutated alix cDNAs were then transferred to the pPGW ( for fly lines ) and pAGW ( for cell lines ) destination vectors by LR recombination The transgenic UASp-GFP-ALIX , UASp-GFP-ALIX-F198D and UASp-GFP-ALIX-I211D Drosophila lines were generated by P-element transformation performed by BestGene Inc . The expression of GFP-ALIX was verified by Western blot analysis . Approximately 1 hour before transfection , 8×106 Dmel cells were seeded in 10 cm plates . The cells were transiently transfected for 48 hours with 2 , 5 µg pAGW ( empty GFP ) or 5 µg pAc-Shrub-GFP , pAGW-ALIX-wt , pAGW-ALIX-F198D or pAGW-ALIX-I211D using FuGene HD according to the manufacturer’s instructions ( Promega ) . Enrichment of mitotic cells was obtained by MG132 treatment ( 25 µM , 5 hours ) as previously described [76] . Cells were used for GFP trap immunoprecipitation analysis performed in line with the protocol provided by the supplier ( ChromoTek ) . The cells were lysed in 200 µl Lysis buffer ( 10 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA , 0 . 5% NP-40 ) supplemented with 1:50 protease inhibitor cocktail ( Roche ) , 1:50 phosphatase inhibitor cocktail 2 ( Sigma-Aldrich ) and 2 mM N-ethylmalemide ( Sigma-Aldrich ) on ice for 30 minutes with extensive mixing every 10 minutes . Nuclei and cell debris were cleared by centrifugation ( 20 , 000g , 10 minutes , 4°C ) , before the lysate was diluted to 1000 µl with Washing buffer ( 10 mM Tris-HCl pH = 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA ) and incubated with pre-washed GFP trap beads ( 30 µl ) for 1 hour at 4°C . The beads and associated proteins were washed three times using Washing buffer and next boiled in SDS sample buffer containing 100 mM DTT for 10 minutes to elute associated proteins . The eluted proteins were subjected to SDS-PAGE , followed by Western blot to detect ALIX , Shrub or GFP . Drosophila tissues were collected and homogenized in ice-cold lysis buffer ( 50 mM Tris pH 8 , 150 mM NaCl , 0 . 5% NP-40 or 50 mM Hepes , 150 mM NaCl , 1 mM EDTA , 1 mM EGTA , 10% glycerol , 1% Triton X-100 , 25 mM NaF , 10μM ZnCl2 ) containing protease inhibitor cocktail ( Complete , EDTA-free , Roche ) . Lysates were cleared by centrifugation for 15 min at 13 , 000 rpm and 4°C . Equal amounts of protein were mixed with Laemmli buffer containing 50 mM DTT , denatured by boiling and subjected to SDS-PAGE and transferred to either nitrocellulose or PVDF membranes . Nitrocellulose membranes were blocked in PBS/5% milk at 4°C over night followed by incubation with primary antibodies diluted in PBS/5% BSA for 1h 30 min or over night . Membranes were then washed three times in PBS/0 . 01% Tween-20 , followed by incubation 1 h with secondary HRP-conjugated anti-rabbit and anti-mouse antibodies ( 1:5000 ) ( Jackson ImmunoResearch ) . Following three further washes in PBS/0 . 01% Tween-20 and one wash in PBS , chemiluminescent ( WestPico , PIERCE ) signal was detected on film ( Amersham Hyperfilm ) . PVDF membranes were blocked ( by drying ) , re-wet in PBS/0 . 01% Tween-20 , incubated with primary antibodies overnight at 4°C , rinsed three times in PBS/0 . 01% Tween-20 , incubated with fluorescently labelled secondary antibodies ( LI-COR Biosciences GmbH ) and washed twice in PBS/0 . 01% Tween-20 and once in PBS followed by scanning using the Odyssey Developer ( LI-COR Biosciences GmbH ) . For RNAi-mediated gene silencing in germ cells , MTD-GAL4 , Nanos-GAL4 , UAS-Dicer; Nanos-Gal4 or Bam-GAL4 drivers were crossed to control ( yv; attP2 ) , TRiP-alix-RNAi , TRiP-shrub-RNAi or TRiP-shrub-RNAi; TRiP-alix-RNAi flies as described . For all RNAi experiments , young female offspring were fed with yeast paste and kept with a couple of males for 2 days at 25°C . Ovaries of 2–4-day-old females were dissected , fixed , and stained as described above . 0–3 day old males were dissected and stained with antibodies as described above to label midbody rings and midbodies ( Cindr ) , the fusome ( α-spectrin ) , hub ( Fasciclin III ) and germ cells ( Vasa ) . Confocal z-stacks of testes tips were acquired at the confocal microscope and mGSC phenotypes were analyzed from z-stacks and three-dimensional reconstructions of z-stacks . mGSC identity was determined based on proximity to the hub . Phenotype scoring was based on the fusome morphology , presence , absence , number and position of Cindr-positive MRs/MBs , Vasa staining and nuclei . For clonal analysis in the follicle cell epithelium , MARCM82 females were crossed to FRT82 and FRT82 , alix3/TM6B , Tb and FRT82 , alix1/TM6B , Tb males . L3 larvae were subjected to two heat-shocks at 37°C for 1 h . Newly hatched females were fed with yeast paste for 2 days in the presence of a couple of males . Ovaries were then dissected and stained to visualize F-actin and nuclei ( Hoechst ) as described above . Genomic rescue constructs ( BAC CH322–119C06 , comprising 20339 bp from 23513227 to 23533565 of chromosome arm 3R ( short-alix-rescue , alix-s ) , and BAC CH321–50C24 comprising 85562 bp from 23500943 to 23586504 of chromosome arm 3R ( long-alix-rescue , alix-l ) in the vector attB-P[acman]-CmR-BW ( http://bacpac . chori . org/home . htm ) were injected into strains y1 w1118; PBac{y+-attP-9A}VK00018 ( BDSC# 9736 , insertion site 53B2 ) and y1 w1118; PBac{y+-attP-3B}VK00037 ( BDSC# 9752 , insertion site 22A3 ) and integrated into predetermined attP docking sites in the genome using PhiC31 integrase-mediated germline transformation . The methodology is decribed in “Versatile P[acman] BAC libraries for transgenesis studies in Drosophila melanogaster” [77] . The injection of the constructs into Drosophila embryos was performed by BestGene ( http://www . thebestgene . com/ ) . Males with integrated constructs were obtained from BestGene , balanced and crossed to generate CH322–119C06/CyO; alix1/TM6B , Tb ( alix-s/CyO; alix1/TM6B , Tb ) , CH322–119C06/CyO; alix3/TM6B , Tb ( alix-s/CyO; alix3/TM6B , Tb ) , CH321–50C24/CyO; alix1/TM6B , Tb ( alix-l/CyO; alix1/TM6B , Tb ) and CH321–50C24/CyO; alix3/TM6B , Tb ( alix-l/CyO; alix3/TM6B , Tb ) stocks . For complementation tests the alix1 and alix3 alleles were crossed to the deficiencies and to each other . In both rescue analyses and complementation tests , young females of the indicated genotypes were collected , fed with yeast paste and kept with a couple of males for 2 days . Ovaries of 2–4 day-old flies were dissected , fixed and stained to visualize F-actin and nuclei and egg chamber phenotypes were quantified as described above . Flies used for fertility tests were 4–7 days old and kept separately with yeast paste for a couple of days before being crossed . Wild type or alix1 mutant virgin females were crossed to wild type or alix1 mutant males as indicated . Eggs were collected on apple juice agar plates for 18 hours three times for each cross in three independent experiments . The eggs were counted after each egg lay to determine the egg lay rate . Hatch rates were determined by quantifying the hatched versus unhatched eggs under a dissecting microscope after eggs had developed for 24–30 hours . The experiments were conducted at 25°C . Embryo collection , permeabilization and fixation were based on the protocol described by Rothwell and Sullivan [78] . The Drosophila melanogaster flies were put on apple juice agar with yeast for egg lay at 25°C overnight . The embryos were dislodged from the agar into a nylon mesh/falcon basket using PBS + 0 . 02% Triton X-100 , and dechorionated by shaking them in a 50% commercial bleach solution until agglutination of the embryos ( 1–3 min ) . The dechorionated embryos were extensively rinsed with PBS + 0 . 02% Triton X-100 , and blotted dry on paper towels . The embryos were transferred from the nylon mesh and to a small flask with 5 mL heptane . An equal amount of 4% formaldehyde in PBS was added , and the two-phase mixture was incubated with vigorous shaking for 17 minutes . The embryos were now between the two phases . The formaldehyde phase was removed and replaced with methanol , and the embryos were gently shaken for 1 minute with gentle heating for removal of the vitelline membrane . The heptane phase was removed along with the embryos still remaining in the interphase . The embryos that sank to the bottom of the flask were washed three times in methanol and stored at -20°C . Immunofluorescent staining of embryos was performed as follows . The embryos were rehydrated by first putting them in 3:4 methanol and 1:4 4% formaldehyde in PBS for 2 minutes , and then 1:4 methanol and 3:4 formaldehyde for 5 minutes . Post fixation was done for 10 minutes in 4% formaldehyde , before the embryos were rinsed six times using PBS with 1% BSA and 0 . 05% Triton X-100 . The embryos were incubated with α-spectrin antibodies ( 1:25 , DSHB ) over night at 4°C . After incubation the embryos were rinsed three times and washed for one hour with PBS with 1% BSA and 0 . 05% Triton X-100 , and then incubated with secondary antibody for two hours . The antibodies were diluted in PBS with 1% BSA and 0 . 05% Triton X-100 . The embryos were again rinsed three times and washed for one hour with PBS with 1% BSA and 0 . 05% Triton X-100 , before being labeled with Hoechst 33342 ( 2 µL/mL ) for 10 minutes , and then rinsed 3 times in PBS to remove detergent . The embryos were mounted using Vectashield ( Vector laboratories ) . For quantifications of mono- and binucleate cells , images of homozygous wild type , alix1 and alix3 mutant stage 16 embryos were captured at the confocal microscope and more than 1000 cells analyzed for each genotype . | Cytokinesis , the final step of cell division , concludes with a process termed abscission , during which the two daughter cells physically separate . In spite of their importance , the molecular machineries controlling abscission are poorly characterized especially in the context of living metazoan tissues . Here we provide molecular insight into the mechanism of abscission using the fruit fly Drosophila melanogaster as a model organism . We show that the scaffold protein ALIX and the ESCRT-III component Shrub are required for completion of abscission in Drosophila female germline stem cells ( fGSCs ) . ESCRT-III has been implicated in topologically similar membrane scission events as abscission , namely intraluminal vesicle formation at endosomes and virus budding . Here we demonstrate that ALIX and Shrub co-localize and interact to promote abscission with correct timing in Drosophila fGSCs . We thus show that ALIX and ESCRT-III coordinately control abscission in Drosophila fGSCs cells and report an evolutionarily conserved function of the ALIX/ESCRT-III pathway during cytokinesis in a multi-cellular organism . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | ALIX and ESCRT-III Coordinately Control Cytokinetic Abscission during Germline Stem Cell Division In Vivo |
DNA looping mediated by transcription factors plays critical roles in prokaryotic gene regulation . The “genetic switch” of bacteriophage λ determines whether a prophage stays incorporated in the E . coli chromosome or enters the lytic cycle of phage propagation and cell lysis . Past studies have shown that long-range DNA interactions between the operator sequences OR and OL ( separated by 2 . 3 kb ) , mediated by the λ repressor CI ( accession number P03034 ) , play key roles in regulating the λ switch . In vitro , it was demonstrated that DNA segments harboring the operator sequences formed loops in the presence of CI , but CI-mediated DNA looping has not been directly visualized in vivo , hindering a deep understanding of the corresponding dynamics in realistic cellular environments . We report a high-resolution , single-molecule imaging method to probe CI-mediated DNA looping in live E . coli cells . We labeled two DNA loci with differently colored fluorescent fusion proteins and tracked their separations in real time with ∼40 nm accuracy , enabling the first direct analysis of transcription-factor-mediated DNA looping in live cells . Combining looping measurements with measurements of CI expression levels in different operator mutants , we show quantitatively that DNA looping activates transcription and enhances repression . Further , we estimated the upper bound of the rate of conformational change from the unlooped to the looped state , and discuss how chromosome compaction may impact looping kinetics . Our results provide insights into transcription-factor-mediated DNA looping in a variety of operator and CI mutant backgrounds in vivo , and our methodology can be applied to a broad range of questions regarding chromosome conformations in prokaryotes and higher organisms .
Looping between two DNA sites , mediated by transcription factors , is a ubiquitous mechanism in prokaryotic transcription regulation [1] . DNA looping brings two distal DNA sites into close proximity , enhancing interactions between transcription factors bound at separate sites or bringing transcription factors close to RNA polymerase at the promoter . Knowing when and how DNA loops in vivo is important to understand the role of DNA looping in gene regulation and cell decision-making; some studies found molecular details of gene regulation have little influence on gene expression [2]–[4] , while others suggested that DNA looping could trigger cell phenotype switching [5] and influence fluctuations in transcription activity [6] . DNA looping was first suggested for the transcription factor AraC ( accession number P0A9E0 ) in the E . coli arabinose operon . Disruption of an AraC binding site ∼280 bp upstream of the promoter reduced AraC-mediated repression nearly 10-fold , indicating a long-range interaction between the promoter and upstream DNA [7] . Subsequently , DNA looping mediated by transcription factors LacI [8] ( accession number P03023 ) , DeoR [9] ( accession number P0ACK5 ) , NtrC [10] ( accession number P0AFB8 ) , GalR [11] ( accession number P03024 ) , and bacteriophage λ repressor CI [12] , [13] was reported . The length of the intervening DNA in these loops can be as short as 58 bp ( lac operon [8] ) or as long as ∼5 kilobases ( deo operon [9] ) . Biochemical , biophysical , and genetic studies have established important roles of DNA looping in transcription regulation . However , transcription-factor-mediated DNA looping on the length scale of a few kilobases in prokaryotic cells has not been directly visualized in vivo , and the in vivo dynamics of DNA looping are difficult to investigate . Chromosome conformation capture ( 3C ) has been used to detect juxtaposition of DNA sites separated by hundreds of kilobases in both eukaryotic and prokaryotic cells [14] , [15] , but high background of interactions at the kilobase scale limits the utility of these methods in studying typical prokaryotic DNA loops [16] . An in vivo imaging method using fluorescent proteins fused to DNA-binding proteins bound to tandem arrays of hundreds of binding sites has been employed to visualize homologous chromosome pairing in yeast induced by double-strand breaks [17]; however , an array of several kilobases of binding sites makes this method unsuitable for studying DNA loops of only a few kilobases . In addition , the long array of tightly bound protein molecules may be detrimental to cells [18] . We developed a two-color , high-resolution imaging method to directly measure the end-to-end separation of two DNA sites 2 . 3 kb apart in live E . coli cells ( Figure 1a ) . This method is based on the ability to precisely determine the location of a specific DNA site in vivo [19] . By expressing a fluorescent protein in fusion with a DNA-binding protein in a cell with only three tandem binding sites ( spanning less than 100 bp ) , the resulting fluorescent spot is diffraction-limited , and the location of the binding site can be determined with sub-diffraction-limited precision by fitting its fluorescence profile to a two-dimensional Gaussian function [20] . By labeling two ends of a DNA segment with two unique sets of binding sequences and co-expressing corresponding fluorescent DNA-binding fusion proteins of different colors , the distance between the two DNA sites can be determined with a precision of a few tens of nanometers . An in vitro experiment employing the same principle measured intramolecular distances using organic dyes [21] , but this approach has not been demonstrated in vivo with comparable resolution using fluorescent proteins . We used our method to probe the mechanisms and dynamics of DNA looping mediated by the bacteriophage λ repressor CI [22] in live E . coli cells and investigate its regulation of transcription from the CI promoter PRM . The λ repressor CI is an essential transcription factor in determining the fate of an E . coli cell infected by the bacteriophage λ . When CI is expressed , it represses lytic promoters to commit to an extraordinarily stable lysogenic state that persists for millions of generations [23]–[25] . However , upon induction by UV irradiation or other specific events , CI degradation can trigger an irreversible switch from lysogenic to lytic gene expression within one cell generation time [26] . The robustness of the λ regulatory circuit has been extensively studied . Among many important features of the system such as promoter-operator arrangement [27] , [28] , CI autoregulation [3] , [29] , [30] , and cooperative binding [31]–[34] , DNA looping between the homologous rightward and leftward operators OR and OL , separated by 2 . 3 kb , was shown to play significant , fate-determining roles in the λ lifecycle [13] , [35] . Cooperative binding of CI dimers at the subsites OR1 and OR2 of OR represses the lytic promoter PR ( reviewed in [36] ) and simultaneously activates CI's own promoter , PRM , by accelerating transcription initiation [37]–[39] . At higher CI concentrations , an additional CI dimer binds to OR3 and represses PRM [40] . As illustrated in Figure 1a , an octameric CI complex ( with or without an additional CI tetramer ) can mediate DNA looping by bridging OR and OL . These higher-order complexes result from interactions between CI dimers bound to subsites at OR123 and OL123 , and were first identified in vitro by ultracentrifugation [41] and later visualized by EM [12] and AFM [42] . Looping dynamics were investigated in vitro using tethered particle motion ( TPM ) [43]–[46] . To gain quantitative insight into the relationship between CI-mediated DNA looping and transcription regulation , thermodynamic models and numerical simulations were developed [33] , [35] , [44] , [47]–[52] . Key parameters in these studies were the free energies of octameric and tetrameric CI interactions that mediate DNA looping [35] . These free energies specify the DNA looping probability at a given condition ( temperature , CI concentration , etc . ) and hence the extent to which distal DNA sites affect each other . To date , DNA-looping probabilities and free energies were either estimated indirectly in in vivo studies by measuring PRM and PR activities in various operator mutants with a priori assumptions of DNA looping states [35] , [49] , [51] or measured using purified components in vitro , where conditions differ from those in a cellular environment [42]–[46] . Consequently , these studies yielded varying estimates for the free energies of DNA looping and the degree to which DNA looping influences PRM activity . Hence , the roles of CI-mediated DNA looping in transcription regulation are still in debate [13] , [35] , [49] , [51] , [53] . In this study , we tracked the apparent separation between the OR and OL sites on a λ DNA segment ( termed OR–OL DNA below ) in real time in live E . coli cells , from which we obtained the first direct estimates of in vivo looping frequencies and kinetics for both wild-type DNA and for DNA carrying mutations in OR3 and OL3 . We also measured corresponding CI expression levels in these strains by counting the number of CI transcripts in individual cells . Applying these independent , in vivo measurements to a thermodynamic model , we were able to obtain looping free energies and quantify the influence of DNA looping on PRM expression . Furthermore , we discuss how the compaction of the E . coli chromosome may impact DNA looping kinetics . The methodology established in this work can be extended to a broad range of questions regarding chromosomal DNA conformation and/or gene activities in prokaryotes and higher organisms .
We inserted the construct shown in Figure 1a into the E . coli chromosome . It contains three tandem tetO sites ( tetO3 ) [54] and three tandem lacOsym sites ( lacO3 ) [55] flanking the wild-type λ lysogen sequence from OR to OL ( including the PR , PRM and PL promoters and the cI , rexA ( accession number P68924 ) and rexB ( accession number P03759 ) genes ) . In this construct , called λWT , CI is expressed from PRM and regulates its own expression . The lacO-binding and tetO-binding proteins LacI and TetR ( accession number P04483 ) were fused with red and yellow fluorescent proteins to generate LacI-mCherry and TetR-EYFP , and were expressed from an inducible plasmid ( Figure 1b ) . With the combination of strong induction , weak ribosome binding sites , and carefully controlled growth , we achieved sufficiently low LacI-mCherry and TetR-EYFP expression levels to detect distinct , diffraction-limited mCherry and EYFP spots in single cells . We then fit the fluorescence intensity profile of each individual spot with a two-dimensional Gaussian function to estimate its centroid position . The average localization precisions for individual spots of LacI-mCherry and TetR-EYFP were 17 and 14 nm , respectively ( Figure S1a ) . Subsequently , we transformed EYFP coordinates into mCherry coordinates using fiducial data to calculate the vector between the mCherry and EYFP spots arising from LacI-mCherry and TetR-EYFP protein molecules bound to the same OR–OL DNA segment . We called this vector ( Figure 1c ) . The magnitude of the vector , , is the two-dimensional projection of the distance between lacO3 and tetO3 onto the image plane; on average , it is proportional to the end-to-end distance between lacO3 and tetO3 in three dimensions . The total error for an measurement , including fitting errors in determining centroid of individual spots ( Figure S1a ) , registration errors in aligning EYFP and mCherry two-color images ( ∼10 nm based upon experiments using fluorescent beads ) , and contributions from local fluorescent background , was on average ∼40 nm ( see below ) . With very low TetR-EYFP and LacI-mCherry expression , it was inevitable that not all lacO3 and tetO3 sites were bound by fusion protein molecules . Furthermore , not all fusion protein molecules were fluorescent due to stochastic chromophore maturation . Figure 2a contains typical data showing that a subset of cells was successfully labeled at both sites . We analyzed all cells having distinct fluorescent spots in both emission channels to calculate . We expected to decrease when DNA between lacO3 and tetO3 looped . To determine whether our two-color imaging method was sufficient to distinguish between looped and unlooped DNA in the crowded intracellular environment , we constructed two control strains ( Table 1 ) . In the positive control λnull , the centers of lacO3 and tetO3 sites are separated by 66 bp ( Figure 1d ) . The outmost lacOsym and tetO sites are separated by less than 40 nm ( Figure S2a ) . The close proximity of lacO3 and tetO3 mimicked permanently looped DNA . In the negative control λΔOL , we inserted the λ sequence from OR up to but not including OL between lacO3 and tetO3 ( Figure 1e ) . The resulting λΔOL DNA has comparable length as the wild-type λ DNA , but CI-mediated DNA looping between OR and OL is abolished . We first examined λnull and λΔOL in two-color fluorescence images to determine whether we could discriminate between looped and unlooped DNA by eye . We obtained at least sixty 20-frame movies ( 100 ms exposures; 2 s total ) for each strain in each of three independent experiments . Typical fluorescence images are shown in Figure 2a and b . Crosstalk between the two emission channels was negligible , as bright mCherry and EYFP spots only appeared in the corresponding channel but not the other . Figure 2c and d show 1 s of typical data for individual λnull and λΔOL spots . Representative movies for the two strains and others discussed below are included as Movies S1 , S2 , S3 , S4 , S5 , S6 . As expected for a permanently looped configuration , the positive control λnull exhibited overlapping EYFP and mCherry spots ( Figure 2c ) . Generally , λΔOL molecules did not exhibit spot separation that was easily identifiable by eye ( Figure 2d ) . However , some λΔOL molecules displayed large displacements between the LacI-mCherry and TetR-EYFP spots that were distinguishable by eye ( Figure 2e ) ; such images were not observed for λnull . Visual inspection of the apparent separation between the LacI-mCherry and TetR-EYFP spots suggested that comparing the end-to-end separation in OR–OL DNAs required a more quantitative approach . We calculated for all OR–OL DNA molecules in the λΔOL and λnull strains that exhibited fluorescent spots in both EYFP and mCherry images . Figure 2f–h shows calculations for movies in Figure 2c–e , respectively , and Figure S3 shows vectors for all movies lasting 0 . 8 s or longer . We then compiled the corresponding probability density distributions ( PDF , , Figure 3a ) and cumulative density distributions ( CDF , , Figure 3b ) of the vector magnitude , . The long-tailed PDF observed for λnull ( Figure 3a ) is consistent with the expected end-to-end distance distribution measured from two spots with a fixed separation when the localization of each spot is subject to Gaussian fitting error [56] . A simple numerical simulation of the end-to-end distance PDF for two sites separated by 22 nm and each subject to 22-nm localization error largely recapitulates the long-tailed distribution ( Figure S2c ) . We found that the distribution for λΔOL was distinctly different from that of λnull ( p<10−3 ) ; the difference was reproduced in three independent experiments ( Figure S1b ) . The mean separations , , were 47 ( N = 1 , 153 ) and 71 nm ( N = 979 ) for λnull and λΔOL respectively ( results and measurement errors summarized in Table 2 ) . Peaks in plots centered at ∼40 nm , reflecting our experimental precision in determining ; that is , OR–OL molecules with below 40 nm could not be distinguished from each other . Hence , it was more meaningful to compare distributions of at large values where distributions differed most prominently . The cumulative probability of being 75 nm or more was ∼40% for λΔOL and only ∼15% for λnull ( Figure 3b ) . Furthermore , two-dimensional distributions of vectors ( Figure S4 ) were clearly wider for λΔOL than for λnull . Thus , by examining distributions , we could distinguish between the looped and unlooped control strains , suggesting that this approach could be used to probe CI-mediated DNA looping . We measured the mean end-to-end separation for λΔOL at 71-nm , much shorter than the ∼200-nm distance expected for B-form DNA with a typical 50-nm in vitro persistence length [57] . While such a result is expected given the many factors known to compact prokaryotic chromosomes [58] , it is possible that nonspecifically bound CI on the λΔOL DNA and/or PRM transcription activity could influence the distribution , as indicated by a series of recent studies in vitro and in higher eukaryotic systems [46] , [59] , [60] . To examine these possibilities , we first compared the distribution of the λΔOL strain to that of a control strain λΔOLPRM−cI−/cItrans ( Table 1 , Figure S5a and b ) . In this control strain , promoter PRM was mutated to abolish transcription and the cI start codon was eliminated , but CI binding to OR was unaffected ( Figure S5c , d , and e ) . In addition , we expressed CI from a plasmid at ∼9 times its level in λWT ( Table S8 ) . We found that the distributions of the λΔOL and λΔOLPRM−cI−/cItrans strains were indistinguishable ( Figure S5a and b ) , demonstrating that the compact λΔOL distribution does not depend on PRM transcription . Furthermore , distributions for the same λΔOLPRM−cI− strain with or without the CI-expressing plasmid were indistinguishable ( Figure S5a and b ) , suggesting that nonspecifically bound CI did not interact with specifically bound CI at OR operator sites to condense DNA in vivo [46] . We next investigated DNA looping in the context of wild-type and mutant OR–OL DNAs . In λWT , the wild-type λ sequence from OR through OL was inserted between lacO3 and tetO3 . CI could bind all OR and OL sites to mediate looping with both octameric and tetrameric CI complexes ( Figure 1a ) . In λOR3− and λOL3− , mutations in OR3 and OL3 essentially eliminated CI binding to these operators at lysogenic CI concentrations ( Table 1 ) [35] , [61] . We measured for these three strains and found that distributions differed significantly from those of the positive and negative controls λnull and λΔOL ( p<10−3 , except p = 0 . 004 for λWT and λnull ) , with and being intermediate to those of the controls ( Figure 3c and d ) . Mean values for the three strains also fell in between those of λnull and λΔOL ( Table 2 ) . The wild-type strain had lower than λOR3− and λOL3− , and its distribution differed from those of the mutant strains with moderate to high significance ( p = 0 . 001 and 0 . 048 for λOR3− and λOL3− , respectively ) ; distributions for λOR3− and λOL3− were indistinguishable from each other ( p = 0 . 493 ) . The trend of λnull<λWT<λOR3−≈λOL3−<λΔOL for was reproduced in three independent experiments ( Figure S1b ) . Assuming that a DNA molecule in the λWT , λOR3− , and λOL3− strains is in either a looped or unlooped state , the intermediate values of the three strains suggested that the fraction of looped DNA molecules ( herein termed looping frequency ) could be estimated by comparing distributions of these strains to those of the looped and unlooped controls λnull and λΔOL . To further investigate whether the observed DNA looping in the λWT , λOR3− , and λOL3− strains could be abolished by eliminating CI cooperative binding rather than by deleting OL , we constructed a control strain λCIG147D ( Table 1 ) . This strain differs from λWT by a CI mutation G147D known to be defective in pairwise cooperative interaction [62] , [63] . Structural evidence suggests that cooperative binding interfaces are shared for pairwise binding to adjacent operator sites and the formation of CI tetramers or octamers via DNA loops [64] . We found that the distribution of the λCIG147D strain was indistinguishable from that of λΔOL ( Figure S5f and g , Table S7 ) . We note that this G147D mutant also diminishes PRM transcription because of its weakened ability to form a CI tetramer at the OR1 and OR2 sites; hence its expression level is lower than that with wild-type CI ( Table S8 ) . Therefore , we constructed another control strain ( λCIG147D/cIG147D , trans ) , in which the CIG147D mutant protein was expressed constitutively at ∼11 times the CI expression level in λWT from a plasmid transformed into the λCIG147D strain ( Table S8 ) . We found that distribution of this strain was indistinguishable from that of the λΔOL and the λCIG147D strains , demonstrating that DNA looping could be abolished by eliminating CI cooperative binding . To quantitatively examine how operator mutations influence DNA looping , we estimated looping frequencies for λWT , λOR3− , and λOL3− by assuming a simple model . In this model , DNA molecule can only exist in one of two states , looped or unlooped , with distributions for each state resembling those of the looped and unlooped controls , λnull and λΔOL , respectively . Therefore , the distribution or for one of the three strains is the linear combination of that of λnull and λΔOL , with their distributions weighted by the looping frequency , :Using this model , we found that the looping frequency was 79% for λWT , and reduced to 53% for λOR3− and 60% for λOL3− ( results with errors summarized in Table 2 ) . The results were indistinguishable within error regardless of whether cumulative or probability density distributions were used , or whether data points from all frames or only the first frame of each molecule's movie were used ( Table S1 ) . The looping frequencies for λOR3− and λOL3− were indistinguishable from each other within error , suggesting a similar role of OR3 and OL3 in loop formation . Reduced looping frequencies of λOR3− and λOL3− compared to λWT suggest that while a CI octamer at OR12 and OL12 is sufficient to loop DNA , the resulting loop can be further stabilized by an additional CI tetramer only if both OR3 and OL3 are intact . To our knowledge , these measurements provide the first quantitative in vivo estimates of DNA looping frequencies that are independent of gene regulation models . In the above analyses , we only utilized , the magnitude of the vector , and discarded information about the direction of and its evolution in time . Looping frequencies estimated from distributions are analogous to equilibrium constants and lack kinetic information . While many DNA molecules only exhibited fluorescent spots in both EYFP and mCherry channels for one or two consecutive frames due to photobleaching , some molecules had fluorescent spots lasting for several consecutive frames in both channels ( Figure 2c–h; also see plots from molecules with many frames in Figure S3 ) . By analyzing how evolves in time , we can obtain additional information about DNA looping kinetics . We calculated the autocorrelation of ( the average dot product of two vectors separated by a time lag ) up to 0 . 5 s for each strain using all movies in which fluorescent spots in both channels lasted two or more frames ( Figure 4a ) . The autocorrelation curves of all strains showed an initial drop of ∼2 , 500 nm2 at the first time lag , corresponding to uncorrelated errors in determining . After the initial drops , all autocorrelation curves showed positive correlation values that were approximately constant at time lags up to 0 . 5 s . The observation of near constant autocorrelation values after the first time lag for all the strains indicated that the conformation of each DNA molecule , characterized by both the magnitude and orientation of , persisted for at least 0 . 5 s . This provides a lower limit for the amount of time it takes for two DNA sites in the relaxed , unlooped state to move relative to each other and potentially form a DNA loop , and thus an upper limit of ∼2 s−1 for the rate of DNA looping . The plateau values are related to the averaged mean end-to-end separations—λΔOL has the highest autocorrelation plateau and λWT , λOR3− , and λOL3− have intermediate values because they contain a mixture of looped and unlooped DNA molecules . Next , we measured average CI expression levels , , in all strains in order to understand to what different extent DNA looping influences PRM regulation . We used single-molecule fluorescence in situ hybridization ( smFISH , [2] , [65] , [66] ) , in which multiple fluorescently labeled oligonucleotides probe targeted nonoverlapping regions of cI mRNA , to count the number of PRM transcripts in individual cells ( Figure 4b and c ) . Given the assumption that the average number of CI molecules translated per PRM transcript is the same in all strains and the observation of indistinguishable cell growth rates ( Figure S6a and b ) , we expected average mRNA expression levels proportional to . The λnull strain does not contain the cI gene and was used as a negative control . All other strains were transcriptionally active . Under our experimental conditions , the false positive rate using the λnull strain was ∼1 transcript per 50 cells , two orders of magnitude below the levels of all other strains; false positives arise when nonspecifically bound probes occasionally co-localize to create a fluorescent spot above the detection threshold . Typical smFISH images of the five strains are shown in Figure 4b . We quantified the number of transcripts in each individual cell by dividing the total intensity of fluorescent spots in each cell by the average intensity of a single-transcript spot ( Figure 4c ) . We then determined in wild-type λ units ( WLU ) by dividing the average number of transcripts in cells of a given strain by the average number of transcripts in λWT cells . We found that deleting OL increased to ∼1 . 4 WLU ( Table 2 ) , indicating that the DNA loop formed between OL and OR in λWT enhances PRM repression . Mutating either OR3 or OL3 further increased to ∼2 . 5 WLU . These observations are consistent with previous observations that although OL3 is 2 . 3 kb away from the PRM promoter , it has as important a role as OR3 in repressing PRM at lysogenic CI concentrations [13] . This suggests that PRM was not strongly repressed by CI binding to OR3 in the absence of a tetrameric interaction with an additional dimer at OL3 . Finally , elevated in λOL3− relative to λΔOL indicated that DNA looping could also activate PRM , which was likely mediated by the binding of a CI octamer at OL12 and OR12 , and was consistent with recent in vivo [49] , [51] and in vitro [53] experiments . We have shown that reduced looping frequencies in λOL3− and λOR3− compared to that in λWT corresponded to increased expression levels of CI in the two strains , and that unlooped λΔOL has a higher expression level than the λWT strain . To establish a quantitative framework that explains all observed relationships between looping and CI expression levels , we refined a thermodynamic model , with which we estimated looping free energies and the degree to which DNA looping changes the activity of PRM . These parameters are important because free energies describe the likelihood of interaction between two distal DNA sites , and changes in promoter activity directly reflect the influence of DNA looping on gene regulation . The thermodynamic approach was first applied to model repression and activation of PRM by CI bound to OR [52] and recently modified to address looping [35] , [44] , [49] , [51] . Our modeling approach is unique in that we used two independent , in vivo measurements , looping frequencies , and corresponding CI expression levels , to refine parameters for DNA-looping free energies and transcription activities . In previous modeling work , DNA-looping free energies were either inferred from PRM and PR expression-level measurements [35] , [49] , [51] or estimated using in vitro data [44] . The thermodynamic model and fixed physical parameters from previous reports we used to estimate PRM expression levels and DNA looping frequencies are essentially identical to the one used to analyze in vivo gene expression experiments [35] . Briefly , we assume that DNA states can be enumerated , that steady-state , in vitro DNA-binding measurements are applicable in vivo , and that mean expression rate , , equals the sum of all products , where is the transcription rate in a particular state and is the probability of the state at a given concentration of free CI dimers :Each state is defined by its free energy , , the number of bound CI dimers , , and the degeneracy , , which is the number of states with the same , , and . The model is described in greater detail in the Materials and Methods section; all states considered are listed in Table S2 . is normalized by the partition function , , so that the sum of all state probabilities is 1 . Following earlier work [49] and considering that the CI-mediated loop is relatively long , we assumed looping free energies to be independent of parallel or antiparallel orientation . Note that loop orientation is important in shorter DNA loops such as those mediated by Gal repressor [67] . We approximated the average CI concentration , , as the concentration at which the degradation rate equaled the production rate . We refined our model to fit seven experimental observables: CI expression levels for λΔOL , λWT , λOR3− , and λOL3− , and the looping frequencies for λWT , λOR3− , and λOL3− . We varied four free parameters: the free energies of forming a CI octamer and tetramer in the DNA loop as defined by Dodd et al . [35] , , and , and the PRM expression rates when OR12 is bound by CI and DNA is either looped ( ) or unlooped ( ) . is the free energy of bringing together OR and OL when both are bound by two adjacent CI dimers to form a CI octamer , resulting in a looped conformation . is the free energy of adding a CI tetramer to a loop already secured by a CI octamer . All other free energies and parameters such as specific and nonspecific DNA binding of CI were fixed at the values used by Dodd et al . [35] . The wild-type CI concentration was fixed to 220 nM ( ∼150 molecules/cell ) based upon our previous experiment in which CI molecules were counted at the single-molecule level in a similar strain at similar growth conditions [3] . The CI degradation rate was fixed to give a half-life equal to the observed 2-h doubling time in our experiments . The four free parameters were adjusted to best fit our experimental measurements of looping frequencies and CI expression levels . Modeled looping frequencies and CI expression rates at different CI concentrations are shown in Figure 5a and b . The best fit estimated and at 0 . 3 and −3 . 2 kcal/mol , respectively , and the CI expression rates at 1 . 9 nM/s and 4 . 5 nM/s for unlooped ( ) and looped ( ) DNA when CI binds OR12 . These results suggest that the DNA looping mediated by only a CI octamer is not strongly favored , while looping mediated by both an octamer and tetramer is the dominant configuration if all six binding sites are bound by CI dimers . Note that a small , positive is consistent with measured looping frequencies greater than 50% for ΔλOL3− and ΔλOR3− , as one unlooped configuration could lead to multiple looped configurations ( Table S2 ) . The higher CI expression rate from the looped configuration suggests that , in the absence of OR3 binding , bringing the distal OL and OR sites together to form a DNA loop activates PRM to 2 . 4 times the unlooped level . To test how sensitive the fitting results were to two fixed parameters that are poorly defined in previous work , we varied CI expression levels and nonspecific DNA binding affinity . We found that across the examined ranges , octameric looping energies , , were consistently near 0 and tetrameric looping energies , , were strongly favorable between −2 . 8 to −4 . 6 kcal/mol ( Table S3 ) . Similarly , CI expression rates and remained close to the original fit values , giving activation ratios between 1 . 7 and 2 . 5 ( Table S3 ) . We also verified that our fit parameters were unique—as shown in Figure 5c and d , the values of fit parameters corresponded to a well-defined minimum in the sum of squared residuals in the four-dimensional ( two free energies and two expression rates ) parameter space ( Figure 5c and d ) . Hence we conclude that the four fit parameters resulted from the model were robust and well defined .
Our estimated looping frequencies of 79% for λWT and greater than 50% for λOR3− and λOL3− are larger than those observed in vitro by TPM and AFM , where looping frequencies at lysogenic CI concentrations were approximately 60% with wild-type operators and 10%–40% in the absence of OR3 and OL3 [42] , [44] , [46] . As looping frequency is directly linked to looping free energy , comparison of values showed the same trend: values estimated in these in vitro experiments were similar to our estimate of −3 . 2 kcal/mol , while in vitro values were 1–2 kcal/mol higher than ours [44] , [46] . Significantly different values likely resulted from differences between naked DNA in an in vitro environment and the compact , protein-decorated E . coli chromosome in the crowded cellular environment . Factors such as supercoiling and nonspecific , “histone-like” DNA-binding proteins could compact DNA and lead to more frequent encounters between OR and OL . Our observation that the unlooped λΔOL DNA was extremely compact ( discussed in more detail below ) was consistent with this view; this level of compaction ( comparable to a polymer with a 3-nm rather than a 50-nm persistence length ) could lead to a 50-fold increase in the rate at which OR and OL encounter each other [68] . The relatively unchanged values could reflect the fact that the entropic and energetic costs of bringing OR and OL together are included in . Our looping frequency estimates confirm what were predicted by in vivo gene expression experiments—DNA was estimated to loop ∼72% of the time for wild-type OR–OL DNA and ∼69% for DNAs similar to our λOR3− and λOL3− constructs [35] . Correspondingly , the and estimated in the in vivo work ( −0 . 5 and −3 . 0 kcal/mol ) [35] compared well to ours ( 0 . 3 and −3 . 2 kcal/mol ) . One important assumption we employed in calculating looping frequencies is that that looped and unlooped λWT , λOR3− , and λOL3− DNA molecules had similar distributions to those of the looped control λnull and unlooped control λΔOL , respectively . It is possible that the unlooped states in the λWT , λOR3− , and λOL3− strains were more compact than that in λΔOL if after a DNA loop breaks OR–OL DNA does not always completely relax before it reforms again . In such a case , looping frequencies estimated using the linear-combination model would be upper limits on the true looping frequencies . Nevertheless , as we show above , our looping frequency estimates broadly agree with expectations from previous studies . Since this simple model only requires one free parameter and gives reasonable results , it is unnecessary to invoke more complicated models . By comparing looping frequencies and corresponding CI expression levels in λWT , λΔOL , λOR3− , and λOL3− , we showed that loop stabilization by the CI tetramer between OR3 and OL3 is important for efficient PRM repression , and that looping mediated by a CI octamer at OR1 and OR2 is important for PRM activation . We note that while it is possible that the presence of tetO3 and lacO3 binding sites flanking OR–OL DNA may influence CI binding and/or transcription , this influence is negligible . This is because CI expression levels in these strains measured using smFISH are comparable to that of a wild-type λ lysogen ( Table S8 ) , and our results are consistent with previous observations [13] , [49] , [51] , [53] . Furthermore , results are directly comparable as all strains used in this study are identical with respect to the presence and positioning of these binding sites . Combining these results in our thermodynamic model , we estimated that CI-mediated DNA looping activates PRM to 2 . 4 times its level when the DNA does not loop . This compares well to earlier estimates of 2–4 fold [49] , and 1 . 6-fold for a high-expression PRM mutant [53] . Another study did not find looping activates transcription , modeling CI-concentration-dependent PR and PRM activities without invoking activation via looping ( by assuming ) [35] . A later study indicated that this discrepancy may have resulted from different constructs used in the earlier study [49] . The molecular basis for DNA loop-enhanced PRM activation is unclear . One possibility is that a CI dimer bound to OR2 interacts with RNA polymerase to a greater extent if it is part of a higher-order CI octamer [53] . Alternatively , a recent work showed that a DNA UP element proximal to OL [49] , [69] enhances CI expression from PRM in looped DNA by contacting the α-C-terminal domain of RNA polymerase [51] . The activation mechanism could be clarified in future experiments measuring both looping frequency and PRM activity while varying operator and UP element sequences and introducing CI mutations affecting operator binding , oligomerization , and RNA polymerase interaction . We estimated the time scale a DNA molecule stays in a particular state by calculating the autocorrelation function of the vector ( Figure 4a ) . The vector was strongly correlated for at least 0 . 5 s , suggesting that a particular DNA conformational state , either compact or extended , persisted for at least 0 . 5 s . This implies an upper limit of 2 s−1 for the rate of loop formation from the extended state . This upper bound of transition rate is in the range of what was observed in a previous TPM experiment , in which looped and unlooped states lasted for tens of seconds [44] , and argues against a significantly faster rate used in a recent computer simulation ( ∼60 s−1 ) [50] . We note that although it is possible that transient CI unbinding does not necessarily lead to immediate and complete DNA conformational relaxation at our measurement time scale , the autocorrelation analysis puts an upper limit for the true transition rate between the looped and unlooped states . The same concern also applies to in vivo 3C and in vitro TPM experiments . Slow transitions between looped and unlooped states imply that low or high expression states resulting from a particular DNA conformation could be long-lived , potentially committing a cell to a particular fate . Supporting this is a recent study that suggested that a single unlooping event could trigger induction of the lac operon [5] . We were unable to obtain time trajectories long enough to clearly identify looped/unlooped transitions for single DNA molecules . Development of brighter , faster maturing , and more photostable fluorescent proteins or in vivo labeling with synthetic fluorophores [70] , [71] will help in increasing the number of measurements made on one DNA molecule , possibly enabling accurate measurement of DNA looping kinetics in vivo . We observed very small end-to-end separation for the unlooped control ( = 71 nm ) . This distance was shorter than expected from modeling the unlooped DNA as a noninteracting worm-chain with an in vitro persistence length of 50 nm [72] , but consistent with the recently observed extreme bendability of short DNA molecules [73] . A noninteracting chain with an equivalent to that of λΔOL would have a persistence length of only 3 nm , which is physically infeasible . Our measurements of indistinguishable conformational distributions in the absence of PRM transcription and the presence of CI overexpression suggest that neither transcription nor nonspecifically bound CI played a major DNA-compacting role in our experiments . Furthermore , C . crescentus chromosomal DNA segments of ∼5 kb were found to be similarly compact and consistent with Brownian dynamics simulations of supercoiled DNA [74] . We attribute the small end-to-end separation observed for λΔOL to the high compaction of the E . coli chromosome in the crowded cellular environment . While the exact molecular mechanisms responsible for compaction remain unclear , previous studies found that in vitro binding of the histone-like HU proteins [75] ( accession numbers P0ACF0 , P0ACF4 ) and in vivo mammalian chromatin packing [76] reduced the apparent persistence length of DNA . Hence , it is possible that nucleoid-associated proteins such as HU may bring distal DNA sites together by protein–protein interactions and/or affect local DNA conformations by introducing bends and relieving torsional strain [77] . Another important factor could be negative supercoiling , which has been shown to compact the chromosomal DNA globally [78] . However , the exact effect of negative supercoiling on a 2 . 3-kb DNA segment is difficult to predict , because negative supercoiling could also introduce extended , plectonemic structures that promote large separations between DNA sites on relatively short length scales [78] . Our two-color , high-resolution method can be applied to examine how chromosomal location , DNA length , genetic background , and growth conditions affect the distance between any two DNA sites on the E . coli chromosome . Furthermore , the spatial organization of the E . coli chromosome can be determined by systematically measuring distributions between DNA sites throughout the chromosome . This method is similar to how chromosome conformation capture was used to generate a 3D model of the C . crescentus chromosome [79] , but with significantly improved spatial resolution and without potential artifacts from fixation .
A plasmid , pS2391 , containing lacO3 and tetO3 ( the tetO2 sequence [54] was used for each repeat in tetO3 ) sites was synthesized by Genewiz , Inc . Segments of λ DNA ( OR through OL for λWT , OR up to but not including OL for λΔOL ) from the wild-type lysogen JL5392 ( a gift from John Little , University of Arizona ) were amplified by PCR . This DNA was sequenced and inserted between lacO3 and tetO3 using the In-Fusion PCR cloning system ( Clontech ) . A kanamycin-resistance cassette flanked by BamHI sites was amplified by PCR and inserted after lacO3 . For strains with mutated operators , mutations r1 [80] , OL3–4 [13] , and cIG147D [62] were introduced to the λWT template via QuikChange ( Agilent ) . A plasmid carrying the PRM−cI− mutations ( Figure S5c ) ( λΔOLPRM−cI− ) was constructed by overlapping PCR mutagenesis using complementary primers carrying the desired mutations , flanked by a forward primer that sits at the EcoRI site on the upstream end of the operon and a reverse primer at the ClaI site in the rexA gene downstream of cI . The 1 . 13 kb PCR product was introduced to the λΔOL plasmid by restriction ligation . This procedure resulted in seven plasmids that were used as templates in subsequent chromosome insertion: pZH105 ( λnull ) , pZH016 ( λΔOL ) , pZH107 ( λWT ) , pZH107r1 ( λOR3− ) , pZH107OL3–4 ( λOL3− ) , pACL006 ( λWTG147D ) , and pACL007 ( λΔOLPRM−cI− ) . Note that we use shorthand names such as λnull here for clarity; corresponding names used in our laboratory are listed in Table S4 . The DNA sequence including lacO3 , the λ DNA segment , tetO3 , and the kanamycin resistance cassette was inserted into the chromosome of E . coli strain MG1655 by λ Red recombination [81] , excising the lac operon , lacI , and all lacO sites . To express the CI protein in trans from a plasmid , we constructed the plasmid pACL18 in which the wild-type cI ORF is driven by a constitutive promoter , PRMc , which has the wild-type −35 ( TAGATA ) and −10 ( TAGATT ) sequences , lacks OR2 , and has a mutated OR1 sequence ( CGCCTCGTGAGACCA ) that eliminates binding by CI . The pRMc–cI fragment was then cloned to the ClaI site of the low-copy vector pACYC184 . The plasmid pACL17 was generated similarly using a template containing the CIG147D mutation . The two-color reporter plasmid pLau53 , which expresses LacI-ECFP and TetR-EYFP polycistronically under the control of the PBAD promoter [82] , was obtained from the Yale Coli Genetic Stock Center . Because the autofluorescence spectrum of live cells is generally strongest at wavelengths around 500 nm [83] , single-molecule imaging of blue-shifted fluorescent proteins such as ECFP is difficult . The red fluorescent protein mCherry , which further benefits from a large Stokes shift , fast chromophore maturation rate , and high brightness relative to other monomeric RFPs [84] , was inserted in place of ECFP . We also created a tandem LacI-mCherry-EYFP reporter , which was used as a fiducial marker , by inserting the linker sequence from the tandem-dimer fluorescent protein tdTomato [84] in between mCherry and EYFP . To accurately localize a fluorescent spot arising from only a few fluorescent protein molecules above the background of unbound molecules within a cell , we reduced the reporter expression level by weakening the ribosome binding sites ( RBSs ) . Weakened RBS sequences were designed using an online RBS calculator [85] . For example , the RBS for TetR-EYFP translation was the consensus AGGAGG Shine-Delgarno sequence in the parent plasmid pLau53 . Our reporter plasmid had an ACCAGG Shine-Delgarno sequence , with a predicted ∼300-fold decrease in the TetR-EYFP translation rate . All sequences including chromosome insertions were verified by sequencing ( Genewiz Inc ) . Reporter plasmids are described in Table 1 . For all experiments reported in this study , cells were grown and imaged at room temperature ( ∼25°C ) in M9 minimal media supplemented with MEM amino acids ( Sigma ) . Cells were grown overnight with 0 . 4% glucose and 50 µg/ml carbenicillin to an optical density ( OD600 ) of 0 . 4 . After centrifugation at room temperature , cells were resuspended at OD600≈0 . 2 with 0 . 4% glycerol plus 0 . 2% L-arabinose and grown for 2 h ( ∼1 cell cycle ) to induce LacI-mCherry and TetR-EYFP expression . Cells were again resuspended at OD600≈0 . 2 with 0 . 4% glucose and grown for another 2 h before immediate observation to allow time for fluorescent protein chromophores to mature . We compared growth rates for the parent strain MG1655 to the experimental strain λnull to determine whether inserting the lacO3 and tetO3 construct into the chromosome and/or inducing expression from the reporter plasmid introduced a significant growth defect . Under induction growth conditions ( ∼27°C , M9 media with 0 . 4% glycerol and 1× MEM amino acids ) starting at OD600≈0 . 1 and observing 8 h of growth , we measured doubling times of 2 . 7 h for MG1655 and 3 . 4 and 3 . 3 h for λnull harboring the reporter plasmid ( in the absence and presence of 0 . 3% L-arabinose , respectively ) , indicating that there is no large growth defect associated with the insertion of the tandem operator sites into the chromosome and/or the expression of TetR-EYFP and LacI-mCherry fluorescent fusion proteins ( Figure S6c ) . In each experiment , samples of all strains were placed on separate gel pads in the same growth chamber . Two sets of at least 30 movies were acquired for each strain , with the second set acquired in the reverse order to minimize any bias possibly introduced by observing some strains in a particular order . All images were acquired within less than one cell doubling time . Cells were put on a gel pad made of 3% low-melting-temperature SeaPlaque agarose ( Lonza ) in M9 with glucose and imaged on an Olympus IX-81 inverted microscope with a 100× oil immersion objective ( Olympus , PlanApo 100× NA 1 . 45 ) and additional 1 . 6× amplification . Images were split into red and yellow channels using an Optosplit II adaptor ( Andor ) and captured with an Ixon DU-895 ( Andor ) EM-CCD with a 13-µm pixel width using MetaMorph software ( Molecular Devices ) . Laser illumination was provided at 514 nm by an argon ion laser ( Coherent I-308 ) , which also pumped a rhodamine dye laser ( Coherent 599 ) tuned to ∼570 nm . A quarter-wave plate ( Thorlabs ) was used to circularly polarize excitation light . Emitted light was split by a long-pass filter , and the red and yellow images were filtered using HQ630/60 and ET540/30 bandpass filters ( Chroma ) . Images were inspected manually using a custom MATLAB script to identify spots that appeared in both EYFP and mCherry images . Images from all strains were displayed in random order without knowing the strain identify to avoid bias in spot selection . Pixel intensities within 3 pixels of the initial spot location were fitted with a symmetric , two-dimensional Gaussian distribution to estimate spot coordinates . The variance of the fit distribution was constrained to be less than 2 pixels . Spot-fitting error was estimated by scrambling residuals from a fit to the fluorescence data in 10 random permutations , adding them to the data , and fitting the resulting images; the reported error for a spot is the standard deviation of the distances between these fits and the initial fit to the raw data . Fitting error distributions are shown in Figure S1a . The LacI-mCherry-EYFP tandem dimer ( Figure S2b ) in which the two fluorescent proteins were directly fused together was used to acquire fiducial control points to transform between the mCherry and EYFP coordinate systems . A projective transform was calculated from the control points using the cp2tform function in MATLAB . We found that relatively simple , global transformations were sufficient to transform coordinates of fluorescent beads ( Tetraspeck , Invitrogen ) with ∼10-nm registration error in our microscope setup , and did not see any further improvement with a locally weighted transformation used in in vitro two-color experiments [21] . This transformation was also used to generate the overlay images in Figure 2 , Figure 3 , and all supplemental movies . Fluorescent beads were not used as fiducial markers because the beads' emission spectra were different from those of the fluorescent proteins . Analysis was restricted to molecules in which mCherry and transformed EYFP coordinates were separated by less than 200 nm . Separations beyond this threshold were rare ( ∼1% of data , see two-dimensional distributions in Figure S4 ) and did not correlate with strain identity in any reasonable way . They possibly arose from data in which cells contained two labeled copies of OR–OL DNA . After transformation into a uniform coordinate system , was calculated from the mCherry and EYFP coordinates and multiplied by an 81-nm pixel size ( resulting from 160× magnification on a CCD with a 13-µm pixel width ) . Probability and cumulative distributions and were calculated for 10-nm bins using the kernel smoothing probability density estimation ( ksdensity ) function in MATLAB , restricting the density to positive values and employing a uniform kernel width small enough to follow empirical cumulative density distributions without any systematic errors . Significant differences between distributions were determined using a two-sample Kolmogorov–Smirnov test; two-tailed Student's t tests of sample means returned smaller , more significant p values . Errors in and were determined by calculating the means of 1 , 000 bootstrapped samples; the reported error is the standard deviation of the calculated means . Looping frequencies were estimated by least squares fitting of 1 , 000 bootstrapped distributions ( control distributions were also randomized on each iteration ) and their error was calculated similarly . Concentration measurements by smFISH followed a previously described protocol [66] . Transcripts from PRM were labeled with a mixture of 42 oligonucleotides labeled with CAL Fluor Red 610 ( Biosearch Technologies ) , 31 of which hybridized to cI ( 11 targeted sequences not found in E . coli and did not cause a problematic level of false positives ) . Table S5 lists all 42 oligonucleotides . Labeled cells were imaged with 561-nm excitation at six imaging planes separated by 200 nm z-depth with negligible photobleaching . For each frame , fluorescent spots were automatically detected and fit to a Gaussian using a custom MATLAB routine . Nearly all molecules appeared in multiple image slices; the slice with the largest fit amplitude was kept . The integrated fluorescence of spots was observed to be quantized with one or a few molecules localized within one diffraction-limited spot . The intensity of one transcript was estimated from the distribution of spot intensities , and the number of molecules contributing to each spot was estimated from this quantization . The number of transcripts in each cell was estimated from the sum of the number of molecules in each spot within that cell . Alternatively , the number of molecules in one cell is proportional to its integrated fluorescence; this measurement provided the same average expression levels within error . The experiment was repeated to ensure that differences in labeling efficiency between samples were not responsible for differences in the number of detected molecules; combined data from both experiments were used for analysis . To generate simulated distributions , we first generated 10 , 000 random radial distances for a chain with a contour length and persistence length from a worm-like , noninteracting chain model using a Gaussian distribution with Daniels' approximation , which is accurate in the regime [72]:Each simulated was projected onto the plane at a random angle to give a distance . Simulated spots were placed at coordinates and . The MATLAB function mvnrnd was then used to simulate normally distributed measurement error with a standard deviation of 22 nm to the coordinates of each simulated spot . This procedure was sufficient to simulate the λnull distribution ( Figure S2c ) using a fixed end-to-end distance of 22-nm ( approximate distance between the centers of the lacO3 and lacO3 sites; Figure S2a ) . Note here that the simulation is simplified in that it assumes that each spot has the same 22-nm localization error . In reality , localization error varies between different spots ( Figure S1a ) and there are other sources of measurement error . These differences may explain the slight deviation of the simulated distribution from the experimental distribution . The same procedure was used to estimate the expected for 2 . 3-kb , B-form DNA with a 50-nm persistence ( ∼200 nm ) as well as the apparently persistence length ( 3 nm ) implied by the 71-nm observed for λΔOL . Additional descriptions of thermodynamic states are listed in Table S3 . Parameter values were determined by first scoring a wide range of parameter values and iteratively searching narrower and more finely grained parameter ranges to manually minimize the sum of the squares of the differences between experimental and modeled values for looping frequency and CI expression level . We then refined this fit by least-squares minimization using MATLAB . This was done using a minimized model that only accounted for states likely to be populated near or above lysogenic CI concentrations ( e . g . , disregarding states in which OR1 and OR2 are unbound by CI ) . Using the same parameters and accounting for all 176 possible states ( 122 unique states accounting for degeneracy ) did not significantly change the fit results . Fitting with this much more complex model gave octameric and tetrameric looping free energies of 0 . 6 and −3 . 3 kcal/mol and unlooped and looped expression rates of 2 . 1 and 5 . 3 nM/min . When determining parameters , rates were expressed in terms of changes in concentration per unit time; we followed earlier work in assuming that in a typical E . coli cell , a single molecule is at a concentration of ∼1 . 47 nM [35] . We do not report any estimate of fitting error; instead , we present only the parameters most consistent with our data and assumptions . Figure 5c and d shows that fit parameters were well-determined at a given combination of wild-type CI concentration and nonspecific binding parameters . As noted in the main text , varying these two parameters changed the absolute best-fit parameters , but did not dramatically change our conclusions . Furthermore , fixed parameters of previous studies were determined in a number of separate experiments employing different methods at temperatures other than 25°C; a rigorous estimate of modeling error would require knowing the error in the measurements of fixed parameters in our experimental conditions . The basal CI expression rate , , was arbitrarily fixed at ; this did not have any significant impact on determining other parameters , as our measurements were all at or above lysogenic , where OR2 is almost always bound by a CI dimer . Additionally , the fraction of free CI dimers was fixed at its value for 150 CI molecules per cell at a given concentration of nonspecific binding sites and nonspecific binding affinity . Fixing the concentration of free CI dimers is a reasonable approximation if ( 1 ) nearly all CI molecules are in dimers and ( 2 ) the number of free nonspecific binding sites is not significantly changed by nonspecifically bound CI dimers . Figure 2a–e , Figure 4b , and Movies S1 , S2 , S3 , S4 , S5 , S6 were prepared using NIH ImageJ [86] . Raw fluorescence image intensities were scaled linearly from the lowest to highest values in region shown . For EYFP/mCherry overlay images , brightfield images were inverted and converted to 8-bit RGB . Fluorescence images were bandpass filtered and background subtracted before being used to generate magenta ( mCherry ) and green ( EYFP ) 8-bit RGB images that were added to the brightfield image . The EYFP images were first transformed in MATLAB using the imtransform function and the same fiducial data that were used to transform EYFP spot locations into mCherry coordinates . For smFISH images ( Figure 4b ) , the value of each pixel is the maximum value of that pixel in six images collected at different z-axis positions . Intensities for all images were scaled linearly from the minimum to the maximum of all pictures ( 117–4 , 840 counts in 16-bit images ) . | One mechanism cells use to regulate gene expression is DNA looping , whereby two distant DNA sites are brought together by regulatory proteins . The looping then either enhances interactions between other regulatory proteins bound at the separate sites or brings those regulatory proteins close to RNA polymerase at the promoter . Recent work in bacteriophage λ has suggested that DNA looping mediated by a transcription factor called λ repressor CI plays a critical role in regulating the expression of λ genes and consequently in determining the fate of the host E . coli bacterial cells . CI-mediated DNA looping has been directly demonstrated in vitro , but it has only been indirectly inferred in vivo . For the current study we developed a method to visualize CI-mediated DNA looping in individual live E . coli cells . We labeled two DNA sites—one each side of the proposed loop—with differently colored fluorescent fusion proteins , allowing us to measure their separation with an accuracy of a few tens of nanometers . Using this method , we directly analyzed CI-mediated DNA looping , providing insight into how transcription factor-mediated DNA looping influences gene regulation in live E . coli cells . Our methodology can be applied to a broad range of questions regarding chromosome conformation in prokaryotes and higher organisms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"nucleic",
"acids",
"gene",
"regulation",
"genetics",
"gene",
"expression",
"dna",
"molecular",
"genetics",
"biology",
"computational",
"biology",
"molecular",
"cell",
"biology",
"synthetic",
"biology",
"biophysics",
"dna",
"transcription"
] | 2013 | Transcription-Factor-Mediated DNA Looping Probed by High-Resolution, Single-Molecule Imaging in Live E. coli Cells |
Thermal adaptation is essential in all organisms . In yeasts , the heat shock response is commanded by the heat shock transcription factor Hsf1 . Here we have integrated unbiased genetic screens with directed molecular dissection to demonstrate that multiple signalling cascades contribute to thermal adaptation in the pathogenic yeast Candida albicans . We show that the molecular chaperone heat shock protein 90 ( Hsp90 ) interacts with and down-regulates Hsf1 thereby modulating short term thermal adaptation . In the longer term , thermal adaptation depends on key MAP kinase signalling pathways that are associated with cell wall remodelling: the Hog1 , Mkc1 and Cek1 pathways . We demonstrate that these pathways are differentially activated and display cross talk during heat shock . As a result ambient temperature significantly affects the resistance of C . albicans cells to cell wall stresses ( Calcofluor White and Congo Red ) , but not osmotic stress ( NaCl ) . We also show that the inactivation of MAP kinase signalling disrupts this cross talk between thermal and cell wall adaptation . Critically , Hsp90 coordinates this cross talk . Genetic and pharmacological inhibition of Hsp90 disrupts the Hsf1-Hsp90 regulatory circuit thereby disturbing HSP gene regulation and reducing the resistance of C . albicans to proteotoxic stresses . Hsp90 depletion also affects cell wall biogenesis by impairing the activation of its client proteins Mkc1 and Hog1 , as well as Cek1 , which we implicate as a new Hsp90 client in this study . Therefore Hsp90 modulates the short term Hsf1-mediated activation of the classic heat shock response , coordinating this response with long term thermal adaptation via Mkc1- Hog1- and Cek1-mediated cell wall remodelling .
Microorganisms inhabit dynamic environments and are continually challenged with environmental stimuli and stresses . Microbial survival depends upon effective environmental response strategies that have been elaborated over evolutionary time . These cellular strategies have been intensively studied in various contemporary model organisms [1] , [2] , [3] , [4] . The emergent paradigm is that cells react to environmental changes via a sense and respond logic: they continuously monitor their environment , and upon encountering a stimulus , mount a cellular response [5] . This is achieved through diverse signalling pathways that drive physiological adaptation to a myriad of environmental stresses that include temperature fluctuations , osmotic , oxidative and weak acid stresses , as well as nutrient limitation [6] , [7] . Fungal pathogens have evolved robust stress responses that enable them to counteract the antimicrobial defences of their host , thereby promoting the colonisation of specific niches . The major fungal pathogen of humans , Candida albicans , is an opportunistic pathogen that has evolved as a relatively harmless commensal of the mucous membranes and digestive tracts of healthy individuals [8] , [9] . C . albicans is a common cause of mucosal infections ( thrush ) and when antimicrobial defences become compromised this yeast can cause life-threatening systemic infections [8] , [10] . Stress responses are critical for survival of C . albicans inside the human body , and genetic inactivation of these responses attenuates virulence of this pathogen [11] , [12] , [13] . However , the regulation of these stress signalling mechanisms has diverged significantly in C . albicans compared with other yeasts [14] . For example , unlike Saccharomyces cerevisiae , Schizosaccharomyces pombe or Candida glabrata [2] , [15] , [16] , C . albicans does not activate a large core transcriptional response [3] . The core transcriptional responses of S . cerevisiae , S . pombe and C . glabrata involve the activation of common sets of stress genes by one particular stress that promote cross-protection to diverse stresses [2] , [15] , [16] . In S . cerevisiae and C . glabrata , this core transcriptional response and hence stress cross-protection are dependent on the transcription factors Msn2 and Msn4 , which activate target genes via stress response elements ( STRE ) in their promoters [17] . In S . pombe , the core stress response is driven largely by the Sty1 stress activated protein kinase ( SAPK: the orthologue of Hog1 in other yeasts ) [15] . In contrast , C . albicans does not mount a broad core transcriptional response to stress , there is limited stress cross-protection in this yeast , and the roles of Hog1 and Msn2/Msn4-like transcription factors have diverged in this pathogen [2] , [3] , [18] , [19] , [20] . Whilst , C . albicans does appear to activate a relatively specialised core transcriptional response to osmotic , oxidative and heavy metal stresses [19] , the consensus view is that this pathogen mounts relatively specific responses to particular environmental challenges . This involves activation of the corresponding signal transduction pathway , subsequent activation of the relevant set of stress genes and requisite changes in cell physiology , morphology and adherence [21] , [22] , [23] . A number of stress regulatory modules have been conserved between C . albicans and other yeasts . For example , the AP1-like transcription factors S . cerevisiae Yap1 [24] , S . pombe Pap1 [25] , C . glabrata CgYap1 [26] , [27] and C . albicans Cap1 [28] play analogous roles in the activation of transcriptional responses to oxidative stress . Also , the Hog1/Sty1 SAPK is conserved in these yeasts , although the orthologues have diverged with respect to the stress responses they regulate . S . cerevisiae Hog1 is primarily involved in responses to osmotic stress , whereas C . albicans Hog1 and S . pombe Sty1 contribute to a diverse range of stress responses [29] , [30] , [31] . Additional mitogen activated protein kinase ( MAPK ) cascades have been conserved between S . cerevisiae and C . albicans . These include the cell wall integrity Mpk1/Slt2 pathway [32] and the cell wall , morphogenesis and pheromone signalling pathways involving the Cek1 and Cek2 MAPKs [33] , [34] . These MAP kinase pathways contribute to thermotolerance in S . cerevisiae [35] , [36] , [37] , although the mechanisms by which they do so remain obscure . These MAP kinase pathways are also important for virulence , as MAP kinase defective C . albicans mutants display attenuated virulence in infection models [11] , [13] , [33] , [38] . The heat shock response is among the most fundamentally important and ubiquitous stress responses in nature . The heat shock transcription factor ( Hsf1 ) which drives this response , is conserved from yeasts to humans [39] , [40] . Indeed , the Hsf1 module and the heat shock response are even conserved in C . albicans , an organism that is obligately associated with warm blooded mammals and hence occupies thermally buffered niches [41] . Furthermore Hsf1 is essential for viability in C . albicans [42] and other yeasts [43] . These observations reflect the fundamental importance of heat shock adaptation in all organisms . Even in the absence of stress , Hsf1 binds as a trimer to canonical heat shock elements ( HSEs ) in the promoters of target heat shock protein ( HSP ) genes [44] , [45] , [46] . When S . cerevisiae or C . albicans cells are exposed to an acute heat shock , Hsf1 becomes hyper-phosphorylated and activated , leading to the transcriptional induction of these target HSP genes , thereby promoting cellular adaptation to the thermal insult [39] , [47] . Many HSPs are molecular chaperones that promote the folding , assembly , or cellular localisation of client proteins [48] . They also minimise the aggregation of unfolded or damaged proteins and often target such proteins for degradation [48] . HSPs are critical for the survival of eukaryotic cells under normal conditions as well as following exposure to an acute heat shock . Indeed our recent exploration of the dynamic regulation of Hsf1 during thermal adaptation has suggested that the Hsf1-HSE regulon is activated even during slow thermal transitions such as the increases in temperature suffered by febrile patients [49] . This explains why the Hsf1-HSE regulon is active in C . albicans cells infecting the mammalian kidney , and why activation of this regulon is essential for virulence of C . albicans [42] . Clearly the Hsf1-HSE regulon is critical for the maintenance of thermal homeostasis , not merely for adaptation to acute heat shocks . Hsp90 has been suggested to play a critical role in regulation of the Hsf1-HSE regulon , contributing to an autoregulatory circuit involving Hsp90 and Hsf1 [49] . In the absence of stress , Hsp90 is generally expressed at relatively high levels [50] , and is thought to repress Hsf1 [49] . However , thermal and other proteotoxic stresses can induce global problems in protein folding that overwhelm the functional capacity of Hsp90 [50] . Under these conditions the repression of Hsf1 by Hsp90 was proposed to be released , allowing Hsf1 to become activated leading to increased HSP90 expression [49] . The repression of Hsf1 by Hsp90 was suggested by the observation that pharmacological inhibition of Hsp90 correlates with HSF1 activation in mammalian cells [51] , [52] . Zou and colleagues demonstrated that HSF1 can be cross-linked to Hsp90 in unstressed HeLa cells , suggesting that HSF1 might interact directly with Hsp90 [52] . Additionally , the trimeric form of human HSF1 has been shown to associate with an Hsp90-immunophilin-p23 complex , and this is thought to repress HSF1 transcriptional activity [53] . Furthermore , HSP90 modulates HSF1 regulation in Xenopus oocytes [54] . Hence HSF1 is known to be a client protein of Hsp90 in metazoan cells . However , although mutations that interfere with Hsp90 function have been shown to derepress the expression of Hsf1-dependent reporter genes in S . cerevisiae [55] , no physical interaction between Hsf1 and Hsp90 has been demonstrated in the fungal kingdom . In this study , we explore the regulatory control of cellular circuitry governing the response to temperature stress through an integrated approach involving specific hypotheses and unbiased screens in C . albicans . We determined that Hsf1 is a client protein of Hsp90 , establishing for the first time in the fungal kingdom that the Hsf1-Hsp90 interaction is critical for the regulation of short term adaptive responses to heat shock . We questioned how cells adapt to heat stress in the longer term by investigating which other signalling pathways contribute to thermotolerance and executing genetic screens for protein kinase mutations that confer temperature sensitivity . This revealed that the Hog1 , Mkc1 and Cek1 MAP kinase pathways contribute to thermotolerance , as does casein kinase signalling . We show that these MAP kinase pathways are not essential for Hsf1 activation . Rather , they contribute to thermal adaptation in the longer term via cell wall remodelling . Hog1 , Mkc1 and Cek1 are client proteins of Hsp90 , and genetic depletion of Hsp90 affects this cell wall remodelling . Therefore , Hsp90 integrates the short term and long term molecular responses that underpin thermotolerance .
All strains are listed in Table 1 , with the exception of the library of C . albicans transposon mutants [56] , [57] , [58] . Strains were grown in YPD ( 1% yeast extract , 2% bactopeptone , 2% glucose ) [59] . To impose an instant heat shock of 30°C–42°C , cells were grown in YPD at 30°C to exponential phase , and mixed with an equal volume of medium that has been pre-warmed at 54°C in flasks which had been pre-warmed at 42°C . Cells were grown at 42°C for the times indicated . Doxycycline was added to YPD medium at a concentration of 20 µg/ml . Geldanamycin was added at 10 µM ( A . G . Scientific , Inc . , San Diego , USA ) , and radicicol at 20 µM ( A . G . Scientific , Inc . ) . For co-immunoprecipitation of Hsf1 and Hsp90 , Hsf1 was tagged with FLAG and Hsp90 was tagged with the tandem affinity purification ( TAP ) tag . The plasmid pACT1pHSF1 , containing the ACT1p-3xFLAG-HSF1 construct [41] was linearized with StuI and transformed into the strains SN95 ( wild type ) or CaLC501 ( HSP90-TAP ) ( Table 1 ) using published procedures [60] . Nourseothricin ( NAT ) resistant transformants were selected on YPD containing 150 µg/mL NAT , and insertion of the ACT1p-3xFLAG-HSF1 cassette confirmed by diagnostic PCR using the primers oLC1117/oLC1628 ( Table S1 ) . Localisation of Hsp90 was achieved by 3′-tagging of one HSP90 allele with GFP in the wild type strain SN95 ( creating CaLC1855 , Table 1 ) . The GFP-NAT cassette was amplified using primers oLC1616/1617 ( Table S1 ) and transformed into SN95 . Proper integration of the cassette in NAT resistant transformants was confirmed by PCR using primer pairs oLC600/756 ( Table S1 ) . To determine Hsf1 phosphorylation status , the pACT1pHSF1 , containing the ACT1p-3xFLAG-HSF1 construct [41] was linearized with StuI and transformed into mkc1 , hog1 , cek1 , cka1 , cka1 , ckb1 , ckb2 mutants ( Table 1 ) [30] , [32] , [33] , [61] . NAT resistant transformants were selected and confirmed by diagnostic PCR and expression of FLAG-Hsf1 in western blots . Cek1 was tagged with the TAP tag at its C-terminus in the wild type strain SN95 ( creating CaLC2287 , Table 1 ) and its derivative CaLC1411 ( creating CaLC2288 , Table 1 ) using a PCR based strategy as described previously [62] . Briefly , the tag and a selectable marker ( ARG4 ) were PCR amplified from pLC573 ( pFA-TAP-ARG4 [63] ) using oligos oLC2292/2251 ( Table S1 ) . 200 µl of PCR product was run through a PCR clean-up and dissolved in 50 µl sterile water and transformed into C . albicans . Correct genomic integration was verified using appropriate primer pairs that anneal ∼500 bp up ( oLC2252 ) or downstream ( oLC2253 ) from both insertion junctions together with oLC1593 ( TAP-R ) and oLC1594 ( ARG4-F ) that target the TAP and the selectable marker ( Table S1 ) . Total soluble protein was extracted and subjected to western blotting using published protocols [64] , [65] . Briefly , mid-log cells were pelleted by centrifugation , washed with sterile water , and resuspended in lysis buffer ( 50 mM HEPES , pH 7 . 5 , 150 mM NaCl , 5 mM EDTA , 1% Triton X-100 ) . An equal volume of 0 . 5-mm acid-washed beads was added to each tube . Cells were mechanically disrupted on a BioSpec ( Bartlesville , OK ) mini-bead-beater for six 30 second intervals , with 1 minute on ice between each cycle . The beads and cell debris were pelleted by high-speed centrifugation and the supernatant removed for analysis . Protein concentration was determined using a Bradford reagent ( Sigma-Aldrich ) assay . Protein samples were mixed with one-sixth volume of 6× sample buffer containing 0 . 35 M Tris-HCl , 10% ( w/w ) SDS , 36% glycerol , 5% β-mercaptoethanol , and 0 . 012% bromophenol blue . Between 2 µg and 30 µg of protein was loaded in wells of an 8% SDS–PAGE gel . Separated proteins were transferred to a PVDF membrane for 1 hour at 100 V at 4°C . Membranes were blocked in 5% milk or 5% bovine serum albumin ( BSA ) in TBS or PBS containing 0 . 1% Tween-20 at room temperature for 1 hour and subsequently incubated in primary antibody as follows . All primary antibodies ( except those against p38 and p44/42 MAPK ) were left on the membrane for one hour at room temperature . The p38 MAPK and p44/42 MAPK antibodies were incubated overnight at 4°C . Membranes were washed with 1×PBS-T or TBS-T and probed for one hour with secondary antibody dissolved in 1×PBS-T and 5% milk or BSA . Membranes were washed in PBS-T and signals detected using an ECL western blotting kit as per the manufacturer's instructions ( Pierce ) . To detect FLAG-Hsf1 , a 1∶25000 dilution of anti-FLAG HRP conjugated antibody ( Sigma , A8592 ) was used in PBS-T+5% milk [PBS 0 . 1% Tween-20 , 5% ( w/v ) milk] . To detect Act1 , an anti-Act1 antibody was used ( Santa Cruz Biotechnology , sc47778 ) at a 1∶1000 dilution in PBS-T+5% milk . To detect Hsp90 , a 1∶10000 dilution of anti-Hsp90 antibody was used ( courtesy of Bryan Larson ) in PBS-T+5% milk . To detect Hsp70 , a 1∶1000 dilution of anti-Hsp70 antibody ( Enzo Life Sciences , ADI-SPA-822 ) was used in PBS-T+5% milk . To detect Mkc1 and Cek1 phosphorylation [22] , [66] , a 1∶2000 dilution of phospho-p44/42 MAPK ( Erk1/2 ) ( Thr202/Tyr204 ) Rabbit mAb was used ( New England Biolabs , Hitchin , Hertfordshire , UK , #4370 ) in TBS-T+5% BSA [TBS 0 . 1% Tween-20 , 5% ( w/v ) BSA] . For Hog1 phosphorylation [30] , a 1∶2000 dilution of phospho-p38 MAPK ( Thr180/Tyr182 ) rabbit mAB was used in TBS-T+5% BSA ( New England Biolabs , Hitchin , Hertfordshire , UK , #9211 ) . To detect Mkc1-6XHis-FLAG , the anti-FLAG-HRP antibody was used as above . To detect total Hog1 an anti-Hog1 antibody ( Santa Cruz Biotechnology , y-215 ) was diluted 1∶1000 in PBS-T+5% milk . TAP-tagged Cek1 was detected using a 1∶5000 dilution of anti-TAP tag rabbit polyclonal antibody ( Thermoscientific , CAB1001 ) in PBS-T+5% milk . C . albicans cultures were grown to mid-log phase ( OD600 = 0 . 5 ) , cells harvested , washed with sterile H20 and resuspended in 1 ml of lysis buffer ( 20 mM Tris pH 7 . 5 , 100 mM KCl , 5 mM MgCl and 20% glycerol , with one protease inhibitor cocktail per 50 ml ( complete , EDTA-free tablet , Roche Diagnostics , Indianapolis , IN , USA ) , 1 mM PMSF ( EMD Chemicals , Gibbstown , NJ , USA ) and 20 mM sodium molybdate ( Sigma Aldrich Co . , St Louis , MO , USA ) ) . Cells were then disrupted by bead beating twice for 4 minutes with a 7 minute break on ice between cycles . Lysates were centrifuged at 1300×g for three 5-minute cycles , recovering the supernatants at each stage . The combined lysate was then cleared by centrifugation at 21 , 000×g for 10 minutes at 4°C and protein concentrations determined using the Bradford assay [67] . Anti-FLAG immunoprecipitations were performed by diluting protein samples to 2 mg/ml in lysis buffer containing 20 mM sodium molybdate and 0 . 2% Tween , and incubating with anti-FLAG M2 affinity agarose ( Sigma Aldrich ) at 4°C overnight as per the manufacturer's specifications . Unbound material was discarded , the beads washed five times with 1 ml lysis buffer containing 0 . 1% Tween , and the bound proteins eluted by boiling in one volume of 2× sample buffer ( 125 mM Tris-HCl , pH 6 . 8 , 5% glycerol , 2 . 5% SDS , 2 . 5% beta-mecarptoethanol , dH2O , bromophenol blue ) . Anti-IgG immunoprecipitations were performed using the same approach , but using rabbit IgG agarose ( Sigma Aldrich ) as per the manufacturer's specifications . Protein samples were then electrophoresed on 8% SDS-PAGE gels . Proteins were then electrotransferred to PVDF membranes ( Bio-Rad Laboratories , Inc . , Hercules , CA , USA ) and blocked with PBS-T+5% milk . Blots were incubated with antibodies against CaHsp90 ( courtesy of Bryan Larsen ) ( 1∶10000 dilution , [68] ) , or FLAG ( 1∶10000 , Sigma Aldrich Co . ) . To monitor gene expression changes in response to tetO-HSP90 depletion , strains SN95 and CaLC1411 were grown overnight at 30°C in YPD while shaking at 200 rpm . Stationary phase cultures were split , adjusted to an OD600 of 0 . 04 where one culture was treated with doxycycline ( BD Biosciences ) , while the other was left untreated . Cells were grown for 7 hours at 30°C . To monitor gene expression changes in response to heat shock , wild type and SN95 cells were grown to mid-log phase , subjected to a 30°C–42°C heat shock and 50 ml was harvested from each culture at the specified time , centrifuged at 3000 rpm for 2 minutes at 4°C , washed once with dH2O before being frozen at −80°C . RNA was then isolated using the QIAGEN RNeasy kit and cDNA synthesis was performed using the AffinityScript cDNA synthesis kit ( Stratagene ) . PCR was carried out using the SYBR Green JumpStart Taq ReadyMix ( Sigma-Aldrich ) with the following cycle conditions: 94°C for 2 minutes , and 94°C for 15 seconds , 60°C for 1 minute , 72°C for 1 minute , for 40 cycles . All reactions were done in triplicate using the following primer pairs: HSP104 ( oLC1620/1621 ) , HSP90 ( oLC754/755 ) , PGA13 ( oLC2256/2257 ) , PMT4 ( oLC2262/2263 ) , RHR2 ( oLC2266/2267 ) . Transcript levels were normalised to ACT1 ( oLC2285/2286 ) ( Table S1 ) . Data were analysed in the StepOne analysis software ( Applied Biosystems ) . The C . albicans transposon insertion mutant library was generously provided by Aaron Mitchell ( Carnegie Mellon University ) [56] , [58] . Strains were inoculated in 100 µl YPD in 96 well plates and incubated overnight at 30°C with shaking at 200 rpm . Cells were then diluted 1∶10 in YPD and incubated at 30°C with shaking at 200 rpm for 4 hours . After 4 hours of exponential growth , cells were then exposed to a one hour 30°C–42°C heat shock by addition of 100 µl pre-warmed YPD at 54°C and incubating at 42°C , and the growth of each culture monitored continuously for 6 hours by measuring the OD600 . Non-heat shocked control cells received 100 µl YPD at 30°C and were incubated at 30°C . This screen was repeated independently three times , and only those mutants that displayed consistent phenotypes were taken for further analysis . The validity of key transposon mutants highlighted by this screen was then confirmed by subjecting corresponding homozygous null mutants to the same screen . Unless otherwise stated , the susceptibilities of strains were determined with the following stressors: heat stress ( 42°C heat shock ) , Calcofluor White ( CFW: 100 µg/ml ) , Congo Red ( CR: 100 µg/ml ) , H2O2 ( 5 mM ) and NaCl ( 1 M ) . All stress assays were performed in YPD . MIC assays were performed in flat bottom , 96-well microtiter plates ( Corning Costar ) . Briefly , assays were set up in 0 . 2 ml/well , with 2× concentrations of NaCl , CFW and CR prepared in 100 µl YPD . Final concentrations of NaCl were 0 , 0 . 25 , 0 . 5 , 1 . 0 , 1 . 5 and 2 M; and for CFW and CR were 0 , 25 , 50 , 100 , 150 and 200 µg/ml . Cell densities of overnight cultures were determined and dilutions were prepared in YPD such that ∼103 cells were inoculated into each well . Plates were placed statically at either 25°C , 30°C , 37°C or 42°C for 48 hours , after which plates were sealed and cells resuspended by agitation . Absorbance was determined at 600 nm using a spectrophotometer ( VERSA max , Molecular Devices ) , and was corrected for background from the corresponding medium . Every strain was tested in duplicate on three separate occasions . MIC data were quantitatively displayed with colour using the program Java TreeView 1 . 1 . 6 ( http://jtreeview . sourceforge . net ) . For stress cross-protection assays , cells were grown overnight in YPD at 30°C with shaking at 200 rpm . These were diluted to an OD600 = 0 . 2 in fresh YPD , and grown for a further 4 hours at 30°C . Cells were then subjected to a 30 minute heat shock at 42°C , incubated at 30°C and stressed for one hour with CFW , CR , H2O2 or NaCl . Cells were diluted , plated onto YPD and viability determined ( CFUs ) . Control cells were not subjected to the prior heat shock . In other experiments cells were exposed to a prior CFW , CR , H2O2 or NaCl stress for one hour at 30°C before being heat shocked at 42°C for 30 min , and then cell viability determined . To test stress sensitivity following Hsp90 depletion , C . albicans tetO-HSP90 cells ( CaLC1411: Table 1 ) were incubated for 7 hours in YPD containing 20 µg/ml doxycycline at a starting OD600 of 0 . 04 . These cells were left untreated or stressed for one hour with CFW , CR , a 30°C–42°C heat shock , H2O2 or NaCl at the concentrations specified above , and CFUs determined . Controls included CaLC1411 cells that were not treated with doxycycline , and parental SN95 cells that were subjected to the same treatments . Growth curves of CaLC1411 ( tetO-HSP90/hsp90 ) were determined in the absence and presence of doxycycline to ensure adequate viability of this strain ( Supplementary Figure S1C ) . To determine kinase activation in response to Hsp90 depletion , wild type SN95 and tetO-HSP90 ( CaLC1411 ) were grown overnight in YPD at 30°C . Cells were diluted in YPD with or without doxycycline at a starting OD600 of 0 . 04 and incubated for 7 hours . Cells were then left untreated or stressed with H2O2 for 10 minutes and NaCl for 12 minutes to determine Hog1 activation . For Mkc1 and Cek1 activation , tagged versions of these proteins in SN95 and tetO-HSP90 ( CaLC1411 ) were grown as above and stressed with CFW for 30 minutes or a 30°C–42°C heat shock for 30 minutes . Mkc1 total levels were assayed using Mkc1-6XHis-FLAG tagged in SN95 and tetO-HSP90 ( CaLC681 and CaLC648 respectively , Table 1 ) . Total Cek1 levels were assayed using Cek1-TAP tagged in SN95 and tetO-HSP90 ( CaLC2287 and CaLC2288 respectively , Table 1 ) . The chitin content of cells was measured as described previously [69] . Briefly , formalin fixed cells were stained with 25 µg/ml Calcofluor White and fluorescence was preserved with Vectashield mounting medium ( Vector Laboratories , Peterborough , United Kingdom ) . All samples were examined by differential interference contrast ( DIC ) and fluorescence microscopy ( 456 nm ) with a Zeiss Axioplan 2 microscope . Images were captured using a C4742-95 digital camera ( Hamamatsu Photonics , Hamamatsu , Japan ) and analysed using Openlab software ( version 4 . 04: Improvision , Coventry , United Kingdom ) . CFW fluorescence was quantified for 50 individual yeast cells from each sample , using region-of-interest measurements . Mean fluorescence intensities were then calculated and expressed as arbitrary units . Exponentially growing Hsp90-GFP ( CaLC1855 , Table 1 ) cells were heat shocked as described above , and 1 ml of cells were removed at 0 , 10 , 60 and 120 minutes post-heat shock . Cells were harvested and washed in 1 ml of 42°C pre-warmed 1×PBS . Supernatant was removed and cells were resuspended in the remaining 50 µl . Of this , 5 µl was placed on a slide , and cells were heat fixed by placing the slide on a 70°C hot plate for 1 minute . Slides were cooled for a few seconds before 4 µl of 1 µg/ml DAPI ( Fluke , Sigma-Aldrich ) were added . Imaging was performed on a Zeiss Imager M1 upright microscope and AxioCam MRm with AxioVision 4 . 7 software . An X-Cite series120 light source with ET green fluorescent protein ( GFP ) and 4′ , 6-diamidino-2-phenylindole ( DAPI ) hybrid filter sets from ChromaTechnology ( Bellows Falls , VT ) were used for fluorescence microscopy . DAPI fluorescence was viewed under the DAPI hybrid filter and GFP-tagged proteins under the GFP filter . For high pressure freezing transmission electron microscopy ( HPF-TEM ) , cells were prepared by high-pressure freezing with a Leica EM PACT2 ( Leica Microsystems ( UK ) Ltd , Milton Keynes ) . After freezing , cells were freeze-substituted in substitution reagent ( 1% OsO4/0 . 1% uranyl acetate in acetone ) with a Leica EM AFS2 . Samples were encapsulated in 3% ( w/v ) low melting point agarose prior to processing to Spurr resin . Additional infiltration was provided under vacuum at 60°C before embedding in TAAB capsules and polymerizing at 60°C for 48 h . Semi-thin survey sections of 0 . 5 µM thickness were stained with 1% toluidine blue to identify areas of best cell density . Ultrathin sections ( 60 nm ) were prepared with a Diatome diamond knife on a Leica UC6 ultramicrotome , and stained with uranyl acetate and lead citrate for examination with a Philips CM10 transmission microscope ( FEI UK Ltd , Cambridge , UK ) and imaging with a Gatan Bioscan 792 ( Gatan UK , Abingdon , UK ) .
An autoregulatory loop , whereby Hsf1 activates HSP90 expression and Hsp90 interacts with and down-regulates Hsf1 , is thought to lie at the heart of heat shock adaptation in fungi . This presumption provided the basis for mathematical modelling of thermal adaptation in C . albicans [49] , but had not been confirmed experimentally . If this presumption is true , one would expect that inhibition of Hsp90 , or Hsp90 depletion would lead to Hsf1 activation [49] . To test this , we examined the impact of the Hsp90 inhibitors , radicicol and geldanamycin [70] , [71] , upon Hsf1 phosphorylation . C . albicans cells ( ML250: Table 1 ) were incubated with radicicol or geldanamycin for up to one hour , and Hsf1 phosphorylation was monitored via the resultant band shift revealed by western blot analysis , as described previously [41] ( Figure 1A ) . These Hsp90 inhibitors induced Hsf1 phosphorylation after one hour , as confirmed by the band shift observed after controlled dephosphorylation with lambda phosphatase . To exclude the possibility that this Hsf1 phosphorylation was induced by a general effect of the drugs upon protein folding , we examined the impact of dithiothreitol and tunicamycin , known inducers of the unfolded protein response in C . albicans [72] . Hsf1 was not activated following treatment with 5 mM dithiothreitol or 4 . 73 µM tunicamycin for 1 hour ( data not shown ) , suggesting that the induction of Hsf1 phosphorylation by radicicol and geldanamycin related to Hsp90 function rather than some general effect upon protein folding . Therefore , Hsp90 inhibits Hsf1 , as predicted . Next , we validated our pharmacological findings using a genetic approach . A doxycycline conditional C . albicans HSP90 mutant ( tetO-HSP90 ) [73] , in which HSP90 expression is independent of Hsf1 , was used to ectopically down-regulate Hsp90 levels ( Figure 1B and S1A ) . Doxycycline treatment led to Hsf1 phosphorylation after approximately 6 hours , at which point Hsp90 levels were reduced by 50% ( Figure 1B and S1A ) . Therefore , ectopic down-regulation of HSP90 caused Hsf1 phosphorylation even in the absence of a heat shock , further reinforcing the hypothesis that Hsp90 inhibits Hsf1 . A key finding from our modelling of thermal adaptation was that this system displays perfect adaptation: i . e . Hsf1 activation returns to basal levels within two hours once cells have adapted to their new ambient temperature [49] . If Hsp90 down-regulates the heat shock response , then one would expect this perfect adaptation to be dependent on Hsp90 . Also , if new Hsp90 synthesis is inhibited after a heat shock , Hsf1 would remain phosphorylated and the system would not adapt to this stress . We tested this by examining Hsf1 phosphorylation levels in doxycycline-treated C . albicans tetO-HSP90 cells after a 30°C–42°C heat shock . As predicted , these Hsp90-depleted cells were unable to recover , as revealed by the maintenance of Hsf1 phosphorylation four hours after the heat shock ( Figure 1C ) . Therefore , perfect thermal adaptation is dependent upon Hsp90 . To further test the impact of depleting Hsp90 on the heat shock response , we looked at induction of HSP104 , a known target of Hsf1 [49] . In wild type cells ( SN95: Table 1 ) HSP104 was up-regulated approximately 10-fold in response to a 30 minute 30°C–42°C heat shock ( Figure 1D ) , which was consistent with previous findings [49] , and HSP104 expression was not affected by doxycycline treatment . HSP104 mRNA levels were elevated in tetO-HSP90 cells even in the absence of doxycycline , presumably because Hsp90 levels are significantly reduced under these conditions ( Figure S1B ) . An additional increase in HSP104 expression was observed after doxycycline treatment when Hsp90 levels were reduced further ( Figure S1B ) . Taken together , these data strongly imply that Hsp90 is a master regulator of the heat shock response . The proposed Hsf1-Hsp90 autoregulatory loop [49] suggests a physical interaction between Hsf1 and Hsp90 . There is some evidence for this in mammalian cells [52] , [53] , but none in fungal systems . Therefore we tested this experimentally by co-immunoprecipitation . First , proteins were extracted from C . albicans cells expressing a FLAG-tagged Hsf1 protein ( CaLC1819 ) , and from control cells in which Hsf1 was not FLAG-tagged ( WT , SN95 ) . Protein extracts were incubated with anti-FLAG beads , and the resulting immunoprecipitates probed for Hsp90 on western blots . Hsp90 was observed reproducibly in immunoprecipitates from the FLAG-Hsf1 expressing strain , but not in those from the untagged control strain ( Figure 2A ) . Reciprocal immunoprecipitations were then performed to test the validity of this apparent Hsf1-Hsp90 interaction . C . albicans cells expressing both TAP-tagged Hsp90 and FLAG-tagged Hsf1 ( CaLC1875 ) were used in these experiments , and cells lacking the FLAG-tagged Hsf1 ( CaLC501: Table 1 ) were used as a control . Hsp90 was immunoprecipitated using IgG beads , which bind the Protein A in the TAP tag . These immunoprecipitates were then probed for the FLAG-tagged Hsf1 , revealing a band of the appropriate mass from cells expressing TAP-tagged Hsp90 , but not from the controls ( Figure 2B ) . This confirmed that Hsp90 interacts physically with Hsf1 . This is the first demonstration of a physical interaction between Hsf1 and Hsp90 in any yeast . An Hsf1-Hsp90 autoregulatory loop , as inferred by the model [49] , [74] , predicts dynamic changes in the Hsf1-Hsp90 interaction during a heat shock . The model predicts that Hsp90 releases Hsf1 following a heat shock , rebinding Hsf1 as the response is down-regulated . We tested this hypothesis by co-immunoprecipitation , examining Hsp90-Hsf1 interactions over a 60 minute period following heat shock . Rather than decreasing when Hsf1 becomes activated ( as was predicted ) , the Hsp90-Hsf1 interaction increased over 60 minutes ( Figure 2C ) during the period when Hsf1 activation is maximal [49] . Therefore , Hsf1 is not released from Hsp90 after heat shock , and this mechanism cannot account for Hsf1 activation . We examined the Hsf1-Hsp90 interaction two hours after heat shock , when the response is down-regulated [49] , and found that the interaction was still increased when compared to untreated samples or a 10 minute heat shock ( Figure S2B , bottom panel ) . This increase in the Hsf1-Hsp90 interaction after a prolonged heat shock ( Figure 2B ) , accompanies the decline in Hsf1 activity and the down-regulation of the heat shock response [49] . To determine whether other components of the Hsp90 chaperone machinery might also interact with Hsf1 , we focused on Hsp70 , which operates within the Hsp90 chaperone system [74] , and has been implicated in Hsf1 regulation . Indeed , the deletion of SSA1 and SSA2 , which encode cytosolic isoforms of Hsp70 derepresses Hsf1 transcriptional activity in S . cerevisiae [75] . However , a physical interaction between Hsp70 and Hsf1 has not been reported in any yeast . Therefore , we re-probed our FLAG-Hsf1 immunoprecipitations with an antibody against Hsp70 . We found that Hsp70 interacts with Hsf1 , but only in response to a prolonged heat shock ( Figure 2C , bottom left panel , and Figure S2A ) . We also re-probed our Hsp90-TAP immunoprecipitations for Hsp70 . Our data suggest that Hsp90 and Hsp70 interact in response to prolonged heat shock ( Figure 2C , bottom right panel and Figure S2B , top panel ) . Given our findings that the Hsf1-Hsp90 interaction strengthens upon heat shock , and that Hsf1 is thought to bind DNA constitutively [44] , one might predict that Hsp90 localises to the nucleus upon heat shock . Indeed , a recent study by Lamoth and colleagues shows Hsp90 nuclear localisation upon a 55°C heat shock in Aspergillus fumigatus [76] . Therefore we followed the localisation of Hsp90-GFP in C . albicans in response to a heat shock ( CaLC1855 , Table 1 ) . Under steady state conditions , Hsp90-GFP was distributed throughout the cell , with no obvious localisation ( Figure 2D ) , and this was also the case 10 minutes after a 42°C heat shock . However , 60 minutes post-heat shock , nuclear accumulation of Hsp90-GFP was clearly evident ( Figure 2D ) , and Hsp90-GFP remained in the nucleus 120 minutes post heat shock ( Figure S2C ) . Therefore the dynamics of the nuclear accumulation of Hsp90-GFP correlated with the dynamics of the Hsf1-Hsp90 interaction . These data reinforce the view that Hsp90 and the Hsp90 chaperone machine plays an important role in the down-regulation of Hsf1 and the heat shock response . The above observations indicate that while Hsp90 down-regulates Hsf1 , Hsf1 activation is mediated by Hsp90-independent mechanisms . How then is Hsf1 activated ? C . albicans Hsf1 is activated by phosphorylation [41] , but the protein kinase responsible for this in yeast remains unknown [52] , [53] . To determine which kinase is responsible for Hsf1 phosphorylation in C . albicans , we exploited the recent availability of the C . albicans transposon insertion kinase mutant collection which was kindly provided by Aaron Mitchell [58] . This collection of mutants which comprises homozygous insertion or deletion mutations in 67 protein kinase genes and 13 protein kinase-related genes , were screened for temperature sensitivity by monitoring their growth following a one hour 30°C–42°C heat shock . Mutants that consistently displayed a >10% growth defect relative to the wild-type controls in three independent screens were considered temperature sensitive , and those that consistently displayed >10% growth acceleration were considered temperature resistant ( Figure 3 ) . The tetO-HSF1 mutant ( CLM60-1: Table 1 ) was always temperature sensitive , and the known temperature sensitivity of mkc1 mutants [32] was consistently recapitulated . ( Mkc1 is the orthologue of S . cerevisiae Slt2 [32] . ) These observations leant weight to the validity of the output from this screen . In this study we focussed on those kinases whose inactivation confers temperature sensitivity . Numerous kinase mutants displayed significant and reproducible temperature sensitivity in our screen , notably the casein kinase subunits as well as components of key MAP kinase pathways such as the cell wall integrity pathway ( Mkc1 , Pkc1 ) , the osmotic stress pathway ( Pbs2 ) and the starvation and cell wall stress pathway ( Ste11 ) ( Figure 3 ) . Several downstream kinases of these pathways were absent from the mutant collection , prompting us to perform a secondary screen of null mutants . This secondary screen confirmed the temperature sensitivity of all of the transposon mutants tested from the primary screen ( Figure 3 ) , including the casein kinase subunits ( Cka1/2 and Ckb1/2 ) and Mkc1 . Furthermore , the secondary screen tested the MAP kinases Cek1 and Hog1 ( which were missing from the collection of transposon mutants ) , thereby confirming the involvement of these pathways in C . albicans thermotolerance ( Figure 3 ) . Our next aim was to test whether any of the kinases identified in the above screen are responsible for Hsf1 phosphorylation in response to heat shock . Hsf1 was FLAG3-tagged at its N-terminus in CAI4 ( MLC67 ) and in each of the following temperature sensitive null mutants: mkc1 ( MLC30 ) , hog1 ( MLC15 ) , cek1 ( MLC21 ) , cka1 ( MLC24 ) , cka2 ( MLC27 ) , ckb1 ( CaLC2259 ) and ckb2 ( CaLC2261 ) ( Table 1 ) . Each of these mutants was then subjected to a 30°C–42°C heat shock , and Hsf1 phosphorylation assayed at 0 , 10 , 30 and 60 minutes post-heat shock ( Figure S3 ) . All mutants displayed similar Hsf1 phosphorylation dynamics to wild-type cells , indicating that none of the kinases alone is essential for Hsf1 phosphorylation . Therefore , there might be functional redundancy with respect to Hsf1 phosphorylation during heat shock . Alternatively , Hsf1 might be phosphorylated by an essential kinase that was not represented in the kinase mutant collection . These observations also suggest that the MAP kinase pathways might contribute to thermal adaptation in C . albicans by mechanisms other than via Hsf1 phosphorylation and the activation of the heat shock regulon . Next we explored how the signalling pathways contribute to thermotolerance . A recent and elegant study by Diezmann and colleagues [62] demonstrated that casein kinase 2 ( CK2 ) regulates Hsp90 phosphorylation and activity . Therefore , the mutations in CK2 subunits ( cka1 , cka2 , ckb1 and ckb2 ) probably reduce thermotolerance by interfering with Hsp90 function . In addition to regulating Hsf1 ( Figures 1 and 2 ) , which is critical for thermal adaptation [49] , Hsp90 has numerous other client proteins , some of which may contribute to thermotolerance [62] . We therefore focussed on the roles of the three MAP kinase pathways that were highlighted by our screen: the cell wall integrity Mkc1 pathway , the osmolarity/oxidative stress Hog1 pathway and the starvation/cell wall response Cek1 pathway . First we monitored the activation of each MAP kinase during heat shock . Wild-type CAI4 cells were subjected to a 30°C–42°C heat shock , proteins extracted at 0 , 10 , 30 and 60 minutes , and MAP kinase phosphorylation probed by western blotting . Each MAP kinase responded differently to the thermal upshift ( Figure 4A ) . Mkc1 was rapidly dephosphorylated before being re-phosphorylated in the longer term . Cek1 phosphorylation levels remained relatively stable . Hog1 was rapidly dephosphorylated , and was not reactivated during the one hour period examined ( Figure 4A ) . These data were entirely consistent with previous studies that have reported effects of temperature on Hog1 , Cek1 and Mkc1 phosphorylation [22] , [30] , [77] , [78] . Significantly , total kinase levels remain unchanged during the 60 minute heat shock . To further validate the effects of heat shock upon these signalling pathways , we examined the expression of key MAP kinase targets: PGA13 is an Mkc1 target [78] , PMT4 is a Cek1 target [79] and RHR2 is a Hog1 target [30] . RNA was extracted from cells 0 , 10 , 30 and 60 minutes after a 30°C–42°C heat shock , and transcript levels assayed by qRT-PCR relative to the internal ACT1 mRNA control . The levels of these targets reflected the activation profiles of the corresponding MAP kinase ( Figure 4B ) . MAP kinase signalling is often represented in terms of linear pathways . However , in C . albicans , as in other organisms there is emerging evidence for cross-talk between these pathways [22] , [80] , [81] , [82] . Therefore we tested whether they interact during thermal adaptation by comparing the phosphorylation status of the terminal kinases in wild type and MAP kinase mutants following a 30°C–42°C heat shock ( Figure 5 ) . First , we examined Cek1 and Hog1 activation in an mkc1 mutant ( Figure 5A ) . Mkc1 inactivation did not affect the responses of either Hog1 or Cek1 . Second , we monitored Hog1 and Mkc1 phosphorylation in a cek1 mutant ( Figure 5B ) . After Cek1 inactivation , the transient dephosphorylation of Mkc1 was inhibited . Hog1 phosphorylation levels still declined over the duration of the heat shock . Third , we tested the effects of Hog1 inactivation upon Mkc1 and Cek1 ( Figure 5C ) . Mkc1 phosphorylation was minimal and remained low when hog1 cells were exposed to the heat shock . In contrast , we observed that Cek1 was hyperphosphorylated in hog1 cells , as reported previously [22] , [81] , [83] . Our data reveal that there are significant interactions between the MAP kinase signalling pathways during thermal adaptation . Given this cross-talk and the modulation of MAP kinase activities after thermal up-shifts , we reasoned that ambient temperature is likely to influence the resistance of C . albicans cells to those stresses normally associated with these signalling pathways . For example , Hog1 signalling mediates osmotic stress adaptation [84] , and the Mkc1 and Cek1 pathways promote resistance to the cell wall stresses Calcofluor White and Congo Red [22] . Therefore we performed MICs to test the effects of ambient temperature upon the resistance of wild-type ( NGY152: Table 1 ) , hog1 , cek1 and mkc1 cells to these stresses ( Figure 6 ) . The mkc1 cells were temperature sensitive , as reported [32] , [84] . Also , as reported previously , hog1 cells were sensitive to NaCl [84] . Furthermore , mkc1 and cek1 cells displayed sensitivity to Calcofluor White [22] , [81] , [83] . These controls faithfully replicated previous observations . We observed that ambient temperature significantly influences the sensitivity of C . albicans cells to cell wall , but not osmotic stresses ( Figure 6 ) . Firstly , wild-type cells were more sensitive to Calcofluor White at lower temperatures ( 25°C and 30°C ) . Secondly , hog1 cells were relatively resistant to Calcofluor White at all temperatures tested , and phenocopied wild type cells on Congo Red . Thirdly , cek1 cells were resistant to Congo Red at most temperatures , but sensitive at 42°C . Clearly ambient temperature significantly affects cell wall stress resistance . These data reinforce the notion of cross-talk between thermal and cell wall stress signalling pathways . We reasoned that the Calcofluor resistance of hog1 cells at low temperatures ( Figure 6 ) might be Cek1 dependent . This is because inactivation of Hog1 led to elevated Cek1 phosphorylation levels ( Figure 5C ) , and Cek1 promotes Calcofluor White resistance ( Figure 6 ) . Hog1 and cek1 mutations are synthetically lethal [81] , and hence we could not examine a hog1 cek1 double mutant . Therefore , instead we tested the phenotype of a hog1/hog1 hst7/hst7 double mutant , in which Cek1 signalling is blocked [81] . As predicted , the inactivation of Cek1 signalling attenuated the Calcofluor White resistance of hog1 cells at low temperatures ( Figure 6 ) . This indicates that Hog1 inactivation promotes cell wall stress resistance at low temperatures via Cek1 signalling . These data reinforce the importance of cross-talk between the MAP kinase signalling pathways and highlight the relevance of this cross-talk for thermal adaptation . In some yeasts , the core transcriptional response to stress underpins the phenomenon of stress cross-protection , whereby exposure to one stress protects the cell against subsequent exposure to an alternative type of stress via the up-regulation of key stress response genes [2] , [4] , [15] , [16] . The core stress response is limited in C . albicans [19] . Nevertheless , it remained conceivable that the effects of ambient temperature upon the resistance of C . albicans to certain stresses might be mediated through stress cross-protection . To test this , mid-exponential C . albicans cells ( NGY152: Table 1 ) were subjected to a 30 minute 30°C–42°C heat shock and subsequently exposed to a cell wall stress ( Congo Red or Calcofluor White ) , osmotic stress ( NaCl ) , or oxidative stress ( hydrogen peroxide ) . Cell wall or osmotic stress resistance was not enhanced by prior exposure to heat shock ( Figure 7A ) . Furthermore , prior exposure to cell wall or osmotic stress resistance did not enhance resistance to a subsequent heat shock ( Figure 7B ) . This was consistent with our observation that ambient temperature does not significantly affect osmotic stress resistance ( Figure 6 ) , and indicated that the influence of ambient temperature upon Calcofluor White resistance is not mediated by stress cross-protection . This was consistent with the divergent core stress response in C . albicans [3] , [19] . We included oxidative stress as a control in the above experiments because a prior heat shock has been reported to protect C . albicans against peroxide stress [3] . The mechanisms by which a heat shock protects cells against a subsequent oxidative stress have not been elucidated . We noted that two uncharacterised genes that are induced by oxidative stress are also up-regulated by heat shock: orf19 . 7882 and orf19 . 7085 [3] , [41] , [85] . Both genes are induced in response to oxidative stress in a Cap1-dependent fashion , and are down-regulated by Hog1 . Therefore we tested whether Hog1 and Cap1 are required for the observed stress cross-protection ( Figure 7C ) . The hog1 mutant ( JC50 ) displayed a comparable increase in survival to wild type cells when cells were pre-treated with a 30°C–42°C heat shock and then exposed to hydrogen peroxide . In contrast , cap1 cells ( JC128: Table 1 ) lost this stress cross-protection . Therefore , the acquired resistance to hydrogen peroxide after exposure to heat shock is dependent on Cap1 . This probably occurs through the Cap1 dependent up-regulation of oxidative stress genes such as orf19 . 7882 and orf19 . 7085 in response to heat shock . What mechanisms are responsible for the cross-talk between thermotolerance and stress adaptation if this is not mediated by stress cross-protection ? Hsp90 is a key regulator of thermal adaptation , regulating its own expression via Hsf1 ( Figure 1 ) . In addition , Hsp90 modulates the activities of multifarious client proteins [62] . This list of Hsp90 interactors in C . albicans includes CK2 subunits [62] , Mkc1 , a well-defined Hsp90 protein client in C . albicans [86] , and Hog1 [62] . Mkc1 and Hog1 orthologues are also known Hsp90 client proteins in S . cerevisiae [87] , [88] . As these protein kinases were identified in our screen as being important for thermotolerance , we reasoned that Hsp90 might play a significant role in coordinating thermal adaptation in C . albicans . According to this hypothesis , changes in ambient temperature are expected to influence Hsp90 availability [88] and this in turn modulates the activity of its client proteins [49] . These client proteins include Hog1 and Mkc1 which , when activated , induce expression of target genes that promote cellular adaptation to elevated temperatures and other stresses . A clear prediction of this hypothesis is that Hsp90 depletion should attenuate the cell's ability to withstand specific stresses . To test this , tetO-HSP90 cells were treated with 20 µg/ml doxycycline for 7 hours , by which point Hsp90 levels were significantly reduced ( Figure S1A and S1B ) and growth was beginning to slow ( Figure S1C ) . These cells were then stressed for one hour with Calcofluor White , Congo Red , a 30°C–42°C heat shock , hydrogen peroxide or NaCl ( Materials and Methods ) . As predicted , Hsp90 depletion significantly attenuated cellular resistance to all of the stresses tested except NaCl when compared to wild type or tetO-HSP90 cells not treated with doxycycline ( Figure 8A ) . To test this further , we examined the effects of the Hsp90 inhibitor geldanamycin upon stress resistance . Although the differences were not as dramatic as for genetic depletion of Hsp90 ( Figure 8A ) , similar effects were observed following geldanamycin treatment ( Figure 8B ) . The fact that Hsp90 depletion did not affect osmotic stress resistance was entirely consistent with our previous findings that ambient temperature did not influence osmotic stress resistance ( Figure 6 ) , and that there was no stress cross-protection for thermal and osmotic stresses ( Figure 7 ) . Clearly Hsp90 influences cellular responses to a range of stresses , not only to heat shock . To determine whether these effects are mediated through its client proteins we assessed the impact of Hsp90 depletion upon the activation of known client proteins , Mkc1 and Hog1 , and the potential client protein , Cek1 ( Figure 9 ) . The basal activation of each MAP kinase was examined in wild type and tetO-HSP90 ( SN95 and CaLC1411: Table 1 ) in the absence of stress , and following the imposition of stress . Mkc1 activation levels were attenuated following Hsp90 depletion . This was the case in the absence of stress and following heat shock or Calcofluor White treatment ( Figures 9A and 9B ) . This corresponds with a decrease in total Mkc1 kinase levels following Hsp90 depletion . In contrast , Hsp90 depletion had no effect on the levels of Cek1 activation in the absence of stress , but led to an increase in Cek1 phosphorylation following Calcofluor White treatment ( Figures 9A and 9B ) . With regard to Hog1 , the basal levels of phosphorylation were maintained following Hsp90 depletion in the absence of stress ( Figure 9C ) , and Hog1 activation was not attenuated in response to osmotic stress although total Hog1 levels decreased ( Figure 9D ) . This was entirely consistent with our other findings , whereby ambient temperature did not affect osmotic tress resistance ( Figure 6 ) and a prior heat shock did not protect cells against a subsequent osmotic stress ( Figure 7 ) . However , Hsp90 depletion blocked Hog1 activation following exposure to hydrogen peroxide ( Figure 9C ) . We conclude that Hsp90 depletion exerts differential effects upon Hog1 , Mkc1 and Cek1 . Furthermore the data are consistent with our prediction that Hsp90 modulates the activities of these MAP kinases . It was conceivable that Cek1 is an Hsp90 client protein . To test this we determined the impact of Hsp90 depletion on Cek1 stability as Hsp90 client proteins are generally destabilised in the absence of Hsp90 [62] , [86] , [89] . Cek1 was TAP-tagged at its C-terminus and the specificity of this tagging was verified by western blotting alongside untagged SN95 and CaLC1411 controls ( Figure S4 ) . The levels of this Cek1-TAP protein were then monitored in the tetO-HSP90 strain CaLC1411 ( CaLC2288 , Table 1 ) and in wild-type SN95 ( CaLC2287 , Table 1 ) cells ( Figure 10 ) . Cek1-TAP protein levels decreased in response to Hsp90 depletion in the absence of stress , as well as in response to heat shock and Calcofluor White treatment ( Figure 10 ) . This suggested that Cek1 is destabilised by Hsp90 depletion and that Cek1 is a client protein of Hsp90 . Therefore the contribution of Cek1 to thermal adaptation appears to be modulated by Hsp90 . Several observations infer a link between stress adaptation and cell wall architecture in C . albicans . For example , the Pkc1/Mkc1 cell wall salvage pathway is activated by certain stresses [58] , [66] , [90] . Also , the Hog1 stress pathway has been implicated in cell wall biosynthesis [81] , partly by regulating chitin synthesis [66] , [69] . If Hsp90 depletion modulates Hog1 , Mkc1 and Cek1 signalling ( Figure 9 ) , and these pathways contribute to cell wall biogenesis , we reasoned that Hsp90 could regulate cell wall architecture . We took two approaches to test this hypothesis . First , we tested the effects of Hsp90 depletion on chitin levels by Calcofluor White staining . Chitin content increased more than two-fold following doxycycline treatment of tetO-HSP90 cells compared to control cells ( Figures 11A and 11B ) . Second , we examined the impact of Hsp90 depletion upon cell wall architecture by transmission electron microscopy ( Figure 11C ) . Cell wall thickness increased two-fold after Hsp90 depletion compared to the controls . Therefore , Hsp90 is essential for normal cell wall structure , providing the first ever link between Hsp90 and cell wall architecture .
Our results illuminate novel functional connections between key cellular regulators required for thermotolerance , and establish distinct roles for Hsp90 in orchestrating short term versus long term mechanisms of thermal adaptation . We provide evidence that Hsf1 is a client protein of Hsp90 for the first time in any fungus , demonstrating that Hsp90 contributes to the short term regulation of thermal adaptation . Furthermore , we identified several key signalling pathways that contribute to thermotolerance in C . albicans . These include the Mkc1 , Hog1 and Cek1 pathways , each of which plays a role in cell wall integrity . These pathways contribute to thermal adaptation in the longer term via cell wall remodelling , and Hsp90 links many of these pathways , as critical components of these pathways are Hsp90 client proteins . Finally , we establish a role for Hsp90 in cell wall biogenesis for the first time in any organism . Overall , we see that Hsp90 drives short term thermal adaptation via down-regulation of Hsf1 , and longer term adaptation through modulation of other client proteins , leading to a more robust cell wall . Hsf1 is known to activate HSP90 expression [41] and Hsp90 has been predicted to regulate Hsf1 [49] . Here we demonstrate that inhibiting Hsp90 in C . albicans , either pharmacologically or genetically , derepresses Hsf1 ( Figure 1 ) , indicating that Hsp90 down-regulates Hsf1 . Furthermore , Hsf1 physically interacts with Hsp90 under steady state conditions ( Figure 2A ) , confirming for the first time in any fungus that Hsf1 is an Hsp90 client . What's more , we have shown that this interaction increases during heat shock , and also leads to the recruitment of Hsp70 , which binds Hsf1 and Hsp90 during a prolonged heat shock ( Figures 2B and S1 ) . This increased interaction correlates with an increase in the levels of both Hsp90 and Hsf1 that occurs during heat shock . Consistent with these findings , we found that Hsp90 accumulates in the nucleus upon a prolonged heat shock ( Figure 2C and Figure S1 ) . Similarly , the S . cerevisiae Hsp70 chaperone Ssa1 has been shown to enter the nucleus one hour after a 42°C heat shock [91] . This is consistent with our observation that Hsp70 binds Hsf1 and Hsp90 one hour after a 37°C–42°C heat shock . Therefore short-term thermal adaptation involves the activation of Hsf1 by Hsp90-independent mechanisms . Hsf1 induces the expression of Hsp90 and other chaperones that promote protein folding and repair proteotoxic damage [39] , [41] . Hsp90 then down-regulates Hsf1 thereby dampening the heat shock response once adaptation is achieved . This regulatory circuit is central to thermal adaptation in C . albicans [41] , [49] and may be conserved across the eukaryotic kingdom as pharmacological inhibition of Hsp90 derepresses Hsf1 orthologues in S . cerevisiae and mammalian systems [52] , [55] . With a view to identifying the kinase responsible for phosphorylating Hsf1 in C . albicans we performed a screen for protein kinase mutants that are temperature sensitive ( Figure 3 ) . This highlighted several critical regulators on key MAP kinase pathways , including the Mkc1 , Hog1 and Cek1 pathways . None of these pathways are essential for Hsf1 phosphorylation ( Figure S3 ) , suggesting that there is functional redundancy with respect to Hsf1 activation during heat shock , or that these pathways act independently of Hsf1 in promoting thermotolerance . These pathways are differentially activated during heat shock ( Figure 4 ) , and there is cross-talk between these pathways under these conditions ( Figure 5 ) . Furthermore , ambient temperature significantly affects the resistance of C . albicans cells to cell wall stresses , and these effects are influenced by Cek1 , Hog1 and Mkc1 ( Figure 6 ) . Each of these MAP kinase pathways is known to contribute to cell wall remodelling [32] , [81] , [83] , [90] , and mutations that interfere with cell wall synthesis are known to confer temperature sensitivity upon C . albicans . For example , the inactivation of certain protein mannosyltransferases of the PMT family , or the deletion of OCH1 can confer temperature sensitivity [92] , [93] . Additionally , deletion of SSR1 , a GPI-anchored cell wall protein causes sensitivity to elevated temperatures [94] . These data strongly suggest that Mkc1 , Hog1 and Cek1 signalling promotes longer term thermotolerance via the maintenance of a robust cell wall ( Figure 12 ) . As an environmentally contingent hub of protein homeostasis and regulatory circuitry , Hsp90 has profound effects on biology , disease , and evolution . Hsp90 modulates the phenotypic effects of genetic variation in an environmentally responsive manner [95] , [96] , [97] , [98] , influencing approximately 20% of observed natural genetic variation and serving to maintain phenotypic robustness and promote diversification [96] . Mkc1 was defined as a protein client of Hsp90 after Mkc1 signalling was shown to contribute to antifungal drug tolerance in C . albicans [86] . Hog1 was subsequently shown to be an Hsp90 client protein after Diezmann and co-workers identified this MAPK in a chemogenetic screen of the C . albicans Hsp90 interactome [62] . However , Cek1 was not highlighted in this screen . We demonstrate here that , like Mkc1 and Hog1 , Cek1 is an Hsp90 client protein ( Figure 10 ) . This is in keeping with a recent study by Taipale and colleagues which demonstrates that Hsp90 binds about 60% of mammalian kinases [99] . Therefore we tested whether Mkc1 , Hog1 and Cek1 signalling is influenced by Hsp90 during thermal upshifts . We found that Hsp90 does influence the activation of these kinases during heat shock and in response to their respective stresses ( Figure 9 ) . Furthermore , Hsp90 depletion influenced the sensitivity of C . albicans cells to specific stresses such as cell wall and oxidative stress , as well as to heat shock ( Figure 8 ) . Clearly Hsp90 modulates Mkc1 , Hog1 and Cek1 signalling and their outputs , and these pathways are known to contribute to cell wall architecture . Therefore , we reasoned that Hsp90 might , in part , influence cell wall remodelling ( Figure 12 ) . We confirmed this hypothesis by demonstrating that Hsp90 depletion significantly increases the chitin content and thickness of C . albicans cell walls ( Figure 11 ) . We note that cell wall robustness does not correlate with cell wall thickness . Indeed Ene and co-workers have recently shown that cell wall architecture is altered by growth on different carbon sources , yielding thinner cell walls that are more robust , leading to increased stress resistance [100] . Therefore , Hsp90 coordinates both short term thermal adaptation ( via Hsf1 down-regulation ) , and long term thermal adaptation ( via its client proteins Mkc1 , Hog1 and Cek1 ) ( Figure 12 ) . Additional pathways that were not highlighted in our screen might contribute to thermotolerance . For example some protein kinases that encode essential functions were missing from the transposon library . Indeed , the protein kinase responsible for Hsf1 phosphorylation remains obscure . It is worth noting that as Hsf1 is activated in response to Hsp90 depletion , and as such it is unlikely that the Hsf1 kinase requires stabilisation by Hsp90 . One must also note that Hsf1 might be phosphorylated by multiple protein kinases , and hence that there may be functional redundancy with respect to Hsf1 phosphorylation . Indeed , NetPhos2 . 0 analysis of Hsf1 suggests multiple phosphorylation sites for multiple kinases . Clearly ambient temperature plays a major role in fungal pathogenicity [42] . Furthermore our data indicate that ambient temperature strongly influences physiological attributes that contribute to fungal pathogenicity , such as a robust cell wall and effective stress adaptation [11] , [13] , [42] , [49] . In addition we show that increases in ambient temperature lead to elevated oxidative stress resistance , but that the reverse is not true ( Figure 7 ) . This observation is consistent with the idea of “asymmetric adaptive prediction” , whereby microbes appear to have “learned” over evolutionary timescales that exposure to one stress is likely to be followed by exposure to a second unrelated stress [101] . As a result , exposure to the first stress results in the activation of an adaptive response that prepares the cell for exposure to the second stress [101] . With respect to C . albicans , the elevated temperatures associated with localised inflammation , appear to protect the fungal cells against the imminent exposure to oxidative stress that will follow exposure to macrophages and neutrophils . In conclusion this study reveals new Hsp90 client proteins that play central roles in the control of cellular adaptation: Hsf1 and Cek1 . Furthermore , this work provides important new insights into the mechanisms by which Hsp90 coordinates short and long term mechanisms that contribute to thermotolerance in a major fungal pathogen . | Candida albicans is one of the most persistent yeast pathogens known to man , causing frequent mucosal infections ( thrush ) in otherwise healthy individuals , and potentially fatal bloodstream infections in immunocompromised patients . C . albicans colonises warm-blooded animals and occupies thermally buffered niches . Yet during its evolution this pathogen has retained the classic heat shock response whilst other stress responses have diverged significantly . We have established that the essential , evolutionarily conserved molecular chaperone , Hsp90 , coordinates thermal adaptation . Hsp90 interacts with and modulates the activity of the heat shock transcription factor , Hsf1 , thereby controlling the expression of heat shock proteins required for the clearance of proteins damaged by proteotoxic stresses . In addition , Hsp90 modulates the activities of key MAP kinase signalling pathways that mediate cell wall remodelling and long term adaptation to heat shock . Loss of any of these factors results in a significant reduction in thermotolerance . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"signal",
"transduction",
"mycology",
"signaling",
"in",
"cellular",
"processes",
"model",
"organisms",
"molecular",
"cell",
"biology",
"cellular",
"stress",
"responses",
"cell",
"biology",
"microbial",
"pathogens",
"yeast",
"and",
"fungal",
"models",
"biology",
"microbiology",
"fungal",
"physiology",
"candida",
"albicans"
] | 2012 | Hsp90 Orchestrates Transcriptional Regulation by Hsf1 and Cell Wall Remodelling by MAPK Signalling during Thermal Adaptation in a Pathogenic Yeast |
Chronic Chagas disease presents several different clinical manifestations ranging from asymptomatic to severe cardiac and/or digestive clinical forms . Several studies have demonstrated that immunoregulatory mechanisms are important processes for the control of the intense immune activity observed in the chronic phase . T cells play a critical role in parasite specific and non-specific immune response elicited by the host against Trypanosoma cruzi . Specifically , memory T cells , which are basically classified as central and effector memory cells , might have a distinct migratory activity , role and function during the human Chagas disease . Based on the hypothesis that the disease severity in humans is correlated to the quality of immune responses against T . cruzi , we evaluated the memory profile of peripheral CD4+ and CD8+ T lymphocytes as well as its cytokine secretion before and after in vitro antigenic stimulation . We evaluated cellular response from non-infected individuals ( NI ) , patients with indeterminate ( IND ) or cardiac ( CARD ) clinical forms of Chagas disease . The expression of CD45RA , CD45RO and CCR7 surface molecules was determined on CD4+ and CD8+ T lymphocytes; the pattern of intracellular cytokines ( IFN-γ , IL-10 ) synthesized by naive and memory cells was determined by flow cytometry . Our results revealed that IND and CARD patients have relatively lower percentages of naive ( CD45RAhigh ) CD4+ and CD8+ T cells . However , statistical analysis of ex-vivo profiles of CD4+ T cells showed that IND have lower percentage of CD45RAhigh in relation to non-infected individuals , but not in relation to CARD . Elevated percentages of memory ( CD45ROhigh ) CD4+ T cells were also demonstrated in infected individuals , although statistically significant differences were only observed between IND and NI groups . Furthermore , when we analyzed the profile of secreted cytokines , we observed that CARD patients presented a significantly higher percentage of CD8+CD45RAhigh IFN-γ-producing cells in control cultures and after antigen pulsing with soluble epimastigote antigens . Based on a correlation between the frequency of IFN-γ producing CD8+ T cells in the T cell memory compartment and the chronic chagasic myocarditis , we propose that memory T cells can be involved in the induction of the development of the severe clinical forms of the Chagas disease by mechanisms modulated by IFN-γ . Furthermore , we showed that individuals from IND group presented more TCM CD4+ T cells , which may induce a regulatory mechanism to protect the host against the exacerbated inflammatory response elicited by the infection .
Infection with the protozoan Trypanosoma cruzi is a major cause of morbidity and mortality in Central and South America , accounting for 12 . 500 deaths per year [1] , [2] . The acute phase of infection is characterized by intense and evident blood parasitemia and may result in death . Upon infection , both the innate and adaptive immune responses lead to the control of parasite levels in the acute phase of the infection , but are insufficient for complete clearance of the T . cruzi . However , this phase is generally followed by an asymptomatic or indeterminate phase , during which there are no clinical symptoms or clear evidence of the presence of the parasite . The asymptomatic phase may last months to decades . Thus , most individuals are infected for life , with parasites persisting primarily in muscle cells , and approximately 30% of the individuals developing cardiac clinical form of Chagas disease [3] , [4] . The scarce parasitism during the chronic phase , the prolonged latent period that precedes morbidity , and the intriguing existence of different clinical forms as well as a clear involvement of the immune response , have led several authors to evaluate the involvement of auto-immune factors in the pathogenesis of the disease . Some authors have pointed out the existence of cross-reactions between host tissues and T . cruzi antigens [5]–[14] . However , the demonstration of the presence of T . cruzi or its antigens by immunohistochemical techniques or of T . cruzi DNA by polymerase chain reaction ( PCR ) in inflamed myocardial tissues suggest that parasite antigens may be necessary to trigger the inflammatory process [15]–[17] . Therefore , both processes may be involved on the development of the severe clinical forms of the disease and may act synergistically as the disease progresses . Although a significant percentage of the patients will develop the severe forms of the disease , a larger proportion remains asymptomatic throughout life . These observations have stimulated several investigators to study the processes involved on the development of severe pathology as well as on the maintenance of the asymptomatic forms of Chagas disease . In fact , many studies have demonstrated that immunoregulatory mechanisms are important for the control of infection , possibly affecting disease morbidity in chronic clinical forms [18]–[20] . We have recently observed that patients with the indeterminate ( IND ) clinical form of chronic Chagas disease have higher percentages of CD4+CD25high T cell population secreting IL-10 expressing FOXP3 [18] . These data suggest that an increase in the secretion of IL-10 by regulatory T cells during the chronic phase of the disease may be associated with protection of the host against the severe pathology induced by type 1 immune response [18] , [21] , [22] . Immunity to T . cruzi is complex , minimally involving a substantial antibody response and the activation of appropriate CD4+ and CD8+ T cell responses . The role of CD8+ and CD4+ T lymphocytes in resistance to T . cruzi and in the severity of clinical disease remains unclear . In our laboratory , we have previously demonstrated that patients with the cardiac clinical form ( CARD ) of chronic Chagas disease elicit a robust immune response against the parasite , with high levels of IFN-γ and low levels of IL-10 [21] , [22] and that this response is associated with severe cardiac pathology . Similarly , Abel et al . suggests that the ability to mount a vigorous IFN-γ-response may be associated with the development of cardiomyopathy [23] . However , Laucella et al . showed that patients with mild disease display substantial amounts of IFN-γ-producing T cells , suggesting that severe disease is prevented , rather than caused , by an immune response dominated by type-1 cytokine production [24] . Additionally , Garg and Tarleton demonstrated that enhancing type 1 immune responses through genetic immunization can substantially reduce the severity of disease in persistently infected mice ( Garg and Tarleton , 2002 ) . Heterogeneity of the response is a hallmark of antigen-specific T cells . CD4+ T cells may develop into T helper cell 1 ( TH1 ) , TH2 , or TH17 cells and likewise become antigen-specific regulatory cells [25]–[27] . According to the model proposed by Lanzavecchia and Sallusto [28] , protective memory is mediated by effector memory T cells ( TEM ) that migrate to inflamed peripheral tissues and display immediate effector function , whereas reactive memory cell development is mediated by central memory T cells ( TCM ) that home to T cell areas of secondary lymphoid organs , have little or no effector function , but readily proliferate and differentiate into effector cells in response to antigenic stimulation . In fact , human TCM are CD45RO memory cells that constitutively express CCR7 and CD62L , two receptors that are also characteristic of naive T cells , which are required for cell extravasation through high endothelial venules and migration to lymphoid organs [29]–[31] . Differently , human TEM are memory cells that have lost the constitutive expression of CCR7 , are heterogeneous for CD62L expression , and display characteristic sets of chemokine receptors and adhesion molecules that are required for homing to inflamed tissues [31] . In addition to the repertoire of cytokine secretion , effector CD4+ T cells exhibit diversity in the homing process , such as migration to peripheral nonlymphoid tissue and transit to lymph node follicles [32] . Furthermore , heterogeneity of CD8+ T cell effector gene expression has also been described [33] , although it is not clear whether this represents physiologically distinct cell fates or simply fluctuation in activation state . Memory T cells are heterogeneous , with central memory cells patrolling secondary lymphoid tissues , recapitulating the surveillance of their naive progenitor , and effector memory cells acting as sentinels at frontline barriers [34] . Although the role and function of effector and memory subsets in protection or pathology and the nature of polarizing signals required for their differentiation are becoming increasingly clear , there are still outstanding questions that need to be addressed , which are mainly related to the mechanism of T cell fate . Many of these questions deal with fundamental uncertainties that are common to many areas of blood cell differentiation , such as the extent of fate diversity , the ontogeny and lineage relationship between opposing and kindred fates , and the degree of natural and therapeutic plasticity at different stages of differentiation . In the experimental mouse model , Martin and Tarleton [35] observed that antigen-specific CD4+ and CD8+ T cells maintain a TEM phenotype during persistent T . cruzi infection . Interestingly , it was observed that TCM CD8+ memory T cells can be generated and maintained despite pathogen persistence in T . cruzi infection . Furthermore , it was demonstrated that complete pathogen clearance through benzonidazole treatment results in stable , antigen-independent and protective T cell memory , despite the potentially exhausting effects of prior long-term exposure to antigen in chronic infection [36] , [37] . On the other hand , Tzelepis and colleagues [38] showed that the differentiation and expansion of T . cruzi-specific CD8+ cytotoxic T cells ( TEM T cells ) is dependent on parasite multiplication . In humans , an increase in total effector/memory CD8+ T cells ( CD45RA−CCR7− ) was observed in CARD patients [39] . However , the role of these subtypes of memory cells is still not completely understood in human Chagas disease . In the present study , we evaluated the profile of peripheral blood subsets of CD4+ and CD8+ T cells expressing naive/memory markers ( CD45RAhigh/ROhigh ) , memory cell subtypes ( TCM−CD45ROhighCCR7+ and TEM− CD45ROhighCCR7− ) and production of cytokines by peripheral blood cells after in vitro stimulation with T . cruzi antigens to evaluate a possible relationship between the presence of these cells and the development of different clinical forms of the disease .
The initial cohort of study subjects was recruited 8 years ago at the Outpatient Referral Center for Chagas Disease of the Hospital das Clínicas , Federal University of Minas Gerais , Brazil . All study participants provided a written informed consent following the guidelines of the Ethics Committee of the Federal University of Minas Gerais . The study protocol complied with the regulations of the Brazilian National Council on Research in Humans and was approved by the Ethics Committee of the Federal University of Minas Gerais under the protocol COEP/UFMG-372/04 . Individuals with systemic arterial hypertension , diabetes mellitus , thyroid dysfunction , renal insufficiency , chronic obstructive pulmonary disease , hydroelectrolytic disorders , alcoholism , previous clinical history suggesting coronary artery obstruction and rheumatic disease , or who were unable to fulfill the study requirements for annual examinations were excluded from the study . Individuals were considered seropositive for T . cruzi infection if two or more of the standard tests performed , indirect immunofluorescence , ELISA or indirect haemagglutination , were positive . Study participants were evaluated annually for a range of clinical and immunological parameters related to Chagas disease [22] . In this study , we investigated the immune response of 23 patients , who fulfilled the protocol described above , all in the chronic phase of the infection . The patients infected with T . cruzi were classified as being in the indeterminate phase of Chagas ( IND ) or having the cardiac ( CARD ) form of the disease as previously reported [18] . Individuals in the IND group ( n = 9 ) ranged from 30 to 68 years of age . These individuals had no significant alterations in the electrocardiography , chest x-ray , echocardiogram , esophagogram and barium enema . The CARD group age ( n = 14 ) ranged from 29 to 73 years , and presented echocardiographic and/or clinical and radiological signs of heart enlargement , with a final diastolic diameter of the left ventricle greater than 55 mm . The cardiac patients that participated in this study were classified as belonging to the group CARD V , as previously reported [18] . Twelve healthy individuals , 29 to 61 years old , from a non-endemic area for Chagas disease , and who had negative serology for Chagas disease , were included in the control group ( NI ) . Epimastigote ( EPI ) antigens were prepared by using the CL strain of T . cruzi as previously described [18] , [21] , [22] . Briefly , EPI were washed three times in cold phosphate-buffered saline ( PBS ) , disrupted by repeated freezing at −70°C and thawing , homogenized at 4 to 6°C in a Potter-Elvejem ( Vir Tis-Precise Wisconsin , USA ) and centrifuged at 20 , 000×g five times at 4°C for 60 seconds , with 30 seconds intervals . The suspension was subsequently centrifuged at 40 , 000×g for 60 minutes in the cold . The clear supernatant was dialyzed for 24 hours at 4°C against PBS , filter sterilized on 0 . 22-µm pore-size membranes , assayed for protein concentration , aliquoted , and stored at −70°C until needed . Whole blood was collected in Vacutainer tubes containing EDTA ( Becton Dickinson , USA ) and 100 µL samples were mixed in tubes with 2 µL of undiluted monoclonal antibodies conjugated with fluorescein isothiocyanate ( FITC ) , phycoerythrin ( PE ) , R-phycoerythrin coupled to the cyanine dye Cy5™ ( PE Cy5 ) or allophycocyanin ( APC ) for the following cell surface markers: CD4 ( RPA-T4 ) , CD8 ( RPA-T8 ) , CD62L ( DREG56 ) , CD45RA ( HI100 ) , CD45RO ( UCHL-1 ) , CCR7 ( 3D12 ) ( all from BD Pharmingen , USA ) . After adding the antibodies , the cells were incubated in the dark for 30 minutes at room temperature . Following incubation , erythrocytes were lysed using 2 mL of FACS Lysing Solution ( BD Biosciences , USA ) and washed twice with 2 mL of phosphate-buffered saline containing 0 . 01% sodium azide . The cells were then fixed in formaldehyde ( 4% ) and permeabilized with saponin buffer ( 0 . 5% ) ( SIGMA , USA ) for 15 minutes . After incubation , the cells were fixed in 200 µL of fixative solution ( 10 g/L paraformaldehyde , 1% sodium-cacodylate , 6 . 65 g/L sodium chloride , 0 . 01% sodium azide ) . Phenotypic analyses were performed by flow cytometry with a FACScalibur flow cytometer ( BD Biosciences , USA ) . Data were collected on 1×105 lymphocytes ( gated by forward and side scatter ) and analyzed using CellQuest software ( BD Biosciences , USA ) . Whole blood was stimulated in vitro with EPI ( 25 µg/mL ) antigens in RPMI 1640 media supplemented with 1 . 6% L-glutamine , 3% antibiotic-antimycotic , 5% of AB Rh-positive heat inactivated normal human serum , for 22 hours at 37°C and 5% CO2 . Control cultures were maintained in culture media for the same period of time . During the last 4 hours of culture , Brefeldin A ( SIGMA , St . Louis , MO , USA ) ( 10 µg/mL ) was added to the cultures [18] . Cultured cells were washed twice in PBS containing 1% bovine serum albumin and stained with monoclonal antibodies specific for the different cell-surface markers , as described above . The cells were then fixed in formaldehyde ( 4% ) and permeabilized in saponin buffer ( 0 . 5% ) for 15 minutes . Finally , the cells were incubated with monoclonal antibodies reactive to IL-10 ( JES3-9D7 ) and IFN-γ ( B27 ) ( both from BD Pharmingen , USA ) . Phenotypic analyses were performed in a FACScalibur flow cytometer ( BD Biosciences , USA ) , and data collected on 1×105 lymphocytes and analyzed using the CellQuest software ( BD Biosciences , USA ) . Lymphocytes were analyzed for their intracellular cytokine expression patterns and frequencies as well as for cell surface markers using Cell Quest software . The frequency of cells was analyzed in four gates for each staining procedure: gate 1 ( R1 ) , lymphocyte gate ( Suplementary Figure S1A ) ; gate R2 , T CD4+ and CD8+ expressing CD45RAhigh and CD45ROhigh lymphocytes ( Suplementary Figure S1B ) ; gate R3 , T CD4+ and CD8+ lymphocytes ( Suplementary Figure S1C ) ; gate R4 , T CD4+ and CD8+ lymphocytes expressing CD45ROhigh/CCR7+ ( Suplementary Figure S1D ) ; gate R5 , T CD4+ and CD8+ CD45RAhigh/CD45ROhigh lymphocytes secreting IFN-γ and IL-10 ( Suplementary Figure S1E ) . Limits for the quadrant markers were always set based on negative populations and isotype controls . Analyses were performed using GraphPad Prism version 4 . 0 software ( GraphPad Software Inc , USA ) . The nonparametric tests Kruskal-Wallis test was used to compare the three clinical groups ( NI×IND×CARD ) , followed by Dunns test to compare all pairs of columns . Mann-Whitney nonparametric test was used to evaluate the significance of the cytokine production and compare the pairs of columns ( NI×IND ) ( NI×CARD ) ( IND×CARD ) , comparing all Differences were considered significant when a p value of less than 0 . 05 was obtained .
The results show that the expression of CD45RAhigh in CD4+ T cells , molecule expressed by naive T cells , was significantly lower ( p<0 . 05 ) in IND patients when compared to NI individuals ex vivo ( Figure 1A ) . We did not observe a significant difference between the infected groups when evaluated ex vivo . Furthermore , the percentage of CD8+CD45RAhigh T cells in CARD patients , after culture in the absence of EPI antigens , was significantly lower ( p<0 . 05 ) when compared to NI individuals ( Figure 1B ) . Similarly , IND and CARD patients had a significantly lower percentage of CD8+CD45RAhigh T cells , p<0 . 05 and p<0 . 01 respectively , when compared to NI individuals in in vitro cultures in the presence of EPI antigens ( Figure 1B ) . Our results also showed that the expression of CD45ROhigh by CD4+ T cells , ( memory lymphocytes ) , was not statistically different between NI , IND and CARD groups ex vivo and in culture of whole blood in the absence and presence of EPI antigens ( Figure 1C and D ) . However , the percentage of CD45ROhigh CD4+ T cells in the IND group was significantly higher ( p<0 . 05 ) than in the NI group after culture without antigen stimulation ( Figure 1C ) . When we evaluated the expression of CD45ROhigh by CD8+ T cells , the data did not reveal statistically significant differences between NI , IND and CARD groups , both ex vivo and in control cultures ( Figure 1D ) . However , the CARD group showed a tendency to have higher percentages of these cells ex vivo when compared to the other groups ( Figure 1D ) . Note worthy , after in vitro culture in the presence or not of EPI antigens , CARD patients presented a decrease in the percentage of these cells ( Figure 1D ) . The presence of intracytoplasmatic cytokines IFN-γ and IL-10 in CD4+ and CD8+ T lymphocytes expressing either CD45RAhigh or CD45ROhigh surface markers was evaluated in NI , IND and CARD groups in the absence or presence of in vitro stimulation by EPI crude extract . The data showed that CARD patients presented a significantly higher percentage ( p<0 . 01 and p<0 , 001 ) of CD8+CD45RAhigh IFN-γ-producing cells in control cultures ( absence of antigen stimulation ) when compared to IND patients and NI individuals . This percentage was also significantly higher in CARD group than in NI individuals ( p<0 . 01 ) after antigen pulsing with EPI ( Figure 2A ) . No significant differences were observed when we evaluated the percentage of CD8+CD45RAhigh cells producing IL-10 in cultures in the absence or presence of EPI antigens ( data not shown ) . Data analysis revealed a significant increase ( p<0 . 01 ) of IFN-γ+ CD4+CD45ROhigh cells in whole blood samples from chagasic patients in comparison to NI individuals , after culture in the presence of EPI antigens ( Figure 2B ) . There were no statistically significant differences among all studied groups evaluated after culture in the absence of antigenic stimulation ( Figure 2B ) . Analysis of IL-10+ CD4+CD45ROhigh T cells showed a significant higher ( p<0 . 05 ) percentage of these cells in CARD patients when compared to NI individuals after culture in the absence of antigenic stimulation ( Figure 2C ) . The results showed that there is a significantly higher percentage of CD4+CD45ROhigh T lymphocytes secreting IL-10 in CARD and IND patients ( p<0 . 001 and p<0 . 01 , respectively ) , after culture in the presence of antigens of EPI , when compared to NI group ( Figure 2C ) . Data analysis of the expression of cytokines IFN-γ and IL-10 by CD8+CD45ROhigh T cells did not show any significant difference between the groups , after in vitro culture ( data not shown ) . Triple labeling of CD4+ and CD8+ T lymphocytes from peripheral blood with the molecules CD45RA , CD45RO and CCR7 allowed us to evaluate the process of recirculation of lymphocytes ( percentage of cells migrating to/from secondary lymphoid organs ) . We classified human CD4+ and CD8+ T cells by using two markers , CD45RO and CCR7 . Three-color flow cytometry analysis demonstrated two major populations of human CD4+ and CD8+ T cells , i . e . CCR7+CD45ROhigh and CCR7−CD45ROhigh . The assessment of the percentage of CD4+CD45ROhighCCR7+ T cells demonstrated that CARD patients have significantly lower percentage ( p<0 . 05 ) of this sub-population ex vivo when compared to NI individuals ( Figure 3A ) . There were no significant differences in the analysis of these cells after culture ( Figure 3A ) . In the analysis of CD4+ cells CD45ROhighCCR7− , we did not find any differences on the percentage of cells between the groups , before or after stimulation ( Figure 3B ) . The expression of the CD45ROhighCCR7+ phenotype by CD8+ cells , IND and CARD groups showed a significantly higher percentage of these cells ( p<0 . 05 and p<0 . 001 ) in comparison to NI individuals in ex vivo ( Figure 3A ) . The evaluation of these cells after culture did not show any statistically significant differences between the studied groups . ( Figure 3A ) . When the percentage of CD8+CD45ROhighCCR7− T cells was assessed , we observed that CARD group presented a higher percentage ( p<0 . 05 ) than the IND group ex vivo . A decrease on this percentage of this cell population was demonstrated after in vitro culture ( Figure 3B ) . However , it was not statistically significant .
