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
|
|---|---|---|---|---|---|
The Tsx gene resides at the X-inactivation center and is thought to encode a protein expressed in testis , but its function has remained mysterious . Given its proximity to noncoding genes that regulate X-inactivation , here we characterize Tsx and determine its function in mice . We find that Tsx is actually noncoding and the long transcript is expressed robustly in meiotic germ cells , embryonic stem cells , and brain . Targeted deletion of Tsx generates viable offspring and X-inactivation is only mildly affected in embryonic stem cells . However , mutant embryonic stem cells are severely growth-retarded , differentiate poorly , and show elevated cell death . Furthermore , male mice have smaller testes resulting from pachytene-specific apoptosis and a maternal-specific effect results in slightly smaller litters . Intriguingly , male mice lacking Tsx are less fearful and have measurably enhanced hippocampal short-term memory . Combined , our study indicates that Tsx performs general functions in multiple cell types and links the noncoding locus to stem and germ cell development , learning , and behavior in mammals .
The X-inactivation center ( Xic ) controls X-chromosome inactivation ( XCI ) and is enriched for genes that produce long noncoding RNAs ( ncRNA ) ( Figure 1A ) [1] , [2] , [3] . Several have been shown to regulate XCI . The 17-kb Xist RNA is induced at the onset of XCI and silences the X-chromosome as the RNA spreads in cis along the chromosome . Xist expression is controlled both negatively and positively [4] , [5] , [6] , [7] , . Xist's antisense partner , Tsix , controls X-chromosome counting and allelic choice , blocking Xist induction on the future active X [8] , [9] , [11] . The upstream locus , Xite , is required to sustain Tsix expression on the future active X during XCI [12] , [13] . Two other noncoding RNAs , Jpx and RepA , function as activators and are required for transcriptional induction of Xist RNA [7] , [14] . Together , these five noncoding genes span <200 kb of sequence . Several studies have suggested that additional regulators of XCI reside in close proximity to this core domain of the Xic [5] , [15] , [16] . Located immediately upstream of Xite is Tsx ( Testes-specific X-linked ) , a gene of some interest owing to its unique location and evolutionary history [16] , [17] , [18] . Tsx is absent in marsupials and arose some 150 million years ago in eutherians during the transition from imprinted to random XCI in mammals . Although absent in marsupials , Tsx partially aligns to genomic sequences in chickens , specifically within three exons of a coding gene of unknown function called Fip1l2 [18] . This finding suggests that vertebrate Fip1l2 is the evolutionary precursor of eutherian Tsx . Because Tsx is only partially conserved among eutherian mammals , it is believed to be a pseudogene in some species such as human , cow , and dog [15] . Nevertheless , human TSX shares significant homology to mouse Tsx , particularly within exons 1 , 3 , 4 , 5 , and 6 [15] . Several Tsx splice variants have been reported in the mouse , including a predominant species of 794-nt transcript with a 432-nt ORF predicted to encode a highly acidic protein of 156 amino acids [17] . Two ORF-less species of 351 and 540 nt have also been described . Curiously , immunostaining with anti-sera raised against Tsx protein suggested exclusive expression in pre-meiotic germ cells in pubertal mouse testes , though Tsx mRNA is not detected until meiosis [17] . In adult animals , Tsx mRNA is predominantly observed in testis [16] , specifically in Sertoli cells [17] . Tsx RNA is not detectable in the female germline , but can be seen in 2-cell mouse embryos and throughout preimplantation development [19] . These previous analyses therefore suggest a male germline protein originating from the Xic , whose expression coincides with or precedes the meiotic period . With respect to XCI , meiotic mechanisms in the male have been proposed to guide imprinting of the paternal X-chromosome in daughter embryos and to thereby direct paternal-specific silencing of X in the early embryo – a form of XCI known as “imprinted XCI” [20] , [21] , [22] , [23] , [24] . Here we examined the idea that Tsx may function during imprinted XCI . Contrary to expectation , generation of a Tsx mouse knockout showed no effects on imprinting . However , in the course of analysis , we learned that Tsx is actually noncoding and uncover several functions for Tsx in mouse development and behavior .
Together with Tsx's location within the RNA-enriched Xic , several observations led us to revisit the question of whether Tsx is a coding gene . First , Tsx's genic ancestry mirrors that of the well-established Xist RNA , which is also proposed to have evolved by pseudogenization of a coding gene ( Lnx3 ) [18] . Second , between mouse and rat , the Tsx cDNAs are 79% identical , yet the 5′ UTR unexpectedly display significantly greater homology ( 89% identity ) [16] , [17] , indicating conservation of noncoding elements within the gene . Third , Tsx lacks a Kozak sequence for translation initiation . Finally , immunostaining with an antiserum against a Tsx peptide produced a tissue staining pattern inconsistent with its RNA profile [17] . Taken together , these findings suggest that Tsx RNA may not be translated . Indeed , Tsx protein has never been isolated from cells , and our continuous search for Tsx peptides has not identified matches in the extensive mouse or human proteomic databases to date ( http://world-2dpage . expasy . org/repository/; http://reprod . njmu . edu . cn/cgi-bin/2d/2d . cgi ) . To investigate the protein-coding potential of Tsx , we cloned the putative Tsx open reading frame ( ORF ) ( 432-nt sequence generating a predicted protein of ∼20 kDa ) in frame with a C-terminal V5 tag in a mammalian expression vector and tested whether it could be translated in a cell-free translation system ( Figure 1B ) . A peptide corresponding to the predicted protein was never observed . Similar results were observed for a hypothetical ORF within the established noncoding RNA , Xist ( ORF of 468 nt located 2 kb downstream of RepA ) . By contrast , ß-galactosidase ( ß-gal ) and luciferase proteins were consistently produced . These results indicated that the Tsx ORF is not translated in vitro . To determine whether the Tsx ORF could be translated in vivo , we transiently transfected human 293T cells and mouse ES cells and then performed Western blot analyses using anti-V5 antibodies . Again , no proteins of predicted size ( ∼20 kDa ) were detected for Tsx and Xist in either cell type , whereas ß-gal was readily detected ( Figure 1C ) . To exclude the possibility that the C-terminal V5 tag was masked and therefore undetectable by Western analysis , we cloned the Tsx and Xist ORFs ( lacking stop codons ) into an expression vector containing an N-terminal HA tag and a C-terminal red fluorescence protein ( RFP ) fusion for live-cell detection of potential protein products ( Figure 1D ) . To increase transcript stability and enhance translation , we also cloned in a 3′ WPRE and a 5′ Kozak sequence , respectively . [Note: Both the Tsx and Xist lack Kozak sequences [16]] . Following transient transfection into 293T cells , RFP was detected from neither the Tsx- nor Xist-transfections ( Figure 1E ) . By contrast , RFP was easily detected from positive controls in which RFP is fused to either GFP or ß-gal . Western blot analysis using an anti-HA antibody confirmed expression of ß-gal∶RFP and GFP∶RFP fusions , but not Tsx∶RFP or Xist∶RFP . We conclude that Tsx does not produce a protein and its transcription produces a long noncoding RNA . We next examined Tsx expression patterns in vivo . Quantitative RT-PCR ( qRT-PCR ) analysis showed that Tsx is widely expressed in adult tissues ( Figure 2A ) . In female mice , highest levels were found in the brain . In male mice , the RNA was expressed at 8-fold lower levels in brain relative to that in female brain . Furthermore , whereas the female gonadal levels were unremarkably low , male gonadal expression was 10- to 100-times higher than in the male brain . High-level expression was generally observed for tissues of the male reproductive tract ( e . g . , epididymus , prostate ) . The testes nonetheless exhibited greatest expression , consistent with previous reports [17] . We then isolated male germ cells from the testes and fractionated them by meiotic stage ( Figure 2B ) . qRT-PCR showed that the premeiotic cell types , Type A and B spermatogonial cells ( AS , BS ) , displayed relatively low-level expression . Expression increased 40-fold during meiosis , specifically in pachytene-stage spermatocytes ( PS ) , suggesting de novo transcriptional induction of Tsx RNA during this stage . Tsx expression then decreased 9-fold in round spermatids ( RS ) , the earliest post-meiotic stage cells , but steady state levels remained elevated relative to AS and BS stages . RNA at this stage could represent either continued but decreased expression of Tsx RNA or retention of RNA synthesized during the pachytene stage . The massive upregulation of Tsx during pachytene is intriguing , given that the sex chromosomes are inactivated during this stage in a process known as ‘meiotic sex chromosome inactivation’ ( MSCI ) [25] . A previous report of X-linked expression during spermatogenesis revealed that only a handful of genes escaped transcriptional silencing during MSCI [24] . Significantly , Tsx is one of the loci that escapes . To determine its subcellular localization pattern , we separated pachytene spermatocytes and round spermatids into nuclear and cytoplasmic fractions . qRT-PCR demonstrated that Tsx RNA could be found in both compartments during pachytene , but it appears to be predominantly nuclear in round spermatids ( Figure 2C ) . Previous analysis showed the Tsx RNA can occur in 794- , 540- , and 351-nt isoforms [17] . To determine which splice variant occurs in each compartment , we cloned and sequenced cDNAs from the nucleus and cytoplasm , and identified more variants than previously observed , with the variations occurring in exon 1 and upstream exons ( Figure 2D; downstream exons appear similar; The Tsx alternative transcript ( Tsx RPS ) originally described as being tissue-specific [17] appeared to be expressed in all cell types , including in male germ cells . We conclude that Tsx RNA exists as multiple RNA species in pachytene spermatocytes . To investigate Tsx function , we generated Tsx knockout mice ( TsxKO ) by homologous targeting in ES cells , blastocyst injection , and production of chimeric mice for germline transmission . The TsxKO allele deletes a 2 kb region encompassing exon 1 and the upstream regulatory region ( Figure 3A ) . Following electroporation into male ES cells , three homologously targeted clones were isolated ( Figure 3B ) and clone 3A8 was used for blastocyst injection . Two chimeric males were obtained and both gave germline transmission of the Tsx2loxNeo allele . We mated Tsx2loxNeo offspring to FLP-expressing mice to remove the Neomycin selection marker ( Figure 3C ) and then crossed the resulting Tsx2lox offspring to Cre-expressing mice to generate TsxKO ( Figure 3D ) . We verified loss of all Tsx RNA isoforms by Northern blot analysis using full-length cDNA probes ( Figure 3E ) and also by RT-PCR analysis using exon-specific primers across all exonic sequences ( data not shown ) . Intercrosses between TsxKO mutant mice ( −/Y , −/+ ) and outcrosses to wildtype C57BL/6J generally produced offspring at expected Mendelian ratios and normal mean litter sizes of ∼7 ( Figure 3F and data not shown ) . All offspring were viable and fertile . However , when −/− females were crossed to −/Y males , we observed a mild but statistically significant sex-ratio distortion favoring females and a lower mean litter size ( Figure 3F ) . Similar findings were observed when −/− females were outcrossed to wildtype males . Intercrossing −/Y males to wildtype females did not reproduce these effects . These results suggested that −/− females have mildly reduced fertility and that female births are slightly favored . The fertility of −/Y males was somewhat surprising , given high-level Tsx expression in wildtype pachytene spermatocytes . We therefore asked if there were measurable consequences of deleting Tsx in the male germline . We isolated testes from littermate animals ( 6 months ) and found that testes from mutant males were smaller than wildtype littermates ( Figure 4A ) . Histological examination of testicular sections revealed no gross abnormalities at 7d , 14d , 2 month , and 1 year ( data not shown ) . To test whether the smaller testis size was due to loss of germ cells , we performed TUNEL staining of paraffin-embedded testicular sections and found elevated apoptosis between mutant and wildtype animals at 7 days , 14 days , and 2 months ( Figure 4B , 4C ) . The difference was greatest at 14d , coinciding with the first wave of pachytene ( meiotic prophase I ) during post-natal male development [26] . At this timepoint , there were five times as many apoptotic germ cells in the −/Y testes . Significant differences were also observed as early as 7d at a time when male germ cells prepare to enter the first wave of meiosis , and also at 2 months when male animals reach adulthood . At 1 year , the difference became insignificant . These results suggested loss of germ cells as the animals first proceed through meiosis . To determine which cell type and what stage of meiosis were affected , we combined TUNEL staining and immunofluorescence for SCP1 ( synaptonemal complex protein 1 ) , a combination that enabled us to distinguish between meiotic and non-meiotic cells and also between meiotic cells of different stages . Tubule sections from 14d mutants revealed that TUNEL-positive cells were located 2 to 3 cell layers in from the basal lamina membrane ( most mature germ cells are found in the center of a tubule cross-section ) and all were SCP1-negative . The TUNEL-positive cells resided in the same cell layer as pachytene spermatocytes and were always positioned next to pachytene spermatocytes that showed robust staining with SCP1 ( Figure 4D ) . Co-staining with Stra-8 [27] showed no overlap between Stra-8-positive and TUNEL-positive cells in 14d tubule sections , thus ruling out pre-meiotic germ cells ( Figure 4E ) . These findings argued that the apoptotic cells were in fact pachytene spermatocytes whose chromatin had become too fragmented for SCP1 staining . Examination of tubules from 2-month old mutant mice showed similar staining patterns ( data not shown ) . To determine if abnormal MSCI could be the cause of germ cell loss , we carried out immunofluorescence with antibodies against H3-K9me3 , HP1-γ , γ-H2AX , and HP1β – chromatin modifications that decorate the XY body during MSCI or post-meiotic sex chromatin ( PMSC ) in round spermatids [24] . No obvious differences between wildtype and mutant cells could be seen ( data not shown ) . We conclude that a subset of pachytene-stage cells inappropriately undergoes apoptosis when Tsx is deleted , but the cause of sub-viability is not clear at present . Because Tsx is expressed at the 2- and 4-cell stages , and also in blastocyst embryos [19] , we examined whether it is also expressed in wildtype ES cells . qRT-PCR revealed that Tsx is expressed in both undifferentiated male and female ES cells and that it is downregulated during differentiation ( Figure 5A ) . Female ES cells consistently demonstrated 5- to 10-times more expression than male ES cells . Given Tsx's expression in ES cells , we next asked whether deleting Tsx affects ES cell growth and differentiation . We derived Tsx−/Y and −/− ES cells from blastocysts resulting from −/−×Tsx−/Y matings and obtained two independent female clones and four independent male clones . The growth rates of undifferentiated Tsx−/Y and −/− ES cells were significantly affected when compared to wildtype cells for multiple independent clones tested ( Figure 5C ) . Similar growth retardation was observed for both mutant male and female clones when they were differentiated into embryoid bodies ( EBs ) . In general , mutant EBs were much smaller during suspension culture ( d0–d4 ) . They also attached poorly to plates during the adherent growth phase ( d4–d8 ) and showed sparse outgrowth compared to wildtype controls ( Figure 5B ) . Although increased cell death was observed in some clones by quantitative assays based on cytotoxicity measurements , cell death could not have been the sole cause of poor growth , as only some male and female mutant clones demonstrated significantly higher cell death ( data not shown ) . Because Tsx resides at the Xic and is immediately proximal to Xite and Tsix , we asked if aberrant XCI may be a cause of the ES cell anomalies . Standard assays for X-inactivation choice did not reveal any skewing of allelic choice ( data not shown ) . However , combined RNA/DNA fluorescent in situ hybridization ( FISH ) showed that Xist RNA was aberrantly upregulated in a small fraction of cells during cell differentiation ( Figure 6A ) . This was observed in both male and female ES cells and in multiple independent clones . Whereas wildtype male cells almost never upregulate Xist RNA , Tsx-deficient male cells showed Xist clusters in 3–5% of differentiating cells on d3 . Similarly , whereas wildtype female cells usually only showed one Xist RNA cloud during differentiation , mutant female cells displayed two RNA foci in 5–10% on d2 . To avoid counting tetraploid cells ( which would have 4 X-chromosomes and appropriately show two Xist RNA clouds ) , Xist RNA FISH was combined with an Smcx BAC DNA FISH to determine X-chromosome number ( Figure 6A ) . Only diploid cells were scored in this assay . It should be noted that in both male and female mutants , the ectopic Xist cluster was relatively loose and generally not as large as those typically seen on the inactive X . We conclude that deleting Tsx partially affects Xist regulation in ES cells . Given the ectopic Xist expression and because Xist is known to be regulated by Xite and Tsix , we next asked if Tsx's effect on Xist may occur through Tsix and Xite by examining expression of Tsix and Xite RNAs in two independent male and female TsxKO ES cell lines . Interestingly , whereas the knockout had little to no significant effect on Xite expression , it caused a significant downregulation of Tsix expression at all timepoints and in both male and female cells ( Figure 6B ) . These results suggest that a Tsx knockout may induce ectopic Xist expression by blunting the upregulation of Tsix during ES cell differentiation . Thus , Tsx may be a positive regulator of Tsix . The observation that Tsx expression in somatic tissues was highest in brain suggested that Tsx may regulate neural processes underlying behavior . Knockout animals displayed no gross neuromuscular defects or disturbances in gait . As a first step toward assessing a possible role for Tsx in cognitive function , we therefore subjected Tsx knockout mice and littermate controls to a series of well-established behavioral paradigms . Exploratory behavior was first analyzed in the open field task , which is sensitive to changes in locomotor activity , stereotypy and anxiety [28] . Tsx knockout and age-matched wildtype littermates displayed similar values for total distance and ambulatory counts , indicating that Tsx deletion does not affect locomotor activity ( Figure 7A , 7B ) . Mice typically explore the periphery of the chamber and avoid the center . Although Tsx knockout mice spent significantly more time at rest in the chamber's central zone ( P<0 . 05; Figure 7C ) , they displayed a normal overall preference for the peripheral relative to the central zone in terms of total occupancy ( Figure S1A ) . To investigate further the possibility of altered anxiety levels , we examined the behavior of Tsx knockout and wildtype littermates in the elevated plus maze task [29] , [30] . The total occupancy and number of entries in the open and closed arms of the maze were similar between knockout and wildtype animals ( Figure S1B , S1C ) , arguing against any generalized changes in anxiety-related behavior as a result of Tsx deficiency . To explore a possible role for Tsx in cognitive function , we next analyzed associative learning and memory for conditioned fear . In contextual fear conditioning , mice learn to associate a particular experimental chamber or “context” with a mild foot shock following a single brief training session , such that subsequent exposure to the context elicits a fear response measured by immobility or “freezing” . In cued fear conditioning , mice learn to associate an experimental tone with a mild foot shock , thereafter displaying a fear response when presented with the tone in a novel context . Contextual fear memory requires both the hippocampus and amygdala , whereas cued fear memory is amygdala-dependent but hippocampus-independent [31] , [32] . Assessment of fear responses at retention intervals of 1 hour and 24 hours after training distinguishes short-term memory and long-term memory for conditioned fear [33] , [34] . Age-matched male Tsx knockout and wildype littermates were subjected to contextual fear conditioning then tested for the memory of the experimental context at 1 hour and 24 hours following the conditioning . Prior to conditioning , knockout and wildtype animals exhibited similarly low baseline levels of freezing , supporting the conclusion from the open field task that Tsx deletion does not affect locomotor activity ( Figure 7D ) . Both groups of mice also displayed similar freezing responses to presentation of the tone and foot shocks during the training sessions . Interestingly , knockout mice exhibited significantly higher levels of freezing when tested at a retention interval of 1 hour ( Figure 7E; wildtype 35% , mutant 53% , p = 0 . 059 ) . This difference was not observed at a retention interval of 24 hours , indicating that knockout mice display a selective enhancement of short-term contextual memory . No differences were detected between knockout and wildtype mice for cued fear conditioning ( Figure 7F ) . Behavioral tests are typically carried out in male mice because estrus cycles in female mice are difficult to synchronize and are known to affect testing . However , because Tsx levels were higher in brain tissue of female mice , we made an effort to test female knockouts . In attempts to control for estrus cycling , female animals were group-housed ( 4 animals per cage ) for 2 weeks prior to testing ( note: co-habitation partially synchronizes estrus cycles; we could not synchronize pharmacologically because exogenous hormones have profound effects on behavior ) . We found no differences between wildtype and Tsx−/− female mice in either the open field test or the fear conditioning tests ( Figure S2 and data not shown ) . There are two potential explanations for this result . First , Tsx may indeed have no effect on female behavior . Second , effects of deleting Tsx may have been masked by fluctuating estrus cycles . From the collective evidence , we conclude that Tsx deletion likely affects specific enhancement of short-term hippocampal memory , at least in male mice .
Here we have described a novel long ncRNA expressed from the Tsx locus , a gene that was previously thought to encode a protein . Tsx RNA is expressed at varying levels from many tissues . Multiple isoforms of the RNA can be observed , with the predominant species being the one encoded by all seven exons , regardless of cell type . The greatest amount and isoform diversity are seen in pachytene spermatocytes , a cell type in which the X-chromosome from which Tsx is expressed is otherwise transcriptionally inactive . Indeed , Tsx is only one of four genes and several microRNAs that escape transcriptional silencing during MSCI in the male germline [24] , [35] . Surprisingly , knocking out Tsx neither affected MSCI nor male fertility . It also did not affect imprinted XCI . However , Tsx mutants showed a number of reproductive defects , including pachytene-specific apoptosis , smaller testes size , mildly reduced litter sizes and a mild sex ratio distortion related to maternal Tsx deficiency . The subtlety of some of these defects may be attributed to the possibility that Tsx is functionally redundant with regards to its meiotic role during male and female germ cell development . Ex vivo , the mutants exhibit poor stem cell growth and both male and female ES cells display aberrant Xist upregulation in a small fraction of cells . We suggest that this may be due to downregulation of Tsix , a known repressor of Xist . Thus , Tsx may be a positive regulator of Tsix and consequently an indirect repressor of Xist . Further investigation is required to pinpoint the mechanism . Aberrant Xist expression may in part explain the poor viability during cell differentiation , but we do not believe that XCI consequences are the sole cause , as the level of ectopic Xist expression is not high . Strangely , in spite of these measurable defects in ES cell growth , knockout male and female embryos are viable , with only a slight sex ratio distortion and reduction in litter size . Tsx mutants must be able to compensate in vivo for the stem cell-related defects seen ex vivo . In vivo , the growth of the epiblast lineage in the ICM niche must provide what culture media cannot . The possibility of functional redundancy with other stem cell regulators can also be entertained here . A most intriguing phenotype to arise from the mutants is the effect on behavior and cognition . We were prompted to examine a possible role for Tsx in brain function based on observation that brain exhibits the highest levels of Tsx expression among somatic tissues examined . Interestingly , our analysis revealed enhanced short-term memory for contextual fear conditioning in Tsx knockout male mice . [No effects were seen in female mice , but the results may have been confounded by varying estrus cycles in the subjects] . Since assessment of learning and memory in fear conditioning depend on fear responses measured by immobility , it was important to consider any potentially confounding effects of Tsx deletion on locomotor activity , anxiety or differential responses to the conditioning regimen . The normal behavior displayed by knockout mice in the open field , elevated plus maze and fear conditioning training sessions argues that Tsx deletion causes a specific alteration in hippocampal memory . Some noncoding RNAs in the brain ( e . g . BC1 , BC200 ) regulate gene expression post-transcriptionally by binding to mRNAs and repressing their translation [36] , [37] . While the specific regulatory functions of Tsx are presently unknown , it is tempting to speculate that Tsx deletion may lead to constitutive derepression of the expression of gene products required for memory acquisition . Further studies will be needed to elucidate the molecular mechanisms by which Tsx may regulate learning and memory in the mammalian brain .
The Tsx ORF ( 480 bp; consisting of 48 bp before the ATG start codon to the position before the stop codon ) was cloned from a mouse testes cDNA library and amplified using primers Tsx cDNA F2 5′-AGCACCCACCTAGACTTGGG-3′ and Tsx exon 7 no stop R 5′-ATCAGTTGGGTTCATGGCAC-3′ . Two different ORFs within the mouse Xist gene were cloned . The Xist ORF lacking a stop codon found within exon 1 ( 469 bp ) was amplified using primers 5′-ATGCTCTGTGTCCTCTATCAGA-3′and 5′-GAAGTCAGTATGGAGGGGGT-3′ , and the Xist ORF ( 489 bp; also lacking a stop codon ) within exon 7 was amplified using primers 5′-ATGTTCTCCTGCATGTTCT-3′ and 5′- GAATACAAGAGAGACACAGA-3′ . The Tsx and Xist exon 1 ORFs were cloned in frame with the V5 epitope of the pEF1/V5-His vectors ( Invitrogen ) . These constructs contain a T7 promoter upstream of the ATG start codon , and 1 µg of each construct was used with the T7 TNT Rabbit Reticulocyte Lysate In-vitro Transcription/Translation kit ( Promega ) , and 1/5 of the reaction mix was run on a 12% SDS-PAGE gel . In addition , the Tsx ORF , Xist exon 7 ORF , ß-galactosidase cDNA , and GFP cDNA ( all lacking stop codons ) were cloned in frame with RFP into the pPS-EF1-LCS-T2A vector ( System Biosciences ) . Mouse ES cells and human 293T cells ( 50% confluent in a 10 cm plate ) were transfected with 24 µg of the various DNA constructs using Lipofectamine 2000 according to the manufacter's protocol ( Invitrogen ) , and cells were harvested 48 hours later . The cell pellets were sonicated three times ( 10 seconds each ) in a lysis buffer containing 10 mM Trish pH 7 . 5 , 150 mM NaCl , 5 mM EDTA , 1% Triton X-100 , 20 mM 2-mercaptoethanol , and 1 mM PMSF . Protein lysates ( 25 µg/lane ) were run on a 10% or 12% SDS-PAGE gel , then transferred to a PVDF membrane . The membrane was blocked with 5% non-fat dry milk in PBS containing 0 . 1% Tween-20 , then incubated with either a mouse monoclonal ( Invitrogen ) or rabbit polyclonal ( Novus Biologicals ) V5 epitope antibody ( diluted 1∶5 , 000 ) or HA epitope antibody ( diluted 1∶10 , 000 ) at 4°C overnight . The next day the membrane was washed four times with PBS/Tween , then incubated with anti-rabbit horseradish peroxidase-labeled secondary antibody ( diluted 1∶20 , 000 ) for 1 hour at room temperature . The membrane was washed again with PBS/Tween , then visualized using the West Pico Chemilluminescence system ( Pierce ) . For the fractionation experiments , cells were resuspended in lysis buffer ( 10 mM Tris pH 8 . 4 , 1 . 5 mM MgCl , 140 mM NaCl ) , then 4 µL of 5% Nonidet P40 ( diluted in PBS ) was added on ice . Aliquots were taken at 5 min , 10 min , 15 min , and 30 min , and centrifuged for 3 min at 2400 rpm . The supernatant ( cytoplasm ) was transferred to a new tube and Trizol-LS was added . The pellet ( nuclei ) was washed 2× with PBS , then Trizol was added . Germ cells from testes of 7d C57/BL/6J ( for isolation of Type A and B spermatogonia ) and 2 month old CD-1 males ( for isolation of pachytene spermatocytes and round spermatids ) were isolated using gravity sedimentation with the STA-PUT device as described previously [24] , [26] . PS and RS fraction purtity was >95% as judged by phase optics ( for qRT-PCR in Figure 2C ) ; RNA for qRT-PCR of Tsx in Figure 2B was also used in Reference 24 . RNA from the fractionated cell populations was isolated using Trizol ( Invitrogen ) . Tissues from 2–4 different wildtype male and female C57BL/6J animals ( 2 months old ) were immediately placed in Trizol after dissection , then homogenized using a Qiagen TissueLyser , and the RNA was cleaned up using the RNeasy Mini kit ( Qiagen ) . All RNA was treated with Turbo DNAse ( Ambion ) following the ‘rigorous’ protocol . Reverse transcription of RNA to cDNA was performed using SuperScript III ( Invitrogen ) , 1 µg RNA and random hexamers . All cDNA was diluted 1∶5 prior to qPCR , except for cDNA from male and female animal tissues , which was not diluted . For real-time PCR analyses , Tsx cDNA ( exons 1–7 ) was amplified using two primer sets: Tsx B10 and Tsx B11 [17] and Tsx 1 5′-ATTAAGCAGGCAGGCAGAAA-3′; Tsx 2 5′-TGCGGTGATTTTCATTTTGA-3′ . For transcript amplification in PS ( Figure 2D ) we used a reverse primer ( Tsx 3 . 5L; within exon 4 ) and 4 different forward primers ( RPS-F , RPL-F , RPLb-F , RPLa-F; upstream of exon 1 ) . ß-actin primers: 5′- CCGTGAAAAGATGACCCAG-3′ ( F ) and 5′-TAGCCACGCTCGGTCAGG-3′ ( R ) . Tsx primers: 3 . 5L 5′-AGCTTGGCAAGTGTCCTC-3′; RPS-F 5′-TACCCTAGCTGAAGGAAAAT-3′; RPL-F 5′- ATGGTTGGAAGATCTAATACCT-3′; RPLa-F 5′-CAACCACTGTCCCCTTCCTA-3′; and RPLb-F 5′-CACCCCAGCAGAGAGAAAAG-3′ . LINE-1 primers: 5′-GTCTGGTGTTTGGACCTCCT-3′ ( F ) ; 5′-CCGACATGTACGACTCCAGA-3′ ( R ) . Standard PCR reactions for Tsx were cycled for 30–32 cycles . Real-time PCR was performed on a Bio-Rad iCycler maching with SYBR-green iQ Mix ( BioRad ) . Standard curves were generated via amplification of 10-fold plasmid serial dilutions , and ß-actin was used for normalization . Tsix was amplified with primers oNS18 and oNS19 [13] and Xite was amplified with primers NGP3 and NGP4 [12] . Values were normalized to expression of ß-actin . The Tsx targeting construct was generated by PCR amplification of three segments of the Tsx gene and upstream region , verified by DNA sequencing , then cloned into a modified version of the ploxP-2FRT-PGKneo vector ( a gift from Dr . David Gordon via the University of Michigan Transgenic Animal Model Core Facility ) that contained a second loxP site between KpnI and BamHI . The upstream Tsx fragment ( a 4 . 8 kB section located 6 . 6 kB upstream of exon 1 ) was cloned into the EcoRI site , the ‘middle’ fragment ( 1 . 8 kB upstream of exon 1 and 160 bp of intron 1 ) was cloned into the BamHI and SalI sites , and the ‘last’ fragment ( intron 1 through intron 4 of Tsx ) was cloned into the XhoI site . In summary , a 2 . 1 kB region encompassing the predicted Tsx promoter region upstream of exon 1 and 160 bp of intron 1 was flanked by loxP sites for Cre-mediated deletion of the Tsx gene . The neomycin selection marker ( under control of the PGK promoter ) was flanked by FRT sites , and was used for positive selection ( 300 µg/mL ) of ES cell clones . A novel PacI site was introduced at the end of the last fragment for linearization prior to electroporation . Male TC1 ES cells ( derived from 129S6/SvEvTac mice ) were electroporated with 20 µg of linearized Tsx targeting construct DNA , and 400 neomycin resistant clones were picked . Genomic DNA was isolated from these clones , and digested with MscI for screening via Southern blotting using an external 1 kB probe ( Tsx7 ) overlapping exon 5 . Tsx probe 7 was generated using primers Tsx 7 F 5′-GCCTCCACTAGCACATGACA-3′ and Tsx 7 R 5′-CCCTCAGTCCTGCCTCTACC-3′ . Three positive clones , containing just one integration of the targeting construct , were obtained , and clone 3A8 was selected for C57BL/6J blastocyst injection at the Brigham & Women's Hospital Transgenic Mouse Facility ( Boston , MA ) . Two chimeric males were obtained following injection , and both transmitted the construct through the germline by matings to C57BL/6J females . Brown female pups were genotyped by Southern blot then mated to male animals expressing FLPe under control of the ROSA26 promoter ( 129S4/SvJaeSor-Gt ( ROSA ) 26Sortm1 ( FLP1 ) Dym/J; Jackson Labs ) mice to remove the neomycin selection marker . Animals were genotyped by Southern blot , digesting DNA with SphI and using an internal 1 kB probe ( Tsx L2 ) , located between exons 3 and 4 . Tsx probe L2 was generated using primers Tsx L2 F 5′-ACATCCCCCATGAAAACTGA-3 and Tsx L2 R 5′-ACCAAAACCAAAACCCAACA-3′ . Positive animals were then mated once to C57BL/6J , pups were genotyped using SphI and probe Tsx L2 , and positive animals were selected for mating with animals expressing Cre-Recombinase under control of the adenovirus EIIa promoter ( B6 . FVB-Tg ( EIIa-cre ) C5379Lmgd/J; Jackson Labs ) . Following Cre-mediated deletion of Tsx , animals were genotyped by Southern blot using BstZ17I digestion and probe Tsx L2 . Positive animals were then outcrossed to C57BL/6J animals for a total of 6 generations . All animals were weaned at 4 weeks , then tagged and tailed for DNA isolation for genotyping . RNA ( 20 µg/lane ) isolated from adult tissues was separated on denaturing 1% agarose-formaldehyde gels for 3 hours at 100V , then transferred to a Hybond-XL membrane overnight . The membrane was pre-hybridized using UltraHyb or buffer ( Ambion ) for 1 hour at 45°C , followed by overnight hybridization of the 32P-radiolabeled probe . The radiolabeled Tsx cDNA probes ( both anti-sense and sense orientations consisting of Tsx exons 1–7 ) , and the ß-actin probe were generated by linear amplification using the Ambion Strip-EZ kit and PCR primers Tsx cDNA F2 ( sense strand ) , Tsx cDNA R2 5′-ATTGGAAGTTTGGCAAGCAA-3′ ( for antisense strand ) , and ß-actin R . The following morning , the membranes were washed twice with low-stringency buffers ( 2×SSC/0 . 1% SDS ) followed by two washes with high-stringency buffers ( 0 . 1×SSC/0 . 1% SDS ) , at 45°C . The membranes were stripped according to the kit instructions then probed for ß-actin . The membranes were visualized by exposing to a phosphoimaging screen . The male testes blot was exposed for 2 hours , and the male lung blot was exposed for 1 day . Histological analysis was carried out on 4% paraformaldehyde-fixed testes that were paraffin-embedded , then sectioned at 5 µM thickness . Testis sections were deparaffinized using two changes of Histoclear ( National Diagnostics ) and hydrated to water by successive 2 min washes in 100% ethanol , 90% ethanol , 80% ethanol , 70% ethanol , and distilled H2O . Slides were then incubated in 10 mM sodium citrate pH 6 . 0 at 100°C for 20 min , followed by 20 min incubation at room temperature . Following unmasking , slides were washed twice in PBS . The TUNEL mixture ( In Situ Cell Death Detection kit , Fluorescein; Roche ) was incubated for 60 min at 37°C . The slides were washed three times in PBS for 5 min , then mounted with Vectashield mounting media containing DAPI . For immunofluorescence co-staining with SCP1 and Stra-8 antibodies , slides after TUNEL were blocked with 5% BSA in PBS-Tween-20 for 20 min at room temperature . The primary antibody was added ( SCP1 at 1∶100 dilution; Stra-8 at 1∶400 dilution ) and slides were incubated overnight at 37°C . The next morning the slides were washed three times in PBS-Tween-20 then incubated with a 1∶500 dilution of goat anti-rabbit Cy3 antibody for 30 min at room temperature . Slides were washed three times in PBS-Tween-20 then mounted with Vectashield containing DAPI . Tsx KO male and KO/KO female animals ( ages 8–12 weeks ) were naturally mated and females were sacrificed at 3 . 5dpc . Blastocysts were flushed out of the uterine horns and plated on gelatinized 15 cm plates containing mouse embryonic fibroblasts ( MEFs ) as described in [38] . The inner cell mass ( ICM ) was dissected five days later , trypsinized in a droplet , and plated onto a fresh well of MEFs . The cells were passaged and expanded until there was sufficient numbers to culture in a T25 flask ( a total of 3–4 passages after ICM dissection ) . The Tsx KO cell lines were genotyped for gender using Zfy , NS18 , and NS19 primers as described previously [39] . For differentiation experiments , ES cells were typsizinized and one million cells were plated ( in triplicate ) in petri dishes in ES medium lacking LIF , using the EB suspension method described previously [8] . Tsx KO cell lines of the same passage number ( spanning passage number 6 , 7 , 8 , 9 ) were used for each differentiation experiment , and four independent differentiations were performed . The medium was changed every two days , and cellular cytotoxicity and viability was determined for 100 , 000 cells at each time point using the MultiTox-Fluor Multiplex Cytotoxicity Assay ( Promega ) . Cell growth of undifferentiated ES cells was determined by plating 150 , 000 cells per well ( in triplicate ) of a 12-well gelatinized plate containing MEFs . The medium was changed daily , and cells were harvested at different time points by trypsinization and counted using a Cellometer ( Nexcelom Bioscience ) . Spontaneous locomotor activity was monitored using a MED-OFA-MS open field test system ( Med Associates , St . Albans , VT ) . The animal was placed in the center of the activity-field arena , which is a transparent Plexiglas cage ( W×D×H; 27×7×20 cm ) equipped with three 16 beam infrared transmitter and sensor arrays to register horizontal and vertical activity . Ambient conditions included moderate levels of illumination and white noise ( 800 lux and 40 dB , respectively ) . The mouse's position and movement is monitored continuously in the horizontal and vertical planes by dual 16-beam infrared beam arrays . The central zone area was defined as 20×20 cm; the left arena was defined as peripheral zone . Total distance traveled , ambulatory time , ambulatory counts , stereotypy time , stereotypy counts , resting time , vertical counts , vertical time , zone entries , zone time , jump counts , jump time , average velocity , and ambulatory episodes were recorded for each test mouse throughout the 60 min . test session . Total distance provides an index of activity , while the proportion of time or distance spent in the center is taken as a measure of anxiety . The elevated plus maze ( Med Associates ) consists of a plus-shaped runway with two horizontal open arms and two horizontal closed arms ( each 6 cm wide×35 cm long ) joined by a 6 cm square center platform . The closed arms are enclosed by 20 cm black polypropylene walls . Mice are placed in the center square and allowed to explore freely under ambient light for five minutes . The number of entries and time spent in each arm is recorded . Open arm entries and occupancy provide an inverse measure of anxiety . The fear conditioning tasks were conducted as described [40] , [41] . Training session consisted of a 3 min exploration period followed by three CS-US pairings separated by 1 min ( foot-shock intensity 0 . 8 mA , duration 0 . 5 s; tone 75 db white noise for 30 sec ) . Context tests were performed in the same training chamber after retention delays of 1 hr and 24 hr . Tone tests were performed in an environmentally altered testing chamber ( different flooring and additional shelter ) 24 hrs following training; baseline freezing was monitored ( 2 min ) prior to phasic presentation of the tone ( 75 db white noise , 3 min duration ) . Baseline freezing was monitored for 2 min prior to phasic presentation of the tone ( 75 db white noise , 3 min duration ) . Mice were trained and tested in conditioning chambers that had a stainless steel grid floor through which footshocks could be delivered ( Med Associates , St . Albans , VT ) . During training and testing sessions , the mouse's position in the chamber is recorded , digitized and analyzed using a video tracking system interfaced with a custom software package . Control and mutant groups consisted of age-matched male and female littermates ( 8–10 weeks of age ) for each analysis . Female animals were group housed ( 4 animals per cage ) for 2 weeks before testing in order to synchronize estrus cycles . Data are presented as group means ± SE . One way and two way ANOVA and Student t-test were used to determine statistically significant differences . For all experiments , the experimenter was blind to genotype . | The X-linked gene Tsx is located within the X-inactivation center and is thought to encode a protein expressed in testis , yet its function is not known . Here we show that Tsx is actually a noncoding RNA , a new member of the large noncoding RNA family expressed from the X . Tsx is abundantly expressed in meiotic germ cells , embryonic stem cells , and brain . Targeted deletion of Tsx generates viable offspring , litter ratios are smaller than expected , X-inactivation is mildly affected ( in embryonic stem cells ) , and male animals have smaller testes due to germ cell apoptosis . Mutant embryonic stem cells are severely growth-retarded and differentiate poorly with elevated cell death . Deletion of this noncoding RNA alters mouse behavior , with animals displaying less fear and enhanced short-term memory . Our study indicates that Tsx performs general functions in multiple cell types and links the noncoding locus to stem and germ cell development , learning , and behavior in mammals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2011 | Tsx Produces a Long Noncoding RNA and Has General Functions in the Germline, Stem Cells, and Brain |
The elbow/no ocelli ( elb/noc ) complex of Drosophila melanogaster encodes two paralogs of the evolutionarily conserved NET family of zinc finger proteins . These transcriptional repressors share a conserved domain structure , including a single atypical C2H2 zinc finger . In flies , Elb and Noc are important for the development of legs , eyes and tracheae . Vertebrate NET proteins play an important role in the developing nervous system , and mutations in the homolog ZNF703 human promote luminal breast cancer . However , their interaction with transcriptional regulators is incompletely understood . Here we show that loss of both Elb and Noc causes mis-specification of polarization-sensitive photoreceptors in the ‘dorsal rim area’ ( DRA ) of the fly retina . This phenotype is identical to the loss of the homeodomain transcription factor Homothorax ( Hth ) /dMeis . Development of DRA ommatidia and expression of Hth are induced by the Wingless/Wnt pathway . Our data suggest that Elb/Noc genetically interact with Hth , and we identify two conserved domains crucial for this function . Furthermore , we show that Elb/Noc specifically interact with the transcription factor Orthodenticle ( Otd ) /Otx , a crucial regulator of rhodopsin gene transcription . Interestingly , different Elb/Noc domains are required to antagonize Otd functions in transcriptional activation , versus transcriptional repression . We propose that similar interactions between vertebrate NET proteins and Meis and Otx factors might play a role in development and disease .
The developing retina of Drosophila melanogaster is a powerful model for studying the specification of cell fates within a retinal mosaic . One important aspect is the localized specification of photoreceptor cell types in response to Wg/Wnt signaling . Wg emanating from the head cuticle specifies specialized ommatidia at the dorsal rim of the developing retina [1]–[3] . Here we show that the Drosophila NET family zinc finger proteins Elbow ( Elb ) and No ocelli ( Noc ) play a crucial role in this process . In fly wing imaginal discs , as well as in a mouse breast cancer model , NET proteins inhibit Wingless ( Wg ) /Wnt signaling [4]–[6] , while expression of the C . elegans homolog TLP-1 is regulated by Wg/Wnt signaling [7] . Hence , the patterning of the ommatidial mosaic in the dorsal periphery of the retina serves as an attractive model to characterize the different roles of NET proteins . The elbow/no ocelli complex of Drosophila is a ∼200 kb locus encoding two closely related proteins belonging to the NET family of zinc finger proteins [8]–[10] ( Figure 1A ) . These proteins are related to Sp1-like transcription factors , but contain only one atypical C2H2 zinc finger with 8 amino acids between the two crucial cysteine residues , making these proteins unlikely to bind DNA directly ( for review: [11] ) . They appear to function as repressors of transcription [12] , [13] and contain several conserved amino acid motifs ( Figure 1C ) . Elb and Noc bind the co-repressor Groucho through a conserved FKPY motif [10] , while other NET family members use different domains for this interaction [6] , [14] . The third highly conserved domain of NET proteins is an N-terminal ‘Sp motif’ , a SPLALLA amino acid sequence shared with the vertebrate Sp1 family of transcription factors , but not with Drosophila Sp1 or Buttonhead ( Btd ) [7] , [11] . Roles for this motif in protein degradation or transcriptional repression have been suggested [15] , [16] . In vertebrates , NET family factors play an important role in the specification of motorneurons [17] , and in the developing hindbrain [18] , [19] and striatum [20] , [21] . Loss of one NET family member , ZNF703 , in humans promotes luminal breast cancer [22] , [23] . In flies , Elb and Noc are important for proximo-distal patterning of the legs and for morphogenesis of tracheal branches [10] , [24] . The Drosophila compound eye is composed of approximately 800 ommatidia ( unit eyes ) , each containing 8 photoreceptor neurons ( called R1 to R8 ) , as well as pigment , cone , and bristle cells [25] . The outer photoreceptors ( R1-6 ) have short axon fibers that terminate in the first optical ganglion , the lamina . The two remaining inner photoreceptors R7 and R8 have light-gathering structures ( rhabdomeres ) that are located in the center of the ommatidium , R7 situated on top of R8 . Their long axonal fibers terminate in a deeper layer of the optic lobe , the medulla [26] . Specification of R7 and R8 depends on the Spalt complex ( Sal ) , encoding two transcription factors , Spalt major ( Salm ) and Spalt related ( Salr ) [27] . The R7 cell type is then induced by Prospero ( Pros ) [28] , while Senseless ( Sens ) determines R8 [29] , [30] . A retinal mosaic arises from the molecular differences between R7/R8 of different ommatidia , creating functional heterogeneity [31] ( Figure 1B ) . At least four ommatidial subtypes can be distinguished in flies . Two subtypes named ‘pale’ ( p ) and ‘yellow’ ( y ) are distributed randomly throughout the eye , with a ratio of 35% ( p ) to 65% ( y ) [32] , [33] . R7 cells in p ommatidia always express UV-sensitive rhodopsin Rh3 , while the underlying R8 express a blue-sensitive rhodopsin Rh5 [34] , [35] . In y ommatidia , R7 express another UV-sensitive rhodopsin , encoded by the rh4 gene , while R8 cells contain a green-sensitive Rhodopsin , Rh6 [36]–[38] . This mosaic of chromatic sensitivities created by p/y ommatidia provides the substrate for Drosophila color vision [39] , [40] . In the dorsal-most third of the adult eye , y ommatidia show an additional specialization by co-expressing both UV Rhodopsins Rh3 and Rh4 in R7 cells [41] . Finally , a narrow band of ommatidia along the dorsal head tissue , called the ‘dorsal rim area’ ( DRA ) , manifests morphological and molecular specializations making these ommatidia ideal detectors for polarized light [42] . Inner photoreceptors R7 and R8 of DRA ommatidia are monochromatic since they both contain the UV Rhodopsin Rh3 [43] . Furthermore , the diameter of their rhabdomeres is enlarged , and the absence of rhabdomeric twist in DRA ommatidia results in high polarization sensitivity [42] , [44] , [45] . As a consequence , DRA ommatidia are both necessary and sufficient for detecting linearly polarized light emanating from the sky [45] , [46] . We have shown that the development of DRA ommatidia in Drosophila depends on Homothorax ( Hth ) a homeodomain transcription factor homologous to vertebrate Meis factors [2] , [47] . Hth is both necessary and sufficient for the specification of DRA ommatidia where it is expressed in both R7 and R8 [2] . Hth always co-localizes in nuclei with its ubiquitous cofactor Extradenticle ( Exd ) , whose nuclear localization depends on Hth ( for review: [48] ) . We have shown that the transcription factor Orthodenticle ( Otd ) is required as an activator of rh3 expression in DRA ommatidia [49] . Otd is expressed in all adult Drosophila photoreceptors , where it acts as a general activator of rh3 and rh5 expression [49] , [50] . Otd also induces repression of rh6 through induction of another homeodomain transcription factor , the repressor ‘Defective proventriculus’ ( Dve ) . In outer photoreceptors ( R1-6 ) , Otd and Dve act in a feedforward loop , resulting in repression of rh3 rh5 and rh6 by Dve in these cells [50] . In inner photoreceptors , Dve is repressed by Spalt ( Sal ) factors , allowing activation of rh3 and rh5 by Otd in ‘pale’ ommatidia . As a consequence , rh3 and rh5 are lost in otd mutants , while expression of rh6 is de-repressed into outer photoreceptors , due to the loss of Dve in these cells and the presence of the rh6-specific activator Pph13 [51] . Alleles of the gene encoding Otd are named ocelliless ( oc ) due to the function of otd in patterning the dorsal head cuticle where ocelli form . Therefore , oc and noc are both required for ocellar development [9] , [52] . Furthermore , both genes show specific , overlapping expression at the anterior pole of the fly embryo [53] , yet their regulatory relationship is not known . The vertebrate homologs of Otd are involved in retinal development , including OTX1 , OTX2 and CRX ( ‘cone rod homeobox’ ) , whose mutations cause retinal degeneration [54] , [55] . Furthermore , OTX1 and OTX2 are required for the specification and regionalization of the forebrain and midbrain , while Otd is required for the development of the anterior brain in flies ( for review: [56] ) . Several aspects of the otd mutant phenotype in flies can be rescued by either CRX or OTX2 in the retina [57] , as well as the developing brain [58] . Furthermore , many aspects of the OTX1 phenotype in mice can be rescued by substitution with the Drosophila Otd protein [59] , demonstrating that the molecular function of these factors is conserved . Here we show that Elb and Noc play an important role in the specification of DRA ommatidia . Loss of both genes together leads to a phenotype identical to the loss of hth: the enlarged rhabdomere diameter of DRA inner photoreceptors is lost , and expression of Rh3 in DRA R8 cells is replaced with Rh6 . Furthermore , the specific R8 marker Senseless ( Sens ) becomes de-repressed in R8 cells of DRA ommatidia . Since Hth expression in the DRA is normal in elb , noc double mutants , this indicates that Hth is unable to execute its DRA-inducing potential . Gain-of-function experiments in combination with site-directed mutagenesis of the three evolutionarily conserved domains in Elb and Noc reveal that an N-terminal SPLALLA motif as well as the unique zinc finger are crucial for the function of Elb and Noc in DRA ommatidia . Furthermore , Elb and Noc can genetically antagonize the activator or repressor functions of Otd in the retina through distinct protein domains . We therefore propose that NET family proteins might interact with Meis as well as Otx family genes , depending on the transcriptional context .