Most mature peripheral T cells are at rest and can be divided into naive and memory cells . This division is based on their response to antigens in a secondary response [40] . In addition to the functional activity , several markers have been identified to allow the distinction of these cell populations . Naive T cells , which have not encountered antigens , express high levels of CD45RA and L-selectin on its surface , and do not express activation markers such as CD25 , CD44 and CD69 [40] . In contrast , memory T cells , which were previously stimulated by an antigen , express high levels of CD45RO and low levels of L-selectin [41] . In this study , we demonstrate that IND and CARD patients have less naive CD4+ and CD8+ T cells , demonstrated by decreased expression of CD45RA , as well as lower levels of CD8+ memory T cells when compared to NI ( control ) individuals . Similar results have been demonstrated by other study , in which patients in the chronic phase of Chagas disease presented the same expression profile of CD45RAlow in both CD4+ and CD8+ peripheral T cells [42] . Moreover , using an experimental model , another study demonstrated that splenic CD8+ T cells from T . cruzi-infected animals show lower expression of CD45RA than cells from non-infected mice [43] . The reduction of naive and increase of memory T cells may occur during the chronic phase of the disease . This might indicate a clonal T cell exhaustion due to continuous antigenic stimulation by persistent parasites , and may be associated with increased disease severity [42] , [44]–[46] . The reduction in the expression of CD45RA is a direct result of the exchange of the CD45 isoform from A to C , possibly leading to an easier association with the TCR complex [47] , [48] . The C isoform allows the cell to be activated by lower antigen stimulation and co-stimulation [49] . The evaluation of CD4+ T cells expressing CD45ROhigh showed a higher percentage of these cells in IND group when compared to NI individuals after culture without antigenic stimulation . In fact , several studies have shown that altering the combination of CD45 isoforms dramatically affects immune function and disease severity in autoimmune models . Available data also show that this mechanism is related to an altered threshold for TCR signaling , altered cytokine production and response . This indicates that manipulating the patterns of CD45 expression or signaling pathways that it modulates might be a potential immunoregulatory strategy [47] , [48] , [50]–[52] . The mechanisms involved on the development of different clinical forms of Chagas disease are still poorly understood , suggesting that multiple factors may be involved in its establishment , such as cytokine production and profile of activation or differentiation status defined by subsets of memory [53] . The most studied cytokines in Chagas disease are IFN-γ and IL-10 . Several studies have described some of the main cell populations involved in the production of these cytokines , and their relationship with pathology or regulation of immune responses during this infection [18] , [22] , [23] , [54]–[60] . Our previous studies suggest a dual role for IFN-γ during human Chagas disease , which is observed during the different stages of the infection ( acute and chronic phases ) or in the presence or absence of treatment [22] , [54] . It has been shown that during the acute phase of experimental mouse infection , IFN-γ participates in the elimination the parasite [55] . Furthermore , the protective role of IFN-γ has also been postulated in humans , as individuals submitted to treatment show a strong cellular response with secretion of high levels of this cytokine after in vitro stimulation of PBMC [54] . On the other hand , studies on human chronic Chagas disease have also shown that IFN-γ production may be harmful to the organism , since its unregulated production in the heart tissue may promote the destruction of the myocardium due to its cytotoxic effects . [23] , [54] , [61] . However , Laucella et al . showed a linkage between increased T . cruzi-specific T cell-mediated IFN-γ responses and decreased disease severity . [24] . Indeed , there are several important differences between all these studies , which might contribute to discrepancies in interpretation . Therefore , a long-term immunological follow-up of T . cruzi-infected patients may provide valuable information on the real role and contribution of IFN-γ responses to parasite control and disease development . In this paper , we evaluate the secretion of IFN-γ and IL-10 by CD4+ and CD8+ naive ( CD45RAhigh ) and memory ( CD45ROhigh ) T cells . We observed that the subjects from the CARD group presented higher percentage of naive CD8+ T cells secreting IFN-γ . This result suggests that naive CD8+ T cells might be associated with the development of the cardiac clinical form of the disease , probably as activated cells . Bourreau et al . [62] , when evaluating the immune response of PBMCs from the NI individuals after in vitro stimulation with L . guyanensis antigens , demonstrated that CD8+CD45RA+ T cells producing IFN-γ are CD62L−CCR7− , while those producing IL-10 are CD4+CD45RA+CCR7−CD62L+ T cells , which are migrating to the inflammatory foci but are not yet activated . Our group has also shown that in Chagas disease , patients from CARD group develop a strong response against T . cruzi antigens , presenting high levels of IFN-γ and low levels of IL-10 [22] . Interestingly , here we observed that chronic patients have more CD4+CD45ROhigh T cells secreting both IFN-γ and IL-10 . Similarly , Antonelli et al . ( 2004 ) [63] demonstrated that PBMC from patients infected with L . braziliensis show high frequencies of memory CD4+ T cells producing both inflammatory ( IFN-γ ) and regulatory ( IL-10 ) cytokines . This result suggests that these cells may not be the main source of these cytokines , since they are producing both IL-10 and IFN-γ cytokines . Therefore , they may not be so relevant on the development of the inflammatory process caused by protozoan infections . Moreover , some studies have shown that memory T cells may be involved in the protection and development of Chagas disease [35]–[37] , [39] . Nonetheless , a detailed evaluation of the cell populations involved in recall responses in human Chagas disease is still necessary . Memory T cells can be divided into two sub-populations , based on their heterogeneity , effector functions and response to the antigen or cytokines . Once activated , a fraction of primed T lymphocytes persists as circulating memory cells that can lead to protection and , upon secondary challenge , result in a qualitatively different and quantitatively enhanced response [31] , [64]–[68] . The central memory ( TCM ) human T cells express CD45RO and also CCR7 and CD62L molecules , two important receptors related to the migration of T cells to peripheral lymphoid organs [30] , [69] . When compared to naive T cells , TCM have higher sensitivity to antigen stimulation , are less dependent on co-stimulation and provide efficient feedback for stimulation of dendritic and B cells . After evaluation of memory subtypes , we observed that chronic patients have more CD4+ and CD8+ TCM cells ex vivo . However , after stimulation with EPI antigens , only the IND group showed more TCM CD4+ T cells . These data suggest that when the overall T cell compartment is eventually driven to exhaustion , it exhibits a low frequency of competent parasite-specific CD8+ T cells and predisposes the subject for disease progression; this profile is presented by CARD patients with persistence of antigens or re-infection by T . cruzi . Recently , using the mouse model of T . cruzi infection , an increase in CD8+ TCM cells was observed during long-term persistent infection [36] . Of note , these CD8+ TCM cells were capable of antigen-independent survival after being transferred into naive mice , and were maintained despite the presence of persistent antigen stimulation . Similarly , Bustamante et al . documented the development of stable , antigen-independent CD8+ T cell memory after benzonidazole-induced cure of chronic infected mice [37] . It has been shown in mice that antigen-specific T cells maintain an effector memory phenotype ( TEM CD8+ T cells ) during persistent T . cruzi infection [35] , [38] , suggesting that these cells may play an important role in the pathogenesis of the Chagas disease . However , the process of cell development and differentiation from naive into TCM or TEM is still not clearly understood . Studies with patients infected with HIV ( human immunodeficiency virus ) or LCMV ( lymphocytic choriomeningitis virus ) suggested that T cells go through the process of differentiation from naive→TCM→TEM [70] , [71] . On the other hand , several authors suggest that the TCM and TEM cells are actually independent subpopulations which develop according to the biological environment , ( eg . : presence of different cytokines or the anatomical region of activation ) are independently maintained [71]–[73] . Finally , other studies on the infection LCMV TCM show that the cells proliferate and/or convert into TEM after re-exposure to the antigen , suggesting an alternative model of CD8+ T cell differentiation: naive→TEM→TCM [74] . However , our data suggest that the type of differentiation is naive→TCM→TEM [70] , [75] . In fact , IND patients seem to have some regulation that prevents TCM cells from becoming TEM , but further studies are still needed to elucidate this question . Previous studies have attempted to clarify the role of different subtypes of memory cells using the experimental mouse model and the CD45RA as a cell marker [39] , [76] . In the current study , we evaluated the memory profile using CD45ROhighCCR7+ phenotype as a marker for TCM and CD45ROhighCCR7− for TEM T cells . In conclusion , our results showed that CARD patients have more naive CD8+ T cells secreting IFN-γ and TEM CD8+ T cells than IND and NI groups . Based on a correlation between the frequency of IFN-γ producing CD8+ T cells in the T cell memory compartment and the chronic chagasic myocarditis , we propose that memory T cells might be involved in the induction of the development of the severe clinical forms of the Chagas disease by mechanisms modulated by IFN-γ . Conversely , some authors have demonstrated that high levels of IFN-γ are inversely correlated with disease severity [24] , [39] . In this way , the role of IFN-γ in human Chagas disease progression is not clear and should be further elucidated by large follow-up studies . Moreover , we demonstrated that individuals from IND group presented higher levels of TCM CD4+ T cells , which could induce immunoregulatory mechanisms to protect the host against the exacerbated inflammatory response elicited by T . cruzi infection . Studies of subtypes of immunological memory are still important for understanding the role of these cells in Chagas disease ( immmunoregulatory or pathogenic ) . Therefore , additional longitudinal studies of changes in CD8+ T cell sub-populations in chronically infected subjects may reveal specific markers for progression to severe disease . | Chagas disease is a parasitic infection caused by protozoan Trypanosoma cruzi that affects approximately 11 million people in Latin America . The involvement of the host's immune response on the development of severe forms of Chagas disease has not been fully elucidated . Studies on the immune response against T . cruzi infection show that the immunoregulatory mechanisms are necessary to prevent the deleterious effect of excessive immune response stimulation and consequently the fatal outcome of the disease . A recall response against parasite antigens observed in in vitro peripheral blood cell culture clearly demonstrates that memory response is generated during infection . Memory T cells are heterogeneous and differ in both the ability to migrate and exert their effector function . This heterogeneity is reflected in the definition of central ( TCM ) and effector memory ( TEM ) T cells . Our results suggest that a balance between regulatory and effectors T cells may be important for the progression and development of the disease . Furthermore , the high percentage of central memory CD4+ T cells in indeterminate patients after stimulation suggests that these cells may modulate host's inflammatory response by controlling cell migration to tissues and their effector role during chronic phase of the disease . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"immunology/immunity",
"to",
"infections"
] | 2009 | Profile of Central and Effector Memory T Cells in the Progression of Chronic Human Chagas Disease |
Two theoretical models dominate current understanding of actin-based propulsion: microscopic polymerization ratchet model predicts that growing and writhing actin filaments generate forces and movements , while macroscopic elastic propulsion model suggests that deformation and stress of growing actin gel are responsible for the propulsion . We examine both experimentally and computationally the 2D movement of ellipsoidal beads propelled by actin tails and show that neither of the two models can explain the observed bistability of the orientation of the beads . To explain the data , we develop a 2D hybrid mesoscopic model by reconciling these two models such that individual actin filaments undergoing nucleation , elongation , attachment , detachment and capping are embedded into the boundary of a node-spring viscoelastic network representing the macroscopic actin gel . Stochastic simulations of this ‘in silico’ actin network show that the combined effects of the macroscopic elastic deformation and microscopic ratchets can explain the observed bistable orientation of the actin-propelled ellipsoidal beads . To test the theory further , we analyze observed distribution of the curvatures of the trajectories and show that the hybrid model's predictions fit the data . Finally , we demonstrate that the model can explain both concave-up and concave-down force-velocity relations for growing actin networks depending on the characteristic time scale and network recoil . To summarize , we propose that both microscopic polymerization ratchets and macroscopic stresses of the deformable actin network are responsible for the force and movement generation .
Cell migration is a fundamental phenomenon underlying wound healing and morphogenesis [1] . The first step of migration is protrusion – actin-based extension of the cell's leading edge [2] . Lamellipodial motility [3] and intracellular motility of the bacterium Listeria monocytogenes [4] are two prominent model systems that in the past decades have added considerably to our understanding of the protrusion based on growth of actin networks . These in vivo systems are complemented by in vitro assays using plastic beads [5] and lipid vesicles [6] that , when coated with actin accessory proteins , move much the same way as the Listeria pathogen . Here we examine computationally the mechanics of growing actin networks . This problem has a long history starting from applying thermodynamics to understand the origin of a single filament's polymerization force [7] . The notion of polymerization ratchet led to the derivation of an exponential force-velocity relation ( Figure S1 in Text S1 ) for a rigid filament growing against a diffusing obstacle [8] . Then , elastic polymerization ratchet model [9] was proposed for flexible actin filaments . This model evolved into tethered ratchet theory , in which a dynamic balance between surface-pushing growing filaments and motion-resisting attached filaments ( Figure 1A ) governs the protrusion [10] . These early theories considered independent single filaments . However , actin filaments do not grow individually , but evolve interdependently as a network by branching sideways from each other [11] . Mathematical treatments and computer simulations of branching and nucleation [12] , [13] of filaments growing against an opposing force , which treated the dendritic actin network as a mechanically rigid body , predicted various force-velocity relations . Those ranged from concave-down ( velocity of protrusion being insensitive to the load up to a threshold and plunging to a stall at a critical opposing force ) to concave-up ( more or less exponential decrease of the velocity with the growing load ) relations ( see Figure S1 in Text S1 ) . These theoretical efforts culminated in detailed agent-based three-dimensional ( 3D ) models of growing networks of rigid filaments propelling Listeria pathogen [14] , [15] . In parallel to these microscopic theories , macroscopic elastic propulsion model [16] , [17] suggested that the curved surface of the pathogen is not merely pushed , but squeezed forward by an elastic stress . This stress is developed from the stretching of the outer layer of actin gel by the growth of the gel near the inner surface ( Figure 1B ) . This model treated the actin network as an isotropic elastic continuum and did not explicitly consider the microscopic mechanism of force generation at the surface . As a result , a concave-up force-velocity relation for the actin-propelled spherical bead was derived [18] , predicting an initial rapid decay with opposing force followed by a region of slower decay of velocity . This prediction was confirmed by using a cantilever setup for beads coated with the actin polymerization activator N-WASP and moving in a pure-protein medium [18] . On the other hand , when the force-velocity relation of an actin network growing against a flat surface was measured using the cantilever method , it was found that the growth velocity was constant at small forces but dropped rapidly at higher forces [19] as predicted by some microscopic ratchet theories . Note that the widely used terminology could be confusing as the elastic propulsion theory is sometimes called mesoscopic rather than macroscopic . Both terms are justified: the macroscopic mechanics is described using continuum theory , but an actin layer of a few microns thin is certainly a mesoscopic system . The model we present is mesoscopic in the sense that it spans from the microscopic level of individual filaments to the macroscopic level of continuous description of an actin gel . The model is also hybrid because it takes into account both local discrete forces and global network stress . We will mostly use the term “hybrid” throughout the paper . The first simple attempt to use hybrid modeling of the lamellipodial edge was recently made in [20] , where the actin network was divided into a semiflexible region near the membrane and a gel-like region at the back . Near the membrane , semiflexible filaments are assumed to produce entropic forces against both the membrane and the gel . In the back , the viscous gel deforms in response to stresses both from frontal filaments and internal contractions , causing retrograde flow . Because the semiflexible region is assumed to be supported by the gel region , the moving speed of the membrane is determined by the coupling between the two regions . This model was able to reproduce both concave-up and concave-down shapes of the force-velocity relation . Since this model considered only a one-dimensional strip of actin gel , it did not address the effects of surface geometry . Besides the force-velocity relation , the non-zero curvatures of the trajectories of motile objects [21] is another important observable . A pioneering microscopic ratchet-based model , which investigates how randomly distributed actin filaments propel a cigar-shaped pathogen , predicted that the resultant bacterial trajectories have curvature values following a Gaussian distribution with zero mean [22] . This conclusion was challenged by a number of studies . One of them showed helical movements that were explained as a result of a non-vanishing torque that arises from a persistent actin-induced off-center force [23] . Another study did not result in helical paths of beads , but rather showed a highly varying curvature of trajectories which has a Gaussian distribution , albeit with a sharp peak at zero curvature [24] . In contrast , a third study indicated that the distribution of the curvatures of trajectories deviated significantly from Gaussian , which was explained by a cooperative breaking of filaments tethered to the bead [25] . All theories used to explain these experiments were microscopic; elastic propulsion model was never applied to these phenomena . Below , we describe observations of ellipsoidal , rather than spherical , beads that cannot be explained by either microscopic or macroscopic model . This , as well as the complex force-velocity relation and curvature distribution described above , hints that perhaps a hybrid model with individual actin filaments pushing from the surface of a macroscopic deformable actin gel can explain the experiments better . Recent experiments and theory [26] , [27] demonstrated that disassembly and breaking of the actin gel are as important as the elastic deformations in generating propulsion . Therefore , we developed a model of a node-spring viscoelastic network representing the actin gel with individual pushing and pulling filaments embedded into the network boundary . Simulations of this in silico hybrid network showed that the combined effects of the macroscopic viscoelastic deformation and microscopic ratchets can explain both concave-up and concave-down force-velocity relations for growing actin networks , bistable orientation of the actin-propelled ellipsoidal beads , and peculiar curvature distributions for the actin-propelled trajectories of the beads .
We developed a two-dimensional ( 2D ) simplification of a 3D hybrid model ( Figure 1C ) , which incorporates both arrays of dynamic actin filaments at the surface-tail interface and the bulk deformable actin gel behind the interface . Filament arrays are embedded into the boundary of the deformable actin gel , which is coarse-grained into a network of nodes interconnected by elastic springs . Individual filament arrays at the surface-tail interface switch between pushing the obstacle surface and attaching to it . The existing filaments are constantly becoming a part of the network and dynamically expanding the actin gel , while nascent filament arrays are created around the surface via a mixture of nucleation and branching processes . The actin network undergoes disassembly , which is treated by removing the nodes and springs at a constant rate , as well as by rupturing crosslinks at a critical stretching force . The deformations of the network as well as the elastic filament forces cause both translational and rotational motion of the bead . The model reproduces the steady motion of beads propelled by treadmilling actin tails behind the beads ( Video S1 ) . Further details about the model assumptions , equations , numerical simulations and model parameters are described in the Materials and Methods and Text S1 . Recently , with our experimental collaborators , we reported observations of the ellipsoidal beads that were uniformly coated with an actin assembly-inducing protein ( ActA ) [28] and moved in the plane between two parallel coverslips ( see the Materials and Methods below ) . Surprisingly , roughly half of the time the beads moved along their long axes , and another half – along their short axes ( Figure 2 , A and B ) , with infrequent switches between these orientations . To see whether the two existing models of actin propulsion can explain this result , we simulated the motion of actin-propelled ellipsoidal beads as described in the Materials and Methods . Elastic theory predicts that squeezing of an ellipsoidal bead introduces a torque orienting the bead with its long axis parallel to the actin tail ( see Figure S2 and Figure S6 in Text S1 ) . In agreement with this prediction , when we decreased the autocatalytic branching of actin and attachment forces , so that the actin gel exerted almost uniform normal stress on the bead surface , the model resulted in a propulsion along the bead's long axis ( Video S2 ) . On the other hand , when we simulated a network of rigid branching filaments pushing the bead , the propulsion was always along the short axis , so the bead moved sideways ( Video S3 ) . This change in the preferred orientation is caused by a subtle bias in how the actin network spreads along the bead surface: if the bead's orientation is skewed relative to the actin tail's axis , filament branching are more likely to happen near the tail-facing flatter surface where there is a higher number of existing filaments . As a result , more filaments push the bead sideways from the actin tail , shifting the filament-contacting region from the curved surface to the flatter one . Eventually , most filaments branch against the flatter part of the surface , orienting the bead with its long axis normal to the tail axis ( see Figure S7 and detailed calculations in Text S1 ) . Thus , the elastic propulsion model predicts that beads only move along their long axes , while microscopic ratchet model predicts that beads only move along their short axes , and neither model can explain the observation . In contrast , the full hybrid model predicts that the bead can move in both orientations due to the combination of the elastic squeezing and the geometric spreading of actin and switch infrequently between them ( Video S1 , Figure 2 , C and D , Figures S5 and Figure S8 in Text S1 ) , in agreement with the observation ( Figure 2 , A and B ) . For more insight into this phenomenon and to generate predictions for experiment , we investigated numerically how the fraction of beads moving with a certain orientation depends on the geometric , mechanical and kinetic parameters . To further test the hybrid model , we simulated the motion of actin-propelled spherical beads ( Figure 3 , A and C ) . We recorded the 2D ‘in silico’ trajectories of the beads and compared them to the experimental observations ( see the Materials and Methods ) . We examined two possible mechanisms for the nucleation of new filaments: autocatalytic branching and spontaneous nucleation . We found that each mechanism alone does not produce the observed motion of the bead ( see Video S4 and Video S5 ) . Only a combination of the two mechanisms leads to realistic motion of the bead ( see Video S6 and details in Text S1 ) . Note that the trajectories are easy to visualize by looking at the actin tails that represent the most recent parts of the trajectories , see Figure 3 , B and D ) . Our typical simulation results ( Figure 3 , A and E , Video S7 ) illustrate that in general the trajectories are mildly curved , as observed in some cases experimentally ( Figure 3B ) . However , in other cases the experimental observations ( Figure 3D ) show that once in a while the beads stop , get surrounded by a dense actin ‘cloud’ , and then break through the cloud and resume movement in a new direction . Indeed , the model predicts that when the detachment rate of actin filaments becomes low and a greater fraction of filaments is attached to the bead surface , beads start to have pulsatory motion due to temporary entrapment by the actin gel ( Figure 3C and Video S8 ) , which occurs frequently in this regime . The explanation is that when filaments detach rapidly and thus do not generate great pulling forces , beads move quickly and can hardly be trapped , but at low detachment rate , beads slow down significantly by the strong pulling forces , which increases their chances to be trapped into the actin gel . Both our simulations and observations from our collaborators show that beads often make sharp turns during their escapement from the surrounding actin gel ( Figure 3 , C and D ) , causing the switching between the low- and high-curvature trajectories . As a result , the trajectories show spatially separated segments of low and high curvatures ( Figure 3F ) . To obtain the distribution of the curvatures of the trajectories , we smoothed the simulated bead's trajectory to remove the high frequency noises and calculated ( see Text S1 for details ) that the curvature distribution is close to Gaussian ( Figure 4A ) for fast-moving beads in the wide range of parameters . This indicates that the turning of the fast-moving bead is likely to be driven by random events in the protruding actin network . When the detachment rate is low , we find that the curvature distribution becomes sharply peaked at zero ( Figure 4B ) , in agreement with both our observation ( Figure 4B ) and previous results [24] . Since the low- and high-curvature trajectories are typically separated in this regime , this sharp peak near zero is due to bead moving in a rapid-and-smooth fashion , while the slowly decreasing distribution at higher curvatures is caused by bead moving in a slow-and-jagged fashion . Furthermore , we find that the distribution is close to a Gaussian at higher curvature , indicating that the highly curved segments of trajectories are also likely to be caused by the random fluctuations in the actin network . We found that the predicted characteristic value of the root-mean-square curvature , ( Figure 4C ) , is of the same order of magnitude as our observations ( Figure S17 in Text S1 ) and available measurements [4] , [24] , [25] . We investigated how the filament attachments affect the value of ( Figure 4C ) and found that is insensitive to for . However , the curvature increases rapidly with for , consistent with the idea that excessive attached filaments cause frequent trapping of the bead leading to highly curved trajectories . We also studied how the bead radius , , affects ( Figure 4D ) and found that decreases as the bead size increases . This result is in agreement with the experimental observations reported in [4] , [25] . Interestingly , this results is also consistent with our experimental observation on the orientation-dependent turning of the trajectories of ellipsoidal beads ( Figure S17 in Text S1 ) : ellipsoidal beads moving along their long-axes are less likely to keep their current direction of motion comparing to those moving along their short-axes . A possible interpretation is that the former are mostly pushed at their sharp ends where the radius of curvature is low . Similar to a spherical bead with small , this will lead to a high in the trajectory and thus will be less likely for the bead to keep the current direction of motion . Together , the above results can be explained as follows: larger beads are propelled by a greater number of filaments , so relative fluctuations in the actin network go down and thus the beads fluctuate less in their motion . These findings suggest that the fluctuation in the number of actin filaments is likely the factor determining the curvature , so we developed a simple model to understand and test such mechanism . Two possible mechanisms may contribute to the turning of beads' trajectory: turning induced by elastic and ratchet torque , and turning induced by actin tail-reorientation ( see Text S1 ) . Because of the symmetry of the spherical bead , the torque-induced rotation found in the ellipsoidal beads is negligible . Our simulations also confirm that a micron-sized spherical bead rarely rotates about its center during its motion . Therefore , the re-orientation of the tail along the bead surface is likely to be the main cause of the trajectory turning . Thus , we consider a simplistic model in which a bead of radius is propelled by randomly distributed filaments at its rear , so the filament number difference between the left and right sides of the bead is on the order of . In other words , out of filaments tend to push the bead off the current direction by an angle while the rest tend to push along the current direction of motion . The change in the direction of motion is expected to be . The typical time over which the directional bias persists is the turn-over time of the actin network , which we estimate in Text S1 . Then , the typical angular velocity of the turning is , and the root-mean-square value of the curvature is One thus expects a linear relation between and with a slope of . To test whether this simple conclusion is correct , we used simulations of the hybrid model to obtain the values of , , and . We plotted the simulation results for as a function of for various values of attachment , detachment , capping and nucleation rates , as well as of actin gel elastic constant , together with the predicted linear relation , and found very good agreement except for low values of the detachment rate ( see Figure 4E , Figure S10 and Figure S11 in Text S1 ) . The higher-than-expected values of obtained from the simulations with low detachment rates are caused by the entrapment of beads into the actin gel , as mentioned above . Thus , macroscopic elastic effects influence the trajectory only in the limiting case of too many attached filaments . Otherwise , stochastic microscopic filament-ratchets are responsible for the curvature of trajectories . Note that in contrast to our results , a non-Gaussian distribution of the curvatures of trajectories of the beads was observed in [25] . According to the model in [25] , the torque balance alone determines the turning of the bead , while in our model both torque and redistribution of actin around the bead determine the trajectory . This difference suggests that the redistribution of actin probably does not play an important role in the experiments in [25] . One possibility is that the actin tail always interacts with a fixed side of the bead in these experiments , which can result from an asymmetric coating of the bead surface by the actin-nucleation promoting factors . Also note that the autocorrelation function of the simulated curvature of trajectories always decays rapidly at a sub-micron distance ( see Figure S12 and details in Text S1 ) . This result differs from the observed long-range correlation of about [24] , which is possibly caused by additional long-ranged bias in the actin network near the bead-tail interface . We simulated growth of an actin pedestal against flat elastic cantilever and force-clamped spherical bead , as in experiments [18] , [19] , respectively ( Video S9 and Video S10 ) . The hybrid model in these cases was used as described above , with the following differences: 1 ) We first generated undeformed node-spring pedestal underneath the surface to be pushed . 2 ) All actin network nodes were free to be positioned by the force balances ( the nodes in the network did not become immobile when they were more than a few microns away from the surface ) except at the very bottom . The layer of the nodes at the very bottom was immobilized . 3 ) The motion of the cantilever or bead was determined by the balance between the pushing/pulling forces from the filaments touching the surface and either a ) the elastic restoring force from the cantilever proportional to cantilever's deflection , or b ) clumped force from the bead . The speed of the cantilever or bead , , was then obtained by dividing the displacement increment of the surface by the time interval . Calibration of the model in these numerical experiments is described in Text S1 . Simulation snapshots are shown in Figure 5 , A and B and Figure S16 in Text S1 . The simulated force-velocity relation predicted by the hybrid model for the flat cantilever is compared to the experimental data [19] in Figure 5C . We scale the cantilever force by , which is the force at half of the maximum cantilever speed and scale to best match the rest of the data . The prediction agrees very well with the observed concave-down force-velocity relation . To quantitatively understand this result , we develop an analytical 1D theory in Text S1 and find that continuing reduction of the network stiffness due to the network disassembly during a long time of the experiment plays an important role in the shape of the force-velocity relation . A network undergoing significant disassembly in the aged gel sections recoils under a high load , reducing both net protrusion rate of the actin network pushing the cantilever and the maximum force that the network can sustain . These factors cause the rapid downturn in the force-velocity relation . Our 1D analytical result ( can be approximated as in relevant parameter range ) is shown in Figure 5C and is in very good agreement with both experimental data and simulation of the 2D hybrid model . We then used the hybrid model to simulate the force-velocity relation for the force-clamped bead . In this case , the force-velocity relation is concave-up , in good agreement with the observations [18] ( Figure 5D , Figure S15 in Text S1 ) . Qualitative explanation for this shape is that the velocities in this experiment were measured on a minute time scale before the network significantly disassembles ( over a few minutes ) . Therefore , the network's recoil is negligible in this case and the force-velocity relation is similar to that of individual filaments . From our 1D calculation for under a constant load ( see Text S1 ) , we find , where is proportional to the disassembly rate constant of the network and is the age of the network when is measured in our simulations , and is the average velocity of individual filaments . This analytical result is also shown in Figure 5D , in very good agreement with the simulation results of the hybrid model . To investigate the effect of the filament attachments to the surface on the force-velocity relations , we varied the value of the attachment rate to change the ratio of the number of attached to the number of pushing filaments , . The simulated force-velocity relations for different ratios are shown in Figure S13 in Text S1 . For both cantilever and force-clamped experiment , we find that increasing the fraction of attached filaments decreases both velocity and stall force without changing the qualitative shape of the force-velocity curve , consistent with the idea that attached filaments counteract the pushing filaments . Finally , to confirm that it is the actin dynamics rather than the shape of the surface that determines the force-velocity relation , we swapped the shapes of the flat cantilever and round bead used in the two experiments . We considered two cases: a slow-growing actin network against a curved surface of a cantilever , and a fast-growing actin network against a flat force-clamped object . The simulation results shown in Figure S13 and Figure S15 in Text S1 illustrate that the force-velocity relations in both experiments remain qualitatively the same ( concave-down and concave-up , respectively ) . Therefore , the shape of the surface does not appear to affect the overall shape of the force-velocity relation .
Complexity of the relation between geometry of the curved surface , molecular pathways of actin polymerization against this surface and resulting force [29] indicates that the actin-based force-generation is a multi-scale phenomenon , understanding of which requires a combination of macroscopic and microscopic mechanisms . We developed such hybrid model of the actin network growing and pushing against rigid surfaces , in which actin filaments interacting directly with the surface are treated as tethered-ratchet filaments , while other filaments are considered implicitly as parts of viscoelastic node-spring network . The elastic propulsion theory predicts that squeezing of the ellipsoidal beads orients them so that motility along the long axes ensues , while geometric effect of spreading of branching actin filaments results in beads moving along their short axes . Separately , the existing theories cannot explain the observed bi-orientation of the beads . Our hybrid model posits that the combination of the elastic squeezing and geometric spreading leads to bi-orientation and reversible switching between two orientations , in agreement with the observations . To test the hybrid theory in the future , we propose to vary the bead geometry and concentrations of actin accessory proteins , thus modulating the network stiffness and interactions with the surface . Our model makes specific , nontrivial and testable predictions ( see Figure 2 , E–G ) for such experiments . The hybrid model reproduces the observed order of magnitude of curvatures of the trajectories in 2D and suggests that switching between the low- and high-curvature trajectories is caused by the temporary entrapment of the beads in the actin gel . The model predicts a Gaussian distribution of the curvatures for fast-moving beads due to random fluctuations of filament numbers and redistribution of actin around the bead's surface . In agreement with observations , our simulations show an additional sharp peak at zero curvature in the curvature distribution for slowly-moving beads . Importantly , the model suggests that elastic effects have little impact on the distribution of trajectory curvatures for fast-moving beads , while for beads that tend to be trapped in the actin cloud due to frequent filament attachments , the elastic effects are responsible for deviations from Gaussian distributions . The hybrid model posits that the qualitative difference between two force-velocity measurements [18] , [19] stems from the characteristic time difference: when the measurement is made over a long time interval [19] , the viscoelastic recoil of the older , aging part of the network near the base of actin pedestal cancels protrusion and causes the concave-down force-velocity relation . On the other hand , when the force is clamped and the experiment is performed over shorter times [18] , the concave-up force-velocity relation is predicted . A possible way to test our model is to use fluorescent speckle microscopy to measure the kymograph of material points of the actin network that move with the recoiling network away from the surface being pushed . We predict the resulting curves for two considered experiments in Figure S14 in Text S1 . Note , that there are alternative explanations for the result [19] . For example , theory in [30] based on a representation of the actin network as a viscoelastic solid could predict a different kymograph . Finally , the model proposes that the shape of the surface does not qualitatively affect the shape of the force-velocity relation . In the present form , our model has a number of limitations . The main one is that due to computational time limitations , we simulated the model in 2D as a simplification of a 3D system . So , rigorously speaking , all our results are applicable to cylindrical , rather than spherical objects . In Ref . [28] , we already attempted the 3D modeling , albeit of an oversimplified model . Preliminary indications from that attempt are that most of the 2D model predictions survive in 3D . However , there are effects of higher dimension: 3D viscoelastic theory and experiment [27] suggest that ellipsoidal beads break through the actin cloud sideways , while [28] reports the observed lengthwise symmetry breaking of the ellipsoidal beads . This problem remains open , and thus more 3D modeling is necessary . In addition , helical and more complex trajectories of actin-propelled beads that have been observed in 3D environments [23] , [24] cannot be captured by our 2D model . Furthermore , our model is coarse-grained and neglects important fine-scale processes such as hydrolysis of ATP bound to polymerized actin [31]–[33] , exact actin branching angles [34] , indirect synergy between capping and branching [35] , molecular details of the nano-scale protrusion [36] and dependence of the branching rate on filament bending [37] . Future incorporation of these details into the model will clarify molecular nature of the mixture of nucleation-based and autocatalytic actin growth posited in the model . Due to these limitations , our model does not capture some observed effects . Notably , the simulations do not reproduce observed hysteresis in the growth velocity of actin networks under force [19] , which likely depends on complex dynamic features of the network [34] , [38] that are not incorporated into our model . Similarly , not reproducing deviations from the Gaussian distribution of the curvatures of trajectories [25] likely means that some inhomogeneities in the distribution of actin nucleation promoting-factors not included into the model play an important role . These inhomogeneities and 3D effects also have to be built into the model to reproduce helical trajectories reported in [21] , [23] . Another open question is relation of our model to other theories of the actin-based propulsion . Those include microscopic models of propulsion by tethered actin filaments [39] , [40] that can in principle be used as boundary conditions for the viscoelastic actin gels and tested by simulations similar to those done here . Two mesoscopic models , very different from ours , were proposed recently . One of them considers excluded volume effects [41] , another is a liquid of dendritic clusters model [42]; both of them successfully reproduce the concave-down force-velocity curve . It is likely that subtle physical effects on which these models are based complement elastic deformations and individual filament ratchet forces of our model . In the future , after including interactions of the filaments with cell membrane [43]–[46] , contractile myosin effects [47] and more adequate actin rheology [48] , our model can be applied to the general problem of cell protrusion .
Motility experiments on ellipsoidal beads were carried out in the lab of J . Theriot as previously described [28] . Briefly , 1- carboxylated polystyrene microspheres ( Polysciences , Warrington , PA ) were placed in a viscoelastic matrix ( 6% polyvinyl alcohol ) , heated to , and stretched uniaxially . The film containing the beads was cooled and dissolved using an isopropanol/water mixture to recover the beads before functionalizing their surfaces with carboxylate . Electron microscopy showed that the beads had average dimensions of , with an average aspect ratio of 2 . 2 . His-tagged ActA was purified and adsorbed on the surface of beads at saturating amounts . ActA-coated beads were then added to Xenopus laevis egg cytoplasmic extract , which was diluted to 40% of the original protein concentration . The slide chamber depth was restricted using 2- silica spherical beads . Note , that the ActA-coated motile beads were contained between two parallel coverslips and restricted from moving perpendicularly to the coverslips , and thus the trajectories of the beads were two-dimensional . All time-lapse sequences taken during the steady-state bead motility were acquired between 2 and 4 h after preparing the slide . Phase-contrast and fluorescence images were acquired as described in [28] . Spherical beads were prepared in the lab of J . Theriot as previously described [5] , which is similar to that for ellipsoidal beads except for the stretching treatment . Bead trajectories were recorded at 10 s intervals . For both experiments , positions and orientations of beads were computed from phase-contrast images and assembled into tracks as described in [28] . Smoothing of the instantaneous angular velocity values of the beads was generated using a weighted average of five nearest neighbors and a cubic equation as described in [28] . The angular velocity fit-in was generated using a seventh-order polynomial function . The curvature was obtained by dividing the resulting angular velocity by the instantaneous linear speed of the bead . In the hybrid model ( Figure 1C ) , arrays of actin filaments interacting directly with the surface of the bead are treated as effective individual filaments , while other ( not in touch with the surface ) filaments are not modeled explicitly but rather treated as the network of elastic springs interconnected by nodes . The model is formulated and all simulations are done in 2D , which is a simplification of a 3D system . We assume that new filaments are created around the surface via a mixture of spontaneous nucleation , which has a spatially uniform rate along the bead surface , and autocatalytic branching processes , which has a rate proportional to the local density of existing filaments ( not necessarily uniform in space ) . Separately , either of these processes produces a defective actin tail ( see Figure S4 and discussion in Text S1 ) . We also assume that newly created filaments immediately anchor to the network at their pointed ends which become new nodes of the network . In the simulations , this step is achieved by connecting each pointed end with undeformed springs to up to 4 neighboring nodes in the network that are within from the pointed end ( see Figure S3 in Text S1 ) . Thus , creation of new filaments dynamically expands the actin network . We treat filaments as elastic springs that are created in an attached and undeformed state . When stretched , attached filaments produce resisting forces that are proportional to their deformations . Attached filaments undergo detachment with a rate that increases exponentially with the load force . After detachment , filaments become free and are able to elongate and produce pushing forces against the obstacle . Free filaments are treated as linear elastic springs with the rest length growing with the polymerization rate . This rate is a function of the load on the barbed end of the filament; the function is given by the individual filament force-velocity relation that follows from the Brownian ratchet theory . The pushing force that a free filament exerts on the surface is computed as follows: at each time step , a virtual ‘penetration’ distance of the barbed end of the rest-length spring , corresponding to the filament , into the bead is computed . The filament is assumed to be deformed by this penetration distance , and respective elastic force is the pushing force . Free filaments can re-attach to the surface and get capped at constant rates . Once capped , the filament is removed from the simulation , since in reality it will stop growing and cannot attach to the surface to exert pulling forces . However , the node corresponding to the pointed end of the filament remains , so this filament effectively becomes a part of the deformable network . We do not track the orientation of individual pushing filaments , but treat them as coarse-grained clusters of actual filaments that always push perpendicularly to the obstacle surface ( see Figure 1D ) . As filaments exert forces on the obstacle , they also apply opposite forces to the elastic network that they are anchored to , causing network deformations ( see Figure 1D ) . Similarly , the stress in the deformed network is transferred to the bead surface through the interacting filaments . The deformation of the network is represented by the motion of nodes and springs in the network , which is obtained by moving all the nodes toward their force-equilibrium positions at each time step . For actin-propelled beads , we assume that the nodes in the network become immobile when they are more than a few microns away from the bead surface , representing the adhesion of the actin tail to the substrate . The bead moves and rotates to satisfy the force and torque balances from the filaments . For the force-velocity measurements , we fix the network at the bottom and allow all the rest nodes to move to reach force balance . The network undergoes disassembly , which is treated by removing the nodes and their connected springs from the network randomly with a rate proportional to the number of existing nodes . We have also included the effect of rupture of crosslinks by introducing a critical stretching force , above which the links break and get removed from the network . During the steady motion of beads , the creation and extinction rates of actin networks balance , causing a treadmilling actin tail behind the bead ( Video S1 ) . Effective viscoelastic behavior of the actin network emerges from the disassembly and breaking of the network . Further details about the model equations and parameters are described in Text S1 . | There are two major ideas about how actin networks generate force against an obstacle: one is that the force comes directly from the elongation and bending of individual actin filaments against the surface of the obstacle; the other is that a growing actin gel can build up stress around the obstacle to squeeze it forward . Neither of the two models can explain why actin-propelled ellipsoidal beads move with equal bias toward long- and short-axes . We propose a hybrid model by combining those two ideas so that individual actin filaments are embedded into the boundary of a deformable actin gel . Simulations of this model show that the combined effects of pushing from individual filaments and squeezing from the actin network explain the observed bi-orientation of ellipsoidal beads as well as the curvature of trajectories of spherical beads and the force-velocity relation of actin networks . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"motility",
"actin",
"filaments",
"biophysic",
"al",
"simulations",
"biology",
"cell",
"mechanics",
"computational",
"biology",
"biophysics",
"biomechanics"
] | 2012 | Mesoscopic Model of Actin-Based Propulsion |
Bacterial spores play an important role in disease initiation , transmission and persistence . In some species , the exosporium forms the outermost structure of the spore and provides the first point of contact between the spore and the environment . The exosporium may also be involved in spore adherence , protection and germination . Clostridium sordellii is a highly lethal , spore forming pathogen that causes soft-tissue infections , enteritis and toxic-shock syndrome . Despite the importance of C . sordellii spores in disease , spore proteins from this bacterium have not been defined or interrogated functionally . In this study , we identified the C . sordellii outer spore proteome and two of the identified proteins , CsA and CsB , were characterised using a genetic and phenotypic approach . Both proteins were essential for the correct formation and positioning of the C . sordellii spore coat and exosporium . The absence of CsA reduced sporulation levels and increased spore sensitivity to heat , sodium hydroxide and hydrochloric acid . By comparison , CsB was required for normal levels of spore adherence to cervical , but not vaginal , cells , with csB mutant spores having increased adherence properties . The establishment of a mouse infection model of the gastrointestinal tract for C . sordellii allowed the role of CsA and CsB to be interrogated in an infected host . Following the oral administration of spores to mice , the wild-type strain efficiently colonized the gastrointestinal tract , with the peak of bacterial numbers occurring at one day post-infection . Colonization was reduced by two logs at four days post-infection . By comparison , mice infected with the csB mutant did not show a reduction in bacterial numbers . We conclude that C . sordellii outer spore proteins are important for the structural and functional integrity of spores . Furthermore , outer spore proteins are required for wild-type levels of colonization during infection , possibly as a result of the role that the proteins play in spore structure and morphology .
Bacillus and Clostridium bacterial species produce spores as a survival mechanism , in response to adverse conditions such as nutrient starvation [1 , 2] . Spores are resistant in environmentally unfavourable conditions and allow bacteria to persist in conditions that do not allow the survival of metabolically active vegetative cells . In both Bacillus and Clostridium species , spores are critical for disease as they are often responsible for disease initiation , transmission and relapse [1–7] . Clostridium sordellii is a spore-forming bacterium that is responsible for severe human and animal diseases , including enteritis , bacteraemia and soft-tissue infections [8–12] . C . sordellii pathogenesis is toxin-mediated with two major toxins , TcsL and TcsH , responsible for host cell cytoskeletal disorganisation and death [12 , 13] . Mutagenesis studies using a TcsL-producing strain have shown that TcsL is essential for C . sordellii pathogenesis since a tcsL mutant was avirulent in an animal infection model while the TcsL-producing parent strain resulted in severe disease and death [14] . C . sordellii spores are important in disease because they are the infectious particle that can initiate infections . Soft tissue infections in injecting-drug users , for example , are likely to be initiated by C . sordellii spores which contaminate black-tar heroin . Speed-balling and skin-popping performed by these drug users can lead to ischemia and necrosis , which are favourable for spore germination and outgrowth [15 , 16] . Spores are also important in initiating diseases that result in post-partum and post-abortive clostridial toxic-shock . The uterus post-pregnancy contains amino acids , progesterone , and an elevated pH , which either trigger or enhance C . sordellii spore germination , resulting in spore outgrowth in the female reproductive tract and infection [17] . C . sordellii-mediated enteric diseases in animals are also likely to result from the ingestion of environmental spores that contaminate food or drinking water [9 , 10] . The C . sordellii spore is a complex structure , made up of an inner spore surrounded by a baggy , balloon-like exosporium [18] . The area between the inner spore and exosporium is known as the interspace in C . sordellii and in some Bacillus species , such as B . anthracis [18 , 19] . In C . sordellii , the inner spore is multilayered and composed of a core , inner membrane , germ cell wall , cortex and coat . The coat itself consists of three layers: the undercoat ( also termed the basement layer [20 , 21] ) , inner coat and outer coat [18] . The role of the inner spore proteins and structures in C . sordellii has not been studied , but in Bacillus subtilis and Clostridioides ( previously Clostridium [22] ) difficile they play a role in spore protection and germination [1 , 2 , 20] . The exosporium surrounding the inner spore is thought to be important as it is the first contact point between the spore and the environment [18 , 23] , however , very little is known about the functional role of this structure . In some species , the exosporium may play a role in spore protection against antibodies , degradative enzymes and host macrophages , and a role in regulating spore adherence to biotic surfaces [24–28] . Specific exosporial proteins may be responsible for some of these roles . For example , in Bacillus anthracis , the exosporial proteins elongation factor-Tu , enolase and arginase are thought to prevent spore opsonization or to provide protection against free radicals within macrophages , thus contributing to the survival of the spore or the germinating bacterium [24 , 25] . Despite the importance of exosporial proteins , the C . sordellii spore proteome has not previously been characterised . Exosporial proteins have been identified in the closely related organism C . difficile , however , the functional role of most of these proteins is not known [29] . Of the few C . difficile proteins that have been characterised , CdeC was found to be essential for the proper assembly of the exosporium and coat layers [30] . Characterisation of spore proteins is required to better understand their functional and structural roles , and these studies are particularly important for exosporial and other outer spore proteins , since they are the first contact point between a host or the environment and the bacterium . However , studying the role of exosporial proteins by their inactivation may result in the absence of or defects in other spore proteins . For example , in B . anthracis mutagenesis of bxpB also results in the mislocalisation of BclA [31] . This is a limiting factor when studying the role of proteins using a mutagenesis approach . In this study , we investigated the C . sordellii outer spore proteome and identified two proteins , CsA and CsB , that are likely to be associated with these spore components . Insertional inactivation of csA and csB resulted in the production of spores with structurally defective coat and exosporial layers , with spores of the csA mutant also exhibiting altered resistance properties . In comparison to wild-type spores , the csB mutation resulted in spores with increased adherence to cervical cell lines , with the cervix being a physiologically relevant anatomical site from which C . sordellii has been isolated [32 , 33] . C . sordellii enteric disease in animals is well described [9–11 , 34]; for this reason , a mouse model of C . sordellii gastrointestinal infection was developed to determine if CsA and CsB play a role in infection and disease . Using this model , we showed that spores from the csB mutant were not cleared as effectively from the gastrointestinal tract and persisted in this niche more effectively than wild-type spores . This study is the first to characterise C . sordellii outer spore proteins via the construction of mutants . The development of a gastrointestinal model of C . sordellii infection , in which spores are administered to reflect what likely occurs in a naturally occurring infection , has allowed the role of C . sordellii outer spore proteins to be examined in the context of the host .
To determine the protein composition of the C . sordellii outer spore , proteins were chemically extracted from the spores of strain ATCC9714 using an alkaline solution without SDS as described in the materials and methods . The absence of SDS prevents the co-extraction of coat proteins with the exosporial proteins [35] and this method has previously been used in B . anthracis and B . cereus to identify the exosporial proteins BclA and ExsM [35 , 36] . Following the extraction procedure , spores were visualized using transmission electron microscopy ( TEM ) to confirm the removal of the exosporium without disruption of the rest of the spore , in line with procedures performed by Redmond et al [37] . Most of the exosporium appears to have been removed with the treatment ( S1 Fig ) . Mass spectrometry was then used to determine the identity of the proteins in the extract ( S1 Table ) and these proteins were assigned to six specific categories ( S2 Table ) based on categories previously used in the literature [29 , 38] . C . sordellii proteins that were orthologues of proteins identified in the C . difficile exosporium were assigned to the same categories as those assigned to C . difficile [29] , which included uncharacterized proteins H477_3144 and H477_2973 . Although several proteins did not have orthologs in C . difficile , exosporial proteins of the same name were identified in previous studies [29 , 38] and were classified as follows: peptidase M20/M25/M40 family proteins ( H477_0435 and H477_0436 ) were classified according to the peptidase M20 family protein identified in the C . difficile spore coat [38]; reverse rubrerythrin-1 ( H477_0313 and H477_0314 ) were classified according to rubrerythrin proteins identified in the C . difficile exosporium spore coat [29 , 38]; the manganese-containing catalase family protein ( H477_3486 ) was placed in the ‘spore assembly’ group together with the manganese containing catalase family protein ( H477_3485 ) that showed homology to CotD of C . difficile [29] . The category assignment of several other proteins was determined with the aid of additional published literature , as follows: putative peptidoglycan binding domain protein ( H477_0372 ) has a SpoIID and peptidoglycan binding domain and may therefore be involved in the degradation of the cortex upon germination [39]; small , acid-soluble spore protein ( SASP ) beta ( H477_4660 ) was identified in the exosporium of B . anthracis [40] , SASPs protect the DNA of the spore and are typically found in the spore core [2 , 40]; coat F domain protein ( H477_1527 ) was identified in the C . difficile exosporium but its role remains unknown [2]; putative amidase domain protein ( H477_1207 ) , cupin domain protein ( H477_1872 ) and fascin domain protein ( H477_5266 ) have not to our knowledge been identified in spores of other species and no putative role could be deduced from the literature . All proteins labeled as ‘putative uncharacterized proteins’ and listed in the ‘unknown putative role’ category either had homology to uncharacterised proteins in the C . difficile or B . cereus proteomes or did not show homology to any protein in these species . C . sordellii proteins categorized as cytosolic proteins were done so because these proteins were the same or similar in name to the cytosolic proteins found in C . difficile or B . cereus [29 , 38] . Cytosolic proteins identified in the exosporium are likely to be mother cell proteins that were trapped in the exosporium during its formation [29] . Two of the uncharacterised C . sordellii outer spore proteins identified in this study , H477_3144 and H477_4099 , showed high peptide sequence coverage in the mass spectrometry analysis and were therefore chosen for further analysis . Specifically , 52 and 21 peptides were identified for protein H477_3144 and 23 and 29 peptides identified for H477_4099 , respectively , in each of the biological replicates ( S1 Table , S2 Table ) . Analysis of the upstream genome regions of H477_3144 and H477_4099 identified two other putative uncharacterized proteins ( H477_3145 and H477_4098 , respectively ) . Sanger DNA sequencing of these regions confirmed errors in the published sequences of gene H477_3145 ( missing nucleotide at position 372 ) and gene H477_4098 ( missing nucleotide at position 432 ) that resulted in a premature stop codon in both sequences [41] . When these errors are corrected , it appears that H477_3144 and H477_3145 , together with the intergenic region , are expressed as a single polypeptide of approximately 45 kDa , designated protein C . sordellii-A ( CsA ) . Similarly , H477_4098 and H477_4099 , together with the intergenic region , were also deemed to encode a single polypeptide of approximately 43 kDa , designated protein C . sordellii-B ( CsB ) . The gene and protein sequences of CsA and CsB are provided in the supporting information ( S1 Appendix ) . To confirm that the proteins were produced as single polypeptides , we performed Western blot analysis using recombinantly expressed CsA and CsB with protein-specific antibodies and showed that CsA and CsB are expressed and can be visualised at the predicted sizes ( S2A and S2B Fig ) . The recombinant proteins were expressed in Escherichia coli with a C-terminus HIS tag resulting in a 46 kDa and 44 kDa protein for recombinant CsA and recombinant CsB , respectively . Western blot analysis on wild-type exosporial extracts using the CsA-specific antibodies showed the presence of one dominant protein species of about 45 kDa ( Fig 1A lane 1 ) . Western blot analysis on wild-type exosporial extracts using the CsB-specific antibodies showed the presence of either two or three dominant protein species ( Fig 1B lane 1 , S2C Fig ) . The lower protein band present when three protein species were seen is at approximately 43 kDa , which is the expected size of CsB ( Fig 1B lane 1 ) . The middle protein band is approximately 45 kDa and the upper protein band is approximately 50 kDa ( Fig 1B lane 1 ) . The presence of multiple protein species for CsB indicates that the protein may be post-translationally modified . Note that there are no alternative start sites in the gene sequence that would result in the production of the other CsB protein species detected . Given that the C . difficile protein CdeC [30] showed homology to H477_3144 in a BLAST search against the C . sordellii genome ( S2 Table ) , we aligned the new full length CsA to CdeC using the BLASTP suite-2sequences program ( parameters: maximum identity ≥ 30% , e-value of ≤ 0 . 005% ) and showed that the two proteins are homologous ( 42% sequence identity ) ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) [42] . The CdeC protein is a 42 kDa cysteine rich protein ( cysteine residue content is 8% ) [30] . CsA is therefore slightly longer than CdeC , with small insertions observed predominantly in the N- and C-terminus of the protein . CsA contains 8 . 5% cysteine residues , also making it a cysteine rich protein . Analysis using the Fold and Function Assignment server ( FFAS , http://ffas . sanfordburnham . org [43] failed to identify any known protein folds , nor were any conserved domains identified . No transmembrane domains were identified using TMpred ( https://embnet . vital-it . ch/software/TMPRED_form . html ) [44] . Blast and FFAS analysis of CsB showed that this protein did not share homology with any characterised protein . The BLAST analysis of the CsB protein did identify homologs in Clostridium bifermentans , Paraclostridium benzoelyticum and Romboutsia ilealis , however , these homologs are also annotated as hypothetical proteins [42] . The cysteine content of CsB is not as high as CsA ( only 7 . 9% ) but this protein does have a high proline content ( 10 . 3% ) . Comparison of the sequences of CsA to CsB shows that the two proteins share only 17 . 8% sequence identity . The genes encoding CsA and CsB were insertionally inactivated using targetron technology and the mutants confirmed by PCR and Southern blotting ( S3 Fig ) using csA- ( S3A Fig ) , csB- ( S3B Fig ) and intron-specific ( S3C Fig ) probes . As expected , the csA-specific probe detected a band of approximately 2 . 3 kb and 4 . 1 kb for the wild-type and the csA mutant , respectively . When the csB-specific probe was used , a band of approximately 3 . 2 kb and 5 . 0 kb was detected for the wild-type and the csB mutant , respectively . These size differences correspond to the inserted targetron element ( approximately 1 . 8 kb ) . The intron-specific probe detected a single band for the mutant strains , confirming the integration of the targetron element . As expected , no intron-specific band was detected in the wild-type DNA . Complementation of the mutants was achieved via homologous recombination through a markerless double cross-over recombination event designed to remove the targetron element from the mutant strains . These revertants were confirmed by Southern blotting ( S3 Fig ) . As expected , Southern hybridization using the csA- and csB-specific probes detected bands of approximately 2 . 3 kb and 3 . 2 kb for the csA complemented and csB complemented strains , respectively . In addition , no bands were detected when DNA from the complemented strains was probed with the intron-specific fragment , indicating the loss of the targetron elements . Western blot analysis on the exosporial proteins from the wild-type and mutant strains , using CsA- and CsB-specific antibodies , showed that CsA and CsB were absent in the respective mutants ( Fig 1A and 1B lane 2 ) with other exosporial proteins still being produced as detected by sodium dodecyl sulfate gel electrophoresis ( SDS-PAGE ) ( S2D Fig ) . Complementation successfully restored the proteins in their respective strains ( Fig 1A and 1B lane 3 ) . As spore formation is a highly regulated process , complementation was carried out in cis for the csA and csB mutants to retain appropriate regulatory control of the genes . Visualisation by electron microscopy showed that the exosporium and inner spore of the csA and csB mutants differed to the wild-type ( Fig 2 ) . Spores of the wild-type strain had an inner spore composed of multiple distinct layers , and the inner spore was located within the centre of a baggy exosporium ( Fig 2A , 2D , 2G and 2J , S4A Fig ) . By comparison , csA mutant spores had an inner spore that was positioned towards one pole of the exosporium ( Fig 2B , 2E and 2H , S4B Fig ) , with 37 out of 40 whole spores analysed by TEM exhibiting this phenotype in comparison to 3 out of 40 wild-type whole spores analysed ( S4D Fig ) . Furthermore , the inner spore layers of the csA mutant did not appear to be as distinct as those of the wild-type strain ( observed in 10 out of 10 spores ) ( Fig 2J and 2K ) , which may indicate an abnormality in these structures . Electron dense bodies , which appeared to originate from the spore coat , were commonly observed in the interspace of the csA mutant ( 8 out of 10 spores observed ) ( indicated by CM , Fig 2K ) , but not in any of the wild-type spores ( 0 out of 10 spores observed ) , and may indicate spore coat instability in the mutant . The length and width of csA mutant spores ( including the inner spore and the exosporium ) appeared to be the same as those produced by the wild-type strain ( S4F–S4H Fig ) . Unlike the csA mutant spores , csB spores did not have an attached exosporium ( Fig 2C , 2F and 2I ) , and this was observed in all spores analysed by TEM ( n = 40 ) . However , putative detached exosporia were seen in the csB spore samples ( Fig 2M , S4E Fig ) , and , as discussed earlier , exosporial proteins could be detected by SDS-PAGE ( S2D Fig ) , suggesting that although csB mutant spores produce an exosporium it is not correctly tethered to the remainder of the spore . Spores of the csB mutant had distinct inner spore layers , but the coat was structurally aberrant in all spores observed ( minimum of 10 out of 10 observed ) . Specifically , the outer spore coat appeared to be absent from spores of the csB mutant , and the inner spore coat appeared to be partially mislocalised in some spores ( Fig 2L ) . In addition , fragments presumably composed of coat material were observed in spores of the csB mutant and may have sloughed off the inner spore ( indicated by CtM , Fig 2M ) . The lengths and widths of the inner spores remained unchanged between the wild-type and mutant spores ( S4F and S4G Fig ) . Note that the mutants exhibited a wild-type spore morphology when complemented in cis with the intact gene ( Fig 2N and 2O ) . Since CsA and CsB appear to be located on the spore surface , we investigated if these proteins were immunogenic by raising antibodies against whole wild-type spores in rabbits . Western blot analysis on exosporial protein extracts from the wild-type , csA and csB mutant strains was performed using these antibodies ( Fig 1C , S2E Fig ) . Though spores of the csB mutant do not have an attached exosporium , detached exosporial material is visible in the spore sample ( Fig 2M , S4E Fig ) and proteins from this material were extracted for the analysis . Prominent protein bands were detected in the wild-type and csA mutant strains but were absent in the csB mutant strain ( Fig 1C , S2E Fig ) . These protein bands were of a similar molecular weight to those previously detected by Western blotting using CsB-specific antibodies against the wild-type exosporial extract ( Fig 1B lane 1 , S2C Fig ) . We performed the same analysis using recombinant CsB and a protein of approximately 44 kDa in size was detected , as expected ( Fig 1D ) . These results suggest that CsB is an immunogenic protein in rabbits . All strains were tested for their ability to produce viable spores in liquid sporulation media over a 72 hour period . There was no difference in the total cell counts ( spores and vegetative bacteria ) between the strains at any time ( Fig 3 ) . Additionally , for each strain , there was no significant increase in the number of spores produced at 24 hours compared to 72 hours post-inoculation , indicating that sporulation was complete for all strains within 24 hours . However , the csA mutant produced 6- to 11-fold less viable spores ( 84%-91% reduction ) compared to the wild-type strain at each time point post-inoculation ( p = 0 . 0087–0 . 0317 ) . The colony forming efficiency of the spores was compared by plating a standardised number of purified spores , as determined by counting spores with a haemocytometer , onto solid media followed by growth at 37°C . No differences in colony forming efficiencies were observed between the strains ( S5A Fig ) . The spores from all strains were then examined for their ability to withstand high temperatures and various chemical treatments . Spores of the csA mutant showed an 15% increase in colony forming efficiency when incubated at 75°C ( Fig 4A ) , a 51% decrease in colony forming efficiency when incubated with hydrochloric acid ( Fig 4B ) and a 29% decrease in colony forming efficiency when incubated with sodium hydroxide ( Fig 4C ) when compared to wild-type spores ( p = 0 . 0286 ) . Spores of the csB mutant did not show sensitivity to any of these treatments in comparison to wild-type spores . Spores of all strains were also incubated with 1 mg/ml lysozyme and 80% ethanol , however , no differences in spore colony forming efficiency were detected between the strains following these treatments ( S5B and S5C Fig ) . Spore colony forming efficiency of the csA mutant in response to heat , hydrochloric acid and sodium hydroxide returned to wild-type levels upon complemention ( Fig 4 ) . C . sordellii causes necrosis and oedema of the female reproductive tract post-birth or post-abortion , perhaps as a result of vaginal tearing which may allow contaminating bacteria to ascend a dilated cervix [12] . C . sordellii has also been isolated from the cervix of women presenting with pelvic infections [32 , 33] and one study identified 0 . 2% of healthy women as asymptomatic vaginal carriers [8] . Here , human cervical ( Ect1/E6E7 and End1/E6E7 ) and vaginal ( VK2/E6E7 ) cell lines were used to investigate if wild-type spores adhered to these physiologically relevant cell lines and if spore abnormalities in the csA and csB mutants altered their adherence properties ( Fig 5 ) . The results showed that 55% of spores isolated from the wild-type and csA mutant adhered to the Ect1/E6E7 and End1/E6E7 cell lines 3 hours post-infection ( Fig 5A and 5B ) . However , spores of the csB mutant were more adherent with a 9- fold and 13- fold increase in adherence to the Ect1/E6E7 and End1/E6E7 cell lines compared to wild-type spores , respectively ( p = 0 . 0286 ) with adherence returning to wild-type levels upon complementation . No difference was detected between the different strains on the VK2/E6E7 cell line , with 55% of spores found to adhere three hours post infection ( Fig 5C ) . C . sordellii has been isolated from the gastrointestinal tract of humans and animals and causes enteric disease in animals [8–11] . In a similar way to C . difficile , it has been hypothesised that changes in the gastrointestinal tract environment or microbiome may induce susceptibility to C . sordellii infection and that the infection source can be internal or external [9–11 , 34] . Although C . sordellii soft tissue infection models have been developed [45–47] , a gastrointestinal tract infection model has not been reported . Furthermore , despite the importance of spores in the infectious cycle , previous C . sordellii animal models have administered vegetative cells and not spores to establish infection in animals and spores have not been monitored during these infections [45–47] . To investigate the role of wild-type and mutant C . sordellii spores in the gastrointestinal tract , a mouse infection model was established . In this model , mice were administered antibiotics to disrupt their gastrointestinal tract microbiota before being orally gavaged with C . sordellii spores ( a minimum of five mice were orally gavaged with each strain ) . Mice were then monitored for spore shedding in their faeces every day post infection , as a surrogate measure of C . sordellii colonisation . With the wild-type strain , the mice had high numbers of faecal spores initially ( mean of 4 x 106 at 1 day post-infection ) with the number of spores decreasing over the duration of the experiment ( mean of 2 x 104 at 4 days post-infection ) ( Fig 6A ) . These results suggest that strong colonisation occurs post-infection after which wild-type levels of C . sordellii are significantly reduced in the gastrointestinal tract . In addition , mice administered wild-type spores did not display disease symptoms of weight loss ( Fig 6B ) , diarrhea and behavioral changes in comparison to uninfected mice . Colonisation of the csA and csB mutants was examined in the mouse infection model in comparison to the wild-type strain . Mice were infected with spores from the wild-type , mutant or complemented strains and faecal spore numbers monitored every day post-infection for the duration of the experiment . Mice infected with csA mutant spores had similar colonisation patterns to mice infected with wild-type spores . Spore shedding from animals infected with these strains steadily decreased over time , resulting in a significant decrease in faecal spore numbers at 4 days ( mean of 2 x 104 for wild-type and csA ) compared to 1 day ( mean of 4 x 106 and 3 x 106 for wild-type and csA , respectively ) post-infection ( p = < 0 . 0001 ) ( Fig 6A ) . By contrast , mice infected with csB mutant spores did not have a significant decrease in faecal spore numbers at 4 days ( mean of 3 x 106 ) compared to 1 day ( mean of 2 x 106 ) post-infection . Colonisation levels were restored to those of the wild-type strain when csB was complemented in cis ( Fig 6A ) . The csA complemented strain was not included in this analysis as there was no difference in the colonisation pattern in mice infected with wild-type and csA mutant spores . Furthermore , the mutations do not appear to affect the pathogenesis of C . sordellii in the mouse infection model as no significant differences in the disease symptoms of weight loss ( Fig 6B ) , diarrhea or behavioural changes were observed between mice infected with mutant or wild-type strains . To determine the variability in spore shedding between mice infected with each strain , the shedding values at every time point for a particular strain were combined and the mean was calculated . The deviation of each shedding value from this mean was determined and a Tukey post-hoc test was performed . This analysis demonstrated that mice infected with the csB mutant exhibited a greater variability between mice in the level of faecal spore shedding compared to mice infected with the wild-type , csA mutant and csB complemented strains ( Fig 6C ) . Finally , note that growth assays showed that growth between the C . sordellii strains was similar ( S6 Fig ) and therefore , these parameters were not responsible for the variable colonisation levels of the strains in the mouse model .
C . sordellii is a human and animal pathogen and the spores produced by this bacterium are important for infection [12 , 15 , 16 , 34] . We recently characterised the C . sordellii spore structure and showed that the outer spore layers consist of a coat which is surrounded by a baggy exosporium [18] . Spore surface proteins are likely to be important for early host interactions and disease and may be important for spore fitness within the host . Mutagenesis of exosporial proteins , such as CdeC in C . difficile and superoxide dismutases and alanine racemase in B . anthracis , resulted in spore structural abnormalities , disease attenuation in a mouse infection model and premature germination , respectively [30 , 48 , 49] . However , C . sordellii spores are structurally dissimilar to spores of C . difficile and B . anthracis . Therefore , to examine the composition and function of the C . sordellii exosporium , we performed mass spectrometry analysis on this extracted fraction and identified numerous proteins ( S1 Table , S2 Table ) . Two of these proteins , CsA and CsB , were characterised further in this study . CsA shared homology to the C . difficile protein CdeC whereas CsB did not have homology to characterised proteins from any other species . Here , we showed that CsA and CsB are important for the correct assembly of the C . sordelli spore and that csB mutant spores appear to be more adherent to human cervical cell lines and to persist in a mouse gastrointestinal infection model . Mutagenesis of the csA and csB genes resulted in structurally aberrant spores and mutagenesis of csB also resulted in spores with altered resistance properties . While wild-type C . sordellii spores had an exosporium that wraps around a centrally located inner spore ( Fig 2A , 2D and 2G ) , csA mutant spores had an inner spore located towards one pole of the exosporium ( Fig 2B , 2E and 2H ) . Thus , it is possible that CsA anchors or stabilises the position of the inner spore within the exosporium . No filaments have been observed connecting the exosporium to the inner spore [18] but it is likely that protein-protein interactions hold the structures together and CsA may play a role in these interactions . Although the spores of the csA mutant have an inner spore that is spatially disorientated , the exosporium appears to be present and associated with the inner spore . This is unlike spores isolated from a C . difficile cdeC mutant , where the exosporium was either partially or completely absent from the inner spore [30] . This result suggests that the CsA and CdeC proteins appear to serve a different structural or functional role within each species despite their homology to one another . Spores isolated from the csB mutant appear to lack an exosporium ( Fig 2C , 2F and 2I ) , however , detached exosporial material was visible in the spore samples ( Figs 2M and S4E ) . Importantly , csB mutant spores did not show any abnormalities in spore colony forming efficiencies ( S5A Fig ) , which requires spores to complete all stages of germination [50] , even though the exosporium plays a role in regulating germination in other species . An examples of the exosporium playing a role in germination is seen in B . anthracis where the exosporial enzyme alanine racemase converts the germinant L-alanine to D-alanine , which inhibits germination [48] . Mutagenesis of the gene encoding alanine racemase resulted in premature germination of the developing spore within the mother cell and more efficient germination of the mature spore in comparison to wild-type spores [48] . In spores of C . difficile the exosporium also appears to play a role in germination . A previous study showed that C . difficile spores which had varying amounts of the exosporium proteolytically removed had higher germination levels when less of the exosporium was present [51] . In this regard , C . sordellii appears to be different to B . anthracis and C . difficile as the spore colony forming efficiency , and consequently spore germination , were unaffected in the csB mutant spores . Alternatively , the detached exosporium present in the spore preparations may be providing the enzymatic functions needed to regulate germination . As well as an abnormal exosporium , csA and csB mutant spores appeared to have spore coat defects . Electron dense material within the interspace region of csA mutant spores was seen , which may represent fragments of the coat that were dissociated from the rest of the spore ( Fig 2K ) . An outer spore coat was absent in all csB mutant spores and the inner coat appeared partially mislocalised in some of these spores ( Fig 2L ) . These results suggest that the csA and csB mutations destabilise the spore coat layers . Mutagenesis of exosporial proteins in other species , such as cdeC in C . difficile , have also resulted in structural abnormalities to the exosporium and coat [30] . The coat is believed to protect spores from lysozyme [52] , however , csA and csB mutant spores were as resistant to lysozyme treatment as wild-type spores ( S5B Fig ) , suggesting that the coat present in these mutants remains impermeable to lysozyme or that another mechanism is conferring this resistance phenotype . In comparison to csA mutant spores , C . difficile cdeC mutant spores appeared to have all coat layers intact but the layers differed in thickness and spores differed in sensitivity to lysozyme in comparison to wild-type spores [30] . In a similar way to csB mutant spores , C . difficile cotA and B . subtilis cotE mutants also lacked an outer spore coat , however , in contrast to csB spores , these spores had increased sensitivity to lysozyme compared to the wild-type and the cotA mutants also had increased sensitivity to ethanol [53 , 54] . Additional inner spore abnormalities were also detected in csA mutant spores . Spores of the csA mutant were more sensitive to hydrochloric acid compared to wild-type spores ( Fig 4B ) . Spores of the csA mutant also had increased sensitivity to sodium hydroxide ( Fig 4C ) . This result may reflect defects in cortex hydrolysis since treating B . subtilis spores with sodium hydroxide appeared to inactivate cortex lytic enzymes , thus preventing the complete germination of spores [55] . In B . subtilis , spores treated with hydrochloric acid subsequently ruptured , suggesting that hydrochloric acid affects the permeability barrier of these spores [55] . The inner membrane of B . subtilis spores may be important in maintaining this permeability barrier and may help to keep the spores dehydrated and resistant to hydrochloric acid , heat and ethanol [55] . Although csA mutant spores had greater sensitivity to hydrochloric acid , they were more resistant to heat and did not show an increased sensitivity to ethanol in comparison to wild-type spores ( Fig 4A and 4B , S5C Fig ) . Similarly , csA mutant spores also differed to C . difficile cdeC mutant spores since cdeC mutant spores were more sensitive to heat and ethanol compared to wild-type spores [30] . These results suggest that the spore structures of C . sordellii , B . subtilis and C . difficile may serve different functions and may therefore respond in different ways to the same treatments . In the csB mutant , the inner coat forms the outermost layer of the spore , not the outer coat as found in wild-type spores . It is therefore likely that spore surface properties are altered in this mutant . In support of this hypothesis , csB mutant spores displayed a greater level of adherence to ectocervical and endocervical cells , but not vaginal cells , compared to wild-type spores ( Fig 5 ) , suggesting that the inner coat surface has a high affinity for cervical cells . By comparison , csA mutant spores displayed the same adherence levels as wild-type spores ( Fig 5 ) . It is possible that the exosporium masks adhesins in the coat that are specific to cervical cells and , consequently , these adhesins are exposed in the csB mutant spores . Alternatively , the exosporium may be responsible for specific binding of spores to cervical cells and , as a result , in the absence of an exosporium the csB mutant spores may bind non-specifically to cervical cells . A similar phenomenon has been seen in cdeC mutant spores , which are missing the outer layer of the exosporium and show higher levels of adherence to gut epithelial cells compared to wild-type spores [30 , 56] . Of note , wild-type spores adhered to both vaginal and cervical cell lines , reflecting the published literature , which reports that C . sordellii has been isolated from both regions of the female reproductive tract [8 , 32 , 33] . Spores of the csB mutant also persisted in the gastrointestinal tract of infected mice ( Fig 6A ) . In a similar way , mutations in the C . difficile genes encoding the spore surface proteins BclA2 and BclA3 resulted in strains that displayed increased colonisation in a mouse model of gastrointestinal infection [57] . Conversely , in B . anthracis , a mutation in the gene encoding a similar spore surface protein , BclA , resulted in a strain with decreased persistence in the host when tested in a pulmonary mouse infection model of anthrax [35 , 58] . In addition to the persistence of the csB mutant in the gastrointestinal tract , spore shedding from this mutant was more variable between mice challenged with this strain compared to mice challenged with the wild-type strain ( Fig 6C ) . The increased persistence of the csB strain may be due to the misassembled spore coat having a greater affinity for the mouse gut epithelium . Given that the spore is less persistent in its native conditions , perhaps the exosporium plays a role in spore transmission . Furthermore , the variation observed in spore shedding for mice challenged with the csB mutant may be due to the variable inner coat presentations of csB mutant spores , which may have differing affinities for the gut epithelium . Given that there were no differences in sporulation or colony forming abilities between the wild-type and csB mutant ( Fig 3 , S5A Fig ) , it is unlikely that these factors played a role in the colonisation patterns observed in the mouse gut model . Note that although CsB was found to induce an adaptive immune response in rabbits ( Fig 1C and 1D ) , the innate immune response to this protein was not examined . For this reason , it was not possible to determine if an altered innate response to infection with the csB mutant compared to the wild-type plays a role in the persistence of this strain in the infection model . It is possible that the absence of CsB affects the ability of the innate immune system to detect and clear this strain from the gastrointestinal tract in comparison to the wild-type strain , leading to the colonisation and shedding patterns detected . It is unlikely that the adaptive response to this protein plays a role in clearance of the wild-type strain in the C . sordellii infection model used here because of the short duration of infection ( 4 days ) , however , such a response may be important and protective during natural infections . The outer structures of spores , including the exosporium , are functionally important because they are the point of contact between the spore and the environment . These structures may also be involved in spore resistance , adherence and in the regulation of germination [24 , 25 , 28 , 48] . Despite their importance , the current understanding of outer spore structures and the role of the proteins that compose these structures is limited , especially in the context of C . sordellii . Here we have shown that the C . sordellii outer spore proteins CsA and CsB are necessary for the correct assembly of both the inner spore and exosporium . CsA is required for wild-type levels of spore resistance to heat , sodium hydroxide and hydrochloric acid , while CsB is required for normal levels of spore adherence to cervical cells . The absence of CsB also allows the csB mutant to persist within an infected host in a gastrointestinal infection model , possibly because of the increased adherence properties of csB spores , however , this needs to be investigated further . The difference in structural abnormalities and resistance properties between spores of the C . sordellii csA mutant and the C . difficile cdeC mutant illustrates that , despite their seeming homology , spore proteins can have variable functions in different species . This finding is not unexpected based on the structural variability between spores from these two species [18] and has been shown before for the ExsY protein from B . cereus and B . anthracis [59 , 60] . Importantly , this result highlights the necessity of studying each spore protein in the cognate species from which it originates . Gaining an understanding of outer spore proteins across a broad cross-section of sporulating bacteria is increasing our knowledge of the various roles that they play , particularly in a structural context and in interactions with the infected host .
All strains of C . sordellii were derivatives of ATCC9714 and cultured as previously described [18] unless otherwise stated . Media were supplemented with one or more of the following antibiotics where appropriate: D-cycloserine ( 250 μg/ml ) , erythromycin ( 10 μg/ml ) , thiamphenicol ( 10 μg/ml ) or anhydrous tetracycline ( 50 ng/ml ) . C . sordellii spores were prepared and stored as outlined previously [18] and were determined by light microscopy to be >99% free of vegetative cells and debris . E . coli strains HB101 , DH5α and TOP10 were grown in 2x YT media with the growth conditions and strain characteristics described previously [61] . E . coli strain BL21 ( C43 ) [62] was grown in LB media . E . coli cultures were supplemented with chloramphenicol ( 30 μg/ml ) , tetracycline ( 10 μg/ml ) or kanamycin ( 20 μg/ml ) where necessary . Outer spore proteins were prepared from two independent biological spore stocks and MS performed independently on each extract . C . sordellii outer spore proteins were extracted from purified spores as previously described [35] . Briefly , spores were incubated at 90°C for 15 min in an extraction buffer containing 8 M urea and 2% 2-mercaptoethanol . The sample was then centrifuged at 13 , 000x g for 10 min and the supernatant , containing the extracted proteins , was dialysed against phosphate buffered saline ( PBS ) ( 137 mM NaCl , 2 . 7 mM KCl , 1 . 4 mM KH2PO4 , 4 . 3 mM Na2HPO4 ) using a molecular weight cut-off between 6000–8000 . The proteins were reduced with 25 mM dithiothreitol , alkylated with 55 mM iodoacetic acid and then trypsin digested in a 1:25 trypsin to protein ratio . The peptides were desalted and concentrated using a peptide trap ( Michrome peptide Captrap ) with 0 . 1% formic acid and 2% acetonitrile at 8–10 μl min-1 . The peptides were then eluted from the column in a linear gradient ( 0%-90% acetonitrile , in 0 . 1% formic acid ) over 80 min using a flow rate of 0 . 5 μl/min . Each sample was then subjected to positive ion nanoflow electrospray MS ( QSTAR , SCIEX at the Australian Proteome Analysis facility ) , operated in information acquisition mode , and a time-of-flight-MS survey scan was acquired ( m/z 380–1600 ) . From the survey scan , the three most intense multiply charged ions ( counts ≥25 ) were selected for MS/MS analysis ( m/z 100–1600 ) . Using ProteinPilot version five ( SCIEX ) , MS data was used to search against the C . sordellii strain ATCC9714 sequence obtained from UniProt ( taxonomy ID 1292036 ) [63] . The protein false discovery rate was set at 1% , and proteins were reported only if at least one unique peptide was identified in both independent biological replicates analysed . Exosporial proteins previously identified in C . difficile by MS [29] were then used to search for orthologs to the outer spore proteins identified by MS in this study . BLASTP ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) was used for this analysis [42] with a maximum identity of ≥ 30% and an e-value of ≤ 0 . 005% to determine orthologs present . The C . sordellii proteins labeled as ‘putative uncharacterised proteins’ were also searched for orthologs in the proteomes of C . difficile ( taxonomy ID 1496 ) and B . cereus ( taxonomy ID 1396 ) to determine any homology to a characterized protein . C . sordellii exosporial proteins categorised as cytosolic proteins were excluded from this analysis , to be consistent with the proteins classified as exosporial proteins in C . difficile [29] . The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [64] partner repository with the dataset identifier PXD009235 . The csA and csB mutants were constructed in strain ATCC9714 using targetron technology as previously described [61] , with the exception that the plasmid pDLL46 was retargeted to give rise to pDLL57 and pDLL58 for the disruption of csA and csB , respectively . Plasmid pDLL46 is a derivative of pMTL9361 [65] with the HindIII and BsrGI sites removed from rep and lacZa and an oriT from RP4 and Tn916 included in this vector . Plasmids pDLL57 and pDLL58 were then conjugated into C . sordellii strain ATCC9714 . Expression of the targetron element was induced by the addition of anhydrous tetracycline ( 50 ng/ml ) with the targetron element inserting after nucleotide 920 on the antisense strand for csA and after nucleotide 742 on the sense strand for csB . PCR and Southern blotting were then used to confirm the correct insertion of the targetron element and Western blot analysis using protein specific antibodies was used to confirm the loss of expression of CsA and CsB in each mutant . The csA mutant was designated DLL5069 and the csB mutant was designated DLL5071 . Complementation of the csA and csB mutants was performed using a markerless double cross-over homologous recombination system , as previously described [66] , to remove the targetron element . Briefly , PCR was used to amplify DNA fragments containing the wild-type csA gene ( primer DLP551 AGCATGCTGTAGATTTATCTGGCGTTTTACAC and primer DLP552 AAAGACGTCTATGGATCTTCAATACTATTCGACC ) or the wild-type csB gene ( primer DLP556 AAAGACGTCTTTTAATTGGTTCACTCCATGTGTC and primer DLP557 AGCATGCGAAACCTCTACTTCACTAGCATTGT ) as well as genomic regions both upstream and downstream of each gene to generate fragment lengths of 2 . 7 and 2 . 8 kb , respectively . The fragments were cloned into the AatII/SphI sites of the clostridial vector pJIR3566 [62] . The correct DNA sequences of the fragments were confirmed by nucleotide sequencing and the plasmids then designated pDLL165 and pDLL167 for csA and csB , respectively . Conjugative transfer of the appropriate plasmid into DLL5069 or DLL5071 from the donor strain E . coli HB101 ( pVS520 ) was performed as previously described [61] . Isolates growing on HIS agar supplemented with D-cycloserine and thiamphenicol were passaged in HIS broths for up to five days and then plated onto HIS agar . Individual colonies were tested for their sensitivity to erythromycin which indicated the loss of the targetron element and plasmids pDLL165 or pDLL167 . Complementation of the mutations was again confirmed by PCR , Southern blot and Western blot analysis . The complemented csA mutant ( DLL5069 ) was named DLL5204 and the complemented csB mutant ( DLL5071 ) was named DLL5208 . All plasmid and genomic DNA isolation and manipulations were performed as outlined previously [61] . Nucleotide sequencing was performed by Micromon ( Monash University , Melbourne , Australia ) using a Prism BigDye terminator cycle sequencing kit according to the manufacturer’s instructions ( Applied Biosystems , MA , USA ) . PCR was used to confirm the insertion of the targetron element in the appropriate genes or to confirm the loss of the targetron element upon complementation: for csA , oligonucleotide primers DLP111 ( GCTATATAGAAACAAATGAAGTTGAAGATG ) and DLP112 ( CTCCATTTGATTTTTTACCATCAGTAA ) were used; for csB , oligonucleotide primers DLP113 ( AGTATTTACTGTAGATCCTAGACCTATGGG ) and DLP114 ( GGAATCTAGCTGTAAAAGGATCACAA ) were used . A 1 . 8 kb size increase in the PCR product indicates the presence of the targetron element . For confirmation by Southern hybridization , purified genomic DNA was digested with either XbaI and AvaII ( csA ) or XbaI and PacI ( csB ) . Southern hybridization procedures were as outlined previously [61] and blots were hybridized either with an intron/ermB-specific probe [61] , a csA-specific probe , generated using oligonucleotide primers DLP111 and DLP112 , or a csB-specific probe , generated using oligonucleotide primers DLP113 and DLP114 . To confirm the correct exosporial protein profiles of both mutants and complemented strains , exosporial proteins were chemically extracted from the spores of these strains [18] and the protein concentration determined using the BCA protein assay kit ( Pierce ) according to the manufacturer’s instructions . Western blot analysis was then performed as outlined previously [61] using 10 μg of protein from each strain which was separated by SDS-PAGE on 12% gels [67] . The Western blots were then probed using CsA- and CsB-specific antibodies . The antibodies were produced by Mimotopes ( Melbourne , Australia ) and generated by immunising rabbits with CsA and CsB specific peptides ( CVFTDGKKSNGDDLDF and CEADDDENHNNHKCCK for CsA , CDRIFDFKCVNQQIPR and CLVVYSAPAEFKHHEK for CsB ) , with each peptide conjugated to a keyhole limpet hemocyanin . The extracted proteins were also visualised by separating 40 μg of protein on a 12% SDS-PAGE and staining with Coomassie Brilliant Blue G-250 ( Sigma Aldrich ) . Western blot analysis was carried out as described for the Western blot analysis using CsA- and CsB-specific antibodies except that proteins were probed with antibodies raised against ATCC9714 whole spores [18] To express CsA and CsB with a C-terminal 6xHis tag , csA was amplified with oligonucleotide primers DLP 143 ( TATACCATGGGCATGAAGGATTTAATGAGAGC ) and DLP 144 ( CGTCTCGAGTTTACAACATTTGTGATTG ) and CsB was amplified with oligonucleotide primers DLP147 ( TATACCATGGGCATGACTAAAAATAATACAGC ) and DLP148 ( CGTCTCGAGTTTTTCGTGATGTTTAAATTC ) . The amplified DNA for CsA and CsB was digested with NcoI and XhoI and cloned into pET-28b ( Novagen ) to give rise to plasmid pDLL82 and plasmid pDLL83 for CsA and CsB , respectively . Plasmids pDLL82 and pDLL83 were individually introduced into strain BL21 ( DE3 ) ( C43 ) and the proteins were overexpressed at the Protein production Unit , Monash University , Australia , using Autoinduction Media ( Overnight Express LB media , Merck ) . Protein purification was performed using nickel affinity chromatography and size exclusion chromatography ( GE Superdex 200 ) . Spores were prepared and visualized for TEM and SEM as previously described [18] All electron microscopy imaging was performed at the Monash CryoEM Ramaciotti Centre , Monash University , Australia . Whole spores that were imaged by TEM were analysed using FIJI software [68] to determine the inner spore and total spore lengths , inner spore width , position of the inner spore in relation to the exosporium and presence of an exosporium attached to the inner spore . For all strains , 10 spores per replicate were analysed from a total of four biological replicates . To determine differences between strains in the inner spore and total spore lengths and inner spore widths , the mean of the measurements for each biological replicate was calculated and this value was used for the statistical analysis . The positioning of the inner spore with respect to the exosporium was determined as follows: at each spore pole , the exosporium was measured as the length from the end of the inner spore at one pole to the end of the exosporium at the same pole . The pole with the longer exosporial length was then used to calculate the percentage of exosporium present at one pole using the formula ( longer exosporial length at spore pole/total spore length ) X 100 . Spores were considered to have an inner spore centrally located within the exosporium if 35% or less of the exosporium was present at one pole . Spores were considered to have an inner spore positioned non-centrally within the exosporium if more than 35% of the exosporium was present at one pole . To compare the viable spores being produced by different strains , HIS broths were inoculated with overnight cultures of C . sordellii to a final OD600 of 0 . 1 . The cultures were grown to an OD600 of 0 . 8 and then added to 20 ml TY broths at a 1 in 50 dilution . A sample of each culture was then taken every 24 hours for 72 hours . Half the sample was plated directly onto HIS agar to obtain total cell counts ( spores and vegetative bacteria ) . The other half was heated at 65°C for 30 minutes to kill vegetative bacteria and then plated onto HIS agar to obtain viable spore counts alone . The ability of spores to form colonies was used as an indicator that viable spores had been produced during the sporulation assay . To determine spore resistance , 1 x 106 spores were treated with ethanol ( 80% v/v ) , hydrochloric acid ( 0 . 5 M ) , sodium hydroxide ( 0 . 15 M ) , lysozyme ( 1 mg/ml ) or heat ( 75°C ) for 30 minutes . With the exception of heat , all treatments were performed at 37°C . All spore stocks were in water with the exception of spores for lysozyme treatment where spores were in PBS . Percentage spore viability ( the ability of spores to form colonies ) following treatment was calculated as follows: ( log10 CFU ml-1 after plating spores on HIS agar post treatment/log10 CFU ml-1 after plating untreated spores on HIS agar ) X 100 . The concentration of spores in all samples was standardized to approximately 6 x 107 spores ml-1 by counting the total numbers of spores present using a hemocytometer ( spores able to and spores unable to form colonies ) . The viable spore count for each sample was then determined by plating spores onto HIS agar and obtaining the CFU ml-1 . Total spore counts and viable spore counts were compared to determine any differences in colony forming efficiencies between spores of the wild-type and mutant strains . Spores were tested for their ability to adhere to or be internalized by the following cell lines: VK2/E6E7 vaginal epithelial cell line ( ATCC CRL-2616 ) , End1/E6E7 endocervical epithelial cell line ( ATCC CRL-2615 ) and Ect1/E6E7 ectocervical epithelial cell line ( ATCC CRL-2614 ) . Cell lines were cultured as previously described [69] , seeded at 2 . 5 x 105 cells ml-1 in 24-well culture plates and grown to 95% confluency . The cells were washed with PBS and 200 μl ( 2 . 5 x 106 cells ml-1 ) of C . sordellii viable spores , resuspended in antibiotic free media , added to each well at an multiplicity of infection of 10:1 and incubated at 37°C in an atmosphere of 5% CO2 . Samples were then processed as described previously [30] to obtain the percentage of bound spores , with the exception that cells were lysed in 0 . 1% triton X-100 in a final volume of 400 μl and samples plated onto HIS agar . Briefly , following incubation , infected cells were washed to remove unbound spores and the cells were lysed to obtain the number of adhered and/or internalized spores . This was compared to infected cells which were directly lysed to obtain the total spore numbers . The percentage of adhered and/or internalized spores was calculated as follows: ( CFU ml-1 adhered and/or internalized spores/ CFU ml-1 total spores ) X 100 . Animal handling and experimentation was performed according to the requirements of the 7th edition of the Australian Code of Practice for the care and use of animals for scientific purposes ( 2004 ) and The Victorian Prevention of Cruelty to Animals Act ( 1986 ) , and was approved by the Monash Animal Research Platform Committee under license number MARP/2014/145 . Groups of five pathogen-free six to eight week old wild type male C57BL/six mice ( Walter and Eliza Hall Institute of Medical Research ) were treated with antibiotics for seven days followed by three days of normal drinking water , as previously described [70] with the exception that gentamicin was administered at 0 . 07 mg/ml . Mice were then infected with 107 viable C . sordellii spores by oral gavage and monitored for disease symptoms of diarrhea , weight loss and behavioral changes . A minimum of five mice were orally gavaged with each strain . Mice were humanely euthanized at the completion of the experiment . Faecal samples were collected daily to monitor for C . sordelii spore shedding . Faecal samples were then resuspended in PBS at a final weight per volume ratio of 100 mg ml-1 , heated at 65°C for 30 minutes to kill vegetative cells and then plated onto HIS agar containing d-cycloserine ( 250 μg/ml ) , kanamycin ( 20 μg/ml ) , streptomycin ( 20 μg/ml ) , trimethoprim ( 20 μg/ml ) and naladixic acid ( 20 μg/ml ) . A Tukey post-hoc test was performed to determine significant weight loss differences between mice and variability in spore shedding between mice infected with each strain . Starter cultures of C . sordellii strains were grown overnight in HIS broths . The broths were then diluted to an OD600 of 0 . 02 in fresh HIS broth and a 200 μl volume was added to wells in a 96-well tray ( Grenier Bio-One ) . Trays were incubated in an anaerobic chamber and removed each hour for a period of nine hours to measure the absorbance on a Tecan plate reader at 600 nm . A Kruskal-Wallis test was performed to determine differences in growth between the strains . GraphPad Prism was used for all statistical analysis . A Mann-Whitney test was performed with a 95% confidence interval unless otherwise stated in which case a Kruskal-Wallis test was performed for the statistical analysis of the C . sordellii growth kinetics or a Tukey post-hoc test was performed for the statistical analysis of the weight loss between mice and the variability in spore shedding between mice infected with each strain in the mouse gastrointestinal tract model of C . sordellii infection . | Clostridium sordellii is a lethal pathogen in humans and animals and its spores are critical for initiating infection . Outer spore proteins in Bacillus and Clostridium species are important for spore fitness and pathogenesis , however , equivalent proteins in C . sordellii have not been identified and their role is unknown . In this study , we characterized two C . sordellii outer spore proteins , CsA and CsB , and showed that these proteins are required for correct spore structure and function . We also established the first mouse model of C . sordellii gastrointestinal tract infection . This model is physiologically relevant as C . sordellii can cause enteric disease in animals . Our model is unique in comparison to previous C . sordellii disease models in that spores , and not vegetative cells , are administered to reflect what likely occurs in naturally occurring infections . Using our model we showed that the absence of CsB allows a csB mutant strain to persist within the host . We also demonstrated that although C . sordellii CsA shares homology with the Clostridium difficile spore protein CdeC , these proteins contribute in different ways to spore phenotypes in the two bacterial hosts , highlighting the necessity of studying spore proteins in the cognate species from which they originate . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"gut",
"bacteria",
"animal",
"models",
"of",
"disease",
"microbiology",
"animal",
"models",
"mutation",
"model",
"organisms",
"gastroenterology",
"and",
"hepatology",
"experimental",
"organism",
"systems",
"bacteria",
"extraction",
"techniques",
"digestive",
"system",
"research",
"and",
"analysis",
"methods",
"clostridium",
"difficile",
"animal",
"models",
"of",
"infection",
"mutant",
"strains",
"microbial",
"physiology",
"protein",
"extraction",
"animal",
"studies",
"bacterial",
"spores",
"mouse",
"models",
"gastrointestinal",
"infections",
"gastrointestinal",
"tract",
"bacterial",
"physiology",
"anatomy",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
] | 2018 | Clostridium sordellii outer spore proteins maintain spore structural integrity and promote bacterial clearance from the gastrointestinal tract |
HIV-1 assembly and release are believed to occur at the plasma membrane in most host cells with the exception of primary macrophages , for which exclusive budding at late endosomes has been reported . Here , we applied a novel ultrastructural approach to assess HIV-1 budding in primary macrophages in an immunomarker-independent manner . Infected macrophages were fed with BSA-gold and stained with the membrane-impermeant dye ruthenium red to identify endosomes and the plasma membrane , respectively . Virus-filled vacuolar structures with a seemingly intracellular localization displayed intense staining with ruthenium red , but lacked endocytosed BSA-gold , defining them as plasma membrane . Moreover , HIV budding profiles were virtually excluded from gold-filled endosomes while frequently being detected on ruthenium red–positive membranes . The composition of cellular marker proteins incorporated into HIV-1 supported a plasma membrane–derived origin of the viral envelope . Thus , contrary to current opinion , the plasma membrane is the primary site of HIV-1 budding also in infected macrophages .
HIV-1 is an enveloped retrovirus that acquires its envelope by budding through limiting membranes . CD4+ T cells and macrophages are the primary targets of HIV-1 and are commonly used to study virus replication in tissue culture . Infected macrophages constitute a long-lived reservoir for HIV persistence and rebound and thus pose a major challenge for HIV clearance from infected individuals ( reviewed in [1–3] ) . Recently , differences in the site of virus assembly and budding have emerged as a major distinguishing feature of HIV-1 infection in macrophages and have been discussed as an important factor in virus persistence and dissemination . In HIV-1-infected primary CD4+ T cells and most cell lines , assembly and budding occur at the plasma membrane [4–8] , possibly involving lipid rafts [6 , 9] . In contrast , early ultrastructural studies implicated intracellular organelles as sites of HIV-1 budding in infected macrophages , such as the Golgi complex and vacuoles [10–13] . More recent immuno-electron microscopy ( EM ) studies of HIV-1-infected primary human macrophages provided strong support for the cell type–specific intracellular budding and proposed late endosomes ( LE ) and multivesicular bodies ( MVB ) as budding compartments in these cells [14 , 15] . These authors reported HIV-1 budding into large intracellular vacuoles bearing markers of LE/MVB , including tetraspanins ( CD63 , CD81 , and CD82 ) , lamp-1 , and MHC-II . Accordingly , extracellular infectious virions were shown to carry tetraspanins , while being largely devoid of most GPI-anchored and cell adhesion proteins tested [14–17] . Topologically , HIV budding and the formation of intralumenal vesicles of the MVB are similar ( away from the cytoplasm ) and share the requirement for ubiquitin conjugation and recruitment of the cellular ESCRT-I and ESCRT-III complexes [18] . Thus , budding of HIV-1 at LE/MVB may be ideally suited to exploit the cellular pathway of exosome formation . These considerations , together with the described EM studies , led to the current view that HIV-1 assembles and buds almost exclusively from late endosomal membranes in macrophages [14 , 17] . The resulting intracellular accumulation of HIV-1 was proposed to be important for pathogenesis and dissemination since HIV-1 can be retained in an infectious state for prolonged periods of time inside macrophages [19] and may be released in a delayed manner similar to secretion of exosomes [20 , 21] . Contrary to the prevailing view of ESCRT-localization , our recent immuno-EM analysis of uninfected and HIV-1-infected primary human T cells and macrophages showed that the analyzed ESCRT-proteins ( HRS , TSG101 , AIP1/ALIX , and VPS4 ) localized to both the endosomal compartment and the plasma membrane [22] . No significant relocalization of ESCRT-proteins was observed in infected cells , even when a high producer T-cell line was analyzed [22] . Our frequent observation of plasma membrane budding in HIV-1-infected primary macrophages prompted us to re-assess HIV-1 budding site localization in macrophages . To this end , we applied a novel ultrastructural approach , which is independent of immunomarker distribution , to distinguish endosomal structures from the plasma membrane . Our results reveal that the plasma membrane is the predominant site of HIV-1 budding also in primary macrophages and suggest a general pathway of HIV-1 morphogenesis .
We revisited HIV-1 assembly and budding in primary human monocyte-derived macrophages ( MDM ) , which had been used in previous studies to demonstrate HIV-1 accumulation within intracellular compartments [10–15] . MDM cultures infected with either of the two macrophage-tropic HIV-1 strains Ba-L or YU-2 showed largely similar replication kinetics ( unpublished data ) . Cells were fixed for EM at the peak of virus release , usually around 12–17 d after infection , and either embedded in Epon before sectioning or processed for cryo-sections followed by immunolabeling . Accumulation of mature HIV-1 particles with typical cone-shaped capsid morphology in large , seemingly intracellular vacuolar structures was readily detected in HIV-1 Ba-L- ( Figure 1A and 1B ) or YU-2- ( not shown ) infected cells . While virus-filled vacuolar structures were often observed , detection of typical HIV-1 budding structures was much less common . Importantly , early ( Figure 1A and 1C ) and late ( Figure 1B ) budding profiles were detected on the plasma membrane ( Figure 1B and 1C; open arrows ) , as well as on membranes that appeared to be intracellular on thin sections ( Figure 1A; open arrow ) . This result was independent of the method of monocyte enrichment or cultivation of MDM ( unpublished data ) . Furthermore , infected and uninfected MDM revealed a complex surface architecture with numerous protrusions and cross-sectioned invaginations suggesting that some apparently intracellular structures may be connected to the cell surface . Immuno-EM using antiserum against the HIV-1 capsid protein readily identified HIV-1 particles in large vacuolar structures with an apparent intracellular localization , both for strain YU-2 ( Figure 1D ) and Ba-L ( Figure 1E and 1F ) . The limiting membrane of these structures was sometimes connected to the extracellular space , indicating that they represented deep plasma membrane invaginations ( Figure 1F ) . Furthermore , virus particles were frequently detected in the extracellular space ( Figure 1G ) . Many virus-filled vacuolar structures contained membrane protrusions similar to the cell surface ( see below ) , and their limiting membrane sometimes displayed structures resembling clathrin-coated pits ( Figure 1H and 1I ) . HIV-1 budding profiles were detected on the limiting membrane of vacuolar structures ( Figure 1H ) and on the plasma membrane ( Figure 1G , 1J , and 1K ) . Budding profiles were often next to structures resembling clathrin-coated pits ( Figure 1H , 1J , and 1K , arrows ) , even when budding occurred into seemingly intracellular vacuoles . The observation of significant plasma membrane budding and of hallmarks of the plasma membrane on the limiting membranes of seemingly intracellular vacuoles prompted us to characterize the origin of these vacuolar structures in more detail . Sections of YU-2- and Ba-L- infected MDM were morphologically indistinguishable , and we therefore used Ba-L-infected MDM throughout the remainder of the study . Previous studies of HIV-1 morphogenesis in infected cells relied on immunolabeling of cellular proteins to identify membrane compartments . To allow an immunomarker-independent identification of the plasma membrane and endosomes , we combined two established EM methods in a novel ultrastructural approach . This involved labeling of endocytic compartments by internalizing BSA-gold followed by staining of the plasma membrane with ruthenium red ( RR ) during fixation . RR is a membrane-impermeant dye , which binds to carbohydrate moieties on the cell surface [23 , 24] . Because of its small size , RR can readily penetrate into membrane invaginations [25] , while fixation at 4 °C prevents its internalization . Upon post-fixation , RR forms an electron-dense precipitate detectable on Epon-sections , and this method has been used to study HIV-1 entry in macrophages [26] . Our previous experiments had shown that over 75% of all endosomal structures of primary human macrophages , including LE , were filled with BSA-gold after 2 h of internalization [22] . Accordingly , the combination of BSA-gold uptake and RR staining should unequivocally identify membrane structures as being of endosomal or plasma membrane origin . This was first tested on uninfected MDM to validate our approach ( Figure 2 ) . Figure 2A shows MDM morphology at low magnification; the smooth cell surface facing to the left was previously attached to the culture dish and is therefore poorly stained with RR . The complex upper surface is facing to the right , with many finger-like surface protrusions clearly stained with RR ( Figure 2B ) . Two neighboring cells form a tight-fitting contact zone with interdigitating membranes ( Figure 2A , arrows ) . Importantly , the limiting membranes of large , seemingly intracellular vacuolar structures that were heterogeneous in size and shape were also stained with RR , thus defining them as plasma membrane–derived ( Figure 2A–2F ) . These structures generally had no apparent connection to the cell surface in the plane of the section and were often found deep inside the cell ( Figure 2B and 2C ) . No RR labeling was observed on the nuclear membrane and on membranes of the endoplasmic reticulum , Golgi complex , or mitochondria . RR-positive vacuoles never contained BSA-gold and BSA-gold-labeled endosomes were always devoid of RR ( Figure 2C–2F ) , confirming the specificity of our approach . BSA-gold-labeled endosomes were generally more electron-dense , often round or oval-shaped , and significantly smaller than the RR-stained vacuoles . The RR-positive structures often enclosed stacked membranes resembling plasma membrane protrusions ( Figure 2D and 2E ) . Morphologically identical RR-negative structures were also observed , sometimes next to RR-positive structures ( Figure 2E ) . They were never filled with BSA-gold ( e . g . , Figure 2E ) and are thus unlikely to be of endocytic origin . Their failure to stain with RR suggested that they are not directly connected to the cell surface , while their morphology strongly suggested a plasma membrane origin . Lack of staining may be explained by inaccessibility to the stain , however , as supported by inefficient staining of the cell surface attached to the substrate ( Figure 2A ) . Next , this approach was used to identify the membranes of HIV-1-containing compartments and of viral budding sites in MDM . Cells infected with HIV-1 Ba-L were fed with BSA-gold followed by RR staining at the peak of virus production ( Figure 3A ) . Infected cells displayed a complex surface organization similar to uninfected MDM with many seemingly intracellular vacuoles ( Figure 3B , arrows ) containing mature HIV-1 with cone-shaped cores . Importantly , the limiting membrane of many of these vacuoles and the surface of the enclosed virions were strongly RR-positive , similar to the cell surface and extracellular virions ( Figure 3B–3I ) . Identical virus-filled structures were observed in cells that had not been starved and filled with BSA-gold prior to fixation , clearly showing that this phenotype was not affected by the pre-treatment ( unpublished data ) . The RR-positive virus-harboring structures contained protrusions reminiscent of the plasma membrane ( Figure 3C ) , sometimes displayed structures resembling clathrin-coated pits ( Figure 3E and 3F ) , and were never labeled with BSA-gold , also indicating that they were cell surface–derived . Figure 3D shows a large RR-positive vacuolar structure with many protrusions and numerous mature HIV-1 particles that is directly connected to the cell surface . RR-positive early and late HIV-1 budding profiles were observed at the cell surface ( Figure 3H ) and on vacuolar structures ( Figure 3I ) , confirming that virus assembly and release did occur on plasma membrane invaginations that appeared intracellular in thin sections . Similar to the observations with uninfected MDM , a subset of virus-harboring vacuolar structures was not stained with RR , while the morphology of these structures was indistinguishable from the corresponding RR-positive compartment ( Figure 3J and 3K ) . HIV-1 budding sites were also observed on such RR-negative membranes ( Figure 3J and 3K ) . These structures were always devoid of BSA-gold and thus did not belong to the endocytic pathway . Gold-filled endosomes were often detected in the vicinity of HIV-1-containing vacuoles ( Figure 3D , 3E , and 3K ) and appeared significantly smaller than virus-filled vacuoles . This was quantified by measuring the surface area of virus-filled vacuoles and endosomes ( Table 1 ) . The vacuolar structures covered 1–3 . 1 μm2 , while endosomes had an average size of 0 . 08 μm2 and were thus ten to 40 times smaller . For a quantitative assessment of virus-specific structures , their localization in infected MDM from three different donors was attributed to one of three compartments depending on the presence or absence of RR and/or BSA-gold . Over 94% of HIV-1-specific structures were detected in compartments devoid of internalized gold . These structures were mostly stained with RR in the case of two donors ( 60% and 80% of all structures , respectively; Figure 4A , donors 1 and 3 ) , while the majority of virus accumulated in RR-negative ( and BSA-gold-negative ) structures in the case of donor 2 . Accumulation of HIV-1 in a certain compartment does not necessarily imply that budding occurs from this compartment , and the localization of viral budding sites was therefore also quantified . The absolute number of budding profiles was relatively small since 91%–97% of all HIV-1-specific structures observed corresponded to free virus particles . Importantly , quantification of budding profiles revealed that 98%–100% localized to BSA-gold-negative structures with the relative distribution between RR-positive ( 40%–60% ) and RR-negative budding profiles being similar to that observed for free viruses ( Figure 4B ) . Although the RR-negative virus-containing structures were morphologically indistinguishable from the RR-positive ones and never contained BSA-gold , they could have represented a population of endosomes that is not part of the active endocytic pathway and therefore inaccessible to the endocytic tracer . The tetraspanins CD63 , CD81 , and CD82 and the lysosomal membrane protein lamp-1 were previously shown to localize to endosomal compartments and to be incorporated into the viral envelope of macrophage-derived HIV-1 particles [14 , 27–29] . Since the presence of these membrane proteins in the HIV-1 envelope has been taken as important evidence that HIV-1 buds into endosomes in MDM [14] , we tested their localization in infected MDM and included CD44 as a plasma membrane marker [30] . As expected , CD63 was predominantly found on membranes of BSA-gold-filled endosomes ( Figure 5A ) . Significant labeling for this marker was also detected on the plasma membrane ( Figure 5B ) as reported previously [17 , 31] . CD63 labeling at the plasma membrane was uneven with areas of higher and lower density ( unpublished data ) . The tetraspanins CD81 and CD82 and the lysosomal protein lamp-1 exhibited a similar labeling pattern as CD63 , albeit with a lower labeling efficiency ( Figure S1 ) . Consistent with the cell surface localization of this marker , CD63-positive budding profiles were observed at the plasma membrane ( Figure 5C ) . Extracellular virions ( unpublished data ) , particles inside vacuolar structures , and the limiting membranes of these structures were positive for CD63 ( Figure 5D and 5E ) , and , to a lesser extent , lamp-1 , CD81 , and CD82 ( Figure S1 ) as previously reported [14] . For each of these markers , labeling of the limiting membrane was similar on all virus-containing structures , suggesting that they do not segregate into separate classes . Importantly , CD44 , which localized almost exclusively to the plasma membrane and was absent from BSA-gold-filled endosomes ( Figure 5F–5H ) , was also readily detected at budding sites ( Figure 5G ) and efficiently incorporated into HIV-1 particles ( Figure 5F and 5H ) . Quantitative analysis of the labeling densities of CD63 ( Figures 6A and S2 ) and CD44 ( Figure 6B ) revealed that they were virtually identical on the plasma membrane and on the limiting membrane of virus-filled vacuolar structures , supporting a cell surface origin of the latter . The labeling density of CD63 in uninfected and infected MDM was 2- to 3-fold higher on endosomes than at the plasma membrane , and the labeling density on virions was intermediate between the two ( Figure 6A ) . In contrast , CD44 was almost exclusively found at the plasma membrane and virus-filled vacuolar membranes and was excluded from gold-filled endosomes ( Figure 6B ) . The CD44 labeling density was identical on the plasma membrane , the limiting membranes of virus-filled vacuolar structures , and the membranes of enclosed HIV-1 particles ( Figure 6B ) , supporting the plasma membrane origin of the particle membrane . To correlate the immuno-EM data with the membrane composition of infectious HIV-1 particles derived from MDM , we immunoprecipitated MDM-derived virus with monoclonal antibodies to CD63 , CD44 , or control antibodies , respectively , as previously described [14 , 32] . Precipitation with anti-CD63 reduced infectivity by ∼40% ( Figure 6C ) , and this was similar for virus derived from MDM of different donors ( Figure S2 ) . These data are consistent with previous studies showing that CD63 is incorporated into infectious HIV-1 from MDM [14 , 16] . An even stronger titer reduction was observed when virus was precipitated with antibodies against the plasma membrane marker CD44 ( ∼80% , Figures 6C and S2 ) . Collectively , these results support the conclusion of predominant plasma membrane budding of HIV-1 in primary human macrophages .
The current view of HIV-1 morphogenesis distinguishes two fundamentally different pathways of virus release . This implies differences in targeting and membrane interaction of viral components , depending on the host cell . HIV-1 budding is readily detected by EM at the plasma membrane of T-lymphocytes , which are morphologically small and characterized by a large nucleus and a small cytoplasmic compartment . In infected macrophages , on the other hand , HIV-1 has long been reported to accumulate within large vacuoles [10–13] , recently described as being of endosomal origin [14 , 15 , 21] . This assignment was based on immunolocalization of cellular proteins in combination with the intuitively appealing concept of common use of the ESCRT-machinery for MVB formation and HIV release at the same site . Here , we report that the seemingly intracellular virus-filled vacuoles are largely continuous with the plasma membrane and conclude that HIV-1 budding occurs predominantly at the plasma membrane also in macrophages . This is mainly based on the following observations: ( i ) the frequent detection of HIV-1 budding sites at the plasma membrane , ( ii ) the predominant HIV-1 budding and accumulation in RR-positive structures , ( iii ) the virtual exclusion of budding sites and HIV-1 particles from BSA-gold-filled endosomes , ( iv ) the size of virus-filled compartments being much larger than that of endosomes , and ( v ) the composition and labeling density of cellular proteins in the viral envelope supporting its plasma membrane origin . It is important to note that the sizes of virus-filled structures in this and previous studies were very similar ( Table 1; [14 , 15] ) , implying that these compartments are the same . Thus , our experimental findings applying previously used techniques are entirely consistent with prior reports , while our interpretation—largely based on an immunomarker-independent approach—is fundamentally different . The incorporation of specific cellular membrane proteins into the HIV-1 envelope has been taken as evidence for the origin of the viral membrane [32 , 33] . This interpretation suffers from the fact that HIV-1 produced from T cells—where the virus buds through the plasma membrane—carries the same marker proteins ( reviewed in [34] ) . The distribution of cellular proteins is generally not exclusive to a single membrane system , however , and CD63 and other tetraspanins , while clearly enriched at LE , are also detected at the plasma membrane [17 , 35] . We found a 2- to 3-fold difference in CD63 density when comparing the endosomal and plasma membrane , with an uneven distribution of CD63 in different regions of the latter . This observation is consistent with two recent studies reporting the presence of tetraspanin-enriched microdomains in the plasma membrane of T cells and cell lines , from which HIV-1 was suggested to bud preferentially [31 , 36] . We observed a higher labeling density for CD63 in the HIV-1 membrane compared to the plasma membrane or membranes of virus-filled structures , consistent with HIV-1 budding from CD63-rich plasma membrane microdomains also in MDM . Plasma membrane budding was supported by the efficient labeling of HIV-1 buds and particles with anti-CD44 , which localized almost exclusively to the plasma membrane . Furthermore , infectious virus was almost quantitatively precipitated by anti-CD44 . This had also been observed in previous studies , where anti-CD44 precipitated >95% of infectious HIV-1 [16 , 37] . Thus , the collective studies on incorporation of cellular membrane proteins into HIV-1 support a plasma membrane origin of the viral envelope , independent of the infected host cell . To define the budding membrane of HIV-1 in infected macrophages in an immunomarker-independent manner , we combined two established EM methods in a novel ultrastructural approach . Endosomes were functionally labeled with BSA-gold and the plasma membrane was stained with the membrane-impermeant dye RR [23 , 24 , 38] . Importantly , RR staining and BSA-gold never colocalized to the same compartment and no RR staining of internal organelles was detected , validating our approach . HIV-1 commonly accumulated in large , seemingly intracellular RR-positive structures , which were thus defined as plasma membrane invaginations . Although budding structures were much more rare , they were also observed at the RR-stained limiting membrane of such seemingly intracellular invaginations ( and not on gold-filled endosomes ) , and the enclosed virus was stained with RR as well . The RR-positive compartment contained the majority of HIV-1 particles and budding sites , but they were also observed at RR-negative structures . We suggest that these virus-filled structures were similar , if not identical , to RR-stained invaginations and were also derived from the plasma membrane , but have not been accessible to RR for technical reasons . RR staining was poor for the substrate-adherent plasma membrane of MDM , and this would thus be expected for structures connecting to this surface as well . Notably , the size and morphology of RR-positive and RR-negative structures was identical , the latter also displayed occasional structures resembling clathrin-coated pits and membrane protrusions , and virus-filled structures were consistently CD44-positive . Furthermore , the density of both CD63 and CD44 was identical for plasma membrane and virus-filled vacuolar structures as would be expected if they represent the same compartment . Finally , labeling of the limiting membrane with antibodies against tetraspanins ( CD63 , CD81 , and CD82 ) was consistent for all virus-containing structures , indicating that these structures comprise a homogenous population . A direct connection of the virus-filled compartment to the cell surface within the plane of the section was only detected in rare cases , which may not be surprising given the complex three-dimensional architecture of the MDM plasma membrane . Images of virus-filled structures in direct membrane continuity with the cell surface have also been published previously [15] , but were interpreted as exocytic vacuoles releasing virus budded at endosomal membranes . We consider this interpretation unlikely because virus-filled structures were devoid of BSA-gold , displayed features of the plasma membrane ( see above ) , and were CD44-positive . Moreover , the large number of RR-positive structures would require a major part of the endomembrane system to be involved in exocytosis at any time , which cannot be reconciled with the main function of macrophages as phagocytic scavengers . Our conclusion of plasma membrane budding of HIV-1 in primary human macrophages is in complete agreement with a report published while this work was under review [39] . These authors based their conclusion on kinetic analyses of Gag trafficking and virus release in the presence of various inhibitors and on targeting mutants of HIV-1 Gag and suggested that the intracellular pool of HIV-1 in macrophages may be due to phagocytosis of virions after their initial release at the plasma membrane [39] . Our ultrastructural analysis of primary human macrophages incubated overnight with large amounts of HIV-1 revealed virtually no viral particles in the endosomal compartment , however , arguing against this interpretation ( unpublished data ) . The experiments presented here showed that uninfected and HIV-1-infected macrophages acquire a complex plasma membrane organization with many protrusions and tightly interdigitated cell contact zones as well as deep invaginations that on sections resemble large intracellular vacuoles . These structures enlarge the surface of the plasma membrane and may be physiologically relevant to sense the environment and to engulf particles to be destroyed . It is tempting to speculate that these highly convoluted parts of the plasma membrane may be different in nature compared to the cell surface that directly faces the growth medium . HIV-1 buds at this complex surface membrane , and infectious virions may be trapped in membrane pockets and accumulate there . Since infected macrophages are known to persist for months , these infectious virions could be released over extended periods of time , consistent with a recent report of prolonged storage of infectious HIV-1 in MDM [19] . Although this scenario is similar to the delayed exosome-release hypothesis [20] , it is likely that the regulation of virus storage and release underlie an entirely different mechanism . Most importantly , however , our study suggests a universal morphogenesis pathway for HIV-1 with virus release occurring predominantly at the plasma membrane , independent of the host cell .
All tissue culture reagents were from GIBCO BRL ( http://www . invitrogen . com ) , unless otherwise indicated . EM chemicals were from EMS ( http://www . emsdiasum . com ) and Uranyl acetate was from Fluka ( http://www . sigmaaldrich . com ) . All Epon embedding solutions , propylene oxide , and RR were from Serva ( http://www . serva . de ) . Bovine Albumine Fraction V ( Biomol , http://www . biomol . com ) coupled to 10 nm gold was prepared as described [40] . Anti-CD63 mAb 1B5 ( used for immuno-EM ) was a kind gift from M . Marsh , anti-CD63 mAb FC5 . 01 was from Zymed ( Invitrogen , http://www . invitrogen . com ) . Anti-lamp-1 mAb H4A3 was from DSHB ( http://www . uiowa . edu/∼dshbwww/info . html ) , anti-CD44 mAb F10-44-2 from Chemicon ( http://www . chemicon . com ) , and rabbit-anti-mouse IgG from ICN Biomedicals ( http://www . mpbio . com ) . Anti-CD81 mAB M38 and anti-CD82 mAB M104 were kind gifts from F . Berditchevski . Anti-nPKCθ IgG2a ( E-7 , purchased from Santa Cruz Biotechnology , http://www . scbt . com ) was used as isotype control . Polyclonal rabbit antiserum against the HIV-1 capsid protein was raised against bacterially expressed and purified protein . Cultures of primary human MDM were routinely prepared from Ficoll gradient-purified peripheral blood mononuclear cells isolated from single , healthy , HIV-1-seronegative blood donors ( DRK Blutspendezentrale , Mannheim , Germany ) by adherence and were differentiated in culture for 6–8 d as described [22] , in the presence or absence of recombinant human M-CSF ( 5 ng/mL; R & D Systems , http://www . rndsystems . com ) . Alternatively , MDM were differentiated from CD14 magnetic bead-selected monocytes from peripheral blood mononuclear cells according to the manufacturer's protocol ( Miltenyi Biotec , http://www . miltenyibiotec . com ) . Details regarding MDM cultivation , infection with the macrophage-tropic HIV-1 strains YU-2 [41] or Ba-L ( purchased from Advanced Biotechnologies , http://www . abionline . com ) , BSA-gold feeding , and RR staining are provided in Figure 3A . Primary human MDM were either left untreated or starved overnight in DMEM with 5% fetal calf serum ( DMEM/5% ) , followed by an additional 2-h starvation in serum-free DMEM prior to BSA-gold feeding . Cells were either washed three times with ice-cold 20 mM EDTA/PBS and fixed directly or incubated for 2 h with BSA-gold ( final OD520 = 10 ) in DMEM/10% containing 10% human AB-positive serum at 37 °C , conditions that have been shown to fill at least 75% of all endocytic compartments in MDM including late endosomes and lysosomes [22 , 29] . BSA-gold-filled cells were placed on ice and washed as above to remove excess BSA-gold and to partially detach the cells from the culture dish prior to fixation . All cells were fixed with 2 . 5% glutaraldehyde in 0 . 1 M ice-cold Na-Cacodylate buffer ( pH 7 . 2 ) containing 0 . 5 mg/ml RR for 1 h , during which time cells were allowed to warm up to room temperature . Cells were then washed with 0 . 1 M Na-Cacodylate buffer ( pH 7 . 2 ) , post-fixed with 2% OsO4 in the same buffer containing 0 . 5 mg/ml RR for 1 h at room temperature before routine embedding in Epon as described [42] . Cryo-sections of fixed cells were prepared as described [43] and immunolabeled as before [42 , 44] . Sections were examined with a Zeiss EM10 ( http://www . zeiss . com ) . The quantification of CD63 and CD44 labeling density was carried out and calculated essentially as described [22 , 42] . Values represent counts from at least two different grids and two independent labeling experiments per cell donor . Immunoprecipitation of HIV-1 from culture medium of HIV-1 Ba-L-infected MDM was essentially done as described [14 , 32] using anti-CD63 ( FC5 . 01 , Zymed ) , anti-CD44 ( F10-44-2 , Chemicon ) , or an IgG2a isotype control antibody ( all at 5 μg/ml ) . After precipitation , the supernatants were analyzed for remaining unprecipitated infectious virus in a standard reporter cell assay as described [45] .
The accession numbers for the entities discussed in this study are from the Swiss-Prot database ( http://www . expasy . org/sprot ) . They include Swiss-Prot entry name CD44_HUMAN , protein name CD44 antigen , gene name CD44 , LHR ( P16070 ) ; Swiss-Prot entry name CD63_HUMAN , protein name CD63 antigen , gene name CD63 , MLA1 , TSPAN30 ( P08962 ) ; Swiss-Prot entry name CD81_HUMAN , protein name CD81 antigen , gene name CD81 , TAPA1 , TSPAN28 ( P60033 ) ; Swiss-Prot entry name CD82_HUMAN , protein name CD82 antigen , gene name CD82 , KAI1 , SAR2 , TSPAN27 ( P27701 ) ; and Swiss-Prot entry name LAMP1_HUMAN , protein name lysosome-associated membrane glycoprotein 1 , gene name LAMP1 ( P11279 ) . | Macrophages are one of the major target cells for HIV-1 infection and play an important role in viral pathogenesis . Previous studies indicated that the pathway of HIV-1 particle morphogenesis is distinct in primary human macrophages , and this has been suggested to play a role in viral persistence . Early reports indicated that HIV-1 accumulates within apparently intracellular vacuolar structures , which were later identified as being of late endosomal origin . Endosomes were therefore suggested to comprise the budding and storage compartment for HIV-1 in primary human macrophages , from which infectious virus can be released in a regulated manner . In the present study , we show that HIV-1 budding occurs predominantly at the plasma membrane also in primary human macrophages . Using electron microscopy , we observed that the cell surface of macrophages displays an unexpectedly complex morphology with many protrusions and deep invaginations . HIV-1 budding occurs primarily at these invaginations that are clearly connected to the cell surface and do not belong to the endocytic compartment . Mature virus particles can remain trapped within such invaginations giving the appearance of an intracellular budding compartment . These results suggest a general pathway of HIV-1 morphogenesis with the plasma membrane as viral budding site . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"viruses",
"infectious",
"diseases",
"cell",
"biology",
"virology",
"immunology",
"homo",
"(human)"
] | 2007 | HIV-1 Buds Predominantly at the Plasma Membrane of Primary Human Macrophages |
Specialized endocrine cells produce and release steroid hormones that govern development , metabolism and reproduction . In order to synthesize steroids , all the genes in the biosynthetic pathway must be coordinately turned on in steroidogenic cells . In Drosophila , the steroid producing endocrine cells are located in the prothoracic gland ( PG ) that releases the steroid hormone ecdysone . The transcriptional regulatory network that specifies the unique PG specific expression pattern of the ecdysone biosynthetic genes remains unknown . Here , we show that two transcription factors , the POU-domain Ventral veins lacking ( Vvl ) and the nuclear receptor Knirps ( Kni ) , have essential roles in the PG during larval development . Vvl is highly expressed in the PG during embryogenesis and is enriched in the gland during larval development , suggesting that Vvl might function as a master transcriptional regulator in this tissue . Vvl and Kni bind to PG specific cis-regulatory elements that are required for expression of the ecdysone biosynthetic genes . Knock down of either vvl or kni in the PG results in a larval developmental arrest due to failure in ecdysone production . Furthermore , Vvl and Kni are also required for maintenance of TOR/S6K and prothoracicotropic hormone ( PTTH ) signaling in the PG , two major pathways that control ecdysone biosynthesis and PG cell growth . We also show that the transcriptional regulator , Molting defective ( Mld ) , controls early biosynthetic pathway steps . Our data show that Vvl and Kni directly regulate ecdysone biosynthesis by transcriptional control of biosynthetic gene expression and indirectly by affecting PTTH and TOR/S6K signaling . This provides new insight into the regulatory network of transcription factors involved in the coordinated regulation of steroidogenic cell specific transcription , and identifies a new function of Vvl and Knirps in endocrine cells during post-embryonic development .
Steroid hormones have a conserved role in the regulation of developmental transitions , growth , metabolism and reproduction in animals [1]-[3] . Specialized endocrine tissues with cell-type specific complements of enzymes that form biochemical pathways mediate the biosynthesis of steroids . In Drosophila larvae , the steroid biosynthetic enzymes are expressed in the prothoracic gland ( PG ) , the endocrine tissue of insects and the major source of the steroid hormone ecdysone . The production of ecdysone in the PG is regulated by a checkpoint control system in response to external and internal signals [2] . These checkpoints allow the endocrine system to assess growth and nutrient status before activating the biochemical pathway that increases the release of ecdysone , which triggers developmental progression . Despite the importance of the coordinated expression in endocrine cells of the steroidogenic enzymes , the PG specific transcriptional regulatory networks that underlie steroidogenic cell function remain unknown . The steroidogenic function of the PG cells is defined by the restricted expression of the genes involved in ecdysone biosynthesis that mediate the conversion of cholesterol to ecdysone . The components of the ecdysone biosynthetic pathway include the Rieske-domain protein Neverland ( Nvd ) [4] , [5] , the short-chain dehydrogenase/reductase Shroud ( Sro ) [6] and the P450 enzymes Spook ( Spo ) , Spookier ( Spok ) , Phantom ( Phm ) , Disembodied ( Dib ) and Shadow ( Sad ) [7]–[12] collectively referred to as the Halloween genes . Ecdysone produced by the PG is released into circulation and converted into the more active hormone , 20-hydroxyecdysone ( 20E ) , in peripheral tissues by the P450 enzyme , Shade ( Shd ) [13] , [14] . The cell-type specific pattern and precise dynamics of the ecdysone titers suggest a tight transcriptional regulation of the biosynthetic enzymes in the PG . This is likely orchestrated by multiple transcription factors working in a network to achieve spatial and temporal control of steroid hormone production during development . The composition of this tissue-specific transcriptional regulation remains largely elusive , although some transcription factors are known to regulate ecdysone production in the PG [15]-[18] . The nuclear receptor DHR4 functions as a repressor of ecdysone biosynthesis in the PG and responds to prothoracicotropic hormone ( PTTH ) mediated activation of the mitogen-activated protein kinase ( MAPK ) pathway [17] . Loss of βFTZ-F1 in the PG has also been associated with reduced expression of phm and dib [18] . The zing-finger protein Without children ( Woc ) is required for ecdysone biosynthesis [19] , although the pathway component regulated by Woc has not been identified . However , it is unclear if Woc , βFTZ-F1 and DHR4 bind directly to the regulatory regions that control expression of the ecdysone biosynthetic genes . In contrast , we recently showed that the transcription factor Broad ( Br ) regulates expression of the genes involved in ecdysone biosynthesis by direct binding to their promoters/enhancers [20] . Although these factors may be important for steroidogenic gene expression , other factors are likely required for the transcriptional regulatory network that defines the PG cell-specific expression of the ecdysone biosynthetic pathway components . We recently characterized cis-regulatory elements required for the expression of phm and dib in the Drosophila PG [20] , including a 69 bp promoter element located in the upstream phm region and a 86 bp region in the third intron of dib . These elements are important for the temporal up-regulation of phm and dib by Br isoform 4 ( Br-Z4 ) that increases the ecdysteroidogenic capacity of the PG and allows the production of the high-level ecdysone pulse that triggers pupariation . To further characterize the tissue-specific regulation of the ecdysone biosynthetic pathway , we analyzed PG specific regulatory elements for the presence of transcription factor binding sites . Here , we report a novel role for Ventral veins lacking ( Vvl ) and Knirps ( Kni ) in regulating ecdysteroidogenesis in Drosophila . The cis-regulatory elements responsible for PG specific expression of spok , phm and dib contain conserved Vvl and Kni binding sites . Expression of vvl is high in the PG compared to the whole animal , while kni expression is less PG-specific . Knock down of vvl and kni in the PG results in larval developmental arrest due to impaired ecdysone production . We show that Vvl and Kni specifically regulate expression of all the ecdysone biosynthetic enzymes through functionally important regulatory sites . Furthermore , we find that Molting defective ( Mld ) specifically regulates enzymes that catalyze early steps in the ecdysone biosynthetic pathway . Our study identifies Vvl as a PG cell-specific transcription factor that underlies steroidogenic cell function . We conclude that Vvl and Kni are involved in the transcriptional regulatory network of the PG that coordinates expression of biosynthetic enzymes required for ecdysone production during Drosophila development .
We analyzed the phm and dib PG specific regulatory elements for transcription factor binding sites . Our in silico search revealed conserved binding sites for the POU-domain transcription factor Vvl and the nuclear receptor Kni in the phm promoter and dib enhancer ( Fig . 1A and S1 ) . Analysis of the phm promoter identified one conserved Vvl site and four Kni sites of which three are highly conserved , indicating that they are important regulatory sites . In support of this , mutations disrupting the Vvl site and one of the conserved Kni sites eliminate PG specific GFP reporter expression [20] . In contrast , mutations in the non-conserved Kni binding sites do not reduce PG expression . The third intron dib enhancer also contains one Vvl site and two Kni sites , both in regions that have been conserved . We also identified a 300 bp PG specific promoter for spok , encoding an enzyme that acts at an early step in the ecdysone biosynthetic pathway [9] . This element located −331 to −32 bp upstream of translation start drives specific PG reporter GFP expression . This spok promoter contains three Vvl and three Kni binding sites , although these sites are less conserved compared to the Vvl and Kni sites identified in the phm and dib regulatory elements . Expression of spok has previously been reported to require Molting defective ( Mld ) , a nuclear zinc finger protein [9] . Since the DNA binding sequence motif for Mld has not yet been characterized , we were unable to examine potential Mld binding sites in the spok promoter . The observation of Vvl and Kni binding sites in the promoter/enhancer of the steroidogenic enzymes prompted us to verify if these transcription factors are expressed in the PG . We performed in situ hybridization on third instar larvae and observed an intense staining of vvl mRNA in the PG ( Fig . 1B ) . Moreover , strong embryonic vvl expression is seen in the primordium of the PG from stage 13 . Importantly , the appearance of vvl in the PG precedes that of the biosynthetic genes which are expressed by stage 15 in the PG primordium [11] , [12] . Although in situ expression of kni was undetectable in the PG of embryos , expression in the PG was observed at the L3 stage ( Fig 1B ) . We also detected expression of vvl in nurse and follicle cells of adult female ovaries ( Fig . S1 ) . Using specific antibodies , we also confirmed that Vvl and Kni are expressed in the PG and that these transcription factors localize in the nucleus ( Fig . 1C ) . Although kni expression was not detected using in situ hybridization in the embryonic PG , expression of Kni was found in the PG at the L2 stage ( Fig . S1 ) . Next , we quantified vvl and kni expression in the ring gland ( an organ with by far the most of its volume constituted by PG cells ) compared to the whole body in order to see if these transcription factors are enriched in the PG ( Fig . 1D ) . As a control , we measured phm expression , which indeed is highly expressed in the PG compared to the whole animal and mld , encoding a factor with a specific role in the PG , but with a broader expression pattern ( Fig . 1D and S1 ) [12] , [21] . Expression of vvl was highly enriched in the ring gland , like phm , while kni expression was less specific to this tissue similar to mld . Based on the potential regulatory role of Vvl and Kni , we next sought to determine if these transcription factors are required for PG expression of the genes involved in ecdysone biosynthesis . We used the PG specific phm-Gal4 ( phm> ) driver and observed that knock down of vvl in the PG using UAS-vvl-RNAi ( vvl-RNAi ) resulted in first instar ( L1 ) arrest ( Fig . 2A ) . Furthermore , RNAi mediated knock down of kni in the PG , by using phm> with a UAS-kni-RNAi ( kni-RNAi ) , led to an L1 and second instar ( L2 ) arrest phenotype . To exclude the contribution of off-target effects , we tested PG specific knock down of vvl and kni using other transgenic RNAi lines that target different regions of the vvl and kni mRNA and found that they produce similar phenotypes ( Table S1 ) . To support this , we also used the P0206-Gal4 ( P0206> ) driver that promotes weak expression in the PG cells [22] . When expression of vvl and kni was reduced using P0206> , development was arrested during later stages compared to when crossed with phm> . Knock down of mld in the PG with phm> driven UAS-mld-RNAi ( mld-RNAi ) also resulted in L1 arrest ( Fig . 2A ) consistent with mutant analysis [9] , [21] . If kni and vvl are involved in specifying the gland during embryonic development , reducing their expression may cause a lack of PG cell differentiation . We used a phm>GFP to label and examine the morphology of the PG in L1 larvae 36 hours after egg lay ( AEL ) . PG cell number and morphology of L1 larvae with reduced expression of vvl , kni or mld in the PG were indistinguishable from the phm>+ control ( Fig . 2A ) . This demonstrates that knock down of these factors does not compromise PG cell fate specification and survival . The developmental arrest indicates that loss of vvl and kni in the PG impair the cellular production of ecdysone . We therefore investigated the ecdysone levels in L1 larvae 36 hours AEL by measuring E75B mRNA expression in the whole animal , which has been used as a readout for ecdysone levels [20] , [23] . Expression of E75B was significantly reduced in mid-first instar phm>vvl-RNAi and phm>mld-RNAi larvae compared to the control ( Fig . 2B ) . This is consistent with the failure of phm>vvl-RNAi and phm>mld-RNAi larvae to molt to the L2 stage . A portion of larvae with knock down of kni in the PG undergoes the L1–L2 transition , suggesting that some of these animals can produce sufficient ecdysone for the L1–L2 molt . Consistent with this observation , knock down of kni in the PG did not lead to a significant reduction of E75B in the mid-first instar . To demonstrate that the observed phenotypes are a result of decreased ecdysone biosynthesis , we tested if ecdysone supplementation could rescue the developmental arrest . Indeed , animals with reduced expression of vvl , kni or mld in the PG were rescued by the addition of 20E to their food , showing that it is the lack of this hormone that is causing the arrest ( Fig . 2A ) . To confirm this , we measured the ecdysteroid titer , which demonstrates that larvae with reduced expression of vvl , kni or mld in the PG have lower levels of ecdysteroids by the mid-first instar compared with the control ( Fig . 2C ) . Taken together , these results indicate that the transcription factors Vvl and Kni , like Mld , are required for ecdysone biosynthesis in the endocrine cells of PG during early larval development . We next investigated if Vvl and Kni regulate the expression of the genes involved in ecdysone biosynthesis . phm>vvl-RNAi and phm>kni-RNAi larvae showed reduction in the expression of phm , dib and sad by the mid-first instar 36 hours AEL compared to the control ( Fig . 3A ) . Knock down of vvl also reduced expression of sro and spok , encoding enzymes believed to work in early steps in the pathway known as the black box [4] , [5] , [9] . However , expression of nvd , encoding a PG specific gene involved in the first step in the biosynthetic pathway , was not significantly reduced in the mid-first instar by knock down of vvl or kni in the PG . This further supports the notion that the PG is specified normally during embryogenesis . Previous studies have indicated that mld mutants have reduced ecdysone levels because of a specific lack of spok expression [9] , [21] . Our knock down results involving mld-RNAi in the PG support that Mld is required specifically for the expression of spok , but not for the later acting products of phm , dib and sad . The binding sites of these factors in the PG specific regulatory elements indicate that Vvl and Kni are involved in a transcriptional network necessary for co-expression of the biosynthetic enzymes . We therefore sought to establish if Vvl and Kni can bind directly to the PG specific regulatory elements by performing a DNA/protein binding assay . For this purpose , we performed electrophoretic mobility shift assays ( EMSAs ) with the conserved sites in the phm promoter since the functional importance of these sites has been confirmed [20] . Radiolabeled DNA oligonucleotide sequences that contained the conserved vvl or kni binding sites in the phm promoter required for PG expression ( Fig . 1A ) formed DNA/protein complexes with nuclear cell extract ( Fig . 3B and C ) . These complexes were outcompeted by unlabeled oligonucleotide sequences containing consensus vvl or kni sites and by the unlabeled phm oligonucleotides containing the vvl or kni site , but not by the unlabeled phm oligonucleotides with mutated vvl or kni binding sites or by an unspecific oligonucleotide sequence . This finding demonstrates that the vvl and kni sites are required for formation of the DNA/protein complex and supports that Vvl and Kni regulate transcription of the genes involved in ecdysone biosynthesis by direct binding to their promoters and enhancers . The data indicate that Vvl and Kni are critical for the steroidogenic activity of the PG during early post-embryonic development . Later during larval development the up-regulation of ecdysone biosynthetic genes and the growth of the gland are required to produce the high-level ecdysone pulse that triggers metamorphosis . To investigate the role of Vvl and Kni during later stages of postembryonic development , we analyzed their expression in third instar ( L3 ) larval ring glands from early ( 72 hours AEL ) , mid ( 96 hours AEL ) and late ( 120 hours AEL ) L3 larvae . In wild type larvae , expression of the steroidogenic genes showed no or little increase from the early to mid L3 , but a dramatic up-regulation in the late L3 ( Fig . 4A ) , coinciding with the high-level ecdysone peak that triggers pupariation 120 hours AEL [20] . While the expression of vvl showed only a minor increase during the L3 stage , a stronger up-regulation of kni and mld was observed . Compared to both vvl , kni and mld , expression of Br-Z4 was highly up-regulated in the late L3 consistent with its role in the temporal up-regulation of the biosynthetic genes important for the high-level ecdysone pulse 120 hours AEL that triggers pupariation [24] . Considering the tissue-specificity and that vvl expression in the PG shows little relation with the ecdysone titer , it seems likely that Vvl is important for the spatial control of ecdysone biosynthetic gene expression in the PG , but not for the temporal regulation during development . On the other hand , kni expression is less PG specific , but show more correlation with the ecdysone titer during L3 . To determine whether Vvl and Kni are only required to set the initial expression of the biosynthetic enzymes during embryonic and early larval development or also to maintain PG expression during late larval stages , we used tub-Gal80ts;phm-Gal4 ( Gal80ts;phm> ) to conditionally induce the UAS-RNAi effect . Gal80ts;phm>vvl-RNAi and Gal80ts;phm>kni-RNAi larvae develop normally at 18°C and were shifted to 29°C to induce the RNAi at different times during larval stages . Development of most Gal80ts;phm>vvl-RNAi , Gal80ts;phm>kni-RNAi and Gal80ts;phm>mld-RNAi larvae was arrested in L3 when larvae were shifted to 29°C 120 hours AEL or earlier , while larvae that were shifted 144 hours AEL or later pupariated normally ( Fig . S2 ) . Because development is slowed down at 18°C , 120 hours AEL is corresponding to the late L2 stage under these conditions [25] . Thus , inducing RNAi in late L2 or earlier causes developmental arrest in L3 , suggesting that it prevents the production of the high-level ecdysone pulse in late L3 that triggers metamorphosis . Since expression of the biosynthetic genes detected in L1 larvae with reduced expression of vvl and kni was measured on RNA extracted from whole animals ( Fig . 3A ) , and normalized to Rpl23 and Rpl32 , ubiquitously expressed ribosomal housekeeping genes , it is possible that reduced PG cell growth could be responsible for the observed decrease in biosynthetic gene expression because of a reduced PG to whole animal size ratio . To exclude this possibility and to test whether Vvl and Kni are required to maintain PG specific expression of steroidogenic genes later during development , we analyzed expression in isolated ring glands from L3 larvae . We first confirmed the efficient reduction of vvl and kni mRNA levels compared to the control ( Fig . S2 ) . Additional analysis showed that expression of all of the steroidogenic enzymes was dramatically decreased in L3 ring glands in which vvl and kni were knocked down compared to the control ( Fig . 4B ) . This demonstrates that the decreased expression of the ecdysone biosynthetic genes in vvl-RNAi and kni-RNAi animals is a consequence of a specific reduction in the transcription of steroidogenic genes , and not reduced glandular growth or a general reduction in transcription . Compared to vvl-RNAi and kni-RNAi , mld knock down had little or no influence on the transcription of the genes encoding enzymes acting in late steps in the biosynthetic pathway . However , spok and nvd levels were strongly reduced in the ring glands of mld-RNAi larvae compared to the control , suggesting that these are direct targets of mld regulation . This indicates that Mld is involved in the transcriptional regulation of the enzymes mediating early biosynthetic conversions of cholesterol . The observation that mld-RNAi also regulates nvd expression may explain why spok overexpression in the PG is insufficient to rescue mld mutants [9] . To examine the influence of vvl and kni knock down on the biosynthetic enzyme level , we measured Phm protein levels in brain-ring gland complexes ( BRGCs ) using immunoblotting analysis . Consistent with the reduced mRNA levels , these results show that Phm protein levels are dramatically reduced in vvl-RNAi and kni-RNAi larvae compared with the control ( Fig . 4C ) . To reinforce that knock down of vvl , kni or mld in the PG impairs ecdysone biosynthesis , we also measured the ecdysteroid levels in L3 larvae . Ecdysteroid levels were reduced in L3 larvae where RNAi mediated knock down of vvl , kni or mld in the PG had been induced in the L2 stage ( Fig . 4D ) . Taken together , the data suggest that the coordinated expression of steroidogenic enzymes in the PG requires Vvl and Kni function . To further corroborate our findings that Vvl and Kni are involved in co-regulating all components in the biosynthetic pathway , we examined whether supplementation of any 20E precursors to the larval growth medium was able to rescue the developmental arrest of vvl or kni RNAi larvae . When fed cholesterol , 7-dehydrocholesterol or 5β-ketodiol , most phm>vvl-RNAi and phm>kni-RNAi animals develop to small L2 larvae ( Fig . 5A and B ) . Since phm>vvl-RNAi and phm>kni-RNAi arrest in L1 and L2 without supplementation , it appears that increasing the amount of substrate for ecdysone synthesis provides some compensation , but not complete rescue , when the pathway activity is reduced . Supporting this notion , providing intermediates further downstream in the pathway gradually increased rescue of phm>vvl-RNAi and phm>kni-RNAi larvae to the L3 stage . In particular , 20E and its precursor ecdysone efficiently rescue phm>vvl-RNAi and phm>kni-RNAi larvae to the L3 stage ( Fig . 5B ) . We then tested whether increased availability of cholesterol substrate is sufficient to promote ecdysone biosynthesis . Indeed , supplementation with cholesterol increased E75E mRNA in wild type larvae and ecdysteroid levels in the control and in larvae with PG specific loss of vvl , kni or mld compared with animals grown on standard food ( Fig . S3 ) . Like rescue of the L1 arrest ( Fig . 5B ) , cholesterol also provided minor rescue of the L3 developmental arrest observed when the RNAi effect was induced in the L2 stage ( Table S2 ) . Increasing cholesterol concentrations only provides minor rescue for loss of vvl and kni . In contrast , we confirmed that it provides complete compensation for loss of Niemann-Pick type C-1a ( npc1a ) ( Fig . S3 ) , which reduces substrate delivery for ecdysone biosynthesis [26] , [27] . These results suggest that the hormone deficiency observed in vvl-RNAi and kni-RNAi larvae is a result of impaired ecdysone pathway activity and not compromised cholesterol substrate delivery , like in phm>npc1a-RNAi larvae . These findings overall indicate that silencing vvl or kni in the PG specifically inhibits synthesis of ecdysone by reducing the activity of the biosynthetic pathway . Supplying the 5β-ketodiol and 5β-ketotriol , but not cholesterol or 7-dehydrocholesterol , rescued mld-RNAi larvae ( Fig . 5B ) , consistent with Mld being required for expression of Nvd and Spok which mediate early steps in the pathway upstream of the 5β-ketodiol . We conclude , that Vvl and Kni are necessary for coordinating the tissue-specific expression of all steroidogenic genes in the endocrine cells of the PG , while Mld specifically regulates genes involved in early steps in the pathway responsible for the conversion of cholesterol to the 5β-ketodiol , an intermediate downstream of the black box reaction ( s ) . Our data demonstrate that Vvl and Kni are specifically involved in transcriptional regulation of ecdysone biosynthetic components . However , when we analyzed the morphology of PG cells with reduced expression of vvl and kni , we found a mild decrease in PG cell size ( Fig . 6A and B ) , indicating that knock down of these transcription factors also influence cellular growth . The major pathways that are thought to control PG cell growth are the PTTH and the insulin/TOR pathways [22] , [28]–[32] . Therefore , we investigated the possibility that Vvl and Kni affect PG cell growth and ecdysone synthesis indirectly by interfering with PTTH and/or insulin/TOR signaling . The neuropeptide , PTTH promotes PG growth and ecdysone synthesis through activation of its receptor Torso , a receptor tyrosine kinase ( RTK ) expressed specifically in the PG [33] . Activation of the insulin receptor ( InR ) , another RTK , in the PG also regulates cell growth and stimulates ecdysone synthesis in response to circulating insulin levels . Although crosstalk between systemic insulin mediated growth regulation and TOR signaling might occur , the TOR pathway cell-autonomously regulates growth in response to cellular nutrient levels [34] . We therefore investigated whether PTTH and insulin/TOR signaling in the gland is affected by knock down of vvl and kni . Analysis of torso transcript levels revealed that , while mld-RNAi larvae have normal torso mRNA levels , expression of the PTTH receptor is reduced in ring glands from L3 vvl-RNAi and kni-RNAi larvae ( Fig . 6C ) . Consistent with down-regulation of the PTTH receptor , we found reduced levels of phosphorylated ERK , an indicator of MAPK activity and PTTH signaling [33] , in BRGCs from vvl-RNAi and kni-RNAi larvae ( Fig . S4 ) . However , unlike the biosynthetic enzymes ( Fig . 3 ) , expression of torso was not reduced in L1 phm>vvl-RNAi larvae 36 hours AEL ( Fig . S4 ) , indicating that torso expression is initiated normally despite the loss of vvl in the PG . When examining the expression of the InR and components mediating insulin signaling , we found reduced expression in vvl-RNAi and kni-RNAi animals of 4EBP that encodes a negative growth regulator depressed by activation of the insulin pathway . Further , levels of akt , which encodes a serine/threonine kinase of the insulin signaling pathway [35] , were increased , while levels of InR were decreased in vvl-RNAi larvae . Increased insulin signaling is generally associated with decreased expression of both 4EBP and InR [36] , [37] . These results imply that loss of vvl and kni increases insulin signaling . The most likely explanation for increased insulin signaling in PG of animals with reduced vvl and kni expression is the low ecdysone levels , which cause a general increase of insulin release from the brain [29] . Thus , the disturbance of insulin signaling in the PG of vvl-RNAi and kni-RNAi animals seems unlikely to account for the PG cell growth reduction . However , we observed a strong transcriptional reduction of the S6 kinase ( S6K ) , an important positive growth regulator downstream of TOR . This suggests that the combined reduction of both PTTH/Torso and TOR/S6K signaling in the PG contributes to the negative influence of vvl-RNAi and kni-RNAi on PG cell growth and ecdysone synthesis . Why does mld knock down not affect PG cell size negatively ( Fig . 6A and B ) ? Since loss of mld does not affect torso expression ( Fig . 6C ) , it is possible that disturbance of the TOR/S6K pathway alone is insufficient to impair growth , especially if this is combined with increased insulin signaling as indicated by the decreased InR and 4EBP mRNA levels in the ring glands of mld-RNAi larvae . Finally , we investigated whether loss of vvl and kni in the PG affects cholesterol substrate delivery for ecdysone synthesis . Surprisingly , we found that , whereas the biosynthetic genes show a strong decrease , npc1a exhibits a dramatic increase in the gland of vvl-RNAi , kni-RNAi and mld-RNAi larvae ( Fig . 6D ) . This finding indicates that up-regulation of npc1a in the PG of vvl-RNAi , kni-RNAi and mld-RNAi larvae reflect a compensatory feedback regulation to maintain cholesterol homeostasis and/or increase substrate delivery to promote steroidogenesis . Down-regulation of biosynthetic activity in vvl-RNAi , kni-RNAi and mld-RNAi larvae reduces cholesterol flux through the ecdysone pathway and may lead to intracellular redistribution of cholesterol to maintain homeostasis through feedback regulation . We therefore explored the possibility that npc1a , which is required for normal cholesterol distribution and availability for steroid synthesis , is controlled by feedback regulation of cholesterol . Expression of npc1a is repressed by cholesterol in wild type larvae ( Fig . S4 ) , indicating that npc1a is feedback regulated . Recently , we showed that ecdysone biosynthesis is controlled by feedback circuits in the PG [20] . We therefore also examined whether ecdysone signaling in the gland is affected by knock down vvl , kni or mld . To test this , we measured mRNA levels of the ecdysone receptor ( EcR ) ring glands isolated from L3 larvae where vvl-RNAi , kni-RNAi or mld-RNAi had been induced in the PG during L2 . Transcript levels of EcR were not affected in ring glands from vvl-RNAi and kni-RNAi larvae ( Fig . S4 ) , indicating that the responsiveness of the PG to ecdysone is not reduced . Taken together , these results suggest alterations of cholesterol uptake and trafficking in the PG when flow through the biosynthetic pathway is impaired .