In a GAL4 enhancer trap screen for genes expressed in adult photoreceptors , we obtained two independent insertions in the elbow/no ocelli complex [8] , [9] , one localized 950 bp upstream of the transcription start of elbow ( also referred to as ‘elbow B’ , elB , or el; [10] , and the other 285 bp upstream of no ocelli ( noc; [24] ) ( Figure 1A ) . We previously showed that these two enhancer traps faithfully reproduce the expression patterns of the genes in which they are inserted , in both tracheae and leg imaginal discs [10] , [24] . When crossed to UAS-GFP , the GFP signal under water immersion could be localized to the adult inner photoreceptors , with additional signal coming from non-neuronal cells , most likely cone cells ( Figure 1D , G ) . GAL4 expression was strikingly similar between the two lines ( Figure 1E , F , H , I ) . Expression was strongest in R7 and R8 cells in the ‘dorsal rim area’ ( DRA ) , as well as in non-DRA R8 cells ( Figure 1E , H ) , while expression in non-DRA R7 cells was much weaker ( Supplemental Figure S1D , E ) . Different expression levels in R7 and R8 did not correlate with p/y-specific rhodopsin subtypes ( Supplemental Figure S1F , G ) . In DRA ommatidia , strong GAL4 expression always co-localized with Homothorax ( Hth ) ( Figure 1F , I ) . Although no expression was detected in adult outer photoreceptors R1-6 ( Figure 1G ) , both GAL4 lines were expressed in larval R3 and R4 photoreceptors in eye-imaginal discs , with an onset 3–4 rows posterior to the morphogenetic furrow ( Supplemental Figure S1A , B ) . No expression in R7 and R8 was detected at that stage . Co-expression with the inner photoreceptor marker Spalt ( Sal ) started around 50% pupation ( Supplemental Figure S1C ) . Thus , elb/noc expression starts in R3 and R4 early in development , but becomes restricted to inner photoreceptor types during mid-pupal development , and is expressed most strongly in adult R7/R8 of DRA ommatidia , as well as non-DRA R8 cells . We first tested single mutants for either elbow or no ocelli , for changes in the Rhodopsin pattern [38] , [39] . Neither homozygous viable elb3 . 3 . 1 null mutants [24] , nor whole mutant eyes generated using the null allele nocΔ64 ( -/- ) [10] , generated with ey-Flip/GMR-hid [60] , showed a Rhodopsin phenotype affecting R7 cells or R8 cells ( Supplemental Figure S2A-F ) . We then used the same technique to generate elb3 . 3 . 1 , noc Δ64 ( -/- ) double mutant eyes ( Figure 2C , D , G ) . Although Rhodopsin expression appeared unaffected in most inner photoreceptors ( Figure 2A–D , Supplemental Figure S2G , H ) , Rh3 expression was no longer detected in R8 cells in the DRA ( Figure 2C ) . Instead , the R8 rhodopsin Rh6 was now expressed in the dorsal-most R8 cells ( Figure 2D ) . Mutants lacking Hth function exhibit the same phenotype of ‘odd-coupled’ expression of Rh3/Rh6 in DRA ommatidia R7/R8 [2] . We stained elb3 . 3 . 1 , noc Δ64 ( -/- ) double mutant eyes for Hth along with Rhodopsins ( Figure 2E–G ) . In these double mutants , Hth expression was unaffected . DRA R7 co-expressed Hth and Rh3 while DRA R8 co-expressed Hth and Rh6 ( Figure 2G ) . This situation was also identical to flies expressing dominant-negative Homothorax ( HthHM ) in all photoreceptors ( Figure 2F ) . Hence , in the absence of elb and noc , DRA ommatidia are mis-specified into ‘odd-coupled’ Rh3/Rh6 ommatidia , despite the presence of Hth ( Figure 2H ) . We therefore concluded that Elb and Noc act downstream of , or in parallel to Hth in the specification of DRA photoreceptor cell fates . DRA R8 cells express an R7 Rhodopsin and therefore lack features of R8 cells , like expression of the R8 transcription factor Sens [61] . Indeed , Sens becomes specifically de-repressed in hth mutant DRA R8 cells [2] . This phenotype correlates with the gain of Rh6 expression in hth mutant DRA R8 cells , since Sens has an inductive effect on rh6 expression [29] , [62] . We stained elb3 . 3 . 1 , noc Δ64 ( -/- ) double mutant retinas for Exd and Sens ( Figure 3A–F ) . Like Hth , Exd was localized to the nuclei of DRA R7 and R8 in wild type ( Figure 3A , B ) , as well as in elb , noc double mutants ( Figure 3D; Supplemental Figure S3A , C ) . However , Sens was now co-expressed with Exd in R8 cells , similar to flies over-expressing dominant-negative HthHM ( Figure 3C ) . Co-expression of Exd and Sens in DRA R8 cells was already visible in pupal retinas ( 50% APF; Figure 3E , F ) , arguing that R8 cells in DRA ommatidia lacking elb and noc became mis-specified before Rhodopsin expression begins . Outside the DRA , expression of inner photoreceptor markers Spalt , Prospero , and Senseless were indistinguishable from wild type ( Supplemental Figure S3B , D–F ) . Unlike the R8 marker Sens , the R7 marker Prospero is not repressed in DRA R7 cells by Hth/Exd [2] . Hence , there was no immuno-histochemical way to tell whether DRA R7 cells had actually changed their fate in elb , noc ( -/- ) mutants . We therefore assessed the morphology of the eye tissue in tangential plastic sections in elb3 . 3 . 1 , noc Δ64 ( -/- ) double mutant eyes ( Figure 3G ) . Wild type DRA ommatidia display an enlarged R7 and R8 rhabdomere diameter [2] , [42] , [45] , [63] . In mutant clones touching the dorsal eye rim , the typical DRA morphology was lost and resembled ommatidia from the main region of the eye . We have previously described the same morphological DRA phenotype in HthB2 ( -/- ) mutant clones [2] . We therefore concluded that all markers of DRA ommatidia were lost in elb3 . 3 . 1 , noc Δ64 double mutants , despite persisting expression of Hth . Over-expression of Hth in all photoreceptors leads to a transformation of the entire retina into DRA ommatidia , with Rh3 expression expanding into all inner photoreceptors , while expression of Rh4 , Rh5 , Rh6 , and Sens is lost [2] . We tested whether Elb and Noc were required for this function of Hth , focusing on expression of Rh6 and Sens as the most reliable markers ( Figure 4A–C ) . First , we confirmed that expression of Rh6 was always lost in transgenic flies expressing a GFP:hth fusion protein directly attached to LGMR ( Figure 4B; Supplemental Figure S4A–C; see Materials and Methods ) . However , when GFP:hth was over-expressed in elb3 . 3 . 1 , noc Δ64 ( -/- ) double mutant eyes , Rh6 was found in R8 in the entire retina ( Figure 4C ) . Rh3 was also detected , while Rh4 and Rh5 were absent ( data not shown ) . Thus , the entire retina was transformed into ‘odd coupled’ Rh3/Rh6 ommatidia , like the ones we had described in the DRA of HthHM mutants . Interestingly , Elb and Noc remained restricted to inner photoreceptors when GFP:hth was over-expressed , but became expressed at high levels in all R7 and R8 ( Figure 4D ) . Taken together , these data suggested that Elb and Noc act downstream of Hth in the specification of DRA ommatidia . As a second test of this epistatic relationship , we took advantage of the fact that the Wg pathway induces DRA ommatidia [1] , [3] . Ectopic activation of the Wg pathway with constitutively active forms of Armadillo ( armS10 , or arm*; [64] ) induces Hth in the entire dorsal eye and transforms it into DRA ommatidia , repressing Rh6 in all dorsal R8 cells [2 ) . We therefore tested the requirement of elb and noc in the dorsal eye using direct GMR-arm* fusions [65] . While Hth was ectopically expressed in all inner photoreceptors of the dorsal eye in GMR-arm* flies , leading to the loss of Rh6 expression from the expanded DRA ( Supplemental Figure S4D ) , we observed co-expression of Rh6 and Hth in R8 throughout the dorsal eye in an elb3 . 3 . 1 , noc Δ64 ( -/- ) mutant background containing GMR-arm* ( Supplemental Figure S4E ) . Rh4 and Rh5 were always excluded from the dorsal eye , while Rh3 expression remained ( not shown ) . Hence , activating the Wg pathway in elb3 . 3 . 1 , noc Δ64 ( -/- ) mutants induced Hth expression throughout the dorsal half of the eye , transforming it into ‘odd coupled’ Rh3/Rh6 ommatidia . Elb and Noc are therefore necessary for the ability of Hth to induce DRA fate in response to activating the Wingless pathway . We also tested whether Hth required Elb and Noc for repression of Sens , when ectopically expressed by generating clones of elb3 . 3 . 1 , noc Δ64 ( -/- ) mutant tissue in pupal retinas expressing LGMR- hth ( Figure 4E ) . In these retinas , the vast majority of R8 cells ( though not all ) strongly expressing Sens were located within the elb3 . 3 . 1 , noc Δ64 ( -/- ) double mutant tissue ( Figure 4E–G ) . Therefore , Hth had lost the ability to repress Sens in the absence of Elb and Noc , which is consistent with a loss of DRA fate [2] . This requirement of Elb/Noc appeared not to be strictly cell autonomous since Sens-positive R8 cells could be observed outside elb , noc mutant clones , although almost always in their direct vicinity ( Figure 4G ) Taken together , we concluded that Hth function in the specification of DRA ommatidia in response Wg pathway activation was dependent on the presence of Elb and Noc . Homothorax is sufficient to induce the DRA fate in all ommatidia when over-expressed [2] . In contrast , over-expression of either Elb or Noc , or both proteins together never resulted in the induction of DRA markers ( Supplemental Figure S5 ) , and the gain-of-function phenotypes will be discussed in more detail below . All NET family zinc finger proteins share several conserved protein domains whose functional significance remains incompletely understood [11] , [14] . We investigated the role of three domains in DRA specification: the conserved Sp/SPLALLA motif at the N-terminus , a more centrally located FKPY motif that was shown to interact with the transcriptional repressor Groucho , and the unusual zinc finger , located C-terminally ( Figure 1C ) . We altered the amino acid sequence of each of these sequences individually using site-directed mutagenesis and placed the resulting cDNA for elb and noc in UAS-vectors . Rescue experiments using these point-mutated versions of elb or noc were not possible since the necessary driver lines elb-GAL4 and noc-GAL4 were ( non-mutant ) enhancer traps inserted within the elb . noc locus . We therefore tested whether these constructs affected DRA specification in a dominant-negative fashion when over-expressed , similar to HthHM . We altered the Sp/SPLALLA motif of both Elb and Noc to TGIVIIV ( Figure 5A; see Materials and Methods ) . Over-expression of UAS-elb[SPLALLA*] in all photoreceptors using LGMR-GAL4 indeed had a dominant negative effect on DRA development . Expression of Rh3 in DRA R8 cells was lost ( Figure 5B ) and , instead , Rh6 was expanded to the dorsal rim of the retina ( Figure 5C ) . In these DRA R8 cells , we detected co-expression of Sens and Exd , which is an indication of DRA fate loss ( Figure 5D ) . Expression of Rh3 , Rh4 and Rh5 was normal outside the DRA in LGMR>elb[SPLALLA*] flies ( Figure 5F ) . Interestingly , Rh6 was expanded into all outer photoreceptors R1-6 , for reasons we will discuss below . To test the involvement of the unique zinc finger motif for DRA specification , we mutated the two crucial Histidine residues in the zinc finger of both Elb and Noc proteins , replacing them with Alanine ( Figure 5G; see Materials and Methods ) , thereby disrupting its ability to chelate zinc [66] , [67] . Over-expression of elb[ZnF*] also resulted in a dominant-negative loss of the DRA: Rh3 was lost in R8 DRA cells and was replaced by Rh6 , resulting in Rh3/Rh6 ‘odd-coupled’ ommatidia ( Figure 5H , I ) . Finally , we altered the conserved , generic Groucho-binding motif present in both Elb and Noc ( Figure 6A; FKPY → IEGS; see Materials and Methods ) . Over-expression of elb[Gro*] had no dominant-negative effect on DRA development ( Figure 6B , C ) , and instead resulted in an unrelated rhodopsin phenotype ( see below ) . We concluded from these experiments that point mutation of either Sp/SPLALLA-motif or zinc finger in Elb resulted in dominant-negative function , possibly by sequestering factors present in the DRA ( possibly Hth ) , through the unaltered parts of the over-expressed inactivated protein . The dominant-negative function of over-expressed , point-mutated forms of Elb might therefore result in inactive protein complexes , similar to what has been described for HthHM [68] . Hence , these results suggested that Sp/SPLALLA domain and zinc finger were crucial for the in vivo function of Elb . Interestingly , the identical point-mutations in the Sp/SPLALLA or zinc finger motifs of the Noc protein did not result in a dominant-negative effect on DRA specification ( Supplemental Figure S6A–F ) . Outside the DRA , over-expression of wild type Elb led to the loss of rh3 and rh5 expression , while rh6 was expanded into outer photoreceptors ( Supplemental Figure S5A–G ) . This phenotype , which was observed both in wild type and in elb , noc double mutants backgrounds was identical to the loss of Otd function in otdUVI mutants [49] , [50] . Gain-of-function experiments using UAS-noc resulted in a similar , but much milder phenotype , with only some outer photoreceptors de-repressing Rh6 ( Supplemental Figure S5K , L ) . Importantly , over-expression of Elb did not repress larval or adult otd expression ( Supplemental Figure S5I ) and expression of elb/noc was unaltered in otdUVI mutants ( Supplemental Figure S4F , G ) . Hence , since elb3 . 3 . 1 , nocΔ64 mutant eyes showed no phenotype outside the DRA , we concluded that these gain-of-function phenotypes might be due to an antagonistic genetic interaction between Elb/Noc and Otd , which is expressed in all photoreceptors [69] ( although it affects Rhodopsin expression in specific photoreceptor subtypes , see discussion ) [49] , [50] . Interestingly , the phenotypes observed outside the DRA when over-expressing different point-mutated versions of Elb and Noc separated different aspects of the otdUVI phenotype: To further investigate Elb/Noc's role as activators or repressors of transcription , and how they antagonize Otd , we generated fusions with the VP16 transcriptional activation domain [70] , as well as the repressor domain of Engrailed [71] ( Figure 7 ) . Both kinds of fusion proteins were placed under UAS control for gain-of-function experiments ( see Materials and Methods ) . Over-expression of VP16:elb with LGMR-GAL4 led to a severe disruption of retinal morphology ( not shown ) that was not observed with UAS-VP16:noc . We therefore focused our analysis on Noc fusion proteins ( Figure 7A–E ) . Over-expression VP16:noc had no effect on R7 Rhodopsins , although Rh3 expression was weak in the DRA ( Figure 7B ) . We confirmed that the DRA markers Hth and Rh3 were correctly expressed in DRA R8 in sevenless ( sev ) mutants ( Figure 7C ) , showing that DRA ommatidia were indeed correctly specified . Rhodopsin expression in R8 outside the DRA was severely disrupted: Rh5 was found expanded to most R8 cells ( figure 7D ) , while Rh6 expression appeared normal ( figure 7E ) . The consequence was co-expression of Rh5 and Rh6 in many R8 cells ( ∼51% ) . In wild type flies , such co-expression is prevented by mutually exclusive genes in p and y R8 [72] , [73] . We therefore concluded that VP16:noc had a direct activating effect on rh5 , without affecting any other rhodopsin . Over-expression of the repressor fusion en[R]:noc ( Figure 7F ) also had no effect on DRA specification ( Figure 7G ) . However , expression of Rh5 was lost , while Rh3 , Rh4 , and Rh6 appeared normal in the rest of the retina ( Figure 7H ) . This R8 phenotype observed with en[R]:noc was therefore the opposite of that observed with the activator fusion VP16:noc , with both Noc fusion proteins specifically affecting rh5 expression in opposite manner . Since Otd directly activates rh3 and rh5 transcription [49] , [50] , Noc might function as a direct antagonist of Otd's function in R8 cells , but not in R7 . Interestingly , we found strong Elb/Noc expression outside the DRA only in R8 cells ( Figure 1E–I; Supplemental Figure S1D–G ) . Mutations in the human Elb/Noc homolog ZNF703 promote metastasis in luminal breast cancer [22] , [23] , [74] . To investigate if NET family protein functions are evolutionarily conserved , we generated UAS-constructs for over-expression of both human NET family proteins in Drosophila ( UAS-ZNF503 , UAS-ZNF703; see Materials & Methods ) . Gain-of-function phenotypes in the retina were very weak , but Rh5 was expanded in R8 cells , leading , in both cases , to co-expression with Rh6 ( Supplemental Figure S7A–E ) . The C . elegans homolog TLP-1 was not active in this assay ( Supplemental Figure S7F-H ) . This co-expression phenotype therefore resembles most closely what we had observed for the over-expression of a VP16:noc , suggesting that genetic interaction of NET family proteins with Otd/Otx proteins is a conserved feature of these factors .
The transcription factors Homothorax ( Hth ) and Extradenticle ( Exd ) have been well characterized as co-factors for Hox genes [48] . Hth/Exd can also act as co-factors for non-Hox transcription factors , like for Engrailed [75] , [76] . Here we showed that loss of both Elb and Noc phenocopies the loss of Hth at the dorsal rim of the retina . All markers of DRA ommatidia are lost in elb , noc double mutants: Rh3 expression and Sens repression in DRA R8 , as well as the DRA-specific inner photoreceptor rhabdomere morphology in DRA R7 and DRA R8 . Our data shows that Elb/noc act downstream of Hth in the specification of DRA cell fates . Elb and Noc are expressed strongly in DRA R7 and R8 . This expression is expanded to all R7 and R8 by ectopic Hth ( but never into outer photoreceptors R1-6 ) , while Hth expression is not affected in elb , noc double mutants . One possibility is that Elb/Noc serve as cofactors for Hth/Exd ( Figure 8A ) , since Hth loses its potential to induce the DRA fate in a double mutant retina . The vertebrate homologs of Elb and Noc function as repressors of transcription [13] . Therefore , aspects of the Hth/Exd and Elb/Noc loss-of-function phenotypes could be due to a direct failure of their complex to repress common target genes . For instance , the de-repression of the R8 marker Sens by dominant-negative hthHM , as well as in elb , noc double mutants could be explained by loss of a repressor complex containing all four proteins . Interestingly , functional antagonism between the Hox/Hth/Exd complex and Sens have been described in the Drosophila embryo [77] . However , in this case the factors were shown to compete for overlapping binding sites in the promoter of the common target gene rhomboid . Gene expression profiling data revealed that the Hox gene Abd-B also directly represses Sens in the embryo using Hth/Exd as cofactors [78] . Elb and Noc might therefore provide a missing link for transcriptional repression of Sens by Hth/Exd . Much work on NET family proteins has focused on functional characterization of their evolutionarily conserved domains . The C-terminus of NET proteins is required for nuclear localization [14] , [79] , as well as for self-association of the zebrafish ortholog Nlz1 , although neither self-association nor heterodimerization with Nlz2 was found to be necessary for wild type function [79] . The ‘buttonhead box’ [80] , a conserved 7–10 amino acid motif which we have not investigated in this study , may be required for transcriptional activation [81] . Deletion of the ‘buttonhead box’ in zebrafish Nlz proteins transformed them into dominant-negatives , an effect that was proposed to be due to reduced affinity to co-repressor Groucho and histone de-acetylases [12] , [79] . Interestingly , deletion of N-terminal sequences , including the Sp/SPLALLA motif also leads to dominant negative proteins [79] . These data are consistent with our findings that a protein with a mutated Sp/SPLALLA motif has a dominant-negative effect on DRA specification . The Sp motif was proposed to mediate transcriptional repression by directly binding to cofactors [15] . It should be noted that both N-terminal Sp/SPLALLA deletion and the VP16 fusions have the same dominant-negative effect for zebrafish Nlz1 [12] . While this is consistent with a pure repressor function of the zebrafish protein , the differences between Sp/SPLALLA mutation and VP16-fusion ( as well as the observation of a phenotype for the Engrailed fusion ) reported in this study hint towards a more complex role of Elb and Noc in transcriptional regulation . We showed that mutation of the conserved zinc finger of Elbow also transforms this protein into a dominant-negative . Usually , multiple zinc fingers are required for DNA binding , suggesting that the NET family zinc finger is a protein-protein interaction domain [11] , [67] . Deletion of the zinc finger from zebrafish Nlz proteins leads to a loss of nuclear localization [79] , and the Nlz1 zinc finger is necessary for transcriptional repression [13] . Although we cannot exclude the possibility that Elb and Noc bind DNA through their zinc finger , it is likely that mutation of the zinc finger either leads to an inactive complex by sequestration of another co-repressor , or that such complex could be trapped in the cytoplasm . Given that mutation of either Sp/SPLALLA motif or zinc finger both lead to a dominant-negative effect raises the possibility that protein binding to both motifs could be necessary for in vivo function , possibly through the formation of higher order transcriptional complexes . Loss of both elb and noc does not result in Rhodopsin phenotypes outside the DRA . However , over-expression of different forms of Elb or Noc recapitulates all Rhodopsin phenotypes observed in otdUVI mutants [49] , [50] . This phenotype might therefore arise from forcing a direct interaction between over-expressed Elb protein and Otd . Little is known about the regulatory relationship between Elb/Noc and Otd . However , the overlapping expression patterns and similar phenotypes for certain alleles of otd named ocelliless , and for no ocelli ( noc ) at the anterior pole of the fly embryo , as well as their common requirement in the morphogenesis of ocelli suggests that these proteins also interact positively outside of the retina . The antagonism we observed might therefore be a dominant-negative effect resulting from sequestration of the Otd protein by over-expressed Elb . Alternatively , different combinations of transcriptional cofactors present between tissues ( for instance DRA versus non-DRA R8 cells ) might decide whether Elb and Noc act in concert with Otd , or as antagonists . In the retina , Otd acts in a ‘coherent feedforward loop’ with Spalt to directly activate transcription of rh3 and rh5 [50] . As a consequence , Rh3 and Rh5 are lost in otd mutants . Furthermore , Otd activates transcription of the repressor Dve , forming an ‘incoherent feedforward loop’ , resulting in repression of rh3 and rh5 in outer photoreceptors . Since rh6 is activated by a distinct factor , Pph13 [51] , loss of Otd leads to a specific de-repression of rh6 into outer photoreceptors [50] . We show that different domains of Elb specifically interfere with different aspects of Otd function in these feedforward loops ( Figure 8B ) . Mutation of the Groucho-binding motif FKPY only abolishes the ability of over-expressed Elbow protein to antagonize Otd function in repressing rh6 in outer photoreceptors , while mutation of the Sp/SPLALLA motif specifically antagonizes Otd function in activating both rh3 and rh5 , without affecting repression of rh6 in outer photoreceptors ( mediated by induction of Dve ) . Furthermore , while the Elb zinc finger is also required for antagonizing the function of Otd in outer photoreceptors , it is also necessary for antagonizing activation of rh3 by Otd , but not rh5 . Hence , these two activator functions of Otd could be separated by mutating the zinc finger . The different Rhodopsin phenotypes caused by loss of Otd can be mapped to different protein domains [51] . Our data therefore reveal specific genetic interactions between the protein domains of Elb/Noc and Otd . Such interactions could be direct or be mediated through additional proteins . For instance , the Otd C-terminus mediates the repression of rh6 in outer photoreceptors [82] , making it a possible interaction domain for Groucho binding to the Elb/Noc FKPY motif . The N-terminus of Otd is necessary for most activation potential on rh3 , while activation of rh5 predominantly maps to the C-terminus [82] . This correlates well with the Rhodopsin-specific phenotypes we see after mutation of Sp/SPLALLA ( affecting rh3 and rh5 ) , or the zinc finger ( affecting rh3 and rh6 ) motifs . Finally , our results using VP16- and en[R]-fusions of Noc show that potentially direct transcriptional effects on rhodopsin genes can only be induced in R8 cells . Both fusion proteins specifically regulate expression of rh5 , while all other rhodopsins remain unaffected . Elb and Noc are both expressed strongly in R8 cells outside of the DRA where they may contribute the repression of Rh5 . The absence of a non-DRA R8 rhodopsin phenotype in elb , noc double mutants , as well as the R8-specific action of VP16:noc could therefore be due to the existence of redundant , R8-specific factors required for Elb/Noc function there , but not for DRA specification . These factors remain unknown , since we found that expression of elb and noc is not altered in homozygous mutants affecting p/y cell fate decisions in R8 cells ( melt and wts , [72]; Supplemental Figure S8 , A–D ) , like in R7 cells ( Supplemental Figure S8E , F ) . Mutations in the human Elb/Noc homolog ZNF703 promote metastasis [6] . We have shown that over-expression of both human NET family proteins UAS-ZNF503 and UAS-ZNF703 in the Drosophila retina result in weak co-expression of Rh5 and Rh6 , resembling over-expression of a VP16:noc protein . It is therefore possible that the genetic interaction of NET family proteins with Otd/Otx proteins is evolutionarily conserved , especially since a central domain of Otd was previously shown to mediate mutual exclusion of Rh5 and Rh6 [82] . Here we present a new role for Drosophila NET proteins in retinal patterning . Both zebrafish homologs of Elb/Noc , Nlz1 and Nlz2 are also required for optic fissure closure during eye development [83] . Furthermore , expression of the Elb/Noc mouse homologue znf503 suggests that NET family genes are involved in the development of mammalian limbs [84] . Given previous reports from Drosophila on the proximo-distal specification of leg segments [24] , it appears that NET family members act in similar processes across species . This raises the possibility that NET proteins serve as evolutionarily conserved modules that have been re-utilized for analogous processes during evolution . Based on our data , their conserved domain structure might be crucial for interacting with transcription factor networks involving conserved families of factors like Otx or Meis . Given their medical relevance in breast cancer , a better understanding of the role NET proteins play in the transcriptional control of tissue patterning will be of great importance .
Primary antibodies used were anti-βGal rabbit polyclonal 1/5000 ( Cappel ) , anti-βGal mouse monoclonal 1/500 ( Promega ) , anti-Homothorax guinea pig polyclonal 1/500 ( R . Mann , Columbia University ) , anti-ElaV mouse or rat monoclonals 1/10 ( Iowa University Hybridoma bank ) , anti-24B10 mouse monoclonal 1/10 ( Iowa University Hybridoma bank ) , anti-Prospero mouse monoclonal 1/4 ( Iowa University Hybridoma bank ) , anti-Senseless guinea pig polyclonal 1/10 ( H . Bellen , Baylor College ) , anti-Rh3 mouse monoclonal 1/100 ( S . Britt , University of Colorado ) , anti-Rh3 chicken polyclonal 1/20 ( T . Cook , University of Cincinnati ) , anti-Rh4 mouse monoclonal 1/100 S . Britt ) , anti-Rh5 mouse 1/100 ( S . Britt , anti-Rh6 rabbit polyclonal 1/1000 [49] . Secondary antibodies were a ) AlexaFluor488 coupled made in goat or donkey , anti-rabbit , mouse , rat or guinea pig ( Molecular Probes ) , b ) Cy3 or TxRed-coupled made in goat or donkey , anti-rabbit , mouse , rat , guinea pig or chicken ( Jackson Immunochemicals ) and c ) Cy5 coupled made in goat or donkey , anti-mouse or rat ( Jackson Immunochemicals ) . | The eyes of many animals contain groups of photoreceptor cells with different chromatic sensitivities that can be arranged in complex patterns . It is of great interest to identify the genes and pathways shaping these ‘retinal mosaics’ which include stochastically distributed groups of cells , versus highly localized ones . In many insect eyes , which are composed of large numbers of unit eyes , or ommatidia , specialized photoreceptors are found only in the dorsal periphery , where they face the sky . These ommatidia are responsible for detecting linearly polarized skylight , which serves as an important navigational cue for these animals . Here we describe how two closely related proteins called Elbow and No ocelli interact with the transcription factors Homothorax and Orthodenticle to correctly specify the polarization detectors at the dorsal rim of the retina of Drosophila melanogaster . Interestingly , all four proteins are evolutionarily conserved from worms to humans , and they appear to be involved in similar developmental processes across species . Furthermore , human homologs of Elbow and No ocelli have been identified as promoters of luminal breast cancer . The newly identified role of these two proteins within a regulatory network might therefore enable new approaches in a number of important processes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"developmental",
"biology",
"developmental",
"neuroscience",
"animal",
"genetics",
"model",
"organisms",
"entomology",
"gene",
"expression",
"genetics",
"molecular",
"genetics",
"cellular",
"types",
"biology",
"sensory",
"systems",
"molecular",
"cell",
"biology",
"zoology",
"neuroscience",
"pattern",
"formation",
"cell",
"fate",
"determination",
"evolutionary",
"developmental",
"biology"
] | 2014 | Genetic Dissection of Photoreceptor Subtype Specification by the Drosophila melanogaster Zinc Finger Proteins Elbow and No ocelli |
The peak shift model predicts that the age-profile of a pathogen's prevalence depends upon its transmission rate , peaking earlier in populations with higher transmission and declining as partial immunity is acquired . Helminth infections are associated with increased immunoglobulin E ( IgE ) , which may convey partial immunity and influence the peak shift . Although studies have noted peak shifts in helminths , corresponding peak shifts in total IgE have not been investigated , nor has the age-patterning been carefully examined across populations . We test for differences in the age-patterning of IgE between two South American forager-horticulturalist populations and the United States: the Tsimane of Bolivia ( n = 832 ) , the Shuar of Ecuador ( n = 289 ) , and the U . S . NHANES ( n = 8 , 336 ) . We then examine the relationship between total IgE and helminth prevalences in the Tsimane . Total IgE levels were assessed in serum and dried blood spots and age-patterns examined with non-linear regression models . Tsimane had the highest IgE ( geometric mean = 8 , 182 IU/ml ) , followed by Shuar ( 1 , 252 IU/ml ) , and NHANES ( 52 IU/ml ) . Consistent with predictions , higher population IgE was associated with steeper increases at early ages and earlier peaks: Tsimane IgE peaked at 7 years , Shuar at 10 years , and NHANES at 17 years . For Tsimane , the age-pattern was compared with fecal helminth prevalences . Overall , 57% had detectable eggs or larva , with hookworm ( 45 . 4% ) and Ascaris lumbricoides ( 19 . 9% ) the most prevalent . The peak in total IgE occurred around the peak in A . lumbricoides , which was associated with higher IgE in children <10 , but with lower IgE in adolescents . The age-patterning suggests a peak shift in total IgE similar to that seen in helminth infections , particularly A . lumbricoides . This age-patterning may have implications for understanding the effects of helminths on other health outcomes , such as allergy , growth , and response to childhood vaccination .
Age-related epidemiological patterns are thought to result from complex interactions between parasite life cycle , exposure to infection , and host-immunity [1] , [2] . Helminth infections show characteristic age-patterning , peaking around puberty and then declining during adulthood [3]–[5] . However , this pattern varies with infection prevalence and intensity , tending to both peak and decline earlier in populations with higher rates of transmission . This “peak shift” is thought to result from the interaction between the rate at which new individuals are infected and the rate at which partial immunity is acquired [1] , [2] . According to this model , when transmission is high infection occurs more quickly , leading to a higher prevalence at a younger age . However , earlier infection also leads to an earlier acquisition of immunity , leading to a decline in prevalence following the peak . Although studies have found age-patterns in helminth prevalences consistent with this hypothesis , few studies have examined whether the age-patterning of immune responses follows similar patterns . Those studies that have examined the age-patterning of immune responses have generally focused on parasite specific immunoglobulins ( IgG , IgA , IgM , and IgE ) [6]–[11] . These have shown age-patterns that resemble the age-specific prevalences of parasites . However , in addition to specific responses , helminth infections are associated with a general shift in the host immune system towards a TH2-biased phenotype , characterized in particular by increased production of total IgE . Although specific responses are thought to participate in protection against infection and thus the generation of peak-shift patterns , an examination of total IgE levels is also critical for understanding the effects of helminths on health and immune function . Due to stimulation of TH2 responses , total IgE is likely to represent the total burden of multiple helminth species infection better than species specific immunoglobulins . Like specific IgE , total IgE levels are elevated in infected individuals and fall with treatment [12] , [13] , and have been shown to correlate with specific IgE for Ascaris lumbricoides and T . trichiura [14] . However , total IgE levels in heavily parasitized individuals remain elevated compared to individuals in industrialized countries for substantial periods of time [15] , suggesting persistent changes in host immune function . Mounting an immune response is energetically costly , necessitating reductions in competing life history demands , including growth , reproduction , and survival [16]–[17] . Higher total IgE levels are associated with poorer growth and shorter adult stature , suggesting a trade-off between growth and investment into immune response [18] . Moreover , the shifting of immune function towards a TH2 phenotype may reduce TH1 responses , decreasing the effectiveness of vaccines or increasing susceptibility to viruses and bacteria [19]–[21] . These effects may depend , in part , on the timing of exposure , as exposure to helminths during critical periods may bias the development of immune function or a child's growth trajectory . Helminths infect more than one seventh of the world's population , and given the peak-shift pattern , a disproportionate number of those infected are schoolchildren [22] . As a consequence , age-patterns in helminth infection and immune response are likely to have significant consequences on growth and development . Although several studies have reported that IgE increases quickly in the first 5–10 years of life and then levels off [23] , [24] , few studies have carefully examined the age-patterning of total IgE and we know of no published studies that have compared age-patterning in IgE across multiple populations . As a marker of helminth infection and TH2-biasing of T-cell responses , an understanding of the age-patterning of total IgE is important for understanding the broader consequences of helminth infections on life history parameters . The current study describes in detail the age-patterning of IgE levels in three populations . These include data from the United States collected by the National Health and Nutrition Examination Survey 2005–2006 ( NHANES ) and data from two populations of South American forager-horticulturalists: the Tsimane of Boliva and the Shuar of Ecuador . First , we test for predicted associations between population mean IgE level and the age-pattern of IgE . Second , using Tsimane data we examine the relationship between age-patterning in IgE and age-patterning in helminth infections .
For Shuar , permission to conduct the study was first obtained from the Federacíon Interprovincial de Centros Shuar ( FICSH ) , the elected representational organization for Shuar affairs . Second , permission was obtained from elected village leaders . Third , a village meeting was held in which a village-level consent form was read aloud , the study explained , questions answered , and a community decision reached about whether to allow the study . Individuals were informed that they could choose not to participate , participate only in individual portions of the study , or participate in the full study . At the time of data collection , individual oral consent was obtained , with individuals able to opt-in or out of individual components of the study ( e . g . , to provide blood spots or not ) . For subjects under age fifteen ( the local age of consent ) both parental consent and child assent were obtained . Oral consent was used for two reasons: 1 ) many Shuar are non or semi-literate or have only a few years of schooling , and 2 ) many Shuar are suspicious or uncomfortable with signing documents due to a history of territorial land disputes and wariness about signed documents leading to ownership conflicts . An independent bilingual Shuar village leader , nurse , FISCH official or assistant was present to translate as needed during group and individual consent and study procedures . The study and consent procedures were approved by the Institutional Review Board ( IRB ) of the University of Oregon . For Tsimane , informed consent was obtained at three levels: 1 ) from the Gran Consejo Tsimane , the local Tsimane government organization that represents Tsimane interests and oversees all projects , 2 ) community officials and participants in village meetings , and 3 ) individual consent during medical visits and before each procedure . After explanation of a formal protocol by bilingual Tsimane assistants , consent forms were signed for literate participants , and verbal approval with fingerprint signature given for non-literate participants . Tsimane consent procedures were approved by the IRBs at the University of New Mexico , University of California , Santa Barbara and the University of Southern California . Tsimane fecal samples were analyzed using two methods . From 2004 to 2008 fecal samples were analyzed for the presence of helminth eggs and larvae by direct identification on wet mounts . As described in greater detail elsewhere [33] , duplicate mounts were prepared with 0 . 9% saline solution and iodine solution , respectively , and examined at 100x and 400x for helminth eggs ( hookworm , A . lumbricoides , and T . trichiuris ) , and larvae ( S . stercoralis ) . Beginning in 2007 , fecal samples were also preserved in 10% formalin solution following direct identification , and later quantitatively analyzed using a modified Percoll ( Amersham Pharmacia ) technique [41] . Of the two methods , the Percoll technique is more sensitive , producing slightly higher detection rates than direct identification ( 59 . 4% vs . 51 . 9% infected ) . These differences may be due to the greater efficiency of the Percoll technique in detecting eggs in fibrous stools and at low-intensities [41] . However , for the present study the differences between the two methods were not qualitatively great enough to justify using only data produced by one method or the other . We therefore aggregated data from the two methods , coding individuals as either infected or not infected if helminths were detected by either method . In total 1 , 495 individuals had Percoll results , with the remaining 3 , 610 having only direct results . Birth dates accurate to the month were available from health clinic and school records for most Shuar children . For Shuar adults , birth dates on government identification were cross-checked with extensive genealogical information collected from multiple informants . Tsimane genealogies were collected during demographic interviews done on individuals over age 18 ( n = 1 , 098 ) . Tsimane ages were estimated based on written records , such as those kept by Catholic missionaries , demographic interviews with independent cross-checking of genealogies and reproductive histories with multiple informants , and the use of photographs of people with known ages [42] . Prior to data analysis , IgE values were converted into international units ( 1 IU = 2 . 4 ng/ml ) . IgE is log-normally distributed in all three populations ( Figure 1 ) , so values were natural log transformed ( lnIgE ) before all analyses . For t-tests , reported means are geometric means calculated by taking the exponential of the mean log values used in the t-test . Descriptive statistics and t-tests were done in PASW Statistics 18 . 0 ( formerly SPSS Statistics , SPSS Inc . ) . All other analyses were done in R 2 . 10 . 1 ( www . r-project . org ) . Generalized additive models ( GAM; [43] , [44] ) were used to examine the non-parametric age pattern of IgE levels for each population . Models were fit with the gam procedure in package mgcv using thin plate regression splines [45] , [46] . Since the cases in each population were not evenly distributed by age , initial basis knots were specified for each population based on even ten-percentiles of the age distribution , allowing knots to be spaced with an equal number of cases between them ( Figure 2 ) . Apart from the basis knots , smoothing parameters were generated automatically according to gam defaults [45] . GAM models included an intercept , a sex factor , a spline for age , and a spline for age-by-sex interaction . In initial models , similar IgE levels at birth were predicted among Shuar and NHANES , with the Shuar model predicting IgE of 7 IU/ml for females and 9 IU/ml for males , and the NHANES model predicting 15 IU/ml for females and 21 IU/ml for males . However , due to the relatively low number of Tsimane under age five , initial Tsimane models were essentially straight lines , with peak IgE predicted at birth . A number of studies have found extremely low IgE levels at birth ( <1 IU/ml ) [47]-[50] , [23] , [51] , [52] , even in infants of mothers with helminth infections and high IgE [53] , [54] . Given the convergence of the other two models and these previous findings , we used dummy cases with age zero and IgE equal to 15 IU/ml to anchor Tsimane models to a similar intercept at birth . Dummy cases were included in GAM models but not in any other statistic . GAM models with a binomial logit-link function were also used to estimate odds-ratios for Tsimane helminth infection by age . Associations between helminth infection and IgE levels were estimated in linear models controlling for infection with other helminths , sex , and age . In addition to GAM , two other methods were used to verify age shapes and compare populations . In the first , a stepwise linear regression was used to identify critical age-related changes in lnIgE for each population . Dummy variables were coded for each unique age indicating whether a case was greater than the given age ( e . g . , [55] ) . Starting from a model with only an intercept and sex term , stepAIC ( package MASS ) was used to enter and remove age variables to minimize model AIC [56] . For the second test we constructed non-linear models composed of linear segments linked together , with model terms representing the point at which the linear segments are stitched together . In this model , model terms directly represent critical ages , such as the age at which the model peaks , so differences in critical ages between populations can be tested using population interaction terms . The basic model is: Where α1 , α2 and α3 are the ages at which the slope changes and α0 equals zero , β1 , β2 , and β3 are the slopes for the segments , and additional terms β0 and βs represent the intercept at age zero and sex effect , respectively . Three logistic functions serve to “turn-on” , “turn-off” , and maintain the value reached at each age transition . The models were fit such that α1 is the initial age where rapid increases in IgE level out , α2 is the age at which IgE peaks for the population , and α3 is the age at which IgE reaches mean adult levels . Models were solved using nls ( package stats ) using the nl2sol algorithm . Initial values specified based on GAM regressions and only individuals under age 50 were used for modeling .
Of the three populations , Tsimane had the highest IgE levels ( geometric mean = 8 , 182 IU/ml ) , followed by Shuar ( 1 , 252 IU/ml ) , and NHANES ( 52 IU/ml ) ( Table 1 ) . IgE distributions were skewed but largely normalized by log-transformation ( Figure 1 ) . All three groups differed from one another in pair-wise comparisons ( all t-test p-values < . 001 after Bonferroni correction ) . In all three populations , males had higher IgE than females . NHANES males had IgE levels 60% higher than females ( 65 . 9 vs . 41 . 3 IU/ml , t = 14 . 08 , p< . 01 ) , while Shuar males had IgE values 29% higher than Shuar females ( 1 , 457 vs . 1 , 129 IU/ml , t = 2 . 15 , p = . 03 ) , and Tsimane males had levels 16% above females ( 8 , 720 vs . 7 , 527 IU/ml , t = 2 . 73 , p< . 01 ) . Upon initial visual examination of the data , age patterns were observed to be non-linear . We therefore used thin plate regression splines in GAM models to examine the age patterning of IgE ( Figure 2A ) . Age terms were significant in all models ( Tsimane: edf = 10 . 94 , F = 43 . 15 , p< . 001; Shuar: edf = 7 . 73 , F = 5 . 35 , p< . 001; NHANES: edf = 8 . 93 , F = 24 . 74 , p< . 001 ) . Despite differences in level , all three populations had similar age-related IgE profiles , characterized by a rapid increase before age five , a peak in the juvenile or adolescent period , and a decrease into adulthood . However , a number of features differ between populations . Principal among these is the age at which IgE initially peaks . Tsimane IgE peaked at 7 . 3 years for males and 7 . 2 years for females . Shuar IgE peaked at 10 . 2 for both sexes . NHANES IgE did not peak until age 16 . 9 for males and 16 . 4 for females . Fitting a linear model to the three population points for each sex suggested that for males the peak age decreases by 1 . 98 years for every one unit increase in population mean lnIgE ( t = 18 . 40 , p = 0 . 04 ) , while for females the peak age decreases by 1 . 76 years per unit increase in lnIgE ( t = 25 . 25 , p = 0 . 03 ) . We next used stepwise linear regression with ordinal age variables to identify ages at which important transitions in IgE level occur and to test the significance of these changes ( Figure 2B ) . Tsimane transitions included an increase at age three ( β = 1 . 06 , t = 1 . 90 , p = 0 . 06 ) followed by a decrease after age nine ( β = −0 . 21 , t = −2 . 26 , p = 0 . 02 ) . For Shuar , there were significant increases after age two ( β = 1 . 07 , t = 2 . 10 , p = 0 . 04 ) and age three ( β = 1 . 18 , t = 3 . 75 , p<0 . 01 ) , and a significant decrease after age eleven ( β = −0 . 41 , t = −3 . 59 , p<0 . 01 ) . In the NHANES sample increases were present in the model after age one ( β = 0 . 48 , t = 3 . 82 , p<0 . 01 ) , age three ( β = 0 . 22 , t = 1 . 60 , p = 0 . 10 ) , age four ( β = 0 . 27 , t = 1 . 90 , p = 0 . 05 ) , age six ( β = 0 . 15 , t = 1 . 67 , p = 0 . 10 ) , and age fifteen ( β = 0 . 13 , t = 2 . 00 , p = 0 . 04 ) , with a decrease after age eighteen ( β = −0 . 19 , t = −2 . 79 , p<0 . 01 ) . Since neither of these models directly tests for differences between populations or allows for the comparison of shape differences in age curves or peaks , we devised a non-linear modeling procedure in which four linear segments are used to model the age profile ( Figure 2B ) . These models include three ages points that correspond to the point at which the rapid increases in early life levels off ( α1 ) , the age at which IgE peaks in the population ( α2 ) , and the age at which IgE reaches adult levels after the peak ( α3 ) . Three slopes ( β1–3 ) describe the change in IgE between age points ( birth – α1 , α1 to α2 , and α2 to α3 ) . A sex term accounts for the difference between males and females ( βs ) . Models were first fit for the three populations independently ( Table 2 ) . Model parameters conformed well to predictions from GAM models , with peak ages ( α2 ) of 8 . 2 , 10 . 0 , and 17 . 9 predicted for Tsimane , Shuar , and NHANES respectively . The ages of initial slope change and final adult level also corresponded to peak ages , with both ages earliest in Tsimane and latest in NHANES . Moreover , all model parameters were highly correlated with IgE levels ( Figure 3 ) . Age terms , initial slopes from age zero , and sex differences all correlated with mean log IgE ( α1: r = −1 . 00 , p<0 . 01; α2: r = −0 . 98 , p = 0 . 13; α3: r = −1 . 00 , p = 0 . 01; β1: r = 0 . 99 , p = 0 . 02; βs: r = −0 . 99 , p = 0 . 09 ) , while the increase between the first peak and the final peak , and the decrease from the final peak to adult levels correlated with untransformed population geometric mean IgE ( β2: 1 . 00 , p = 0 . 06 , β3: −1 . 00 , p<0 . 01 ) . To compare populations on these terms we first attempted to fit models with population interaction terms for each parameter . However , this model , with 21 parameters , was too complex for the model algorithms and the data available , and failed to fit . Instead we simplified the models based on the relationship between model parameters and population mean IgE levels . In the first of these models we included parameter by population IgE interaction terms ( Table 3 , Model 1 ) . This model verified interactions between population IgE and all model parameters , with each one unit increase in log IgE associated with a 0 . 37 year decrease in the age of the initial slope change , a 1 . 70 year decrease in the age of peak IgE , and a 4 . 10 year decrease in the age at which levels dropped to adult mean values . The initial rate of increase in IgE was also significantly related to IgE mean levels indicating bother faster and earlier acquisition of high IgE in the Tsimane and secondarily the Shuar . In the second , third , and forth models we tested population differences in the ages at which slopes change , using population factor terms rather than interactions with population IgE . In Model 2 all three ages were left independent and the Shuar were used as a contrast group , since they lie between Tsimane and NHANES . In this model NHANES α1 and α2 were significantly later than Shuar ages , and α3 was later but with marginal significance . Although all three Tsimane ages were early than Shuar ages , none were significantly so , although all three were significantly earlier than NHANES ages when the model was run with NHANES as the contrast group ( not-shown ) . Given the strong correlation between all three ages and mean IgE levels , we suspected that multicollinearity between terms might be reducing parameter significance . We therefore examined how α1 , α2 , and α3 might be included as functions of a single age term . By fitting linear models to the parameters in Table 3 we found that α1∼1 . 47+α2×0 . 22 , and α3∼0 . 90+α2×1 . 50 . We used these terms in Model 3 , removing α1 and α3 . In this model with a single age term to describe the shape , ages in both Tsimane and NHANES were significantly different from Shuar ages , with the overall age shape shifted earlier in Tsimane and later in NHANES . Overall , 57% of Tsimane participants were infected with at least one helminth species , with hookworm ( 45 . 3% ) and A . lumbricoides ( 19 . 88% ) the most prevalent , and S . stercoralis ( 5 . 6% ) and T . trichiura ( 3 . 2% ) less common ( Table 4 ) . In order to compare the IgE age-pattern with helminth infection patterns , we examined likelihood of helminth infection by age in the Tsimane sample using logistic GAM models ( Figure 4 ) . By sex the only significant difference was in A . lumbricoides infection , with women being more likely to be infected ( 22% vs . 18% , χ2 = 15 . 5 , p< . 001 ) . By age , the odds-ratio of hookworm infection is highest in adults over age 45 , but also has a small peak at age 12 . 8 . In contrast , the odds-ratio for infection with A . lumbricoides peaks sharply at age 8 . 1 and then declines , mirroring the IgE age-pattern more closely . The odds-ratio for infection with S . stercoralis peaks somewhat later , around age 24 . 9 . The odds of T . trichiura infection is essentially flat with respect to age , reflecting the low prevalence of T . trichiura . Overall , the odds-ratio for having any type of helminth infection peaks at age 11 . 1 and then declines until age 45 , at which point it increases again . Odds-ratios closely mirror actual prevalences by age group ( Table 4 ) . For infected individuals we also examined whether egg/larva burden showed age-patterning . The only significant age-pattern was a slight decline in hookworm burden with age up to about age thirteen ( not shown ) . Other egg/larva burdens did not show age-patterning independent of changes in detection prevalence . We examined the association between helminth infection and IgE levels in our Tsimane sample using regression models to control for co-infection status , and with the sample divided by age group ( Figure 5 ) . Hookworm infection was significantly associated with higher IgE levels in 11–20 year-olds ( β = 0 . 44 , t = 2 . 85 , p<0 . 01 ) , individuals over forty ( β = 0 . 25 , t = 3 . 35 , p<0 . 01 ) , and in the overall sample ( β = 0 . 23 , t = 4 . 08 , p<0 . 01 ) . A . lumbricoides infection was significantly associated with higher IgE levels in individuals ≤10 years-old ( β = 0 . 41 , t = 1 . 96 , p = 0 . 05 ) , but with significantly lower IgE levels in 11–20 year-olds ( β = −0 . 45 , t = −2 . 45 , p = 0 . 02 ) . Although non-significant , T . trichiura infection showed a pattern similar to A . lumbricoides in those 10 and younger ( β = 0 . 84 , t = 1 . 41 , p = 0 . 16 ) . S . stercoralis was positively associated with IgE levels only considering the overall sample ( β = 0 . 32 , t = 2 . 24 , p = 0 . 03 ) . From the total sample , 459 individuals had IgE levels and full Percoll egg/larva counts . Of these , 195 were positive for hookworm , 89 for A . lumbricoides , 18 for S . stercoralis , and 16 for T . trichiura . Examining infected individuals only , egg/larva counts were not significantly correlated with lnIgE , either in the overall sample or with the sample dived by age .
We report on the age patterning of IgE in three populations: U . S . residents , Ecuadorian Shuar , and Bolivian Tsimane . The highest known IgE levels are found among lowland South American populations [18] . Tsimane IgE levels fit this pattern and resemble the levels of other South American groups with low levels of market integration ( e . g . , [14] , [57] , [58] ) . Tsimane levels are significantly higher than typical values in the United States , even for individuals reporting high levels of allergic symptoms ( based on NHANES data , analysis not shown ) . In contrast , despite inhabiting a similar neotropical environment , Shuar display lower IgE , resembling other South Americans living in rural areas [12] , [13] , [59] . However , Shuar IgE was also significantly higher than NHANES values . Although many studies have reported elevated IgE levels in populations infected with parasites such as helminths and malaria , very few have carefully characterized the age-patterning of IgE , and none that we are aware of has tested for a peak shift . A number of studies have noted that IgE is very low at birth , but increases rapidly in the first five years of life [23] , [24] , [52] , [60] . However , most of these studies have been conducted in North America or Europe , and most report that IgE is relatively stable after age five or six without characterizing the degree of stability . One of the few studies to report detailed age profiles found an initial increase to age nine , a slight decrease , and then a second peak at age fifteen in Croatian children [61] . The shape of the increase , with an initial peak and then a second peak , is remarkably similar to the age profiles seen in this study for the NHANES and Shuar sample . Our results suggest that IgE does reach an initial plateau between ages three and five , but continues to increase slowly before reaching higher peaks at age seventeen in the U . S , age ten in the Shuar , and age seven in the Tsimane . It is important to note that although we report the peaks for simplicity , the overall shape of the pattern is more important than the peak itself . This includes a faster rate of increase at an earlier age , an earlier peak , and an earlier decline to adult levels . The age-patterns we report in this study are consistent with mathematical models for what is known as the peak shift [1] , [2] , [62] . The peak shift model predicts that immunity will develop earlier in populations with higher exposure and transmission rates and subsequently decline earlier as cohorts acquire partial immunity . The peak shift hypothesis was formulated with regard to helminth infections . Typically , helminth infections peak just before or during adolescence [1] , [4] , [5] , [22] . Although data on helminth infections was only available for one of our three populations , we suggest that the IgE levels and peak ages reported in this study for Shuar and Tsimane are likely the consequence of high helminths loads since helminths are a primary cause of elevated IgE in rural populations . The IgE patterns reported also match expectations from helminth infections . The lack of helminth data for the NHANES participants may also not be much of a limitation , as what studies exist support the assumption that helminths among US residents are likely to be much less prevalent than among either Shuar or Tsimane . There are few recent estimates , but in 1972 Warren estimated that 4 . 0 million Americans were infected with A . lumbricoides , 2 . 2 million with T . trichiuris , 0 . 7 million with hookworm , and 0 . 4 million with S . stercoralis [63] . Given the US population in 1972 , these are prevalences of 1 . 9% , 1 . 0% , 0 . 3% , and 0 . 2% respectively . Hotez revises Warren's estimate for S . stercoralis to a current estimate of 68–100 , 000 or 0 . 05% of the 2008 population [64] . Similarly , of 216 , 275 stool samples sent to state laboratories in 1987 , only 0 . 8% were positive for A . lumbricoides , 1 . 2% for T . trichiuris , 1 . 5% for hookworm , and 0 . 4% for S . stercoralis [65] . A similar study examined 2 , 896 samples sent to state laboratories in 2000 and found that 0 . 4% were positive for A . lumbricoides [66] . These estimates are clearly much lower than the prevalences we report for Tsimane and the prevalences reported for other ethnic groups living near the Shuar , enough so that the exact prevalence is not critical for interpreting our findings . Due to TH2 biasing , total IgE may be a better index of total helminth load than specific IgE levels . However , the lack of parasite-specific IgE in these data sets is also a limitation in that we cannot state how much parasite-specific IgE contributes to total levels . It may be that Shuar and Tsimane differ less in the total helminth prevalences than they do in prevalences of particular helminth species . Using helminth infection data for the Tsimane we were able to examine associations between helminth species and total IgE . We found that the overall age-pattern for IgE in the Tsimane resembled the age-pattern for A . lumbricoides infection . A . lumbricoides infection was associated with higher IgE levels in children age 3–10 , but with lower IgE levels in 11–20 year-olds . These data suggest that this species may contribute more to the age-pattern of IgE than others . Hookworm and Strongyloides infection were also associated with higher IgE in the overall sample , but showed less age-pattern in association . Future studies will need to investigate this in more detail by examining specific-IgE and extending into other populations . The association between IgE and A . lumbricoides is consistent with other studies showing that total IgE is correlated with specific IgE to A . lumbricoides [14] . The positive association between A . lumbricoides and IgE in participants under age ten and the negative association after age ten may also suggest that IgE conveys partial immunity to A . lumbricoides [67] , [68] . Other studies that [69] have failed to find increased resistance with higher IgE may have not taken this age-pattern into account . Other parasites , such as Plasmodium falciparum , also raise total IgE levels [70] . However , malaria is unlikely to be an important factor for the populations studied in this paper . Although malaria is present in parts of Shuar territory , it is not present in the villages where the data for this paper were collected , and very few individuals in the area report having had it . Malaria also appears to be absent from the Tsimane territories , with no Tsimane reporting malaria in extensive health interviews . Finally , in all three populations IgE levels were higher in males . Although noted in many studies ( e . g . , [23] ) , the reason for this sex difference is not entirely clear . The only significant sex difference in helminth infections was in A . lumbricoides , with slightly more women being infected . Due to the higher IgE in males , it is tempting to hypothesize that this is due to increased resistance in males . However at present this is merely supposition . It is just as likely that Tsimane women are infected more frequently because they spend more time in direct contact with children , who themselves have the greatest number of A . lumbricoides infections . In addition to its importance for theoretical models describing the epidemiology of infections , an understanding of the age patterning of IgE may have public health implications . In populations with higher parasite transmission rates , exposure triggers an elevation of IgE at earlier ages . More rapid and heavy investment in earlier immunocompetence may be favored with high exposure , even at the expense of other investments , such as growth . In Shuar children high IgE levels are associated with increased stunting [18] . It seems plausible that insults to growth may be most pronounced in populations in which peak infection rates occur during critical growth periods , such as early adolescence . Additionally , the timing of infection may affect the development of immune function in other ways , for example by affecting the TH1/TH2 balance , with consequences for the later development of allergy [71] . Although these hypotheses remain to be tested , they suggest that interventions might be developed with the specific goal of shifting infection peaks toward less critical ages . | Infection with parasitic worms , known as helminths , alters the immune system , causing individuals to produce high levels of a type of antibody known as immunoglobulin E ( IgE ) . IgE is typically very low in western populations , but is many times higher where helminth infections are common , particularly indigenous populations in South America . Helminths infect more than one seventh of the world's population . Since helminths tend to infect people at younger ages in areas where they are more common , a disproportionate number of those affected are schoolchildren . In this paper we examine IgE levels in two indigenous South American groups in comparison to levels in the United States . In these groups we find that IgE levels are not only higher , but that they also reach their highest levels at earlier ages in more infected populations . This finding is important since effects on immune function , including IgE production , may have additional consequences if they occur at young ages , changing the development of allergy and asthma , growth , and response to vaccines . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"social",
"and",
"behavioral",
"sciences",
"immunology",
"anthropology",
"parasitic",
"diseases",
"helminth",
"infection",
"hookworm",
"infection",
"neglected",
"tropical",
"diseases",
"ascariasis",
"infectious",
"diseases",
"biological",
"anthropology",
"hookworm",
"immune",
"response",
"trichuriasis",
"clinical",
"immunology",
"immunoglobulins"
] | 2011 | Evidence for a Peak Shift in a Humoral Response to Helminths: Age Profiles of IgE in the Shuar of Ecuador, the Tsimane of Bolivia, and the U.S. NHANES |
To search for virulence effector genes of the rice blast fungus , Magnaporthe oryzae , we carried out a large-scale targeted disruption of genes for 78 putative secreted proteins that are expressed during the early stages of infection of M . oryzae . Disruption of the majority of genes did not affect growth , conidiation , or pathogenicity of M . oryzae . One exception was the gene MC69 . The mc69 mutant showed a severe reduction in blast symptoms on rice and barley , indicating the importance of MC69 for pathogenicity of M . oryzae . The mc69 mutant did not exhibit changes in saprophytic growth and conidiation . Microscopic analysis of infection behavior in the mc69 mutant revealed that MC69 is dispensable for appressorium formation . However , mc69 mutant failed to develop invasive hyphae after appressorium formation in rice leaf sheath , indicating a critical role of MC69 in interaction with host plants . MC69 encodes a hypothetical 54 amino acids protein with a signal peptide . Live-cell imaging suggested that fluorescently labeled MC69 was not translocated into rice cytoplasm . Site-directed mutagenesis of two conserved cysteine residues ( Cys36 and Cys46 ) in the mature MC69 impaired function of MC69 without affecting its secretion , suggesting the importance of the disulfide bond in MC69 pathogenicity function . Furthermore , deletion of the MC69 orthologous gene reduced pathogenicity of the cucumber anthracnose fungus Colletotrichum orbiculare on both cucumber and Nicotiana benthamiana leaves . We conclude that MC69 is a secreted pathogenicity protein commonly required for infection of two different plant pathogenic fungi , M . oryzae and C . orbiculare pathogenic on monocot and dicot plants , respectively .
Rice blast , caused by an ascomycete fungus Magnaporthe oryzae , is the most severe fungal disease of rice throughout the world [1] . Genetic studies of this pathogen over the last two decades have made the Magnaporthe-rice pathosystem an excellent model for investigating fungus-plant interactions . Plants are equipped to sense evolutionarily conserved microbial molecular signatures , collectively called Pathogen-Associated Molecular Patterns ( PAMPs ) or Microbe-Associated Molecular Patterns ( MAMPs ) , and activate PAMP-Triggered Immunity ( PTI ) [2]–[4] . Pathogens are capable of inhibiting PTI on their host plants by delivering virulence effector proteins into host cells [5]–[9] . In M . oryzae , effector secretion machinery has recently been elucidated [10]–[12] . A Golgi-localized P-type ATPase-encoding gene , MgAPT2 is required for exocytosis during plant infection . Further analysis suggested that MgAPT2 is involved in secretion of a range of extracellular enzymes as well as an AVR effector for the rapid induction of host defense responses in an incompatible reaction in rice cultivar IR-68 [10] . Another study demonstrated that M . oryzae mutants with a defect in an ER chaperone-encoding gene , LHS1 , have reduced activities of extracellular enzymes and secretion of AVR-Pita1 [13] , [14] blocking Pi-ta R-gene-mediated hypersensitive response . The contribution of LHS1 to protein translocation and secretion of proteins , including effectors , revealed the importance of ER chaperones for successful disease development by rice blast fungus [12] . Live-cell imaging revealed development of the biotrophic interfacial complex ( BIC ) , a structure that accumulates fluorescently labeled effectors secreted by invasive hyphae ( IH ) . The examined BIC-localized secreted proteins were translocated into rice cytoplasm . By contrast , a biotrophy-associated secreted protein BAS4 , which uniformly outlines the IH , was not translocated into the host cytoplasm [11] . These results suggest that BIC represents the site of effector translocation in rice blast disease [11] . Several effector protein genes have been cloned and characterized from M . oryzae but all of them were avirulence ( AVR ) effectors with no virulence function elucidated to date [13]–[20] except for a recently identified virulence effector protein , Slp1 [21] . Slp1 accumulates at the interface between the fungal cell wall and the rice plasma membrane , can bind to chitin , and is able to suppress chitin-induced plant immune responses , including generation of reactive oxygen species and plant defense gene expression [21] . Several effector candidates were identified by using interaction transcriptome in the biotrophic invasion of M . oryzae [22] . In the paper the authors have identified a known effector PWL2 as well as 58 candidate effectors showing >10-fold increase in the expression in the biotrophic invasive hyphae relative to control mycelia using M . oryzae oligoarrays . Four of these candidates were confirmed to be fungal biotrophy-associated secreted proteins [22] . However , virulence function of all the candidates has not been elucidated , and comprehensive gene disruption analyses of the candidates have not been carried out . Therefore , in this study we employed a large-scale disruption analysis of M . oryzae secreted protein genes to search for novel virulence effectors . Whole-genome draft sequence of M . oryzae was published for the isolate 70-15 , a laboratory strain [23] . The genome assembly consists of 37 . 8 Mb nucleotides encoding 11 , 109 predicted protein coding genes . We recently retrieved 1 , 306 putative secreted protein genes from the predicted proteome of 70-15 [20] . From these , a total of 78 genes expressed in the fungus were disrupted and analyzed . We found that disruptants of the 77 genes did not show change in pathogenicity as compared to the wild-type strains . Disruption of only one gene , MC69 , showed a severe reduction in pathogenicity . Further analysis showed that MC69 protein is involved in the full pathogenicity of M . oryzae after the penetration stage of infection .
To search for effector protein genes of Magnaporthe oryzae , we carried out a large-scale targeted gene disruption analysis of the 78 putative secreted protein genes that are expressed during infection ( Table 1 ) . Initially we selected 1 , 306 putative secreted protein genes as described previously [20] . The 78 genes subjected to functional study were selected on the basis of their confirmed expression in the pathogen at the early stages of infection . We focused on secreted protein genes involved in the two stages of infection of M . oryzae . One is the appressorium formation stage and the other is the biotrophic invasion stage . Treatment with cyclic AMP ( cAMP ) induces appressorium formation on hydrophilic surface [24] . By SAGE ( Serial Analysis of Gene Expression ) of cAMP-treated conidia on hydrophilic membrane 6 h after the start of treatment [25] , we identified several pathogenicity genes , e . g . MPG1 , MAS1 and MAC1 , already characterized [26]–[28] and thought to be involved in pathogen-host interaction . Therefore , we assumed that a part of effector genes should be expressed during the appressorium formation . To achieve a high efficiency tag-to-gene annotation , we established SuperSAGE method that extracts a 26-bp tag from each cDNA [29] . SuperSAGE of the cAMP-treated M . oryzae strain 70-15 has been done in this study to search for novel effector candidates ( Table S1 ) . Furthermore , we also used the SuperSAGE data of invasive hyphae for searching new effectors ( Supplemental Data Set 1 in [20] ) . Indeed , this SuperSAGE analysis revealed that two AVR effector genes , AVR-Pia and AVR-Pii were expressed at the stage of invasive hyphae ( Supplemental Data Set 2 in [20] ) . To investigate the function of the effector candidate genes , we generated disruption mutants for each of the selected 78 genes ( Table 1 ) in M . oryzae by TAG-KO method [30] , [31] . To assess the virulence of each mutant , conidial suspension of each mutant was sprayed onto seedlings of a barley cultivar Nigrate , which is susceptible to the wild type M . oryzae . Blast phenotypes of barley infected by KO mutants of all the genes except for MC69 gene were the same as that infected by wild-type strain Ina72 . Similarly the 77 mutants did not show reduced virulence in a susceptible rice cultivar Shin No . 2 . By contrast , we observed a dramatic reduction in disease symptoms on barley cotyledons and the susceptible rice cultivar , Shin No . 2 , inoculated with all of the three independent mc69 mutant lines ( Figure 1A ) . Consequently , we identified the MC69 gene ( MGG_02848 . 6 ) as required for pathogenicity of M . oryzae after a large-scale targeted gene disruption analysis of the 78 putative secreted protein genes . In summary , we found that targeted disruption of MC69 affected pathogenicity of M . oryzae and disruption of the other 77 genes had no effect on its pathogenicity . To investigate the physiological and molecular function of MC69 in detail , we generated MC69 disruptants in M . oryzae strain Ina72 by targeted gene disruption as described above . Colony growth , color and the production of conidia were the same as the wild-type strain ( Figure S2A and B ) . We observed a remarkable reduction in disease symptoms on barley and rice inoculated with the mc69 mutants compared to those inoculated with the wild type strain 4 and 7 days after inoculation suggesting an important role of MC69 in fungal pathogenicity ( Figure 1A ) . Subsequently , we performed a detailed phenotypic analysis of the mc69 mutants . The mc69 mutants exhibited a defect in appressorium-mediated penetration in rice leaf sheath cells but neither in conidial germination nor appressorium formation ( Figure 1B and C ) . We studied 200 appressoria of each mc69 mutant , 97∼99% with failed penetration ( no visible hyphae ) and 1∼2% with post-penetration blockage . Therefore , we conclude that MC69 is required for appressorial penetration and pathogenicity of M . oryzae . Although most of appressoria formed by the mc69 mutants could neither penetrate nor produce infectious hyphae in the inoculated rice leaf sheath cells , we further analyzed invasive growth at 50 appressorial penetration sites by rating the hyphal growth from the level 1 ( low ) to 4 ( high; see Materials and Methods , Figure 1D ) . In Ina72 WT , 84% of penetration sites showed invasive growth levels 3 or 4 , by contrast in the mc69 mutant infectious growth within the inner epidermal tissue was relatively limited ( levels 2 and 3 ) 32 hours after inoculation , suggesting that the loss of MC69 also affects infectious growth to some extent at the post-penetration stage ( Figure 1D ) . To test whether the observed phenotypes of the mc69 mutants were solely caused by disruption of MC69 , an intact copy of MC69 was introduced into the mc69 mutant mc69-87 ( Figure S1B ) for complementation . The MC69-reintegrated strain showed normal appressorial penetration rate and the strain developed blast disease symptoms on barley and rice leaves with a similar extent to the wild type ( Figure 1E and F ) . These results demonstrate that disruption of MC69 gene caused defects in appressorial penetration and development of blast symptoms by M . oryzae . MC69 was found in the SuperSAGE list of the cAMP-treated conidia ( Table S1 ) . MC69-EST was found in MGOS databases for mycelium , conidia , germinated conidia and appressoria [32] . RL-SAGE tags of MC69 were also found in the fungus grown on a minimum medium for three days [33] . These data suggest that MC69 is constitutively expressed in M . oryzae . To investigate the expression pattern of MC69 in detail , we produced an M . oryzae strain Ina72 harboring a vector containing the MC69 promoter fused with a reporter protein gene mCherry ( MC69p::mCherry; Figure 2A ) generating WT+mCherry . The mCherry fluorescence was observed in all morphological stages with enhanced fluorescence in conidia before germination ( 0 h ) and matured appressoria ( 12 h after incubation ) on glass coverslips under confocal laser-scanning microscope ( Figure 2B ) . To determine the mode of expression and spatial localization of the MC69 protein , a construct MC69p::MC69::mCherry was prepared ( Figure 2D ) and used for transformation of the mc69 mutant generating mc69+MC69::mCherry , and the mCherry fluorescence was then observed on glass coverslips ( Figure 2E ) . The transgenic M . oryzae mutant mc69 expressing MC69::mCherry restored pathogenicity ( Figure 2G ) , showing that the fusion protein MC69::mCherry is functional for infectivity . The mCherry fluorescence was detected in all the developmental stages like WT+mCherry , but the intensity of fluorescence in the strain mc69+MC69::mCherry was weaker than that of the strain WT+mCherry presumably because of secretion and diffusion of MC69::mCherry fusion protein ( Figure 2B and E ) . To observe the fluorescence of WT+mCherry and mc69+MC69::mCherry in the infected tissues , we inoculated these conidial suspensions to rice leaf sheath . The mCherry fluorescence was detected in the invaded hyphae 24 and 48 hours after inoculation with WT+mCherry , but not detected in that inoculation with mc69+MC69::mCherry ( Figure 2C and F ) . These results suggest that the MC69 gene is expressed throughout infection: conidia , infection-related morphogenesis and subsequent growth stage . The fluorescence from MC69::mCherry fusion proteins was not detected in the invaded hyphae in planta presumably because they have been secreted . To obtain direct evidence that the MC69 protein is actually produced in the invasive hyphae , we made mc69 mutant harboring a construct MC69p::MC69::HA ( mc69+MC69::HA ) or a MC69p::MC69::3xFLAG ( mc69+MC69::3xFLAG ) to perform immunodetection of MC69::HA or MC69::3xFLAG proteins in planta , respectively . Both of mc69+MC69::HA and mc69+MC69::3xFLAG restored appressorial penetration and invasive growth in rice leaf sheath ( Figure S3A and C ) , showing that the fusion proteins MC69::HA and MC69::3xFLAG are functional for pathogenicity . To clarify whether the MC69::HA and MC69::3xFLAG are expressed or not , mc69+MC69::HA- and mc69+MC69::3xFLAG-infected rice leaf sheath extracts were analyzed by SDS-PAGE gel blot analysis . We extracted total protein at 24 and 48 hours after leaf sheath inoculation . Note that at 24 hours after inoculation , most conidia develop appressoria but hyphal invasion is still limited , whereas at 48 hours extensive hyphal growth develops . Both of HA- and 3xFLAG-tagged MC69 were detected only faintly 24 hours after inoculation , but these proteins were abundant 48 hours after inoculation ( Figure S3B and D ) indicating that MC69 protein was indeed produced in invasive hyphae . Furthermore , we tried to express an AVR effector gene , AVR-Pia [18] , [20] from the MC69 promoter to see whether AVR-Pia avirulence function is supported by MC69 promoter in rice plant harboring Pia R-gene . We hypothesized that only when MC69 promoter allows expression of AVR-Pia in invasive hyphae , a sufficient amount of AVR-Pia protein would be translocated inside rice cells to be recognized by Pia , NBS-LRR-type cytoplasmic R-proteins [34] . We performed transformation of the M . oryzae isolate Ina86-137 ( that lacks AVR-Pia and can infect rice cultivars possessing Pia R-gene ) [20] with a construct MC69p::AVR-Pia . We used the wild-type stain ( Ina86-137 WT ) and a transformant harboring an intact copy of AVR-Pia ( +AVR-Piap::AVR-Pia ) [20] as negative and positive controls , respectively . In contrast with the Ina86-137 WT , +AVR-Piap::AVR-Pia and the transformants harboring MC69p::AVR-Pia ( +MC69p::AVR-Pia-1 and -2 ) failed to cause disease in the rice cultivar Sasanishiki possessing Pia ( Figure S4A ) . Both Ina72-WT and the transformants successfully infected rice cultivar Shin No . 2 that lacks Pia , suggesting that their inability in infecting cv Sasanishiki was caused by Pia-AVR-Pia interactions and that the MC69 promoter is expressed during invasive growth . Active transcription of the AVR-Pia in the transformants was confirmed by RT-PCR ( Figure S4B ) [20] . Khang et al . ( 2010 ) reported that secreted fluorescent effectors preferentially accumulate in biotrophic interfacial complexes ( BICs ) at the invasive hyphae-rice cell interface [11] . By fusing nuclear localization signal ( NLS ) to the fluorescent effectors to facilitate visualization of translocation , they also showed that the two BIC-localized secreted proteins , PWL2 and BAS1 were translocated into rice cytoplasm [11] . To test translocation of MC69::mCherry in rice cells , we added a modified small NLS from simian virus large T-antigen [35] at the C terminus of the MC69::mCherry fusion downstream of the PWL2 promoter ( PWL2p::MC69::mCherry::NLS ) and transformed M . oryzae strain Sasa2 with the construct . A transformant M . oryzae harboring PWL2p::PWL2::mCherry::NLS was used as positive control . PWL2::mCherry::NLS exhibited significant fluorescence in BIC and in nuclei of invaded host cells at successful infection sites 24 , 27 and 32 hours after inoculation ( Figure 3A ) whereas MC69::mCherry::NLS did not show fluorescence in nuclei of the invaded rice cells , but showed weaker fluorescence in BIC than that of PWL2::mCherry::NLS ( Figure 3B ) . To eliminate the possibility that the NLS influences BIC localization , a transformant strain harboring PWL2p::MC69::mCherry was inoculated to rice leaf sheath . The result showed that MC69::mCherry was also detected in the BIC ( Figure 3C ) . In addition , we observed mCherry fluorescence with different pinhole settings to compare the signals in the BIC among the three strains 27 hours after inoculation . The result showed that BIC accumulation signals of MC69::mCherry::NLS and MC69::mCherry were significantly weaker than that of PWL2::mCherry::NLS ( Figure S5 ) . These results suggest that the MC69 does not translocate into the infected rice cells , but localizes in BIC , however the accumulation level of MC69 in BIC is significantly lower than that of PWL2 . MC69 homologs were found in other filamentous fungi Colletotrichum orbiculare ( AB669186 ) , Glomerella graminicola ( EFQ29542 ) , Verticillium albo-atrum ( EEY15898 ) , V . dahliae ( EGY20943 ) , Neurospora crassa ( XP_965292 ) , N . tetrasperma ( EGO52621 ) , Myceliophthora thermophila ( XP_003659994 ) , Podospora anserina ( XP_00190740 ) , Grosmannia clavigera ( EFX05010 ) , Fusarium oxysporum ( EGU75378 ) , Gibberella zeae ( XP_388669 ) , Trichoderma atroviride ( EHK44387 ) , T . virens ( EHK23962 ) , Metarhizium acridum ( EFY93067 ) , M . anisopliae ( EFY97094 ) and Cordyceps militaris ( EGX95034 ) ( Figure S6 and S7 ) . However , these amino acid sequences did not contain known domains/motifs that would allow the prediction of their function . Nevertheless , MC69 homologs contain two conserved cysteine residues in the mature protein region C-terminal to the signal peptide ( Figure S6 ) . A software DISULFIND ( http://disulfind . dsi . unifi . it/; [36] ) predicted that the two cysteine residues in mature MC69 can form a disulfide bond ( Figure 4A ) . To test whether these cysteines are necessary for MC69 function , mutant alleles of MC69 were generated in which each or both of C36 and C46 were replaced with alanine ( Figure 4B ) . Mutant alleles with one amino acid replacement ( MC69 ( C36A ) ; MC69 ( C46A ) ) or two replacements ( MC69 ( C36A , C46A ) ) were expressed in the mc69 mutant ( mc69+MC69 ( C36A ) , mc69+MC69 ( C46A ) or mc69+MC69 ( C36A , C46A ) ) . In all cases , appressorial penetration rate and blast symptoms on barley and rice were slightly restored , but still significantly reduced as compared to the wild type ( Figure 4C and D ) . In addition , we further analyzed invasive growth rating of the 50 appressorial penetration sites . Infectious growth of mc69+MC69 ( C36A ) , mc69+MC69 ( C46A ) and mc69+MC69 ( C36A , C46A ) within the inner epidermal tissue was slightly restored as compared to the mc69 mutant ( Figure S8 ) . These results indicate that C36 and C46 , presumably involved in disulfide bond , are necessary for MC69 to exert its pathogenicity in M . oryzae . To see whether C36 and C46 are important for MC69 secretion/localization , spatial localization of the MC69 ( C36A ) protein was tested by transforming mc69 mutant with a construct MC69p::MC69 ( C36A ) ::mCherry , resulting in mc69+MC69 ( C36A ) ::mCherry ( Figure 5A ) . We inoculated conidial suspension of the strain to rice leaf sheath to observe the mCherry fluorescence in the infected tissue . The mCherry fluorescence was detected in appressoria but not in the invaded hyphae 24 and 48 hours after inoculation ( Figure 5B ) . The result suggests that the MC69 ( C36A ) ::mCherry protein was secreted into the plant and diffused below the detection limit like MC69::mCherry ( Figure 2F and 5B ) . To clarify whether the MC69::mCherry and MC69 ( C36A ) ::mCherry are secreted or not , extracellular proteins secreted by Magnaporthe after liquid culture were analyzed by SDS-PAGE gel blot analysis . We used the wild-type strain expressing mCherry under MC69 promoter ( WT+mCherry; Figure 2B and C ) as negative control . Western blot analysis ( Figure 5C ) revealed the presence of mCherry-tagged MC69 and mCherry-tagged MC69 ( C36A ) in the culture medium . Faint signals of cleaved mCherry were observed as well for the transformants mc69+MC69::mCherry and mc69+MC69 ( C36A ) ::mCherry . The molecular weight ( MW ) of the fusion proteins was around 30 kDa , in line with the predicted MW of mature MC69::mCherry . These data strongly suggest that both of MC69 and MC69 ( C36A ) are secreted to the medium , and C36 is not important for MC69 secretion . It could be possible that mutation of the Cys residues may impact pathogenicity by reducing the stability of the protein after secretion in planta . To investigate whether pathogenicity function of MC69 is conserved in M . oryzae , we produced mc69 disruptants in other two Japanese field isolates of TH68-141 and Hoku1 , in addition to the isolate Ina72 . We used the MC69 knockout vector used for Ina72 to generate mc69 disruptants of both TH68-141 and Hoku1 isolates . Generated mc69 disruptants of the two isolates showed a reduced pathogenicity on barley leaves as compared to the wild type strains ( Figure 6 ) , indicating the importance of MC69 in virulence of TH68-141 and Hoku1 . The whole genome sequence of 70-15 , a well-studied laboratory strain of M . oryzae , was published [23] . We found that 70-15 showed poor virulence as compared to the Japanese strains in the previous study . It caused intermediate responses in all of the 13 tested rice cultivars: infection caused reddish lesions of various sizes , but they did not further develop into typical susceptible brown spindle-shaped necrotic lesions [20] . To investigate whether MC69 is required for pathogenicity in 70-15 , MC69 gene disruption analysis was performed in the 70-15 background ( Figure S1 ) . Two independent MC69-KO lines ( mc69-119 and mc69-31 ) and wild-type 70-15 were sprayed onto barley cotyledons and rice leaves . The barley and rice infected by mc69 mutants showed much weaker symptoms as compared to the 70-15-infected plants ( Figure 7A ) , indicating the importance of MC69 in pathogenicity of 70-15 . Appressorial penetration rates of the mutants in rice leaf sheath cells were significantly lower than that of 70-15 but the rates of germination and appressorium formation were same with the wild type ( Figure 7B ) . In addition , we further analyzed invasive growth rating of 50 appressorial penetration sites . Infectious growth of the mc69 mutants within the inner epidermal tissue was restricted as compared to the wild type ( Figure S9 ) . However , the colony growth and conidiation of the mutants on oatmeal agar media were similar to the wild type ( Figure S2C and D ) . Thus , the mc69 mutants of 70-15 have a defect in appressorial penetration and development of blast symptoms , which is similar to the phenotype of the mc69 disruptants of Ina72 . These results suggest that MC69 is commonly required for appressorial penetration and subsequent colonization in various M . oryzae strains . The importance of MC69 in M . oryzae raised a possibility that MC69 orthologs are also involved in pathogenicity of other fungal pathogens . To assess this point , we investigated whether MC69 ortholog is involved in pathogenicity of the cucumber anthracnose fungus C . orbiculare ( Figure S6 ) . A gene homologous to MC69 was isolated from C . orbiculare in this study . The isolated gene , designated CoMC69 , comprises 220 bp interrupted by an intron and encodes a predicted protein of 54 amino acids ( Figure 8A ) . Intron/exon organization in MC69 orthologs in filamentous fungi indicated that most of them have one intron only followed by an exon ( 140–156 bp ) except for the genes in T . virens and Gibberella zeae ( Figure S10 ) . First exons in all genes encode a common region containing two conserved cysteine residues in the mature proteins ( Figure S6 ) . To investigate whether CoMC69 is involved in the pathogenicity of C . orbiculare , we produced CoMC69 disruption mutants . The plasmid pCBGDMC69 was designed to replace the CoMC69 gene in the wild-type strain 104-T through double crossover homologous recombination ( Figure S11A and B ) . The colony morphology and conidiogenesis of Comc69 mutants grown on PDA medium were similar to that of 104-T ( Figure S11C and data not shown ) . We next investigated their pathogenicity on host cucumber leaves . Conidial suspensions from the Comc69 mutants were spotted on detached cucumber leaves and incubated for 7 days . The Comc69 mutants exhibited clear reduction in lesion development in comparison with the wild-type strain 104-T ( Figure 8B ) . C . orbiculare 104-T is able to infect Nicotiana benthamiana , which is not closely related to cucumber [37] . The Comc69 mutants also exhibited reduced pathogenicity on N . benthamiana ( Figure 8C ) . These results indicate that CoMC69 is required for pathogenicity of C . orbiculare , suggesting conserved roles of the MC69 proteins in pathogenicity of both M . oryzae and C . orbiculare . To investigate the gene expression of CoMC69 in plant infection of C . orbiculare , we generated C . orbiculare strains carrying a reporter plasmid containing the 1 . 4 kb 5′ upstream region of CoMC69 fused with mCherry . As a result , we found the mCherry fluorescence in appressoria and primary intracellular hyphae of the transgenic C . orbiculare , indicating the expression of CoMC69 in the plant infection stage of C . orbiculare ( Figure 8D ) .
In this study , we show that MC69 , a novel secreted protein of Magnaporthe oryzae , is essential for successful appressorial penetration and blast symptom development in rice and barley cultivars . The MC69 gene ( MGG_02848 . 6 ) resides on chromosome VII of the M . oryzae genome . The MC69 protein comprises 54 amino acids and is predicted to harbor a putative N-terminal secretion signal peptide ( Figure 4B ) . MC69 seems to be a solitary gene without any paralogs in the genome . It lacks known sequence motifs associated with enzymatic function . Although MC69 homologs were found in other filamentous fungi ( Figure S6 ) , their functions are also not known . Expression of MC69 was observed in mycelia , conidia and all stages of infection ( Figure 2 ) . The mc69 disruptants were unable to invade plant cells to establish compatible interaction with the host plant ( Figure 1 , 6 and 7 ) . However in mc69 mutants , other phases of infection-related development such as conidial germination and appressorium formation were unaffected ( Figure 1B and 7B ) . MC69 of M . oryzae , which is predicted to comprise 38 amino acids after cleavage of the signal peptide , contained no known functional domains . Therefore it is unlikely that MC69 has an enzymatic function . The two cysteine residues ( C36 and C46 ) conserved among the MC69 homologs may be involved in disulfide bridge formation and were shown to be important for pathogenicity function of MC69 ( Figure 4 and S6 ) . Pep1 is a novel effector protein from the corn smut fungus Ustilago maydis that is essential during penetration . Disruption mutants of pep1 are not affected in saprophytic growth and develop normal infection structures , but are arrested during the penetration of epidermal cells of maize leaves . In addition , two of the four cysteine residues in Pep1 were shown to be essential for the virulence function [38] . The authors consider that the importance of two of the four cysteine residues for secretion of Pep1 to make a compact structure with disulfide bridge structure of Pep1 [38] . To address this possibility we observed localization of MC69::mCherry and MC69 ( C36A ) ::mCherry in M . oryzae . We were unable to detect red fluorescence in infectious hyphae and the fusion proteins were detected in the culture filtrate of both strains ( Figure 2 and 5 ) . These results indicated that the substitution of cysteine residues of MC69 did not affect secretion , but affected pathogenicity of M . oryzae . On the other hand , disruption of a total of 77 secreted protein genes in M . oryzae did not affect its pathogenicity within our experimental condition . Since there is no systematic bias in our selection of secreted protein genes for disruption , we extrapolate that 77/78 = 99% of secreted protein genes do not show clear reduction in pathogenicity even after knockout . Several secreted avirulence ( AVR ) effector genes have been isolated from M . oryzae , including PWL effectors [16] , [19] , AVR-Pita [13] , [14] , AVR1-CO39 [15] , AVR-Piz-t [17] , AVR-Pia , AVR-Pii and AVR-Pik/km/kp [18] , [20] , but the virulence functions of the genes are still unknown . In fact , the AVR-Pita effector is dispensable for virulence on rice [14] , [39] . According to our results and available information on M . oryzae AVR effectors , we hypothesize two mutually non-exclusive possibilities: ( 1 ) virulence contribution of most of effectors is too small to be detected by conventional assays; ( 2 ) effectors have redundant activities and more than one effector participate in the same virulence pathway . A recent report of M . oryzae indicates that the fungus overcomes the first line of defense ( PAMP-Triggered Immunity ) by secreting an effector protein , Slp1 during invasion of new rice cells [21] . There are several reports for secreted effectors of other fungal pathogens . Pep1 of U . maydis was described above . Several hydrophobins or repellent genes that encode secreted proteins of U . maydis were examined for their roles in virulence . Single knock-outs of these genes did not affect virulence , but a double knockout of the repellent-encoding gene Rsp1 and Hum3 ( a gene encoding a protein containing both , a hydrophobin domain and a repellent region ) were arrested at an early stage of penetration . This indicates that Rsp1 and Hum3 are effectors with a partly redundant virulence function during the early stages of infection [40] . We speculate a similar situation may occur in M . oryzae . It would be a good way to focus on effector candidates exhibiting higher similarities and knockout or silence multiple genes simultaneously , to identify the multiple effectors that act redundantly . We showed that MC69 was required for pathogenicity of the additional three strains of M . oryzae in addition to the strain Ina72 , indicating conserved roles of MC69 in M . oryzae . To investigate whether the pathogenicity function of the MC69 ortholog in the well-studied dicot fungal pathogen , we isolated an ortholog , CoMC69 from the cucumber anthracnose fungus , Colletotrichum orbiculare . Notably , in C . orbiculare the deletion of CoMC69 reduced pathogenicity on the hosts cucumber and Nicotiana benthamiana leaves ( Figure 8 ) . Phylogenetic analyses were performed with M . oryzae MC69 ( MoMC69 ) , with 16 homologs from other phytopahogenic ( C . orbiculare , Glomerella graminicola , Verticillium albo-atrum , V . dahliae , Grossmannia clavigera , Fusarium oxysporum and Gibberella zeae ) , entomopathogenic ( Metarhizium acridum and M . anisopliae ) , caterpillar killer ( Cordyceps militaris ) , fungal parasite ( Trichoderma atroviride and T . virens ) or saprophytic ( Neurospora crassa , N . tetrasperma , Myceliophthora thermophila and Podospora anserina ) fungi ( Figure S7 ) . MoMC69 and CoMC69 are closely related to the Verticillium wilt pathogens V . albo-atrum , V . dahliae and the cereal plants anthracnose fungus Glomerella graminicola . A conserved motif containing the two cysteine residues showed high homology among all MC69 homologs ( Figure S6 ) . Thus , it will be interesting to determine whether genes orthologous to MoMC69 also contribute to the pathogenicity of various plant , fungus , entomo or caterpillar pathogenic fungi . However , MC69 orthologs also occur in saprophytes suggesting that a possibility that the primary function of the protein is in relation to the structure or function of the fungus itself , and that the function must be intact for the fungus to succeed as a pathogen . We generated transgenic rice overexpressing MC69 to examine susceptibility to M . oryzae wild-type strain or the mc69 mutant infection . However , overexpression of MC69 in rice neither enhanced the pathogenicity of M . oryzae wild-type strain nor complemented the pathogenicity deficiency of the mc69 mutant ( data not shown ) . We hypothesize three possibilities why overexpression of MC69 did not affect M . oryzae wild-type strain and the mc69 mutant infection . One possibility is that the localization of MC69 in the infection sites of M . oryzae in rice cells is important . We produced M . oryzae transformant harboring PWL2p::MC69::mCherry . After inoculation of the strain to the rice leaf sheath , mCherry fluorescence was detected in biotrophic interfacial complex ( BIC ) ( Figure 3C ) . However , MC69::mCherry fusion protein expressed by MC69 promoter was not detected in BIC ( Figure 2F ) . It might be because MC69 promoter activity was weaker than PWL2 promoter activity or PWL2 promoter leads MC69::mCherry to BIC accumulation but MC69 promoter did not . To test these possibilities , mc69 mutant expressing MC69::EGFP fusion protein downstream of MC69 promoter has been produced because EGFP fluorescence was relatively stronger than mCherry fluorescence . When the transformant was inoculated to the rice leaf sheath , MC69::EGFP was shown to be accumulated to the BIC ( data not shown ) . These results indicate that BIC localization of MC69 is important for virulence of M . oryzae . Therefore , ectopic overexpression of MC69 in rice neither enhanced pathogenicity of wild-type strain nor complemented deficiency of mc69 mutant of M . oryzae in trans because of the MC69 protein would not be localized in BIC . The second possibility is that post translational modification of MC69 protein in M . oryzae might be different from that in planta even though the secreted MC69::mChrerry protein shows an expected molecular size ( Figure 5C ) . The third possibility is that MC69 affects the physiology of the fungus but does not directly affect the physiology of the plant so that expression of MC69 in rice did not complement the defect in mc69 mutant of M . oryzae . Khang et al . ( 2010 ) demonstrated that BIC-localized secreted proteins PWL2 and BAS1 were translocated into the rice cytoplasm but a secreted protein BAS4 , which uniformly outlines the invasive hyphae , was not [11] . Interestingly , when M . oryzae transformants secreted fluorescent MC69 fusion protein during epidermal cell invasion , the fluorescent protein was observed in BICs ( Figure 3 ) . To investigate whether this feature is specific to MC69 or not , we generated M . oryzae strains expressing fluorescence-labeled version of four putative secreted proteins ( HMM14 , MC55 , GAS1 or GAS2 ) with NLS . HMM14 and MC55 are the secreted protein genes studied here ( Table 1 ) and GAS1 and GAS2 are secreted protein genes involved in virulence of M . oryzae [28] . The generated strains were inoculated to rice leaf sheaths , and localization of each protein was investigated . The result showed that all four proteins were localized to the BICs , and both GAS1 and GAS2 were then translocated into the rice cytoplasm , which is similar to PWL2 . By contrast , HMM14 and MC55 were not translocated into the rice cytoplasm like MC69 ( data not shown ) . However , fluorescent signals of BIC accumulations of MC69::mCherry::NLS , MC69::mCherry , HMM14::mCherry::NLS and MC55::mCherry::NLS were significantly weaker than that of PWL2::mCherry::NLS , GAS1::mCherry::NLS and GAS2::mCherry::NLS ( Figure 3 , S5 and data not shown ) . These finding indicate that BIC accumulation level of secreted proteins might be important for translocation to the infected rice cells . A virulence effector Slp1 sequesters chitin oligosaccharides to prevent PAMP-triggered immunity in rice , thereby facilitating rapid spread of the fungus within host tissue [21] . Slp1 contains two putative LysM domains , which have previously been shown to bind carbohydrates [41] . An effector known as Ecp6 that also contains LysM domains was identified from a fungal pathogen Cladosporium fulvum that causes leaf mold of tomato [42] . Another effector AVR4 of C . fuluvum binds to chitin present in fungal cell walls and that , through this binding AVR4 can protect these cell walls against hydrolysis by plant chitinases [43] . Growth of the Δpep1 mutants of U . maydis are arrested during penetration of the epidermal cell and elicit a strong plant defense response such as formation of large papillae , induction of strong cell wall autofluorescence , H2O2 accumulation and defense related gene expression [38] . We tried to elucidate the roles of MC69 by addressing differences in H2O2 accumulation and expression of defense related genes in rice infected by mc69 mutant and wild type M . oryzae . However , results showed no difference between mc69 and wild type so that we have no evidence that MC69 suppresses plant defense responses at the moment ( data not shown ) . Taken together , we demonstrated that MC69 has a pathogenicity function ( required for the fungus to be a pathogen ) , but its function has yet to be elucidated .