Drosophila developmental progression is dictated by tightly regulated ecdysone pulses released from the PG . Like any cell specialized for steroid biosynthesis , the PG expresses a set of enzymes that mediate steps in the conversion of cholesterol into steroids . The tissue-specific expression of these enzymes is key to the specialization of the cells that endows the PG with the competence to produce ecdysone . The transcriptional control mechanism underlying such regulation is likely orchestrated by a regulatory network of transcription factors . Here , we identify two transcription factors Vvl and Kni that are required for the expression of the biosynthetic enzymes in the ecdysone producing PG cells . Vvl is a POU domain transcription factor which has multiple important functions during Drosophila development . Mutations in vvl cause embryonic lethality with defects in the development of the trachea and the nervous system [38]–[41] . Moreover , Vvl is required for wing vein development and is involved in innate immunity by regulation of the expression of antimicrobial peptides [42] , [43] . We show that Vvl is expressed in the PG during late embryogenesis and in the larval stages . One important characteristic of Vvl is that it maintains its own expression by autoregulation [44] . Once activated , Vvl maintains its expression and likely also the expression of the ecdysone biosynthetic genes in the PG . Knock down of vvl in the PG reduces the expression of all genes in the biosynthetic pathway , showing that Vvl is required for maintaining expression of all pathway components . Together with the high expression of Vvl in the gland , this suggests that Vvl is a master transcriptional regulator involved in specifying the genetic program that dictates PG cell identity including its tissue-specific expression of steroidogenic enzymes . It is interesting to note that human chromosome 6 deletions that affect POU3F2 , a homolog of Vvl , have been associated with hypogonadotropic hypogonadism and adrenal insufficiency [45] , [46] , making it possible that Vvl is a conserved regulator of steroid biosynthesis . The gap gene kni is known for its role in embryonic segmentation patterning and development of the trachea and wing vein [47]–[51] similar to vvl . Kni is a nuclear receptor with a zinc-finger motif that is unlikely to be ligand activated since it lacks a ligand-binding domain . Our data show that Kni is required for expression of the genes involved in ecdysone biosynthesis in the PG , suggesting that Kni functions as an activator in this situation . Although Kni is generally considered a short-range repressor [52] , it is required to activate hairy expression in stripe 6 during embryogenesis [53] . Thus , Kni may act either as a repressor or as an activator in a context-dependent manner . In mammals , nuclear receptors are also key regulators of steroidogenic target genes encoding P450 enzymes [54]–[56] . Although Vvl and Kni specifically control genes in the steroidogenic pathway , other targets of these factors could also be important for ecdysone synthesis in the PG . During development the continuous growth of the PG cells and endoreplication of DNA is important to scale its hormone production to the capacity required for developmental progression . We found that both vvl-RNAi and kni-RNAi larvae have mildly reduced PG nuclei and cell size , which is likely to contribute to the reduced ecdysone levels in these animals . Kni has been shown to suppress endoreplication activity in the gut by regulating cell cycle genes [48] . This is in contrast to our observation indicating that loss of kni results in a reduction in the nuclei size , and hence , reduced polyploidy of the PG cells . Instead our results indicate that loss of vvl and kni reduces activity of PTTH/Torso and TOR/S6K signaling , two major pathways that promote growth and stimulate ecdysone biosynthesis [30] , [31] , [33] , [57] . However , loss of vvl and kni had no effect on torso expression in the mid-first instar . This indicates that these factors are not required for the initial setting of torso expression , but for the maintenance of high torso expression during development . In tracheal cells , Vvl is required to maintain expression of the RTK breathless , but not for activating its initial expression [42] , [58] . It is unclear how the transcription of the biosynthetic enzymes fluctuates during the low level ecdysone peaks in L1 and L2 , before the induction of the steroidogenic pathway by PTTH stimulation [17] . Unlike PTTH/Torso , Vvl and Kni are required in the PG during L1 and L2 for the transition to the L3 stage , which suggests that Vvl and Kni are important for the proper transcription of the biosynthetic enzymes throughout larval development . Altogether , these data suggest that in addition to being required to initiate and maintain expression of the biosynthetic enzymes , Vvl and Kni play an indirect role important for ecdysone production by enabling PG cells to be competent to respond to PTTH and by regulating the TOR/S6K pathway . In contrast , Vvl and Kni are not required for normal expression of EcR in the gland , indicating that feedback regulation of ecdysone biosynthesis is not influenced by knock down of these factors [20] . In contrast to the transcription factor Br-Z4 involved in positive feedback regulation , which is strongly induced in the PG during late L3 to up-regulate expression of the biosynthetic pathway components , PG expression of vvl shows little relation with the high-level ecdysone peak that triggers pupariation . Taken together these data suggest that Vvl is required for maintaining PG specific expression ( i . e . spatial control ) , while temporal regulation during development is controlled by other factors such as Br-Z4 . Furthermore , our results confirm that Mld is required for PG expression of spok [9] , but we also found that it controls Nvd , an enzyme that acts upstream of Spok in the biosynthetic pathway [4] . Thus , our data suggest that Mld is a specific regulator of the two early enzymes Nvd and Spok , while its function is not important for biosynthetic reactions that are downstream of Spok and the black box reaction ( s ) and the responsiveness of the PG to PTTH . Our data show that Vvl and Kni are required in the PG during post-embryonic development to maintain PG specific expression of the ecdysone biosynthetic genes . During embryogenesis , vvl expression appears in the PG primordium by stage 13 , after the embryonic ecdysone pulse ( stage 8–12 [12] ) that is required for morphogenesis and differentiation of the embryo . During early embryonic development where the PG primordium is not yet formed , the spatial expression patterns of kni and vvl ( Fig . 1B ) are different from the biosynthetic genes essential for the embryonic ecdysone pulse [8] , [9] , [11] , [12] , [59] . This suggests that Vvl and Kni regulate the biosynthetic genes in the PG , but not during early embryonic development . Consistent with this notion , vvl and kni mutants differentiate the embryonic cuticle [42] , [60] , unlike the ecdysone deficient mutants that are unable to produce the embryonic ecdysone peak [11] , [59] . In adult females , the ovaries are believed to be the source of ecdysone consistent with expression of the ecdysone biosynthetic genes in the nurse and/or follicle cells [11]–[13] . In adult females , we find that vvl is expressed in both nurse and follicle cells , suggesting that Vvl may be involved in regulating expression of the ecdysone biosynthetic genes in the adult stage . Interestingly , we observed that loss of vvl , kni or mld results in dramatic increase of npc1a expression in the PG . Npc1a is highly expressed in the PG where it is required for uptake and intracellular trafficking of cholesterol for steroidogeneis [26] . Larvae with loss of npc1a exhibit a punctuate pattern of sterol accumulation in the PG cells , indicating defects in cholesterol transport within the cells . Normally cholesterol is taken up as low density lipoproteins ( LDLs ) and trafficked within endosomes to the lysosomes where hydrolysis releases free cholesterol that is delivered to the plasma membrane and endoplasmic reticulum ( ER ) [61] where the first step in the conversion of cholesterol to ecdysone likely takes place . Why is npc1a up-regulated in the PG when ecdysone synthesis and pathway activity is impaired ? It seems unlikely that Vvl , Kni and Mld are all involved in repression of npc1a . The block of flux through the biosynthetic pathway in the PG of vvl-RNAi , kni-RNAi and mld-RNAi animals may change intracellular cholesterol pools in the gland and affect feedback regulation to maintain cholesterol homeostasis . Our results indicate that npc1a is regulated by cholesterol suggesting that the up-regulation of npc1a may be part of a feedback regulatory response to changes in cellular cholesterol levels . This may indicate a compensatory mechanism to redistribute cholesterol by increasing storage of cholesterol esters and/or efflux to reduce free cholesterol levels when ecdysone biosynthesis is blocked . Moreover , npc1a is regulated by Br [27] , a factor induced by EcR in the PG [20] , implying that ncp1a may also be regulated by ecdysone feedback . Our study shows that cholesterol availability is an important parameter for ecdysone biosynthesis . Interactions between cholesterol and ecdysone feedback mechanisms may therefore be important for coordinating the supply cholesterol with the rate of steroidogenesis . A key aspect of steroidogenesis is regulating the tissue-specific expression of the biosynthetic enzymes . We have shown here that the transcription factors , Vvl and Kni , are required for the coordinated expression of ecdysone biosynthetic genes in the PG . The transcriptional activation by Vvl and Kni is likely mediated by direct binding to cis-regulatory elements responsible for PG specific expression . This identifies an important new role for Vvl and Kni during post-embryonic development in the gene regulatory network of the steroid hormone producing cells in Drosophila .
The following Drosophila strains were used in this study: w1118 , UAS-vvl-RNAi ( #110723 ) , UAS-kni-RNAi ( #2980 ) , UAS-mld-RNAi ( #101867 ) and UAS-npc1a-RNAi ( #105405 ) from the Vienna Drosophila RNAi Center ( VDRC ) ; UAS-vvl-RNAi ( #26228 ) , UAS-kni-RNAi ( #34705 ) , tub-Gal80ts and UAS-CD8-GFP ( UAS-GFP ) from the Bloomington Drosophila Stock Center ( BDSC ) ; phm22-Gal4 ( phm-Gal4 ) [9] and P0206-Gal4 [29] . A transgenic line phm-291-4B ( phm-GFP ) with a 69 bp phm promoter in a pH-stinger GFP reporter vector generated in [20] was used to collect ring glands by dissection for analyzing the development expression profile in the gland . Flies were raised on standard cornmeal food under a 12∶12 hour light:dark cycle . For experiments involving staged or timed larvae , flies were allowed to lay eggs at 25°C for 2–4 hours on apple juice agar plates supplemented with yeast paste in a humidified chamber . After 24 hours , 25 L1 larvae were collected and transferred to vials containing standard food . For experiments using tub-Gal80ts , eggs deposited at 25°C were immediately transferred to 18°C and 25 larvae were transferred to vials containing food 48 hours later . Images of phenotypes were captured with an Olympus SZX7 camera and analyzed using AxioVision software ( Zeiss ) . Characterization of the PG-specific spok element was done as described [20] by generating transgenic animals with constructions of 5′-UTR spok elements in a pH-stinger GFP reporter vector . EMSA was carried out as previously described [20] . DNA oligonucleotide sequences ( Table S3 ) were designed to cover Vvl and Kni binding sites in the phm promoter based on in silico analysis using Transfac and Jaspar databases . Oligos containing Vvl ( Vvl-wt ) or Kni ( Kni-wt ) consensus binding sites and oligos with mutations that disrupt the Vvl ( Vvl-mut ) or Kni ( Kni-mut ) binding sites were adapted from [62] , [63] . The complementary oligonucleotides were annealed and labeled at the 5′-end labeling by [γ32P]ATP ( Perkin Elmer ) using T4 polynucleotide kinase ( Fermentas ) and purified using Microspin G-25 columns ( GE Healthcare ) . The EMSA reaction was performed on ice by mixing Drosophila S2 cell nuclear extracts ( Active Motif ) , dialysis buffer ( 25 mM Hepes pH 7 . 6 , 40 nM KCl , 0 . 1 mM EDTA , 10% glycerol ) , gelshift buffer ( 25 mM Tris-HCl pH 7 . 5 , 5 mM MgCl2 , 60 mM KCl , 0 . 5 mM EDTA , 5% Ficoll 400 , 2 . 5% glycerol , 1 mM DTT and protease inhibitors ) and poly ( dI-dC ) ( Invitrogen ) . The reaction mixture was supplemented and incubated with 25-50-fold molar excess of unlabeled competitor nucleotides before adding the radiolabeled probe . After incubation the mixture was supplemented with gelshift loading buffer and run on a 5% non-denaturing polyacrylamide gel and dried on a slab gel dryer ( Savant ) followed by exposure onto a phosporimager screen . The image was acquired using a Storm 840 scanner ( Molecular Dynamics ) and processed with ImageQuant software version 5 . 2 . Tissue dissections were performed in PBS followed by fixation in 4% formaldehyde for 20 minutes at room temperature . For this study , the following primary antibodies were: mouse anti-GFP 1∶200 ( Clontech , #632380 ) ; rabbit anti-Phm 1∶200 [18]; rat anti-Kni , 1∶1000 [64] and rat anti-Vvl 1∶1000 [65] . Tissues were incubated over night with primary antibodies at 4°C . Fluorescent conjugated secondary antibodies used were goat anti-mouse Alexa Fluor 488 ( A11001 , Invitrogen ) , goat anti-rabbit Alexa Fluor 555 ( A21429 , Invitrogen ) and goat anti-rat Alexa Fluor 555 ( A21434 , Invitrogen ) . Secondary antibodies were diluted 1∶200 and incubated for two hours at room temperature . DAPI was used in 1∶500 for nuclei staining . Confocal images were captured using Zeiss LSM 710 laser scanning microscope and processed using ImageJ ( NIH ) . Images of mid-first instar PG morphology were obtained by confocal imaging of live L1 larvae ( 36 hours AEL ) mounted in 80% glycerol . Preparation and synthesis of 3β , 14α-Dihydroxy-5β-cholest-7-en-6-one ( 5β-ketodiol ) and 3β , 14α , 25-Trihydroxy-5β-cholest-7-en-6-one ( 5β-ketotriol ) were previously described [8] . For the steroid feeding rescue experiment , 30 mg of dry yeast was mixed with 57 µl H2O and 3 µl ethanol or supplemented with 3 µl of the following sterols dissolved in ethanol: 20E ( Sigma; 450 µg ) , ecdysone ( Sigma; 100 µg ) , cholesterol ( Sigma; 45 µg ) , 7-dehydrocholesterol ( Sigma; 200 µg ) , 5β-ketodiol ( 450 µg ) , or 5β-ketotriol ( 280 µg ) . Thirty larvae were transferred to the yeast paste on an apple juice agar plate and allowed to develop in a humid chamber at 25°C . The phenotype of the larvae was scored at day 5 prior to pupariation of w1118 control for rescue to the L3 stage . For other experiments with cholesterol supplementation of the food , standard cornmeal was supplied with cholesterol ( Sigma ) dissolved in ethanol to a final concentration of 40 µg/ml . Digoxigenin ( DIG ) -labeled antisense RNA probes were synthesized using DIG RNA labeling mix ( Roche ) and T3 ( Fermentas ) , T7 ( Fermentas ) or SP6 ( Roche ) RNA polymerase according to the manufacturer's instructions . For the kni probe , an EST clone GH19318 [66] was used as a templates . For the vvl probe , a portion of vvl gene was amplified by PCR with cDNA derived from w1118 larvae and the following primers: vvl_PA_CDS_F ( 5′-ATGGCCGCGACCTCGTACATGAC-3′ ) and vvl_PA_CDS_R ( 5′-CTAGTGGGCCGCCAACTGATGC-3′ ) . For the mld probe , a portion of mld gene was amplified by PCR with the plasmid mld-pUAST [21]; a gift from S . M . Cohen and the following primers: mld_CDS_1_F ( 5′-ATGAGTGCCAACCGAAGAAGACG-3′ ) and mld_CDS_1_R ( 5′-CATCTGAGATTGGTCATGAGATTGTACTTGAGG-3′ ) . PCR products containing the vvl and mld fragments were subcloned into SmaI-digested pBluescript II SK ( - ) and pCRII-Blunt-TOPO ( Invitrogen ) , respectively , and then used as the templates for synthesizing RNA probes . Fixation , hybridization and detection were performed as previously described [8] , [67] . For gene expression experiments using the whole animals , 30 L1 larvae or 4 L3 larvae were used for each replicate . For analysis of ring gland expression , 10–15 ring glands were dissected in PBS and directly transferred to RNA lysis buffer . RNA was extracted using the RNeasy mini kit ( Qiagen ) and DNase treated to avoid genomic DNA contamination according to the manufacturer's instructions . RNA was quantified using a NanoDrop ( Thermo Scientific ) and the integrity was assessed using agarose gel electrophoresis . Total RNA was used for cDNA synthesis with the SuperScript III First-Strand Synthesis kit ( Invitrogen ) . Primers were designed using the Primer3 software [68] ( Table S4 ) . Relative gene expression was analyzed using a Mx3000P qPCR System ( Agilent Technologies ) with the QuantiTect SYBR Green PCR Kit ( Qiagen ) according to the manufacturer's instructions as described [10] , [33] , [69] . All reactions were subjected to 95°C for 10 min , followed by 45 cycles of 95°C for 15 sec , 60°C for 15 sec and 72°C for 15 sec . Dissociation curve analysis was applied to all reactions to ensure the presence of single specific PCR product . Non-reverse transcribed template controls and non-template controls were included to check for background and potential genomic contamination . No product was observed in these reactions . Efficiencies were calculated for each primer pair from standard curves generated from serial dilutions of a mix of cDNA from all control samples . PCR efficiencies were always close to 100% , which was therefore used as the standard in all calculations . Expression of target genes was normalized to reference gene , Rpl23 and Rpl32 . We confirmed that these reference [32] , [33] , [70]–[74] are stably expressed across tissues and experimental conditions , by comparing Rpl23 and Rpl32 mRNA levels in cDNA synthesized from equal amounts of RNA extracted from different tissues and developmental stages ( Fig . S2 ) . Reference gene stability determined using qBASE Plus ( Biogazelle NV , Zwijnaarde , Belgium ) was within the recommended limits ( M = 0 . 274 and CV = 0 . 095 ) . For definition of these stability factors see [75] . For ecdysteroid measurements , ecdysteroids were extracted from whole animals as described [24] . Briefly , whole larvae were rinsed in water and stored at −80°C . Samples were homogenized in 0 . 5 ml methanol and the supernatant was collected following centrifugation at 14 , 000 g . The remaining tissue was re-extracted first in 0 . 5 ml methanol then in 0 . 5 ml ethanol . The pooled supernatants were evaporated using a SpeedVac and redissovled in ELISA buffer ( 1 M phosphate solution , 1% BSA , 4 M sodium chloride and 10 mM EDTA ) . ELISA was performed according to the manufacturer's instructions using a commercial ELISA kit ( ACE Enzyme Immunoassay , Cayman Chemical ) that detects ecdysone and 20-hydroxyecdysone with the same affinity [76] . Standard curves were generated using 20E ( Sigma ) . Absorbance was measured at 405 nm on a plate reader , ELx80 ( BioTek ) using the Gen5 software ( BioTek ) . Four brain-ring gland complexes were dissected in cold PBS and transferred to 20 µl Laemmli Sample Buffer ( Bio-Rad ) supplemented with 2-mercaptoethanol . Samples were boiled for 5 minutes , centrifuged at 14 , 000 g and 10 µl supernatant were loaded on a 4–20% polyacrylamide gradient gel ( Bio-Rad ) followed by transfer onto a PVDF membrane ( Millipore ) . Primary antibodies used were; mouse anti-α-tubulin , 1∶5 , 000 ( T9026 , Sigma Aldrich ) , rabbit anti-Phm , 1∶1 , 000 [18] and rabbit anti-phospho-ERK , 1∶1 , 000 ( 9101 , Cell Signaling Technology ) . Secondary antibodies were goat anti-mouse IRDye 680RD , 1∶10 , 000 ( 926-68070 , LI-COR ) and goat anti-rabbit IRDye 800CW , 1∶10 , 000 ( 926–32211 , LI-COR ) . The blot was scanned on an Odessey Fc ( LI-COR ) and the software , Image Studio for Odessey Fc , was used for image processing and protein quantification . The statistical differences between data sets were calculated using two-tailed Student's t-test and error bars represent standard error of the mean ( s . e . m . ) . | Steroid hormones play important roles in physiology and disease . These hormones are molecules produced and secreted by endocrine cells in the body and control sexual maturation , metabolism and reproduction . We found transcriptional regulators that underlie the specialized function of endocrine steroid-producing cells . In the steroid-producing cells of the fruit fly Drosophila , Ventral veins lacking ( Vvl ) and Knirps ( Kni ) turn on all the genes required for steroid production . When Vvl or Kni were inactivated in the cells where the hormone is made , the genes involved in steroid production were not activated . Because of the reduced steroid production , the juvenile larvae failed to develop and undergo maturation to adulthood . Inactivation of Vvl and Kni also reduces endocrine cell growth by disturbing their response to growth promoting signals . Genetic variations in humans with the loss of a homolog of Vvl have been associated with disorders caused by insufficient steroid production . Together with the fact that Vvl is highly expressed in the steroid-producing cells of Drosophila , this suggests that Vvl may be a conserved master regulator of steroid production . Our findings provide insight into the network of factors that control endocrine cell function and steroid hormone levels that could have implication for human diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"gene",
"function",
"animal",
"models",
"signal",
"transduction",
"developmental",
"biology",
"model",
"organisms",
"animal",
"genetics",
"organism",
"development",
"cell",
"biology",
"cell",
"growth",
"gene",
"expression",
"genetics",
"biology",
"and",
"life",
"sciences",
"cell",
"processes",
"molecular",
"genetics",
"molecular",
"cell",
"biology",
"morphogenesis",
"pattern",
"formation",
"research",
"and",
"analysis",
"methods"
] | 2014 | Transcriptional Control of Steroid Biosynthesis Genes in the Drosophila Prothoracic Gland by Ventral Veins Lacking and Knirps |
Cycles involving covalent modification of proteins are key components of the intracellular signaling machinery . Each cycle is comprised of two interconvertable forms of a particular protein . A classic signaling pathway is structured by a chain or cascade of basic cycle units in such a way that the activated protein in one cycle promotes the activation of the next protein in the chain , and so on . Starting from a mechanistic kinetic description and using a careful perturbation analysis , we have derived , to our knowledge for the first time , a consistent approximation of the chain with one variable per cycle . The model we derive is distinct from the one that has been in use in the literature for several years , which is a phenomenological extension of the Goldbeter-Koshland biochemical switch . Even though much has been done regarding the mathematical modeling of these systems , our contribution fills a gap between existing models and , in doing so , we have unveiled critical new properties of this type of signaling cascades . A key feature of our new model is that a negative feedback emerges naturally , exerted between each cycle and its predecessor . Due to this negative feedback , the system displays damped temporal oscillations under constant stimulation and , most important , propagates perturbations both forwards and backwards . This last attribute challenges the widespread notion of unidirectionality in signaling cascades . Concrete examples of applications to MAPK cascades are discussed . All these properties are shared by the complete mechanistic description and our simplified model , but not by previously derived phenomenological models of signaling cascades .
Covalent modification cycles are one of the major intracellular signaling mechanisms , both in prokaryotic and eukaryotic organisms [1] . Complex signaling occurs through networks of signaling pathways made up of chains or cascades of such cycles , in which the activated protein in one cycle promotes the activation of the protein in the next link of the chain . In this way , an input signal injected at one end of the pathway can propagate traveling through its building-blocks to elicit one or more effects at a downstream location . Examples of covalent modification are methylation-demethylation , activation-inactivation of GTP-binding proteins and , probably the most studied process , phosphorylation-dephosphorylation ( PD ) [1] , [2] . In such cycles , a signaling protein is activated by the addition of a chemical group and inactivated by its removal . This protein is modified in turn by two opposing enzymes , such as a kinase and a phosphatase for PD cycles . In the absence of external stimulation , a cycle exists in a steady state in which the activation and inactivation reactions are balanced . External stimuli that produces a change in the activity of the converting enzymes , shifts the activation state of the target protein , creating a departure from steady state which can propagate through the cascade . The advantages of these cascades in signal transduction are multiple and the conservation of their basic structure throughout evolution , suggests their usefulness . A reaction cascade may amplify a weak signal , it may accelerate the speed of signaling , can steepen the profile of a graded input as it is propagated , filter out noise in signal reception , introduce time delay , and allow alternative entry points for differential regulation [3]–[5] . Intracellular signaling through cascades of biochemical reactions has been the subject of a great number of studies ( e . g . , [2] , [6] for reviews ) . Theoretical investigations have been motivated by the increased need for developing an abstract framework to understand the vast amounts of experimental data now available . This whole field of research is further motivated by the hope of characterizing pathways that are deregulated in diseases such as cancer and to define targets to combat these diseases [7] . Since the stimuli a cell receives are varied and complex , cascades do not operate in isolation , but rather the integration of stimuli depends on crosstalk between pathways . Another crucial property of signaling cascades is their ability to integrate information by transmitting the effects downstream and also feedback upstream . In spite of a few decades of intense work on signaling cascades , no models have ever been built that exhibit crosstalk with backwards and forwards transmission of a lateral input from another cascade , except when ad hoc feedback is explicitly added to the cascade model . Our model , built from first principles , naturally exhibits these characteristics and therefore inspires novel interpretations of experimental data . A well studied example of a cascade of activation-inactivation cycles is the cascade of protein kinases . In this case , the basic signaling unit is a PD cycle , whose activating kinase is the phosphorylated protein of the previous cycle . Many proteins contain several phosphorylation sites , allowing for great versatility of regulation . Such is the case , for example , for the mitogen-activated protein kinase ( MAPK ) cascade , which is widely involved in eukaryotic signal transduction [3] , [8]–[10] . For the sake of simplicity , in this article we will mostly consider cascades composed of simple , 2-state activation-inactivation cycles . However , the equations corresponding to the MAPK cascade are also derived and some of their properties compared with those of the simpler cascades . Even though our results are valid in general , for covalent modification cycles , we will employ the nomenclature associated with PD cycles , i . e . the converting enzymes will be referred to as kinase/phosphatase . The focus of our study is to refine the mathematical modeling of cascades of covalent modification cycles , such us the one depicted in Figure 1 . Several mathematical descriptions have been developed to describe such cascades using ordinary differential equations . Typically , those descriptions are built up starting with a model for a single cycle , which is then phenomenologically incorporated into a cascade of cycles . A well known model for describing the single cycle was introduced by the pioneering work of Goldbeter and Koshland ( GK ) [11] . The GK model considers the concentration of the target protein to be in large excess over those of the converting enzymes , thereby reducing the description to a single equation per cycle . The model obtained in this way was then phenomenologically extended to a cascade of individual GK cycles . Here , by the designation “phenomenological” we mean that , in the cascade , the forward coupling between the GK cycles is chosen as simply as possible , but not strictly deduced from first principles . This phenomenological framework extension of the GK model will be denoted as the GK-like model . The GK-like model has been used by several authors to describe the dynamics of signal transduction [9] , [12]–[16] . For particular limiting cases , the GK-like model can be simplified further , which results in a model where the inter-converting reactions follow linear rate laws with first-order rate constants . This description was studied in several key papers [17]–[19] , and we will refer to it as the linear rates model . The concept of a “cascade” in the study of transduction pathways is appealing because of its modular structure . What is especially appealing is the possibility of defining the cascade state by only one variable per module . As mentioned above , since the building blocks of the GK-like model are the well-studied GK cycles , they involve only one equation per cycle . A different approach however , is to deal with the dynamics of the cascade of Figure 1 by considering the complete set of biochemical reactions and by writing the corresponding equations without any upfront approximations . This was accomplished , for example , for the case of the MAPK cascade [8] . We will refer to this approach as the mechanistic model . For the purposes of this paper , we will consider that the mechanistic model represents a complete description of the system under study ( event though we recognize that , in reality , it is not a hypothesis-free model ) . In this article , starting from the mechanistic description of a cascade composed of an arbitrary number of cycles , we derive a consistent approximation under which the cascade is described with one variable per cycle . It turns out that in this derivation , referred to as a reduced mechanistic description , the phenomenological GK-like model is not recovered . At first sight , our new approximation differs slightly from the previously derived description for signaling cascades . However , it involves qualitatively different dynamics from the GK-like model , yet it is in very good agreement with the complete mechanistic description when the approximation conditions are fulfilled . The main difference between our simplified mechanistic description and the phenomenological one is the appearance of an intrinsic feedback from each unit to the preceding one , caused by the fact that in each cycle there is sequestration of part of the activated protein of the previous step . The new description of the cascade predicts the existence of damped oscillations along the chain , a phenomenon that cannot be observed using the previous phenomenological description . Interestingly , a corollary of our model is that if a particular unit in the middle of the chain receives an input–a common event , given the high degree of crosstalk between signaling pathways–then our reduced mechanistic description predicts that this perturbation is able to travel both forwards and backwards . This “bicistronic” propagation , which may be critical for effective eukaryotic signaling , is not possible within the GK-like description either . Our model provides a suitable framework for future experiments that investigate crosstalk and bicistronic propagation of signals .
In Text S1 we derive in detail the new class of model equations obtained as an approximation of the mechanistic model . The goal of our approach is to reduce the number of variables in the complete system by bringing into play hypothesis that allow us to use the quasi-steady state approximation . Three key dimensionless parameters are defined to facilitate the analysis: ( 5 ) εi and ηi are ratios of total amounts of proteins . εi is the ratio of total phosphatase over total targeted protein . ηi is defined as the total targeted protein in one cycle over the corresponding amount in the next cycle in the cascade , or , equivalently , the ratio of total kinase over total targeted protein . The parameter µi is the ratio of the kinetic rates of product formation in both the activation and the inactivation reactions ( see reactions in Equation 1 ) . Using a standard singular perturbation analysis , we have found that the state of each biochemical cycle can be described by a single variable defined as xi = yi*+ci+1 , which is the natural slow variable describing the total kinase i available at a given time for the phosphorylation in cycle i+1 . This reduction is only valid if the total phosphatase in the cycle is much lower than the total targeted protein , i . e . , in the limit εi«1 . The other parameters must satisfy µi ηi∼εi . The dynamics of xi is described by the differential equation: ( 6 ) with the following conservation equation from which yi has to be extracted: ( 7 ) x0 = S is the normalized input signal and yn+1 = 0 . In Equation 6 , Vi = ( k′i µi ηi ) / ( ε k′ ) and V′i = ( εi k′i ) / ( ε k′ ) , where ε k′ is a typical number representing the set of εi k′i ( i = 1 , … , n ) , e . g . the arithmetical or the geometrical average over this set . In the conservation equation ( Equation 7 ) , the notation O ( εi ) is just a reminder that this equation is written in the lowest order in εi , as is also the case for the differential equation for xi . In Text S1 , we discuss an improvement of this conservation equation which takes into account the first correction in εi . Although this extension does not alter the new properties discussed below , its numerical integration is easy and it increases the accuracy of the approximation . The reduced system given by Equations 6–7 seems to be , in principle , equivalent to the GK-like model given by Equation 4 . However , two main features make it significantly different . First , in our novel system , termed the reduced mechanistic model , the conservation equation depends on the variable of the previous cycle . Second and more interesting , the denominator of the negative term in Equation 6 is now a function of the next variable yi+1 , in contrast to the GK-like model . This function has the appearance of an effective Michaelis-Menten coefficient K′eff , i = K′i ( 1+yi+1/Ki+1 ) , which is a typical way to indicate competitive inhibition in enzyme kinetics [22] . In the context of activation-inactivation cycles , a similar type of equation was obtained by Salazar and Höfer in the systematic study of a single cycle taking into account the competition between kinase and phosphatase to bind the same target protein [23] . In that case , an effective Michaelis-Menten coefficient appears also in the negative term of Equation 4 , but with the form K′eff , i = ( 1+yi/Ki ) . In our study , however , the competition is induced by the next substrate yi+1 , and this precisely describes a negative feedback from cycle i+1 on cycle i: the higher the level of xi+1 , the smaller yi+1 and , therefore , the larger the value of the negative term in Equation 6 . This modified denominator reflects the influence of the downstream step on the state variables of one given cycle . It is not a detail of the formalism . It has consequences upon the dynamics and on the properties of the signaling pathway , as it will be demonstrated in the following sections . Moreover , we will see that , since our system arises from a controlled approximation of the mechanistic model , the dynamics of both models can be made comparable . In the limit ηi∼εi«1 , one retrieves the simple conservation law xi+yi≈1 . However , we note that even in that limit and due to K′eff , i , our resulting system is not equivalent to the GK-like model . Notice that ηi∼εi«1 is the closest we can be to the hypothesis behind the GK-like model , where it is considered that the concentration of the targeted protein is in large excess over those of the converting enzymes . In our description , the converting enzymes for unit i are E′iT ( phosphatase ) and Yi−1 , T ( kinase ) . Taking the limit ηi«1 together with the fact that the targeted protein of each cycle is the activating protein of the next one , results in increasing protein concentrations as the cascade proceeds . Even though this is not the usual condition in signaling cascades , examples could arise where this limit is suitable . As a possible relevant example , the concentrations reported for the MAPK cascade go from nM in the first unit to µM in the second and third ones [8] . In addition to the limit ηi∼εi«1 , our perturbation scheme encompasses situations where the total protein does not necessarily increase along the cascade . We then allow ηi∼1 , for all or for some index i , as long as µi ηi∼εi , which results in the limit µi∼εi«1 . Since µi = ki/k′i , this limit requires that the phosphatase of that cycle be much more active that the corresponding kinase . In this limit however , the conservation law remains as expressed in Equation 7 and no further simplification can be made . As a result , in this limiting case , the first term in Equation 6 depends on the variable describing the previous step in a different ( and more complicated ) way compared to Equation 4 . Finally , we notice that our description enables a reduction of the cascade equations with mixed hypothesis concerning the enzymatic reactions . For example , we could have µ1∼ε1 and η1∼1 for the first cycle , µ2∼1 and η2∼ε2 for the second cycle , etc . Or even µi∼εi½ and ηi∼εi½ for all or for some index i . In Text S3 , we present the extension of the reduced mechanistic model for a cascade involving double-phosphorylation . Notwithstanding that these equations are more complicated than Equations 6–7 , the distinguishing feature is maintained: each level in the cascade is subject to influence from the following level which , in the appropriate xi variable , can be identified as a negative feedback . In the current study we analyze mostly static properties of these more complicated equations and compare them to those of Equations 6–7 , while a more exhaustive characterization will be presented in a future article . In this section we report on dynamic and static properties of the new chain equations ( Equations 6–7 ) , when studied by numerical simulations , and compare them to those of previous cascade models . We will consider both short ( n = 3 ) and long chains ( n = 10 , or 15 ) , respectively . In all the figures we plot yi* , the level of active protein , obtained from xi in Equations 6–7 ( see Text S4 for a comparison between variables xi and yi* ) . In this section , each parameter in the reduced mechanistic model is considered to be homogeneous throughout the chain , i . e . , the parameters do not depend on the index i characterizing the position of a particular unit in the chain . The homogeneity assumption implies that Vi≡V = µη/ε and V′i≡V′ = 1 . Parameter S indicates the level of input stimulation the chain receives . The parameter K′ is chosen by considering the relation K′ = K/µ . We have performed numerical simulations with other parameter relationships and the properties reported below are not critically dependent upon that choice . The control parameters are , then , V , K , µ , and η . Since V′ = 1 , the range of V values of interest lies around 1 . The initial condition for all the numerical simulations considered is , at t = 0 , xi = 0 ( and yi = 1 ) for every i . In this section we apply the reduced mechanistic model to a well-known signaling pathway , the mitogen-activated protein kinase ( MAPK ) one [3] , [8]–[10] , [25] , [26] . We first base our description on a particular published set of parameters for this pathway [8] . Importantly , the results obtained are not qualitatively modified by variations of the selected values in the ranges suggested in the literature [8] . Moreover , they are not modified by choosing different sets of parameters [9] , [25] , [26] , as described in the Text S6 . It is well know that the MAPK cascade consists of three levels , the second and the third ones involving a double-phosphorylation mechanism . In this section we consider both the MAPK cascade and a simpler case , a 3-unit chain where each unit is a 2-state cycle . Starting with the published set of parameters ( see [8] and also [10] , for a summary ) , we have computed the parameters involved in the reduced mechanistic description and listed them in Table 1 . As described in Text S6 , there are some extra parameters for the case involving double-phosphorylation , that are designated ν , K* , and K″ and take the values of 1 , 0 . 25 , and 0 . 25 , respectively . According to Table 1 , the conditions under which the reduced model is valid are only partially satisfied , ηµ∼ε for the first unit but ηµ∼10 ε for the second and third ones . Even for these conditions and since the focus of this section is in steady states , the reduced mechanistic model provides a description that is in excellent agreement with the complete mechanistic one . In Figure 6 we plot the normalized stimulus-response curves for a 3-unit chain , either with single-phosphorylation in all the units ( A ) or with single-phosphorylation in unit 1 and double-phosphorylation in units 2 and 3 ( B ) , i . e . , the case corresponding to the MAPK cascade . Both cases are characterized by the parameters in Table 1 . The input stimulus was taken to be the concentration of E1T , the total amount of kinase for the first unit in the cascade ( corresponding to MAPK kinase kinase in B ) . E1T , related to the parameter S we have used as input in the previous section , was varied over a wide range . The outcomes were obtained by both the complete mechanistic and the reduced mechanistic models and the results are indistinguishable for the scales of the figure ( black , blue , and red filled lines for y1* , y2* , and y3* , respectively ) . For completeness , we are also including the corresponding outcomes obtained by the GK-like model ( dotted lines ) . In order to compare the steepness in the responses , we have computed the apparent Hill coefficient nH ( [8] ) for each curve , as indicated in the legend . As expected , nH increases through the chain . Moreover , nH is also considerably reduced when comparing GK-like model's predictions with the predictions of both mechanistic and reduced mechanistic models ( which are , as already mentioned , undistinguishable ) . As explained in the section dealing with stimulus-response curves , these differences could be due to the fact that both the mechanistic and the reduced mechanistic descriptions take into account “sequestration” in the enzymatic reactions [3] . We have mentioned that the good agreement between the mechanistic and reduced mechanistic descriptions regarding the prediction of steady states is due to the conservation law , Equation 7 , taking into account the first correction in εi ( see Text S1 ) . If that correction is not considered , differences could appear in the steady states predicted by the mechanistic model and the reduced mechanistic one . However , and for the parameters in Figure 6 , the predicted values of nH are not modified by removing the εi correction in the conservation law or , even by , removing the ηi correction as well ( i . e . , using a conservation law of the form xi+yi = 1 ) . These results strongly indicate the robustness of the new equations regarding the “ultrasensitivity” characteristics of the cascade . In Figure 6B , the mechanistic and reduced mechanistic models' outcomes and corresponding Hill coefficients recover published results [8] . Comparing figures A and B , we also confirm that the chain involving double-phosphorylation responds in a steeper manner than the one with only single-phosphorylation , as expected from previous work [27] . In Figure 7 we show the outcome of stimulating the 3-unit chain as indicated in the schemes close to each panel: the input stimulus to the cascade was taken to be the concentration of E′3T , the total amount of phosphatase for the last unit in the cascade ( corresponding to MAPK in B ) . E′3T was varied over its suggested range of variation [8] . Increasing the amount of phosphatase produces a decrease in the response curve y3* ( red filled line ) , as expected . Interestingly , our new reduced model ( Equations 6–7 ) , as well as the complete mechanistic description , predict that this perturbation on the third level of the chain is propagated backwards: the variation in y2* is actually a decrease due to a higher sequestration of free y2* by the next step in the chain caused , in turn , by the increased demand of y3 . This result is exhibited by both cascades in Figure 7 ( the one involving only single-phosphorylation and the one with double-phosphorylation in units 2 and 3 ) and we call it “reverse” stimulus-response curves . As stated before , this result is obtained with both the mechanistic and the reduced mechanistic descriptions , with realistic parameters associated with a well studied signaling pathway , such as MAPK . The insets in both figures indicate that is not necessary to vary parameter E′3T over a wide range to observe this property , rather it is clearly seen by changing it only by a factor 5 around its suggested concentration ( 0 . 12 µM ) , where a 20% variation in y2* is observed , a value that is high enough to be detected experimentally ( meaning that it is most likely not contained within the error of the experiment ) . Due to the parameters characterizing this particular pathway , the effect is not propagated to y1* ( black filled line ) , but this fact does not have to be generalized ( see Text S6 ) . The dotted horizontal lines in Figures 7A and 7B are the GK-like prediction for the response curve y2*: within that phenomenological description , a particular level in the cascade is not at all influenced by what happens in a downstream unit . However , this well known property of unidirectional influence in a signaling chain , which is embodied by the appellation of “cascade” , is shown here not to be guaranteed in general signaling cascades . In Text S6 we extend the results in this section concerning “reverse” stimulus-response curves , for different sets of published parameters on the MAPK cascade . A modular response analysis ( MRA ) [28] was applied to determine the network architecture of the cascade in the context of the new model equations ( Equations 6-7 ) . MRA has recently been proposed as a tool to characterize the interactions between “modules” in a cellular regulatory network , having the advantage of allowing direct experimental implementation . As a matter of fact , the negative sign of the Jacobian element ∂x ˙i/∂xi+1 indicates that the ( i+1 ) th level of the cascade exerts a negative effect in what concerns variable xi . This effect ( what we have called “negative feedback” ) is intrinsic , as opposed to “explicit” negative feedback which is sometimes considered in models of signaling pathways [9] , [12] , [13] . MRA is , then , an appropriate approach to test this bidirectional structure and to estimate the relative strength of the backward interaction , as compared with the forward coupling in a signaling cascade . As a result of applying MRA , a matrix of local response coefficients r is obtained . An element rij in this matrix describes how the state of the variable associated with module j directly influences the state of the variable associated with module i . More precisely , a response coefficient rij lower/greater than 1 means that a relative change in module j is attenuated/amplified in module i by a factor rij ( i . e . , Δxi/xi = rij Δxj/xj ) . A zero response coefficient indicates no direct effect between the involved modules , whereas a negative response coefficient means inhibition . In this way , the matrix r provides an interaction map to characterize the type and strength of the interactions between the modules in a cellular regulatory network . Indeed , if the rate of change of variable xi is denoted by the function fi , it easily can be shown that: ( 8 ) meaning that rij corresponds to a scaled version of the Jacobian matrix ∂fi/∂xj ( evaluated in the steady state ) . Moreover , it was proven that the local response matrix r can be obtained from another matrix named global response matrix , Rp , that has the advantage of being accessible experimentally [28] . For example , the element ( i , j ) of this matrix can be obtained by perturbing a parameter pj affecting only module j and computing the relative changes induced on the steady state of xi , namely ( Δxi/xi ) /Δpj . For more details about the broad scope of the method , we refer the reader to the cited reference and references therein . Using notations and concepts from the literature [28] , we apply the MRA method to a 3-unit cascade involving only single-phosphorylation and characterized by the parameters in Table 1 [8] . There are three modules in this network as described by Equations 6-7 , each of them corresponding to the three successive levels in the cascade and characterized by a single variable xi . Figure 8A contains the matrix of local response coefficients r . This matrix was obtained both by direct computation of the scaled Jacobian matrix ( Equation 8 ) and by simulating experimental perturbations to the cascade , then computing the global response matrix , Rp , and finally obtaining r , as described previously [28] ( details of second calculation not shown ) . Using MRA , the “theoretical” and “experimental” outputs were in perfect agreement and the results are displayed in Figure 8A . The structure of matrix r is tridiagonal , meaning that the first level in the cascade does not directly influence the third one ( r31 = 0 ) , and viceversa ( r13 = 0 ) . Coefficients r21 and r32 are positive , representing the positive effect of each level in the cascade to the subsequent one . Interestingly , r12 and r23 are both negative , indicating an inhibitory effect from unit ( i+1 ) to unit i . The resulting connections between the units in the cascade are summarized in the scheme in Figure 8A . To understand these results in more depth , we have studied how the coefficients in the matrix in Figure 8A depend on the parameters characterizing the cascade . For example , Figure 8B shows coefficients r21 , r32 , r12 , and r23 versus the parameter E1T . r31 and r13 are zero ( data not shown ) , r21 and r32 are positive , and r12 and r23 are negative , throughout the range where E1T was varied . Depending on the value of E1T , each of the nonzero rij could be less or greater than 1 and the relative strength of the backward and forward couplings for a given pair of modules , e . g . |r12/r21| , could exhibit large variations . Similar curves have been reported in the literature for signaling cascades [29] , but lacking the information about r12 and r23 , which have always been considered to be zero in previous papers . Studies like the one in Figure 8B help us understand and also predict the degree of backwards coupling as a function of the parameters in the model . One utility of this work is as a starting point of a more systematic study on how to enhance or attenuate that coupling in the cascade , the subject of our ongoing work . Interestingly , the interaction map characterizing the connectivities between variables xi ( matrix r ( xi ) in Figure 8A ) shows strong differences when compared to the one computed for the “free” enzyme variables yi* ( matrix r ( yi* ) in Figure 8C . Although an explicit set of differential equations is not written for the variables yi* , the matrix r ( yi* ) can be calculated using the “experimental”' method described in the literature [28] . The result in Figure 8C is the average of four outputs and the corresponding error ( standard error of the mean ) is lower than 4% . As indicated in the reconstructed topology close to the matrix , r12 and r23 are now positive ( as are r21 and r32 ) , r31 is zero , and r13 is negative , indicating an inhibitory coupling from variable y3* to variable y1* . The matrix r ( yi* ) is consistent with the results in Figure 7A ( and also those in Text S6 ) : in other words , the response in y2* goes in the same direction as the one in y3* ( whereas plotting variables xi indicates a decrease in x3 and an increase in x2 , data not shown ) . Experimental data concerning the application of MRA to the MAPK cascade are now available in the literature [30] , showing non zero r21 and r32 coefficients ( and also non zero r31 and r13 coefficients ) . The interpretation of non zero r31 and r13 was proposed in terms of the usual “explicit” positive or negative feedbacks which are sometimes considered in models of signaling pathways [9] , [12] , [13] . From this perspective , the explanation for the non zero r12 and r23 coefficients was , at least regarding r23 and based on experimental evidence , that not only is y2* able to phosphorylate y3 , but y3*can phosphorylate y2 as well [30] , [31] . Our results however , suggest that the non vanishing backward coefficients ( r12 , r13 , r23 ) can be accounted for , at least partly , by the natural “implicit” feedback which can exist in a signaling cascade . A quantitative correlation between these recent experimental results and our predictions is not possible at this time . In the published experiments , the MAPK cascade is not isolated but embedded in the complex cellular machinery; therefore , the measured connectivities could involve proteins external to the cascade itself and it would be premature to establish the connection with our simplified model . Nevertheless , the work in [30] suggests a direction for the type of experiments that could validate our results .
The main contribution of this work is to propose a new one-variable per cycle model for signaling cascades of covalent modification cycles , consistent with a mechanistic complete description . Our model reveals new and biologically relevant properties of such cascades . These properties are characterized completely for the case of single-phosphorylated cascades . Furthermore , single and doubly-phosphorylated cases are compared regarding their stimulus-response curves , while a more exhaustive characterization of the scheme involving double phosphorylation will be presented in a future article . The scheme in Figure 1 , which has been employed by many groups , is suggestive of the concept of a “cascade” . From a systemic point of view , a cascade is a system composed of units , the output of which is successively an input to the next unit . Based on this structure , powerful concepts from control theory can be applied successfully to the study of signaling cascades [14] . Although these concepts have proven its utility in many contexts , this kind of schematic representation implicitly conveys the idea that a signaling cascade is only a feed-forward chain in which signal transmission is analogous to a domino effect [32] , [33] . Our study sheds a different light on this system , showing that this schematic representation can be misleading , since it turns out that each unit is actually coupled not only to the following one but also to the previous one , and interesting dynamics can arise from these interactions . Our initial motivation for developing a new one-variable description of signaling cascades , was the following observation . The main assumption underlying the GK description of a single cycle is that the concentration of the target protein is in large excess compared to those of the converting enzymes . Holding the same assumption over a cascade of units would mean that the target proteins are in higher and higher concentration as the cascade progresses , since they act as the transforming enzyme for the following cycle . To our knowledge , this important issue has not been remarked upon in the literature , except for a brief comment in the work of Millat et . al . ( [20] , page 11 ) . In order to get more insight into this point , we have sought special limiting cases for which the mechanistic and the GK-like model are in good agreement . However , it turns out that the dynamics of the signaling cascade described by the mechanistic and the GK-like models cannot be compared consistently . The fundamental reason for this mismatch is that a careful perturbation analysis applied to the mechanistic model provides a different set of equations . We note that in search for an adequate set of hypothesis leading from the mechanistic equations to the model given by Equations 4 , we have studied an alternative scheme in which the modified protein Yi* is not directly the kinase of the next reaction . Instead , we studied the case where Yi* activates that kinase . This scheme was suggested by the work of Goldbeter [12] . The resulting equation ( see Text S7 ) is fundamentally different from the GK-like model . In reality , no set of assumptions can give rise to the GK-like model as a limiting case of our model . Our mathematical method relies on the standard quasi-steady state assumption ( QSSA ) , which can be applied under well defined conditions to elicit a clear separation between the slow and fast dynamics of the mechanistic model . Under this standard QSSA framework , our analysis shows that a good slow variable for which evolution equations can be written is the sum of the free activated enzyme which is available in the ith cycle plus the amount of this protein which is captured by the next inter-converting cycle . The idea of working with a mixed variable xi can be further generalized by considering the “total” variable corresponding to the total amount of activated enzyme found not only as free molecules or bound to the next substrate , but also complexed with the reverse enzyme E′i . In fact , this choice is the key ingredient of the method called the “total” quasi-steady state approximation ( tQSSA ) which has been proved to be a simple but most efficient extension of the standard QSSA [34] . The application of this extended framework to the description of the signaling cascade of Figure 1 is concerned with our current research . In the same context , other authors have recently applied the tQSSA method to the study of small networks of GK cycles [35] . These systems do not form cascades , but involve a more complicated coupling between the units . Nevertheless , their results show that indeed the tQSSA method is successful in obtaining a reduced set of equations , with one variable per cycle , which faithfully reproduces the dynamics of the network for a large range of system parameters . Even in the less extended QSSA framework , the conditions under which the model is valid are made clear . Under such conditions , our new model is indeed in perfect agreement with the complete mechanistic model ( Figure 2 ) . Those conditions are expressed in terms of three key parameters ( Equation 5 ) we have defined to simplify the study . Even though the phenomenological equations , Equation 4 , are appealing because of their simpler form and modular nature , we could not find any set of assumptions that would enable us to recover those descriptions . Our simplified model reveals properties of signaling cascades that were either hidden by the complex structure of the complete mechanistic model or lost in the simplified phenomenological descriptions . It was stated that the reduced mechanistic model is valid whenever these two conditions are satisfied: εi«1 and µi ηi∼εi . The study of the performance of the new approximation ( Figure 2 and the corresponding computed errors ) makes it clear that even when those conditions are satisfied only moderately , the new model is still robust in approximating the complete description . As an example , we have computed a 5% error for ε = 0 . 1 , η = 1 , and µ = 0 . 5 ( meaning µη∼5ε ) . Moreover , we have observed that the steady state predictions of the reduced model are highly accurate . Therefore the properties of signaling cascades we are unveiling thanks to the new reduced model , are not restricted by a tight relationship between concentrations and reaction rates hard to achieve in in vivo or in vitro conditions . All the novel properties of a signaling cascade reported in this paper are linked , as previously mentioned , to the negative feedback from each unit to the previous one . This backward negative feedback can produce damped temporal oscillations in the chain , or it can create amplified “pathway” oscillations in the steady states of the cascade . Interestingly , it can also transduce a signal both forward and backwards . Given the multi-branched complex nature of many signal transduction pathways , this finding could have wide implications and can help focus further experimental investigation . It has recently been reported that the 3-level MAPK cascade has autonomous oscillations without any kind of added explicit feedback [36] . Following a systematic numerical exploration of the corresponding mechanistic model [8] , the authors provide a qualitative description of the mechanism responsible for these sustained oscillations . Their explanation strongly suggests the necessity of a bistable behavior at the second or third levels of the cascade , thus requiring double-phosphorylation at these stages [37] . Consistent with their findings , we have observed only damped oscillations in the dynamics of the single-phosphorylated cascade ( Equations 6–7 ) , which has been the main focus of the present work . Interestingly , preliminary numerical simulations of our reduced doubly-phosphorylated cascade model ( Text S3 ) , indicates that these autonomous oscillations are recovered in the simplified description . The stimulus-response curves of the new model were also investigated ( Figure 5 ) . They have the usual sigmoidal shape characteristic of ultrasensitive responses; however , they exhibit lower steepness when compared with the output of the GK-like model . This result corroborates the conclusions stated in the work of Blüthgen et al . [3] , where an analysis of the effect of sequestration was conducted . This effect is partially mitigated by double-phosphorylation ( Figure 6 ) , as expected from the literature [27] . To further characterize the new model within realistic conditions , we have studied it subject to different sets of published parameters corresponding to a well-known signaling pathway , such as the MAPK one ( Figures 6 and 7 , and Text S6 ) . We have found that the ability of the model to transduce a signal both forward and backwards is widespread and that the effect is of enough magnitude to allow experimental verification . Finally , we have applied a modular response analysis to determine the network architecture of the cascade described by the new model equations ( Figure 8 ) . This well-known approach enables not only to test the bidirectional structure of the cascade , but also to estimate the relative strength of the backward interaction . In summary , our findings do not at all weaken the importance of previous models like the GK-like models and those with linear rates . To the contrary , the results of our model provide a different approach to deal with a simple one-variable per cycle model to describe signaling cascades . We hope that our contribution will help in the understanding of existing models for signaling cascades , will improve the description of available data , and will inspire both theoretical and experimental investigation .
All the ODEs were integrated in MATLAB 7 ( Mathworks , Natick , MA ) . The stimulus-response curves were obtained using MATCONT , a MATLAB package for numerical bifurcation analysis of ODEs . The symbolic calculations were done using the Symbolic Math Toolbox in MATLAB . | Cellular signaling is carried out by a complex network of interactions . A structure that is found commonly in signaling pathways is a sequence of on-off cycles between two states of the same protein , referred to as a cascade . By analyzing and reducing the basic kinetic equations of this system , we have constructed a new mathematical model of an intracellular signaling cascade . It is widely accepted that information travels both outside-in and inside-out in signaling pathways . Conversely , cascades , even while being main components of those pathways , have been so far understood as structures where signal transmission occurs in a manner analogous to a domino effect: the information flows in only one direction . Adding explicit connections linking a particular level with an upstream location has been the way bidirectional propagation has been explained so far . In other words , up to now , unidirectional cascades would allow bidirectional propagation only when embedded in more complicated circuits . The proposed model shows that a cascade can naturally exhibit bidirectional propagation without invoking extra re-wiring . This result inspires novel interpretations of experimental data; since signaling pathways are usually reconstructed from such data , this outcome could have far-reaching implications in the understanding of cell signaling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/signaling",
"networks",
"biochemistry/cell",
"signaling",
"and",
"trafficking",
"structures",
"biochemistry/theory",
"and",
"simulation",
"cell",
"biology/cell",
"signaling"
] | 2008 | A Hidden Feedback in Signaling Cascades Is Revealed |
Long non-coding RNAs regulate various biological processes such as dosage compensation , imprinting , and chromatin organization . HOTAIR , a paradigm of this new class of RNAs , is localized within the human HOXC gene cluster and was shown , in human cells , to regulate HOXD genes in trans via the recruitment of Polycomb Repressive Complex 2 ( PRC2 ) , followed by the trimethylation of lysine 27 of histone H3 . We looked for the presence of Hotair in mice to assess whether this in trans mechanism was conserved , in particular at the developmental stages , when Hoxd genes must be tightly regulated . We show that the cognate mouse Hotair is poorly conserved in sequence; and its absence , along with the deletion of the HoxC cluster , has surprisingly little effect in vivo , neither on the expression pattern or transcription efficiency , nor on the amount of K27me3 coverage of different Hoxd target genes . We conclude that Hotair may have rapidly evolved within mammals and acquired a functional importance in humans that is not easily revealed in mice . Alternatively , redundant or compensatory mechanisms may mask its function when studied under physiological conditions .
Genomes contain a large number of RNAs , which do not encode any protein [1]–[5] . While some of these non-coding RNAs such as XIST , TSIX and AIR associate with epigenetic modifying complexes [6]–[11] , the functions of others remain poorly understood . Many of the predicted long non coding RNAs ( lincRNAs ) are thought to be spliced and polyadenylated , thus resembling protein coding RNAs [12]–[15] and have been proposed to impact on gene regulation [16] , [17] . Recent studies have shown that distinct lincRNAs are involved in diverse biological processes such as dosage compensation , imprinting or cancer metastasis [10] , [18]–[20] . More specifically , they may function at the interface between DNA and its epigenetic regulation by targeting remodeling complexes to their target sites [21] . HOTAIR , one such lincRNA located within the human HOXC cluster , regulates HOXD cluster genes in trans via the recruitment of PRC2 , a silencing complex responsible for the deposition of trimethyl groups on lysine 27 of histone H3 ( H3K27me3 ) [10] . Knock-down of HOTAIR in human fibroblasts induced gain of expression of different members of the HOX family , associated with a loss of K27me3 decorating part of the HOXD locus in these cells [10] . In addition , HOTAIR has been shown to co-immunoprecipitate with members of the PRC2 complex such as SUZ12 and EZH2 , but not with the putative PRC1 member YY1 , suggesting a primary role in the initiation of silencing , rather than in its maintenance [6] , [10] , [21] . Subsequent studies have suggested that distinct sub-domains of HOTAIR are essential for the binding of either EZH2 , or of LSD1 and that HOTAIR functions as a bridge to bring both complexes together . In the absence of these two binding domains , the epigenetic functionalities of this lincRNA are indeed completely abrogated [21] . Altogether , these results indicate that human HOTAIR is an important regulator of the HOX epigenetic landscape in skin fibroblasts . Given both the importance of this lincRNA in adult tissues and the critical dynamics of H3K27 trimethylation for the early control of Hoxd gene activation [22] , we investigated its role in developing mouse embryos . Here , we describe the mouse Hotair cognate lincRNA and show that its complete depletion in vivo has no severe effect upon Hoxd gene activation , neither during early trunk development , nor in the course of limb morphogenesis , two sites where HOTAIR was seen expressed .
We first looked for the presence of Hotair in the mouse genome . Because the human RNA locates between HOXC12 and HOXC11 , i . e . within a region of very high micro-synteny amongst all vertebrates , we performed a pair-wise sequence alignment with the cognate mouse DNA segment , using the rVISTA software [23] . Alignment of the entire mouse Hoxc11 to Hoxc12 region with the human genome revealed various domains of strong sequence homology ( Figure 1A ) . Expectedly , the Hoxc11 and Hoxc12 exons are highly conserved , with more than 95% homology between the mouse and human sequences . However , the intergenic region between Hoxc11 and Hoxc12 showed more variability , with some peaks of conservation , but also segments close to random variability , as previously described [15] , [24] . Sequence alignment revealed that the human HOTAIR lincRNA most likely has a mouse ortholog RNA , referred to as AC160979 . This EST ( mHotair from now onwards ) is indeed located at the expected micro-syntenic position and exhibits partial homology with human HOTAIR . mHotair derives from the Vega Protein Coding Annotation and corresponds to the UCSC gene based on RefSeq AK035706 transcript . However , and even though mHotair and HOTAIR are clearly cognate transcripts , several important differences were scored . First , while the RefSeq annotation of HOTAIR indicates six exons , mHotair derives from two exons only . The second half of the first exon of mHotair seems to match exon 4 of HOTAIR , whereas the second exon clearly matches exon 6 of HOTAIR ( Figure 1A ) . Blasts of the first three human exons against the mouse Hoxc11 to Hoxc12 intergenic region did not give any significant homology . Secondly , the level of sequence similarity between different exons is highly variable . The first exon of mHotair , which is 234 base pairs long , shows significant conservation ( >80% over >100 bp ) with the human sequence . However , the second exon , which is 1770 bases long , is poorly related to the human sequence and shows conservation higher than 75% only in a sub-domain of ca . 400 bp . Altogether , this large exon , which contains the LSD1 binding region of HOTAIR , is rather poorly conserved in its mouse counterpart , ranging from 50 to 70% homology . In addition , human HOTAIR contains several binding sites for the SET domain containing PRC2 component EZH2 , responsible for the histone H3 methyltransferase activity ( HMTase ) of this enzyme , which are absent from mHotair . Although it is unclear as to whether the primary nucleotide sequence or the tertiary RNA structure is involved in binding EZH2 , it nevertheless suggests that the function of this RNA in mice is not identical to that described for its human cognate . Transcriptome analyses by deep sequencing confirmed that mHotair was most likely encoded by two exons only , instead of six in humans ( see below ) . Hox genes are clustered at four different genomic loci ( HoxA , B , C and D ) and are crucial in organizing the metazoan body plans . They encode transcription factors , which work in various combinations to allocate morphogenetic identities to groups of cells . To properly coordinate their transcription , these contiguous genes are activated following a collinear regulatory strategy , whereby genes positioned at the 3′ end of the cluster are activated earlier in time and more anteriorly , whereas more 5′ located genes are activated later in time and more posteriorly [25] . This sequential activation in time and space thus follows the physical positions of genes along their respective clusters . This property , which may in part depend upon chromatin modifications [22] also applies either to transgenes , when introduced into the gene clusters , or for non coding intergenic transcripts , regardless of their sense of transcription . These non-coding transcripts associated with Hox genes were proposed to regulate the collinear opening and maintenance of the epigenetic status of the cluster [5] . We looked at the expression of mHotair by whole mount in situ hybridization ( WISH ) on developing mouse embryos at embryonic day 11 . 5 , 12 . 5 and 13 . 5 , and compared with the expression of Hoxc11 , the gene located immediately 3′ from the mHotair promoter . The mHotair probe was selected within the region showing the highest conservation with the human ortholog ( Figure 1A ) , i . e . the middle half of the second exon , such as to compare as accurately as possible with previously published data where the distantly related human HOTAIR sequence was used as a probe for WISH on mouse embryos [10] . Experiments using sense and antisense probes confirmed that mHotair is solely transcribed from the opposing canonical Hox DNA strand , as is human HOTAIR . As expected from its position within the ‘posterior’ part of the HoxC cluster , mHotair expression was scored in posterior and distal sites . It was readily detected in E11 . 5 embryos with marked staining in the posterior part of the hindlimbs , in the genital bud and in the tail . At E12 . 5 , the expression pattern was mainly restricted to the posterior aspect of the intermediate part of the hindlimbs , as well as to the genital bud , whereas it became barely detectable at E13 . 5 . In parallel experiments , Hoxc11 transcripts showed a comparable distribution , yet with stronger signals at all three stages ( Figure 1C and 1D ) , in agreement with previously published data . Given the strong similarities of expression patterns between mHotair and its closest 3′ neighbor Hoxc11 , we concluded that mHotair is expectedly regulated in coordination with other posterior Hoxc genes . mHotair expression , however , was quite distinct from that reported in similar staged mouse embryos when using a human HOTAIR probe [10] . Human HOTAIR was shown to act in trans by tethering Polycomb Repressive Complex 2 ( PRC2 ) to a subset of its targets , amongst which the HOXD locus [10] , [21] . HOTAIR thus acts as a scaffold for the repression of a number of genes in this region via the recruitment of these silencing proteins , with a particular impact on the expression levels of human HOXD8 , HOXD9 , HOXD10 , HOXD11 and HOXD13 , while having no impact neither on the HOXA , nor on the HOXB and HOXC clusters [10] . To investigate whether this mechanism was conserved throughout mammals , we looked at the expression of these potential target genes in the absence of mHotair . We used a full deletion of the HoxC cluster whereby all Hoxc genes and intergenic transcripts are missing ( Figure 2A ) [26] . We isolated HoxCDel/Del embryos at embryonic day 13 . 5 ( E13 . 5 ) , derived from a cross between heterozygous animals , and dissected them into four distinct pieces; the forebody , hindbody , forelimbs and hindlimbs . We performed quantitative RT-PCR analyses on these various samples using wild type and heterozygous littermates as controls for homozygous mutant samples . As expected , mHotair was detected neither in HoxCDel/Del mutant embryos , nor in forebody samples of all three genotypes , which we used as negative controls . In the three other samples , mHotair transcripts were scored , though at very low levels . However , no difference was noted in the expression levels of the presumptive mHotair targets Hoxd8 , Hoxd9 , Hoxd10 , Hoxd11 or Hoxd13 ( Figure 2B ) . The expression level of Hoxd12 remained unchanged too , as well as those of Evx2 and Lunapark , two neighboring genes largely co-regulated with Hoxd genes [27] . A change in the expression of different Hoxd genes could nevertheless remain unnoticed , should a spatial shift in their transcript patterns occur , rather than variations in their RNA steady state levels . We thus performed in situ hybridization on mutant animals to reveal the distribution of Hoxd10 transcripts , which was reported as the main HOXD target for a HOTAIR-mediated de-repression in human cells . At all three stages examined ( E11 . 5 , E12 . 5 , E13 . 5 ) , Hoxd10 transcripts showed wild type patterns in mutant animals ( Figure 2C and 2D ) . Taken together , these observations indicate that mHotair has little or no detectable regulatory effect in trans over Hoxd cluster genes in mice , at least in these conditions . HOTAIR was reported to regulate several HOXD genes by tethering PcG proteins ( the PRC2 complex ) to the posterior HOXD cluster [10] , [21] . Knock-down of HOTAIR in human fibroblasts indeed showed a decreased trimethylation of lysine 27 on histone H3 , in particular at the HOXD locus , with the strongest effect observed over the region between HOXD3 and HOXD8 . Since a loss of H3K27me3 may not necessarily be translated into a detectable increase in Hoxd gene transcription in mouse embryos , we investigated the chromatin status of the HoxD locus in mutant animals . We used chromatin immunoprecipitation ( ChIP ) on E13 . 5 embryos , a stage at which mHotair is transcribed ( see below ) , followed by quantitative RT-PCR to quantify the enrichment of H3K27me3 over the gene cluster . Here again , the parallel loss of both HoxC and mHotair alleles did not significantly alter the amount of K27me3 covering this presumptive target locus ( Figure 3A and 3B ) . From this set of experiments , we concluded that although human HOTAIR might be essential for the recruitment of PRC2 and subsequent tri-methylation of H3K27 in cultured fibroblast , its role in the regulation of mouse Hoxd genes in embryo seems to be minor , if any , at least at this developmental stage . As the reported effects of human HOTAIR were not observed in the absence of the mouse counterpart in vivo , we derived mouse embryonic fibroblast ( MEFs ) from E13 . 5 embryos , either heterozygous or homozygous mutant for the HoxC cluster , to try and better match the conditions wherein HOTAIR's functions had been originally elucidated . We quantified both the amount of transcription of different Hoxd genes and the enrichment of H3K27me3 at this locus . Results obtained with MEFs heterozygous for the deletion of the HoxC cluster were indistinguishable from those obtained from MEFs lacking both copies of HoxC and mHotair . Analyses of both lines of MEFs gave similar amounts of Hoxd gene transcripts and no significant variations was scored in the enrichments of H3K27me3 marks , indicating that the presence of mHotair is not critical for the regulation of Hoxd genes in this context ( Figure 3C and 3D ) . To assess the global impact of mHotair on the gene regulation , we looked at the transcriptomes of those tissues where mHotair was clearly transcribed at E13 . 5 in our whole mount in situ hybridization , namely the hindbody , the hindlimbs and the genital bud . Embryonic tissues were micro-dissected and total messenger RNA isolated from both control and HoxC mutant animals and sequenced using an Illumina Genome Analyzer . Nearly 15 million high quality single reads were mapped on the mouse mm9 genome , using Tophat [28] and visualized using the integrative genome viewer [29] . In this way , we could confirm that , as annotated in RefSeq , mHotair is a two-exons transcript initiating from the opposite strand of the canonical HoxC genes , at least in this context . No additional 5′ located exons were used , unlike in human . We compared mutant and wild type transcript profiles genome wide and observed significant changes . These modifications , which may reflect direct or indirect targets either of mHotair , or of Hoxc gene products , were either up- or down regulated and broadly distributed over all gene ontology categories . Hox genes were included , along with housekeeping genes and genes from unrelated structures and functions ( Figure 4A and 4B ) . We looked at the HoxD cluster and the strongest variation in steady-state level of transcripts was observed for Hoxd8 , Hoxd9 , Hoxd10 and Hoxd11 , as previously reported for HOTAIR in human cells , though the amplitudes were significantly lower ( Figure 4A ) . While these results appeared at first to somehow correlate with the reported effect of human HOTAIR on this gene cluster , Rinn et al . [10] observed a substantial increase in expression of these genes by down-regulating HOTAIR by a factor of two thirds , whereas we detected a maximum of three-fold difference in the complete absence of this lincRNA . To assess whether these differences could be partly explained by the relatively low expression of mHotair at this particular stage ( E13 . 5 ) or a dilution effect , we isolated RNA from the same set of tissues , i . e . hindbody , hindlimbs and genital bud , from E11 . 5 embryos and quantified the RNAs by reverse transcription PCR . Differences in absolute expression levels of the different Hox genes analyzed were comparable to those obtained in our RNA-seq experiment at E13 . 5 , suggesting that the observed effects of mHotair and HoxC deletions on gene regulation are reproducible , at least between these two developmental stages ( Figure 4D ) . The discrepancies between our results and those reported previously may reflect a dilution effect due to only few cells expressing mHotair in our samples . However , we also observed a slight up-regulation of Hoxd1 , Hoxd3 and Hoxd4 and , surprisingly , our mutants exhibited no change in Hoxd13 transcripts ( Figure 4B and 4D ) , neither in downstream-located non coding RNAs , a region significantly up-regulated in previous work . Also , we observed a similar de-repression of Hox genes belonging to other clusters , with Hoxa7 and Hoxb9 showing comparable up-regulations ( two fold , Figure 4B ) , unlike previously reported . Of note , a substantial increase of transcripts matching the second exon of Hoxc4 , i . e . the most 3′ part remaining after the deletion of the HoxC gene cluster . This unexpected burst likely reflects the presence of ‘posterior-acting’ regulation , which are now re-routed towards this sequence , in the absence of the intervening HoxC cluster , as describe in similar contexts [30] . Taken together , while these observations support a general , though rather moderate , effect of removing the HoxC gene cluster , including mHotair , in the posterior part of the developing embryo , transcriptome analyses confirmed the difficulty to attribute to mHotair the same regulatory capacities during embryonic development , than those associated to its human counterpart in cultured fibroblasts . Even though the structure of mHotair showed substantial differences with its human ortholog , we looked for additional evidence of a potential role as a molecular scaffold to bridge PcG proteins to their target sites . We assessed whether or not the group of genes that displayed a clear transcriptional de-repression in HoxC mutant animals was enriched in genes known to recruit PRC2 in ES cells , i . e . in conditions where Hox clusters are covered by H3K27me3 . We applied a stringent cut-off with a significance window of 1 kb and obtained 263 genes up-regulated in the mutant sample , whereas 105 genes were down-regulated . We looked at which fraction of these genes represented known PcG targets , as defined by binding to SUZ12 [31] . Of the 263 genes defined as up-regulated in the HoxC null mice , only 35 ( 13% ) had been determined as being bound by SUZ12 in ES cells ( Figure 4C ) . Likewise , out of a total of 105 genes down-regulated , only 16 were bound by SUZ12 ( 15% ) , a figure that was down to 8 . 6% after Hoxc genes were removed from the list ( since they are deleted in the mutant ) ( Figure 4C ) .
The importance of long non-coding RNAs ( lincRNAs ) for gene regulation has been recently emphasized in many different contexts . One of the paradigms of this novel class of transcripts is the human HOTAIR RNA , which is encoded from within the HOXC gene cluster and acts in trans to regulate HOXD target genes via the recruitment of PRC2 and further tri-methylation of H3K27 [10] . Interestingly , the mouse counterpart shows little sequence conservation with HOTAIR . While such lincRNAs are known to be moderately conserved in sequence between different species , sequence alignment between the mouse and human HoxC clusters reveals that the DNA fragments included in both HOTAIR and mHotair are amongst the less conserved within the Hoxc12 to Hoxc11 DNA interval , as if they would correspond to the less constrained sequences in terms of evolution . Yet some intron-exon borders are conserved , as well as the direction of transcription , which suggests that the mouse HoxC cluster does contain a genuine cognate HOTAIR RNA . Interestingly , the first three exons of HOTAIR seem to be absent from mHotair , which appears to contain two exons only , a first exon related to the fourth exon of HOTAIR , followed by a larger exon 2 , related to the large sixth exon of HOTAIR . Even though an increase in the number of sequence reads may reveal the presence of either additional , poorly spliced 5′ located exons or alternative start sites , mHotair is thus quite distinct in structure from its human cognate . Such a divergence may underlie important differences in function since the first three exons of HOTAIR ( absent from mHotair ) contain binding sites for EZH2 . Likewise , the LSD1 binding sequences , localized at the 3′ extremity of human HOTAIR , is part of the least conserved DNA sequence within mHotair exon 2 ( below 70% conservation ) . Altogether , based on DNA sequence analyses , it is difficult to reconcile the structure of mHotair with the potential function previously attributed to HOTAIR , even though binding of both EZH2 and LSD1 proteins may mostly rely on tri-dimensional structures rather than upon specific RNA sequences . This conclusion was re-enforced by the expression analyses during mouse development , which revealed patterns different from those previously reported when a human probe was used to assess the presence of mouse transcripts [10] . As expected , mHotair is expressed very much like the neighboring Hoxc11 gene , i . e . in parts of the proximal hindlimbs , in the posterior part of the body and in the emerging presumptive external genital organs . We think that this discrepancy in expression patterns can be explained by the very low sequence conservation between the human RNA antisense probe and the mouse target RNA . Coordinated expression of RNA or transgenes introduced within Hox gene clusters has been reported in several instances [32] and illustrates the strong global regulation that controls these groups of genes . Non-Hox promoters located in- or introduced into- these loci tend to adopt the shared expression specificities and thus behave like their nearest neighbors . The genetic ablation of mHotair , under physiological condition , confirmed the apparent difference between the functions of this lincRNA in mice and humans . Firstly , Hoxd genes expression remained moderately affected in most tissues analyzed , as assessed by quantitative PCR , in situ hybridization and RNA-seq , in particular in those tissues of the developing body where steady-state levels of mHotair were the highest . Secondly , the group of genes that was either up- or down-regulated in the absence of mHotair , as scored by transcriptome analyses , did not particularly overlap with known PcG targets as described in ES cells , nor was it enriched in any of the GO terms . Thirdly , no significant difference was scored in the amount of H3K27me3 decorating the HoxD locus , neither by using embryonic tissues , nor when assessing MEFs derived from homozygous null fetuses . This latter point may reflect the fact that mouse Hotair lacks most of the cognate human 5′ RNA fragment , which was shown to be necessary for the binding of EZH2 [21] . Although we cannot exclude that the deletion of mHotair may have induced a subtle effect upon Hox gene expression , these genes would need to be affected much more severely for a phenotypic outcome to be observed , as animals heterozygous for a deletion of the entire HoxD cluster are virtually of wild type phenotype [33] . Therefore , only a robust impact of mHotair on Hoxd genes regulation would make this lincRNA a candidate regulator of these developmental genes in mice , at least at the time when critical changes in chromatin status are observed [22] . How can we explain this unexpected difference in the functional importance of cognate non-coding RNAs in two mammalian species where both the structure and function of Hox genes appear to be highly conserved ? First , our mutant configuration not only lacks mHotair , but also all Hoxc genes as well as the potential mouse ortholog of FRIGIDAIR , another lincRNA located within the HoxC cluster [3] and whose deletion could counterbalance the effect of removing mHotair . However , our transcriptome dataset indicates that mFrigidair , if present in the mouse genome , is not transcribed at detectable level in our posterior body sample , unlike mHotair , which makes this possibility unlikely . Also , there is no evidence supporting a strong effect of HOXC proteins over Hoxd genes regulation . If any , this effect would need to exactly compensate for a potential effect of mHotair such that the situation in the mutant samples would look like wild type . Secondly , the function of mHotair could be restricted to a limited number of cells within the expression domains of Hox genes , in which case our selection of a rather large piece of tissue would reduce the sensitivity of our functional assays via a dilution effect , which would not occur in cultured fibroblasts . While this is a serious possibility , it would imply that only a small subset of Hox positive cells would be ‘exposed’ to mHotair , questioning its general importance in the recruitment of PRC2 during development . Alternatively , human HOTAIR may be required for HOXD gene regulation at later stages and in different contexts , rather than in the early recruitment of PRC2 over the HOXD cluster . As for all other posterior Hox genes , Hoxc11 and Hoxc12 expression is restricted towards the posterior part of the developing body in early mouse embryos . It is nonetheless conceivable that mHotair be transcribed subsequently , in a tissue or organ where it may have a functional importance , such as in foreskin fibroblasts where its function was originally described . This would imply that the recruitment of PRC2 and subsequent tri-methylation of H3K27 over Hoxd cluster genes would be achieved by different mechanisms in different contexts or , at least , by using various components to recruit PRC2 . Another possibility is that mHotair and HOTAIR may have importantly diverged and no longer share any functional similarity . Non-coding RNAs are generally rather poorly conserved in sequences amongst different species and this possibility may not be overtly surprising . The fact that RNA sequences present in HOTAIR and associated with the binding of either EZH2 or LSD1 do not seem to be present in mHotair supports this view . However , this would be difficult to reconcile with HOTAIR being a key player in the regulation of HOX genes in human , since this gene family has been the paradigm of the structural and functional conservation of genetic circuitries in vertebrates , not talking about mammals . Alternatively , mHotair may have a genuine function in organizing the chromatin landscape over Hox genes , but its deletion in vivo could activate redundant or compensatory pathways still allowing proper PcG-mediated silencing to occur , a mechanism absent from cultured human fibroblasts . Silencing of Hox genes during early development must be tightly achieved , to prevent precocious activation leading to mis-identification of structures . Yet this repression will have to be easily reversed subsequently , in the many different contexts where these genes will be activated . Whether or not this epigenetic versatility would be best implemented by redundant silencing mechanisms or by a preponderant strategy relying upon PRC2 dependent tri-methylation of H3K27 is difficult to evaluate . In both cases , mHotair may be recruited to the HoxD cluster to help this silencing to be established , in those regions where it is expressed . However , our results argue against this mechanism being a fundamental process in Hox gene silencing , in particular as these gene clusters are tightly covered by PcG proteins and decorated by tri-methylated H3K27 in all embryonic contexts analyzed so far where these genes must be repressed , i . e . mostly in tissues where mHotair transcripts were below our detection level .
All experiments involving living animals were authorized by- and carried out following- the swiss legal framework . Mice carrying a deletion of the HoxC gene cluster were published previously [26] . They were purchased from the RIKEN BioResource Center ( BRC ) , in Japan . Heterozygous mice were crossed to obtain wild type , heterozygous and homozygous mutant embryos . Genotyping was performed on individual yolk sacs with the following primers: Sequences alignments between the mouse and human HoxC loci were performed using the pairwise Lagan analysis from the Vista website [23] . Mid-day of vaginal plug was considered as E0 . 5 . Embryos were dissected in PBS and fixed overnight at 4° in 4% PFA . Whole mount in situ hybridization was performed according to standard protocols . The decreasing signal intensity observed for the oldest processed embryos is partially due to the somewhat lower permeability of the probe , along with tissue differentiation . Mutant , heterozygous and wild type animals were processed simultaneously to ensure identical conditions . The Hoxd10 probe was as previously described [34] . The murine Hotair and Hoxc11 probes were PCR-subcloned into pGEM-T Easy vector ( Promega ) , sequence verified , linearized and in vitro transcribed with either SalI-T7 ( antisense ) or NcoI-SP6 ( sense ) , using the DIG RNA Labeling Mix ( Roche ) . Chromatin immunoprecipitation followed by quantitative reverse transcription was performed as previously described [35] . Briefly , cells or tissues were fixed for 15 minutes in 1% formaldehyde , washed three times in cold PBS and stored at −80° before being processed using polyclonal anti-H3K27me3 antibody ( Millipore , 17-622 ) . Mouse embryonic fibroblasts were derived from heterozygous crosses of E13 . 5 embryos using standard protocols . Cells were cultured in MEF culture conditions in DMEM supplemented with 10% FBS . Isolated lines were first genotyped using tissues from the embryos and subsequently confirmed with DNA extraction procedures . Passage No 4 MEFs were used for further experiments . The posterior parts of embryos including the hindlimbs , the genital bud and the developing tail at day 11 . 5 and the forebody , hindbody , forelimbs and hindlimbs at day 13 . 5 , were dissected and stored in RNAlater ( Qiagen ) until genotyped . Cells or tissues were first disrupted and homogenized using a Polytron ( kinematic ) before RNA was extracted using the RNeasy Microkit ( Qiagen , 74034 ) , followed by qRT-PCR with SYBR Green . Mean values derive from two ( MEFs ) or four ( tissues ) biological replicates , processed in triplicates and normalized to a housekeeping gene ( Rps9 ) . The most posterior parts of fetuses at day E13 . 5 were dissected , including the hindlimbs , the genital bud and the developing tail , and total RNA was extracted as for expression analysis . Wild type and mutant samples were deep sequenced using the Illumina Genome Analyzer . Reads were mapped onto the mouse mm9 genome using Tophat and visualized with the integrative genome viewer ( mean value of 25 bp windows ) . Mis-regulated genes were identified using a 200 bp binning approach across the genome . Significance was measure by the presence of probes showing a difference between wt and mutant profiles greater than 6 over at least 5 probes ( 1 kb ) . | Long non-coding RNAs ( lincRNA ) have recently become a new paradigm for gene regulation via chromatin remodeling in a variety of biological processes , including during embryonic development . HOTAIR , a human lincRNA localized within the HOXC cluster was shown to help silence HOXD cluster genes in trans , through the recruitment of the Polycomb Repressive Complex ( PRC2 ) . In this paper , we investigated the role of the murine mHotair lincRNA and report that both its structure and function are quite different from that described in human cells . Deletion of mHotair in vivo has little effect on the transcriptional regulation and chromatin modification of mouse Hoxd genes , and embryonic transcriptomes did not reveal any particular effect upon genes reported as targets of PRC2 in ES cells . Our results indicate that the function of this RNA in mice is distinct from that reported for human cell lines , pointing to a rapid evolution of this lincRNA . Alternatively , redundant mechanisms may mask the function of mHotair in physiological conditions or this lincRNA may be required in a very restricted and specialized cell type . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"genetics",
"biology",
"genomics",
"evolutionary",
"biology",
"genetics",
"and",
"genomics"
] | 2011 | Structural and Functional Differences in the Long Non-Coding RNA
Hotair in Mouse and Human |
In eukaryotes , hundreds of mRNAs are localized by specialized transport complexes . For localization , transcripts are recognized by RNA-binding proteins and incorporated into motor-containing messenger ribonucleoprotein particles ( mRNPs ) . To date , the molecular assembly of such mRNPs is not well understood and most details on cargo specificity remain unresolved . We used ASH1-mRNA transport in yeast to provide a first assessment of where and how localizing mRNAs are specifically recognized and incorporated into mRNPs . By using in vitro–interaction and reconstitution assays , we found that none of the implicated mRNA-binding proteins showed highly specific cargo binding . Instead , we identified the cytoplasmic myosin adapter She3p as additional RNA-binding protein . We further found that only the complex of the RNA-binding proteins She2p and She3p achieves synergistic cargo binding , with an at least 60-fold higher affinity for localizing mRNAs when compared to control RNA . Mutational studies identified a C-terminal RNA-binding fragment of She3p to be important for synergistic RNA binding with She2p . The observed cargo specificity of the ternary complex is considerably higher than previously reported for localizing mRNAs . It suggests that RNA binding for mRNP localization generally exhibits higher selectivity than inferred from previous in vitro data . This conclusion is fully consistent with a large body of in vivo evidence from different organisms . Since the ternary yeast complex only assembles in the cytoplasm , specific mRNA recognition might be limited to the very last steps of mRNP assembly . Remarkably , the mRNA itself triggers the assembly of mature , motor-containing complexes . Our reconstitution of a major portion of the mRNA-transport complex offers new and unexpected insights into the molecular assembly of specific , localization-competent mRNPs and provides an important step forward in our mechanistic understanding of mRNA localization in general .
In eukaryotes , directional transport and localization of mRNA is widely used to regulate gene expression on a temporal and spatial level . mRNA localization is involved in diverse processes such as inducing cellular asymmetry , guiding key events during embryonic development , and supporting synaptic plasticity [1]–[4] . For these processes , motor-containing mRNPs usually translocate translationally silent transcripts from perinuclear areas to their subcellular destination . After anchoring , mRNA translation is activated and encoded proteins are produced [2]–[4] . To date , the basic principles underlying the incorporation of mRNAs into translocating particles , the specific roles of mRNP-core factors , and subsequent mRNA localization are not well understood . Thus , a detailed analysis is required on how mRNA-translocation particles assemble and how RNA-cargo specificity is achieved . A study on the localization of Vg1 and VegT RNPs in Xenopus oocytes already showed marked differences between the nuclear and cytoplasmic protein composition of these RNPs [5] . By using immunoprecipitation experiments on cell extracts and antibody stainings , the authors demonstrated that factors sequentially join the maturing complex on its path from the oocyte nucleus to the cytoplasmic vegetal pole . Comparably little biochemical data are available explaining how mRNPs recognize their cargo-transcripts in a specific way . Available studies only revealed a 3–7-fold higher affinity for localizing mRNAs when compared to non-localizing RNAs [6] , [7] . Because such rather small differences are unlikely to explain the highly selective transport of RNAs observed in vivo , it seems obvious that essential information on the assembly of specific transport complexes is missing . The goal of the present work was to understand such general principles by in vitro reconstituting a major part of a core mRNA-transport complex from yeast and complementing in vivo studies . During mitosis of Saccharomyces cerevisiae , ASH1 mRNA is transported as part of large mRNPs from the mother cell to the daughter cell [4] , [8] . ASH1 mRNA contains four cis-acting regions , termed zip-code elements E1 , E2A , E2B , and E3 , mediating mRNA incorporation into the mRNPs and its subsequent localization [9] , [10] . After mRNP anchoring , ASH1 mRNA is translated in the daughter cell . The protein product Ash1p acts as a repressor of mating-type switching exclusively in the daughter cell [11] , [12] . In addition to ASH1 mRNA , more than 30 transcripts are localized by this transport complex [13]–[15] . Their incorporation into the mRNP is thought to be mediated by its core RNA-binding protein She2p [16]–[18] . She2p is an unusual RNA-binding protein [19] . It interacts with ASH1 mRNA already in the nucleus at the site of transcription [20] and escorts it into the cytoplasm [7] , [16] , [21] . After nuclear export , the She2p-ASH1 mRNA complex binds to the stable cytoplasmic co-complex of the myosin-adapter She3p and the type V myosin motor Myo4p [17] , [18] , [22]–[26] . Recent actin-gliding assays with mRNPs purified from yeast extracts showed that a core complex consisting of Myo4p , She3p , She2p , and a shortened ASH1-E3 RNA element has motile activity [27] . In addition to this cytoplasmic core mRNP , the RNA-binding proteins Puf6p and Khd1p associate with the ASH1 mRNA-She2p complex and are required for efficient ASH1-mRNA localization in vivo [28]–[30] . Whereas the nucleo-cytoplasmic shuttling protein Puf6p binds to ASH1-E3 , cytoplasmic Khd1p interacts with a region encompassing ASH1-E1 RNA . Both proteins are involved in translational repression during transport [28]–[30] . Among all these factors , She2p is the only RNA-binding protein known to bind to all four zip-code elements of the ASH1 mRNA [9] , [10] , [17] , [31] and to be part of the core-transport complex [27] . She2p is therefore believed to mediate the specific incorporation of localizing mRNAs into the ASH1 mRNP . To date , it is unclear how She2p mediates this function . Since the ASH1-E3 zip code alone can mediate efficient mRNA localization in vivo [27] , [31] , [32] , we concentrated our reconstitution studies on this element . Because Khd1p does not interact with the E3 element [32] , we excluded this protein from our study . We found that She2p and Puf6p bind to zip-code elements only with moderately higher affinity than to unrelated RNAs . This raised the question of when and how specificity for mRNP transport is achieved . We further discovered that the myosin adapter She3p is a previously uncharacterized RNA-binding protein . It interacts directly with She2p also in absence of RNA . Addition of zip code containing RNAs results in the formation of highly specific ternary complexes . This specific mRNP has an at least 60-fold higher affinity for zip-code RNAs . Furthermore , incorporation of the RNA increases the interaction between She2p and She3p to a similar extent . Since we find She3p exclusively in the cytoplasm , we propose that highly specific mRNA recognition occurs after nuclear export and is coupled to the assembly of mature , motor-containing transport complexes .
Previous studies showed that She2p ( Gene ID: 853728 ) binds to all four zip-code elements of the ASH1 mRNA ( Gene ID: 853650 ) [9] , [10] . It also binds to the zip-code element of EAR1 mRNA ( Gene ID: 855207 ) [13] and to the 5′ zip-code element ( WSC2N ) of WSC2 mRNA ( Gene ID: 855438 ) [13] . The equilibrium-dissociation constants ( Kd ) of She2p binding to these elements are all in the nanomolar range ( Figure 1A , left table ) [33] . The strongest binding with a Kd of 0 . 10 µM was observed for the most 3′-located E3 element of the ASH1 mRNA and the weakest binding for the EAR1 zip-code element ( Kd = 0 . 77 µM ) . No binding was detected to an unstructured Poly-A20 RNA or the MS2-stem-loop RNA [33] . However , we had previously also observed considerable She2p binding to the stem-loop containing RNA HIV-1 TAR ( 16 mer ) ( Kd = 0 . 91 µM; Figure 1A , left table ) [7] . We therefore wondered whether this binding to an unrelated RNA is an exception or constitutes a more general feature of She2p . When repeating RNA filter-binding experiments with HIV-1 TAR RNA ( 16 mer ) , a longer HIV-1 TAR RNA ( 57 mer ) , and the U1 snRNA hairpin II as controls , we observed Kds ranging from 1 . 1 µM to 1 . 6 µM ( Figure 1A , right table ) . These data indicate that the affinity of She2p to zip-code elements is only about 2–10-fold higher than to unrelated stem-loops . In vivo , however , only few mRNAs are associated with She2p [13]–[15] . Thus , the low in vitro specificity of She2p binding is unlikely to explain the high selectivity for localizing mRNAs in the cell . One possible explanation for the high specificity in vivo is that another , more selective RNA-binding protein mediates specific binding to localizing transcripts . A recent study suggested that Puf6p ( Gene ID: 852107 ) , which binds to the ASH1-E3 zip-code element , also interacts directly with She2p in vivo [21] . Thus , Puf6p could potentially bind specifically to ASH1 mRNA and simultaneously to She2p , mediating cargo specificity for the transport complex . We purified Puf6p to near homogeneity ( see Materials and Methods ) and confirmed its integrity by size-exclusion chromatography and circular dichroism spectroscopy ( Figure S1 ) . Then we assessed the specificity of Puf6p binding to RNAs by electrophoretic mobility shift assays ( EMSA ) and observed a rather indiscriminate binding of Puf6p to RNAs ( Figure 1B ) . Also the concentration-dependent increasing size of the shifted RNA complexes hints at an unspecific binding of additional Puf6p molecules to RNA . We repeated these experiments with ASH1-E3 RNA in the presence of excess amounts of either specific ( ASH1 E3 ) or unspecific ( EAR1 ) unlabeled competitor RNA . The EAR1 zip-code element was chosen as an unspecific competitor because it lacks a PUF-consensus site [29] . Complex formation of ASH1-E3 RNA and recombinant Puf6p was competed for by excess amounts of ASH1-E3 RNA , but only poorly by EAR1 RNA ( Figure 1C ) . Thus , in vitro Puf6p displays a limited preference for RNAs with PUF-consensus sites . In order to find out if the combination of She2p and Puf6p results in synergistic RNA binding , we also performed EMSAs with both proteins and RNAs . In none of these experiments did we observe any synergistic binding or increased specificity ( Figure 1D ) . To assess the previously suggested direct interaction of Puf6p with She2p in vitro [21] , we performed pull-down experiments with the respective recombinant proteins . Using GST-Puf6p and She2p , no stable interaction was observed in absence of RNA ( Figure 1E ) . Next , we investigated whether Puf6p and She2p interact in an RNA-dependent manner . In order to obtain the large amounts of RNA required for these experiments , we used a recently established expression strategy , in which RNAs of interest are recombinantly expressed in fusion with tRNA [34] , [35] . Wild-type ASH1-E3 RNA and a shortened control ASH1-E3 RNA , which fails to bind She2p ( unpublished data ) and lacks predicted Puf6p binding sites [29] , were expressed and purified in fusion to the anticodon stem of a methionine tRNA ( E3-118-tRNA and E3-33-tRNA , respectively; Figure S2 ) . When adding E3-118-tRNA , we observed both Puf6p and She2p in the same complex , whereas no interaction was detected with the control E3-33-tRNA ( Figure 1E ) . Thus , the interaction between Puf6p and She2p observed in vitro appears to be indirect . Furthermore , from the sum of these experiments we can conclude that Puf6p is unlikely to mediate cargo-specificity to the transport complex . In the cytoplasm , the cargo adapter She3p ( Gene ID: 852427 ) joins the She2p:RNA complex . She3p is thought to connect specifically pre-bound mRNAs to the motor Myo4p ( Gene ID: 851204 ) [17] , [18] , [22]–[26] . However , the weak RNA-binding specificities we observed for She2p and Puf6p appear inconsistent with this model . Therefore , we assessed the role of She3p in complex formation . In pull-down experiments with His-tagged She3p and in size-exclusion chromatography , we observed a co-purification of She2p ( Figures 2A and S3A , B ) . Subsequent surface plasmon resonance ( SPR ) experiments with amine-coupled She3p yielded a Kd of 1 . 6±0 . 2 µM ( Figure 2B ) and fast complex dissociation ( unpublished data ) , suggesting low complex stability . Since previous data indicated that She3p might be able to join the co-complex of She2p and RNA [17] , we also tested their assembly with recombinant proteins by size-exclusion chromatography . She2p and She3p indeed co-eluted efficiently with zip code containing E3-118-tRNA at high molecular weight ( Figure 2C ) . In contrast , the control E3-33-tRNA did not elute in co-complex with both proteins . Reconstitution could be further expanded by the addition of the C-terminal half of Myo4p ( Figure S3C ) , which interacts with She3p . In summary , these observations show that She2p , She3p , and zip-code RNA form a ternary complex that can be reconstituted with recombinant proteins in vitro . More importantly , it suggests that only functional zip-code RNAs join the She2p:She3p complex . Because Puf6p co-purifies with She2p and ASH1 mRNA and colocalizes with ASH1 mRNA at the bud tip , it is considered to be associated with the actively transported ASH1 mRNPs [28] , [29] . Thus , Puf6p should also associate with the recombinant She2p:She3p:E3-RNA complex . We performed in vitro pull-down experiments with either His-tagged She3p or GST-tagged Puf6p and analyzed the co-elution with other complex factors . Using either strategy , Puf6p co-purified with She2p , the zip-code element-containing E3-118-tRNA , and She3p ( Figure S4A–C ) . In contrast , in absence of RNA or when the PUF consensus-lacking E3-33-tRNA was used , Puf6p did not form a detectable complex with She2p and She3p ( Figure S4A , B ) . These results indicate that Puf6p and the She2p:She3p co-complex can indeed bind simultaneously to the ASH1-E3 element . It supports the previously reported role of Puf6p as translational repressor of the actively transporting ASH1 mRNP [28] , [29] . To understand the specific role of She3p in ternary complex assembly , we also tested if She3p alone binds to RNA . In EMSAs She3p indeed bound to zip-code RNAs ( Figure 3A ) but also to HIV-I TAR RNA ( Figure 3B ) with affinities comparable to She2p [7] , [33] . Thus , we consider She3p to be a previously undetected , rather unspecific RNA-binding protein . Remarkably , it fails to show sequence similarity to known RNA-binding domains . As with other RNA-binding proteins tested in this study , She3p alone lacks RNA-binding specificity suitable to explain the highly specific transport of a subset of mRNAs in vivo . Next , we assessed whether the cytoplasmic She2p:She3p complex shows higher specificity for zip-code RNAs than the individual proteins . Whereas the interaction of She2p with zip-code RNAs was barely detectable by EMSAs ( Figure S5A , B ) , She3p binding was observed even at low nanomolar concentrations ( Figure S5C ) . In order to detect potential synergisms in binding , we performed EMSAs with increasing concentrations of She2p and a constant , low concentration of She3p ( 25 nM ) . At these concentrations , She2p alone did not show any mobility shift ( Figure S5A , B ) , whereas She3p yielded a faint shift for the ASH1-E3 element ( Figure S5C ) . In the event of synergistic RNA binding by both proteins , a distinct and stronger band shift corresponding to the ternary complex would be expected . EMSAs with She2p , She3p , and ASH1-E3 RNA showed indeed strong binding even at the lowest experimental She2p concentration of 5 nM ( Figures 3C and S5D ) . Subsequent Western blot analysis confirmed that She2p is present in the detected RNA-protein complexes ( Figure S5E ) . A similar increase in affinity was also observed with the EAR1 zip-code element ( Figure 3D ) and the ASH1-E1 , E2A , and E2B zip-code elements ( Figure S5F–H ) . Based on these EMSAs , we estimated the Kds for ternary complex formation with all four ASH1 zip-code elements and the EAR1 zip-code element to be around 25 nM . When repeating these EMSAs with unspecific HIV-1 TAR RNA , no complexes were observed even at experimental She2p concentrations of 1 . 5 µM ( Figure 3E ) . Thus , it can be estimated that the She2p:She3p co-complex assembles with all tested zip code containing RNAs with at least 60-fold higher affinity . It should be noted that these numbers only provide a rough estimate . They nevertheless clearly indicate that this complex is highly specific . In order to confirm the synergistic effect observed in EMSAs by a different approach , we also performed filter-binding assays . These experiments with the ASH1-E3 and EAR1 zip-code elements also yielded a clear synergistic effect whenever combinations of She2p and She3p were added ( Figure S6A , B ) . Although She2p and She3p are both RNA-binding proteins , it is possible that in the ternary complex only one of them directly contacts the RNA . In such a scenario the effect of the other protein would be limited to an allosteric influence on the RNA-interacting protein . In order to test this possibility , we performed UV cross-linking experiments with She2p , She3p , and radioactively labeled ASH1-E3 RNA . In order to optimize the experimental conditions , we employed a shortened E3 RNA of 51 nucleotides ( Figure S2 ) that still mediated synergistic RNA binding with She2p and She3p ( Figure S5I ) . Following UV cross-linking and denaturing gel electrophoresis , we observed distinct bands for She2p ( Figure 3F , lane 3 ) and for She3p ( Figure 3F , lanes 5–6 ) . Those bands for She2p and She3p were both detected also when the ternary complex was subjected to UV cross-linking ( Figure 3F , lanes 7–8 ) . Since UV light only cross-links direct atomic interactions between proteins and nucleic acids , this experiment indicates a direct RNA binding of She2p and She3p within the ternary complex . She3p does not share sequence homology to known protein structures . In order to identify features in She3p that mediate RNA binding , She2p interaction , and ternary complex formation , we analyzed different deletion fragments of She3p . We noticed that a fragment consisting of the very C-terminal 72 amino acids , termed She3p ( 354–425 ) ( Figure 4A ) , could be expressed in Escherichia coli as a stable protein . This fragment bound RNA and She2p but failed to mediate synergistic formation of the ternary complex ( Figure 4B–E and Table 1 ) . A slightly longer fragment of 92 amino acids ( Figure 4A ) , termed She3p ( 334–425 ) , showed slightly stronger binary interactions with She2p and RNA ( Figure 4B , C and Table 1 ) . In contrast to the shorter C-terminal fragment , She3p ( 334–425 ) also supported synergistic RNA binding ( Figure 4D , E and Table 1 ) . Interestingly , these additional 20 amino acids in the longer C-terminal She3p fragment harbor residues previously described to be important for ASH1-mRNA localization in vivo [36] . In order to obtain further details on this minimal ternary complex , we performed UV cross-linking experiments followed by RNase treatment , trypsin digestion , enrichment via TiO2 , and mass-spectrometric analysis . This recently developed technique [37] , [38] allows for an unambiguous identification of protein fragments that directly contact RNA . By using this technique , we identified one peptide in She2p and one peptide in She3p ( 334–425 ) that were cross-linked to ASH1 E3-51 RNA ( Figures 4F and S7 ) . In She2p the cross-linked region is part of a small helix , termed helix E , which protrudes at right angles from the body of the structure ( Figure 5A–D ) [19] , [33] . The cross-linked peptide in She3p contains amino acids that are part of the 20 amino acids long fragment identified in this study to be required for synergistic RNA binding with She2p ( residues 334–353; Figure 4D–E ) . Thus , the cross-linking/mass-spectrometry experiments assign a direct molecular function , i . e . RNA binding , to two distinct regions within the ternary complex . Knowing that the C-terminal 92 amino acids of She3p are functionally important , we mutated conserved residues in this region and tested whether synergistic RNA binding with She2p is affected . Two sets of mutations have already been described in the C-terminus of She3p to affect ASH1-mRNA localization in vivo [36] . The first of these mutant proteins , She3p ( S343E S361E ) , did not show a significant defect in synergistic RNA binding ( Table 1 and Figure S8A , G ) , whereas the other mutant She3p ( S348E ) showed a slight reduction in the synergism ( Table 1 and Figure S8B , F ) . Since the observed defect is rather weak , it seems likely that both sets of mutations also affect a different step during RNA localization . We also generated two additional point mutations at conserved amino acid positions of She3p . Whereas She3p ( R341E ) did not show any defect in vitro ( Table 1 and Figure S8C ) , the mutant She3p ( L364A V367A ) showed significantly reduced synergistic RNA binding with She2p ( Table 1 and Figure S8D ) . Further analysis showed that She3p ( L364A V367A ) alone exhibits impaired She2p binding but wild-type-like RNA binding ( Table 1 and Figure S8H , I ) . Thus , we identified a region in She3p required for direct She2p interaction and synergistic RNA binding with She2p . We also generated mutant versions of She2p and tested them for interaction with She3p and RNA . Two structural regions of She2p show high sequence conservation [19] but have not yet been assigned to a specific function . One region is the helix E of She2p that we already identified to UV-cross-link to zip-code RNA ( Figures 4F and 5B–D ) , indicating that it is part of the RNA-binding surface of She2p . The second conserved region is the protease-sensitive , very C-terminus of She2p ( Figure 5A ) . In the modeled tetrameric structure of She2p , the C-terminus is located at the dimer-dimer interface , close to the previously assigned RNA-binding surface [33] . We created She2p deletion mutants of both regions and confirmed their integrity by size-exclusion chromatography ( Figure S9A ) . When performing pull-down experiments with His-tagged She3p , She2p ( ΔC ) could be co-eluted ( Figures 6A , S9B , and Table 1; compare with wild-type She2p in Figure 2A ) . In contrast , no binding was observed with She2p ( ΔhE ) ( Figures 6A , S9B , and Table 1 ) . SPR with amine-coupled She3p showed almost wild type-like binding by She2p ( ΔC ) ( 2 . 0±0 . 2 µM ) , whereas no binding was observed with She2p ( ΔhE ) even at experimental concentrations of 150 µM ( Figure 6A; compare with Figure 2B ) . We conclude that She2p requires its protruding helix E for the interaction with She3p . In contrast , the very C-terminus of She2p is dispensable for She3p binding . Next we assessed ternary complex formation with both She2p mutants by size-exclusion chromatography . Whereas She2p ( ΔC ) still forms stable complexes with She3p and E3-118-tRNA , no ternary complexes assembled with She2p ( ΔhE ) ( Figures 6B and S9C ) . When using the control RNA E3-33-tRNA , no ternary complex was observed for any of these proteins . We also used EMSAs to study synergistic RNA binding of the She2p mutants with She3p . Compared to wild-type She2p , the She2p ( ΔC ) mutant showed an about 5-fold reduced complex formation with She3p and the E3 zip-code element ( Figure 6C; compare with Figure 3C ) . Since the binary interaction of She2p ( ΔC ) with She3p is largely unaffected ( compare Figure 6A with Figure 2B ) , deletion of the C-terminus mainly impairs the synergistic RNA binding with She3p . No mobility shift was observed with She2p ( ΔhE ) ( Figure 6C ) , which confirms that helix E plays an essential role in the formation of the ternary complex . We also analyzed ternary complex formation of these She2p mutants with the EAR1 zip-code RNA and observed defects comparable to the ones obtained with ASH1-E3 RNA ( Figure S9D , compare with Figure 3D ) . Finally , we analyzed the effects of both mutations on the RNA binding of She2p alone . Since in our hands RNA binding by She2p is too transient for detection by EMSAs ( Figure S5A , B ) , these interactions were tested by filter-binding assays [19] , [33] . In these experiments , both mutants showed reduced binding to zip-code RNAs ( Figure S9E ) . Interestingly , She2p ( ΔhE ) still bound unrelated RNAs like wild-type She2p ( Figure S9E ) , suggesting a role of helix E in specific RNA recognition . Defects in RNA-binding by She2p ( ΔC ) were much more pronounced for ASH1 zip-code elements than for EAR1 mRNA . These observations are consistent with respective defects in ternary complex formation . Furthermore , they demonstrate that selective reduction of RNA binding , as observed for She2p ( ΔC ) , impairs synergistic RNA binding with She3p . Deletion of the helix E of She2p abolishes synergistic complex formation . Because this deletion constitutes a considerable alteration of the overall shape of the protein , we generated more subtle point mutations around this region ( Figure 5A–D ) . Whereas the mutant versions She2p ( E183A D184A G185A ) and She2p ( T191A D192A ) affect the flexible loop region at the tip of helix E ( dotted line in Figure 5D ) , She2p ( E195A L196A ) and She2p ( Q197A E198A I199A ) altered amino acids that directly interact with helix E and are part of a joint surface landscape with this helix ( Figure 5C , D ) . She2p versions with mutations in the flexible loop region did not show any defect in synergistic RNA binding with She3p ( Table 1 and Figure S10A , B ) . In contrast , the helix E–affecting mutants She2p ( E195A L196A ) and She2p ( Q197A E198A I199A ) displayed reduced synergistic RNA binding with She3p ( Figure 6D and Table 1; compare with Figure 3C ) . With both mutants the interaction with She3p was abrogated and RNA binding was moderately reduced ( Figure S10C–E ) . We used circular dischroism spectroscopy to confirm that both mutant proteins adopt the alpha-helical fold observed for the wild-type protein ( Figure S10F ) . To analyze how She2p mutations affect the formation of stable translocation complexes in vivo , we performed co-immunoprecipitation experiments of Myc-tagged She2p followed by Western blot analyses against HA-tagged She3p . Experiments were performed in she2Δ background . Wild-type She2p-Myc efficiently co-precipitated She3p ( Figure 7A , IP lanes ) , indicating that assembled translocation complexes can be detected . In contrast , co-immunoprecipitation with She2p-Myc ( ΔhE ) or She2p-Myc ( ΔC ) did not show any She3p interaction above background levels ( Figure 7A ) . Thus , these experiments together with our reconstitution studies show that synergistic mRNA binding of She2p and She3p is important for the assembly of mRNPs in vivo . We also investigated She2p ( ΔhE ) localization by antibody staining in she2Δ cells . As expected , She2p ( ΔhE ) completely failed to localize to the bud tip ( 0% , n>300 cells; Figure 7B–D ) . We further tested localization of the C-terminally truncated version of She2p , which moderately affects ternary complex formation in vitro but exhibits wild-type-like binding to She3p ( Table 1 ) . The She2p ( ΔC ) mutant also failed to localize to the bud tip above background levels ( i . e . 1 . 0% , n>300 cells; Figure 7B–D ) . In addition , we transformed a Δshe2 strain with a plasmid expressing She2p ( ΔhE ) under its endogenous promoter and assessed ASH1-mRNA localization by in situ hybridization . In contrast to wild-type She2p , we did not observe any ASH1-mRNA localization in response to She2p ( ΔhE ) expression ( Figure 7E; n>100 cells ) . Thus , disruption of ternary complex formation as observed in vitro also results in abolished ASH1-mRNA localization in vivo . A current model of ASH1-mRNP assembly suggests an early assembly of She2p with ASH1 in the nucleus [7] , [16] , [20] , [21] . Although previous studies did not hint at a nuclear role of She3p [17] , [18] , [21]–[23] , [36] , this has not been tested directly . It is therefore a remaining possibility that the ternary complex also plays a functional role in the nucleus . A well-established way of testing nuclear shuttling of proteins that travel together with mRNA is the use of the temperature-sensitive nuclear export mutant mex67-5ts . At restrictive temperature , shuttling proteins such as She2p accumulate in the nucleus [16] . Immunofluorescence microscopy with mex67-5ts cells revealed wild-type-like localization of She3p at permissive temperature ( Figure 8A ) . Also at restrictive temperature She3p remained in the cytoplasm , indicating that it does not shuttle into the nucleus . Because at restrictive temperature She2p and other shuttling factors are retained in the nucleus , mRNA localization is abolished [16] . In addition , transport factors like Myo4p and She3p become delocalized in the cytoplasm ( Figure 8A ) [16] , serving as an internal control for the efficient block of nuclear export . In summary , this experiment confirms the previous assumption that She3p is an exclusively cytoplasmic protein . The finding also supports our notion that the highly specific ternary complex only forms in the cytoplasm during the assembly of the mature transport complex . Recent chromatin-immunoprecipitation ( ChIP ) experiments showed that She2p associates co-transcriptionally with RNA polymerase II [20] in a gene-unspecific manner . A subsequent comparison of She2p ChIP occupancy in presence or absence of RNase also indicated a selective co-transcriptional binding of She2p to localizing mRNAs . Our studies on She3p localization using mex67-5ts strains yielded no signs of nuclear accumulation of She3p ( Figure 8A ) . It is possible , however , that a minor fraction of She3p enters the nucleus , which escapes visualization by immunostaining . If She3p is present in the nucleus , the ternary complex described in this study should also form and stabilize co-transcriptional RNA binding of She2p . In order to test this possibility , we performed ChIP experiments with wild-type and Δshe3 cells . A comparison of the She2p occupancy at the ASH1 locus and three control loci in wild-type cells and in Δshe3 cells showed no significant difference ( Figures 8B and S11 ) . As previously shown by Shen and colleagues [20] , treatment with RNase resulted in a reduced enrichment at the ASH1 locus ( Figures 8B and S11 ) . However , in contrast to Shen and colleagues , we also observed an RNase-dependent reduction of She2p recruitment to genes , which do not encode for localizing mRNAs ( Figures 8B and S11 ) . This RNase-dependent reduction of She2p occupancy is the same in wild-type and Δshe3 cells . In summary , these data show that She3p does not contribute to co-transcriptional She2p recruitment and is therefore unlikely to play an active role in the nucleus .
To date , all components of the ASH1 mRNP including the core factors connecting mRNA to the myosin motor have been identified [4] , [8] , [39] . However , largely due to the lack of quantitative interaction studies , the molecular mechanisms leading to specific mRNA recognition and mRNP assembly remain ambiguous . To our knowledge , we performed the first in vitro reconstitution of the core of an mRNA-transport complex . We followed the path of mRNP assembly from the nucleus to the cytoplasm and assessed binding affinities , specificities , and synergisms in complex assembly . Key findings were confirmed by complementing in vitro and in vivo experiments . Since She2p binds ASH1 mRNA co-transcriptionally and escorts it until it is anchored at the bud tip [16]–[18] , [21] , it had been assumed that this protein is responsible for the specific incorporation of zip code containing mRNAs into localizing mRNPs . We found that neither She2p nor Puf6p , which is the other ASH1-E3 element-interacting protein present in the nucleus , bind zip-code RNAs with high specificity ( Figure 1A–C ) . Because in a previous study a more specific recognition of ASH1-E3 RNA by Puf6p had been suggested [29] , we performed additional control experiments to ensure that Puf6p is folded and active in our experiments ( Figure S1 ) . Whereas RNA-binding assays with unlabeled competitors are virtually identical in both studies , we also observed strong Puf6p binding to unrelated control RNAs in standard EMSAs . Although we arrive at a somewhat different estimation of the binding specificity of Puf6p , it seems clear from both studies that RNA binding by Puf6p is less specific than by other Pumilio/FBP family members [40] . It might be that Puf6p requires an additional co-factor to achieve the specificity reported for its in vivo function . Previous immunoprecipitation experiments with yeast extracts showed that the Puf6p:She2p complex is RNase-resistant , suggesting that both proteins might interact directly [21] . With recombinant proteins , we recapitulated an RNA-mediated interaction of Puf6p and She2p , but not a direct protein-protein binding ( Figure 1E ) . These results suggest that ASH1 mRNPs are rather RNase-insensitive , stable particles that allow for co-purification of both factors even after RNase treatment . When performing in vitro RNA-binding assays with Puf6p and She2p , we also observed that their combination failed to bind synergistically or with higher specificity to zip-code RNAs ( Figure 1D ) . Together , these experiments suggest that a complex consisting of ASH1 mRNA , She2p , and Puf6p is able to assemble in the nucleus with moderate specificity ( Figure 9 ) . Although unlikely , we cannot exclude the requirement of other factors not yet implicated in the assembly of specific nuclear mRNPs . In the cytoplasm , She3p is thought to link the RNA:She2p complex to the myosin motor Myo4p [17] , [18] . The prevailing model proposes that besides Myo4p , the translationally repressed ASH1-E3 complex harbors She2p , She3p , and Puf6p . To date , it had not been shown that such a complex really assembles on the RNA . In our in vitro assays , we reconstituted this Puf6p-containing complex in an ASH1-E3 RNA-dependent manner ( Figure S4 ) . Whether Khd1p behaves in a similar manner remains to be shown . A previous study reported that E . coli–expressed , crude-purified GST-She3p causes a supershift in an EMSA with GST-She2p and ASH1-E3 RNA [17] . From this single experiment , it remained unclear whether this supershift is based on protein-protein or protein-RNA interactions and whether the presence of She3p has implications on RNA-binding affinity and specificity . Diverging models have been proposed on the function of She3p [17] , [18] , [36] , including a scenario where She2p is not required for cytoplasmic ASH1-mRNA transport [36] . When assessing molecular interactions of She3p , we first identified a direct and specific binding to She2p with a Kd of 1 . 6 µM ( Figure 2A , B ) . This interaction suffices to explain the previously described supershift by She3p in EMSAs [17] . It is also consistent with a previous report where She3p was immunoprecipitated with a She2p mutant that lacks RNA-binding capacity [41] . However , the modest affinity and transient nature of the interaction between She2p and She3p and the low RNA-specificity of She2p appears insufficient to explain the specific localization of ASH1 mRNA observed in vivo . More surprisingly , we identified the myosin adapter She3p as a new , rather unspecific RNA-binding protein ( Figure 3A , B ) with affinities comparable to She2p and Puf6p . Database searches failed to identify any known RNA-binding motif in She3p . However , the most important finding for a mechanistic understanding of specific cargo recognition and mRNP assembly was our observation that She2p and She3p together form a highly specific ternary complex with all zip-code elements of localizing mRNAs tested in this study . EMSAs demonstrated that the She2p:She3p complex has at least 60-fold higher Kds for zip-code RNAs over unrelated RNA stem-loops ( Figures 3C–E and S5F–H ) . In addition , this RNA interaction stabilizes the rather weak binary She2p:She3p complex ( Figure 2B ) to a similar extent . Since for the control HIV-1 TAR RNA no binding was observed even at the highest experimental concentration ( Figure 3E ) , the true difference in Kd might be even higher . These findings were confirmed by a total of 12 mutations in She2p and She3p , in which for seven of them an impaired complex formation was observed ( Table 1 ) . Two of these She2p mutants were also tested in vivo , where they showed a total loss of RNA localization ( Figure 7 ) . Interestingly , four of the mutations in She2p and She3p with a defect in synergistic complex formation also showed defects in both binary interactions , RNA and protein binding . This observation suggests that both interactions are spatially and mechanistically intertwined to allow for the synergism described in this study . This conclusion is further supported by UV cross-linking experiments , which demonstrate that She2p and She3p both directly bind to RNA in the ternary complex ( Figures 3F , 4F , and S7 ) . No known RNA-binding motif can be identified in She3p . We nevertheless found a C-terminal fragment of 92 amino acids to be sufficient for synergistic RNA binding with She2p . Within this 92 amino acids long fragment , we mapped residues that are of functional importance for the synergism . Our subsequent UV cross-linking/mass-spectrometry experiment with She2p , RNA , and the C-terminal fragment of She3p confirmed that a functionally important subfragment of 20 residues is indeed involved in RNA binding ( Figures 4F and S7 ) . These cross-linking experiments also showed that the helix E of She2p directly interacts with RNA . The latter finding confirms our observations on the functional importance of the helix E and assigns a direct molecular function , i . e . RNA binding , to this protein region . A surface plot of all She2p-surface regions that are important for RNA binding shows a large continuous RNA-interaction surface ( Figure S12 ) . In a recent study , a She3p-dependent ASH1 mRNA-transport system was reported in Candida albicans [42] . Although in this yeast species no clear She2p homolog could be identified , ASH1 mRNA is transported in a fashion similar to S . cerevisiae . Thus , the question arose of how cargo binding is achieved . Our identification of S . cerevisiae She3p as an RNA-binding protein suggests that C . albicans She3p could mediate at least part of the RNA binding for mRNP assembly and transport . A sequence alignment of She3p from different yeast species reveals that the C-terminal half of this protein shows great differences between species with and without She2p in their genomes ( Figure S13 ) . As shown in this study , the C-terminal part of S . cerevisiae She3p binds to She2p ( Figure 4B ) [17] , [18] as well as to RNA ( Figure 4C ) and is required for synergistic RNA binding with She2p ( Figure 4D–F ) . It therefore appears likely that the She2p-lacking species have optimized the C-terminal She3p sequence for an interaction with a different RNA-binding protein or even for a more specific RNA-binding by She3p itself . Previous studies suggested that She3p acts only in the cytoplasm [17] , [18] , [21]–[23] , [36] . However , this assumption had not been rigorously tested before . We used a nuclear export mutant to show that She3p indeed does not shuttle into the nucleus ( Figure 8A ) . To scrutinize whether a small sub-fraction of She3p might play a role in the nucleus , we also analyzed the recently reported co-transcriptional recruitment of She2p to chromatin by ChIP experiments [20] . Since we observed no significant difference of She2p occupancy in wild-type and Δshe3 strains , a nuclear role of She3p could be further excluded ( Figure 8B ) . Shen et al . also reported a reduction of She2p chromatin binding in ChIP experiments after RNase treatment selectively at open reading frames ( ORFs ) of localizing mRNAs . The authors concluded that part of the chromatin-associated She2p interacts selectively with localizing mRNAs already during transcription . However , we could not confirm this observation . Although we also detected a reduction of She2p-dependent enrichment after RNase treatment , this effect was observed for all transcripts ( Figure 8B ) . Thus , our ChIP experiments suggest rather unspecific She2p association with nascent transcripts . They further indicate that She3p does not play a functional role in the nucleus . In summary , our and previous data suggest the following model: First , She2p binds co-transcriptionally to RNA polymerase II and to nascent transcripts ( Figure 9A ) . After transcription , nuclear mRNAs are bound by She2p and Puf6p with only limited specificity ( Figure 9B ) , followed by a nuclear export of both proteins together with mRNAs ( Figure 9B , C ) . In the cytoplasm , She2p and localizing mRNAs form a highly specific co-complex with myosin-bound She3p ( Figure 9C , D ) . This transport complex mediates the translocation of cargo mRNAs to the bud cell ( Figure 9D ) , where after anchoring at the bud tip translation is activated . On one hand , She2p and She3p function together exclusively in the cytoplasm to select zip-code RNAs . On the other hand , the mRNA cargo itself substantially stabilizes the She2p:She3p interaction . Because this interaction brings together the She2p-dependent pre-mRNP with the cytoplasmic motor complex , we conclude that it is the mRNA cargo itself that triggers joining of all components into the mature transport complex . This interpretation is fully consistent with the observation that RNase-treated ASH1-mRNPs do not have motile activity in vitro [27] . We propose that coupling of specific mRNA recognition and assembly of stable transport complexes constitutes a critical quality control step to ensure that only target mRNAs are transported . Previous publications reported only moderate in vitro selectivity for localizing mRNAs in yeast and Drosophila ( e . g . between 3- and 7-fold higher Kd for localizing RNAs ) [6] , [7] . Thus it remained ambiguous what difference in affinity to localizing and non-localizing RNAs might be required for specific mRNA transport and how highly specific mRNP assembly is achieved in vivo . In comparison , the ternary complex formation described in this study shows an unprecedented selectivity for zip code containing RNAs . In our understanding , this observation gives a more realistic example for the cargo specificity required for mRNA localization . It also demonstrates that co-complexes of transport factors might play a much more important role in the recognition of transcripts than previously assumed . Last but not least , our finding that the cargo RNA itself triggers the incorporation of all protein-core factors into one mature transport complex provides a new mechanistic paradigm for the assembly of RNA-localization complexes . An important question arising from this study is whether synergistic binding to RNA cargo is a more general feature in eukaryotes . For instance , during the oocyte-to-embryo transition of Drosophila development , the RNA-binding protein Staufen is involved in the localization of bicoid mRNA . In vitro , Staufen yielded strong binding to specific as well as to control RNAs with extensive secondary structures [43] . In contrast , in vivo injection experiments in Drosophila embryos showed that Staufen-containing mRNPs only form and localize efficiently when its native target , the bicoid 3′UTR , is injected [44] . Thus , it might well be that those Drosophila complexes also require mRNP assembly for specific mRNA binding and localization . In the past , in vitro reconstitution of molecular assemblies has been very successful to provide new insights into biological processes as diverse as transcription and membrane fusion [45] . By showing that specific mRNA recognition and assembly of stable cytoplasmic transport complexes is coupled , we demonstrate that in vitro reconstitution is also well suited to advance our mechanistic understanding of mRNA-transport .
Detailed information on plasmids , oligonucleotides , and yeast strains are found in Tables S1–S3 . Information on cloning and generation of yeast strains are summarized in Text S1 . She2p and Puf6p were expressed as GST-fusion in E . coli BL-21 Star ( DE3 ) cells ( Invitrogen ) and purified using standard techniques [19] , including affinity , ion-exchange , and size-exclusion chromatography columns . His-She3p ( 334–425 ) and His-She3p ( 354–425 ) were expressed in fusion with either GST or MBP in E . coli BL-21 Star ( DE3 ) cells and purified via affinity and size-exclusion chromatography . The GST-tag was cleaved with PreScission protease ( GE Healthcare ) during purification , unless stated otherwise . Full-length His-She3p and His-She3p point mutants were co-expressed with She2p using the Bac-to-Bac system ( Invitrogen ) . Recombinant baculovirus was amplified in Sf21 cells and used to infect High Five insect cells . Cells were cultured for 60–70 h . After sonication , His-She3p was purified using HisTrap , HiTrap Q , HiTrap Heparin , and Superose 6 10/300 GL columns ( GE Healthcare ) . She2p and nucleic acids were removed by extensive washing with 1 M NaCl-containing buffer . At the last step of purification , RNA-binding proteins were assessed by measuring the ratio of A260 to A280 to exclude contaminations by nucleic acids . Myo4p tail was purified as previously described [46] . RNAs for filter-binding assays and EMSAs were produced by either in vitro transcription or total chemical synthesis ( Table S4 ) . In vitro–transcribed RNAs were purified using native PAGE . For pull-down assays and analytical size-exclusion chromatography experiments , ASH1-E3 RNA was fused to the E . coli initiator tRNA ( Met ) ( E3-33-tRNA , E3-77-tRNA , and E3-118-tRNA; Table S4 ) , expressed in E . coli JM101 cells ( New England Biolabs ) , and purified by ion exchange chromatography essentially as described [34] , [35] . RNA filter-binding assays were essentially performed as described [33] . Serial She2p dilutions were incubated with 0 . 5 nM of radiolabeled RNA in Dot-Blot buffer ( pH 7 . 4 ) ( 20 mM HEPES ( pH 7 . 5 ) , 150 mM NaCl , 2 mM MgCl2 , 2 mM DTT , and 30 µg/ml yeast tRNA ) . Four-fifths of the reaction mixture was applied on the nitrocellulose membrane and radioactivity retained was measured by phosphoimaging . Kd calculation was performed by plotting the fraction of bound RNA versus the protein concentration and applying the Langmuir isotherm . The following serial dilutions were used: wild type–She2p binding to HIV-I TAR ( 16 bases , 57 bases ) or U1snRNA hairpin II: 0 to 12 µM; She2p-mutant binding to bud-localizing RNAs: 0 to 16 µM; She2p ( ΔC ) -binding to ASH1-E2A element: 0 to 32 µM; She2p-mutants binding to HIV-I TAR ( 16 bases ) , U1snRNA hairpin II , poly ( A ) 20 RNA: 0 to 32 µM; and She2p-F195A-L196A and She2p-Q197A-E198A-I199A to ASH1-E3: 0 to 2 µM . Standard deviations were calculated from at least three independent experiments . To show synergistic RNA-binding effects , serial She2p dilutions ( 0 to 320 nM for ASH1-E3 , 0 to 2 . 67 µM for EAR1 ) were incubated with 25 nM She3p and 0 . 5 nM 32P-RNA in Dot-blot buffer ( pH 7 . 9 ) and treated as described above . For data analysis , the RNA-binding signal of 25 nM She3p alone was subtracted as background . Signal intensities of She2p:She3p:RNA complexes were plotted against the respective She2p concentrations and normalized to the relative signal intensities of the corresponding She2p:RNA complexes . In a 20 µl reaction , protein was incubated with 5 nM of 32P-labeled RNA oligonucleotide in HNMD-buffer ( 20 mM HEPES ( pH 7 . 8 ) , 200 mM NaCl , 2 mM MgCl2 , 2 mM DTT ) supplemented with 4% ( v/v ) glycerol and 30 µg/ml ( EMSAs with She2p and She3p ) or 100 µg/ml yeast tRNA ( EMSAs with She3p or Puf6p only ) for 30 min at 20°C . For competition experiments , excess of unlabeled competitor RNA was added to the reaction after 30 min preincubation and incubated for another 15 min at 20°C . Additionally , HNMD-buffer supplemented with 100 µg/ml yeast tRNA was used . Protein:RNA complexes were resolved by PAGE ( native 4% gel in 0 . 5× TBE running buffer , 90 V for 45 min at 20°C ) . Gels were scanned with a Storm Scanner and analyzed using the software ImageQuant . In 20 µl reactions , recombinant She2p and She3p at indicated amounts were incubated with 5 nM 32P-labeled ASH1-E3-51 RNA for 20 min at 20°C in Dot-Blot buffer ( pH 7 . 9 ) . Subsequently , UV cross-linking was performed for 15 min on ice , using a Spectrolinker XL-1500 ( Spectroline ) with an average intensity of 2 , 500 µW/cm2 . 18 µl of each cross-linked sample were resolved by SDS-PAGE . Gels were scanned and analyzed using the software ImageQuant . 100 µg of purified ternary complex ( ASH1-E3-51 RNA , full-length She2p , and His-She3p ( 334–425 ) ) in 200 µL HNMD-buffer were cross-linked for 10 min at 254 nm . After digestion of the sample with RNase A , RNase T1 , and trypsin , cross-linked peptides were enriched on a TiO2 column [37] , [38] and analyzed by Nano-LC-ESI-MS . Data analysis for the identification of cross-linked peptides was carried out essentially as described earlier [37] . A detailed description is provided in Text S1 . Chromatography was performed with a Superose 6 10/300 GL column . 20 µM wild-type or mutant She2p , 30 µM She3p and 8 . 5 µM RNA ( E3-33-tRNA or E3-118-tRNA ) were preincubated for 5 min at 20°C . Then 200 µl samples were loaded on the column in HNMD-buffer ( flow rate: 0 . 5 ml/min ) . Fractions were analyzed by SDS PAGE ( Coomassie blue staining ) and 2% agarose gels ( in 1× TBE ) stained with GelRed DNA Stain ( Biotium ) . In a volume of 100 µl , 7 . 5–10 µM of each protein and , if applicable , 5 µM ASH1-E3-tRNA-fusion were incubated in HNMD-buffer with 50 µl resin ( nickel sepharose for His-She3p; glutathione sepharose for GST-Puf6p ) for 30–60 min at 4°C on a rotating wheel . Binding reactions were pelleted and washed four times with 200–500 µl of HNMD-buffer , followed by a final washing step with 50 µl . For His-tagged She3p , HNMD-buffer for binding and washing was supplemented with 30–50 mM imidazole . Bound proteins were eluted with 50 µl of elution buffer ( nickel pull-downs: HNMD-buffer supplemented with 750 mM imidazole; GST pull-downs: 30 mM reduced glutathione ) . One-tenth of the input , 1/5 of the final wash , and 1/5 of the elution fraction were analyzed by SDS-PAGE and Coomassie blue staining . RNA content was analyzed with 2% agarose gels ( in 1× TBE ) stained with GelRed DNA Stain ( Biotium ) . All experiments were performed using a Biacore 3000 system with a CM-5 chip ( Biacore ) in HNMD-buffer . She3p was covalently bound to the chip surface by standard amine coupling . For measurements with She2p ( wt ) and She2p ( ΔC ) concentrations from 0 . 123 to 30 µM were used . She2p ( ΔhE ) binding was probed with up to 150 µM protein . All binding signals were below 300 response units . Experiments were performed as duplicates and double referencing was applied . Kds were derived from steady-state measurements , applying the Langmuir isotherm . In situ hybridization was performed using TexasRed-conjugated antisense oligonucleotides against ASH1 mRNA [47] . Immunostaining was performed either using polyclonal anti-She2p antibody [7] and goat anti-rabbit antibody coupled to AlexaFluor488 or using 3F10 anti-HA antibody ( Roche ) directed against She3-6×HA and goat anti-rat coupled to Alexa594 . DNA was stained with Hoechst Stain Solution ( Sigma ) . Cells were inspected with an Olympus BX60 or Zeiss CellObserver fluorescence microscope and images acquired on a Hamamatsu OrcaER or Zeiss MRM CCD camera using the Openlab 4 . 0 software ( Improvision ) or Axiovision ( Zeiss ) software . In each experiment , at least 100 budding yeast cells were counted ( see text ) . Expression levels of She2p were analyzed by Western blotting using the monoclonal anti-She2p antibody She2p-1C3 [33] . Co-immunoprecipitation of Myc-tagged She2p was performed using monoclonal anti-Myc antibody ( 9E11 , Acris Antibodies ) coupled to magnetic Protein G beads ( Invitrogen ) [33] . For quantification , Western blots of two independent experiments were analyzed using the LAS-3000 mini system and Multi Gauge software ( FUJIFILM ) . Background signals from Δshe2 lanes were subtracted from respective Western blot signals . Subsequently , She2p signals from the supernatant fractions were subtracted from input fractions and used to normalize co-immunoprecipitated She3p against the bead-bound fraction of She2p-Myc3 . ChIP experiments were performed as previously described [48] with the following exceptions . Yeast strains containing a TAP-tagged version of She2p were grown in 100 ml YPD medium to mid-log phase ( OD600∼0 . 8 ) . To increase the fraction of cells with ASH1 transcription , yeast cultures were treated with nocodazole ( final concentration: 15 µg/ml ) for 2 h and washed once with 100 ml YPD as described [20] before cross-linking with formaldehyde ( final concentration: 1% ) . The additional RNase treatment step in the ChIP protocol required a reduction of the cross-linking time from 20 to 5 min . For better comparison , ChIP experiments without RNase treatment were treated identically . The RNase treatment was performed essentially as described [49] . Briefly , the chromatin was treated with 7 . 5 U of RNase A and 300 U of RNase T1 ( RNase Cocktail Enzyme Mix; Ambion ) . After incubating at room temperature for 30 min , immunoprecipitations were performed . Expression levels of She2p-TAP were analyzed by Western blotting using anti-TAP antibody ( Sigma ) and monoclonal antibody anti-She2p-1C3 [33] . The experiments as well as the analyses were performed as previously described [48] , [50] . Input and immunoprecipitated ( IP ) samples were assayed by qPCR to assess the extent of protein occupancy at different genomic regions . Primer pairs were designed for three different open reading frame ( ORF ) regions of ASH1 , a distinct region within the ORF regions of PMA1 and FBA1 , the promoter , coding and terminator regions of ADH1 , as well as for a heterochromatic control region of chromosome V . The PCR efficiencies were determined prior to ChIP experiments and were in the range of 95%–100% . The sequences and the location of the primers used in this study are shown in Table S5 . | In eukaryotes , the majority of cells are asymmetric and a way to establish such polarity is directional transport of macromolecules along cytoskeletal filaments . Among the cargoes transported , mRNAs play an essential role , as their localized translation contributes significantly to the generation of asymmetry . To date , hundreds of asymmetrically localized mRNAs in various organisms have been identified . These mRNAs are recognized by RNA-binding proteins and incorporated into large motor-containing messenger ribonucleoprotein particles ( mRNPs ) whose molecular assembly is poorly understood . In this study , we used the well-characterized process of ASH1-mRNA transport in Saccharomyces cerevisiae to address the question of how localizing mRNAs are recognized and specifically incorporated into mRNPs . Surprisingly , we found that the previously implicated mRNA-binding proteins She2p and Puf6p do not bind to cargo mRNAs with high specificity . Instead , the cytoplasmic motor-adapter protein She3p is responsible for synergistic cargo binding with She2p and for the stable incorporation of specific localizing mRNA into the transport complex . We propose that the specific recognition of localizing mRNAs happens at the very last step of cytoplasmic mRNP maturation . Other organisms might employ similar mechanisms to establish cellular polarity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/morphogenesis",
"and",
"cell",
"biology",
"molecular",
"biology/translational",
"regulation",
"molecular",
"biology/rna-protein",
"interactions",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines"
] | 2011 | A Cytoplasmic Complex Mediates Specific mRNA Recognition and Localization in Yeast |
Recent studies have identified thousands of regions in the genome associated with chromatin modifications , which may in turn be affecting gene expression . Existing works have used heuristic methods to investigate the relationships between genome , epigenome , and gene expression , but , to our knowledge , none have explicitly modeled the chain of causality whereby genetic variants impact chromatin , which impacts gene expression . In this work we introduce a new hierarchical fine-mapping framework that integrates information across all three levels of data to better identify the causal variant and chromatin mark that are concordantly influencing gene expression . In simulations we show that our method is more accurate than existing approaches at identifying the causal mark influencing expression . We analyze empirical genetic , chromatin , and gene expression data from 65 African-ancestry and 47 European-ancestry individuals and show that many of the paths prioritized by our method are consistent with the proposed causal model and often lie in likely functional regions .
Discerning the genetic and molecular basis of complex traits is a fundamental problem in biology . Genome-wide association studies have revealed that the majority of variants associated with disease lie in noncoding regulatory sequences [1 , 2] . Identifying the target genes of these variants and the mechanisms through which they act remains an open problem [3] . Recent efforts to systematically characterize how genetic variation impacts more granular molecular phenotypes have yielded thousands of single nucleotide polymorphisms ( SNPs ) that associate with local and distal histone modifications—termed histone quantitative trait loci ( hQTLs ) [4–7] . Furthermore , recent studies have identified many expression quantitative trait loci ( eQTLs ) that co-localize with hQTLs , implying there may exist a shared genetic influence on epigenetic traits and gene expression [8–11] . Therefore , one proposed mechanism by which regulatory variants may affect gene expression and thereby impact traits is through changes in chromatin state [10] . However , this putative chain of causality whereby the effects of SNPs on expression are mediated by chromatin modifications has yet to be established . This is further compounded by the complex space of plausible causal directions connecting transcription factor binding , DNA methylation , chromatin variation , and gene expression . Since laboratory experiments are very costly , there is a need for statistical methods that can accurately prioritize the causal SNP and chromatin mark within an implicated region under a plausible causal model . However , even if the causal direction is given , pinpointing the exact SNP and mark within a genomic region is very challenging due to the confounding effects of linkage disequilibrium ( LD ) among SNPs and correlations among marks [5 , 6 , 10 , 12–14] . Methods to investigate the relationships between the genome , the epigenome , and expression have largely focused on quantifying the overlap between hQTLs and eQTLs [10 , 14 , 15] . Previous studies have sought to identify hQTLs by selecting the SNP with the strongest p-value for association to a local chromatin mark and to local gene expression [10 , 14 , 15] . Moreover , various methods exist for the fine-mapping of SNPs that may be concurrently affecting two traits , including eCAVIAR [16] and Coloc [17] . Although these methods can be applied to jointly analyze SNP , chromatin , and expression data , they do not model the causal path whereby SNPs impact expression through chromatin alteration . Here we propose a fine-mapping framework , pathfinder , that explicitly models the hierarchical relationships between genome , chromatin , and gene expression to predict both the causal SNP and the causal mark within a gene region that are influencing expression of a given gene . Our framework assumes a causal model where a SNP impacts a chromatin which in turn alters gene expression . In our framework we refer to a “causal” SNP as any SNP that disrupts inter-individual variation of chromatin state either through a direct biological mechanism ( e . g . , chromatin accessibility ) or indirectly through an unobserved biological mechanism . Similarly , we refer to a “causal” chromatin mark as either a mark that biologically alters expression or that tags an underlying epigenetic regulatory mechanism of expression . Our framework takes as input the strength of association ( as quantified through the standard Z-scores ) between all SNP/mark pairs and all marks to expression as measured in a given set of individuals . To explicitly account for the correlation structure among SNPs and marks , we use a Matrix-variate Normal distribution to model all Z-scores jointly . By construction , this allows our probabilistic model to assign posterior probabilities for each SNP , mark , and path ( where paths include all possible SNP-mark combinations ) to be causal in the region . A key advantage of our approach is that it produces well-calibrated posterior probabilities for causality . Thus , pathfinder can be used to prioritize variants and marks for validation experiments . In simulations we compare against several existing methods , demonstrating that pathfinder outperforms alternative approaches with respect to both accuracy and calibration . This is largely because our comparators do not take into account mark-expression associations . In some cases , these additional associations may help distinguish between two potentially causal paths that have comparable evidence for causality . For example , in cases where a SNP is associated with expression of a local gene and is also associated with two local chromatin marks , knowledge of the impact of each mark on gene expression may help distinguish between two possible paths for causality . Finally , we analyze genotype , chromatin and expression data from 65 African-ancestry and 47 European-ancestry individuals . We show that the top causal SNPs proposed by pathfinder tend to lie in more functional regions and disturb more regulatory motifs than expected by chance . We also present evidence that most of the top paths reported by pathfinder demonstrate consistency with our proposed sequential model , thus strengthening the case for our method’s applicability to empirical biological data .
Here we introduce a hierarchical statistical method for fine-mapping of causal SNPs and chromatin marks ( e . g . , histone modifications ) that may be concordantly influencing gene expression within a genomic region . We build upon previous insights that a vector of Z-scores is well-described by a Multivariate Normal ( MVN ) distribution parameterized by LD [13 , 18 , 19] to model association statistics between chromatin marks and gene expression . We analyze all chromatin peaks across four mark types ( DHS , H3K4me1 , H3K4me3 , and H3K27ac ) jointly in the same framework; we refer to a “mark” as a chromatin peak at a particular location , and “mark types” as DHS , H3K4me1 , H3K4me3 , and H3K27ac . To simultaneously take into account both SNP LD and the correlations between chromatin marks , we use the Matrix-variate Normal distribution to jointly model association statistics between all SNPs and marks within a region . Our method takes as input SNP-mark and mark-expression associations within a region centered around a particular gene , as well as correlations among all SNPs ( LD ) and correlations among all considered marks . Pathfinder enumerates over all possible causal paths , considering one causal SNP and one causal mark for each path , and outputs a posterior probability for each path to be causal , which can subsequently be used to prioritize SNPs and marks for validation . We compute marginal probabilities for individual SNPs ( or marks ) to be causal by summing the posterior probabilities over all paths that contain the SNP ( or mark ) . For simplicity , in this work we refer to a “causal” mark as a mark that either causally drives inter-individual variation of gene expression or is correlated to an underlying causal mechanism ( e . g . transcription factor binding ) , though it may not be biologically causal for expression . The advantage of our method over existing approaches is that it integrates mark-expression associations which may help distinguish between two paths with otherwise comparable evidence for causality . We illustrate a scenario in Fig 1 . Consider a genetic region where SNP g1 has a strong association with two local marks h1 and h2 , as well as a significant association with gene expression . Using only SNP-mark and SNP-expression effects , we are unable to discern whether SNP g1 influences expression through mark h1 or h2 . However , if we consider mark-expression effects , we see that mark h1 has a strong association with gene expression where mark h2 does not . This additional information helps support the hypothesis that there is a causal path from SNP g1 to mark h1 to gene expression . We used simulations to compare pathfinder’s performance against alternative methods with respect to SNP- , mark- , and path-finding efficiency as well as the calibration of its posterior probabilities . We generated genetic , chromatin , and gene expression data for 10 , 000 50kb regions , each centered around a single gene , over 100 individuals , using SNP LD and mark correlations derived from 65 Yoruban ( YRI ) individuals ( see Methods ) . We define a “mark” as an individual peak location for any mark type in the dataset ( DHS , H3M4me1 , H3K4me3 , or H3K27ac ) . For each gene , we randomly assigned a single causal pathway from one SNP to one mark to gene expression . We then ran our methods on all regions individually and assessed their ability to correctly prioritize the true causal path in each region ( Methods ) . We compare against an independent fine-mapping approach ( whereby we fine-map SNP-mark associations and mark-expression associations independently and take the product of the resulting probabilities to produce posterior probabilities for paths ) , a Bayesian network analysis [20] , a naive ranking ( where we rank SNP-expression and mark-expression associations to prioritize SNPs and marks within a region; for path-finding , we rank the product of these two ) , a formal colocalization method [17] , and finally , against overlaps between eQTLs and hQTLs within a region centered around a gene of interest ( see Methods ) . Unlike the first four approaches , the overlap methods do not produce rankings , but yield candidate sets of causal SNPs , marks , and paths . For this reason , we present these results in a separate analysis using an alternative metric for comparison . We find that pathfinder has consistently better performance than the other ranking approaches with respect to all three features—SNP- , mark- , and path-mapping within a region ( Fig 2 ) . For example , association ranking , Coloc , Bayesian network analysis , and independent fine-mapping accumulate 55% , 62% , 47% , and 13% of the top paths on average in order to recapture 90% of the causal paths , whereas our method only requires 8% of the top paths . Note that SNP-expression association ranking is equivalent to running a basic eQTL analysis , which does not take into account chromatin data , in order to identify causal SNPs . A similar improvement in accuracy was observed for the size of the credible sets , defined as the number of SNPs required to capture a given percentage of the causal variants ( S1 Table ) . Next , we evaluated pathfinder’s performance compared against standard analyses that investigate overlaps between hQTLs and eQTLs within a genomic region . In such experiments , the variant with the strongest association to each local chromatin mark is selected , as well as the variant with the strongest association to local gene expression . In addition , marks are filtered to ensure a 10% FDR ( see Methods ) . This produces a set of candidate marks , as well as one candidate SNP per mark , and one SNP deemed causal for gene expression in the region . Implicitly , the overlap of these variants suggests a set of candidate SNPs , marks , and paths for the region . For the same set sizes , pathfinder identifies 96% of the causal marks versus 74% in the standard overlap approach ( Fig 3 ) . SNP-finding accuracy is comparable between the two methods . We next assessed the calibration of the posterior probabilities for causality output by pathfinder . Our method has slightly deflated credible sets for SNP- and path-finding , but well-calibrated credible sets for mark-finding ( Fig 4 ) . In contrast , the independent fine-mapping approach has consistently inflated credible sets—that is , it captures more causal paths than expected , but also has drastically larger credible set sizes . For example , when accumulating 90% of the posterior probabilities over all regions , pathfinder captures 88% of the true causal paths within the top 380 candidate paths , whereas independent fine-mapping captures 94% of the causal paths within the top 1026 candidate paths . Similar outcomes were attained for the 50% and 99% credible sets ( S1 Fig ) . Overall , pathfinder’s credible sets are less biased and narrower than those obtained through the independent fine-mapping approach . Finally , we investigated the effects of simulation and method parameters on pathfinder’s accuracy . Firstly , we varied the causal SNP and mark effect sizes such that the variance explained of mark and gene expression ranged from 0 . 1 to 0 . 5 . As anticipated , increased heritability leads to better performance ( See Fig 5A–5C ) . Secondly , in order to assess the impact of SNP LD and mark correlations on SNP- and mark-finding performance , we stratified our existing simulations based on the mean correlation of the causal SNP or mark to all other SNPs or marks ( See Fig 5D–5I ) . We grouped our simulations into three categories: low , medium , and high correlations . As anticipated , SNP-finding performance decreases slightly as SNP LD increases . Notably , mark-finding performance is actually improved at higher SNP LD . This is due to the redundancy in information about SNP-mark associations at the causal mark when these effects are exhibited across multiple correlated SNPs . SNP- and mark-finding performance , however , do not seem to be significantly affected by mark correlations in our simulations—at least not at the level of variation exhibited in our data . In addition to stratifying our existing simulations by LD , we also assessed the impact of using European rather than African LD in the same regions , as European LD is known to be more extensive . Here we retained the YRI mark and expression data in order to isolate the effect of SNP correlations . The credible set sizes computed from the CEU dataset do not substantially differ from those obtained in YRI ( S2 Table ) . This result demonstrates that the more extensive LD observed in European individuals will not significantly affect pathfinder’s performance . Thirdly , we evaluated the effect of the prior variance tuning parameter on fine-mapping performance ( See Fig 5J–5L ) . The prior variance is an estimate of the variance explained by the causal SNP and mark in the region , as we do not know a priori what the causal effect sizes are . We show that the optimal range for the prior variance parameters is between 5 and 10 , in simulations with a variance explained of 0 . 25 on both levels . Overall , performance does not seem to change drastically in response to variations in the prior variance , even significantly outside of this optimal range . Our hierarchical model makes several key assumptions that may sometimes be violated in empirical data . Firstly , pathfinder assumes that a single causal SNP and a single causal mark are driving the associations within a region , where in reality there may exist multiple true causal SNPs or marks [13 , 19] . Secondly , pathfinder assumes that SNP effects on gene expression are mediated by a chromatin mark , which may not be the case in real data . We therefore assessed the performance of our method when these two assumptions are violated in various ways , diagrammed in Fig 6 . First , we investigate violations 1–3 , which include multiple causal pathways throughout the region . Path-mapping accuracy , measured by the proportion of causal paths identified , is reduced in all three scenarios ( Fig 6 ) . Note that the number of causals identified does not necessarily decrease , but rather the proportion , as there are more causal paths in each region . SNP- and mark-finding accuracy under these violations are also compromised , but with two notable exceptions . In the multi-causal-SNP scenario , mark-finding accuracy increased in comparison with the single-SNP simulations; for example , only 8% of marks were selected ( versus 18% in the single causal simulations ) to capture 90% of the causal marks . In the multi-causal-mark scenario , SNP-finding accuracy increased . Intuitively , this is due to the redundancy in the signal that is captured by the Matrix-variate Normal distribution . We next investigate violations 4–5 , in which an additional SNP or mark influences gene expression directly . We observe in these two scenarios that performance is reduced for SNP- , mark- , and path-finding , but not drastically . For example , in order to capture 90% of the causal paths , pathfinder must select on average 25% and 28% of paths under violations 4 and 5 , respectively ( compared with 15% under standard simulations ) . Because anti-correlated marks ( e . g . activating and repressing marks ) often tend to act in the same region , we also assess pathfinder’s behavior specifically when two marks have opposite effects on expression . As expected , pathfinder’s performance does not decline in the presence of anti-correlated peaks ( S2 Fig ) . Finally , we discuss pathfinder’s performance under violations where the causal order is modified ( violations 6–7 ) . Under violation 6 , where a single causal SNP affects gene expression directly , which in turn affects a single mark , pathfinder actually captures a higher proportion of the affected marks and overall paths . For example , in order to capture 90% of the causal paths , pathfinder must select on average only 3% of the top-ranked paths ( compared with 15% under standard simulations ) . In violation 7 , where the SNP has independent effects on the mark and the gene expression , we show that pathfinder’s accuracy in finding the causal mark and path is significantly reduced . Note that in this case , the “path” is not truly a path but a SNP/mark pair , as effects of the SNP on mark and gene expression are independent . Our power in distinguishing between these two models depends on the prior variance explained parameter . Under violation 7 , the variance explained in gene expression by the causal mark is much smaller than expected , thus reducing our confidence in the true causal configuration . We conclude that under the SNP→expression→mark violation , pathfinder will identify causal paths very confidently even if they do not follow the assumed SNP→mark→expression model . Therefore a high posterior probability for a path may not be sufficient evidence for causality . On the other hand , when SNP effects on mark and expression are independent , pathfinder is less likely to produce false positives . For these reasons , we recommend a pre- or post-filtering step to retain only those regions that show some prior evidence for the SNP→mark→expression model using a conditional analysis or partial correlation approach ( Methods ) . For completeness , we also assess existing methods under these simulations ( S3 Fig ) . Most notably , the simple association-ranking approach shows a distinct improvement under violations 6 and 7 , in which SNPs have a direct effect on gene expression . This is expected as pathfinder assumes the causal effect to be mediated by chromatin . A similar improvement can be observed for Coloc under violation 7 , in which the SNP affects both chromatin and gene expression directly . We evaluated the behavior of our hierarchical fine-mapping method when applied to empirical data . We performed these analyses on data from 65 YRI individuals whose genotypes were obtained through 1000 Genomes , and whose PEER-corrected H3K4me1 , H3K4me3 , H3K27ac , DHS , and RNA expression levels in lymphoblastoid cell lines ( LCLs ) were obtained from [10] . In each region , we analyzed all four mark types jointly ( H3K4me1 , H3K4me3 , H3K27ac , and DHS ) by including all peaks spanning the region for each mark type . Each peak of each mark type was therefore treated as a single chromatin mark . We filtered the 14 , 669 regions using a two-step regression analysis to yield 1 , 317 regions that showed evidence for the sequential model of SNPs affecting histone marks which in turn affect gene expression ( see Methods ) . pathfinder’s runtime scales approximately as s3t3 , where s and t are the number of SNPs and marks within a region , respectively . On average , each 50kb region contained 160 SNPs and 25 marks . Most runs were completed in under a few minutes . The most dense region contained 331 SNPs and 66 marks and took approximately 21 minutes ( S4 Fig ) . In Table 1 , we report the average 50% , 90% , and 99% credible set sizes produced when running pathfinder on real data . We compare against basic eQTL mapping , where we fine-map SNPs to gene expression ignoring chromatin data . We show that the credible set sizes are significantly narrower when running pathfinder with all three levels of data , consistent with our findings in simulations . For example , eQTL mapping requires an average of 45 . 3 SNPs in order to capture 90% of the posterior probability for SNP causality , whereas pathfinder only requires 28 . 4 SNPs . If we define a gene to be fine-mapped if 99% of the posterior probability mass for SNP causality is contained within the top 10 SNPs or fewer , then standard eQTL mapping fine-maps 46 of the genes in our data , whereas pathfinder fine-maps 73 of the genes . Notably , pathfinder also requires only 1 . 8 marks on average in order to capture 90% of the posterior probability for causal marks . In 82% of the regions where the top two marks capture more than 90% of the posterior probability , these two marks are two distinct peaks of the same mark type . The mean variance explained observed in the top path chosen by pathfinder , across all conforming regions , were 0 . 38 ( s . e . 0 . 01 ) for the SNP-mark effect and 0 . 20 ( s . e . 0 . 01 ) for the mark-expression effect ( S5 Fig ) . These effects are reasonably consistent with the 25% variance explained we used in simulations at each level ( see Simulations ) . The correlation between the SNP-mark and mark-expression effect size magnitudes in the top selected paths across all regions was 0 . 03 ( p = 0 . 400 ) . That is , the strength of the SNP-mark effect did not seem to correlate with the strength of the mark-expression effect . We assessed the relative impacts of each type of histone mark by computing the proportion of probability mass assigned to each mark type in aggregate over all regions ( S3 Table ) . H3K4me3 is the most informative mark type in this data , capturing 31% of the total probability mass despite being the least prevalent of all four mark types , constituting only 13% of all marks . We also report the size of pathfinder’s credible sets when applied to empirical CEU data rather than YRI in Table 2 . These two datasets are not directly comparable , as the types of epigenetic marks and their quantities differ substantially . Nonetheless , we demonstrate that pathfinder’s performance on the CEU dataset does not drastically diverge from its behavior in YRI . Data pre-processing strategies such as PCA and PEER correction may substantially impact the number of mark-expression correlations that are retained [21] . We find that credible set sizes for PEER-corrected data are narrower , giving a slight but significant improvement in performance ( S4 Table ) . As our pre-filtering step was designed to preserve only regions in which SNP effects on gene expression are mediated by chromatin , we expected a large majority of the analyzed regions to show evidence for this mechanism . To confirm this , we investigated whether the top paths prioritized by our method demonstrate consistency with this causal model . We defined a set of top paths as those which were ranked first in a region and whose posterior probabilities for causality were assigned by pathfinder to be greater than 0 . 1 . This resulted in 480 total top paths . Out of 480 top paths , only 12 had a significant ( p < 0 . 05/480 ) partial correlation between SNP and gene expression after controlling for chromatin . However , 193 paths had a significant partial correlation between SNP and chromatin after controlling for gene expression . This finding suggests that the top paths are more consistent with the SNP→mark→expression model than with a SNP→expression→mark model . Next we examined the relationship between the product of the effect sizes between SNP-mark and mark-expression against the overall SNP-expression association ( Fig 7 ) . We expect this relationship to be correlative; if truly mediated by the mark in question , the overall SNP-expression effect size should be proportional to the product of the two contributing effect sizes . Note that we weight our correlation by the reported posterior probability for each path , such that the paths we have more confidence in will contribute more to this metric . We find a high correlation ( r = 0 . 91 ) between these effect size vectors for our top paths , as compared with a correlation of r = 0 . 36 when running the same analysis on random paths within each region . This result indicates that pathfinder is identifying many pathways that are likely to be following its causal model . In Table 3 , we list the top ten paths prioritized by pathfinder across all real data regions . Most SNPs implicated in these paths are known to alter several regulatory motifs and often lie in an enhancer region or a promoter region of the genes whose expression they affect . 59% ( s . e . 2% ) of the SNPs implicated in the top paths fall into active ChromHMM states ( 1–7 ) in LCLs , including active TSS , flanking active TSS , transcription at gene 5’ and 3’ , strong transcription , weak transcription , genic enhancers , and enhancers . Only 47% ( s . e . 2% ) of random paths fall into these active states ( p = 0 . 001834 ) . Moreover , on average , SNPs in the top paths disturbed 5 . 35 ( s . e . 0 . 26 ) regulatory motifs , whereas random SNPs chosen at the same regions only disturbed 4 . 40 ( s . e . 0 . 20 ) motifs on average ( p < 0 . 001 ) . We did not , however , observe a similar change in transcription factor binding affinity at these motifs ( δ = 5 . 26 vs δ = 5 . 27 , ( p = 0 . 511 ) ) . As an example , in Fig 8A–8D , we display the genomic context for the top region reported by pathfinder , including average mark signals for DHS , H3K4me1 , H3K4me3 , and H3K27ac , stratified by genotype , in a 4kb region centered around the TSS of the NDUFA12 gene . The implicated SNP lies within the NDUFA12 TSS . Fig 8E plots the gene expression signal against that of the top mark , stratified by genotype . In S6 Fig , we show associations for the top region reported by pathfinder , spanning a 50kb region centered around the NDUFA12 TSS . Next we examined the spatial relationships between the SNP , mark , and TSS implicated in the top paths reported by pathfinder ( Fig 9 ) . SNP to mark and mark to TSS distances were significantly lower in our selected paths compared with randomly chosen paths at the same regions . The average distance from SNP to mark in pathfinder’s top paths was approximately 11 . 7kb , compared to 15 . 3kb in randomly chosen paths ( p < 0 . 001 ) . The average distance from mark to TSS in selected paths was approximately 8 . 6kb , compared to 9 . 7kb in randomly chosen paths ( p = 0 . 026 ) . SNP to TSS distances were not significantly different in top versus random paths ( p = 0 . 108 ) , with top SNPs lying on average 11 . 7kb away from the TSS and random SNPs lying 12 . 4kb away . 5% of top SNPs lied within 2kb of the TSS while 15% lied within 2kb of the corresponding peak . 23% of peaks in the top paths lied within 2kb of the gene TSS . S7 Fig displays all three distances where top paths are stratified by mark type . To further validate the top paths chosen by pathfinder , we determined the extent to which SNPs in these paths overlap with eQTLs that have been identified in LCLs using the larger scale Geuvadis data set [22] . 21% of the top paths contained SNPs that were also identified as eQTLs from the Geuvadis data set . In comparison , when randomly choosing paths at the same regions , only 11% overlapped with eQTLs ( p < 0 . 001 ) . Simply choosing the SNP with the highest association with gene expression in each region ( equivalent to standard eQTL-mapping ) resulted in an overlap of 24% with existing eQTLs . These results contradict the improvement in accuracy demonstrated in simulations when using pathfinder . We suspect this discrepancy is due either to imperfect locus ascertainment ( i . e . , a number of loci may include SNPs that directly affect gene expression rather than indirectly through chromatin ) or the fact that the Geuvadis eQTLs were also selected using standard fine-mapping approaches and we may thus expect a stronger agreement between the two resulting eQTL sets . We also investigated the extent to which pathfinder’s top SNPs overlap with eQTLs that have been experimentally validated through differential expression in an LCL dataset [23] . Here , we define the set of validated eQTLs to be those whose p-values for differential expression passed a threshold of 0 . 01 . We find that 2 . 2% ( or 13 ) of pathfinder’s top SNPs overlap with this validated set , where choosing the SNP with the highest association with gene expression in each region resulted in an overlap of 2 . 3% ( also 13 SNPs ) . Finally , we investigated whether any of the top paths reported by pathfinder could be found within GWAS hit regions for various autoimmune diseases , as our data were collected from LCLs . These autoimmune diseases included Celiac disease , Crohn’s disease , PBC ( Primary Biliary Cirrhosis ) , SLE ( Systemic Lupus Erythematosus ) , MS ( Multiple Sclerosis ) , RA ( Rheumatoid Arthritis ) , IBD ( Irritable Bowel Disease ) , and UC ( Ulcerative Colitis ) . We restricted to GWAS hits with variants associated to the trait with p < 5 × 10−8 . We found that 19 of our 480 top paths were contained in a GWAS-implicated region . In Table 4 , we report the paths that localized within autoimmune GWAS regions . In order to determine whether our top paths are truly enriched in GWAS regions , we established how many of these paths appear in an equivalent number of random regions that have not been implicated by an autoimmune GWAS . We centered each random region around a SNP that was matched for a similar MAF and LD score as the GWAS tag SNP . We ran this analysis 100 times to define a null distribution for the number of top paths found in a background region . We found that 19 out of 480 top paths was not a significant enrichment ( p = 0 . 44 ) .
In this work we proposed a hierarchical fine-mapping framework that integrates three levels of data—genetic , chromatin , and gene expression—to pinpoint SNPs and chromatin marks that may be concordantly influencing gene expression . A key contribution of our approach is the ability to model the correlation structure in the association statistics using a Matrix-variate Normal distribution . Our approach is superior to existing methods , demonstrating the advantage of using a probabilistic approach that takes into account the full sequential model . Moreover , pathfinder produces well-calibrated posterior probabilities , and is thus a reliable method for the prioritization of SNPs and marks for functional validation . We conclude by addressing some of the limitations of our method . Most notably , our method is based upon the SNP→mark→expression assumption . In many genomic regions that show simultaneous evidence for SNP to mark and SNP to gene expression effects , this model will not necessary hold true . In simulations , we show that under the SNP→expression→mark violation , pathfinder may identify causal paths very confidently , leading to false positives under the proposed model . When a SNP is in fact independently influencing a mark and gene expression , pathfinder is less likely to produce false positives . However , the risk of mis-appropriating our method in this way can be reduced by requiring genomic regions to show evidence for our causal model . We recommend a pre-filtering step before running pathfinder on real data that we outline in Methods . In our empirical data analyses , we demonstrate that this two-step regression robustly filters out non-conforming regions . We also acknowledge that , though there are multiple lines of evidence for SNPs influencing expression through local hQTLs , recent works have also emphasized the importance of interactions with distal hQTLs . Thus , developing a systematic way to incorporate data in distal regions with evidence for interactions with a local eQTL would be a fruitful direction . Moreover , pathfinder assumes that the true causal SNP and mark within a region are present in the data , which may not always be the case . In this scenario , pathfinder will instead place its confidence in the SNP or mark that best correlates with the missing causal SNP or mark in question . Similarly , many epigenetic marks are not themselves causal for gene expression , but are simply correlated to a causal event ( e . g . , transcription factor binding ) . It is also often the case that multiple marks at promoter and enhancer regions are concordantly acting to impact gene expression . In these cases , individual marks are not necessarily causal in themselves , but may be viewed as a cause for inter-individual variation or simply correlated to a causal factor . In this light , pathfinder aims to identify the epigenetically modifying region so that it can be tested experimentally and/or characterized functionally ( for example , to identify the effector transcription factor ) . We also note that pathfinder currently uses an approximation whereby the observed Z-score at the causal SNP is used to estimate the true NCP at the causal SNP ( Methods ) . We leave this to be addressed in future work; this correction will likely further improve the calibration of our method’s credible sets . We note that pathfinder only uses individuals for which we simultaneously have genetic , chromatin , and gene expression measurements , thus ignoring eQTL data that has been measured in larger sample sizes . However , eQTL data from larger samples could potentially be used as a prior for expectation of SNP causality or perhaps for validation after running pathfinder on real data . Finally , although our analyses showed that H3K4me3 marks are the most informative for fine-mapping , small data set sizes analyzed in this work prohibit us in making definitive conclusions on which mark is most useful leaving such avenues for future work .
For each individual , let h be the signal value for the causal histone mark and G be their vector of genotypes at a region containing s SNPs . Let E be the individual’s mRNA expression level for the gene at this region and H be a vector representing all t marks at the region , which contains h . Here we analyze all individual peak locations across all available mark types in a joint framework . As such , each of t individual marks represents one peak location for a particular mark type . Our causal framework can be modeled as: h = G β g + ϵ g ( 1 ) E = H β h + ϵ h ( 2 ) where ϵ g ∼ N ( 0 , 1 - σ g 2 ) and ϵ h ∼ N ( 0 , 1 - σ h 2 ) . The vector βg represents the allelic effects on the causal histone mark whose entries will be non-zero only at the causal SNP . The vector βh represents the histone mark effects on expression levels whose entries will be non-zero only at the causal histone mark . σ g 2 and σ h 2 represent the variance explained at the SNP-mark and mark-expression levels . We simulated data for 100 individuals over 10 , 000 50KB regions , using genotypes and LD from 65 YRI individuals obtained through 1000 Genomes [26] . SNP and mark correlations in our simulations were taken from the true correlations exhibited in these regions derived from these individuals . To determine causal status , we randomly chose one SNP and one mark to be causal in each region , thus defining a causal path through the data . Subsequently , we standardized genotypes and simulated values for chromatin marks and gene expression over all 100 individuals . In order to simulate correlations between histone marks as observed in our empirical data , we drew mark values from an MVN as N ( H i n d , ϵ g Σ h ) , where the means , Hind = HcΣh , c , represent the induced values on non-causal marks due to correlations with the causal mark . The mean mark values for the causal mark were generated for each of the 100 individuals as Hc = βgGc , where Gc is the genotype of the individual at the causal SNP , the effect size βg was drawn from a normal distribution , N ( 0 , σ g 2 ) , with variance set to the desired variance explained by SNPs on marks σ g 2 = 0 . 25 , with the error term ϵg set to 1 - σ g 2 . Finally , the individuals’ values for gene expression are computed as E = βhHc + ϵh , where Hc is the causal mark value as computed from the MVN , the effect size βh was set to the desired variance explained from mark to expression σ g 2 = 0 . 25 , with the remaining error term given by N ( 0 , 1 - σ g 2 ) . For simulations in which there were multiple causal SNPs or marks , we randomly drew m or p , the number of causal SNPs or marks , from a binomial distribution where the expected number of causals per region was set to 1 . However , we only included simulations with two or more causals . For multi-causal-SNP simulations , we then randomly selected m causal SNPs in the region and simulated chromatin marks and gene expression as described previously , but drew the effect sizes of each SNP as N ( 0 , σ g 2 / m ) , such that the total expected variance explained remained at 0 . 25 . For multi-causal-mark simulations , we randomly selected p causal marks in the region and simulated chromatin marks by defining the means , Hc , of each causal mark independently as described for the single-causal simulations . We then computed gene expression by drawing the effect size , βh , of each causal mark from N ( 0 , σ g 2 / p ) such that the total expected variance explained remained at 0 . 25 . We benchmark our method against five alternative approaches . Firstly , we compare against the standard overlap analysis whereby hQTLs and eQTLs are independently identified within a region centered around a gene . We follow the protocol outlined in [14] . In this experiment , we computed the best SNP association in each region with every mark measured in the region as well as with the gene expression value for that region . We determined adjusted p-values for each top association by performing permutation tests . We then accounted for multiple testing at the mark level by determining the minimum FDR at which each adjusted p-value would be considered significant . This was estimated via the qvalue package [27] . This procedure resulted in a set of significant SNP-mark associations , as well as one SNP-expression association within the region , as only the top SNP association is retained for each biological phenotype . We then evaluated the number of causal SNPs , marks , and paths that were ultimately included in these candidate sets . Secondly , we compared against the approach of independently fine-mapping the two levels of data ( SNP-mark and mark-expression ) , and multiplying together pairs of posterior probabilities to produce probabilities of causality for paths . For these independent fine-mapping experiments , we used a simple approach that assumes a single causal variant , approximating posterior probabilities for causality directly from Z-scores [28] . In addition , we compared against a basic ranking approach , where we independently computed SNP-mark , mark-expression , and SNP-expression associations for every SNP and mark within a region . For SNP and mark prioritization , we simply produced a ranking of the SNP-expression and mark-expression posterior probabilities for causality , respectively . For path prioritization , we produced a ranking of the product of SNP-mark and SNP-expression posterior probabilities . We next compared against a bayesian network model which computes directed association strengths between all possible pairs of nodes in a given network [20] . The method takes as input raw genotype and phenotype values . As nodes , we included all SNPs and marks , as well as the gene expression value , within a region . We allowed only for node pairings directed from SNP to mark or from mark to gene expression . For SNP and mark prioritization , we ranked association strengths over all directed SNP-expression edges and mark-expression edges , respectively . For path prioritization , we produced a ranking of the product of SNP-mark and mark-expression strengths . Finally , we compared against Coloc , which is designed to identify SNPs that are likely to be causal for multiple traits at once . Specifically , Coloc outputs a posterior probability that a SNP is causal for two arbitrary traits simultaneously . We adapted Coloc for our purposes by running the method on all SNPs independently . For each SNP , the two given traits were ( 1 ) gene expression , and ( 2 ) a mark value . Thus , we ran Coloc independently for all SNP-mark combinations . This produced a set of posterior probabilities indicating , for each SNP-mark combination , the likelihood that the SNP is causal for both the mark value and gene expression simultaneously . For path prioritization , we ranked these probabilities over all SNP and mark combinations . For SNP and mark prioritization , we marginalized over all marks and SNPs , respectively , producing posterior probabilities for each SNP and mark to be causal independently . The real data analyses were done on 65 YRI individuals whose genotypes were obtained through 1000 Genomes and standardized . PEER-normalized [29] H3K4me1 , H3K4me3 , H3K27ac , DHS , and RNA expression marks in lymphoblastoid cell lines ( LCLs ) for these individuals were obtained from [10] . For each gene in the dataset , we computed associations for every SNP-mark , SNP-gene , and mark-gene pair within a 50kb window centered around the gene TSS . On average , each region contained 160 SNPs and 25 marks ( across the four mark types—H3K4me1 , H3K4me3 , H3K27ac , and DHS—whose peak values we analyzed together in each region ) . Overall , from 14 , 669 50kb regions , we filtered for regions that exhibited evidence for our sequential model where SNPs affect chromatin marks , which in turn affect gene expression . Specifically , for each region we performed a two-stage regression where we first regressed gene expression on all chromatin marks , and ( 2 ) regressed the proportion of expression explained by the chromatin marks on each SNP . If at least one SNP had a low p-value for association ( p < 0 . 05/n . snps ) to the proportion of gene expression explained by chromatin data , we kept this region for our real data analysis . After this filtering procedure , we retained 1 , 317 regions . We obtained motif annotations from HaploReg [25] and ChromHMM annotations from the NIH Roadmap Epigenomics Consortium [30] . When comparing annotations of top prioritized paths with those of random paths , we established corresponding background paths by choosing a random SNP/mark combination at every region where a top path was reported . For GWAS analyses , we explored regions whose tag SNP was associated to an autoimmune trait with p < 5 × 10−8 . Associations were obtained from recent literature for eight autoimmune phenotypes [31–36] . For each of pathfinder’s top reported paths , we determined whether the corresponding SNP was contained within any of the GWAS regions in our dataset . In order establish a null distribution for this statistic , we ran the same analysis for random regions in the genome not overlapping with the GWAS regions in our dataset . Specifically , for each GWAS region , we randomly selected a SNP in the same chromosome matched for MAF ( ϵ = 0 . 01 ) and LD score ( ϵ = 0 . 001 ) with the GWAS tag SNP . We established a window around this matched SNP corresponding to the window size of the GWAS region . Finally , we determined the number of top paths that fell within these random regions . We repeated this experiment 100 times to establish the null distribution of this measurement and calculated a p-value using a Z-test . | Genome-wide association studies ( GWAS ) have revealed that the majority of variants associated with complex disease lie in noncoding regulatory sequences . More recent studies have identified thousands of quantitative trait loci ( QTLs ) associated with chromatin modifications , which in turn are associated with changes in gene regulation . Thus , one proposed mechanism by which genetic variants act on trait is through chromatin , which may in turn have downstream effects on transcription . In this work , we propose a method that assumes a causal path from genetic variation to chromatin to expression and integrates information across all three levels of data in order to identify the causal variant and chromatin mark that are likely influencing gene expression . We demonstrate in simulations that our probabilistic approach produces well-calibrated posterior probabilities and outperforms existing methods with respect to SNP- , mark- , and overall path-mapping . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"genome-wide",
"association",
"studies",
"dna-binding",
"proteins",
"social",
"sciences",
"simulation",
"and",
"modeling",
"probability",
"distribution",
"mathematics",
"genome",
"analysis",
"epigenetics",
"molecular",
"genetics",
"chromatin",
"research",
"and",
"analysis",
"methods",
"economic",
"models",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"histones",
"molecular",
"biology",
"economics",
"probability",
"theory",
"biochemistry",
"cell",
"biology",
"normal",
"distribution",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"computational",
"biology",
"human",
"genetics"
] | 2018 | Methods for fine-mapping with chromatin and expression data |
Neurocysticercosis ( NCC ) is an infection of the brain with the larval cyst of the tapeworm , Taenia solium . Cysticidal treatment induces parasite killing resulting in a post inflammatory response and seizures , which generally requires corticosteroid treatment to control inflammation . The nature of this response and how to best control it is unclear . We investigated the anti-inflammatory effects of pretreatment with etanercept ( ETN ) , an anti-tumor necrosis factor agent , or dexamethasone ( DEX ) , a high potency corticosteroid , on the post treatment inflammatory response in naturally infected pigs with neurocysticercosis after a single dose of the cysticidal drug praziquantel ( PZQ ) . We followed the methods from a previously developed treatment model of NCC in naturally infected swine . The four study groups of infected pigs included 3 groups treated with PZQ on day 0: PZQ-treated alone ( 100 mg/kg PO; n = 9 ) , pretreated with dexamethasone ( DEX , 0 . 2 mg/kg IM administered on days -1 , +1 and +3; n = 6 ) , and pretreated with etanercept ( ETN , 25 mg IM per animal on days -7 and 0; n = 6 ) . The fourth group remained untreated ( n = 3 ) . As measured by quantitative RT-PCR , ETN pretreatment depressed transcription of a wide range of proinflammatory , regulatory and matrix protease encoding genes at 120 hr post PZQ treatment in capsules of cysts that demonstrated extravasated Evans Blue ( EB ) ( a measure of blood brain barrier dysfunction ) compared to animals not receiving ETN . Transcription was significantly depressed for the proinflammatory genes tumor necrosis factor ( TNF ) -α , and interferon ( IFN ) -γ; the inflammation regulating genes cytotoxic T-lymphocyte-associated protein ( CTLA ) 4 , interleukin ( IL ) -13 and transforming growth factor ( TGF ) -β; the tissue remodeling genes matrix metalloprotease ( MMP ) 1 and 9 , tissue inhibitors of metalloproteases ( TIMP ) 1 and 2 , and the genes regulating endothelial function vascular endothelial growth factor ( VEGF ) 1 , angiopoietin ( Ang ) 1 , Ang 2 , and platelet endothelial cell adhesion molecule ( PECAM ) -1 . In contrast , transcription was only modestly decreased in the DEX pretreated pigs compared to PZQ alone , and only for TNF-α , IL-6 , IFN-γ , TGF-β and Ang1 . IL-10 was not affected by either ETN or DEX pretreatments . The degree of inflammation , assessed by semi-quantitative inflammatory scores , was modestly decreased in both ETN and DEX pretreated animals compared to PZQ treated pigs whereas cyst damage scores were moderately decreased only in cysts from DEX pretreated pigs . However , the proportion of cysts with EB extravasation was not significantly changed in ETN and DEX pretreated groups . Overall , TNF-α blockade using ETN treatment modulated expression of a large variety of genes that play a role in induction and control of inflammation and structural changes . In contrast the number of inflammatory cells was only moderately decreased suggesting weaker effects on cell migration into the inflammatory capsules surrounding cysts than on release of modulatory molecules . Taken together , these data suggest that TNF-α blockade may provide a viable strategy to manage post-treatment pericystic inflammation that follows antiparasitic therapy for neurocysticercosis .
Neurocysticercosis ( NCC ) , an infection of the central nervous system ( CNS ) by the larval stage ( cysticercus ) of the parasitic cestode Taenia solium , is a major cause of epilepsy in developing countries and a serious public health burden [1–4] . The disease is endemic to regions across the world where pigs are raised and allowed to roam freely with access to human waste [1 , 5] . The occurrence of seizures and other symptoms of NCC depend on the number , location and distribution of cysticerci , the intensity of brain inflammation and the degenerative stage of the parasite , resulting in a wide variety of manifestations [2 , 6] . A notable feature of T . solium infections is that viable cysts provoke minimal or no host-directed inflammatory responses . However , degenerating cysts or cysts damaged by anthelmintic treatment provoke inflammatory responses that can have pathological consequences on brain tissues surrounding the dying parasite [2 , 5 , 7] . Consequently , inflammation around degenerating cysts in the brain parenchyma generally results in seizures , whereas inflammation in the subarachnoid spaces causes diffuse and/or focal arachnoiditis frequently resulting in hydrocephalus , infarctions and nerve entrapments . Cysts in the ventricles commonly cause hydrocephalus due to mechanical obstruction of cerebrospinal fluid ( CSF ) outflow or to ventriculitis and scarring [1 , 8] . The pathological inflammatory response induced by cysticidal drugs can interfere with treatment . Although corticosteroids are almost universally used to suppress inflammation and control symptoms , the ideal regimen for the safe and effective use of corticosteroids or other anti-inflammatory agents in multicystic or complicated NCC has not been determined . As a result , the dose , duration and type of corticosteroid used are frequently based on the individual practitioner’s experience or preference [5] . A better understanding of the acute inflammatory responses induced by treatment is necessary to formulate simple , safe and more effective treatment measures . Studies of human and animal models of NCC indicate that inflammatory mediators produced by innate and adaptive immune cells play an important role in regulating inflammation both locally and systemically [9–16] . We previously demonstrated that expression of mediators of inflammation such as tumor necrosis factor ( TNF ) -α , interleukin ( IL ) -6 and interferon ( IFN ) -γ was up regulated following anthelmintic treatment around cysts that displayed disruption of blood brain barrier integrity [17] . These findings suggested points of attack to suppress specific pathways controlling treatment-induced inflammation to avoid the serious adverse effects of global immunosuppression associated with corticosteroids . In the present study we focused on the TNF-α pathway of inflammation because of its importance in this infection . Changes in expression of genes encoding a number of inflammatory mediators and regulatory factors following treatment with praziquantel were determined in pericystic brain tissue from infected pigs following blockade of TNF-α with etanercept ( ETN ) , a competitive inhibitor of TNF-α , and compared to corresponding tissues from a group of PZQ-treated pigs pretreated with corticosteroids and a control group of PZQ-treated pigs who did not receive any pretreatment .
Twenty-four T . solium-infected outbred pigs , confirmed by a positive tongue examination for cysts , were obtained in Huancayo , Peru , a town in a region of Peru endemic for cysticercosis . Four healthy outbred uninfected pigs purchased in Lima , Peru served as a source of tissues to normalize the gene expression assays; they did not receive any treatment . The four study groups included: untreated ( U ) , anthelmintic treatment with praziquantel ( PZQ , 100 mg/kg; P ) , dexamethasone ( DEX , and PZQ; DP ) and etanercept ( ETN and PZQ; EP ) . The experimental design , including treatment and sample collection schedule is shown in Fig 1 . Pigs were housed in the animal facility of the San Marcos Veterinary School . A hundred and twenty hours after administration of PZQ , the pigs were anesthetized with ketamine ( 10 mg/kg , intramuscular injection ) and xylazine ( 2 mg/kg , both from Agrovetmarket SA , Peru ) , for an intravenous catheterization and infusion of Evans Blue ( EB ) and euthanized with sodium pentobarbital ( 25 mg/kg kg every 30 min for two hours , intravenous injection; Montana SA , Peru ) . The study protocol and procedures were reviewed and approved by the Comité Institucional de Ética para el Uso de Animales–CIEA ( Institutional Ethics Commitee for the Use of Animals ) of the Veterinary School of San Marcos University in Lima , Peru ( Protocol numbers 006 for Universidad Nacional Mayor de San Marcos and 62392 for Universidad Peruana Cayetano Heredia ) . The Comité is registered in the Office for the Wellbeing of Laboratory Animals of the Department of Health and Human Services of the National Institutes of Health with Policy Number A5146-01 . All procedures used at the Veterinary Medicine Faculty of Universidad National Mayor de San Marcos ( FMV-UNMSM ) adhere to the International Guiding Principles for Biomedical Research that Imply the Use of Animals by the Council of International Organizations of Medical Sciences ( CIOMS ) , Geneva , 1985 . As shown in Fig 1 , infected pigs were treated with a single dose of praziquantel ( PZQ , 100 mg/kg PO , on day 0; Montana SA , Peru ) , pretreated with DEX ( 0 . 2 mg/kg IM , on days -1 , +1 and +3; Química Farvet , Mexico ) , ETN ( 25 mg/pig on days -7 and day 0 , when PZQ treatment was also administered; Amgen , CA ) or no drugs and sacrificed 120h later ( n = 21 , total ) . Three untreated infected pigs were used as controls . Two hours before euthanasia , pigs were anesthetized and infused with 2% EB ( 80 mg/kg; Sigma-Aldrich , St . Louis , MO ) in saline solution ( NaCl 0 . 85% , Laboratories Baxter , Colombia , Baxter del Peru ) by intravenous injection via the carotid artery . Just after euthanasia , the pigs were perfused with chilled saline solution containing heparin ( NaCl 0 . 85% with 10 U of heparin/mL ) and the brains were immediately removed to collect specimens . Brains were sliced into 10-mm thick sections on dry ice . The presence of unambiguously blue and clear stained cysts and tissues was documented by gross examination . Tissues surrounding visible cysts ( pericystic “capsules” , histologically consisting of collagen and cellular infiltrate ) were sampled from each brain ( ~1 mL fragments ) and either placed in RNALater solution ( Invitrogen , Gaithersberg , MD ) for RNA extraction or fixed , together with the cyst , in 10% formalin ( PBS pH 7 . 2 with 3 . 7% formaldehyde ) for histological examination . More cysts were selected for histology than for RNA studies because of relatively limited resources for the latter ( See S1 Table ) . Paraffin sections were processed using standard procedures and stained with hematoxylin and eosin ( HE ) and Masson’s Trichrome stain ( MT ) . Only capsules containing a whole cyst or a cyst wall surrounded by host tissue were selected for histological examination . We focused on the blue cysts capsules for qPCR analysis for two main reasons . Firstly , our previous studies had shown that significant upregulation of inflammatory and regulatory genes was apparent in blue capsules and not in clear capsules [17 , 18] . Secondly , the proportion and number of clear capsules following PZQ treatment was dramatically reduced , resulting in unacceptably large variance in the parameters studied and making statistical inference unreliable or not possible ( See S2 Table ) . Histological examination and analysis were performed exactly as reported previously [17 , 18] . Briefly , a low power image of a section of whole cyst was first examined to assess the circumferential proportion associated with pericystic inflammation . Higher power examination was then employed to determine the proportion of the cyst circumference showing each inflammatory stage ( IS ) present around each cyst . The classification of the inflammatory stages followed the schema described by Álvarez et al [19] and Londoño et al [20] , and adapted by us in a previous study [18] . According to this scheme , stage categorization of the inflammatory infiltrate was semi-quantitatively determined based on the average number of cells per high power field , and the thickness and location of type I collagen fibers around the cysts and in pericystic capsules [17] . Using these measurements , IS 1 to 4 represented increasing severity of inflammatory reaction and pathology . Similarly , cyst wall damage was categorized into four stages ( Damage Score: DS0-DS3 by severity of tissue disruption in the cysts , as outlined by Londoño et al . [20] . Composite inflammatory ( IS ) and damage scores ( DS ) were determined for each cyst using the formula: Composite IS or DS score = Sum [ ( Score x % of cyst circumference ) x 100] . The percentage of circumference was rounded off to increments of 25% ( i . e . , 0 , 25 , 50 , 75 or 100% ) . IS and CD scores from the untreated ( U ) and PZQ-treated ( P ) groups pooled data from the experimental groups in the current experiments and with those from previous experiments of identical design published elsewhere [17] . qPCR was performed on subsets of cysts selected from cysts with unambiguously clear or blue capsules . The distribution of cysts from each experimental group is shown in S1 Table . Fragments of brain tissue ( 50–100mg ) containing only cyst capsules ( no cyst ) were homogenized in 1 ml TRIzol ( Invitrogen , Gaithersberg , MD ) for standard RNA chloroform extraction . cDNA was generated from 1 μg of total RNA using the High-Capacity cDNA Reverse Transcription Kit with multiscribe RT polymerase and random primers ( Applied Biosystems , Foster City , CA ) in 20-μl reactions by incubation for 10 min at 25°C followed by 60 min at 37°C , 5 min at 95°C on a thermocycler ( MJ Research PTRC-200 , BioRad-MJ Research , Hercules , CA ) . Real time PCR ( qPCR ) was performed in 10-μl reaction volumes using the Taqman Gene Amplification System ( Applied Biosystems ) with commercially available primer-probe combinations using conditions recommended by the manufacturer . We used 18S rRNA as a control gene to validate RNA integrity and primer probe pairs for porcine TNF-α , IL-6 , IFN-γ , IL-10 , CTLA4 , IL-13 , TGF-β , MMP1 , MMP9 , TIMP1 , TIMP2 , VEGF , PECAM1 , Ang1 and Ang2 genes . qPCR reactions , run in triplicate , used the following cycling parameters: preincubation of 2 min at 50°C and 10 min at 95°C followed by 40 cycles of 15 sec at 95°C and 1 min at 60°C , on an AB StepPlusOne cycler ( Life Technologies , Grand Island , NY ) . Results for each gene were expressed as relative to the expression of 18S rRNA using the X-fold value defined by the 2-ΔΔCT formula [21] . The number of cysts analyzed for each marker differed within a given experimental group , because limitations on the amount of RNA extracted prevented us from analyzing each cyst for all the desired markers . Non-parametric statistics , Mann-Whitney U test for two groups and Kruskal-Wallis test for multiple groups , were calculated using Prism® software ( Graphpad , San Diego , CA ) for comparisons of the above parameters between infected pigs that were not treated or treated with one of the two anti-inflammatory agents ( PZQ or ETN ) , and between clear cysts and those with EB staining ( blue cysts ) . Corrections for multiple comparisons were applied for pairwise comparisons . Differences with p-values of <0 . 05 were considered statistically significant . Two-way contingency analysis , for example , comparing differences in proportions of cysts between PZQ-treated and PZQ plus DEX-pretreated pigs were performed by the Fisher’s exact test .
All infected animals had cysts scattered throughout the brain . On gross examination , pericystic capsules were found either to be clear or stained blue due to extravasated EB . As observed previously [22] , the proportion of capsules that had EB extravasation was increased 120h after PZQ treatment ( Fig 2A; p<0 . 001 , Fisher’s test ) . The increase in proportion of blue capsules after PZQ treatment was not reduced by pretreatment with DEX or ETN ( Fig 2A; p>0 . 05 , Fisher’s test ) . A total of 339 cysts were collected from all experimental animals , of which 233 cysts with intact pericystic capsules were examined histologically for both IS and DS . To determine if pretreatment with DEX or ETN affected the degree of inflammatory infiltration around cysts or damage to cyst walls , we compared the IS and DS among the experimental groups . No significant differences were noted in the IS or DS in clear capsules ( Fig 2B and 2C ) , likely due to the high variability in the scores . However , IS and DS scores for blue capsules were higher in PZQ-treated pigs compared to untreated pigs . Pretreatment with DEX or ETN reduced the IS of blue cysts compared to PZQ alone ( Fig 2B ) . Analysis of cyst wall damage associated with PZQ treatment revealed that DEX pretreatment , but not pretreatment with ETN , resulted in a decrease in the DS . ( Fig 2C ) These findings suggest dissociation between the regulation of inflammation by TNF-α , and the induction of cyst wall damage . A smaller subset of cysts than those examined histologically was analyzed for expression levels of genes for molecules involved in tissue inflammation ( S1 Table ) . Comparison of three markers of inflammation , TNF-α , IL-6 and IFN-γ , revealed that 120-h after PZQ treatment there was a significant up regulation of all three markers in blue capsules , as expected ( Fig 3A–3C , p<0 . 05 for each ) . However , TNF-α blockade with ETN prior to PZQ treatment resulted in a profound decrease in the expression of TNF-α compared to PZQ alone ( Fig 3A; ~10-fold decrease to untreated baseline , p<0 . 0005 ) . A smaller , but significant ( p<0 . 005 ) decrease was observed in DEX pretreated pigs . IL-6 and IFN-γ were similarly inhibited in blue capsules by ETN pretreatment ( Fig 3B and 3C ) . In clear capsules that lack the disruption of vascular integrity , gene expression of pretreatment with ETN or DEX could not be compared to PZQ alone ( see S1 Fig ) . Remarkably , expression of TNF-α was lower in cysts from ETN pretreated pigs than from pigs that did not receive PZQ ( Fig 3A; p<0 . 05 , Mann-Whitney U test ) . There was no significant corresponding inhibition of IFN-γ in the clear cysts ( S1 Fig ) . We next analyzed the effect of ETN and DEX pretreatment on counter regulatory pathways . In our previous study , we found variable inhibition of regulatory molecules , so that CD25 and CTLA4 transiently decreased at 48h post PZQ treatment , whereas IL-10 showed persistent decrease 48h post treatment and later [17] . In the present study , IL-10 gene expression was inhibited from baseline after ( 120-h ) PZQ treatment , but neither DEX nor ETN reversed the PZQ-induced inhibition ( Fig 4A ) . In contrast , the expression of three other regulatory molecules , CTLA4 , IL-13 and TGF-β , was significantly decreased by ETN ( but not by DEX ) pretreatment ( Fig 4B–4D ) . While the inhibitory effect of ETN on gene expression of CTLA4 and TGF-β was only apparent around blue cysts , inhibition of IL-13 expression was also significantly decreased around clear cysts in pigs pretreated with ETN compared to DEX ( S1 Fig ) . Unfortunately , there were insufficient numbers of clear cysts in the PZQ alone treated pigs for valid statistical inferences for comparisons with ETN or DEX pretreatment . Genes associated with tissue remodeling and also in granuloma formation , such as matrix metalloprotease ( MMP ) 2 and MMP9 , as well as their regulators , the tissue inhibitors of metalloproteases ( TIMP ) 1 and TIMP2 , are up regulated following PZQ treatment in rodent models of NCC [23 , 24] and in infected pigs [17] . Analysis of these genes in the present study revealed profound down regulation of MMP2 , MMP9 , TIMP1 and TIMP2 in blue cyst capsules from ETN-pretreated pigs compared to pigs who received PZQ alone ( Fig 5A–5D ) . These observations are consistent with the global inhibition of gene expression that appears to accompany TNF-α blockade . Interestingly , DEX pretreatment did not significantly down regulate transcripts that were markedly inhibited by the TNF-α blockade . Expression of genes involved in endothelial activation such as PECAM1 , and angiogenesis such as VEGF , angiopoietin 1 and 2 ( Fig 5E–5H ) are commonly involved in the pathogenesis of parasitic infections [25–27] . Our prior findings that showed increased vascular leakage in the cyst capsules following PZQ treatment of pigs [17 , 18] also suggest that endothelial integrity and function may be involved in the resulting inflammatory pathology . We found that ETN pretreatment resulted in a significant inhibition of transcription of all these molecules , which were up regulated by PZQ treatment alone . Taken together , these data reveal that TNF-α blockade resulted in transcription inhibition of a diverse range of inflammatory and regulatory pathway molecules and suggest an important role for TNF-α in regulating the inflammation that predictably follows PZQ treatment . We also determined gene expression levels for the genes discussed above in tissues around cysts from the four experimental groups that did not have EB leakage and , therefore , were not demonstrating disruption of the BBB ( Fig 6 ) , referred to as “clear” cysts . The total number of clear cysts was low because the PZQ treatment , which the three experimental groups received , appears to induce BBB leakage in the large majority of cysts . Although there were decreases in gene expression of some inflammatory markers ( e . g . , TNF-α , IFN-γ ) in clear cysts from DEX and ETN-treated pigs that were similar to those seen in blue cysts , other inflammatory markers showed increased expression in DEX- and ETN-treated pigs ( e . g . , IFN-γ and Ang2; Fig 6C and 6G ) . These differences between the blue and clear cysts probably reflect the lower level of tissue penetration of ETN in the latter due to an intact BBB , that may influence it’s effectiveness in the tissues .
In T . solium-infected humans and pigs , inflammation around dying cysts occurs frequently and predictably within a week of cysticidal treatment in humans and in animal models of NCC [28–33] . In humans , the evidence for this manifest as an increased incidence of headaches , seizures and other neurologically based symptoms associated with the development of gadolinium enhancement and edema around some cysts seen on MRI imaging within the first week after initiation of cysticidal treatment . Limited histopathological examination of brain tissues around degenerating cysts in patients post-treatment or with untreated “degenerated” cysts that shows infiltration of inflammatory cells [5] additionally supports the concept of treatment induced inflammation . Previous studies performed by ourselves and others demonstrated similar post-treatment inflammatory reactions in pigs [17 , 18 , 22 , 33–35] . In a practical sense , these detrimental side effects of antiparasitic treatment complicate medical treatment of NCC because they are a cause of morbidity and need to be prevented and controlled with corticosteroids , the use of which is associated with a variety of side effects [1 , 5 , 36 , 37] . In a review focused on murine models of cysticercosis , one author discussed the concept of using approaches other than corticosteroids to inhibit inflammation associated with cysticidal treatment [38] . In the CNS , TNF-α produced by migrated peripheral immune cells or by microglia and astrocytes in the presence of inflammation [39 , 40] plays an important role in inducing and maintaining inflammation that occurs in NCC . Evidence for this includes high levels of TNF-α reported in CSF samples from patients [12 , 41–47] and increased expression of the TNF-α gene in pericystic capsules in pigs treated with PZQ [17 , 48 , 49] . TNF-α has also been shown to be a key cytokine in maintaining inflammation in other inflammatory diseases involving the CNS , such as some causes of meningitis and autoimmune encephalitis [50–56] . Therefore , we hypothesized that blockade of TNF-α should mitigate post-treatment inflammation in NCC . Our data show that TNF-α blockade during PZQ treatment resulted in a broad and unique pattern of inhibition of gene expression for inflammation promoting proteins , regulating molecules , a number of pathways of tissue remodeling substances and molecules that modulate endothelial cellular function . Notably , not all the genes tested were inhibited by ETN: for example , IL-10 expression in PZQ plus ETN treated pigs did not differ significantly from pigs receiving PZQ alone or PZQ plus DEX . In contrast , DEX administration prior to PZQ treatment significantly inhibited only TNF-α , IL-6 , IFN-γ , TGF-β and Ang1 relative to PZQ treatment alone ( Figs 3 , 4 and 6 ) . The lower anti-inflammatory responses to DEX compared to ETN was surprising , since corticosteroids are known for their potent and global immunosuppressive activity [57–59] . However , pigs are known to be relatively insensitive to the immunosuppressive effects of corticosteroids [60] . The effectiveness of ETN-mediated inhibition of multiple pathways of inflammation and tissue remodeling that is demonstrated by these data suggest a significant role for TNF-α in post-PZQ inflammatory responses An interesting finding in this study was an apparent dissociation between the effects of ETN on gene expression for inflammatory mediators and regulators ( Figs 3 and 4 ) and its effect on cellular infiltration in pericystic tissues ( Fig 2B ) . TNF-α triggers a cascade of inflammatory cytokines , but also promotes endothelial cell contribution to local inflammation via the display of different combinations of adhesion molecules for leukocytes , including E-selectin , intercellular adhesion molecule-1 ( ICAM-1 ) and vascular cell adhesion molecule-1 ( VCAM-1 ) in a distinct temporal , spatial and anatomical pattern [61 , 62] . In combination with the release of chemokines ( including IL-8 , MCP-1 and CCL2 ) [63] , these responses lead to recruitment of different populations of leukocytes , so blockade of TNF-α would normally be expected to inhibit cellular recruitment . However , our data ( Fig 2B ) reveal a weak , albeit significant , reduction in scores signifying only a small decrease in cellular infiltration . The reason for this apparent dissociation in the two functional properties of TNF-α in this model is unclear , and may relate to our use of a human TNF-α blocker in pigs or possibly a differential effect of TNF-α concentration on the two processes . Interestingly , the effect of TNF-α blockade on parasite damage , as reflected in the cyst wall damage scores ( Fig 2C ) , suggests that the inhibition of measured proinflammatory , regulatory and other molecules did not inhibit damage to cysts caused by PZQ , as was found with DEX pretreatment . ETN , a licensed biologic , has been used for TNF-α blockade for over 20 years [64 , 65] and has shown remarkable efficacy as an anti-inflammatory agent in rheumatoid arthritis and inflammatory bowel diseases [62 , 66]; its safety profile is well known . Our data demonstrate that TNF-α blockade induces potent suppression of post-treatment pericystic inflammation in a natural infection model of NCC . The inhibitory effect of TNF-α in this model was comparable to that of DEX , a potent inhibitor of inflammation in many settings . This study provides proof of principle that TNF-α blockade , used alone or as a steroid-sparing agent , may be a viable strategy for management of post-PZQ pericystic inflammation . | Infection of the brain with larvae of the tapeworm Taenia solium is called neurocysticercosis ( NCC ) , a disease with varied and serious neurological symptoms . Therapy requires antiparasitic drugs and corticosteroids to prevent seizures caused by treatment due to inflammation around dying parasites . The gene expression of the proinflammatory molecule tumor necrosis factor alpha ( TNF-α ) is increased in NCC . We treated three groups of naturally infected pigs with an antiparasitic drug: one group was also pretreated with an anti-TNF-α inhibitor , the second one with a corticosteroid , and the third was not pretreated . All pigs were infused with Evans blue dye ( EB ) , which leaks where the blood brain barrier is damaged by inflammation around cysts . We compared the expression of several genes involved in inflammation , healing and fibrosis and regulation of vascular function in tissues surrounding cysts . In inflamed samples showing leaked EB , the inhibition of TNF-α suppressed nearly all the genes assessed , and this suppression was significantly stronger than the moderate decrease caused by corticosteroid pretreatment on most of the genes . On microscopic examination , the inflammation observed was slightly decreased with both pretreatments in relation to the group that was not pretreated . We believe that the inflammatory route that includes TNF-α should be further explored in the search for better management of inflammation directed to degenerating cysts . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"inflammatory",
"diseases",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"pig",
"models",
"gene",
"regulation",
"immunology",
"vertebrates",
"animals",
"mammals",
"animal",
"models",
"developmental",
"biology",
"signs",
"and",
"symptoms",
"pharmaceutics",
"experimental",
"organism",
"systems",
"molecular",
"development",
"research",
"and",
"analysis",
"methods",
"swine",
"inflammation",
"gene",
"expression",
"immune",
"response",
"immune",
"system",
"eukaryota",
"diagnostic",
"medicine",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"amniotes",
"organisms"
] | 2017 | TNF-α blockade suppresses pericystic inflammation following anthelmintic treatment in porcine neurocysticercosis |
Neuronal excitability relies on inward sodium and outward potassium fluxes during action potentials . To prevent neuronal hyperexcitability , potassium ions have to be taken up quickly . However , the dynamics of the activity-dependent potassium fluxes and the molecular pathways underlying extracellular potassium homeostasis remain elusive . To decipher the specific and acute contribution of astroglial Kir4 . 1 channels in controlling potassium homeostasis and the moment to moment neurotransmission , we built a tri-compartment model accounting for potassium dynamics between neurons , astrocytes and the extracellular space . We here demonstrate that astroglial Kir4 . 1 channels are sufficient to account for the slow membrane depolarization of hippocampal astrocytes and crucially contribute to extracellular potassium clearance during basal and high activity . By quantifying the dynamics of potassium levels in neuron-glia-extracellular space compartments , we show that astrocytes buffer within 6 to 9 seconds more than 80% of the potassium released by neurons in response to basal , repetitive and tetanic stimulations . Astroglial Kir4 . 1 channels directly lead to recovery of basal extracellular potassium levels and neuronal excitability , especially during repetitive stimulation , thereby preventing the generation of epileptiform activity . Remarkably , we also show that Kir4 . 1 channels strongly regulate neuronal excitability for slow 3 to 10 Hz rhythmic activity resulting from probabilistic firing activity induced by sub-firing stimulation coupled to Brownian noise . Altogether , these data suggest that astroglial Kir4 . 1 channels are crucially involved in extracellular potassium homeostasis regulating theta rhythmic activity .
Astrocytic processes enwrap more than half of CA1 hippocampal synapses to form tripartite synapses [1 , 2] . Perisynaptic astroglial processes are enriched in ionic channels , neurotransmitter receptors and transporters , enabling astrocytes to detect neuronal activity via calcium signaling [3] and ionic currents with various components , such as glutamate and GABA transporter [4–7] or potassium ( K+ ) [8–10] . Thus astrocytes regulate neuronal activity through multiple mechanisms , involving signaling or homeostasis of extracellular space volume , glutamate , GABA or K+ levels [11] . Interestingly , membrane depolarization was the first activity-dependent signal identified in glial cells and was attributed to K+ entry across their membrane [10] . Such K+ entry was suggested to contribute to K+ spatial buffering , consisting in glial uptake of excess extracellular K+ ( [K+]o ) , redistribution via gap-junction astroglial networks and subsequent release at sites of low [K+]o [12] . Modeling studies have mostly investigated astroglial regulation of [K+]o during pathological conditions to clarify its impact on aberrant neuronal activity . In particular astrocytes , by regulating [K+]o , have been shown to contribute to initiation and maintenance of epileptic seizures [13–15] , as well as to the severity of ischemia following stroke , with a neuroprotective or neurotoxic role , depending on [K+]o [16 , 17] . In addition , experimental data suggest that several K+ channels or transporters contribute to astroglial K+ clearance , such as inward rectifier 4 . 1 and two pore K+ channels ( Kir4 . 1 and K2P , respectively ) or Na/K ATPases [18 , 19] . Remarkably , recent work suggest that Kir4 . 1 channels play a prominent role in astroglial regulation of [K+]o [20–23] . However , the mouse model used to draw these conclusions , i . e . conditional Kir4 . 1 knockout mice directed to glial cells ( GFAP-Cre-Kir4 . 1fl/fl mice , Kir4 . 1-/- ) , exhibits several limitations: 1 ) Kir4 . 1 channels are not specifically deleted in astrocytes , but also in other glial cells such as oligodendrocytes or retinal Müller cells [22]; 2 ) astrocytes are severely depolarized [21 , 22]; 3 ) Kir4 . 1-/- mice die prematurely ( ~3 weeks ) and display ataxia , seizures , hindleg paralysis , visual placing deficiency , white matter vacuolization and growth retardation [22] , highlighting that chronic deletion of Kir4 . 1 channels induces multiple brain alterations and possibly compensations . Thus , the specific and acute contribution of astroglial Kir4 . 1 channels to [K+]o and to the moment to moment neurotransmission is still unclear . To decipher the acute role of astrocytes in controlling K+ homeostasis and neuronal activity , we built a tri-compartment model accounting for K+ dynamics between neurons , astrocytes and the extracellular space . We quantified K+ neuroglial interactions during basal and high activity , and found that Kir4 . 1 channels play a crucial role in K+ clearance and astroglial and neuronal membrane potential dynamics , especially during repetitive stimulations , and prominently regulate neuronal excitability for 3 to 10 Hz rhythmic activity .
To model K+ ions dynamics during neuronal activity , we built a biophysical model that includes three compartments: the neuron , the astrocyte and the extracellular space ( Fig . 1A ) . As performed in several studies [13 , 16 , 24] , the neuron is approximated by a single compartment conductance-based neuron containing Na+ and K+ voltage-gated channels , enabling action potential discharge . The associated neuronal membrane potential is coupled with the dynamics of intracellular and extracellular Na+ and K+ levels via the dependence of the neuronal currents to the Nernst equation . The ion concentrations depend also on the activity of neuronal and astroglial Na/K ATPases , which maintain resting [K+]i by balancing K+ and Na+ fluxes . Similarly , the astrocyte is approximated by a single compartment conductance-based astrocyte containing Kir4 . 1 channels , which are inward rectifier K+ channels strongly expressed in astrocytes that generate dynamic K+ currents [25] . In the model , neurons and astrocytes are separated by a homogenous extracellular space compartment . The model is based on balancing ionic fluxes between the three compartments ( Fig . 1B ) . The model starts with the induction of a synaptic current ( Iapp , see Materials and Methods ) . This current is the initial input of a classical Hodgkin-Huxley model , which describes the neuronal membrane potential dynamics ( entry of Na+ and exit of K+ ) . Released extracellular K+ is taken up by astrocytes through Kir4 . 1 channels and Na/K ATPases ( Fig . 1B and Materials and Methods ) . Because Kir4 . 1 channels are strongly involved in K+ uptake [22] , we fitted the I-V curve of K+ ions through Kir4 . 1 channels using equation 22 ( see materials and methods ) and predicted the I-V curve at various values of [K+]o ( Fig . 1C ) . We obtain that K+ fluxes through Kir4 . 1 channels vanish around astrocytic resting membrane potential ( ~-80 mV ) and are outward during astrocytic depolarization for a fixed [K+]o ( 2 . 5 mM , Fig . 1C ) . However , they become inward when [K+]o increases ( 5–10 mM , Fig . 1C ) . Using this model , we shall investigate quantitatively the contribution of Kir4 . 1 channels to K+ uptake in relation to neuronal activity associated with different [K+]o . To validate our tri-compartment model , we compared simulation results with electrophysiological recordings . To account for the synaptic properties of CA1 pyramidal neurons , we generated a synaptic current ( Iapp ) using the depression-facilitation model ( equation 1 ) ( see Materials and Methods with input f ( t ) = δ ( t ) ) ( Fig . 2A , E , I ) . We first investigated responses to single stimulation . Using the Hodgkin-Huxley model , this synaptic current induces a firing activity ( S1A Fig . ) , resulting in a ~ 0 . 9 mM increase of [K+]o within 300 milliseconds , which slowly decayed back to baseline levels during 10 seconds ( S1B Fig . ) . The extracellular K+ dynamics was associated in our model with a small astrocytic depolarization of ∆V = −1 . 35 mV ( equations 22 , 23 , 25 ) ( Fig . 2C ) . Using electrophysiological recordings of evoked field excitatory postsynaptic potential ( fEPSP ) by a single stimulation of Schaffer collaterals in acute hippocampal slices ( Fig . 2B ) , we measured astroglial membrane potential depolarization and found that it reached ~ 1 . 3 mV ( 1 . 3 ± 0 . 2 mV , n = 6 ) ( Fig . 2C ) , confirming the result of our simulation . After validating the responses of the tri-compartment model to basal stimulation , we investigated the impact of trains of stimulations on the dynamics of astroglial membrane potential . During tetanic stimulation ( 100 Hz for 1 second ) , variations in neuronal membrane potential described by the Hodgkin-Huxley equation show a bursting activity during ~ 1 second ( S1C Fig . ) . This is associated with a depolarization of astrocytic membrane potential of ~ 5 mV , which lasts ~ 6 seconds ( Fig . 2G , H ) and an increase in [K+]o that reaches a peak value of 4 . 4 mM ( S1D Fig . ) . For repetitive stimulations ( 10 Hz for 30 seconds ) , the neuron exhibited firing activity during the whole stimulation ( S1E Fig . ) . This was associated with an astroglial depolarization of ~ 12 mV ( Fig . 2K ) and an increase in [K+]o peaking at 6 . 9 mM after 17 . 5 seconds of stimulation ( S1F Fig . ) . Although the stimulation lasted 30 seconds , the astrocytic depolarization started to decay after 17 seconds ( Fig . 2K ) . The kinetics of astroglial membrane potential dynamics obtained with the numerical simulations are comparable to the results obtained with electrophysiological recordings performed in individual astrocytes during single stimulation ( rise time: 48 . 4 ms for numerical stimulation , 42 ± 19 ms n = 6 for experiments; time of peak: 740 ms for numerical simulation , 730 ± 60 ms n = 6 for experiments; decay time: 3 . 67 s for numerical simulation , 4 . 50 s ± 0 . 2 n = 6 for experiments , Fig . 2D ) , tetanic stimulation ( rise time: 610 ms for numerical simulation , 491 ± 122 ms n = 5 for experiments; time of peak: 1 . 07 s for numerical simulation , 1 . 05 s ± 0 . 25 n = 5 for experiments; decay time: 4 . 18 s for numerical simulation , 4 . 55 s ± 0 . 45 n = 5 for experiments , Fig . 2H ) and repetitive stimulation ( rise time: 1 . 5 s for numerical simulation , 1 . 27 s ± 0 . 18 n = 5 for experiments; time of peak: 6 . 8 s for numerical simulation , 5 . 2 s ± 0 . 9 n = 5 for experiments; decay time: 7 . 95 s for numerical simulation , 8 . 3 s ± 0 . 4 n = 5 for experiments , Fig . 2L ) . These data show that the dynamics of astroglial membrane potential changes obtained from numerical simulations and from electrophysiological recordings are similar . Thus our model captures the key players sufficient to mimic the evoked astroglial membrane potential dynamics observed experimentally in different regimes of activity . We investigated the dynamics of the K+ cycle between neurons , extracellular space and astrocytes induced by neuronal activity to decipher the time needed to restore basal extracellular and intra-neuronal K+ levels . We studied K+ redistribution induced by single , tetanic ( 100 Hz , 1 s ) and repetitive ( 10 Hz , 30 s ) stimulations , and found that the general behavior of K+ dynamics was divided into three phases ( phases 0 , 1 and 2; Fig . 3 ) . During phase 0 ( t = 0 to t1 ) , neuronal K+ is released in the extracellular space ( peak [K+]o during phase 0: 0 . 9 mM at 300 ms for single stimulation; 1 . 9 mM at 1 . 3 s for tetanic stimulation; 4 . 4 mM at 30 s for repetitive stimulation , Fig . 3A , D , G ) . Compared to basal K+ levels in each compartment ( at t = 0 ) , the relative transient evoked increase in K+ concentration is prominent only in the extracellular space ( ~+37% for single stimulation , +76% for tetanic stimulation and +168% for repetitive stimulation , Fig . 3B , E , H ) . During phases 0 and 1 , released K+ is then mostly buffered by astrocytes ( ~80 to 99% at the end of phase 1 ) during the different regimes of activity ( time t2 ( at the end of phase 1 ) for single stimulation: 8 . 2 s; tetanic stimulation: 8 . 7 s; repetitive stimulation: 34 . 2 s , Fig . 3C , F , I ) . The astroglial net K+ uptake increases with the activity-dependent [K+]o transient rises ( S2A–C Fig . ) evoked by the different regimes ( S2D Fig . ) . Neurons slowly re-uptake only ~5–10% of their released K+ at the end of phase 1 ( Fig . 3C , F , I ) . Remarkably , although [K+]o increases with the strength of stimulation ( from 0 . 9 to 4 . 4 mM , Fig . 3A , D , G and S2A–C Fig . ) , the time needed for astrocytes to buffer the released K+ is not proportional to [K+]o rises ( Fig . 3C , F , I ) , as shown by the phase diagram illustrating astroglial K+ uptake as a dynamic function of activity-dependent changes in [K+]o evoked by the different stimulations ( S2D Fig . ) , but is to the square root of [K+]o ( equation 22 ) . In addition , at the end of phase 1 , [K+]o is almost back to baseline levels , whereas intra-astroglial K+ levels reach their peak value ( Fig . 3C , F , I ) . Finally during phase 2 ( t2 to end ) , astroglial buffered K+ is slowly redistributed back to neurons , which ends the K+ cycle . The long-lasting phase 2 is marked by an inversion of K+ fluxes in astrocytes , suggesting moderate K+ release by astrocytes over time . Indeed , K+ redistribution to neurons depends on K+ release through Kir4 . 1 channels , which is limited by the low outward rectification of these channels ( Fig . 1C ) . Altogether , these data suggest a slow , but dynamic and efficient astroglial clearance capacity for the different regimes of activity . To study quantitatively the acute and selective role of astroglial Kir4 . 1 channels in neuroglial K+ dynamics , we inhibited the Kir4 . 1 current in our tri-compartment model . Because Kir4 . 1-/- mice display altered synaptic plasticity compared to wild type mice [22 , 26] , we recalibrated the synaptic current ( Iapp ) parameters τrec and τinact in equations 1 , 2 ( see Table 1 ) for the facilitation-depression model to get an optimal fit to the recorded postsynaptic responses [26] . Another change in the model consisted in setting at zero both the Kir4 . 1 current and the leak term . In addition , to compensate for the loss of K+ fluxes through astroglial Kir4 . 1 channels , we added in equation 27 a constant K+ flux to maintain [K+]o at an equilibrium value of 2 . 5 mM . Consequently , the astrocytic membrane potential displayed no change during stimulation , in agreement with electrophysiological recordings [21 , 22] . The numerical simulations show that inhibition of astroglial Kir4 . 1 channels leads to higher transient peak increase in [K+]o during repetitive and tetanic stimulation compared to control conditions ( Fig . 4E , F , I , J ) , while no difference is observed for single stimulation ( Fig . 4A , B ) . In addition , for all regimes of activity , the rise and decay times of the [K+]o were increased when Kir4 . 1 channels were inhibited ( single stimulation , control: rise time 136 ms , decay time 3 . 4 s; Kir4 . 1 inhibition: rise time 232 ms , decay time 4 . 2 s; tetanic stimulation , control: rise time 638 ms , decay time 4 s; Kir4 . 1 inhibition: rise time 753 ms , decay time 6 s; repetitive stimulation , control: rise time 6 . 8 s; Kir4 . 1 inhibition: rise time 20 . 2 s , Fig . 4B , F , J ) . Finally , Kir4 . 1 channel inhibition only slightly increased neuronal firing induced by single stimulation ( Fig . 4C , D ) and tetanic stimulation ( Fig . 4G , H ) , while it had major effect on neuronal excitability during repetitive stimulation ( Fig . 4K , L ) . Indeed , although firing frequency was only slightly increased during the first 8 seconds of repetitive stimulation when [K+]o reached 10 mM ( Fig . 4I ) , action potential amplitude and firing rate then progressively decreased due to neuronal depolarization ( from-33 mV to-19 mV after 14 and 30 seconds of stimulation , respectively ) , suppressing neuronal firing after 14 seconds of stimulation ( Fig . 4K ) . Altogether , these data show that astroglial Kir4 . 1 channels are prominently involved in K+ buffering during high level of activity , and thereby have a major impact on neuronal resting membrane potential controlling firing during trains of stimulations . To investigate the effect of astroglial Kir4 . 1 channels on endogenous physiological rhythmic activity , we generated probabilistic firing induced by sub-firing stimulation coupled to neuronal Brownian noise ( Fig . 5A , B ) . To simulate the firing activity , we generated a sub-firing periodic stimulation ( 5 ms squared stimulus ) , which defines the applied synaptic intensity in our tripartite compartment model ( Fig . 5A ) , and added a Brownian noise in the neuronal membrane potential ( equation 21 , Fig . 5B ) . Such stimulation induces an increase in [K+]o ( Fig . 5C ) , and thus firing over time ( Fig . 5D-E ) . We found that astroglial Kir4 . 1 channels had no effect on the firing probability ( computed over 100 simulations ) for basal ( 0 . 1 Hz , Fig . 5F ) , low ( 1 Hz , Fig . 5G ) and high ( 50 Hz , Fig . 5K ) frequency stimulations . However , Kir4 . 1 channels directly regulate the firing probability for 3 and 5 Hz stimulations after 7 and 12 s of sub-firing stimulation , respectively ( Fig . 5H , I ) . In contrast , Kir4 . 1 channels regulate only transiently the firing probability induced by 10 Hz stimulation ( Fig . 5J ) . These data suggest a prominent and specific involvement of astroglial Kir4 . 1 channels in regulation of firing during theta rhythmic activity .
Several models have investigated extracellular K+ regulation of neuronal activity , including glial uptake mechanisms [13–17 , 24 , 33 , 34] . To study seizure discharges and spreading depression , a first tri-compartment model including the neurons , astrocytes and extracellular space was proposed [24] , although the astrocytic membrane potential was not taken into account , and K+ accumulation in the interstitial volume was controlled by a first-order buffering scheme that simulated an effective glial K+ uptake system . With such model , after evoked firing , it took ~17 s for the neuronal membrane potential to return to resting values , via activation of Na/K ATPases . The model also predicted that elevated [K+]o have a key role in the initiation and maintenance of epileptiform activity . In our study , we accounted for the astroglial modulation of K+ buffering capacity regulated by its membrane potential , and found that the biophysical properties of astrocytic membranes including Kir4 . 1 channels were sufficient to account for the long-lasting clearance of extracellular K+ . Interestingly , we confirm that alteration in K+ clearance leading to an extracellular K+ accumulation induces epileptiform activity , and show specifically that Kir4 . 1 channel acute inhibition leads to such pathological bursting activity during repetitive stimulation . A similar tri-compartment model has been simplified as a one-dimensional two-layer network model to study how neuronal networks can switch to a persistent state of activity , as well as the stability of the persistent state to perturbations [13] . In this model , Na+ and K+ affect neuronal excitability , seizure frequency , and stability of activity persistent states . In particular , the quantitative contribution of intrinsic neuronal currents , Na/K ATPases , glia , and extracellular Na+ and K+ diffusion to slow and large-amplitude oscillations in extracellular and neuronal Na+ and K+ levels was revealed . In the model , the estimated [K+]o during epileptiform activity are comparable to the ones observed experimentally [35 , 36] . Although this model does not account for astroglial Kir4 . 1 channels , it shows that a local persistent network activity not only needs balanced excitation and inhibition , but also glial regulation of [K+]o [15] . Finally , a model accounting for the extracellular space and astroglial compartments has quantified the involvement of several astroglial ionic channels and transporters ( Na/K ATPase , NKCC1 , NBC , Na+ , K+ , and aquaporin channels ) in the regulation of firing activity [34] . To account for K+ dynamics between neurons , astrocytes and the extracellular space , we built for the first time a tri-compartment model , where we included neuronal voltage-gated channels , Na/K pumps and astrocytic Kir4 . 1 channels according to their biophysical properties , as well as membrane potential of astrocytes . Because functional expression of voltage-gated calcium channels on hippocampal mature astrocytes in situ in physiological conditions and its impact on astrocytic functions is still a matter of debate [37] , such channels were not included in our model . However , many other astroglial K+ channels ( such as two pore domain K+ channels ( K2P ) ( TWIK-1 , TREK-1 , TREK-2 and TASK-1 ) , inward rectifier K+ channels ( Kir2 . 1 , 2 . 2 , 2 . 3 , 3 . 1 , 6 . 1 , 6 . 2 ) , delayed rectifier K+ channels ( Kv1 . 1 , 1 . 2 , 1 . 5 , 1 . 6 ) , rapidly inactivating A-type K+ channels ( Kv1 . 4 ) , calcium-dependent K+ channels ( KCa3 . 1 ) ) , but also other channels , transporters or exchangers ( such as Cx hemichannels , Na+/K+/Cl- co-transporter ( NKCC1 ) K+/Cl- exchanger , glutamate transporters ) [16 , 38 , 39] could also play a role in the regulation of activity-dependent changes in [K+]i or [K+]o . Functional evidence of the contribution of these channels , transporters or exchangers in astroglial K+ clearance is actually scarce , although K2P channels have been suggested to participate in astroglial K+ buffering [40] , while NKCC1 were recently shown in hippocampal slices not to be involved in activity-dependent K+ clearance [41] . Similarly , adding slower timescale K+ dependent conductances in the neuron model could modulate the slow redistribution of K+ to neurons , and thus the duration of the neuroglial potassium cycle , and is of interest to implement in future development of the model . In our study , the aim was to simplify the system to capture in the model the minimal set of astroglial channels and pumps accounting for our experimental data related to activity-dependent changes in astroglial membrane potential . In addition our tri-compartment model , as most existing models , did not account for the complex multiscale geometry of astrocytes and neurons . Incorporating in our current model additional astroglial and neuronal channels , as well as complex cell geometry is of particular interest to identify modulatory effects of other specific channels and of microdomain geometry on the neuroglial potassium cycle . In accordance with previous studies , where Kir4 . 1 channels were chronically deleted genetically in glial cells [20 , 21 , 23] , we found that acute inhibition of Kir4 . 1 channels leads to altered regulation of extracellular K+ excess and affects the kinetics of [K+]o ( Fig . 4I , J ) . However , in contrast to these studies , we found that Kir4 . 1 channel inhibition also alters significantly [K+]o peak amplitudes during repetitive stimulation , suggesting that Kir4 . 1-/- mice may display some compensatory mechanisms attempting to maintain extracellular K+ homeostasis . In addition , our model reveals that specific and acute inhibition of Kir4 . 1 channels slows down , but does not abolish , astroglial uptake of excess K+ during single , tetanic and repetitive stimulations , confirming that astroglial Na/K ATPases , included in our model , also contribute to K+ clearance [41] . Contrary to action potentials , characterized by a very fast dynamics in the order of a few milliseconds , astroglial K+ buffering lasts tens of seconds . As shown in the present study , most of extracellular K+ released by neurons is first cleared by astrocytes through Kir4 . 1 channels . To determine the factors controlling the slow timescale of astroglial K+ clearance , we focused on Kir4 . 1 channels . Because the astroglial leak conductance ( equation 23 ) is six times smaller than the Kir4 . 1 channel conductance , we neglected it . The dynamics of astrocytic membrane potential VA is described by equation 23 , where the membrane capacitance is CA ≈ 15 pF and the maximal Kir4 . 1 channel conductivity isGKir≈60pS . In that case , using equation 23 , the time constant of Kir4 . 1 channel-mediated return to equilibrium of astroglial membrane potential τA is defined as We obtain the following approximation τA ≈ 0 . 6s using equation 23 and the parameters of table 1 . This time constant is consistent with the fitted exponential decay time obtained in our simulations and experiments for a single stimulation where we obtained τ ≈ 0 . 7s . However , simulations for stronger stimulations indicate an increase of τ to approximatively 4 seconds ( tetanic stimulation ) and 9 seconds ( repetitive stimulation ) . This increase in clearance duration is due to the dependence of the Kir4 . 1 current to [K+]o , as illustrated by the IV relation ( Fig . 1C ) and described in equation 22 . The Nernst potential VKA increases for strong stimulations ( tetanic and repetitive ) , which slow down the kinetics of astrocytic membrane potential VA through the term 1+exp ( VA−VKA−V2AV3A ) in equation 22 . We conclude that the slow time scale of K+ clearance is in part due to the availability of Kir4 . 1 channels at low and high [K+]o . This clearance timescale is much longer than the glutamate clearance rate of τglu ≈ 15 ms that we previously reported [42] . Moreover , the redistribution of K+ released by neurons during the different regimes of activity shows that the higher the activity , the lower the proportion of released K+ remains transiently in the extracellular space . This suggests that Kir4 . 1 channels have a strong uptake capacity , especially for high regimes of activity ( [K+]o up to 5–6 mM ) . Remarkably , our model reveals that astroglial Kir4 . 1 channels strongly regulate neuronal firing induced by high stimulation regime such as repetitive stimulation . Kir4 . 1channels are crucially involved in regulation of [K+]o during this regime of activity , most likely because such stimulation triggered long-lasting neuronal release of K+ ( 20 mM over 30 seconds , Fig . 3G ) resulting in a sustained , but moderate increase in [K+]o ( >6 mM for ~22 s , Fig . 4I , J ) , compared to the neuronal release . These data suggest that during repetitive stimulation , astrocytes can buffer up to ~14 mM of [K+]o and thereby preserve neuronal firing . However , astroglial Kir4 . 1 channels slightly impact neuronal firing induced by single and tetanic stimulations , probably because these stimulations only triggered transient neuronal K+ release ( 0 . 9 mM over 300 ms ( Fig . 3A ) and 1 . 9 mM over 1 . 3 s ( Fig . 3D ) , respectively ) , resulting in a short and small increase in [K+]o ( >2 . 7 mM for ~450 ms for single stimulation ( Fig . 4A , B ) , and >3 . 5 mM for 1 . 5 s for tetanic stimulation ( Fig . 4E , F ) ) . Nevertheless , we show a prominent and specific involvement of astroglial Kir4 . 1 channels in probabilistic firing activity induced by 3 to 10 Hz sub-firing stimulations ( Fig . 5 ) , suggesting a key role of these channels in sustained theta rhythmic activity . Interestingly , these data imply that Kir4 . 1 channels can contribute to fine tuning of neuronal spiking involving low , but long-lasting , increase in [K+]o . Thus besides gliotransmission , regulation of [K+]o by Kir4 . 1 channel provides astrocytes with an alternative active and efficient mechanism to regulate neuronal activity . Several studies have reported decreased Kir4 . 1 protein levels and Kir functional currents in sclerotic hippocampus from epileptic patients [43–46] . Whether these changes are the cause or the consequence of epilepsy is still an open question . However , Kir4 . 1-/- mice display an epileptic phenotype [22 , 47] and missense mutations in KCNJ10 , the gene encoding Kir4 . 1 , have been associated with epilepsy in humans [48 , 49] . These data thus suggest that impairment in Kir4 . 1 function leading to alterations in [K+]o dynamics , as shown in our study , may cause epilepsy . Remarkably , dysfunction of [K+]o regulation by Kir4 . 1 channels is likely involved in other pathologies , since it contributed to neuronal dysfunction in a mouse model of Huntington’s disease [50] and the presence of antibodies against Kir4 . 1 channels in glial cells was recently found in almost 50% of multiple sclerosis patients [51] . Thus astroglial Kir4 . 1 channels may well represent an alternative therapeutic target for several diseases .
Experiments were carried out according to the guidelines of European Community Council Directives of January 1st 2013 ( 2010/63/EU ) and our local animal committee ( Center for Interdisciplinary Research in Biology in Collège de France ) . All efforts were made to minimize the number of used animals and their suffering . Experiments were performed on the hippocampus of wild type mice ( C57BL6 ) . For all analyses , mice of both genders and littermates were used ( PN19–PN25 ) . Acute transverse hippocampal slices ( 400 μm ) were prepared as previously described [42 , 52–54] from 19–25 days-old wild type mice . Slices were kept at room temperature ( 21–23°C ) in a chamber filled with an artificial cerebrospinal fluid ( ACSF ) composed of ( in mM ) : 119 NaCl , 2 . 5 KCl , 2 . 5 CaCl2 , 1 . 3 MgSO4 , 1 NaH2PO4 , 26 . 2 NaHCO3 and 11 glucose , saturated with 95% O2 and 5% CO2 , prior to recording . Acute slices were placed in a recording chamber mounted on a microscope including infra-red differential interference ( IR-DIC ) equipment , and were bathed in ACSF perfused at 1 . 5 ml/min . ACSF contained picrotoxin ( 100 μM ) , and connections between CA1 and CA3 regions were cut to avoid epileptic-like activity propagation . Extracellular field and whole-cell patch-clamp recordings were obtained using glass pipettes made of borosilicate . Astroglial and postsynaptic responses were evoked by Schaffer collateral stimulation ( 0 . 05Hz ) in the CA1 stratum radiatum region with glass pipettes filled with ACSF ( 300–700 kΩ ) . Astrocytes from stratum radiatum were recognized by their small soma size ( 5–10 μm ) , very low membrane resistance and hyperpolarized resting membrane potentials ( ≈- 80 mV ) , passive properties of their membrane ( linear I-V ) , absence of action potentials , and large coupling through gap junctions . Field excitatory postsynaptic potentials ( fEPSPs ) were obtained in 400 μm slices using pipettes ( 4–6 MΩ ) located in the stratum radiatum region . Stimulus artifacts were suppressed in representative traces . Whole-cell recordings were obtained from CA1 astrocytes , using 4–6 MΩ glass pipettes containing ( in mM ) : 105 K-Gluconate , 30 KCl , 10 HEPES , 10 Phosphocreatine , 4 ATP-Mg , 0 . 3 GTP-Tris , 0 . 3 EGTA ( pH 7 . 4 , 280 mOsm ) . Prolonged repetitive stimulation was performed for 30 s at 10 Hz . Post-tetanic potentiation was evoked by stimulation at 100 Hz for 1 s in the presence of 10 μM CPP ( ( Rs ) -3- ( 2-Carboxypiperazin-4-yl- ) propyl-1-phosphonic acid ) . Recordings were performed with Axopatch-1D amplifiers ( Molecular Devices , USA ) , at 10 kHz , filtered at 2 kHz , and analyzed using Clampex ( Molecular Devices , USA ) , and Matlab ( MathWorks , USA ) softwares . The data represent mean ± SEM . Picrotoxin was from Sigma and CPP from Tocris . We present here the biophysical model we have built to describe K+ dynamics during neuronal activity and specifically the role of astroglial Kir4 . 1 channels . After Schaffer collateral stimulation , excitatory synapses release glutamate molecules that activate postsynaptic neurons . We modeled this step by classical facilitation/depression model [55] . The resulting postsynaptic activity triggers ionic release in the extracellular space and a change in the astrocytic membrane potential through ion uptake . We used the average neuronal potential and mass conservation equations for ionic concentrations to model changes in astrocytes . We have built a tri-compartment model , which accounts for: 1 ) the neuron , 2 ) the astrocyte and 3 ) the extracellular space . We included voltage gated channels , Na/K pumps and astrocytic Kir4 . 1 channels . To account for the stimulation of Schaffer collaterals that induce a postsynaptic response in the CA1 stratum radiatum region , we used a facilitation-depression model [55–57] . where f is the input function . For a single stimulation generated at time tstim , f ( t ) = δ ( t-tstim ) . A stimulation instantaneously activates a fraction Use of synaptic resources r , which then inactivates with a time constant τinac and recovers with a time constant τrec In the simulations , at time t = tstim , r and e respectively decreases and increases by the value User . The synaptic current Iapp is proportional to the fraction of synaptic resources in the effective state e and is given by Iapp = Asee ( the parameter Ase is defined in table 1 ) . We used the following definitions for the input function f: The dynamics of the neuronal membrane potential , VN , follows the classic Hodgkin Huxley ( HH ) equations [58] . with rate equations αn ( VN ) =0 . 01 ( VN+10 ) exp ( 0 . 1 ( VN+10 ) ) −1 ( 12 ) βn ( VN ) =0 . 125exp ( VN/80 ) ( 13 ) αm ( VN ) =0 . 1 ( VN+25 ) exp ( 0 . 1 ( VN+25 ) ) −1 ( 14 ) βm ( VN ) =4exp ( VN/18 ) ( 15 ) αh ( VN ) =0 . 07exp ( VN/20 ) ( 16 ) βh ( VN ) =1exp ( 0 . 1 ( VN+ 30 ) ) +1 ( 17 ) Vrest is the resting membrane potential and VKN and VNaN are respectively the K+ and Na+ equilibrium potentials and are given by the Nernst equations VNaN=RTFln ( Na0NaN ) ( 18 ) VKN=RTFln ( K0KN ) ( 19 ) where Na0 and NaN are respectively the extracellular and neuronal sodium concentrations , and K0 and KN are respectively the extracellular and neuronal K+ concentrations that may vary as we shall describe below . We complete the description of all the neuronal currents with a leak current IlN=glN ( VN−VlN ) ( 20 ) which stabilizes the membrane potential at its resting value . Finally , the neuronal membrane potential satisfies the equation CNdVNdt=− ( INa+IK+IlN+Iapp ) ( 21 ) where Iapp is the synaptic current derived from equation 1 . To account for the K+ dynamics in astrocytes , we modeled the Kir4 . 1 channel according to its biophysical properties [59] and I-V curve [60] . The total astroglial current IKir depends on the membrane potential , the extracellular ( K0 ) and the astrocytic ( KA ) K+ concentrations , and is approximated by IKir=GKir ( VA−VKA−VA1 ) ( K01+exp ( VA−VKA−VA2VA3 ) ) ( 22 ) where VKA is the Nernst astrocyte K+ potential , VA , the astrocyte membrane potential , K0 is the extracellular K+ concentration and VA1 ( an equilibrium parameter , which sets Kir current to 0 at-80 mV ) , VA2 and VA3 are constant parameters calibrated by the I-V curve ( Fig . 1C , [60] ) , as detailed below . The second term of equation 22 describes the dependence of IKir to the square root of K0 [60–64] and to the steady state open/close partition function of Kir4 . 1 channels according to the Boltzmann distribution [59] , which includes dynamic variations of potassium Nernst potential during neuronal activity . Adding a leak current IlA = glA ( VA—VlA ) , which stabilizes the astrocyte membrane potential at—80 mV , the astrocyte membrane potential VA satisfies the equation CAdVAdt=− ( IKir+IlA ) ( 23 ) where IKir is defined by relation 22 . We fitted the Kir4 . 1 channel I-V curve ( equation 22 ) using the experimental recordings for the Kir4 . 1 channel ( 3 mM [K+] ( Fig . 4 in [60 , 65] ) . We first obtained that VA1 = ( VrestA − 26ln ( 3/145 ) ) = −14 . 83 mV where VrestA = −80mV ( potential for which the current is zero ) . We then used the Matlab fitting procedure for a single exponential with formula 22 changed to ( V−VA1−26ln ( 3/145 ) ) 3I with ( V from-100 to 20 mV ) to get that VA2 = 34 mV and VA3 = 19 . 23 mV ( table 1 ) . Varying [K+]o by 0 . 5 mM did not affect significantly the Kir4 . 1 channel I-V curve , confirming its robustness . The K+ resting concentrations in neurons and astrocytes are maintained by Na/K pumps that balance the outward K+ and inward Na+ fluxes . The associated pump currents ipump , k ( index k = N for the neuron , k = A for the astrocyte ) depend on the extracellular K+ K0 and intracellular Na+ concentrations ( NaN for the neuron and NaA for the astrocyte ) and follow the same equation as [66] , ipump , k=imaxk ( 1+7 . 3K0 ) −2 ( 1+10Nak ) −3 for k=N , A ( 24 ) where imaxk is a constant ( table 1 ) . We converted the different electrogenic neuronal and astrocytic channel currents into ionic fluxes [13] . A current I across a membrane induces a flow of charge i equals to δQ = I per unit of time . The corresponding change in extracellular concentration is given by I/ ( qNAVol0 ) , where q = 1 . 6 * 10–19C is the charge of an electron , NA the Avogadro number and VolN , VolA andVol0 are the neuronal , astrocytic and extracellular volume respectively . To model the ionic concentration dynamics , we converted the currents INa , IK and IKir to the corresponding ionic fluxes iNa , iK and iKir We describe in the following paragraphs the equations for the ionic concentrations in the three compartments ( neuron , extracellular space and astrocyte ) . To determine the system of equations for the K+ fluxes , we use the mass conservation law for the extracellular K0 , the neuronal KN and the astrocytic KA K+ concentrations . The extracellular K+ K0 increases with the neuronal current IK ( see equation 8 ) , which is here converted to iK ( ion flux ) , but it is also uptaken back into neurons with a flux 2 ipumpN ( the factor 2 is described in [67] and into astrocytes as the sum of the two fluxes 2 ipumpA plus iKir . Similarly , we obtain the equations for the neuronal and astrocytic K+ to balance the various fluxes . Finally , we get To study quantitatively the acute and selective role of astroglial Kir4 . 1 channels in neuroglial K+ dynamics , we inhibited the Kir4 . 1 current in our tri-compartment model . We thus set at zero both the Kir4 . 1 current and the leak term . To compensate for the loss of K+ fluxes through astroglial Kir4 . 1 channels , we added in equation 27 a constant K+ flux to maintain [K+]o at an equilibrium value of 2 . 5 mM . This constant K+ flux in astrocytes could be mediated by various channels or transporters such as two pore domain potassium channels ( K2P such as TWIK-1 , TREK-1 , TREK-2 and TASK-1 ) , delayed rectifier potassium channels ( Kv1 . 1 , 1 . 2 , 1 . 5 and 1 . 6 ) , rapidly inactivating A-type potassium channels ( Kv1 . 4 ) , glutamate transporters or connexin43 hemichannels . However , since TASK-1 [68] and Cx43 hemichannels [69] are thought to be active in basal conditions , they are more likely to mediate such flux . Similarly to the K+ dynamics , the equations for the Na+ fluxes are derived using the balance between the neuronal , astrocytic and extracellular concentrations . However , the main differences are that the pump exchanges 2 K+ for 3 Na+ ions , leading to the coefficient 3 in front of the pump term . In addition , to stabilize the sodium concentrations , we added two constant leak terms iNalA and iNalN ( values given in table 1 ) , as classically used [24] , dNa0dt=iNa+ iNalN+3ipumpN+3ipumpA+ iNalA ( 28 ) dNaNdt= ( −iNa−3ipumpN− iNalN ) Volo VolN ( 29 ) dNaAdt= ( −iNalA−3ipumpA ) Volo VolA ( 30 ) Numerical simulations . Simulations , numerical integrations and fitting computations were performed in Matlab . We used Runge Kunta fourth order method for the simulations , which were numerically stable . We used a time step of ∆t = 0 . 1 ms ( simulations were repeated with smaller time step to check whether numerical accuracy was affecting results ) . The leak currents parameters were adjusted to stabilize the model at the resting membrane potentials ( - 70 mV and - 80 mV for neurons and astrocytes respectively ) and resting concentrations ( neuronal [K+] and [Na+]: 135 mM and 12 mM , respectively; extracellular [K+] and [Na+]: 2 . 5 mM and 116 mM , respectively; astrocytic [K+] and [Na+]: 135 mM and 12 mM , respectively ) . The parameters for the Hodgkin Huxley equations were also adjusted to these concentrations . Approximation of time constants . Time constants τ of simulated extracellular K+ transients were fitted to curves using a single exponential ( e−tτ ) ( Fig . 4B , F , J ) . For all the fits obtained on the numerical simulation curves , we obtained an error estimation R-square ≥ 0 . 97 . Time constants τ of experimental and simulated astroglial membrane potentials were calculated by computing the rise and decay times between 20% and 80% of the maximal peak amplitude responses ( Fig . 2D , H , L ) . All time constants τ were fitted to curves using a single exponential ( e−tτ ) . For all the fits obtained on the numerical simulation curves , we obtained an error estimation R-square ≥ 0 . 97 . Approximation of facilitation/depression model parameters . To account for the synaptic properties of CA1 pyramidal neurons following single , tetanic and repetitive stimulations , we generated a synaptic current using the depression-facilitation model ( equation 1 ) where Iapp depends on the input functions fs ( t ) ( equation 4 ) , fTT ( t ) ( equation 5 ) and fRs ( t ) ( equation 6 ) , respectively ( Fig . 2A , E , I ) . The synaptic current parameters were fitted to experimental recordings [26] by matching the time of maximal peak amplitude of fEPSP with the one of Iapp in control conditions ( τ = 300 ms , τinact = 200 ms ) . The parameters for the Kir4 . 1 inhibition condition in the model were extracted from our experimental results on Kir4 . 1 glial conditional knockout mice [26] and are given by τrec = 500 ms , τinact = 160 ms . When Kir4 . 1 channels are inhibited ( model ) or knocked-out ( experiment ) , the maximal peak amplitudes of the applied synaptic currents in the model ( Iapp ) and fEPSPs recorded experimentally are increased compared to control conditions [26] . We imposed an initial input at various frequencies ( 0 . 1 , 1 , 3 , 5 , 10 , 50 Hz ) . Each input is generated by a sub-firing square current lasting 5 ms ( Iapp ) . In addition , we added a Brownian noise of amplitude σ = 0 . 68 pA2 ms-1 to induce neuronal membrane potential fluctuation ( equation 21 ) , which amplitude ( 1 mV ) was chosen to induce a probabilistic firing of 0 . 2 , matching the CA1 pyramidal cells synaptic release probability p = 0 . 2 ( probability to induce a postsynaptic response in equation 1 ) [70] . Using the tri-compartment model , we simulated at various frequencies a quantity that we called the observed firing probability defined empirically at time t as the time dependent ratio of the number of spikes observed at time t to the total number of simulations . | Neural excitability relies on precise inward and outward ionic fluxes . In particular , potassium ions , released by neurons during activity , have to be taken up efficiently to prevent hyperexcitability . Astrocytes , the third element of the synapse , play a prominent role in extracellular potassium homeostasis . Thus unraveling the dynamics of the neuroglial potassium cycle during neurotransmission and the underlying molecular pathways is a key issue . Here , we have developed a tri-compartment model accounting for potassium dynamics between neurons , astrocytes and the extracellular space to quantify the specific and acute contribution of astroglial Kir4 . 1 channels to extracellular potassium levels and to the moment-to-moment neurotransmission . We demonstrate that astroglial Kir4 . 1 channels are sufficient to account for the slow membrane depolarization of astrocytes and crucially contribute to extracellular potassium clearance during basal and high activity . We also show that astrocytes buffer in less than 10 seconds more than 80% of the potassium released by neurons , leading to recovery of basal extracellular potassium levels and neuronal excitability . Remarkably , we found that Kir4 . 1 channels also prominently regulate slow 3 to 10 Hz rhythmic firing activity . Altogether , these data show that Kir4 . 1 channels acutely regulate extracellular potassium and neuronal excitability during specific patterns of activity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | The Neuroglial Potassium Cycle during Neurotransmission: Role of Kir4.1 Channels |
Translationally Controlled Tumor Protein ( TCTP ) controls growth by regulating the G1/S transition during cell cycle progression . Our genetic interaction studies show that TCTP fulfills this role by interacting with CSN4 , a subunit of the COP9 Signalosome complex , known to influence CULLIN-RING ubiquitin ligases activity by controlling CULLIN ( CUL ) neddylation status . In agreement with these data , downregulation of CSN4 in Arabidopsis and in tobacco cells leads to delayed G1/S transition comparable to that observed when TCTP is downregulated . Loss-of-function of AtTCTP leads to increased fraction of deneddylated CUL1 , suggesting that AtTCTP interferes negatively with COP9 function . Similar defects in cell proliferation and CUL1 neddylation status were observed in Drosophila knockdown for dCSN4 or dTCTP , respectively , demonstrating a conserved mechanism between plants and animals . Together , our data show that CSN4 is the missing factor linking TCTP to the control of cell cycle progression and cell proliferation during organ development and open perspectives towards understanding TCTP’s role in organ development and disorders associated with TCTP miss-expression .
The correct implementation of organs with unique shape , size and function is fundamental to the development of all multicellular organisms and is the result of coordinated cellular processes requiring key molecular actors . One such key player in eukaryotes is the Translationally Controlled Tumor Protein ( TCTP ) . TCTP was discovered in the 1980’s as a protein positively regulated at the translational level in many tumors [1 , 2] . TCTP is a highly-conserved protein found in all eukaryotes . TCTP was reported to be involved in several cellular processes , including cell proliferation , cell growth , malignant transformation , apoptosis and protection against various cellular stresses [3–8] . In plants as in animals , TCTP loss of function leads to embryonic lethality , because of slower cell cycle progression and reduced cell proliferation [6] and reduced cell proliferation associated with excessive cell death [9 , 10] , respectively . The fact that TCTP loss of function is lethal demonstrates its major role in eukaryote development , but also hampered the full comprehension of its exact roles . By performing embryo rescue , we generated the first TCTP full knockout adult organism which allowed us to demonstrate that TCTP controls cell cycle progression by regulating G1/S transition and that this role is conserved between plants and animals [6] . However , how TCTP controls the G1/S transition and cell proliferation remains unknown in plants as in animals . To better understand such a role , we searched for TCTP interacting proteins and identified that TCTP interacts physically with CSN4 , one of the eight subunits of the Constitutive Photomorphogenesis 9 ( COP9 ) Signalosome ( CSN ) , initially discovered in plants [11 , 12] and conserved in eukaryotes [13] . CSN regulates the activity , the assembly and/or subunits stability of CULLIN-RING ubiquitin Ligase ( CRLs ) complexes , a major class of the E3 ligase complexes in eukaryotes [14] involved in the polyubiquitination of proteins targeted to degradation [15] . The proper functioning of CRL has been shown to be absolutely required for various functions associated with the control of cell cycle , transcription , stress response , self-incompatibility , pathogen defense and hormones and light signaling [16–18] . CSN mainly regulates CLR activity through the removal of the post-translational modification RUB/NEDD8 ( Related to UBiquitin/Neural Precursor Cell Expressed , Developmentally Down-Regulated 8 ) from its CULLIN ( CUL ) subunit [18] . The eight subunits of CSN are mandatory for the deneddylation activity [19–21] . Here , we show that TCTP interacts physically and genetically with CSN4 to control cell cycle progression . Our data demonstrate that downregulation of TCTP or CSN4 both leads to retarded G1/S transition , slower cell cycle progression and delay in plant development . Consistent with these data , knockout of TCTP is associated with increased CUL1 deneddylation . Conversely , over-accumulation of TCTP leads to accelerated cell cycle and plant development , associated with over-accumulation of neddylated CUL1 . We also show that in Drosophila , the downregulation of dTCTP or dCSN4 is associated with defects in cell proliferation . Moreover , similar to our observations in plants , dTCTP downregulation is accompanied by increased CUL1 deneddylation . These data suggest that TCTP interacts with CSN4 and such interaction acts on the CUL1 neddylation status and affects CRL complex activity early during cell cycle progression influencing organ development both in plants and animals .
To gain insights into how TCTP controls cell cycle progression , we searched for its interacting proteins . Immunoprecipitation followed by mass spectrometry ( IP/MS ) experiments were performed using protein extracts from Arabidopsis line expressing 35S::AtTCTP-GFP . Wild-type ( WT ) Col-0 and free-GFP-overexpressing plants lines ( 35S::GFP ) served as controls . Among the proteins that co-immunoprecipitated with TCTP-GFP , but were absent in the control samples , we identified AtCSN4 , a subunit of COP9 Signalosome , that was previously shown to be involved in the regulation of cell cycle progression [22] . To confirm this interaction , we generated line AtTCTPg-GFP that expresses a TCTP-GFP fusion under the control of its own promoter and line AtTCTPg-GFP/35S::AtCSN4-Flag , that , in addition , expresses AtCSN4-Flag under the control of the constitutive CaMV 35S promoter ( 35S::AtCSN4-Flag ) . Anti-GFP antibody was used to immunoprecipitate the TCTP-GFP fraction from total proteins extracted from 10 days-old seedlings ( S1A Fig ) , or from mature green seeds harvested from green siliques ( S1B Fig ) . The presence of CSN4 ( 46 kDa ) was then revealed in the immunoprecipitated fraction . AtCSN4 was found in the AtTCTP-GFP enriched fraction from both 10 days-old seedlings and mature green seeds ( S1A and S1B Fig , respectively ) . No trace of AtCSN4 was detected in WT Col-0 used as control and treated in parallel . Next , we confirmed that endogenous AtCSN4 is able to co-immunoprecipitate with AtTCTP . For this , we performed immunoprecipitation on total proteins extracted from inflorescence of AtTCTPg-GFP and 35S::AtTCTP-GFP . As shown in Fig 1A , CSN4 co-immunoprecipitated with AtTCTP-GFP in both lines ( Figs 1A and S1C ) . The interaction was further confirmed by pulling-down AtCSN4 . Using plant lines 35S::AtCSN4-GFP , overexpressing AtCSN4 , and 35S::AtCSN4-GFP/35S::AtTCTP , overexpressing both proteins , we show that TCTP co-immunoprecipitates with AtCSN4-GFP from both plant lines ( Figs 1B and S1D ) . Finally , to address if Drosophila TCTP also interacts with CSN4 , we performed co-immunoprecipitation experiments using plant lines expressing Drosophila dTCTP in tctp knockout genetic background ( line 35S::dTCTP-GFP ) . Previously we demonstrated that dTCTP was able to fully complement loss-of-function tctp [6] . As shown in Fig 1A , CSN4 co-immunoprecipitated with AtTCTP-GFP and also with dTCTP-GFP ( Figs 1A and S1C ) , thus suggesting conservation of the TCTP/CSN4 interaction between plants and animals . Next , we investigated the sub-cellular localization of CSN4 compared to TCTP . In agreement with co-immunoprecipitation data , co-expression of AtTCTP-mCherry and AtCSN4-GFP in tobacco leaf epidermal cells showed that these proteins co-localize in planta ( Fig 1C ) . To further investigate AtTCTP-AtCSN4 interaction , we performed Bi-molecular Fluorescence Complementation ( BiFC [23] ) experiments by protein fusion to split YFP moieties and co-expression in tobacco cells . The data demonstrate that AtTCTP and AtCSN4 interact in planta ( Fig 1D ) , confirming the co-immunoprecipitation results . Similar BiFC complementation results were obtained regardless if AtTCTP and AtCSN4 were fused to YFPCter or YFPNter ( Fig 1D ) . BiFC experiments also showed that both AtTCTP and AtCSN4 were able to form homodimers ( S2 Fig ) . However , the localization of AtTCTP-AtCSN4 heterodimer was distinct from that of the AtTCTP or AtCSN4 homodimers , thus corroborating TCTP-CSN4 in vivo interaction . No signal was observed when one of the vectors was empty ( S2 Fig ) . Taken together these data demonstrate that AtTCTP and AtCSN4 interact physically in vivo . To understand the biological significance of the AtTCTP-AtCSN4 interaction , we performed genetic interaction analyses using knockdown or overexpressor lines for both proteins . Previously , we showed that while AtTCTP knockout leads to embryo lethality , its down-regulation via RNAi ( RNAi-AtTCTP lines ) gave viable plants that nevertheless showed developmental defects [6] . Similar to plants mutants for AtTCTP , plants knockout for AtCSN4 showed severe developmental defects during early seed germination leading to seedling death in few days after germination ( S3 Fig ) , thus in agreement with Dohmann et al . [22] . To overcome this difficulty , we generated knockdown plants via expression of a RNAi directed against AtCSN4 ( line RNAi-AtCSN4 ) . Western blot analyses confirmed the downregulation of AtTCTP and AtCSN4 in the corresponding RNAi lines ( S4 Fig ) . RNAi-AtCSN4 exhibited significant delay in rosette development compared to Col-0 , thus a similar phenotype as RNAi-AtTCTP plants ( Fig 2A–2C ) . Measurements of growing rosette diameters from 8 days post-germination until bolting , confirmed that these genotypes are impaired in development and that the difference in growth starts to be visible as early as 8 days after germination ( Fig 2C ) . Similar to RNAi-AtTCTP plants , the inflorescence stems of AtCSN4-RNAi plants were shorter and the plants exhibited a dwarf phenotype ( Figs 2B and S5 ) . Additionally , RNAi-AtCSN4 plants had very short internodes between siliques resulting in a “bushy” plant phenotype ( Figs 2B and S5 ) . Plants overexpressing AtCSN4 ( line 35S::AtCSN4 ) had normal development and adult plants had similar size as WT Col-0 plants ( Fig 2 ) . Plants overexpressing AtTCTP ( line 35S::AtTCTP ) exhibited accelerated growth and reached adult size earlier than the wild-type ( Fig 2 ) , in agreement with previously reported data [6] . The double overexpressor line ( 35S::AtCSN4/35S::AtTCTP ) exhibited no additive effect and behaved as the single overexpressor line 35S::AtTCTP ( Fig 2 ) . Crosses between RNAi-AtTCTP and AtCSN4 overexpressor lines on the one hand ( line RNAi-AtTCTP/35S::AtCSN4 ) and between RNAi-AtCSN4 and AtTCTP overexpressor lines on the other hand ( 35S::AtTCTP/RNAi-AtCSN4 ) , yielded plants with delayed growth compared to the WT , thus a phenotype similar to the single RNAi-AtTCTP and RNAi-AtCSN4 knockdown plants , respectively ( Fig 2A–2C ) . We should note that 35S::AtTCTP/RNAi-AtCSN4 plants grew a little slower than RNAi-AtCSN4 , and RNAi-AtTCTP /35S::AtCSN4 plants grew a little faster than RNAi-AtTCTP plants . However , adult plants of the double transformant lines were indistinguishable from simple RNAi plants , demonstrating similar dwarf phenotype ( Fig 2B ) . These data show that overexpression of AtTCTP or AtCSN4 could not compensate the developmental anomalies induced by knockdown of AtCSN4 or AtTCTP , respectively , and suggest that TCTP requires functional CSN4 to control growth . Despite several attempts , by plant genetic crossing or by genetic transformation , we were unable to generate the RNAi-AtTCTP/RNAi-AtCSN4 double knockdown line . It is likely that double TCTP/CSN4 knockdown leads to plant lethality . To explore the cause of the growth defects observed in RNAi-AtTCTP and RNAi-AtCSN4 plants , we analyzed cell division and cell expansion profiles . Previously , we demonstrated that downregulation of AtTCTP and the resulting decrease in plant organs size are correlated with reduced cell proliferation activity [6] . To explore if AtCSN4 down-regulation is also associated with defects in cell proliferation , we performed kinematic analysis of leaf growth on plantlets grown in vitro [6 , 24] . Previously , kinematic of leaf growth analyses demonstrated that downregulation of AtTCTP results in smaller leaves compared to the WT , due to slower cell proliferation [6] . Similarly , in line RNAi-AtCSN4 a significant reduction in leaf area was observed compared to the WT Col-0 , starting from seven days after germination and the reduction was maintained during the whole observation period ( Fig 3A ) . The reduction in leaf size correlated with about 20% decrease in cell number starting from day seven after germination and was maintained all along the observation period ( Fig 3C ) , while no differences in cell size were observed ( Fig 3B ) , thus again similar to RNAi-AtTCTP [6] . These data demonstrate that , similar to AtTCTP , the leaf growth defects associated with AtCSN4 knockdown ( RNAi-AtCSN4 ) correlated with the decrease in cell number , while cell size remained unchanged . Using the slope of the log 2–transformed number of cell per leaf [25] , we calculated the cell division rate in RNAi-AtCSN4 plants . We observed that the cell division rate in RNAi-AtCSN4 was slower than in WT Col-0 plants ( Fig 3D ) in the early stage of leaf development where the cell division activity is higher [26] . Previously , we reported a similar tendency for RNAi-AtTCTP lines [6] . Determination of the number of newly produced cells per hour during leaf development showed that the lower cell division rate in RNAi-AtCSN4 line resulted in less newly produced cells at the beginning of leaf development ( S6 Fig ) . Interestingly , the cell division rate was maintained at a higher level for longer time in the RNAi-AtCSN4 , which indicates that a compensation mechanism likely exists . However , this was not enough to compensate the delay in leaf development , and therefore RNAi-AtCSN4 leaves stayed smaller than WT ( Fig 3A ) . Later in leaf development , cell division in both plants was reduced to a very low level ( Fig 3D ) . We investigated root growth in both RNAi-AtTCTP and RNAi-AtCSN4 lines . Similar to leaf growth , we also observed reduced root growth in developing seedlings of both lines ( S7A Fig ) . Both RNAi-AtTCTP and RNAi-AtCSN4 roots were shorter compared to WT starting as early as 3 days after germination ( S7A Fig ) . Similar to leaves and roots , petals of RNAi-AtCSN4 were also smaller in size compared to WT Col-0 ( S7B Fig ) . The petal size reduction was associated with reduced cell number , thus a phenotype similar to that observed in RNAi-AtTCTP ( S7B Fig ) [6] . These data corroborate the results obtained in leaves and demonstrate that , like for AtTCTP , downregulation of AtCSN4 leads to reduced organ size as a result of altered cell proliferation . However , conversely to leaves , we observed that in petals of both RNAi-AtCSN4 and RNAi-AtTCTP the defects in cell proliferation were associated with an increase in cell size ( S7B Fig ) , suggesting a compensation mechanism in the petals . AtTCTP overexpression ( line 35S::AtTCTP ) resulted in petals with increased size , but cell size was not affected , thus in agreement with Brioudes et al . [6] . AtCSN4 overexpression ( line 35S::AtCSN4 ) did not affect petal development , ( S7B Fig ) , thus corroborating the observed normal development of plants overexpressing AtCSN4 ( Fig 2 ) . Similar to observation during rosette development , overexpression of AtTCTP or AtCSN4 in RNAi-AtCSN4 and RNAi-AtTCTP , respectively , could not compensate for the petal developmental defects of the RNAi lines ( S7B Fig ) . Line overexpressing both AtTCTP and AtCSN4 showed similar phenotype to 35S::AtTCTP ( S7B Fig ) . These data together show that the downregulation of AtCSN4 leads to slower cell proliferation associated with reduced organs size , thus a phenotype similar to that observed in AtTCTP mutant plants . To further identify the origin of this reduced cell proliferation activity , we investigated cell cycle progression in tobacco BY-2 cells down-regulating or overexpressing NtCSN4 ( line RNAi-NtCSN4 and 35S::NtCSN4 , respectively ) or NtTCTP ( RNAi-NtTCTP , 35S::NtTCTP , respectively ) . Both , CSN4 and TCTP proteins accumulated at low levels in RNAi BY-2 cells and over-accumulated in overexpressor BY-2 cells , as demonstrated by Western blot analysis ( S8 Fig ) . Wild-type BY-2 cells and BY-2 cells under- or over-accumulating TCTP ( RNAi-NtTCTP , 35S::NtTCTP ) or CSN4 ( RNAi-NtCSN4 , 35S::NtCSN4 ) were synchronized using aphidicolin . Cell cycle progression was followed every 2 hours after aphidicolin release ( AAR ) using flow cytometry ( Fig 3E ) . Normal progression of cell cycle over time was observed in wild-type BY-2 cells with rapid reduction of G1 cells ( 2N ) and a concomitant increase of G2 cells ( 4N ) over the first 5 hours AAR , followed by a decrease of G2 cells with mitosis ending at about 13h AAR . In agreement with previously reported data [6] , the G1/S transition in RNAi-NtTCTP BY-2 cells occurred with 4 hours delay compared to the wild-type and this delay was maintained all along the cell cycle ( Fig 3E ) . Similarly , RNAi-NtCSN4 BY-2 cells also showed a slower cell cycle progression with about the same 4 hours delay at the G1/S transition compared to wild-type BY-2 cells ( Fig 3E ) . Like for RNAi-NtTCTP , the 4 hours delay of cell cycle progression in RNAi-NtCSN4 was maintained until the end of the cell cycle , thus corroborating the kinematic of growth data obtained in Arabidopsis leaves ( Fig 3 ) . These data together demonstrate that the downregulation of either NtTCTP or NtCSN4 leads to comparable delays in cell cycle progression and that such delay occurs at G1 and/or early S phase . Conversely to BY-2 cells knockdown for NtTCTP , BY-2 cells overexpressing NtTCTP ( 35S::NtTCTP ) entered G1/S transition about 2 hours earlier than wild-type BY-2 ( Fig 3E ) . However , these cells completed their first cell cycle the same time as WT BY-2 cells , at 13h AAR . This indicates that 35S::NtTCTP BY-2 cells has longer S or M2 phase to compensate faster G1 . In agreement with the absence of cell proliferation and developmental defects in 35S::AtCSN4 Arabidopsis line ( Figs 2 and S7B ) , no significant difference in cell cycle progression was observed in 35S::NtCSN4 BY-2 cells , compared to the wild-type ( Fig 3E ) . To further explore at which step of the cell cycle TCTP and CSN4 precisely act and to confirm BY-2 results directly in planta , we performed cumulative EdU labeling in the root tips of the different Arabidopsis lines to estimate S-phase length and total cell cycle length . Results show that , similar to observations in BY-2 cells , total cell cycle length in root tips was increased by 4 hours , both in RNAi-AtTCTP and RNAi-AtCSN4 lines , while S-phase length remained unchanged ( Table 1 ) . Likewise , no substantial changes were observed in cell cycle length in 35S::AtTCTP and 35S::AtCSN4 lines ( Table 1 ) , again consistent with the results obtained in BY-2 cells ( Fig 3C ) . Interestingly , S-phases lengths again remained unchanged , suggesting that the observed cell cycle slow-down after the rapid progression through G1/S phase seen in 35S::NtTCTP BY2 cells , did not affect S-phase but probably G2/M phase . All these data together with the TCTP-CSN4 co-immunoprecipitation and the in vivo interactions studies suggest that TCTP and CSN4 interact physically to control G1/S transition during cell cycle progression . CSN4 is one of the eight subunits of the CSN that regulates CRL activity via the post-translational RUB/NEDD8 modification of its CUL subunit . To investigate the biological significance of AtTCTP/AtCSN4 interaction , we evaluated the neddylation status of the Arabidopsis CULLIN1 ( CUL1 ) , a major target of CSN . Using an antibody against CUL1 , we were able to distinguish free CUL1 and the neddylated CUL1 ( CUL1NEDD8 ) forms ( Figs 4A–4C and S9A–S9C ) . By growing plants on medium with increasing concentrations of MLN4924 , a drug that inhibits CUL neddylation [27] , we confirmed that the observed protein bands indeed correspond to CUL1 and CUL1NEDD8 ( Figs 4C and S9C ) . In WT plants , we observed about 8–10 times more CUL1NEDD8 than free CUL1 while at 50μM MLN4924 we observed that almost all CUL1 were non-neddylated ( Figs 4C and S9C ) . Moreover , we demonstrate that WT inflorescence contains more neddylated CUL1 , while in WT seedlings non-neddylated CUL1 is more present ( Figs 4B and S9B ) . Inflorescence of a weak mutant of csn4 contains almost exclusively neddylated CUL1 ( Fig 4B ) in accordance with previous results [28] . Next , CUL neddylation status was analyzed in two independent tctp knockout lines ( tctp-1 and tctp-2 ) , in 35S::AtTCTP , RNAi-AtCSN4 and 35S::AtCSN4 ( Figs 4A and S9A ) . For both tctp-1 and tctp-2 , we observed a drastic decrease ( Fig 4A ) to complete absence ( S9A Fig ) of the CULNEDD8 , with a concomitant increase of free CUL1 , leading to a drop in CUL1NEDD8/CUL1 ratio in knockout lines ( Figs 4A and S9A ) . Overexpression of AtTCTP led to a slight increase in CUL1NEDD8 ( Fig 4A ) in agreement with the fact that although 35::AtTCTP plants grow faster , fully adult plants showed a phenotype similar to the wild-type [6] ( Fig 2 ) . These data suggest a role for AtTCTP in the regulation of CUL1 neddylation status , which in turn influences the activity of CRL complexes . Only small changes of CUL1NEDD8/CUL1 ratio were observed in RNAi-AtCSN4 plants ( Fig 4A ) . This is likely due to the fact that flowers already accumulated high level of neddylated CUL1 ( Fig 4B ) and that in RNAi-AtCSN4 lines we do not have full obliteration of AtCSN4 ( S4A and S4B Fig ) . In 35S::AtCSN4 , no decrease of CUL1NEDD8 was observed and the CUL1 neddylation status was similar to wild-type ( Fig 4A ) . This is in agreement with previous studies showing that overexpression of only one subunit of the COP9 complex does not modify its deneddylation activity [29] . The data also corroborate the fact that CSN4 overexpression does not affect cell cycle progression as well as organ and plant development ( Figs 2 and 3E and S7 ) . We demonstrated that CUL1 neddylation is affected in tctp mutants , suggesting that CRLs function in general might be affected . The best characterized SCF/CRL in Arabidopsis is SCFTIR1 implicated in auxin perception and signaling [30 , 31] . Therefore , we investigated if auxin responses are affected in tctp mutant . Auxin homeostasis reporter DR5rev::GFP and auxin efflux reporter PIN1::PIN1-GFP [32] were introgressed into tctp knockout plants . Similar to WT embryos , in tctp embryos PIN1::PIN1-GFP accumulated in apical cell lineage at globular and transition stage , then at heart stage the pattern switches to a PIN1 localization in the future cotyledons and across the provasculature ( S10A Fig ) . In agreement with these data , in tctp mutant embryos the accumulation pattern of auxin homeostasis reporter DR5rev::GFP was similar to that of WT embryos . Moreover , the accumulation pattern of auxin homeostasis reporter DR5rev::GFP in response to exogenous treatment with synthetic auxin 2 , 4D was also similar in WT and tctp embryos ( S10B Fig ) . These data show that the modification in CUL1 neddylation associated with AtTCTP mutation do not affect the auxin pathway nor the signaling via SCFTIR1 , indicating that the role of TCTP-CSN4 interaction in regulating CUL neddylation and CRL activity is specific to cell cycle . Previously , we demonstrated that the role of TCTP in regulating cell proliferation was conserved between plants and animals . Furthermore , we demonstrated that AtCSN4 is able to interact with dTCTP ( Figs 1A and S1C ) . Therefore , we investigated if the TCTP-CSN4 pathway has similar roles in Drosophila development as observed in plants . Using the UAS/GAL4 system [33] , first , we generated flies in which the expression of dTCTP was silenced via the expression of a dTCTP RNAi under the control of the eye-specific promoter eyless ( line ey>dTCTPi ) . eyless promoter drives gene expression in the eye imaginal disk anterior to the furrow at the time when eye cell progenitors are actively dividing [34] . Interference with dTCTP using the eyless promoter led to a significant size reduction of the eye , in agreement with Brioudes et al [6] ( Fig 5A ) . To study the role of CSN4 in the Drosophila eye , we used the P-element recessive lethal lines CSN4k08018 from the UCLA URCFG collection . Previous analysis of CSN4k08018reported a rough eye associated with loss of photoreceptor and patterning defects [35 , 36] . To further analyze the loss of dCSN4 function , we used the mosaic Tomato/GFP-FLP/FRT method in combination with the caspase inhibitor p35 [36 , 37] . In this system , the absence of red fluorescence ( tdTomato ) marks dCSN4-/- mutant photoreceptors in a background where all photoreceptors express GFP in the adult Drosophila eye . This method allowed the observation of mosaic eyes , in which regions where the dCSN4 gene have been inactivated can be detected by the absence of the tomato reporter ( Fig 5B , middle panel ) . dCSN4 inactivation led to a strong loss of photoreceptors as seen by the lack or diffuse GFP staining in dCSN4-/- mutant clones of Drosophila adult retina ( Fig 5B , left panel ) . Importantly , no effector caspase ( dcp-1 , death caspase-1 ) staining was detected in dCSN4-/- mutant clones in third instar eye discs ( Fig 5C ) . Moreover the loss of photoreceptor in dCSN4-/- mutant clones was not rescued by the expression of the caspase inhibitor p35 ( Fig 5B , lower panels ) , a protein that prevents apoptosis [38] . This indicates that the loss of photoreceptors in dCSN4-/- mutant clones is not due to increased apoptosis but rather impaired proliferation as in dTCTP mutants . Next , we examined the expression of the neuronal marker ELAV , which is progressively acquired in differentiating photoreceptor posterior to the morphogenetic furrow in third instar eye discs [34] . We observed a delay and a reduction of ELAV staining in dCSN4-/- mutant clones compared to wild type ( Fig 5D and 5E ) . The delay in the acquisition of ELAV marker in dCSN4-/- mutant clones could be the consequence of a delay in the proliferation of dividing photoreceptor progenitors as previously described in dachsund mutant that allows a tight coordination of proliferation and differentiation [39] . To explore whether the effect of TCTP on CUL1 neddylation is also conserved between plants and animals , we investigated CUL1 neddylation in Drosophila knockdown for dTCTP under the control of TUBULIN constitutive promoter ( line tub>dTCTPi ) . Interference with dTCTP under the TUBULIN constitutive promoter ( tub>dTCTPi ) led to severe larval developmental defects and to growth arrest after the first larvae instar ( Fig 5F ) . We evaluated the CUL1 neddylation status in tub>dTCTPi larvae knockdown for dTCTP exhibiting severe developmental delay , as compared to the wild-type flies and we observed a drastic decrease of the CULNEDD8 with a concomitant increase of free CUL1 ( Fig 5G ) . These data strongly suggest that , like in Arabidopsis , knockdown of dTCTP or dCSN4 led to impaired cell proliferation . Moreover , knockdown of dTCTP in Drosophila also affects CUL1 neddylation status in a similar manner than in Arabidopsis , suggesting a conserved role of TCTP in the control of CUL neddylation between plants and animals . Importantly , the fact that AtCSN4 interacts with dTCTP ( Fig 1C ) suggests that , similar to plant AtTCTP , Drosophila dTCTP controls CUL1 neddylation likely via its interaction with CSN4 .
In plants and in animals , TCTP is known to be implicated in many cellular processes , but its mode of action is largely unknown . Previously , we demonstrated that in Arabidopsis and in Drosophila , TCTP has a conserved role in the control of organ growth by regulating cell proliferation . We showed that TCTP regulates cell cycle progression more specifically at the G1/S transition [6] . To gain insight into the pathway by which TCTP fulfills this function , we identified its interactors in Arabidopsis . Here , we establish a functional relationship between TCTP and CSN4 , one of the eight subunits of the COP9 Signalosome , a complex conserved among eukaryotes [12] . Like tctp null mutants , csn4 mutants are not viable [22] ( and this study ) . We therefore analyzed partial loss-of-function lines using RNAi . Both RNAi-AtTCTP and RNAi-AtCSN4 Arabidopsis lines display dwarf phenotype and reduced organ size due to decreased cell proliferation , and more specifically to a delay in the G1/S transition as evidenced by tobacco BY-2 cell synchronization and EdU incorporation assays in Arabidopsis root meristematic cells . The ability of AtTCTP and AtCSN4 to interact and the similarity of the phenotypes of RNAi lines suggest that they could function in the same pathway , which was corroborated by our genetic analyses . No additive phenotypic effect was observed in the double overexpressor and the down-regulation of AtCSN4 was epistatic on AtTCTP overexpression , and reciprocally , AtCSN4 overexpression did not fully rescue the developmental defects of RNAi-AtTCTP plants . However , we observed that although the RNAi-AtTCTP plants that overexpress AtCSN4 exhibited delayed development compared to the wild-type , they grew faster than RNAi-AtTCTP plants . This could be due to the fact that CSN4 is likely involved in other biological process required for plant development , separately from AtTCTP [40] . We were unable to generate RNAi-AtTCTP/RNAi-AtCSN4 double knockdown line , likely because these plants are not viable . This could be due to the fact that both simple RNAi lines are only partially loss of function , and that simultaneous down-regulation of AtTCTP and AtCSN4 leads to defects comparable to what is observed in tctp or csn4 knockout lines . However , we cannot rule out the possibility that AtTCTP and AtCSN4 could also be separately involved in several other biological processes required for plant development [3 , 4 , 6 , 40] . This is supported by the fact that AtTCTP and AtCSN4 proteins are not always co-localized in the cell ( Fig 1C ) . Thus , TCTP/CSN4 interaction is most likely required at a defined time points during cell proliferation , but both proteins have additional functions independently of their interaction . Using synchronized BY-2 cells and EdU incorporation assays , we determined that both RNAi-AtTCTP and RNAi-AtCSN4 present similar delays at the G1/S transition , reinforcing the idea that the two proteins act in the same pathway . On the other hand , adding more AtTCTP leads to accelerated cell cycle progression . The fact that no effect on cell cycle progression was observed in plants over-accumulating AtCSN4 , suggests that AtTCTP is likely the limiting factor to control cell cycle progression in the AtTCTP-AtCSN4 pathway . CSN4 is part of the COP9 complex known to control CRL via neddylation status of their CULLINS ( CUL ) subunits [17] . We therefore asked whether TCTP and CSN4 might control CUL neddylation status . Indeed , we observed that both AtTCTP and dTCTP misexpression strongly impacts CUL1NEDD8/CUL1 ratio supporting an elegant hypothesis relative to how TCTP controls G1/S transition . In fact , it is conceivable that via its interaction TCTP could sequester CSN4 , preventing its association with the COP9 complex specifically at the G1/S transition . It is well established that the lack of one of the eight COP9 sub-units is sufficient to suppress the deneddylation activity [19 , 20 , 30 , 41] . Moreover , in Drosophila it was previously reported that during development , CSN4 functions to maintain self-renewal of stem cells , and the switch from self-renewal to differentiation requires the sequestration of CSN4 from the CSN by the protein Bam [42] . Another subunit , CSN5 was also demonstrated to be sequestered by the small protein Rig-G , in order to negatively regulate SCF-E3 ligase activity in mammalian cells [43] . We can thus imagine that a similar scenario exists between TCTP and CSN4 to drive cell cycle through the G1/S transition and maintain cell proliferation . This is the first time that a sequestration mechanism to regulate CSN activity is described in plants . This model is also consistent with the phenotypic defects triggered by TCTP deficiency ( Fig 6 ) . Franciosini et al . [44] suggested that during embryo maturation COP9 became deactivated , and subsequently reactivated at germination . It is possible that the role of TCTP during embryogenesis is to sequester CSN4 and prevent assembly and activity of COP9 complex during the G1/S transition . Moreover , it was demonstrated that CUL1 neddylation is increased during normal embryo development , thus the reduced neddylation in tctp embryos could explain their delayed development and death [6 , 44] . To our surprise , although AtCSN4 down-regulation induced severe developmental defects , we were not able to detect strong modification in CUL1NEDD8/CUL1 ratio in RNAi-AtCSN4 lines . This is probably due to the fact that CUL1NEDD8 level is already too high in the inflorescence to see a small increase due to reduced level of CSN4 accumulation and also to additional functions of CSN4 [16 , 18 , 40] independently of its interaction with TCTP . Also , because the CUL1NEDD8/CUL1 homeostasis is known to be highly dynamical process that is strictly regulated during plant development [44] , it is possible that AtCSN4 down-regulation in our RNAi lines is not strong enough to have measurable CUL1NEDD8 ratio changes at the studied developmental stages . However , we assumed that the developmental phenotypes observed in this line were due to compromised COP9 complex function . Indeed , previous work on dominant negative , weak alleles or RNAi lines of different CSN subunits demonstrated that the developmental defects , similar to those observed in our RNAi-AtCSN4 line , were due to disturbance of CSN activity [28 , 31 , 41 , 44] . TCTP down-regulation leads to a decrease of the CUL1NEDD8/CUL1 ratio . Since CSN4 is involved in the deneddylation of CULs , its down-regulation would be expected to increase this ratio , as described in csn4 mutant [22] . TCTP and CSN4 thus appear to act antagonistically on CULs neddylation , and one could therefore expect RNAi-AtCSN4 and RNAi-AtTCTP lines to display opposite phenotypic defects instead of the similarities we report here . However , the effect of neddylation on CRLs activity is extremely complex: csn mutants that produce more neddylated CUL are expected to increase CRL activity resulting in positive effect on development . However , all csn mutants are delayed in their development . On the other hand , treatment of plants with MLN924 , a drug that inhibits neddylation resulting in the accumulation of deneddylated CUL1 , also leads to plant lethality [27] , thus a similar phenotype as when TCTP is mutated . It is now well established that in addition to the CULNEDD8/CUL ratio , the temporal kinetics of CUL and NEDD8 association/dissociation controls CRL activity [45–48] . Indeed , to be fully active , CRLs need to undergo a cycle of neddylation and de-neddylation [14 , 49] and inactivation of factors with opposite roles on this cycle can thus result in the same cellular defects . On the other hand , CRL acts both on positive and negative cell cycle regulators and the timing of the degradation of these regulators has to be tightly coordinated [16] . Thus , opposite perturbation of CRL activity by TCTP or CSN4 downregulation can lead to similar phenotypes . Together , our results provide evidence for the role of TCTP and CSN4 in the control of cell cycle regulation via the modification of CUL1 neddylation status that affects CRL activity ( summarized in Fig 6 ) . However , the nature of the CRLs and their targets that account for this role on cell cycle regulation remains to be established . In plants , the mechanisms controlling the G1/S transition are poorly known . However , it seems that there are close similarities compared to animals . Indeed , overexpression of CKI results in G1 arrest of the cell cycle in plants [50] . Furthermore , ICK2/KRP2 , a CKI related protein , and other central cell cycle regulators such as E2Fc , a central transcription factor controlling G1/S transition in plants , must be degraded via the ubiquitin/26S proteasome pathway to allow the cells to go through the G1/S transition [51 , 52] . Although it was suggested that AtCSN4 is implicated in the G2/M transition [22] , it has also been reported that G1/S phase specific core cell cycle genes were overexpressed in csn4 mutant [22 , 53] . Furthermore , in human cells , it was reported that the downregulation of different CSN subunits can affect cell cycle in opposite ways , and the resulting perturbation of COP9 activity affect both G1/S and G2/M transition [54] . These published data suggest that CSN is likely equally important for G1/S and G2/M transitions in both plants and animals , and thus corroborate our findings here . They are also consistent with the fact that CUL deneddylation affects the activity of multiple CRL complexes . Our data show that TCTP regulates G1/S transition , in agreement with Brioudes et al [6] . It is likely that TCTP/CSN4 interaction is specifically interfering with CSN function at G1/S . In this scenario , TCTP controls cell cycle through sequestration of CSN4 , leading to CSN assembly impairment that in turn impacts neddylation status and stability of CRL complexes . CRLs have major role in several biological processes , among which hormone transduction and signaling are well studied [17] . The fact that no changes were observed in auxin flux and homeostasis in tctp knockout embryo during development is in favor of the conclusion that TCTP/CSN4 interaction likely controls CRLs specifically involved in cell cycle control but not in auxin signaling . It was reported that perturbation of CUL neddylation do not necessarily affect the activity of all CLRs . In mice , deletion of CSN8 did not affect SCF/CRL function in general , as a subset of CRLs maintained their capacity to degrade their substrate [55] . Similarly , in Drosophila csn4 and csn5 mutants , some but not all SCF/CRLs implicated in circadian rhythm maintenance are affected [56] . Therefore , it is likely that in tctp embryos , only a specific subset of CRLs implicated in cell cycle progression is affected , resulting in arrest of development , while auxin signaling remains normal . This is also supported by the fact that conversely to tctp , auxin signaling mutants are able to complete embryo development and produce mature seeds [57] . Previously , we demonstrated that TCTP function in the regulation of cell proliferation is conserved between plants and animals [6] . Arabidopsis AtTCTP and Drosophila dTCTP share only 38% amino acids identity , but many of the essential amino acids and domains known to be required for TCTP functions are conserved between plant and animal TCTPs [3 , 6] . Previously we showed that AtTCTP and dTCTP were able to homodimerize , but also to dimerize with each other in vivo [6] , thus another argument that despite the overall relatively divergent protein sequences , their function is conserved . In support of a conservation and importance of TCTP/CSN4 interaction in animals , we show that , similar to dTCTP loss of function , dCSN4 loss of function result in a loss of photoreceptors in adult Drosophila retina accompanied by a delayed acquisition of neuronal identity , which requires a tight coordination with cell proliferation in the developing eye disc . Furthermore , we show that , like for AtTCTP , down-regulation of Drosophila dTCTP also led to a decrease of CUL1NEDD8 . Moreover , co-immunoprecipitation results show that dTCTP is able to interact with AtCSN4 , which indicates that TCTP/CSN4 interaction is likely conserved between plants and animals . In summary , our data provide evidences that TCTP functions as a key growth regulator by controlling cell proliferation together with CSN4 . We propose that TCTP could sequester CSN4 to control CUL1 neddylation status and thus CRL activity ( Fig 6 ) , and this role is conserved between plants and animals . These data add a new piece to resolve the puzzle of developmental biology processes by connecting two evolutionary conserved pathways , TCTP and COP9 , for cell proliferation . Our work will help future studies to better understand growth disorders and malignant transformation , associated with TCTP miss-expression .