All isolates of M . oryzae used in this study are stored at the Iwate Biotechnology Research Center . Fungal strains used were the wild-type strains 70-15 , Ina72 , TH68-141 , Hoku1 , Sasa2 and Ina86-137 [20] . To obtain protoplasts , hyphae of M . oryzae strains were incubated for 3 days in 200 mL of YG medium ( 0 . 5% yeast extract and 2% of glucose , w/v ) . Protoplast preparation and transformation were performed as described previously [44] . Hygromycin- or bialaphos-resistant transformants were selected on plates with 300 µg ml−1 of hygromycin B ( Wako Pure Chemicals , Osaka , Japan ) or 250 µg ml−1 of bialaphos ( Wako Pure Chemicals ) . C . orbiculare ( Berk . & Mont . ) Arx ( syn . C . lagenarium [Pass . ] Ellis & Halst . ) strain 104-T ( MAFF240422 ) was used as the wild-type strain . All C . orbiculare strains were maintained on 3 . 9% ( w/v ) PDA ( Difco Laboratories , Detroit , MI ) at 24°C . Preparation of protoplasts and transformation of C . orbiculare were performed according to a method described previously [45] . Mycelia of M . oryzae were grown on oatmeal agar medium ( 30 g l−1 oatmeal , 5 g l−1 sucrose and 16 g l−1 agar ) . To enhance conidia formation , the fungus was first grown on oatmeal agar medium for 9 days at 25°C , and then exposed to Black Light Blue light ( Toshiba FS20S/BLB 20W; Toshiba , Tokyo , Japan ) for 4 days at 22°C , after aerial hyphae of the colonies had been washed away with sterilized distilled water . Conidia of M . oryzae were suspended in 50 mM cAMP to a final density of 1×106 conidia ml−1 . This suspension was then poured onto dialysis membranes ( 2 . 5 ml of suspension/25 cm2 membrane surface; Spectra/Por , cutoff 1 , 000 Da; Spectrum Medical Industries , Terminal Annez , LA ) and incubated at 25°C in dark [25] . Total RNA was extracted from germinating conidia incubated for 6 h on dialysis membranes , as described below . 32 sheets of membranes containing the germinating conidia were crushed and homogenized in liquid nitrogen with mortar and pestle . The homogenate was transferred to a centrifuge tube containing 40 ml of TRI Reagent ( SIGMA-ALDRICH , St . Louis , MO ) , homogenized by vigorous shaking and incubated at room temperature for 5 min . Then 8 ml of chloroform was added , homogenized by vigorous shaking for 15 sec and incubated at room temperature for 3 min . After centrifugation at 1000× g for 15 min at 4°C , the upper aqueous phase was transferred to a new centrifuge tube , and the total RNA was precipitated by the addition of 20 ml of isopropanol after incubation at room temperature for 10 min . The pellet was rinsed with 70% ethanol . SuperSAGE library was made from total RNA as described [46] , [47] . Di-tag fragments were sequenced by the 454 FLX sequencer ( 454 Life Sciences ) . Each 26-bp tag sequence was used for BLASTN search against M . oryzae 70-15 genome sequence . A total of 23 , 491 tags to comprising 26-bp sequence were recovered . Number of tags for each of putative secreted protein genes of M . oryzae is given in Table S1 . To construct the gene-disruption vector pGPSMC69-44 , a 8 . 7-kb fragment containing the MC69 gene amplified with the primers MC69S1 ( 5′- TTATGACGGGAGCACAGGCACAGCACAC-3′ ) and MC69AS1 ( 5′- TGGCCGACGTTGTGCTCTTTCAGTTCCT-3′ ) was cloned into pCR-XL-TOPO to generate pXLMC69 using the TOPO XL PCR Cloning Kit ( Invitrogen , Carlsbad , CA ) . MC69 was mutated using an adaptation of the TAG-KO method using pGPS-HYG-CAM [30] , [31] . The pXLMC69 containing MC69 was used as the target . An insertion was formed within the coding region of MC69 ( at 11 amino acids ) in pXLMC69 , which resulted in pGPSMC69-44 ( Figure S1 ) . For complementation assay of an mc69 mutant with MC69 , a 5 . 7-kb fragment containing MC69 was amplified with the primers NMU1 ( 5′-ATAAGAATGCGGCCGCTGATTCTCAATGCCCTCTGTCCTTT-3′; the NotI site is underlined ) and MC69AS1 . The PCR product was digested with NotI and XbaI ( exists in the middle of the PCR product after 1 . 1-kb far from the poly A signal recognition site of MC69 ) to generate 3 . 1-kb fragment containing MC69 , and ligated to the same restriction sites of which carries the bialaphos-resistant ( bar ) gene [48] , creating pCB1531-MC69 . To substitute a cysteine residue at 36 amino acids in MC69 by alanine , single point mutation was introduced in plasmid pCB1531-MC69 using a primer BMC36AU4 ( 5′- CAGGTCACCAGACGACGCCTTCTTTGG -3′; mutation site is underlined and the BstEII site is indicated in italics ) . A 1 . 7-kb fragment containing a half 3′-terminal part of the MC69 ORF and terminator was amplified with the primers BMC36AU4 and M13F ( 5′- CGCCAGGGTTTTCCCAGTCACGA-3′ ) . The PCR product was digested with BstEII and XbaI , and exchanged to the BstEII/XbaI fragment of pCB1531-MC69 , generating pCB1531-MC69 ( C36A ) ( Figure 4B ) . To substitute a cysteine residue at 46 amino acids in MC69 by alanine , single point mutation was introduced in plasmid pCB1531-MC69 using a primer MC46AU5 ( 5′-CGTCACGCCGCAAGGCGCCGGGTATGTTCTGGG-3′; mutation site is underlined ) . A 1 . 7-kb fragment was amplified with the primers MC46AU5 and M13F , and the PCR product was used as a template for another PCR with the primers BMC46AU4 ( 5′- CAGGTCACCAGACGACTGCTTCTTTGGTGTCGTCACGCCGCAAGGCGCCGG-3′; mutation site is underlined and the BstEII site is indicated in italics ) and M13F . The PCR product was digested with BstEII and XbaI , and exchanged to the BstEII/XbaI fragment of pCB1531-MC69 , generating pCB1531-MC69 ( C46A ) ( Figure 4B ) . To substitute two cysteine residues at 36 and 46 amino acids in MC69 by alanine , double point mutations were introduced in plasmid pCB1531-MC69 using a primer BMC36&46AU4 ( 5′- CAGGTCACCAGACGACGCCTTCTTTGGTGTCGTCACGCCGCAAGGCGCCGG-3′; mutation sites are underlined and the BstEII site is indicated in italics ) . A 1 . 7-kb fragment was amplified with the primers MC46AU5 and M13F , and the PCR product was used as a template for another PCR with the primers BMC36&46AU4 andM13F . The PCR product was digested with BstEII and XbaI , and exchanged to the BstEII/XbaI fragment of pCB1531-MC69 , generating pCB1531-MC69 ( C36A , C46A ) ( Figure 4B ) . For construction of the MC69-EGFP gene fusion vector pCB1531-MC69-EGFP , a 1 . 7-kb fragment containing MC69 gene was amplified with the primers NMU1 and XMG5L1 ( 5′-GCTCTAGACCACCACCACCACCTTTGGCAGGTCCGCGAAGAGGG-3′; XbaI site is indicated in italics ) which was designed with five glycine codons ( underlined ) as a spacer peptide between MC69 and EGFP . The PCR product encoding MC69-Gly5 was digested with NotI and XbaI , and exchanged to the NotI/XbaI fragment of the Tef promoter in pBAGFP [49] , generating pCB1531-MC69-EGFP . For construction of the MC69-mCherry gene fusion vector pCB1531-MC69-mCherry , a 0 . 7-kb mCherry cDNA fragment was amplified with the primers XmU1 ( 5′- GCTCTAGACATGGTGAGCAAGGGCGAGG-3′; XbaI site is underlined ) and BmL1 ( 5′- CGGGATCCTACTTGTACAGCTCGTCCAT-3′; BamHI site is underlined ) using pmCherry ( Clontech , Mountain View , CA ) as a template . The PCR product was digested with XbaI and BamHI , and exchanged to the XbaI/BamHI fragment of EGFP cDNA in pCB1531-MC69-EGFP , generating pCB1531-MC69-mCherry ( Figure 2D ) , a 1 . 6-kb fragment of MC69 promoter ( MC69p ) and MC69 ( C36A ) ORF was amplified with the primers NMU1 and XMG5L1 . The PCR product was digested with NotI and XbaI , and exchanged to the NotI/XbaI fragment of MC69p-MC69 in pCB1531-MC69-mCherry , generating pCB1531-MC69 ( C36A ) -mCherry ( Figure 5A ) . A 1 . 4-kb fragment of MC69p was amplified with the primers NMU1 and XMpL2 ( 5′- GCTCAGACCTTCGTAGGCCTGGAACGAGACGCTTCC-3′; XbaI site is underlined ) . The PCR product was digested with NotI and XbaI , and exchanged to the NotI/XbaI fragment of MC69p-MC69 in pCB1531-MC69-mCherry , generating pCB1531-MC69p-mCherry ( Figure 2A ) . A 0 . 6-kb fragment of PWL2 promoter was amplified the primers Ppwl2-5′ ( 5′-GAGGAGAAGCGGCCGCGTTAACAACGCGGTGTAAAGATTC-3′; NotI site is underlined ) and Ppwl2-3′ ( 5′-GAGAGGAGAAGGATCCACTAGTTCTAGATTTGAAAGTTTTTAATTTTAAAAAGAGATTTTCCGAG-3′; BamHI-SpeI-XbaI sites are underlined ) . The PCR product was digested with NotI and BamHI , and exchanged to the NotI/BamHI fragment of Tefp-EGFP in pBAGFP [49] , generating pCB-Ppwl2 . mCherry cDNA fragment was amplified with primers mCherry-N ( 5′-GAGAGGAGAAGGATCCAGATCTCTCGAGACCATGGTGAGCAAGGGCGAGGAG-3′; BamHI-BglII-XhoI sites are underlined ) and mCherry-C ( 5′-GAGAGGAGAAGAATTCGCTAGCGTCGACCTTGTACAGCTCGTCCATG-3′; EcoRI-NheI-SalI sites are underlined ) . The PCR product was digested with BamHI and EcoRI , and introduced into pCB-Ppwl2 , to produce pCB-Ppwl2-mCherry . pCB-Ppwl2-mCherry was digested with SalI , fill in with Klenow fragment and performed self-ligation , generating pCB-Ppwl2-mCherry-stop . A modified SV40 NLS-coding double stranded fragment [35] was produced annealing with the oligos mSV40NLS ( 5′-TCGACGGTCCAGGTGGAGCTGGACCAGGTAGAAAGAGGCCACCAAAGAAAAAGAGAAAGGTAGATTATGGAGCTTAAG-3′; SalI protruding end is underlined ) and c-mSV40NLS ( 5′-AATTCTTAAGCTCCATAATCTACCTTTCTCTTTTTCTTTGGTGGCCTCTTTCTACCTGGTCCAGCTCCACCTGGACCG-3′; EcoRI protruding end is underlined ) . The annealed product was introduced into pCB-Ppwl2-mCherry , to produce pCB-Ppwl2-mCherry-NLS . From 1 µg of total RNA of cAMP-treated M . oryzae strain 70-15 , single-stranded cDNA was synthesized by using oligo ( dT ) primer and ReverTra Ace reverse transcriptase ( Toyobo , Osaka , Japan ) . A 0 . 4-kb PWL2 cDNA fragment was amplified from the total cDNA as a template with the primers PWL2-N ( 5′-GAGAGGAGAATCTAGAAAAATGAAATGCAACAACATCATCCTC-3′; XbaI site is underlined ) and PWL2-C ( 5′-GAGAGGAGAAGGATCCCATAATATTGCAGCCCTCTTCTC-3′; BamHI site is underlined ) . The PCR product was digested with XbaI and BamHI , and introduced into pCB-Ppwl2-mCherry-NLS , generating pCB-Ppwl2-PWL2-mCherry-NLS ( Figure 3A ) . A 0 . 2-kb MC69 cDNA fragment was amplified from the total cDNA with the primers XMU2 ( 5′-GCTCTAGAAAATAAAAATGAAGGCCGCT-3′; XbaI site is underlined ) and XML1 ( 5′-CCGCTCGAGTTTGGCAGGTCCGCGAAGAGGGCCGC-3′ , XhoI site is underlined ) . The PCR prodct was digested with XbaI and XhoI , and introduced into pCB-Ppwl2-mCherry-NLS and pCB-Ppwl2-mCherry-stop , to produce pCB-Ppwl2-MC69-mCherry-NLS and pCB-Ppwl2-MC69-mCherry , respectively ( Figure 3B and C ) . To make the MC69-HA gene fusion vector pCB1531-MC69-HA , HA-tagged full cDNA of MC69 ( MC69HA ) was amplified from the total cDNA with the primers XMU2 and BMHAL1 ( 5′- CGggatccTCAAGCATAATCTGGAACATCGTATGGATAACCACCTTTGGCAGGTCCGCGAAGAGGGCCGC-3′; BamHI site and HA tag sequence are indicated in lower cases and italics , respectively ) which was designed with two glycine codons ( underlined ) as a spacer peptide between MC69 and HA tag . The PCR product was digested with XbaI and BamHI , and exchanged mCherry gene at the same sites of pCB1531-MC69p-mCherry , generating pCB1531-MC69-HA ( Figure S3 ) . MC69-3xFLAG gene fusion construct pUC57-MC69-3xFLAG was custum-synthesized ( GenScript , Piscataway , NJ ) . MC69-3xFLAG was amplified from pUC57-MC69-3xFLAG with the primers SMU2 ( 5′-GACTAGTGAAAATAAAAATGAAGGCCGCTTTCGTTCTCGC-3′; SpeI site is underlined ) and BFL1 ( 5′- CGGGATCCTCACCCATCATGATCCTTGTAATCG-3′; BamHI site is underlined ) . The PCR product was digested with SpeI and BamHI , and exchanged mCherry gene at XbaI and BamH sites of pCB1531-MC69p-mCherry , generating pCB1531-MC69-3xFLAG ( Figure S3 ) . For construction of the MC69p::AVR-Pia epression vector pCB1531-MC69p-AVR-Pia , a 0 . 3-kb fragment containing AVR-Pia gene was amplified from pCB1004-pex22 [20] with the primers XP22U2 ( 5′-GCTCTAGACAAAATGCATTTTTCGACAATTTTC-3′; XbaI site is underlined ) and BP22L2 ( 5′-CGGGATCCTAGTAAGGCTCGGCAGCAAGCC-3′; BamHI site is underlined ) . The PCR product was digested with XbaI and BamHI , and exchanged mCherry gene at the same sites of pCB1531-MC69p-mCherry , generating pCB1531-MC69p-AVR-Pia ( Figure S4 ) . CoMC69 was isolated from genome of C . orbiculare 104-T by PCR using degenerate primers designed in amino acid sequences of MC69 homologs in fungal pathogens including C . graminicola . To construct the gene replacement vector pGDCOMC69 , the 3 . 0-kb fragment containing the 5′ flanking region of CoMC69 was amplified by PCR with the primers COMC5S ( 5′-ATAAGAATGCGGCCGCCCAGTGCTTTGTCATGTTGC-3′; NotI site is underlined ) and COMC5AS ( 5′-CCCAAGCTTCGCTGGTTGCGAAGAATGCG-3′; HindIII site is underlined ) . The amplified fragment was digested with NotI and HindIII , and introduced into pCB1636 [48] , which contained the hph gene , to produce plasmid pCB5MC69 . The 3-kb fragment that contained the 3′ flanking region of CoMC69 was amplified by PCR with the primers COMC3S ( 5′-GAAGGGCCCCCGGTCACCACGCATGTGTGATACG-3′; ApaI site is underlined ) and COMC3AS ( 5′-GGGGTACCACGTGTGCACTCTTAAGGAG-3′; KpnI site is underlined ) . The amplified fragment was digested with ApaI and KpnI , and introduced into pCB5MC69 to produce pGDCOMC69 ( Figure S11A ) . To generate the reporter construct pBATCoMC69pro-mCherry , the 1 . 4 kb 5′ upstream region of CoMC69 and mCherry were amplified using PCR with the two primer sets , ( i ) CoMC69pro-NotI-f ( 5′-ATAAGAATGCGGCCGCGTCTTTCGTCTTTTCGGTCT-3′; NotI site is underlined ) and CoMC69pro-BamHI-r ( c ) ( 5′-CGGGATCCCGTGTCGATGTATTTGTTGTG-3′; BamHI site is underlined ) , and ( ii ) mCherry-BamHI-f ( 5′-GCGGATCCATGGTGAGCAAGGGCGAGGAGGATAAC-3′; BamHI site is underlined ) and mCherry-EcoRI-r ( c ) ( 5′-CCGGAATTCTTACTTGTACAGCTCGTCCATGCC-3′; EcoRI site is underlined ) , respectively . The amplified fragments were introduced into each corresponding site of pBAT [49] , resulting in pBAT-CoMC69pro-mCherry . Barley leaf and rice leaf inoculation were performed as follows: conidial suspension ( 1×105 conidia ml−1 ) containing Tween 20 ( 0 . 01% in final concentration ) was sprayed onto susceptible barley cotyledons ( cv . Nigrate ) and rice seedlings ( cv . Shin No . 2 or cv . Sasanishiki ) of the fourth leaf stage . Inoculated plants were placed in a dew chamber at 27°C for 24 h in the dark , and then transferred to the growth chamber with a photoperiod of 16 h . A rice leaf sheath inoculation test was performed according to the method described previously [50] . To investigate the function of appressorium-mediated penetration of the inner epidermal tissue of rice leaf sheath , penetration hyphae were stained with lactophenol-trypan blue and destained in saturated chloral hydrate as described previously [51] . Invasive growth rating of the 50 appressorial penetration sites in rice leaf sheath cells were scored 32 h after inoculation . Invasive growth were classified into 4 levels: Level 1 , invasive hypha length is shorter than 10 µm with no branch; Level 2 , invasive hyphae length is 10–20 µm with 0–2 branches; Level 3 , invasive hyphae length is longer than 20 µm and/or with more than 2 branches within one cell; Level 4 , invasive hyphae are spread more than one cell ( Figure 1D ) . To test fungal pathogenicity of C . orbiculare , conidial suspensions of tested C . orbiculare strains ( approximately 5×105 conidia/ml ) were spotted onto detached leaves of cucumber or N . benthamiana . Germinated conidia and appressoria were observed on glass coverslips , and invaded hyphae were observed in epidermal cells of rice leaf sheath . mCherry fluorescence was observed using an Olympus FluoView FV1000-D confocal laser-scanning microscope ( Olympus , Tokyo , Japan ) equipped with a Multi argon laser , a HeNe G laser , a 40× UPlanSApo ( 0 . 9 numerical aperture ) and a 60× UPlanFLN ( 0 . 9 numerical aperture ) objective lens . To assess fluorescent signal in the reporter strains of C . orbiculare , conidia of the reporter strain were inoculated on the lower surfaces of cucumber cotyledons . Detection of mCherry fluorescence was performed using an Olympus FluoView FV500 confocal laser-scanning microscope ( Olympus ) with a Nikon 60× PlanApo ( 1 . 4 numerical aperture ) oil-immersion objective ( Nikon , Tokyo , Japan ) . Samples were mounted in water under cover slips and excited with the He/Ne laser . We used diachronic mirror DM488/543/633 , SDM630 beam splitter , and emission filter BA560-600 . Conidial suspension ( 1×105 conidia ml−1 ) was injected into rice ( cv . Shin No . 2 ) leaf sheath and placed in a dew chamber at 25°C for 32 h in the dark . The infected leaf sheaths were ground in liquid nitrogen , thawed in X µl of extraction buffer ( 250 mM Tris-HCl pH 7 . 5 , 2 . 5 mM EDTA , 0 . 1% ascorbic acid ( w/v ) , 1 mM PMSF , 0 . 01% PI cocktail ( v/v ) ( SIGMA-ALDRICH ) , 0 . 1% Triton X-100 ( v/v ) ) for X mg sample , vortex for 10 min at 4°C , and centrifuged at 15 , 000× g for 20 min at 4°C in a microcentrifuge . The crude extracts ( 15 µl per lane ) were separated on a 10–20% precast e-PAGEL ( ATTO , Tokyo , Japan ) and the proteins were transferred on to Immobilon Transfer Membranes ( Millipore , Billerica , MA ) . The blots were blocked in 2% ECL Advance Blocking Agent ( GE Healthcare , Buckinghamhire , UK ) in TTBS ( 10 mM Tris-HCl , pH 7 . 5 , 100 mM NaCl , 0 . 1% Tween 20 ( v/v ) ) for 1 h at room temperature with gentle agitation . For immunodetection , blots were probed with anti-HA ( 3F10 ) -HRP ( Roche , Mannheim , Germany ) or anti-FLAG M2-HRP ( SIGMA-ALDRICH ) in a 1∶10 , 000 dillution in TTBS for 2 h . After washing the membrane for 10 min three times , the reactions were detected using an ECL Advance Western blotting detection reagents ( GE Healthcare ) and a Luminescent Image Analyzer LAS-4000 ( Fujifilm , Tokyo , Japan ) . M . oryzae strains were cultured in 20 ml YG medium at 25°C at 120 rpm for 48 h . The culture was filtrated with Miracloth ( Merck , Darmstadt , Germany ) , concentrated and desalted by ultrafiltration with Amicon Ultra-15 ( 10K ) ( Millipore ) . The culture filtrates ( 20 µg of protein per lane ) were separated on a 12 . 5% SDS-PAGE gel and the proteins were transferred on to Immobilon Transfer Membranes ( Millipore ) . The blots were blocked in 5% nonfat dry milk in TTBS for 1 h at room temperature with gentle agitation . For immunodetection , blots were probed with Living Colors DsRed Polyclonal Antibody ( Clontech ) in a 1∶1 , 000 dilution in TTBS for 2 h . After washing the membrane with TTBS for 10 min three times , Anti-Rabbit IgG HRP conjugate ( Promega W401B ) ( Promega , Madison , WI ) in a 1∶10 , 000 dilution in TTBS was used as secondary antibody and incubated for 1 h at room temperature with gentle agitation . After washing the membrane for 10 min three times , the reactions were detected using an ECL Western blotting detection reagents ( GE Healthcare ) and a Luminescent Image Analyzer LAS-4000 ( Fujifilm ) . Sequence data of MC69 and the homologs from this article can be found in the GenBank/EMBL data libraries accession number MGG_02848 . 6 ( Mo ) , AB669186 ( Co ) , EFQ29542 ( Gg ) , EEY15898 ( Va ) , EGY20943 ( Vd ) , XP_965292 ( Nc ) , EGO52621 ( Nt ) , XP_003659994 ( Mt ) , XP_00190740 ( Pa ) , EFX05010 ( Gc ) , EGU75378 ( Fo ) , XP_388669 ( Gz ) , EHK44387 ( Ta ) , EHK23962 ( Tv ) , EFY93067 ( Mac ) , EFY97094 ( Man ) , and EGX95034 ( Cm ) . | Magnaporthe oryzae causes the most devastating fungal disease in rice . M . oryzae secretes a plethora of effector proteins , including several avirulence proteins which are known to be recognized by host resistance proteins activating innate immunity . However , the effectors that are required for virulence activity have not been identified in M . oryzae to date except for an effector protein , Secreted LysM Protein 1 ( Slp1 ) that was recently identified . We performed a large-scale disruption analysis of M . oryzae effector candidates and identified a small protein MC69 , which is secreted by the fungus during infection . When MC69 is absent , pathogenicity is severely reduced after penetration into the host cells . Furthermore , deletion of the MC69 orthologous gene in Colletotrichum orbiculare reduced its pathogenicity in the host plants cucumber and Nicotiana benthamiana . Thus , MC69 is conserved in ascomycete fungi and is crucial for establishing compatibility . This is the first report of a single secreted protein that is indispensable for pathogenicity in both monocot and dicot pathogenic fungi . How MC69 contributes to pathogenicity or virulence is unknown but it could be required for the fungus to be a pathogen or might be a classical effector that acts on plant target molecules . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"mycology",
"plant",
"science",
"fungi",
"plant",
"biology",
"plant",
"pathology",
"botany",
"biology"
] | 2012 | Large-Scale Gene Disruption in Magnaporthe oryzae Identifies MC69, a Secreted Protein Required for Infection by Monocot and Dicot Fungal Pathogens |
The differentiation of cells into distinct cell types , each of which is heritable for many generations , underlies many biological phenomena . White and opaque cells of the fungal pathogen Candida albicans are two such heritable cell types , each thought to be adapted to unique niches within their human host . To systematically investigate their differences , we performed strand-specific , massively-parallel sequencing of RNA from C . albicans white and opaque cells . With these data we first annotated the C . albicans transcriptome , finding hundreds of novel differentially-expressed transcripts . Using the new annotation , we compared differences in transcript abundance between the two cell types with the genomic regions bound by a master regulator of the white-opaque switch ( Wor1 ) . We found that the revised transcriptional landscape considerably alters our understanding of the circuit governing differentiation . In particular , we can now resolve the poor concordance between binding of a master regulator and the differential expression of adjacent genes , a discrepancy observed in several other studies of cell differentiation . More than one third of the Wor1-bound differentially-expressed transcripts were previously unannotated , which explains the formerly puzzling presence of Wor1 at these positions along the genome . Many of these newly identified Wor1-regulated genes are non-coding and transcribed antisense to coding transcripts . We also find that 5′ and 3′ UTRs of mRNAs in the circuit are unusually long and that 5′ UTRs often differ in length between cell-types , suggesting UTRs encode important regulatory information and that use of alternative promoters is widespread . Further analysis revealed that the revised Wor1 circuit bears several striking similarities to the Oct4 circuit that specifies the pluripotency of mammalian embryonic stem cells . Additional characteristics shared with the Oct4 circuit suggest a set of general hallmarks characteristic of heritable differentiation states in eukaryotes .
How differentiated cell types are epigenetically maintained through repeated cell division is a topic of intensive study [1] , [2] , both for its role in basic developmental processes [3] and its relevance to the advancement of human stem cell therapeutics [4] . However , as a basic model of differentiation , stem cell systems have several drawbacks , such as the vast number of distinct cell types , the difficulty of isolating large homogeneous cell populations , and the challenge of genetic manipulation . A much simpler example of epigenetic inheritance of differentiated cell states is found in Candida albicans , the most prevalent human fungal pathogen . This eukaryote forms two distinctive types of cells , white and opaque , that differ strikingly in their appearance [5] ( Figure 1A and 1B ) , competency to mate [6] , and the human tissues to which they are likely best suited [7]–[11] . Each cell type is heritably maintained through many cell divisions , with switching back and forth between the two cell types occurring stochastically , only once every 104 generations . The low rate of switching makes it easy to obtain large populations of homogeneous cells of each type . Furthermore , it is relatively straightforward to manipulate the genes of C . albicans , which has allowed dissection of both the regulation underlying the switch and the functions of downstream genes that are ultimately responsible for conferring the specific attributes of each cell type [12]–[16] ( for reviews , see [17] , [18] ) . A master regulator of the white-opaque switch , White Opaque Regulator 1 ( Wor1 ) , forms interlocking feedback loops with two other transcription regulators ( Czf1 and Wor2 ) . The three regulators are up-regulated in opaque cells compared to white cells and together are responsible for the establishment and maintenance of the opaque cell type [13] . The white state is maintained by the transcription regulator Efg1 , which is down-regulated in opaque cells [13] , [19] . The expression of more than 400 genes was previously found to differ between the two cell types [20] , [21] , but subsequent genome-wide chromatin immunoprecipitation ( ChIP-Chip ) experiments indicated that Wor1 directly bound only 58 of these genes [13] . Much of this discordance may be due to indirect regulation; indeed , Wor1 itself controls a large number of transcriptional regulators that may direct the differential expression of additional genes . However , it was much more difficult to explain the observation that only 30% of all Wor1-bound regions flank at least one differentially expressed transcript . Are the other Wor1 binding sites simply non-functional ? Do they act only on more distal transcripts and/or only in response to certain environmental cues ? Does Wor1 also play a non-regulatory role , helping to maintain chromosome structure via these binding sites ? Although we investigate this issue specifically in C . albicans , we note that discordance between binding ( determined by ChIP ) and regulation ( based on RNA analysis ) has frequently been observed in the circuits of a broad range of organisms [22]–[26] . To better resolve the relationship between the binding of a master regulator of differentiation and differential expression of its direct targets between cell types , we performed massively-parallel strand-specific sequencing of RNA from white and opaque cells . Applying several novel algorithms to the resulting dataset and merging these results with the existing ORF-based gene annotation , we first annotated the C . albicans transcriptome . This revealed that thousands of transcripts overlap another transcript on the opposite strand , demonstrating widespread presence of anti-sense transcription in this yeast , as in the model yeast Saccharomyces cerevisiae [27] , [28] . With the new annotation we found that the abundance of 1 , 306 transcripts differed between white and opaque cell types , a 3-fold increase over the number identified previously by microarray . We next revisited the poor correspondence between Wor1 binding and differential expression and found a remarkable improvement in concordance . Thus , a large fraction of the Wor1 bound regions previously lacking proximity to a differentially expressed gene , and therefore also lacking obvious function , can now be assigned the function of regulating previously invisible or inaccurately-measured transcripts . Our analysis of the Wor1 circuit revealed several unusual properties . For example , the targets of Wor1 have abnormally long upstream intergenic regions and un-translated regions ( UTRs ) . We show here that many of these long UTRs are cell-type-specific ( that is , the transcript length is differentially regulated ) and thus may function to bring additional layers of regulation to the differentiation circuit . A meta-analysis of the Oct4 circuit [29]–[31] , which governs the pluripotency and differentiation of mouse embryonic stem cells , reveals many of these same “unusual” properties . These surprising similarities across vast evolutionary distances , combined with many other shared features , suggest that several hallmarks of cell differentiation circuits exist broadly across eukaryotes .
To characterize the transcriptomes of white and opaque cells , we sequenced the poly ( A ) fraction of RNA extracted from replicate white and opaque cell cultures ( Materials and Methods and Figure 1B ) , expecting to find messenger RNAs , polyadenylated non-coding RNAs , and abundant non-polyadenylated transcripts that persist through the purification steps . Importantly , the sequencing libraries were prepared using an approach that preserves the genomic strand from which the sequenced RNA fragments were originally transcribed ( see Materials and Methods and Figure S1 ) [32] . Our sequencing runs yielded 29–136 million 50-base sequence reads per sample , which were subsequently aligned to a filter database ( containing , e . g . , rDNA sequences ) and then to the Candida albicans genome ( build Ca21 ) and a database of previously annotated splice junctions ( Materials and Methods and Figure S2 ) . An overview of the results is depicted in Figure 1C . The majority of reads from each sample ( 60–68% ) was successfully aligned , allowing detection of 93–95% of previously annotated exons with mean 50–200x sequence coverage ( i . e . , the number of reads aligned across a genomic position ) . 37–47% of positions were covered by an alignment in the strand-specific genome , and 423–904 deletions , which represent both splice junctions and deletion polymorphisms relative to the haploid reference genome , were detected ( Mitrovich et al . [33] , in preparation ) . On the whole , we have obtained more than sufficient sequence depth from these samples to build the first transcript annotation for C . albicans . Our RNA-Seq dataset allows us the first opportunity to define a true transcript annotation for C . albicans , which until now has had a gene annotation based primarily on computationally-predicted open reading frame ( ORF ) sequence boundaries and generally not informed by experimental data . We first developed a general computational approach ( Figure 2A ) that can define a new transcript annotation by combining an existing annotation ( in this case the ORF-based annotation ) with evidence found in RNA sequence data for un-translated regions ( UTRs ) and entirely novel transcripts . This effort included the development of new methods for the de novo identification of splice junctions and transcriptionally active regions ( TARs ) , which are based on gapped read alignments and clusters of sequence coverage , respectively ( Materials and Methods , Figure S3 , and Mitrovich et al . [33] , in preparation ) . We applied these methods to a single dataset produced by combining the reads from all four RNA sequence libraries , reasoning that ( 1 ) combining the datasets at this stage would be more powerful and straightforward than combining four separate annotations further downstream , and ( 2 ) the different datasets were sufficiently similar to one another . This is supported by the high reproducibility of biological replicates ( r = 0 . 95−0 . 99; Figure S5 ) and the observation that most exons , when expressed in both cell types , appear to extend to roughly the same boundaries . Rather than providing a completely de novo gene annotation ( as for S . cerevisiae in Yassour et al . [34] , for example ) , we sought to leverage the existing ORF-based annotation to provide an updated annotation in which existing transcripts , if expressed , were augmented with 5′ and 3′ UTRs , and new , isolated clusters of expression ( i . e . , those not overlapping an annotated exon on the same strand ) were added to the annotation as novel TARs ( nTARs ) . Thus , we devised a method to merge the splice junction and TAR-finding output with the existing ORF-based annotation ( Materials and Methods and Figure S4 ) and applied it to our datasets , resulting in the new C . albicans transcript annotation ( Tables S1 , S2 , S5; summarized in Figure 2B ) . The new transcript annotation contains 23% more transcripts ( N = 7 , 823 ) covering 13% more of the genome ( 76 . 1% versus 63 . 6% ) than the old annotation . We estimate that roughly 1 , 048 of these transcripts are non-coding because they do not contain a canonical ORF that is at least 120 nucleotides long ( i . e . , encoding a peptide at least 40 amino acids long ) , which increases the number of non-coding RNAs ( ncRNAs ) annotated in C . albicans by nearly 500% . However , there are also a large number of new coding transcripts ( i . e . , transcripts that contain putative ORFs encoding peptides 40 or more amino acids long ) , leading to an estimated 9% increase in the number of coding transcripts . Many of these ORFs may have been missed in previous annotations due to their short length ( 91% are shorter than 100 amino acids ) and , in some cases , due to lack of conservation in other species . It is likely that some of the ORFs defined here by our arbitrary length cutoff are not translated into protein . On the whole though , the number of putative ORFs at least 40 amino acids long found in novel transcripts ( N = 561 ) is significantly higher than expected by chance ( median N = 453; P-value <0 . 0001 by simulation; Materials and Methods ) , suggesting that many are translated into protein . As detailed in the next section , at least 18 of these short , novel ORFs are likely to serve an important function in opaque cells . In the new transcript annotation 5′ and 3′ UTRs of median length 99 and 136 bases were defined for 5 , 465 and 5 , 768 transcripts , respectively . These estimates are longer than estimates of 5′ and 3′ UTR length based on tiling arrays ( 68 and 91 in David et al . [35] ) , but closely resemble those based on RNA-Seq data ( 111 and 142 in Yassour et al . [34] ) for the related model yeast , Saccharomyces cerevisiae . Finally , 50% of transcripts in the new annotation overlapped transcripts from the opposite strand by at least 1 bp and 31% did so across more than 10% of their length , indicating that , as in other eukaryotes [27] , [28] , [36] , there is widespread antisense transcription in C . albicans . This observation underscores the importance of sequencing RNA in a strand-specific manner . Overall , the new transcript annotation described here represents a dramatic revision from previous annotations that microarrays were designed to assess . Using this new annotation we revisited the differences in gene expression between white and opaque cells . We determined which of the 7 , 823 newly defined transcripts were differentially expressed between white and opaque cell types by employing a likelihood ratio test [37] . We required a 2-fold or greater change in expression and false discovery rate ( FDR ) of 10−4 or less , which resulted in a set of 1 , 306 differentially-expressed transcripts ( Table S3 ) . As expected , we find strong ( 50-fold ) up-regulated expression of WOR1 , the gene that encodes a master regulator of white-opaque switching ( Figure 3A ) . As predicted by a previous study [14] , WOR1 has an unusually long 5′ UTR ( 1 , 978 bp , compared to the genome-wide median length of 99 bp ) . Unexpectedly , the lower WOR1 expression in white cells is associated with increased expression on the strand opposite this long UTR , suggesting an alternative internal antisense promoter is active and may be repressing WOR1 expression in white cells . To confirm the quality of these data we compared them directly to data generated using microarrays that are commonly used to study gene expression in C . albicans . We hybridized the same samples used for RNA sequencing ( Materials and Methods ) and examined the fold-change measurements produced by each technology for all previously annotated transcripts ( Figure 3B ) . We found a strong overall correlation ( r = 0 . 79 ) , which , as noted in other comparisons of RNA-Seq and microarray data , is stronger for high abundance transcripts ( r = 0 . 89 ) than it is for low abundance transcripts ( r = 0 . 71 ) , which are generally more accurately measured by RNA-Seq [32] , [37] , [38] . The 1 , 306 differentially expressed transcripts found here represent a 3-fold increase in the number observed by microarray [21] , which is partly attributable to the fact that 37% of these transcripts are novel ( N = 488 ) and thus were not probed on previous microarrays . Novel transcripts are unexpectedly frequent amongst the set of white-opaque differentially-expressed transcripts ( N = 488 versus 218 expected; χ2 P-value = 10−89 ) , a provocative observation we can not yet entirely explain , but which suggests an important role for non-coding transcripts and short proteins in the white-opaque circuit . In any case , this observation emphasizes the importance of “hypothesis-free” approaches to measuring gene expression . The remaining differentially-expressed transcripts , not recognized as such by microarray ( N = 376 ) , may be explained by the documented , improved sensitivity and dynamic range of RNA-Seq [38] , [39]; indeed , these transcripts not discovered by microarray have 2-fold lower average abundance than those that were , as estimated by RPKM ( reads per kb of transcript per million uniquely aligned reads ) . We were especially interested in the 488 novel differentially expressed transcripts , which fall into three major classes: ( 1 ) antisense transcripts , ( 2 ) isolated transcripts that encode proteins , and ( 3 ) isolated non-coding transcripts . We discuss these three classes in turn . We found 213 novel transcripts that overlap another transcript on the opposite strand across at least one third of their length . NTAR_364 is a particularly informative example of a differentially expressed novel transcript that overlaps another transcript on the opposite strand ( Figure 3C ) . The gene opposite NTAR_364 is STE4 , which encodes the β subunit of the heterotrimeric G protein complex required for mating [40] , [41] . Mating is a process specific to opaque cells [6] , and accordingly , NTAR_364's 14-fold down-regulation is inverse to STE4's 8-fold up-regulation in opaque cells . The anti-correlated expression of these two overlapping transcripts strongly suggests a mechanism in which NTAR_364's expression acts to repress expression of STE4 . There is ample precedent for this type of regulation in eukaryotes and bacteria [42]–[45] . To determine the prevalence of such mechanisms in C . albicans , we examined the expression profiles of all 759 such sense-antisense transcript pairs , filtering down to the subset of 44 pairs in which both transcripts are significantly changed and at least one transcript is coding ( Figure 3D ) . Our expectation was that we would observe strong anti-correlated differential expression across all such pairs if these mechanisms are prevalent and a lack of correlation if they are not . Instead , we found a modest and significant anti-correlation ( r = −0 . 25; P-value = 0 . 05; Figure 3D ) . Sense–antisense pairs in which one member is differentially-expressed are 2-fold more likely , than expected by chance , to have the second member differentially-expressed in the opposite direction ( 17% versus 8%; χ2 P-value = 10−4 ) . These results suggest that some , but not all , anti-sense transcripts act to repress the steady-state abundance of their sense counterpart . Despite the lack of perfect anti-correlation , there are several transcript pairs that , like the STE4-NTAR_364 pair mentioned , are considerably differentially-expressed in opposite directions ( Figure 4 ) , which strongly suggests a regulatory function for the novel antisense transcripts involved . The second major class of novel , differentially-expressed transcripts contains those that are isolated in the genome and code for protein . In total , we identified 224 novel differentially expressed transcripts that do not overlap a transcript on the opposite strand . Sixty-nine of these transcripts encode a putative protein at least 40 amino acids long . Amongst these is a group that clusters into three genomic locations and encodes a large family of novel , short ORFs ( Figure 5A , Figure S6A and S6B ) . Eighteen of the 24 ORFs in this family are encoded by transcripts that are opaque-specific , including NTAR_1179 . 2 , which with 287-fold higher abundance in opaque cells is the third most differentially-expressed transcript genome-wide . Using a combination of BLAST and PSI-BLAST against fungal genomes and eukaryotic protein sequence databases , we identified 46 members of this family ( see sequence alignments in Figure 5B and Figure S6C ) , 24 from C . albicans and 22 from its closest known relative , Candida dubliniensis . Homologs could not be identified in any other species , further underscoring the potential importance of these genes to opaque-cell differentiation , since these two yeast species are the only two known to switch between distinct white and opaque forms [46] . The neighbor-joining phylogeny inferred for these ORFs ( Figure 5C and Figure S6D ) indicates that most were present and similarly clustered in the common ancestor of C . albicans and C . dubliniensis . Computational predictions of secondary structure [47] indicated there are likely three β sheets followed by two α helices in these proteins ( Figure 5B ) and the structure prediction server I-TASSER [48] found a putative bacterial hemolysin ( PDB ID: 3HP7 ) to be the closest structural analog . Finally , 155 of the isolated , differentially-expressed transcripts do not appear to code for protein . At this time it is difficult to assess their functions in a purely computational manner; thus , their roles in the white-opaque switch await experimental characterization . In all three classes of novel transcripts we observe examples in which the master regulator Wor1 is bound adjacent to or overlapping the differentially expressed transcripts ( Figure 3C and Figure 5A ) , suggesting that these novel antisense and isolated transcripts are directly regulated by Wor1 binding . Thus , they may form a key , but heretofore unknown , part of the circuit . To assess the concordance between Wor1 binding and differential expression of nearby transcripts more globally we compared the previous ORF-based and our new RNA-Seq-based gene annotations to regions identified as Wor1-bound in chromatin immunoprecipitation-on-tiling microarray ( ChIP-Chip ) experiments [13] . We first associated Wor1-bound regions with adjacent genes using both the new and the old annotations ( Figure S7 ) , and then evaluated both the frequency with which Wor1 binding flanked at least one differentially expressed gene and the frequency with which Wor1-bound genes were differentially expressed ( Figure 6 ) . We also compared measurements of differential expression from three different platforms: ( a ) hybridization to spotted PCR-product microarrays ( reported previously by Tsong et al . [21] ) , ( b ) hybridization to custom-designed Agilent 8x15k microarrays ( reported here ) , and ( c ) strand-specific RNA-Seq ( also reported here ) . The pairing of the new transcript annotation with the RNA-Seq measurements of differential expression ( Figure 6 , first row ) clearly yields the strongest concordance between Wor1 binding and differential expression: 65% of Wor1-bound regions are associated with at least one differentially expressed transcript . This represents a greater than 2-fold improvement in concordance over a previously published association [13] , in which only 30% of bound regions were observed to flank at least one differentially expressed transcript ( Figure 6 , last row ) . In this previous association , differential expression of transcripts was measured by spotted PCR-product arrays designed to assay only transcripts in the old annotation . The concordance between binding and differential expression improves incrementally with the use of better microarray platforms ( 38–40%; Figure 6 , rows 5–6 ) and with RNA-Seq-based expression measurements computed using the old transcript annotation ( 48–51%; Figure 6 , rows 3–4 ) . However , by far the best concordance is found when RNA-Seq-based expression measurements are computed using the new transcript annotation . Thus , the dramatically improved association of master regulator binding and cell type-specific expression observed here is attributable to both the novel transcripts and the improved expression measurements provided by RNA-Seq . The fact that the WOR1 gene has a 2 kb long 5′ UTR and about 6 kb of Wor1-bound intergenic DNA upstream of it ( Figure 3A ) suggests that this master regulator of white-opaque switching is under complex regulation . We next examined whether other transcripts in the circuit have similar properties . It was previously noted that Wor1-bound intergenic regions are , on average , 5-fold longer than typical intergenic regions ( median 3 , 390 bp for Wor1-bound genes versus 623 bp genome-wide ) [13] . However , given the substantial changes we have made to the gene annotation , it was unclear whether this length bias would remain; in particular , it seemed plausible that some of the unusually long “intergenic” regions may actually contain , and thus be due to , previously unannotated long UTRs . We find that while genome-wide intergenic length is , on average , more than 2-fold shorter in the new annotation ( new median length = 262 bp ) , the intergenic regions bound by Wor1 are still , on average , 5-fold longer than expected by chance ( new median length = 1346 bp; Mann-Whitney P-value = 10−80; Figure 7A ) . Unexpectedly , we also found that 5′ UTRs of Wor1-bound genes are 58% longer than expected ( median 157 bp in the circuit versus 99 bp genome-wide; Mann-Whitney P-value = 10−20; Figure 7B ) and 3′ UTRs in the circuit are 22% longer than expected ( median 166 bp in the circuit versus 136 bp genome-wide; Mann-Whitney P-value = 10−6; Figure 7C ) . The unusually long UTRs found in the Wor1 circuit and the apparent change in UTR length at WOR1 ( Figure 3A ) motivated us to look more generally into changes in promoter usage and transcriptional termination between cell types , as reflected in changes in 5′ and 3′ UTR length , respectively . We devised a simple method to isolate putative cases of UTR length change , reasoning that a change in UTR length for a given transcript could be detected as a change in the apparent expression of the UTR that is significantly less than or greater than what was measured for the transcript's coding region . We required a minimum 2-fold difference in fold-change between UTR and coding region and a χ2 P-value less than 10−5 ( Materials and Methods ) . Using these criteria , we identified 145 transcripts with at least one UTR apparently changing length between white and opaque cells ( Table S4 ) . Visual inspection revealed that not all these cases are straightforward to interpret; however , many are , and these provide several examples for further study ( Figure 7D–7F ) . Most of the cases identified here are changes in 5′ UTRs ( N = 111; 77% ) , which likely reflects an emphasis on the usage of alternative promoters as a means of differentiating the two cell types . One of the transcripts , EFG1 , is a regulator of white-opaque switching and was previously shown to exhibit different 5′ UTR lengths in white and opaque cells [49] . EFG1 and 26 other transcripts with significant 5′ UTR changes are also associated with Wor1 binding nearby their genomic loci ( observed frequency = 24%; expected = 10%; χ2 P-value = 10−8 ) . For several of these transcripts , such as ORF19 . 2049 ( Figure 7D ) and EFG1 ( Figure 7E ) , the UTR is shorter in opaque cells and Wor1 is bound in opaque cells between the apparent white- and opaque-preferred transcription start sites , suggesting a direct regulatory mechanism . Other examples , such as PPS1 ( not shown ) and ORF19 . 7060 ( Figure 7F ) , are probably not directly related to Wor1 binding , but may instead involve mechanisms related to the transcription of antisense genes . Comparing Wor1 binding to gene expression revealed an additional feature of Wor1-controlled transcripts: direct binding of Wor1 within a transcribed region ( rather than upstream of it ) is associated with strong down-regulation of the bound transcript in opaque cells . The non-coding transcript NTAR_913 provides a clear example of this phenomenon ( Figure 7G ) . Genome-wide , we found 89 cases in which a transcript overlaps a Wor1-bound region by more than 50% , and the expression of such transcripts is frequently white-specific ( Figure 7H ) . This observation suggests the prominence of an underappreciated mode of gene regulation in which a transcription regulator may repress transcription via direct binding to the transcribed region . Given the unusual characteristics of the WOR1 locus and Wor1's target genes , we next examined whether other examples of heritable cell differentiation circuits exhibited similar features . One of the most studied transcription circuits is that of Oct4 , which governs the differentiation and pluripotency of mammalian embryonic stem ( ES ) cells [1] , [50] . Oct4 is a master regulator of mammalian cell types in the same sense that Wor1 is a master regulator of Candida cell types: Oct4 expression is required to maintain the pluripotent ES cell type [51] , and Oct4's over-expression in other cell types , along with additional factors , returns them to the ES cell state [2] , [52] . Although much is known about this circuit , we could not find any previous reports on the general properties of the circuit ( e . g . , relative UTR length of Oct4-bound genes ) . To determine if the unusual properties of the Wor1 circuit in Candida are shared with the Oct4 circuit , we performed a meta-analysis of publicly-available data , including ChIP-Seq-based Oct4 binding data [30] , [31] and microarray-based profiles of gene expression during stem cell differentiation [29] ( Materials and Methods ) . We discovered that the Oct4 circuit of mice does indeed share “unusual” characteristics with the Wor1 circuit of Candida . Intergenic regions bound by Oct4 are 33% longer than expected by chance ( median 23 kb in the circuit versus 17 kb genome-wide; Mann-Whitney P-value = 10−3 ) and are 2-fold longer than expected if they also flank a transcript that is differentially expressed during differentiation ( median 34 kb in the differentially-expressed circuit; Mann-Whitney P-value = 10−4; Figure 8A ) . 5′ UTRs and 3′ UTRs are also longer than expected ( 161 and 1048 bp in the circuit versus 137 and 727 bp genome-wide; Mann-Whitney P-values = 10−5 and 10−12 , respectively; Figure 8B and 8C ) , but the relative magnitude of length bias for 5′ versus 3′ UTRs ( +18% and +44% , respectively ) is flipped relative to that observed in the Wor1 circuit ( +58% and +22% , respectively ) . Unfortunately , the appropriate data are not yet available to determine whether UTR lengths are frequently changing between cell types in the Oct4 circuit of mice as they are in the Wor1 circuit of Candida .
By sequencing the transcriptomes of white and opaque cells ( Figure 1 ) and applying a novel computational approach ( Figure 2A ) , we have provided the first transcript annotation for C . albicans ( Figure 2B ) , the most prevalent human fungal pathogen . This new view of the C . albicans transcriptional landscape includes over a thousand newly discovered transcripts , some of which are transcribed antisense to previously annotated genes , but many of which are entirely isolated from other genes . A subset of these transcripts codes for proteins , some of which are specific to Candida species and may function in host-pathogen interactions . Overall , the new view of gene expression in C . albicans is reminiscent of that provided by recent sequencing of the transcriptome of another yeast species , S . cerevisiae [28] , [34] , [38] , but with two important differences . First , we have captured a more faithful depiction of the transcriptome by using a method that measures expression across entire genes in a strand-specific fashion . Second , relative to the model organism S . cerevisiae , the transcriptome of C . albicans was poorly characterized prior to RNA sequencing . Our analysis dramatically expands the view of transcription in this yeast , resulting in annotations for hundreds of new coding and non-coding transcripts and thousands of UTRs . The revised annotation and expression data allowed us to examine , at unprecedented resolution , the differences between two cell types . White and opaque cells are specified by one of the largest known transcriptional circuits in C . albicans; as discussed in the introduction , each cell type is heritable for many generations and switching between them is epigenetic . Our principle findings are summarized as follows: In addition to the conclusions listed above , a comparison of the RNA-seq data from C . albicans to those determined in other species reveals some important differences and similarities . With the new strand-specific data presented here we were able to systematically examine changes in the expression of sense and antisense transcripts . The high frequency of antisense transcripts combined with the weak anti-correlated expression of transcripts in sense-antisense pairs ( Figure 3D ) suggests that while transcriptional interference mechanisms likely control transcription rates in some cases , antisense transcription may also play a different role in this yeast , perhaps acting post-transcriptionally via RNAi mechanisms Genome-wide anti-correlated expression of sense-antisense pairs was previously observed in S . cerevisiae [27] , but in that study the anti-correlation across all sense-antisense pairs was stronger than what we observed here . It is possible that the difference between species is related to the loss of mechanisms for post-transcriptional control by antisense transcripts in S . cerevisiae , but not in C . albicans [54] . Thus , whereas C . albicans may use antisense transcripts for a mix of transcriptional and post-transcriptional regulation , antisense transcription in S . cerevisiae may function primarily to regulate sense transcripts through transcriptional interference . Finally , we note several striking mechanistic similarities between the Wor1 circuit that governs white-opaque switching in yeast and the Oct4 circuit that controls the pluripotency and differentiation of mammalian embryonic stem cells . In both systems , differentiation is controlled by a series of master transcription regulators arranged in interlocking feedback loops , the differentiation process requires long periods of time relative to the cell division time , and the differentiated states are “remembered” through many cell generations [1] , [17] , [18] , [50] . In each system , hundreds of binding sites for the master regulator were thought to be “non-functional” [25] , though , as we have shown here for the yeast system , many of these instead are likely to impart cell-type specific expression to previously unannotated transcripts . In addition , amongst the direct targets of the master regulators is an abundance of genes that encode transcription regulators themselves [13] , [29] , [55] and genes with unusually long upstream intergenic regions ( compare Figure 7A to Figure 8A ) and abnormally long UTRs ( compare Figure 7B and 7C to Figure 8B and 8C ) . It seems likely that the latter two characteristics reflect a large number of regulatory inputs to genes of these circuits . The expanded upstream regions may also allow the formation of more complex tertiary chromatin structures involved in gene regulation [56] , [57] . Regardless of their function , they are clearly identifiable landmarks of both circuits . We have also shown here that many of the long UTRs are regulated , in the sense that they are longer in one cell type and shorter in the other . Finally , it appears as though non-coding RNAs are an important component of both circuits [31] . Taken together , these findings suggest an unexpected level of sophistication is required to maintain distinct cell types through many cell divisions—whether in a relatively simple fungal system with only two cell types , or in a complex mammalian developmental system involving numerous differentiated tissues .
White cells of mating type a/a were selected by growth of C . albicans strain QMY23 [58] , a derivative of the sequenced strain SC5314 , on sorbose medium [59] . Opaque cell lines were then isolated following spontaneous cell-type switching . Liquid cultures of white or opaque cells ( two samples of each , referred to throughout the manuscript as white and opaque replicate #1 and white and opaque replicate #2 ) were grown at 23°C in SC medium [60] supplemented with 100 mg/l uridine to an OD600 of 1 ( log phase growth ) . Samples ( 5 ml ) were collected by centrifugation ( 5 min , 2000 g , 4°C ) , and pellets frozen in liquid nitrogen . Total RNA was extracted from frozen pellets as described [61] . For each sample , poly ( A ) RNA was isolated from 50 µg of total RNA by two rounds of purification using a Poly ( A ) Purist MAG kit ( Ambion ) . To construct libraries suitable for SOLiD System sequencing ( Figure S1 ) , each poly ( A ) -selected RNA sample ( 150–300 ng ) was fragmented in a 10 µl volume by incubation with 1 unit of RNase III and 1X reaction buffer ( Ambion ) for 10 minutes at 37°C . Fragmented RNA was then immediately diluted to 100 µl and purified using a RiboMinus Concentration Module ( Invitrogen ) following manufacturer's protocol , with the following modifications: sample was initially mixed with 100 µl Binding Buffer and 250 µl ethanol , column was washed only once with 500 µl Wash Buffer , and purified sample was eluted in 20 µl water . RNA fragmentation was confirmed and sample quantified using an Agilent 2100 Bioanalyzer , with an RNA 6000 Pico Chip , following manufacturer's protocol . 50 ng fragmented RNA was dried by vacuum centrifugation at low heat , then suspended in 3 µl water . An amplified cDNA library was constructed using components from the SOLiD Small RNA Expression Kit ( Ambion ) . Hybridization and ligation of Adaptor Mix A to the fragmented RNA and reverse transcription were carried out according to manufacturer's protocol , but with 18 h ligations and no RNase H treatment . cDNA was brought up to 100 µl and purified using a Qiagen MiniElute PCR Purification Kit , following manufacturer's protocol . Half of the eluted cDNA was mixed with an equal volume of loading dye ( 95% formamide , 0 . 5 mM EDTA , 0 . 025% each bromophenol blue and xylene cyanol FF ) , heated to 95°C for 3 min , then cooled immediately on ice . Sample was run on a 7 cm denaturing 7M urea/1X TBE/6% polyacrylamide gel at 180V for 17 min , then stained with SYBR Gold Nucleic Acid Gel Stain ( Invitrogen ) . DNA was visualized by UV-illumination , and material between 100–200 nt excised by scalpel . The excised region was further cut into 4 vertical strips ( such that each represented the same DNA size distribution ) . Amplification was performed directly on gel strips again using components from the SOLiD Small RNA Expression Kit ( Ambion ) . Two 100 µl PCR reactions were performed , each with one gel strip , 1X PCR Buffer , 0 . 2 mM dNTP mix , 2 µl AmpliTaq DNA Polymerase , and 2 µl SOLiD PCR Primer Sets 1 , 2 , 3 or 4 ( for white and opaque sample replicates #1 and white and opaque sample replicates #2 , respectively ) . Reactions conditions were 95°C ( 5 min ) ; 16 cycles of 95°C ( 30 sec ) , 62°C ( 30 sec ) , and 72°C ( 30 sec ) ; 72°C ( 7 min ) . The two amplification reactions were pooled and purified using a PureLink PCR Micro Kit ( Invitrogen ) following manufacturer's protocol , but combining two sequential elutions . To ensure appropriate size distributions ( >75% of product >140 bp ) , products were assayed using a Bioanalyzer DNA 1000 chip; yields ranged from 360–1140 ng . Templated beads were generated for sequencing using standard manufacturers' protocols . Beads from the first pair of white and opaque libraries ( “Replicate #1” ) were deposited onto a full slide with 8 other barcoded libraries not presented here . Beads from the second pair of white and opaque libraries ( “Replicate #2” ) were deposited onto two quadrants of a slide each . Massively parallel ligation sequencing was carried out to 50 bases using Life Technologies SOLiD System V3 and following the manufacturer's instructions . For microarray analysis , we used aliquots of the same total RNA samples used to generate the WT libraries ( replicate #2; discussed above ) . Aminoallyl-labeled cDNAs were synthesized using 5 µg of total RNA in 50 µl reverse transcription reactions with 250U SuperScript III Reverse Transcriptase ( Invitrogen ) , as described previously [58] . The cDNA samples were dried in a speed-vac to ≤9 µl total . Samples were then brought to 9 µl with water and supplemented with 1 µl of fresh 1M Na Bicarbonate , pH 9 . 0 . Cy3 and Cy5 dyes were prepared by re-suspending Amersham mono-reactive dye packs ( Cat . #PA23001 and PA25001 ) in 10 µl DMSO , and 1 . 25 µl of either Cy3 or Cy5 were added to each sample . Labeling reactions were incubated for one hour at room temperature in darkness . Dye-coupled cDNA samples were purified by adding 800 µl of Zymo DNA binding buffer ( Zymo Research ) to each sample and loading onto Zymo-25 columns . The remainder of the purification was performed as per the manufacturer's directions , and the samples were eluted with 40 µl of water . For each competitive hybridization , 0 . 2 µg each of Cy3 and Cy5 labeled cDNA were combined in 25 µl final volume of water , incubated at 95°C for 3 min , cooled to room temperature , mixed with 25 µl of Agilent 2x GE hybridization buffer ( HI -RPM ) , and loaded onto individual “blocks” ( 40 µl each ) on Agilent 8x15k custom gene expression microarrays . Hybridization was carried out at 65°C for 16 hours and the arrays were washed with Agilent wash buffers as per the manufacturer's recommendations . Whole transcriptome reads were aligned to a modified version of the Assembly 21 release of the Candida albicans genome [62] . As this is a haploid assembly , known single nucleotide variation between alleles from the most recent diploid assembly ( Assembly 19 , [63] ) was mapped to Assembly 21 , and the genome sequence was modified to reflect these ambiguous positions , allowing expressed sequences from either allele to be aligned equivalently . Alignment was performed with Life Technologies' SOLiD Whole Transcriptome Pipeline [32] , [64] . This software is open-source and freely available ( http://solidsoftwaretools . com/gf/project/transcriptome/ ) . An overview of the alignment strategy is presented in Figure S2 . In all the analyses of gene expression presented here , only reads that were both uniquely and fully aligned were considered . A “uniquely and fully” aligned read is defined as a read with a max-scoring alignment to the genome ( 1 ) scoring at least 31 ( alignment score is calculated with a match score of +1 and a mismatch score of −2 ) , ( 2 ) scoring at least 9 higher than any of the other alignments of that read to the genome , and ( 3 ) at least 40 bp long . All sequence data have been deposited at the MIAME compliant Gene Expression Omnibus ( GEO ) database at the National Center for Biotechnology Information ( http://www . ncbi . nlm . nih . gov/geo ) and are accessible through accession number GSE21291 . Known and novel splice junctions were identified by looking for sets of read sequences whose alignments share a gap ( specifically , a deletion relative to the reference ) with the same genomic start and end coordinates . We determined empirically that by requiring at least 5 such reads , and considering only deletions of at least 50 nucleotides , we captured , and thus validated , 85% of the 421 known junctions , while also predicting 158 novel junctions or deletions . False positives were filtered from this set by requiring matches to splice motifs and by removing deletions caused by obvious artifacts ( e . g . , cleavage and polyadenylation junctions ) , yielding 45 new introns in total . The details of this method are provided elsewhere in Mitrovich et al . ( In preparation ) [33] . A TAR is a region of the strand-specific genome exhibiting a cluster of sequence coverage , most often representing the presence of an exon . We employed a sliding window approach to identify such clusters on each strand of the C . albicans genome . The approach is described in depth in the manual for Life Technologies' Novel Transcribed Region ( NTR ) finder ( http://solidsoftwaretools . com/gf/download/docmanfileversion/138/693/NTR_Finder_Manual_v1 . 1 . pdf ) . Briefly , a window of specified size is scanned base-by-base across the genome , average sequence coverage is calculated within each window , and windows with average coverage greater than a specified cutoff are marked . A set of contiguous marked regions in the genome is then joined and trimmed from each end to better fit the coverage profile , forming a putative TAR ( pTAR ) . We used the NTR finder to perform TAR-finding on the combined dataset of all four sequence libraries presented in this work . TAR-finding was performed with many different parameter sets ( i . e . , different values chosen for the size of the window and the minimum average coverage required for the marking of a region ) and it was determined that a window size of 125 and minimum average coverage of 20 were optimal for reproducing the previously annotated TARs ( aTARs ) , with the expectation that the pTARs would be slightly larger than the aTARs because the existing annotations were ORF-based only and thus did not include UTR definitions ( Figure S3 ) . Other parameters were kept fixed: min-score = 25 , trimming-fraction = 0 . 01 , min-overlap = 0 . 9 . The existing transcript annotation ( Ca21 ) , which is primarily based on putative ORF sequences , was downloaded from the Candida Genome Database ( http://www . candidagenome . org/ ) and the exons defined therein were used as our aTARs . In merging the pTARs with aTARs to define a new transcript annotation , we found that in addition to this optimal pTAR set ( pTAR_opt_set , with parameters window-size = 125 , min-window-coverage = 20 , min-score = 25 , trimming-fraction = 0 . 01 , and min-overlap = 0 . 9 ) , a more fragmented pTAR set produced from a smaller window size ( pTAR_frag_set , with parameters window-size = 10 , min-window-coverage = 20 , min-score = 25 , trimming-fraction = 0 . 01 , and min-overlap = 0 . 9 ) was also helpful ( see below ) . We also experimented with Hidden Markov Model ( HMM ) approaches to finding pTARs ( not shown ) , but found that the models we trained did not perform better than the simpler sliding window approach taken here . In fact , they tended to perform much worse , which may simply reflect that we did not find the best way of modeling the segmentation problem . Rather than providing a completely de novo transcript annotation [34] , we sought to leverage the existing annotation to provide an updated transcript annotation in which existing ORF-encoding regions , if expressed , were augmented with 5′ and 3′ UTRs and isolated TARs ( i . e . , those not overlapping an aTAR on the same strand ) were added to the transcript annotation as novel TARs ( nTARs ) . Thus , we employed a set of rules that merged the pTAR_opt_set with the aTARs in the previous transcript annotation ( Ca21 , from the Candida Genome Database [65] ) to form a new set of transcript annotations . The rules are most concisely described diagrammatically in Figure S4 . For transcripts found to contain one or more splice junctions , the internal exon coordinates defined by reads spanning those splice junctions are used in place of those defined by the pTARs ( i . e . , splice junction-derived coordinates override these purely coverage-based coordinates ) . The more fragmented pTAR_frag_set was used to define transcript boundaries in cases where two or more aTARs were overlapped by a single pTAR ( scenario ‘f’ in Figure S4 ) , which typically happens when transcripts are positioned very close to one another on the same strand . In such cases , if a pTAR was found in the more fragmented set that overlapped the edge of one aTAR without also overlapping the edge of the other aTAR , this pTAR was used to define the UTR of the overlapping aTAR in the new annotation . We performed 10 , 000 rounds of simulation to determine whether the 561 nTARs containing an ORF of length 40 amino acids or longer was more than expected by chance . In each round , 1 , 443 regions with the same size distribution as the 1 , 443 nTARs were chosen randomly in a strand-specific fashion from regions of the genome not covered by ORFs in the previous annotation ( i . e . , the Ca21 ORF-based annotation ) . The median number of ORFs found per round was 453 . 561 or more ORFs were not found in any round of the simulation ( P-value <0 . 0001 ) . For each transcript model ( in either the new or old annotation ) , reads that uniquely aligned to the genome within its exons or across its splice junctions were counted . One pseudo-count was added to this sum and the resulting modified raw transcript count was converted to a normalized measurement of abundance by normalizing for transcript length and total number of uniquely aligned reads in the sample ( i . e . , RPKM; reads aligned per kb of transcript per million uniquely aligned reads ) [39] , [66] . The fold-change of each transcript between cell types was then computed by dividing its mean RPKM across opaque cell replicates by its mean RPKM across white cell replicates . We employed a recently proposed likelihood ratio test combined with a fold-change cutoff to define sets of differentially expressed transcripts [37] . Specifically , a false discovery rate ( FDR ) less than or equal to 10−4 and an absolute fold-change greater than or equal to 2 defined a set of 1306 differentially expressed transcripts using the new transcript annotation and a set of 824 using the old annotation . RPKM , fold-change estimates , P-values and FDRs for each transcript can be found in Table S3 . Microarray data were normalized and differentially expressed transcripts were identified using limma v2 . 16 . 5 [67] in R v2 . 8 . 1 . Background correction was performed with the “normexp” method and an offset value of 50 . Normalization was then performed within arrays using the “loess” method and between arrays using the “quantile” method . Finally , differential expression of transcripts between white and opaque cells was determined on our dye-swapped replicate arrays using the “lmFit” and “eBayes” methods , which produced fold-change estimates and Benjamini-Hochberg multiple test-corrected P-values for each probe on the array . For each transcript , only the expression value given by the probe with the highest average expression value ( i . e . , AveExpr value ) was used in downstream analysis . As with the analysis of the RNA-Seq data , we applied an adjusted P-value cutoff of 10−4 and required an absolute fold-change greater than or equal to 2 . This defined a set of 512 differentially expressed transcripts . Wor1-bound regions were identified as peaks of binding enrichment in the Wor1 ChIP-Chip data using the “Extract peaks from Data Set ( s ) ” utility of MochiView v1 . 311 [68] . The algorithm is described in detail in the MochiView manual . Briefly , a smoothing function is applied to the log2 enrichment values of the Wor1 ChIP-Chip tiling arrays followed by the application of an algorithm to detect local regions of maximal enrichment ( i . e . binding peaks ) , which are assigned a P-value using permutation testing . Note that this algorithm is not based on deconvolution of binding events using shearing profiles – in the case of the Wor1 ChIP-chip data , the binding peaks are atypically broad and varied , and thus tend to confound deconvolution-based algorithms . Peak extraction was applied independently to the normalized ChIP-Chip data derived from antibodies targeting the N- and C-terminus of Wor1 [13] . Peak-finding significance thresholds were kept at their default values ( P≤0 . 001 in the Wor1 ChIPs of wild-type cells and P>0 . 05 in the Wor1 ChIPs of wor1ΔΔ controls ) , though the amount of sampling was increased 10-fold from default to improve significance estimates . The minimum value for peak inclusion/consideration was set to 0 . 25 . All other settings were kept at their default values . It was subsequently determined that the union of Wor1-bound regions defined independently from the N- and C-terminal datasets gave the best concordance with microarray-based and RNA-seq-based gene expression measurements of differential expression . Thus , the 504 Wor1-bound regions used throughout this work result from taking the union of Wor1-bound regions generated from the N- and C-terminal ChIP-Chip datasets . For the purposes of comparing Wor1 binding to differential expression , Wor1-bound regions were associated with nearby divergently transcribed transcripts as depicted in Figure S7 . For the purposes of calculating the distribution of intergenic lengths “in the Wor1 circuit” a slightly different approach was taken to associate Wor1-bound regions with nearby transcripts than described above . In this case , Wor1-bound regions that fall within intergenic regions were associated with all divergent transcripts within 1 kb and intergenic regions that associated with one or more such transcripts were determined to be “in the Wor1 circuit” . This approach avoids the problem of length correction required under the null model that binding sites are distributed randomly throughout the genome ( i . e . , that longer intergenic regions are inherently more likely to have random binding ) . Similarly , to avoid length bias when determining the distribution of 5′ and 3′ UTR lengths “in the Wor1 circuit” , we only considered Wor1-bound regions that resided in the intergenic space immediately upstream of the transcript , thereby avoiding the possibility that random binding to the longer UTRs themselves would drive artificial UTR length discrepancies . Putative cases of UTR length change between cell types were isolated by comparing changes in UTR expression to changes in coding sequence ( CDS ) expression between the cell types . We first calculated differential expression ( in white versus opaque cells ) independently for the 5′ un-translated , coding , and 3′ un-translated regions of each coding transcript . The number of reads aligned within each region of a transcript was counted in the merged set of alignments from each cell type ( i . e . , the two biological replicates for each cell type were combined ) and a single pseudocount was added . The counts for the opaque cell type , whose dataset had 4% more uniquely aligned reads overall , were normalized by the ratio of uniquely aligned reads in the datasets of the two cell types ( i . e . , they were multiplied by a constant factor of 0 . 96 ) . Fold-changes were calculated for each transcript region by dividing the normalized count in opaque by the count in white cells . We then scanned for UTRs whose expression changed more or less than their corresponding coding sequence , as determined by a χ2 test of independence comparing the observed , normalized UTR counts to the expected counts in the two cell types . The expected count for each CDS region in each cell type was calculated by redistributing the total reads counted across cell types for the corresponding UTR in a fashion proportional to the fold-change calculated for the CDS . To ensure accurate fold-change estimates for the CDS regions , only transcripts with a CDS that had at least 50 reads aligned in at least one cell type were considered . By also requiring a minimum 2-fold absolute difference in fold-change values for the UTR and CDS regions and a χ2 P-value less than 10−5 , we identified 145 transcripts with putative UTR length changes ( Table S4 ) . The analysis of transcript features in the Oct4 circuit was performed on publicly available data . Lists of Oct4-bound regions in mouse ES cells determined independently by Chen et al . [25] and Marson et al . [31] were downloaded from supplemental tables provided by these groups in their respective publications . The intersection of bound regions from these two sources was taken to define a high confidence set of Oct4-bound regions that was used for all further analysis . Gene expression measurements of differentiating mouse ES cells were downloaded from a supplemental table provided by Loh et al [29] . For the purposes of our analysis , we considered transcripts that were significantly ( multiple test-corrected P-value ≤10−4 ) up- or down-regulated across the 18 profiling experiments ( median fold-change of at least 1 . 5 ) to be differentially expressed between cell types . Mouse transcript annotations were downloaded from the UCSC Genome Browser ( http://genome . ucsc . edu/ ) and are based on alignments of RefSeq transcripts to assembly mm8 of the mouse genome sequence [69] . The distribution of intergenic lengths “in the Oct4 circuit” was calculated as described above for the Wor1 circuit , except that in the mammalian circuit transcripts could be up to 10 kb away from an Oct4-bound region . We allow a longer distance here since intergenic regions are overall much longer in mouse and because regulation is generally expected to occur over longer distances . The distribution of 5′ and 3′ UTR lengths “in the Oct4 circuit” was calculated as described above for the Wor1 circuit . | The differentiation of cells into distinct cell-types , each of which is “remembered” for many generations , underlies the development of both healthy and cancerous tissues . Such differentiation , however , is not restricted to multi-cellular organisms: “white” and “opaque” cells of the unicellular fungal pathogen Candida albicans are two heritable cell-types , each thought to be adapted to unique niches within their human host . Here we examine the differences between these two cell-types by sequencing their RNA contents and subsequently reconstructing and comparing their gene expression profiles . We know that the transcription factor Wor1 plays a central role in mediating these expression differences . As with many other transcriptional regulators , however , a major unresolved issue is the apparent discordance between the genomic locations to which Wor1 binds and whether neighboring genes are differentially expressed . Here we resolve this discordance , showing that hundreds of Wor1 binding sites , previously without apparent function , actually flank differentially-expressed genes that were undiscovered , or not measured accurately , before . Additionally , we find that transcripts regulated by Wor1 have many unusual properties , several of which we also observe for transcripts regulated during the development of mammalian embryonic stem cells , suggesting they may be general hallmarks of cell differentiation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"cell",
"biology/gene",
"expression",
"developmental",
"biology/stem",
"cells",
"computational",
"biology/transcriptional",
"regulation",
"genetics",
"and",
"genomics/gene",
"expression",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"developmental",
"biology/cell",
"differentiation",
"molecular",
"biology/bioinformatics",
"genetics",
"and",
"genomics",
"genetics",
"and",
"genomics/epigenetics",
"developmental",
"biology/developmental",
"molecular",
"mechanisms",
"computational",
"biology/systems",
"biology"
] | 2010 | The Transcriptomes of Two Heritable Cell Types Illuminate the Circuit Governing Their Differentiation |
Leishmania major is an endemic vector-borne disease in Morocco that causes zoonotic cutaneous leishmaniasis ( ZCL ) , especially in arid pre-Saharan regions where its unique vector and reservoir are Phlebotomus papatasi and Meriones shawi , respectively , and may cause epidemics . In late 2017 , the Zagora province , an endemic focus for ZCL in southern Morocco , had CL outbreak . The main objective of our investigation was to analyze the epidemiological features of this latest ZCL outbreak . We analyzed epidemiological features of this latest ZCL outbreak . The Regional Delegation of Health , Zagora , recorded 4 , 402 CL patients between October 2017 and end of March 2018 . Our findings showed that 24 municipalities were affected and majority ( 55 . 1% ) of infected cases belonged to the Tinzouline rural municipality . Majority of patients were females ( 57 . 2% ) . While all age group patients were affected , those aged <10 years were the most affected ( 42 . 1% ) . During this outbreak over 5 days in December 2017 , we conducted a survey in Tinzouline and recruited and sampled 114 CL patients to confirm CL diagnosis by parasitological ( direct examination and culture ) and molecular ( ITS1-PCR ) methods and identify the etiological agent of infection using ITS1-PCR-RFLP and sequencing . We completed a detailed questionnaire including clinical and epidemiological data for each patient and found 72 . 8% of patients presenting multiple lesions ( ≥2 ) , with an average number of lesions of 5 . 16 ± 0 . 5 . Lesions were more prevalent in the upper limbs , with the most common type being the ulcerocrusted lesion ( 60 . 5% ) . We detected no associations between lesion type and patients’ sex or age . Among 114 clinically diagnosed CL patients , we confirmed 90 . 35% and identified L . major as the species responsible for this outbreak . Self-medication using various products caused superinfection and inflammation of lesions and complicated the diagnosis and treatment . Thus , ZCL remains a major public health problem in the Zagora province , and commitment of all stakeholders is urgently required to implement a sustainable regional control program .
Leishmaniases , a spectrum of diseases that are caused by several species belonging to the genus Leishmania , are transmitted to humans and other mammals by phlebotomine sandflies . Although they cause relatively low mortality , they are responsible for considerable morbidity [1] . A dermal infection known as cutaneous leishmaniasis ( CL ) , which is caused mostly by Leishmania major , L . tropica , and L . infantum , in the Middle East and North Africa , is the most common form of leishmaniasis and one of the so-called neglected diseases , which mainly affects the world's poorest populations [2 , 3] . The incidence of CL , endemic in 87 countries , is constantly on the rise owing to many environmental and socioeconomic factors [4 , 5] . According to the WHO , 90% of the recorded CL cases worldwide are from 12 countries: Afghanistan , Algeria , Brazil , Colombia , Iraq , Iran , Morocco , Peru , Sudan , Syria , Tunisia , and Yemen [6] . In Morocco , three species—L . major , L . tropica , and L . infantum—are responsible for CL . L . major , the principal etiological agent of zoonotic cutaneous leishmaniasis ( ZCL ) , has been known to exist in Morocco since 1914 . Six decades later , Rioux and Peter identified the first epidemic foci in the southeastern part of the kingdom [7] . ZCL is largely confined to arid pre-Saharan regions , where the unique vector and reservoir of L . major are Phlebotomus papatasi ( Sandfly ) and Meriones shawi ( Shaw's Jird ) , respectively [7 , 8] . The rodent M . shawi , a Gerbillidae infesting the oasis-village complex in the southern Morocco was found to be parasitized by L . major since 1982 [9] . In this animal , discrete chronic lesions are almost found on the edge of the ear and rarely on the tail . The infection remains localized to the skin for several months or even years; it is often necessary to wait for the pre-mortem stage to see cutaneous or visceral metastases [7] . Moreover , in the arid areas , the burrows of these rodents provide microhabitats suitable for sand flies breeding and larvae development [9 , 10] . A closely related association between the eruption of M . shawi population and the incidence of ZCL was reported in endemic areas in Morocco [11] . Typically , the foci of ZCL are either palm groves or periurban areas with unhygienic habitats [12] . Clinical manifestations of ZCL are particularly diverse and pleotropic , ranging from a single self-limiting lesion to multiple disfiguring lesions , with , in extreme cases , as many as 30 lesions [13] . The establishment of a primary Leishmania infection and the development of a dermal disease are currently believed to depend on the parasite’s genetic background , host immune response , and factors related to the sandfly . All these components interact in a close relationship to produce different clinical forms [14 , 15] . According to the available records from the Moroccan Ministry of Health , between 2000 and 2016 , a total of 31 , 354 cases of ZCL were reported [16] . In Morocco , ZCL appears to shift between alternating endemic and epidemic cycles , with the latter involving brutal outbreaks lasting two or three years and the former involving long remission periods of about five years or more , during which most affected people are children and newcomers [16 , 17] . Following outbreaks , however , the deceleration of control measures creates a higher risk of severe epidemics in endemic areas . The province of Zagora , a pre-Saharan region in southern Morocco , is an endemic focus for ZLC , where the last outbreak between 2008 and 2010 included 4 , 437 cases [18] . Late in 2017 , reports emerged on a new outbreak with a high incidence in the region . In our study , we aim to understand the epidemiological characteristics of this latest outbreak and to shed light on the features of the ZCL outbreak in order to enable more focused and engaged control procedures to prevent new outbreaks in the future .
We conducted this study in the province of Zagora , in the region of Drâa-Tafilalet in southern Morocco ( Fig 1 ) , a territory dominated by the chain of Anti-Atlas Mountains . The climatic environment is generally part of the Saharan bioclimatic stage , with very low average annual rainfall , decreasing from the north to the south , from 100mm at Agdz to 60 mm at Zagora . The rainy periods fall between September and May , with 30–40 rainy days annually . Tinzouline , a rural municipality in the province of Zagora , is situated at an altitude of 1 , 051m at around 30°30′26 . 1″N 6°06′07 . 9″W , with a total population of about 15 , 000 inhabitants . This municipality is characterized by severe winters ( temperature ranging from −1°C to −7°C ) and hot summers ( 40°C ) . Located 35km away from Zagora , this village was the scene of a ZCL outbreak in 2017; the first cases began to appear from October 2017 . Besides Tinzouline , 23 other municipalities were affected to different extents by ZCL infections ( Fig 1 ) . Between October 2017 and March 2018 , the regional health authority of Zagora province recorded 4 , 402 patients with ZCL; from these records , we extracted the age , sex , and address of each patient . From the onset of the outbreak , the CL cases were diagnosed by parasitological method ( direct microscopic examination ) . Thereafter , giving the large number of infected people , the passive and active case-detection of CL was based only on clinical features . We carried out descriptive statistics , a Chi-square ( χ2 ) test , and Fisher’s exact test using Prism 7 software ( GraphPad Software , Inc . , La Jolla , CA , USA ) and SPSS v . 20 ( SPSS , Inc . , Chicago , IL , USA ) . Statistical significance was defined at P < 0 . 05 . The 95% confidence intervals ( CI ) for the positivity rates were calculated using the Wilson method [19] , implemented in the online program EpiTools epidemiological calculators [20] . During a period of five days in December 2017 , we recruited 114 patients at the health center of Tinzouline . All recruited patients presented skin lesions clinically suggestive of CL , and never treated by Glucantime injection . Pregnancy women , patients presenting chronic illness ( eg , blood pressure issues , diabetes , etc . ) were not skin-sampled . We interviewed each patient using a structured questionnaire comprising all the information about the patient ( patient’s code , age , gender , address , and travel history ) and the disease ( the onset of the lesion , diagnosis , history of treatment , and number and location of lesions ) . We consolidated all completed questionnaires for data input and analysis . We performed patient sampling for Leishmania cultures and smear staining by dermal scraping of the lesion’s edge . For molecular CL diagnosis and identification of Leishmania species , we used the swab sampling method , which is painless and simple to perform . We took swab samples by gently rubbing over the skin lesion approximately five times and then storing them at −20°C until DNA extraction . We fixed and stained all lesion smears with absolute methanol and Giemsa ( Avicenne Group , Casablanca , Morocco ) , respectively , for CL direct diagnosis through microscopic examination . We analyzed all slides twice using a 100x immersion objective . We cultured Leishmania on an RPMI 1640 medium ( Biowest , Nuaillé , France ) supplemented with 2mM L-glutamine ( Eurobio , Les Ulis , France ) , 10% fetal bovine serum ( Biowest , Nuaillé , France ) , and 1% penicillin/streptomycin ( 100U/mL penicillin and 100μg/mL streptomycin; Biowest , Nuaillé , France ) , followed by incubation at 25°C . We performed DNA extraction from a cotton swab . We placed each swab in a 1 . 5mL centrifuge tube containing 250μL of a lysis buffer ( 50mM NaCl , 50mM Tris , and 10mM EDTA; pH = 7 . 4 ) , 1% SDS , and 100 μg/mL proteinase K . After incubating the lysates overnight at 60°C , we subjected them to phenol–chloroform extraction , followed by ethanol precipitation , as described elsewhere [10] . We then quantitatively determined the DNA samples using NanoDrop ( Thermo Fisher Scientific , Waltham , MA , USA ) before dilution to a final concentration of 50ng/μL . We used LITSR and L5 . 8S primers to amplify ITS1 according to the protocol described by Schonian et al . in 2003 [15] , using a negative control ( without DNA ) for each PCR run . In order to identify Leishmania species , we subjected the positive PCR products of 350bp to enzymatic restriction by HaeIII ( New England Biolabs , Hitchin , UK ) for 2h at 37°C . We analyzed RFLPs using electrophoresis on a 3% agarose gel containing ethidium bromide , using a 100bp DNA size marker ( HyperLadder 100bp Plus; Bioline , London , United Kingdom ) . We compared the restriction profiles to the profiles of Moroccan strains previously identified by sequencing as L . infantum , L . major , and L . tropica . We directly sequenced 10 randomly chosen ITS1-PCR products to confirm our PCR-RFLP identification results . We purified them using the Exonuclease I/Shrimp Alkaline Phosphatase ( GE Healthcare , Chicago , IL , USA ) and then sequenced them using BigDye Terminator v3 . 1 Cycle Sequencing Kit ( Applied Biosystems , Foster City , CA , USA ) and an ABI Prism 3130 DNA automated sequencer ( Applied Biosystems ) . Phylogenetic analysis of 10 L . tropica ITS1 sequences , generated in this study , was performed to confirm the identification of L . major as etiological agent , responsible of the outbreak occurred in Zagora province . In addition to our ten ITS1 sequences , other ITS1 L . major sequences and additional Leishmania spp . retrieved from GenBank database were used . For data analysis we used the MEGA version 7 ( http://www . megasoftware . net ) ; phylogram was constructed using the Maximum likelihood algorithm with the Jukes-Cantor model . The tree topology was supported by 1000 bootstrap replicates . Written informed consent was obtained from all the adults who participated in the study . Consent for inclusion of young children , was obtained from parents or guardians . The study and the protocols were approved by the Ethics Committee for Biomedical Research ( CERB ) of the Faculty of Medicine and Pharmacy , Rabat , Morocco .
Between October 2017 and the end of March 2018 , the Regional Delegation of Health , Zagora , clinically diagnosed and recorded a total of 4 , 402 patients with skin lesions from Zagora as cases of ZCL . All patients were freely treated with Glucantime ( meglumine antimoniate ) according to the Moroccan Ministry of Health guidelines . 24 municipalities were affected to different extents; the most affected area during this CL outbreak was the rural municipality of Tinzouline , with 55 . 1% ( n = 2424 ) of the total CL cases , followed by Bouzeroual and Bni Zoli , with 19% ( n = 841 ) and 10% ( n = 443 ) , respectively; the 21 other municipalities recorded a range of 5% ( n = 212 ) to 0 . 02% ( n = 1 ) of the total cases reported ( Fig 1 ) . Almost all ZCL cases were from rural municipalities , with only 2 . 7% from the urban city of Zagora ( Fig 1 ) . Among the 4 , 402 patients with ZCL , 2 , 517 cases were females ( 57 . 2% ) and 1 , 885 were males ( 42 . 8% ) ( Table 1 ) , a statistically significant gender difference ( χ2 = 90 . 73 , P<0 . 0001 ) . Patients with ZCL were aged between two months and 99 years , with an age median of 13 years ( interquartile range: 7–27 years ) , and the majority of the patients were under 10 years old ( Table 1 ) .
Since leishmaniasis became a notifiable disease in Morocco in 1996 , the number of cases in the province of Zagora has been counted by the hundreds . Between 2002 and 2003 , a lull of two years was noted , with 22 and 52 cases , respectively [16] . The next outbreak was triggered in 2008 with 1 , 421 cases , increasing to 1 , 882 cases in 2009 and 1 , 134 cases in 2010 . A very significant steady drop in the number of leishmaniasis cases to 48 cases in 2015 followed this epidemic , before the latter outbreak reported in this work erupted during the last quarter of 2017 . The abruptness of these epidemiological waves and the length of the interepidemic silences cannot be explained without referring to the periarid rainfall regime . This climatic type is characterized by long periods of drought , interspersed intermittently with violent storms that mark the biological cycles profoundly . With the rain , vegetation is the first to react , showing a strong push , followed by the proliferation of rodents and vectors [7 , 21] , after which L . major , hitherto sporadic , multiplies actively . The epizootic precedes the epidemic [7] . Indeed , the inhabitants of Tinzouline , the most affected municipality , reported that the density of the M . shawi rodent reservoir population had become very high; the sylvatic cycle occurring at poor dwellings in different areas within the municipality . Deceleration of the control measures implemented in this province , following the decrease in ZCL cases in the years preceding this outbreak , could also be a risk factor responsible for this outbreak ( e . g . , dump pits , open sewerage , and cattle manure in the vicinity of dwellings [microfocus] ) . Apart from causes , such as environmental conditions , socioeconomic status , and human behaviors , the increase in human leishmaniasis prevalence is mainly attributed to several demographic risk factors , commonly including sex , age , household design , and construction material [2 , 3] . The affected age range is reported to depend on the intensity of transmission ( force of infection ) to which populations are exposed [22] . Our data revealed that all age groups were affected . However , children under 10 years old displayed the highest rate of infection ( 42% ) , while groups above 31 years old showed the lowest rate of infection ( 9 . 7% ) . In established endemic areas , the prevalence of CL was reported to increase generally with age up to 15 years , after which it stabilized , probably reflecting the progressive buildup of immune protective status [2] . Our analysis showed that both sexes were affected; however , women were more affected than men , with the most important difference according to gender observed in groups aged 20 years and above . This factor is related to behavioral patterns that increase the exposure of people to the vector , in our case P . papatasi , the proven vector of L . major in North Africa and the Middle East [23] . During the hot summer nights characterizing this pre-Saharan region , men are known to stay and sleep outdoors ( i . e . , on terraces ) , unlike women , who are often indoors . As P . papatasi is highly endophilic and anthropophagic [24] , women are more susceptible to being bitten by the vector , which may explain the predominance of females infected by ZCL in this study . Regardless of the reasons behind this fact , ZCL may cause cosmetic disfigurement and permanently disfiguring scars , which may create lifelong stigma and impact women’s lives [25 , 26] . In studies carried out in Moroccan rural ZCL foci , substantial gender differences related to the perceived burden and psychosocial consequences of ZCL were reported , in which women seemed to be more strongly stigmatized and the emotional well-being of young single women with facial lesions was strongly affected by ZCL scars [27 , 28] . The emotional representations associated with ZCL among women were also demonstrated to be correlated with the loss of self-esteem and feelings of inferiority [29] . Our clinical analysis of the 114 patients with ZCL showed that the ulcerocrusted lesion form was the most frequent compared to the papulonodular and ulcer forms; the size of the lesion was also very variable . However , what caught our attention was the striking number of lesions observed , in some cases reaching up to 35 lesions . The lesions frequently appeared severely inflamed and superinfected because most patients attempted to treat themselves using different kinds of spices ( salt , chilies , etc . ) , herbs ( henna , white wormwood ) , sand , soil , or toxic products ( tar , bleach water , tobacco , and used engine oil ) , leading to superinfection of lesions and complication of the biological diagnosis and treatment . Clinical manifestations of Leishmania infections depend on multifactorial parameters , such as human genetic susceptibility and the genetic background of the parasite . The factors related to the vector may also affect the CL manifestations [30] . By taking multiple blood meals and multiple inoculations , the vector increases its capacity to transmit parasites , resulting in multiple lesions on the susceptible host , which can lead to disfiguring scars; these forms are often difficult to treat and require specialized advice [31] . Lesions were more common in the exposed parts of the body ( face , upper and lower limbs ) , which appear to be more prone to sandfly bites . However , the location of predilection for L . major infection was the upper limbs . Unlike CL due to L . tropica , where most of the lesions are on the face [32 , 33] , cutaneous leishmaniasis due to L . major is more common in the extremities [34–36] . In endemic ZCL foci , despite the decrease in the number of CL cases , the vector and rodent control measures should be vigorously maintained . In addition to control measures , awareness campaigns for a better knowledge of the disease should be conducted regularly in order to avoid exposure to infections , as well as self-medication , which is responsible for most cases of superinfection and both clinical and therapeutic complications . Finally , operational research and collaboration among researchers , clinicians , veterinarians , and public health authorities is required to establish a suitable strategy for the control of ZCL and to prevent future outbreaks . | Zoonotic cutaneous leishmaniasis ( ZCL ) caused by Leishmania major is endemic in Morocco , especially in the arid pre-Saharan regions . Its main characteristics are rurality and proneness to epidemics . Zagora province an endemic focus for ZCL has been recently the scene of an important CL outbreak . Data analysis of 4 , 402 CL patients recorded in Zagora province , from October 2017 till Mars 2018 showed that 24 municipalities were affected to different extents; the important percent of cases belonged to Tinouzline , a rural municipality . While all age group patients were affected , children under 10 years old were the most affected . Majority of patients were female; the disfiguring CL scars often lead to the stigmatization of women and impact on their psychosocial lives . Lesions were diverse and pleiotropic ranging from a single lesion to multiple disfiguring sores , with , in extreme cases , as many as 35 lesions . Leishmania major was identified as the Leishmania species responsible of this outbreak . Self-medication by using various products caused superinfection and inflammation of lesions and complicated the diagnosis and treatment . Thus , ZCL remains a major public health problem in Zagora province , and commitment of all stakeholders is urgently required to implement a sustainable regional control program . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"morocco",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"geographical",
"locations",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"age",
"groups",
"protozoans",
"signs",
"and",
"symptoms",
"leishmania",
"neglected",
"tropical",
"diseases",
"africa",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"infectious",
"diseases",
"zoonoses",
"bioinformatics",
"lesions",
"biological",
"databases",
"protozoan",
"infections",
"people",
"and",
"places",
"eukaryota",
"diagnostic",
"medicine",
"sequence",
"databases",
"database",
"and",
"informatics",
"methods",
"leishmaniasis",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"organisms"
] | 2019 | Epidemiological features of a recent zoonotic cutaneous leishmaniasis outbreak in Zagora province, southern Morocco |
Mitochondria must buffer the risk of proteotoxic stress to preserve bioenergetics , but the role of these mechanisms in disease is poorly understood . Using a proteomics screen , we now show that the mitochondrial unfoldase-peptidase complex ClpXP associates with the oncoprotein survivin and the respiratory chain Complex II subunit succinate dehydrogenase B ( SDHB ) in mitochondria of tumor cells . Knockdown of ClpXP subunits ClpP or ClpX induces the accumulation of misfolded SDHB , impairing oxidative phosphorylation and ATP production while activating “stress” signals of 5′ adenosine monophosphate-activated protein kinase ( AMPK ) phosphorylation and autophagy . Deregulated mitochondrial respiration induced by ClpXP targeting causes oxidative stress , which in turn reduces tumor cell proliferation , suppresses cell motility , and abolishes metastatic dissemination in vivo . ClpP is universally overexpressed in primary and metastatic human cancer , correlating with shortened patient survival . Therefore , tumors exploit ClpXP-directed proteostasis to maintain mitochondrial bioenergetics , buffer oxidative stress , and enable metastatic competence . This pathway may provide a “drugable” therapeutic target in cancer .
The control of protein homeostasis , or proteostasis , occupies a central , evolutionary-conserved role in organismal integrity and flexible adaptation to environmental “stress” [1] . This pathway involves mechanisms of chaperone-directed protein ( re ) folding [2] as well as removal of aggregated or misfolded proteins via proteolytic degradation [3] . Defects in either process impair organelle function , in particular the endoplasmic reticulum ( ER ) [4] and mitochondria [5] , activating an unfolded protein response ( UPR ) that may culminate in cell death and tissue damage [6] . There is also evidence that a heightened proteostatic threshold can contribute to disease , in particular cancer , by buffering the risk of proteotoxic stress associated with the biosynthetic needs of transformed cells . Accordingly , molecular chaperones of the heat shock protein-90 ( Hsp90 ) family , including Hsp90 [7] and its homolog , TNFR-associated molecule-1 ( TRAP-1 ) [8] , become overexpressed in mitochondria of tumor cells compared to normal tissues [9] and preserve the folding and activity of key effectors of organelle homeostasis [10] . In turn , the heightened proteostatic environment prevents the emergence of a mitochondrial UPR [11] , antagonizes cyclophilin D-dependent apoptosis [9] , and maintains bioenergetics [10] , including oxidative phosphorylation [12] , correlating with unfavorable disease outcome in cancer patients [13] . What has remained unclear , however , is whether chaperone-directed protein folding is the sole mechanism of mitochondrial proteostasis in cancer [10] . In this context , mitochondria contain an evolutionary-conserved , ATP-dependent unfoldase-peptidase protein complex , ClpXP [14] , which mediates proteolytic removal of misfolded proteins [15] . There is evidence that this pathway may regulate a mitochondrial UPR [16 , 17] and contribute to human disease pathogenesis [18] . In this study , we investigated mechanisms of mitochondrial proteostasis as a potential driver of tumor progression .