T-DNA insertion knockout lines tctp-1 ( SAIL_28_C03 ) , tctp-2 ( GABI_901E08 ) as well as the RNAi-AtTCTP , 35S::AtTCTP , 35S::AtTCTP-GFP , 35S::dTCTP-GFP and the AtTCTPg-GFP lines harboring the AtTCTP genomic sequence including the promoter region , exons , introns , and the 3′ UTR region in which the GFP was inserted in frame with AtTCTP ( At3g16640 ) , have been previously described [6] . csn4 knockout line ( Salk_043720C ) was provided by NASC . The weak allele mutant of AtCSN4 was kindly provided by C . Bellini ( Umea University , Sweden ) . AtCSN4-RNAi lines: The DNA fragment corresponding to the ORF of AtCSN4 ( At5g42970 ) without start codon ATG was cloned into the vector pB7GWIWG2D ( II ) [58] under the control of the CaMV 35S constitutive promoter . The resulting construct was then used to transform A . thaliana Col-0 plants . Lines are referred to as RNAi-AtCSN4 . AtCSN4-GFP overexpressing line: The DNA fragment corresponding to the ORF of AtCSN4 was cloned into the pK7WGF2 [58] under the control of the CaMV 35S promoter . Lines are referred to as 35S::AtCSN4 . The above lines were then crossed to generate the 35S::AtTCTP/RNAi-AtCSN4 , the 35S::AtCSN4/RNAi-AtTCTP and the 35S::AtTCTP/35S::AtCSN4 lines . AtTCTPgGFP/35S::AtCSN4-Flag line: The DNA fragment corresponding to the ORF of AtCSN4 was cloned into pEarleyGate202 vector containing Flag motif [59] . The resulting construct was then used to transform Arabidopsis line AtTCTPg-GFP that harbors pTCTP::TCTPg-GFP construct [6] . Arabidopsis thaliana Columbia-0 ( Col-0 ) was used as the wild-type ( WT ) for all experiments . Embryo rescue of homozygous tctp-1 and tctp-2 embryos was performed as described previously [6] . During all steps , embryos from wild type siliques were used as control . PIN1::PIN1-GFP/TCTP+/- and DR5rev::GFP/TCTP+/- were generated by crossing tctp-2+/- with PIN1::PIN1-GFP or DR5rev::GFP , respectively . Embryo from white seeds and green seeds were then isolated [6] and observed under LSM 710 confocal microscope ( Zeiss ) . In vitro culture of Arabidopsis embryo and hormone treatments with 2 , 4D were performed as previously described [60] . All seedlings were grown in culture chambers under shorts-day condition ( 8h/16h day/night at 22°C/19°C ) for 4 weeks , then transferred to long-day condition ( 22°C , 16h/8h light/dark ) to promote flowering , with light intensity of 70 μEm-2 sec-1 . All flies were maintained on standard corn/yeast medium at 25°C . eyless>dTCTPi line: Expression of dTCTPi was carried out using the GAL4/UAS expression system [33] . We crossed eyless-GAL4 line ( Bloomington Drosophila Stock Center ) with UAS::dTCTPi [10] and analyzed the F1 adult progeny . tub>dTCTPi line: Expression of dTCTPi was carried out using the GAL4/UAS expression system . We crossed tub-GAL4 line ( Bloomington Drosophila Stock Center ) with UAS::dTCTPi [10] and analyzed the F1 larvae . We generated dCSN4 mosaic clones in the eye using Tomato/GFP-FLP/FRT method [36 , 37] . The following fly strains were used: P{neoFRT}42D P{lacW}CSN4k08018/CyO ( Kyoto Center ) , ey-FLP; FRT42D , Rh1-tdTomato[ninaC]/CyO; GMR-p35 , Rh1-GFP/TM6B and ey-FLP; FRT42D , Rh1-tdTomato 94 . 1/CyO; UAS-GFP . Living adult flies were anesthetized using CO2 and embedded in a dish containing 1% agarose covered with cold water , as described [37] and imaged using a Leica SP5 upright confocal microscope using a water immersion objective . Photoreceptors were marked by Rh1-GFP or Rh1-tdTomato . Mosaic clones were also generated using ey-FLP; FRT42D Ubi-GFP and eye discs analyzed at third instar larvae as previously described [61] . Dissected eye discs were stained with a rat anti-ELAV ( Developmental Hybridoma Bank , 1/10 ) or anti-Dcp-1 ( Cell Signaling , 1/300 ) and a rabbit anti-GFP ( Invitrogen , 1/400 ) and imaged using a LSM800 Zeiss confocal microscope . RNAi-NtTCTP and 35S::NtTCTP-GFP BY-2 cell lines have been described previously [6] . RNAi-NtCSN4 and 35S::NtCSN4-GFP BY-2 cell lines: DNA corresponding to NtCSN4 ORF was amplified using primers BY2-CSN4-F ( CACCATGGAGAGTGCGTTCGCTAGTG ) and BY2-CSN4-R ( CTAGACAGGAATAGGGAGCCCCTTCT ) and cloned into the vector pK7GWIWG2 or pK7WGF2 , respectively [58] . The resulting constructs were used to transform tobacco BY-2 ( N . tabacum L . cv . Bright Yellow-2 ) cell suspension as previously described [62] . BY-2 cells were grown in the dark at 25°C constant temperature and with agitation at 150 rpm . BY-2 cells were synchronized using aphidicolin ( Sigma ) as previously described [6] . Samples were collected every two hours and flow cytometry analyses were performed essentially as previously described [6] . Fluorescence intensity of stained nuclei was measured with MACSQuant VYB flow cytometer ( BD Bioscience ) , using 405nm excitation blue laser . DNA content analysis was performed using FlowJo , LLC software version 10 . Seeds of the relevant lines were germinated on half strength MS . Five days after germination , plantlets were transferred to EdU ( 10 μM , Sigma-Aldrich ) supplemented medium and harvested after 3h , 6h , 9h and 12h of incubation . Experiments were performed as described [63] . The percentage of EdU positive nuclei increases linearly with time , and follows an equation that can be written as P = at + b where P is the percentage of EdU positive nuclei and t is time . Total cell cycle length is estimated as 100/a , and S phase length is b/a . The parameters of the equation and their confidence intervals were estimated with the R statistics software using the least-square method . Kinematic of leaf growth was performed as previously described [24] on the two first initiated-leaves of Col-0 WT and RNAi-AtCSN4 plants grown in vitro . Leaf size as well as number and size of abaxial epidermal cells were determined starting of day 3 and until day 16 after germination . The average cell division rates were determined by calculating the slope of the Neperian Logarithmic-transformed number of cells per leaf , which was done using five-point differentiation formulas [25] . The number of newly produced cells were calculated by 72h time period . Rosette diameter was determined starting of 8 days after germination , until bolting . Rosette area was measured every 3 days with a caliper . Each measure was performed using 18 plants for each genotype . Experiments were performed on two independent transformation events and in three biological replicates . To investigate root growth , seeds were germinated on half strength MS and 5 day-old plantlets were transferred to a new plate and grown vertically for 6 days . Root length of each plant was measured using the Fiji software at day 0 , 3 and 6 after transfer . Petal area , petal cell number and size measurements were performed as previously described [64] . Briefly , petals were cleared overnight in a solution containing 86% ethanol and 14% acetic acid followed by two incubations of 4h each in ethanol 86% . Petals were dissected and photographed using Leica MZ12 stereomicroscope . Cells from cleared petals were observed with a Nikon Optiphot 2 microscope with Nomarski optics . Petal area and cell density i . e . number of cells per surface unit , were determined from digital images using ImageJ software ( U . S . National Institutes of Health ) . Co-localization experiments: cDNA fragments corresponding to the coding sequences of AtTCTP and AtCSN4 were PCR amplified and then fused to RFP and GFP , respectively . Resulting constructs were used to infiltrate leaf epidermal cells of Nicotiana benthamiana plants as previously described [65] . Bimolecular Fluorescence Complementation ( BiFC ) experiments: cDNA fragments corresponding to the coding sequences of AtTCTP and AtCSN4 were PCR amplified and then cloned into the pBiFP1 , pBiFP2 , pBiFP3 and pBiFP4 vectors [66] using the Gateway technology . Resulting constructs were used to infiltrate leaf epidermal cells of Nicotiana benthamiana plants as previously described [65] . BiFC was observed four days post-infiltration using a LSM 710 Confocal microscope ( Zeiss ) . Co-immunoprecipitation experiments using AtTCTP-GFP or AtCSN4-GFP as bait were performed using the μMACS GFP Isolation Kit ( Miltenyi Biotec ) . Three biological replicates were performed for each sample . Wild-type Col-0 and 35S::GFP plants were used as controls . Tissues of 10 days-old seedlings , mature seeds harvested from green siliques or inflorescences were ground to a fine powder in liquid nitrogen using mortar and pestle . The tissue powder ( 200 mg ) was resuspended with 1ml pre-cooled ( 4°C ) Miltenyi lysis buffer complemented with one tablet of cOmplete Mini EDTA-free Protease Inhibitor Cocktail ( Roche ) for 5 ml of lysis buffer . Cellular extracts were incubated on ice 10 min and then centrifuged 10 min at 21 000g ( 4°C ) . The supernatants were incubated with 50μl anti-GFP antibody coupled to magnetic μMACS microbeads for specific isolation of GFP-tagged protein during 1 hour at 4°C on orbital shaker . Microbeads were bound to magnetic columns and washed as described by the manufacturer , before elution of GFP-tagged proteins and bound proteins . Eluted proteins were analyzed by Western blot . For the immunoprecipitation followed by mass spectrometry ( IP/MS ) we visualized proteins by silver staining of the SDS-PAGE gel ( ProteoSilver Plus Silver Stain Kit , SIGMA ) . Tissue ( plants or cell culture ) were grinded and total proteins were extracted using « Plant Total Protein Extraction Kit » ( Sigma ) . Protein from Drosophila larvae were extracted using approximately 10 larvae in 100 μl of protein extraction buffer ( 20 mM HEPES pH:7 , 5 , 100 mM KCl , 5% Glycerol , 10 mM EDTA , 0 , 1% Tween , 1 μM DTT , 1 μM PMSF , 5 μl/ml Protease Inhibitor Cocktail ( Sigma ) , 5 μl/ml Phosphatase Inhibitor ( Sigma ) ) . After centrifugation for 10 min at 160000 g , proteins contained in the supernatant were dosed using Bradford method[67] . Proteins were analyzed by Western blot using antibodies directed against AtTCTP ( 1/500 dilution ) [6] , AtCUL1 ( 1/2000 dilution; Enzo LifeScience ) , dCUL1 ( 1/500 dilution; Thermo Scientific ) , AtCSN4 ( 1/2000 dilution; Enzo LifeScience ) , Flag ( 1/1000 dilution; Sigma-Aldrich ) ; α-Tubulin ( 1/1000 dilution; Sigma ) or GFP ( 1/2000 dilution; Roche ) . IRDye 800CW and IRDye 680RD ( 1/10 000 dilution; LI-COR ) were used as secondary antibodies and the signal was revealed using Odyssey Clx imaging system and signal intensity was quantified using the Image Studio Lite software ( LI-COR ) . HRP conjugated anti-mouse or anti-rabbit IgG were used as secondary antibodies ( 1/5000 dilution ) and the signal was revealed using Clarity Western ECL substrate and the ChemiDocTouch imaging system ( Biorad ) . Intensity of the bands was quantified using ImageJ software ( U . S . National Institutes of Health ) . | During organism development , the correct implementation of organs with unique shape , size and function , is the result of coordinated cellular processes , such as cell proliferation and expansion . Deregulation of these processes affect human health and can lead to severe diseases . While plants and animals have largely diverged in several aspects , some biological functions , such as cell proliferation , are conserved between these kingdoms . Previously we reported that the Translationally Controlled Tumor Protein ( TCTP ) , a highly-conserved protein among all eukaryotes , positively regulates cell proliferation and this role is conserved between plants and animals . In agreement with these data , animals TCTP was reported to highly accumulate in tumor cells , and thus represents a target for cancer research and therapies . To discover how TCTP regulates cell proliferation , we conducted studies to identify factors acting in the TCTP pathway . Using the model plant Arabidopsis , we identified that TCTP fulfil its role by interacting with CSN4 , a subunit of the conserved COP9 complex . TCTP interferes with the role of COP9 to regulate the downstream complex CRL known to control cell proliferation in eukaryotes . We further demonstrate that this role is conserved in the fly Drosophila , thus corroborating the conservation of TCTP pathway between plants and animals . We believe that , the data here will provide exciting perspectives , beyond plant research , that will help understand developmental disorders associated with TCTP misfunction , such as cancer . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"plant",
"anatomy",
"medicine",
"and",
"health",
"sciences",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"brassica",
"animals",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"plant",
"science",
"experimental",
"organism",
"systems",
"embryos",
"eyes",
"plants",
"flowering",
"plants",
"drosophila",
"research",
"and",
"analysis",
"methods",
"arabidopsis",
"thaliana",
"embryology",
"cell",
"proliferation",
"animal",
"studies",
"head",
"leaves",
"insects",
"arthropoda",
"eukaryota",
"plant",
"and",
"algal",
"models",
"cell",
"biology",
"anatomy",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"organisms"
] | 2019 | TCTP and CSN4 control cell cycle progression and development by regulating CULLIN1 neddylation in plants and animals |
The oxidative stress theory of aging postulates that aging results from the accumulation of molecular damage caused by reactive oxygen species ( ROS ) generated during normal metabolism . Superoxide dismutases ( SODs ) counteract this process by detoxifying superoxide . It has previously been shown that elimination of either cytoplasmic or mitochondrial SOD in yeast , flies , and mice results in decreased lifespan . In this experiment , we examine the effect of eliminating each of the five individual sod genes present in Caenorhabditis elegans . In contrast to what is observed in other model organisms , none of the sod deletion mutants shows decreased lifespan compared to wild-type worms , despite a clear increase in sensitivity to paraquat- and juglone-induced oxidative stress . In fact , even mutants lacking combinations of two or three sod genes survive at least as long as wild-type worms . Examination of gene expression in these mutants reveals mild compensatory up-regulation of other sod genes . Interestingly , we find that sod-2 mutants are long-lived despite a significant increase in oxidatively damaged proteins . Testing the effect of sod-2 deletion on known pathways of lifespan extension reveals a clear interaction with genes that affect mitochondrial function: sod-2 deletion markedly increases lifespan in clk-1 worms while clearly decreasing the lifespan of isp-1 worms . Combined with the mitochondrial localization of SOD-2 and the fact that sod-2 mutant worms exhibit phenotypes that are characteristic of long-lived mitochondrial mutants—including slow development , low brood size , and slow defecation—this suggests that deletion of sod-2 extends lifespan through a similar mechanism . This conclusion is supported by our demonstration of decreased oxygen consumption in sod-2 mutant worms . Overall , we show that increased oxidative stress caused by deletion of sod genes does not result in decreased lifespan in C . elegans and that deletion of sod-2 extends worm lifespan by altering mitochondrial function .
The oxidative stress theory of aging proposes that reactive oxygen species ( ROS ) generated by normal metabolism cause damage to macromolecules within the cell and that the accumulation of this damage over time leads to cellular dysfunction and eventually organismal death [1]–[3] . The majority of ROS present in the cell is thought to be generated in the mitochondria . In order to counteract this process , cells have a number of defense mechanisms which serve to detoxify ROS . Superoxide dismutase ( SOD ) is a detoxification enzyme that converts superoxide to hydrogen peroxide , which can subsequently be converted to water [4] . The oxidative stress theory of aging predicts that loss of SOD activity should result in increased sensitivity to oxidative stress , since the organism would be less able to detoxify ROS . This should , in turn , result in a shortened lifespan . This is essentially what is observed in yeast , flies and mice for both cytoplasmic SOD ( SOD1 , CuZnSOD ) and mitochondrial SOD ( SOD2 , MnSOD ) . In yeast , knocking out sod1 has been shown to decrease clonal and replicative lifespan [5] , [6] and accelerate chronological aging [7] , [8] . In flies , knocking out Sod1 decreases lifespan [9] . In mice , targeted inactivation of Sod1 results in high oxidative stress and a 30% decrease in lifespan [10] . For SOD2 , yeast knockouts show decreased chronological and replicative lifespan [6]–[8] . Reduction of Sod2 in flies by either RNA interference ( RNAi ) or genetic deletion results in marked reductions in lifespan [11] , [12] . In mice , Sod2 knockouts exhibit high degrees of oxidative stress and neonatal or perinatal lethality [13] , [14] . In contrast , loss of extracellular SOD ( SOD3 , EC-SOD ) does not appear to impact lifespan despite an increased sensitivity to hyperoxia [15] . Thus , in support of the oxidative stress theory , the effect of deleting Sod1 or Sod2 in all three model species is increased oxidative stress and decreased lifespan or early lethality in the case of Sod2 mice . In contrast , lifespan in C . elegans may be relatively unaffected by decreased sod expression . Using an RNAi approach to knockdown either sod-1 or sod-2 , Yang et al . showed a mild decrease in lifespan with sod-1 RNAi but no effect of sod-2 RNAi , despite the fact that both knockdowns resulted in increased sensitivity to paraquat and an increase in oxidatively damaged proteins [16] . However , it is possible that if the RNAi did not completely abolish sod expression then the remaining low level of SOD activity is sufficient for normal lifespan . Here , we examine the effect of eliminating SOD on lifespan and sensitivity to oxidative stress in C . elegans and thereby test the oxidative stress theory of aging . Whereas most organisms have only three SODs ( one cytoplasmic , one mitochondrial and one extracellular ) , C . elegans has five sod genes [17] . sod-1 , sod-2 and sod-4 encode the primary cytoplasmic , mitochondrial and extracellular SODs respectively [18]–[22] ( equivalent to Sod1 , Sod2 and Sod3 in mice ) . In addition , sod-3 is expressed in the mitochondrial matrix and sod-5 is expressed in the cytoplasm , thereby providing C . elegans with two cytoplasmic and two mitochondrial SODs [18] , [23] . By examining C . elegans mutants with deletions in each of the five sod genes , we find that elimination of individual sod genes can increase sensitivity to oxidative stress but does not decrease lifespan . Furthermore , we find that sod-2 mutant worms are long-lived and propose that their lifespan extension is due to an alteration of mitochondrial function .
The oxidative stress theory of aging predicts that increasing oxidative stress should result in decreased lifespan . To test this hypothesis , we assessed the lifespan of worms lacking each of the five individual sod genes in C . elegans ( the location and size of the mutation for each allele tested are shown in Figure S1 ) . For sod-1 , sod-2 and sod-5 we assessed two independent alleles . We found that lifespan was not affected by the disruption of sod-1 , sod-3 , sod-4 or sod-5 ( Figure 1A , C–E; all lifespan data is included in Table S1 ) . This is particularly surprising in the case of sod-1 since SOD-1 accounts for the majority of SOD activity in the cell [23] . In addition we found that deletion of sod-2 resulted in a significant increase in lifespan ( Figure 1B ) . This is also a surprising result given that SOD-2 is the primary SOD present in the mitochondrial matrix and the mitochondria is a major site of superoxide production in the cell . To ensure that the lifespan extension in sod-2 mutant worms resulted from the deletion of the sod-2 gene we generated heteroallelic mutants . This was accomplished by crossing sod-2 ( gk257 ) males with sod-2 ( ok1030 ) hermaphrodites and following lifespan in the male offspring ( since these must be cross progeny ) and by crossing sod-2 ( gk257 ) males with either dpy-17 ( control ) or sod-2 ( ok1030 ) ;dpy-17 hermaphrodites and following the lifespan of the resulting non-dumpy hermaphrodite offspring . In both cases , we found that heteroallelic sod-2 ( gk257 ) /sod-2 ( ok1030 ) mutant worms lived significantly longer than their corresponding controls ( Figure S2 ) . Since the loss of individual SODs failed to decrease lifespan , we next sought to determine whether the deletion of individual sod genes had an impact on oxidative stress . As it is currently not possible to accurately quantify the levels of ROS in worms , we used paraquat and juglone to assess sensitivity to oxidative stress as has been described in previous experiments [24]–[26] . Both of these compounds are reduced upon entry into the cell and are thought to induce oxidative stress by generating superoxide from oxygen during their subsequent reoxidation [27] , [28] . To assess paraquat sensitivity , we examined the survival of 7 day old adult worms on plates containing 4 mM paraquat . We found that sod-1 mutant worms were very sensitive to paraquat with all of the worms dying within one or two days ( Figure 1F ) . We also found that sod-2 and sod-3 mutant worms were more sensitive to paraquat than wild-type worms although not as sensitive as sod-1 mutant worms ( Figure 1G , H ) . In contrast , sod-4 and sod-5 mutant worms showed similar survival to wild-type worms ( Figure 1I , K ) In order to confirm our observation of increased sensitivity to oxidative stress , we assessed sensitivity to juglone . One day old worms were transferred to plates containing 240 µM juglone and survival was monitored for the following 10 hours . As with the paraquat assay , both sod-1 deletion strains showed markedly increased sensitivity to oxidative stress as no worms survived to the 4 hour time point ( Figure 1L ) . The sod-2 deletion strains also showed increased sensitivity to juglone which was not as severe as the sod-1 mutants ( Figure 1M ) . In contrast , deletion of sod-3 , sod-4 or sod-5 did not make worms significantly more sensitive to juglone-induced oxidative stress ( Figure 1N–P ) . Next , we assessed sensitivity to paraquat during development by exposing eggs to plates containing 0 . 2 mM paraquat and determining the latest developmental stage attained for each strain . While exposure to paraquat slowed development in all strains , including wild-type N2 worms , we found that all of the sod deletion mutants except for sod-2 were able to develop to adulthood ( Figure S3 ) . The sod-2 mutants were found to arrest at the L1 stage . Thus , sod-2 mutant worms are the most sensitive of all of the sod deletion mutants to oxidative stress during development while sod-1 mutant worms are the most sensitive in adulthood . In order to confirm the absence of individual sod expression in the deletion strains , we assessed the levels of each of the five sod mRNAs by qRT-PCR ( quantitative real time reverse transcription polymerase chain reaction ) . Importantly , we also sought to determine whether the deletion of individual sod genes is compensated for by the upregulation of other sod genes . In all strains , the deletion mutation resulted in decreased levels of the corresponding mRNA ( in most cases no mRNA was detected ) ( Figure 2A ) . In both sod-2 mutant strains , sod-1 , sod-3 and sod-4 mRNAs were significantly elevated and there was a trend towards increased expression of sod-5 ( Figure 2A ) . Similarly , sod-3 mutant worms showed increased expression of sod-1 and sod-4 mRNA and a trend towards increased expression of sod-2 and sod-5 ( Figure 2A ) . One of two sod-5 mutants showed significantly increased expression of sod-1 mRNA , while sod-1 and sod-4 mutant worms showed no significant changes in mRNA levels of the other four sod mRNAs ( Figure 2A ) . Overall , we observed some compensatory upregulation of other sod mRNAs in the sod deletion strains but the degree of upregulation was small , generally 2-fold or less . The fact that sod-3 mutant worms showed a similar upregulation of sod mRNA as sod-2 mutant worms suggests that the lifespan extension in the sod-2 mutants does not result from the observed upregulation of sod mRNA . Since a compensatory increase in SOD expression could also occur at the translational level , we examined the level of SOD-1 and SOD-2 protein in the sod deletion mutants ( antibodies to the other SOD proteins are currently not available ) . We observed no SOD-1 protein in sod-1 deletion mutants or SOD-2 protein in sod-2 deletion mutants ( Figure 2B ) . As with mRNA expression we did not observe a dramatic upregulation of SOD-1 or SOD-2 in any of the sod deletion mutants ( Figure 2C ) . Although the magnitude of compensatory upregulation of other sod genes , when present , was small , it is possible that this could have accounted for the normal or extended lifespan we observed in the sod single deletion mutants . To investigate this possibility , we sought to determine whether elimination of a second sod gene would shorten the lifespan of the sod single deletion mutants . Accordingly , we generated a panel a sod-sod double mutants consisting of all of the double mutants for sod-1 and sod-2 , since these are the major contributors to SOD activity in the cytosol and mitochondria respectively , and sod-3; sod-5 ( this mutant lacks both of the “extra” sod genes found in C . elegans ) . Examining the lifespan of sod-1 double deletion mutants revealed that deletion of sod-3 , sod-4 or sod-5 did not shorten the lifespan of sod-1 mutant worms ( Figure 3A ) . In contrast , sod-1;sod-2 mutant worms lived significantly longer than wild-type N2 worms ( Figure 3A ) . Among the sod-2 double deletion mutants , all of the worms maintained the extended lifespan seen in sod-2 single deletion mutants indicating that in no case is the upregulation of another sod gene entirely responsible for the long life observed in sod-2 mutant worms ( Figure 3B ) . Finally , we found that sod-3;sod-5 mutant worms had a similar lifespan to wild-type worms ( not shown ) . Next , we examined the sensitivity to oxidative stress among the sod-sod double mutants using both paraquat and juglone . Examining the survival of one day old adult worms on 4 mM paraquat plates , we found that all of the sod-1 double mutants , including the long-lived sod-1;sod-2 mutant worms , had decreased survival compared to N2 worms ( Figure 3C ) . Among the sod-2 double mutants , sod-2;sod-3 mutant worms were hypersensitive to paraquat , while sod-2;sod-4 and sod-2;sod-5 mutant worms appeared to be only mildly more sensitive than N2 worms ( Figure 3C ) . A similar pattern of sensitivity to oxidative stress was observed on juglone plates . All of the sod-1 double mutants as well as sod-2;sod-3 mutant worms were more sensitive to juglone than N2 worms ( Figure 3D ) . There was also a trend towards decreased survival in the remaining double mutant strains ( Figure 3D ) . Overall , the sod-sod double mutants showed increased sensitivity to oxidative stress but normal or extended longevity . Thus , we did not observe any correlation between sensitivity to oxidative stress and lifespan . We also examined sod mRNA expression levels in the sod-sod double mutant worms ( Figure S4 ) . As with the sod single deletion mutants , we observed some compensatory upregulation of other sod genes but the magnitude of this increase was small and failed to rescue the observed increase in sensitivity to oxidative stress . Based on our finding that even the elimination of two sod genes together does not shorten the lifespan of C . elegans , we assayed lifespan in a selection of sod triple mutants . To eliminate the possibility that the reason why sod-1 and sod-2 mutants of C . elegans do not show decreased lifespan is because C . elegans has duplicate SODs in the cytoplasm and mitochondria , we generated sod-1;sod-3;sod-5 and sod-2;sod-3;sod-5 triple mutants to model sod-1 and sod-2 knockouts in species with only three sod genes . We also generated sod-1;sod-2;sod-4 worms which lack the primary cytoplasmic , mitochondrial and extracellular sod genes . Examination of worm lifespan revealed that sod-1;sod-3;sod-5 mutant worms live as long as wild-type worms while both sod-2;sod-3;sod-5 and sod-1;sod-2;sod-4 mutant worms live significantly longer than wild-type ( Figure 4A–C ) . This clearly indicates that the normal lifespan observed in sod-1 worms does not result from the overlapping expression of sod-3 in the mitochondria or sod-5 in the cytoplasm . Since sod-2;sod-3;sod-5 triple mutant worms do not survive as long as sod-2 single deletion mutants , it is possible that the mild upregulation of sod-3 and sod-5 may contribute to the increased lifespan of sod-2 mutant worms . However , the fact that similar upregulation of sod mRNA in sod-3 mutant worms does not result in extension of lifespan and that upregulation of sod-3 and sod-5 in sod-2 mutant worms is insufficient to prevent increased levels of oxidative damage ( see below ) suggests that other mechanisms are involved in the long life of sod-2 mutant worms . In order to investigate possible mechanisms of lifespan extension in sod-2 mutant worms , we generated double mutants with genes known to extend lifespan which are representative of different lifespan extending mechanisms including daf-2 ( insulin/IGF signaling ) [29] , clk-1 ( decreased mitochondrial function ) [30] , isp-1 ( decreased mitochondrial function ) [25] , eat-2 ( dietary restriction ) [31] and glp-1 ( germ-line ablation ) [32] . Deletion of sod-2 did not extend the lifespan of daf-2 worms ( Figure 5A ) . clk-1 worms showed a marked extension of lifespan when sod-2 was deleted ( Figure 5B ) . In contrast , sod-2 deletion greatly shortened the lifespan of isp-1 worms such that isp-1;sod-2 worms had a shorter lifespan than wild-type N2 worms ( Figure 5C ) . eat-2;sod-2 mutant worms showed a small increase in lifespan compared to eat-2 worms ( Figure 5D ) . Finally , deletion of sod-2 in glp-1 worms resulted in a modest increase of mean but not maximum lifespan ( Figure 5E ) . Overall , we found that sod-2 deletion had the greatest impact on the lifespan of mutants which exhibit extended longevity as a result of alterations in mitochondrial function . Based on our finding that sod-2 deletion interacts with long-lived mutants with altered mitochondrial function and the fact that SOD-2 is localized to the mitochondria , we hypothesized that deletion of sod-2 extends lifespan by decreasing mitochondrial function . In C . elegans a number of genes have been identified that affect mitochondrial function and at the same time increase lifespan [25] , [30] , [33] , [34] . Although these genes do not necessarily interact and the exact mechanism of lifespan extension is unclear , these mutants are generally grouped together since it is believed that the alteration of mitochondrial function is the key to their long life . In addition to impaired mitochondrial function and extended longevity , these mutants , sometimes referred to as Mit mutants , are characterized by slow development , slow defecation rate and decreased brood size . Accordingly , we quantified the development , brood size and defecation rate of sod-2 mutant worms and compared this with two prototypes of this class of mutants - clk-1 and isp-1 worms [25] , [30] , [35] . We also examined clk-1;sod-2 and isp-1;sod-2 double mutants to determine if the loss of sod-2 enhanced the phenotypes observed in clk-1 and isp-1 worms . Examination of post-embryonic development ( PED ) revealed that sod-2 , clk-1 and isp-1 worms all developed slower than wild-type worms ( Figure 6A ) . On a clk-1 and isp-1 background , sod-2 deletion resulted in further increase in PED time ( Figure 6A ) . Examination of defecation cycle length revealed a slow rate of defecation in sod-2 , clk-1 and isp-1 worms compared to wild-type N2 worms ( Figure 6B ) . Deletion of sod-2 had opposite effects on defecation cycle length in clk-1 and isp-1 worms . sod-2 deletion further lengthened the defecation cycle of clk-1 worms but shortened the defecation cycle length of isp-1 worms ( Figure 6B ) . Self-brood size was decreased in sod-2 , clk-1 and isp-1 worms and deletion of sod-2 further decreased the brood size in clk-1 and isp-1 worms ( Figure 6C ) . Finally , a comparison of lifespan between these strains revealed that isp-1;sod-2 worms were short-lived , sod-2 and clk-1 worms were long-lived and clk-1;sod-2 and isp-1 worms were very long-lived ( Figure 6D ) . Clearly , sod-2 deletion mutants exhibit the key characteristics of extended longevity mitochondrial mutants and modulate these phenotypes in clk-1 and isp-1 worms . The phenotypic similarity of sod-2 mutant worms to extended longevity mitochondrial mutants as well as the ability of sod-2 deletion to alter these characteristic phenotypes in clk-1 and isp-1 worms suggests that sod-2 extends lifespan through a similar mechanism . Based on this hypothesis , we would predict that mitochondrial function would be altered in sod-2 mutant worms . To assess this , we measured whole worm oxygen consumption , which has previously been shown to be decreased in both clk-1 and isp-1 worms [16] , [25] . We found that oxygen consumption in one day old adult worms was significantly decreased in sod-2 mutant worms compared to wild-type worms ( Figure 7A ) . We have previously reported that both clk-1 and isp-1 worms exhibit decreased levels of oxidatively damaged proteins [16] . Here , we find that sod-2 worms are hypersensitive to oxidative stress , suggesting that there may be an increase in oxidatively damaged proteins in these worms . To determine the level of oxidative damage in sod-2 mutant worms , we quantified carbonylated proteins in sod-2 mutant worms and wild-type worms . We found that sod-2 mutant worms had significantly more oxidative damage than wild-type worms ( Figure 7B ) . In order to determine whether the sensitivity of sod-2 mutant worms to paraquat and juglone was the result of a specific sensitivity to oxidative stress or a sign of general weakness , we examined the ability of sod-2 worms to withstand heat stress [36] or osmotic stress [37] . After exposure to 35 degree Celsius heat stress for a period of 6 hours or 9 hours , we found that sod-2 mutant worms survived as well as wild-type worms ( Figure 7C ) . Similarly , exposing sod-2 mutant worms to osmotic stress on 500 mM NaCl NGM plates for 20 hours revealed that sod-2 mutant worms survive osmotic stress as well as wild-type N2 worms ( Figure 7D ) . Combined these results suggest that the sensitivity of sod-2 mutant worms to oxidative stress is a specific sensitivity resulting from their decreased ability to detoxifying ROS .
Since its origins in 1956 , the oxidative stress theory of aging has been extensively tested in multiple organisms both by observing variations in natural populations and through genetic intervention [1]–[3] . Thus far there have been many experiments that support this theory , but also experiments which challenge the notion that molecular damage from ROS leads to aging ( reviewed in [38] ) . In this paper , we find that the effect of sod deletion on lifespan in C . elegans is unique from other organisms . In concordance with the oxidative stress theory of aging , yeast , flies and mice lacking either cytoplasmic or mitochondrial SOD show either decreased lifespan or lethality ( in the case of Sod2 knockout mice ) [5] , [7] , [9] , [10] , [12]–[14] while mice lacking extracellular SOD live as long as wild-type mice [15] . Here , we demonstrate that none of the sod deletion mutants in C . elegans show decreased lifespan . One possible explanation for this discrepancy is the fact that C . elegans has five sod genes , rather than three , including two cytoplasmic SODs and two mitochondrial SODs . To eliminate this explanation , we show that the lifespan of sod-1;sod-3;sod-5 and sod-2;sod-3;sod-5 triple mutants , which model sod-1 and sod-2 deficient organisms in species with only three sod genes , is not decreased . Another possible explanation for why C . elegans sod mutants exhibit a normal lifespan would be compensatory upregulation of other sod genes . In support of this hypothesis , we observed sod mRNA upregulation in sod-2 and sod-3 mutant worms as well as one of two sod-5 mutants . However , the magnitude of this upregulation was small ( 2-fold or less ) and we observed no significant sod upregulation in sod-1 or sod-4 mutant worms which also exhibit a normal lifespan . These results are in general agreement with studies of Sod knockouts in flies and mice where either no change in other SOD activity is reported [12] , [15] , [39]–[41] or changes with magnitudes of less than 50% [10] , [14] , [42] . Although the compensatory upregulation of other sod genes was small in magnitude and not present in all sod deletion mutants , it is possible that this small increase contributed to the normal or extended lifespans observed in these strains . It is also possible that changes in SOD protein levels or activity contributed to the preservation of lifespan in these strains . To investigate these possibilities , we used the genetic approach of generating sod double and triple mutants . The loss of an additional sod gene did not decrease the lifespan in sod-1 mutant worms , nor did it revert the lifespan of sod-2 mutant worms to wild-type . This indicates that the increased expression or activity of any single other sod gene is not responsible for the normal lifespan observed in sod-1 mutant worms or extended lifespan observed in sod-2 mutant worms . Similarly , all of the sod triple mutants were able to live at least as long as wild-type worms . The lack of lifespan shortening in worms with multiple sod genes deleted is in concordance with studies in mice where the loss of extracellular SOD [43] or the loss of glutathione peroxidase ( another ROS detoxifying enzyme ) and one copy of Sod2 [38] does not further decrease lifespan in Sod1 knockout mice . The lack of additive effects between different compartments can be explained by the inability of superoxide to cross biological membranes [44] , [45] . A compartment specific effect of genes involved in ROS detoxification on lifespan has also been observed in C . elegans with genes encoding catalase , where deletion of peroxisomal catalase ( ctl-2 ) results in decreased lifespan while deletion of cytoplasmic catalase ( ctl-1 ) has no effect on lifespan [46] . A comparison of our results for lifespan and sensitivity to oxidative stress reveals no correlation . None of the sod single or double mutants exhibited a shortened lifespan despite many strains showing markedly increased sensitivity to oxidative stress . Most strikingly , sod-1;sod-2 mutants show the highest sensitivity to oxidative stress in combination with the longest lifespan . Similarly , our laboratory has recently shown that decreasing levels of sod-1 or sod-2 by RNAi increases paraquat sensitivity and oxidative damage to proteins in N2 as well as in multiple long-lived strains ( daf-2 , clk-1 and isp-1 ) yet does not decrease lifespan in these strains [16] . Initial experiments examining sensitivity to oxidative stress in long-lived worms indicated that increased resistance to oxidative stress occurs with increased lifespan [24] , [47] , [48] . It was also found that when longevity was selected for in flies , increased longevity was accompanied by resistance to oxidative stress [49] . More recently , in the reverse experiment examining the lifespan of worms that were resistant to paraquat-induced oxidative stress , it was found that only 84 of 608 RNAi treatments that increased stress resistance also increased lifespan [50] . Similarly , examination of the relationship between paraquat resistance and lifespan in 138 lines of flies revealed only a weak positive correlation [51] . In mice , Sod1 knockouts show increased sensitivity to oxidative stress and decreased lifespan [10] , [39] , mice heterozygous for the targeted inactivation of Sod2 showed a normal lifespan despite increased oxidative damage [52] , while Sod3 knockout mice show increased sensitivity to oxidative stress and a normal lifespan [15] . Combined with our results , it appears that the correlation between sensitivity to oxidative stress and lifespan is weak at best . SOD2 is the primary , and normally sole , SOD present in the mitochondrial matrix . Since the mitochondria is a major source of superoxide within the cell and superoxide is not able to pass through membranes[44] , [45] , SOD2 may be the most critical SOD within the cell for decreasing superoxide-induced damage . This conclusion is supported by findings that decreasing or eliminating SOD2 expression affects lifespan more than elimination of SOD1 or the extracellular SOD . In flies , eliminating SOD1 reduces lifespan from about 60 days to 11 . 8 days [9] while eliminating SOD2 decreases lifespan to less than 1 day [12] . Similarly , Sod1 knockout mice show a 30% decrease in lifespan living an average of 20 . 8 months [10] , while Sod2 knockout mice exhibit either peritnatal or neonatal lethality [13] , [14] . In contrast to what is observed in other species , we find that sod-2 deletion in C . elegans results in extended lifespan . While these worms show small but significant increases in sod-1 , sod-3 and sod-4 mRNA expression , deleting sod-1 , sod-3 or sod-4 in sod-2 mutant worms does not revert their lifespan to wild-type suggesting that this upregulation of sod expression is not responsible for the lifespan increase in sod-2 mutant worms . Our observation of decreased lifespan in sod-2;sod-3;sod-5 mutant worms compared to sod-2 mutant worms suggests the possibility that upregulation of sod-3 and sod-5 partially contributes to the extended lifespan observed in sod-2 mutant worms . However , the fact that we observe similar upregulation of other sod genes in sod-3 mutant worms without the lifespan extension supports the conclusion that the mild compensatory upregulation of other sod genes is not responsible for the long life of sod-2 mutants . Furthermore , the fact that sod-2 mutant worms show increased oxidative damage indicates that the upregulation of sod-3 and sod-5 is not sufficient to reduce mitochondrial oxidative stress in sod-2 mutant worms . To gain insight into the mechanism of lifespan extension in sod-2 mutant worms , we examined the effect of sod-2 deletion on other mutants with extended longevity . sod-2 deletion did not extend lifespan in daf-2 worms , which extend lifespan through the insulin-IGF1 pathway [29] but did result in a modest extension of lifespan in eat-2 worms , which extend lifespan through caloric restriction [31] and glp-1 worms , which extend lifespan through the germ-line ablation [32] . In clk-1 worms , which extend lifespan by decreasing mitochondrial function [30] , [53] , deletion of sod-2 resulted in a 15 day increase in mean lifespan . In contrast , sod-2 deletion decreased the lifespan of isp-1 worms by 25 days despite the fact that isp-1 also extends lifespan through via a decrease in mitochondrial function [25] . The clear interaction of sod-2 deletion with mutants that extend lifespan through alterations in mitochondrial function suggested the possibility that sod-2 also increases longevity through a similar mechanism . In C . elegans a number of mutants have been identified by genetic deletion or RNAi that affect mitochondrial function and extend lifespan [25] , [30] , [33] , [34] , [54] . In addition to decreased mitochondrial function and extended lifespan , the group of mitochondrial mutants also share a number of characteristic phenotypes such as slow rate of development , slow rate of defecation and decreased brood size [25] , [35] , [55] . Phenotypic characterization of sod-2 mutant worms demonstrates that sod-2 mutants exhibit all of the phenotypes of the extended longevity mitochondrial mutants including slow development , slow defecation rate , decreased brood size , decreased mitochondrial function and increased lifespan . While we have previously shown that clk-1 and isp-1 worms have decreased levels of oxidatively damaged proteins [16] , here we find that sod-2 mutant worms exhibit an increase in oxidative damage . The fact that all three strains have a long lifespan suggests that both high and low levels of oxidative damage are compatible with long life . Moreover , the fact that oxidative damage in clk-1 and isp-1 worms can be increased to a level that is significantly greater than wild-type worms without diminishing the long life of these two strains suggests that the low levels of oxidative damage in clk-1 and isp-1 worms does not contribute to their extended longevity [16] . In addition to those genes which impair mitochondrial function and increase lifespan , there are at least two mutations , mev-1 [56] and gas-1 [57] , which decrease mitochondrial function and decrease lifespan . While it is currently uncertain why these mutations have a different effect on lifespan compared to the extended longevity mitochondrial mutants , it appears that there are at least two ways in which decreasing mitochondrial function can lead to decreased lifespan . First , the severity of the mutation can be incompatible with long life . This has recently been demonstrated using an RNAi dilution series against genes involved in mitochondrial function [55] . These authors find that RNAi against the same gene can increase lifespan at low concentration ( i . e . mildly inhibited mitochondrial function ) and decrease lifespan at high concentration ( i . e . severely inhibited mitochondrial function ) . In our work , we hypothesize that isp-1;sod-2 worms are another example whereby the overall mitochondrial function in the double mutant worm is severely affected leading to a shortened lifespan . Second , the decreased lifespan can be the result of the way in which mitochondrial function is altered . For example , RNAi targeted against any of the four subunits of electron transport chain complex II results in decreased lifespan [58] while RNAi targeted against proteins in any other complex of the electron transport chain can result in increased lifespan [33] . Furthermore , recent work examining the effect of an RNAi dilution series against mev-1 indicates that it is not the severity of this mutation that prevents it from extending lifespan , since mev-1 RNAi failed to increase the lifespan of wild-type worms at any concentration [55] . Examination of clk-1;sod-2 double mutants shows that sod-2 deletion enhances all of the mitochondrial mutant phenotypes of clk-1 worms . However , a different pattern is observed with isp-1 worms , where sod-2 deletion further slows development and decreases brood size but quickens defecation towards wild-type and decreases lifespan below wild-type N2 worms . We propose that the reason for the different effects of sod-2 deletion on clk-1 and isp-1 worms results from differences in the initial degree of mitochondrial function . Based on previous measurements of oxygen consumption , respiration is only mildly impaired in clk-1 worms [16] , [53] while it is more than 50% reduced in isp-1 worms [25] . By comparing the other phenotypes of N2 , clk-1 and isp-1 worms it can be seen that as mitochondrial function decreases , development time gets longer , defecation gets slower , self brood size decreases and lifespan increases . However , according to mitochondrial threshold theory , once a certain threshold of mitochondrial dysfunction is reached , the cell is no longer able to compensate and lifespan decreases [59] . This theory was recently explored in C . elegans through the use of an RNAi dilution series to show that progressively decreasing mitochondrial function resulted in increased lifespan only until a certain threshold after which lifespan began to decrease [55] . Based on these findings , we propose a model in which the shortened lifespan that we observe in isp-1;sod-2 worms results from the sod-2 deletion pushing mitochondrial function past the threshold at which the organism is able to compensate for the lost mitochondrial function and accordingly lifespan is decreased ( Figure 8 ) . Similarly , we propose that the increased lifespan in clk-1;sod-2 worms results from the sod-2 deletion reducing the mitochondrial function to a level similar to isp-1 worms . In line with our demonstration of decreased oxygen consumption in sod-2 mutant worms , deletion of sod-2 has also been shown to decrease mitochondrial function in mouse models [60]–[62] . Although the first of the extended longevity mitochondrial mutants was identified more than a decade ago [30] , [35] , [54] , the precise mechanism by which these mutants extend lifespan is still unresolved . Nonetheless , a number of potential mechanisms have been suggested [63] . Future studies will need to more precisely define how mitochondrial mutants , such as sod-2 , extend lifespan and to determine how C . elegans is able to cope with reduced SOD activity . It will be particularly interesting to examine the interaction between SODs and other proteins involved in ROS detoxification ( catalases , peroxidases , thioredoxins , peroxiredoxins ) in order to obtain a more complete understanding of the relationship between oxidative stress and lifespan .
The following strains were used in these experiments: N2 ( wild-type ) , sod-1 ( tm776 ) , sod-1 ( tm783 ) , sod-2 ( gk257 ) , sod-2 ( ok1030 ) , sod-3 ( tm760 ) , sod-4 ( gk101 ) , sod-5 ( tm1146 ) , sod-5 ( tm1246 ) , clk-1 ( qm30 ) , eat-2 ( ad1116 ) , daf-2 ( e1370 ) , isp-1 ( qm150 ) , glp-1 ( e2141 ) . Strains obtained from external sources were outcrossed with our N2 worms for 5-10 generations . For these experiments the following double and triple mutant strains were generated: sod-1 ( tm783 ) ;sod-2 ( ok1030 ) , sod-1 ( tm783 ) ;sod-3 ( tm760 ) , sod-1 ( tm783 ) ;sod-4 ( gk101 ) , sod-1 ( tm783 ) ;sod-5 ( tm1246 ) , sod-2 ( ok1030 ) ;sod-3 ( tm760 ) , sod-2 ( ok1030 ) ;sod-4 ( gk101 ) , sod-2 ( ok1030 ) ;sod-5 ( tm1246 ) , sod-3 ( tm760 ) ;sod-5 ( tm1246 ) , clk-1 ( qm30 ) ;sod-2 ( ok1030 ) , eat-2 ( ad1116 ) ;sod-2 ( ok1030 ) , daf-2 ( e1370 ) ;sod-2 ( ok1030 ) , isp-1 ( qm150 ) ;sod-2 ( ok1030 ) , glp-1 ( e2141 ) ;sod-2 ( ok1030 ) , sod-1 ( tm783 ) ;sod-2 ( ok1030 ) ;sod-4 ( gk101 ) , sod-1 ( tm783 ) ;sod-3 ( tm760 ) ;sod-5 ( tm1246 ) , and sod-2 ( ok1030 ) ;sod-3 ( tm760 ) ;sod-5 ( tm1246 ) . All of the sod deletions were confirmed by PCR . All strains were maintained at 20°C . Lifespan studies were completed at 20°C with a minimum of 3 independent trials and an initial number of 80 worms per strain per trial . Initial lifespan assays for sod single deletion mutants , sod-sod double deletion mutants and sod-2 double mutants with genes in known pathways of lifespan extension were completed on normal NGM plates . As some sod double mutant strains bagged extensively subsequent lifespan studies were completed on plates containing 100 µM FUDR ( Sigma ) . Results obtained on NGM plates were all repeated and confirmed on FUDR plates . Survival plots shown represent pooled data from multiple trials on FUDR plates . For glp-1 and glp-1;sod-2 lifespan analyses worms were grown at 25°C and then transferred to 20°C at adulthood . Paraquat and juglone sensitivity assays were completed in triplicate with 30–40 worms per strain per trial at 20°C . To assay paraquat sensitivity , 7 day old adult worms were transferred to plates containing 4 mM paraquat ( Sigma ) and survival was monitored daily . Initially , paraquat assays were performed on 1 day old adult worms . However , by day 3 of adulthood , paraquat causes most of the worms to have internal hatching of progeny ( bagging ) such that more worms die of this than of paraquat toxicity . Juglone sensitivity was assessed in 1 day old adult worms on plates containing 240 µM juglone ( Sigma ) . For this assay , plates were made fresh on the day of the assay as the toxicity of juglone decreases rapidly over time . Survival was monitored for 6 to 10 hours . To assess the ability of worms to develop under oxidative stress , a minimum of 40 eggs were placed on plates containing 0 . 2 mM paraquat and seeded with OP50 bacteria . RNA was isolated from young adult worms using TRIZOL reagent ( Invitrogen ) . Subsequently , 1 µg of RNA was converted to cDNA using the Quantitect Reverse Transcription kit ( Qiagen ) . 1 µl of the resulting cDNA preparation was used for quantitative real-time PCR using the Quantitect SYBR Green PCR kit and a Biorad iCycler RT-PCR machine . Primer sequences for sod mRNAs were previously validated [64] . A combination of three control primer sets ( cdc-42 , pmp-3 and Y45F10D . 4 ) were used as has been previously described [65] . Results represent the average of three independent biological samples , each of which was amplified in triplicate . Eggs were collected and allowed to hatch over a period of 3 hours . After 3 hours , L1 worms were transferred to a new plate and monitored for development to an adult worm . Results are the average of at least three independent trials with 20 worms per trial . Defecation cycle length in young adult worms was measured as the average time between consecutive pBoc contractions . Results represent a minimum of 3 trials with 10 worms per trial . To determine the average number of progeny produced by each strain , L4 worms were placed on individual NGM plates . Worms were transferred daily until egg laying ceased and the total number of live progeny produced was counted . Gravid adult worms were collected from five to ten 100 mm NGM plates and bleached to recover eggs . Eggs were allowed to hatch in M9 buffer over a period of 5 days before L1 worms were transferred to NGM plates . At adulthood worms were collected in M9 buffer , washed free of bacteria and oxygen consumption was measured using a Clark electrode for a 10 minute period . Subsequently worms were pelleted and frozen for protein quantification . Proteins were quantified using a bicinchonic acid protein assay kit ( Thermo Scientific ) according to the manufacturer's protocol . Western blotting for SOD proteins was completed as described previously [16] . Antibody dilutions were as follows: SOD-1 ( 1∶1000 ) , SOD-2 ( 1∶1000 ) , tubulin ( 1∶10 , 000 ) . Levels of protein were compared in three independent samples of one day old adult worms . Oxidative damage was assessed using an Oxyblot assay kit ( Millipore ) to detect carbonylated proteins as previously described [16] . In this assay carbonyl groups are derivatized to 2 , 4-dinitrophenylhydrazone ( DNP-hydrazone ) which can then be detected by western blotting with a DNP specific antibody . The Oxyblot assay was completed according to the manufacturer's protocol using 10 samples of N2 worms and 8 samples of sod-2 mutant worms ( Millipore ) . 9 µg of protein lysate was loaded in each lane . Quantification of carbonylated proteins was achieved by taking the ratio of DNP staining to tubulin staining . Heat stress experiments were based on previously developed protocols [36] . Briefly , young adult worms on NGM plates were incubated at 35 degrees Celsius for a period of 6 or 9 hours . Worms were then transferred to a 20 degree Celsius incubator . Two days later the percentage of worms surviving was determined . Osmotic stress experiments were also done according to previously developed protocols [37] . Young adult worms were transferred to NGM plates containing 500 mM NaCl . After 20 hours , worms were washed off salt plates in M9 buffer containing 300 mM NaCl and transferred to normal NGM plates . After one day of recovery , the percentage of worms surviving was determined . Results for both stress assays are the average of three independent trials . Survival plots were compared using the log-rank test . The maximum lifespan of a given strain was measured as the average of the lifespan of the ten longest living worms . A student's t-test was used to compare maximum lifespan between strains . Significance between strains for paraquat and juglone sensitivity assays were assessed by ANOVA . Oxygen consumption results were compared by student's t-test . Error bars show standard deviation . | In this paper , we examine the oxidative stress theory of aging using C . elegans as a model system . This theory proposes that aging results from the accumulation of molecular damage caused by reactive oxygen species ( ROS ) . To test this theory , we examined the effect of deleting each of the five individual superoxide dismutase ( SOD ) genes on lifespan and sensitivity to oxidative stress . Since SOD acts to detoxify ROS , the oxidative stress theory predicts that deletion of sod genes should increase oxidative stress and decrease lifespan . However , in contrast to yeast , flies , and mice , where loss of either cytoplasmic or mitochondrial SOD results in decreased lifespan , we find that none of the sod deletion mutants in C . elegans exhibits a shortened lifespan despite increased sensitivity to oxidative stress . Surprisingly , we find that sod-2 mutant worms have extended lifespan and even worms with the primary cytoplasmic , mitochondrial , and extracellular sod genes deleted can live longer than wild-type worms . By examining genetic interactions with other genes known to extend lifespan and by comparing the phenotype of worms lacking sod-2 to that of known long-lived mitochondrial mutants such as clk-1 or isp-1 , we provide evidence that the loss of sod-2 extends lifespan through alteration of mitochondrial function . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics",
"developmental",
"biology/aging"
] | 2009 | Deletion of the Mitochondrial Superoxide Dismutase sod-2 Extends Lifespan in Caenorhabditis elegans |
Transgenic crops producing insecticidal toxins from Bacillus thuringiensis ( Bt ) are commercially successful in reducing pest damage , yet knowledge of resistance mechanisms that threaten their sustainability is incomplete . Insect resistance to the pore-forming Cry1Ac toxin is correlated with the loss of high-affinity , irreversible binding to the mid-gut membrane , but the genetic factors responsible for this change have been elusive . Mutations in a 12-cadherin-domain protein confer some Cry1Ac resistance but do not block this toxin binding in in vitro assays . We sought to identify mutations in other genes that might be responsible for the loss of binding . We employed a map-based cloning approach using a series of backcrosses with 1 , 060 progeny to identify a resistance gene in the cotton pest Heliothis virescens that segregated independently from the cadherin mutation . We found an inactivating mutation of the ABC transporter ABCC2 that is genetically linked to Cry1Ac resistance and is correlated with loss of Cry1Ac binding to membrane vesicles . ABC proteins are integral membrane proteins with many functions , including export of toxic molecules from the cell , but have not been implicated in the mode of action of Bt toxins before . The reduction in toxin binding due to the inactivating mutation suggests that ABCC2 is involved in membrane integration of the toxin pore . Our findings suggest that ABC proteins may play a key role in the mode of action of Bt toxins and that ABC protein mutations can confer high levels of resistance that could threaten the continued utilization of Bt–expressing crops . However , such mutations may impose a physiological cost on resistant insects , by reducing export of other toxins such as plant secondary compounds from the cell . This weakness could be exploited to manage this mechanism of Bt resistance in the field .
Insecticidal protein toxins of the Cry1A family produced by certain strains of the gram-positive bacterium Bacillus thuringiensis ( Bt ) are highly active against many Lepidoptera but nontoxic to most other animal species . Transgenic cotton producing Cry1Ac and transgenic maize producing Cry1Ab have been grown commercially since 1996 and offer protection against some major pests , including species in the genera Heliothis , Helicoverpa , Ostrinia , and Pectinophora [1] , [2] . After ingestion and solubilization in the alkaline midgut lumen of the caterpillar , the protoxin is cleaved by digestive proteases to yield an active 60 kDa toxin which interacts with high-affinity binding sites on the brush border epithelium , eventually oligomerizing to form a transmembrane pore , leading to lysis of epithelial cells [3] , [4] . Additional mechanisms of toxicity involving an adenylyl cyclase/PKA signaling pathway have also been described [5] . High toxin concentrations are lethal , lower toxin concentrations inhibit larval growth in a dose-dependent manner . The binding targets are critical in determining the range of species on which the toxin is active [6] , and reduction or loss of binding is an important mechanism of genetically based resistance in the target pest species [7] . The most common type of Bt toxin resistance ( “Mode 1” ) [8] which has evolved in field populations of Plutella xylostella in response to sprays of formulated Bt toxins [9] and in laboratory-selected strains of other Lepidoptera [10]–[12] is characterized by recessive inheritance , >500-fold resistance to at least one Cry1A toxin , much less resistance to Cry1C , and greatly reduced binding of Cry1A toxins to target sites in the midgut membrane . Several cases of resistance to Bt crops in field populations of insect pests have also been reported , but the genetic basis of resistance has not been identified in any of these cases [2] , [13] , [14] . Genetic mutations linked to Cry1A resistance have been identified in laboratory strains , but their role in Mode 1 resistance is still not fully understood . The mutations most commonly found in Cry1A-resistant strains inactivate a gene encoding a 12-cadherin-domain protein of about 1750 amino acids , expressed in the larval midgut [15]–[17] . These mutations confer resistance to Cry1A toxins including Cry1Ac , but do not block irreversible Cry1Ac binding to midgut membranes as measured by in vitro assays [12] , [18] , [19] . Conversely , Mode 1 resistance in the NO-QA strain of P . xylostella which includes loss of Cry1Ac binding [20] is determined by a single gene that segregates independently from the 12-cadherin-domain protein gene [21] . What could account for the apparent independence of resistance-conferring cadherin mutations and resistance-conferring loss of irreversible membrane binding ? There is evidence for a multi-step mechanism that could offer an explanation . Bravo et al . [22] have proposed that activated Cry1A toxin monomers first bind to an extracellular membrane-proximal domain of the 12-cadherin-domain protein . The toxin undergoes a conformational change , facilitating proteolytic cleavage of the Domain I helix α1 from the toxin N-terminus by a yet-uncharacterized protease . The resulting “clipped” toxin monomers subsequently assemble into a oligomeric pre-pore structure in solution , which binds reversibly to several other membrane-bound proteins , and finally inserts irreversibly into the membrane [22] . Thus absence of the cadherin protein in resistant strains would slow the rate of monomer clipping and oligomerization of the active pore structure , but not directly affect the subsequent irreversible binding and insertion of the pore into the membrane . This would predict that even higher levels of resistance could be attained by interfering with the later binding steps in this sequential binding model . To test this idea , it would be useful to examine toxin binding to membranes of resistant strains , with or without the cadherin protein . The first Bt-resistant cadherin mutation was identified [17] as the resistant allele 4r of the previously-mapped gene BtR-4 [23] in the YHD2 strain of the cotton pest Heliothis virescens . This strain had evolved >10 , 000 fold resistance in response to laboratory selection by diet-incorporated Cry1Ab and Cry1Ac toxin over four years [10] . Insertion of an LTR retrotransposon into the coding sequence of the 12-cadherin-domain protein defines the 4r allele , and resulted in a truncated 622-amino acid protein lacking the last 7 cadherin domains , membrane-proximal toxin binding region , transmembrane domain , and cytoplasmic domain . The absence of the 12-cadherin-domain protein from the midgut membranes of YHD2 was confirmed with antibodies [19] . The first binding measurements on homozygous resistant 4r4r YHD2 published in 1995 showed greatly reduced binding of midgut epithelial brush border membrane vesicles ( BBMV ) to Cry1Aa , but surprisingly no reduction in Cry1Ab or Cry1Ac binding [18] . YHD2 was subsequently selected to even higher levels of resistance , and later studies published in 2002 and 2004 showed a loss of membrane binding by Cry1Ab and Cry1Ac also , as well as a reduction in their pore-forming ability [19] , [24] . This suggested the existence of a second gene ( which we named BtR-6 ) with a mutant allele 6r in the more resistant strain responsible for its increased resistance and decreased binding affinity to Cry1Ac . In order to test the hypothesis that a separate mechanism affecting later steps in toxin binding existed in this more resistant strain , we sought to identify BtR-6 and the nature of the 6r allele by map-based cloning . We first isolated the two resistance mechanisms into separate strains and characterized their toxin-binding properties . We then used these strains in a series of backcrosses that were assayed for resistance using a sublethal , growth-inhibition bioassay . Fine-scale linkage mapping identified a cluster of ABC transporter genes , one of which showed an inactivating mutation in the most resistant strain . This implicates the ABC transporter family for the first time in the mode of action of Bt Cry1A toxins , and offers an explanation for Mode 1 resistance that is compatible with the sequential binding model .
In order to synthesize strains that were homozygous for different combinations of resistant and susceptible alleles at the BtR-4 and BtR-6 loci ( 4r vs 4s , 6r vs 6s ) , we used a combination of progeny testing on Cry1Ac-containing diet and marker-assisted selection of parents with a PCR ( polymerase chain reaction ) test diagnostic for 4r [25] . These strains were then maintained on artificial diet containing the highest concentration of Cry1Ac that would allow the same larval growth rate as toxin-free diet . Strain YHD3 was homozygous resistant for both genes ( 4r4r/6r6r ) , had a resistance level similar to the newer YHD2 , and was reared on 200 µg/ml Cry1Ac . YFO was 4r4r/6s6s and could be reared on at most 5 µg/ml Cry1Ac . YEE was 4s4s/6r6r and was reared on 50 µg/ml Cry1Ac . Fully susceptible strains CNW and JEN were 4s4s/6s6s and were reared on toxin-free diet; their growth rate was reduced 50% by only 0 . 064 µg/ml Cry1Ac . Qualitative in vitro binding studies with Cry1Aa , Cry1Ab , and Cry1Ac using BBMV ( Figure 1 ) showed that the doubly homozygous susceptible JEN strain ( 4s4s/6s6s ) bound all three toxins as expected . The doubly homozygous resistant YHD3 strain ( 4r4r/6r6r ) bound to none of the three , similar to YHD2 in 2002 [24] and 2004 [19] . The two intermediately resistant strains showed a complementary pattern: YFO ( 4r4r/6s6s , the hypothesized genotype of the older YHD2 strain ) had lost only the ability to bind Cry1Aa , similar to YHD2 in 1995 [18] . YEE ( 4s4s/6r6r ) still bound Cry1Aa but failed to bind Cry1Ab and Cry1Ac , a pattern that has not been previously reported ( Figure 1 ) . Thus homozygosity for the 6r allele but not 4r is correlated with loss of Cry1Ac binding . We explored the genetic basis of these resistance and binding differences by linkage mapping using a larval growth bioassay with Cry1Ac conducted on backcrosses as done previously [17] . Backcrosses to YHD3 using F1 ( YHD3 x YFO ) mothers were first screened with a panel of RFLP markers to identify the linkage group containing BtR-6 , by exploiting the absence of crossing-over during meiosis in female Lepidoptera . A probe previously mapped to linkage group 2 ( LG2 ) with similarity to a microsomal glutathione transferase ( GenBank HM150720 ) showed a highly significant association with resistance as measured by larval weight on the Cry1Ac diet . This confirmed that BtR-6 was genetically distinct from the two previously mapped resistance genes in this species , BtR-4 ( the 12-cadherin-domain protein ) on LG9 [17] and BtR-5 on LG10 [26] . Neither LG9 nor LG10 had a significant association with resistance in these crosses ( in the backcross to YFO , all progeny are BtR-4r4r ) . The significant effect of LG2 was confirmed in backcrosses to YEE using F1 ( YEE x YFO ) mothers , which were also segregating at BtR-6 . Ribosomal protein genes RpP0 , RpS5 , RpL8 , RpL10A , and RpL30 also mapped to LG2 in H . virescens , indicating homology with Chromosome 15 ( Chr15 ) of the domesticated silkmoth Bombyx mori , where these same genes had been mapped by recombinational [27] and cytogenetic [28] methods . We localized BtR-6 relative to marker genes along LG2 using recombinational mapping in backcrosses with F1 males , which do undergo crossing-over during meiosis . In the first step , markers were chosen from H . virescens and Helicoverpa armigera cDNA clones homologous to genes that had been genetically mapped to Chr15 in B . mori . The second step at a finer scale used genes physically mapped to Chr15 after the assembled B . mori genome sequence was made available to the public in April 2008 . The linkage map of LG2 in H . virescens was entirely collinear with the genetic and physical maps of Chr15 of B . mori ( Figure 2 ) . BtR-6 was localized within the interval between markers b7730 and b7793 , showing zero recombinants out of a total of 1060 informative progeny from 3 sets of mapping families that had been reared on Cry1Ac-containing diet . The physical map of this region in B . mori contains 10 predicted genes , nine of which showed expression in B . mori larval midgut as indicated by microarray studies , and which also had homologs in cDNA libraries constructed from midgut tissue of larval H . armigera ( Table S2 ) . A PCR product corresponding to B . mori predicted gene BGIBMGA007793 was amplified from H . virescens midgut cDNA , and used to screen BAC libraries of H . virescens and its sister species H . subflexa . The latter library yielded a positive clone which was sequenced ( GenBank Accession No . GQ332573 , Figure 3A ) , revealing a cluster of three genes with high sequence similarity to ABC transporters ( ABCC1 , ABCC2 , and ABCC3 ) , in the same orientation as the corresponding region in B . mori ( Figure 3B ) . Genomic sequence comparison of the ABCC2 gene from YHD3 ( GenBank GQ332572 ) and YFO ( GenBank GQ332571 ) strains revealed a 22-bp deletion in exon 2 occurring only in YHD3 ( Figure S2 ) . The same deletion was found in RT-PCR products from YHD3 larval midgut cDNA . The frameshift generated by this deletion predicts a truncated 99-residue protein from YHD3 mRNA . In contrast , the full-length ABCC2 protein of 1339 amino acids predicted from the YFO or H . subflexa sequence ( 97% amino acid identity ) has all the features of the bipartite structure of ABC transporters , with six transmembrane segments and a large cytoplasmic ATP-binding domain in each half [29] ( Figure 4 , Figure 5 ) . A PCR assay using primers flanking the exon 2 deletion region was used to determine genotypes of individual backcross progeny ( Figure S1 ) . Only those larvae with two copies of the exon 2 deletion allele grew rapidly on Cry1Ac-containing diet . The deletion was present in all YHD3 and YEE individuals tested; no YFO or CNW individuals had the deletion . This 22-bp deletion is taken to define the 6r allele of BtR-6 in YHD3 . We used PCR analysis of archival DNA samples to investigate whether an increase of the 6r allele frequency occurred concomitantly with the decrease in Cry1Ac binding affinity of YHD2 over the years . DNA from parents of YHD2 backcrosses conducted in March 1993 [23] yielded a 6r allele frequency estimate of 14% and an expected 2% frequency of 6r6r homozygotes; not high enough to appreciably reduce the binding to Cry1Ac [18] . Thus the 6r allele was present although rare in the YHD2 strain as early as 1993 . When we screened for 6r in DNA that had been isolated in December 2002 from the YHD2 larvae whose BBMV showed a loss of Cry1Ab and Cry1Ac binding [19] , we found that the frequency of 6r had increased to 100% . Thus a loss of Cry1Ab and Cry1Ac binding was correlated with an increase in 6r within YHD2 over approximately 100 generations while the Cry1Ac resistance level as measured by bioassay also increased . This correlation also extended to other strains . DNA samples from the Cry1Ac-resistant KCBhyb strain had a 6r frequency of 5%; membranes from these larvae retained Cry1Ab and Cry1Ac binding , and binding of Cry1Aa only was dependent on the BtR-4 genotype [19] . Both the Cry1Ac-resistant strain CxC with a 6r allele frequency of 0% and the Cry1Ac-susceptible strain YDK with 6% retained Cry1Aa , Cry1Ab and Cry1Ac binding [19] ( Table S3 ) .
Recent research has shown that the mode of action of Bt toxins is more complex than originally envisioned . Cry toxins may induce cell death by interacting with the 12-cadherin-domain protein without forming pores [5]; responses to Cry toxins may involve intracellular signal transduction pathways that protect cells against pore forming toxins [30] , [31] . Yet a major feature of Cry1A toxin action in Lepidoptera is the formation of pores in the plasma membrane leading to cell disruption by colloid-osmotic lysis [32] . At high enough concentrations , Cry toxins can eventually insert and form pores in planar lipid bilayer membranes devoid of any other protein [33] , [34] . However , these toxins have evolved to interact with a series of host proteins in the midgut membrane to form pores much faster and at much lower concentrations . These interactions are toxin- and host-specific , e . g . Cry1A toxins are active against certain Lepidoptera , but not Diptera or Coleoptera . Interfering with one or more of these steps may confer resistance , such that higher concentrations of toxin are required to achieve the same mortality endpoint . Identifying the molecular changes that accompany resistance is a useful first step to posing hypotheses about the mode of toxin action . Based on the mapping results and binding correlations described here , we hypothesize that the ABCC2 protein participates in the mechanism of Cry1Ab and Cry1Ac toxicity by binding and facilitating insertion into the membrane , in an extension of the multi-step model of Bravo et al . [35] . In the first step of this model , reversible toxin binding to the 12-cadherin-domain protein accelerates the formation of clipped toxin monomers which are more competent to form the oligomeric pre-pore structure in solution . Evidence supporting this mechanism includes the enhanced toxicity of Cry1Ab or Cry1Ac toxin when fed to larvae along with a peptide fragment from the toxin-binding domain of the cadherin protein [36] . This fragment itself binds to Cry1Ab and Cry1Ac [36] , and accelerates the rate of formation of a 250 kDa oligomer of Cry1Ac [37] . Additional evidence is provided by the elevated potency of “pre-clipped” Cry1Ab or Cry1Ac monomers generated by recombinant methods , which lack the α1 helix [38] . These modified Cry1AbMod and Cry1AcMod toxin monomers rapidly form oligomers in the absence of the cadherin , and are more potent than unmodified toxins against Cry1Ac-resistant Pectinophora gossypiella with cadherin mutations [38] , although possessing similar properties in most other respects [39] . According to this model , absence of the 12-cadherin-domain protein confers a certain level of resistance to Cry1Ab or Cry1Ac by slowing down the formation of oligomers , not by stopping it completely . Evidently oligomerization of these two toxins can occur in the absence of cadherin binding , but at a slower rate; since higher concentrations of Cry1Ab or Cry1Ac are still capable of killing resistant insects with cadherin mutations . Moreover , even if the cadherin functions in accelerating the “clipping” of Cry1Ab and Cry1Ac toxin monomers , this does not rule out a separate role in additional binding events . The cadherin appears to be the major binding protein for Cry1Aa; as BBMV from strains lacking it have also lost their ability to bind Cry1Aa [18] , [19] . Furthermore , presence of the cadherin appears to be necessary and sufficient for binding of BBMV to Cry1Aa , but not Cry1Ab or Cry1Ac ( Figure 1 ) . The 12-cadherin-domain protein from B . mori also binds to Cry1Aa [40] , but experiments on the effect of the cadherin on oligomerization have not yet been conducted on Cry1Aa . Therefore , cadherin binding may play more than one role , depending on the toxin . In the second binding step in the hypothesized mode of action [35] , toxin oligomers bind to the soluble ectodomains of membrane-associated glycosylated proteins such as aminopeptidase N ( APN ) [41] , [42] , alkaline phosphatase [43] , [44] , P252 glycoprotein [45] , or BTR-270 glycoprotein [46] . These proteins are GPI-anchored and enriched in lipid rafts , and disruption of lipid rafts by cholesterol depletion reduces pore formation by Cry1Ab [47] . Experimental cleavage of GPI anchors removes APN from the BBMV surface and reduces the amount of Cry1Ab toxin inserted into the membrane [22] . Massive shedding of GPI-anchored proteins by the action of endogenous phospholipase C has been shown to occur in response to toxin consumption [48] , which might be a defense mechanism against the second binding step , but so far this has not been observed to occur in any resistant strains . Toxin binding to these glycoproteins appears to be predominantly reversible; e . g . binding of Cry1Ac to purified APN exhibits measurable on- and off-kinetics by surface plasmon resonance [49]–[51] . No single glycoprotein appears to be essential for Cry1A toxicity; e . g . mutants of Cry1Ac which eliminate binding to a 115 kDa APN only result in a two-fold decrease in toxicity [52] . RNA interference directed against midgut APNs produces a measurable but slight decrease of toxicity [53] , [54] . Therefore the main significance of Cry1A toxin binding to these glycoproteins seems to be to increase the concentration of the pre-pore oligomer at the membrane surface , increasing the probability of eventual insertion by some other mechanism . The final binding step proposed here involves interactions of the oligomeric toxin pre-pore structure with the ABCC2 protein . ABC transporters cycle between closed and open configurations as they transport small molecules out of the cell , driven by binding and hydrolysis of ATP by the intracellular nucleotide-binding domains . A recently determined structure for the ABCB1 P-glycoprotein shows that in the closed configuration , the extracellular loops between the transmembrane domains completely cover the channel opening , resulting in a large internal cavity facing the cytoplasm [29] . In this pretransport state , the small molecule to be transported is located within the internal cavity . Binding of ATP by the two cytoplasmic nucleotide-binding domains causes their dimerization and a large conformational change resulting in the open state , in which several hydrophobic surfaces of the channel are transiently exposed to the outside of the cell while the small molecule is expelled [29] . Hydrolysis of the ATP restores the ABC protein to the closed configuration . We hypothesize that Cry1Ab and Cry1Ac toxins , as pre-formed oligomers or possibly also as monomers , bind to the open configuration of ABCC2 and that this facilitates subsequent membrane insertion . The resistance conferred by BtR-6r would thus be due to the absence of this binding site for Cry1Ab and Cry1Ac . Direct toxin binding assays with the membrane-integrated ABCC2 protein would be required for evaluation of this hypothesis . To our knowledge , ABC transporters have not yet been suggested as binding targets for Bt toxins . Failure to detect them may be due to the under-representation or absence of integral membrane proteins in 1-D or 2-D gels used in ligand binding studies with labelled toxin [55] , [56] . Failure to isolate them could be due to the general difficulty of isolating membrane proteins . The midgut proteins from Lepidoptera previously isolated on the basis of binding to Cry1A-toxin-immobilized affinity columns [42] , immunoprecipitation [57] , [58] or preparative gel electrophoresis [46] all have a large ectodomain projecting into the lumen available for binding , and are readily solubilized , being attached to the membrane by a GPI anchor or a single transmembrane domain . The predicted structure of ABCC2 , however , presents only 6 small loops ( of 19 , 5 , 5 , 43 , 5 , and 5 residues respectively ) projecting into the lumen , which connect the 6 α-helices of each of the two transmembrane domains buried in the lipid bilayer ( Figure 4 , Figure 5 ) . Cry1A toxins are known to bind to carbohydrate residues of glycoproteins , but none of the 6 loops of ABCC2 have predicted glycosylation sites . If the toxin binds primarily to the hydrophobic interior of the channel , then methods stringent enough to solubilize the ABC protein would likely disrupt this interaction . If confirmed , the role of an ABC transporter in Bt toxin action proposed here could have implications for the management of Cry1Ac resistance in field populations of H . virescens and other lepidopteran pests currently controlled by Bt-cotton or Bt-maize . We emphasize that as no attention has been paid to ABC transporters in Bt resistance previously , we do not know whether this or similar mutations occur in the field in H . virescens or any other species . However , the genetic basis of field-evolved resistance to Bt sprays by Plutella xylostella [21] , [59] and Trichoplusia ni [60] , and to Bt crops by Helicoverpa zea , Spodoptera frugiperda , and Busseola fusca [2] has not yet been identified , and these strains should be examined for ABC transporter mutations . We do not know whether H . virescens larvae homogozygous for the ABCC2 mutation can survive on cotton , with or without Cry1Ac toxin . Developmental arrest in the last larval instar of the YHD2 strain feeding on non-transgenic cotton was observed prior to 1993 [25] , when BtR-4r was nearly fixed , indicating a strong fitness cost to the cadherin mutation; but at that time BtR-6r was still at a very low frequency . We do not know how ABCC2 mutations would respond to selection for Cry1A-toxin resistance in the field . In India , China , and many other countries , the predominant varieties of Bt-cotton still produce the single toxin Cry1Ac , thus selection for Cry1Ac resistance is strong . The Bt-cotton currently used in the USA and Australia produces Cry2Ab in addition to Cry1Ac; the different modes of action of these two toxins are thought to produce a “redundant killing” effect whereby selection for resistance to either single toxin is greatly weakened . However , we do not know whether ABC transporter mutations confer cross-resistance to Cry2Ab . The binding targets of Cry2Ab are unknown and ABC proteins have not yet been investigated as candidates . Moreover , Cry2Ab resistance is detectable using F2 screens in Australian populations of Helicoverpa armigera [61] and H . punctigera [62] , and the molecular basis of the resistance mechanism involves binding site alterations in both species [63] . The biological function of ABCC2 is unknown , but its similarity to multidrug resistance proteins suggests that it could export small hydrophobic toxins from midgut epithelial cells for eventual elimination in the feces . Homozygous deletions of ABCC2 as seen in the YEE and YHD3 strains have no obvious effect on insects consuming artificial toxin-free diet in the laboratory . However , plant secondary compounds that deter herbivory or poison the herbivore would be encountered by larvae consuming plants in nature , affecting Bt-susceptible and resistant insects in different ways . If exported by an active ABCC2 in Bt-susceptible insects , they could potentiate the Bt-toxin by increasing the proportion of time the channel is in the open state , exposing the hydrophobic inner surfaces to toxin binding . There is evidence for an effect of different plant tissues with different amounts of secondary compounds on the potency of Cry1Ac [64] . Additionally , by imposing a fitness cost on Bt-resistant insects they could select against resistance alleles encoding defective variants of the ABCC2 protein that fail to export them . For example , Bt-resistant Pectinophora gossypiella is more sensitive to the cotton secondary compound , gossypol [65] . Even a slight fitness cost of ABCC2 mutations would be effective in delaying the increase of resistance alleles , the goal of the high-dose/refuge strategy mandated by the US Environmental Protection Agency . PCR-based DNA diagnostics for specific ABCC2 mutants shown to be present in the field could be useful in supporting the continued success of this strategy by monitoring resistance alleles in field populations of insect pests .
All crosses used virgin adults of Heliothis virescens in single-pair matings . Resistant strain YHD2 was crossed to the susceptible strain CNW ( July 2001 ) and F1 offspring were intercrossed . F2 progeny were reared on artificial diet [66] containing 0 . 2 µg/ml Cry1Ac toxin for 10 days , individually weighed , and transferred to toxin-free diet for rearing to adulthood . The 10-day weights were used as an indication of the ability to resist the growth-inhibiting effect of this sublethal Cry1Ac concentration , due to the presence of different combinations of resistant and susceptible alleles at the two resistance genes BtR-4 and BtR-6 . F2 adults from the top third of the weight distribution were intercrossed to form the YHD3 strain , which was subjected to selection on Cry1Ac-containing diet over 25 generations , eventually attaining the same resistance level as the parent YHD2 . It was maintained on artificial diet with 200 µg/ml Cry1Ac . To develop the YFO strain , F2 adults from the middle third of the weight distribution were repeatedly backcrossed in single-pair matings to the susceptible CNW strain . Parents were scored for the presence of the BtR-4r allele by PCR using the primers SF1 , SR2 , and RR3 ( Figure S1 ) [25] , after collection of fertile eggs . Only progeny of parents that still carried the BtR-4r allele were retained for subsequent matings . Larvae of these generations were reared on toxin-free diet to avoid any toxin-based selection of resistance alleles . After 6 generations of backcrossing and PCR screening , YFO adults were intercrossed and subsequent generations made homozygous for BtR-4r , after which the strain was raised on 5 µg/ml Cry1Ac . The YEE strain was developed by intercrossing the F2 from the lower third of the weight distribution and subsequent generations , and keeping only progeny of parents with the lowest frequency of BtR-4r alleles as detected by PCR . Larvae of this strain were reared on diet with 5 µg/ml Cry1Ac toxin to select for BtR-6r alleles . After no parents were found to carry BtR-4r alleles , the YEE strain was maintained on 50 µg/ml Cry1Ac . As YFO was homozygous BtR-4r4r and YEE was subsquently shown to be homozygous BtR-6r6r , the ABCC2 mutation permits larvae to consume 10 times as much Cry1Ac without growth retardation as does the cadherin mutation . All strains showed equivalent growth in the laboratory on artificial diet with no toxin . Backcross larval progeny were tested by rearing on a sublethal concentration of Cry1Ac in artificial diet [17] , allowing normal growth ( i . e . equivalent to susceptible individuals on non-Bt diet over the same time period ) in individuals homozygous resistant for the gene segregating in the cross , but suppressing growth in heterozygotes . Larvae were weighed to the nearest 1 mg after 7 days; backcross size distributions were strongly bimodal consistent with segregation of a single major resistance gene ( Figure S3 ) . All larvae were then transferred to toxin-free diet and reared to adults for DNA extraction . Polymorphisms at genetic marker loci were scored using RFLPs ( restriction fragment length polymorphisms ) visualized by Southern blots of restriction-digested genomic DNA , or scored by screening for intron size polymorphisms by PCR using primers placed in adjacent exons . Three series of interstrain crosses were used to generate backcross families ( BRX ) segregating at BtR-6 . In BRX28 ( February 2006 ) and BRX35 ( December 2007 ) , F1 progeny from crosses between YHD3 ( 4r4r 6r6r ) and YFO ( 4r4r 6s6s ) were backcrossed to YHD3 . Backcross progeny were expected to be 4r4r 6r6r or 4r4r 6r6s; they were tested on 25 µg/ml Cry1Ac . In BRX36 ( June 2008 ) , F1 progeny from crosses between YFO and YEE ( 4s4s 6r6r ) were backcrossed to YEE . Backcross progeny were expected to be 4s4s 6r6r , 4r4s 6r6r , 4s4s 6r6s , or 4r4s 6r6s . To minimize the effect of segregation of the 4r allele , which is recessive at high concentrations , backcross progeny were tested on 50 µg/ml Cry1Ac and otherwise treated as in the other two series . The linkage analysis strategy exploited the absence of crossing-over during meiosis in female Lepidoptera [67] . Female-informative backcrosses ( with F1 mothers ) were examined first to verify that segregation of LG2 markers correlated with larval weight . Male-informative backcrosses ( with F1 fathers in which crossing-over occurs ) were then used to estimate linkage relationships among LG2 markers and resistance as measured by larval weight on Cry1Ac-containing diet . For RFLP analysis , DNA was isolated from adults using phenol and chloroform , digested with HindIII or PstI , electrophoresed on 0 . 8% agarose gels , and transferred to Hybond N+ filters for probing with 32P-labelled probes . RFLP probes for LG2 markers were generated from H . virescens or Helicoverpa armigera cDNA probes previously mapped to LG2 , or from genes mapped to Bombyx mori Chromosome 15 . These were used to search EST databases of H . virescens and H . armigera by BLAST , or to design degenerate PCR primers for amplification and sequencing from H . virescens cDNA or gDNA . Intron size polymorphisms in some markers were scored by agarose gel electrophoresis of PCR products generated using primers positioned in adjacent exons . Three strategies were used to screen B . mori Chromosome 15 for markers that could be used in mapping , in a sequential approach to narrow the interval containing BtR-6 . First , sequence information from the RAPD-based linkage map of Yasukochi et al . [27] , [68] was used in BLASTN searches of the wgs section of GenBank ( http://www . ncbi . nlm . nih . gov ) to identify whole-genome-shotgun contigs produced by the first [69] and second [70] genome assemblies , and these in turn were screened for conserved coding sequences present in the H . armigera and H . virescens cDNA libraries . This approach was limited by small contig size and frequent occurrence of chimeric contigs . Second , a BAC-walking strategy was employed using BAC-end sequences deposited in the gss section of GenBank [71] . BAC ends occuring in contigs were identified by BLASTN to gss , the other end was obtained by a text search using the BAC clone name , and used to identify the contig in which it occurred by BLASTN to wgs . Third , when the third genome assembly [72] was made available to the public on-line on SilkDB ( http://silkworm . swu . edu . cn/silkdb/ ) [73] and Kaikobase ( http://sgp . dna . affrc . go . jp/index . html ) [74] , predicted genes in the genome browser view were used . Serial numbers of BGI predicted genes are represented here by the last four digits; e . g . b7795 for BGIBMGA007795 . These approaches were successful because of the high degree of evolutionary conservation of gene order among Bombyx and Heliothis for this linkage group . Recombinants were identified by reference to parental and grandparental genotypes and tallied by hand in order to guide the direction of search for additional markers . The final linkage map was constructed using 20 markers and 1060 offspring using the program Mapmaker3 [75] with Haldane centimorgans . A Macintosh PowerBook running the MacPort implementation of the unix version was used , as the MS-DOS version of this program running under Windows crashed with our dataset . After BtR-6 was localized within the interval between markers b7730 and b7793 showing zero recombinants , the linkage map of B . mori Chr15 was examined and found to also have zero recombinants out of 190 informative progeny in the corresponding region [71] . The physical map of this region in B . mori contains 10 predicted genes [72] , nine of which showed expression in B . mori larval midgut as indicated by microarray studies [76] and also had homologs in cDNA libraries constructed from midgut tissue of larval H . armigera ( Table S2 ) . Actively feeding early fifth-instar H . virescens larvae were chilled on ice and dissected , ( May 2007 ) . Tracheae , Malpighian tubules , peritrophic matrix and food bolus were removed and the midgut tissue was rinsed briefly in ice-cold phosphate-buffered saline ( PBS ) . Brush border membrane vesicles ( BBMV ) were prepared by the Mg2+ precipitation method according to Wolfersberger et al . [77] . The final BBMV pellet was resuspended at a protein concentration of 1 mg/ml in PBS ( determined by the BCA protein assay with BSA as standard , Bio-Rad ) and stored at −80°C until use . Brush border membrane enrichment was estimated by measuring the aminopeptidase activity using L-leucine-p-nitroanilide as a substrate . Typical enrichment of the leucyl-aminopeptidase activity in the BBMV preparation was between 5 and 6 fold compared to the initial midgut homogenate . E . coli strains harboring individual Cry1Aa , Cry1Ab , or Cry1Ac genes cloned into pKK223-3 were obtained from the Bacillus Genetic Stock Center ( Ohio State University ) . Cry1A protoxins were prepared according to Lee et al . [78] , and were activated by trypsin at a trypsin/protoxin ratio of 1/50 ( w/w ) at 37°C for 1 h . Activated toxins were further purified by anion exchange chromatography using a 1 ml RESOURCE Q column ( GE Healthcare ) . For toxin biotinylation , 0 . 5 mg of purified toxins was incubated ( 1∶30 molar ratio ) with NHS-Biotin ( Sigma ) for 30 min at room temperature . To remove excess biotin , samples were run through a 5 ml HiTrap desalting column ( GE healthcare ) . Qualitative binding assays were performed by incubating 2 . 5 nM of each biotinylated Cry1A toxin with BBMV ( containing 20 µg protein ) for 1 h at room temperature . Then , BBMV were pelleted by centrifugation ( 13 , 000 g , 10 min , 4°C ) and washed three times with PBS to remove unbound toxin . The final pellet was resuspended in SDS-PAGE sample buffer , boiled for 5 min , and proteins were resolved on a 10% SDS-PAGE gel . Toxin binding was revealed by western blot using streptavidin-HRP ( Sigma ) and ECL ( GE Healthcare ) . The homologous competition experiment was performed as described above except that biotinylated toxin and BBMV were incubated in the presence of a 200-fold excess of the corresponding unlabeled Cry1A toxin . High-density filters for a BAC library of H . virescens [79] were obtained from the Texas A&M BAC Center ( http://hbz7 . tamu . edu ) , and high-density filters for a BAC library of H . subflexa were obtained from the Clemson University Genomics Institute ( CUGI , http://www . genome . clemson . edu ) . These were screened by hybridization using a 32P-labelled 236-bp PCR product amplified from H . virescens larval midgut cDNA using primers Ha-ABC2-U14-F1 ( 5′ AACAA TCGTT ACCTG ATGGC GT ) and Ha-ABC2-U14-R2 ( 5′ AGGAT TGGTA TCGAA AAATC TCATT AC ) for the H . subflexa filters , and a 252-bp PCR product from the nachbar gene using primers Ha-bgi07733-F7 ( 5′ GAACT TGGGA CCTAC AGGTG GTAT ) and Ha-bgi07733-R10 ( 5′ GCAGC ATTAC GGATA TTAAT TTCAA C ) . The H . virescens filters yielded two positive clones , and the H . subflexa filters 20 positive clones , which were obtained from CUGI and re-screened by PCR with primers Hs-BACscr02-F1 ( 5′-CACCG GCTCA ACACC ATCAT ) and Hs-BACscr02-R2 ( 5′-GTCCT TGGCC ATGCT GTAGAA ) . Clone HS_Ba 89F08 was chosen and shot-gun sequenced at the Max Planck Institute for Chemical Ecology , Department of Entomology and deposited in GenBank as GQ332573 . Primers designed from the H . subflexa sequence ( Table S1 ) were used to amplify the ABCC2 gene in overlapping fragments from genomic DNA; sequence from YHD3 was deposited as GQ332572 and from YFO as GQ332571 . Alignment of exon 2 of the YHD3 and YFO sequence revealed a 22-bp deletion in the former , causing a frameshift and resulting in a predicted stop codon after residue 99 ( Figure S2 ) . Conceptual translations of the ABCC2 coding sequence from H . subflexa and the YFO allele of H . virescens were subjected to analysis for conserved domains by blastp to the Conserved Domain Database of NCBI ( http://www . ncbi . nlm . nih . gov/cdd ) and for transmembrane topology by the server ( http://phobius . sbc . su . se/ ) for the prediction program Phobius [80] . Potential glycosylation sites were screened for using the CBS Prediction Servers ( http://www . cbs . dtu . dk/services/ ) ; none were found in the sequences examined . Conserved domains , predicted transmembrane domains and extracellular loops are depicted on a sequence alignment of ABCC2 from H . virescens , H . subflexa , B . mori , and homologues from Drosophila melanogaster and Mus musculus ( Figure 4 , Figure 5 ) . PCR with primers eU02-F1 and eiT02-R10 ( Table S1 ) flanking the region containing the 22-bp deletion in the 6r allele were used to genotype individuals ( Figure S1 ) used in previous mapping crosses and binding studies . YHD2 strain individuals from March 1993 are the adults used in crosses to map BtR-4 from which DNA was still available; no binding data are available from that generation . No DNA was available from individuals in the binding studies of Lee et al . [18] in 1995 . All other samples come from binding studies of Jurat-Fuentes et al . in 2004 [19] in which midguts were dissected from individual larvae in December 2002 , the genotypes at BtR-4 were determined by PCR , and midguts from individuals with the same BtR-4 genotypes were pooled for binding analysis as shown in Figure 2 of that publication [19] ( Table S3 ) . | Crystal toxin proteins from Bacillus thuringiensis ( Bt ) make ideal bioinsecticides because of their high potency against certain insects and lack of activity against most other species . Transgenic cotton and maize expressing pore-forming Cry1A Bt-toxins are now widely used in agriculture , enabling substantial reductions in the use of chemical insecticides . However this has greatly increased the selection pressure in pest populations for toxin resistance . Preventing or delaying the development of this resistance is a high priority , to avoid a replay of the onset of insecticide resistance brought on by dependency on chemical pesticides . Because the molecular details of Bt mode of action are still not fully understood , insect strains collected from the field and selected to high levels of resistance in the laboratory are useful in discovering the obstacles the toxin must overcome before it finally forms the pore and kills the insect . We used a genetic approach to explore a poorly understood step in the toxin mode of action , which is blocked in an extremely resistant strain of an important cotton pest . As well as providing the tools to diagnose this type of resistance when it appears in the field , this discovery suggests factors that may counteract its eventual spread . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/animal",
"genetics",
"cell",
"biology/cellular",
"death",
"and",
"stress",
"responses",
"biotechnology/plant",
"biotechnology",
"biochemistry/biomacromolecule-ligand",
"interactions"
] | 2010 | An ABC Transporter Mutation Is Correlated with Insect Resistance to Bacillus thuringiensis Cry1Ac Toxin |
Noise is a prevalent and sometimes even dominant aspect of many biological processes . While many natural systems have adapted to attenuate or even usefully integrate noise , the variability it introduces often still delimits the achievable precision across biological functions . This is particularly so for visual phototransduction , the process responsible for converting photons of light into usable electrical signals ( quantum bumps ) . Here , randomness of both the photon inputs ( regarded as extrinsic noise ) and the conversion process ( intrinsic noise ) are seen as two distinct , independent and significant limitations on visual reliability . Past research has attempted to quantify the relative effects of these noise sources by using approximate methods that do not fully account for the discrete , point process and time ordered nature of the problem . As a result the conclusions drawn from these different approaches have led to inconsistent expositions of phototransduction noise performance . This paper provides a fresh and complete analysis of the relative impact of intrinsic and extrinsic noise in invertebrate phototransduction using minimum mean squared error reconstruction techniques based on Bayesian point process ( Snyder ) filters . An integrate-fire based algorithm is developed to reliably estimate photon times from quantum bumps and Snyder filters are then used to causally estimate random light intensities both at the front and back end of the phototransduction cascade . Comparison of these estimates reveals that the dominant noise source transitions from extrinsic to intrinsic as light intensity increases . By extending the filtering techniques to account for delays , it is further found that among the intrinsic noise components , which include bump latency ( mean delay and jitter ) and shape ( amplitude and width ) variance , it is the mean delay that is critical to noise performance . As the timeliness of visual information is important for real-time action , this delay could potentially limit the speed at which invertebrates can respond to stimuli . Consequently , if one wants to increase visual fidelity , reducing the photoconversion lag is much more important than improving the regularity of the electrical signal .
The phototransduction cascade consists of a series of chemical reactions which convert light inputs into usable current signals at the retina . As it serves as the front end to the visual system , it should be capable of extracting information from environmental inputs with high accuracy [1] . However , the cascade mechanism involves small numbers of molecules and therefore randomly executing reactions . Consequently , the electrical representation of each photon is non-identical and subject to inevitable variability . The signal is therefore degraded during photoconversion and the cascade said to be intrinsically noisy [2] . However , the cascade is not responsible for all the signal variability measurable at the retina . The light input itself introduces randomness through the variable timing of photons . Since the cascade has single photon sensitivity [3] this extrinsically introduced noise is also transferred across the cascade . This paper focuses on disentangling the relative contributions of these various noise sources . The noise fidelity of phototransduction is critical . According to the data processing inequality [4] , the information obtainable from the output of this process cannot be improved upon by later neuronal computations . As a result , whatever limits cascade performance may also delimit overall visual performance . An important and as yet unresolved question has been the determination of the relative influence of extrinsic photon noise as opposed to the intrinsic degradation caused by cascade reactions in representing the light stimuli . Since photon noise falls with intensity the relative noise quantification is necessarily a function of light intensity [5] . If in certain intensity regimes photon noise dominates , then the cascade can be considered sufficiently reliable for purpose at these luminance settings . If at other intensities the cascade noise is found limiting then one can query what characteristic of the phototransduction machinery is fundamentally stopping further intrinsic noise reduction . This relative noise problem remains unsolved not only in invertebrate vision but also in vertebrate rods [6] , and has parallels in other sensory systems like olfaction . While there are important structural and dynamical differences between invertebrate vision and rods ( fundamentally different types of electrical signalling despite similar cascades [7] ) and olfaction ( photons are replaced by odour molecules [8] ) they are all characterised by an overall motif known as G protein signalling . Hence , clarifying the invertebrate intrinsic-extrinsic noise relation can have useful and far reaching consequences for sensory analysis . Previous researchers have focussed on quantifying noise limits by estimating the comparative contributions of photon and cascade noise on the electrical signal variance [9] or by applying linearised filter approximations to the cascade to calculate relevant noise spectra [10] . However , such approaches generally do not account for the point process or time ordered nature of this problem . These unmodelled problem dynamics may be especially important in phototransduction given that it is sensitive to individual photons and that its output is used in real time higher visual processing . This paper will address these issues by applying and adapting Snyder filtering techniques [11] to data generated from the archetypal phototransduction system model of Drosophila melongaster . The Snyder filter is the point process analogue of the well known Kalman filter [12] . It provides a mathematically natural way of obtaining causal minimum mean squared error estimates of hidden Markov variables signalled via discrete observation events . Algorithms for extracting data from transduced photons will be developed and used in conjunction with Snyder based stimuli reconstruction techniques to show that photon noise dominates in low light while mean cascade delay limits system performance under bright conditions . This work will clarify and decompose the noise sources which dominate phototransduction while highlighting the importance of keeping the analysis discrete and causal . The invertebrate phototransduction cascade consists of a series of stochastic molecular reactions . The Drosophila cascade is particularly important as an analysable and testable archetype , not only for invertebrate phototransduction , but also for more general G protein signalling motifs [13] . G proteins are important signal transducing proteins that are ubiquitous across many biological processes . It is therefore critical to understand the properties and performance of the Drosophila cascade . Such an understanding would not only provide general insight into major biological signalling strategies but could also contribute useful theory , applicable to artificial visual systems aiming at high performance with minimal processing . Early vision in Drosophila involves light impinging on the rhabdomere of a photoreceptor , which is composed of about 30 , 000 microvilli [13] . Microvilli are essentially semi-autonomous processing units which absorb photons locally and produce quantum bumps ( QBs ) in response . The reaction set occurring in each microvilli involves photon absorption by Rhodopsin which leads to activation of the G protein via a guanosine diphosphate to guanosine triphosphate exchange . This results in activation of phospholipase C which liberates secondary messengers . There are only a few G protein and phospholipase C molecules per microvillus [14] . The secondary messengers activate the first light sensitive channel after a variable delay ( 15-100ms ) [13] . This delay corresponds to the time required for the messengers to cooperatively overcome an effective channel activation threshold . Activation of a single transient receptor potential ( TRP ) or TRP-like ion channel results in an influx of Ca2+ which rapidly activates the remaining channels in the microvilli ( 15-20 total ) via a positive feedback mechanism . This generates a current which is quickly deactivated via a negative feedback Ca2+ based loop [13] . The effect of these regulated loops is the production of an essentially discrete , unitary photon signal or QB . A single QB codes for a single absorbed photon at a microvillus . The QBs generated by each stimulated microvillus sum linearly to produce the macroscopic photoreceptor response at the receptor cell membrane . As light intensity increases the QBs become faster and smaller ( adaptation ) which ensures that the cell dynamic range is properly utilised . Despite these changes it appears that photoreceptors are linear event counters even at daylight luminance values [14] . The entire cascade , from input photon to output QB is therefore described by transformations involving small numbers of discrete molecules or channels . This results in randomly timed reactions and probabilistic interactions that lead to process variability that is visible in the resulting QB signal [14] . The QB is a non-linear and stochastic electrical depolarisation that codes for an absorbed photon . Each QB has non-identical and randomly distributed latency , amplitude and duration . These stochastic QBs lead to variable representations of identical light stimuli which reduces the reliability of event representation . The variable latency between photon absorption and bump signalling is controlled by the time required for the second messenger to accumulate beyond a dynamic threshold [15] . The variable amplitude and width of the QB are determined by the sequential Ca2+ feedback loops . Evidence suggests that the QB latency is uncorrelated with the QB waveform and highly supports the conclusion that different and independent cascade reaction sets are responsible for these characteristics [16] . Consequently , one can think of QB shape distortion and latency as two independent and important manifestations of mechanistic cascade noise . However , these sources of randomness , which are collectively termed cascade noise , are not the only forms of intrinsic noise [17] . Spontaneous ( in the absence of light ) activation of G proteins , Rhodopsin or single TRP channels can lead to spurious ( false positive ) QBs which are indistinguishable from photon stimulated QBs . While a possible issue at high temperature-low intensity settings , this dark noise source is negligible under the conditions investigated here and often the cascade suppresses these events via inbuilt molecular threshold signalling techniques [15] . Additionally a conceptual equivalent to false negative QBs also exists . When a light stimuli is presented to the cascade , microvilli respond independently to incident photons and produce respective QBs . Due to cascade stochasticity ( random reaction timings ) , not every photon becomes a QB but instead there is a quantum capture efficiency , QE= no . effective photons no . incident photons . Experiments , however , indicate that often QE is very high especially at low intensities where it can be close to 1 [18] . Hence the terms intrinsic and cascade noise will be used synonymously , and QE = 1 assumed . Supplements S3 Text , S4 and S5 Figs show that QEs of at least 0 . 66 have negligible noise impact . Since there is a clear and intimate link between intrinsic noise and the cascade machinery , any noise analysis will require the incorporation of a physiological model that describes the full reaction set from Rhodopsin absorption to TRP channel opening . This work makes use of the Drosophila phototransduction model developed by Nikolic et al [19] . This model simulates all known reactions and mechanisms of fly phototransduction within a discrete , Poisson framework , uses known biochemical parameter estimates , and allows for multi-photon inputs . It was particularly chosen for this analysis since it provides a complete stochastic simulation of the cascade reactions , does not use common mass action ( continuity ) approximations and attempts to maintain a rich and unsimplified dynamical description of the process . Most importantly , it places emphasis on getting the noise distributions correct and in keeping with experimental data . Further , the model can be modified to account for other G protein based cascades , thus providing flexibility for future work [20] . Further information on the Nikolic model , including a mathematical description and a visualisation of the intrinsic noise component distributions , are provided in Supplements S4 Text , S7 and S8 Figs . This paper will investigate the relative noise source problem and quantify the main limitations on the reliability of early vision . Previous work on this problem has taken two main viewpoints . The first uses deviations from Poisson statistics as a measure of noise contributions . For a constant light intensity , photon absorption follows a Poisson distribution so that photon noise results in a photon catch variance to mean ratio of 1 . If the measured ratio is above 1 then any excess variance must be due to intrinsic noise [9] . The results of this analysis suggest that at low intensities cascade noise contributed 50% and at high intensities 90% of all noise . This method provides a clear scheme for assessing relative noise sources that is easily interpretable . Unfortunately , it is not easily generalised to non-constant or pseudorandom light sources and depends on the intensity being low enough so that QBs are easily distinguishable ( countable ) . Moreover , this scheme neglects causality which asserts that the phototransduced output at any time cannot depend on inputs after that time . Failing to factor in this information constraint can lead to spurious performance evaluations [21] . The second posits that the cascade aims to maximise ( mutual ) information transfer and calculates signal to noise ratios and channel capacities under a noisy , dynamically range-limited Gaussian channel assumption [22] . Relative noise sources are then described by how much they alter these ratios and are treated additively through their spectral densities [23] . This approach , by keeping analysis within the frequency-domain is analytically tractable and affords a convenient dissection of optimal sensory noise filtering . However , neither is the noise additive [24] nor neuronal dynamics Gaussian [25] , making these methods at best approximate . Additionally , photons are point events and their resulting QB streams are countable sums of impulse responses . As a result the phototransduction system is necessarily discrete and the transfer of information from the environment to the retina is actually via a Poisson channel . This insight is a key motivator for the approach taken in this work . The stark coding differences between Poisson and Gaussian channels were noted by Ghanem et al in [26] and suggest that conducting the phototransduction analysis under a Gaussian ( or continuity ) assumption can be misleading . Additionally , these mutual information based studies neglect time order and use random variable based definitions of capacity [4] . In reality , the noisy phototransduction input is a stochastic process and the causal capacity should be expressed as in Lestas et al [21] instead . Consequently , these signal to noise ratio based schemes are probably only reliable at very high intensities where Poisson-Gaussian approximations are valid , and likely break down at lower luminance values , where the discrete nature of the system dominates . In contrast to the above methods , this work directly and explicitly treats the time ordered , discrete and stochastic nature of the relative noise problem with point processes , integrate-fire algorithms and causal filters . No continuity or linearity approximations are made and a causal mean squared error distortion is used as the noise performance measure; in place of the usual signal to noise or variance to mean ratios . Getting the performance metric right is important , because application of different measures , even on the same data set , as noted in the work of Grewe et al [27] , can lead to different and often misleading results . Since phototransduction should maximally preserve environmental information , using causal capacity based metrics would seem natural . However , embedding causality in this way is difficult . It is known from information theory that distortion functions ( which directly calculate estimation error ) provide a useful dual to capacity [4] . In this interpretation the concept of maximising information transfer is replaced with that of minimising distortion between a true and estimated stimuli . The distortion quantifies the difference between the true stimuli , denoted x ( t ) and some estimator of x ( t ) . Note that x ( t ) is also referred to as the hidden state . Discreteness and causality are then directly accounted for by choosing an estimate that incorporates the naturally ordered and point-based structure that information takes when transmitted across a Poisson channel . This work chooses the causal conditional mean as a suitable estimator . If the causal information available until time t are the observations D 0 t then this estimator is x ^ ( t ) = E [ x ( t ) | D 0 t ] . This expectation integrates x ( t ) with respect to its conditional distribution given the causal data . The conditional mean estimator has a fundamental link to the ordered mutual information across a Poisson channel [28] and minimises a broad class of distortion functions . Among these is the function: E[ ( x ( t ) −x^ ( t ) ) 2 ] ( the expectation is now over state trajectories ) . This distortion is termed the minimum mean squared error ( MMSE ) and is chosen as the performance metric for this work . If some other estimator is used instead then this distortion is simply called the mean squared error ( MSE ) . Any estimate which is a function of D 0 t alone will incorporate causality and discreteness . However , it is only when it takes the form of x ^ ( t ) , the conditional mean , that it will minimise the MSE . Note that these square error distortions provide the additional advantage of not being sensitive to the removal of constant signals , which is a characteristic of the adaptation response in phototransduction [29] . Consequently the cleanest approach , involving the least approximations is to optimally and causally reconstruct the input intensity stimuli and directly calculate the MMSE . While mathematically more difficult than the Gaussian and variance approaches , comparing MMSEs between light intensity reconstructions subject to various noise sources is the most appropriate and quantifiable way of measuring relative noise contribution . The importance of the approach presented here may extend beyond the phototransduction noise problem . Any real-time system which receives information sequentially and in packets is subject to the constraints of causality and discreteness . Developing and adapting techniques that can incorporate these often unmodelled dynamics can therefore have wide ranging importance . Accounting for these constraints has appeared as a concern , explicitly in molecular biochemistry [21] [30] , and implicitly in general neuroscience [31] . In both cases information can be interpreted as being transferred over timing ( Poisson ) channels , with packets representing either molecular events or spikes and the real time estimation problem involving either molecular population inference or neuronal stimuli reconstruction . This work sits precisely within this framework . Bobrowski et al [32] is the only other work ( to the authors’ knowledge ) to have ( implicitly ) dealt with causality and dicreteness . They showed that causal Snyder filtering could be used to obtain real time MMSE reconstructions of neuronal stimuli without the need for common approximations such as time discretisation . Using these techniques they estimated noisy , dynamic stimuli from discrete spiking streams . This paper extends their work by i ) developing algorithms for the causal estimation of dynamic stimuli that are now no longer observable through the Poisson events they modulate , but must instead be inferred from noisy and potentially continuous waveforms based on those events , and ii ) explicitly showing why factoring discreteness , non-linearity and causality is important for relative noise analysis . In this setting photons are interpreted as analogues to information bearing spikes and it is shown that is is possible to estimate complex light inputs from noisy outputs without the need for common linearity , continuity or Gaussian approximations . Since the relative noise problem quantifies the contributions of both extrinsic and intrinsic noise , it requires both an environmental light model and a phototransduction simulator . Visual neuroscience literature has often taken two approaches to light inputs ( both in experiments and theoretical analyses ) when investigating sensory performance [33] . The first is to repeatedly present simple inputs , which are easy to generate and control ( for example short , constant intensity flashes ) , in the hope that they elicit simple responses that allow more transparent analysis of the sensory system . Averaging over the repetitions should then lead to meaningful results . The second involves using more complex , non-repeated stimuli that represent input forms that the sensory system is likely to encounter in normal operation . These naturalistic stimuli should reveal more intricate properties of the system as it is supposedly evolutionarily designed for such inputs . The downside of these is the loss of analytical tractability [33] . This works attempts to meld the advantages of both approaches by using simple but non-trivial , non-repeated light models that share some of the characteristics of naturalistic stimuli yet are still amenable to analysis . Specifically , a flickering , continuous time , discrete space Markov process description for environmental light intensity , λ ( t ) is used . This intensity modulates a Poisson process that produces a discrete photon stream , denoted as P 0 t over the domain [0 t] . Time , t , is measured in ms with λ ( t ) having units of ms−1 . These models are analytically treatable with Markov modulated Poisson theory . Naturalistic stimuli are known to have i ) stochastic fluctuations which can be large , ii ) relatively slow dynamics compared to the visual response time and iii ) long correlation times [34] . The Markov modulated stimuli used here intrinsically possess these first two properties . The third property is approximated and simplified within the Markov assumption of the model . There are two key differences , however , between the light models used here and the general naturalistic light stimuli often found in the literature . The first is that most stimuli are in terms of velocities and contrast and applied to motion detection ( optic flow ) . The models in this paper are solely in terms of light intensity . Velocities and contrasts are not appropriate for this work as the paper treats light at a single photoreceptor so that no motion signal is present [35] . This property is therefore not relevant to this work . Secondly , naturalistic stimuli would likely encompass a broad range of intensity distributions depending on ambient conditions . As a result some of these would be dissimilar to those used here . However , given the rich range of intensity functions obtainable by appropriate choices of Markov parameters , this is not a true limitation . This work focuses on switching , bimodal stimuli . The fundamental formulation of this stimuli ( called the interrupted model ) is given in the models section of this paper . More complex and general bimodal light models are described in Supplement S2 Text . This photon stream produced by the light model is the input to the phototransduction model . The Nikolic model described previously is used as the cascade simulator . It produces a QB output stream Q 0 t which , at best , can only contain the information present in P 0 t ( by the data processing inequality [4] ) . By comparing estimates of the light intensity given P 0 t or Q 0 t one obtains an understanding of the noise deterioration at both the front and back end of the cascade respectively . To make the problem non-dimensional the normalised intensity , x ( t ) = λ ( t ) α , α > 0 ms−1 is used ( it is explicitly described by the Markov states of the light model ) . The causal conditional mean estimates of x ( t ) given P 0 t , the raw photon stream , and Q 0 t , the quantum bump stream , are denoted x ^ ph ( t ) and x ^ qb ( t ) respectively . The comparison across the cascade is made with the MMSE indices: mseph=E[ ( x ( t ) −x^ph ( t ) ) 2 ] which only measures photon noise ( the noise floor ) and mseqb=E[ ( x ( t ) −x^qb ( t ) ) 2 ] which characterises the combined impact of photon and cascade noise . These estimates of x ( t ) assume a spatial integration of the output from all stimulated photoreceptors . It is well know that QBs linearly sum over the low and medium intensity range up to hundreds of photons per second [36] . Consequently for the range of intensities investigated this assumption is valid . Values of x ^ ph and mseph within this Poisson-Markov framework can be directly obtained from an optimal ( MMSE ) non-linear filter , known as the Snyder filter [11] . The Snyder filter is a Bayesian technique for inferring the posterior distribution of a hidden Markov state given modulated Poisson observations and a prior on the hidden states . Integrating this posterior gives the conditional mean estimator that is used in the MMSE distortion function . The Snyder filter is exact in that it makes no approximations or simplifications on either the state or observation dynamics . It solves a differential-difference equation on the state posterior leading to a deterministic continuous solution trajectory between observation ( event ) times with discontinuous updates at the observed random event points . The continuous component of the solution restarts after every event update and the overall posterior evolution falls within the framework of piecewise deterministic Markov processes . The equations behind the filter and their adaptation for this work are given in Supplement S1 Text . In general the Snyder filter does not require a Markov chain description of the state process . The inference procedure developed here may therefore be a useful blueprint for analyses involving other types of light models . For details of its general formulation and its application to other types of state estimation problems in biology see [11] and [32] [37] respectively . In contrast , there is no known method for calculating x ^ qb and mseqb since it is difficult to describe the cascade noise within a tractable analytical framework . This paper develops algorithms for estimating and bounding x ^ qb and mseqb . By comparing these estimates to x ^ ph and mseph meaningful conclusions will be drawn about the relative impact of intrinsic and extrinsic noise on the normalised intensity input .
Two main Markov light models are used in this work . The first is a fundamental , symmetric , stochastic on-off light switch which emits photons according to an interrupted Poisson process . It is also known as the random telegraph signal . The second is a multi-state Markov model which has a bimodal Gaussian type state distribution and represents a complex light source that flickers between two extreme intensities while also possessing small light changes about the extreme modes . The bimodal model can be thought of as a generalisation of the interrupted model to higher state spaces with additional switch frequencies . The mathematical details of the former follow . The bimodal model and its qualitative similarities and differences to the interrupted source are described in Supplements S2 Text , S1 , S2 and S3 Figs . The interrupted light source with state transition rates set at k > 0 ms−1 is shown in Fig 1 . This model has the maximum entropy rate of any 2 state stationary Markov source and is therefore the most complex model ( from an information theory perspective ) for its state size [4] . To better compare and quantify the behaviour of light models , two dimensionless parameters are introduced . The first , β = α k , characterises the number of photons produced on average per on state and can be understood as a relative ( normalised ) light intensity descriptor . Higher β indicates more photons per state and hence more usable information . The second , γ = k - 1 100 , measures the number of QBs ( with a reference maximum width of 100ms ) that can fit into the on time of the underlying Markov chain . This parameter basically describes the relative flicker of the light stimuli . Higher γ implies a relatively slower flickering light model , which should be easier to estimate . Thus , all light models will be compared in terms of their ( β , γ ) setting with increases in either likely to lead to better MSE . For more complex stimuli , such as the bimodal model , k is defined as the smallest death rate on the chain . Given the stream of photons produced by the interrupted model , an appropriate Snyder filter can be formulated to obtain the posterior state distribution and then the conditional mean estimator . The Snyder filter description usually involves a coupled differential-difference equation for each Markov state posterior component . These dynamical equations are given in Supplement S1 Text . Since the interrupted model has only 2 states and probabilities must sum to 1 , the complete Snyder equation set can be reduced to a single expression in terms of the conditional mean estimate x ^ ph . When solved with the dummy variables A = β−1 + 2−1 , B = β - 2 + 2 - 2 and C = tanh - 1 ( A - 1 B ) and initial condition x ^ ph ( 0 ) = 1 ( assumes a photon at t = 0 ) the following conditional state estimator results: x ^ ph ( t ) = A − B tanh ( B α t + C ) for 0 ≤ t ≤ s − ( 1 ) x ^ ph ( s + ) = 1 ( 2 ) The equations above are given until some time s , when a new photon arrives , with s− and s+ indicating time infinitesimally before and after the photon . This hybrid solution provides useful insight into the behaviour of the optimal MMSE estimator . On every photon x ^ ph discontinuously jumps to a value of 1 ( eq 2 ) and then continuously decays until the next photon arrives ( eq 1 ) with a minimum possible value of lim t → ∞ x ^ ph ( t ) = A - B where 0 ≤ A − B ≤ 0 . 5 . The estimator trajectory is reset on every event ( local time becomes 0 again ) . The inter-event solution at different β is given in Fig 2 . For the interrupted model , the mean time between photons is 2 α ms [38] , while the mean time between on-off switches is 1 k ms . Higher β implies more available information and lower mseph values . This effect is observable in the decay curves of x ^ ph ( t ) . At low β , due to less certainty about the state , the scheme relatively quickly decays to a steady value approaching 0 . 5 . As β → 0 , x ^ ph ( t ) → A - B = 0 . 5 , and mseph → 0 . 25 . This is no better than simply choosing the best constant estimate of x ^ ph ( t ) = E [ x ( t ) ] . At high β the decay is slower ( relative to the normalisation time 2 α ) and the estimate stays close to 1 sufficiently long such that if another photon would occur their would be little decay between them . However , the decay is sufficiently fast relative to the switching time so that if no photons occur for an adequate amount of time then the quick fall to A − B ≈ 0 accurately reflects the state switch . This is especially true for very high β where photons essentially delineate the boundaries of the of the x ( t ) trajectory such that staying close to 1 between photons and decaying to 0 in the absence of photons will , as β → ∞ , result in mseph = 0 . The more generalised filter solution for any Markov model and an illustration of the resulting trajectory for the bimodal model can be found in Supplements S2 Text and S2 Fig respectively .
While the Snyder filter computes the optimal estimate x ^ ph exactly , there is no known equivalent method for obtaining x ^ qb , the MMSE estimator given the bumps . One contribution of this paper is an algorithm that allows calculation of a good substitute for x ^ qb , denoted x ^ qb * , using the QB data stream . This non-MMSE estimator would then provide a useful upper bound on the combined distortion resulting from both intrinsic and extrinsic noise . The known relationship of the estimators can then be summarised as mse ph ≤ mse qb ≤ mse qb * . The quantity mse qb * - mse ph upper bounds the noise introduced by the cascade since mseph is precisely the distortion due only to the photon input variability . The scheme developed here for deriving x ^ qb * ( t ) is called the integrate-fire-Snyder algorithm . The integrate-fire part of the name reflects its similarity to a standard neuronal model [39] . Consider a summed QB stream Q 0 T for some maximum observation time , T , obtained by applying the photon stream P 0 T to the Nikolic phototransduction simulator [19] . The algorithm uses Q 0 T to first derive an estimated photon stream P ^ 0 T . This is then applied to the Snyder filter to obtain x ^ qb * ( t ) . While the standard Snyder filter cannot achieve MMSE estimates given a distorted photon stream , as long as P ^ 0 T is sufficiently close to P 0 T in some sense then mse qb * should not be too much larger than mseqb . The integrate-fire part of the algorithm generates a good P ^ 0 T from the QB signal by minimising the dissimilarity between the real and estimated photon streams . This is done by treating photons as spikes and optimising a spike distance metric across a training data set . The Victor-Purpura Dspike [40] measure , which calculates the optimal transformation from one train to the other via shifts , deletions or insertions of spikes ( all of which have an associated transformation cost ) , was used . Dspike performed well and possessed clear minima . In contrast , optimising directly using the calculated mse qb * was found to be computationally problematic . Let P 0 T and Q 0 T be partitioned into training sets Pa and Qa and test sets Pb and Qb . The training set is used to optimise the mapping from the real photon stream to the estimated one . The test set ensures the designed mapping properly captures the overall relationship and is neither over-trained nor under-trained . The training stage optimises a constant ζ = m ∫ Q a d t ∫ P a d t where the integral is over the time frame of the training set and m is the varying parameter of interest . Here ζ is interpreted as a multiple of the average charge integral of a single photon QB response . The estimated time of the nth photon in the testing stage occurs at the first time t: ∫Qb dt > nζ , with the integral over the test data set . The method fires photons every time the integrated QB signal crosses an optimised threshold , ζ . Fig 3 shows the spike metric optimisation curves , which have a clear minimum at m ≈ 1 . The best mse qb * values were obtained when the shift cost of Dspike goes to 0 . At this setting the metric gives the difference in the photon count between the real and estimated stream . This can be confirmed by defining a photon count cost on the training set as μ = | ∫ P a d t - ∫ P ^ a d t | . Here ∫Pa dt = 〈Qa〉−1∫Qa dt is the integral of the true photon stream from the training set and ∫P^adt=m〈 Qa 〉−1∫Qadt is the integral of the estimated photon stream using the integrate-fire method with parameter m . The quantity 〈Qa〉 = ζm−1 is the average QB charge per photon , from the training set . Substituting these expressions: μ = ( 1 − m ) mζ−1∫Qa dt . This is minimised when m = 1 for any QB set . This optimisation will therefore produce good results as long as the QB charge per photon is similar in the training and test sets . This condition is guaranteed for sufficiently large test and training sets . Thus , using the optimal m = 1 , the integrate-fire converts some input photon stream into an estimated one every time ∫Qb dt crosses an integer multiple of 〈Qa〉 . The resulting stream is then fed into the Snyder filter to produce x ^ qb * ( t ) . The use of a perfect integrator ( with m = 1 ) may suggest that errors from previously processed QBs ( due to their area deviating from average ) would persist with observation time and influence the signalling of upcoming QBs . Often in neuronal modelling this error memory problem is solved by adding a term that continuously drains some of the accumulated charge . This makes the integrator ‘leaky’ [39] . However , given the variation of QB area around its mean size it appeared that these memory errors negated each other and lead to the inference of the correct number of photons . Since training with Dspike indicated that it was more important to get the number than the position of the inferred photons correct; perfect integration worked for this problem . It is not clear that making the scheme leaky would improve performance . If used naively , for example , then losing charge would lead to less effective photons being signalled ( false negatives ) which could deteriorate performance . However , it could improve performance if m and the leaky time constant , ϵ , are jointly optimised during training so that the number of photons signalled remains the same as in the perfect integrator case . The value of ϵ might then help get the positions of estimated photons closer to those of the real stream . This could only strengthen the following results by tightening the MSE upper bound on the true mseqb . The integrate-fire-Snyder scheme was compared to standard supervised machine learning techniques ( Gaussian Processes [41] ) and optimal linear filtering methods ( least squares Finite impulse response [42] ) which directly estimated a x ^ qb * from the same Q 0 T . The former attempted to fit a covariance function while the latter minimised the square error on a series of filter coefficients . Both methods used variable window sizes ( also known as fading memory , this describes how much of the input history affects the output ) to get the best performance . These methods worked directly on the QB data without estimating photons . The optimal window size for both schemes was found to be about 5 ( 2 α ) . A representative simulation of the resulting MSE upper bounds from these techniques is given in Fig 4 . The integrate-fire achieved a lower mse qb * than all of these machine learning schemes . Thus , it appears that estimating a photon stream and applying the Snyder filter leads to appreciably better results than methods which directly estimate x ( t ) from the QB waveform . This figure also shows the relationship between the integrate-fire estimator and other techniques which also estimated photons for use with a Snyder filter . The ‘Pure threshold’ scheme produces estimated photons whenever a simple electrical signal threshold is exceeded and the ‘Gradient’ one does so when a forward and backward gradient shift is observed on the QB waveform . The integrate-fire outperformed the threshold scheme but had comparable MSE to the gradient one . However , the gradient scheme is not a robust technique , making integrate-fire the best overall choice . An overview of the general integrate-fire-Snyder estimation process is given in Fig 5 . To illustrate the importance of using non-linear filtering techniques for this analysis , consider the optimal ( MMSE ) linear estimate given P 0 T , denoted x ^ ph l , with distortion mse ph l . For the interrupted model , this linear MMSE is obtained by applying a continuity approximation to the system state dynamics [43] [44] . This leads to a differential equation with steady state solution ( see Supplement S1 Text ) : mse ph l = 1 β ( 1 + β 2 - 1 ) ( 3 ) The central importance of the non-dimensional parameter β is now obvious . Comparison of the optimal linear estimate given true photons , mse ph l , the optimal non-linear estimate given the true photons , mseph , and the integrate-fire-Snyder estimate ( at several γ ) from the resulting bumps , mse qb * , are given in Fig 6 . The significant difference between mse ph l and mseph indicates that some of the MSE is being contributed by the inefficiency of the linear filter as an estimator ( since it does not fully model the system dynamics ) . This additional error can likely cloud or misrepresent the true relative impact of different noise sources . The fact that mse ph l is above mse qb * at several γ suggests that these approximations reduce the ability to resolve the more subtle contributions of cascade variability . Consequently , linearity or continuity approximations can seriously distort one’s view of phototransduction noise performance . As concluded above , keeping analysis exact and non-linear is necessary to properly resolve the contributions of various cascade noise sources . Comparing the integrate-fire-Snyder results with filtering solutions that only feature specific components of the cascade noise should achieve this resolution . Such analysis allows one to identify the major ( independent ) intrinsic noise sources clearly . These mainly include QB shape ( width and height ) variation and latency noise ( see Introduction ) . The impact of shape noise was assessed by elimination and the application of the integrate-fire-Snyder to a deterministic version of the Nikolic model that featured QBs of fixed size and shape . More detail on the differences between stochastic and deterministic Nikolic models is given in Supplement S4 Text . QB response latency can be decomposed into a ( deterministic ) mean delay and a jitter term describing variation around the mean . The effect of the total QB latency was isolated by applying the Snyder filter to photons that have been empirically delayed according to the physiological latency distribution embedded within the Nikolic model . The empirically delayed MSE upper bounds are given in Supplements S4 and S5 Figs . To quantify the relative impact of jitter and mean delay a Snyder filter that optimally reconstructs light intensity given a deterministically delayed photon stream was developed . If the original photon stream up to time t is P 0 t , then denote the delayed stream P 0 t - τ as M 0 t with delay τ ms . The optimal filtering problem focuses on finding the posterior probability vector P ( x ( t ) | M 0 t ) . Let the conditional estimate x ^ d ( t ) = E [ x ( t ) | M 0 t ] . Since introduction of the delay does not alter the inter-event spacing of the observed events then the usual Snyder filter can be applied to obtain the estimate x ^ ( t - τ ) using M 0 t as the observed process . Transforming x ^ ( t - τ ) into the desired optimal estimate x ^ d ( t ) is possible using P ( x ( t - τ ) | M 0 t ) , which is produced by the original Snyder filter . Since there is no data ( no observed events ) the evolution from the delayed to the desired posterior is achieved using the Chapman-Kolmogorov equations , with initial condition t − τ . R describes the transition rates of the x ( t ) Markov chain ( see Supplement S1 Text for more details ) . This gives P ( x ( t ) | M 0 t ) = P ( x ( t - τ ) | M 0 t ) e τ R which is used to calculate the expectation x ^ d ( t ) . The solution for the interrupted model between and at event times is provided in eqs 4 and 5 respectively . x ^ d ( t ) = 1 2 ( 1 - e - 2 τ k ) + x ^ ( t - τ ) e - 2 τ k for 0 ≤ t ≤ s + ( 4 ) x ^ d ( s + ) = 1 2 ( 1 + e - 2 τ k ) < 1 if τ > 0 ( 5 ) This solution is optimal and not an upper bound like the integrate-fire-Snyder mse qb * curves . When compared to the standard Snyder solution of eqs ( 1 ) and ( 2 ) , the delayed solution features a higher minima and a lower maxima . This shows the increased uncertainty in this problem as the solution lies closer to the uninformed estimator E [ x ( t ) ] = 0 . 5 . Observe that lim β → 0 x ^ d ( t ) = lim β → 0 x ^ ph ( t ) = 0 . 5 . The main results from preceding sections are combined into Fig 7 . Results are averaged over 10 independent photon-QB streams of 8000 photons long with error bars showing the maximum and minimum MSE across the runs . Flickering models at γ = [5 , 10 , 20 , 30] are investigated . At γ > 30 the relative curve behaviour is unchanged and at γ < 5 the model switches too fast for sensible inference ( it is beyond the bandwidth of the phototransduction process ) . The β range ( 1-200 ) was chosen similarly . Lower β means photons are produced too slowly relative to light switches making inference pointless ( hence why MSE settles near 0 . 25 ) while higher β is not any further informative . The ‘Snyder , photon noise’ curves give mseph , or the noise floor of the model while ‘Snyder , delay’ refers to msed ( photon noise and a deterministic delay ) . The remaining curves are integrate-fire-Snyder upper bounds . The ‘all noise’ ones correspond to mse qb * ( fully stochastic Nikolic model ) while the ‘delay’ ones refer to the integrate-fire-Snyder applied to QBs from the deterministic Nikolic model . This last curve set is used to confirm the subsequent conclusions and involves applying the integrate-fire-Snyder algorithm to the Nikolic model with all its variability switched off . Consequently , the QB stream has deterministic bump shapes and sizes and there is no latency jitter . As a result , only the mean of the stochastic delay acts in this model . These 4 sets of curves provide an overall perspective of noise performance with relative light intensity and illustrate the main conclusions that will follow . At low β all curves converge , while at higher β the cascade noise separates the ‘Snyder , photon noise’ ones from the rest . The ‘Snyder , photon noise’ curves are exactly the MMSEs achievable at the front of the cascade , and cannot be bettered . The ‘Snyder , delay’ curves are hard lower bounds , in the sense that they provide a lower limit on the remaining error when all sources of cascade variability are removed . Since this setting also applies to the deterministic Nikolic model then the ‘Integrate-fire , delay’ is an upper bound on the impact of the mean QB delay . The precise MMSE curve under this setting would lie between these bounds . The ‘Integrate-fire , all noise’ curves are an upper bound on all photon and cascade based distortion . All 3 of these sets of bounds remain close at higher β suggesting mean delay is the key intrinsic noise component . The relative behaviour of the MSE curves with β , remains unchanged as γ changes . Higher MSE values are observed at lower γ as a relatively faster flickering model ( less QBs are available per state ) is more difficult to estimate . In all simulations τ = 43 . 3 ms is used to match the mean delay from the Nikolic QB latency distribution . This analysis was also repeated using more complex and realistic 16 state bimodal light models . Similar results were obtained and the relative behaviour of the MSE curves found to be consistent with the above descriptions ( see Supplements S2 and S3 Text , S4 and S5 Figs ) . In the bimodal model higher frequency small amplitude fluctuations are superimposed on a lower frequency flickering ( modal switching ) between high and low light levels ( the modes ) . This gives a characteristic 1 f like power spectrum behaviour over a frequency , f , range of almost two decades . The precise exponent was f−1 . 23 . The integrate-fire-Snyder algorithm is a surprisingly simple yet effective way of reconstructing light intensity from the bumps . Given these attributes , one wonders if it could potentially be implemented within a biologically sensible neuronal network framework . The algorithm is composed of two main parts; i ) an integrate-fire estimator that gives P ^ 0 t and ii ) the Snyder filter that leads to the state estimate x ^ qb * ( t ) . The integrate-fire part involves the accumulation of QB charge until some spike optimised threshold , ζ is crossed . Once ζ is exceeded a point event is emitted to indicate a suspected photon . This description is standard in neuronal modelling with the build up of charge usually leading to the release of an action potential or spike [39] . While the integrate-fire mechanism is biologically implementable , the optimisation of its threshold ( which depends on the parameter m ) requires a supervisory learning technique that uses true photon times . In reality , the invertebrate visual system does not have access to this data . Unsupervised learning procedures may therefore be necessary . If light intensity is low , unsupervised integrate-fire training should be easily realisable . In this regime QBs are distinct and far apart so the integration threshold is obvious . In the high intensity regime where QBs merge together , a more sophisticated way of reducing the data into effective photons is needed . This would likely involve finding clusters ( which code for sets of QBs ) within the data . Standard algorithms exist for extracting clusters and structure from input data . These can be implemented in modular , Hebbian learning networks [45] which are meant to resemble actual spiking neuronal computational behaviours . Hence , it should be possible to train the integrate-fire scheme in a biologically sensible manner , especially as only a single parameter must be learnt . However , given that threshold optimisation found that the integrate-fire should only signal photons every time the average bump charge is crossed , training may not be necessary . The average QB area could likely be a parameter that is naturally encoded in the cascade biology . The second part of the algorithm uses standard Snyder filtering . The filter solution involves an instantaneous update at the event time which serves as the initial condition for a decaying inter-event trajectory . For all Markov chain state models it can be shown that the inter-event Snyder solution can be expressed as a linear equation set in un-normalised probabilities that must then be normalised [46] ( see Supplement S2 Text ) . Using this formulation the inter-event trajectory can be seen as only involving weighted combinations of exponentially decaying functions . Bobrowski et al [32] provided strong evidence that these linear Snyder computations can be achieved with a recurrent neural network that is physiologically realisable and which uses simple operations such as weighted sums . This involves a graphical structure that treats incoming spikes as inputs from a sensory layer and encodes the causal posterior components , with another layer that has weights derived from the Markov state transition rates . Having a neuronal population code for posterior probabilities has already been shown to be biologically plausible [47] . Further , it is known that recurrent networks of integrate-fire neurons are able to represent distributions given noisy inputs [48] . As a result , all parts of the complete integrate-fire-Snyder estimation algorithm possess the potential for possible biological implementation .
The above analyses show that as light intensity increases , the dominant noise source transitions from being extrinsic to intrinsic . Furthermore , among intrinsic noise sources , it seems that the performance deterioration caused by QB shape variability is small compared to that introduced by the mean cascade delay . As a result , the relative magnitude of the mean QB arrival time , given a photon , to the mean photon inter-arrival time is a key indicator of noise performance . Such an event timing based description of noise fits well into the Poisson channel framework in which this estimation problem lies . The relatively small impact of QB shape noise is particularly significant . Previous research using signal to noise ratios found QB amplitude variation responsible for limiting cascade reliability at all frequencies [10] . This diametrically opposite conclusion is likely evidence of why it is important to get the performance metric right [49] . Furthermore , work based on the Poisson-variance approach [9] predicted that at least half of the total noise ( extrinsic and intrinsic ) was always due to the cascade . This directly belies the extrinsic-intrinsic noise transition described here and elucidates the importance of maintaining a point process approach which properly accounts for signal causality . The msed curve exactly describes the optimal distortion given photon noise and only a fixed cascade delay . It therefore provides a distortion lower bound by describing a hypothetical scenario in which all ( independent ) sources of variability have been removed from the cascade . The mse qb * curve is derived from the integrate-fire-Snyder algorithm and upper bounds the unknown MMSE given the QB stream . Simulations found that these two bounds were generally in close agreement . This suggests that the integrate-fire-Snyder is likely a good means of converting QBs into estimated photons , over a wide range of intensities . In fact , as shown in Fig 4 , the algorithm outperformed several standard machine learning techniques , achieving lower mse qb * values on the same data . This not only reinforces the integrate-fire-Snyder as a good estimation method but also suggests that , generally , it may be beneficial to reduce a QB stream to an estimated photon stream before performing inference . This seems reasonable as QBs , being responses to individual photons , must have a discrete information structure embedded within their noisy waveforms . Note that , since the integrate-fire-Snyder makes few assumptions on its input and output signals , it may also have applications in more general systems with discrete-sum informational structures . Work by van Steveninck and Laughlin [50] treated the transduction process as a filtering operation and postulated that one could recover the exact photon times by inverting the filter , subject to fundamental performance limitations . The integrate-fire-Snyder algorithm fits exactly into this framework , albeit in a non-linear point process setting , and thus provides a good answer to this postulation . The algorithm performance suggests that generating a good photon event estimation is sufficient to achieve good mse qb * and that more complex techniques are unnecessary . However , the use of simplifying techniques that do not fully model process dynamics can also be misleading . There exists a set of ( α , k ) for which the noisy integrate-fire-Snyder estimate , mse qb * , outperforms the linear MMSE estimate given the true photon data , mse ph l ( see Fig 6 ) . Any linear estimator using the QB data must do even worse . Performing the analysis under a linear or continuity approximation therefore gives a wrong impression of the noise floor for the system ( since the MSE with only photon noise is so high ) and will not resolve the more subtle noise source contributions . This results in a spurious picture of phototransduction noise performance . For a more detailed analysis of the dangers of such approximations , when applied to systems which are naturally discrete and causal see Parag and Vinnicombe [51] [30] . For all the stimuli investigated , photon noise was found to be limiting at low light intensities . This is likely due to the photon inter-arrival time being much larger than the delays and widths of QBs . Given this dominance , one can easily convert QBs into estimated photons using naive threshold based methods ( see the ‘pure threshold’ curve of Fig 4 ) , as the cascade deterioration has negligible impact on inference . Further , as mse qb * ≈ mse ph in this region , the true MMSE given QBs can be well approximated with the exact , standard Snyder solution . As intensity increases all MSE curves fall due to the availability of more information . However , cascade based noise becomes visible as the mse qb * curve significantly separates from the mseph one . This divergence increases with normalised intensity , β . At very high intensities photon noise contributes almost nothing and mseph → 0 as β → ∞ . A major part of this work involved trying to infer the relative contributions of the various intrinsic noise sources . As noted previously , intrinsic noise is composed of dark noise ( false positive QBs ) , QB shape noise ( amplitude and width variability ) , QB latency ( mean delay and jitter ) and a quantum capture efficiency , QE < 1 ( an analogue to false negative QBs ) . These components are independent of each other and can therefore be treated in isolation . Dark noise is negligible under the conditions of this work and was excluded from analysis . QE is usually high in real invertebrate visual systems and it can be treated as an additional noise component on top of the cascade noise ( which involves QB shape and latency ) . Moreover , simulations found results to be robust to the loss of up to 1 3 of all input photons ( QE = 0 . 66 ) ( see Supplements S3 Text , S4 and S5 Figs ) . The integrate-fire-Snyder adapted to lower QEs by simply reducing its trained firing threshold ζ , which encodes how much QB charge is believed to represent a photon . As a result , QE can be largely ignored . The relative noise problem in the medium-high intensity regime was therefore reduced to one of disentangling the relative impact of QB shape and latency . This was initially investigated by comparing the mse qb * bounds achievable with different noise components excluded from the Nikolic model . If shape noise is removed then the observed QB stream is equivalent to the input photon stream distorted according to the physiological QB latency distribution . This stream was described as empirically delayed . The optimal filter for empirically delayed photons is computationally intractable . However , it was found that applying the standard Snyder filter to the empirically delayed photons gave an MSE upper bound that was in close agreement with the fully noisy mse qb * ( see Supplements S3 Text , S4 and S5 Figs ) . This hinted that shape noise could be less important than latency . To test this hypothesis it was necessary to construct complementary lower bounds on achievable noise performance . The QB latency profile can be well described by a normal distribution that matched its mean [52] . Since normal distributions are parameterised solely in terms of their mean and variance , QB latency may then be considered as essentially and independently composed of a mean delay and jitter . Separating the impact of each component therefore makes sense . The mean delay was analysed , because i ) unlike jitter it can be examined within an exact inference framework , and ii ) fixed delays or dead times have a clear and important impact on causally constrained real time systems such as vision . Additionally , there is some evidence in the literature for dead times and fixed delays being both irreducible and critical to high intensity phototransduction noise performance [52] . This motivated the development of the optimal deterministically delayed Snyder filter , which yielded the msed curve . This curve provided a lower bound on the remaining error when all cascade variability , including jitter , was removed . A close correspondence between msed and mse qb * was observed across all the light models investigated . This confirmed QB latency as the dominant cascade noise source and suggested that the mean delay in generating a QB from a photon input is critical to cascade noise performance . In fact , when plotted , the most visible difference between x ^ ph ( t ) and x ^ qb * ( t ) is a time shift which is of the order of the mean delay . For an illustration of this difference on a more complex light model see Supplement S6 Fig . The cascade therefore essentially encodes discrete events as delayed electrical depolarisations . As a result , by elimination , the actual shape of the latency distribution is not very important and the variable shape of the QB largely insignificant . This conclusion was further investigated by applying the integrate-fire-Snyder to the Nikolic model with all forms of variability turned off . At this setting QBs are of deterministic shape and size , and only the mean delay is acting on the cascade . The resulting MSE curves lay close to and between msed and mse qb * . This correspondence validated the performance deteriorating dominance of mean delay . Thus to a first approximation , the complex set of cascade reactions can be replaced with a pure delay on the photon inputs . This conclusion contradicts the work in [9] and reiterates the importance of maintaining a causal , discrete approach . This is especially the case here since pure delays do not affect the common ( acausal ) mutual information and would largely be neglected in such analyses . However , this work asserts that such delays are important , especially as they may limit the ability of an invertebrate to respond rapidly to stimuli . The dominance of QB delay on cascade noise means that QB shape can be optimised for other performance goals , such as achieving a dynamic range that maximises the input amplitude representation , without affecting cascade accuracy . This shape-latency decoupling allows for more flexible cascade functionality and may underly why experiments have found the QB waveform and latency to be uncorrelated [16] . Cascade delay appears to be the single largest intrinsic noise component . One may therefore wonder why nature has not further optimised phototransduction to reduce its impact . Reducing delay would not only improve accuracy but also increase visual bandwidth . According to Eckert and Zeil [53] faster phototransduction requires more mitochondria and hence is much more energy intensive . Consequently , improving latency will only be feasible if the organism is willing to devote more energy to visual processes . This explains why flies that must perform more demanding visual tasks , such as Coenosia , have faster cascades than Drosophila [54] [1] . The dominant effect of cascade delay also clarifies why visual light adaptation , which involves the production of faster QBs , is helpful at high intensities . In this regime the latency becomes much more critically limiting and its impact must therefore be , at least partly , countered [14] . A key feature of the adaptation response is a reduced mean delay but a mostly fixed jitter [52] . This seems to support the idea that mean delay is the major source of cascade noise . Moreover , even if energy constraints were not limiting , significantly improving QB latency and mean delay would still be unlikely . The latency relates to the time taken for a sufficient amount of G proteins and phospholipase C molecules to become activated such that the threshold for opening the TRP channels and generating a normal QB is achieved [15] . A noteworthy reduction in cascade latency would require a reduction in the threshold for opening the TRP channels . However , this would increase the sensitivity of the cascade to spontaneous phospoholipase C and G protein activations which would lead to an increase in dark noise [55] . Thus , mechanistically reducing latency to improve cascade accuracy may simply result in a trade between the limiting noise source and possibly lead to no overall improvement . This paper has outlined a useful scheme for extracting data from QBs and used this to characterise the main sources of noise in invertebrate phototransduction . By making use of the photon counting nature of the visual system , and point process theory , a consistent and accurate measure of relative noise was achieved that extends the stimuli reconstruction methodology of Bobrowski et al [32] and clarifies the until now contradictory results between the work of van-Steveninck and Bialek [10] and those of Lillywhite [3] and Laughlin [9] . Furthermore , this work pinpointed mean delay as the key cascade noise source and emphasised the dangers of using continuity approximations for inherently discrete random systems . Lastly this research fulfils the request of Grewe et al [27] for a sensible metric ( causal MMSE ) which can give a ‘complete description of the system [photoreceptor] performance’ . | The invertebrate phototransduction system captures and converts environmental light inputs into electrical signals for use in later visual processing . Consequently , one would expect it to be optimised in some way to ensure that only a minimal amount of environmental information is lost during conversion . Confirming this requires an understanding and quantification of the performance limiting noise sources . Photons , which are inherently random and discrete , introduce extrinsic noise . The phototransduction cascade , which converts photons into electrical bumps possessing non-deterministic shapes and latencies , contributes intrinsic noise . Previous work on characterising the relative impact of all these sources did not account for the discrete , causal , point process nature of the problem and thus results were often inconclusive . Here we use non-linear Poisson process filtering to show that photon noise is dominant at low light intensity and cascade noise limiting at high intensity . Further , our analysis reveals that mean bump delay is the most deleterious aspect of the intrinsic noise . Our work emphasises a new approach to assessing sensory noise and provides the first complete description and evaluation of the relative impact of noise in phototransduction that does not rely on continuity , linearity or Gaussian approximations . | [
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
] | [
"invertebrates",
"markov",
"models",
"particle",
"physics",
"applied",
"mathematics",
"social",
"sciences",
"light",
"neuroscience",
"animals",
"electromagnetic",
"radiation",
"simulation",
"and",
"modeling",
"algorithms",
"photons",
"mathematics",
"vision",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"probability",
"theory",
"physics",
"psychology",
"eukaryota",
"cell",
"biology",
"microvilli",
"elementary",
"particles",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"sensory",
"perception",
"organisms",
"phototransduction"
] | 2017 | Point process analysis of noise in early invertebrate vision |
The parasite Entamoeba histolytica is the etiological agent of amoebiasis and phagocytosis plays a key role in virulence of this organism . Signaling pathways involved in activation of cytoskeletal dynamics required for phagocytosis remain to be elucidated . Phagocytosis is initiated with sequential recruitment of EhC2PK , EhCaBP1 , EhCaBP3 and an atypical kinase EhAK1 after particle attachment . Here we show that EhARPC1 , an essential subunit of the actin branching complex Arp 2/3 is recruited to the phagocytic initiation sites by EhAK1 . Imaging , expression knockdown of different molecules and pull down experiments suggest that EhARPC1 interacts with EhAK1 and that it is required during initiation of phagocytosis and phagosome formation . Moreover , recruitment of EhARPC2 at the phagocytosis initiation by EhAK1 is also observed , indicating that the Arp 2/3 complex is recruited . In conclusion , these results suggests a novel mechanism of recruitment of Arp 2/3 complex during phagocytosis in E . histolytica .
Phagocytosis plays a critical role in invasion and pathogenesis of the parasite Entamoeba histolytica , the causative agent of amoebiasis , and a major cause of morbidity and mortality in developing countries . Phagocytosis is an essential process in E . histolytica as blocking this process leads to an inhibition of cell proliferation and pathogenicity [1 , 2] . The pathways coupling phagocytosis initiation signals to actin dynamics have been studied quite extensively in some model systems [3–5] . In mammalian and other systems a number of proteins that bind and regulate actin nucleation , polymerization , bundling , and branching have been identified and characterized . Arp2/3 complex is one of the main group of molecules required for actin dynamics . It comprises of seven subunits , Actin related protein 2 ( Arp2 , 44KDa ) , Actin related protein 3 ( Arp3 , 47KDa ) , ARPC1 ( 40KDa ) , ARPC2 ( 35KDa ) , ARPC3 ( 21KDa ) , ARPC4 ( 20KDa ) and ARPC5 ( 16KDa ) . There are multiple ways by which Arp 2/3 complex is recruited at the site of actin dynamics . Some of the examples are , interaction with VCA domain of activated NPFs ( nucleation protein factors ) [6] , direct binding of Arp2/3 complex to vinculin ( an integrin associated protein ) during cell migration at the sites of integrin-mediated adhesions and membrane protrusions and binding of F-actin [7] , direct binding to cortical actin associated protein ( cortactin ) [8] and recruitment through WAVE2 complex during T-cell activation [9] . Though NPFs , such as WAVE , WASP and WASH are thought to be involved in activation of Arp 2/3 complex during phagocytosis , in some situations ( e . g . uptake of Yersinia pseudoparatuberculosis ) the pathway through N-WASP is by passed and Arp 2/3 complex is directly activated by Rac-1[10] . We can conclude from this discussion that there are multiple ways by which Arp 2/3 complex gets activated in different systems . However , there is not much information available about the pathway ( s ) and regulatory steps during phagocytosis in E . histolytica . Two approaches were used to understand the mechanism of phagocytosis in E . histolytica . In one approach sequence similarity searches were used to identify putative homologous proteins that are known to participate in phagocytosis in other systems [11] . In the second approach , phagosome proteome of E . histolytica was analysed using mass spectrometry [12–14] . A summary of the published results is shown in the S1 Table . Sequence similarity analysis identified all subunits of Arp 2/3 complex in E . histolytica , and a few of the proteins known to be involved in recruitment and activation of this complex during actin dynamics [11] . Some of these are; a homolog of WASP protein containing a conserved VCA domain , a homolog of MIM which also contains a VCA domain , eight formins , filamins and alpha actinins . Not all of these proteins were consistently found in phagosome proteome . For example , WASH homolog was identified in only one of the experiments . Moreover , it is not clear if these homologs carry out the same function in E . histolytica as is known as in other systems , since experimental evidence to this effect is still not available in E . histolytica . Since many proteins involved in initiation or scission of phagosomes are lost either before or soon after phagosome formation ( for example EhCaBP1 [15] ) , the lack of participation of a molecule in phagocytosis cannot be inferred based only on its absence in phagosome proteome . Some of the actin modulating proteins have been studied at functional level in E . histolytica . Foremost among these are Formins . Formin 1 and 2 were found to colocalize with actin in pseudopodia , cell division sites and in both pinocytic and phagocytic vesicles suggesting that these may be involved in cell division , pinocytosis and phagocytosis [16] . EhFormin1 was also shown to modulate actin polymerization through its formin homology 2 domain . EhRho1 is thought to regulate directly activating this protein [17] . In general coactosins belong to ADF/cofilin family and are known as F-actin severing and depolymerising proteins [18] . On the contrary , Ehcoactosin was found to stabilize F-actin [19] . EhFLN ( previously known as EhABP-120 ) , a filamin protein is recruited at the plasma membrane via PI ( 3 ) P and phosphatidic acid ( PA ) . When the d100 region of EhFLN required for binding to PA , was over-expressed it increased the amoebic motility suggesting its role in actin dynamics [20] . In addition other actin binding proteins such as 16 kDa and EhNCABP166 have been partially characterized . On down regulation of expression of the former , an inhibition in cell motility was observed [21] . EhNCABP166 is present in both cytosol and nucleus and is thought to be involved in phagocytosis and cell motility [22 , 23] . It is clear from the discussion that there are a number of likely components of actin dynamics machinery . However , the participation of specific components in different cellular processes has not been worked out in detail . We have been investigating molecular mechanisms that are involved in the initiation of phagocytosis using red blood cell ( RBC ) uptake as a system . Our major effort has been to identify molecules that are needed for initiation of a protein complex at the site of particle attachment leading to phagocytic cup formation , and channeling actin dynamics for progression of phagocytic cups to phagosomes . The nature of the primary signal generated after the attachment has not been elucidated so far , though there is some evidence to suggest that the GPI anchor present in Gal/GalNAc lectin complex may be the key transducing component [24 , 25] . Some of the other proteins shown to be involved in phagocytosis in E . histolytica are phagosome-associated transmembrane kinase [26] , serine-rich E . histolytica protein [27] , EhPAK [28] , and EhCaBP5 [29] . It is not clear how and in which stages these molecules participate in the phagocytic process . For example , cell surface molecules PATMK and SREHP are suggested to be involved in erythrophagocytosis but it is not clear whether they are the receptor for particles , or are the initiator molecules required for transducing signal immediately after particle attachment . Our studies have shown that the primary signal helps to enrich a C2 domain protein kinase , EhC2PK , at RBC-attachment sites [30] . EhC2PK recruits calcium binding protein EhCaBP1 [15 , 30] , which in turn brings in atypical kinase EhAK1 at the site of attachment [31] . Another calcium binding protein EhCaBP3 is independently recruited forming a multimeric complex [32] . All these proteins have different roles during progression of phagocytic cups to phagosomes . While both EhC2PK and EhCaBP1 leave phagocytic cups before closure of cups , EhAK1 is found in just closed cups before scission and only EhCaBP3 is present in the phagosomes after scission ( mature phagosomes ) [15 , 30–32] . Nearly twenty proteins were found to interact with EhCaBP1 as determined by affinity chromatography and mass spectrometry [30] . Among these only EhARPC1 , a member of the Arp2/3 complex , was found to be a potential molecule that may couple EhCaBP1-EhC2PK mediated signaling with actin dynamics . Arp2/3 complex proteins EhARPC1 and EhARPC2 were also found when EhAK1 binding proteins were analyzed by mass spectrometry [31] . Absence of other actin modulating proteins in these two pull down experiments suggests that EhARPC1 and EhARPC2 may have important role in the phagocytic pathway mediated by EhCaBP1-EhC2PK . In this report we describe the role of EhARPC1 , one of the subunits of Arp2/3 complex , in the phagocytosis of RBC in E . histolytica . Our results show that it is recruited to the phagocytic cups through EhAK1 and participates in phagocytosis . We also show that another subunit of Arp 2/3 complex EhARPC2 is recruited at the cups , suggesting the presence of Arp 2/3 complex at the phagocytic site . Our results provide a basis for coupling of actin dynamics to the initial signaling system activated on attachment of a RBC to the cell membrane . The pathway described by us is novel and has not been seen in any other system so far .
Sequence analysis of EhARPC1 showed maximum identity with p41 subunit of Arp2/3 complex from a number of species . Identity ranged from 32% with Saccharomyces cerevisiae to 41% with human . Multiple alignment of amino acid sequences of homologs from different species displayed conservation across the full length of the protein , higher towards N-terminal region than C-terminal ( S1A Fig ) . The p41 subunit of Arp 2/3 complex of all organisms contain conserved WD40 repeats . Therefore , it was not surprising to find probable WD40 repeats ( amino acids 50 to 181 ) in the amoebic protein as well . However , the WD40 repeat containing region in the amoebic protein was longer than the corresponding region of human protein ( 48 to 89 amino acids ) or yeast ( 51 to 92 amino acids ) but was similar to that of Dictyostelium discoideum protein ( 50–180 amino acids ) ( S1B Fig ) . Moreover , “Arm” region ( C-terminal sequence ) in p41 subunit of S . cerevisiae that is required for binding WASP , is absent in EhARPC1 [33] ( S1C Fig ) . It appears from sequence analysis that EhARPC1 may have diverged functionally from the human or the yeast homologs . Arp2/3 complex comprises of seven subunits and EhARPC1 is one of its members , as mentioned before sequence analysis has suggested presence of all seven subunits in E . histolytica only EhArp2 and EhArp3 showed maximum identity with Arp2 and Arp3 of D . discoideum and EhARPC5 , was found to be maximally diverged from that of other species ( S2 Table ) . We carried out immunofluorescence imaging for determining cellular distribution of EhARPC1 using antibody raised against recombinant protein ( specificity of the antibody is shown in S2A Fig ) . As a membrane marker , antibody against the E . histolytica pan membrane marker EhTMKB1-9 was used [34] . F-actin was visualised using TRITC-phalloidin . We found EhARPC1 in the cytoplasm , some parts of the membrane and in F-actin rich areas ( Fig 1A ) . In order to investigate if EhARPC1 accumulates in some sites more than others , a quantitative analysis of the images was carried out . We estimated strengths of co-localization using fluorescent signals of a pair of stains analysed using Pearson’s correlation coefficient ( PCC ) ( Fig 1B ) . We found preferential localization of EhARPC1 in actin rich areas ( r = 0 . 898 ) and much less in plasma membrane ( r = 0 . 483 ) . A considerable amount of fluorescent signal of EhARPC1 originated from the cytoplasm . The results suggest that EhARPC1 is likely to be a cytoplasmic protein that gets recruited at F-actin-rich sites . Localization of Arp2/3 complex in model systems has also revealed preferential recruitment at the leading edges of lamellipodia in mammalian cells [35] and at the actin patches in S . cerevisiae [36] . The association of EhARPC1 with actin was further validated using E . histolytica cells expressing GFP-EhARPC1 . The distribution of GFP fluorescence was similar to that seen by antibody staining ( Fig 1C ) . Fluorescence signals strongly localized with F-actin enriched areas ( r = 0 . 886 ) , and a significant amount of fluorescence came from the cytoplasm as revealed by analysis of distribution across the whole cell ( S2B and S2C Fig ) . The results suggest that GFP tagged protein behaved in the same way as native protein , residing mainly in cytoplasm but getting recruited to F-actin rich areas ( Fig 1C ) . Binding of phagocytic ligands to the cell surface , triggers directed reorganization of cytoskeleton underneath the binding site , resulting in pseudopod formation . Pseudopods extend around the particle , fuse and then separate out from the membrane to form phagosomes . In order to investigate if EhARPC1 may be involved in pseudopod , and subsequent phagosome formation , we first measured enrichment of EhARPC1 at the pseudopods by live cell imaging of E . histolytica trophozoites expressing GFP-EhARPC1 . We clearly saw enrichment of EhARPC1 at the moving ( or leading ) edge of amoebae , marked by arrowhead in images ( Fig 2A and S1 Movie ) . This was confirmed by quantitative analysis of the images ( Fig 2B ) . The time taken to complete a psuedopod formation and retraction was found to be 90ms± 20ms , indicating that the process is extremely rapid . We further investigated the involvement of EhARPC1 in phagocytosis using a number of different approaches . Imaging was used to localize EhARPC1 during uptake of RBCs by E . histolytica . Cells were incubated for different times with RBCs , followed by fixing and staining with indicated antibodies . Images were analysed by quantifying the events in 25 cells ( Fig 2C ) . The panel in Fig 2D shows representative images of cells displaying different stages of phagocytosis , such as early phagocytic cup ( marked by 1 ) , late phagocytic cup ( marked by 2 ) , closure of cups before scission ( marked by 3 ) and mature phagosome ( marked by 4 ) . It is clear from the figure that EhARPC1 is recruited early after initiation of phagocytic cups and it stays till cups close , but have not yet undergone scission . It is not present in phagosomes , that is , it leaves during the process of scission ( Fig 2D ) . Similar patterns were visualized when experiments were carried out using fluorescent labeled RBCs and GFP-EhARPC1 ( S2D and S2E Fig respectively ) . We also visualized RBC uptake in GFP-EhARPC1 expressing cells by time lapse imaging ( S2 Movie ) . Generally , a cycle of phagocytosis is completed within 240-260ms after attachment of RBC . The data is shown as snap shots of a complete cycle at intervals ( Fig 3A ) . Quantitative analysis of the data is shown in Fig 3B . GFP-EhARPC1 was present during cup formation and was found till the process of scission starts . It was not found in early phagosomes . We could clearly observe RBC attached to the surface of E . histolytica cells ( marked by a yellow colored arrow in DIC images ) . EhARPC1 was also found to involve in phagocytosis of other ligands , such as mammalian cells . This was visualized by observing enrichment of EhARPC1 during phagocytosis of Chinese hamster ovary cells ( CHO ) labeled with cell tracker blue dye ( Fig 4A ) , and by time lapse imaging ( snap shots Fig 4B and S3 Movie ) . EhARPC1 was observed at the phagocytic cups from the start of the phagocytosis till closure of phagocytic cups . Interestingly we observed that EhARPC1 was present just after the membrane fusion event but not when phagosome got separated from the membrane . The average time taken by the whole process was found to be 2s based on EhARPC1 enrichment . In order to understand the role of EhARPC1 in the context of some of the other molecules that have been identified as part of the phagocytosis pathway in E . histolytica ( EhCaBP1 , EhC2PK , EhCaBP3 , EhAK1 ) , pairwise staining was carried out and extent of co-localisation during phagocytosis was quantified using PCC ( Fig 5A and 5B ) . We found all five molecules ( actin , EhCaBP1 , EhCaBP3 , EhC2PK and EhAK1 ) in phagocytic cups along with EhARPC1 . However , both EhCaBP1 and EhC2PK were not present in cups just closed before scission from membrane , and EhCaBP3 was the only molecule present in mature phagosomes . Therefore , it appears that EhARPC1 behaves like EhAK1 in terms of its association with phagocytic machinery . Both of these molecules leave phagosomes before or immediately after scission takes place . Summary of all observations are shown schematically in Fig 5C . All these results suggest that EhAK1 and EhARPC1 may participate in a similar manner during phagocytosis . We further demonstrated the involvement of EhARPC1 in cytoskeleton dynamics by determining the effect of EhARPC1 down regulation on the rate of phagocytic cup formation , amoebic motility and proliferation . Down regulation of EhARPC1 expression was achieved by over expressing the gene in antisense orientation in a tetracycline-inducible manner [37–39] . The results are shown in Fig 6A . The level of down regulation achieved was 50% in the presence of 30 μg/ml tetracycline . On the other hand , EhARPC1 protein increased by 60% in cells expressing the gene in sense orientation ( Fig 6B ) . Phagocytosis of RBCs was measured in these cells using a colorimetric assay ( Fig 6C ) . All comparisons were made against cells carrying either the vector alone , or with the gene construct in the absence of tetracycline . When EhARPC1 was over expressed in the sense cell line there was an increase in RBC uptake by 30% in 10 min . However , it was reduced by 70% in the antisense cell line in presence of tetracycline . We stained these cells with phalloidin and with antibodies against EhARPC1 and representative pictures are shown in Fig 6D and images depicting population of downregulated cells is shown in S3A Fig . In antisense expressing cells phagocytic cups were visible only after 5 min of incubation with RBC . We rarely observed phagosomes in these cells . In comparison , many phagocytic cups were visible in cells over expressing EhARPC1 by 1 min . We carried out quantitative analysis of the images by observing 25 cells ( in triplicates ) and enumerated number of phagocytic cups and phagosomes in these cells . By 5 min , cups and phagosomes were found to be only 11% and 7% of the control cells in antisense cells respectively . The results clearly showed that compared with wild type cells the rates of both cup and phagosome formation were significantly reduced in cells expressing antisense EhARPC1 . On the other hand , cups and phagosomes increased by 60% and 40% respectively in cells over expressing EhARPC1 by 5 min ( Fig 6E and 6F ) . Similar results were obtained when CHO cells were used for phagocytosis assay . Phagocytic cup formation was reduced significantly in the antisense cell line ( S3B Fig ) . Further , cell motility and proliferation were also reduced in cells where EhARPC1 expression was down regulated by antisense RNA as compared to TOC vector alone in presence of tetracycline ( S4 and S5 movies respectively ) . The level of reduction observed in antisense cells in case of proliferation was 40% ( S3C Fig ) . From this data we can conclude that EhARPC1 is required for a number of process including phagocytosis in E . histolytica . EhARPC1 was identified as EhCaBP1-binding protein in an affinity screen , as previously mentioned . In order to validate binding of EhARPC1 to EhCaBP1 directly we incubated GST-tagged EhARPC1 with EhCaBP1 in the presence and absence of Ca2+ . Glutathione-Sepharose was used to pull down the complex and the presence of EhCaBP1 was determined by using a specific antibody . The result is shown in Fig 7A . No EhCaBP1 was found in the pull down material either in the presence or absence of Ca2+ . As a positive control we used GST-tagged EhC2PK which directly binds EhCaBP1 [15] , and could pull down EhCaBP1 both in the presence and absence of Ca2+ as expected . However , when we immunoprecipitated the complex from E . histolytica cell lysate using anti EhARPC1 antibody , we clearly observed EhCaBP1 in the pull down ( Fig 7B ) , indicating indirect interaction between the two proteins . We therefore tested if other EhCaBP1-binding proteins [30 , 31] could act as a bridge between EhCaBP1 and EhARPC1 . Recombinant EhAK1 and EhC2PK were used to test this in an in vitro pull down experiment . We incubated GST-tagged EhAK1 with full length His-tagged EhARPC1 and the pull down material was analysed by western blotting using anti-his antibody . A band corresponding to EhARPC1 was clearly seen in the pull-down ( Fig 7C ) . However , no pull-down of EhARPC1 was observed using GST-tagged EhC2PK in a similar experiment ( Fig 7C ) . The interaction of EhAK1 with EhARPC1 was further demonstrated by immunoprecipitation using a specific antibody and whole cell lysate . EhAK1 antibody indeed precipitated EhARPC1 from total lysate ( Fig 7D ) . The results suggest a direct interaction of EhARPC1 with EhAK1 , and indirectly with EhCaBP1 . We further investigated the binding between EhARPC1 and EhAK1 in order to understand the nature of interaction between these two molecules . For this , His-tagged fragments of EhAK1 containing either the kinase or the SH3 domain were generated as shown in Fig 7E . Although EhARPC1 is not known to be a multidomain protein , the molecule was divided into two parts , N terminal ( Nter , containing WD 40 repeats ) and C-terminal ( Cter ) . Each fragment was fused to GST tag as outlined in Fig 7E . These fragments were used in the in vitro binding assay . GST-tagged full length EhARPC1 , Nter and C-ter showed strong interaction with full length EhAK1 and weak binding with KD . The Nter fragment was able to bind to SH3 domain of EhAK1 as shown in Fig 7F . Since SH3 domains [40] and WD40 repeats [41] are known to be involved in protein-protein interaction , it is likely that EhARPC1 is recruited through the SH3 domain of EhAK1 through Nter . We then studied if EhCaBP1 and its interacting partner EhAK1 may be involved in recruitment of EhARPC1 to the phagocytic cups . In order to demonstrate this , we visualized the sub-cellular localization of EhARPC1 in RBC phagocytosing cells carrying antisense constructs of either EhAK1 or EhCaBP1 in the presence and absence of tetracycline ( Fig 8A and 8B ) . The results displayed that EhARPC1 was not enriched at the site of RBC attachment in these cell lines grown in presence of tetracycline upto 5 min of incubation with RBC ( Fig 8A and 8B ) . Quantitative analysis indicated 60% reduction in the level of EhARPC1 signal at RBC attachment sites in EhAK1 antisense cells in presence of tetracycline as compared to level of EhARPC1 signal at the phagocytic cup in EhAK1 antisense cells in absence of tetracycline ( Fig 8C ) . Similar results were obtained when EhCaBP1 levels were down regulated in antisense cell line in presence and absence of tetracycline ( Fig 8D ) . In order to rule out any down regulation of EhARPC1 protein expression in EhAK1 antisense cells , we investigated the levels of EhARPC1 in EhAK1 antisense cell lines in presence of different concentrations of tetracycline . While there was a decrease in the level of EhAK1 and EhCaBP1 on increasing tetracycline concentration ( Fig 8E and S4A Fig respectively ) , we did not observe any change in the level of EhARPC1 in these cells ( Fig 8F and S4A Fig ) . Image depicting the population of downregulated EhAK1 and EhARPC1 cells is shown in S4B Fig . These results suggest that EhARPC1 recruitment at the phagocytic cups requires EhAK1 and on decreasing concentration of these molecules time taken to phagocytose RBC increase significantly . Many protein kinases bind their cognate substrates . Since binding between EhARPC1 and the KD of EhAK1 was observed , we investigated if EhARPC1 is one of the substrates of EhAK1 . The results are shown in Fig 9A . When purified KD of EhAK1 was incubated with EhARPC1 in presence of phosphorylation buffer and γ-32P-ATP , bands corresponding to phosphorylated form of EhARPC1 and autophosphorylated kinase were observed . No radioactive band was visible in the reaction were kinase dead mutant ( K85A-EhAK1 ) was used as the enzyme . This suggests that in addition to actin [31] , EhARPC1 is one of the substrates of EhAK1 . The role of this phosphorylation event is not clear and is currently being investigated . It is clear from the results presented so far that amoebic Arp 2/3 subunit EhARPC1 is recruited to the phagocytic site through EhAK1 . Since Arp 2/3 is a complex of seven proteins it is likely that the whole complex may be recruited through EhARPC1 subunit . To demonstrate this , we investigated if another subunit of the complex might also be similarly recruited . ARPC2 has been used in the past as a representative subunit of the Arp 2/3 complex for determining presence of the entire complex [7 , 42] . Moreover , crystal structure of bovine Arp2/3 complex and cross linking studies has revealed that Arpc1 and Arpc2 subunits are present in close proximity in the complex [43–45] . Therefore , we chose EhARPC2 , the 34KDa subunit of the EhArp2/3 complex for further study , sequence alignment and specificity of anti-EhARPC2 antibody S5A and S5B Fig respectively . We first investigated possible interaction of EhARPC1 and EhARPC2 using an in vitro pull down approach . EhARPC1 antibody was able to pull EhARPC2 from total amoebic lysate ( Fig 9B ) . Moreover , anti EhARPC2 antibody was also able to pull down EhARPC1 suggesting that these two proteins are present in the complex and may be interacting with each other , as seen in other systems . We carried out fluorescence imaging during RBC uptake to localize EhARPC2 in relation to EhARPC1 ( Fig 9C ) . Like EhARPC1 , EhARPC2 was also found in phagocytic cups , and in just closed phagosomes , but not in phagosomes after scission . Quantitative analysis of the images displayed co-localization of both molecules at phagocytic sites ( Fig 9D ) . Further we did not observe significant enrichment of EhARPC2 in RBC attachment sites of trophozoites expressing antisense EhAK1 . Most of the stain was found in the cytoplasm ( Fig 9E ) . The accumulation of EhARPC2 at RBC attachment sites in trophozoites expressing antisense EhARPC1 was also investigated . No significant enrichment at the site was observed suggesting that EhARPC2 is not recruited independent of EhARPC1 ( S5C Fig ) . The results suggest that EhARPC2 is also recruited to the phagocytic site through EhAK1 and that EhARPC1 and EhARPC2 interact with each other . Therefore , the Arp2/3 complex is likely to be recruited to phagocytic sites through EhAK1 .
Phagocytosis is a multifactorial and multistep process that is initiated on attachment of a particle and completed after phagosomes are formed and separated from plasma membrane . Attachment of the particle leads to activation of downstream signaling cascade which ultimately causes the tethering of actin filaments to the plasma membrane and generation of force required for the pseudopod protrusion . Therefore , one of the major objectives of the phagocytic signaling system is to initiate actin dynamics . The mechanism of coupling of the signaling system with that of actin dynamics has been worked out in a few systems . It appears that one of the key steps that initiates actin filament mesh is recruitment of proteins that are involved in nucleation , polymerization , bundling , and branching of actin . A few systematic studies were carried out to identify proteins that may be involved in phagocytosis . One approach used sequence similarity based identification of E . histolytica genome encoded homologs of known actin dynamics proteins with the assumption that some of these may be involved in actin dynamics during phagocytosis [11] . In the second approach proteins present in phagosomes under different conditions were characterized [12–14] . Though there is always the possibility that molecules that are involved in the early phase of phagocytosis may not be there by the time phagosomes are isolated , it is likely that the majority of the proteins involved at different phases may still be present in the proteome . We have used a different approach for identification of relevant proteins , particularly since we are interested in deciphering the pathway mediated by EhCaBP1 . Our approach involved identification of proteins that bind EhCaBP1 and EhAK1 [30 , 31] . EhARPC1 turned out to be one of the common proteins observed in most of the analysis described above ( as shown in S1 Table ) . Since EhARPC1 is part of the Arp 2/3 complex proteins that are thought to be key regulators of actin dynamics we selected this protein for further studies . Involvement of other actin dynamics modulating proteins cannot be ruled out at present . Transient participation by these proteins may not likely be reflected in the proteome composition . The involvement of other pathways for phagocytosis in E . histolytica is an open question . Our laboratory has investigated the sequence of events that are initiated on attachment of a particle destined for phagocytosis in E . histolytica [15 , 30–32] . Though RBCs were mainly used as phagocytic particle , the proposed pathway was also found to operate during phagocytosis of mammalian cells [46] . The pathway unraveled so far proposes that upon particle attachment EhC2PK accumulates at the site , followed by EhCaBP1 and EhAK1 . EhCaBP3 is independently recruited to the phagocytic sites . EhCaBP1 , EhCaBP3 and EhAK1 bind actin and manipulate actin polymerization and/or bundling [15 , 30–32] . However , these proteins are unlikely to be involved in actin nucleation and formation of directed branched filaments . In this report we show that the EhARPC1 and EhARPC2 of E . histolytica Arp 2/3 complex , are recruited to the macromolecular complex at the phagocytosis initiation site , and that this recruitment is through EhAK1 . The results presented here show that the proposed phagocytic pathway involving EhARPC1 is also involved in the uptake of other cells , such as mammalian cells . The mechanism presented here for recruitment of Arp 2/3 complex proteins in E . histolytica has not been seen in any other system so far . EhARPC1 , the p41 subunit of Arp 2/3 complex was first identified as EhCaBP1 binding protein through a proteomic screen [30] . However , results presented here clearly show that EhARPC1 binds EhAK1 , and it interacts with EhCaBP1 through EhAK1 . Moreover , our results suggest that N-terminal domain consisting of WD40 repeats of EhARPC1 binds mainly the SH3 domain of EhAK1 . Since SH3 domains are known to be involved in protein-protein interaction and act as recruiters of various molecules at signaling sites it was not surprising to find this domain playing a possible role in EhARPC1 recruitment [40] . A number of evidences suggest that the interaction between EhAK1 and EhARPC1 is important for RBC uptake in these cells; EhARPC1 is not recruited to the phagocytic cups on down regulation of EhAK1 expression and similar pattern of distribution of EhARPC1 and EhAK1 during phagocytosis , that is , present in phagocytic cups and just closed phagosomes , but not in phagosomes after scission , unlike EhC2PK , EhCaBP1 and EhCaBP3 . Similar observations were also made in D . discoideum using GFP-tagged Arp3 and p41Arc . These proteins were present just after closure of phagosomes , but not after phagosomes were separated [47] . This is consistent with the role of Arp 2/3 complex in generating the necessary force for extension of pseudopods . Once phagosomes are formed there is no need for proteins that are involved in actin branching and extension . On phagosome maturation , actin molecules are removed paving the way for vesicles to fuse with other compartments [48 , 49] . We have also looked into the possibility that EhARPC1 recruitment at the cups is needed for progression of cups to become phagosomes . This was done by observing the effect of expression knockdown of EhARPC1 on amoebic phagocytosis . A significant reduction in phagocytic rate suggested that recruitment of EhARPC1 is necessary both for initiation ( delay in cup formation ) , and completion of phagocytosis ( delay in phagosome formation ) . ARPC1 is an important subunit of the Arp 2/3 complex . In S . cerevisiae null mutants of all Arp 2/3 complex proteins except Arpc1 survive , suggesting that this is the only essential subunit of Arp 2/3 complex [50] . Similar studies in A . thaliana have also shown that T-DNA insertion mutants of Arp2 , Arp3 and ARPC5 did not have major defects in development and were still viable , but ARPC1 was essential [51] . However , Arp2 was also found to be essential in D . discoideum , and E . histolytica homolog was able to complement D . discoideum protein suggesting that some of these proteins from E . histolytica may be functionally equivalent to that from other systems [52] . Our sequence analysis did reveal that EhARPC1 may have different properties compared to ARPC1 from other organisms unlike Arp2 subunit , which may be more conserved . It is likely that Arp 2/3 complex is recruited to phagocytic sites through EhARPC1 . In order to show if the whole complex is likely to be present and not just EhARPC1 , we have used EhARPC2 or p34 subunit as a marker of Arp 2/3 complex . This subunit of Arp 2/3 complex has been used as a representative of Arp 2/3 complex in many studies [7 , 42] . A number of observations , such as imaging , and pull down suggest that both EhARPC1 and EhARPC2 interact with each other and colocalize at the phagocytic cups . We have also shown that EhARPC2 is not recruited independent of EhARPC1 . Down regulation of EhARPC1 expression abolished EhARPC2 accumulation at the site of phagocytosis . Therefore , it appears that EhAK1 recruits both EhARPC1 and EhARPC2 to the phagocytic site . Additional evidence in support of this comes from our previous study where we had observed that EhAK1 could pull down both EhARPC1 and EhARPC2 [31] . Our data show that during phagocytosis coupling of the signaling system to actin dynamics is likely to be mediated through recruitment of EhARPC1 and EhARPC2 , components of Arp 2/3 complex through EhAK1 . The role of ARPC1 ( also known as Arc 40 ) in the recruitment of Arp 2/3 complex has also been shown in yeast [33 , 50] . Here , the ARPC1/Arc40 protein binds the VCA domain of WASP activators and helps Arp 2/3 complex recruitment through an “Arm” region at the C-terminal of Arc40 . However , in our study we find that EhARPC1 is likely to be recruited by binding with SH3 domain of EhAK1 . We have also seen EhAK1 kinase domain binding EhARPC1 . Since many protein kinase substrates are known to interact with respective kinases [53] , it was speculated that EhARPC1 may be a substrate of EhAK1 . This was verified experimentally and EhAK1 dependent phosphorylation of EhARPC1 was observed . In our previous study actin was shown to be a major substrate of EhAK1 based on use of total amoebic cell lysate [31] . In view of the results presented here it appears that though actin may be the major substrate , EhAK1 may also be phosphorylating other proteins , as minor substrates . The significance of EhARPC1 phosphorylation in relation to mechanism of phagocytosis is currently being explored . ARPC1 is known to be phosphorylated in mammalian systems by p-21 activated kinase . This phosphorylation is required for optimal cell motility upon stimulation with growth factors and may play a critical role for localization of this subunit with rest of the complex [54] . Phosphorylation of ARPC1 by an alpha kinase like kinase has not been observed before . It is also possible that EhARPC1 has a functional role independent of being part of Arp 2/3 complex . This may likely explain the essential nature of this protein unlike other components of Arp 2/3 complex . Whether EhARPC1 also has other functions in E . histolytica is totally an open question . In conclusion a novel mechanism of recruitment of Arp 2/3 complex to the phagocytic machinery in E . histolytica is proposed . This suggested mechanism is distinctly different from all other mechanisms proposed so far . Since this parasite has a high phagocytic rate it may have evolved novel mechanisms to meet its requirement for rapid actin dynamics .
Both mice and rabbits used for generation of antibodies were approved by the Institutional Animal Ethics Committee ( IAEC ) , Jawaharlal Nehru University ( IAEC Code No . : 18/2010 ) . All animal experimentations were performed according to the National Regulatory Guidelines issued by CPSEA ( Committee for the Purpose of Supervision of Experiments on Animals ) , Ministry of Environment and Forest , Govt . of India . E . histolytica strain HM-1: IMSS trophozoites and all transformed strains were maintained and grown in TY1-S-33 medium supplemented with 125 μl of 250 U ml− 1penicillin G ( potassium salt from Sigma ) and 0 . 25 mg ml− 1streptomycin per 100 ml of medium as described before [37] . The transformants containing tetracycline inducible system and GFP ( a constitutive expression system ) were grown in the presence of 10 μg ml− 1of hygromycin B or G418 . The cells were first grown for 48 h ( 60–70% confluent ) and then 30 μg ml− 1 tetracycline or 20 μg ml− 1 G418 was added to the medium for 36 h for induction . E . histolytica was transfected by electroporation . Briefly trophozoites were collected from log phase cultures and washed with PBS followed by incomplete cytomix buffer ( 10 mM K2HPO4/KH2PO4 ( pH 7 . 6 ) , 120mMKCl , 0 . 15mM CaCl2 , 25 mM HEPES ( pH 7 . 4 ) , 2 mM EGTA , 5 mM MgCl2 ) . The washed cells were then re-suspended in 0 . 8 ml of complete cytomix buffer ( incomplete cytomix containing 4 mM adenosine triphosphate , 10 mM glutathione ) containing 200 μg of plasmid DNA and subjected to two consecutive pulses of 3 , 000 V cm− 1 ( 1 . 2 kV ) at 25 mF ( Bio-Rad , electroporator ) . The transfectants were initially allowed to grow without any selection . Drug selection was initiated after 2 days of transfection in the presence of 10 μg ml−1G418 ( for constitutive expression vectors ) or hygromycin B ( for tetracycline inducible vector ) | E . histolytica is the causative agent of amoebiasis and leads to morbidity and mortality in developing countries . It is known to phagocytose immune and non-immune cells , epithelial tissue , erythrocytes and commensal bacteria . The high rate of phagocytosis in this protist parasite provides a unique system to study the signaling cascade that is activated after attachment of the particle to the cell surface . The major objective of the signaling pathway is to generate force for uptake of the particle and this is done through stimulating cytoskeleton to form appropriate structures . However , the molecular mechanism of the same is still largely unknown in E . histolytica , though this pathway has been characterized in many other systems . We have been investigating this pathway by using red blood cells as a particle and have identified different molecules required during the initial stages of phagocytosis . In this study we demonstrate the mechanism by which actin cytoskeleton branching complex EhARP2/3 is recruited at the site of erythrophagocytosis and show that the recruitment is through an atypical alpha kinase EhAK1 . A number of different approaches , such as pull down assay , conditional suppression of EhAK1 expression and imaging were used to decipher this pathway . Therefore this study provides a mechanism by which actin dynamics couples to the initial signaling system , activated on attachment of RBC to the cell receptors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | The Entamoeba histolytica, Arp2/3 Complex Is Recruited to Phagocytic Cups through an Atypical Kinase EhAK1 |
Cholera is the prime example of blood-group-dependent diseases , with individuals of blood group O experiencing the most severe symptoms . The cholera toxin is the main suspect to cause this relationship . We report the high-resolution crystal structures ( 1 . 1–1 . 6 Å ) of the native cholera toxin B-pentamer for both classical and El Tor biotypes , in complexes with relevant blood group determinants and a fragment of its primary receptor , the GM1 ganglioside . The blood group A determinant binds in the opposite orientation compared to previously published structures of the cholera toxin , whereas the blood group H determinant , characteristic of blood group O , binds in both orientations . H-determinants bind with higher affinity than A-determinants , as shown by surface plasmon resonance . Together , these findings suggest why blood group O is a risk factor for severe cholera .
Cholera is a severe diarrheal disease caused by the pathogen Vibrio cholerae [1] . It originated in the Ganges delta , but is now widespread across the world . Caused by contaminated drinking water , severe cholera outbreaks often occur in war zones or in the aftermath of natural disasters . Well remembered is the recent outbreak in Haiti following the earthquake in 2010 –the worst in recent history . Every year , several million people fall ill from cholera and one hundred thousand die [2] . Individuals with blood group O have a particularly high risk of severe symptoms and hospitalization [3–8] , but the molecular causes of this association are not well understood . However , individuals with blood group O are not more susceptible to cholera infection than those with other blood groups , pointing towards the critical role of cholera virulence factors , rather than bacterial colonization [3 , 4] . V . cholerae is a gram-negative bacterium capable of colonizing the human small intestine . In response to environmental stimuli , the bacterium produces and secretes its main virulence factor , the cholera toxin ( CT ) [9] . CT is an AB5 protein toxin , which consists of one toxic A-subunit ( CTA ) and five identical B-subunits arranged in a doughnut-shaped ring ( CTB ) [10] . The B-pentamer itself is non-toxic and , in fact , part of a well-known vaccine ( Dukoral ) . It is responsible for binding to the host cell receptors as a first step of pathogenesis . The primary receptor of the cholera toxin is the GM1 ganglioside , which is expressed in distinct membrane microdomains of small-intestinal epithelial cells [11 , 12] . CTB can bind five GM1 gangliosides simultaneously , resulting in one of the strongest carbohydrate-protein interactions known [13] . After receptor binding , the toxin is internalized by endocytosis , where it hijacks the cells’ endogenous pathways , ultimately inducing the watery diarrhea that is characteristic of cholera [14] . There are two major biotypes of V . cholerae , classical and El Tor , which produce cholera toxins with a slight sequence variation ( cCT and ET CT ) [15 , 16] . cCT and ET CT differ in only two residues of the B-subunit , His18/Tyr18 and Thr47/Ile47 ( classical/El Tor ) . El Tor V . cholerae is reported to have the strongest blood group association [17] . This biotype has been dominating the current pandemic , which has been ongoing since 1961 . However , since 2001 a new hybrid biotype has taken over , which has the biochemical characteristics of El Tor V . cholerae , but produces a cholera toxin with the classical sequence [18] . ABO ( H ) histo-blood group antigens are carbohydrates expressed on the surface of red blood cells and epithelial cells , as glycolipids or glycoproteins . The smallest determinant is the H determinant , a trisaccharide characteristic of blood group O . It can be converted into the A- and B-determinants by enzymatic addition of N-acetylgalactosamine ( GalNAc ) or galactose ( Gal ) , respectively ( see Fig 1 for carbohydrate structures and nomenclature ) . Blood group antigens are also expressed in bodily secretions such as mucus or as free lactose-derived oligosaccharides in human milk ( referred to as ‘human milk oligosaccharides’ or HMOs , to distinguish these from ‘blood group antigens’ or BGAs ) [19–21] . Despite the identification of the cholera toxin as the likely culprit of cholera blood group dependence [7] , the literature has been inconclusive regarding the toxin’s capacity to bind blood group antigens . While the cholera toxin was reported to bind A- and B-active glycolipids and glycoproteins on intestinal brush border membranes [22 , 23] , microtiter well assays suggested that only a chimera of CTB and the homologous heat-labile enterotoxin from enterotoxigenic E . coli ( LTB ) , but not CTB itself , could bind blood group antigens [24] . Two crystal structures of LTB and the CTB/LTB chimera in complex with HMOs identified a possible blood group antigen binding site on the lateral side of the toxins , distinct from its primary binding site ( for GM1 ) [25 , 26] . Recent studies , by surface plasmon resonance ( SPR ) , isothermal titration calorimetry ( ITC ) and saturation transfer difference NMR ( STD NMR ) , confirmed that HMOs or other blood group antigen derivatives can indeed bind to CTB , however , none of these studies used unmodified blood group determinants [27–29] . Furthermore , they came to different conclusions: While HMOs were found to bind to CTB of both biotypes , but with different kinetics [27] , a custom-synthesized β-glycoside of the B Lewis-y oligosaccharide only bound to cCTB , but not to ET CTB [28] . The binding to ET CTB was restored by mutating Ile47 to Thr47 , suggesting a prominent role of this residue in blood group antigen recognition . The current study was designed to elucidate the molecular causes of cholera blood group dependence . Here , we present high-resolution X-ray crystal structures of both variants of CTB , in complexes with the unmodified blood group determinants of blood groups A and O , with matching SPR data .
Binding of the different blood group oligosaccharides to cCTB and ET CTB was further characterized by SPR ( Fig 3 ) . CTB was immobilized on the sensor chip , and the oligosaccharides were injected in increasing concentrations . H-tetra-BGA was found to bind both biotypes equally well ( KD = 1 . 1/1 . 5 mM for cCTB/ET CTB ) , whereas A-penta-BGA bound cCTB significantly stronger than ET CTB ( KD = 2 . 2/> 30 mM for cCTB/ET CTB ) . In contrast , A-penta-HMO exhibited similar binding affinities ( KD = 2 . 7/4 . 6 mM for cCTB/ET CTB ) , comparable to previously reported values ( KD = 1 . 2 mM , as determined by SPR for both cCTB and ET CTB [27]; KD = 5 . 2/6 . 4 mM for cCTB/ET CTB , as determined by STD-NMR [29] ) . However , the new experiments showed no differences in the kinetics; all ligands exhibited the same quick association and dissociation , and no regeneration of the SPR chips was required . Preliminary experiments with B-penta-BGA ( Elicityl SA ) were also performed and showed the same phenomenon as for A-penta-BGA ( KD > 30 mM ) , suggesting a strong effect of the N-acetyl group present in blood group antigens . In addition , we performed preliminary SPR experiments with a blood group A determinant of type-1 structure , which showed very weak to no binding to CTB of both biotypes . Blood group antigens with type 2 core structures are well represented on small intestinal glycoproteins [31] , while type 1 structures are predominantly found on glycolipids [32] . We conclude that the toxins prefer type 2 over type 1 structures , in line with previous experiments [24 , 33] .
CTB of both V . cholerae biotypes were found to bind blood group determinants or HMOs at a secondary binding site on the lateral face of the toxin ( Fig 2A ) . The same site was identified in previous studies for A-penta-HMO binding to LTB or a CTB/LTB chimera [25 , 26] . However , the oligosaccharides exhibited unexpected orientations: both A-penta-BGA and A-penta-HMO bound to CTB in the opposite orientation compared to the other toxin structures , and H-tetra-BGA bound in both orientations . Complementary SPR analysis revealed that while both cCTB and ET CTB bound to H-tetra-BGA with similar affinity ( 1 . 0 and 1 . 5 mM , respectively , which is comparable to H-tetra-HMO , KD = 0 . 5 mM for ET CTB [27] ) , they differed greatly in their binding affinity to A-penta-BGA ( with ET CTB binding very weakly ) . Similar observations were previously made using ITC , involving a β-glycoside of B-penta-BGA [28] . The millimolar KD values observed for the ABH blood group oligosaccharides may seem modest , but multivalent binding of the pentamer is expected to enhance binding avidity significantly . This has been demonstrated for GM1 [34] , and also for Shiga toxins , which bind the GD3 ganglioside with millimolar affinity , but achieve nanomolar avidity upon multivalent binding [35] . Summarizing , our experiments revealed that ET CTB can strongly distinguish between blood group A and H determinants , which both contain a GlcNAc residue at the reducing end , where human milk oligosaccharides contain unmodified Glc . In the ET CTB crystal structure , only three of ten binding sites showed sufficient electron density to warrant modeling of the entire A-penta-BGA ligand , and even then , the GalNAc residue characteristic of blood group A was disordered ( Fig 2E ) . Notably , the GalNAc residue is close to Tyr18 and not Ile47 , the substitution of which restored binding activity [28] . This was in contrast to expectations , since the ligand bound in the opposite orientation compared to the toxin structures published previously [25 , 26] . In the crystal structure of ET CTB in complex with A-penta-HMO , the GalNAc residue is , however , well defined by electron density ( Fig 2D ) , strongly suggesting that interference of the reducing end N-acetyl group with Ile47 is responsible for the weaker binding affinity of A-penta-BGA versus A-penta-HMO and ET CTB versus ET CTB I47T/cCTB/LTB . In addition , we observe a conformational change of Gln16 upon binding of A-penta ( BGA or HMO ) , which suggests that this residue interferes with binding in its native conformation . The trend appears clear: The cholera toxin binds blood group H-determinants preferentially over A-determinants , suggesting that stronger binding correlates with more severe disease . The relationship is especially pronounced for the El Tor biotype , which shows the strongest blood group dependence [17] . Individuals with the ‘secretor’ phenotype , who also express ABO ( H ) blood group antigens on the surface of intestinal epithelial cells and mucins , experience less severe symptoms than ‘non-secretors’ [36] . How can this be explained ? The small intestinal mucus layer is known to serve a protective function , by blocking and removing intruding pathogens and toxins . Stronger binding may hence be expected to correlate with enhanced protection against disease rather than with increased toxicity . However , binding to the mucus layer can serve as a double-edged sword: by binding potent invaders , the invaders can also gain entrance to the host cells . Moreover , with the toxins likely being released in close proximity to the host cells [37 , 38] , the main barrier may already be overcome . But are the measured affinities at all biologically relevant ? For one , only the H-determinant , but not the A-determinant , can bind to the toxin in two alternative orientations , likely contributing to its increased affinity . Both orientations should in principle be possible in the context of glycoconjugates , as long as they are β-glycosidically linked , avoiding clashes with the toxin . In this regard , we note that there is a dominance of the α-anomer of the reducing end GlcNAc in the X-ray structures , mimicking the Gal 4-OH group , while the relevant glycoconjugates are β-glycosidically linked ( S2 Fig ) . This may influence binding affinities , and the actual differences between blood group determinants with fixed β-linkage may be even greater . Also , we need to consider that the blood group determinants are terminal epitopes of glycoconjugates , and are connected to the protein/lipid by carbohydrate linkers . The lengths of these linkers can also have an effect on the binding to CT , with shorter linkers preventing the non-reducing end of the glycoconjugates from reaching to blood group binding site of the toxin . Finally , not all blood group antigens exhibit the second fucose residue ( Fucα3 ) , which is likely critical for binding in orientation 2 . It is very difficult to predict glycosylation based on the expression profiles of glycosyltransferases , and only limited data exist regarding the distribution of glycoconjugates in human tissues [19 , 39] . However , it is known that secretor fluids are characterized by enhanced fucosylation [40 , 41] , hence secretors should be able to bind a considerable fraction of the blood group antigens in orientation 2 . In addition , it has recently been shown that fucosylated structures promote CTB binding and cellular uptake in colonic epithelial cells [42] . Therefore we expect that the discrimination between blood group determinants observed in vitro resembles the situation in the human body . The scenario may be as follows ( Fig 4 ) : After colonization of V . cholerae in the human small intestine , the bacteria secrete their main virulence factor , the cholera toxin . The small intestine is well suited for invasion , because of the limited thickness of its mucus layer , as compared to , for example , the colon . At the Peyer’s patches , the mucus layer is particularly thin . Moreover , after efficient intestinal colonization , the bacteria are already in close contact with the epithelium [37 , 38] . The secreted toxin either binds directly to the receptors on the epithelial cells or to glycoconjugates in the human mucus layer—among those , if present ( in secretors ) , the ABH blood group antigens . Also non-secretors are likely to exhibit suitable docking structures ( such as Lewis-x ( Galβ4[Fucα3]GlcNAcβ ) , which comprises part of H-tetra-BGA ) , but only secretors contain ABH histo-blood group antigens in their mucus and on intestinal epithelial cells [19 , 39] . Low-affinity multivalent receptor interactions with quick association-dissociation dynamics are likely essential for efficient intoxication , to ensure that the toxin reaches the epithelial cells and does not get stuck on its way ( and risk being expelled by mucus shedding ) . Once the CT reaches the epithelial cells , it may either bind to the GM1 ganglioside , to other functional receptors [43–45] , or be engulfed by an alternative mechanism involving the turn-over of brush border membranes [46–48] . Here , it is interesting to note that based on the data presented here , the CT appears to be able to bind to blood group antigens and GM1 simultaneously . However , until the toxin is strongly bound to the cell surface , it may still diffuse away from the cells through the mucus layer . The flow around the small intestinal villi is poorly understood , but simulations suggest that during intestinal contractions , mixing and absorption around the mucosa is increased [49] . Diffusion of the toxin away from the epithelium towards the lumen of the gut is more likely to occur in secretors with blood group A or B than in those with blood group O , which bind the CT more strongly , via their H-antigens . The toxin is therefore expected to be able to enter host cells more effectively in secretors with blood group O than in those with blood groups A or B , with the effect being most pronounced for the El Tor biotype . How this compares to non-secretors is unknown , but it is conceivable that the steric interference observed for blood group A and B determinants is the main protecting factor from severe disease . Such a scenario would explain why secretors with blood groups A , B or AB are relatively protected , while those with blood group O experience more severe symptoms , as observed by Arifuzzaman et al . [50] . With ≈80% of the human population being secretor-positive , individuals with blood group O would overall be more prone to severe disease , as observed [7 , 8 , 17] . Likewise , they would experience stronger protection from vaccination [17] . The scenario further explains why the El Tor biotype is more discriminatory than the classical biotype [17] . However , one should keep in mind that in recent years , a new hybrid biotype evolved , which expresses a toxin with the classical sequence [18] . This enables more effective intoxication also of individuals with blood group A or B , promoting the bacteria’s propagation . With respect to the human milk oligosaccharides , we note that previous SPR data [27] showed different binding kinetics for H-tetra-HMO and A-penta-HMO , with A-penta-HMO exhibiting significantly slower association/dissociation rates . In contrast , all ligands used in the present study showed fast binding and release . The ligands used were obtained from different sources , but were otherwise identical . While the previous study was performed with compounds isolated from the feces of breast-fed infants ( Isosep AB ) , the SPR results presented herein are based on ligands produced by bacterial fermentation ( Elicityl SA ) . In both cases , A-penta-HMO samples were only > 90% pure , suggesting that residual impurities may have contributed to the different binding profiles . Indeed , the SPR binding profile in our earlier report ( Fig 1CE in [27] ) suggests two separate binding events upon closer inspection . GM1 and GM1-derived oligosaccharides have earlier been shown to be present in human milk [51 , 52] , and are known to strongly bind CTB [13 , 53] . The presence of trace amounts of these would explain the observed binding profiles . Preliminary ELISA experiments ( S4 Fig and S1 Methods ) support such an interpretation . While cell biological studies are needed to establish the molecular basis of cholera blood group association , the picture becomes increasingly clear . The additional GalNAc ( or Gal ) residue in blood group A ( or B ) antigens weakens binding to the CT through steric interference . Moreover , only the symmetric blood group H ( but not A ) determinants can bind to the CT in two alternative orientations . This enhances the binding capacity and the biological potency of H-antigens . While in endemic regions like Bangladesh , the blood group profile of the population has been shaped by evolution to minimize losses of human lives [3 , 29 , 54] , most of the world is less well adapted . With climate change , water-borne diseases like cholera are predicted to be aggravated worldwide [55 , 56] , with significant socioeconomic costs [57] . The molecular insights obtained in this investigation may help to stem the tide .
The nucleotide sequence of the classical CTB gene ( Genbank: AAC34728 . 1; coding for amino acid sequence 22–123 ) was codon-optimized for expression in Escherichia coli and synthesized by GeneArt ( Life Technologies ) in a standard pMA-T plasmid . The gene contained the N-terminal LTB signal sequence directing the protein to the periplasmic space , as well as flanking NdeI and BamHI restriction sites . After excision of the gene , it was ligated into a pET-21b ( + ) vector ( Novagen ) , and transformed by heat shock into competent E . coli BL21 ( DE3 ) cells for protein expression . The cells transformed with the pET-21b ( + ) -cCTB plasmid were grown in LB medium supplemented with 0 . 1 mg/ml ampicillin at 37°C until an OD600nm of ~0 . 5 was reached . Before induction with 0 . 5 mM IPTG , the temperature was lowered to 25°C and the protein was expressed for 16–20 hours . The cells were harvested by centrifugation and re-suspended in ice-cold sucrose solution ( 20 mM Tris pH 8 , 25% ( w/v ) sucrose , 5 mM EDTA ) , and incubated on ice for 15 minutes . The supernatant was removed by centrifugation at 8500 × g for 20 minutes , and the pellet was re-suspended in periplasmic extraction buffer ( 5 mM MgCl2 , 0 . 1 mg/ml lysozyme ) . The periplasmic fraction was separated from the cell debris by centrifugation at 8500 × g for 20 minutes , and dialyzed against PBS . cCTB was purified by affinity chromatography using D-Gal-sepharose ( Thermo Scientific ) , and was eluted with 300 mM D-Gal ( AppliChem ) in PBS . The protein was concentrated and applied to a Superdex75 size-exclusion chromatography column ( GE Healthcare ) , where the buffer was exchanged to 20 mM Tris pH 7 . 5 , 100 mM NaCl . The purified protein was dialyzed against Tris-storage buffer ( 20 mM Tris pH 7 . 5 , 200 mM NaCl ) , concentrated to 3–10 mg/ml and stored at -80°C . Vibrio sp . 60 containing the gene for ET CTB was grown in LBS medium ( LB medium with 15 g/l NaCl ) at 30°C , supplemented with 0 . 1 mg/ml ampicillin , until an OD600nm of ≈0 . 2 . The culture was induced with 0 . 5 mM IPTG and expressed for 16–20 hours . ET CTB is naturally secreted into the growth medium , and the cells were removed by centrifugation at 40 , 000 × g . The supernatant was applied to a D-Gal-sepharose affinity column , and ET CTB was eluted using 300 mM D-Gal in PBS . After concentration , the protein was further purified by size-exclusion chromatography ( in 20 mM Tris pH 7 . 5 , 100 mM NaCl ) and dialyzed against 20 mM Tris pH 7 . 5 , 200 mM NaCl . The purified protein was concentrated to 3–10 mg/ml , and stored at -80°C . CTB and oligosaccharide ligands were mixed at a molar ratio of 1:10 ( B-subunit:ligand ) two hours prior to the crystallization setups . For the initial screening performed on an Oryx4 crystallization robot ( Douglas Instruments , UK ) , the sitting-drop vapor-diffusion technique was used , with a CTB concentration ranging from 3 . 0 to 8 . 5 mg/ml at 20°C . For the optimization setups , the hanging-drop vapor-diffusion technique was used with the same protein concentration range , also at 20°C . All oligosaccharides were purchased from Elicityl-Oligotech ( Elicityl SA , Crolles , France ) , with the product numbers GLY035-5 ( A-penta-BGA: GalNAcα3[Fucα2]Galβ4[Fucα3]GlcNAc ) , GLY048 ( H-tetra-BGA: Fucα2Galβ4[Fucα3]GlcNAc ) and GLY067 ( A-penta-HMO: GalNAcα3[Fucα2]Galβ4[Fucα3]Glc ) . Crystals of the ET CTB complexes were grown in 0 . 1 M Tris-bicine pH 8 . 5 , 9–12 . 5% ( each ) PEG1000/PEG3350/MPD , 30 mM ( each ) CaCl2/MgCl2 ( optimization of hits from the Morpheus crystallization screen; Molecular dimensions ) . Crystals of the cCTB complexes were obtained from either 0 . 1 M Tris-bicine pH 8 . 5 , 9–12 . 5% ( each ) PEG1000/PEG3350/MPD , 30 mM ( each ) CaCl2/MgCl2 , or 0 . 1 M MES-imidazole pH 6 . 7 , 25% PEG4000 , 5% PGA-LM ( optimization of hits from the PGA-LM crystallization screen; Molecular Dimensions ) . Diffraction quality crystals were obtained by microseeding , using microseed stocks prepared by crushing small crystals from initial hits with a seed bead . Most of the CTB complexes crystallized in space group P212121 , with two pentamers in the asymmetric unit , from a MORPHEUS-derived condition , except for ET CTB + H-tetra-BGA , which crystallized in P21 , and cCTB + A-penta-BGA , which crystallized from the PGA-LM-derived condition , with one pentamer in the asymmetric unit ( space group P212121 ) . Unit cell parameters are listed in Table 1 . Diffraction data for all complexes were collected at ESRF beam lines ID29 [58] or BM30A-1 ( automated data collection by MASSIF-1 ) . Data collection and refinement statistics are listed in Table 1 . Data were processed and scaled using XDS [59] and Aimless from the CCP4 software suite [60] , and the structures solved by molecular replacement using MOLREP [61] or Phaser [62] , taking the atomic coordinates of the 1 . 25 Å crystal structure of CTB as a search model for molecular replacement ( PDB: 3CHB [63] ) . The search model was prepared with the program CHAINSAW [64] , removing ligands , water molecules and alternative conformations , as well as pruning non-conserved residues down to the γ atom . After molecular replacement , rigid body refinement was performed with REFMAC5 [65] , using 5% of the data for cross-validation ( Rfree ) . The structures were subsequently manually optimized using Coot [66] and further refined with REFMAC5 . Automated restrained refinement and manual rebuilding were repeated in an iterative manner until no more significant changes in R-factors were observed . The ET CTB + H-tetra-BGA structure exhibited twinning ( twin fraction 0 . 1 ) , and was consequently refined using intensity-based twin refinement as implemented in REFMAC5 . The highest-resolution structure , cCTB + H-tetra-BGA ( 1 . 08 Å ) , was refined with anisotropic B-factors . In all structures , the sequence was corrected by substituting His94 for Thr94 , and in the ET CTB structures , in addition His18 and Thr47 were replaced by Tyr18 and Ile47 , respectively . After refinement of the protein alone , ordered buffer components and ions were modeled , before adding water molecules in several cycles . The blood group or human milk oligosaccharides were included last , to ensure unbiased and conformationally optimal modeling . The ligands were prepared with MarvinSketch ( ChemAxon . com ) , using isomeric SMILES strings for the individual monosaccharides . The final ligand PDB and library files were created using PRODRG [67] , and were checked for errors using pdb-care ( glycosciences . de/tools/pdbcare [68] ) . The oligosaccharides were only modeled into the binding sites if clear electron density of at least three residues was present . Occupancies were refined by comparing the B-factors of the ligands with those of the interacting protein atoms , and taking the difference Fourier maps into account . The structures were validated using Coot [66] , Molprobity [69] , and figures prepared using PyMol ( Schrödinger LLC ) . SPR analyses were performed on a Biacore T100 biosensor system ( Biacore Life Sciences , GE Healthcare , Uppsala , Sweden ) at the Infrastructural Centre for Analysis of Molecular Interactions , University of Ljubljana , Slovenia . The interaction studies were carried out at 25°C in HBS-EP running buffer ( 10 mM Hepes pH 7 . 4 , 150 mM NaCl , 3 mM EDTA , 0 . 05% ( v/v ) surfactant P20 ) ; and the carbohydrate ligands A-penta-BGA , H-tetra-BGA , A-penta-HMO were solubilized in the same buffer . ET CTB and cCTB were diluted in 10 mM Na-acetate pH 5 . 5 and immobilized by amine coupling to a CM5 sensor chip to a response of 4000–7000 RU . For the first experiments with cCTB , H-tetra-BGA and A-penta-BGA were injected over the surfaces with an initial flow rate of 20 μl/min for 120 s ( followed by a 60 s dissociation phase and allowing 60 s of stabilization prior to the next injection ) , at increasing concentrations of the carbohydrate ligand ( 20 μM to 10 mM ) . The experiment was repeated with two injections per concentration at 10 μl/min for 120 s . H-tetra-BGA and A-penta-BGA ( 58 μM to 30 mM ) were injected over the ET CTB chip with an initial flow rate of 20 μl/min for 120 s . The experiment with H-tetra-BGA was repeated with two injections per concentration ( 29 μM to 15 mM ) at a flow rate of 10 μl/min for 120 s . Finally , both ligands were tested against both toxin biotypes using three randomized injections per concentration ( 313 μM to 10 mM ) , at a flow rate of 5 μl/min for 30 s . A-penta-HMO was tested at a separate occasion , using the same chip for cCTB ( quality tested by a repetition of H-tetra-BGA ) , but a newly immobilized chip for ET CTB ( ≈6000 RU ) . The graphs for A-penta-HMO are therefore not comparable to the rest . The experiments with A-penta-HMO ( 39 μM to 20 mM ) were performed with a flow rate of 10 μl/min for 120 s , with three separate runs . No regeneration was required for any of the interactions . The dissociation constants ( KD ) of the interactions were calculated by using a steady-state affinity model in the Biacore T100 evaluation software . Preliminary experiments were performed with B-penta-BGA ( GLY38-5 , Galα3[Fucα2]Galβ4[Fucα3]GlcNAc ) ( Elicityl SA ) and A-penta-BGA-type-1 ( GLY35-4 , GalNAcα3[Fucα2]Galβ3[Fucα4]GlcNAc ) ( Elicityl SA ) . The binding of B-penta-BGA was measured with a single concentration series from 39 μM to 20 mM ( 5 μl/min , 60 s ) for cCTB , and from 39 μM to 10 mM ( 5 μl/min , 60 s ) for ET CTB , while A-penta-BGA-type-1 was only manually tested at 1 mM and 5 mM concentrations for cCTB , and at 1 mM for ET CTB . | Cholera is a severe diarrheal disease that kills a hundred thousand people per year . With climate change , the number of cases is predicted to increase to millions . Individuals with blood group O are particularly at risk . Here we report high-resolution crystal structures of the native cholera toxin of both major biotypes , in complexes with relevant blood group determinants . These structures , in combination with quantitative binding data , shed light on cholera blood group dependence . Understanding the molecular basis of this association is expected to be of considerable importance , for example for developing new vaccination strategies that take this information into account . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"milk",
"body",
"fluids",
"toxins",
"pathology",
"and",
"laboratory",
"medicine",
"crystal",
"structure",
"immune",
"physiology",
"immunology",
"tropical",
"diseases",
"condensed",
"matter",
"physics",
"blood",
"groups",
"toxic",
"agents",
"toxicology",
"bacterial",
"diseases",
"neglected",
"tropical",
"diseases",
"crystallography",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"cholera",
"solid",
"state",
"physics",
"mucus",
"proteins",
"antigens",
"hematology",
"physics",
"biochemistry",
"blood",
"anatomy",
"electron",
"density",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] | 2016 | High-Resolution Crystal Structures Elucidate the Molecular Basis of Cholera Blood Group Dependence |
Telomerase , the enzyme that maintains telomeres , preferentially lengthens short telomeres . The S . cerevisiae Pif1 DNA helicase inhibits both telomerase-mediated telomere lengthening and de novo telomere addition at double strand breaks ( DSB ) . Here , we report that the association of the telomerase subunits Est2 and Est1 at a DSB was increased in the absence of Pif1 , as it is at telomeres , suggesting that Pif1 suppresses de novo telomere addition by removing telomerase from the break . To determine how the absence of Pif1 results in telomere lengthening , we used the single telomere extension assay ( STEX ) , which monitors lengthening of individual telomeres in a single cell cycle . In the absence of Pif1 , telomerase added significantly more telomeric DNA , an average of 72 nucleotides per telomere compared to the 45 nucleotides in wild type cells , and the fraction of telomeres lengthened increased almost four-fold . Using an inducible short telomere assay , Est2 and Est1 no longer bound preferentially to a short telomere in pif1 mutant cells while binding of Yku80 , a telomere structural protein , was unaffected by the status of the PIF1 locus . Two experiments demonstrate that Pif1 binding is affected by telomere length: Pif1 ( but not Yku80 ) -associated telomeres were 70 bps longer than bulk telomeres , and in the inducible short telomere assay , Pif1 bound better to wild type length telomeres than to short telomeres . Thus , preferential lengthening of short yeast telomeres is achieved in part by targeting the negative regulator Pif1 to long telomeres .
Telomerase is a specialized reverse transcriptase that extends the G-strand of telomeric DNA using its RNA subunit as a template . In Saccharomyces cerevisiae , telomerase consists minimally of Est2 , the catalytic reverse transcriptase , TLC1 , the templating telomerase RNA , and Est1 and Est3 , two telomerase accessory subunits that are both essential for telomerase action in vivo . In addition , yeast telomerase requires Cdc13 , the sequence specific TG1-3-binding subunit of the CST ( Cdc13-Stn1-Ten1 ) complex that has dual roles in protecting telomeres from degradation and recruiting telomerase to DNA ends ( reviewed in [1] ) . Yeast telomerase is regulated by both the cell cycle and telomere length . Telomerase-mediated telomere lengthening occurs only in late S/G2 phase , even though telomerase activity is present throughout the cell cycle [2 , 3] . Although Est2 and TLC1 are telomere associated even in G1 phase when telomerase is not active , Est1 [4] and Est3 [5] come to the telomere primarily in late S/G2 phase . Short telomeres are preferentially lengthened by telomerase [6 , 7] , a pattern that can be explained by higher levels of telomerase binding to short telomeres late in the cell cycle [8 , 9] . Both the Tel1 checkpoint kinase [8] and Tbf1 [10] , a telomere structural protein that binds to the sub-telomeric DNA of some telomeres , target telomerase to short telomeres . Tel1 itself binds preferentially to short telomeres [8 , 11] , as does the Mre11-Rad50-Xrs2 ( MRX ) complex [12] , which recruits Tel1 to telomeres [8] . Short telomeres have reduced levels of Rif2 [8 , 12] , a telomere structural protein that negatively regulates telomerase [13 , 14] . Because Rif2 and the MRX complex compete with each other for telomeric DNA binding [15] , in the absence of Rif2 , Tel1 no longer binds preferentially to short telomeres [12] . The S . cerevisiae Pif1 is the founding member of a helicase family that exists in virtually all eukaryotes ( reviewed in [16] ) . Pif1 was first identified because of its important role in maintaining mitochondrial DNA [17] . However , there are two forms of Pif1 depending on whether the first or second AUG is used to start translation , one destined for mitochondria and one localized to nuclei [18 , 19] . Nuclear Pif1 inhibits telomerase at both telomeres and double strand breaks ( DSBs ) [18–20] . Thus , telomeres are longer and the rate of de novo telomere addition to DSBs is greatly elevated in pif1 mutant cells . Checkpoint-mediated phosphorylation of Pif1 is required for Pif1 inhibition of telomerase at DSBs but not at telomeres [21] . In vivo and in vitro , Pif1 uses its ATPase activity to displace telomerase from DNA ends [22] . Pif1 also has more general roles in chromosome maintenance: it facilitates replication and suppresses DNA damage at G-quadruplex motifs , cooperates with Dna2 to process long Okazaki fragments , is needed for stability of mitochondrial DNA , and promotes break-induced replication ( BIR ) ( reviewed in [16]; also , [23–25] ) . Here , we show that Pif1 acts similarly at DSBs and telomeres in that its presence was associated with lower levels of telomerase binding at DSBs ( as it is at telomeres [22] ) . Using an assay that monitors telomeric DNA addition at individual telomeres [7] , we find that Pif1 reduced both the frequency and processivity of telomere addition . Moreover , telomerase was no longer bound preferentially to short telomeres in Pif1-deficient cells . By two assays , Pif1 bound preferentially to long telomeres . Together , these data can be explained if Pif1 is more likely to remove telomerase from those telomeres that are least in need of lengthening .
Pif1 uses its ATPase activity to evict telomerase from telomeres [22] , which can explain how it suppresses telomere lengthening . Pif1 also inhibits telomere addition to DSBs [19 , 20] . To determine if Pif1 affects telomerase binding to DSBs , we used chromatin immuno-precipitation ( ChIP ) to monitor the association of Est1 and Est2 to an induced DSB in the presence and absence of Pif1 ( Fig 1 ) . These experiments were carried out in a strain with a galactose-inducible HO endonuclease and an HO recognition site ~13 kb from the end of chromosome VII-L , the only accessible HO site in the strain ( Fig 1A ) [26] . We used a strain with an 80-bp tract of TG1-3 telomeric DNA ( TG80 , Fig 1A grey box ) adjacent to the HO site to increase the rate of de novo telomere addition [26] . Cells also expressed a Myc-tagged protein ( Est2 , Est1 , or Cdc13 ) . Experiments were carried out in PIF1 and pif1-m2 versions of the strain , where pif1-m2 cells express wild type ( WT ) levels of mitochondrial Pif1 and reduced nuclear Pif1 [18] . Although pif1-m2 retains some nuclear function , we used it because pif1-m2 cells progress through the cell cycle similarly to WT cells , unlike pif1Δ cells which progress more slowly owing to their reduced mitochondrial function [19] . In addition , pif1Δ cells are very hard to synchronize . The efficiency of cleavage at the HO site , which was monitored by Southern blotting , was not affected by Pif1 levels ( ~65–80% cutting in both WT and pif1-m2 cells; S1 Fig ) . HO cleavage was induced in WT or pif1-m2 cells that were first arrested in late G1 phase with Δ-factor . After HO action , cells were released from the G1 phase arrest . In both strains , ChIP samples were taken in G1 arrested cells both before ( “0—gal” time points ) and after HO induction and throughout the synchronous cell cycle that occurred upon removal from Δ-factor ( 0–90 min time points; 24°C ) . Samples from each time point were also assessed by FACS to determine cell cycle position , which demonstrated that pif1-m2 and WT cells moved similarly through the cell cycle ( Fig 1B ) . ChIP samples were quantified using real-time PCR and normalized to input DNA . In this and other ChIP experiments , results are presented as fold enrichment of binding to the DSB compared to binding to a control site ( ARO1 ) . We examined binding of Est2 ( Fig 1C ) , Est1 ( Fig 1D ) and Cdc13 ( Fig 1E ) to the HO break site in WT and pif1-m2 cells . None of the three proteins was associated with the HO recognition site before HO induction ( Fig 1C–1E , 0—gal time point ) . In both strains , Est2 binding to TG80-HO was at background levels in G1 and early S phase ( Fig 1C , WT closed squares; pif1-m2 , open squares , 0–30 min time points ) . In both strains , high levels of Est2 binding to the DSB were detected from mid-S phase through the end of the cell cycle ( 45–90 min time points ) . However , average Est2 binding was over four times higher in pif1-m2 compared to WT cells ( Fig 1C , open circles ) . Est1 showed a similar pattern of binding to the DSB , occurring at background levels in G1 and early S phase with strong binding from mid-S phase to the end of the cell cycle ( Fig 1D ) . Although Est1 binding was significantly higher in pif1-m2 versus WT cells from mid-S to the end of the cell cycle , the difference was fairly modest ( ~1 . 5–2 . 5-fold over WT levels; see Fig 1 legend for p values ) . As reported previously [12] , Cdc13 showed strong association with TG80-HO after HO induction . Cdc13 enrichment at TG80-HO was ~100 to 150-fold over background during G1 phase and steadily increased to ~400- to 600-fold over background by the end of the cell cycle . However , unlike Est2 and Est1 , Cdc13 binding to TG80-HO was not affected significantly by reduced Pif1 ( Fig 1E , open squares ) . We conclude that Pif1 affects telomere addition to DSBs by reducing telomerase binding to breaks , as it does at telomeres . This effect is specific for telomerase , as binding of Cdc13 was unaffected by Pif1 at either telomeres [22] or DSBs ( Fig 1E ) . Pif1-mediated removal of telomerase could regulate telomere length by affecting the frequency of telomere addition or telomerase processivity ( or both ) . To distinguish among these possibilities , we used the Single Telomere Extension ( STEX ) assay [7] , which analyzes lengthening at individual telomeres at nucleotide resolution over a single cell cycle in diploid PIF1 and pif1Δ cells ( Fig 2 ) . Freshly dissected pif1Δ spore clones were used quickly after dissection from a heterozygous diploid to minimize the negative effects of mitochondrial DNA loss on growth rate . Telomerase-deficient tlc1Δ PIF1 or tlc1Δ pif1Δ spore clones ( recipient cells ) were mated to telomerase proficient TLC1 PIF1 or TLC1 pif1Δ cells ( donor cells ) , respectively , to restore telomerase activity in the resulting diploids . To distinguish recipient from donor telomeres , URA3 was integrated adjacent to the left telomere of chromosome VII in the recipient strain . Mating efficiencies and cell cycle progression , which were determined by flow cytometry ( Fig 2A ) , were similar in the two crosses ( Fig 2A ) . ~93% of the recipient cells mated in both crosses ( Fig 2A ) and replication of the diploid genomes occurred by 5 h post-mixing ( 15 . 5% out of an expected 16 . 7% of cells had 4C DNA content at this time; see Fig 2A legend ) . Thus , the null status of PIF1 did not affect growth rates or mating efficiency during the small number of generations required to obtain sufficient cells for this experiment . The URA3-marked VII-L telomeres derived from recipient cells were amplified , cloned , and sequenced . Because yeast telomerase adds irregular TG1-3 repeats [27] , pre-existing telomeric DNA can be distinguished from the newly added sequence [7 , 28] . Prior to mating , we measured telomere lengthening in the parental haploid tlc1Δ PIF1 ( Fig 2B , top ) and tlc1Δ pif1Δ ( Fig 2B , bottom ) spore clones to obtain an estimate of the frequency of lengthening in the absence of telomerase , which likely occurs via recombination . In both strains , many of the short telomeres ( ≤50 bps ) were lengthened in the absence of telomerase . Among longer telomeres ( ≥50 bps ) , the fraction of telomeres lengthened ( 3 . 7% and 4 . 8% ) by telomerase-independent events and the average amount of telomeric DNA added ( 39 and 35 nt ) were similar in , respectively , tlc1Δ and tlc1Δ pif1Δ haploid cells . To limit analysis to telomeres lengthened by telomerase , we considered only telomeres longer than 50 bps for further analysis . Next , we prepared DNA from the diploid cells at the end of the first post-zygotic S phase in both WT ( Fig 2C , top ) and pif1Δ ( Fig 2C , bottom ) cells . The elongation frequency for telomere VII-L in tlc1Δ pif1Δ/TLC1 pif1Δ was 28% for telomeres over 50 bp . In contrast , only 7 . 2% of the telomeres were lengthened in WT cells , a significant difference ( see Fig 2 legend for p values ) . Moreover , the average length of telomeric sequence added to individual telomeres in tlc1Δ pif1Δ/TLC1 pif1Δ cells was 72 nt , a value significantly greater than 45 nt that was added in PIF1 cells ( see Fig 2 legend for p values ) . The value obtained in PIF1 WT ( tlc1Δ/TLC1 ) cells in our experiments was the same as reported previously for STEX in tlc1Δ/TLC1 cells [7] . Taken together , these data suggest that Pif1 reduces both the frequency and processivity of telomerase action . In earlier work , we did not see a difference in Est2 binding to bulk telomeres in asynchronous cells [12] . To explore the effects of Pif1 on telomerase binding , we combined an inducible short telomere system [6] with ChIP in cells going synchronously through S phase ( see [8] and Fig 3 ) . The experimental strains used in these experiments had a modified chromosome VII-L in which a cassette containing telomeric repeats flanked by recognition sites ( FRT ) for the Flp1 site-specific recombinase was positioned directly adjacent to the telomere ( Fig 3A left , experimental strain ) , as well as an integrated copy of a galactose-inducible FLP1 . The control strains were isogenic to the experimental strains except that the cassette at modified chromosome VII-L had no telomeric DNA between the two FRT sites [6] ( Fig 3A , right , control strain ) . In both strains , FLP1 expression was induced by adding galactose , which caused recombination between the two FRT sites . In the experimental strain , FLP action resulted in loss of the telomeric DNA between the sites , thus generating a short ~100 bp telomere at chromosome VII-L that is preferentially lengthened by telomerase for multiple cell cycles [6] . The telomere adjacent to the control cassette is maintained at the strain characteristic length ( ~300 bp in WT cells ) regardless of whether or not it is acted upon by Flp1 ( Fig 3A ) . Both the experimental and control strains expressed a Myc-tagged protein ( Est1 , Est2 or Yku80 ) . These experiments were carried out in parallel in WT and pif1-m2 versions of both the experimental and control strains . Cells were synchronized , and ChIP samples were prepared and quantified as described for the DSB experiments . Galactose was added to G1-arrested cells to induce expression of Flp1 , which caused recombination at the VII-L telomeric region in all strains , resulting in either a short VII-L telomere ( experimental strain ) or a VII-L telomere that was the same length as the other telomeres in the cell ( control strain ) . The efficiency of recombination was similar in all strains ( S2A Fig ) . After Flp1 action , cells were released from G1 arrest ( time 0 ) , and samples were removed for analysis throughout the first synchronous cell cycle ( 15 to 90 min ) . The efficiency of recombination , which was monitored in each experiment ( S2A Fig ) , was equivalent ( ~75% ) in WT and pif1-m2 cells for both the experimental and control strains . Using the same system , previous studies found that Est2 and Est1 binding was approximately two times higher on the short VII-L telomere versus the WT length VI-R telomere in the same cells or the WT length VII-L or VI-R telomeres in the control strain [8 , 9] . For this study , we recapitulate the preferential binding of Est2 and Est1 to the short VII-L telomere in WT cells ( Figs 3B and S3 ) . Notably , preferential Est2 binding to the shortened VII-L telomere was lost in pif1-m2 cells ( Fig 3B , upper panel , open circles ) , even though the amount of Est2 binding to the unaltered VI-R telomere in the same cells was unchanged ( Fig 3B lower panel , see legend for p values ) . In addition , levels of Est2 binding were the same in pif1-m2 and WT versions of the control strain at the VII-L telomere ( Fig 3C , upper panel ) . Similar results were obtained for Est1 ( S3 Fig ) . To determine if Pif1 acts preferentially on telomerase binding or alternatively affects all telomere-binding proteins , we examined Myc-tagged Ku80 telomere association as a function of Pif1 ( Fig 3D ) . Consistent with earlier results [8] , Yku80 bound robustly and equally well to short ( Fig 3D , top panel , filled diamonds ) and WT length ( Fig 3D , bottom panel , filled diamonds ) telomeres in WT cells . In contrast to Est2 and Est1 , levels of Yku80 telomere binding in pif1-m2 cells were not significantly different from WT binding at either the short ( Fig 3D , top panel , open diamonds ) or WT length telomeres ( Fig 3D , lower panel , open diamonds ) . Thus , higher telomerase binding to short telomeres is Pif1 dependent . This result is unlikely to be an artifact of telomeres being longer in pif1-m2 than in WT cells , as the level of binding of telomerase to the VI-R telomere in both the control and experimental strains and to the VII-L telomere in the control strain was similar in WT and pif1-m2 cells ( Fig 3 ) . Likewise , levels of Yku80 binding were not affected by the presence of Pif1 ( Fig 3D ) . In WT cells , telomerase preferentially binds [8 , 9] and lengthens [6 , 7] short telomeres . However , this binding preference was not detectable in pif1 cells ( Figs 3B and S3 ) . Because Pif1 uses its ATPase activity to remove telomerase from DNA ends [22] , these results could be explained if Pif1 preferentially removes telomerase from WT length and/or long telomeres . This model predicts that the average length of telomeres that are Pif1-associated will be longer than the average for bulk telomeres . To test this possibility , we determined the average size of telomeric DNA in an anti-Pif1 immuno-precipitate ( Fig 4 ) . We epitope-tagged a mutant version of Pif1 , Pif1-K264A , in which the invariant lysine in the Walker A box is mutated to alanine . Although Pif1-K264A is helicase dead in vivo and in vitro [18] , it binds single-stranded DNA as well as WT Pif1 [22] . We used Pif1-K264A because it gives a stronger signal in ChIP than WT Pif1 , probably because it is trapped at its binding sites [29] . However , pif1-K264A cells have long telomeres , and the experiment to determine the lengths of Pif1-bound telomeres must be done in cells with WT telomere length . Because the effects of pif1-K264A on telomere length are recessive [18] , we did the experiment in a heterozygous diploid ( pif1-K264A-Myc/PIF1 ) . As a control , we monitored the lengths of telomeres in an anti-Yku80 immuno-precipitate in a pif1-K264A/PIF1 diploid background . Chromatin was prepared from both diploid strains and processed for PCR amplification either before ( input samples ) or after ( ChIP samples ) immuno-precipitation . After C-tailing the purified DNA , PCR primers were used to amplify either the VI-R ( Fig 4A ) or XV-L ( Fig 4B ) telomeres . The PCR-amplified DNA in the input and immuno-precipitated samples was gel separated ( Fig 4A and 4B ) , and the average sizes of the DNA in these samples were determined using an AlphaImager 3400 Molecular Weight Analysis program ( Fig 4A , VI-R; B , XV-L ) . For both the VI-R and XV-L telomeres , the average size of telomeric DNA in the anti-Pif1 immuno-precipitate was significantly longer in the ChIP samples compared to input DNA ( the average size in bps ± SD of telomeric DNA in the ChIP samples was 337 . 1± 2 . 0 for VI-R and 346 . 2 ± 3 . 4 for XV-L; the average size of input telomeric DNA was 272 . 7 ± 1 . 1 for VI-R and 274 . 0 ± 2 . 0 for XV-L; Fig 4C and 4D; see Fig 4 legend for p values ) . As expected for the strain expressing Yku80-Myc [12] , the average sizes in bps of telomeric DNA in the ChIP and input samples were indistinguishable ( averages ± SD for ChIP samples were 273 . 3 ± 0 . 8 and 275 . 1 ± 1 . 3 , respectively , for telomeres VI-R and XV-L ) . Thus , Pif1 binds preferentially to long telomeres . This effect is specific for Pif1 because Yku80 binding showed no length preference in the same assay ( Fig 4A–4D ) . To extend the finding that Pif1 binds preferentially to longer telomeres , we monitored Pif1-Myc binding to short and WT length telomeres using the inducible short telomere system . In both the experimental and control strains , Pif1 telomere binding occurred in S phase . However , in different experiments , the peak of Pif1 binding to a given telomere occurred at different times in S phase . Therefore , for each telomere in both strains , the ChIP binding is presented as average binding from three independent experiments in G1 phase ( pooling 0 , 15 and 30 min time points ) , S phase ( pooling 45 and 60 min time points ) , and G2 phase ( pooling 75 and 90 min time points ) ( Fig 4E and 4F ) . In both , the experimental and control strains , Pif1 binding to all telomeres was low in G1 and G2 phase . In the experimental strain , Pif1 binding was similarly low in S phase to short VII-L . In contrast , binding of Pif1 to the WT length VI-R telomere in both the control and experimental strains and to the WT VII-L telomere in the control strain was three times higher in S phase ( Fig 4E and 4F ) . Thus , using a very different assay , Pif1 binds more robustly to longer telomeres .
In S . cerevisiae and mammals , short telomeres are preferentially elongated by telomerase [6 , 7 , 30] . This preference is explained in part by proteins that bind preferentially to short telomeres , such as the MRX complex [12] and Tel1[8 , 9 , 11] . The presence of these ( and other ) proteins in combination with low levels of Rif2 result in the preferential binding of telomerase to short telomeres [8 , 12 , 15] . Here , we show that telomere binding of Pif1 , a negative regulator of telomerase , is also length dependent with longer telomeres having higher binding ( Fig 4 ) . In vivo and in vitro , Pif1 removes telomerase from DNA ends without affecting binding levels of telomere structural proteins [22] . Here we show that Pif1 acts similarly at DSBs ( Fig 1 ) because in the absence of Pif1 , Est1 ( Fig 1D ) and especially Est2 ( Fig 1C ) bound at higher levels to an induced DSB . In contrast , Cdc13 binding to the break ( Fig 1E ) was not Pif1-sensitive . Thus , the ability of Pif1 to inhibit telomere addition to spontaneous DSBs [19 , 20] can be explained by Pif1 removing telomerase from these breaks . Pif1-mediated removal of telomerase could affect the frequency of telomerase action , its processivity , or its preference for short telomeres . We used two different methods to determine the impact of Pif1 on these events . STEX experiments indicate that telomerase is more processive in vivo in the absence of Pif1 , as telomerase added an average of 72 nt ( pif1Δ ) versus 45 nt ( WT ) to the VII-L telomere in a single cell cycle ( Fig 2 ) . Consistent with this finding , Pif1 reduces telomerase processivity in vitro [22] , and telomeres are longer in pif1-m2 and pif1Δ cells [19] . STEX also revealed that the fraction of telomeres acted upon by telomerase in vivo was almost four times higher in pif1Δ ( 28% ) versus WT ( 7 . 2% ) cells ( Fig 2 ) . In contrast , in vitro , Pif1-dependent release of Est2 from telomeric oligonucleotides increases the fraction of elongated oligonucleotides by freeing Est2 from its original substrate [22] to which it is otherwise tightly bound [31] . The fact that the fraction of telomeres lengthened in the presence of Pif1 is lower in vivo and higher in vitro is likely explained by the high concentration of telomeric oligonucleotides in vitro versus the small number of telomeres in cells . In addition , as Est1 and Est3 binding to telomeres is limited to a short period late in S phase [4 , 5] , telomerase action occurs only during a narrow window of the cell cycle [2 , 3] while the standard in vitro system is Est1 and Est3-independent [32] . To determine if Pif1 affects the preferential lengthening of short telomeres , we used an assay in which a single short telomere is induced in cells with otherwise WT length telomeres [6] . This assay was used previously to show that in WT cells , Est2 and Est1 bind preferentially to short telomeres [8 , 9] . However , in pif1-m2 cells , we see similar levels of Est2 and Est1 at short and WT length telomeres ( Figs 3B and S3 ) . This loss of preferential binding of Est2 and Est1 to the short VII-L telomere in pif1-m2 cells is unlikely to be a consequence of all telomeres being longer in pif1-m2 cells [18 , 19] . Even though the average length of the “short” VII-L telomere ( ±SD ) ( 196 ± 11 . 7 bp ) was longer in pif1-m2 versus WT ( 124 . 8 ± 7 . 2 bps ) cells ( S2B Fig ) , 196 bp is still short enough to be preferentially lengthened by telomerase in WT cells [7] . Moreover , levels of Est2 binding to the control VI-R telomere were not affected by Pif1 ( nor was Yku telomere binding ) ( Fig 3C and 3D ) . Thus , our results cannot be explained by its being more difficult to ChIP Est2 to longer telomeres . Finally , in the control strains the difference in the average post-recombination lengths of the VII-L telomeres in WT versus pif1-m2 cells was even larger ( 71 bp difference in experimental versus 78 bp in control strains ) ( S2B Fig ) , yet the levels of Est2 association in the control strains were not PIF1-sensitive ( Fig 3C and 3D ) . Together with the STEX results , our data demonstrate that Pif1 contributes to the preferential targeting of telomerase to short telomeres . If Pif1 binds more readily to long telomeres , it could explain how Pif1 contributes to preferential telomerase activity at short telomeres . To test this possibility , we determined the average length of Pif1-associated telomeres . For both , the VI-R and XV-L telomeres , Pif1-associated telomeres were about 70 bps longer than bulk telomeres ( Fig 4D ) . In contrast , using the same assay , Yku80-associated telomeres were the same length as bulk telomeres in the two strains . In addition , in S phase , Pif1 bound better to WT length telomeres than to a very short telomere ( Fig 4E and 4F ) . These data argue that Pif1 contributes to the selective lengthening of short telomeres by binding to and removing telomerase preferentially from longer telomeres . This finding is interesting in light of an in vitro study that used single molecule analyses to show that Pif1 was better able to displace telomerase from substrates with long TG1-3 single-strand tails [33] . If longer telomeres have longer G-tails in vivo , the two observations may be linked . Rif2 , another negative regulator of yeast telomerase , also binds to a greater extent at wild type than to short telomeres [12] . Thus , preferential lengthening of short telomeres is achieved by proteins like Pif1 and Rif2 that act preferentially at longer telomeres and proteins like MRX , Tel1 , and Tbf1 that act preferentially at shorter telomeres . It is tempting to speculate that Rif2 recruits Pif1 preferentially to long telomeres . Pif1 efficiently binds to and unwinds G4 structures in vitro and suppresses DNA damage at G4 motifs in vivo ( reviewed in [16]; also , [23–25] ) . Thus , a unifying model for Pif1 action is that it inhibits telomerase by unwinding telomeric G4 structures . However , intra-molecular G4 structures inhibit telomerase [34] . Thus , it is difficult to attribute the inhibitory effects of Pif1 on telomerase to its G4-unwinding activity . However , the presence of G4 structures might stimulate Pif1’s ability to displace telomerase as it does Pif1 unwinding of duplex DNA [35] . Another possibility is that Pif1 uses its ATPase activity to displace the protein subunits of telomerase from DNA ends , the type of protein eviction activity attributed to the closely related S . cerevisiae Rrm3 and the S . pombe Pfh1 helicases during chromosome replication [36 , 37] . Although we can not rule out this model , it seems unlikely as Pif1 does not affect Cdc13 [22] or Yku ( Fig 3D ) telomere binding . Rather we favor a model where Pif1 removes telomerase from telomeres by disrupting the telomerase RNA-telomeric DNA intermediate as Pif1 unwinds RNA/DNA hybrids very efficiently [38 , 39] . In summary , the Pif1 helicase affects multiple aspects of the telomerase reaction: it reduces telomerase processivity , the frequency of telomere elongation , and the preference of telomerase for short telomeres . All of these effects could be a result of Pif1 binding preferentially to ( and hence telomerase removal from ) telomeres that are longer than average in length . Because yeast telomerase is not abundant , with fewer telomerase complexes than telomeres [5 , 40 , 41] , this regulation is important to ensure that those telomeres most in need of lengthening receive telomerase .
Strains and primers are presented in , respectively , Tables 1 and 2 . Experiments were carried out in RAD5+ versions of W303 , unless otherwise indicated . Deletions eliminated entire ORFs . The pif1-m2 and pif1-K264A strains were made as in [19] . Epitope tagging to generate Est1-G8-Myc9 , Est2-G8-Myc18 , Cdc13-Myc9 , Yku80-G8-Myc18 , Pif1-Myc13 and Pif1-K264A-Myc13 was carried out as described [8 , 29 , 42 , 43] . Each Myc-tagged protein was expressed from its endogenous locus and promoter . Except for the experiments using Myc-Pif1-K264A , the tagged protein was the only form of the protein in cells . STEX strains contained URA3 adjacent to the VII-L telomere [44] . The strains used to analyze telomerase binding as a function of telomere length contained a galactose-inducible FLP1 and a modified chromosome VII-L for excision of telomere-adjacent DNA [6] . Strains for DSB experiments had a galactose-inducible HO gene and a modified chromosome VII-L with the HO endonuclease recognition site between ADH4 and MNT2 [26] . A heterozygous ( pif1-K264A/PIF1 ) diploid strain in which either Pif1-K264A or Yku80 was epitope tagged was used to determine the lengths of Pif1-associated telomeres . For experiments using galactose , cells were grown in 3% raffinose prior to induction in 1% galactose . For HO experiments , cells were maintained on media lacking lysine to prevent leaky expression from the HO gene . Recipient cells were prepared from freshly dissected spore clones from JP192 ( tlc1Δ::HIS3 PIF1 ) or JP223 ( tlc1Δ::HIS3 pif1Δ::TRP1 ) and grown to OD660 0 . 2 . Next , 6 x 107 recipient cells were mated with 1 x 107 of either WT or pif1Δ::TRP1 donor cells freshly dissected from JP192 or JP223 . Mating mixtures were filtered onto four membranes ( Microfil S , Millipore ) , placed onto pre-warmed YEPD plates , and incubated at 30°C for 3 h [7] . Cells were resuspended in 30 ml synthetic media lacking histidine ( YC-his ) and then added to 90 ml of fresh YC-his . Fifty ml resuspended mating cells were taken immediately for analysis; the remaining cells were incubated at 30°C with shaking at 160 rpm . After 2 h ( 5 h after initial mixing ) , an additional 50 ml were taken for analysis . Mating efficiency was monitored by flow cytometry by determining the proportion of cells containing >2N DNA content after correcting for the ratio of recipient to donor cells in the mating mixture ( i . e . , 100% mating of recipient cells results in 16 . 7% of total cells with >2N DNA content ) . Reported experiments had >85% mating efficiency . DNA was isolated using Masterpure Yeast DNA Purification Kits ( Epicentre ) . Telomere PCR was performed using 400 ng of DNA as described [7] , except that primers TEL7L and PolyG18 , specific to the URA3-marked telomere VII-L , were used for amplification ( Table 2 ) . PCR products were ligated into the pDRIVE vector ( Qiagen ) as per the manufacturer’s instructions . Plasmid inserts were sequenced ( Genewiz , South Plainfield , NJ ) using the SP6 primer . Sequences were aligned by eye and analyzed using the contig assembler of Vector NTI ( Invitrogen ) . Where indicated , Fisher’s Exact tests ( which test potential relationships between categorical variables ) and Mann-Whitney U tests ( non-parametric tests of significance for two independent sample groups ) were used to determine statistical significance . Strains used for analysis of protein association at either a short telomere or a double strand break were grown overnight at 30°C in rich media plus 3% raffinose to OD660 0 . 15 . Synthetic Δ-factor was added ( final concentration 160 μM ) , and cells were incubated at 30°C for ~2 h or until > 90% of cells were unbudded ( G1 phase ) . Samples were then taken for flow cytometry , Southern blot analysis , and ChIP ( “0-gal” time point ) . Dry galactose ( Sigma ) was added to the rest of the culture ( final concentration 1% ) to induce either the FLP recombinase or the HO endonuclease; cells were incubated at 30°C for 3 h [8] . Cells were filtered and then resuspended in YEPD with a final concentration of 160 μM Δ-factor and incubated for 15 min at 24°C . The cells were filtered again and resuspended in rich medium containing glucose ( no Δ-factor ) and 170 μg/ml P6911 Protease ( Sigma ) and released into the cell cycle at 24°C . Samples were taken every 15 min for one cell cycle and analyzed by flow cytometry , Southern blot , and ChIP [45] . ChIP was performed as described [46] , except that cells were lysed by adding 10 μl 5 mg/ml Zymolyase 100T ( MP Biomedicals ) to thawed samples , which were then incubated at 37°C for ~10 min or until visual inspection showed >90% lysis . Myc-tagged proteins were immunoprecipitated using an Δ-Myc monoclonal antibody ( Clontech ) . Immunoprecipitated DNA was quantitated by real-time PCR on an iCycler iQ Real-Time PCR Detection System ( Bio-Rad Laboratories ) as in [45] , except that primers were used to amplify either a region adjacent to a short telomere on chromosome VII-L ( primers FRT+ and FRT- [8] ) and WT telomere VI-R ( primers VI-R+ and VI-R- ) or a region adjacent to an HO recognition site ( primers FTM7 and RTM7 [26] ) . ChIP samples were normalized to inputs; data are presented as fold enrichment relative to a non-telomeric control site on chromosome IV ( ARO1 , using primers RT-ARO+ and RT-ARO ) . Each synchrony was repeated a minimum of three times to obtain an average enrichment value . Error bars represent the standard deviation from three or more independent experiments . A two-tailed students t-test was used to determine statistical significance ( p≤0 . 05 ) . After ChIP from asynchronously growing diploid strains , telomere PCR was performed with minor modification of the methods in [12] . Briefly , ChIP samples were C-tailed according to manufactures instructions using T4 polynucleotide kinase ( Invitrogen ) . PCR conditions were 58°C annealing for 30 s and 72°C extension for 45 s . The PCR products were separated on a 1 . 8% ( w/v ) MetaPhor ( Lonza ) agarose gel . A subset of the DNA was cloned and sequenced to establish that it was indeed telomeric DNA: of the 90 sequenced clones , 89% were telomeric DNA while the remainder had little or no insert DNA . The AlphaImager 3400 Molecular Weight Analysis program was used to determine the average telomere length in each sample . Experiments were performed in three biological replicates; telomere length was determined at two different telomeres ( VI-R and XV-L ) . | Telomerase , the enzyme that maintains telomeres , preferentially lengthens short telomeres . The baker’s yeast Pif1 DNA helicase inhibits both telomerase-mediated lengthening of existing telomeres and the formation of new telomeres at double strand breaks . By virtue of its ATPase activity , Pif1 reduces the level of telomerase binding to telomeres . Here , we report that the association of the telomerase subunits Est2 and Est1 at a DNA break was increased in the absence of Pif1 , suggesting that Pif1 affects telomere length and new telomere formation by similar mechanisms . In cells lacking Pif1 , Est2 and Est1 no longer bound preferentially to short telomeres , a larger fraction of telomeres was lengthened and the amount of telomeric DNA added per telomere was increased compared to wild type cells . Furthermore , by two different assays , Pif1 bound preferentially to long telomeres in vivo . Thus , preferential lengthening of short telomeres is achieved in part by targeting Pif1 , a negative regulator of telomerase , to long telomeres . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | The Pif1 Helicase, a Negative Regulator of Telomerase, Acts Preferentially at Long Telomeres |
Schools of fish and flocks of birds can move together in synchrony and decide on new directions of movement in a seamless way . This is possible because group members constantly share directional information with their neighbors . Although detecting the directionality of other group members is known to be important to maintain cohesion , it is not clear how many neighbors each individual can simultaneously track and pay attention to , and what the spatial distribution of these influential neighbors is . Here , we address these questions on shoals of Hemigrammus rhodostomus , a species of fish exhibiting strong schooling behavior . We adopt a data-driven analysis technique based on the study of short-term directional correlations to identify which neighbors have the strongest influence over the participation of an individual in a collective U-turn event . We find that fish mainly react to one or two neighbors at a time . Moreover , we find no correlation between the distance rank of a neighbor and its likelihood to be influential . We interpret our results in terms of fish allocating sequential and selective attention to their neighbors .
Collective motion phenomena such as swarming , flocking and schooling behavior have been observed in a large variety of animal species ranging from bacteria to humans [1] . Several theoretical models have been proposed to explain how such large scale coordination patterns emerge from “microscopic level” interaction rules among individual animals [2–7] . These models have been instrumental in improving our understanding of collective motion in real animal groups by providing an indication of which interaction mechanisms are sufficient to reproduce realistic patterns of collective behavior . In particular , most models agree on the fact that two types of interaction are responsible for maintaining group cohesion to achieve coherent collective motion: attraction and alignment . More recent improvements in remote sensing and video-tracking technologies [8–10] have made possible to automate data collection and test directly theoretical models against highly resolved empirical movement data in various species . Generally , these studies have confirmed the importance that attraction and alignment behavior play in the formation and maintenance of collective movement patterns [11–15] . However , there is a less clear scientific consensus about how these interaction rules are implemented in the sensory-motor responses of individuals . This lack of agreement underscores the importance of answering the following question: how do individuals mediate interactions with multiple neighbors ? [16] . Specifically , theoretical studies have postulated a number of factors that are likely to affect the probability and intensity of interactions: distance ( metric neighborhood ) [2–7] , position rank ( topological neighborhood ) [17] , projected size ( visual neighborhood ) [18–20] , and spatial arrangement around a focal individual ( Voronoi neighborhood ) [13] . Each of these different definitions of influential neighborhood is supported to some extent by computational models and empirical observations . Rather than siding with one or more of the proposed neighborhood definitions , we adopt a fully data-driven approach with minimalist modeling assumptions . The simplest hypothesis consists of assuming that fish copy the actions of their neighbors , but not instantaneously: the fish reaction takes time to process sensory information and to trigger the appropriate behavioral response . Those assumptions impose a temporal constraint given by the sequential occurrence of the perception of the neighbors’ actions , and the movement response [21 , 22] . We thus assume that animals following a particular neighbor in a new direction are subject to a time-delay when copying the heading of influential neighbors . Considerable work has already appeared on the identification of these time-delays . The delays with which individuals align with each other have in fact been exploited to determine social hierarchies in animal groups , as shown , e . g . , for pigeon flocks [23] , where the leadership network is constructed with link weights given by the delay for which pairwise angle correlation is maximal . Improvements on how to identify such delays from movement data have proposed the use of time-dependence in pairwise angle correlation [24] . A computational analysis , based on similarities between trajectories ( Fréchet distance ) , has also been proposed and implemented in a visual analytic tool [25] . A different approach has made use of a time-ordering procedure on the pairwise angle correlation to determine temporary leader/follower relations in foraging pairs of echolocating bats [26] . The analysis of the bat trajectories was instrumental in identifying transient leadership and coupling it to sensory biases of the species . However , only pairs of individuals were considered and group influence on individual behavior was not investigated . Since identifying influential neighbors is key to unravel the mechanisms of interaction , there is a need in collective behavior studies to establish transient leadership from the dynamics of the individual trajectories . One way to bridge this gap consists of determining who are those influential individuals whose heading is being copied more closely by others , how many of such influential neighbors exist , and where are located in the group . Fish have the ability to choose not only when to copy the heading of another individual , but also the extent to which this heading is copied , that is the similarity and the pace at which fish match the trajectory’s curvature of another individual [11 , 27] . The closer two ( or more ) fish are to this matching , the more aligned they are ( even if with some delay ) , and the more faithfully they are following the movement path of the transient leader . Here , we introduce a procedure that allows us to identify the influential neighbors of fish moving in a group , and we test it along a series of experiments in groups of two and five individuals of the freshwater tropical fish Hemigrammus rhodostomus swimming in a ring-shaped tank ( see details in Materials and methods ) . In this set-up , fish swim in a highly synchronized and polarized manner , and can only head in two directions , clockwise or anticlockwise , regularly switching from one to the other . We base our procedure for identifying influential neighbors on time-dependent directional correlations between fish , focussing our analysis on the interactions that occur during these collective U-turns . Indeed , during U-turns , fish have to make a substantial change of direction to reverse their heading , making easier the extraction of the correlation resulting from the direct interactions between individuals rather than other incidental correlations , e . g . , their channeled motion in the ring-shaped tank . Moreover , as correlation does not imply causal influence , we need to control for potential spurious correlations . We do so by constructing a null model of collective U-turns to show that the patterns of interaction observed in the experiments are not due to random processes .
Hemigrammus rhodostomus performs burst-and-coast swimming behavior that consists of sudden heading changes combined with brief accelerations followed by quasi-passive , straight decelerations [15] . Moreover , fish spend most of their time swimming in a single group along the wall of the tank . Fish regularly change their position within the group [28] , so that every individual fish can be found at the front of the group . A typical collective U-turn event starts with the spontaneous turnaround of a single fish ( hereafter called the initiator ) , mostly located at the front of the group [28] . This sudden change of behavior triggers a collective reaction in which all the other individuals in the group make a U-turn themselves , so that , after a short transient , all individuals adopt the same final direction of motion as the initiator . Overall , we analyzed 1586 U-turns of which 1111 were observed in groups of 2 fish and 475 in groups of 5 fish . Fig 1 shows two examples of collective U-turns in groups of N = 2 ( left column , panels ABC ) and N = 5 fish ( right column , panels DEF; see also supplementary S8 Fig and supplementary S1 and S2 Videos in the Supplementary Information ) . Fig 1A shows a first fish F1 ( red color ) swimming close to the upper-left region of the tank , followed by a second fish F2 ( purple color ) at a distance d12 ≈ 8 . 5 cm , swimming in the same direction . Right before the U-turn starts ( Fig 1A ) , fish F1 reduces its speed ( circles become closer to each other ) , the distance d12 decreases ( to ≈ 5 . 1 cm ) , and F2 also reduces its speed . Then , both fish perform a change of direction which lasts about 1 second and during which fish F2 clearly follows fish F1 ( see the corresponding circles at each instant of time in Fig 1B ) . Once the U-turn is completed ( Fig 1C ) , F1 accelerates again , and so does F2 , which also adopts the direction of motion of F1 . The distance d12 increases again ( ≈ 9 . 5 cm ) , due to the larger velocities , and remains of the same order along the depicted trajectory . The situation is less clear when we try to describe collective U-turns in larger groups . Fig 1D , 1E and 1F show a collective U-turn for the case where N = 5 . Before the U-turn , fish F2 ( orange ) seems to be the fish that the rest of the group follows , the first circle of its trajectory being the most advanced one in the direction of motion . In fact , a position order can be inferred from Fig 1D: F2 , F3 , F5 , F1 and F4 . However , it is rather complicated to extract from Panel E a precise information about which fish is the initiator of the U-turn , in which order the other fish follow , and therefore , who is influencing whom , especially if time-delays and reaction times are taken into account . The same happens with the information about fish’s positions after the U-turn , provided by Panel F . In order to describe rigorously the individual behavior of the N fish during a U-turn , we introduce the angle ϕi ( t ) as an instantaneous measure of the direction of motion of a fish Fi; see Fig 2 . We assume that the instantaneous heading of a fish Fi can be defined in terms of the velocity vector v → i ( t ) , so that v → i = ( cos ϕ i , sin ϕ i ) ∥ v → i ∥ . The heading of a fish ϕi allows us to characterize the angle of incidence of the fish relative to the wall , θwi = ϕi − ψi , where ψi is the angle formed by the position vector of the fish with the horizontal line ( see Fig 2 ) . The angle of incidence θwi is an individual measure that doesn’t depend on the heading of another fish . When a fish Fi is swimming along the wall , the value of θwi is around ±90° ( we choose , by convention , the positive sign for the anticlockwise angle ) . In our experiments , most of the time the absolute value of the angle of incidence is close to 90°; equivalently , |sin ( θwi ( t ) ) | ≈ 1 . When the motion is perpendicular to the wall , the incidence is zero if the fish points towards the wall ( θwi = 0° ) , and maximal if the fish points towards the center of the tank ( θwi = 180° ) ; in both cases , sin ( θwi ( t ) ) = 0 . The change of sign of angle θwi can serve as an indicator that a U-turn has taken place . In fact , this allows us to delimit the individual U-turns with precision and , consequently , to determine the start and the end of a collective U-turn . We define the start and end times ts , i and te , i of the individual U-turn of fish Fi in terms of the absolute value of the angle of incidence , |θwi ( t ) | . Once a U-turn has been detected , we obtain the time ts , i at which |θwi ( t ) | has decreased ( from approximately 90° ) below a given threshold θ ¯ s , and the time te , i at which |θwi ( t ) | has increased again and is above another given threshold θ ¯ e ( see Materials and methods for more details ) . Thus , the start of a collective U-turn is determined by the time ts at which the first individual U-turn starts , while the end of a collective U-turn is given by the time te at which the last individual U-turn finishes . That is: t s = min i = 1 , … , N { t s , i } , t e = max i = 1 , … , N { t e , i } . ( 1 ) For each collective U-turn , we have made a convenient time shift so that ts = 0 . Then , te denotes not only the end time but also the duration of the collective U-turn . We also introduce an instantaneous measure of how similar the direction of motion of individual fish are across the group . We define the instantaneous group polarization P ( t ) as the following function of normalized fish velocity vectors: P ( t ) = 1 N ∥ ∑ i = 1 N e → i ( t ) ∥ , ( 2 ) where e → i = v → i / ∥ v → i ∥ . When all the fish have the same direction then the polarization is maximal and P ( t ) = 1 . The minimum value P ( t ) = 0 is reached instead when the velocity vectors cancel . Figs 3 and 4 depict the two U-turns introduced in Fig 1 , in terms of the polarization P ( t ) and the sine of the angle of incidence of each fish with respect to the outer wall θwi ( t ) . The duration of the two illustrated collective U-turns is te = 0 . 94 s for N = 2 and te = 1 . 5 s for N = 5 . For both group sizes , the group polarization ( Figs 3B and 4B ) before and after the U-turn is quite close to 1 , showing that before and after the collective U-turn , all individual fish maintain essentially the same common direction . During the U-turn , the polarization decreases , describing a sharp V-form with a minimum at P ( t ) ≈ 0 . 27 for N = 2 and P ( t ) ≈ 0 . 60 for N = 5 . The minimum is reached at approximately half the duration of the collective U-turn , tm = ( ts + te ) /2: tm = 0 . 47 s for N = 2 and tm = 0 . 75 s for N = 5 . Figs 3C and 4C show the change of direction individually for each fish in both U-turns: from anticlockwise to clockwise direction for N = 2 , and vice versa for N = 5 . Fig 3C clearly indicates that at t ≈ 0 . 3 s , the fish F1 has almost completed its individual U-turn , while F2 has just started to change direction: sin ( θw2 ( 0 . 3 ) ) ≈ 0 . 98 , while sin ( θw1 ( 0 . 3 ) ) ≈ −0 . 5 . In Fig 4C , a similar ordering can be inferred from the times of departure from the bottom line at ordinate sin ( θwi ) = −1 + δ , where δ > 0 is a small parameter with respect to the range of ordinate values; we used δ = 0 . 1 . Thus , the order is 2-3-1-5-4 . However , the order in which individual fish change the sign of their angle of incidence θwi is different , 2-1-3-5-4 , and also different is the arrival order to the top line at ordinate sin ( θwi ) = 1 − δ: 2-5-1-4-3 . Moreover , some of these departure and arrival times are almost identical ( see , e . g . , F1 and F4 ) , and the behavior of the fish during the U-turn is completely different . These difficulties in establishing a consistent order show that another criterion is necessary to identify the relation of influence between fish . We have based our criterion to decide if a fish is an influential neighbor of another fish on the average value of the time-dependent directional correlation between the two fish along a time window . For each pair of fish Fi and Fj , we define the directional correlation Hij as a function of the heading of Fi evaluated at time t and the heading of Fj evaluated at a delayed time t − τ , where τ is the time-delay [26]: H i j ( t , τ ) = e → i ( t ) · e → j ( t - τ ) . ( 3 ) The function Hij ( t , τ ) is in fact the cosine of the angle formed by the headings e → i ( t ) and e → j ( t - τ ) , and is a measure of how aligned is fish Fi at time t with fish Fj at time t − τ . The values of Hij ( t , τ ) are between −1 ( when fish swim in opposite directions ) and 1 ( when fish have the same direction ) , and equals zero when fish have perpendicular directions . By averaging Hij ( t , τ ) along a time-window of length ( 2w + 1 ) Δt , we are able to quantify how much the focal fish Fi is copying the moving direction of its neighbor with a time-delay τ by means of the following function [26] C i j ( t , τ , w ) = 1 2 w + 1 ∑ k = - w w H i j ( t + t k , τ ) , ( 4 ) where tk = kΔt ( the time-step in our experiments is Δt = 0 . 02s ) . The time-window parameter length w has been determined by means of a sensitivity analysis ( pairwise similarity matrix ) , finding that w = 2 yields the more satisfactory results; see Section “Parameter selection” in Materials and methods and S5 Fig . The average directional correlation Cij ( t , τ , w ) allows us to characterize a fish Fj as an influential neighbor of a focal fish Fi at time t with time-delay τ , if the value of Cij ( t , τ , w ) is larger than a given threshold Cmin . Details on how w and Cmin are obtained are given in Sections “Optimal setting parameters for influential neighbors identification” and “Parameter selection” in Material and Methods . Fig 5 shows the directional correlation H12 and its time-average C12 between fish F1 and F2 along the collective U-turn depicted in Fig 3 . Left ( resp . right ) panels aim to indicate the alignment of fish F1 ( resp . F2 ) at each time t with respect to the alignment of fish F2 ( resp . F1 ) at an earlier time t − τ . Panels A and C show respectively that for all τ , there is always an interval of time during which H12 ( t , τ ) ≈ −1 and C12 ( t , τ ) ≈ −1 ( dark region ) , meaning that for all time-delays there is always an interval of time in which fish have opposite directions . Moreover , the larger the time-delay , the wider the black region where the direction of F1 is opposite to the direction of F2 at the previous time . On the other hand , the figures of the directional correlation of F2 with F1 , especially Panel D , show a connected region in which the correlation C21 ( t , τ ) remains positive and above the threshold ( yellow in the figure ) around τ ≈ 0 . 42 s where H21 ≈ 1 during all the time interval [−0 . 5 , 2 s] . This strongly suggests that , during this time interval , F2 is copying the behavior of F1 with a 0 . 42 s time-delay , denoted τ2 , 1 for this specific U-turn . Thus , one can consider that F1 is influencing F2 with time-delay τ2 , 1 , while F2 is not influencing F1 in this specific case . This influence dynamics is illustrated in Fig 3D by drawing an arrow at time t from Fj to Fi when Fj satisfies the condition Cij ( t , τ , w ) > Cmin for being an influential neighbor of Fi at time t , which in turn receives this influence and responds by copying the exhibited heading with a time-delay τ . Using the same procedure for the N = 5 case depicted in Fig 4 , we draw Fig 6 that shows F1 copying F2 with a time-delay τ1 , 2 ≈ 0 . 5 s ( Panels A and E ) . F1 also copies F3 and F5 with , respectively , τ1 , 3 ≈ 0 . 2 s ( Panels B and F ) and τ1 , 5 ≈ 0 . 1 s ( Panels D and H ) , but it doesn’t copy F4 ( Panels C and G ) . The influential neighbors of F1 are thus F2 , F3 and F5 , at different times and with different time-delays . We have calculated the rest of the correlations for all pairs of fish ( see S1 Fig for an overview of all the heading correlations ) . As for the N = 2 case , these relations are illustrated by arrows going from the influential neighbors to the reacting fish in Fig 4D . The specific behavior of H . rhodostomus , namely , the successive alternation of bursts and coasts [15] , leads us to ask whether these abrupt changes of acceleration and speed can provide information that other fish could use to adjust their own movement . To address this aspect we study whether there is any correlation between the bursting activity of one fish at time t and the fact that this fish is an influential neighbor of another fish shortly after time t . A burst corresponds to a brief phase of acceleration during which most changes in fish heading occur [15] . Panels E in Figs 3 and 4 show the bursting activity of each fish Fi , i = 1 , … , N , and that of its influential neighbors . For each fish Fi , we draw a dot at time t and ordinate i if fish Fi is displaying a burst precisely at time t . Dot color at ordinate i corresponds to fish Fi’s color . The absence of a dot at a given time denotes that the fish is in a coasting phase at that time . A second row of colored dots is drawn at ordinate i − 0 . 5 for some values of t when two conditions are met: ( 1 ) Fish Fi is being influenced at those times by one or more fish Fj , j ∈ {1 , … , N} , j ≠ i , whose identity is given by the color of the dots , and ( 2 ) the influential fish Fj was bursting when it was influencing Fi at time t − τ earlier . If Fi has more than one influential neighbor at time t , the dot drawn at time t in row i − 0 . 5 has the color of the Fj fish with the highest index j . In Fig 3E , red dots at i = 1 mean that fish F1 is bursting at those time-steps and coasting at the other time-steps , and red dots at i − 0 . 5 = 1 . 5 indicate that , first , F1 is the influential fish of F2 at those time-steps , and second , F1 was bursting when it was earlier influencing F2 . In turn , there are two possible reasons to explain the absence of red dots at i − 0 . 5 = 1 . 5 for certain time values: either F2 has no influential neighbor , or F1 was coasting . To assess which of the two explanations is valid , one needs to look at Fig 3D . For example , the absence of dots at i − 0 . 5 = 1 . 5 during 0 . 57 s and 0 . 62 s is due to F2 having no influential neighbors , while the absence of dots in the same row between 0 . 75 s and 0 . 81 s results from the fact that F1 , which is the influential neighbor of F2 , is in a coasting phase at time t − τ ( in this example the delay was found to be τ = 0 . 42 s ) . Fig 3E shows that the bursting activities of both the focal fish and its influential neighbor are not directly correlated , suggesting that the primary source of information for fish to adjust their movements is the distance , orientation and angular position of their neighbors [15] . The same conclusion is obtained for N = 5 . By focusing on fish F2 for example , Fig 4E shows that there is no systematic overlap between the yellow dots at i = 2 and those at i − 0 . 5 for i ≠ 2 , suggesting that the correlation between the bursting activity of a fish and that of their influential neighbors is marginal . For all U-turns , we have counted the number of frames in which a fish is an influential neighbor , that is , the number of frames where the above described condition for identifying influential neighbors is met . When there are only two fish , a fish is found to be the influential neighbor 30% of the time spent in a U-turn . In groups of five fish , this proportion grows up to 62% . We have counted the number of influential neighbors Nif a fish Fi has during a U-turn in groups of five fish , finding that in most cases , a fish has only one or two influential neighbors ( for 58% of the time spent in a U-turn Nif = 1 or 2 ) ; see Fig 7A . The most frequent case is Nif = 1 ( 43% ) . Having more than one influential neighbor is frequent ( 19% ) , but less than having no influential neighbors ( 38% ) . The cases where there are more than two influential neighbors are negligible ( less than 4% of the total time spent in U-turns ) . For each fish Fi , we have calculated the respective distance dij ( t ) at which the other N − 1 fish Fj are from Fi during the U-turns , thus establishing a rank order among the neighbors influenced by Fi . We have then compared the influence of close neighbors with those of distant neighbors , finding no correlation between the distance rank of a neighbor and the influence it exerts on the focal fish . This is shown in Fig 7B , where we have depicted the distribution of the distance rank of influential neighbors with respect to a focal fish . The figure shows that fish spent the same proportion of time ( ≈ 25% ) being an influential neighbor of a focal fish independently of their distance rank . In other words , influential neighbors are not necessarily the closest ones . When trying to identify events of causal influence by means of correlations , it is crucial to keep in mind that correlation does not imply causation . We thus have controlled the effects of potential chains of influence , where e . g . fish F1 is highly correlated with F3 not because F1 is directly influencing F3 , but because F1 is influencing fish F2 , which in turn is influencing F3 . To check the impact of these chains of influence on our results , we have removed from our data all the pairwise influence data that correspond to the following situation: if F1 is influenced by both F2 and F3 and F2 is simultaneously influenced by F3 ( or F3 is influenced by F2 ) , then we removed the pairwise correlation ( focal fish , influential neighbor ) corresponding to ( F1 , F2 ) ( or ( F1 , F3 ) ) . After removing 7172 out of 69703 data points and recomputing the results with the remaining data , we found that our results remain practically unchanged . We have also calculated the position rank that each fish occupies in the group during a collective U-turn , finding that influential neighbors are mostly located in the front region of the group: 32% in the leading most advanced position , and 20% in the second place; see Fig 7C . Noticeably , influential neighbors can be found in the back of the group ( in 29% of the cases in the fourth or fifth position ) , and even in the last position ( a non-negligible 13% of cases ) . We also paid attention to the order in which each fish starts its individual U-turn during a collective U-turn , finding that influential neighbors are those that most frequently turn earlier ( 32% of the cases ) , and that this relation decreases linearly; see Fig 7D . It is again noticeable that influential fish can be found to be the last turning fish ( in 8% of the cases ) . The apparently surprising fact that influential fish can be found in the back of the group and that the last fish turning can be an influential fish is due to the anisotropic perception of fish and their relative orientations during U-turns . But these findings have to be understood in the light of our specific time-dependent characterization of influential neighbor . If , for instance , F1 turns first and influences F2 , F2 will turn with some time-delay after F1 . Then , when F2 is at half of its individual turning process , F2 can be rotating in the same direction as F1 in such a way that F1 , influenced by F2 , slightly adjusts its direction . We would then say that F2 , which is the last turning fish , has influenced F1 , the first turning fish . In order to compare different collective U-turns , we define a normalized time t ¯ = ( t - t s ) / ( t e - t s ) in terms of the actual time t and the starting and ending time of each U-turn , so that the duration of a U-turn is now t ¯ = 1 . Thus , t ¯ = - 1 corresponds to a time as long as the U-turn duration previous to the start of the U-turn , and t ¯ = 2 corresponds to a time as long as the U-turn duration after the end of the U-turn . We have calculated the instantaneous value of the average speed V ( t ) = 〈 ∥ v → ( t ) ∥ 〉 , the average group polarization P ( t ) = 〈P ( t ) 〉 and the average number of influential neighbors N ( t ) = 〈 N if ( t ) 〉 . Here , angle brackets refer to the average across all fish in the U-turn along a time-window containing the collective U-turn . Fig 8A and 8B show respectively the time evolution of V ( t ) and P ( t ) during the collective U-turns in groups of 5 fish . The description of the specific U-turn presented in Fig 4 is also valid for the general case: the speed decreases before the U-turn ( from V ( - 1 ) ≈ 150 mm/s to V ( 0 ) ≈ 115 mm/s ) , it reaches a minimum at half the U-turn duration t ¯ = 0 . 5 ( V ( 0 . 5 ) ≈ 70 mm/s ) , and it then grows to a higher value after the U-turn ( V ( 1 . 5 ) ≈ 165 mm/s ) . A very similar behavior was found in groups of 2 , 4 , 8 and 10 fish of the same species in [28] . At the same time , the polarization is very high and almost constant outside the U-turn ( P ( t ¯ ) ≈ 0 . 95 ) , and exhibits a perfect V-shape during the U-turn , with the high values ( P ( t ¯ = { 0 , 1 } ) ≈ 0 . 93 ) reached at exactly the instants where the start and end of the U-turn takes place t ¯ = 0 and t ¯ = 1 , and the minimum value ( P ( 0 . 5 ) ≈ 0 . 48 ) at the middle of the U-turn . As expected , the average group polarization P ( t ¯ ) significantly decreases during the U-turn to almost half the value it has outside the U-turn . Right after reaching this minimum , there is a sharp increase of speed and polarization as more fish adopt the new direction of motion . Fig 8C shows that before the U-turn the average number of influential neighbors N ( t ) increases until a maximum value is reached right before the start of the U-turn ( N ( - 0 . 1 ) ≈ 1 . 45 ) . During more than one half of the U-turn , N ( t ) decreases until a minimum ( N ( 0 . 6 ) ≈ 0 . 8 ) , and grows again beyond the end of the U-turn until a second maximum ( N ( 1 . 2 ) ≈ 1 . 6 , twice the height of the minimum ) . After that , all fish have completed their U-turns and N ( t ) decreases again . When the polarization is very high , the time-delay with which influential neighbors are detected is often too small in comparison with biologically realistic reaction times τR , so that these influential neighbors are not taken into account ( we used τR = 0 . 04 s; see Section “Optimal setting parameters for influential neighbors identification” in Materials and methods ) . This is the reason why the average number of influential neighbors N ( t ) appears to be smaller in regions outside the U-turn , than when the U-turn is just about to start ( t ¯ ≈ - 0 . 1 ) or slightly after its end ( t ¯ ≈ 1 . 2 ) . Meanwhile , the decrease of N ( t ) in the middle of the U-turn has a different origin: once a fish has started to turn around , there is no real need of updating its alignment according to all its neighbors . That fish can safely reverse its motion by keeping the alignment with only one of those neighbors and even not paying attention to them for some period of time . Another indicator of how fish make decisions while turning is how frequently a focal fish pays attention to other individuals . We define the relative variation of the number of influential neighbor per fish Nif ( t ) between two successive time-steps as follows: η ( t ) = | N if ( t + Δ t ) − N if ( t ) | N if ( t ) , ( 5 ) denoting by Δt the time-step between frames ( Δt = 0 . 02 s ) . We have depicted the time-evolution of the average 〈η ( t ) 〉 in Fig 8D , finding that 〈η ( t ) 〉 remains essentially constant before , during and after the U-turn event , the amplitude of its variation being smaller than 10% of the signal ( 0 . 007 and 0 . 08 , respectively ) . Since the average number of influential neighbors N ( t ) is smaller when fish are engaged in the U-turn than right before or right after the U-turn , a constant average 〈η ( t ) 〉 suggests that fish adjust their heading more frequently during the U-turn than outside the U-turn . Indeed , in the middle of a U-turn , no real common direction of motion exists ( P ( t ) ≈ 0 . 5 ) , that is , there is a high diversity of headings , so that fish have to frequently update their direction by paying attention to different neighbors . We are now interested in determining the dynamical spatial organization of the influential neighbors of a focal fish . The relative state of a fish Fj with respect to a focal fish Fi is characterized by several parameters: the relative position of the neighbor u → i j = u → j - u → i , where u → i is the vector position of Fi in cartesian coordinates , the distance between them d i j = ∥ u → i j ∥ , the viewing angle of Fj relative to the direction of Fi [26] , which is the angle θij with which Fi perceives Fj ( note that θij is not necessarily equal to θji ) , the relative velocity v → i j = v → j - v → i , and the relative heading ϕij = ϕj − ϕi . All these quantities are time-dependent . We have calculated their average value for all the U-turns in a uniform spatial grid of square cells to facilitate the interpretation of the vector field of these continuous variables . Each square cell , of side 20 mm , shows the average of the arbitrarily different number of values contained in the cell . Fig 9A shows the density map of the relative position of the influential neighbor with respect to the focal fish when N = 2 . The intensity of color is proportional to the frequency of occupation of the grid cell , showing that the influential neighbor is mostly located in front of the focal fish and at a distance of one to three body lengths from the focal fish . The same information is quantified in Panel B with a heat map in polar coordinates , highlighting the most frequent location of the influential neighbor . The average relative velocity 〈 v → i j 〉 is shown in Fig 9A ( arrows ) , superimposed to the density map . The vector field shows that when the influential neighbor is in front of or behind the focal fish ( sin〈θij〉 ≈ 0 ) , both fish move at similar speed although the focal fish is a little bit faster ( the small black arrows are pointing in the opposite direction to the red one ) and the difference in heading is also small . However , when the influential neighbor is on the sides of the focal fish , relative speed and heading difference tend to vary more as the distance between them increases . The distributions of distances dij and exposure angles θij between a focal fish and its neighbors are depicted in Panels C and D of Fig 9 respectively . We find , on the one hand , that their most frequent separation is 62 . 6 mm ± 29 . 7 mm ( mean and standard deviation of histogram in Fig 9C ) , a value that is consistent with previous results where it was shown that the behavioral reactions of a fish depend on the angular position of its neighbors , as a consequence of the anisotropic perception of the environment [15] . On the other hand , the distribution of the exposure angle of fish Fj to the focal fish Fi is narrower when Fj is influencing Fi than when Fj is a neighbor of Fi , not necessarily influencing Fi . As both distributions are centered on θij = 0 , this shows that Fj is more frequently located in front of Fi when Fj is an influential neighbor of Fi than in the case when Fj is just a neighbor of Fi . Fig 10 shows similar results for groups of N = 5 fish . Influential neighbors are more frequently located in front of the focal fish ( although with a slight shift to the right; see Panels A and B ) and at a mean distance of 67 . 5 mm ± 40 . 6 mm ( Panel C ) . In turn , the velocity field has a smaller intensity and is much more homogeneous than in the case where N = 2 . A slight asymmetry can also be observed ( not noticed when N = 2 ) with fish located in front and slightly to the right of the focal fish having a higher velocity than those located elsewhere . Moreover , the distribution of exposure angles is more dispersed than in the case of two fish , meaning that influential neighbors are exposed to the focal fish with a larger diversity of angles , something that is simply due to the higher number of fish . The difference in the homogeneity of the velocity field between groups of 5 and 2 individuals is not necessarily the result of averaging over a larger number of individuals . Although averaging over fish data pairs may reduce the uncertainty in the extracted parameter values , it is well-known that the level of homogeneity in the direction of motion of the school increases with group size [29] . But one also ought to consider that specific values of delay and curvature the individuals adopt during the U-turns could help to limit variability in coordinating the group . Some theoretical studies support this idea: simplified models of velocity alignment with additive noise have shown semi-analytically the existence of delay and rate of turn values that minimise the fluctuations in the variance of the individual speed [30] , and flocking models of self-propelled particles have also shown that delay can be tuned to increase stability and alignment of the group [31] . Finally , we have analyzed the variation of the time-delay τ as a function of both the distance between the focal fish and its influential neighbors dij and the difference of heading ϕij , finding that in both cases N = 2 and N = 5 , the time-delay increases with respect to both the distance dij and the heading difference ϕij ( see Fig 11 ) . This result can be understood because during a U-turn the fish speed is decreasing and two fish can display larger reaction times the more separated they are and the less aligned they are . As already mentioned in the introduction , establishing causal influence on the basis of correlation measures requires controlling for spurious effects . Although our experimental data correspond to a specific collective behavior in which individuals influence each other , the relatively short time-windows over which cross-correlation are averaged and the use of several parameters through sensitivity analysis can weaken the accuracy of our results . To demonstrate that the particular detections of influential neighbors are not purely due to chance , we generated random artificial U-turns events by bootstrapping the data and applying the same procedure used to analyze collective U-turns in our experiments . The null model is built for groups of 5 fish , for which our experimental data provide M = 2375 individual trajectories ( 5 × 475 collective U-turns ) . For every fish Fi , i = 1 , … , M , the trajectory is rotated so that the individual turning point of the fish ( where sin ( θwi ) = 0 ) is located in the upper part of the tank , by randomly sampling the new angular position ψi in the interval [π/2 − ξ , π/2 + ξ] , where ξ is a small angle ( we used ξ = π/12 ) . Similarly , the time scale of each fish is shifted by sampling the instant of turning in the time interval [−ζ , ζ] , where ζ is a short time ( we have used ζ = 1 s ) . Then , five trajectories are randomly sampled , each one from a different randomly sampled collective U-turn , and mirrored if necessary so that the five individual U-turns are done in the same direction , clockwise or anti-clockwise . This way , the five fish of the artificial U-turn make their individual U-turn approximately at the same place and approximately the same time . For more details , see the section “Null model” in Materials and methods . We have produced 1000 artificial collective U-turns; S9 Fig shows a collection of 10 of them . The results of our analysis are shown in red in Figs 7 and 8 . As expected , they reveal clear differences between artificial and experimental U-turns . Fig 7A shows that in artificial U-turns the proportion of time during which a focal fish has no influential neighbor is more than 63% of the time , while in the experiments it was less than 39% . The analysis also reveals that in artificial U-turns a focal fish has one influential neighbor for less than 28% of the time , while in the experiments , the proportion raises to 43% . Similarly , Fig 8C shows that the average number of influential neighbors N ( t ) = 〈 N if ( t ) 〉 is much smaller in artificial U-turns ( ≈ 0 . 4 ) than in real U-turns , where N ( t ) is almost always greater than 1 . Note that the increase of N ( t ) during U-turns in artificial data is the consequence of the channeled motion of fish by the corridor . Moreover , the variation of N ( t ) along time , including the transients preceding and following the U-turn , decreases in artificial U-turns while it remains constant and with a higher value in experiments . Fig 7B shows that distance rank has no significant effect on which fish is the influential one , both in experiments and in artificial U-turns . The decreasing number of influential neighbors comes from the fact that the tank is circular and the method we use . If the tunnel had been a straight corridor , we should have detected no decrease in our null model . However , in a circular tank , because of the geometrical constraints imposed by the curvature , even when two fish are both swimming in the same direction ( i . e . , clockwise or anti-clockwise ) , as the distance between fish increases , our method will detect a decrease of correlation . While Fig 7C confirms that influential neighbors are slightly more often ranked in the first position of the group , this effect is much more pronounced in the experiments . In fact , Figs 7B , 7C and 7D and 8A and 8B show that the selected null model satisfactorily reproduces the typical spatiotemporal behavioral patterns of real U-turns: the position and turning ranks are almost identical , as well as the variation of the average speed and the average group polarization , although the V-shape of the average polarization in real U-turns is significantly sharper than in artificial U-turns . An additional , albeit expected , result of our null model is the homogeneous ( isotropic ) spatial distribution of “influential neighbors” , while in real collective U-turns influential neighbors are mostly located in front of the focal fish; see S10A and S10B Fig , compared with Fig 10A and 10B .
By sharing information with other group members , schooling fish and other collectively moving animals can potentially improve their navigational accuracy ( e . g . the many wrongs principle [32] ) , take better decisions ( e . g . to avoid a predator [33] ) , or improve their abilities to sense the environment [34] . However , there are both physical and practical reasons why information is expected to be shared with only a few neighbors . Physical reasons involve material limitations , such as visual occlusions . Practical reasons often refer to trade-offs between sharing information , so that the group collectively selects a direction of motion , and deciding independently [35 , 36] . Assuming that correlations between fish behavior rely to some extent on a causal influence , our analysis reveal that in groups of H . rhodostomus , during a collective U-turn , at any moment in time each fish only pays attention to a small number of neighbors whose identity regularly changes . We also find that the phases during which a focal fish is affected by one or two influential neighbors are interspersed with other phases during which its movement appears uninfluenced by the movement of neighbors . Moreover , influential fish are mostly located in front of the focal fish . The distance between a focal fish and its influential neighbors is about two body-lengths and the relative exposure angle is smaller than 60 degrees . Our results bring insights on the way information on the neighborhood is processed by fish . Instead of having a synchronous update based on a fixed number of neighbors ( topological neighborhood ) or on all neighbors located within a fixed distance ( metric neighborhood ) , our results suggest an asynchronous updating that does not depend on the distance between a focal fish and its influential neighbors . A similar asynchronous updating scheme has been previously introduced by Bode et al . [37] in a flocking model showing that it can give rise to emergent topological interactions consistent with the measures done on starling flocks [38] . It is however worth noting that our experiments , performed on small group sizes , may have prevented us from detecting any influence of the distance , since each of the four neighbors are located between one and three body lengths . In larger groups of fish moving in an unconstrained space , we expect the effective neighborhood of fish to result from the interplay between an asynchronous updating on a small number of neighbors and a modulation of the strength of interactions with the distance between fish [15] . Previous studies on the number and the spatial arrangement of influential neighbors led to different results depending on the species and on the procedure used to analyse the data . The work by Ballerini et al . [39] provides evidence that each bird within a starling flock ( Sturnus vulgaris ) coordinates its motion with a fixed number of closest neighbors , irrespective of their distance , while in mosquitofish ( Gambusia holbrooki ) , one single nearest neighbor was sufficient to account for the large majority of the observed interaction responses [12] . In barred flagtails ( Kuhlia mugil ) , it has been shown that different kinds of neighborhoods ( Voronoi neighborhood and the k nearest neighbors ( k ≈ 6 ∼ 8 ) were compatible with experimental data in a tank [13] . Our study points to a low number of influential neighbors . There are multiple possible explanations for the differences in the number of interacting neighbors found across the scientific literature . ( i ) It is possible that different animal groups interact with different numbers of neighbors . ( ii ) Temporal factors are also important [37] , as interactions can be integrated in time to produce effectively larger neighborhoods . Here , we propose a third explanation ( iii ) based on the consideration that interaction responses such as attraction , alignment and avoidance are qualitatively different mechanisms that rely on different sensory-motor responses and , consequently , on different interacting neighborhoods . In particular , attraction and repulsion require to process information about the position of neighbors , while alignment is intrinsically a response dependent on orientation and velocity . These different interactions are likely to rely on different neural circuits ( motion and form are typically processed by different brain areas in many animal groups [40 , 41] ) and hence might depend on different sets of influential neighbors: for instance , a focal individual could avoid collisions with its Voronoi neighbors , be attracted towards a different neighborhood of visually salient individuals and only process alignment information for one or two selected neighbors . It might also depend on different sets of influential neighbors: for instance a focal individual could avoid collisions with its Voronoi neighbors , be attracted towards a different neighborhood of visually salient individuals and only process alignment information for one or two selected neighbors . It is thus natural to suggest that influential neighbors are intrinsically associated with different interaction mechanisms , which might also explain why fish point to different neighborhoods . Our method for identifying influential neighbors is based on the computation of the time-dependent directional correlation between a focal fish and its neighbors . Of course , correlation does not imply causation , so that inferring causal influence between fish from directional correlation requires an extremely cautious methodology . The methodology we proposed here is based on two solid procedural cornerstones . First , the data used in our study were carefully selected from a clearly recognizable behavior , the collective U-turns , where influence from neighbors undoubtedly exists , and thus should be , to some extent , responsible for a fundamental part of the correlations detected by our method . Time-delay between individuals’ direction choices has already been used to measure the interactions between group members in animal flocking . Specifically , Nagy et al . [23] used correlation delay times to reconstruct flight hierarchies in flocks of pigeons . Their approach consisted in integrating delay times over the entire trajectory to obtain a “leadership mark” for each individual . Our assumption is instead that the time-delay results from the individuals’ behavior and their environment , which varies in time depending on the information being gathered . To detect the response delay of each individual , we have instead followed the approach employed in [26] that allows for a change of delay over time . In fact , it is easy to show that the time delay between the same pair of fish is not constant , as revealed by our analysis of pair of fish ( see Material and methods ) . Applying Nagy et al . ’ method to different subsets of data in the same experiment , we found that the time delays between the same pair of fish vary substantially ( see S2 Fig ) . The second methodological cornerstone is provided by the results of the null model that clearly show that the correlations we detected come from causal influence between neighbors and not from spurious random coincidences . The results of the null model also confirm that distance rank has no effect . Identifying the number and position of influential neighbors is an essential step towards reconstructing behavioral cascades of information propagation across a group . Our method provides an accurate basis for mapping interaction network that does not rely on any assumption about the channel ( e . g . , vision , sound or hydrodynamic interactions ) mediating information transfer . We are confident that by adopting our technique to map interactions in different species and different experimental contexts we will gain a much more detailed understanding of the distributed information processing taking place in fish schools .