Previous studies have shown that a pool of the inhibitor-of-apoptosis ( IAP ) protein survivin [19] localizes to mitochondria and contributes to the stability of oxidative phosphorylation Complex II subunit succinate dehydrogenase B ( SDHB ) [20] . To further characterize this pathway , we carried out a proteomics screen for additional survivin-associated molecules in mitochondria , using prostate adenocarcinoma PC3 cells as a tumor model ( S1 Methods ) . In this screen , molecules associated with mitochondrial survivin comprised regulators of organelle trafficking ( Rab11 ) , assembly of respiratory chain complexes ( PTCD1 ) , oxidative stress ( GPX4 ) , mitoribosomal transferase activity ( NSUN4 ) , ubiquinone biosynthesis ( COQ7 ) , and the AAA+ peptidase subunit of the ClpXP complex , ClpP ( Fig 1A ) . Consistent with these results , survivin immune complexes precipitated from mitochondria of PC3 cells contained ClpP as well as TRAP-1 ( Fig 1B ) [20] . Reciprocally , ClpP co-immunoprecipitated with survivin and TRAP-1 in tumor mitochondria ( Fig 1C ) . In addition , ClpX and ClpP mutually associated with each other as well as with SDHB in co-immunoprecipitation experiments ( Fig 1D ) , consistent with the assembly of a survivin-ClpXP [14]-SDHB complex in tumor mitochondria . When analyzed for submitochondrial distribution , survivin , ClpP , and TRAP-1 co-localized within the organelle matrix in PC3 cells ( Fig 1E ) . In addition , a fraction of TRAP-1 and survivin localized to the inner mitochondrial membrane ( Fig 1E ) , in agreement with recent observations [20] . To begin investigating the role of a survivin-ClpXP complex in mitochondria , we next exposed PC3 cells to YM155 , a small molecule survivin suppressant currently examined in the clinic as an anticancer drug . In these experiments , YM155 treatment was associated with reduced expression of ClpP ( Fig 1F ) . As an independent approach , we next silenced survivin using our validated small interfering RNA ( siRNA ) sequences [20] . Survivin knockdown significantly reduced the levels of ClpP in mitochondria of PC3 cells , especially under conditions of hypoxic stress ( Fig 1G ) . In contrast , YM155 treatment ( Fig 1F ) or siRNA silencing of survivin ( Fig 1G ) did not affect the expression of mitochondrial chaperones Hsp70 or Hsp60 , or the voltage-dependent anion channel ( VDAC ) . Next , we used cycloheximide block experiments to quantify changes in ClpXP stability after survivin targeting . In these experiments , YM155 treatment ( S1A and S1B Fig ) or siRNA silencing of survivin ( S1C and S1D Fig ) resulted in accelerated turnover of ClpP , compared to control siRNA transfectants . This regulation occurred at the protein level because survivin knockdown did not affect ClpP mRNA expression ( S1E Fig ) . In reciprocal experiments , transfection of PC3 cells with a survivin variant engineered to selectively accumulate in mitochondria [20] increased ClpP levels in mitochondria ( Fig 1H ) . In contrast , targeting survivin with YM155 ( Fig 1F ) or siRNA knockdown ( S1C and S1D Fig ) or , conversely , over-expression of mitochondrial survivin ( Fig 1H ) did not appreciably affect ClpX levels . This suggests that ClpP is the primary binding partner of survivin in the ClpXP complex , consistent with the results of the proteomics screen ( Fig 1A ) . In mitochondria , survivin associates with the oxidative phosphorylation Complex II subunit , SDHB , and regulates its stability [20] . Therefore , a role of ClpXP in this process [15] was next investigated . First , siRNA silencing of ClpP resulted in increased accumulation of SDHB in mitochondria , whereas COX-IV levels were not affected ( S2A Fig ) . We next looked at the folding status of SDHB in these conditions . We found that siRNA silencing of ClpP caused the accumulation of detergent-insoluble , i . e . , misfolded , mitochondrial Complex II , whereas subunits of oxidative phosphorylation Complex I , III , IV , or V were not affected ( Fig 2A ) . When individual Complex II subunits were examined , ClpP knockdown selectively induced misfolding and aggregation of SDHB at different detergent concentrations ( Figs 2B , 2C and S2B ) . In contrast , ClpP depletion did not affect the solubility of SDHA or other mitochondrial proteins , including COX-IV ( Fig 2A and 2B ) , VDAC , or citrate synthase ( CS; S2B Fig ) , compared to control transfectants . Aggregation and misfolding of SDHB , but not SDHA , after ClpP knockdown was observed under different detergent conditions ( NP-40; S2C and S2D Fig ) . In addition , siRNA silencing of ClpX similarly induced accumulation of aggregated or misfolded SDHB ( Figs 2D and S2E ) . Next , we asked whether SDHB misfolding induced by ClpXP targeting affected mitochondrial respiration . siRNA silencing of ClpX or ClpP did not affect mitochondrial Complex I activity in PC3 cells , compared to control siRNA transfectants ( Fig 3A ) . In contrast , knockdown of ClpX or ClpP significantly reduced Complex II activity in PC3 cells ( Fig 3B ) as well as other prostate cancer cell types , including C4-2 ( S3A Fig ) and DU145 ( S3B Fig ) . This response was specific because ClpX or ClpP targeting had no effect on mitochondrial Complex III ( S3C Fig ) or Complex IV ( S3D Fig ) activity , whereas Complex V function was increased in ClpP- but not ClpX-silenced PC3 cells ( S3E Fig ) . Functionally , impaired Complex II activity after ClpXP targeting reduced oxygen consumption ( Fig 3C ) , increased the NAD/NADH ratio ( Figs 3D and S3F ) , and lowered overall ATP production ( Fig 3E ) in prostate cancer cells . Markers suggestive of compensatory glycolysis , including glucose consumption ( S3G Fig ) or lactate production ( S3H Fig ) , were increased after ClpP- but not ClpX-silencing ( S3G and S3H Fig ) . Consistent with defective bioenergetics , PC3 cells silenced for ClpP or ClpX exhibited increased phosphorylation of the energy sensor 5′ adenosine monophosphate-activated protein kinase ( AMPK ) , whereas total AMPK levels were not affected ( Fig 3F ) . In turn , AMPK phosphorylation coupled to downstream activation of autophagy , as determined by increased LC3-II conversion ( Fig 3G ) , appearance of punctate LC3 staining by fluorescence microscopy ( Fig 3H and 3I ) , accumulation of autophagy markers p62 and Beclin-1 , and increased phosphorylation of ULK1 on its AMPK target site , Ser555 ( Fig 3J ) . Based on these data , we next looked at the downstream consequences of defective mitochondrial respiration in ClpXP-targeted cells . siRNA silencing of ClpP or ClpX resulted in increased total cellular superoxide production in prostate cancer cells ( Fig 4A–4C ) . This response was also associated with heightened production of mitochondria-specific reactive oxygen species ( ROS ) , compared to control transfectants ( Fig 4D ) . To rule out potential off-target effects , we next generated clones of PC3 and DU145 cells with stable shRNA knockdown of ClpP or ClpX ( S4A Fig ) . Stable depletion of ClpXP did not affect total mitochondrial content ( S4B Fig ) or mitochondrial membrane potential ( S4C Fig ) . In contrast , stable knockdown of ClpP increased mitochondrial ROS production ( Fig 4E and 4F ) , with hyperoxidation of mitochondrial peroxiredoxin III ( Prx III ) , a marker of oxidative damage ( Fig 4G ) . Total Prx III levels were not affected ( Fig 4G ) . In contrast , stable shRNA silencing of ClpX did not significantly modulate mitochondrial ROS production ( Fig 4E and 4F ) or Prx III hyperoxidation ( Fig 4G ) , suggesting that other mechanisms compensated for ClpX loss in the stable cell line . Treatment with the antioxidant N-acetylcysteine ( NAC ) or mitochondrial ROS scavenger MitoTempo alone did not rescue Prx III hyperoxidation after ClpXP knockdown ( S4D Fig ) . Conversely , the combination of NAC plus MitoTempo reversed the hyperoxidation of Prx III in these cells ( S4D Fig ) . Functionally , we next asked if oxidative stress induced by ClpXP targeting was important for autophagy induction in these settings . Consistent with this possibility , the combination of NAC plus MitoTempo significantly attenuated LC3-II conversion in ClpP- or ClpX-silenced cells ( Fig 4H and 4I ) , and this response was amplified by the autophagic flux inhibitor hydroxychloroquine ( Fig 4H ) . To complement the results obtained with tumor cell lines , we next looked at a potential differential expression of ClpXP in human cancer . First , ClpP was prominently upregulated in breast adenocarcinoma MCF-7 cells , compared to non-tumorigenic breast epithelial MCF-10A cells ( Fig 5A ) . Similarly , ClpP was highly expressed in primary tissue samples of human prostatic adenocarcinoma , but not normal prostate epithelium ( Fig 5B ) . For comparison , another protease involved in mitochondrial protein quality control , LonP1 , was expressed in normal prostate but not prostate cancer ( Fig 5B ) . When applied to tissue extracts of patient-derived tumor samples , an antibody to ClpP reacted with a predominant single band by western blotting ( S5A Fig ) , reinforcing its specificity . Immunohistochemical staining of a universal cancer tissue microarray demonstrated that ClpP was overexpressed in virtually every human malignancy examined ( Fig 5C ) , with intense cytoplasmic staining in the tumor cell population ( Figs 5D and S5B ) . ClpP expression in cancer patients was independent of grade ( colon adenocarcinoma and CNS tumors ) , Gleason score ( prostate adenocarcinoma ) , histotype ( lung cancer ) , or aggressive versus indolent lymphomas ( S5C Fig ) . Conversely , ClpP levels were increased in histotypes of breast adenocarcinoma compared to normal epithelium ( S5C Fig ) . In addition , ClpP became more prominently expressed in metastatic non-small cell lung cancer ( NSCLC ) , compared to non-metastatic lesions , with the highest levels in brain-metastatic NSCLC ( Figs 5E , 5F and S5D ) . In contrast , no difference in the percentage of ClpP-positive cells was observed in primary or metastatic NSCLC ( Fig 5G ) . Consistent with these results , bioinformatics meta-analysis of public databases ( PrognoScan ) revealed that ClpP expression correlated with poorer outcome in 9 out of 14 analyzed datasets ( 64%; S2 Table and Fig 5H–5J ) , whereas ClpP levels were associated with better prognosis ( S2 Table ) in only one dataset ( Melbourne ) . Importantly , high levels of ClpP expression were associated with shortened distant metastasis-free survival ( S2 Table ) in patients with breast adenocarcinoma ( Fig 5H ) and uveal melanoma ( Fig 5I ) and in abbreviated relapse-free survival in lung adenocarcinoma ( Fig 5J ) . Based on these results , we next asked how ClpXP influenced tumor progression . In a first series of experiments , knockdown of ClpP or ClpX partially reduced tumor cell proliferation ( Figs 6A and S6A ) and inhibited colony formation ( Figs 6B , 6C and S6B ) , a marker of tumorigenicity . The effect of ClpXP silencing on tumor cell proliferation was cell-type-specific and more pronounced after knockdown of ClpP compared to ClpX ( S6A Fig ) . In addition , silencing of ClpP or ClpX minimally reduced proliferation of non-metastatic breast adenocarcinoma MCF-7 cells , and non-tumorigenic breast epithelial MCF-10A cells were not affected ( S6A Fig ) . When characterized in sensitive PC3 cells , knockdown of ClpXP resulted in lower levels of cyclins A , B1 , and D1 ( Fig 6D ) , reduced number of BrdU-positive cells ( S6C Fig ) , and accumulation of cells with G1 DNA content ( S6D Fig ) , consistent with cell cycle arrest . Next , we looked at the mechanism ( s ) of ClpXP regulation of tumor cell proliferation . First , silencing of ClpP or ClpX in MCF-7 cells had little to no effect on mitochondrial Complex II activity ( S6E Fig ) and oxygen consumption rate ( OCR , S6F Fig ) , thus mirroring the marginal sensitivity of these cells to ClpXP targeting ( S6A Fig ) . In addition , ClpP or ClpXP knockdown did not reduce Complex II activity ( S6G Fig ) or OCR ( S6H Fig ) in non-tumorigenic MCF-10A cells , further linking impaired mitochondrial respiration induced by ClpXP targeting to reduced tumor cell proliferation . Based on these results , we next asked if aberrant ROS production associated with ClpXP targeting interfered with tumor cell proliferation . Consistent with this possibility , the combination of antioxidants NAC plus MitoTempo ( MT ) restored tumor cell proliferation ( Fig 6E ) and colony formation ( Fig 6F ) in ClpP-silenced PC3 cells . Similarly , transfection of the ROS scavenger MnSOD rescued the defect of cell proliferation induced by ClpP knockdown ( Fig 6G ) . Finally , ClpXP depletion further sensitized tumor cells to “stress stimuli , ” with more sustained inhibition of tumor cell proliferation mediated by oxidative stress ( H2O2 , Fig 6H ) , high galactose:glucose ratios ( Fig 6I ) , or amino acid deprivation ( 50% amino acids , Fig 6J ) . We next reconstituted ClpP-silenced PC3 cells with siRNA-insensitive ClpP cDNA to test the specificity of these findings . Re-expression of ClpP under these conditions rescued the phosphorylation of Akt , Src , and p70S6K in ClpP-silenced cells ( Fig 6K ) . In addition , re-expression of ClpP reversed the induction of autophagy ( S555 phosphorylation of ULK1 ) , AMPK signaling ( ACC phosphorylation ) , and oxidative stress ( Prx III hyperoxidation ) induced by ClpP silencing ( Fig 6K ) . A potential participation of ClpXP in other tumor traits was investigated next , and we focused on cell motility , which requires mitochondrial bioenergetics [20] . siRNA knockdown of ClpP or ClpX inhibited directional PC3 cell migration in a wound closure assay ( Fig 7A and 7B ) , and suppressed tumor cell invasion across Matrigel-coated Transwell inserts ( Figs 7C and S7A ) . This response was specific because reconstitution of ClpP-silenced cells with a ClpP cDNA restored tumor cell motility in a wound closure assay ( Fig 7D ) . Mechanistically , the combination of antioxidants NAC plus MitoTempo rescued the defect of tumor cell invasion mediated by ClpP knockdown ( Fig 7E ) , demonstrating that increased ROS production in these settings was responsible for the inhibition of cell motility . Next , we mapped the signaling requirements of ClpXP regulation of tumor cell motility . In these experiments , siRNA silencing of ClpP attenuated Akt ( S473 ) phosphorylation ( S7B Fig ) and lowered the expression of cell motility effectors Caveolin-1 ( Cav1 ) and Axl ( S7B Fig ) . ClpXP knockdown was marginally effective ( S7B Fig ) , and no changes in Cav1 mRNA levels were observed in control or ClpP siRNA transfectants ( Fig 7F ) . Exposure of tumor cells to the oxidative stressor H2O2 mimicked this response , causing hyperoxidation of Prx III and loss of Cav1 expression ( S7C Fig ) . Accordingly , siRNA silencing of Cav1 in PC3 cells was sufficient to reduce the levels of phosphorylated Akt ( S473 ) ( S7D Fig ) and suppressed tumor cell migration and invasion , compared to control transfectants ( S7E Fig ) . To validate a role of Cav1 in this cell motility pathway , we next reconstituted ClpP- or ClpX-depleted cells with a Cav1 cDNA . Re-expression of Cav1 in these settings restored Akt phosphorylation ( Fig 7G ) and rescued the defect of tumor cell invasion ( Fig 7H ) after ClpP , but not ClpX knockdown . In contrast , reconstitution of ClpXP-targeted cells with an Akt cDNA had no effect ( S7F Fig ) . Finally , we asked if ClpXP regulation of tumor cell motility was important for metastasis in vivo . Intra-splenic injection of PC3 cells transfected with control shRNA gave rise to extensive metastatic dissemination to the liver of immunocompromised mice within 11 d of injection ( Fig 7I and 7J ) . In contrast , stable silencing of ClpX or ClpP in these cells suppressed the size , number , and extent of liver metastases at the same time interval ( Fig 7I and 7J ) .
In this study , we have shown that the unfoldase-peptidase ClpXP forms a complex with survivin and the Hsp90-like chaperone TRAP-1 in mitochondria of tumor cells . This interaction maintains protein quality control and function of the oxidative phosphorylation Complex II subunit SDHB . Accordingly , genetic targeting of ClpXP causes the accumulation of misfolded SDHB , resulting in impaired bioenergetics , oxidative damage , and activation of stress signals , including autophagy . ClpXP is dramatically upregulated in primary and disseminated human tumors , correlates with shortened patient survival , and mechanistically supports tumor cell proliferation , cell motility , and heightened metastatic competence in vivo . Extensively studied in bacteria [21] and proposed as a regulator of cell persistence [22] , the ClpXP multimolecular complex [23] comprises a ClpX subunit that functions as an ATPase-directed unfoldase for unstructured protein substrates and an internal caseinolytic peptidase , ClpP , which degrades the translocated , unfolded peptides [14] . Although this proteasome-like arrangement is conserved in mammalian cells [15] , the data presented here suggest that the ClpP and ClpX subunits may not have completely overlapping function ( s ) in tumor mitochondria , especially with respect to Akt activation and Cav1-dependent tumor cell motility . It is possible that these different responses reflect individual mitochondrial proteins independently regulated by ClpP or ClpX , or , alternatively , the coupling of individual ClpXP subunits to separate regulators of downstream signaling . The interaction between TRAP-1 [10] and ClpXP [14] described here brings together the two main mechanisms of proteostasis: chaperone-regulated protein folding ( TRAP-1 ) and proteolytic removal of misfolded molecules ( ClpXP ) in a single , functional continuum . A mitochondrial pool of survivin contributes to this proteostasis network , potentially as a scaffolding protein that binds both ClpP ( this study ) and TRAP-1 [20] , and contributes to the stability of the complex . Consistent with the earlier elucidation of a TRAP-1 proteome in tumor mitochondria [10] , a key substrate of the proteostasis network identified here was the iron-sulfur SDHB subunit of oxidative phosphorylation Complex II . The biochemical requirements of mitochondrial respiration and electron transport chain are well understood [24] , but the possibility that these activities may depend on a carefully orchestrated balance of protein folding/protein clearance , especially in the proteotoxic-prone environment of tumor mitochondria [5] , has not been widely considered . Consistent with this possibility , perturbation of the mitochondrial proteostasis network by TRAP-1 targeting [10] or ClpXP knockdown ( this study ) caused SDHB degradation or , conversely , accumulation of misfolded or aggregated SDHB . The exquisite specificity of this response , in which none of the other oxidative phosphorylation complexes are affected , highlights a potential unique propensity of SDHB to protein misfolding , especially in tumor mitochondria , or , alternatively , a more stringent requirement of protein quality control to enable efficient Complex II activity . Irrespective , the loss of SDHB due to defective proteostasis profoundly affected mitochondrial bioenergetics , with decreased oxygen consumption , loss of ATP production , and a phenotype of cellular “stress , ” characterized by activation of a mitochondrial and ER UPR [11] , AMPK phosphorylation [25] , stimulation of autophagy [12] , and loss of mechanistic target of rapamycin ( mTOR ) signaling [20] . As shown here , an important mediator of these responses was the increased production of mitochondrial ROS and the ensuing oxidative damage likely associated with a defective electron transport chain [24] . This model is consistent with other data in the literature that homozygous deletion of TRAP-1 caused increased ROS production , DNA damage , and reduced cell proliferation [26] , whereas TRAP-1 overexpression is protective against oxidative stress [27 , 28] . Here , antioxidants that include a mitochondrial ROS scavenger , MitoTempo , prevented the activation of autophagy and rescued the defects of cell proliferation and tumor cell invasion associated with ClpXP targeting , establishing a causal role of mitochondrial oxidative stress in these tumor traits . Although a role of TRAP-1 in tumor progression is recognized [29] , the possibility that ClpXP-directed proteostasis [14] may be also exploited in cancer has been proposed only recently [30] . In that study , the ClpXP subunit ClpP was found overexpressed in a subset of human acute myeloid leukemias , and pharmacologic or genetic targeting of ClpP impaired mitochondrial oxidative phosphorylation , resulting in leukemia cell killing [30] . The data presented here extend these observations and demonstrate that ClpXP-directed proteostasis is exploited in most human cancers , correlating with shortened patient survival . Where the two studies diverge , however , is in the mechanistic underpinning of the proposed pathway . Here , ClpXP was recognized as a pivotal component of a proteostasis network that , together with TRAP-1 [10] and survivin [20] , ensures mitochondrial homeostasis in tumors . In our hands , and at variance with recent findings [30] , targeting ClpXP only partially reduced tumor cell proliferation and in a cell-type-specific manner , with no measurable effect on tumor cell viability . Instead , we found that ClpXP was required to support directional tumor cell migration , invasion , and heightened metastatic dissemination in vivo . Mechanistically , this pathway involved increased phosphorylation of key cell motility kinases Akt and Src , and reconstitution experiments in ClpP-silenced cells identified the membrane microdomain adapter caveolin-1 [31] as a novel , oxidative , stress-regulated mediator of tumor cell motility . As most epithelial tumors rewire their metabolism toward glycolysis [32] , a role of mitochondrial bioenergetics in cancer has been controversial , and SDHB itself has been at times dubbed as a “tumor suppressor . ” On the other hand , oxidative phosphorylation remains an important energy source in most cancers [33] , fueling critical disease traits such as tumor repopulation after oncogene ablation [34] and drug resistance [35] . The data presented here reinforce this model and establish a key requirement of mitochondrial integrity for tumor cell motility and metastatic competence in vivo . This conclusion fits well with other evidence that oxidative phosphorylation is required for membrane lamellipodia dynamics , turnover of focal adhesion complexes , and phosphorylation of cell motility kinases [20 , 36] , supporting tumor cell invasion and metastatic dissemination , in vivo [25 , 37] . In the context of a proteostasis network , ClpXP may contribute to this response by efficiently removing misfolded or aggregated SDHB molecules to preserve Complex II bioenergetics as well as buffering organelle oxidative stress . In summary , we have shown that the unfoldase-peptidase complex ClpXP [14] is universally exploited in human cancer and contributes to a mitochondrial proteostasis network that controls metabolic reprogramming and downstream signals of tumor cell proliferation , motility , and metastatic competence in vivo . There is now considerable interest in targeting unique features of tumor metabolism , including mitochondrial functions [38] , as a novel approach to cancer therapy . In this context , proof-of-concept studies have demonstrated that small molecule targeting of mitochondrial Hsp90s [25] or ClpXP [30] is feasible and produces potent anticancer activity in preclinical models . Together , this suggests that therapeutic inhibition of the mitochondrial proteostasis network described here may provide a viable strategy to disrupt key requirements of tumor progression .
All patient-related studies were reviewed and approved by an institutional review board at Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico Milan , Italy . A cohort of 53 patients with single brain metastases who underwent surgical resection for curative purposes between 2010 and 2015 was retrieved from the archives of Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico ( Milan , Italy ) and arranged in a tissue microarray ( TMA ) , as described [39] . The predominant primary cancer histotype of metastatic cases was non-small cell lung cancer ( NSCLC , n = 44 ) . Nineteen NSCLC patients for which complete 5-y follow-up records were available were included in the study . During follow-up , 9 NSCLC patients developed metastatic disease to the brain , whereas 10 patients had no evidence of metastasis ( S1 Table ) . A Cancer Universal TMA ( CaU-TMA ) representative of 13 different cancer types ( 10 cases for each tumor type ) was described previously [40] . Mitochondrial protein folding assays were performed as previously described [10] . Briefly , mitochondrial fractions were isolated from PC3 cells transfected with control non-targeting siRNA or ClpP- or ClpX-directed siRNA after 72 h and suspended in equal volume of mitochondrial fractionation buffer containing increasing concentrations of NP-40 ( 0% , 0 . 05% , 0 . 1% , 0 . 2% , 0 . 5% , or 2% ) or CHAPS ( 0% , 0 . 1% , 0 . 25% , 0 . 5% , 1% , or 2 . 5% ) . Samples were incubated for 25 min on ice with vortexing every 5 min , and detergent-insoluble protein aggregates were isolated by centrifugation at 20 , 000 g for 20 min , separated on SDS polyacrylamide gels , and analyzed by western blotting . The various prostate cancer cell types were transfected with control non-targeting siRNA or ClpP- or ClpX-directed siRNA for 48 h or 72 h and analyzed for ATP generation ( BioChain cat No . Z5030041 ) or oxygen consumption ( ENZO Lifesciences cat . No . ENZ-51045-1 ) , as described previously [10] . In other experiments , fresh culture medium containing dialyzed FBS was harvested after 2 h and examined for lactate production ( Abcam cat No . ab65331 ) . The quantification of NAD+ or NADH was measured by enzymatic NADH recycling assay according to the manufacturer’s instruction ( BioVision Cat No . K337-100 ) . Briefly , PC3 cells ( 6 x 105 ) transfected with control non-targeting siRNA or ClpP- or ClpX-directed siRNA were harvested after 72 h , and cell pellets were disrupted by two cycles of freezing and thawing in NADH/NAD+ extraction buffer . Soluble fractions were collected by centrifugation at 20 , 000 g for 5 min and processed for removal of NADH-consuming enzymes using YM-10 ( Millipore ) . For assessment of NADH content , the cycling assay was performed after decomposition of NAD+ by heating at 60°C for 25 min , whereas the decomposition step was omitted for determination of total NAD+/H content . Various prostate cancer cell types were analyzed for changes in oxidative phosphorylation complex activity using Abcam reagents ( Cat . no . ab109721—Complex I , ab109908—Complex II , ab109905—Complex II/III , ab109909—Complex IV ) and Cayman reagent ( 701000—Complex V ) using isolated lysed mitochondria , as described [10] . Briefly , tumor cells were transfected with control non-targeting siRNA or ClpP-or ClpX-directed siRNA and validated for protein knockdown by western blotting , and 2 μg of mitochondrial extracts from each condition were assayed for citrate synthase ( CS ) activity ( ScienCell ) . Aliquots of mitochondrial lysates with comparable CS activity were applied for determination of the individual mitochondrial complex function . Relative complex activities were calculated by determining the change in absorbance over time in the linear range of the individual measurements . To detect total ROS , ClpP- or ClpX-silenced prostate cancer cell types were incubated with 2 . 5 μM of CellROX Green Reagent ( Invitrogen ) for 30 min at 37°C , according to the manufacturer’s instructions . After three washes in PBS , pH 7 . 4 , cells were harvested and analyzed on a FACS Calibur flow cytometer , with the CellROX Green Reagent signal in FL1 . Intact cells were gated in the FSC/SSC plot to exclude small debris . The resulting FL1 data were plotted on a histogram . Superoxide production by mitochondria was visualized by fluorescence microscopy , as described previously [36] . Briefly , 1 . 5 x 104 cells were grown on high optical quality 8-well μ-slides ( Ibidi ) and stained with MitoSOX Red mitochondrial superoxide indicator ( 5 μM , 10 min ) in complete medium , followed by extensive washing in warm complete medium . Stained cells were imaged with a 40X objective on a Nikon TE300 inverted time-lapse microscope equipped with a video system containing an Evolution QEi camera and a time-lapse video cassette recorder . The atmosphere was equilibrated to 37°C and 5% CO2 in an incubation chamber . Phase and red fluorescence ( TRITC filter cube , excitation wavelength: 532–554 nm , emission wavelength: 570–613 nm ) images were captured . To quantitate superoxide levels , files were imported into Image J and masks were manually created around the periphery of the cell based on the phase image and subsequently applied to the TRITC channel to measure intensity . A minimum of 100 cells were analyzed in each independent experiment to obtain mean values . The various tumor cell types suspended in 0 . 1% BSA/RPMI and seeded ( 1 x 105 cells ) in the upper compartment of 8 μM pore diameter BD Transwell membranes ( BD ) were quantified for cell migration , as described [20] . For cell invasion , the Transwell membranes were coated with Matrigel . In all experiments , NIH3T3 conditioned medium was placed in the lower compartment as a chemoattractant [25] . After 18 h incubation at 37°C , the Transwell membranes from each insert were recovered and cells on the upper side ( non-migratory ) were scraped off the surface . Cells on the lower side of the membrane were fixed in methanol , rinsed in water , and mounted on glass slides with Vectashield medium containing DAPI ( Vector Laboratories ) . Migrated cells on each membrane were counted in five different fields at 20x magnification by fluorescence microscopy . All experiments involving animals were carried out in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . Protocols were approved by the Institutional Animal Care and Use Committee ( IACUC ) at The Wistar Institute . A liver metastasis models was performed essentially as described previously [25] . Briefly , PC3 cells stably transfected with control pLKO or ClpP- or ClpX-directed shRNA at 80% confluency were suspended in PBS , pH 7 . 4 , and 50 μl containing 1 × 106 cells were injected in the spleen of 6–8 wk-old male NOD SCID gamma ( NSG , NOD . Cg-Prkdcscid Il2rgtm1Wjl/SzJ ) mice ( Jackson Laboratory ) . Spleens were removed the first day after injection to minimize potentially confounding effects on metastasis due to variable growth of primary tumors . Animals were sacrificed 11 d after injection of the tumor cells , and their livers were resected , fixed in formalin , and paraffin-embedded . Serial liver sections 500 μm apart ( n = 15 per each condition ) were stained with hematoxylin and eosin and analyzed histologically . Metastatic foci were quantified by morphometry and expressed as number and surface areas of metastatic tumor growth compared to total surface area , as described [25] . Data were analyzed using the two-sided unpaired t or chi-square tests using a GraphPad software package ( Prism 6 . 0 ) for Windows . Data are expressed as mean ± SD or mean ± SEM of replicates from a representative experiment out of at least two or three independent determinations . A p value of <0 . 05 was considered statistically significant . | As the powerhouse of the cell and a pivotal hub for oxidative stress , mitochondria must tightly control the state of the proteins that they contain , quickly eliminating misfolded , aggregated , or otherwise damaged proteins . Here , we show that tumor mitochondria manage their set of proteins by assembling an integrated network of protein homeostasis , or proteostasis , that controls both the folding and degradation of proteins . This protein complex is formed by the unfoldase-peptidase ClpXP , survivin , and the Hsp90-like chaperone TRAP-1 , and regulates the function of the oxidative phosphorylation Complex II subunit succinate dehydrogenase B ( SDHB ) . We find that interference with this process impairs energy production , promotes oxidative stress , and shuts down critical downstream signals important for tumor cell proliferation , invasion , and metastatic dissemination in vivo . Our results suggest that the mitochondrial proteostasis network may offer therapeutic opportunities in advanced disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"urology",
"cell",
"motility",
"medicine",
"and",
"health",
"sciences",
"gene",
"regulation",
"metabolic",
"processes",
"carcinomas",
"cancers",
"and",
"neoplasms",
"genitourinary",
"tract",
"tumors",
"oxidative",
"phosphorylation",
"oncology",
"prostate",
"cancer",
"adenocarcinomas",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"small",
"interfering",
"rnas",
"lung",
"and",
"intrathoracic",
"tumors",
"gene",
"expression",
"prostate",
"diseases",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"non-small",
"cell",
"lung",
"cancer",
"biology",
"and",
"life",
"sciences",
"energy-producing",
"organelles",
"non-coding",
"rna",
"metabolism"
] | 2016 | The Mitochondrial Unfoldase-Peptidase Complex ClpXP Controls Bioenergetics Stress and Metastasis |
Vγ9/Vδ2 T cells are a minor subset of T cells in human blood and differ from other T cells by their immediate responsiveness to microbes . We previously demonstrated that the primary target for Vγ9/Vδ2 T cells is ( E ) -4-hydroxy-3-methyl-but-2-enyl pyrophosphate ( HMB-PP ) , an essential metabolite produced by a large range of pathogens . Here we wished to study the consequence of this unique responsiveness in microbial infection . The majority of peripheral Vγ9/Vδ2 T cells shares migration properties with circulating monocytes , which explains the presence of these two distinct blood cell types in the inflammatory infiltrate at sites of infection and suggests that they synergize in anti-microbial immune responses . Our present findings demonstrate a rapid and HMB-PP-dependent crosstalk between Vγ9/Vδ2 T cells and autologous monocytes that results in the immediate production of inflammatory mediators including the cytokines interleukin ( IL ) -6 , interferon ( IFN ) -γ , tumor necrosis factor ( TNF ) -α , and oncostatin M ( OSM ) ; the chemokines CCL2 , CXCL8 , and CXCL10; and TNF-related apoptosis-inducing ligand ( TRAIL ) . Moreover , under these co-culture conditions monocytes differentiate within 18 hours into inflammatory dendritic cells ( DCs ) with antigen-presenting functions . Addition of further microbial stimuli ( lipopolysaccharide , peptidoglycan ) induces CCR7 and enables these inflammatory DCs to trigger the generation of CD4+ effector αβ T cells expressing IFN-γ and/or IL-17 . Importantly , our in vitro model replicates the responsiveness to microbes of effluent cells from peritoneal dialysis ( PD ) patients and translates directly to episodes of acute PD-associated bacterial peritonitis , where Vγ9/Vδ2 T cell numbers and soluble inflammatory mediators are elevated in patients infected with HMB-PP-producing pathogens . Collectively , these findings suggest a direct link between invading pathogens , microbe-responsive γδ T cells , and monocytes in the inflammatory infiltrate , which plays a crucial role in the early response and the generation of microbe-specific immunity .
The immune system has evolved to survey the body constantly for potentially hazardous structures . In order to initiate an appropriate defense , sentinel cells need to encounter ‘danger’ signals derived from invading microbes or stressed tissue [1] . Microbial signals comprise pathogen-associated molecular patterns ( PAMPs ) that are invariant among a broad range of organisms , allow self/non-self discrimination , and are detected by germline-encoded pattern recognition receptors [2] , [3] . By contrast to this innate immune recognition , the adaptive immune response is mediated via somatically rearranged and clonally distributed antigen receptors on B cells and αβ T cells . Unconventional T cells expressing γδ T cell receptors ( TCRs ) do not easily fit into this scheme , as they integrate features of innate and adaptive immune recognition [4]–[6] . In humans and higher primates , Vγ9/Vδ2 T cells comprise a small lymphocyte population in peripheral blood ( typically 0 . 5–5% of all T cells [7] ) that shows a striking propensity for expansion in many infections [8] . With their unique specificity for the low molecular weight compound HMB-PP [9] , Vγ9/Vδ2 T cells are specialized in targeting a distinctive and vital metabolite shared by a broad range of bacteria ( and some protozoan parasites ) that is absent in all higher eukaryotes including humans [10] , [11] . Selection and peripheral amplification of public Vγ9–Jγ1 . 2 clonotypes during early childhood appears to ensure rapid , innate-like responses of Vγ9/Vδ2 T cells to invading pathogens in later life [12] , [13] . HMB-PP is 10'000 times more active in vitro than any other physiological compound [14] , and the potential of microbial pathogens to stimulate Vγ9/Vδ2 T cells correlates with their ability to produce HMB-PP [15] , [16] . Still , Vγ9/Vδ2 T cells may in vivo also respond to lower activity agonists such as isopentenyl pyrophosphate and dimethylallyl pyrophosphate released locally by necrotic host cells , and may thus alert the immune system to invading pathogens as well as to tissue damage and progressing tumors [8] , [17] . Importantly , HMB-PP and isopentenyl pyrophosphate do not require presentation by human leukocyte antigen ( HLA ) class I and II molecules or CD1 [18] , supporting a sentinel function of Vγ9/Vδ2 T cells [19] , [20] . Activated Vγ9/Vδ2 T cells display distinct natural killer ( NK ) cell-like functions and directly eliminate infected and transformed cells , a feature that is successfully being exploited in immunotherapy trials in cancer patients [21] , [22] . Another intriguing innate-like aspect is their potential to act as professional antigen-presenting cells ( APCs ) , which includes uptake and processing of antigens and the induction of antigen-specific αβ T cell responses [23]–[25] . However , there is a paucity of data on Vγ9/Vδ2 T cells from anatomical locations other than blood and secondary lymphoid tissue , and information on the response of Vγ9/Vδ2 T cells in acute infection is especially sparse , not least due to the absence of HMB-PP-reactive γδ T cells in small animal models [4]–[6] , [8] . Studies in severe combined immunodeficiency mice reconstituted with human peripheral blood mononuclear cells ( PBMC ) suggested that Vγ9/Vδ2 T cell may mediate rapid clearance of intraperitoneal infections , by enhancing monocyte-mediated killing of bacteria through production of interferon ( IFN ) -γ and tumor necrosis factor ( TNF ) -α [26] . Vγ9/Vδ2 T cells do not reside at common sites of pathogen entry , such as skin , lung , gastrointestinal or urinary tract , and it is unclear under which conditions they are recruited to peripheral tissues [27] . Infections trigger the local production of inflammatory chemokines , which control the composition of the cellular infiltrate [28] , [29] . Importantly , a change in local chemokines is an essential factor in the transition from the neutrophil-driven immediate response to the T and B cell-driven later response to infection . The majority of circulating Vγ9/Vδ2 T cells and monocytes expresses the chemokine receptors CCR2 and CCR5 and displays similar migration characteristics [27] , [30]–[32] . Vγ9/Vδ2 T cells are thus well equipped for instant relocation from the circulation and co-recruitment to inflammatory processes , which explains their accumulation at sites of infection [33] . We have here examined whether the joined extravasation of Vγ9/Vδ2 T cells and monocytes in response to pathogens is of functional relevance . Our present findings demonstrate a rapid and HMB-PP-dependent crosstalk between Vγ9/Vδ2 T cells and monocytes in the presence of microbes , leading to highly activated γδ T cells on one hand , and to monocyte differentiation into inflammatory dendritic cells ( DCs ) on the other hand . This interaction establishes conditions that support further recruitment of effector cells in acute infection , enhance local phagocyte activity , and create antigen-presenting cells ( APCs ) for the initiation of microbe-specific adaptive immunity . Importantly , these events translate directly to episodes of acute bacterial peritonitis and suggest that γδ T cells play a pivotal role in the immediate response to infection .
Optimum γδ T cell stimulation with the microbial metabolite HMB-PP or related compounds requires the presence of accessory cells including monocytes [18] , [34]–[36] . Here , freshly isolated Vγ9/Vδ2 T cells rapidly responded to HMB-PP in the presence of autologous monocytes as judged by induction of activation markers and expression of cytokines ( data not shown ) . However , we noticed that stimulation of Vγ9/Vδ2 T cells with HMB-PP also had a pronounced reciprocal effect on the co-cultured monocytes , and led to cluster formation and the appearance of elongated cells within 18 hours ( Fig . 1A ) . These spindle-shaped monocytes were not observed in unstimulated monocyte-γδ T cell co-cultures , or in pure monocyte cultures incubated overnight with HMB-PP or other bacterial compounds , such as lipopolysaccharide ( LPS ) or peptidoglycan ( PGN ) ( Fig . S1A ) . In addition , this remarkable effect was also not seen in monocytes treated overnight with granulocyte/macrophage colony-stimulating factor ( GM-CSF ) +IL-4 , a cytokine combination giving rise to monocyte-derived DCs within a week of culture [37]; or with macrophage colony-stimulating factor ( M-CSF ) , which induces differentiation toward macrophages [38] ( Fig . S1A ) . In many cases , monocyte clusters formed around γδ T cells , suggesting their engagement in tight cellular interactions ( Fig . S1B ) . In support of substantial crosstalk between these cells , the majority of monocytes survived in co-cultures with Vγ9/Vδ2 T cells in the presence of HMB-PP , whereas typically only <30% of all monocytes cultured in HMB-PP alone or together with resting Vγ9/Vδ2 T cells were still viable after two days . This ‘conditioning’ with Vγ9/Vδ2 T cells plus HMB-PP also resulted in monocyte forward/side scatter profiles that agreed with the morphologic features seen in the corresponding cultures ( Fig . 1B ) . Collectively , these data not only confirm that monocytes provide ‘feeder’ qualities for optimum stimulation of Vγ9/Vδ2 T cells with HMB-PP but that they unexpectedly also receive reciprocal differentiation signals , which are more potent than , and distinct from , any other in vitro stimulus tested . Next , we examined the phenotypic changes that occurred in monocytes during overnight co-culture with Vγ9/Vδ2 T cells and HMB-PP ( abbreviated here as γδHMB-PP ) . γδHMB-PP-activated monocytes down-regulated surface CD14 within 18 hours ( Fig . 2A ) , similarly to monocytes treated with GM-CSF+IL-4 but unlike monocytes treated with M-CSF ( Table S1 ) . At the same time , HMB-PP-stimulated Vγ9/Vδ2 T cells induced an up-regulation of the APC markers CD40 , CD86 , and HLA-DR on monocytes , to levels that were comparable , or even superior , to those seen with GM-CSF+IL-4 or with M-CSF . Importantly , neither resting Vγ9/Vδ2 T cells nor HMB-PP alone showed this activity . Cross-titrations confirmed a highly selective and dose-dependent effect of HMB-PP-stimulated Vγ9/Vδ2 T cells on monocytes . Down-modulation of CD14 was readily observed at an HMB-PP concentration of 0 . 1 nM , and at a ratio of 1 γδ T cell per 500 monocytes ( Fig . 2B ) . Induction of APC markers in monocytes occurred with comparable efficiencies , as illustrated for CD40 . Of note , induction of mRNAs for CD40 , CD86 , and HLA-DR in γδHMB-PP-activated monocytes was very rapid and already pronounced after 4 . 5–6 hours ( Fig . 2C ) . Finally , protein expression levels for APC markers on γδHMB-PP-activated monocytes after only 18 hours of co-culture readily exceeded those observed on fully differentiated monocyte-derived DCs or macrophages ( Fig . 2D ) . Collectively , these data demonstrate that γδHMB-PP-activated monocytes undergo a rapid and substantial differentiation program toward an APC phenotype . The phenotypic changes seen in γδHMB-PP-activated monocytes might have stemmed from contact-dependent monocyte-γδ T cell interactions . Hence , we added neutralizing antibodies against major integrin components to co-cultures , concomitantly with HMB-PP . Blocking of CD11a or CD18 abrogated morphologic changes and cluster formation , while antibodies against CD11b or CD49d showed no such effect ( Fig . 3A ) , indicating a crucial involvement of lymphocyte function-associated antigen-1 ( LFA-1 , CD11a/CD18 ) but not macrophage antigen-1 ( Mac-1 , CD11b/CD18 ) or very late antigen-4 ( VLA-4 , CD49d/CD29 ) . As a consequence of broken cell clusters , monocytes were not transformed into APCs in the absence of LFA-1 contacts , since anti-CD11a and anti-CD18 , but not anti-CD11b or anti-CD49d , antibodies inhibited CD14 down-modulation , and up-regulation of CD40 and CD86 ( Fig . 3B ) . Collectively , these data demonstrate that the morphological changes and the rapid acquisition of APC markers by monocytes require cell-cell interactions , leading to activation of Vγ9/Vδ2 T cells in the presence of HMB-PP and reciprocal activation of monocytes . The possible contribution of soluble factors was examined in transwell experiments by measuring the response of monocytes in the lower chamber to molecules released by monocyte-γδ T cell co-cultures in the upper chamber . In this setting , HMB-PP-stimulated but not unstimulated co-cultures produced soluble factors that crossed the separating membrane and down-modulated CD14 as well as induced expression of CD40 , CD86 , and HLA-DR . These factors included IFN-γ and TNF-α , since addition of neutralizing antibodies against IFN-γ and soluble TNF-α receptor ( sTNFR ) reversed the effects on CD14 and CD40 expression in transwell cultures ( Fig . S2; and data not shown ) . We corroborated these findings in HMB-PP-stimulated monocyte-γδ T cell co-cultures , where addition of anti-IFN-γ and sTNFR , but not anti-GM-CSF or anti-IL-4 , inhibited morphologic changes and monocyte survival ( Fig . 4A; Fig . S2 ) . γδ T cell-derived IFN-γ and TNF-α also appeared to be the major regulators of cell surface CD14 , CD40 , and CD86 expression , whereas GM-CSF and IL-4 had only minor effects ( Fig . 4B ) . Still , a cocktail of blocking reagents against IFN-γ , TNF-α , IL-4 , and GM-CSF inhibited the γδ T cell-induced down-modulation of CD14 and up-regulation of CD40 and CD86 on monocytes more than just the combination of anti-IFN-γ+sTNFR ( Fig . S3 ) . Of note , both recombinant IFN-γ+TNF-α and GM-CSF+IL-4 promoted cell survival in pure monocyte cultures , down-modulated CD14 , and induced APC marker expression . In addition to CD40 , CD86 , and HLA-DR , γδHMB-PP-activated monocytes expressed CD83 , CD206 ( mannose receptor ) and CD209 ( DC-SIGN ) ( Fig . 4C ) , while CD205 ( DEC-205 ) and CD207 ( langerin ) were absent ( Table S1 ) . Of note , GM-CSF+IL-4 but not IFN-γ+TNF-α mimicked the effect of HMB-PP-stimulated Vγ9/Vδ2 T cells on CD206 and CD209 expression in pure monocytes ( Fig . S4 ) , and anti-GM-CSF+anti-IL-4 blocked the γδ T cell-induced expression of CD206 and CD209 in co-cultures ( Fig . 4C ) . Neutralization of IFN-γ and TNF-α increased the levels of CD206 and CD209 even further , in line with reports showing that acquisition of ‘classical’ DC markers such as CD1a and CD209 by monocytes is dependent on IL-4 and counteracted by IFN-γ [39] . Collectively , these data demonstrate that the rapid acquisition of DC-like features by monocytes is mediated in part through the HMB-PP-driven release of IFN-γ , TNF-α , GM-CSF , and IL-4 by Vγ9/Vδ2 T cells . Microbial sensing by DCs induces the release of a plethora of factors , the combination of which reflects the type of the encountered microorganism and the quality of the T cell response required for the control of this particular pathogen . We therefore examined the cytokine profile of γδHMB-PP-activated monocytes . Addition of HMB-PP to monocyte-γδ T cell co-cultures not only induced TNF-α production in γδ T cells ( data not shown ) , it also led to expression of significant levels of TNF-α in co-cultured monocytes ( Fig . 5A ) . TNF-α mRNA was rapidly induced in γδHMB-PP-activated monocytes and abundantly present after 4 . 5–6 h , at levels even exceeding those in the co-cultured γδ T cell population ( Fig . 5B ) . Similarly as described above for CD14 and APC markers , TNF-α expression was highly dependent on the concentration of HMB-PP and the number of γδ T cells in the co-cultures ( Fig . 5C ) . These data suggest a positive feed-back mechanism in the inflammatory response , where both monocytes and γδ T cells rapidly produce large amounts of TNF-α . Two other inflammatory cytokines expressed by γδHMB-PP-activated monocytes were IL-6 and oncostatin M ( OSM ) , which were detectable in the supernatants of HMB-PP-stimulated co-cultures ( Fig . 5D ) ; IL-6 was confirmed to be expressed in monocytes by intracellular staining ( data not shown ) . Expression of TNF-α and IL-6 but not OSM by γδHMB-PP-activated monocytes could be blocked by anti-IFN-γ+sTNFR ( Fig . 5A , 5D ) , and mimicked by addition of recombinant IFN-γ+TNF-α to pure monocyte cultures ( data not shown ) . By contrast , the same monocyte-γδ T cell co-culture conditions did not lead to induction of IL-1β , IL-10 , IL-12 , IL-23 , or IL-27 , as assessed by protein and/or mRNA expression analyses ( Table S1 ) . Intriguingly , although γδHMB-PP-activated monocytes expressed high levels of CD40 , engagement of this receptor by soluble trimeric CD40L did not result in detectable levels of IL-12 or IL-23 but led to the release of substantial amounts of IL-6 ( data not shown ) . Finally , neither HMB-PP-stimulation of monocyte-γδ T cell co-cultures nor treatment of monocytes with recombinant IFN-γ and/or TNF-α was able to induce significant levels of nitric oxide in monocytes ( data not shown ) [40] . However , these co-culture conditions led both γδ T cells ( data not shown ) and monocytes ( Fig . 5D ) to express the pro-apoptotic effector molecule TNF-related apoptosis inducing ligand ( TRAIL ) [41] . Collectively , these data demonstrate that crosstalk of monocytes and HMB-PP-stimulated γδ T cells leads to the rapid generation of a highly inflammatory milieu . The function of APCs is largely governed by their recruitment and relocation properties . Of note , HMB-PP-stimulated monocyte-γδ T cell co-cultures were a rich source of chemokines with known functions in acute infections , as shown for CXCL8 ( IL-8 ) and CCL2 ( MCP-1 ) , which target neutrophils and monocytes , respectively , and CXCL10 ( IP-10 ) , one of three IFN-γ-inducible chemokines that control the recruitment of effector T cells . All three chemokines were rapidly induced in monocyte-γδ T cell co-cultures in the presence of HMB-PP and could be blocked by anti-IFN-γ+sTNFR ( Fig . 6A ) . Circulating monocytes express a series of chemokine receptors , including CXCR4 and the prototype monocyte receptors CCR2 and CCR5 [28] . Here , γδHMB-PP-activated monocytes down-modulated surface expression of CXCR4 , CCR2 , and CCR5 ( Fig . 6B; and data not shown ) , which may have resulted from chemokine-mediated receptor internalization [42] . Intriguingly , the chemokine receptor CCR7 , which enables mature DCs and naïve T cells to co-localize within the T-zone of lymph nodes , was not induced under these conditions ( Fig . 6B ) but was readily detected when further microbial stimuli such as LPS ( Fig . 6C ) or PGN ( not shown ) were added to HMB-PP-stimulated monocyte-γδ T cell co-cultures . Collectively , these data demonstrate that γδHMB-PP-activated monocytes produce cytokines and chemokines typically associated with inflammatory sites but lack factors such as CCR7 that are mobilized in response to Toll-like receptor ( TLR ) ligands . The chemokine profile of γδHMB-PP-activated monocytes and their capacity to switch from inflammatory chemokine receptors to CCR7 is reminiscent of immature DCs undergoing maturation and acquisition of lymph node homing properties [43] . The resemblance of γδHMB-PP-activated monocytes with inflammatory DCs prompted us to examine their ability to present antigens and induce αβ T cell activation . First , we established that monocytes co-cultured with γδ T cells retained endocytic activities , as assessed by uptake of soluble proteins , Lucifer yellow and dextran , and phagocytosis of bacteria ( data not shown ) . These experiments ruled out the presence of inhibitory factors associated with DC maturation . Next , we studied the activation of αβ T cells in response to autologous monocytes presenting Mycobacterium tuberculosis purified protein derivative ( PPD ) , which requires uptake and processing , and Staphylococcus aureus superantigen toxic shock syndrome toxin-1 ( TSST-1 ) , which is loaded directly onto cell surface HLA class II [23] . γδHMB-PP-activated monocytes induced a significant expansion of PPD-specific T cells from 5- ( and 6- ) carboxyfluorescein diacetate succinimidyl ester ( CFSE ) -labeled , naïve responder αβ T cells , as evidenced by the appearance of proliferating CD45RO+ αβ T cells after only 4 days of culture; this αβ T cell expansion was not seen in the presence of freshly isolated monocytes or monocytes co-cultured with resting Vγ9/Vδ2 T cells ( Fig . 7A ) . Of note , neutralization of IFN-γ and TNF-α during co-culture of monocytes with HMB-PP-stimulated Vγ9/Vδ2 T cells prevented this response , which fully agrees with the observed inhibition of APC marker expression ( Fig . 4 ) . Similarly , γδHMB-PP-activated monocytes served as APCs for the expansion of TSST-1-specific CD45RO+ Vβ2+ cells from CFSE-labeled , naïve responder αβ T cells ( data not shown ) . Generation of TSST-1-specific αβ T effector cells was very robust when using γδHMB-PP-activated monocytes as APCs , as evidenced by the large proportion of IFN-γ-producing Vβ2+ T cells , but was largely absent with control monocytes ( Fig . 7B ) . As opposed to IFN-γ , IL-17 was not induced in TSST-1-specific responder T cells by γδHMB-PP-activated monocytes , whereas γδHMB-PP-activated monocytes treated with LPS or PGN were able to do so ( Fig . 7B; and data not shown ) , confirming that TLR ligands have a direct impact on the generation of microbe-specific Th17 cells [44] . Collectively , these data suggest that monocytes turn into Th1 cell-inducing inflammatory DCs in the presence of HMB-PP-stimulated Vγ9/Vδ2 T cells and possess the capacity for further development into lymph node seeking ( CCR7+ ) Th17 cell-inducing DCs upon contact with microbes via TLR signaling . In order to extend our study to situations of disease , we examined peritoneal effluent cells from individuals on continuous ambulatory peritoneal dialysis ( PD ) ( Table S2 ) , where bacterial infection and associated inflammation remain a frequent complication . The peritoneal catheter of PD patients allows convenient , repeated , and non-invasive sampling of recruited leukocytes , and provides unique access to inflammatory scenarios in vivo [45] . Under stable , i . e . non-inflamed conditions , peritoneal effluent cells consisted mainly of CD3+ lymphocytes and CD14+ monocytes . Importantly , Vγ9/Vδ2 T cells represented a minor but detectable fraction of peritoneal leukocytes ( 0 . 06–0 . 28% ) . As these values were similar to the activation threshold in our titration experiments with peripheral monocyte-γδ T cell co-cultures ( Fig . 2B ) , we tested whether peritoneal γδ T cells interact with peritoneal monocytes in a similar way . Indeed , addition of HMB-PP led to cluster formation and the appearance of larger , activated cells within 18 hours . This morphological change was significantly inhibited by anti-IFN-γ+sTNFR or anti-CD11a+anti-CD18 neutralizing antibodies ( Fig . S5A ) . The Vγ9/Vδ2 T cells in these peritoneal cultures showed a dose-dependent response to HMB-PP that was already apparent at 0 . 1 nM , as judged by expression of CD25 , CD69 , and TNF-α ( Fig . S5B ) . Moreover , HMB-PP at concentrations of >10 nM led to expansion of Vγ9/Vδ2 T cells ( i . e . , in the absence of exogenously added cytokines ) ( Fig . S5B ) . In the same cultures , monocytes showed a corresponding dose-dependent response to HMB-PP at 0 . 1 nM and higher , as judged by increased forward scatter , down-modulation of CD14 , up-regulation of CD40 and CD86 , and induction of TRAIL and TNF-α ( Fig . S5C ) . Collectively , these data demonstrate that Vγ9/Vδ2 T cells are present in the peritoneal cavity and able to induce rapid monocyte differentiation in the presence of minute quantities of HMB-PP . In addition to the situation in non-infected individuals , Vγ9/Vδ2 T cells were also detectable during episodes of PD-associated bacterial peritonitis . As expected from early stage infections , neutrophils represented the vast majority of peritoneal cells , while the number of Vγ9/Vδ2 T cells varied considerably ( from <0 . 01% up to 7 . 3% of total leukocytes ) and could amount to several million cells in the over-night effluent ( Fig . 8A ) . Compared to stable , non-infected controls , the proportion of Vγ9/Vδ2 T cells among peritoneal CD3+ T cells and among total peritoneal effluent cells was considerably augmented in patients with acute peritonitis as a result of infection with the HMB-PP+ bacteria E . coli , Leclercia , or Bacteroides ( Fig . 8A ) . This was not the case in patients infected with HMB-PP− staphylococci or streptococci , which confirms in vivo that human Vγ9/Vδ2 T cells selectively recognize microbial pathogens capable of synthesizing HMB-PP ( Table S3 ) [8]–[11] , [14]–[16] . Peritoneal Vγ9/Vδ2 T cell frequencies and CD69 expression levels in HMB-PP+ peritonitis patients were consistently higher than in the peripheral blood of the same patients ( data not shown ) , implying local recruitment/proliferation and activation in the peritoneal cavity . In addition , highest numbers of activated , CD69+ Vγ9/Vδ2 T cells were seen at the earliest time-points post-infection ( p . i . ) and steadily declined over the following days to the background values seen in HMB-PP− bacterial infections and in non-infected individuals ( Fig . 8A ) , underscoring an immediate as opposed to long-lasting involvement of HMB-PP-responsive γδ T cells in bacterial infections . Finally , we examined peritoneal effluent for the presence of a series of soluble factors ( IL-6 , TNF-α , TRAIL , CXCL10 , and OSM ) that we found prominently expressed in HMB-PP-stimulated monocyte-γδ T cell co-cultures . Remarkably , all five proteins could be detected in patients samples and were higher in HMB-PP+ infections ( E . coli , Neisseria , Proteus , Pseudomonas , unspecified coliform and coryneform bacteria ) than in HMB-PP− staphylococcal or streptococcal infections on day 1 p . i . ( Fig . 8B; 9 . 75 and 7 . 38 pg/ml OSM , n . s . ) . Moreover , and in agreement with the gradual reduction in peritoneal γδ T cell numbers , cytokine levels similarly declined over the following days p . i . ( Fig . 8C; and data not shown ) . Collectively , these data imply that Vγ9/Vδ2 T cells play a significant role in PD-associated acute infection caused by HMB-PP+ producing pathogens .