Our experiments have been approved by the Ethics Committee for Animal Experimentation of the Toulouse Research Federation in Biology N°1 and comply with the European legislation for animal welfare . Hemigrammus rhodostomus ( rummy-nose tetras , Fig 12A ) were purchased from Amazonie Labège ( http://www . amazonie . com ) in Toulouse , France . Fish were kept in 150 L aquariums on a 12:12 hour , dark:light photoperiod , at 27 . 7°C ( ±0 . 5°C ) and were fed ad libitum with fish flakes . The average body length of the fish used in these experiments was 31 mm ( ± 2 . 5 mm ) . The experimental tank ( 120 × 120 cm ) was made of glass and was set on top of a box to isolate fish from vibrations . The setup was placed in a chamber made by four opaque white curtains surrounded by four LED light panels to provide an isotropic lighting . A ring-shaped corridor was set inside the experimental tank filled with 7 cm of water of controlled quality ( 50% of water purified by reverse osmosis and 50% of water treated by activated carbon ) heated at 28 . 1°C ( ±0 . 7°C ) ( Fig 12B ) . The corridor was made of a vertical circular outer wall of radius 35 cm and a circular inner wall with a conic shape of radius 25 cm at the bottom , so that the effective width of the corridor available to fish for swimming ranges from 10 cm at the bottom to 12 cm at the surface . The conic shape was chosen to avoid the occlusion on videos of fish swimming too close to the inner wall . Fish were randomly sampled from their breeding tank for a trial and were used at most in only one experiment per day . Groups of 2 or 5 fish were introduced in the experimental tank and acclimatized to their new environment for a period of 10 minutes . Their behavior was then recorded for one hour by a Sony HandyCam HD camera filming from above the setup at 50 images per second in HDTV resolution ( 1920x1080p ) . We performed 10 trials for each group size of 2 and 5 fish . The positions of each fish on each frame were tracked with idTracker 2 . 1 [10] . Fish were sometimes misidentified by the tracking software , for instance when two fish were swimming too close to each other for a long period of time . In those cases , the missing positions were corrected manually . All sequences with 50 consecutive missing positions or less were interpolated . Larger sequences of missing values were checked by eye to determine whether interpolating was reasonable or not; if not , namely the trajectory doesn’t look like a straight line , then merging positions with closest neighbors were considered . Time series of positions were converted from pixels into meters . The origin of the coordinate system was set to the center of the ring-shaped tank . Body orientation of fish were measured using the first axis of a principal component analysis of the fish shapes detected by idTracker 2 . 1 . Since the experiments were performed in an annular setup , the direction of rotation can be converted into a binary value: clockwise or anti-clockwise . We choose the anti-clockwise direction as the positive values for angular position . Before a U-turn event , all fish move in the same direction , say clockwise . Then , one fish , not necessarily the one located at the front of the group , changes its direction of motion to anti-clockwise direction . After a short transient , the other fish of the group display the same direction change , from clockwise to anti-clockwise . We defined the whole process of changing direction as a collective U-turn ( see examples in Fig 1 and in S8 Fig ) . After data extraction and pre-processing , we found 1111 and 475 collective U-turns in groups of 2 and 5 fish , respectively . The duration distribution of collective U-turns in groups of 2 fish is shown in S3 Fig while the results for groups of 5 fish are shown in S4 Fig . Most of the collective U-turns last between 1 and 3 seconds , while the individual turning time usually lasts between 0 . 4 and 1 second . The procedure used to define an individual U-turn for a fish Fi is as follows: we first determine the time tm , i at which the sign of the angle of incidence of fish Fi changes sign ( from negative to positive or vice versa ) . Then , starting from tm , i , we reverse time step by step until the first time at which the absolute value of the angle of incidence is higher than a threshold θ ¯ s , i is reached . We denote this time by ts , i . Similarly , we start again from tm , i and go forward step by step until the first time at which the absolute value of the angle of incidence is higher than a second threshold θ ¯ e , i is reached . We denote this time by te , i . To determine the values of the thresholds θ ¯ s , i and θ ¯ e , i , we first compute the moving average of the angle of incidence over a period of 50 time steps ( 1s in real time ) , before and after the middle point tm , i , with a window of 5 time steps ( 0 . 1s in real time ) , respectively . Then we set the threshold values as the maximum values of the absolute moving average . Doubling the length of the period of time over which the average is computed , or doubling the width of the window , do not affect the results . Finally , the time at which the collective U-turn starts ( resp . ends ) is defined by min { t s , i } i = 1 N ( resp . max { t e , i } i = 1 N ) . The relative position of a fish Fi in a group of N fish is calculated by projecting the vector position of the fish u → i on the average group velocity vector z → = ( 1 / N ) ∑ i = 1 N v → i . This allows us to define a group centroid in the direction of z → , with respect to which the fish are ranked: the first fish in the group is the fish whose projection on z → is the most advanced one in the direction of motion of the group ( given by z → ) , the second fish in the group is the second most advanced , and so on . Relative distance between fish are not taken into account when establishing the rank . Four parameters are used to identify influential neighbors: the time-delay τ , the window size w , the correlation threshold Cmin above which individuals are supposed to be interacting , and the threshold ε for selecting more than one influential fish . The time delay must be specified along the whole trajectory of the focal fish: it is thus a series of values { τ k * } k = 0 M , where M is the number of time-steps or frames in the individual U-turn . The parameters Cmin , ε and w are in turn given for all time and for all fish by means of a sensitivity analysis described in the next section . Assume by now that the three values Cmin , ε and w are known , and denote by Fi the focal fish and by Fj one of its neighbors . Then , the series of time-delays { τ k * } k = 0 M i is built recursively as follows ( actually only w is required to extract the time delays ) . Denote by Γi ( tk ) the highest value of the pairwise directional correlation Cij of the velocity of fish Fi at time tk with the velocity of Fj at each time-step in the range of the previous ( τ k - 1 * + 1 ) time-steps R k = [ 0 , τ k - 1 * + 1 ]: Γ i ( t k , w ) = max τ r ∈ R k { C i j ( t k , τ r , w ) } . ( 6 ) Then , the time-delays τ k * , k = 1 , … , Mi , are determined by the smallest value of the time-delay τr ∈ Rk where Γi ( tk , w ) reaches its maximum . For t1 , the maximum correlation is reached at C i j ( t 1 , τ 1 * , w ) , for some time-delay τ 1 * ∈ R 1 = [ 0 , τ 0 * + 1 ] . We set τ 0 * = 50 for the initial value of the recurrence . For the rest of time-delays τ k * , k = 2 , … , Mi , the size of Rk is based on the assumption that if , at some time t , Fi copies the behavior that Fj displayed at a previous time t − τ , then , after time t , Fi will not copy the behavior that Fj displayed at any time earlier than t − τ . Time-delays obtained with more complicated and time consuming procedures such as the time-ordered technique developed in [26] or through the similarity analysis based on Fréchet distances [25] would in principle produce similar values . Fig 13B shows the distribution of time-delays obtained with this procedure in groups of two fish . The distribution is clearly bimodal with a first peak when τ = 0 and a second one around τ = 0 . 4 s . Considering a reaction time threshold of 50-100 ms for a fish to integrate information and reach a decision [42] , we cannot attribute small values of time-delays to situations where the behavioral decision of the focal fish has been influenced by its neighbors . This is confirmed by the analysis of the spatial distribution of the extracted time-delays ( Fig 13A ) , where we show that the lowest average values of τ are found mostly when the neighbor was behind the focal fish , in a zone with the lowest perception [15] , while the highest values of τ > 0 . 4 s are found when the neighbor is located in front of the focal fish . This has lead us to consider in our analyzes only situations where τ > τR = 0 . 04 s . Although the time-delays { τ k * } k = 0 M are determined once w is known , they also strongly depend on Cmin and ε , as the value of these three parameters must be fixed at the same time . This is done by means of a sensitivity analysis in which we have tested the following 40 combinations of parameter values: w ∈ {0 , 1 , 2 , 3 , 4} , ε = {3 , 5} , and Cmin ∈ {0 . 995 , 0 . 99 , 0 . 95 , 0 . 5} . Each combination ( Cmin , ε , w ) gives rise to four histograms like those depicted in Fig 7 . These histograms constitute the solution of our method of analysis , and can be characterized by a vector S → ( C min , ε , w ) in 19 dimensions: ( i ) the 5 proportions of the number of influential neighbors in groups of 5 fish , ( ii ) the 4 proportions of their distance rank , ( iii ) the 5 proportions of their position rank , and ( iv ) the 5 proportions of their turning rank . This allows us to determine how similar are the results arising from two combinations ( Cmin , ε , w ) and ( C min ′ , ε ′ , w ′ ) , by computing the cosine similarity of the two vectors S → ( C min , ε , w ) and S ′ → ( C min ′ , ε ′ , w ′ ) . The cosine similarity of two vectors a → and b → , denoted cos sim ( a → , b → ) , is the cosine of the angle between these two vectors . Thus , two colinear vectors are such that cos sim ( a → , b → ) = ± 1 independently of their magnitude , while two perpendicular vectors are such that cos sim ( a → , b → ) = 0 . In our case , the components of the vectors are positive , so cos sim ( S → , S ′ → ) ≥ 0 for all ( Cmin , ε , w ) and ( C min ′ , ε ′ , w ′ ) . Moreover , as the components are proportions , colinearity implies identity , both in direction and magnitude . Thus , cos sim ( S → , S ′ → ) = 1 means that both results are identical , while cos sim ( S → , S ′ → ) = 0 means that they differ as much as possible . S5 Fig shows the cosine similarity matrix for the 40 combinations we have tested . Note that the matrix is symmetric with respect to the diagonal , where cos sim ( S → , S → ) = 1 . Except for Cmin = 0 . 5 , all similarity values are in the thin range [0 . 96 , 1] , showing that all combinations yield practically the same results . The higher dissimilarity is found in the white-yellow lines , where one of the combinations is ( Cmin , ε , w ) = ( 0 . 5 , 3 , 2 ) . The selection of parameter values is thus done as follows . We choose w = 2 , which corresponds to the higher dissimilarity regions . The selected time window size is sufficiently large so that the jagged nature of the movement data is smoothed out but not too large so that the actual turns gets washed out from the data . Using ε = 3 or ε = 5 yields very similar results and we have arbitrarily chosen ε = 3 . The selection of Cmin is done by a specific procedure , which consists in calculating the number of data points that remain available for our analysis for each value of Cmin . S6 and S7 Figs exhaustively demonstrate that the larger Cmin is , the less data points remain available , and vice versa . We might be prone to choose a sufficiently small Cmin in order to get the maximum number of data points . However , according to our definition of influential neighbor , Cmin should be sufficiently large to select only the real influential neighbors . We have thus chosen the highest value which provides a sufficiently large number of data points , that is , the largest value before the fall of the number of data points in S11 Fig , Cmin = 0 . 95 . This value preserves 61% ( 23830 ) and 76% ( 69703 ) of data points for N = 2 and N = 5 respectively . We want to design artificial collective U-turns in groups of 5 fish where all fish perform an individual U-turn at more or less the same place and more or less the same time , and in the same direction ( clockwise or anti-clockwise ) . Fish must coincide in time and space to constitute a “group” , but individual U-turns must happen in an absolutely independent way . Correlations at hand in this paper are thus reduced to a minimum , while preserving the general aspect of a group of fish changing direction . Our experimental data provide us with 5 × 475 = 2375 trajectories of individual fish , which we have conveniently normalized and combined to build 1000 groups of 5 fish changing direction in the same spatiotemporal interval . This is done as follows . The whole trajectory of a fish Fi during a U-turn takes place in an interval of time [ts , i , te , i] , where ts , i is the instant at which the individual U-turn of fish Fi starts , and te , i is the time at which the individual U-turn ends . See the paragraph above Eq ( 1 ) . The trajectory of fish Fi in radial coordinates is given by { ( ρ i ( t k ) , ψ i ( t k ) ) } k = 1 N i , ( 7 ) where ρi ( tk ) is the radius ( distance of the fish from the center of the tank ) , ψi ( tk ) the already defined angle position ( computed anticlockwise as positive ) , and Ni is the number of time-steps tk in the trajectory . Denote by Ti the instant at which fish Fi effectively turns , i . e . , Fi is perpendicular to the wall: sin ( θwi ( Ti ) ) = 0 . In well defined individual U-turns as the ones we are using in our data , this happens only once per U-turn . Accordingly , ( ρi ( Ti ) , ψi ( Ti ) ) denotes the fish position at time Ti . Although we would like to have absolutely uncorrelated fish , it would not make sense to use groups of trajectories that do not reproduce a consistent U-turn , e . g . , if one fish makes its U-turn much later than another , or on the other side of the tank . We thus try to decorrelate fish trajectories as much as possible , while preserving at the same time the typical spatiotemporal shape of real collective U-turns . The decorrelation of all individual U-turns is done with the following two steps: The artificial collective U-turn is thus built as follows: Then , the fish of reference of the artificial U-turn will make its individual U-turn at time ζref ∈ [−1 , 1] s and position ( ρref ( Tref ) , π/2 + ξref ) . The other four fish Fj will make their individual U-turn at time ζj ∈ [−1 , 1] s and position ( ρj ( Tj ) , π/2 + ξj ) respectively , for j = 1 , … , 5 , j ≠ ref . We have depicted in S9 Fig a set of artificial U-turns for comparison with the real U-turns shown in S8 Fig . Note that in these figures the time-scale has been shifted again so that collective U-turns start at t = 0 s . | Schooling fish exhibit impressive group-level coordination in which multiple individuals move together in a seamless way . This is possible because each individual in the group responds to the movement of other group members . But how many individuals does each fish pay attention to ? Which are the influential neighbors ? It is necessary to answer these questions in order to understand how directional information propagates across a group . Our research shows that in the rummy-nose tetra species there is a limited number of influential neighbors which are not necessarily the closest ones . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"velocity",
"collective",
"animal",
"behavior",
"fish",
"swimming",
"medicine",
"and",
"health",
"sciences",
"classical",
"mechanics",
"reaction",
"time",
"vertebrates",
"mathematical",
"models",
"neuroscience",
"biological",
"locomotion",
"animals",
"cognitive",
"neuroscience",
"animal",
"behavior",
"zoology",
"research",
"and",
"analysis",
"methods",
"behavior",
"mathematical",
"and",
"statistical",
"techniques",
"physics",
"eukaryota",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"cognitive",
"science",
"organisms",
"motion"
] | 2017 | Identifying influential neighbors in animal flocking |
Paracoccidioides brasiliensis causes paracoccidioidomycosis , one of the most prevalent systemic mycosis in Latin America . Thus , understanding the characteristics of the protective immune response to P . brasiliensis is of interest , as it may reveal targets for disease control . The initiation of the immune response relies on the activation of pattern recognition receptors , among which are TLRs . Both TLR2 and TLR4 have been implicated in the recognition of P . brasiliensis and regulation of the immune response . However , the role of TLR9 during the infection by this fungus remains unclear . We used in vitro and in vivo models of infection by P . brasiliensis , comparing wild type and TLR9 deficient ( −/− ) mice , to assess the contribution of TLR9 on cytokine induction , phagocytosis and outcome of infection . We show that TLR9 recognizes either the yeast form or DNA from P . brasiliensis by stimulating the expression/production of pro-inflammatory cytokines by bone marrow derived macrophages , also increasing their phagocytic ability . We further show that TLR9 plays a protective role early after intravenous infection with P . brasiliensis , as infected TLR9−/− mice died at higher rate during the first 48 hours post infection than wild type mice . Moreover , TLR9−/− mice presented tissue damage and increased expression of several cytokines , such as TNF-α and IL-6 . The increased pattern of cytokine expression was also observed during intraperitoneal infection of TLR9−/− mice , with enhanced recruitment of neutrophils . The phenotype of TLR9−/− hosts observed during the early stages of P . brasiliensis infection was reverted upon a transient , 48 hours post-infection , neutrophil depletion . Our results suggest that TLR9 activation plays an early protective role against P . brasiliensis , by avoiding a deregulated type of inflammatory response associated to neutrophils that may lead to tissue damage . Thus modulation of TLR9 may be of interest to potentiate the host response against this pathogen .
Paracoccidioides brasiliensis is a causative agent of paracoccidioidomycosis ( PCM ) , one of the most prevalent systemic mycosis in Latin America [1] . One of P . brasiliensis biological hallmarks is its particular temperature-dependent morphological dimorphism . This fungus switches from the environmental non-pathogenic mycelial/conidial form at ambient temperatures to the pathogenic multiple budding yeast form , with an high variability of cell sizes , when exposed to temperatures similar to those of the mammalian host [2] , [3] , [4] . The mechanism of infection by P . brasiliensis entails the inhalation of airborne conidia that , when in the lung and exposed to host temperatures , undergo a complex morphological switch to the pathogenic yeast form [1] , [5] . PCM is divided in two different forms: the acute or sub-acute form and the chronic form , depending on the natural course of infection and clinical manifestations of the patient [1] , [6] . The clinical manifestations rely on the virulence of P . brasiliensis infecting strain , the degree and type of immune response triggered , the tissues infected , and importantly , on intrinsic characteristics of the host [7] , [8] . Despite the fact that a large number of individuals are exposed to the fungus , only a minority develops the disease , suggesting that , for the majority of the population , a protective immune response is developed [6] , [9] . Therefore , the understanding of the protective characteristics of the immune response to P . brasiliensis is of interest as it may reveal targets for disease control . The initiation of the immune response relies on the activation of the innate immune system upon recognition of pathogen-associated molecular patterns ( PAMPs ) [10] . This recognition is mediated by the family of pattern recognition receptors ( PRRs ) that is composed by a large number of receptors in immune cells [11] , [12] . Activation of PRRs culminates with the expression of several immune mediators , including pro- and anti-inflammatory cytokines and also with the activation of a series of microbicidal mechanisms that aim at eliminating the pathogen [13] . The most widely recognized type of PRRs are the toll-like receptors ( TLRs ) [14] . Over the past few years , several studies have demonstrated a relevant role for TLRs in the recognition of fungal pathogens , such as P . brasiliensis , Candida albicans , Aspergillus fumigatus , and Cryptococcus neoformans [15] , [16] , [17] , [18] , [19] , [20] , [21] . The role of MyD88 , an adaptor protein used by all TLRs ( with the exception of TLR3 ) , during P . brasiliensis infection remains controversial , with some authors reporting that this protein is not essential for an effective response against P . brasiliensis [22] , and others claiming that MyD88 is important for the activation of innate fungicidal mechanisms and for the induction of the effector and regulatory cells of the adaptive immune response [23] . A role for both TLR2 and TLR4 in the recognition and internalization of P . brasiliensis has been reported in human monocytes and neutrophils [24] . In a model of experimental PCM , TLR2 deficiency leads to increased Th17 immunity associated with diminished expansion of regulatory T cells and increased lung pathology due to unrestrained inflammatory reactions [25] . In contrast , P . brasiliensis recognition by TLR4 leads to an increased production of Th17 cytokines , enhanced pro-inflammatory immunity , and impaired expansion of regulatory T cells , resulting in a more severe form of infection [26] . However , the involvement of TLR9 in P . brasiliensis infection has not yet been addressed . Several lines of evidence suggest that TLR9 may play a role in infection by P . brasiliensis , similarly to what is already described for other pathogenic fungi , such as A . fumigatus , C . albicans and C . neoformans [27] , [28] , [29] , [30] , [31] , [32] . Firstly , P . brasiliensis DNA is known to have large numbers of CpG motifs [33] , the natural ligand to TLR9 [34] . Secondly , due to the fungus multinucleated nature [35] , [36] , a high amount of DNA is expected to be released upon cell death during infection . Thirdly , previous studies show that , in in vitro models , P . brasiliensis DNA increases the phagocytic index of macrophages , whereas in in vivo models of P . brasiliensis infection , TLR9 activation may act as a Th1-promoting adjuvant in a time/concentration dependent-manner [33] , [37] . In this study , we investigated the role of TLR9 in the recognition of P . brasiliensis , and its influence on the infective process and evolution of the disease . Our results show that TLR9 recognizes P . brasiliensis , playing a major regulatory role during early times of in vivo infection with its absence making the host more prone to increased liver pathology and premature death , mainly mediated by neutrophils .
The strain ATCC 60855 of P . brasiliensis registered at the American Type Culture Collection ( Rockville , MD ) was used throughout this experiment . Yeast cells were maintained at 37°C by subculturing in brain heart infusion ( BHI ) ( Duchefa ) solid media supplemented with 1% glucose and gentamicin ( 50 µg/mL ) . For both in vitro and in vivo assays , yeast cells were grown in BHI liquid medium supplemented with 1% glucose and gentamicin ( 50 µg/mL ) at 37°C with aeration on a mechanical shaker ( 220 rpm ) . Cell growth was monitored for 148 h by microscopic counting using a Neubauer's Chamber and cells were collected during the exponential growth phase ( 72 h of growth , 1 . 65±0 . 8×107 cells/mL ) for all the experimental assays . This study was approved by the Portuguese national authority for animal experimentation Direção Geral de Veterinária ( ID: DGV 594 from 1st June 2010 ) . Animals were kept and handled in accordance with the guidelines for the care and handling of laboratory animals in the Directive 2010/63/EU of the European Parliament and of the Council . Eight-week-old C57BL/6 mice were obtained from Charles River ( Barcelona , Spain ) and eight-week-old TLR9−/− ( generated in a C57BL/6 background ) were kindly provided by P . Vieira ( Pasteur Institute of Paris , France ) . Mice were housed under specific-pathogen-free conditions with food and water ad libitum . Liver from dying mice were harvested in the time-period comprehending 48 h post-infection , fixed in 3 . 8% phosphate-buffered formalin and embedded in paraffin . Light-microscopy studies were performed on tissue sections stained with hematoxylin and eosin ( HE ) as previously described [38] . The histological analysis was performed by the presence of necrotic areas and the type of inflammatory infiltrate ( when present ) in each field of 10× objective . C57BL/6 WT and TLR9−/− mice were made neutropenic by treatment with the monoclonal antibody ( MAb ) RB6-8C5 , as previously described [39] , [40] . Briefly , mice were injected i . v . in the lateral tail vein with 200 mg of MAb RB6-8C5 , and 6 h later infected i . v . with 1×106 P . brasiliensis yeast cells grown to the exponential phase in BHI liquid medium and treated as indicated before . Mice were monitored as indicated before during the first 2 days of infection . Bone marrow-derived macrophages ( BMDMs ) from C57BL/6 WT and TLR9−/− mice were prepared as described previously [38] , [41] , [42] . BMDMs were seeded in 24-well plates at 5×105 cells/well and kept at 37°C and 5% CO2 atmosphere . Cells were challenged with 5 µg of P . brasiliensis DNA extracted from exponentially growing cells . In other experiments , BMDMs from C57BL/6 WT and TLR9−/− mice seeded as described above were challenged with P . brasiliensis yeast cells grown to the exponential phase , with late-stationary growing cultures or with P . brasiliensis lysed cells , using a 2∶1 multiplicity of infection ( MOI; yeast/BMDMs ratio ) for 24 h . Prior to infection , fungal cells were washed 3 times with LPS-free PBS , passed through a syringe to eliminate cell clumps , and submitted to Neubauer counting procedures ( mother and bud cells were considered as individual counts ) . Supernatants from stimulated BMDMs were collected 24 h post-infection and stored at −80°C for later cytokine analysis . BMDMs from C57BL/6 WT and TLR9−/− prepared as previously described [38] were seeded in 24-well plates at 5×105 cells/well , kept at 37°C and 5% CO2 atmosphere and infected with P . brasiliensis yeast cells grown to the exponential phase using a 2∶1 multiplicity of infection ( MOI; yeast/macrophage ratio ) for 3 h . The wells were then washed three times with PBS to eliminate non-phagocytized cells , BMDMs were lysed with sterile H2O and the remaining phagocytized yeast cells were collected to determine total phagocytosis , using a Neubauer's Chamber . For the experiments , P . brasiliensis cells treated with DNase were also used . Briefly , after collecting and washing , P . brasiliensis yeast cells grown to the exponential phase were incubated at 37°C for 1 h with 20 µg of DNaseI ( Ambion ) . As control , heat-inactivated DNase ( 70°C for 1 h ) was used . Cytokine levels were measured in serum collected from infected animals or in supernatants of infected cell cultures by capture enzyme-linked immunosorbent assay ( ELISA ) ( eBioscience ) . The ELISA procedure was performed according to the manufacturer's protocol , and absorbances were measured with a Bio-Rad 680 Micro-plate Reader . Quantification of neutrophils and mononuclear cells influx to the peritoneal cavity during the i . p . infection with P . brasiliensis was performed by flow cytometry ( FCM ) on a BD LSR II flow cytometer . Cells collected from the peritoneal cavity were stained , using specific antibodies for CD11b , CD11c , GR-1 and Ly6G to distinguish neutrophils and mononuclear cells populations , and a minimum of 100 , 000 cells per sample was acquired at low/medium flow rate . Offline data was analyzed with the flow cytometry analysis software package FlowJo 7 . 6 . 1 . Total RNA ( 1 µg ) was isolated according to TRIzol methodology ( Invitrogen ) . For liver samples , small portions were homogenized in between microscopy slides , suspended in 1 mL of TRIzol reagent and kept at −80°C for later analysis . For the i . p . experiment around 5×105 cells from the peritoneal exudates were used , suspended in 250 uL of TRIzol reagent and stored at −80°C for later analysis . RNA integrity was checked by the presence of clear 18S and 28S rRNA bands in agarose gel electrophoresis . The absence of DNA contamination in the samples was confirmed by the absence of PCR amplification of the ubiquitin gene in the isolated RNA . Total RNA ( 1 µg ) was reverse transcribed using the iScript cDNA Synthesis kit ( Bio-Rad ) following manufacturer's instructions and 1 µL of cDNA used as a template for real-time quantification using the SsoFast EvaGreen SuperMix ( Bio-Rad ) following manufacturer's instructions . Real-time quantification was carried out on a CFX96 Real-Time System ( Bio-Rad ) using threshold cycle ( Ct ) values for ubiquitin transcripts as the endogenous reference . The primer sequences were designed and synthesized by TIB Mol . Biol . and were as follows: UBQ forward , TGG CTA TTA ATT ATT CGG TCT GCA T; UBQ reverse , GCA AGT GGC TAG AGT GCA GAG TAA; IL-10 forward , TTT GAA TTC CCT GGG TGA GAA; IL-10 reverse , GCT CCA CTG CCT TGC TCT TAT T; IL-17 forward , CTC AGA CTA CCT CAA CCG TTC CA; IL-17 reverse , TTC CCT CCG CAT TGA CAC A; TNF forward , GCC ACC ACG CTC TTC TGT CT; TNF reverse , TGA GGG TCT GGG CCA TAG AAC; MIP2 forward , CTC AGT GCT GCA CTG GT; MIP2 reverse , AGA GTG GCT ATG ACT TCT GTC T; IL-6 forward , TCG TGG AAA TGA GAA AAG AGT TG; IL-6 reverse , TAT GCT TAG GCA TAA CGC AC TAG . All measurements were performed in triplicate . A single melting peak was obtained for each gene analyzed in all samples . Data is reported as the mean ± standard error of the mean ( SEM ) and all assays were repeated at least three times . All statistical analysis was performed using the GraphPad Prism Software version 5 . 01 . For the experiments comparing two groups ( see Fig . 1A and B , 4 and 5 ) , a two-tailed unpaired Student t test was performed . Welch's correction was applied when making multiple comparisons . The One Way ANOVA test was performed in data presented in Fig . 1C using Turkey's multiple comparison post-test . The survival curves , representative of three independent experimental infections ( Fig . 2 , n = 21 mice ) , are represented using the Kaplan-Meier estimator , and Gehan-Breslow-Wilcoxon test was applied . For all data analysis statistical significance was considered at the level of 0 . 05 ( 2-tailed , 95% confidence interval ) .
Since previous studies have reported the recognition of fungal DNA by TLR9 [43] , we questioned if P . brasiliensis triggers TLR9-mediated responses in macrophages . For that , we stimulated BMDMs generated from wild-type ( WT ) and TLR9−/− mice with purified P . brasiliensis DNA . Our results showed that purified P . brasiliensis DNA induced the secretion of TNF-α and IL-6 by BMDMs in a TLR9 dependent way ( Fig . 1A ) . Next , we tested if TLR9 could also recognize P . brasiliensis DNA in the yeast cellular context and whether this depended on the cellular physiological stage and integrity . We stimulated WT and TLR9−/− BMDM with either P . brasiliensis cells from exponential or late-stationary growing cultures or with lysed cells . We found that BMDMs stimulated with P . brasiliensis yeast-form produced TNF-α in a TLR9 dependent manner ( Fig . 1B ) , although the response was higher for late-stationary growing and lysed cells . In addition to cytokine expression , TLR triggering also associates with the onset of phagocytosis [44] , [45] . We next investigated if TLR9 activation by the yeast-form of P . brasiliensis impacted phagocytosis . The percentage of phagocytosis of P . brasiliensis by BMDMs was significantly reduced when P . brasiliensis yeast cells were treated with DNase or when TLR9−/− BMDMs were used ( Fig . 1C ) . When heat inactivated DNase was used , the percentage of phagocytosis by WT macrophages was similar to that observed when no treatment was performed , whereas TLR9−/− macrophages maintained the phagocytic profile , indicating that DNase is not interfering with phagocytosis . Thus , our data indicate that TLR9 activation is required for maximal P . brasilisensis phagocytosis . Given the in vitro impact of P . brasiliensis recognition by TLR9 , we next sought to investigate a role for this receptor during the course of an in vivo experimental infection . WT and TLR9−/− mice were intravenously infected with P . brasiliensis and the survival rates followed over time . We found that 24 days post-infection 100% of TLR9−/− mice had succumbed , whereas for WT mice this was only observed 31 days post-infection ( Fig . 2 ) . Furthermore , the estimated mean survival for WT mice was of 13 . 6±2 . 0 days , while for TLR9−/− mice it was reduced to 6 . 8±1 . 4 days ( p<0 . 001; Fig . 2 ) . Even more striking was the observation that during the first 48 h post-infection , TLR9−/− infected mice showed severe impaired mobility , with a significant number of animals being humanely sacrificed as a result . As shown in Fig . 2 , during this early period the mortality of the TLR9−/− mice was approximately 43% , while WT mice showed no signs of disease . Thus , our data highlight an unexpected protective role of TLR9 during the early phases of P . brasiliensis infection . In view of the key protective role observed for TLR9 during the early stages of P . brasiliensis infection , we next investigated several parameters that could be associated with the premature death of TLR9−/−- infected animals . Since during the initial period of infection ( the first 48 h ) differences in the lungs of infected animals are difficult to assess , we performed a histological analysis of the liver of the infected animals . We found that whereas the livers of infected WT animals showed a normal structure , those of TLR9−/− hosts presented small areas of granulocytes/neutrophil infiltrates and of necrosis ( Fig . 3 ) . To dissect further if the histological differences observed were associated to the intensity of the immune response between WT and TLR9−/− animals , we assessed cytokine expression in the liver and blood of infected animals . In the liver , the expression of both pro- and anti-inflammatory cytokines , namely TNF-α , IL-6 and IL-10 , was increased in the absence of TLR9 ( Fig . 4A–E ) . Likewise , a significant increase of circulating levels of IL-6 was observed in TLR9−/−-infected mice ( Fig . 4G ) . Moreover , a trend towards higher levels of circulating TNF-α was detected ( Fig . 4F ) whereas those of IL-10 were below detection limit ( data not shown ) . To further validate if the absence of TLR9 correlated with an exacerbated immune response to P . brasiliensis , and to study differential patterns of cellular recruitment in WT versus TLR9−/− mice , we used a model of i . p . infection The analysis of the peritoneal exudates revealed that the cytokine expression was increased in the absence of TLR9 ( Fig . 5A–E ) . Of notice , a marked increased expression of both MIP-2 and IL-17 , known to be associated with neutrophil recruitment [46] , [47] , [48] , was found in the absence of TLR9 ( Fig . 5C , D ) . Consistently , in TLR9−/− mice we found an increased recruitment to the peritoneal cavity of both mononuclear cells ( macrophages and dendritic cells ) and neutrophils , with a special significance of the latter ones ( Fig . 5F–H ) . Thus , the data obtained for peritoneal infection recapitulates that obtained for intravenous infection . Considering the high influx of neutrophils to the peritoneal cavity of TLR9−/− mice , which correlated with the high expression of MIP-2 and IL-17 found in the liver and peritoneal cavity , we next evaluated if this could be the detrimental factor during the infectious process of TLR9−/− hosts . For this purpose , we depleted neutrophils in WT and TLR9−/− mice prior to infection with P . brasiliensis , using a specific monoclonal antibody ( MAb RB6-8C5 ) . I . p . injection of MAb RB6-8C5 abrogated neutrophils from 6 h to up to 48 h post-injection ( data not shown , [40] ) . As shown in figure 6 , depletion of neutrophils reverted the highly susceptible phenotype observed in TLR9−/− mice during the first two days of infection . Depletion of neutrophils in WT mice had no influence on the mice survival during the first 48 h post-infection ( Fig . 6 ) . Therefore , a high neutrophil recruitment to the site of infection appears to be responsible for the death observed in the absence of TLR9 during the early stages of P . brasiliensis infection .
Since the discovery of TLRs and their major role in host recognition of conserved molecular structures from microorganisms , particularly those of invading pathogens , enormous advances have been made in comprehending how the immune system responds to pathogenic organisms . It is well established that the first phase of an immune response involves innate immune mechanisms [49] , namely those triggered by TLR activation [10] , [50] , [51] . These receptors are present in neutrophils , monocytes and macrophages , cells that besides their phagocytic activity are crucial for the signaling and amplification of the response against the pathogen [43] . The knowledge on PRR recognition and activation of an efficient immune response against P . brasiliensis has progressively increased , mainly in what concerns TLR2 , TLR4 , and Dectin-1 [24] , [52] . However , a role for TLR9 during P . brasiliensis infection has only been addressed in the context of vaccination [37] , despite the evidences for the involvement of this TLR in other fungal infections [27] , [28] , [30] , [53] . Considering the high DNA content and number of unmethylated CpG oligonucleotides of this multinucleated fungus [33] , a role for this receptor during P . brasiliensis infection is likely . We herein demonstrate that P . brasiliensis purified DNA activated TLR9 in macrophages , leading to the expression of cytokines . This is in line with previous studies showing that fungal DNA is recognized by TLR9 . A . fumigatus DNA stimulates the production of pro-inflammatory cytokines in mouse and human dendritic cells [30] . Similarly , murine dendritic cells express IL-12p40 and CD40 upon stimulation with DNA from C . neoformans [27] . Human monocytes and macrophages from TLR9−/− mice were described to produce less IL-10 than cells from control mice when stimulated with C . albicans [28] . Despite TLR9 activation by purified P . brasiliensis DNA , upon macrophage stimulation with P . brasiliensis yeast cells grown to the exponential phase , TLR9 was activated in a lesser extent . This finding is likely due to the low exposure of DNA in this case , as when yeast cells on late stationary phase or lysed cultures were used , TLR9 recognized P . brasiliensis more prominently , leading to the production of pro-inflammatory cytokines by macrophages . Even though intact P . brasiliensis yeast cells did not fully activate TLR9 to induce the production of pro-inflammatory cytokines , in the absence of this receptor macrophages showed a decreased ability to phagocyte P . brasiliensis . Our findings are in line with previous studies showing that P . brasiliensis DNA activates macrophages , promoting their capacity to phagocyte P . brasiliensis [33] . Altogether , our data indicate that TLR9 triggering affected the overall responsiveness of macrophages to P . brasiliensis . Furthermore , during infection , the release of DNA from P . brasiliensis is expected to occur following fungal cell death , thus suggesting that TLR9 activation in vivo is very likely to happen . To assess the role of TLR9 in P . brasiliensis infections , we used in vivo models of infection . As our results show , TLR9 is crucial for mice survival in early times of infection ( first 48 h ) . Remarkably , we found that , in mice showing severe signs of disease , lack of TLR9 increased the expression of pro-inflammatory cytokines in the liver . Earlier studies on immune responses to other organisms have revealed that an excessive inflammatory response , mainly due to high levels of TNF , can lead to premature death of the host cells [54] , [55] , [56] , [57] . In the context of P . brasiliensis infection , it was demonstrated that an excessive inflammatory response may be detrimental rather than protective . A more severe disease development in mice susceptible to P . brasiliensis was associated with the presence of increased IL-12 and IFN-γ levels in the lungs , suggesting that the production of pro-inflammatory mediators does not always correlate with immunoprotection [58] . In addition to high TNF expression , we also found increased expression of MIP-2 and IL-17 in the liver and cells from the peritoneal cavity of P . brasiliensis-infected TLR9−/− mice . A detrimental role of IL-17 during P . brasiliensis infection was previously described as TLR2 and TLR4 deficiency associate with an increase of Th17 responses , lung pathology and more severe forms of infection [25] , [26] . Both IL-17 and MIP-2 have been previously associated with neutrophil recruitment [47] , [48] , [59] , [60] . In line with this , we observed a high influx of granulocytes/neutrophils into the peritoneal cavity of TLR9−/− i . p . infected animals . This enhanced neutrophil recruitment could thus be contributing to the detrimental response observed in P . brasiliensis-infected TLR9−/− animals . Indeed , upon transient neutrophil depletion , TLR9−/− mice survived during the first 48 hours post-infection , resembling the phenotype of WT mice . Therefore , the susceptibility profile observed for TLR9−/− mice early after infection with P . brasiliensis likely associates with an exacerbated neutrophil recruitment to the site of infection and/or with a particular detrimental phenotype of neutrophils . Although neutrophils are crucial during acute inflammatory response and subsequent resolution of fungal infection , in some situations , due to excessive release of oxidants and proteases , these cells may be responsible for injury to organs and fungal sepsis [61] , [62] . In addition to fungal infection , neutrophilia can also be harmful to the host in the context of other infections , such as M . tuberculosis and P . aeruginosa [63] , [64] , [65] , [66] . Our data implicating TLR9 in the phagocytic activity of macrophages raises the hypothesis that , in the absence of TLR9 , less P . brasiliensis cells are phagocytized . Thus , in the absence of this receptor , the extracellular P . brasiliensis cells may contribute to an exacerbated recruitment of neutrophils , which can result in a deregulated immune response . Since a significantly higher recruitment of dendritic cells and neutrophils is observed upon infection of TLR9−/− hosts , it is possible that the higher cytokine expression observed in this scenario results from the fact that more producing cells are present . Therefore , a fine tuned balance is required during infection with P . brasiliensis , in order to protect the host from infection . Several mechanisms must operate to achieve this balance , as is the case reported for TLR2/TLR4 activation [25] . It is also important to refer that P . brasiliensis can be triggering TLR9-independent mechanisms in macrophages , as supported in the literature [67] , [68] . One cannot rule out the hypothesis of a parallel activation of other receptors together with TLR9 . Several studies refer that Dectin-1 interaction with TLR9 results in a sinergistyc induction of IL-10 , TNF-α , IL-2 , IL-6 and IL-23 and down-regulation of IL-12 [69] , [70] . Studies with A . fumigatus reported a link between TLR2-mediated recognition and the phagocytic response [71] , whereas internalization of TLR2 with A . fumigatus phagossome was demonstrated [72] . As it has been shown for TLR2/TLR4 , it is also possible that TLR9 signaling is involved in the instruction of appropriate regulatory T cell responses , a hypothesis currently under investigation . Overall , this study highlights the relevant role of TLR9/neutrophils for the nature of the immune response to P . brasiliensis , and paves the way to the development of new preventive/therapeutical strategies , as resistance patterns of the host can be of great value on the comprehension of susceptibility and pathogenesis of P . brasiliensis infections . | Paracoccidioides brasiliensis is an etiological agent of paracoccidioidomycosis , one of the most prevalent systemic mycosis in South America , affecting over 10 million people . One of the hallmark features of this fungus is its multinucleated nature , where a single P . brasiliensis cell can have from one to over 50 nuclei . Thus , fungal cell death may culminate in the release of large amounts of DNA and in the activation of the host innate immune system via toll like receptor ( TLR ) 9 , a member of the TLR family . In this study , we evaluated the contribution of TLR9 activation during the infection by P . brasiliensis using an experimental model of P . brasiliensis infection of wild-type and TLR9−/− mice . We found that the lack of this receptor results in tissue damage and increased expression of cytokines , culminating in a higher death rate of TLR9−/− mice early post infection . This was mainly associated to neutrophils , as depletion of these cells , during the first 48 hours post-infection , converted the early response of TLR9−/− mice to infection to that of wild-type mice . Our work provides evidence for the protective effects of TLR9 activation during early stages of P . brasiliensis infection , highlighting the relevance of a balanced innate immune response to avoid tissue damage . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"clinical",
"immunology",
"genetics",
"immunology",
"biology",
"microbiology",
"clinical",
"genetics"
] | 2013 | TLR9 Activation Dampens the Early Inflammatory Response to Paracoccidioides brasiliensis, Impacting Host Survival |
In Timor-Leste there have been intermittent and ineffective soil-transmitted helminth ( STH ) deworming programs since 2004 . In a resource-constrained setting , having information on the geographic distribution of STH can aid in prioritising high risk communities for intervention . This study aimed to quantify the environmental risk factors for STH infection and to produce a risk map of STH in Manufahi district , Timor-Leste . Georeferenced cross-sectional data and stool samples were obtained from 2 , 194 participants in 606 households in 24 villages in the Manufahi District as part of cross sectional surveys done in the context of the “WASH for Worms” randomised controlled trial . Infection status was determined for Ascaris lumbricoides and Necator americanus using real-time quantitative polymerase chain reaction . Baseline infection data were linked to environmental data obtained for each household . Univariable and multivariable multilevel mixed-effects logistic regression analysis with random effects at the village and household level were conducted , with all models adjusted for age and sex . For A . lumbricoides , being a school-aged child increased the odds of infection , whilst higher temperatures in the coolest quarter of the year , alkaline soils , clay loam/loam soils and woody savannas around households were associated with decreased infection odds . For N . americanus , greater precipitation in the driest month , higher average enhanced vegetation index , age and sandy loam soils increased infection odds , whereas being female and living at higher elevations decreased the odds of infection . Predictive risk maps generated for Manufahi based upon these final models highlight the high predicted risk of N . americanus infection across the district and the more focal nature of A . lumbricoides infection . The predicted risk of any STH infection is high across the entire district . The widespread predicted risk of any STH infection in 6 to 18 year olds provides strong evidence to support strategies for control across the entire geographical area . As few studies include soil texture and pH in their analysis , this study adds to a growing body of evidence suggesting these factors influence STH infection distribution . This study also further supports that A . lumbricoides prefers acidic soils , highlighting a potential relatively unexplored avenue for control . ClinicalTrials . gov ACTRN12614000680662 .
Soil-transmitted helminth ( STH ) infections are predominantly a disease of poverty , typically prevalent in poor tropical and subtropical regions . [1] In 2010 , the four main species of human STH were estimated to infect 1 . 5 billion people worldwide; with whipworm ( Trichuris trichiura ) , roundworm ( Ascaris lumbricoides ) and the two main species of hookworm ( Ancylostoma duodenale and Necator americanus ) respectively infecting 464 . 6 million , 819 . 0 million , and 438 . 9 million people . [2] The majority of STH infections are insidious for hosts living in impoverished conditions , and have a demonstrable impact on an individual’s health and wellbeing . [1 , 3] A . lumbricoides and T . trichiura are thought to contribute to malnutrition , whereas hookworm and T . trichiura infections have been associated with iron-deficiency anaemia . [4 , 5] All four main human STH species have been associated with impaired childhood growth . [5–7] Anaemia and malnutrition may have a negative long-term impact on an individual’s health and productivity , [6 , 7] with resulting economic ramifications contributing to the cycle of poverty . [3] To reduce the impact of STH on communities , comprehensive control strategies are required . An impediment to implementing cost-effective control programs is a lack of accurate information detailing the geographic distribution of STH infections . [8] Since the transmission of infective STH relies upon favourable environmental conditions , environmental factors may be used to identify high risk areas . [9] Over the past two decades , geographic information systems ( GIS ) coupled with remotely sensed environmental data have been used to identify areas of high STH infection risk in several countries , [10–15] allowing governments to develop cost-effective targeted STH control strategies . [16] . Due to constrained resources , there has been little useful geographical information obtained about STH distributions in Timor-Leste in the decade following the restoration of independence in 2002 , making it difficult to target interventions to the locations where they are needed the most . In 2005 the Timor-Leste Ministry of Health ( MoH ) in conjunction with the World Health Organization began a mass drug administration ( MDA ) program to control STH , however this program ceased in 2008 due to funding shortages . [17] Since then there has been limited distribution of albendazole to children presenting at mobile health units that attempt to provide regular health care to remote populations . In 2012 , a national cross-sectional study found 29 . 0% of children in grades 3–5 were infected with STH , ranging from 4% in the Manatuto district to 55% in the Dili district . [17] High STH levels in 2012 , following years of intermittent control strategies , suggest the MoH programs have been inadequate in controlling STH . [17] In 2013 , the MoH developed an integrated national plan targeting neglected tropical diseases , including a seven year MDA program aimed at reducing STH infections and eliminating lymphatic filariasis , due to begin in 2014 . [18] However , limited resources have delayed program implementation , with the MoH re-starting the program in 2015 in a limited number of districts . [19] Given the limited resources available in Timor-Leste , cost-effective methods to identify populations at high risk of STH infections are needed . In 2012 , the ‘WASH for Worms’ ( W4W ) randomized controlled trial was initiated in the Manufahi district , Timor-Leste , aiming to evaluate the effectiveness of a community-based water , sanitation and hygiene ( WASH ) programme in reducing parasitic infections above gains achieved through mass chemotherapy alone . [19] In the context of the W4W trial , cross-sectional surveys were conducted at baseline on 24 villages , involving the collection of stool samples to assess parasitic infection status , as well as demographic data and geographic coordinates for each participating household . [19] Using this cross-sectional data , this study aimed to describe and predict the distribution of STH infection in Manufahi district using environmental variables . This study specifically focused on A . lumbricoides and N . americanus , as analysis has shown these species of STH to be most prevalent in the study area . [20] .
As part of the baseline cross-sectional surveys , informed written consent from study participants aged ≥18 years and parents/guardians of children <18 years was obtained . Additionally , informed written assent was sought from children 12–17 years old . Informed consenting illiterate participants provided an ink thumb print in lieu of a signature . [19] Ethics approval for the W4W RCT was granted from: The Australian National University Human Ethics Committee ( protocol: 2014/311 ) , The Timorese Ministry of Health Research and Ethics Committee ( reference: 2011/51 ) , and The University of Queensland Human Research Ethics Committee ( project number: 2011000734 ) . The W4W trial is registered with the Australian and New Zealand Clinical Trials Registry ( ACTRN12614000680662 ) . Manufahi is one of 13 districts in Timor-Leste . The district is rural and largely agrarian , with farmers predominantly using traditional small scale subsistence farming methods . [21 , 22] The climate is tropical , with a relatively cool dry season between July–November and a humid wet season between December–May ( these seasons reportedly vary ) . [23] Based on WorldClim data from 1950−2000 , [24] the average annual precipitation for Manufahi was approximately 1900 mm , whilst the average annual temperature was 24 . 5°C . The terrain is highly variable , ranging from flat coastal plains in the south composed of predominantly clay soil , to mountain ranges in the north with predominantly sandy loam and clay soils . [25] Villages involved in the cross-sectional baseline survey of the W4W trial were selected using detailed criteria . [19] Specifically , all selected villages were rural , had low access to WASH , and a willingness to participate in the W4W trial . [19] . All residents from the selected villages were eligible to participate in the cross-sectional survey provided they were present , greater than or equal to one year old , and they or a parent or guardian ( if under 18 ) provided informed consent . [19] Baseline parasitological surveys were undertaken in the enrolled villages prior to any interventions , and were staggered over an 18 month period between May 2012 and October 2013 . [19] At baseline , standardised questionnaires were administered to each participant inquiring about demographic factors and questions pertaining to WASH practices . [19] Additionally , standardised household and village questionnaires were administered to collect household and village characteristics , with global positioning system devices ( Garmin handheld GPS , Garmin Ltd ) used to determine the exact geographical position of each household . [19] Each participant was asked to provide one faecal specimen , which was separated into 2–3 ml aliquots and preserved in 15 ml centrifuge tubes containing 6ml of potassium dichromate ( 5% ) for molecular analysis conducted at the QIMR Berghofer Medical Research Institute . Multiplex real-time polymerase chain reaction ( qPCR ) was used to identify and quantify the following STH infections: Ascaris spp . , N . americanus , Ancylostoma spp . , and Trichuris trichiura . [26] Given the W4W trial had molecular evidence that the vast majority of Ascaris spp . infections were indeed A . lumbricoides infections , [27] Ascaris spp . infections will herein be referred to as A . lumbricoides infections . All eligible participants in the W4W cross-sectional baseline surveys were included in this analysis where age , sex and infection status were available . Of the 2 , 827 residents present and eligible to participate in the baseline cross-sectional survey , 99 . 1% ( 2 , 802/2 , 827 ) consented , and 78 . 5% ( 2 , 219/2 , 827 ) had infection status determined using qPCR ( due to not all participants providing stool ) . Of those with infection status determined ( 2 , 219 ) , sex status was missing for 0 . 99% ( 22/2 , 219 ) , and age for 0 . 1% ( 3/2 , 219 ) . After restricting the sample , 2 , 194 people in 606 households were in the study population for the analysis presented here . The following environmental data were obtained for analysis: temperature , precipitation , elevation , soil texture and pH , landcover and vegetation ( Table 1 ) . All environmental variable processing was conducted in the GIS ArcMap 10 . 3 ( ESRI , Redlands , CA ) , unless otherwise specified . Long-term average annual and seasonal temperature ( average , maximum and minimum ) and precipitation variables were created using data retrieved from WorldClim at 1 km spatial resolution . [24] These layers were produced by using a thin-plate smoothing spline algorithm to interpolate data collected from global weather station sources between 1950 − 2000 , [24] and have been validated for Timor-Leste by Molyneux et al . [28] Seasonality ( e . g . wettest quarter ) was assigned based upon the WorldClim monthly averages for Manufahi . [24] . Average vegetation indices ( normalized difference vegetation index ( NDVI ) and enhanced vegetation index ( EVI ) ) , landcover , elevation and slope variables were created using data courtesy of the NASA EOSDIS Land Processes Distributed Active Archive Center ( LP DAAC; see https://lpdaac . usgs . gov/ ) . 16 day NDVI/EVI ( MOD13Q1 ) data were obtained from January 1st 2012 to December 31st 2013 , whilst yearly composite landcover ( MCD12Q1 ) data were only available for 2012 . [29 , 30] Prior to developing long-term average variables for vegetation and landcover , the MODIS Reprojection Tool ( version 4 . 1 ) [31] was used to convert data into a usable projection system for use in ArcGIS , and pixels underwent quality assessment . Landcover classes were defined using the International Geosphere Biosphere Programme ( IGBP ) classification system , which provides 11 natural vegetation classes , three non-vegetated classes , and three urban/agricultural classes . [32] The Advanced Spaceborne Thermal Emission and Reflection Radiometer ( ASTER ) global digital elevation model ( GDEM ) data at 30m spatial resolution were used , [33] and smoothed using Sun’s denoising algorithm . [34] This smoothed data were also used to determine the slope ( in degrees ) . Soil pH and texture variables were created based upon data from the 1960’s ‘Os Solos De Timor’ study , which sampled soil layers from over 285 locations across Timor-Leste . [35] Soil data were accessed from the Seeds of Life Timor website ( http://seedsoflifetimor . org/ ) . [25] Soil pH was classified using the United States Department of Agriculture ( USDA ) classification system . [36] . To be used in the univariate and multivariate analysis described below , all environmental variables were obtained in a 1 km radius around each household , and land attributes ( soil texture , pH and landcover ) also obtained for each household point ( HHP ) . One kilometer buffer variables for EVI , NDVI , temperature , rainfall , elevation and slope were determined based on the median value within a 1 km radius of each household . The median value was thought to provide a more accurate and representative measure of exposure over an area thought to represent the average range of movement of participants from the HHP . [37] For land attributes , each household was assigned the land attribute category covering the majority of area within a 1 km radius . However , in some villages the land attributes differed markedly across a 1 km radius . Therefore , to account for potential micro-climates around the household , categorical land attribute variables were also obtained for each HHP . When determining soil texture and pH for each HHP , data were missing for some households in Lalmamir , Datina , Sarin and Ahiklatun villages . In these cases , the soil texture and pH were assigned based on the geographically closest soil type . All analysis was conducted using the R Statistical Software , version 3 . 2 . 0 . [38] Some variables were transformed to improve model stability and interpretation of results , as outlined below . Elevation was expressed per 100 m , precipitation in centimetres , and EVI/NDVI in 0 . 01unit increments . As the categorical variable , soil texture , had several categories , some of which had a low sample size , univariable multilevel mixed-effects logistic regression with random-effect terms at the village and household levels was run for this variable against each outcome . As loam and clay loam categories had very similar odds ratios and it made sense to combine them based on composition , they were combined . This produced a soil texture variable with five groups: ‘clay/clay loam’ , ‘loam’ , ‘sandy clay’ , ‘sandy loam’ and ‘variable’ soil type . To prevent highly collinear variables from being in the same model , variables were grouped into domains within which many variables were collinear . To determine domain groupings , the Pearson’s correlation coefficient ( r ) for each pair of continuous variables was assessed , and separate domains created for each categorical variable . Specifically , these domains were: elevation/temperature , rainfall/slope , EVI/NDVI , landcover , soil type and soil pH , sex and age . Univariable mixed-effects logistic regression with random effects at the village and household level was conducted for each independent variable . Additionally , for continuous independent variables , univariable models including orthogonal quadratic transformations of continuous independent variables were also considered , to allow for the possibility of a non-linear relationship between each continuous environmental variable and the probability of infection . Multivariable multilevel mixed-effects regression with random-effect terms at the village and household levels accounting for age and sex were developed by including the independent variable from each domain with the lowest Akaike Information Criterion ( AIC ) . Two variables from separate domains were not allowed to enter the same model if the pair’s r was greater than 0 . 90 . To create the final model for each helminth , working models were simplified for parsimony by dichotomising categorical variables if this reduced the AIC , followed by removing variables with p > 0 . 25 on the Wald test . To determine if the final models adequately accounted for spatial autocorrelation , semivariograms were created based on the village random intercepts for each given model . Semivariograms describe how data are related with respect to distance and direction , with a semivariogram presenting the similarity between observations ( semivariance ) at different separation distances . [39] Omni-directional semivariograms based on the village random intercepts were developed in R using the geoR package . [40] Lack of distinctive spatial patterns in these semivariograms ( S1 Fig ) supported that the final models adequately accounted for spatial clustering at the village level . Separate predictive risk maps for A . lumbricoides and N . americanus in the highest risk age and sex groups were created based upon the parameter estimates of the various environmental and demographic variables included in the final model for each species . For the A . lumbricoides model , the coefficient for children aged 6-<18 years was included , and for the N . americanus model , the coefficient for adults ( ≥18 year old ) was included . For display purposes ( i . e . , map visualization ) smoothing was used . In this context , missing predictions arising from missing soil data ( texture and pH ) were completed by allocating the mean value of the neighboring cells in a 2 by 2 pixel area . To develop the risk map for any STH in 6-<18 year olds , separate predictive risk maps for A . lumbricoides and N . americanus were developed as described above; however , the models used to create these maps incorporated the coefficient for children aged 6-<18 years , and separate risk maps were generated for both males and females ( adding the coefficient for females in the latter ) . Missing predictions were handled in the same manner as described above . For each sex , the predicted risk maps for A . lumbricoides and N . americanus in 6–18 year olds was then summed , and the product of the risk maps subtracted in order to address coinfections . This manner of addressing coinfections assumes that individuals infected with one species were no more or less likely to be infected with the second species than uninfected individuals . The predicted risk maps for any STH in males and females were then averaged , assuming equal distribution of both sexes , to produce an overall predicted risk map of any STH in 6 to 18 year olds . Predictions were then categorized based on the World Health Organisation ( WHO ) thresholds for deworming school aged children , namely low risk ( <20% ) , moderate risk ( ≥20-<50% ) and high risk ( ≥50% ) [41] .
The study population ( n = 2 , 194 ) was 51 . 3% female , and ranged in age from 1–92 years ( median 19 . 0 years ) ( Table 2 ) . The number of participants varied between villages , ranging from 29–318 ( mean 91 . 4 participants ) . Overall , the unadjusted prevalence for A . lumbricoides was 24 . 2% ( ranging from 0 . 0–79 . 6% between villages ) and 60 . 7% for N . americanus ( ranging from 39 . 2–94 . 5% between villages ) . Since the age structure varied across the villages , the age standardised prevalence ( ASP ) of A . lumbricoides and N . americanus was determined for each village . A . lumbricoides ASP was heterogeneous across villages , ranging from 0% –82 . 0% ( Fig 1A ) . N . americanus ASP was higher overall , ranging from 36 . 2–94 . 7% ( Fig 1B ) . Results of univariable mixed-effects logistic regression analysis for A . lumbricoides and N . americanus are detailed in S1–S4 Tables . For A . lumbricoides , the environmental variable with the lowest AIC in each group was: mean temperature in the coldest quarter , average NDVI , slope , soil pH at HHP ( three groups ) , soil texture at HHP and landcover at a 1 km buffer ( S1 and S2 Tables ) . The AIC’s of univariable models with only the linear environmental term ( S1 Table ) did not meaningfully differ ( < 4 points , guided by the rule of thumb reported by Burnham and Anderson [42] ) to the respective AIC’s of models containing both the linear and the quadratic terms ( S2 Table ) , and thus for parsimony , linear forms of continuous environmental variables were included in the working multivariable model . For N . americanus , the environmental variable with the lowest AIC in each group was: elevation ( quadratic form ) , average EVI , mean precipitation in the driest month , predominant soil pH ( three groups ) and landcover in 1km buffer of household , and soil texture ( five groups ) at the HHP ( S3 and S4 Tables ) . Although models including the quadratic forms of temperature and elevation lowered the AIC by approximately 10 points ( S4 Table ) , they were unstable ( large standard errors ) , and so the continuous and categorical forms of elevation were considered for model selection ( S3 Table ) . The full and final environmental multivariable models for A . lumbricoides and N . americanus infections are detailed in Tables 3 and 4 respectively . The AIC of the working A . lumbricoides model was improved by binarising soil texture and landcover variables by combining several categories , and removing NDVI . For N . americanus , the model was improved by binarising soil texture and removing soil pH and landcover . The continuous form of elevation was used in the final N . americanus model as this form of elevation resulted in the model with the lowest AIC . For A . lumbricoides , the final model following variable elimination indicated temperature in the coldest quarter , slope , soil pH , soil texture , landcover and age were associated with the odds of infection in the study site ( Table 3 ) . Increasing temperatures in the coldest quarter of the year were associated with a decreased odds of infection ( OR: 0 . 74 , 95% Confidence Interval [CI]: 0 . 60–0 . 92 , p = 0 . 006 ) . Clay loam/loam soils around the household were associated with reduced infection odds compared to other soil types ( OR: 0 . 22 , 95% CI: 0 . 09–0 . 54 , p < 0 . 001 ) . Having predominantly woody savanna in a 1 km radius of households significantly lowered infection odds compared to other vegetation types ( OR: 0 . 51 , 95% CI: 0 . 28–0 . 93 , p = 0 . 027 ) . Alkaline soils were associated with lower odds of infection compared to acidic soils ( OR: 0 . 27 , 95%CI: 0 . 08–0 . 96 , p = 0 . 043 ) . After adjusting for environmental factors , a non-linear association between age and infection odds was evident , as 6–<18 year olds had significantly higher infection odds compared to 1–<6 year olds ( OR: 1 . 68 , 95% CI: 1 . 13–2 . 49 , p = 0 . 010 ) , and no significant difference between 6-<18 year olds and ≥18 year olds ( OR: 0 . 77 , 95% CI: 0 . 53–1 . 12 , p = 0 . 173 ) . The random effect terms included in the model ( village , and household ) reveal there was greater variation at the village level ( variance: 0 . 77 ) than the household level ( variance: 0 . 55 ) . For N . americanus , the final model , after variable elimination , suggests elevation , precipitation , average EVI , soil texture , sex and age were associated with infection ( Table 4 ) . Greater elevations were associated with reduced odds of infection ( OR: 0 . 91 , 95% CI: 0 . 85–0 . 97 , p = 0 . 004 ) , whilst higher precipitation in the driest month of the year significantly increased infection odds ( OR: 6 . 54 , 95% CI: 2 . 94–14 . 57 , p < 0 . 001 ) . Higher EVI values were associated with increased infection odds ( OR: 1 . 07 , 95% CI: 1 . 00–1 . 15 , p = 0 . 040 ) . Having predominantly sandy loam soils in a 1 km buffer of the household was associated with greater than double the odds of infection compared to other soil types ( OR: 2 . 50 , 95% CI: 1 . 56–4 . 01 , p < 0 . 001 ) . Sex was significantly associated with infection , with females having a significantly reduced odds of infection compared to males ( OR: 0 . 34 , 95% CI: 0 . 27–0 . 43 , p < 0 . 001 ) . In relation to age , compared to 1–<6 year olds , 6–<18 year olds had 4 . 98 times higher infection odds ( 95% CI: 3 . 54–7 . 02 , p < 0 . 001 ) , and adults had 9 . 37 times higher odds ( 95% CI: 6 . 68–13 . 16 , p < 0 . 001 ) . For the random effect terms , village and households , there was greater variation at the household level ( variance = 0 . 69 ) than the village level ( variance = 0 . 09 ) . Fig 2A and 2B present the predicted risk of infection with A . lumbricoides and N . americanus in males in the highest risk age groups ( 6–<18 years and ≥18 years respectively ) . Note , risk predictions for other age and sex categories will have an identical spatial distribution , but with an overall lower mean . In Fig 2A , the predicted risk of infection with A . lumbricoides was higher in the central and northern parts of Manufahi , which have lower temperatures , more acidic soils and higher elevation and precipitation compared to the lower risk southern coastal regions . The sudden change from low to medium/high predicted risk is due to a combination of decreased temperatures and the soils in the region being more acidic and of variable soil types ( predominantly sandy clay and sandy loam ) , all of which are associated with an increased odds of A . lumbricoides infection ( Table 3 ) . In Fig 2B , the predicted risk of N . americanus was high across the district , with the highest predicted risk concentrated in the central region . The coastal southeastern area had a lower risk , likely due to having higher temperatures , lower elevations , lower precipitation , and predominantly clay soils ( Table 4 ) . Some missing data are shown in the maps–these are locations where soil texture and pH data were unavailable , and where the infilling using the mean of the nearest 2 x 2 pixels was unable to provide an estimate due to there being no neighbouring pixels with prediction estimates . Fig 2C depicts the predicted risk of any STH in 6-<18 year olds , based on WHO prevalence thresholds for deworming in school aged children . Fig 2C highlights that the majority of 6-<18 year olds in Manufahi have a high predicted risk of infection ( ≥50% ) , with the remainder having moderate predicted risk ( ≥20 to <50% ) .
Our results support that A . lumbricoides and N . americanus infections favoured different environmental conditions in the study area , influencing the predicted geographical distribution of human infection with these species in Manufahi district . Loam and clay loam soils were associated with lower odds of A . lumbricoides infection whilst sandy loam soils were associated with increased odds of N . americanus infection , compared to other soil types . The finding that sandy loam soils were favourable for N . americanus is biologically plausible [9] and in accordance with a number of epidemiological studies . [43–45] The porous environment of sand provides a favourable environment for hookworm survival , offering drainage during wet conditions to prevent hookworm larvae from being waterlogged , whilst also enabling hookworm larvae to migrate downwards during hot and dry conditions to prevent dessication . [9] The finding that loam and clay loam soils were negatively associated with A . lumbricoides infection is in contrast to the available literature which suggests clay and loam soils are more favourable for A . lumbricoides egg development compared to sandy soils . [46 , 47] This was an unanticipated finding , but one explanation could be residual confounding , such as other soil properties , exposure to ultraviolet radiation , and human interaction . [8 , 9] . Contrasting relationships with temperature/elevation were observed for A . lumbricoides and N . americanus , suggesting that different species favour different climatic conditions . Higher temperatures in the coldest quarter of the year were associated with reduced odds of A . lumbricoides infection , suggesting A . lumbricoides favours cooler conditions in tropical areas . This supports epidemiological and biological evidence that A . lumbricoides is prone to desiccation in hot conditions . [9 , 48] In contrast , the odds of infection with N . americanus were higher at lower elevations that have slightly warmer temperatures , potentially reflecting the ability of hookworm to survive in warmer temperatures due to their motility . [9] This negative relationship has been well documented by studies across Africa . [49–52] . Having predominantly woody savanna in a 1 km radius of households was unfavourable for A . lumbricoides infection compared to other vegetation types assessed . Woody savannas have an understorey plant system , and a forest canopy that covers between 30–60% of the environment . [53] This reasonably high level of vegetation , shade , and thus moisture , should in theory be conducive to A . lumbricoides survival and development . [9] This finding may therefore reflect micro-climates ( e . g . soil properties ) , human movement patterns , or possibly livelihoods and related occupational exposures , which may potentially reduce transmission in areas with woody savanna . Alkaline soils were associated with reduced odds of A . lumbricoides infection , suggesting A . lumbricoides favours acidic conditions . This finding is in agreement with Chammartin et al . who found acidic soils ranging between pH 5 . 35 and 5 . 65 were associated with increased odds of A . lumbricoides infection [11] . Although not evident in this study , Chammartin et . al also found a negative association between pH and the risk of hookworm infection , [11] which is supported by a laboratory study that reported that N . americanus favours acidic conditions . [54] Combined , these studies provide evidence to support further research into using products that reduce soil acidity as part of interventions to control STH , namely lime , which is used routinely in agriculture to increase productivity . [55 , 56] . Greater precipitation in the driest month of the year was associated with increased odds of N . americanus infection , supporting studies that find hookworm requires a minimum level of soil moisture or humidity throughout the year to survive and develop . [9 , 15 , 57] The positive association between EVI and N . americanus infection is consistent with Saathoff et al . [44] As EVI is a measure of vegetation and thus a proxy for environmental moisture and shade , this finding further supports that hookworm require sufficient soil moisture for survival and transmission . [9 , 54] . The age trends identified in this study were consistent with the available literature and earlier analysis . [20] School-aged children ( 6–<18 years ) had a greater odds of A . lumbricoides infection compared to pre-school aged children ( 1–<6 years ) and adults ( ≥18 years ) . [37 , 58–60] On the other hand , the odds of N . americanus infection increased with age . [1 , 7 , 58 , 61 , 62] Males were found to have a higher odds of N . americanus infection , whilst no association was evident for A . lumbricoides . The higher odds of hookworm infection in males may reflect occupational roles or behavioural factors that could increase exposure of males to infection . [63] . The predictive risk maps highlight that 6 to 18 year olds were at high to medium risk of having a STH infection in a large portion of Manufahi . Given that STH infections may have immediate and long term health and economic impacts , [3] the widespread predicted risk of STH in Manufahi provides evidence to support that a district wide STH control strategy is needed . Ideally any such strategy would include regular deworming with antihelminthic drugs such as albendazole or mebendazole integrated with WASH interventions . [64] Importantly , studies suggest co-infected individuals are at a higher risk of helminth related morbidity . [65 , 66] This is thought to be because of the morbidity associated with each species , [5–7] and because co-infected individuals generally have heavier infections of each worm species . [65–69] Our maps suggest that the central area of Manufahi is likely to have a higher presence of co-infections . The study has several strengths . Since the W4W trial is a randomised controlled trial , individual and household data were collected in a systematic manner using trial protocols , [19] reducing the likelihood of information bias . Importantly , misclassification was minimized in this study , as the W4W trial employed the use of qPCR to detect STH infections , which is a more sensitive technique compared to standard microscopy . [26] The use of qPCR and including entire communities ( as opposed to school-age children only ) may explain the higher infection levels reported here , compared to the 2012 national survey . As household GPS coordinates were available , the highest resolution open-access environmental data available could be used to determine individual exposures as accurately as possible , minimising measurement error and improving the accuracy of environmental parameter estimates . Our analysis included soil pH and soil texture data , which are often not available for spatial studies . We found both soil pH and soil texture variables were associated with the likelihood of STH infection , supporting the wider use of these variables in developing explanatory and predictive models for STH . Future studies should therefore endeavour to collect information on soil texture , pH and other properties ( e . g . ions , carbon to nitrogen ratio etc . ) to better define risk and potentially offer additional strategies for STH control . Our analysis however did have some limitations . The soil data used were collected in the 1960s and alongside missing data it may be inaccurate at the time of the study . Socio-demographic and behavioural factors were not included in the analysis as data were unavailable or outdated for all of Manufahi , meaning risk maps could not incorporate these variables . A separate analysis of behavioural and sociodemographic factors in the W4W trial provides detailed information on these factors . [20] Moreover , the W4W trial did not utilise a spatial sampling frame , as spatial analysis was not the main purpose of the trial . A spatial sampling frame would have been more efficient and provided a more uniform coverage of the study area . Furthermore , the “smoothed” maps were produced making the following assumptions: to address coinfections , we assumed that individuals infected with one species were as likely to be infected with a second species as uninfected individuals , and to produce the overall predicted risk map of any STH in 6–<18 year olds , we assumed equal distribution of both sexes in each prediction location . If different species of STH cluster in the same individual we might have overestimated the total prevalence of infection . Additionally , the actual prevalence in each location might be affected by an unequal distribution of the sexes but given that we did not have data on sex distributions in non-sampled locations we were unable to incorporate this in the predictions . However , we believe that neither issue was likely to change significantly the conclusions of the study . It should also be noted that the risk maps are based on the environmental and demographic variables included in the model . There will be other factors that influence the distribution of disease risk . Some of these factors vary at the village and household levels , and some of this variance would have been captured in the random effects . These maps therefore show the relative difference in risk of each location based on their environmental characteristics , adjusted for any confounding related to age and sex , and robust to clustering of the data at the village level . For the scale of the analysis ( a single district of Timor-Leste ) it is reasonable to assume the relationship between disease risk and the environmental characteristics is uniform across the region . Lastly , “smoothing” of missing values for soil texture and pH data may limit the accurate representation of risk in the small number of locations where these data were missing . In conclusion , in the Manufahi district , STH infection is strongly associated with several environmental factors . The widespread predicted risk of any STH infection across the study area provides good evidence that control strategies are required for the entire geographical area . Our research combined with other epidemiological and biological studies suggests that products that increase soil pH , such as liming , could be investigated as a potential avenue for control . | The majority of soil-transmitted helminth ( STH ) infections have long-term ramifications on an individual’s health and productivity , contributing to malnutrition , anaemia , and impaired childhood growth . In Timor-Leste there have been intermittent and ineffective STH control programs since 2004 . When resources are constrained , having information on the geographic distribution of STH is important to ensure limited resources are targeted to areas most in need . In this study we predicted the risk of Ascaris lumbricoides , Necator americanus and any STH infection in the Manufahi district of Timor-Leste , providing the first available risk maps for STH infection in any part of the country . To achieve this , we assessed the relationship between selected environmental factors and STH infection to create multivariable models accounting for potential clustering of infections in villages and households . Using these multivariable models , we predicted the risk of A . lumbricoides and N . americanus infection across the district based on the environmental variables . We found that the predicted risk of infection with N . americanus in the age group older than 18 years of age and with any STH in 6–18 year olds was high and widespread across Manufahi , supporting the need for control strategies across the entire geographical region . Our study also highlighted the focal nature of A . lumbricoides infection and its relationship with soil pH . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"helminths",
"grasslands",
"tropical",
"diseases",
"hookworms",
"geographical",
"locations",
"plant",
"communities",
"parasitic",
"diseases",
"animals",
"necator",
"americanus",
"plant",
"science",
"ascaris",
"ascaris",
"lumbricoides",
"neglected",
"tropical",
"diseases",
"plant",
"ecology",
"necator",
"soil",
"science",
"agriculture",
"timor-leste",
"people",
"and",
"places",
"helminth",
"infections",
"asia",
"ecology",
"nematoda",
"biology",
"and",
"life",
"sciences",
"soil-transmitted",
"helminthiases",
"agricultural",
"soil",
"science",
"organisms",
"terrestrial",
"environments"
] | 2017 | An environmental assessment and risk map of Ascaris lumbricoides and Necator americanus distributions in Manufahi District, Timor-Leste |
Cell-fate asymmetry in the predivisional cell of Caulobacter crescentus requires that the regulatory protein DivL localizes to the new pole of the cell where it up-regulates CckA kinase , resulting in a gradient of CtrA~P across the cell . In the preceding stage of the cell cycle ( the “stalked” cell ) , DivL is localized uniformly along the cell membrane and maintained in an inactive form by DivK~P . It is unclear how DivL overcomes inhibition by DivK~P in the predivisional cell simply by changing its location to the new pole . It has been suggested that co-localization of DivL with PleC phosphatase at the new pole is essential to DivL’s activity there . However , there are contrasting views on whether the bifunctional enzyme , PleC , acts as a kinase or phosphatase at the new pole . To explore these ambiguities , we formulated a mathematical model of the spatiotemporal distributions of DivL , PleC and associated proteins ( DivJ , DivK , CckA , and CtrA ) during the asymmetric division cycle of a Caulobacter cell . By varying localization profiles of DivL and PleC in our model , we show how the physiologically observed spatial distributions of these proteins are essential for the transition from a stalked cell to a predivisional cell . Our simulations suggest that PleC is a kinase in predivisional cells , and that , by sequestering DivK~P , the kinase form of PleC enables DivL to be reactivated at the new pole . Hence , co-localization of PleC kinase and DivL is essential to establishing cellular asymmetry . Our simulations reproduce the experimentally observed spatial distribution and phosphorylation status of CtrA in wild-type and mutant cells . Based on the model , we explore novel combinations of mutant alleles , making predictions that can be tested experimentally .
The asymmetric localization of proteins is critical for cell and/or tissue development in eukaryotic systems as diverse as S . cerevisiae [1] , C . elegans [2] , A . thaliana [3] , and D . melanogaster [4] . For years , spatial organization of cellular components was thought to be an exclusive feature of eukaryotes , but advances in microscopy and protein labeling over the past two decades have dispelled this notion [5] . The localization of cellular components—including lipids , DNA , RNA and proteins–is also an integral feature of prokaryotic cells; observed to play a role in the growth , function and survival of many bacteria , including E . coli [6] , B . subtilis [7 , 8] , V . cholerae [9] , S . flexnerii [10 , 11] . However , with roughly 10% of its proteins having the potential to localize [12] , Caulobacter crescentus serves as the model bacterium to study subcellular localization of proteins in prokaryotes . In Caulobacter , the non-uniform distribution of proteins is visibly manifested in the asymmetric division cycle that gives rise to two morphologically and functionally distinct daughter cells [13–15] . Furthermore , subcellular localization of macromolecules influences many physiological attributes of Caulobacter cells , such as growth [16 , 17] , cell shape [18 , 19] , morphogenesis [20] , differentiation [21 , 22] , stringent response [23 , 24] , and cell division [25] . Caulobacter shares many regulatory genes with other species of alpha-proteobacteria , including species that are of importance to agriculture and medicine , such as the nitrogen-fixing Sinorhizobium meliloti , the plant pathogen Agrobacterium tumefaciens , and the mammalian pathogens Rickettsia prowazekii and Brucella abortus [26 , 27] . While mounting evidences show causal links between protein localization and cell function in these bacteria [20 , 28–34] , the underlying molecular mechanisms that enable the cell to use subcellular protein gradients to achieve complex cellular behavior are not completely understood . The bacterium Caulobacter crescentus undergoes asymmetric division to give rise to two non-identical daughter cells , called a stalked cell and a swarmer cell . The sessile and replication-competent stalked cell is anchored to the substratum , while the motile but replication-quiescent swarmer cell swims to a new locale , before shedding its flagellum and differentiating into a stalked cell . This dimorphism enables the bacterial population to disperse and survive in the low-nutrient , aquatic environments where Caulobacter is naturally found [15] . The precursor to asymmetric division is the predivisional cell , which is characterized by a stalk at one pole and nascent swarmer apparatus at the opposite pole . The swarmer , stalked and predivisional cells represent three distinct developmental stages that define the Caulobacter cell cycle . Progression through this cycle is dictated by the phosphorylation status of the master regulator CtrA , which serves as a transcription factor for nearly 100 genes [35] . In particular , by regulating expression of the hemimethyltransferase , CcrM , CtrA controls the methylation state of DNA in stalked and predivisional cells [36–40] , and by binding to the origin of replication , the phosphorylated form of CtrA ( CtrA~P ) inhibits DNA replication in swarmer cells [41] . A gradient of CtrA phosphorylation is established in predivisional cells , with CtrA~P high in the swarmer end and low in the stalked end . As a result , one daughter cell inherits the phosphorylated form of CtrA , and the other daughter cell inherits the unphosphorylated form . Subsequently , different sets of proteins are expressed in the two cells , culminating in distinct swarmer and stalked cell morphologies . CtrA and another response regulator , DivK , are at the termini of two phosphotransfer modules: DivJ-PleC-DivK and DivL-CckA-CtrA , see Fig 1A . PleC and CckA are bifunctional histidine-modifying enzymes that may act as either kinases or phosphatases for their respective response regulators , DivK and CtrA [42 , 43] . DivJ is the main kinase for phosphorylating DivK , while the role of DivL is to up-regulate CckA’s kinase activity [44] . DivL is a tyrosine kinase , but its kinase activity is not involved in the up-regulation of CckA [44]; how DivL promotes CckA activity is still unknown . An important step in the pathway is the inhibition of DivL by binding to DivK~P [45] . In a swarmer cell , DivJ is absent , PleC is a phosphatase , and DivK is unphosphorylated . Consequently , DivL is actively up-regulating CckA kinase activity [46] , which in turn maintains CtrA in its phosphorylated form , thereby inhibiting DNA replication in the swarmer cell [41] . The introduction of DivJ during the swarmer-to-stalked transition enables the phosphorylation of DivK , triggering a pathway that culminates in the dephosphorylation of CtrA~P in stalked cells [47–49] . Therefore , at the molecular level , swarmer and stalked cells can be distinguished based on which response regulator—CtrA or DivK—is phosphorylated . The schematic in Fig 1A suggests that DivL and CtrA cannot be concurrently phosphorylated during the cell cycle . Therefore , it was surprising to find that both response regulators are phosphorylated in predivisional cells ( Fig 1B ) [14 , 20 , 50] . While the level of DivK~P remains fairly constant in the stalked and predivisional stages [51] , CtrA~P level changes sharply from lowest in the stalked cell stage to peak activity in the predivisional stage [43 , 52] . How does CtrA~P avoid DivK~P dependent inhibition only in predivisional cells ? Recent experimental observations [45 , 46] have shown that the reason for phosphorylation of CtrA in the predivisional cell is the restoration of DivL activity ( up-regulating CckA kinase ) after it localizes to the new pole . In contrast , cells that are unable to localize DivL to the new pole fail to localize CckA or activate its kinase function . Taken together , these observations suggests that ( a ) phosphorylation of CtrA in predivisional cells is the result of failure of DivK~P to inhibit DivL , and ( b ) the location of DivL determines whether or not it can be inhibited by DivK~P . From the wiring diagram in Fig 1 it is not immediately apparent how the inhibitory interaction between DivK~P and DivL—which is required for the swarmer-to-stalked transition early in the cell cycle—is circumvented in the predivisional stage of the cell cycle simply because DivL relocates to the new pole . One possible explanation for DivL reactivation is that , by localizing to the new pole , DivL comes in proximity to the bifunctional enzyme PleC , which may dephosphorylate and inactivate any incoming DivK~P , thus rendering it unable to inhibit DivL [45] . We shall refer to this informal model as ‘protection by dephosphorylation’ . ( We refer to verbal explanations of experimental findings as “informal models” to distinguish them from the mathematical models we explore in this paper . ) The protection-by-dephosphorylation model assumes that new-pole PleC is functioning to dephosphorylate DivK~P . There are contrasting opinions as to whether PleC in a predivisional cell is a phosphatase [53] or a kinase [54–56] . Clearly , to understand how DivL relocation influences its reactivation requires knowledge of whether PleC is functioning primarily as a phosphatase or a kinase at the new pole of a predivisional cell . To the best of our knowledge , experimental methods to measure the functional status of PleC ( or other bifunctional kinases ) at a specific subcellular location have yet to be developed . Therefore , we have undertaken a mathematical modeling approach to address this question . In earlier work , we described the temporal dynamics of the DivJ-PleC-DivK and DivL-CckA-CtrA network by a formal mathematical model ( in terms of ordinary differential equations ) that was consistent with the principles of thermodynamics , biochemical kinetics , and allostery [57] . In this paper , we extend our temporal model to include spatial aspects of protein localization ( Methods and S1 and S2 Tables ) . Our objective is to use computational analysis to understand how the cell exploits dynamic localization of key enzymes , DivL and PleC , to regulate signal transduction and drive differentiation events . It is important to note that , of the six proteins that are a part of this model , there is limited understanding of how DivJ [48 , 58] , PleC [20 , 53 , 58–60] , DivL [61 , 62] , and CckA [52 , 61 , 62] are localized . Hence , we have refrained from a mechanistic description of the localization of these proteins in our current model . Instead , we enforce a set of rules for the localization of these four proteins . Given these rules , our reaction-diffusion equations determine the spatial location of DivK and CtrA and how the interact with DivJ , PleC , DivL and CckA . In particular , our equations determine whether PleC and CckA are functioning ( in particular times and places ) as kinases or phosphatases and , as a result , whether DivK and CtrA are phosphorylated or not . We first investigated , for physiologically relevant values of diffusion constants , whether PleC functions as a kinase or phosphatase in the predivisional cell . Based on our simulation results , we favor the conclusion that PleC is a kinase in the predivisional cell , in contradiction to the protection-by-dephosphorylation model . As an alternative mechanism , we propose an ‘inhibitor-sequestration’ model , in which PleC kinase binds to and sequesters incoming DivK~P , thus rescuing DivL from inactivation . To test the feasibility of the inhibitor-sequestration idea ( within the framework of our mathematical model ) , we alter the localization profiles of relevant proteins and compare the predictions of our model equations to experimentally observed localization of proteins in wild-type and mutant strains of Caulobacter . Finally , we make experimentally verifiable predictions regarding the distribution and phosphorylation profiles of CtrA and DivK .
Our reaction-diffusion model ( S1 Table ) is based on the mechanism proposed in our earlier paper for regulation of bifunctional histidine kinases [57] . Readers should consult that paper for the rationale behind the reaction kinetics in S1 Table . The temporal model in [57] is expanded to include diffusion of proteins along the long axis of a Caulobacter cell and the localization of specific proteins to the poles of the cell during specific stages of the cell cycle . Because we are interested in protein patterns along the long axis and at the poles of the cell , it is sufficient to formulate the model for one spatial dimension . The governing partial differential equation ( PDE ) for a generic chemical species takes the form: ∂C∂t= Reaction terms+ D· ∂2C∂x2 where C ( x , t ) is the concentration of species C at location x and time t . By discretizing the spatial dimension into n = 100 compartments of equal length h = L/n , where L is the total length of the Caulobacter cell , and using a central difference scheme to approximate the Laplacian operator , we convert each PDE into a set of ordinary differential equations ( ODEs ) . In our notation , Ci is the concentration of species C in compartment i ( 1 ≤ i ≤ 100 ) where dCidt= Reaction terms + D ⋅Ci+1−2Ci+Ci−1h2 Because Caulobacter cells are elongating as a result of new cell wall material being added uniformly along the long axis [16] , we assume that each compartment grows exponentially in time as dhdt= kgrowth ⋅h Since the molecules being investigated cannot diffuse across the cell wall , we implement no-flux boundary conditions at x = 0 and x = L by adding two additional compartments , i = 0 and i = 101 , and insisting at every time step that C0 = C1 and C101 = C100 . The reaction and diffusion rate constants for the wild type and mutant cells are provided in S2 and S3 Tables . A complete understanding of the mechanisms governing localization of DivJ , PleC , DivL and CckA is lacking at this stage , and our model does not attempt to offer one . We enforce the localization of these four kinases based on experimentally observed distributions in wild-type and mutant cells [20 , 48 , 52 , 58 , 62] . We do this by defining rates of binding and unbinding of species C to docking proteins in compartment i as follows: dCibdt= pi⋅ kbinding ⋅Cif−kunbinding ⋅Cib + other terms where pi is an indicator function that takes the value of 1 or 0 , Cib is the concentration of the localized form and Cif is the concentration of the freely diffusing form of a generic protein in compartment i of the cell . The values of the indicator functions for DivJ , PleC , DivL and CckA are provided in S4 Table . The full set of ODEs were simulated in MATLAB using the ode15s solver [63] . The spatiotemporal distribution plots in the figures were generated using the python library Matplotlib [64] . The colors indicate the concentration gradient from zero ( blue ) to the maximum value of protein concentration ( red ) during the cell cycle . A disadvantage of such a plot is that a shallow gradient can be interpreted as significant changes in protein activity and localization . On the other hand , a very steep gradient can result in underestimation of fluctuations in protein activity and localization occurring at the lower range of concentration values . To avoid these problems and to make comparison between wild-type and mutant simulations more convenient , the color bar for each simulation indicates the concentration gradient from zero ( blue ) to maximum wild-type concentration Cwt_max ( red ) . A summary of all our simulation results is provided in S5 Table . The MATLAB code used to simulate the model is available at https://github . com/subkar/PleC_DivL_Spatial
While the dual roles of CckA as a phosphatase and kinase are acknowledged and understood [43–46 , 62] , there are contrasting opinions regarding the function of the histidine kinase/phosphatase , PleC , in predivisional cells . In vitro experiments revealed that the kinase form of PleC is up-regulated by DivK~P [42] . Based on this important finding , an informal model was proposed [42] , suggesting that the DivJ-dependent increase in the level of DivK~P that occurs during the swarmer-to-stalked transition induces PleC to become a kinase . As a kinase , PleC phosphorylates PleD , which in turn initiates a pathway for stalked-cell development [65] . Furthermore , the informal model suggests that , in the predivisional cell , PleC remains a kinase until cytokinesis separates PleC and DivJ into separate compartments [54–56] . The alternate view–that PleC is a phosphatase in predivisional cells [53]–is part of the protection-by-dephosphorylation model developed to explain reactivation of DivL and phosphorylation of CtrA in the predivisional stage [45] . As CtrA~P level falls in the stalked cell , its inhibition of DNA replication is lifted . Within the timeframe of the natural Caulobacter cell cycle , the single chromosome replicates only once to give two chromosomes . However , if cell division is blocked and the cell grows further , only the chromosomal origin that is proximal to the old pole begins a second round of replication to give a third chromosome [66] . This observation suggests a gradient of CtrA~P is established in the predivisional cell , with high concentration at the new pole ( the incipient swarmer pole ) and low concentration at the old pole ( bearing the stalk ) . Mathematical modeling [46] has shown that such a phosphorylation gradient can only be established if CckA functions as a kinase at the new pole and a phosphatase at the old pole . This scenario requires DivL to be active at the new pole , i . e . , unbound to DivK~P . Intriguingly , DivK~P is found to co-localize with DivL and PleC at the new pole . How does DivL remain unbound and active ( up-regulating CckA kinase ) while in close-proximity to DivK~P ? The protection-by-dephosphorylation model suggests that predivisional PleC at the new pole is a phosphatase that is continuously working to dephosphorylate DivK~P and create an inhibition-free ‘protection zone’ for DivL [45] . Of these informal models , the first ( in support of PleC kinase ) is necessary to explain stalk formation , while the second ( in favor of PleC phosphatase ) posits conditions that are to be satisfied for replicative asymmetry . While the first model lays out changes of PleC function throughout the cell cycle , the second only addresses the function of PleC in the predivisional cell . Since PleC-dependent phosphorylation of PleD is required for development of the stalk , it is fair to assume that PleC is a kinase in the stalked cell . Hence , the difference in the two informal models can be narrowed down to the suggested function of PleC at the new pole of the predivisional cell ( Fig 2 ) . We use our spatiotemporal mathematical model to simulate changes of PleC function during the course of the cell cycle , in order to test the two conflicting theories . In swarmer cells , PleC is localized at the old pole and functions as a phosphatase [20 , 58] . During the swarmer-to-stalked transition , PleC becomes a kinase [42 , 54 , 55] before it is cleared from the old pole in mature stalked cells [58] . Later in the cell cycle , PleC localizes at the new pole ( the incipient swarmer pole ) . In our model , the transition from newborn swarmer cell to compartmentalized predivisional cell takes 150 minutes , a doubling time that is consistent with growth in poor medium [14 , 39 , 67 , 68] . Although there is considerable variation in the timing of various developmental transitions [53] , the Caulobacter cell cycle appears to be robust to these variations . Therefore , in this deterministic model of cell cycle transitions , we assign time intervals for each stage of the cell cycle from generic descriptions of cell cycle progression [14] . Hence , in our model , the cell is in the swarmer stage for the first 30 min of the cell cycle and then in the stalked stage for the next 60 min ( t = 30–90 min ) . PleC is localized at the old pole during the swarmer stage ( t = 0–30 min ) . We assume the PleC remains at the old pole for the first 20 min ( t = 30–50 min ) of the stalked stage , because PleC and DivJ are known to be co-localized there for a short time in the developing stalked cell [54 , 55] . Although this 20-minute window may be on over-estimate , we chose it so that one can clearly see the transition of PleC to the kinase form before it is cleared from the stalked cell . ( A shorter residency time does not qualitatively affect our simulation results; see S1 Fig ) . For t = 50–90 min , PleC is delocalized before it relocates to the new pole to define the start of the predivisional stage of the cell cycle ( t = 90–150 min ) . We enforce compartmentalization at t = 120 min by preventing the diffusion of proteins across the mid-cell line . To distinguish between pre- and post-compartmentalized stages , we refer to these two stages as early predivisional ( t = 90–120 min ) and late predivisional ( t = 120–150 min ) . To initiate the swarmer-to-stalked transition , we localize DivJ to the old pole at t = 30 min ( Fig 3A ) . The resultant surge in the level of DivK~P up-regulates the kinase form of PleC . In our simulations , when PleC translocates to the new pole , it continues to function as a kinase . Fluorescence-loss-in-photobleaching ( FLIP ) experiments show that DivK shuttles from pole to pole in about 5 seconds , indicating that the diffusion coefficient of DivK is 20–100 μm2 min-1 [53] . In our simulations , we assume that DDivK = 100 μm2 min-1 for both DivK and DivK~P . In the absence of cytokinesis , DivK diffuses freely throughout the cell , from the new end , occupied by PleC , to the old end , containing DivJ . As a result , DivK~P , phosphorylated by DivJ at the old pole , is able to interact with new-pole PleC and induce it to become a kinase . We investigated whether a smaller value of DDivK would permit PleC to be a phosphatase in predivisional cells . For PleC to be a kinase in an incipient stalked cell and regain phosphatase activity in the predivisional cell , DDivK had to be 1000-fold smaller than our estimated value ( S2A Fig ) . Our simulations support the notion that , provided DDivK is sufficiently small , the phosphatase form of PleC can create a “protection zone” for DivL by dephosphorylating DivK in the vicinity of the new pole ( S2B Fig ) . Consequently , a gradient of CtrA~P can be established to enforce replicative asymmetry ( S2C Fig ) . However , unlike the distribution pattern observed in experiments [20] , we find that DivK~P no longer localizes at the new pole of predivisional cells . This aberrant result , combined with the fact that DivK~P would have to diffuse significantly more slowly than estimated from the experiments in [53] , leads us to conclude that new-pole PleC cannot function primarily as a phosphatase . Hence , the protection-by-dephosphorylation model may not be the correct explanation for DivL reactivation in predivisional cells . If new-pole PleC is not acting as a phosphatase in early predivisional cells , how is DivL protected from DivK~P-dependent inhibition ? Simulation results from our temporal model ( S3 Fig ) show that , in the process of up-regulation of the kinase form of PleC by its allosteric ligand DivK~P , a significant fraction of DivK~P is bound to PleC kinase . In contrast , DivK does not form a complex with PleC phosphatase that is prevalent in swarmer cells ( S3 Fig ) . This simulation result is in agreement with fluorescence-resonance-energy-transfer ( FRET ) microscopy measurements that show interaction between DivK and PleC at the new pole [53] . These results would make sense if new-pole PleC is primarily a kinase , since DivK~P is an allosteric ligand that needs to remain bound to PleC to maintain it in the kinase form . Based on these observations , we suggest that polar localization of DivK~P at the new pole of a predivisional cell may be a result of PleC being in the kinase form . In the predivisional cell , DivK~P is localized at both old and new poles [20] . Apart from PleC kinase , the only other recognized binding partners for DivK~P are DivJ and DivL [44] . DivJ accounts for old-pole localization of DivK~P , but DivJ’s absence from the new pole implies that it is not a binding factor for new-pole DivK~P . DivL is present at the new pole , but we require that DivL should not be bound to DivK~P , because DivL is actively up-regulating CckA kinase at the new pole . By ruling out DivJ and DivL and refraining from invoking an unidentified binding partner , we conclude that PleC kinase is the binding partner for new-pole DivK~P . Further , we hypothesize that PleC kinase outcompetes DivL for DivK~P binding . Instead of functioning as a phosphatase and dephosphorylating DivK~P , we predict that PleC is a kinase that sequesters DivK~P away from DivL . We speculate that DivK~P sequestration ( rather than DivK~P dephosphorylation ) may be the real reason for a “protection zone” for DivL at the new pole , so that DivL regains the ability to up-regulate CckA kinase there , prior to cytokinesis . Using our model of the DivJ-PleC-DivK + DivL-CckA-CtrA network , we studied the plausibility of our inhibitor-sequestration hypothesis . As in the case of PleC , there is limited information on the mechanism behind the dynamic localization of DivL . Hence , we force DivL to spread throughout the cell initially ( from t = 0 to t = 90 min ) , and later to localize at the new pole in the early predivisional stage ( t = 90–120 min ) . Under these circumstances , our simulations show that: ( 1 ) DivL is active ( unbound to DivK~P ) in the swarmer stage ( Fig 3B ) ; ( 2 ) during the swarmer-to-stalked transition , the level of active DivL falls , as it binds with DivK~P; and ( 3 ) in the early predivisional stage ( t = 90–120 min ) , DivL is localized at the new pole in the active form , even though DivK~P is present at the same location . These simulations confirm our hypothesis that , at the new pole of an early predivisional cell , where PleC , DivK and DivL co-localize , PleC kinase sequesters DivK~P , allowing DivL to be reactivated and CckA to be a functional kinase . Why does inhibitor sequestration fail to protect DivL in the stalked cell ( t = 50–70 min ) , when PleC is transiently localized at the old pole as a kinase ? We reason that the spatial separation of PleC kinase and DivL in the stalked cell stage means that most DivL molecules lie outside the protection zone created by PleC-dependent inhibitor sequestration . Moreover , by spreading over the entire length of the cell , DivL can be inhibited by a lower concentration of DivK~P per unit length of the cell . Our simulations demonstrate that Caulobacter cells can program different cell fates by regulating DivL inhibition through spatial reorganization of DivL and PleC . CckA localization is governed by both PopZ [69] and DivL [61] . Although the spatial distribution of CckA shows cell-to-cell variability , the consensus opinion is that CckA protein is spread uniformly throughout the cell in the swarmer stage , followed by old-pole localization in the stalked cell , followed by a bipolar distribution in the early predivisional stage [52 , 70] . According to our model simulations , the spatiotemporal distribution of CckA activity reflects changes in local concentrations of free DivL . In the swarmer cell , when CckA is uniformly distributed , 70% of CckA is in the kinase conformation ( Figs 3C and 4A ) . In the stalked cell ( t = 30–90 min ) , DivL inactivation results in an increase in the phosphatase fraction of CckA , even as total CckA localizes to the old pole . Finally in the early predivisional cell ( t = 90 min ) , a second focus of CckA co-localizes as a kinase with reactivated DivL , while old-pole CckA remains in the phosphatase form ( Fig 3C ) . At the early predivisional stage , less than half ( ~36% ) of total cellular CckA is in the kinase conformation ( Fig 4A ) . Importantly , however , this 36% is localized in the incipient swarmer half ( new pole ) of the early predivisional cell ( Fig 3C ) . It has been suggested that replicative asymmetry is a result of differential phosphorylation and dephosphorylation of CtrA across the length of the predivisional cell [46] . Synthesis and degradation of CtrA has no bearing on the CtrA phosphorylation gradient . For these reasons , we do not account in our model for changes in the total amount of CtrA protein by means of transcriptional [37 , 40 , 71] and proteolytic [72–74] controls . We assume that CtrA is synthesized in the unphosphorylated form at a constant rate , while CtrA and CtrA~P are degraded at a rate proportional to their individual concentrations . Fluorescence-recovery-after-photobleaching ( FRAP ) experiments indicate that CtrA has a diffusion coefficient of 60–600 μm2 min-1 , while modeling studies suggest that , in order to obtain a gradient of CtrA~P , the rate constants for phosphorylation and dephosphorylation must be faster than the inverse diffusive time scale ( kctr_phos = kctr_kin >> 2D/L2 ) [46] . In our model , we assume DCtrA = DCtrAP = 100 μm2 min-1 and kctr_phos = kctr_kin = 600 min-1 . Our simulations show that CtrA~P is dephosphorylated during the swarmer-to-stalked transition . Once CckA localizes as a kinase at the new pole , a gradient of CtrA~P is established across the cell body ( Fig 3C ) . In order to simulate compartmentalized predivisional cells ( t = 120–150 min ) , we use end-point concentrations from the early predivisional stage ( t = 120 min ) as initial conditions in a simulation of a growing cell with a diffusion barrier at mid-cell . Since DivJ and PleC are in separate compartments , the level of new pole PleC kinase in the swarmer compartment begins to decline and DivK becomes unphosphorylated ( Fig 4B ) . The concentrations of the kinase and phosphatase forms of CckA in the swarmer and stalked compartments at t = 150 min are almost identical to their concentrations during the swarmer ( t = 0–30 min ) and stalked stages ( t = 30–90 min ) , respectively ( Fig 4A ) . Finally , CtrA in the swarmer compartment of the late predivisional cell is phosphorylated , while stalked-compartment CtrA is unphosphorylated . While replicative asymmetry does not require cytokinesis , our simulations show that compartmentalization reinforces cell fate asymmetry by deactivating PleC kinase and dephosphorylating DivK~P in the swarmer compartment . The main assumption of our model is that PleC kinase and DivL compete for binding DivK~P . In simulations of a wild-type cell , PleC kinase at the new pole of a predivisional cell outcompetes DivL for binding DivK~P , thus allowing DivL to remain in its active conformation . Overexpression of DivK should undermine this mechanism , since there will be sufficient DivK~P to bind both PleC kinase and DivL . Upon increasing DivK synthesis by four-fold ( ksyn = 0 . 2 min-1 ) , we see an increase in the level of PleC kinase in both the stalked and predivisional cell compared to the wild-type simulation ( Fig 5A ) . Excess DivK~P in the over-expressing cell binds to and inhibits DivL . Consequently , in the predivisional cell , new-pole CckA does not convert to its kinase form , and CtrA does not get re-phosphorylated . In our temporal model [57] , we showed that the phosphatase-to-kinase transition of PleC is thermodynamically more favorable when DivK~P , not unphosphorylated DivK , is the allosteric ligand . However , in vitro experiments demonstrated that DivK need not be phosphorylated to up-regulate PleC kinase [42] . We reasoned that the concentration of DivK in vivo is within a range that requires DivK to be phosphorylated in order to induce the PleC kinase conformation . In this case , the swarmer-to-stalked transition is controlled by a bistable PleC switch that is flipped to the kinase state by the action of DivJ . When the rate of synthesis of DivK was increased 5–10 fold , our temporal model predicted that PleC kinase would be up-regulated even in the absence of DivJ-dependent phosphorylation of DivK . In this scenario , PleD would be phosphorylated throughout the cell cycle , thus committing the cell to obligate stalked cell morphology . In our current spatiotemporal model , a four-fold increase in the rate of DivK synthesis results in the cell being unable to enter the predivisional stage of the life cycle . However , because PleC is a phosphatase early ( t = 0–50 min ) and a kinase later ( t = 50–120 min ) in the simulation , the cell is predicted to retain distinct swarmer and stalked stages ( Fig 5A ) . Increasing the rate of synthesis of DivK by eight-fold ( ksyn = 0 . 4 min-1 ) , we find ( Fig 5B ) that PleC is a kinase even in the absence of DivJ ( t = 0–30 min ) . DivK is phosphorylated and localized to the old pole , while the level of free and active DivL falls to a third of its wild-type concentration . Throughout the cell cycle , CckA is a phosphatase and CtrA is unphosphorylated . These simulation result are consistent with the experimentally observed phenotype of a divK overexpressing strain , namely , chromosome accumulation , and significant reduction in the level of CtrA~P and autophosphorylated CckA [44 , 50] . In addition , our model suggests that the cell shows a graded response to divK overexpression . A four-fold increase in DivK sees the cell retain swarmer-to-stalked transition , but lose predivisional cell asymmetry on account of being unable to rephosphorylate CtrA . Further increase ( eight-fold ) restricts the cell to a stalked-only morphology , since PleC becomes a kinase and DivK is phosphorylated independent of DivJ . divL mutant cells that cannot localize DivL at the new pole in the predivisional stage also fail to localize and activate CckA [61] . In contrast , in the swarmer stage , uniformly distributed DivL is actively up-regulating CckA . This lack of a causal relationship between localization and activation across two stages of the cell cycle indicates that new-pole localization alone cannot account for DivL activity . Rather , it appears that the act of localizing to the new pole protects DivL from inhibition by DivK~P in the predivisional cell . This may seem counter-intuitive at first , since DivL is positioned in close proximity to its inhibitor DivK~P in the early predivisional cell . We contend that the spatial segregation of PleC and DivL in stalked cells enables DivK~P to bind and inhibit DivL . In contrast , DivL is protected in predivisional cells by co-localized PleC kinase , which sequesters the DivL-inhibitor , DivK~P . To test this hypothesis , we simulate a scenario where DivL and CckA are uniformly distributed ( delocalized ) throughout the cell cycle . Consistent with our reasoning , the simulations show ( Fig 6A ) that free and active DivL is present only in the swarmer stage when no DivK~P is present . Despite the presence of PleC kinase at the new pole during the predivisional stage , DivK~P is able to inhibit DivL as it is diffusely spread over the entire cell . Consequently , new-pole CckA is a phosphatase , and a CtrA~P gradient is not established . Based on this simulation result , we propose that new-pole localization of DivL is not solely responsible for its activation . Instead , it is the co-localization of DivL and PleC kinase that enables DivL to up-regulate CckA kinase . The situation of delocalized DivL and CckA has been encountered in experiments where DNA replication is blocked , and , as in our simulations , these experiments reveal that CckA retains phosphatase activity and CtrA remains unphosphorylated [62] . At the other extreme , we simulated a mutant in which DivL is always localized at one of the poles ( Fig 6B ) . In this mislocalization mutant , DivL is initially present at the old pole of swarmer and stalked cells ( t = 0–90 min ) , and later relocates to the new pole in predivisional cells ( t = 90–120 min ) . In comparison to the wild-type case , the level of free DivL is higher for a further 20 min , from t = 0 to t = 50 min , which is also the duration of PleC localization at the old pole . While the PleC phosphatase-to-kinase transition occurs as usual ( t = 30 min ) , DivL inactivation and the consequent dephosphorylation of CtrA is delayed by a period consistent with the co-localization of PleC kinase and DivL . Once PleC is delocalized ( t = 50 min ) , DivL activity drops , CckA reverts back to a phosphatase , and CtrA is dephosphorylated . Essentially , the novel mutant is characterized by a delay in G1-to-S transition , which otherwise occurs concurrent to the swarmer-to-stalked transition . To further emphasize this point , we simulated the case where PleC and DivL are retained at the old pole throughout the stalked stage , before being redistributed to the new pole in the predivisional cell ( Fig 6C ) . PleC becomes a kinase in the stalked and predivisional cell . However , since DivL and PleC are always co-localized in this novel mutant , CckA remains a kinase and CtrA is phosphorylated during all stages of the cell cycle . Overall , our simulations suggest that Caulobacter cells exploit DivL and PleC localization to fashion two separate phosphorylation profiles for CtrA in stalked and predivisional stages of the cell cycle . Furthermore , a uniform distribution of DivL is essential for the temporal coupling of the swarmer-to-stalked transition with the G1-to-S transition . If the kinase form of PleC is required for inhibitor sequestration , then it follows that pleC mutants without kinase function would be ill-equipped to resolve the stalked and predivisional stages of the cell cycle . This class of mutants includes ΔpleC , pleCH610A , pleCF778L and divKD90G . The first two mutations ( ΔpleC and pleCH610A ) produce non-motile , pili-less and stalk-less cells [20] . pleCF778L mutants are similar to wild-type cells , except that they produce underdeveloped stalks [42] . On the other hand , divKD90G mutant cells are arrested in G1 and lose morphological asymmetry [53] . We sought to explain the physiology of these mutants from the perspective of our inhibitor-sequestration hypothesis . To simulate ΔpleC mutant cells , we set the rate constant for PleC synthesis to zero ( S3 Table ) . In this case , the cell completes the G1-to-S transition but fails to progress any further ( Fig 7A ) . Devoid of PleC kinase , PleD will not be phosphorylated , resulting in the stalk-less phenotype . In the predivisional cell , the kinase form of PleC is not available to sequester DivK~P . Early in the cell cycle , DivK~P localizes to the old pole by binding to DivJ , and later DivK shows moderate bipolar localization through binding to DivL at the new pole . Consequently , DivL remains inactive at the new pole of predivisional cells , and the cell does not exhibit replicative asymmetry . pleCH610A encodes a mutant protein that is both phosphatase- and kinase-negative ( K-P- ) . For reasons unknown , PleCH610A is not released from the old pole of stalked and predivisional cells [20] . Hence , we enforce bipolar localization of PleC in our simulations of this mutant ( S4 Table ) . While the enzyme is incapable of auto-phosphorylation or phosphotransfer , it is unknown if PleCH610A retains its ability to undergo allosteric modification . We parameterize the pleCH610A mutant such that PleCH610A undergoes allosteric modifications by DivK but is unable to auto-phosphorylate or change the phosphorylation status of DivK ( S3 Table ) . Under these assumptions , our simulations show that PleC transitions from an inactive phosphatase form to an inactive kinase form during the swarmer-to-stalked transition ( Fig 7B ) . Although the inactive kinase form of PleC sequesters DivK~P ( according to our assumptions ) , DivL is nonetheless inactivated . The reason why inhibitor sequestration fails in this case is the inability of PleCH610A to dephosphorylate DivK . As a result , the level of DivK~P is high in these mutant cells , and the excess DivK~P molecules bind to and inactive DivL . The failure to dephosphorylate new pole DivK~P , which is characteristic of ΔpleC and pleCH610A mutants , is also reproduced in our simulations . The two mutant alleles pleCH610A and pleCF778L are similar in all aspects except that PleCF778L retains its phosphatase activity . Hence , in our simulations , new-pole PleCF778L , like PleCH610A , can acquire the kinase conformation and sequester DivK~P . Unlike PleCH610A however , the ability of PleCF778L to dephosphorylate DivK means that the level of DivK~P is low enough to allow competition between DivL and PleC kinase . In this scenario , DivL remains active and up-regulates CckA kinase . Hence , even though PleCF778L kinase is unable to phosphorylate DivK , our simulations show that the inactive kinase form of PleC is present at the new pole , where it fulfills the important role of sequestering DivK~P ( Fig 7C ) . Based on our results , we propose that the kinase form of PleC has two independent and important functions to ensure normal progression through the Caulobacter cell cycle . Firstly , the kinase activity of PleC is required to phosphorylate PleD and initiate stalk development . Secondly , the kinase conformation enables PleC to bind and sequester DivK~P , an effect that is essential for DivL activity and the replication-asymmetry of the predivisional cell . Under conditions of replication inhibition , CckA and DivL fail to localize at the new pole [62] . Because divKD90G mutant cells arrest in G1 [53] , we assume that they fail to localize CckA and DivL to the new pole . Hence , in our simulations of this mutant strain ( S4 Fig ) , we enforce DivL and CckA to be delocalized in the predivisional cell . The molecular defects of the divKD90G allele are the inability of DivKD90G to up-regulate PleC kinase [42] and its reduced efficiency in binding and inhibiting DivL [45] . Consequently , in our simulations , PleC kinase level in the stalked and predivisional cell stages falls to a tenth of its wild-type maximum . The experimentally observed unipolar localization of DivK in divKD90G mutant cells [53] is also reproduced in our simulations . Our results suggest that the lack of a new pole focus of DivKD90G is due in part to its ineffective binding to DivL and also to the state of its second binding partner , PleC , which is in the phosphatase form and thus not strongly bound to unphosphorylated DivKD90G . The unipolar localization of DivK in divKD90G cells further supports the notion that new-pole PleC in wild-type cells is in the kinase form . Even though PleC kinase is unavailable to sequester DivKD90G~P , DivL retains much of its activity , since the DivKD90G mutant protein is an ineffective inhibitor of DivL . The level of active DivL remains high until t = 50 min , and only shows moderate decrease during the period when PleC is delocalized ( t = 50–90 min ) . As a consequence , CtrA remains phosphorylated throughout the cell cycle , which explains the G1-arrest observed in experiments [72] .
In earlier work , we proposed a mechanism for the DivK-dependent allosteric regulation of PleC kinase [57] . Our mathematical model of the proposed mechanism , based on elementary chemical reactions , showed that the transition of PleC activity from phosphatase to kinase might function as a bistable switch flipped by DivJ . We believe bistability of the PleC switch ensures a robust and irreversible transition from swarmer to stalked cell morphology . Based on our simulations of mutant phenotypes such as divKD90G and pleCF778L , we predicted that the PleC kinase form is essential for stalked cell development . While the model itself was focused on understanding temporal dynamics during the window of the swarmer-to-stalked transition , we speculated that PleC at the new pole of predivisional cells is a kinase . Without an accurate spatiotemporal model however , we could not predict the effects that diffusion and differential localization might have on the behavior of the molecular switch or its physiological impact on the development of different stages of the asymmetric division cycle . In this paper we present a spatiotemporal model of the network of coupled signaling pathways , DivJ-PleC-DivK and DivL-CckA-CtrA , which determine the phosphorylation status of CtrA in predivisional cells and hence the replicative asymmetry of the incipient swarmer and stalked cells . Our model extends earlier efforts to model various aspects of cell cycle control in Caulobacter . Spatiotemporal models focused solely on the DivJ-PleC-DivK [75] pathway or the CckA-CtrA [46] pathway have been developed . However , these models simulated only the predivisional stage of a non-growing Caulobacter cell . Hence , the proposed mechanisms investigated by these models could not be validated against the behavior of wild-type and mutant cells at other stages of the cell cycle . Other models have captured various temporal aspects cell cycle regulators [57 , 76 , 77] , without considering spatial localization of the proteins . The reaction-diffusion model described here captures the spatiotemporal dynamics of the DivJ-PleC-DivK + DivL-CckA-CtrA network in a Caulobacter cell that grows from a newborn swarmer cell to the late predivisional stage . We use the model is to investigate our “inhibitor sequestration” hypothesis for generating a CtrA~P gradient in the predivisional cell , and to validate our hypothesis against the phenotypes of wild-type and mutant Caulobacter cells at every stage of the cell division cycle ( S5 Table ) . As with any dynamical modeling approach , trade-offs must be made in terms of the molecular details to be included in / neglected from the model , so that the phenomenon under study is accurately described without the model becoming unwieldy to optimize or constrain with experimental data . All the proteins under study in our model are subject to transcriptional regulation and are part of complex localization mechanisms; aspects that were not included in this study . For instance , the bifunctional histidine kinases ( PleC and CckA ) and their partners ( DivJ and DivL ) change their locations in the Caulobacter cell during succeeding stages of the division cycle by mechanisms that are poorly understood at present . Since an accurate mechanistic model of the localization of these proteins is not essential to answering the key questions proposed in this paper , we have enforced a set of localization-rules for these proteins and used our model to predict the consequences of these experimentally observed localization patterns on the phosphorylation states of the response regulators , DivK and CtrA . Other important aspects that have been excluded from the present model are the transcriptional regulation of CtrA [40] and the ClpXP proteolytic machinery that degrades CtrA during the G1-to-S transition [78–80] . Since we were only concerned with changes in the phosphorylation status of CtrA as a measure of DivL and CckA activity , we chose to ignore the regulated synthesis and degradation of CtrA in our model . In any case , the cell cycle proceeds normally in mutant strains containing non-degradable but phosphorylable forms of CtrA [81] , indicating that ClpXP-dependent degradation is non-essential as long as CtrA activity is modulated via phosphorylation . The level of total CtrA remains roughly constant during the replication-division cycle of these normally behaving mutant strains [81] , an observation that we approximate by using basal rates for CtrA synthesis and degradation . Hence , we believe that our simulation results and the qualitative conclusions we draw would remain the same were we to include regulated CtrA synthesis and degradation . We propose that PleC performs three distinct functions that are crucial to proper progression through the Caulobacter cell cycle . In the swarmer cell , PleC is a phosphatase that maintains DivK in its dephosphorylated state . In stalked cells , PleC is a kinase that phosphorylates PleD . Interestingly , the kinase conformation enables PleC to perform a third function—that of binding to DivK~P , which permits reactivation of DivL , activation of CckA at the new pole , and phosphorylation of CtrA in the swarmer-half of a predivisional cell . Our model reproduces the expected distribution pattern of CtrA~P in predivisional cells of pleC and divK mutants ( Fig 8A ) . Each of these mutants can be defined in terms of the loss of one or more of the three distinct functions of PleC—namely , the auto-phosphorylation and phospho-transfer activities of the kinase form ( K ) , the catalytic activity of the phosphatase form ( P ) , and the DivK~P binding capability of the kinase form ( B ) . A CtrA~P gradient from the new end ( high ) to the old end ( low ) is observed in wild-type cells ( K+P+B+ ) that have all three functions . Mutants that are defective only in the kinase activity ( K-P+B+ ) can establish a CtrA~P gradient , as in the case of pleCF778L . In this mutant , inhibitor sequestration is still operative , enabling DivL reactivation at the new pole of predivisional cells . In pleCH610A ( K-P-B+ ) mutant cells , the phosphatase activity is impaired as well , resulting in an increased level of DivK~P , which binds both PleC and DivL . Therefore , a CtrA~P gradient cannot be established , despite the retention of DivK~P’s binding function . In ΔpleC mutant cells , inhibitor sequestration is absent and DivK~P is also elevated , resulting in dephosphorylation of CtrA in the stalked and predivisional stages . We conclude that although auto-phosphorylation and phospho-transfer are dispensable , the phosphatase function of PleC and the DivK~P sequestration role of PleC kinase are required for replicative asymmetry . The morphological transition that occurs during stalked-cell development requires PleC to be active as a kinase , whereas the G1-to-S transition requires deactivation of DivL . The two processes are coupled because PleC kinase phosphorylates DivK , and DivK~P binds to and inactivates DivL . Our simulations show that mislocalizing PleC has no bearing on its functional change from phosphatase to kinase . On the other hand , co-localizing DivL with PleC kinase in the stalked cell does not allow DivK~P to inhibit DivL; hence , CtrA~P level stays high and DNA replication is inhibited . This simulation result prompts us to suggest that the G1-to-S transition requires DivL to be uniformly distributed in the cell membrane . We predict that the G1-to-S transition would be uncoupled from the morphological swarmer-to-stalked transition in a novel mutant where DivL always co-localizes with PleC . At cytokinesis , the cell must partition the phosphorylated forms of CtrA and DivK into separate compartments . This can be achieved only if PleC in the swarmer compartment switches back to the phosphatase form and dephosphorylates DivK~P . In earlier work , we demonstrated that the phosphatase-to-kinase transitions are robust to small changes in the level of DivJ [57] . Compartmentalization creates a situation where DivJ is completely absent in the PleC-containing swarmer compartment . Devoid of its signal kinase , PleC shows a decline in kinase level while DivK gets dephosphorylated . Our model is able to reproduce the defect in swarmer progeny development of ΔpleC and pleCH610A mutants ( Fig 8B ) . Given the absence of a functional phosphatase , DivK~P level remains high in the swarmer compartment of the mutant cells . Fig 9 summarizes the localization and functional status of proteins , as proposed in our model . At the molecular level , the stalked cell is distinguished from the swarmer cell by the kinase DivJ , which localizes at the old pole and initiates the swarmer-to-stalked transition . Hence , DivJ can be considered as a cell-fate determinant protein [23] . However , there is no counterpart for DivJ that initiates the transition from stalked to predivisional cell . Instead , the cell recycles the same components—namely , DivJ-PleC-DivK and DivL-CckA-CtrA—to program the predivisional stage of the cell cycle . Based on our simulation results , we make the case that rapid diffusion of DivK~P does not permit PleC to be a phosphatase prior to cytokinesis . The co-localization of DivL and PleC kinase at the new pole of the predivisional cell enables the effective sequestration of DivK~P by PleC kinase , thus allowing both DivK and CtrA to be phosphorylated at the same stage of the cell cycle . Here , we use our mathematical model to offer further insight into how Caulobacter crescentus exploits spatial localization to temporally regulate the activity of DivL , ultimately giving rise to cell fate asymmetry . | The aquatic bacterium , Caulobacter crescentus , divides asymmetrically into a non-motile “stalked” cell that stays at its place of birth , and a motile “swarmer” cell that disperses to a different locale . Prior to cell division , the cell passes through a “predivisional” stage , when it has a stalk at its “old” end and a flagellum at its “new” end . These spatiotemporal changes in morphology are driven , in part , by changes in subcellular localization of signaling proteins . To understand how the cell exploits protein localization to generate distinct cell fates , we formulated a mathematical model of the spatiotemporal dynamics of six regulatory proteins ( DivJ , DivK , PleC , DivL , CckA and CtrA ) during the Caulobacter cell cycle . Contrary to some suggestions , our model predicts that PleC functions as a kinase during the predivisional stage of the cell cycle . Further , we show that spatial separation of DivL and PleC kinase in the stalked stage is required for inactivation of DivL and for initiation of DNA synthesis . Later , co-localization of DivL and PleC kinase at the new pole of the cell restores DivL activity in the swarmer-half of the cell , resulting in the establishment of replicative asymmetry in the predivisional stage of the cell cycle . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Dynamical Localization of DivL and PleC in the Asymmetric Division Cycle of Caulobacter crescentus: A Theoretical Investigation of Alternative Models |
Neuropathic pain has been little studied in leprosy . We assessed the prevalence and clinical characteristics of neuropathic pain and the validity of the Douleur Neuropathique 4 questionnaire as a screening tool for neuropathic pain in patients with treated leprosy . The association of neuropathic pain with psychological morbidity was also evaluated . Adult patients who had completed multi-drug therapy for leprosy were recruited from several Bombay Leprosy Project clinics . Clinical neurological examination , assessment of leprosy affected skin and nerves and pain evaluation were performed for all patients . Patients completed the Douleur Neuropathique 4 and the 12-item General Health Questionnaire to identify neuropathic pain and psychological morbidity . One hundred and one patients were recruited , and 22 ( 21 . 8% ) had neuropathic pain . The main sensory symptoms were numbness ( 86 . 4% ) , tingling ( 68 . 2% ) , hypoesthesia to touch ( 81 . 2% ) and pinprick ( 72 . 7% ) . Neuropathic pain was associated with nerve enlargement and tenderness , painful skin lesions and with psychological morbidity . The Douleur Neuropathique 4 had a sensitivity of 100% and specificity of 92% in diagnosing neuropathic pain . The Douleur Neuropathique 4 is a simple tool for the screening of neuropathic pain in leprosy patients . Psychological morbidity was detected in 15% of the patients and 41% of the patients with neuropathic pain had psychological morbidity .
Leprosy is a chronic granulomatous disease caused by Mycobacterium leprae that principally affects skin and peripheral nerves . [1] Leprosy is still present throughout the tropics and sub-tropics . Worldwide 249 , 007 new cases were registered in 2008 with India registering 134 , 184 . [2] Leprosy affects peripheral nerves causing enlargement , sensory loss and motor weakness , and nerve fibres in the skin causing loss of sensation in affected skin sites . The M . leprae infection is treated with multi-drug therapy ( MDT ) and all patients receive either dual or triple drug therapy for up to 12 months . MDT is highly effective with a relapse rate of 1% . New nerve damage is treated with steroid therapy , but only about 50% of patients will have improvement in nerve function after a course of steroid treatment . [3] Leprosy is also complicated by further episodes of inflammation affecting skin and nerves . These may be Type 1 reactions associated with delayed type hypersensitivity which cause inflammation affecting skin and nerves . Type 2 or erythema nodosum leprosum ( ENL ) reactions are associated with immune complex deposition and systemic inflammation is seen with involvement of skin , nerves , eyes , bones and testes . Inflammation of nerves is prominent in leprosy pathology . [4] Neuropathic pain is defined as “pain arising as a direct consequence of a lesion or disease affecting the nervous system” . [5] This may be due to nerve damage at a peripheral or central level . Neuropathic pain typically persists after the primary cause has resolved . Neuropathic pain is characterised by positive and negative symptoms including pain , hypoesthesia to touch , tingling , electric shocks and pins and needles . The diagnosis of neuropathic pain rests on clinical judgement , a relevant clinical history and clinical neurological examination . [6] The patients should complain of pain that is not generated by a stimulus and it must be in one or more regions related to affected nerves ( anatomic plausibility ) . A diagnosis of probable neuropathic pain can be made if 1 ) clinical examination shows positive or negative sensory signs confined to innervation territory of the lesioned nervous structure or if 2 ) diagnostic tests can confirm lesion or disease explaining the neuropathic pain . If both of these criteria are met , the diagnosis of definite neuropathic pain can be made . [7] Several diagnostic questionnaires have been developed to alert the clinician about the possibility of neuropathic pain , for example the Douleur Neuropathique 4 ( DN4 ) . The DN4 consists of seven items related to sensory descriptors and three related to signs which require a simple clinical examination . [8] It examines the positive and negative sensory symptoms of neuropathic pain , such as evoked pain and hypoesthesia to touch . The DN4 has a sensitivity of 83% and specificity of 90% . [8] Little is known about neuropathic pain in leprosy patients . 26 . 4% of 358 newly presenting leprosy patients from a referral centre in Brazil had neuropathic pain . [9] 29% of 96 Ethiopian patients who had been treated for leprosy more than 10 years earlier had neuropathic pain , which was “severe” in 43% . [10] The clinical presentation of neuropathic pain in leprosy patients can be continuous or intermittent and occur at a single or multiple locations . Studies in India and Brazil found most patients with neuropathic pain had a “glove and stocking” symptom distribution , characteristic of polyneuropathy [9] , [11] , and in the Indian study neuropathic pain was associated with nerve tenderness . [11] Several studies have shown increased prevalence of psychological morbidity in patients with leprosy . In some series , up to 65% of patients have psychological morbidity , depression being the most common . [12] , [13] , [14] Chronic neuropathic pain has also been associated with psychological and quality of life comorbidities , including circadian rhythm disturbance , anxiety and depression . [15] So leprosy and neuropathic pain may have separate and perhaps synergistic contributions to psychological morbidity . Non-psychotic mental disorders are defined in ICD-9 as “disorders in which the symptoms are distressing to the individual and recognized by him or her as being unacceptable” . [16] The General Health Questionnaire-12 ( GHQ-12 ) is used to detect the presence of non-psychotic psychological morbidity . It is a self-reporting questionnaire intended for use in a primary health care setting . [17] It consists of 12 questions , asking patients about their general level of happiness , experience of depressive and anxiety symptoms , and sleep disturbance over the last four weeks . [15] This cross-sectional study was designed to assess the prevalence and characteristics of neuropathic pain in leprosy patients who have completed MDT . From previous studies [11] , [18] , we expected a prevalence of approximately 15% , because the patients were recruited at a referral centre . This would be higher than in the general population [19] , [20] , and lower than in patients with diseases in which neuropathic pain is the main feature , such as diabetes or stroke . [21] , [22] . Neuropathic pain was assesed using clincial examination . We also assesd pain using the DN4 questionnaire . Psychological morbidity was assesed using the GHQ-12 . With this study design we were able to asses the prevalence of neuropathic pain . We also tested the utility of the DN4 as a screeening tool for neuropathic pain in this group . We also predicted that we would find increased psychological morbidity in the patients with neuropathic pain .
Between July and August 2008 , we enrolled 101 patients ( 73 male [72 . 3%]; 28 female [27 . 7%] ) from Bombay Leprosy Project clinics in the state of Maharashtra , India . Every patient ( estimate n = 150 ) over the age of 16 who attended the clinics during the study period and had taken a full course of MDT was invited to participate . We did not collect reasons for non participation in the study as we should have done . This sample size was used because we estimated that we would detect about 20 patients with neuropathic pain . Leprosy was diagnosed when a patient had one of the following: skin lesions typical for leprosy; and/or thickened peripheral nerves; and/or acid fast bacilli on slit skin smears . [23] Patients with leprosy were then classified using the 1998 WHO classification in which patients are classified as paucibacillary ( PB ) if they have up to five skin lesions and as multibacillary ( MB ) if they have five or more skin lesions . [23] The Ridley-Jopling classification was made on the basis of the clinical features and bacterial index . [24] All patients were given a blank body chart ( Figure 1 ) and asked to draw in any areas of pain . The clinician used the same chart to draw patches or skin lesions . A clinical history was then taken . Data was collected on past medical history , diagnosis of leprosy , leprosy antibiotic treatment , leprosy reactions ( Type 1 , ENL , Neuritis ) , past and current treatments for leprosy reactions , including corticosteroids , thalidomide , azathioprine and chloroquine . This was followed by a clinical evaluation and the patients completed the DN4 and GHQ-12 questionnaires . The DN4 was translated into Hindi and Marathi , and the GHQ-12 was translated to Hindi and simultaneously translated to Marathi when needed . Nerve enlargement and tenderness was clinically evaluated by palpation of the main peripheral nerves ( great auricular , ulnar , median , lateral popliteal and posterior tibial ) , and defined as present or absent . Motor function was tested in the following nerves by assessing the muscle they supply: facial nerve ( orbicularis oculi ) ulnar nerve ( abductor digiti minimi ) , median nerve ( opponens pollicis ) and common peroneal nerve ( extensor hallucis longus ) . Sensory testing was done using a series of Semmes-Weinstein monofilaments ( MF ) ( 0 . 05 gm / 0 . 2 gm / 2 gm / 4 gm / 10 gm / 300 gm ) to assess punctate static light touch perception in skin areas supplied by the ulnar , median and posterior tibial nerves . [25] Perception was assessed using a Yes/No response . Sensory impairment was defined as present when the MF threshold increased by three or more levels ( filaments ) on any site , or two levels on one site and at least one level on another site , or one level on three or more sites for one nerve [26] , with a lower limit of 0 . 2 gm MF perception on the hands , face and leprosy skin lesions , and for 2 gm on the foot . The WHO disability criteria , which has three levels , 0 , 1 and 2 , was used to document disability . Zero is scored when no disability is present , 1 when loss of sensation in hand or foot is present , and 2 when there is visible damage or disability including deformity , lagophthalmos or hand or foot ulcer . [25] Both DN4 and GHQ-12 questionnaires were completed by the patient , with the clinician asking the questions since many patients were not functionally literate . When completing the DN4 questionnaire , if the patient had any of the symptoms , the area in which s/he had it was recorded beside each symptom . In these cases a clinical assessment of the area was undertaken , evaluating ‘hypoesthesia to touch’ by applying light touch with finger on the painful area and a non-painful area simultaneously , and ‘hypoesthesia to pinprick’ using the 300 g MF . Dynamic mechanical allodynia was evaluated using a brush on the painful area and patients responded Yes/No to the evoking of pain . For the GHQ-12 , the patients were asked how they had felt during the past four weeks . If they had any symptoms , they were asked what cause they attributed to them . Interpretation of the answers is based on a four point response scale scored using a bimodal method ( symptom present: ‘not at all’ = 0 , ‘same as usual’ = 0 , ‘more than usual’ = 1 and ‘much more than usual’ = 1 ) . A score of four or more indicates the presence of a disorder , without diagnosing the exact pathology . [17] Ethical approval was received from the London School of Hygiene and Tropical Medicine Ethics Committee prior to commencing the study . Ethical approval in India was received from the Bombay Leprosy Project Managing Committee . On recruitment , patients were given an explanation of the study , which was translated to Hindi or Marathi when needed before being invited to sign the consent form which was also in Hindi and Marathi . No patient was paid to participate in the study . A database was created using a Microsoft Excel spreadsheet , and analysed using Stata 10 . 1 . A χ2 test was then used comparing all the variables to the outcome variables neuropathic pain , psychological morbidity , and presence of disability . For those variables with a significant p value a Mantel-Haenszel test was performed ( Table 1 ) . Missing data related mainly to aspects of the patients medical history .
Sixty nine patients ( 68 . 3% ) had ongoing skin involvement with patches , nodules , ulcers or infiltration . Twenty four ( 23 . 8% ) had painful skin lesions , of whom 9 ( 37 . 5% ) had neuropathic pain . Forty six patients ( 45 . 5% ) had nerve enlargement and 19 ( 18 . 8% ) had nerve tenderness . Nerve enlargement was present in one nerve only in 14 patients ( 30 . 4% ) and in multiple nerves for 32 ( 69 . 6% ) . Nerve enlargement was found most often in the ulnar nerve ( 65 . 22% ) , then the lateral popliteal ( 43 . 5% ) , posterior tibial ( 39 . 1% ) and greater auricular ( 37 . 0% ) nerves . Nerve tenderness was most commonly elicited in the ulnar nerve ( 68 . 4% ) , followed by the posterior tibial ( 57 . 9% ) , lateral popliteal ( 31 . 6% ) and great auricular ( 15 . 8% ) nerves . Ulnar nerve motor impairment was detected in 35 patients ( 34 . 7% ) . Combining motor and sensory testing , 65 patients had ulnar nerve impairment ( 64 . 4% ) and 42 ( 41 . 6% ) median nerve impairment . Seventy patients ( 69 . 3% ) had plantar sensory impairment . Sensory testing detected impairment in the areas supplied by the lateral and medial plantar ( 69 . 3% ) , ulnar ( 45 . 5% ) , median ( 38 . 6% ) and trigeminal ( 6 . 9% ) nerves . Twenty two patients had neuropathic pain ( 21 . 8% ) by the clinical definition , and Table 2 gives the features of the pain . Neuropathic pain occurred in areas supplied by ulnar ( 6 ) , lateral popliteal ( 6 ) , plantar medial and lateral ( 4 ) , posterior tibial ( 2 ) , sural ( 1 ) , trigeminal ( 1 ) , tibial anterior ( 1 ) and median ( 1 ) nerves respectively . Eleven patients had pain in one limb only and 11 had pain affecting both limbs , either arms or legs . Twenty one patients with neuropathic pain had sensory impairment ( 95 . 5% ) and 11 ( 50 . 0% ) had motor impairment . Nine ( 40 . 9% ) had painful leprosy skin lesions in addition to the areas of neuropathic pain associated with nerve trunks or cutaneous nerves . Patients described sensory symptoms as numbness ( 19 ) , tingling ( 15 ) , burning sensation ( 13 ) , electric shocks ( 11 ) , pins and needles ( 10 ) , painful cold ( 6 ) and three had itching and these symptoms overlapped in individual patients . Eighteen patients had hypoesthesia to touch in the painful area , 16 had hypoesthesia to pinprick , and one had allodynia . Twenty eight ( 27 . 7% ) patients scored 4 or more on the DN4 questionnaire whilst 22 had neuropathic pain by the clinical definition . None of the patients clinically diagnosed with neuropathic pain scored less than 4 on the DN4 . These six patients experienced some of the symptoms listed in the questionnaire occasionally in a certain area , but without pain or discomfort . One patient had pain with no anatomical plausibility . In this group the DN4 used for screening for neuropathic pain had a sensitivity of 78 . 6% and a specificity of 100% . Fifteen ( 14 . 9% ) patients had psychological morbidity , mainly anxiety , mild or incipient depression , which they associated with having leprosy . Forty one percent of the patients with neuropathic pain had psychological morbidity . Some patients who scored less than 4 on the GHQ-12 explained that although they had some symptoms , such as lack of sleep or increased strain , they felt that these were not associated with their leprosy . Forty four patients ( 43 . 6% ) had disability , eight had grade 1 ( 7 . 9% ) , and 36 grade 2 ( 35 . 6% ) . Of the 22 patients with neuropathic pain , 11 ( 50% ) had disability grade 0 , five had grade 1 ( 22 . 7% ) and six grade 2 ( 27 . 3% ) . Neuropathic pain is significantly associated with psychological morbidity ( p = 0 . 0001 ) , nerve enlargement ( p = 0 . 016 ) , nerve tenderness ( p = 0 . 0003 ) , trigeminal nerve impairment ( p = 0 . 0194 ) , and painful skin patches ( p = 0 . 0335 ) ( Table 3 ) . No significant association was found between neuropathic pain and sex , age , type of leprosy ( WHO or Ridley-Jopling ) , presence or type of reaction , type of treatment for leprosy , other treatments ( prednisolone , thalidomide , chloroquine and azathioprine ) or disability . Psychological morbidity was also significantly associated with nerve tenderness ( p = 0 . 024 , OR = 3 . 74 , 95% CI 1 . 09–12 . 77 ) .
The study demonstrates a high prevalence of neuropathic pain in patients who have completed treatment for leprosy and are generally not accounted for in leprosy statistics . The main symptom was numbness ( 86 . 4% ) followed by tingling ( 68 . 2% ) , which differs from the findings from other studies . The DN4 is a useful and simple tool for the screening of neuropathic pain , which makes it appropriate for leprosy field work . Its high specificity in the study is important in its further application outside research environment , considering the chronic implications of neuropathic pain . The relationship of neuropathic pain with psychological morbidity found in this study further increases the magnitude of the problem . It will be important to continue research in this field in order to attend the needs of these patients in terms of prevention , diagnosis and treatment . | Neuropathic pain has only recently been recognised as a complication of leprosy . We assessed 101 treated leprosy patients in Mumbai and found that 22 of them had neuropathic pain . The pain occurred as numbness 86% , tingling 68% , and decreased sensation to light touch 81% . This pain was significantly associated with nerve enlargement and tenderness , which suggests that ongoing inflammation may be important in causation . A questionnaire-based screening tool ( Douleur Neuropathique 4 ) for detecting neuropathic pain has been developed and validated in other patients groups . We are the first group to have used the DN4 as a screening tool in leprosy patients and found that it worked well , detecting 78% of patients with no inappropriate diagnoses . There is also an increasing recognition that leprosy is associated with psychological morbidity . Neuropathic pain is also associated with psychological morbidity . We also assessed psychological morbidity using the 12-item General Health Questionnaire and found that neuropathic pain and psychological morbidity are associated with leprosy patients . Leprosy patients with neuropathic pain thus have a double hit for psychological morbidity . Clinicians looking after leprosy patients should warn their patients about neuropathic pain and assess their patients for psychological morbidity . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"dermatology",
"public",
"health",
"and",
"epidemiology",
"mental",
"health",
"science",
"policy",
"evidence-based",
"healthcare",
"neurological",
"disorders",
"anesthesiology",
"and",
"pain",
"management",
"infectious",
"diseases/bacterial",
"infections",
"neuroscience"
] | 2011 | Neuropathic Pain and Psychological Morbidity in Patients with Treated Leprosy: A Cross-Sectional Prevalence Study in Mumbai |
TNFα is a pleiotropic pro-inflammatory cytokine with a key role in the activation of the immune system to fight viral infections . Despite its antiviral role , a few viruses might utilize the host produced TNFα to their benefit . Some recent reports have shown that anti-TNFα therapies could be utilized to treat certain viral infections . However , the underlying mechanisms by which TNFα can favor virus replication have not been identified . Here , a rhabdoviral infection model in zebrafish allowed us to identify the mechanism of action by which Tnfa has a deleterious role for the host to combat certain viral infections . Our results demonstrate that Tnfa signals through its receptor Tnfr2 to enhance viral replication . Mechanistically , Tnfa does not affect viral adhesion and delivery from endosomes to the cytosol . In addition , the host interferon response was also unaffected by Tnfa levels . However , Tnfa blocks the host autophagic response , which is required for viral clearance . This mechanism of action provides new therapeutic targets for the treatment of SVCV-infected fish , and advances our understanding of the previously enigmatic deleterious role of TNFα in certain viral infections .
Tumor necrosis factor alpha ( TNFα ) is one of the main pro-inflammatory cytokines produced in response to a broad type of bacterial , viral and fungal infections [1] . TNFα has a crucial role in activating and orchestrating the immune response in order to protect the host organism from pathogens . TNFα deregulation can promote susceptibility to pathogens by impairing pathogen clearance and , ultimately , promoting maintenance of infection and death . When specifically talking about viral pathogenesis , TNFα has been shown to inhibit the replication of certain viruses such as hepatitis B virus ( HBV ) and the varicella zoster virus ( VZV ) [2] . In addition , anti-TNF therapies to treat autoimmune diseases exacerbate the infection produced by virus such as herpes simplex virus ( HSV ) , Epstein-Barr virus ( EBV ) , cytomegalovirus ( CMV ) and human papillomavirus ( HPV ) [3] . It is not surprising that due to the key role of TNFα in the host protection to viral infections , some viruses have developed different ways to interfere with the TNFα pathway [4] . In contrast , it seems that a few viruses might utilize the host produced TNFα to their benefit . Interestingly , human immunodeficiency virus 1 ( HIV-1 ) infection induces TNFα expression . These increased TNFα levels in serum correlates to increased viral replication [5] . In accordance to that , TNFα inhibitors are able to impair HIV-1 replication [6] , and anti-TNFα treatments have been proposed to combat HIV-1 infection in combination with other therapies [7] [5] . Similarly , neutralization of TNFα decreases virus production in CMV-infected macrophages [8] . The ability of TNFα to favor virus replication has also been demonstrated for non-mammalian viruses , such as the spring viremia of carp virus [9] , a fish rhabdovirus infecting cyprinids [10 , 11] . Moreover , intraperitoneally SVCV-infected adult fish , in which recombinant TNFα was administrated simultaneously , has shown a higher mortality rate than fish injected with the virus alone . The mechanism explaining how TNFα facilitates viral infection and its deleterious effects in the host has not yet been proposed . Since zebrafish is a cyprinid susceptible to SVCV infection , and TNFα can exacerbate SVCV infection , we chose this amenable infection model to investigate how a virus might utilize host produced TNFα to their benefit . To that end , we analyzed the role of zebrafish TNFα ( Tnfa ) in i ) the key steps of SVCV pathogenesis: virus adhesion , fusion , and replication; and ii ) in the antiviral host response , such as interferon production and autophagy . The results showed that Tnfa signaling through its receptor Tnfr2 inhibits autophagy , leading to impaired viral clearance in SVCV-infected cells . This mechanism of action provides new therapeutic targets for the treatment of SVCV-infected fish , and advances our understanding of the previously enigmatic deleterious role of TNFα in certain viral infections .
As in most infections , Tnfa is up-regulated in response to SVCV infection [12] . Unexpectedly , this up-regulation rather than help to control the infection , has a deleterious role in adult zebrafish [9] . To further study this phenomena , we first investigated whether or not Tnfa was also able to enhance SVCV replication both in vivo and in vitro . For that , we pre-incubated the zebrafish embryonic fibroblast cell line , ZF4 , which expresses both Tnfrs [9] , with zebrafish recombinant Tnfa or interferon 1 ( Ifn1 , also known as Ifnphi1 ) for 4 hours and , subsequently , the treated cells were infected with SVCV . At 24 hours post-infection ( hpi ) , viral replication , measured as the presence of transcript of the nucleoprotein that forms the SVCV capside ( N protein ) , was evaluated by RT-qPCR ( Fig 1A ) . N protein transcripts significantly increased in Tnfa-treated cells and significantly decreased in Ifn1-treated cells ( Fig 1B ) , suggesting that Tnfa enhances and Ifn1 decreases viral replication in vitro . Both insufficient and excess Tnfa have been shown to promote susceptibility to mycobacterial infection [13] . We then asked whether endogenous rather than exogenous Tnfa was beneficial or detrimental to the host during SVCV infection . The percentage of animals that survived at 7 days post-infection ( dpi ) was significantly higher in Tnfa-depleted larvae when compared to controls ( Tnfa expressing larvae ) ( 55% versus 30% , respectively ) ( Fig 1C and 1D and S1A and S1B Fig ) . The survival percentage of control and Tnfa-depleted uninfected larvae was 100% in both cases . In accordance to these results , qPCR analysis of embryos harvested at 48hpi showed that the highest levels of viral replication ( measured as the amount of SVCV N protein mRNA in infected animal tissues ) ( Fig 1E ) , and virus particles ( measured as the amount of negative sense RNA encoding SVCV G glycoprotein in infected animal tissues ) ( Fig 1F ) , were found in control larvae . These results were further confirmed in larva forced to express Tnfa RNA , which showed drastic increased susceptibility to SVCV ( Fig 1G and S1C and S1E Fig ) . Together , these results indicate that Tnfa enhances SVCV replication and pathogenesis in vivo . TNFα exerts its activity through the binding and activation of two receptors , TNFR1 and TNFR2 ( Tumor necrosis factor receptor 1 and 2 , respectively ) [14] . Tnf receptors are expressed early during zebrafish development [15] , and they both have important roles for the clearance of viral infections [16] . To further dissect the contribution of Tnfa signaling in SVCV pathogenesis , we performed loss-of-function experiments for both Tnfa receptors using specific antisense morpholinos ( MOs ) [15] in SVCV-infected embryos ( Fig 2A and S1F–S1I Fig ) . Tnfr2-depleted larvae were distinctly more resistant to SVCV infection compared to their control siblings ( 60% versus 30% , respectively ) ( Fig 2B ) , while Tnfr1-depleted larvae showed a slightly , but statistically significant , reduced survival compare to their control siblings ( Fig 2B ) . This result was supported by increased , or decreased , SVCV replication in Tnfr1- and Tnfr2-depleted larvae , respectively ( Fig 2C ) . Accordingly , the presence of viral genomes was also higher in Tnfr1-depleted larvae and lower in Tnfr2-depleted larvae at 48 hpi ( Fig 2D ) . In addition , larva forced to express a RNA encoding a dominant negative ( DN ) form of Tnfr2 , which is lacking the entire intracellular signaling domain and extinguishes Tnfr2 signaling by trimerization with endogenous Tnfr2 [15] , showed increased resistance to SVCV ( Fig 2E and S1D and S1E Fig ) . Overall , these results suggest that Tnfa facilitates SVCV replication through Tnfr2 signaling . Interferon is one of the most powerful antiviral cytokine [20] . It has been shown that interferons can act synergistically with TNFα to suppress virus replication [21 , 22] . However , TNFα and interferon can have antagonistic roles in certain cells such as human fibroblast-like synoviocytes [23] . Therefore , we decided to investigate if the enhancing role of TNFα in SVCV replication was the result of impairing the interferon response during SVCV infection . ZF4 cells were pre-treated with Tnfa and/or Ifn1 for 4 hours . Subsequently , these cells were infected with SVCV alone or in combination with Tnfa . After 24 hours , qPCR analysis was performed to detect the expression of antiviral host genes and viral N protein transcript ( Fig 4A ) . The addition of Tnfa before or in combination with SVCV did not alter the transcript levels of the genes encoding major antiviral effectors , such as myxovirus ( influenza ) resistance b ( Mxb ) , radical S-adenosyl methionine domain containing 2 ( Rsad2 ) , Mxc and protein kinase containing Z-DNA binding domains ( Pkz ) compared to untreated cells ( Fig 4B and 4C and S2A and S2B Fig ) . In contrast , Ifn1-treated cells showed drastically increased levels of the transcripts for the same host genes ( Fig 4B and 4C and S2A and S2B Fig ) . Cells were then pre-incubated with 2 different dilutions of Ifn1 ( 1/100 and 1/500 ) , alone or in combination with Tnfa , and subsequent SVCV infection was performed . Both Ifn1 dilutions were able to increase the RNA levels of mxb , though these levels were unaffected by the simultaneous addition of Tnfa , in both uninfected and infected cells ( Fig 4D ) . Altogether , these results indicate that Tnfa does not antagonize the antiviral role of Ifn1 during SVCV infection . To verify that Ifn1 was indeed interfering with SVCV replication in ZF4 cells , SVCV replication was quantitated by RT-qPCR analysis of the N protein transcripts in SVCV-infected ZF4 cells pre-treated with Tnfa and Ifn1 ( Fig 4A and 4E ) . While the N protein mRNA levels were up-regulated in Tnfa-treated cells , they were down-regulated in the Ifn1 , as well as in the Tnfa/Ifn1 combination ( Fig 4E ) . This finding suggests that Ifn1 has a protective role against SVCV replication and that TNFα does not antagonize the antiviral role of Ifn1 during SVCV infection . Since autophagy is an efficient antiviral mechanism in response to many viral infections including SVCV [24 , 25] , we asked if Tnfa could interfere with the autophagy-mediated clearance of SVCV by host infected cells . ZF4 cells were incubated with Tnfa for 4 hours and autophagy levels were assessed by cellular LC3 distribution . Cells were treated with autophagy modulators , such as 3-Methyladenine ( 3MA ) and rapamycin ( Rapa ) to respectively inhibit or enhance autophagy . As expected , autophagy ( red puncta indicating L3C recruitment ) was clearly diminished in 3MA-treated cells and , in contrast , strongly increased in Rapa-treated cells ( both in number and size of the autophagosomes ) ( Fig 5A ) . Interestingly , Tnfa treatment diminished autophagosome formation suggesting that Tnfa inhibits autophagy ( Fig 5A ) . In contrast , cells treated with heat-inactivated Tnfa ( control Tnfa , CTnfa ) did not affect autophagy ( Fig 5A ) . After autophagy induction , the cytosolic soluble form of LC3 ( LC3-I ) is conjugated to phosphatidylethanolamine to form LC3-phosphatidylethanolamine conjugate ( LC3-II ) , which is recruited to autophagosomal membranes . This process is conserved among vertebrates and present in mammals [26] and in fish [25] . Therefore , LC3-II/LC3-I ratio is commonly used to quantify autophagosome formation by western-blot ( WB ) [27] . In order to quantify autophagy formation in Tnfa-treated cells , western-blot for LC3 was performed from lysates of Tnfa-treated ZF4 cells at 4 hours post-treatment ( Fig 5B ) . The LC3-II/LC3-I ratio decreased by 2-fold in Tnfa-treated cells , but was unaltered in heat-inactivated Tnfa ( CTnfa ) , indicating the negative impact of Tnfa in autophagy ( Fig 5B ) . As expected , the LC3-II/LC3-I ratio decreased after 3MA treatment and increased after Rapa addition ( Fig 5B ) . To further verify the Tnfa-mediated down-regulation of autophagy , ZF4 cells were pre-treated with Rapa for 4 hours and , subsequently , Tnfa was added for 4 hours and western-blot for LC3 was performed using the cell lysates ( Fig 5C ) . The addition of Tnfa to Rapa-treated cells led to a 2-fold reduction in the autophagy activity compared to cells treated with Rapa alone ( Fig 5D ) . As expected , incubation of Tnfa alone reduced the autophagy activity by 2-fold compared to untreated cells ( Fig 5D ) . All together , these results indicate that Tnfa reduces autophagy in ZF4 cells . Moreover , these data suggest a role for Tnfa as a potent effector in reverting autophagy after this process has been initiated . We have previously demonstrated that autophagy has a protective role during SVCV and viral hemorrhagic septicemia virus ( VHSV ) infection [25] . To investigate whether Tnfa-mediated reduction of autophagy impairs SVCV clearance , ZF4 cells were pre-incubated with Tnfa prior to SVCV infection . As shown in the diagram of Fig 6A , the virus foci forming units ( ffu ) were first detected by immunofluorescence against N protein alone ( Fig 6B ) , or in combination with LC3 ( Fig 6D ) or P62 ( Fig 6E ) . To evaluate whether these ffu correlated with the infective viral particles , the SVCV present in the supernatant ( viral yield ) was also isolated and titrated by plaque forming units ( PFU ) ( Fig 6A and 6C ) . The ffu number increased in Tnfa-treated cells compared to untreated and CTnfa-treated cells ( Fig 6B ) . However , no differences on the foci size were found between these two treatments ( Fig 6B ) . As expected , 3MA increased the ffu number , while Rapa decreased it ( Fig 6B ) . Supernatant from cells treated with Tnfa contained 2 . 5-times more infective viral particles ( 4 . 5x105 pfu/ml ) than un-treated ( 1 , 8x105 pfu/ml ) , or CTnfa-treated cells ( 1 , 4x105 pfu/ml ) ( Fig 6C ) . As expected , 3MA-treatment also increased the SVCV pfu/ml ( 5 , 5x105 ) , while Rapa significantly decreased it ( 2 , 7x104 pfu/ml ) ( Fig 6C ) . Notably , although viral particles colocalization with LC3 and P62 puncta was hardly observed in control cells , probably reflecting the rapid degradation/loss of immunogenicity of the virus , it was nicely observed in cells treated with Tnfa ( Fig 6D and 6E ) . Taken together , these results demonstrate that Tnfa impairs viral clearance through the inhibition of the autophagy response in infected cells . To analyze the impact of Tnfa on the regulation of the host autophagic response to SVCV infection , we used a GFP-LC3 transgenic line that allows a real-time visualization of autophagy activity [28] . Morpholino-dependent Tnfa depletion resulted in increased basal autophagy in whole larvae , observed at low magnification as an increased fluorescence due to LC3 aggregation ( Fig 7A and 7B ) . As expected , Rapa treatment also increased autophagy ( Fig 7B ) . Moreover , as predicted from the previous in vitro data , SVCV-induced autophagy [25] was highly potentiated by depleting endogenous Tnfa [12] ( Lopez-Munoz et al . , 2010 ) ( Lopez-Munoz et al . , 2010 ) ( Lopez-Munoz et al . , 2010 ) ( Fig 7A and 7B ) . These results were confirmed by western blot analysis of the LC3-II/LC3-I ratio where depletion of Tnfa in infected larvae increased autophagy ( Fig 7C ) . Therefore , Tnfa inhibits autophagosome formation during viral infection in vivo .
Although the administration of anti-TNFα therapies normally aggravates viral infections , there are a few reports suggesting that TNFα inhibition could be beneficial for the treatment of certain viral infections [5] . However , the mechanism by which viruses manipulate the host-produced TNFα for their own benefit had never been determined . Here , we have used the zebrafish as an infection model to examine in vitro and in vivo the mechanisms by which TNFα enhances viral pathogenesis . We utilized the previously established viral infection model of SVCV in zebrafish , in which excess Tnfa had already been reported to increase viral susceptibility [9] , to dissect the possible negative role of TNFα for the host during SVCV infection . Our studies demonstrate that Tnfa enhances SVCV replication through its receptor Tnfr2 . Mechanistically , Tnfa does not alter SVCV binding to the cells , its escape from the endosome to the cytosol , or the Ifn-mediated antiviral response . In contrast , Tnfa inhibits autophagy both in vitro and in vivo , leading to decreased viral clearance and , consequently , to a higher susceptibility to the infection . The increased survival of Tnfa- and Tnfr2-depleted larvae infected with SVCV demonstrates that Tnfa signaling through Tnfr2 has a deleterious effect in the host during SVCV infection . These results are further confirmed by the increased susceptibility of larvae forced to express Tnfa , confirming previous studies using recombinant Tnfa [9] , and by the increased resistance of larvae forced to express a DN form of Tnfr2 . The fact that the percentage of survival of Tnfr1-depleted larvae is slightly reduced compared to control larvae suggests that Tnfa signaling through Tnfr1 might have some protective role against SVCV infection . Furthermore , the observation that Tnfr2 depletion leads to a much higher larval survival than Tnfa depletion ( 70% versus 55% , respectively ) , further supports dual roles for Tnfa during viral infection , being protective signaling through Tnfr1 and detrimental signaling through Tnfr2 . However , the overall effect of Tnfa during viral infection is predominately harmful for the host . Therefore , we need to be aware that the manipulation of each Tnf receptor leads to different outputs than Tnfa depletion alone . Thus , the use of specific TNF receptor inhibitors , rather than TNFα neutralizing drugs , could prove to be beneficial for the treatment of TNFα-related pathologies [15 , 29] . The potential protective role of signaling through Tnfr1 during SVCV infection still remains unexplored , and further experiments should be performed to investigate this phenomenon . Since this is the first study conducted to address the enhancing role of TNFα in viral pathogenesis , we decided to dissect the essential steps occurring during viral infection in order to identify which of them , if any , were affected by TNFα . These steps include virus adherence to the cell , release from the endosome to the cytosol , replication and new viral particle formation . Our studies demonstrate that Tnfa slightly reduces SVCV binding to the ZF4 cells yet the fact that this modest reduction is also observed when Tnfa is added simultaneously with the SVCV , suggests that Tnfa could be physically interfering with the SVCV rather than deterring its adhesion through the TNFα activation pathway . In addition , we also demonstrate that Tnfa does not affect the SVCV capability to escape from the endosome to the cytosol . Here , we have characterized the possible interference of TNFα in two key antiviral cell mechanisms that restrict virus replication: interferon response [30] and autophagy [31] . Our studies demonstrate that while Tnfa does not alter the interferon response during SVCV infection , it is able to diminish the viral-induced autophagic cell response in vitro and in vivo . Although TNFα has generally been linked to an up-regulation of autophagy [32–35] , it has also been shown that , in certain contexts , TNFα up-regulates mTOR activity through NF-κB , leading to autophagy inhibition [36] . In agreement with this , we provide evidences that Tnfa inhibits autophagy , which leads to increased viral susceptibility . Interesting , TNFα can also have a dual role in viral infection by promoting cell survival or cell death depending on the expression and activation balance of its receptors [37] . Although further studies should be conducted to address whether the TNFα/TNFR2 axis indeed inhibits autophagy through the activation of NF-κB , this is quite plausible since Tnfr2 mainly regulates NF-κB activation in zebrafish larvae [15 , 29] . It would be of interest to inhibit TNFα , or potentially TNFR2 , in SVCV-infected carps for the treatment of this viral disease that produces abundant losses in aquaculture worldwide . In addition , it would be advantageous to manipulate the activation of TNF receptors in those viral infections in which autophagy plays an antiviral role , such as HSV1 , HIV-1 , Sindbis virus , chikungunya virus and West Nile virus [38] . It is important to emphasize that anti-TNFα therapies have already been suggested to be helpful for the treatment of some of these aforementioned viral infections , such as HIV-1 [5] . This therapeutic approach could have important health implications for the treatment of these devastating viral infections since , to date , there are no available treatments for the majority of them .
The experiments performed comply with the Guidelines of the European Union Council ( 86/609/EU ) and the Spanish RD 53/2013 . Experiments and procedures were performed as approved by the Bioethical Committee of the University of Murcia ( approval number #537/2011 ) . The fish cell line ZF4 ( zebrafish embryonic fibroblast ) was purchased from the American Type Culture Collection ( ATCC , #CRL-2050 ) . Cells were maintained at 28°C in a 5% CO2 atmosphere in RPMI-1640 Dutch modified ( Gibco ) cell culture medium containing 10% fetal bovine serum ( FBS ) ( Sigma , F6178 ) , 1 mM pyruvate ( Gibco ) , 2 mM L-glutamine ( Gibco ) , 50 μg/mL gentamicin ( Gibco ) and 2 μg/mL fungizone ( Gibco ) . The SVCV isolate 56/70 ( kindly provided by Dr . P . Fernández-Somalo , Laboratorio Central de Veterinaria , MAGRAMA ) was propagated in ZF4 cells at 22°C as previously described [39] . Supernatants from SVCV-infected cell monolayers were clarified by centrifugation at 4 , 000 × g for 30 min and kept in aliquots at −80°C . Clarified supernatants were used for the experiments . The virus stock was titrated in 96-well plates by limit-dilution ( 50% tissue culture infectious dose ( TCID50 ) /ml ) [40] . The zebrafish ( Danio rerio H . ) AB strain was obtained from the Zebrafish International Resource Center ( ZIRC , https://zebrafish . org/home/guide . php ) . The transgenic line Tg ( CMV:EGFP-map1lc3b ) zf155 ( GFP-LC3 for simplification ) was previously described [28] . Fish were mated , staged , raised , and processed as previously described [41] . In vitro-transcribed RNA of wild type Tnfa and DN Tnfr2 [15] was obtained following manufacturer’s instructions ( mMESSAGE mMACHINE kit , Ambion ) . Morpholinos were diluted in DEPC-treated water at a concentration of 0 . 3 mM ( Standard-mo , Gene Tools ) 0 . 5 mM ( Tnfa-MO , 5’-GCAGGATTTTCACCTTATGGAGCGT-3’ [42] , 0 . 65 mM ( Tnfr1-mo , 5’-ctgcattgtgacttacttatcgcac-3’ [15] , 0 . 3 mM ( Tnfr2-mo , 5’-ggaatctgtgaacacaaagggacaa-3’ [15] . Morpholinos and RNA were mixed in microinjection buffer and microinjected into the yolk sac of one-cell-stage embryos using a microinjector ( Narishige ) ( 0 . 5–1 nl per embryo ) . The same amount of MOs and/or RNA were used in all experimental groups . The efficiency of the MOs was checked by RT-PCR [15 , 42] . Groups of 20–40 wild type or GFP-LC3 transgenic zebrafish larvae of 3 days post fertilization ( dpf ) were challenged at 26°C by bath immersion in 5 ml of filtered egg water ( 60 mg/ml sea salts in distilled water ) containing ~109 TCID50 ( 50% tissue culture infectious dose ) /ml SVCV . Twenty four hours later , the solution containing the larvae was diluted by adding 35 ml of egg water and the larvae were monitored every 24 hours for 8 days for clinical signs of disease and mortality . Fifteen pooled larvae were collected at 48 hpi in 250 μl Trizol ( 15 larvae ) for gene expression studies . For in vivo visualization of autophagy activity , GFP-LC3 transgenic larvae were anesthetized at 48 hpi ( 5 dpf ) with 0 . 16 mg/ml tricaine and mounted in 1% low melting point agarose supplemented with 0 . 16 mg/ml tricaine . Images of the whole larvae were then taken using a Leica MZ16F fluorescence stereo microscope . As positive control , 48 hpf larvae were treated with 1 μM Rapa ( Calbiochem ) for 72 h . The SVCV infectivity in vitro was evaluated by two different methods , RT-qPCR and foci forming unit assays . To detect SVCV by RT-qPCR , ZF4 cells were cultured in 25 cm2 flasks at 80% confluence and treated with 100 ng/ml of zebrafish recombinant Tnfa [9] or Ifn1 ( dilutions 1/100 or 1/500 ) [43] for 4 hours at 28°C and 5% CO2 . Subsequently , the media was removed , cells were washed twice with the cell media containing 2% FBS and infected with SVCV ( multiplicity of infection ( MOI ) of 10−3 ) in the presence or in absence of Tnfa ( 100 ng/ml ) at 22°C for 24 hours . Afterward , the media was removed , RNA extracted , cDNA obtained and qPCR carried out as below indicated . Two different sets of primers ( S1 Table ) were used for SVCV detection: i ) to quantify virus replication a primer pair amplifying the mRNA of N protein of SVCV and ii ) to quantify the amount of viral genomes ( negative sense RNA ) , a primer pair designed to detect the negative sense RNA encoding the gen of SVCV G protein . For foci forming unit assays a previously developed methodology [44] with minor modifications was used . Briefly , ZF4 cells , grown on 96-well plates , were treated with 0 . 1 μg/ml or 1 μg/ml Tnfa , 1 μg/ml heat inactivated ( C Tnfa ) , 1 μM RAP or 10 mM 3MA at 28°C for 4 hours . After incubation , cell culture medium was removed and cells were infected with SVCV ( multiplicity of infection ( MOI ) of 10–2 ) at 22°C . Two hours post-infection , the supernatants from infected cell cultures were removed to eliminate non-bound virus , cell media containing 2% FBS added and plates further incubated for 24h . On the one hand , supernatants form infected cells were harvested and stored at -80°C for viral tritation to determine the virus yield as below indicated . On the other hand , cell were fixed with a solution of 4% formaldehyde ( Sigma , F1635 ) for 15 min , washed with PBS and further fixed with cold methanol ( −20°C ) for 15 min . Fixed cells were stained with a monoclonal antibody to SVCV ( Teknokroma Analítica S . A . monoclonal antibody anti-SVCV ) at 4°C for 24h [45] . After washing with PBS and cell monolayers were incubated with a FITC-labelled rabbit anti-mouse antibody ( SIGMA ) diluted 1/500 and incubation was continued for 30 min . Stained SVCV infected cell foci were then viewed and photographed with an inverted fluorescence microscope ( Nikon Eclipse TE2000-U; Nikon Instruments , Inc . , NY ) provided with a digital camera ( Nikon DS-1QM , Nikon Instruments , Inc . , NY ) . At least , three different assays , each in duplicated , were performed Virus titers in the supernatants of SVCV infected cells in the presence or absence of Tnfa were determined by a plaque forming units assay [39] and expressed as plaque forming units ( PFU ) per ml . Briefly , different dilutions of each supernatant ( from 10−3 to 10−9 ) were added to ZF4 cell monolayers , grown on 24-well plates at 22°C for 2 hours . Then , culture media was removed and the infected cell monolayers covered with a solution of RPMI-1640 cell culture medium with 2% FCS and a 2% aqueous solution of methyl cellulose ( Sigma ) . Cell plates were incubated at 22°C for 5 days and then the media with methyl cellulose was removed . Finally , wells were stained with crystal violet-formalin and plaques counted . To analyze whether or not TNFα impairs the binding of SVCV viral particles to target cells , ZF4 cells grown in 25 cm2 flasks at 80% confluence , were treated with TNFα ( 100 ng/ml ) for 4 hours at 28°C . The media was then removed , cells were washed twice with the cell media containing 2% FBS and infected with SVCV ( 10−3 MOI ) in the presence or absence of TNFα ( 100 ng/ml ) for 30 minutes at 4°C to allow virus binding/attachment but not its endocytosis . Afterward , the media was removed , cells washed twice with cell media containing 2% FBS , RNA extracted and cDNA obtained . By means of qPCR using specific primers ( S1 Table ) the presence of SVCV G protein in the surface of the infected cells ( viral binging ) was evaluated . ZF4 cells , grown on 96 well-plates , infected with SVCV ( MOI of 10−2 ) . Two hours post-infection , the supernatants from infected cell cultures were removed to eliminated un-bound virus and fresh cell culture medium 2% FBS was added . After 24h of incubation at 22°C , the cell culture medium was removed and the SVCV-infected cell monolayer treated with Tnfa ( 100 ng/ml ) for 45 min . The cells were then washed and the membrane fusion triggered by incubating the cells with fusion medium [44] at pH 6 for 30 min at 22°C . After that , cell monolayers were washed and subsequently incubated with fusion medium at pH 7 . 5 for 2 h at room temperature . Finally , cells were fixed with cold methanol ( -20°C ) for 15 min , dried and stained with Giemsa ( 5 mg/ml in PBS ) . Cells were viewed and photographed with an inverted fluorescence microscope ( Nikon ) provided with a digital camera ( Nikon DS-1QM ) . At least , three different assays , each in duplicated , were performed . ZF4 cells were grown on 24-well plates in culture medium supplemented with 10% FBS at 28°C . After 24 h , the different treatments ( 1 μM RAP , 10 mM 3MA , Tnfa ( 100 ng/ml ) or CTnfa ( 100 ng/ml ) were added . After 4 hours of incubation , culture media was removed and cell monolayers were resuspended in 500 μl of PBS with a cocktail of protease inhibitors ( Sigma ) . Cells were then processed to a frozen/thawed cycle 4 times and protein concentration adjusted before loading protein samples onto the gel . Samples were then loaded in Tris—Glycine sodium dodecyl sulfate 17% polyacrylamide gels under reducing conditions and the electrophoresis performed at 100 V for 90 min . The proteins in the gel were then transferred to nitrocellulose membranes ( BioRad ) for 75 min at 100 V in transfer buffer ( 2 . 5 mM Tris , 9 mM glycine , 20% methanol ) . The membranes were then blocked with 8% dry milk , 0 . 05% Tween-20 in PBS . Then , the membranes were incubated with the primary antibody microtubule-associated protein 1 light chain-3 ( LC3 ) -I/LC3-II , a polyclonal antibody anti-LC3A/B ( Cell Signaling Technology ) diluted 1000-fold in PBS containing 5% BSA and 0 . 1% Tween-20 as indicated by the manufacturer . Membranes were then washed 3 times with PBS containing 0 . 05% Tween-20 for 15 min before incubation with GAR-Po in 0 . 5% milk in PBS for 90 min . After the last 3 washes with PBS containing 0 . 05% Tween-20 , the peroxidase activity was detected by using ECL Select chemiluminescence reagents ( Amersham Biosciences , RPN2232 ) and revealed by exposure to X-ray . Protein bands were analyzed by densitometry using the Totalab Software . Analysis of LC3-I and LC3-II bands was performed and calculated as relative to the actin intensity band . Results are presented as the ratio of LC3-II/LC3-I from 3 independent experiments . After 4 hours of incubation with the different treatments ( 1 μM RAP , 10 mM 3MA , Tnfa ( 100 ng/ml ) or CTnfa ) , monoloayers were fixed with a solution of 4% formaldehyde ( Sigma ) for 15 min , washed with PBS and further fixed with cold methanol ( −20°C ) for 15 min . Cell monolayers were then incubated overnight at 4°C with the anti-LC3 or anti-p62 ( Abcam ) antibodies in dilution buffer ( PBS with 0 . 03% Triton X 100 [Sigma] ) and 5% of albumin from bovine serum ( BSA , Sigma ) . To visualize LC3 and p62 , monolayers were washed again and incubated with appropriate secondary antibodies ( in dilution buffer ) for 1 h . To visualize nuclei , cells were stained with 1 μg/mL of 4′-6-Diamidino-2-phenylindole ( DAPI ) for 10 min . Cell monolayers were finally washed for another 3 times . Cells were viewed and photographed with an inverted fluorescence microscope ( Nikon Eclipse TE2000-U; Nikon Instruments , Inc . , NY ) provided with a digital camera ( DS-1QM , Nikon Instruments , Inc . , NY ) . Total mRNA was extracted from pooled larvae or ZF4 cells with TRIzol Reagent ( Life Technologies ) and purified using the PureLink RNA Mini Kit ( Life Technologies ) following the manufacturer’s instructions . Isolated RNAs were stored at −80°C until used . The purified mRNA was treated with DNase I , amplification grade ( 1 unit/μg RNA; Invitrogen ) . SuperScript III RNase H− ReverseTranscriptase ( Invitrogen ) was used to synthesize the first strand of cDNA with an oligo-dT18 primer from 1 μg of total RNA at 50°C for 50 minutes . Real-time PCR was performed with an ABI PRISM 7500 instrument ( Applied Biosystems ) using SYBR Green PCR Core Reagents ( Applied Biosystems ) . Reaction mixtures were incubated for 10 minutes at 95°C , followed by 40 cycles of 15 seconds at 95°C and 1 minute at 60°C , and finally by 15 seconds at 95°C , 1 minute 60°C and 15 seconds at 95°C . For each mRNA , gene expression was normalized to the ribosomal protein S11 ( rps11 ) content in each sample using the Pfaffl method [46] . In all cases , the PCR was performed with triplicate samples and repeated with at least two independent samples . The primers used are shown in S1 Table . Data are shown as mean ± SEM of at least three separate assays for gene expression experiments . Data were analyzed by ANOVA and a Tukey multiple range test to determine differences between groups , while the differences between two samples were analyzed by the Student t test . Log-rank ( Mantel-Cox ) Test was used for the survival curves . | Tumor necrosis factor alpha ( TNFα ) is one of the main pro-inflammatory cytokines produced in response to a broad type of infections [1] . Although TNFα has a crucial role in protecting the host organism from pathogens , its deregulation can promote susceptibility to pathogens by impairing pathogen clearance and , ultimately , promoting maintenance of infection and death . In addition , some viruses might utilize the host produced TNFα to their benefit . Thus , anti-TNFα therapies could be utilized to treat certain viral infections . However , the underlying mechanisms by which TNFα can favor certain virus replication have not been identified . Here , we have used a viral infection model in zebrafish to identify the mechanism of action by which TNFα has a deleterious role for the host to combat certain viral infections . Our results demonstrate that Tnfa does not affect viral ability to infect host cells or to antagonize the main host antiviral pathway , namely the interferon pathway . However , Tnfa impairs viral clearance by blocking the host autophagy response , which is usually used by host cells to degrade unnecessary or dysfunctional cellular components , and that we found to be critical to eliminate intracellular viral particles . This mechanism of action provides new therapeutic targets for the treatment of SVCV-infected fish in aquaculture and probably to other viral infection affecting cattle industry and human . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"death",
"autophagic",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"antiviral",
"immune",
"response",
"viral",
"transmission",
"and",
"infection",
"cell",
"processes",
"immunology",
"microbiology",
"vertebrates",
"animals",
"animal",
"models",
"osteichthyes",
"developmental",
"biology",
"model",
"organisms",
"research",
"and",
"analysis",
"methods",
"fishes",
"proteins",
"gene",
"expression",
"viral",
"replication",
"immune",
"response",
"biochemistry",
"zebrafish",
"cell",
"biology",
"virology",
"interferons",
"genetics",
"metamorphosis",
"biology",
"and",
"life",
"sciences",
"larvae",
"organisms"
] | 2016 | TNFα Impairs Rhabdoviral Clearance by Inhibiting the Host Autophagic Antiviral Response |
Cerebral autoregulation ( CA ) is an important vascular control mechanism responsible for relatively stable cerebral blood flow despite changes of systemic blood pressure ( BP ) . Impaired CA may leave brain tissue unprotected against potentially harmful effects of BP fluctuations . It is generally accepted that CA is less effective or even inactive at frequencies >∼0 . 1 Hz . Without any physiological foundation , this concept is based on studies that quantified the coupling between BP and cerebral blood flow velocity ( BFV ) using transfer function analysis . This traditional analysis assumes stationary oscillations with constant amplitude and period , and may be unreliable or even invalid for analysis of nonstationary BP and BFV signals . In this study we propose a novel computational tool for CA assessment that is based on nonlinear dynamic theory without the assumption of stationary signals . Using this method , we studied BP and BFV recordings collected from 39 patients with chronic ischemic infarctions and 40 age-matched non-stroke subjects during baseline resting conditions . The active CA function in non-stroke subjects was associated with an advanced phase in BFV oscillations compared to BP oscillations at frequencies from ∼0 . 02 to 0 . 38 Hz . The phase shift was reduced in stroke patients even at > = 6 months after stroke , and the reduction was consistent at all tested frequencies and in both stroke and non-stroke hemispheres . These results provide strong evidence that CA may be active in a much wider frequency region than previously believed and that the altered multiscale CA in different vascular territories following stroke may have important clinical implications for post-stroke recovery . Moreover , the stroke effects on multiscale cerebral blood flow regulation could not be detected by transfer function analysis , suggesting that nonlinear approaches without the assumption of stationarity are more sensitive for the assessment of the coupling of nonstationary physiological signals .
Cerebral blood flow ( CBF ) is regulated to provide adequate blood supply to brain . One of the important CBF control mechanisms is cerebral autoregulation ( CA ) [1] . Involving dilation and constriction of cerebral arterioles through myogenic and neurogenic regulation , CA allows to maintain relatively stable CBF despite changes of systemic blood pressure ( BP ) [2]–[7] . Impaired CA leads to more dependence of CBF on BP , leaving brain tissue unprotected against the potentially harmful effects of BP fluctuations , as demonstrated in cerebromicrovascular disease associated with diabetes [8]–[10] , or after ischemic stroke [11]–[14] and brain injury [15]–[17] . A widely accepted concept of CBF regulation is that CA is less effective or even inactive at high frequencies ( >∼0 . 1 Hz ) or at small time scales ( <∼10 seconds ) , thus leading to a passive dependence of CBF on BP at small time scales [18] . Many studies supported this concept [19]–[22] but all were exclusively based on transfer function analysis ( TFA ) that utilizes the Fourier transform to quantify the relationship between BP and cerebral blood flow velocity ( BFV; recorded by transcranial Doppler ) [18] . The TFA assumes stationary signals while physiological signals including BP and BFV are highly nonstationary , displaying complex fluctuations at different time scales with varying amplitude and period even during baseline conditions [23] , [24] . Moreover , TFA assumes a linear relationship between two signals while it is well known that CA leads to a nonlinear pressure-flow interaction [18] . Thus , TFA may render unreliable or even misleading results [25] , [26] . To better understand CBF regulation at different time scales or frequencies , we introduce a new analytical tool termed intrinsic multiscale pressure-flow analysis ( IMPFA ) that has no assumptions of linearity and stationarity . Based on the empirical decomposition analysis [27] , [28] , this method extracts intrinsic oscillations of BFV and BP at multiple time scales and provides a pressure-flow spectrum to quantify dynamic BFV-BP interaction ( see Methods ) . Using this new method , we aimed to establish the multiscale relationship between spontaneous BP fluctuations and BFV fluctuations in old adults , and to determine the long-term effects of ischemic stroke on CBF regulation . Normal CA function is characterized by a faster recovery of the BFV than BP ( i . e . , BFV has advanced phases compared to BP ) and can be estimated by specific phase shifts between BP and BFV oscillations [29] , [30] . We hypothesize that autoregulation is a continuous process operating over a wide range of time scales ( from a few heart beats to about one minute ) , and that the multiscale autoregulation can be quantitatively assessed from the phase shift between baseline BFV and BP fluctuations . We further hypothesize that stroke leads to permanent CA impairment and affects BFV-BP phase relationship at multiple time scales .
Both BFV and BP signals showed complex fluctuations across a wide range of time scales . Figure 1 showed all intrinsic oscillatory components of a BP signal that were obtained in the first step of the proposed IMPFA using the empirical mode decomposition ( see Methods ) [27] , [28] . Figure 2 shows two of these oscillatory components and the corresponding BFV components in the frequency range of ∼0 . 03–0 . 06 Hz ( corresponding to cycle length of ∼16 . 7–30 seconds ) and ∼0 . 1–0 . 2 Hz ( corresponding to cycle length of ∼5–10 seconds ) , respectively . The amplitude and period of oscillations were highly variable among different cycles even in the same component ( see Text S1; Figure S1 ) . The variability may be due to nonstationary properties of the BP and BFV signals as well as noise and artifacts in the recordings ( see Methods ) . In the non-stroke group , the mean BFV-BP phase shift was positive for all tested time scales . Such phase shift would be expected during an active cerebrovascular regulation that leads to a faster recovery in BFV compared to BP fluctuations ( see simulation results below ) . The value of BFV-BP phase shift was generally larger at low frequencies ( or large time scales ) and smaller at high frequencies ( mixed model ANOVA p<0 . 0001 ) ( Figure 3A ) , e . g . , BFV-BP phase shift in the non-stroke group decreased from 54 . 8±4 . 3° ( Mean±SE ) at 0 . 02–0 . 1 Hz to 10 . 7±3 . 2° at 0 . 3–0 . 38 Hz ( Figure 3B ) . Similar frequency dependence was also observed in the stroke group , i . e . , smaller phase shift at higher frequencies ( p<0 . 0001 ) . We note that the phase shift value in non-stroke group remained relatively constant within the frequency range of 0 . 02–0 . 1 Hz while the stroke group showed a maximal phase shift at ∼0 . 09 Hz ( post hoc p<0 . 05 ) . BFV-BP phase shift was consistently smaller in the stroke group compared to the non-stroke group over the entire examined frequency range of 0 . 02–0 . 38 Hz ( p<0 . 0001 ) ( Figure 3A ) . For instance , at 0 . 02–0 . 1 Hz , BFV-BP phase shift in the stroke group was 42 . 5±4 . 1° ( SE ) ( ∼12° smaller than the value in the non-stroke group ) ; at 0 . 3–0 . 38 Hz , BFV-BP phase shift was −5 . 4±3 . 2° ( 16° smaller than the value in non-stroke group ) which was statistically indistinguishable from zero ( Wilcoxon signed-rank test p>0 . 35 ) . The mixed mode ANOVA indicated that there was no significant interaction between effects of group and frequency ( p>0 . 8 ) , i . e . , the group difference was not significantly dependent on frequency . Repeating the same statistical analysis at four separated frequency bins ( 0 . 02–0 . 1 Hz , 0 . 1–0 . 2 Hz , 0 . 2–0 . 3 Hz , 0 . 3–0 . 38 Hz ) confirmed reduced phase shift in stroke patients at all frequencies ( Figure 3B ) . We note that the p value for the group difference was smaller for bins at higher frequencies ( smaller p value indicate more significant group difference ) . This difference in p value might be partially explained by the fact that there were generally more cycles for a better estimate of mean BFV-BP phase shift at higher frequencies . In this case-control study , age , sex , BMI , mean BP , and baseline CO2 were matched between the non-stroke and stroke groups ( Table 1 ) . Thus , the observed group difference in BFV-BP phase shift was independent of these variables . To test whether the reduced BFV-BP phase shift in the stroke group was associated with the changes in cerebrovascular resistance and/or CO2 vasoreactivity ( Table 1 ) , we repeated the statistical analysis with including these variables as covariates in the mixed models . We found that neither variable had significant influences on BFV-BP phase shift ( p>0 . 12 for both variables ) while the effects of group and frequency persisted . Using the similar approach , we found that the observed group difference in BFV-BP phase shift was also independent of mean BP ( p>0 . 7 ) , hypertension or normtension ( p>0 . 2 ) , mean BFV ( p>0 . 3 ) , and mean heart rate ( p>0 . 08 ) . Moreover , the difference between the stroke and non-stroke groups remained ( p<0 . 0001 ) when only normotensive subjects ( 13 stroke and 23 non-stroke subjects ) were included in the analysis . Within the stroke subjects , BFV-BP phase shifts were not different between the stroke and non-stroke sides at all tested frequencies ( p>0 . 7 ) . Indeed phase shifts of the stroke and non-stroke sides were highly correlated ( p<0 . 0001 , r = 0 . 74; Figure 4 ) . These observations remained the same for 13 stroke subjects without hypertension ( non-significant side effect: p>0 . 8; side correlation: p<0 . 0001 , r = 0 . 85 ) . In addition , BFV-BP phase shifts at all tested frequencies were not significantly correlated with the time period after the stroke for both stroke ( p>0 . 3 ) and non-stroke sides ( p>0 . 8 ) . To demonstrate that BFV-BP phase shift is an autoregulation measure , we examined BFV changes in response to oscillatory BP fluctuations utilizing the Aaslid-Tiecks model [29] , [31] . This model was originally used to simulate CBF or BFV recovery in response to sudden BP drop for different degree of cerebral blood flow regulation that is characterized by an autoregulation index ( ARI ) . Ranging from 0 to 9 , ARI indicates how quickly resistance can be adjusted and BFV can be restored in response to BP change , i . e . , ARI = 0 for no autoregulation or no recovery of BFV and ARI = 9 for best autoregulation [29] , [31] . In this study , we used a sinusoidal waveform to simulate BP oscillations ( Figure 5 ) . As expected , the Aaslid-Tiecks model showed that BP oscillations led to BFV oscillations with a sinusoidal waveform at the same frequency of BP ( Figure 5A ) . BFV-BP phase shift was zero when ARI = 0 ( red lines in Figures 5A ) , indicating a passive dependence of BFV on BP . At a fixed frequency of BP oscillations , BFV-BP phase shift increased monotonically when ARI increased from 0 to 9 ( Figure 5B , 5C ) . Such a relationship between BFV-BP phase shift and ARI remained at all tested frequencies from 0 . 02–0 . 38 Hz ( Figure 5 ) . These results are consistent with the well-accepted belief that a large BFV-BP phase shift indicates a better autoregulatory function with quicker resistance adjustment . The simulations also showed that BFV-BP phase shift for any ARI>0 was generally larger at lower frequencies and become smaller at higher frequencies ( Figure 5 ) . The similar frequency dependence was also observed in human data ( Figure 3 ) . Thus , it is important to specify frequency of BP/BFV oscillations when comparing BFV-BP phase shift for assessment of differences or changes in CA . In addition , this frequency-dependent relationship is a key feature indicating that BFV-BP phase shift is not caused by a simple time delay between BP and BFV ( see Text S2; Figure S2 ) .
There is clear evidence that perfusion is only relatively stable and BP fluctuations can induce CBF variations at different frequencies . For instance , Claassen et al . showed that repeated squat-stand maneuvers could induce oscillations at a designed frequency in both BP and BFV in healthy young individuals with normal autoregulation [32]; and we also showed BFV during baseline conditions possessed oscillations that matched to BP oscillations over a wide range of time scales ( 0 . 02–0 . 38 Hz; Figures 1–2 ) . Indeed , these BFV variations in response to BP fluctuations can provide important information about cerebrovascular control system . In this study , we focused on one of the useful biomarkers derived from BFV and BP variations , namely BFV-BP phase shift . This measure can reflect dynamic CBF regulation via adjustment of cerebrovascular resistance ( Figure 5 ) although it can also be affected by other vascular properties such as absolute levels of baseline cerebrovascular resistance and compliance [33] . The physiological understanding of CBF ( or BFV ) regulation at different time scales is still debated and a high-pass filter-cybernetic model is often used to describe the coupling between BP and CBF/BFV [34] . This model predicts that a very slow oscillation in BP ( frequency approaching zero ) will generate an oscillation in BFV with very small amplitude and an advanced phase close to 90° , while a fast oscillation in BP will be completely transmitted to a BFV oscillation with phase lag close to zero . This frequency-dependent feature is well demonstrated in our human data ( Figure 3 ) and our simulations using the Aaslid-Tiecks model ( Figure 5 ) , as well as in previous studies using the TFA [5] , [32] , [34] , [35] . It is important to note that the high-pass model did not suggest a particular cutoff frequency or time scale of CA function . Many CA studies using TFA did consider the BFV-BP interaction over a wide range of frequencies including >0 . 1 Hz [5] , [19] , [21] , [32] , [33] , [36] . However , most of these studies refute that BFV-BP relationship at frequencies >0 . 1 Hz is useful in term of detecting physiological and pathological changes in cerebrovascular system [18] . These TFA-based studies might underlie the widely accepted concept of the active CA region at low frequencies . Using the proposed IMPFA , we showed in this study that the effect of stroke on the BFV-BP interaction persists at high frequencies ( 0 . 1–0 . 38 Hz ) as well as at low frequencies ( ∼0 . 02–0 . 1 Hz ) , leading to a reduction of BFV-BP phase shift at multiple time scales . Additionally , our simulation results based on the Aaslid-Tiecks model confirmed that the association between BFV-BP phase shift and the degree of autoregulation remained at all tested frequencies from 0 . 02–0 . 38 Hz ( Figure 5 ) . These results indicate that the BFV-BP phase shift at high frequencies may also provide insights into CBF regulation . A possible concern on the multiscale flow-pressure coupling is whether small phase shift at high frequencies was simply caused by a time delay between BFV and BP recordings . Since BFV was measured from arteries in brain and BP was measured from finger , the pulse transit time could be different for the two locations . The possible time delay would artificially induce certain BFV-BP phase shift that contains not much physiological information . This is unlikely because the phase shift in both simulation and experimental data became smaller at higher frequencies while a simple time delay would predict a larger phase shift at higher frequencies ( see Text S2; Figure S2 ) . However , a time delay may still contribute to BFV-BP phase shift , especially at frequencies >0 . 3 Hz in control subjects ( Figure S3 ) , leading to an overestimation of the phase shift . Such time-delay effect may potentially contribute to the observed reduction of BFV-BP phase shift in stroke subjects at the high frequencies since stiffer arteries in these populations can lead to a reduced time delay . Furthermore , the frequency-dependent pressure-flow relationship can also be described by a mechanical model without active control of cerebrovascular resistance — Windkessel model [33] , [37] , [38] , in which the cerebral arterial bed is composed of both resistance and compliance elements . In this model , the compliance element can lead to phase advance of flow oscillations , and the magnitude of the resultant phase shift is smaller at higher frequencies . Therefore , further studies are required to formally determine how the mechanical properties of vasculature ( vascular compliance and stiffness ) contribute to pressure-flow phase shift and its reduction in stroke , especially at high frequencies . To ensure a physiologically meaningful and reliable estimate of BFV-BP phase shift , we only studied frequencies up to 0 . 38 Hz . This choice was mainly based on the following two considerations . There were not enough matched BFV-BP oscillations at frequencies between ∼0 . 38 Hz and ∼1 Hz because there were not enough BP oscillatory components over the frequency range . In addition , the potential time-delay effect as discussed above would have stronger effect on the phase shift estimation at >0 . 38 Hz . For instance , the estimated time delay between BP and BFV recordings in control subjects was generally ∼50 ms ( see Text S2 and Figure S3 ) , which would lead to an artificial phase shift of ∼7° at 0 . 38 Hz and ∼18° at 1 Hz . Multiple control mechanisms are involved in the CBF regulation from the cellular level to the neurovascular unit to regional blood flow in main vascular territories . These feedback mechanisms , including myogenic [2] , metabolic [3] , [4] , [39] , [40] , endothelial [41] , and neurogenic regulations [5] , [6] , can operate at multiple time scales from seconds to minutes [42] . For instance , stretch of vascular muscles resulting from changes of intravascular pressure can induce vasodilatation or constriction through myogenic [2] and endothelial [7] responses within a few heartbeats; cholinergic dilation of cerebral blood vessels is also a rapid process that is engaged within neurovascular unit comprising of a neuron , astrocyte and a microvessel [6]; and sympathetic modulation is involved in modulating overall vascular tone at larger time scales ( >20 seconds ) [5] . These mechanisms are integrated within the CA process to accommodate and redistribute CBF locally and regionally in response to changes of neuronal activity , variations of arterial BP , and other physiological stimuli . Future studies are required to test whether these complex mechanisms are capable of affecting BFV-BP coupling across the wide range of frequencies from 0 . 02 to 0 . 38 Hz or other vascular mechanisms ( different from cerebral autoregulation ) are responsible for flow-pressure phase interaction at high frequencies ( >0 . 1 Hz ) . There is accumulating evidence that influences of stroke on the cerebrovascular system evolve in time and space and that they may extend to regions distant form infarct site [43]–[46] . Our recent study supported this notion , showing impaired vascular reactivity in larger areas of brain including vascular territories that are distant from the infracted site and were not damaged by acute infarction [47] . Such long-term and dynamic effects of stroke on cerebrovascular system may explain our finding of similar degradation of CBF regulation in both stroke and non-stroke sides in the stroke patients . Vascular changes associated with other cerebromicrovascular diseases than stroke such as hypertension and diabetes can also lead to impaired autoregulation [8]–[10] , [14] and increase risk for stroke [48] . Thus , it is possible that the observed impaired autoregulation in both stroke and non-stroke sides may be caused by certain vascular complication other than stroke that may precede stroke onset or may be one of the causes of ischemic stroke . To explore such possibility , we have checked many vascular measures in this study including heart rate , blood pressure , mean BFV , cerebrovascular resistance , and CO2 reactivity . Though many of these variables showed significant differences between the stroke and non-stroke groups , none of them could account for the observed group difference in BFV-BP phase shift . For instance , we showed that the group difference persisted when excluding hypertensive subjects , suggesting that the group difference was not caused by the possible effect of hypertension or antihypertensive medications . Note that such results did not exclude the possible effects of antihypertensive drugs on CBF regulation via their influences on the autonomous nervous system , which shall be addressed in future studies . Overall , our results strongly suggest that the global degradation of autoregulation in the stroke group reflects more likely the long-term effect of stroke . However , in order to formally prove or refute this hypothesized mechanism , it is necessary to examine autoregulation before , immediately following and after stroke for the same subjects . Such data are not available in this study and future longitudinal and prospective studies are warranted . We note that the stroke subjects were studies at quite different time after stroke insults ( i . e . , 0 . 5–30 . 9 years ) . This is not an ideal approach considering the fact that the time course of the long-term effect of stroke on cerebral blood flow regulation is not clear . In this study , we did not find significant association between BFV-BP phase shift and time after stroke ( stroke side: p>0 . 3 , non-stroke side: p>0 . 8 ) . These preliminary results suggest that the observed impairment of cerebral blood flow regulation may occur within the 6 months following stroke insults . Since this is a pilot study with a small sample size , future large-scale longitudinal and prospective studies are needed to determine the long-term impact of stroke on cerebral blood flow regulation . Most of previous studies of CA at different time scales ( frequencies ) utilized the TFA [18] . This traditional approach is based on the Fourier transform , assuming that BP and BFV signals are stationary and are composed of superimposed sinusoidal oscillations of constant amplitude and period at a pre-determined frequency range [26] . Recently , it has been realized that physiological signals are intrinsically nonstationary , and that traditional analysis with the stationary assumption may be unreliable or even invalid [25] , [26] . In addition , the TFA assumes a linear relationship between two signals and this assumption is often verified or refuted by checking one TFA-derived measure , namely coherence that ranges from 0 to 1 . The mean TFA coherence for the BFV-BP relationship is smaller than ∼0 . 75 at all frequencies <0 . 5 Hz and even smaller at frequencies <0 . 1 Hz ( close to or less than 0 . 5 ) ( see Text S3; Figure S4 ) . The low coherence at <0 . 1 Hz is believed to reflect cerebral autoregulation that leads to a nonlinear BFV and BP interaction . Such a belief would indicate that TFA-derived BFV-BP phase shift and gain at low frequencies are not valid although these TFA measures have been widely used to assess cerebral autoregulation and its change under physiological and pathological conditions [5] , [19] , [21] , [32] , [33] . Therefore , the interpretation of TFA results should deserve more careful considerations and further theoretical studies shall be conducted to resolve the current contradiction in the interpretations of TFA results . To better handle nonstationary signals , the multimodal pressure-flow analysis ( MMPF ) was introduced and has been successfully applied to identify altered CA in hypertension , diabetes , stroke , and brain injury [8] , [14] , [17] , [49] . Though the MMPF is also based on the empirical mode decomposition , several notable concerns remain for this analysis . First only a single BP mode and its corresponding BFV mode were selected and used to quantify the BFV-BP phase interaction while all other components of BP and BFV are ignored . Thus , the rich multiscale dynamic information in BFV and BP fluctuations is not fully examined in the MMPF . Second , the MMPF estimates BFV-BP phase shift by averaging all oscillatory cycles that could have different frequencies due to the intrinsic nonstationary feature of physiological signals ( Figure 2 and Figure S1 ) . Depending on the cycle frequencies in the selected mode , the estimated mean phase shift can vary because it depends on frequency ( Figure 3 ) . Moreover , artifacts caused by either data acquisition ( missing data or outliers ) or nonstationarity in the data often exist and contaminate each oscillatory component extracted by the EMD , affecting the performance of the MMPF [26] . Thus , the performance of the MMPF is limited because BFV-BP phase shift is frequency dependent ( Figure 3 ) and nonstationarity is an intrinsic property of many physiological signals ( Figure 2 ) . The proposed IMPFA overcomes many limitations of the MMPF and TFA by examining the phase shift of intrinsic cycle-by-cycle BFV-BP oscillations at different time scales . As compared to the MMPF , the IMPFA uses a spectrum to describe frequency-dependent phase interaction between BP and BFV oscillations , thus providing more dynamic information in a more accurate manner . Moreover , the IMPFA is designed to better account for nonstationarities and noise in BP and BFV recordings by filtering out data without matched BFV-BP cycles . We also performed the TFA analysis in this study but the TFA-derived BFV-BP phase shift did not reveal any stroke effect ( p>0 . 07 for both sides ) ( see Text S3; Figure S4 ) . This might be due to influences of nonstationarities and noise in BP and BFV signals which can introduce significant variations and random errors in the TFA results ( see Text S3 ) [26] . In addition to the EMD-based approaches , many sophistical analyses derived from modern concepts and techniques of nonlinear dynamics such as synchronization , wavelet transform , and adaptive filtering have been gradually applied to biological and physiological research [50]–[52] . The current and previous findings strongly suggest that these nonlinear approaches without the assumption of stationarity are more suitable for the assessment of complex physiological interactions including the BP and BFV coupling . In summary , we demonstrated a multiscale regulation in cerebral blood flow during supine resting conditions and showed a long-term effect of stroke that alters the regulation in both hemispheres , thus compromising the ability to counteract perturbations imposed on brain tissue during perfusion pressure fluctuations . With impaired cerebral blood flow regulation , pressure fluctuations are transmitted into cerebral vasculature , exposing brain tissue to potential harmful variations of perfusion and limiting the delivery of blood flow to areas of increased metabolic demands . Thus , our findings highlight the potential importance and benefit of reliable and non-invasive CBF regulation monitoring for the management and daily care of stroke patients . These findings also provide new insights into cerebral blood flow regulation and also raise a new challenge to the modeling of the cerebral autoregulation function . Different control mechanisms such as myogenic , metabolic , and neurogenic controls are involved in the CBF regulation . Future studies are needed to test whether each of these control mechanisms contributes to flow-pressure coupling in a specific time scale range or at all time scales , and to determine which mechanisms affected by stroke are responsible for the altered multiscale cerebral blood flow regulation in these patients .
All data were previously collected in the Syncope and Falls in the Elderly Laboratory at Beth Israel Deaconess Medical Center ( BIDMC ) . All participants provided informed consent and research protocols were approved by the local Institutional Review Board . To test our hypotheses , we studied 79 participants ( 50–85 years old ) with 39 stroke patients and 40 age- and sex-matched non-stroke subjects . Subjects were recruited from the community-living older people via advertisement in local newspapers . All subjects were screened with a medical history , physical examination , standard battery of autonomic tests and routine blood and urine chemistries . All Stroke subjects had chronic large artery hemispheric MCA infarcts documented on MRI or CT during the acute phase . Neurological and functional status of stroke patients was assessed by NIHSS ( mean±SE: 2 . 6±0 . 4 ) and a Modified Rankin Scale ( 1 . 1±0 . 2; <4 for all patients , indicating ability to walk ) . Studies were conducted at 0 . 5–30 . 9 years ( mean = 6 . 1 years ) after stroke when these patient were clinically stable . The 40 non-stroke subjects had no clinical history of stroke , no known carotid stenosis , and no focal deficits on neurological examination . Twenty-six of the stroke patients and 17 of the non-stroke subjects had hypertension that was defined as use of antihypertensive medications or systolic BP >140 mm Hg or diastolic BP>85 mm Hg on 24 hour BP monitoring . Antihypertensive medications were tapered and withdrawn for 3 days prior to the study with home BP monitoring . We excluded subjects with intracranial or subarachnoid hemorrhage on MRI or CT , diabetes mellitus , clinically significant arrhythmias , uncontrolled hypertension ( systolic BP>180 mm Hg and/or diastolic BP>100 mm Hg; or subjects taking ≥3 antihypertensives ) , morbid obesity , contralateral carotid stenosis >50% cases ( for stroke patients ) , or any contraindications to MRI . There were no significant differences in age , body mass index ( BMI ) , mean BP , and CO2 level between the stroke and non-stroke groups ( Table 1 ) . Baseline heart rate was higher and the mean BFV was lower ( for both stroke and non-stroke sides ) in the stroke group compared to the non-stroke group . Within the stroke subjects , mean BFV did not show significant difference between the stroke and non-stroke sides ( p>0 . 6 ) . The experiment was performed between ∼10AM–11:30AM ( at least 2 hours after the last meal ) after subjects stayed overnight in the inpatient room of the Harvard Clinical and Translational Science Center at BIDMC . Before the test , Subjects were resting comfortably in supine position in a quiet environment for at least 20 minutes . Then data were collected for at least 5 minutes during a baseline condition when subjects remained awake and relaxed in the horizontal and supine position . Following the baseline , CO2 vasoreactivity was assessed by performing a 3-minute hyperventilation test and a 3-minute test of rebreathing air in a bag with 5% CO2 . Changes in systemic BP were continuously assessed by measuring beat-to-beat BP waveforms from a finger using a Finapres device ( Ohmeda Monitoring Systems , Englewood CO ) [18] . BFV was simultaneously measured from left and right middle cerebral arteries using transcranial Doppler ultrasonography system ( PMD150 Spencer Technologies , Inc . , WA ) . Doppler probes were positioned to achieve maximal BFV , and stabilized using a 3-D holder . During the data collection , subjects were instructed to minimize movement and to breathe at their normal respiratory frequency . Thus , the diameter of the insonated artery , insonation angle and plasma hematocrit remained relatively constant such that changes in BFV reflect changes in CBF , and the BFV-BP phase relationship can represent the CBF-BP phase interaction in the territory of the insonated vessel [53]–[55] . The electrocardiogram was measured from a modified standard lead II using a Spacelab Monitor ( SpaceLab Medical Inc . , Issaquah , WA ) . End-tidal CO2 values were also recorded from the face mask ( Capnomac Ultima , Ohmeda Monitoring Systems , Englewood , CO ) . Data were continuously recorded at a sampling frequency of 500 Hz and was re-sampled to 50 Hz for data analysis . Note that the time course of the CO2 change in cerebral arteries during breathing tests could vary considerably between subjects , depending on the lung function and blood gas transport mechanisms of individuals . To account for such individual difference , we selected a 30-second window during the hyperventilation test when BFV reached minimum and a 30-second window during the CO2 rebreathing test when BFV reached the maximum . The mean BFV levels in the two 30-seconds windows were used to calculate the CO2 vasoreactivity . To quantify the coupling between CBF ( or BFV ) and systemic BP at different frequencies , we introduced an intrinsic multiscale pressure-flow analysis ( IMPFA ) that is based on theories of nonlinear dynamics without the assumption of linearity and nonstationarity . The IMPFA quantifies dynamic phase relationship between intrinsic BP and BFV oscillations at different frequencies . The analysis includes three steps: ( i ) decompose BP and BFV signals into multiple intrinsic oscillatory modes each within a narrow frequency band ( Figure 1 ) ; ( ii ) identify matched individual BP and BFV cycles from all oscillatory modes ( Figure 2 ) ; and ( iii ) calculate BFV-BP phase shift for each matched BFV-BP cycle , assign individual cycles to different frequency bins based on cycle length , and calculate mean BFV-BP phase shift in each frequency bin . ( i ) The first step was fulfilled using the empirical mode decomposition ( EMD ) ( see details in Text S4 ) [27] , [28] , which allows the decomposition of a complex nonstationary signal into multiple empirical modes with each mode representing a frequency-amplitude modulation in a narrow band ( Figure 1 ) . Unlike the Fourier transform , the EMD is a nonlinear adaptive decomposition processes without assuming the shapes of waveforms . Thus , the resultant BP ( or BFV ) components are true intrinsic oscillatory functions embedded in the complex fluctuations . ( ii ) For each BP mode and its corresponding BFV mode , instantaneous BP and BFV phases at all time points were obtained using the Hilbert transform . Then the BP mode and the BFV mode were divided into individual cycles with each cycle corresponding to a phase increment of 360° , e . g . , from 0° to 360° and from 360° to 720° ( Figure 2 ) . Not all EMD-derived BP and corresponding BFV cycles necessarily reflect true underlying BFV-BP interactions . This can be caused by influences of noise or artifacts in the recordings such as artifacts in BFV signals when subjects were talking or when head movements affected the insonation angle of the TCD probe , missing data during the calibration of Finapres device , and changes in BP signals due to finger movement . Thus , we introduced the following criteria to exclude BP-BFV cycles that were possibly contaminated by noise and artifacts: The target of Criteria 1 and 2 is BFV or BP cycles that contain extra oscillations at higher frequencies while Criterion 3 is aimed for BFV ( or BP ) changes that were unrelated to BP ( BFV ) changes . For all matched BFV-BP cycles we used for the estimation of BFV-BP phase shifts , the difference in frequencies based on BP and BFV was 0 . 00067±0 . 0002 Hz ( SE ) . ( iii ) For each matched BFV-BP cycle , the start and end points of the cycle were based on the determined BP cycle , and phase shift was calculated by averaging all instantaneous phase differences between BFV and BP components in the cycle . All matched BFV-BP cycles were pooled and divided into non-overlapped frequency bins based on cycle length . There were 18 frequency bins with size of 0 . 02 Hz that cover the frequency range from 0 . 02 Hz to 0 . 38 Hz . Mean BFV-BP phase shift in each frequency bin was calculated from all cycles in the bin . Descriptive statistics were used to summarize data . One-way analysis of variance was used for between-group comparisons of age , body mass index , mean heart rate , and mean BP , mean BFV , CO2 , cerebral resistance , and CO2 vasoreactivity . To assess the effects of frequency , group and their interaction on BFV-BP phase while accounting for possibly different or missing data points in certain bin ( s ) for different subjects , a mixed model ANOVA with subject nested in group as a random factor was performed ( JMP-9 . 0 SAS Institute , Cary , NC ) . A similar mixed model was used to assess the potential difference between stroke and non-stroke sides in the stroke patients . In addition , possible influences on BFV-BP phase shift of age , sex , BMI , heart rate , mean BP , CO2 , cerebral resistance , CO2 vasoreactivity were also explored using the mixed model . | Cerebral autoregulation is an important mechanism that regulates blood supply to brain tissue to match metabolic demands during daily activities . Impaired cerebral autoregulation increases the dependence of cerebral blood flow on systemic blood pressure , and is associated with fatal outcomes in patients after brain injury and acute ischemic stroke . Reliable and noninvasive assessment of cerebral autoregulation is still a major challenge in medical diagnostics and clinic studies , mainly because blood pressure and flow are intrinsically nonstationary ( possessing complex oscillations/fluctuations with varying amplitude and frequency ) while traditional methods for assessment of the pressure-flow dependence assume stationary signals . We propose a new computational technique that is based on nonlinear theories without the assumption of stationary signals . This technique allows us to detect the degradation of cerebral autoregulation in patients with mild ischemic stroke even at >6 months after the insult . The degradation was present in both stroke and non-stroke sides and was accompanied by an altered pressure-flow interaction over a wide range of frequencies from 0 . 02–0 . 38 Hz . Our results challenges the traditionally accepted functional region of autoregulation ( <∼0 . 1 Hz ) . The observed long-term influences of stroke highlight the importance of reliable monitoring of cerebral blood flow regulation for the management and daily care of stroke patients . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"integrative",
"physiology",
"anatomy",
"and",
"physiology",
"cardiovascular",
"mathematics",
"circulatory",
"physiology",
"stroke",
"medical",
"physics",
"biology",
"nonlinear",
"dynamics",
"cardiovascular",
"system",
"physics",
"cerebrovascular",
"diseases",
"ischemic",
"stroke",
"neurological",
"disorders",
"neurology",
"physiology"
] | 2012 | A Nonlinear Dynamic Approach Reveals a Long-Term Stroke Effect on Cerebral Blood Flow Regulation at Multiple Time Scales |
In laboratory yeast strains with Sir2 and Fob1 function , wild-type NAD+ salvage is required for calorie restriction ( CR ) to extend replicative lifespan . CR does not significantly alter steady state levels of intracellular NAD+ metabolites . However , levels of Sir2 and Pnc1 , two enzymes that sequentially convert NAD+ to nicotinic acid ( NA ) , are up-regulated during CR . To test whether factors such as NA might be exported by glucose-restricted mother cells to survive later generations , we developed a replicative longevity paradigm in which mother cells are moved after 15 generations on defined media . The experiment reveals that CR mother cells lose the longevity benefit of CR when evacuated from their local environment to fresh CR media . Addition of NA or nicotinamide riboside ( NR ) allows a moved mother to maintain replicative longevity despite the move . Moreover , conditioned medium from CR-treated cells transmits the longevity benefit of CR to moved mother cells . Evidence suggests the existence of a longevity factor that is dialyzable but is neither NA nor NR , and indicates that Sir2 is not required for the longevity factor to be produced or to act . Data indicate that the benefit of glucose-restriction is transmitted from cell to cell in budding yeast , suggesting that glucose restriction may benefit neighboring cells and not only an individual cell .
Calorie restriction ( CR ) extends lifespan and healthspan in several model organisms [1 , 2] . Although inconsistent results have been obtained on lifespan extension in primates , the beneficial effects on healthspan are widely observed [3 , 4] . Thus , dissecting the underlying mechanism of how CR contributes to health is of substantial interest . Two different lifespan paradigms are commonly employed in the yeast Saccharomyces cerevisiae . Replicative lifespan ( RLS ) , a model for understanding aging of dividing cells , is defined as the number of divisions that a yeast mother cell undergoes prior to senescence [5 , 6] . Chronological lifespan , considered to be more relevant to post-mitotic cells , measures the duration of cell viability during stationary phase [7] . In yeast , CR is achieved by reducing glucose concentration from 2% to 0 . 5% or below [8] . The SIR2 gene , which encodes an NAD+-dependent protein lysine deacetylase , and functional NAD+ salvage genes were shown to be required for CR-mediated RLS extension in strain backgrounds containing a wild-type FOB1 gene [8–10] . Moreover , addition of nicotinamide riboside ( NR ) extends yeast lifespan concomitant with increased Sir2-dependent functions and elevated intracellular NAD+[11] . CR-mediated changes in NAD+ metabolites were proposed to exist and underlie the longevity benefit of CR [10 , 12] . However , sensitive and quantitative liquid chromatography mass-spectrometry ( LC-MS ) methods [13 , 14] have been developed to measure the NAD+ metabolome during normal and CR conditions [13] . Though all NAD+ metabolites were increased by addition of nicotinic acid ( NA ) —a condition that extends lifespan—the levels of intracellular NAD+ metabolites did not change upon CR [13] . Careful experiments have established that Sir2 and CR work in parallel pathways [15] and that exogenously added nicotinamide ( Nam ) , initially proposed to function as a Sir2 inhibitor in high glucose [10] , blocks the longevity benefit of CR in yeast strains without Sir2 [16] . To dissect the complexities of CR , we developed methods to quantify the enzymes that participate in the reactions of the NAD+ metabolome and discovered that Sir2 and Pnc1 , which successively convert NAD+ to Nam and NA , are up-regulated during CR [17] . If the effects of glucose restriction on NAD+ metabolism were to promote conversion of NAD+ to NA , one might expect to see a change in levels of intracellular NAD+ metabolites , such as NA . However , existing NAD+ metabolome data are inconsistent with such intracellular changes [13] . Previous experiments indicated that when yeast cells are provided with extracellular Nam , it is imported , converted intracellularly to NA by Pnc1 , and then exported to the culture media in a manner that maintains intracellular NAD+ homeostasis [18] . We therefore questioned whether glucose restriction might result in increased net conversion of NAD+ to NA accompanied by export of NA . If this were the case , then young mother cells might export NA in order to use it later in life . In this study , we aimed to test the hypothesis that glucose-restricted mother cells export metabolites , such as NA , for replication of older cells . By performing a modified RLS assay in which aged mother cells were moved away from their original locations to new locations on the same plate , we discovered that moving mother cells diminished the longevity benefit of glucose-restriction . However , supplementation with NA , NR , or conditioned medium from glucose-restricted cells restores the longevity benefit . This longevity-promoting activity is dialyzable and does not require Sir2 for production or action . Taken together , our data suggest that glucose restriction benefits entire colonies rather than single , glucose-restricted cells . Though a yeast colony is clonal and largely contains cells of identical genotype , this mechanism seems to bear some features of altruism [19–21] .
To test the hypothesis that conditioned medium from glucose-restricted cells is essential for CR-mediated lifespan extension , we modified the RLS paradigm in a manner that permits evaluation of conditioned medium . Yeast mother cells were arrayed on 2% , 0 . 5% , or 0 . 2% glucose-containing yeast extract/peptone/dextrose ( YPD ) plates and were subjected to the typical routine of daughter cell removal in every generation . When most mother cells had produced 15 generations of daughters , half of the aged mother cells were moved to fresh locations on the same plate—to avoid bias in moving mother cells , the mothers to be moved were chosen prior to aging . As shown in Fig . 1 , yeast mother cells cultured on 2% glucose YPD plates were unaffected by migration to new plate locations . In contrast , cells that were restricted to 0 . 5% or 0 . 2% glucose obtained a 20%–30% increase in lifespan with respect to 2% glucose-grown cells only if they remained in their original plate locations . Mother cells lost the longevity benefit of CR if they were moved to new plate locations with the same restricted concentrations of glucose . Replicative longevity is exhibited by yeast strains that are either genetically deleted for glucose and growth control pathways , termed CR-mimetic strains , or by longevity pathways that run in parallel to CR [15] . If moving mother cells diminishes longevity owing to loss of a CR-induced factor , then CR-mimetic strains should show extended lifespan on 2% glucose plates that is diminished by moving mother cells . As shown in S1A Fig . , the control strain BY4742 is unaffected by moving when cultured on 2% glucose YPD plate . However , this strain exhibited a longevity benefit on 0 . 2% glucose media that was lost upon migration . Moreover , as shown in S1B Fig . , deletion of sch9 [22] , tor1 [22] , or hxk2 [15] extended lifespan on 2% glucose in a manner that was diminished by 67% upon moving . As shown in S1C Fig . , consistent with the idea that deletion of fob1 and overexpression of SIR2 produce extensions in RLS that are not related to CR [15] , these strains exhibited 30% and 20% extended lifespan compared to the BY4742 control strain on 2% glucose whether mothers were moved or allowed to remain on their original plate locations . These results indicate that wild-type yeast mother cells on 2% glucose media and strains with CR-unrelated longevity pathways are unaffected by moving while glucose-restricted and CR-mimetic mother cells lose a longevity benefit upon moving . These data suggest that CR mother cells are either hypersensitive to physical movement or rely on a component of conditioned media for longevity . Because mother cells on 2% glucose media did not have their lifespan degraded by migration to fresh plate locations , it seemed unlikely that yeast mother cells are hypersensitive to migration . Instead , we suspected that a longevity-promoting substance was left behind in conditioned media . Since the enzymes to convert NAD+ to NA were up-regulated in CR mother cells [17] and NA supplementation to aged mother cells could function to elevate NAD+ synthesis and Sir2 activity , as precedented by the effects of NR on yeast cell RLS in high glucose [11] , we aimed to test NA as a candidate longevity factor . To test whether NA could complement the loss of RLS in migrated CR mother cells , we performed an RLS experiment at three concentrations of glucose in which one-third of the mother cells at 15 generations were moved to fresh locations on the same plate , and one-third of the mother cells were migrated to fresh locations of media containing 0 . 5 mM NA—as in all similar experiments , the mothers to be moved were preselected in order to avoid experimental bias . The data indicate that mother cells grown in 0 . 5% and 0 . 2% glucose consistently lose the longevity benefit of CR . However , supplementation with NA was sufficient to maintain the longevity benefit of CR despite a move ( Fig . 2 ) . Because NR metabolism was reported to be essential for CR-mediated lifespan extension [23] , we also tested whether NR could provide the same longevity benefit . As shown in S2 Fig . , NR supplementation provided similar rescue activity as NA . These data defeat the idea that CR mother cells are hypersensitive to migration . Moreover , these data indicate that extracellular supplementation of NA or NR is capable of restoring lifespan extension to migrated CR mother cells . Interestingly , addition of NR extends RLS of yeast mother cells grown in 2% glucose when provided at initiation of aging [11] . However , as shown in S2 Fig . and Fig . 2 , when mother cells were aged for 15 generations on 2% glucose , addition of NA or NR to fresh plate locations was incapable of rejuvenating these cells . Since glucose restriction up-regulated expression of enzymes that convert NAD+ to NA [17] and NA addition maintained RLS , NA emerged as an obvious candidate transmissible factor , which might mediate the longevity benefit of CR . To address this question , we prepared 30 ml cultures of yeast at final OD600 nm of 0 . 5 in which the glucose concentrations were 2% , 0 . 5% , and 0 . 2% . The three conditioned media samples plus a 30 ml sample of YP without glucose in which yeast cells were not grown were analyzed by LC-MS as described [14] . As shown in S3 Fig . , the NA concentration in YP media and in each of the three conditioned media samples was between 37 µM and 43 µM . NR concentrations were below the detection limit ( <0 . 04 µM ) in all conditions examined . This experiment does not provide experimental support for NA as a molecule that is exported by glucose-restricted cells . However , it is possible that isolated glucose-restricted cells export NA and that the effect is masked by simultaneous vitamin import and export in a flask-grown culture , which contains a mixture of aged and young mother cells . To test whether a flask-grown culture of glucose-restricted cells contains a factor that allows lifespan to be extended despite migration , we inoculated wild-type yeast in 2% or 0 . 2% glucose-containing YPD media , allowed cells to grow until the glucose was undetectable and then collected and lyophilized the media . To control for the RLS effects of salt or other medium components , nonconditioned YP media without glucose were also lyophilized as a control . Lyophilized media , reconstituted in water , were applied to sections of 2% and 0 . 2% glucose-containing YP plates on which the modified RLS assay was performed . As shown in Fig . 3 , when mother cells began their RLS on 2% glucose media , they had a short lifespan whether they remained in place , were moved to nonconditioned media , or were moved to media from 2% glucose-grown yeast . When mother cells began their RLS on 0 . 2% glucose media and were moved to nonconditioned media , they lost 20% of their lifespan , just as though they had been moved to new plate locations without supplementation . However , when glucose-restricted mother cells were moved to new plate locations supplemented with concentrated media from glucose-restricted cells , lifespan was maintained and , in fact , extended by about 10% . These data indicate that glucose restriction may result in production of a transmissible factor required for extension of lifespan . Since CR-conditioned media could restore and provide an increased longevity benefit to migrated CR mother cells , we wished to distinguish between a longevity factor in CR-conditioned media and the absence of a harmful factor in 2% glucose-conditioned media . We therefore moved yeast mother cells every generation on both 2% and 0 . 2% glucose plates so that yeast mother cells were always maintained in fresh environments . As shown in S4 Fig . , lifespan for yeast mother cells grown on 2% glucose was not changed whether kept in the same locations or always-fresh environments . This result eliminated the possibility of a diffusible harmful factor produced by non-moved mother cells in high glucose . In contrast , the lifespan of yeast mother cells on 0 . 2% glucose dropped significantly compared to non-moved mother cells ( S4 Fig . ) . Because CR mother cells moved at 15 generations retained a 7% benefit ( Fig . 1 ) while CR mother cells moved in every generation obtained no benefit , it appears that prolonged exposure to a longevity factor helps CR mother cells survive old age . We therefore collected conditioned media and nonconditioned YP media without glucose and removed small molecules with 3 . 5 kDa cutoff dialysis cassettes . After dialysis , media samples were lyophilized , suspended in water , and applied to plates . As shown in Fig . 4 , dialyzed conditioned media from 2% glucose or 0 . 2% glucose-grown cells were inactive at altering RLS . Because dialyzed conditioned media from high glucose-grown yeast did not gain a lifespan extending activity upon dialysis , there is no evidence of a harmful substance in the 2% glucose conditioned media . However , loss of RLS-extending activity in low glucose-grown culture media subjected to dialysis suggests a low molecular weight transmissible factor . Sir2 activity was initially proposed to be crucial for CR-mediated lifespan extension in yeast [8] . However , Sir2 is not required for CR-mediated lifespan extension in strains lacking FOB1 [15] . Because the observation that CR-induces up-regulation of Sir2 and Pnc1 [17] led to the idea that young mother cells might export a longevity factor , we tested whether Sir2 is required for this process . Consistent with previous reports [15] , we observed a 20%–30% lifespan extension for the sir2 fob1 double mutant yeast strain ( KK144 ) on 0 . 5% or 0 . 2% glucose compared to lifespan on 2% glucose ( Fig . 5A and 5B ) . However , moving the double mutant yeast strain to fresh locations completely negated the CR benefit ( Fig . 5A and 5B ) . Further , we prepared conditioned media from the sir2 fob1 double mutant yeast strain grown in 2% or 0 . 2% glucose . The conditioned media were then applied to sections of 2% and 0 . 2% glucose-containing YP plates on which the modified RLS assay was performed . As shown in Fig . 5C and 5D , glucose-restricted wild-type mother cells moved to plate locations supplemented with conditioned media from glucose-restricted sir2 fob1 double mutant yeast showed similar lifespans as wild-type cells that were not moved . Because glucose-restricted conditioned media from sir2 mutant yeast have longevity factor activity , Sir2 is dispensable for producing , exporting , and utilizing this activity .
The underlying mechanism of how CR extends lifespan in yeast remains incompletely understood despite mutational data in support of requirements for functional NAD+ salvage [8–10] . Multiple studies have shown that levels of NAD+ and/or levels of the collection of intracellular NAD+ metabolites are not greatly altered by CR in flask-grown cultures of yeast [9 , 10 , 12 , 13 , 24] . However , Sir2 and Pnc1 , which successively convert NAD+ to Nam and NA , are increased in protein expression in CR [17] . Because NA does not increase intracellularly during CR [13] , we considered whether NA or another compound might be exported by a young mother cell and taken up later in order to survive old age . Our data indicate that media from glucose-restricted mother cells are necessary ( Fig . 1 ) and sufficient ( Fig . 3 ) for CR-induced lifespan extension and this requirement is independent of Sir2 to produce or respond to the activity ( Fig . 5 ) . Moreover , we showed that moving to new plate locations only affect CR-mediated lifespan extension while showing no effect on CR-unrelated long-lifespan strains ( S1 Fig . ) . Interestingly , microfluidic technologies have recently been applied to characterize yeast RLS . Whereas , the lifespan extension exhibited by fob1 deletion is readily apparent in microfluidic assays [25 , 26] , the extension of lifespan due to CR disappeared [27] . Though the latter microfluidic assays led to questions about whether CR-induced lifespan extension is real [27] , microfluidic dissection does not , by its nature , maintain the micro-environment of CR mother cells because media are continuously streamed away , much like our experiment in which mother cells are moved in every generation ( S4 Fig . ) . This result provides additional literature support for the existence of a transmissible longevity factor . CR has been shown to extend yeast lifespan by regulating vacuolar acidity [28] . If compartmental pH differs as a function of glucose , it is possible that extracellular pH value is also altered . However , a recent study has shown that buffering the pH of culture media does not extend yeast RLS [29] . This suggests that CR-induced RLS transmission is not simply a function of pH . Like conditioned media from glucose-restricted cells , provision of NA or NR allowed yeast mother cells to be migrated from their original plate locations in a manner that preserved an increase in lifespan ( S1 Fig . and Fig . 2 ) . Export of NA has been previously observed [18] and extension of lifespan with an NAD+ precursor vitamin is precedented by the effect of NR [11] . However , we could not detect a significant change in the levels of these vitamins in flask-grown cells . Because acetylation of telomeric histone H4 in yeast has been shown to correlate with aging [30] , it is possible that the transmissible activity directly or indirectly inhibits this process . This could account for why NA and NR have a mimetic activity though the endogenous mechanism would not require Sir2 . Cell to cell transmission of the beneficial effect of CR was not anticipated . Though the cells in a colony are clonal , the activity suggests characteristics that could be termed altruistic , especially if there are costs to transmit the activity . Alternatively , the activity may simply spill over from cell to cell and produce a community benefit without substantial cellular costs . In a cell-autonomous context , the ability of damaged yeast cells to undergo programmed cell death has been termed altruistic because it produces surviving cells with reduced damage [21 , 31] . Chemical characterization of the transmissible mediator of CR-promoted lifespan extension is now a major focus area .
Yeast strains used in this study are listed in Table 1 . Medium used for RLS analysis was YP ( 2% bacto peptone , 1% yeast extract ) supplemented with filter-sterilized glucose at final concentrations of 2% , 0 . 5% , or 0 . 2% . Single colonies of wild-type or sir2 fob1 yeast strains were inoculated in 5 ml 2% glucose-containing YPD media and allowed to grow until OD600 nm reached 0 . 5 . Cells were then inoculated into 50 ml 2% or 0 . 2% glucose-containing YPD media at an initial OD600 nm of 0 . 01 and were grown until glucose was undetectable by Glucose Colorimetric Assay kit ( Cayman Chemical ) . After centrifugation , culture supernatants were transferred to new tubes . Some samples were dialyzed twice with 3 , 500 MWCO Slide-A-Lyzer Dialysis cassettes ( Thermo Scientific ) against 1 liter non-conditioned YP media without glucose at 4°C . All conditioned media were lyophilized and resuspended in water at 1/10 volume of the original media samples . Note that when conditioned medium is provided to yeast mother cells in a RLS analysis , conditioned medium is always derived from a culture grown at the same concentration of glucose . Experiments were carried out as described [5] with some modification ( S5 Fig . ) . To blind the experiments , the plates and/or supplementations were prepared by other laboratory members and coded . In brief , 60 cells were arrayed on one part of a YP plate with 2% , 0 . 5% , or 0 . 2% glucose . Prior to any cell division , 30 mother cells were assigned the group that would be migrated when these cells reach 15 generations ( 20 generations for BY4742 background cells ) . Such mother cells were moved to new locations on the same plate . For NA-supplementation experiments , 90 mother cells were arrayed on one part of the YP plate with 2% , 0 . 5% , or 0 . 2% glucose . After scoring the plate into three sectors , 30 mother cells were assigned to groups that would not be moved , moved to new locations on the same plate on a sector to which 200 µl of water was applied , or moved to new locations on a sector to which 200 μl of 0 . 5 mM NA had been applied . Migrations were effected after mother cells reached 15 generations . Conditioned media experiments were performed in a similar manner . Ninety mother cells were arrayed on plates with either 2% or 0 . 2% glucose . Thirty mother cells were moved to a sector to which 200 μl of 10×-concentrated non-conditioned YP had been applied . Thirty mother cells were moved to a sector to which 200 μl of 10×-concentrated conditioned media ( 10× CM ) from cells grown in the same glucose concentration was applied . In all cases , the author performing RLS was blinded with respect to plate conditions . Survival curves were plotted with all data collected from four independent experiments ( 120 mothers in total ) , and RLS data were plotted as percent increases in median RLS compared to non-moved mother cells aging in 2% glucose media . Data are means and standard deviations from independent RLS experiments and statistical data were analyzed by one-way ANOVA . Raw data for all RLS experiments are available in S1 Data . | Though calorie restriction extends lifespan and healthspan in multiple model organisms , the intrinsic mechanisms remain unclear . In budding yeast Saccharomyces cerevisiae , manipulation of nicotinamide adenine dinucleotide ( NAD+ ) —a central metabolic cofactor—can restrict or extend replicative lifespan , suggesting that NAD+-dependent targets might be mediators of extended longevity . However , although treating cells with the NAD+ precursor nicotinamide riboside extends lifespan , intracellular NAD+ metabolites levels are not altered by glucose restriction . This suggests the potential involvement of extracellular factors in replicative lifespan extension . Here we show that though yeast cells display a longevity benefit upon glucose restriction , these cells surprisingly lose the longevity benefit if moved from their local environment to fresh glucose-restricted media . They are , however , able to regain the longevity benefit , despite the change in environment , if the new environment is supplemented with conditioned medium from glucose restricted cells . Our results suggest that calorie restriction-induced longevity is not cell autonomous and , instead , appears to be transmitted from cell to cell in S . cerevisiae via a dialyzable extracellular factor . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Calorie Restriction-Mediated Replicative Lifespan Extension in Yeast Is Non-Cell Autonomous |
Experimental evolution of the plant pathogen Ralstonia solanacearum , where bacteria were maintained on plant lineages for more than 300 generations , revealed that several independent single mutations in the efpR gene from populations propagated on beans were associated with fitness gain on bean . In the present work , novel allelic efpR variants were isolated from populations propagated on other plant species , thus suggesting that mutations in efpR were not solely associated to a fitness gain on bean , but also on additional hosts . A transcriptomic profiling and phenotypic characterization of the efpR deleted mutant showed that EfpR acts as a global catabolic repressor , directly or indirectly down-regulating the expression of multiple metabolic pathways . EfpR also controls virulence traits such as exopolysaccharide production , swimming and twitching motilities and deletion of efpR leads to reduced virulence on tomato plants after soil drenching inoculation . We studied the impact of the single mutations that occurred in efpR during experimental evolution and found that these allelic mutants displayed phenotypic characteristics similar to the deletion mutant , although not behaving as complete loss-of-function mutants . These adaptive mutations therefore strongly affected the function of efpR , leading to an expanded metabolic versatility that should benefit to the evolved clones . Altogether , these results indicated that EfpR is a novel central player of the R . solanacearum virulence regulatory network . Independent mutations therefore appeared during experimental evolution in the evolved clones , on a crucial node of this network , to favor adaptation to host vascular tissues through regulatory and metabolic rewiring .
Bacterial plant pathogens constitute a major threat to crop production . In addition , disease emergence can occur through rapid adaptation of many pathogens to new hosts [1 , 2] . Understanding how pathogens are adapting to new hosts is crucial for unraveling the mechanisms that drive disease emergence . One way to study evolution of pathogen adaptation to new hosts is to conduct experimental evolution of the pathogen in a given host over hundreds of generations [3–6] . The combination of experimental evolution with whole-genome sequencing has enabled the characterization of the mutations underlying the within-host fitness gain in various host-pathogen systems [4 , 6] . In a previous work , aiming at investigating the genetic bases of host adaptation in the bacterial plant pathogen Ralstonia solanacearum , we performed an experimental evolution of strain GMI1000 in different host plants [6] . R . solanacearum is the causal agent of bacterial wilt disease . It is recognized as one of the most destructive bacterial plant diseases affecting an unusually large host range of more than 250 plant species around the world , mainly in warm tropical climates [7 , 8] . R . solanacearum is a soilborne pathogen that infects the plants through the roots , invades the xylem vessels and spreads to aerial parts of the plant through the vascular system where it multiplies extensively and produces large amounts of exopolysaccharide ( EPS ) [9] . To cope with these various soil and plant microenvironments , R . solanacearum has evolved a complex regulatory network that senses key signals and triggers important physiological changes via global shifts in gene expression [10] . At the center of this virulence regulatory network is the global regulator PhcA , a LysR-type transcriptional regulator that controls expression of many genes [11] . Experimental evolution of the GMI1000 strain was conducted by serial passage experiments ( SPE ) from one plant individual to another in order to maintain the pathogen population in the same host for over 300 bacterial generations [6] . Five independent lineages of experimentally evolved clones were generated by conducting five parallel SPEs . Phenotypic analysis of the experimentally evolved clones demonstrated that almost 80% of them had an increased fitness in their experimental host compared to the ancestral GMI1000 clone . However , no increase in disease symptom rate was observed whether the experimental plant was a susceptible host or a tolerant host ( on which R . solanacearum grows asymptomatically ) . Interestingly , genomic sequence analysis of the evolved clones revealed single nucleotide polymorphisms ( SNPs ) in the efpR gene in three of the five lineages evolved on bean , a tolerant host [6] . A reverse genetic approach in the GMI1000 strain confirmed that the SNPs detected in the efpR gene were associated with fitness gain on bean . This gene encodes a putative transcription regulator protein of 113 amino acids belonging to the HTH_XRE ( Helix-Turn-Helix_Xenobiotic-Response-Element ) superfamily . The efpR gene is highly conserved in all the 38 R . solanacearum strains representative of the species complex diversity sequenced to date , with more than 90% protein identity between all strains . This gene is also present in other β-proteobacteria such as Burkholderia , Pandoraea and Bordetella species with more than 80% protein identity . However , the function of the EfpR protein remains unknown . The aim of the present study was to determine the functional impact of the mutations in the efpR gene , in order to understand how these mutations promote bacterial fitness in planta . We performed a transcriptomic profiling of the efpR-deleted mutant to assess the impact of EfpR on gene expression and the nature of its targets . We used a system biology approach using the recently published model of R . solanacearum [12] to predict the putative phenotypes associated with the large set of genes differentially expressed . Model predicts were used to drive functional analyses which then revealed that EfpR affects several metabolic pathways and important virulence traits such as motility and EPS production . We demonstrated that the SNPs selected in the efpR gene during the evolution experiment significantly alter the efpR expression . This work highlights the importance of EfpR as a global regulator which coordinates the expression of virulence functions to the metabolic status of the bacterial cell .
We previously identified three SNPs in the efpR gene from bacterial populations evolved on bean plants and these mutations were associated to the fitness gain on this host [6] . We investigated whether mutations in efpR also occurred in populations evolved in plant species other than bean . In a previous work , mutations in efpR were not detected in three independent populations propagated on tomato ( populations A , B and E ) and in three independent populations propagated on cabbage ( populations A , B and C ) [6] . Here , we conducted further analyses in 14 additional independent populations ( populations C and D propagated on tomato , populations D and E propagated on cabbage , five independent populations propagated on eggplant and five independent populations propagated on melon ) . This analysis was conducted using a similar PCR / sequencing approach on 96 clones randomly isolated from each population as previously described [6] . Three new SNPs in efpR resulting in non-synonymous mutations were detected: a R97Q mutation in tomato[population C] , a R98W mutation in eggplant[population A] , and a F44C mutation in melon[population B] . This observation suggested that mutations in efpR were not solely associated to a fitness gain on bean , but also on additional hosts . We tested this hypothesis by determining if the efpR mutation D49N selected on bean also conferred a fitness gain when bacteria were inoculated on other hosts ( tomato and cabbage ) , using the same conditions as previously reported for bean [6] . Competition assays were performed using the ΔefpR mutant ( GRS704 ) , the allelic mutant efpRD49N ( GRS886 ) and the complemented strain ( GRS705 ) ( Table 1 ) . The competitive index ( CI ) was determined for each strain after two passages in plant stem ( see Materials and Methods ) . The results showed that both the ΔefpR mutant and the allelic mutant efpRD49N had a mean CI significantly superior to the mean CI with the complemented strain in tomato ( Wilcoxon test , p-value = 0 . 016 and 0 . 016 , respectively ) and in cabbage ( Wilcoxon test , p-value = 0 . 028 and 0 . 016 , respectively ) ( S1 Fig ) . These results demonstrated that both the deletion of efpR and the bean-efpR allele ( efpRD49N ) also improve bacterial fitness on other plants in addition to bean . The global transcriptomes of the WT ( wild-type strain GMI1000 ) and the ΔefpR mutant were determined using an RNA-sequencing approach of strains grown in minimal media supplemented with 20 mM glutamate . All samples rendered between 10 and 15 million of Ralstonia-mapped reads , with the exception of one replicate for the ΔefpR mutant ( i . e . GRS704 . R1 ) , which rendered 37 million mapped-reads . Comparative analysis of RNA-seq data was conducted by considering differentially expressed genes ( DEGs ) as those producing a Benjamini-Hochberg ( BH ) adjusted p-value <0 . 05 and presenting an absolute Fold Change ( |FC| ) between strains >2 . Using these cutoff values , we found that the total number of DEGs in the ΔefpR mutant compared to the WT strain was 877 genes , including 522 up- and 355 down-regulated genes ( S1 Table ) . Those genes represent 17% of all GMI1000 predicted genes [13] . One third of the DEGs encoded hypothetical proteins , including 148 up- and 160 down-regulated genes , and 1/15 of the DEGs encoded predicted transcription factors , including 34 up- and 27 down-regulated genes ( S1 Table ) . However , these proportions were not significantly more than the random expectation based on the proportion of hypothetical proteins or transcription factors in the GMI000 genome . Among the predicted transcription factors that were differentially expressed in the ΔefpR mutant , the RSc3149 gene , which was one of the highest up-regulated genes ( Fold Change = 8 . 12 ) , is an EfpR homolog with 79% protein identity with the EfpR protein sequence . In order to get insights into the biological functions regulated by EfpR we performed a Gene Ontology ( GO ) enrichment analysis on DEGs ( see Materials and Methods ) . Among the GO terms significantly enriched , metabolism-related terms were particularly abundant ( Fig 1 and S2 Fig ) . Specifically , genes involved in several catabolic pathways , like amino acid and organic acid catabolism , were enriched in the EfpR-dependent up-regulated DEGs ( Fig 1 ) . In addition , genes involved in secondary metabolism ( siderophore and thiamine cofactor biosynthesis ) and in plant associated functions were also enriched in the EfpR-dependent up-regulated DEGs ( Fig 1 ) . In the EfpR-dependent down-regulated DEGs , genes involved in fatty acid biosynthesis and transporters were found to be enriched , as well as genes involved in cell cycle functions ( S2 Fig ) . The GO enrichment analysis indicated that metabolic pathways are particularly enriched in the EfpR regulon . Thus , we took advantage of the recently reconstructed genome-scale metabolic model of R . solanacearum [12] to infer the causal relationship between the EfpR-dependent DEGs ( EfpR-DEGs ) and metabolic processes . We first conducted phenotype simulations using flux balance analysis with the EfpR-DEGs as constraints ( see Materials and Methods ) . A total of 16 phenotypes were tested in 7 different environments ( see S2 Table for the list of phenotypes and environments tested ) . This analysis revealed that 13 out of the 16 phenotypes tested were affected by the EfpR-DEGs ( Fig 2A ) . Among them are several virulence-related phenotypes like EPS production , swimming and twitching motility , and secretion of type III effectors , but also other biological functions like biomass production ( Fig 2A ) . Only 3 phenotypes were not found to be affected by the EfpR-DEGs: the production of specific secondary metabolites , the biosynthesis of storage macromolecules and energy production ( Fig 2A ) . Since the GO analysis revealed enrichment in catabolic processes , we also tested the involvement of EfpR-DEGs in various catabolic capacities . We thus tested the catabolism of 58 substrates known to encompass around 90% of the substrates used by R . solanacearum [12] . This analysis revealed that the EfpR-DEGs were involved in the catabolism of 19 compounds out of the 58 tested ( 31% ) ( Fig 2B ) . These substrates were mainly amino acids and organic acids , in agreement with the GO enrichment analysis ( Fig 2B ) . Among the 19 compounds for which catabolism was affected by the EfpR-DEGs , 15 were gained in the ΔefpR mutant due to up-regulated genes ( Fig 2B ) . These predictions further suggest a role of EfpR in catabolic repression . In order to validate the phenotypes predicted to be EfpR-dependent , we conducted an analysis of the catabolic capacities of the ΔefpR mutant by using metabolic phenotype microarrays . This revealed that the ΔefpR mutant possessed a wider substrate usage capacity than the WT strain . Indeed the mutant was able to significantly better catabolize 65 substrates , but could also use 12 substrates not catabolized by the WT strain in these conditions ( S3 and S4 Figs , S3 and S4 Tables ) . The metabolic capacities of the complemented strain were similar to those of the WT strain . Among the carbon and nitrogen sources that were poorly used by the WT strain , gamma-Amino-n-Butyric Acid ( GABA ) and L-Glutamate were significantly better used as carbon sources , likewise Sodium Nitrite , L-Serine and L-Pyroglutamic acid were significantly better used as nitrogen sources by the ΔefpR mutant ( Fig 3 ) . Even among the carbon and nitrogen sources that were not used by the WT strain , this analysis revealed that the ΔefpR mutant possessed a high efficiency to catabolize L-Alanine and L-Histidine as carbon sources , Adenine as nitrogen source and L-Proline both as carbon and nitrogen sources ( Fig 3 ) . The ability to catabolize L-Glutamate , L-Proline , L-Histidine and GABA was confirmed by investigating bacterial growth of the ΔefpR mutant and the complemented strain in comparison to the WT strain in minimal media supplemented with one of these four carbon sources . The results showed that the growth rate of the mutant on these four carbon sources was significantly increased compared to the WT ( Welch t-test , p-value <1 . 32e-3 , 3 . 49e-4 , 6 . 80e-6 , and 1 . 06e-2 , respectively ) and the complemented strains ( Welch t-test , p-value <1 . 04e-4 , 3 . 69e-4 , 1 . 06e-6 and 1 . 07e-2 , respectively ) ( Fig 4 ) . All these data suggested that EfpR plays a major role in a catabolic repression operating in the bacterial cell . The EfpR-DEGs analysis predicted that motility is affected in the mutant ( Fig 2A ) . Motility is considered as one of the key factors determining the ability of the bacteria to colonize the roots of their host plants [15 , 16] . We compared the swimming and twitching motilities between the WT strain , the ΔefpR mutant and the complemented strain . Comparison of the swimming halo diameters after 24 , 36 , 48 , 60 and 72 hours between the different strains showed that the ΔefpR mutant had a significant wider mean diameter during the whole growth kinetic compared to the WT ( Mann-Whitney test , p-value = 4 . 47e-8 , 3 . 97e-8 , 3 . 69e-8 , 4 . 31e-8 and 4 . 84e-8 , respectively ) and compared to the complemented strain ( Mann-Whitney test , p-value = 3 . 44e-8 , 3 . 26e-8 , 4 . 25e-8 , 5 . 31e-8 and 5 . 32e-8 , respectively ) ( Fig 5B ) . After 72 hours , the mean diameter of the ΔefpR mutant was 44 . 55 ± 1 . 43 mm while the mean diameters of the WT and the complemented strains were 29 . 9 ± 1 . 86 mm and 26 . 85 ± 1 . 28 mm , respectively ( Fig 5A ) . From these data , we could estimate a speed of movement by swimming motility for each strain . The mean speed for the ΔefpR mutant was 0 . 81 ± 0 . 03 mm . h-1 , significantly higher than the mean speed for the WT and the complemented strains ( Welch t-test , p-value <2 . 2e-16 ) , which were 0 . 54 ± 0 . 04 mm . h-1 and 0 . 49 ± 0 . 02 mm . h-1 , respectively . Similar results were observed for twitching motility assays . The ΔefpR mutant had a significantly wider mean diameter than the WT strain ( Welch t-test , p-value <1 . 93e-8 , 1 . 36e-8 and 1 . 19e-9 after 24 , 28 and 33 hours , respectively ) ( Fig 6 ) . After 33 hours , the mean diameter of the ΔefpR mutant was 1961 ± 402 μm while the mean diameters of the WT and the complemented strains were 1443 ± 211 μm and 1568 ± 112 μm , respectively ( Fig 6A ) . Again , from these data , we estimated the mean twitching motility speed for the ΔefpR mutant at 120 ± 29 μm . h-1 , significantly higher than the mean speed for the WT and the complemented strains ( Mann-Whitney test , p-value <1 . 16e-12 and 2 . 61e-5 , respectively ) , which were 86 ± 9 μm . h-1 and 97 ± 6 μm . h-1 , respectively . The EfpR-DEGs analysis predicted that EPS production is affected in the ΔefpR mutant ( Fig 2A ) . In addition , the colonies of the ΔefpR mutant were found to be less mucoid in appearance than those of the WT after 48 hours on agar plates , suggesting a decreased EPS production in the ΔefpR mutant . We quantified the flux of EPS production in cell culture supernatant in minimal medium with L-Glutamate as sole carbon source . The kinetics of hexosamine released in the ΔefpR mutant in comparison to the WT strain , the complemented strain and the Δeps mutant were quantified ( see Materials and Methods ) . This analysis showed that the ΔefpR mutant produced 6 . 4 and 8 . 2 times less EPS than the WT and complemented strains , respectively ( Welch t-test , p-value <3 . 84e-3 and 1 . 08e-3 , respectively ) ( Fig 7 ) . However , EPS biosynthesis was not completely abolished since the ΔefpR mutant still produced EPS compared to the Δeps mutant strain ( Fig 7 ) . In order to investigate the impact of the SNPs that appeared during experimental evolution on bean [6] , we performed metabolic phenotype microarrays with the three efpR allelic mutants carrying the efpR base substitution as detected in the ‘bean’ evolved clones , namely the allelic mutants efpRP93Q ( GRS884 ) , efpRL2L ( GRS885 ) and efpRD49N ( GRS886 ) ( Table 1 ) . Similarly to the ΔefpR mutant , microarray phenotyping revealed that the three allelic mutants possessed a wider substrate usage capacity than the WT strain , being able to significantly better catabolize 39 , 42 and 38 substrates , with 4 , 5 and 5 newly used substrates compared to the WT strain , respectively ( S3 and S4 Figs , S3 and S4 Tables ) . The metabolic profiles of these three allelic mutants were similar to each other , but were intermediate to those of the ΔefpR mutant and the WT strain ( Fig 3 , S3 and S4 Figs ) . The ability to catabolize L-Glutamate , L-Proline , L-Histidine and GABA was also confirmed by investigating bacterial growth of the three allelic mutants in comparison to the WT strain on minimal media supplemented with one of these four carbon sources ( Fig 4 ) . The results showed that the growth rates of the three allelic mutants on these four carbon sources were significantly increased compared to the WT and complemented strains ( Welch t-test , p-value <0 . 05; Fig 4 ) . However , these growth rates were not significantly different from the growth rate of the ΔefpR mutant ( Welch t-test , p-value >0 . 05; Fig 4 ) . Only the efpRL2L mutant had a growth rate on L-Histidine significantly lower than the ΔefpR mutant ( Welch t-test , p-value = 0 . 025; Fig 4 ) . The virulence of the ΔefpR mutant , the complemented strain and the three efpR allelic mutants were tested on susceptible tomato plants in comparison to the virulence of the WT strain . Virulence assays were conducted on the susceptible tomato variety used during experimental evolution [6] and not on bean plants which remain asymptomatic after infection with strain GMI1000 . Fig 8 represents the log10 ( hazard ratio ) , the hazard ratio being the wilting rate for plants inoculated with the WT strain divided by the wilting rate of plants inoculated with mutant or complemented strains ( see Materials and Methods for details ) . Five biological repeats were conducted , each repeat investigating the wilting of 16 tomato plants . The obtained log10 ( hazard ratio ) for the ΔefpR mutant and for the three allelic mutants efpRP93Q , efpRL2L , efpRD49N were significantly higher than zero ( one sample t-test , p-value = 0 . 004 , 0 . 0083 , 0 . 0381 and 0 . 0287 , respectively ) ( Fig 8 ) . This result indicated that virulence of these four mutants was decreased on tomato plants compared to the WT strain , whereas the virulence of the complemented strain was not significantly different from the WT strain ( one sample t-test , p-value = 0 . 06 ) ( Fig 8 ) . Abundance of efpR mRNA in the three allelic efpR mutants was measured by qRT-PCR . Interestingly , this analysis revealed that the evolved allelic mutations , P93Q , L2L and D49N , induced a drastic reduction ( >10 fold ) in efpR mRNA level in vitro compared to the WT and complemented strains ( Fig 9 ) . This decrease of efpR expression was also true for three original ‘bean’ evolved clones carrying mutation in the efpR gene ( Fig 9 and Table 1 ) . Finally , since PhcA also controls EPS production , twitching and swimming motilities [14–17] , similarly to EfpR , we tested if the expression of efpR was dependent on PhcA . For that purpose , the efpR mRNA abundance was measured in a phcA mutant in vitro by qRT-PCR . This analysis revealed that the efpR mRNA abundance was similar in the phcA mutant and in the WT strain , thus demonstrating that the PhcA regulator didn’t control the expression of efpR in these conditions ( Fig 9 ) .
In this work , we investigated the role of the efpR gene , a gene involved in adaptation of R . solanacearum to various host plants , discovered during an experimental evolution where bacteria were maintained on plant lineages for more than 300 generations [6] . This former study showed that the fitness gain of the evolved bacteria was associated to SNPs within the efpR sequence . A similar fitness gain was obtained by deleting efpR . In the present study , we provide evidence that EfpR is mainly a repressor of gene expression ( 60% of the DEGs ) , suggesting that the mutations in efpR selected during the experimental evolution led to the de-repression of biological functions contributing to the observed fitness gain . Indeed , our data showed that EfpR regulates , directly or indirectly , a large number of genes ( about 17% of the WT genes displayed a differential expression in the ΔefpR mutant ) . As a consequence , EfpR controls many metabolic pathways and virulence-associated traits . Similar changes in global gene expression through mutations affecting regulatory network components have been observed in evolution experiments conducted with bacteria and appear to be a first and rapid adaptive response to novel environmental conditions [18] . Based on the transcriptomic profiling of the ΔefpR mutant and flux balance analysis using the genome-scale model of R . solanacearum , we predicted several candidate functions to be controlled by EfpR . Phenotypic and functional characterization of the ΔefpR mutant first confirmed that this mutant has an expanded metabolic versatility since it is able to catabolize 65 carbon and nitrogen substrates better than the WT strain , including 12 substrates that the WT strain is unable to use in our conditions . Interestingly , this increased substrate usage pattern matches well with the list of carbon sources identified in the xylem of tomato plants [19] . Indeed , many amino acid identified in tomato xylem fluids , including L-Glutamate , L-Proline , L-Histidine and GABA , are better catabolized by the ΔefpR mutant . These results therefore establish that EfpR behaves as a global catabolic repressor in R . solanacearum , directly or indirectly down-regulating multiple metabolic pathways . However , we show that EfpR also controls functions associated to bacterial pathogenesis beyond central metabolism . An efpR mutant produces much less EPS than the WT strain , and EPS has been reported to be required for full virulence on plants [9] . This could explain the reduced pathogenicity of the ΔefpR mutant when inoculated on tomato plants . From the list of DEGs in the ΔefpR mutant , it is probable also that additional factors directly or indirectly linked to virulence are under the control of EfpR . Twitching and swimming motility were also shown to be repressed by EfpR , thus raising the hypothesis that the efpR mutant may have an increased invasive capacity during plant infection . In order to get a better understanding of the impact of the efpR mutations that occurred during the evolution experiment , the three different SNPs from the evolved clones in bean were re-created in the WT strain . Interestingly , whatever the position of the SNP in the efpR gene , the three allelic mutants all shared very similar phenotypic characteristics that were very close , even if not identical , to the ΔefpR mutant . Similar enhanced growth rates were observed for the allelic and deleted mutants in presence of various carbon substrates and the allelic mutants were able to significantly catabolize a wider number of carbon and nitrogen substrates than the WT strain , even if this number was less than for the ΔefpR mutant . As observed with the ΔefpR mutant , virulence of the three allelic mutants was decreased on tomato plants . These results indicate that the efpR allelic mutants almost behave as loss-of-function mutants , which was further confirmed by showing that the three different SNPs all lead to a strong decrease in the expression of efpR ( Fig 9 ) . The residual efpR expression detected in the allelic mutants could explain the minor difference observed between the catabolic capacities of the allelic and deleted mutants . Bacterial adaptation to novel habitats through loss-of-function mutations has recently been reported [20] , with evidence that such mutations can favor adaptation to novel environments through regulatory and metabolic rewiring . Our findings support the view that the enlarged metabolic capacities of the efpR allelic mutants provide the fitness advantage occurring during growth in the plant xylem of different species and explains why such efpR mutations were rapidly fixed in the populations during the experimental evolution [6] . In addition , the strong impairment of efpR mutants to produce EPS could also contribute to this fitness gain since EPS biosynthesis represents a significant metabolic cost for the bacterial cell [12] . It cannot be excluded that the increased motility of efpR mutants also provides a competitive advantage over the WT strain by enabling a faster colonization of the host vascular system . Modeling of the predicted EfpR structure reveals that the mutations detected in bean , tomato , eggplant and melon during the evolution experiment occurred in two main domains of the protein ( S5 Fig ) : two SNPs ( F44C and D49N ) lead to changes in positions in the HTH domain just adjacent to the DNA binding site whereas three other SNPs are located in the exterior domain , inducing changes in charged amino acids ( R97Q , R98W ) or on the probable folding of the domain ( P93Q ) . This cluster of mutations in this protruding domain is suggestive of a role in interacting with another protein or co-factor and/or that it is critical for EfpR stability . A third SNP is a synonymous mutation occurring in the second codon , presumably altering the gene functionality [21] . The experimental evolution scheme developed by Guidot and collaborators [6] was based on serial passages of bacterial populations which were collected from the xylem of infected stems at the onset of wilting symptoms and then stem-inoculated on the next plant . This experimental procedure therefore avoided the root infection step , which is the natural mode of entry of the soilborne R . solanacearum in its hosts . Our results showed that the efpR mutants are significantly reduced in virulence when using ‘natural’ soil drenching infection assays ( Fig 8 ) . This probably explains why efpR mutations which occurred several times during the experimental evolution might be counter-selected in nature . It is tempting to speculate that EfpR plays an important role during steps of the R . solanacearum life cycle ( such as in the soil or during root infection ) which were not reproduced during the evolution experiment and which may be advantageous for the pathogen in its natural context . Altogether the results presented in this study establish that EfpR is a novel player of the complex regulatory network controlling R . solanacearum virulence in response to multiple environmental signals [11] . EfpR is presumably a key component of this network which tightly links the bacterial metabolism to virulence . Beyond EPS production and motility , the prediction of phenotypes affected by EfpR-DEGs suggests that additional functions such as biofilm formation , secondary metabolism and other virulence-associated proteins are also controlled by EfpR . The pleïotropic phenotype of the efpR mutant is reminiscent of the phenotype of another key R . solanacearum virulence regulatory gene , phcA , identified more than 25 years ago [22] . As efpR , a phcA mutant is defective for EPS production and many other virulence factors , and displays an hypermotile phenotype [22] . More recently , a phcA mutant was also found to have an expanded metabolic versatility [12] , being able to metabolize an even wider repertoire of metabolic substrates than the efpR mutant . This phenotypic relatedness incited us to explore the existence of functional links between PhcA anf EfpR . Our results indicated that the expression of efpR is not significantly altered in a phcA mutant ( Fig 9 ) and that phcA does not appear to be differentially regulated in the efpR mutant background . This suggests that if there is interplay between these two central regulators , this should operate at a post-transcriptional level . Future works will aim at understanding how PhcA anf EfpR jointly orchestrate a tight coupling between catabolic pathways ( supporting in planta growth ) and virulence functions , in a mechanism which appears to be driven by a resource allocation trade-off . The discovery of the role played by EfpR in this process also highlights how the experimental evolution approach helps to uncover a key regulatory component of R . solanacearum pathogenicity and adaptation to hosts .
R . solanacearum strains used in this study are described in Table 1 . Strains were grown in complete BG medium or in MP minimal medium at 28°C [23] . The pH of the MP medium was adjusted to 6 . 5 with KOH . For agar plates , BG medium was supplemented with D-Glucose ( 5 g liter-1 ) and triphenyltetrazolium chloride ( 0 . 05 g liter-1 ) . The MP medium was supplemented with various carbon sources ( L-Glutamate , D-Glucose , L-Proline , gamma-Amino-n-Butyric Acid ( GABA ) or L-Histidine ) at 20 mM final concentration . When needed , antibiotics were added to the media at the following final concentrations ( mg liter-1 ) : spectinomycin , 40; gentamicin , 10 . The plants used in this study were tomato ( Solanum lycopersicum var . Super Marmande ) , eggplant ( S . melongena var Zebrina ) , cabbage ( Brassica oleracea var . Bartolo ) and melon ( Cucumis melo var . Vedrantais ) . Four- to five-week-old plants were used for the inoculations . Plant experiments were conducted in a growth chamber under the following conditions: 75% humidity , 12h light 28°C , 12h darkness 27°C . Mutations in efpR in populations evolved in plant species different from bean was determined by PCR amplification and Sanger sequencing using the primers RSc1097-L and RSc1097-R ( S5 Table ) from cell suspension of 96 clones randomly isolated in each population as described in Guidot et al . [6] . To monitor the in planta fitness of the efpR mutants , we conducted competition assays as previously described [6] . Briefly , a mixed inoculum at a 106 CFU/ml concentration , was prepared , containing equal CFU of the mutant ( spectinomycin or gentamicin resistant ) and the WT strain , and 10 μl was used to inoculate the stem of plants . After 5 days for the tomato susceptible plants and after 15 days for the cabbage tolerant plants , the bacteria were recovered from the stem , serial diluted and plated onto BG medium with and without spectinomycin or gentamicin . The competitive index ( CI ) , which is the mutant/WT ratio recovered from the plant stem divided by the mutant/WT ratio in the inoculum , was determined after two passages in plant stem ( S1 Fig ) . A total of five replicates were performed for each strain . Base substitution mutants at the efpR locus were created using the natural co-transformation ability of R . solanacearum and following an adapted protocol of the multiplex genome editing by natural transformation ( MuGENT ) described in Dalia et al . [24] . Briefly , PCR primers were designed to amplify 3kb arms of homology from either side of the SNPs ( S5 Table ) . Then , high fidelity PCR has been performed on the genomic DNA of clones evolved in bean . The PCR products were then concentrated by sodium acetate precipitation and served as unselected products . The selected product was carried by an insertional plasmid pAGSI ( a pRCG-GWY-based plasmid [25] ) linearized by ScaI . Natural transformation of R . solanacearum was performed , as previously described [26] , using 300 ng of the selected marker and 1 μg of unselected marker added to 50 μl of competent cells . R . solanacearum transformants were selected after growth on BG medium supplemented with gentamicin and then the co-transformants were validated by PCR amplification and Sanger sequencing using the primers RSc1097-L and RSc1097-R ( S5 Table ) . The resulting efpR allelic mutants were called efpRP93Q , efpRL2L and efpRD49N mutants in the present paper ( Table 1 ) . Total RNA was extracted from the WT GMI1000 strain and the ΔefpR mutant strain growing in MP medium supplemented with L-Glutamate 20 mM at comparable cell densities ( OD600nm ~ 0 . 5 ) . Three biologically independent experiments were conducted for the GMI1000 strain and two for the ΔefpR mutant . Before RNA extraction , the bacterial culture was first stopped by mixing 1 ml ethanol/phenol ( 95:5 ) to 20 ml of culture during 3 min using a vortex . The culture was then centrifuged at 4000 x g for 10 min at 4°C and the pellet was resuspended in 200 μl H2O RNase free for total RNA extraction . Total RNA was isolated and depleted of ribosomal RNAs as previously described [27] . The oligonucleotide sets used for the ribosomal RNA depletion were specifically designed to target R . solanacearum rRNAs ( S5 Table ) . Oriented paired-end RNA sequencing ( 2x100 bp ) was carried out by Fasteris ( Fasteris SA , Plan-les-Ouates , Switzerland ) , using an Illumina Hiseq 2000 and the procedures recommended by Illumina , with adaptors and amplification primers designed by Fasteris . The size of selected inserts was 150–250 bp . Libraries were sequenced in paired-end . Two technical repeats per RNA sample were performed . Read pairs were mapped using the glint software ( http://lipm-bioinfo . toulouse . inra . fr/download/glint/ ) with parameters set as follows: matches ≥40 nucleotides , with ≤3 mismatches , only best-scoring hits taken into account . Ambiguous matches ( same best score for several read-pairs ) were removed . Finally , between 8 . 2 and 23 . 4M non ambiguous read-pairs were obtained . Mapped reads were imported into R environment . The package HTSFilter was used to eliminate very low-expressed genes from the analysis . A total of 5157 out the predicted 5307 genes were thus kept in . R package DESeq2 was used to normalize and complete the differential analysis by conducting the built-in Wald test [28] . The p-values thus obtained were adjusted for multiple comparisons using the Benjamini-Hochberg ( BH ) method [29] . Genes with a BH-adjusted p-value <0 . 05 and Fold Change |FC|>2 between strains were taken into further consideration in this work . We collected Gene Ontology from the Gene Ontology ( GO ) Consortium website ( http://geneontology . org/ ) and assigned the GO annotation to R . solanacearum proteome using InterProScan [30] . Enrichments analysis of GO terms filed under Biological Process were separately conducted for up- and down-regulated genes using the R package topGO . Terms producing a p-value <0 . 05 after the Fisher exact test were taken into further consideration . GO enrichment visualization was performed using Cytoscape 3 . 2 . 1 and plug-in the Biological Networks Gene Ontology [31] . Predictions of bacterial phenotypes were performed by Flux Balance Analysis [32] using the genome-scale metabolic network model of R . solanacearum GMI1000 [12] . Simulations were conducted using the software FlexFlux [33] . Predictions of the phenotypes feasibility were done by optimizing the metabolic fluxes corresponding to each phenotype using the DEGs as constraint . Two sets of DEGs constraints were used corresponding to the up-regulated genes set and the down-regulated genes set . Indeed , the states of the genes in the DEGs sets were set to 0 . This simulated the behavior of genes which are not expressed or deleted . Thus , by doing so , we assessed the essentiality of the DEGs sets for the various phenotypes . Predictions were conducted considering various environmental conditions as constraints . These environmental conditions correspond to different composition of the surrounding of the cells , in vitro or in planta . Hence , uptake of a substrate was possible , i . e . exchange fluxes can be >0 , only if the compounds are present in the environment . A set of 7 environmental conditions were tested ( S2 Table ) . Then , the probability of the phenotypes to be affected by the EfpR-DEGs sets was assessed by calculating the number of environments in which the phenotype was affected on the number of environments tested . Prediction of bacterial catabolism was performed by Flux Balance Analysis as described previously . The objective function optimized was the ATP hydrolysis flux ( Non-growth associated maintenance ) . The environmental conditions tested correspond to the Biolog phenotype microarray environments PM01 and PM02 . In those conditions only one carbon substrate is available allowing simulating a total of 190 catabolic capabilities corresponding to more than 90% of the R . solanacearum versatility [12] . Phenotypic microarrays were performed using Biolog Phenotype Microarray plates PM01 , PM02 and PM03 according to the constructor protocol with the following modifications . Before inoculation of Biolog fluid IF-0 , the cells were collected from a static culture on plate containing agar BG medium and resuspended in 15 ml sterile H2O supplemented with D-glucose at 20 mM with an OD600nm between 0 . 3 and 0 . 5 . The bacterial cells were then incubated during 6h at 28°C and 180 rpm agitation . This step , which starved the bacteria for nitrogen , phosphorus and sulfate , was found to reduce the background observed in PM03 . D-glucose at 20 mM concentration was used as carbon source for the inoculation of the PM03 plate . The measurements were recorded on Omnilog reader ( Biolog ) during 82 . 5 hours . Three biological replicates were performed . Data were analyzed and statistic calculated using the R software package opm [34] . Overnight cultures grown at 28°C and 180 rpm shaking in minimal medium supplemented with L-Glutamate 20 mM were used to inoculate three replicates of 200 μl of fresh minimal medium supplemented with L-Glutamate , L-Proline , GABA or L-Histidine 20 mM with an initial OD600nm at 0 . 05 . Bacterial growth was performed in 96-well microplates and monitored using a microplate spectrophotometer ( FLUOstar Omega , BMG Labtech , Offenburg , Germany ) . The incubation temperature was fixed at 28°C and the microplates were shaken continuously at 700 rpm using the linear-shaking mode . Measures of OD600nm was performed every 5 min during 50 hours . Three biological repeats were performed . Differences between in vitro growth rates were tested using a Welch t-test performed in the R statistical software . Bacteria were grown in BG medium for 1 day at 28°C . The cultures were then diluted and plated on BG agar medium supplemented with 20 mM Glucose to have isolated colonies . The colony diameter was compared after 24 , 28 and 33 hours of incubation at 28°C . Colonies were examined for twitching motility under a stereomicroscope ( Leica MZ FLIII ) . Colonies with layered edges and multiple irregular projections ( ‘spearheads’ ) are typical of migration of cells over the agar surface by twitching motility ( Fig 6A ) [15] . Digital images were acquired using a colour CCD camera ( Leica DFC420 ) and saved as uncompressed TIFF files . The experiment was repeated twice , with twenty colonies for each strain per replicate . Differences between colonies diameters were tested using a Welch t-test performed in the R statistical software . Bacteria were grown in MP minimal medium supplemented with L-Glutamate 20mM for 1 day at 28°C . The bacterial culture ( diluted to 1 . 108 CFU/ml ) was stab-inoculated into the agar of freshly poured MP minimal medium semi agar ( 0 . 3% ) plates supplemented with L-Glutamate 20 mM . The diameter of the swimming halo was then measured after 36 , 48 , 60 and 72 hours of incubation at 28°C . The experiment was repeated twice , with ten plates for each strain per replicate . Differences between swimming halo diameters were tested using a Mann-Withney test performed in the R statistical software . Bacteria were grown in MP minimal medium supplemented with L-Glutamate 20mM for 1 day at 28°C then centrifugated ( 5000 x g ) during 15 min to remove EPS . Pellets were resupended in fresh MP minimal medium supplemented with L-Glutamate 20mM and then diluted to have an OD600nm = 0 . 2 . After 0 , 4 . 33 , 6 . 5 , 8 . 83 and 11h of incubation at 28°C , 2ml of culture has been filtrated ( 0 . 22 μM ) and EPS production of R . solanacearum cells was assayed by the Elson Morgan method as described in Peyraud et al . [12] . Bacteria were grown overnight in BG liquid medium with appropriate antibiotics . Virulence tests were conducted on 16 4-week-old tomato plants grown in 30x30 cm trays filled with compost . The 16 tomato plants were inoculated by soil drenching with 500 ml of a 5 . 107 CFU/ml bacterial suspension . Plant symptoms were scored daily using a disease-index scale ranging from 0 ( no symptoms ) to 4 ( complete wilting ) as described in Poueymiro et al . [35] . This virulence test was repeated five times for each strain . In order to compare the virulence of the efpR mutants , the complemented strain and the WT strain on tomato plants , the disease scoring was transformed into binary data , with a disease index below 3 corresponding to 0 and a disease index equal to or higher than 3 corresponding to 1 . This transformation was performed in order to construct survival curves and to apply the survival analysis statistical protocols [36] . The hazard is defined as the slope of the survival curve and is a measure of how rapidly plants are dying . The hazard ratios comparing the survival curve of each efpR mutant or complemented strain to the survival curve of the WT strain were calculated by Graphpad Prism 5 . 0 software and the log10 ( hazard ratio ) transformation was used to analyze the data as previously described [37] . Total RNA were isolated using TRIzol Reagent ( life technologies ) followed by RNeasy MiniElute Cleanup Kit ( Qiagen ) . To avoid contamination by genomic DNA each sample was treated with the TURBO DNA-free Kit ( life technologies ) . The reverse transcription was performed on 1 μg of total RNA using the Transcriptor Reverse Transcriptase ( Roche ) with random hexanucleotides primers . Quantitative PCRs were performed on a Roche LightCycler480 using The LightCycler 480 SYBR Green I Master ( Roche ) . Cycling conditions were as follows: 95°C for 5 min , 45 cycles at 95°C for 15 s , 60°C for 20 s and 72°C for 20 s . The specificity of each amplicon was validated with a fusion cycle . The mean efficiency of each amplicon group was determined using the LinRegPCR software [38] . The expression of efpR was normalized using the geometric average of three selected reference genes ( RSc0403 , RSc0368 and RSp0272 ) for each sample [39] . All kit and reagents were used following the manufacturer’s recommendations . The primer sets used in the experiments are listed in S5 Table . | Among plant pathogens of major economic and food crops , Ralstonia solanacearum , the causal agent of bacterial wilt , is recognized as one of the most destructive plant bacterial diseases . In addition , the emergence of new pathotypes , more aggressive and adapted to new hosts , has been reported . During an evolution experiment of R . solanacearum , where bacteria were maintained on plant lineages for more than 300 generations , we demonstrated that several single mutations in the regulatory gene efpR were associated with fitness gain on plants . However , the function of the EfpR regulator was totally unknown . In this work , we provided evidence that EfpR controls several metabolic pathways and important virulence traits of R . solanacearum . We then demonstrated that the single mutations selected in the efpR gene during the evolution experiment strongly alter the efpR expression , and thus enlarge the metabolic capacities of the bacterial cell . Altogether , our study reveals that EfpR is a novel key component of the complex regulatory network of the R . solanacearum cell , tightly linking the bacterial metabolism to virulence in response to multiple environmental signals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"bacteriology",
"organismal",
"evolution",
"microbiology",
"cloning",
"plant",
"science",
"genome",
"analysis",
"microbial",
"evolution",
"crops",
"plant",
"pathology",
"molecular",
"biology",
"techniques",
"plants",
"research",
"and",
"analysis",
"methods",
"legumes",
"genomics",
"crop",
"science",
"beans",
"plant",
"bacterial",
"pathogens",
"tomatoes",
"gene",
"ontologies",
"molecular",
"biology",
"fruits",
"agriculture",
"phenotypes",
"gene",
"identification",
"and",
"analysis",
"plant",
"pathogens",
"genetics",
"mutation",
"detection",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"evolutionary",
"biology",
"bacterial",
"evolution",
"organisms"
] | 2016 | Enhanced in planta Fitness through Adaptive Mutations in EfpR, a Dual Regulator of Virulence and Metabolic Functions in the Plant Pathogen Ralstonia solanacearum |
In nature the prevalence of Leishmania infection in whole sand fly populations can be very low ( <0 . 1% ) , even in areas of endemicity and high transmission . It has long since been assumed that the protozoan parasite Leishmania can manipulate the feeding behavior of its sand fly vector , thus enhancing transmission efficiency , but neither the way in which it does so nor the mechanisms behind such manipulation have been described . A key feature of parasite development in the sand fly gut is the secretion of a gel-like plug composed of filamentous proteophosphoglycan . Using both experimental and natural parasite–sand fly combinations we show that secretion of this gel is accompanied by differentiation of mammal-infective transmission stages . Further , Leishmania infection specifically causes an increase in vector biting persistence on mice ( re-feeding after interruption ) and also promotes feeding on multiple hosts . Both of these aspects of vector behavior were found to be finely tuned to the differentiation of parasite transmission stages in the sand fly gut . By experimentally accelerating the development rate of the parasites , we showed that Leishmania can optimize its transmission by inducing increased biting persistence only when infective stages are present . This crucial adaptive manipulation resulted in enhanced infection of experimental hosts . Thus , we demonstrate that behavioral manipulation of the infected vector provides a selective advantage to the parasite by significantly increasing transmission .
Parasites exhibit myriad adaptations to ensure their survival and transmission from host to host . These include manipulation , where infection elicits a specific behavioral response from the host of selective benefit to the parasite [1–3] . For vector-borne diseases like leishmaniasis there are two hosts to consider: the invertebrate host and the mammalian host . These come together at a critical point in the parasite life cycle , transmission . Any manipulation by the parasite that increases both the contact with and biting of the mammalian host will combine to increase the probability of transmission . The ability of parasites ( nematodes , protozoa , bacteria , or viruses ) to influence the feeding of their arthropod vectors has long been observed [2–4]; however , in the majority of examples , neither the manipulator molecules involved nor the effect on transmission , has been demonstrated . The protozoan Leishmania is a parasite of both humans and animals transmitted by the bite of female phlebotomine sand flies . The World Health Organization has estimated that 12 million people are infected with Leishmania and 350 million are at risk of infection [5] . However , despite the medical importance of this tropical parasitic disease , surprisingly little is known of the interaction between Leishmania parasites and the feeding behavior of the sand fly vector . After being ingested by the sand fly along with the blood , Leishmania amastigotes transform to motile promastigote forms and undergo a complex development , limited to the gut of the sand fly [6] . In all Leishmania species , the final phase of development results in the colonization of the junction between the sand fly anterior midgut and foregut at the stomodeal valve , a structure that regulates blood intake during feeding . It is here that the parasites undergo their final multiplication and differentiation into highly motile mammal-infective metacyclic promastigote forms [6] , a process termed metacyclogenesis . Recent work on transmission of leishmaniasis demonstrated that metacyclic promastigotes of Leishmania mexicana are regurgitated from the midgut of the sand fly vector accompanied by a viscous gel-like material of parasite origin [7] . This promastigote secretory gel ( PSG ) is a potent parasite virulence factor , and , together with sand fly saliva , significantly enhances cutaneous infections when co-delivered into the skin of the mammalian host [7 , 8] . The major component of the gel responsible for disease exacerbation is filamentous proteophosphoglycan ( fPPG ) , a very high molecular mass glycoprotein unique to Leishmania [9] . In its quest for transmission , Leishmania can also considerably modify the sand fly gut environment by damaging the stomodeal valve [10 , 11] and physically blocking the gut with fPPG , which forms the 3-D matrix of the parasite-derived gel [7 , 12] . The combination of these two events results in the blockage of the anterior midgut with a plug of Leishmania promastigotes and their gel , which distends and permanently holds open the already eroded valve . This is hypothesized to adapt the sand fly for transmission by promoting the reflux of both parasites and gel during blood feeding , dubbed the “blocked fly hypothesis” [13] . The resulting blockage of sand flies , first noted in the early twentieth century [14] , interferes with feeding and limits the volume of blood a fly can obtain [15–17]; and , probably as a consequence , infected flies have been noted to probe the skin more frequently and spend more time feeding [15–18] . Although the observation that infected sand flies can experience difficulty with feeding is not new , the connection between this behavior and the likelihood of increased parasite transmission was not made until the late 1970s when Killick-Kendrick and co-workers [16] re-analyzed the data of Chung et al . [15] . They showed that Phlebotomus chinensis were more likely to transmit Leishmania donovani to hamsters when flies probed and took no blood compared to those infected flies that took a blood meal . This demonstrated that Leishmania could manipulate the feeding ability of the sand fly to promote its own transmission success; however , neither the extent of the manipulation nor the mechanisms that underlie it have been investigated since . These various previously described effects of the parasite on the sand fly host promote host-vector contact and transmission , and are thus potentially advantageous to the parasite , but are useful by-products of infection rather than being examples of true manipulation . Moreover , it is important to bear in mind that host contact and blood feeding do not come without risks and present a considerable danger to both vector and parasites alike when the host defends itself [19] , so it is not clear if any of these effects are selectively advantageous . In such a scenario , the timing of transmission opportunities is critical in order to minimize the risk of being killed and to promote the survival of the vector long enough to complete a parasite's development to infectiousness [1 , 2 , 19 , 20] . In this study , we investigate the behavior of Leishmania-infected sand flies . We describe an elegant and novel example of behavioral manipulation in which the timing of parasite development is linked to the feeding persistence of the sand fly vector and demonstrate that such a strategy enhances Leishmania transmission .
The two known parasite-mediated events that directly enhance transmission of Leishmania are metacyclogenesis and secretion of fPPG . Since either or both could exert some pathogenic and/or manipulative effect on the sand fly host , we investigated their kinetics in detail . Sand fly infections initiated with tissue-derived amastigotes of both L . mexicana ( experimental combination ) and Leishmania infantum ( natural combination ) in Lutzomyia longipalpis were produced , and the generation of metacyclic promastigotes and fPPG synthesis was examined ( Figure 1 ) . The onset of metacyclogenesis was more rapid and the numbers of metacyclic promastigotes higher in L . mexicana than in L . infantum ( Figure 1A ) . This was mirrored in the quantity and appearance of fPPG ( Figure 1B ) , being first detected on day 2 versus day 4 in L . mexicana– and L . infantum–infected flies , respectively . To test this apparent linkage further , the rate of development was deliberately modulated by infecting flies with L . mexicana axenic amastigotes in either exponential or stationary growth phase [21] . Amastigotes in either growth status can transform to and grow as promastigotes , but those in exponential growth phase accomplish this transition more quickly . Infection of flies with exponential phase axenic amastigotes accelerated transformation and differentiation compared to infection with stationary phase amastigotes , causing both metacyclic promastigotes ( Figure 1C ) and fPPG ( Figure 1D ) to appear earlier and in greater amounts . Thus , these data demonstrate a direct association between metacyclogenesis and fPPG secretion . The development of transmissible Leishmania infections in the sand fly gut seems likely to exact a fitness cost on the fly host . For example , the considerable distension of the gut by the gel-like plug [17] , damage to the stomodeal valve [10 , 11] , and the diversion of nutrient resources to parasite growth [22] each have potential for causing pathology to sand flies . We examined the effect of infection on the longevity of Lu . longipalpis . Both L . mexicana and L . infantum infections were found to significantly reduce longevity under normal lab conditions ( log rank test: p < 0 . 0001 ) , from a median of 11 d for controls to 9 d for both L . mexicana– and L . infantum–infected flies . Although little is known regarding the survival of flies in the wild , in some respects the natural environment may be more stressful and hazardous than the conditions experienced in laboratory insectaries . Therefore , the survival of flies exposed to a combination of stress conditions was also investigated ( Figure 2B ) . Again , infections with L . infantum or L . mexicana were both found to reduce sand fly longevity compared to controls . Stress conditions by themselves reduced the longevity of uninfected control flies from a median of 11 d to 7 d , but caused an even more marked reduction in both groups of infected flies than under benign conditions ( L . mexicana normal versus stressed: median longevity 9 d versus 5 d , p < 0 . 0001; L . infantum normal versus stressed: median longevity 9 d versus 6 d , p < 0 . 0001 ) . A final test of longevity was made by infecting flies with either exponential phase– or stationary phase–cultured L . mexicana amastigotes ( Figure 2C and 2D ) . Flies with accelerated Leishmania development derived from exponential phase amastigotes experienced significantly earlier mortality than those infected with stationary phase amastigotes , and the pathogenicity of these infections was exacerbated when stress conditions were applied ( logarithmic versus stationary , normal conditions: median longevity 5 d versus 7 d , p < 0 . 0001; logarithmic versus stationary , stressed conditions: median longevity 3 d versus 5 d , p < 0 . 0001 ) . The conclusion of these experiments is that infection of Lu . longipalpis with Leishmania exerts a fitness cost as expressed by reduced longevity , and that this is linked to parasite developmental rate . Next we examined possible effects of Leishmania on fecundity , since infection of certain vectors ( for example , malaria in mosquitoes ) can reduce reproductive fitness by causing production of fewer eggs [2] . In mosquitoes , it has been suggested that this may be a compensatory mechanism to ameliorate reduced longevity by re-diverting resources that would otherwise be used for egg production and egg laying , enabling the parasite to complete development to infective sporozoite forms . However , when we examined this possibility , none of the Leishmania infections described here ( L . mexicana/L . infantum , stressed or unstressed ) had any significant impact on immediate reproductive fitness as measured by egg production ( Table S1 ) , indicating that this is not a strategy used by Leishmania in this vector . The data described above and presented in Figures 1 and 2 indicate that an interesting balance exists in the Leishmania–sand fly relationship: more aggressive and quickly developing infections will lead to increased numbers of metacyclics and fPPG ( e . g . , Figure 1C and 1D ) , thereby increasing the probability of transmission; but this will also exert a greater fitness cost on the vector ( e . g . , Figure 2C ) , thereby reducing the probability of transmission . Thus , although parasite developmental rate does vary between parasite–vector combinations and can also be influenced by experimental or environmental conditions in the laboratory and field , a priori there is no obvious selective benefit to the parasite of developing more quickly or more slowly . In the face of these potentially conflicting influences on transmission , any alteration of vector feeding behavior by the parasite that will independently increase the probability of transmission will be of clear selective advantage . Such behavior might include factors encouraging a fly to find a host and begin feeding and those that encourage a fly to continue to feed and/or re-feed . Also the defensive behavior of the mammalian host itself is important , as this will influence the success of vector blood feeding [23] . Therefore , the timing of vector feeding with respect to parasite development is a crucial behavior , since successful transmission of parasites requires that their vectors survive at least long enough for the parasites to complete their development to an infective stage ( the extrinsic incubation period ) . In the rodent malaria–mosquito model of Plasmodium yoelii in Anopheles stephensi , infected mosquitoes experienced an increase in feeding persistence once the extrinsic incubation period was completed [20] . Feeding persistence is the repetition of feeding attempts when prevented from feeding , i . e . , a behavioral mechanism in the face of simulated host defensive behavior . Such biting persistence of a vector to resume feeding after interruption is considered an important parameter of vectorial capacity , which theoretically could promote transmission to multiple hosts [24] . Therefore , we investigated whether infection of Lu . longipalpis increased their feeding persistence . Control or infected flies were offered an anaesthetized mouse but were prevented from continued feeding by gentle brushing of the antennae after probing . When the assay was complete , each fly was dissected to determine the number of metacyclic promastigotes present . The results showed that both L . mexicana– and L . infantum–infected flies displayed a positive correlation between persistence and number of metacyclic promastigotes/fly ( Figure 3A ) . When the timing of persistence was examined in L . infantum–infected sand flies there was no increase apparent in the early phase of infection , up to day 9 , but thereafter a significant increase compared to control flies ( Figure 3B ) . The increase in feeding persistence during days 10–11 occurred concurrently with the increase in metacyclic promastigotes in L . infantum–infected flies ( Figure 1A ) . A similar result was observed in L . mexicana–infected flies , and crucially the increase in persistence was apparent earlier ( Figure 3C ) as was the appearance of metacyclic promastigotes ( Figure 1A ) . Thus , there was a strong correlation between increased feeding persistence and the accumulation of infective forms . To directly test this association , we again used L . mexicana axenic amastigotes to modulate parasite developmental rate . Sand flies infected with exponential phase amastigotes showed an early onset of increased persistence compared to those infected with stationary phase amastigotes ( Figure 3D ) . In these experiments , the exponential phase amastigote infections increased the persistence of the sand fly 22-fold and 7-fold on day 7 compared to uninfected control and stationary phase amastigote–infected flies , respectively ( average persistence ± SE: blood control , 13 s ± 11 s; stationary-L . mexicana , 89 s ± 20 s; exponential-L . mexicana , 290 s ± 72 s ) . These results demonstrated for the first time that Leishmania can control this critical aspect of sand fly feeding behavior . We also determined the duration and frequency of feeding attempts in uninterrupted sand flies . With regard to duration , flies with mature L . infantum and L . mexicana infections ( i . e . , with metacyclic promastigotes and fPPG in their anterior midgut ) took an average of 1 . 3 and 2 . 4 times longer ( total time ) to feed on anaesthetized mice compared to uninfected flies ( average feeding time ± SE of blood control , 322 s ± 40 s; L . infantum–infected , 419 s ± 44 s; L . mexicana–infected , 759 s ± 52 s ) , and in accordance with the blocked-fly hypothesis , these infected flies were more likely to obtain only a partial meal of blood ( Tables S2–S4 ) . Both of these results confirm the reduced feeding capacity of infective sand flies . With regard to frequency of biting , a possible consequence of an insufficient blood meal is that this would encourage a fly to attempt to feed more frequently . However , interestingly , when this was tested we found that infected flies exposed to unconscious mice did not demonstrate increased probing; rather , they remained feeding for longer until either they finally engorged ( often partially ) or they gave up completely . This shows that both the presence of an infection in the fly and the interruption of biting are required to reveal the observed alterations to vector feeding behavior . This makes sense , because under natural conditions potential hosts are obviously not anesthetized and sand flies can give a noticeable , even painful bite , so host defensive behavior is very likely in such cases . Further , even where the bite is not noticeable interruption could easily occur through the normal activity of the host disturbing the feeding activity of the sand fly . Importantly , for the increase in feeding persistence of infected flies to be of adaptive value to the parasite the behavior must lead to increased transmission . During a transmission attempt interruption to blood feeding can arise either from an insufficient blood flow to the bite site or from the activation of host defensive behavior [2 , 23] . In both scenarios a partial blood meal is likely to promote a fly to re-feed in order to obtain sufficient nutrients for egg-maturation , and as indicated above , persistence is further enhanced in the case of a Leishmania-infected sand fly . Although re-feeding on the same host would clearly benefit the parasite , re-feeding on multiple hosts is particularly advantageous as this will increase the basic reproduction number ( R0 ) of the infection ( the number of new cases of a disease that arises from a single infection ) . Such behavior has been proposed with malaria-infected mosquitoes [24] , but there is only circumstantial evidence for this to apply to Leishmania-infected sand flies [25 , 26] . Therefore , we modified the interruption-behavioral assay by including a second anaesthetized mouse in the cage to assess the probability of the fly choosing to feed on a different host when interrupted ( Figure 4 ) . Flies were tested at 5 , 7 , and 10 d post-infection to provide infections at different stages of metacyclogenesis ( beginning , middle , and end ) for both Leishmania species . The results showed that infection with both L . infantum and L . mexicana promoted feeding on multiple hosts , whereas uninfected flies in most cases gave up completely when interrupted . Furthermore , this behavior correlated with metacyclogenesis , because exponential phase L . mexicana infections exhibited an increased and earlier tendency to initiate a feed on a new host earlier than infections initiated with stationary phase L . mexicana amastigotes ( p < 0 . 005 ) . These data demonstrated that persistent feeding behavior can lead to an increase in the number of hosts that become infected . The final issue we considered was whether this behavior could also lead to an increased number of parasites per infected host . As a preliminary experiment , we first investigated whether there was any inherent difference in infectivity of metacyclic promastigotes obtained from exponential phase or stationary phase–initiated fly infections ( Figure 5A ) . These populations proved to be of equal infectivity when isolated from flies and needle-injected into BALB/c mice; therefore , any difference in outcome cannot be ascribed to a difference in the quality of the parasites . To assess the direct consequence of vector behavioral manipulation for Leishmania transmission and infection we repeated the persistence assay with individual exponential and stationary phase L . mexicana amastigote–infected flies on individual BALB/c mice; however , in this experiment the mice were maintained to monitor the evolution of the resulting cutaneous lesions ( Figure 5B ) . Day 7–infected flies were used because of the large differences in feeding persistence demonstrated in previous experiments ( Figure 3D ) . This experiment revealed a significant difference in lesion pathology between the two groups of mice ( p < 0 . 05 ) ; the more persistent exponential phase–infected flies generating more aggressive infections ( average ± SE fly persistence: exponential = 193 s ± 94 s; stationary = 35 s ± 16 s ) . This was also reflected in the final parasite burdens for the two groups of mice ( 3 . 72 ± 0 . 64 ×108 versus 2 . 28 ± 0 . 82 ×108 amastigotes/lesion ) . From the above we know this difference was due to the delivery of parasites by the sand fly and not from any qualitative differences in parasite infectivity ( Figure 5A ) . In a parallel experiment where infected flies were allowed uninterrupted feeding on mice , the courses of the resulting infections were not significantly different ( Figure 5C ) , and the accompanying parasite burdens were very similar . This is an important control because it shows that the infections in the two populations of flies are equally capable of infecting mice , but what is different is the persistence of the flies in the face of interruption ( Figure 5B ) . Collectively , these data showed for the first time that persistent vector feeding behavior can lead to both an increase in the numbers of parasites transmitted per host and the number of hosts infected . Is this an example of adaptive or nonadaptive manipulation of vector behavior ? It is notoriously difficult to dissect these apart; however , the current data satisfy the criteria described by Poulin [27] and Thomas et al . [1] , enabling us to conclude that the behavioral manipulation is adaptive . First , it is complex—the manipulation relies on the successful establishment and development of the parasite within the sand fly and this requires many intrinsic barriers to be overcome [6] . Second , it shows evidence of “purposive design”—the behavior is exhibited when the parasite is ready for transmission . Third , the fPPG gel appears to have arisen independently in other Leishmania–sand fly combinations [12] and has interesting similarities with the polysaccharide biofilm production and mechanism of Yersinia pestis ( plague ) transmission by the flea [28 , 29] . Fourth , and most importantly , we have shown that the manipulation directly increased the fitness of the parasite through enhanced transmission . This manipulation was shown to correlate closely with metacyclogenesis of the parasites and the accumulation of the fPPG gel that accompanies this process . fPPG is a potent virulence factor for Leishmania that benefits both the transmission of parasite from the sand fly and infection of the mammalian host [7] . Previously , it was suggested that the attachment of Leishmania parasites in contact with the foregut might interfere with the function of mechanoreceptors that detect blood flow , and this might explain the reduced ability of flies to take a blood meal [4 , 30] . However , in view of the results presented here , we propose that the functioning of these mechanoreceptors is impaired by fPPG , which plays the role of manipulator molecule . This in turn may promote the hunger state and the persistence of the fly , or alternatively , increase the threshold blood volume at which blood-seeking behavior is inhibited . Currently , we are applying a genomics approach to investigate the effect of infection on sand flies [31] , which may enable us to identify genes involved in the manipulation . Further , we ultimately aim to examine the role of manipulation in a field setting . In this study , we report that under experimental conditions Leishmania parasites can manipulate the feeding behavior of the sand fly to promote its own transmission . Our data reveal that under conditions of interruption , the feeding persistence of the infected sand fly increased in parallel with the development of the parasite , the number of transmission stages ( metacyclic promastigotes ) , and the accumulation of fPPG . As a consequence , flies with the highest number of metacyclic promastigotes were also found to be more likely to initiate a second feed on naïve hosts , and the lesions generated from the bites of individual infected flies demonstrated that this form of behavioral manipulation directly results in enhanced parasite transmission . This study shows that Leishmania transmission is the product of the following: the physical blockage of the gut with fPPG that ensures regurgitation of infective forms [7 , 12 , 13]; the subsequent exacerbation of infection in the mammalian host through the action of fPPG and vector saliva [7 , 8]; and now the manipulation of feeding behavior according to the presence of infective forms available for transmission . These factors highlight the close coevolution between the Leishmania parasite and its sand fly host . This study significantly improves our understanding of Leishmania transmission , and these findings can now be incorporated into both epidemiological and experimental models of infection .
L . mexicana ( MNYC/BZ/62/M379 ) or L . infantum ( syn . Leishmania chagasi ) ( MHOM/BR/76/M4192 ) were cultured as previously described [22 , 32] . Formalin-fixed samples of parasites were counted using a Neubauer hemocyctometer , and their developmental morphology assessed through measurement of parasites from Giemsa-stained smears , using the system of Rogers et al . [17] . Five-day-old Lutzomyia longipalpis ( Jacobina strain ) female sand flies were infected with L . mexicana or L . infantum amastigotes through an artificial membrane feeding system at a density of 2 × 106 amastigotes/ml in fresh rabbit blood [33] . L . mexicana amastigotes were harvested from the rump lesions of female BALB/c mice , and L . infantum amastigotes were isolated from the spleens of BALB/c mice . In some experiments , L . mexicana amastigotes were axenically cultured prior to use [34] . Blood-fed flies were separated and maintained under a 12-h light:dark cycle at 28–30 °C , 80%–95% relative humidity , and supplied with 70% ( w/v ) sucrose ad libitum . Flies were denied the opportunity to lay eggs to minimize post-oviposition mortality [16] , and all dissected flies with mature infections were observed to contain eggs ( gravid ) . In some experiments , female flies were stressed on a daily basis by incubating them at 16 °C without a sugar meal for 2 h , and by forcing them to fly during a 30-min period by agitating the cage . To assess sand fly survival , cages were monitored daily for the occurrence of new deaths . Statistical analysis of survival data was performed using the log rank test [35] . Pooled infected sand fly homogenates were immunoblotted to detect fPPG using monoclonal antibody AP3 as previously described [7] . The method of Anderson et al . was used [20] with slight modifications . Feeding persistence was defined as the amount of time ( up to a maximum of 10 min ) that a sand fly continued to land and insert their mouthparts after being interrupted . BALB/c mice that were age and weight matched were used as the source of the blood meal in each experiment . The mice were anesthetized and placed into a 25-cm3 netted cage; their bodies were screened with netting except for their right leg , to introduce an element of searching during feeding . Care was taken to keep the same orientation and position of the mice for each exposure . Sand flies were released into the cage singly , and the time taken for each fly to find the right paw and insert their mouthparts for a feed was recorded . A sand fly was removed and recorded as a “no-feed” when 10 min after release had elapsed without it initiating a feed . Flies that began feeding within this time were allowed 1 min before interruption by gently brushing the antennae . The disturbance was repeated , allowing 10 s of feeding before the next disturbance , until the sand fly gave up or took longer than 1 min before re-feeding . The total time a fly spent attempting to re-feed was recorded as the measure of feeding persistence . Following each observation , sand flies were transferred to a glass vial , knocked down on ice , and midguts were dissected , homogenized , and the parasites quantified as described [17] . A modified form of the persistence assay was used , in which flies were assayed individually for their decision to feed on an alternative host in the face of interruption . Two anesthetized BALB/c mice were kept 10 cm apart in a 25-cm3 cage and exposed to single flies . When a fly began feeding on a mouse it was allowed a further minute before it was interrupted . Following this , the fly was interrupted every 10 s if it tried to remain feeding on the same mouse . Feeding interruption was carried out until either the fly initiated a feed on the opposite mouse , or until 15 min elapsed . Flies were prevented from feeding by gentle brushing of the antennae between feeding attempts . Flies were dissected after the assay to confirm infection where appropriate . Female BALB/c mice were infected either by injection of 103 metacyclic promastigotes via needle into the dorsal surface of the right hind foot or by allowing single infected sand flies to bite at the same site [7] . Lesion development was monitored by measuring the swelling of the right foot with Vernier callipers and subtracting the width of the contralateral uninfected foot . At the end of experiments , mice were humanely euthanized , and parasite burdens in the feet determined either by direct counting via hemocytometer or by limiting dilution . All procedures involving animals were approved by a local Animal Welfare Committee and performed in accordance with United Kingdom Government ( Home Office ) and EC regulations . Sequential measurements for each individual mouse were used to calculate the area under the curve for a plot of lesion thickness against time [36] . The distribution of values did not show evidence of non-normality using the Shapiro-Wilk test and therefore parametric analysis was performed ( t tests ) . The null hypothesis was rejected if p < 0 . 05 . Statistical analysis was performed using the StatsDirect software package version 2 . 3 . 1 . | Leishmaniasis is a human parasitic disease that can result in destructive skin lesions or sometimes in fatal visceral involvement if left untreated . Infections are acquired via the bites of female sand flies , small blood-feeding insects . In this experimental study using mice , we show that Leishmania parasites can manipulate the behavior of sand flies so that infected flies become more tenacious , returning to feed if interrupted more readily and delivering more bites than uninfected flies . This change in behavior occurs only when the parasites have become infective to humans and are ready for transmission . At the same time , the parasites in the sand fly secrete a viscous gel that blocks up the gut and mouthparts and interferes with feeding . The result of this manipulation is an increase in the severity of disease and number of infections . These findings help us to better understand how human infection occurs , which is important when trying to devise ways of controlling this disease . It is currently estimated that at least 12 million people have leishmaniasis , with 2 million new infections every year . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases",
"none",
"(leishmania)",
"microbiology",
"mus",
"(mouse)",
"insects"
] | 2007 | Leishmania Manipulation of Sand Fly Feeding Behavior Results in Enhanced Transmission |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.