We have employed an in vitro model , composed of autologous Vγ9/Vδ2 T cells , monocytes , and the microbial metabolite HMB-PP , that mimics local conditions at early stage infections while leaving out components of the immediate ( neutrophils ) and adaptive ( T and B cells ) immune response . We hypothesized that our model would shed light on the role of Vγ9/Vδ2 T cells in the control of anti-microbial immunity . Our present findings demonstrate a reciprocal interaction between γδ T cells and monocytes within 18 hours , leading to the rapid induction of a remarkable differentiation program in monocytes . The γδ T cell-mediated effects on monocytes included enhanced survival; production of inflammatory cytokines and chemokines; and the development of DC-like characteristics , as evidenced by morphologic features , expression of APC markers , uptake and processing of antigens , and induction of antigen-specific αβ T cell responses . Of note , these inflammatory DCs were not generated when HMB-PP was omitted from the monocyte-γδ T cells co-cultures , illustrating the importance of TCR-triggering in γδ T cells for initiation of monocyte differentiation . Also , these changes in monocytes did not require the addition of exogenous cytokines and occurred within 18 hours , in contrast to what was previously reported in co-cultures with NK cells or NKT cells [46] , [47] . While the mechanisms of driving monocyte differentiation toward DCs may be similar between γδ T cells and NK cells and involve the same factors , in those studies co-culture periods of up to 6 days with IL 15-activated NK cells or excess cloned NKT cells were required . In our own experiments , we observed an intimate and rapid monocyte-γδ T cell crosstalk that occurred within a few hours , at subnanomolar HMB-PP concentrations , and at ratios as low as 1 γδ T cell per monocyte . These conditions resemble early stages of infections where monocytes outnumber γδ T cells , and where low numbers of bacteria produce trace amounts of HMB-PP . Of note , we previously estimated lysates of E . coli [48] and L . monocytogenes [16] bacterial cells to contain 200–300 nM HMB-PP . Thus , we conclude that our experimental setup adequately mirrors an early aspect in acute anti-microbial defense and implies a role for inflammatory DCs , generated by short-term culture of monocytes in the presence of Vγ9/Vδ2 T cells and HMB-PP , in pathogen clearance and induction of microbe-specific T cell responses . At the same time , activated Vγ9/Vδ2 T cells may acquire APC features themselves and contribute to the transition from the innate to the adaptive phase of the immune response [23]–[25] . The pronounced effect of the microbial metabolite HMB-PP on Vγ9/Vδ2 T cells at low nanomolar concentrations is remarkable but remains elusive . Possible scenarios include direct binding of HMB-PP to the TCR , alone or in the context of a ‘presenting’ molecule on accessory cells , or the generation of a secondary Vγ9/Vδ2 T cell-specific ligand in response to HMB-PP [49]–[52] . Also , it is not clear if invading pathogens release free HMB-PP into the microenvironment , or whether HMB-PP becomes only accessible upon processing of bacteria by phagocytic cells . Studies with the intracellular pathogen Mycobacterium tuberculosis suggest that uptake of whole bacteria by monocytes , macrophages , or DCs is required for the recognition of HMB-PP , highlighting an essential role for monocytic cells in the activation of Vγ9/Vδ2 T cells [53]–[55] . It is conceivable that neutrophils , which are the first cells to be mobilized in response to bacterial infections , may also contribute to Vγ9/Vδ2 T cell activation . Irrespective of the underlying mechanism , the extremely potent yet highly selective activation of Vγ9/Vδ2 T cells by a single microbial compound is reminiscent of bona fide PAMPs and other ‘danger’ signals that our innate immune system has learnt to recognize in order to mount robust anti-microbial responses [1]–[3] . The generation of inflammatory DCs during co-culture with HMB-PP was the result of a combination of direct cell-cell contact and cytokines released by activated γδ T cells . Macroscopic changes in the co-cultures included LFA-1-dependent cell clustering as well as polarized monocyte spreading , and inhibition of integrin function prevented the formation of inflammatory DCs , in agreement with the role of LFA-1 in conjugate formation between γδ T cells and other cells [18] , [56] . Transwell assays and experiments with neutralizing antibodies revealed the importance of Vγ9/Vδ2 T cell-derived soluble factors in monocyte activation , foremost IFN-γ , TNF-α , GM-CSF , and IL-4 , which are co-expressed by the same Vγ9/Vδ2 T cell population upon stimulation with HMB-PP [36] . However , no cytokine combination fully substituted for HMB-PP-stimulated Vγ9/Vδ2 T cells , and complete inhibition was not achieved with a cocktail of blocking reagents against IFN-γ , TNF-α , GM-CSF , and IL-4 , implying additional soluble and/or cell-associated components in promoting monocyte activation , including IL-13 , CD40L , and integrin ligands [31] , [36] . Thus , we conclude that HMB-PP-stimulated Vγ9/Vδ2 T cells are perfectly equipped to interact with monocytes by providing a whole range of factors necessary for monocyte survival and differentiation into inflammatory DCs . Clearly , these findings differ from the reported effect of γδ T cell-derived IFN-γ and TNF-α on monocyte-mediated killing of bacteria [26] or on the maturation of immature DCs [55] , [57]–[59] . The overall outcome of our in vitro co-cultures was a milieu rich in inflammatory mediators that gave rise to cells with characteristics of immature DCs . We did not detect IL-1β , IL-12 , IL-23 , or IL-27 in the culture supernatants , a fact that may be explained by the absence of TLR ligands or other ‘danger’ signals . Despite uniform expression of CD40 on γδHMB-PP-activated monocytes , signaling through this receptor by means of co-culture with CD40L+ Vγ9/Vδ2 T cells or activation with soluble CD40L did not lead to substantial levels of IL-12 or IL-23 , further documenting the paramount importance of synergism with microbial products [60] . The inflammasome-controlled processing of IL-1β is induced by numerous microbial ligands [61] , and the absence of IL-1β ( and IL-23 ) in our co-cultures could explain the observed inability of inflammatory DCs to induce the differentiation of naïve αβ T cells into Th17 cells [44] . In agreement , we were able to ‘correct’ this deficit by adding LPS or PGN to our co-cultures , which resulted in inflammatory DCs capable of inducing Th17 cells . Finally , the lymph node homing receptor CCR7 was completely absent in inflammatory DCs but became expressed upon treatment with LPS or PGN . We therefore conclude that our co-culture system allows a view at the cellular cross-talk between two types of simultaneously recruited immune cells that occurs in the absence of TLR signaling . This model will be useful for studying the parameters that determine the ‘quality’ of microbe-specific αβ T cell responses ( Th1 , Th2 , Th17 , Tfh , Treg ) by treating inflammatory DCs with defined microbial compounds or whole pathogens . It is conceivable that our model of γδHMB-PP-induced inflammatory DCs is more relevant to induction of adaptive immunity in response to acute infections as opposed to monocyte-derived DCs that require one week of in vitro culture for development [37] . Recent studies in mice provided compelling evidence that circulating monocytes are able to develop into DCs in vivo [62] , [63] . Murine Gr-1+ CCR2+ monocytes ( which correspond to the CD14high human blood monocytes used in the present study ) are recruited to sites of infection where they differentiate into inflammatory DCs [64] , [65] . The factors responsible for the conversion of blood monocytes into DCs in humans and the kinetics by which they act are largely unknown , where access to tissue material at early stage infections is limited . Evidently , inflammatory chemokines , e . g . those produced under the settings of bacterial invasion , will selectively recruit blood monocytes expressing the corresponding chemokine receptors while factors provided by local tissue-resident cells and microbes will affect survival and differentiation of recruited monocytes . The powerful monocyte-γδ T cell crosstalk reported here translates directly to episodes of bacterial peritonitis . Our data demonstrate that Vγ9/Vδ2 T cells in peritonitis patients infected with HMB-PP+ bacterial species were always highest at the earliest time points when infected peritoneal samples could be retrieved ( day 1–2 p . i . ) , and then gradually declined over the next few days . In contrast , Vγ9/Vδ2 T cell frequencies and total numbers in patients infected with HMB-PP− bacterial species did not vary from the stable , non-inflamed situation , and remained constant over the course of one week after the onset of infection . Most remarkably , this initial peak and rapid resolution of Vγ9/Vδ2 T cell numbers in HMB-PP+ peritonitis clearly preceded the delayed influx of αβ T cells ( day 3–4 p . i . ) yet overlapped with the early wave of neutrophils ( day 1–2 p . i . ) that was observed previously in PD-associated peritonitis patients [45] , [66]–[68] . An attractive hypothesis would predict that some of the mediators we identified in our in vitro model control neutrophil recruitment and function ( IFN-γ , IL-6 , OSM ) [66] , [67] as well as neutrophil turnover ( TRAIL ) [69] . Accordingly , we could demonstrate that responses to HMB-PP of unfractionated peritoneal cells were indistinguishable from those seen in our in vitro model , as evidenced by cytokine production ( including IFN-γ , TNF-α , and TRAIL ) and induction of APC markers on monocytes . Thus , Vγ9/Vδ2 T cells are ideally positioned to contribute to immediate infection control and to support microbe-specific adaptive immune responses in patients with HMB-PP+ bacterial infections . Ongoing studies are designed to examine the monocytic infiltrate and DC subsets in acute bacterial peritonitis and detect phenotypic and functional differences between HMB-PP+ and HMB-PP− infections . Of note , disproportionate monocyte-γδ T cell crosstalk may result in excessive production of inflammatory mediators , possibly explaining why episodes of HMB-PP+ peritonitis are associated with a 2–3fold increased risk of PD technique failure ( removal of the peritoneal catheter , transfer to hemodialysis , and/or patient death ) and implying a role for γδ T cells in the nature and severity of the inflammatory response to pathogens ( our unpublished observations ) . In summary , our in vitro and ex vivo data support a model where Vγ9/Vδ2 T cells bridge innate and adaptive immune mechanisms in response to infection with HMB-PP-producing pathogens . At the earliest stage of infection , microbial products activate local macrophages and tissue cells to produce neutrophil-specific ( CXCL8 and related chemokines ) and monocyte/γδ T cell-specific chemokines ( CCL2-5 ) [70] . Freshly recruited Vγ9/Vδ2 T cells interact with monocytes and become activated by microbial-derived HMB-PP , which in turn leads to substantial cytokine secretion and the generation of inflammatory DCs . Following antigen-uptake and processing , newly generated DCs upregulate CCR7 in response to microbes , and relocate to the draining lymph nodes where they instruct microbe-specific effector T cells . Thus , Vγ9/Vδ2 T cells set the stage for early adaptive immune responses while invading microbes determine the choice of play .
This study was conducted according to the principles expressed in the Declaration of Helsinki and under local ethical guidelines ( Bro Taf Health Authority , Wales ) . The study was approved by the South East Wales Local Ethics Committee under reference number 04WSE04/27 . All patients provided written informed consent for the collection of samples and subsequent analysis . 43 patients on PD for ≤5 years were recruited from the Peritoneal Dialysis Unit , Cardiff University School of Medicine ( Table S2 ) . Diagnosis of acute peritonitis was based on the presence of abdominal pain , a cloudy peritoneal effluent with >105 leukocytes per ml , and a positive microbiological culture . The day of the first appearance of leukocytes in the effluent was defined as day 1 post-infection . The causative organisms of bacterial peritonitis were divided into HMB-PP− ( Enterococcus , Staphylococcus , Streptococcus ) and HMB-PP+ species ( Bacteroides , Corynebacterium , Escherichia , Leclercia , Neisseria , Proteus , Pseudomonas , and other coliform or coryneform bacteria ) , in accordance with the distribution of the non-mevalonate pathway of isoprenoid biosynthesis across their genomes ( Table S3 ) [8] , [11] . Patients with bacterial episodes of peritonitis were uniformly treated with a standard regime of ciprofloxacin and vancomycin according to the guidelines of the International Society for Peritoneal Dialysis ( ISPD ) . Peritoneal cells were harvested from chilled overnight dwell effluents [45]; cell-free supernatants were stored at −70°C . PBMC were isolated from peripheral blood using Lymphoprep ( Axis-Shield ) . Monocytes ( >99% ) were purified from PBMC of healthy volunteers using anti-CD14 microbeads ( Miltenyi ) . Vγ9/Vδ2 T cells ( 99 . 04±0 . 65% Vγ9+ , mean±SD ) were purified from CD14-depleted PBMC using monoclonal antibodies ( mAbs ) against Vγ9-PE-Cy5 ( Immu360; Beckman-Coulter ) and anti-PE microbeads ( Miltenyi ) . Untouched bulk αβ T cells ( >95% ) and naïve CD4+ T cells ( >95% ) were purified from γδ T cell-depleted PBMC using the pan-T cell isolation kit II and the naïve CD4+ T cell isolation kit ( Miltenyi ) , respectively . Immature DCs were derived from monocytes over 6 days in 50 ng/ml GM-CSF and 10 ng/ml IL-4 ( Peprotech ) ; mature DCs were obtained from DCs by adding 100 ng/ml LPS ( Sigma ) for 15 h . Macrophages were derived from monocytes over 6 days in 50 ng/ml M-CSF ( Peprotech ) . Cells were analyzed on a four-color FACSCalibur supported with CellQuest ( BD Biosciences ) , using mAbs against pan-TCRγδ ( 11F2 ) , TCR-Vδ2 ( B6 . 1 ) , CD3 ( SK7 , UCHT1 , HIT3a ) , CD4 ( RPA-T4 ) , CD14 ( MOP9 ) , CD25 ( M-A251 ) , CD45RA ( HI100 ) , CD45RO ( UCHL-1 ) , CD69 ( FN50 ) , CD83 ( HB15e ) , CD86 ( 2331 ) , HLA-DR ( L243 ) , CCR5 ( 2D7 ) , and CXCR4 ( 12G5 ) ( all from BD Biosciences ) ; TCR-Vβ2 ( MPB2D5 ) , TCR-Vγ9 ( Immu360 ) , CD40 ( mAB89 ) , CD206 ( 3 . 29B1 . 10 ) , and CD207 ( DCGM4 ) from Beckman Coulter; CD209 ( 120507 ) and CCR2 ( 48607 . 211 ) from R&D Systems; CD205 ( DEC-205 ) from eBioscience; TCR-Vδ1 ( TS8 . 2 ) from Endogen; and rat anti-CCR7 ( 3D12 ) from Dr . M . Lipp ( Max Delbrück Center for Molecular Medicine , Berlin , Germany ) ; together with appropriate isotype controls and secondary reagents . For detection of intracellular cytokines , brefeldin A ( Sigma ) was added to cultures at 10 µg/ml 4 hours prior to harvesting . Surface-stained cells were labeled using the Fix&Perm kit ( eBioscience ) and mAbs against IFN-γ ( 45 . 15 ) , TRAIL ( RIK-2 ) ( BD Biosciences ) , TNF-α ( 188 ) ( Beckman Coulter ) , IL-6 ( AS12 ) , IL-10 ( JES3-9D7 ) , and IL-17 ( 64DEC17 ) ( eBioscience ) . Monocytes in co-cultures were identified based on their appearance in forward/sideward scatter , lack of CD3 expression and residual expression of CD14; γδ T cells were gated on CD3+ Vγ9+ lymphocytes . The medium used was RPMI-1640 with 2 mM L-glutamine , 1% non-essential amino acids , 1% sodium pyruvate , 50 µg/ml penicillin/streptomycin , 50 µM β-mercaptoethanol , and 10% fetal calf serum ( Invitrogen ) . Monocytes were co-cultured with γδ T cells at a ratio of 5–10 monocytes per γδ T cell in the presence of 10 nM synthetic , i . e . LPS-free HMB-PP [14] , with no additional stimuli . Monocytes incubated with γδ T cells or HMB-PP alone served as controls . In transwell experiments , monocytes were separated from monocyte-γδ T cell co-cultures by 0 . 4 µm pore polycarbonate membranes ( Fisher Scientific ) . Alternatively , monocytes were cultured with 10 ng/ml IFN-γ , 20 ng/ml TNF-α , 50 ng/ml GM-CSF , 50 ng/ml M-CSF , 10 ng/ml IL-1β , 10 ng/ml IL-4 , or 50 ng/ml IL-6 ( Peprotech ) ; 100 ng/ml trimeric CD40L with 1 µg/ml enhancer ( Alexis ) ; 100 ng/ml LPS from Salmonella abortus equii ( Sigma ) ; 5 µg/ml PGN from Staphylococcus aureus ( Sigma ) ; or combinations of which . Blocking reagents used were anti-IFN-γ ( 25718 ) , anti-GM-CSF ( 3209 ) , anti-IL-4 ( 3007 ) , and anti-CD40L ( 40804 ) from R&D Systems; anti-CD11a ( TS1/22 ) , anti-CD11b ( OKM1 ) , and anti-CD18 ( TS1/18 ) from Dr . R . Pardi ( DIBIT-Scientific Institute San Raffaele , Milano , Italy ) ; anti-CD49d ( HP2/1 , HP2/11 ) from Dr . F . Sánchez-Madrid ( Hospital Universitario de la Princesa , Madrid , Spain ) ; and sTNFR p75-IgG1 fusion protein ( etanercept , Enbrel® ) from Amgen; alone or in combination at 10 µg/ml each . Monocytes were activated as described above except that the co-cultured γδ T cells were irradiated at 12–19 Gray . Freshly isolated or pre-activated monocytes were washed and used as APCs for autologous bulk TCRαβ+ T cells or naïve CD4+ T cells , at a ratio of 5–10 αβ T cells per monocyte; for proliferation assays , responder αβ T cells were pre-labeled with CFSE ( Molecular Probes ) . TSST-1 ( Toxin Technology ) was pulsed directly onto APCs at 1 ng/ml for 1 h; PPD ( Statens Serum Institut ) was added to the T cell cultures at 1 µg/ml . Monocytes were co-cultured with γδ T cells as described above except that in some assays γδ T cells were purified using biotinylated pan-TCRγδ mAbs ( 11F2; BD Biosciences ) and anti-biotin microbeads ( Miltenyi ) , and labeled with PKH26 ( Sigma ) . Photographs were taken from live cultures at a magnification of 200× , using a Leica DM IRBE inverted microscope with a Hamamatsu ORCA-ER camera supported with OpenLab 3 . 1 . 7 ( Improvision ) . Images were processed with Photoshop 6 . 0 ( Adobe ) . γδ T cells and monocytes from 4 . 5–6 hours co-cultures with or without HMB-PP were sorted to >99 . 5% purity each on a MoFlo machine ( Cytomation ) , using mAbs against Vγ9 , Vδ2 , CD3 , CD4 , and CD14 . Freshly isolated γδ T cells and monocytes served as controls . Total RNA was isolated using Trizol ( Invitrogen ) and reverse transcribed with SuperScript II in the presence of 500 µg/ml random hexamer primers and 100 mM dNTPs ( Invitrogen ) . Real-time PCRs were run on ABI Prism 7000 and 7900HT systems , using 2×ABI master mix ( Applied Biosystems ) , 0 . 9 µM forward and reverse primers , and 0 . 25 µM 5′-FAM and 3′-BHQ1 labeled probes ( Microsynth ) . Primer sequences are listed in Table S4; amplification efficiencies were between 1 . 92 and 2 . 0 ( R2>0 . 95 ) . Relative gene quantification was performed in duplicate using the 2−ΔΔCT method . Results were expressed as expression levels relative to 1 , 000 copies of cyclophilin A . Cell-free peritoneal effluents and culture supernatants were analyzed using ELISA kits for IL-1β , IL-6 , IL-12p70 , IL-17 , IL-27 , CXCL10 , TRAIL , and OSM ( R&D Systems ) ; IFN-γ , TNF-α , CXCL8 , and CCL2 ( BD Biosciences ) ; and IL-23 ( eBioscience ) . Monocyte-derived nitric oxide was assessed using the Total Nitric Oxide and Nitrate/Nitrite Parameter Assay Kit ( R&D Systems ) . All samples were measured in duplicate on a Dynex MRX II reader . Data were analyzed using two-tailed Student's t-tests ( GraphPad Prism 4 . 0 ) , with differences considered significant as indicated in the figures: * , p<0 . 05; ** , p<0 . 01; *** , p<0 . 001 . | As antibiotic resistance is spreading and posing a significant threat in many bacterial diseases , there is a need for a better understanding of host responses to infection . The precise role of an enigmatic subset of human immune cells , so-called Vγ9/Vδ2 T cells , in early infection still remains to be unveiled . These cells respond to a common molecule shared by the majority of bacterial pathogens and appear to be quickly drawn to sites of acute inflammation , where they will encounter invading microbes in the context of other immune cells , mainly granulocytes and monocytes . We here observed an unexpected interplay between microbe-activated Vγ9/Vδ2 T cells and monocytes that attracts further effector cells , enhances the activity of scavenger cells , and promotes the development of microbe-specific immunity . These findings not only improve our insight into the complex cellular interactions in early infection but may also suggest new therapies by modulating immune responses to improve host defenses and to resolve inflammatory activities . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"immunology/cellular",
"microbiology",
"and",
"pathogenesis",
"microbiology/immunity",
"to",
"infections",
"immunology/immune",
"response",
"microbiology/innate",
"immunity",
"immunology/innate",
"immunity",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis",
"nephrology/dialysis",
"and",
"renal",
"transplantation",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections",
"microbiology/medical",
"microbiology",
"immunology/leukocyte",
"activation"
] | 2009 | A Rapid Crosstalk of Human γδ T Cells and Monocytes Drives the Acute Inflammation in Bacterial Infections |
Highly active antiretroviral therapy ( HAART ) suppresses human immunodeficiency virus ( HIV ) replication to undetectable levels but cannot fully eradicate the virus because a small reservoir of CD4+ T cells remains latently infected . Since HIV efficiently infects only activated CD4+ T cells and since latent HIV primarily resides in resting CD4+ T cells , it is generally assumed that latency is established when a productively infected cell recycles to a resting state , trapping the virus in a latent state . In this study , we use a dual reporter virus—HIV Duo-Fluo I , which identifies latently infected cells immediately after infection—to investigate how T cell activation affects the estab-lishment of HIV latency . We show that HIV latency can arise from the direct infection of both resting and activated CD4+ T cells . Importantly , returning productively infected cells to a resting state is not associated with a significant silencing of the integrated HIV . We further show that resting CD4+ T cells from human lymphoid tissue ( tonsil , spleen ) show increased latency after infection when compared to peripheral blood . Our findings raise significant questions regarding the most commonly accepted model for the establishment of latent HIV and suggest that infection of both resting and activated primary CD4+ T cells produce latency .
Once highly active antiretroviral therapy ( HAART ) became available in 1995 , HIV infection was transformed from a deadly disease into a chronic lifelong condition [1] . The antiretroviral drugs used in HAART target multiple stages of the viral lifecycle , which can reduce patient viremia to undetectable levels [2–4] . However , HAART cannot eradicate HIV [5] because infected individuals harbor a small reservoir of latently infected cells that contain a transcriptionally silent but reactivatable provirus [6] . Because this latent reservoir prevents viral eradication , there is an urgent need to study and better understand the mechanisms of latency . HIV infection primarily targets CD4+ T cells , and the most extensively studied latent reservoir resides within resting CD4+ T cells [7–9] . During infection , HIV enters a target cell and reverse-transcribes its genomic viral RNA into a double-stranded cDNA that then enters the nucleus and integrates into the host genome , where it becomes controlled by the host transcriptional machinery . In most cases , integration of the viral genome leads to productive infection , in which viral genes are transcribed followed by virion production . However , in rare instances , latency occurs instead of productive infection and is characterized by a provirus that produces little-to-no viral transcripts [10] . Because the latently infected cell is not producing viral proteins , it escapes the viral cytopathic effects and is ignored by the immune system . Furthermore , since antiretroviral drugs only target active viral replication , they are ineffective against latent proviruses . Latent HIV is primarily found within memory CD4+ T cells , which have a long half-life in vivo [11 , 12] , allowing latent virus to persist within infected individuals for decades [13] . However , when latently infected memory CD4+ T cells encounter an antigen or are exposed to specific cytokines or chemokines , proviral transcription is activated , leading to productive infection [8 , 14] . This “reactivation” is likely the cause of viral rebound after a patient stops HAART , and it explains why infected individuals must take antiretroviral drugs for life . HIV latency has proven difficult to study because latently infected cells are very rare in vivo ( ~1 in 1 × 106 cells ) [11] , and they cannot be distinguished from uninfected cells [15] . Despite these challenges , several in vitro latency models exist , which have led to important observations about how latently infected cells are maintained and reactivated ( reviewed in references [16 , 17] ) . However , it is not clear how the latent reservoir is established because current technologies only quantify latently infected cells by reactivating them from latency . We recently developed a dual reporter virus , HIV Duo-Fluo I , that can distinguish between cells that are productively infected , latently infected , or uninfected , and allows us to purify each population [18] . Using this new reporter virus , we can study the kinetics of HIV latency immediately after infection by employing two separate fluorescent markers: an LTR-driven eGFP marker ( productive infection ) and an LTR-independent mCherry marker ( latent infection ) driven by an EF1α promoter ( Fig 1A ) . It should be noted that we use the term “productive infection” here and throughout the manuscript to indicate an infection resulting in the expression of the LTR-driven GFP reporter . Since the virus used in this manuscript is env-deficient , these infections are not truly productive . However , they are behaving like a productive infection in terms of virus expression levels . Using this dual reporter virus , we have studied how HIV latency is established with a unique focus on the role of T cell activation . Based primarily on in vitro evidence , it is generally accepted that HIV predominantly replicates in activated CD4+ T cells [19–22] . Conversely , resting CD4+ T cells present several barriers to HIV infection ( reviewed in reference [23] ) , as they do not support efficient nuclear import [24] or integration of the viral cDNA [22 , 25] . However , the most notable obstacle to infection of resting CD4+ T cells occurs at the stage of reverse transcription [26 , 27] . Resting CD4+ T cells do not support reverse transcription nearly as efficiently as activated cells because , at least in part , they contain the restriction factor SAMHD1 [28 , 29] . Additionally , in vivo , most HIV-infected resting CD4+ T cells exhibit a memory phenotype , suggesting that they arose from the infection of previously activated CD4+ T cells . Based on this evidence , a leading theory postulates that latency is established from infected activated CD4+ T cells that revert back to a resting memory state . According to this model , the transition to a resting memory state is associated with a decrease in NFκB and pTEFb activity , two critical factors for HIV transcription , and with a concomitant silencing of the HIV genome [30] . However , for this type of latency to occur , the infected cell would have to survive the virus-induced cytopathic effects and the host immune response that usually kill productively infected cells very quickly ( cells survive ~1 day ) [31 , 32] . Another possibility is that infection occurs at a “sweet spot” in the trajectory that activated CD4+ T cells taken from full activation to a fully rested state . This sweet spot would be characterized by permissivity for HIV reverse transcription and integration but not for HIV transcription [33] . Interestingly , previous studies have reported that resting CD4+ T cells can be directly infected , with the strongest evidence coming from in vivo and ex vivo studies of both SIV and HIV infection [34–39] . Most studies that show resting CD4+ T cells can be directly infected have been performed with cells isolated from primary lymphoid tissues . In vivo studies have found that resting CD4+ T cells in lymphoid tissue harbor viral RNA [35] , and ex vivo studies have shown that directly infecting resting CD4+ T cells from lymphoid tissue results in productive infection [40] . Strikingly , a subsequent study found that resting CD4+ T cells in ex vivo lymphoid cells isolated from tonsillar tissue can support HIV infection , but purified CD4+ T cells isolated from that same lymphoid tissue could not [41] , suggesting that the lymphoid tissue microenvironment is critical for rendering resting CD4+ T cells permissive to HIV infection . Indeed , several lymphoid tissue–associated factors , including cytokines [42] , chemokines [43] , extracellular matrixes [44] , and cell surface markers [45] , enhance HIV infection in resting CD4+ T cells . Therefore , HIV latency may be established by the direct infection of resting CD4+ T cells when they are exposed to soluble factors that do not induce classic T cell activation . In this study , we use the dual reporter virus , HIV Duo-Fluo I , to investigate the role of T cell activation on the establishment of HIV latency in primary CD4+ T cells . We also use HIV Duo-Fluo I to explore the theories of how HIV latency is established; namely , whether it occurs through infection of activated CD4+ T cells that return to a resting state or through the direct infection of resting CD4+ T cells . We find that both resting and activated primary CD4+ T cells can support both productive and latent infection . In the case of activated T cells , the latent state is established very early in the infection and is not significantly influenced by the return of that activated cell to a resting state . We further observed that the fraction of cells that become latent ( latent/productive ) is higher in resting CD4+ T cells than in activated CD4+ T cells .
The literature is replete with conflicting reports on whether resting CD4+ T cells can be infected by HIV , either productively or latently [23] . Many studies indicate that resting CD4+ T cells are refractory to productive HIV infection but can become permissive to infection after treatment with certain cytokines or chemokines that do not induce classic T cell activation [46 , 47] . To test the permissivity of resting CD4+ T cells to our HIV Duo-Fluo I virus , we isolated total CD4+ T cells from peripheral blood of uninfected donors via depletion of all non-CD4+ T cells ( negative selection ) . These cells did not express CD69 or CD25 ( Fig 1B ) and can therefore be considered resting . These cells were either left untreated or were treated with the cytokine IL-7 or the chemokine CCL19 for 72 h prior to infection . Stimulation with IL-7 or CCL19 slightly increased CD25 expression , such that 0 . 58% and 1 . 38% of cells expressed CD25 , respectively ( Fig 1B ) . As a positive control , resting CD4+ T cells were stimulated with αCD3/αCD28 activating beads in the presence of IL-2 for 72 h prior to infection , which led to significant expression of both CD69 and CD25 activation markers ( Fig 1B ) . Both untreated and treated cells were spinoculated with HIV Duo-Fluo I for 2 h at 37°C and then returned to culture in the presence of IL-2 . Productive infection ( GFP+ and mCherry+/GFP+ ) and latent infection ( GFP-/mCherry+ ) were monitored daily by flow cytometry for 6 days following infection ( S1 Fig ) . Compared to αCD3/αCD28-stimulated cells at 6 days postinfection , untreated resting CD4+ T cells showed significantly lower levels of HIV infection but , nonetheless , permitted both productive and latent infection ( Fig 1C ) . Importantly , productive and latent infection of resting CD4+ T cells over the 6-day time-course was not the result of replication-competent virus being present in our HIV Duo-Fluo I viral stocks ( S2 Fig ) [48] . Despite minimally affecting T cell activation , both IL-7 and CCL19 treatment led to an increase in HIV infection compared to untreated resting CD4+ T cells ( Fig 1C and 1D ) , suggesting that the permissibility of resting CD4+ T cells can be enhanced without undergoing classic T cell activation , which agrees with previous studies [47 , 49] . However , we show how productive and latent infection is distributed within resting CD4+ T cells after such treatments ( Fig 1D ) . Treating resting CD4+ T cells with CCL19 increased productive infection 2-fold over untreated cells , while IL-7 treatment produced a 5-fold increase in productive infection . Latent infection increased by 3-fold after treatment with either CCL19 or IL-7 . In addition to cytokine and chemokine treatment , we also investigated the role of the human protein SAMHD1 in restricting HIV infection within resting CD4+ T cells . To do this , we infected resting CD4+ T cells with HIV Duo-Fluo I containing Vpx , which was provided in trans as a fusion protein with Vpr . Vpx is a lentiviral accessory protein encoded by HIV-2 that degrades SAMHD1 and thereby allows the virus to infect many cell types—resting CD4+ T cells [28] , dendritic cells , monocytes , and macrophages [50 , 51]—that are usually off limits to HIV-1 because of a SAMHD1-imposed post-entry block . Infecting untreated resting CD4+ T cells with Vpx-containing HIV Duo-Fluo I increased their infection levels over untreated cells infected with HIV Duo-Fluo I alone ( Fig 1C and 1D ) . This increase in infection correlates with SAMHD1 protein down-regulation mediated by Vpx ( S3A Fig ) and has little to do with T cell activation ( S3B Fig ) . As such , levels of productive infection increased significantly and were almost comparable to those of activated CD4+ T cells , whereas levels of latent infection were more than 2-fold greater than in resting untreated cells infected with HIV Duo-Fluo I alone . Overall , this increase in infection suggests that knocking down SAMHD1 in resting CD4+ T cells biases the cells toward productive infection , an observation that is confirmed by calculating the ratio of latently infected to productively infected cells ( Fig 1E ) . Based on the ratio of latently infected to productively infected cells , IL-7–treated resting CD4+ T cells and activated CD4+ T cells also support more productive infection than latent infection . Conversely , untreated and CCL19-treated resting CD4+ T cells support more latent infection than productive infection . To ensure that infecting resting CD4+ T with HIV Duo-Fluo I did not lead to any silent infection events—in which viral integration occurred but failed to produce expression of either fluorescent marker—we sorted the uninfected populations ( GFP-/mCherry- ) of both untreated and treated CD4+ T cells by FACS at 6 days post-infection . These cells were then stimulated with αCD3/αCD28 activating beads in the presence or absence of the integrase inhibitor , raltegravir , to distinguish between the reactivation of any pre-integration latent virus and post-integration latent provirus that might be present ( S4 Fig ) . As a control , we isolated ( via FACS ) latently infected cells ( GFP-/mCherry+ ) from infected cells that were pretreated with αCD3/αCD28 activating beads and subjected them to the same treatments as the uninfected cells . We analyzed reactivation of the latent virus by flow cytometry 48 h after stimulation . All cell populations contained some reactivatable pre-integration latent virus ( Fig 1F ) ; the highest levels were observed in CCL19- and IL-7-treated populations and in the untreated population infected with Vpx-containing HIV Duo-Fluo I . Untreated resting CD4+ T cells infected with the HIV Duo-Fluo I virus alone showed the lowest levels of reactivatable pre-integration latency , followed closely by activated CD4+ T cells . By analyzing the reactivatable post-integration latent provirus , we found that the uninfected cell population isolated from activated CD4+ T cells contains very little reactivatable provirus ( 0 . 26% , Fig 1F ) compared to the initial latent population identified in activated CD4+ T cells after infection ( 1 . 01% , Fig 1D ) . These findings suggest that HIV Duo-Fluo I can efficiently identify latently infected cells within activated CD4+ T cells . Conversely , uninfected cells isolated from untreated resting CD4+ T cells contained 0 . 40% reactivatable provirus , but only 0 . 14% latently infected cells were identified in this cell population after the initial infection . This finding suggests that HIV Duo-Fluo I identifies only a fraction of latently infected cells within resting CD4+ T cells . Similarly , IL-7–treated resting CD4+ T cells and untreated resting CD4+ T cells infected with the Vpx-containing HIV Duo-Fluo I both contained over 1% reactivatable provirus within their isolated uninfected cell populations , which was more than twice the size of the latent cell populations identified in these respective populations after the initial infection ( Fig 1D ) . This suggests that IL-7 treatment and SAMHD1 knockdown lead to silent infection events in resting CD4+ T cells , and , thus , an underestimation of the latently infected cell population . Interestingly , CCL19 treatment of resting CD4+ T cells produced 0 . 2% reactivatable provirus from the isolated uninfected cell population , while the initial latent cell population after infection was 0 . 47% . Overall , these data suggest that HIV Duo-Fluo I can be used to identify latently infected cells within activated CD4+ T cells , but may underestimate the percentage of latently infected cells within resting CD4+ T cells . Because HIV replicates most efficiently in activated CD4+ T cells [23] , and the largest in vivo latent reservoir is within resting memory CD4+ T cells , we next investigated whether HIV latency is preferentially established in CD4+ T cells that become infected as they transition from an activated to a resting state . To do this , we isolated total CD4+ T cells from peripheral blood of uninfected donors and stimulated the cells with αCD3/αCD28 activating beads in the presence of IL-2 for 3 days ( Fig 2A ) . We then removed the αCD3/αCD28 activating beads and allowed the cells to return to a resting state in the presence of IL-2 for 20 days . We infected the CD4+ T cells with HIV Duo-Fluo I at peak activation ( day 4 ) and every 5 days thereafter as they transitioned back to resting . As indicated by expression of the activation markers CD69 and CD25 , the cells transitioned from active to resting over the 20 days and remained >60% viable ( Fig 2B ) . Maximal activation occurred at day 4 with 79% of cells expressing both CD69 and CD25 ( Fig 2C ) . By day 9 , however , 31% of cells no longer expressed CD69 or CD25 , and by day 24 , 92% of cells no longer expressed either activation marker . Despite most cells losing CD69 expression by day 14 ( <1% CD69+ ) , a small fraction of cells continued to express CD25 throughout the experiment—with 8% of cells still CD25+ at day 24—suggesting that while most CD4+ T cells had returned to a resting state by day 24 , a small population was still transitioning back to a resting state . Others have observed similar expression profiles while trying to return activated CD4+ T cells to a resting state [52 , 53] . We analyzed productive infection and latent infection by flow cytometry at 3 days postinfection for each time point ( Fig 2D ) , and we determined the average of each cell population from three donors ( Fig 2E ) . Infection at day 4 , when the CD4+ T cells were maximally activated , produced the highest levels of both productive and latent infection . As the cells returned to a resting state , the levels of both productive infection and latent infection decreased , suggesting that HIV most effectively infects CD4+ T cells when they are at their highest activation state; as CD4+ T cells stop expressing the activation markers CD69 and CD25 , they become less permissive . However , the ratio of latent infection to productive infection steadily increased from 0 . 010 at day 4 when the cells where most active to 0 . 025 at day 24 when the cells exhibited a more resting phenotype ( Fig 2F ) . Therefore , while activated CD4+ T cells support the most robust infection , latent infection is more likely to occur relative to productive infection in cells that are resting or are transitioning back to a resting state . To explore another possible way latency is established , we next investigated whether productively infected activated primary CD4+ T cells can return to a resting state and contribute to the latent reservoir . We isolated CD4+ T cells from uninfected donor blood and activated them with αCD3/αCD28 activating beads in the presence of IL-2 for 3 days . At this point , we spinoculated the cells with HIV Duo-Fluo I for 2 h at 37°C ( Fig 3A ) . After infection , cells were kept in an activated state by returning them to culture in the presence of activating beads and IL-2 . Four days postinfection , CD4+ T cells were sorted to isolate three distinct populations: uninfected ( GFP-/mCherry- ) , productively infected ( GFP+/mCherry- & GFP+/mCherry+ ) , and latently infected ( GFP-/mCherry+ ) cells . After sorting , a small fraction of each population was used to measure HIV integration via Alu-gag PCR ( S5 Fig ) , while the majority of each population was cultured with IL-2 and allowed to return to a resting state over an 11-day period . The activation state of each cell population was monitored by the expression of the activation markers CD69 and CD25 ( Fig 3B ) , as well as changes in cell size ( S6 Fig ) . The activation markers CD69 and CD25 were maximally expressed at day 4 , when the cells were infected with HIV Duo-Fluo I ( Fig 3B ) . After sorting , each cell population began to lose both CD69 and CD25 expression , and the cell size of each population began to shrink ( S6 Fig ) . However , none of the distinct cell populations fully returned to resting during the 11-day period . The uninfected cell population contained 21 . 6% of CD69 and CD25 double-negative cells at 11 days post-activation compared to 16 . 6% for the latently infected population and just 9 . 03% for the productively infected population . This indicates that the uninfected population is returning to a resting state more quickly than either of the infected cell populations . Additionally , the productively infected cell population contained 12 . 2% CD69 and CD25 double positive cells at 11 days post-activation compared to 5 . 58% for the latently infected cell population and just 2 . 44% for the uninfected cell population , indicating that the productively infected CD4+ T cells are returning to a resting state at a slower rate than either latently infected or uninfected CD4+ T cells . As such , latently infected cells are less activated than productively infected cells . While activated cells were returning to a resting state over the 11-day period , they experienced changes in their infection profiles ( productive or latent infection ) ( Fig 3C ) . A small percentage of the uninfected cell population became infected , which most likely reflects pre-integration latency [18] . More interestingly , 98 . 8% of the productively infected cell population continued to express GFP throughout the 11 days , with 1 . 18% of the cells no longer expressing GFP . Based on these data , only a small fraction of productively infected cells have returned to a resting state over the course of the 11 days , and potentially contribute to the latent reservoir . Over the course of the 11 days , the latently infected cell population began to display two distinct phenotypes . First , over 6% of the latently infected cell population spontaneously reactivated as exhibited by their GFP expression ( Fig 3C ) . Second , over 60% of the latently infected cells lost expression of mCherry , in addition to not expressing GFP . By analyzing the activation marker expression profile for these distinct populations ( S7 Fig ) , we found that those cells that lost mCherry expression had higher levels of CD69/CD25 double-negative cells ( 24 . 9% ) than the latently infected cells that continued to express mCherry ( 12 . 8% ) and those that spontaneously reactivated ( 3 . 23% ) . Taken together , these data demonstrate that the majority of activated CD4+ T cells that become latently infected revert back to a resting state , while also silencing the EF1-α-driven mCherry expression . After allowing the sorted cell populations to return to a resting state for 11 days , we stimulated each population with αCD3/αCD28 activating beads to reactivate any latent provirus ( Fig 3D ) . In the uninfected cell population , stimulation produced a small amount of reactivatable provirus which , again , is likely due to pre-integration latency [18] . Interestingly , stimulation of the productively infected cell population did not lead to reactivation of the ~1% of cells that no longer expressed GFP . Lastly , stimulation of the latently infected cell population produced only a small amount of reactivatable provirus ( ~2% ) . However , we observed a large shift from latently infected cells that no longer express mCherry to latently infected cells that do express mCherry , as evidenced by a 20% increase in mCherry+ cells after αCD3/αCD28 stimulation ( Fig 3D ) . These results indicate that latently infected cells that lose mCherry expression over time are not dying but instead returning to a resting state . Lastly , based on evidence that resting CD4+ T cells within lymphoid tissue can support HIV replication [34–39] , we wanted to investigate how latency is established within ex vivo lymphoid tissue using our HIV Duo-Fluo I virus . We isolated CD4+ T cells from tonsillar and splenic tissues , as well as from peripheral blood from uninfected donors . Because lymphoid organs contain over 98% of the body’s CD4+ T cells , and are the primary sites of HIV replication , we also isolated total lymphoid cells from tonsillar and splenic tissues from uninfected donors in the form of human lymphoid aggregated cultures ( HLACs ) , which closely mimic the conditions encountered by HIV , in vivo [36] . In addition , we isolated total peripheral blood mononuclear cells ( PBMCs ) from uninfected donors . CD4+ T cells , PBMCs , and HLACs were either left untreated or stimulated with αCD3/αCD28 activating beads in the presence of IL-2 for 3 days and then spinoculated with HIV Duo-Fluo I for 2 h at 37°C . We assessed expression of the activation markers CD69 and CD25 before and after stimulation ( Fig 4A ) . PBMCs and CD4+ T cells isolated from peripheral blood expressed very little CD69 or CD25 and were thus considered resting cells . However , CD4+ T cells isolated from tonsillar and splenic tissues , as well as HLACs from these tissues , highly expressed the early activation marker CD69 ( 38% and 42% , respectively ) but expressed low levels of the intermediate activation marker CD25 . Thus , CD4+ T cells isolated from lymphoid tissue are not resting cells , but they are also not fully activated . After stimulation with αCD3/αCD28 activating beads in the presence of IL-2 for 3 days , CD4+ T cells from all three tissues expressed high levels of both activation markers , reflecting classic T cell activation . However , upon stimulation , CD4+ T cells from peripheral blood achieved higher activation levels than CD4+ T cells isolated from either lymphoid tissue . Lastly , expression of CD69 and CD25 among αCD3/αCD28-stimulated PBMCs and HLACs was consistently lower than in purified CD4+ T cells , and may reflect the size of the CD4+ T cell population within each culture . Levels of productive and latent infection were analyzed by flow cytometry 72 h post-infection ( Fig 4B and 4C ) . Untreated CD4+ T cells from peripheral blood , which expressed no activation markers , produced very little productive infection ( 0 . 47% ) . Despite expressing moderately high levels of CD69 , untreated CD4+ T cells isolated from tonsillar tissue did not give rise to significantly higher levels of productive infection ( 0 . 68% ) compared to CD4+ T cells from peripheral blood . However , untreated CD4+ T cells from splenic tissue—which expressed CD69 at levels comparable to those of CD4+ T cells isolated from tonsillar tissue—did show an increase in productive infection ( 1 . 8% ) as compared to untreated CD4+ T cells from peripheral blood . Levels of latent infection in untreated CD4+ T cells isolated from both lymphoid tissues were at least 2-fold greater than those observed in untreated CD4+ T cells from peripheral blood ( Fig 4C ) , suggesting that CD4+ T cells within lymphoid tissue are more likely to become latently infected . Overall , untreated PBMCs and untreated HLACs from both lymphoid tissues displayed lower levels of productive infection than untreated CD4+ T cells isolated from each tissue ( Fig 4C ) . Infection of untreated PBMCs resulted in a 3 . 5-fold decrease in productive infection as compared to untreated CD4+ T cells from the blood , while infection of untreated HLACs from tonsil resulted in a 4-fold decrease in productive infection as compared to untreated CD4+ T cells from the same tissue . Untreated HLACs from the spleen resulted in an 11-fold decrease in productive infection as compared to untreated CD4+ T cells from the same tissue . In contrast , the levels of latent infection did not change between untreated total lymphoid cells and untreated CD4+ T cells from each tissue ( Fig 4C ) . In addition , despite differences in activation levels , infection of untreated HLACs from both lymphoid tissues did not result in an increase of either productive or latent infection as compared to untreated PBMCs . When analyzing the ratio of latent infection to productive infection , we found that HIV Duo-Fluo I infection of all untreated cells from the three different tissues was at least 5-fold more likely to result in latent infection than their αCD3/αCD28-stimulated counterparts ( Fig 4D ) . Thus , activated cells exhibit a higher propensity for productive infection , while resting cells exhibit a higher propensity for latent infection . Additionally , infecting untreated total lymphoid cell populations results in more latent infection than when infecting purified untreated CD4+ T cells from the same tissue , suggesting that co-culture of CD4+ T cells with other lymphoid cells promotes latent infection .
The role that T cell activation plays in establishing HIV latency within CD4+ T cells is still not fully understood . HIV replication is clearly most efficient in activated CD4+ T cells [19–22] , and the largest in vivo latent reservoir is within memory CD4+ T cells [11 , 12] . This evidence suggests that HIV latency is established in one of two ways: 1 ) Activated CD4+ T cells become productively infected but survive viral cytopathic effects and evade elimination by the immune system long enough for the cell to transition to a resting memory state; or 2 ) CD4+ T cells that are transitioning from an activated to a resting memory state are infected by HIV while the cellular environment can still support integration of viral cDNA but cannot support proviral transcription . However , studies have shown that both naive and memory CD4+ T cells contain integrated viral DNA [34] , and that direct infection of resting CD4+ T cells in lymphoid tissue results in productive infection [40] . These findings suggest that HIV latency can also be established in another way: direct infection of resting CD4+ T cells . In this study , we show that all three scenarios can produce latent HIV infection . We further show that HIV latency can be established in activated CD4+ T cells without them first returning to a resting state . Additionally , infecting activated CD4+ T cells is more likely to result in productive infection , while infecting resting CD4+ T cells is more likely to result in latent infection . Finally , HIV latency is more likely to occur in resting lymphoid cell aggregates than in resting CD4+ T cells cultured alone . Using primary CD4+ T cells isolated from the blood of uninfected donors , we demonstrate that infecting resting and activated CD4+ T cells with our HIV Duo-Fluo I virus causes both productive and latent infection in the two populations . In activated CD4+ T cells , HIV latency is established within the first few days of infection and does not require the cell to return to a resting state . We showed this previously [18] , as did another group that developed a similar dual-reporter virus [54] . That construct uses a different LTR-independent promoter ( CMV ) than our EF1α promoter , and it places the LTR-driven eGFP cassette in the Gag region , while ours replaces the Nef open reading frame . Despite these differences , both dual-reporter viruses can detect latent infection events in activated CD4+ T cells early after the initial infection , and these latently infected cells can be reactivated by different stimuli . Additionally , we sorted these latently infected cells and showed that they still express significant amounts of both CD69 and CD25 activation markers; the cells only stop expressing these markers as they are allowed to return to a resting state . As they return to a resting state , latently infected CD4+ T cells also stop expressing the EF1α-driven mCherry marker , suggesting that as these cells return to resting , both promoters become silenced , perhaps by packaging into heterochromatin [55] . This means that HIV latency may be established after activated CD4+ T cells are initially infected , and it is these cells , potentially , that survive and return to a resting memory state , thus significantly contributing to the latent pool . How HIV latency is established in activated CD4+ T cells immediately after infection is still unknown , but it may arise from stochastic viral gene expression [56–59] . Our studies also suggest that activated CD4+ T cells that become productively infected can contribute to the latent pool as they return to a resting state . In our studies , these cells did not return to a completely resting state because so many of the cells died ( S6 Fig ) , a likely consequence of viral cytopathic effects . However , the data clearly indicate that a small population of productively infected cells starts to return to a resting state and as they do , they lose expression of the LTR-driven GFP marker . However , when these cells were then stimulated with αCD3/αCD28 , they failed to express GFP , suggesting that they could not be reactivated by CD3/CD28 stimulation , though it is possible that other reactivating agents could work . Finally , it is important to note that these productively infected CD4+ T cells that did eventually shut down LTR-driven GFP expression , did so in a culture dish . It remains to be seen , in vivo , if productively infected CD4+ T cells can survive long enough to return to a resting state and contribute to the latent pool , or if activated CD4+ T cells that become latently infected immediately after infection are the major contributors . Finally , infecting activated CD4+ T cells produces more productively infected cells than latently infected cells , while infecting resting CD4+ T cells produces more latently infected cells . These results reflect that HIV replicates more efficiently in activated CD4+ T cells , but they also show that resting CD4+ T cells can support HIV infection , at least up to the point of viral integration . In resting primary CD4+ T cells , we show that both productive and latent HIV infection can be achieved , though at levels much lower than those seen in activated CD4+ T cells . The infection kinetics in resting CD4+ T cells seem to be slower than in activated cells , since peak infection was not reached until 6 days after infection , while activated cells reached peak infection 4 days after infection ( S1 Fig ) . These results agree with other’s findings [22 , 60] . Also in agreement with previous findings , resting CD4+ T cells were made more permissive to HIV infection when exposed to the chemokine CCL19 , which increases the ability of resting CD4+ T cells to support latent infection [47] . However , our data demonstrate that CCL19 also increases permissibility to productive infection , although its overall effect on resting CD4+ T cells increases latent infection . Interestingly , the cytokine , IL-7 , which increases permissibility of resting CD4+ T cells to productive HIV infection , also increased both productive and latent infection in resting CD4+ T cells in our study . Lastly , infecting untreated resting CD4+ T cells with a Vpx-containing virus significantly increased productive infection but only modestly increased latent infection . Interestingly , infecting resting CD4+ T cells with our HIV Duo-Fluo I virus produced a significant amount of silent infection events , in which expression of both fluorescent proteins was silenced , camouflaging latently infected cells within our uninfected population . In fact , the isolated uninfected population of resting CD4+ T cells contained more silently infected cells than the number of latently infected cells that were identified via the mCherry fluorescent marker after the initial infection . This was true for all untreated and treated resting CD4+ T cells , except CCL19-treated cells , and was highest in IL-7-treated cells and untreated resting CD4+ T cells infected with a Vpx-containing virus . The reasons for this are unclear . Within resting CD4+ T cells , viral integration occurs in regions of the host genome that are unfavorable for viral gene expression [61] , and studies also suggest that latently infected cells are more likely to contain provirus in or near heterochromatin [62 , 63] . Integration into such regions would be unfavorable not only for LTR-driven gene expression but also for EF1α-driven gene expression . In the presence of Vpx , SAMHD1 cannot inhibit HIV reverse transcription , allowing the virus to bypass one of the major obstacles to replication in resting CD4+ T cells . Therefore , integration of the viral cDNA may occur more readily in these unfavorable heterochromatic regions . Treating cells with IL-7 , which signals through the JAK/STAT pathway [64] , may produce a similar situation . Lastly , previous studies have reported that resting CD4+ T cells can only be infected by HIV in the context of total lymphoid cell aggregates [41] . However , our results show that infecting untreated resting CD4+ T cells ( alone ) and untreated resting total lymphoid cells from peripheral blood and lymphoid tissue all produced productive and latent populations . Although , we did find that latent infection is more likely to occur in total resting lymphoid cell aggregates than in resting CD4+ T cells alone . The reasons for this are still unclear , but recent studies have shown that co-culture of resting CD4+ T cells with myeloid dendritic cells [65] , or co-culture of resting CD4+ T cells with endothelial cells [66] , enhances HIV latency , further proving that the lymphoid environment plays an important role in how HIV latency is established within resting CD4+ T cells . Overall , our studies show that HIV infection can occur in both resting and activated CD4+ T cells , such that infection of resting cells more often results in latent infection and infection of activated cells more often results in productive infection . Based on our data , we now have a better understanding of the contribution that each infected cell type makes to the latent reservoir . Our study underscores why we must consider both resting and activated CD4+ T cells when investigating how HIV latency occurs .
Pseudotyped HIV Duo-Fluo I viral stocks were generated by co-transfecting ( using the standard calcium phosphate transfection method ) HEK293T cells with a plasmid encoding HIV Duo-Fluo I and a plasmid encoding HIV-1 dual-tropic envelope ( pSVIII-92HT593 . 1 ) . We generated a Vpx-containing HIV Duo-Fluo I pseudotyped virus by co-transfecting HEK293T cells with the HIV Duo-Fluo I plasmid , the pSVIII-92HT593 . 1 plasmid , and a plasmid encoding a Vpr-Vpx fusion protein ( pSIV3+ , generously donated by Warner Greene ) . Supernatants were collected after 72 h and filtered through a 0 . 45 μM membrane to clear cell debris , and were then concentrated by ultracentrifugation ( 76 , 755 x g ) for 2 h at 4°C . Concentrated virions were resuspended in complete media and stored at -80°C . Virus concentration was estimated by p24 titration ( HIV-1 alliance p24 ELISA kit , Perkin-Elmer ) . Primary CD4+ T cells and peripheral blood mononuclear cells ( PBMCs ) were purified from healthy donor blood ( Blood Centers of the Pacific , San Francisco , CA , USA and Stanford Blood Center ) . CD4+ T cells were isolated by negative selection using the RosetteSep Human CD4+ T Cell Enrichment Cocktail ( StemCell Technologies ) . PBMCs were purified by Histopaque-1077 density gradient . Purified resting CD4+ T cells and PBMCs from peripheral blood were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum ( FBS ) , L-glutamine ( 2 mM ) , penicillin ( 50 U/ml ) , and streptomycin ( 50 mg/ml ) . Human lymphoid aggregate cultures ( HLACs ) were purified using tonsillar or splenic tissue from uninfected donors ( Cooperative Human Tissue Network ) as previously described [67] . CD4+ T cells were isolated from HLACs by negative selection using the EasySep Human CD4+ T Cell Enrichment Kit ( StemCell Technologies ) . HLACs and CD4+ T cells isolated from splenic and tonsillar tissues were cultured in RPMI 1640 supplemented with 20% heat-inactivated FBS , 100 mg/ml gentamicin , 200 mg/ml ampicillin , 1 mM sodium pyruvate , 1% nonessential amino acids ( Mediatech , Manassas , VA , USA ) , 2 mM L-glutamine , and 1% fungizone ( Invitrogen , Indianapolis , IN , USA ) Purified resting CD4+ T cells were either left untreated or treated for 3 days with 20 ng/ml IL-7 ( R&D Systems ) or 100 μM CCL19 ( R&D Systems ) . Purified CD4+ T cells isolated from peripheral blood and tonsillar and splenic tissues , as well as PBMCs and HLACs , were stimulated with αCD3/αCD28 activating beads ( Life Technologies ) at a concentration of 1 bead/cell in the presence of 30 U/ml IL-2 ( PeproTech ) for 3 days . All cells were spinoculated with either HIV Duo-Fluo I alone or Vpx-containing HIV Duo-Fluo I at a concentration of 100 ng of p24 per 1 × 106 cells for 2 h at 1 , 200 × g at 37°C . After spinoculation , all cells were returned to culture in the presence of 30 U/ml IL-2 , except for CD4+ T cells pre-stimulated with αCD3/αCD28 activating beads , which were placed back in culture with the αCD3/αCD28 activating beads and 30 U/ml IL-2 . Uninfected cells were stained in fluorescence-activated cell sorting ( FACS ) buffer ( phosphate buffered saline supplemented with 2 mM EDTA and 2% FBS ) with αCD69-PE and αCD25-APC ( eBioscience ) and fixed in 1% paraformaldehyde . Infected cells were stained in FACS buffer with αCD69-V450 and αCD25-APC/Cy7 ( BD Biosciences ) and fixed in 1% paraformaldehyde . Data were collected on a FACS Caliber and a FACS LSRII ( BD Biosciences ) , and analyses were performed with FlowJo software ( TreeStar ) . Untreated and treated CD4+ T cells from Figs 1F and S4 were sorted with a FACS AriaII ( BD Biosciences ) based on their GFP and mCherry fluorescence at 6 days post-infection , and they were placed back in culture with or without 30 μM Raltegravir ( National AIDS Reagent Program ) . CD4+ T cells stimulated with αCD3/αCD28 activating beads in the presence of 30 U/ml IL-2 from Fig 3 were sorted based on their GFP and mCherry fluorescence at 4 days post-infection . Untreated resting primary CD4+ T cells infected with either HIV-Duo-Fluo I alone or Vpx-containing HIV Duo-Fluo I were lysed 6 days post-infection in radioimmunoprecipitation assay buffer ( 150 mm NaCl , 1% Nonidet P-40 ( vol/vol ) , 0 . 5% AB-deoxycholate ( vol/vol ) , 0 . 1% sodium dodecyl sulfate ( SDS ) ( vol/vol ) , 50 mm Tris-HCl ( pH 8 ) , 1 mm DTT , and EDTA-free Protease Inhibitor ( Calbiochem ) . Cell lysates were used for SDS-polyacrylamide gel electrophoresis ( SDS-PAGE ) immunoblotting analysis . The primary antibodies used were rabbit polyclonal anti-SAMHD1 ( Sigma-Aldrich , Cat# SAB2102077 ) and monoclonal anti-β-actin ( A5316 , Sigma-Aldrich ) . DNA was prepared after cell sorting of uninfected , productively infected and latently infected cell populations using the DNeasy Kit ( Qiagen ) . Real-time PCR was used to detect total HIV DNA , β-globin , and integrated HIV DNA as previously described [68] . | The study of HIV latency has been hindered because there are few latently infected cells in vivo , and we cannot distinguish latently infected cells from uninfected cells prior to reactivation of the latent provirus . In general , HIV latency is quantitatively studied by reactivating latently infected cells after latency has been established . However , this practice limits the investigation of how latency is established and how latent provirus can be reactivated . Our recently developed dual reporter virus , HIV Duo-Fluo I , can identify latently infected cells early after infection . In this study , we use HIV Duo-Fluo I to investigate how T cell activation affects the outcome of HIV infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | HIV Latency Is Established Directly and Early in Both Resting and Activated Primary CD4 T Cells |
In recent years , there has been an increasing interest in immunomodulatory therapy as a means to treat various conditions , including infectious diseases . For instance , Toll-like receptor ( TLR ) agonists have been evaluated for treatment of genital herpes . However , although the TLR7 agonist imiquimod was shown to have antiviral activity in individual patients , no significant effects were observed in clinical trials , and the compound also exhibited significant side effects , including local inflammation . Cytosolic DNA is detected by the enzyme cyclic GMP-AMP ( 2’3’-cGAMP ) synthase ( cGAS ) to stimulate antiviral pathways , mainly through induction of type I interferon ( IFN ) s . cGAS is activated upon DNA binding to produce the cyclic dinucleotide ( CDN ) 2’3’-cGAMP , which in turn binds and activates the adaptor protein Stimulator of interferon genes ( STING ) , thus triggering type I IFN expression . In contrast to TLRs , STING is expressed broadly , including in epithelial cells . Here we report that natural and non-natural STING agonists strongly induce type I IFNs in human cells and in mice in vivo , without stimulating significant inflammatory gene expression . Systemic treatment with 2’3’-cGAMP reduced genital herpes simplex virus ( HSV ) 2 replication and improved the clinical outcome of infection . More importantly , local application of CDNs at the genital epithelial surface gave rise to local IFN activity , but only limited systemic responses , and this treatment conferred total protection against disease in both immunocompetent and immunocompromised mice . In direct comparison between CDNs and TLR agonists , only CDNs acted directly on epithelial cells , hence allowing a more rapid and IFN-focused immune response in the vaginal epithelium . Thus , specific activation of the STING pathway in the vagina evokes induction of the IFN system but limited inflammatory responses to allow control of HSV2 infections in vivo .
Virus infections may cause acute and chronical diseases , and there is therefore a need for development of efficient treatments . Significant improvements have been made in the development of therapeutics that target specific viral molecules , such as the HIV reverse transcriptase and the hepatitis C virus NS5A protein [1 , 2] . Despite this , satisfactory treatments are not available for many virus infections , and there is also a need for treatments acting in a broader manner , and which are less sensitive to viral development of resistance . Herpes simplex virus ( HSV ) -2 is the leading cause of genital ulcers [3 , 4] with an estimated 500 million infected people globally [5] . Although intensively pursued , all attempts at making an effective anti-HSV2 vaccine have failed [4] . The current standard treatment is acyclovir or derivatives , which target the viral thymidine kinase [6] , and although generally efficient if treatment is initiated early , there are reports of development of resistance in immunosuppressed patients receiving long-term treatment [6] . The need for new and better anti-HSV2 treatments is underpinned by several factors , including the ability of this virus to cause neonatal herpes [3] , the role of HSV2 in amplifying HIV-transmission , which has been reported to account for up to half of all new transmissions in areas of high HSV2 seroprevalence [7 , 8] , and the recently reported association to increased rates of autism-spectrum disorders [9] . In addition to directly targeting the virus , antiviral treatments can stimulate host immune responses . Previously tested experimental immune modulatory therapies for virus infections have mainly focused on agonists for Toll-like receptors ( TLRs ) . The TLR7-agonist imiquimod and the mixed TLR7/8-agonist resiquimod induce interferon ( IFN ) α , and imiquimod is the first approved topically active TLR7 agonist used to treat human papilloma virus ( HPV ) , but has failed to show significant efficacy against HSV2 infection [10–12] , although cases have been reported with benefit of imiquimod 5% cream for treatment of herpes labialis and genital herpes [13 , 14] . Furthermore , pretreatment of mice with oligodeoxynucleotide TLR9 agonists has been shown to lower the viral load in the brain in an HSV-1 encephalitis model [15] . Common to the aforementioned TLRs is their role in innate recognition of foreign nucleic acids in the endosomal compartment [16] . However , TLRs are mainly expressed in leukocytes , and to a much lesser extent in epithelial cells [17] . Another protein involved in nucleotide sensing is the cyclic GMP-AMP synthase ( cGAS ) , which is localized in the cytoplasm , and hence is a sensor of mislocalized endogenous or exogenous DNA [18] . We and others have shown that cGAS plays an intrinsic role in mounting protective immune responses against DNA viruses , including HSV-1 [19–21] . When cGAS senses dsDNA in the cytosol , it produces the second messenger 2’3’-cyclic GMP-AMP ( 2’3’-cGAMP ) which activates the adaptor protein Stimulator of IFN-genes ( STING ) on the ER [18 , 22 , 23] . STING dimerizes and traffics to the ER-Golgi intermediate compartment ( ERGIC ) where it recruits the TANK-binding kinase 1 ( TBK1 ) , which in turn phosphorylates IFN-regulatory factor 3 ( IRF3 ) that translocates as a dimer to the nucleus where it initiates transcription of type I IFN genes [24] . Type I IFNs are secreted cytokines , which work in auto- and paracrine manners via the IFNα receptor ( IFNAR ) to upregulate IFN-stimulated genes ( ISGs ) that target specific steps in the viral life cycle to inhibit replication [24] . In addition to the action of 2’3’-cGAMP inside the DNA-stimulated cell , this CDN is also able to exert effects inside other cells through at least two distinct mechanisms , either juxtacrinely by diffusing through gap junctions [25] or endocrinely by being packaged into newly forming virions [26 , 27] . Serum contains the ectonucleotide pyrophosphatase/phosphodiesterase ( ENPP ) 1 , which metabolizes 2’3’-cGAMP [28] , thus hindering extensive circulation and endocrine stimulation by cell-free 2’3’-cGAMP . However , free 2’3’-cGAMP ( and other cyclic dinucleotides ( CDNs ) ) has the capacity to pass the cell membrane and activate STING [29] . Therefore , CDNs could potentially stimulate immune responses locally and over longer distances . The use of STING agonists for treatment of disease has been tested in a series of models . The small molecule DMXAA ( vadimezan ) was shown to have anti-tumor effects in a mouse model [30] before its mechanism of action as a STING ligand was discovered [31] , and it was tested in a clinical phase III efficacy trial of treatment of advanced non-small-cell lung cancer [32] . It was subsequently shown that DMXAA does not stimulate human STING , as opposed to murine STING , due to a single amino acid difference between human and murine STING at a residue mediating interaction between murine STING and DMXAA [31 , 33] . Another interesting difference between STING in humans and mice is that human STING is activated more potently by 2’3’-linked CDNs ( produced in metazoan cells ) than 3’3’-linked CDNs ( produced in non-metazoan cells ) , while murine STING is activated to comparable degrees by the two types of CDNs [29] . In the past few years , STING ligands have been investigated for anti-tumor activity in a variety of mouse models [34 , 35] , anti-inflammatory effects in an experimental autoimmune encephalitis mouse model [36] , as well as for potential as adjuvants in vaccines [37–41] . However , whether STING agonists could have antiviral therapeutic effects has remained under-explored . In this study , we investigate the antiviral action of natural and non-natural CDNs in a murine model of genital HSV2 infection . We demonstrate that different CDNs stimulate the IFN pathway to varying degrees , and that intraperitoneal ( i . p . ) and local delivery stimulates rapid IFN response with associated antiviral function . Lastly , we show that local application of a CDN to the vaginal mucosa confers full protection against genital HSV2 infection in both wild type and immunodeficient cGas-/- mice . The effects of CDN treatment are superior to TLR7 and 9 agonists , based on high antiviral activity , IFN-biased response with low TNF expression , and targeted stimulation of epithelial cells .
To test whether STING agonists can induce antiviral response against HSV2 , we treated human keratinocytes ( HaCaT ) with five different STING agonists . Keratinocytes are specialized epithelial cells and are the primary cells involved in clinical lesions caused by HSV2 . The STING agonists tested were DMXAA ( 5 , 6-dimethylxanthenone-4-acetic acid , vadimezan ) , 2’3’-cGAMP; 3’3’-cyclic di-adenosine monophosphate ( 3’3’-c-di-AMP ) , and two cyclic dinucleotide analogs , namely the Rp , Sp isomer of the bisphosphorothioate analog of 2’3’-cGAMP ( 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) ) , and 3’3’-cyclic adenosine monophosphate- inosine monophosphate ( 3’3’-cAIMP ) . As expected , DMXAA did not induce expression of the ISGs viperin and ISG15 or phosphorylation of STAT1 ( Fig 1A ) , as measured 6 h after stimulation . For the CDNs , we observed phosphorylation of STAT1 , upregulation of viperin , and lower levels of STING , indicating its activation and subsequent degradation ( Fig 1A ) . Despite the observed phosphorylation of STAT1 and induction of viperin , we did not find detectable levels of type I IFN in the supernatants of HaCaT cells stimulated with 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) ( Fig 1B ) . In sharp contrast to HaCaT cells , the human monocyte-like cell line THP-1 responded to 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) stimulation with a very strong induction of type I IFN production , but limited induction of ISGs , and was also less responsive to IFNα or -β treatment ( Fig 1B , 1D–1F ) . Since HaCaT cells were highly responsive to IFNα or -β treatment ( Fig 1E and 1F ) , we cannot exclude that these cells did induce type I IFN levels below the detection limit , which stimulated significant IFN signaling and gene expression . Of note , for the data shown in Fig 1B , 1D and 1H , cells were not permeabilized before treatment with 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) , thus suggesting direct cellular entry without additional delivery systems . Collectively , the data suggest that monocytes produce much higher levels of type I IFN in response to STING agonists than keratinocytes , which however are more responsive to IFN stimulation than monocytes . To investigate whether the CDN stimulation could protect against an HSV2 infection , we treated HaCaT cells with 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) either 24 h or 30 min before infection . In cells treated with CDN 24 h prior to infection , we observed strongly reduced levels of the major capsid protein , viral protein 5 ( VP5 ) ( Fig 1C ) , indicating a protection of the cells from infection . The reduced accumulation of viral proteins was also seen in cells treated with high dose of 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) 30 min before infection . Since , THP-1 cells produced high levels of type I IFN upon CDN stimulation , we wanted to test whether this could contribute to antiviral activity in a setting with crosstalk between different cell types . Therefore , supernatants from 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) -treated THP-1 cells were transferred to WT and IFNAR2-/- HaCaT cells , which were subsequently infected with HSV2 ( Fig 1G ) . Interestingly , treatment with the supernatants from stimulated THP1 cells led to a strong reduction of HSV2 replication in WT , but not IFNAR2-/- HaCaT cells ( Fig 1H and 1I ) . Collectively , these data suggest that CDNs differentially stimulate keratinocytes and monocytes , and that crosstalk between the cell-types enables the full antiviral response , which is dependent on type I IFNs . To determine the effect and potency of STING agonists in animals , we injected the compounds i . p . into wild type ( WT ) mice in equimolar doses ( Fig 2 ) . At 6 h post treatment , we observed phosphorylation of STAT1 and upregulation of viperin and ISG15 ( Fig 2A ) . This was most pronounced in the mice treated with the non-endogenous CDNs , and to a much lesser extent with 2’3’-cGAMP . Very strong responses were observed in serum , spleen , and the vagina ( epithelial surface ) , whereas systemic treatment with the compounds gave rise to very low responses in the brain ( Fig 2A–2C ) . The lack of a response to DMXAA is explained by the low dose used in this setting , as we found that DMXAA potently stimulated an IFN response when the dose was increased ( S1A Fig ) . Further analysis of the induction of expression of IFNs and ISGs by systemically delivered STING agonists confirmed that the non-endogenous CDNs potently stimulated the IFN pathway in the vagina and the spleen , and that DMXAA and 2’3’cGAMP were much less potent in vivo ( Fig 2C ) . We also found that CDNs stimulated modest induction of the nuclear factor ( NF ) κB-induced genes A20 , Il6 , and Tnfa ( S1B Fig ) , thus suggesting that direct activation of the STING pathway in mice does not lead to strong activation of the NFκB pathway . Furthermore , mice deficient for STING did not respond to 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) treatment , underlining that STING is the target for the molecules tested ( S1C Fig ) . However , since an NFκB responsive reporter gene was readily activated by STING agonists in THP1 cells ( S1D Fig ) , we are reluctant to conclude that NFκB is not involved in the gene induction program induced upon STING activation . To ensure that the small , but significant , 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) -induced IFN response we observed in the CNS was not derived from blood left over in the vasculature following harvest of brain tissue , we injected WT mice i . p . with 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) and performed a whole-animal transcardial perfusion of PBS to remove any residual blood from the peripheral vasculature before taking out the brain of the animals . Following this procedure , we still found a significant upregulation of the expression of Ifnb and Mx1 in the brain ( Fig 2D ) . This indicates that the IFN response in the brain tissue originates from resident cells , suggesting that among the CDNs tested , 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) may cross the blood-brain-barrier and exert an effect directly in the CNS . Taken together , these data suggest that systemic administration of non-endogenously produced CDNs yields an IFN response in several different tissues , including lymphoid organs , epithelial surfaces , and the CNS . To investigate how the STING agonists affected the IFN response in the tissue , we chose to look more closely at the vagina , which is one of the main portals of entry for viral infections . Using immunohistochemical staining of vaginal tissues , we found that systemic delivery of 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) induced strong expression of the ISG viperin by cells in the stroma ( Fig 3 ) . Furthermore , epithelial cells located close to the basal membrane were positive for viperin , while epithelial cells located to the luminal side had less expression . Vaginal tissue from mice infected locally with HSV2 showed strong viperin expression in cells located to the area of infection . Few cells co-expressed viperin and HSV2 antigens , but cells around the HSV2-infected cells expressed high levels of viperin . This included a range of cell types , including epithelial , stroma , and potentially also immune cells . To explore whether the upregulation of ISGs in the vaginal mucosa could play a role in control of infection in vivo , we set up three different treatment regimens with i . p . injections of 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) before or after HSV2 infection . The regimens were ( CDN treatment relative to infection ) : one pretreatment , one post-treatment , or two post-treatments ( Fig 4A ) . When treating WT mice , the CDN pretreatment and two-time post-treatments showed significantly improved survival as compared to non-treated controls , and the one-time post-treatment group suggested improved survival , but the data did not reach statistical significance ( p = 0 . 06 ) ( Fig 4B ) . We also tested cGas-/- mice , since they cannot produce 2’3’-cGAMP upon immune recognition of HSV2 , but can be stimulated by external delivery of STING agonists . When treating cGas-/- mice , however , only the pretreated group had a 100% overall survival , while almost all the post-treated mice eventually succumbed to the infection , although the two-times post-treatment led to significantly improved survival , ( Fig 4C ) . The differences in the effects of two-time post treatment on survival between WT and cGas-/- mice may be explained by the role of endogenous activation of STING signaling upon genital HSV2 infection , as reported previously [42] . All groups receiving CDN treatment , regardless of the genotype , had significantly lower viral loads in the vagina when compared to their respective controls ( Fig 4D ) . Taken together , these data demonstrate that activation of the STING pathway with the non-endogenous CDN 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) mounts a potent antiviral defense , the efficacy of which depends on the timing and repetition of treatment as well as the immunocompetence of the mouse . Given the clear antiviral action of CDNs in the vagina together with the systemic immune activation after i . p . treatment , we wanted to evaluate whether direct administration of CDNs to epithelial surfaces would also stimulate a protective response and with less systemic effects . For these experiments , we used 3’3’-cAIMP , which showed in vivo responses very similar to 2’3’-cGAMP ( PS ) 2 ( Rp/Sp ) and has a high solubility , thus allowing 250 μg to be applied in 20 μl saline . Six hours after application of 3’3’-cAIMP to the vaginal mucosal surface , we observed very strong staining for viperin throughout the epithelial layer ( Fig 5A ) . Interestingly , this was in contrast to mice receiving systemic CDN treatment , where viperin staining was strongest in cells localizing to the stroma and inner epithelium ( Fig 3 ) . In vaginal tissue from the mice receiving local CDN treatment , a strong upregulation of both Ifnb was observed and modest induction of Tnfa , Il6 , and A20 , ( Fig 5B ) . In contrast to this , no IFN signature was observed in the spleen , which has a very high expression of STING ( Fig 2A ) . However , we did detect elevated IFNβ protein levels in serum from some , but not all the mice treated with 3’3’-cAIMP in the vagina ( Fig 5C ) . To examine the antiviral response stimulated by locally administered CDN , WT or cGas-/- mice were treated in the vagina with 3’3’-cAIMP and infected with HSV2 . Following this treatment , both genotypes showed complete protection against disease , and no detectable virus in the vaginal lavage ( Fig 5D–5F ) . Staining of tissue sections from 3’3’cAIMP-treated HSV2-infected mice failed to reveal virus-infected cells , which was readily seen in the absence of CDN treatment ( Figs 5G and 3 ) . Finally , we examined how long the CDN treatment could be separated from HSV2 infection temporally . Local treatment with 3’3’-cAIMP up to 72 h prior to HSV2 infection conferred full protection against genital herpes and suppression of viral replication ( Fig 5H and 5I ) . Taken together , these data suggest that local administration of a STING agonist to a mucosal surface confers local protection against viral infections with only mild systemic activation of the IFN system . Previous attempts to take advantage of innate immune activation to control genital herpes infections , showed promising results in mice , but only limited clinical effects were obtained , and clear inflammatory side effects were observed [10 , 11] . TLR7 and 9 are mainly expressed on leukocytes , notably plasmacytoid dendritic cells , while HSV2 replicates in the vaginal epithelium . Therefore , we were interested in directly comparing induction of ISG expression by STING agonists relative to TLR agonists . We first treated HaCaT cells with imiquimod , ODN2216 or 3’3’-cAIMP and examined for expression of ISGs . While none of the TLR agonists induced ISG expression , this was observed following CDN treatment ( Fig 6A ) . Next , we were interested in examining how the different agonists affected induction of the IFN effector proteins , the ISGs , in the vaginal epithelium . First , we observed in agreement with previous results that ODN1826 , similarly to 3’3’-cAIMP , totally blocked genital HSV2 replication in mice , while imiquimod has a partial effect [43 , 44] ( Fig 6B ) . Interestingly , when vaginal tissue sections from mice treated locally with the TLR agonists for 6 h were stained for the ISG viperin , we observed only very sporadic positive staining , whereas the epithelium exposed to 3’3’-cAIMP was highly positive for viperin ( Fig 6C ) . At later time points after treatment with TLR agonists , more extensive viperin positive staining was observed in the vagina , but with more focal staining patterns in both the epithelium and subepithelial areas ( Fig 6D ) . At 6 h post treatment , significantly elevated levels of Ifnb and Mx1 mRNA were observed in the vaginal tissue after treatment with 3’3’-cAIMP , but not TLR agonists ( Fig 6E and 6F ) . Despite this , the TLR9 agonist induced Tnfa expression to an extent comparable to what was observed after STING activation ( Fig 6G ) . These results demonstrate that STING agonists directly targets epithelial cells , and that vaginal administration induces highly IFN-focused local responses . By contrast , TLR agonists do not directly target epithelial cells , and hence induce vaginal IFN responses with a delayed kinetics , and a stronger associated inflammatory gene expression profile .
Upon infection by a pathogenic microorganism , the body possesses numerous mechanisms to mount defenses that will clear the infection without causing major tissue damage . However , even in individuals with apparently full immune-competence , many infections are not well controlled , but can be cleared by anti-microbial treatments . In immunocompromised individuals , continuous treatment with antimicrobials provides an opportunity for microbes to develop drug resistance , thus hindering microbial control , and increasing the risk of severe disease [6] . Therefore , in recent years there has been increasing interest in not only directing therapy towards the infecting pathogen , but also in developing therapies that stimulate the immune system in a manner targeted towards elimination of the infections [45 , 46] . STING is an adaptor protein involved in DNA-activated innate immune signaling , the importance of which we have only just begun to unravel [47 , 48] . STING is involved in innate defense against infections with a range of different DNA-containing pathogens , such as viruses [49] , intracellular bacteria [50 , 51] , and protozoa [52] , but it is also implicated in autoinflammatory diseases , most notably STING-associated vasculopathy with onset in infancy [53] . Here , we show that it is possible to exploit the former role of STING in antiviral defense by treatment with STING agonists . We found that STING agonists confer protection against genital HSV2 infection in mice when applied systemically or locally . The therapeutic effects were more pronounced in cGas-/- mice as compared to WT mice , suggesting that endogenous activation of the STING pathway in WT mice did confer some degree of protection . The antiviral action of the CDNs could be mimicked when applying supernatants from CDN-stimulated THP-1 cells ( human monocyte-like cell line ) onto WT HaCaT cells ( human keratinocyte cell line highly permissive for HSV-1 ) , but not IFNAR-deficient HaCaT cells . This suggests that CDNs target a range of cell types and activate antiviral cross-talk between different cell types . The suggestion of an effect on different cell types led us to test the STING agonists in an in vivo model . We confirmed previous findings [51 , 54] , that CDNs induce a robust type I IFN response , and that they do so in a STING-dependent manner , with much less induction of genes stimulated by the NFκB-pathway . Moreover , we found that a panel of non-endogenous CDNs were superior to 2’3’-cGAMP in stimulating IFN responses in vivo while having comparable activity in vitro . Furthermore , 2’3’-cGAMP did not induce IFN responses in the brain , whereas the non-endogenous CDN 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) induced type I IFN in all tested tissues . This could indicate an ability of this CDN to better protect neurons from neuroinvasive viruses , although intrathecal type I IFN-synthesis has also been linked to neurotoxicity [55–58] . A mechanistic explanation as to why the non-endogenous CDNs were more potent than the endogenous 2’3’-cGAMP could be that the non-endogenous CDNs are resistant to hydrolysis by phosphodiesterases in the body . ENPP1 has been shown to efficiently degrade 2’3’-cGAMP , and has been found in the plasma as well as on the outside of the cell membrane [28] . Since 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) , has been shown to be a competitive inhibitor of ENPP1 [28] , treatment with CDN would also potentially amplify the response to endogenously produced 2’3’-cGAMP . The results are consistent with the model shown in S2 Fig . With respect to the current knowledge on STING-directed therapy against disease , DMXAA was shown to have antitumor effects even before its mechanism of action through STING was known [30 , 31 , 59] . CDNs have also been investigated for anticancer effects which was shown to be mediated by endothelially-produced type I IFN and correlated with the generation of more CD8+ T cells [35] . CDNs have been investigated as potential adjuvants on mucosal surfaces as they induced Th1 , Th2 and Th17 cells [38 , 40 , 60] . The induction of type I IFN by CDNs activates dendritic cells [51 , 54] , thus promoting maturation of CD4+ T cells [61 , 62] . However , recent papers from the Manel and Poltorak groups [63 , 64] suggest that CDNs could impair T cell-based immunity , which is known to be essential for full clearance of many viruses [65 , 66] . Therefore , the characterization of STING agonists in therapy should also include the impact on adaptive immunity . Previous studies have tested the effects of TLR agonists in models for genital herpes . While TLR3 and TLR9 agonists led to efficient protection against disease in mice , a TLR7 agonist was less efficient [67–69] . However , in the guinea pig model for genital herpes , the TLR7 agonist imiquimod potently induced protection . At the mechanistic level , all the tested TLR agonists work in a manner dependent on type I IFN [70] , and at least for the TLR9 agonist , the antiviral action was independent of IFNγ [67 , 71] . In addition , the TLR agonists potently induce expression of inflammatory cytokines with potential adverse effects for patients . Moreover , there are data suggesting that imiquimod also exerts biological activities through stimulation of the adenosine receptor A1-pathway , including upregulation of cystatin A [67 , 72] . Despite the promising data on TLR agonists in animal model systems [67–69] , and case reports on clinical effects of imiquimod treatment in individual patients [13 , 14] , imiquimod treatment failed to demonstrate significant effect in a clinical trial on genital herpes [10 , 11] . In addition , TLR agonists showed significant inflammatory side effects [10 , 11] . The results of our study together with the existing literature suggests that STING and TLR agonists target different cell types and have different modes of action . While , TLR7/9 agonists mainly target plasmacytoid dendritic cells [73 , 74] , and hence rely on recruitment to the infected tissue , STING agonists act directly on epithelial cells to induce IFN/ISGs . In addition , TLR agonists are stronger inducers of inflammatory cytokines , such as TNFα , than STING agonists are . Collectively , this suggests that local treatment with STING agonists leads to a rapid IFN/ISG response aimed at the epithelial cells , and with limited inflammation . It should also be mentioned that STING agonists potently induce autophagy and different types of programmed cell death , which have been reported to have antiviral activity [75 , 76] . Therefore , STING agonists have potential as a novel therapeutic option for treatment of genital herpes , due to their small molecular size and targeted type I IFN-biased response , which in addition induces type III IFN , autophagy , programmed death pathways , and low levels of inflammatory cytokines . The treatment potential includes acyclovir-resistant HSV , which is a problem in immunocompromised patients permanently treated with nucleoside analogs [6] . Finally , future studies should address the clinically important question on the impact on herpesvirus latency and reactivation . In conclusion , we report that natural and non-natural STING agonists induce expression of type I IFNs and ISGs in human immune and tissue cells , and that this stimulates type I IFN-dependent intercellular cross-talk . This was found to stimulate anti-HSV2 activity both in vitro in human cells and in vivo in a mouse model for genital herpes . Most notably , mucosal administration of the STING agonist 3’3’-cAIMP evoked a local IFN response , which conferred total protection against genital herpes , even in a highly permissive mouse strain . Such data highlight the potential for immunotherapy in treatment of virus infections and suggest STING-directed therapy to hold a potential that should be further explored .
Mice were age-matched ( 6–8 weeks of age ) female C57BL/6J ( WT ) , STINGgt/gt ( STING-Goldenticket; C57BL/6J-Tmem173gt/J ) and cGAS-/- ( B6 ( C ) -Mb21d1tm1d ( EUCOMM ) Hmgu/J ) , and all experiments were carried out at Aarhus University . Vero cells were obtained from ATCC ( CLL-81TM ) and were grown in DMEM with 5% FCS supplemented with L-glutamine and antibiotics . HaCaT cells ( ATCC HB-241 ) were grown in DMEM and THP-1 cells ( ATCC TIB-202 ) in RPMI with 10% FCS and supplemented as described above . All STING and TLR agonists were supplied by InvivoGen: 5 , 6 , -dimethylxanthenone-4-acetic acid ( DMXAA , vadimezan ) , 2’3’-cyclic guanosine monophosphate-adenosine monophosphate ( 2’3’-cGAMP ) , 3’3’-cyclic di-adenosine monophosphate ( 3’3’-c-di-AMP ) , the Rp , Sp-isomers of the bisphosphorothioate analog of the mammalian cyclic dinucleotide 2’3’-cGAMP ( 2’3’-cGAM ( PS ) 2 ( Rp/Sp ) ) , 3’3’-cyclic adenosine monophosphate-inosine monophosphate [77] , imiquimod , ODN1826 , and ODN2216 . STING agonists were diluted in PBS ( with 10% DMSO for DMXAA ) and administered by local application or i . p . injections . STING agonists were added to the cell medium . In some experiments , the cells were pretreated with digitonin to permeabilize the cell membrane . In brief , medium was removed and cells were permeabilized with digitonin ( 5 μg/mL ) and buffer ( 0 . 2% BSA , 50 mM HEPES , 100 mM KCl , 3 mM MgCl2 , 0 . 1 mM DTT , 85 mM sucrose , 1 mM ATP , 0 . 1 mM GTP , 2 . 2 mM NaOH ) together with the respective drug for 10 minutes . New antibiotics-free medium was then supplied . TLR agonists were diluted in PBS . Mock treatment was performed with PBS in an appropriated volume . One preparation of HSV2 333 strain was used for all experiment and produced as previously described [41] . Cells were infected with an MOI of 0 . 1 . The mice were pretreated with a subcutaneous ( s . c . ) injection of 2 mg Depo-Provera ( medroxyprogesterone acetate; Pfizer ) . Five days later , the mice were anesthetized and inoculated intravaginally with 20 μL HSV2 ( 6 . 7×104 p . f . u . ) suspended in IMDM . The mice were then placed on their backs for 10 min . Vaginal washes were collected 48 h post infection ( p . i . ) by washing with 40 μL of IMDM and dilution to a final volume of 200 μL . In the survival experiments , infected mice were weighed and monitored for disease symptoms daily and euthanized when they reached humane endpoints: weight loss >10% , severe inflammation with ulceration in the genitoanal region , paresis of the hind limbs , or hunchback with antisocial behavior and facial expressions of pain [78] . The mouse was anesthetized and the right heart atrium was punctured . A feeding needle was used to infuse PBS , 10 mL/min , through the left ventricle , clearing the circulation of blood . Successful perfusion was confirmed by the color change of the liver from deep red-brown to pale brown . Sera from mice were analyzed for IFNμ and -β by Bioluminescent LumiKine Xpress ELISA ( luex-mifnb and luex-mifna; InvivoGen ) as described by the manufacturer . Human type I IFN was assessed by HEK-Blue™ IFN-α/β reporter cells by the manufacturer guide lines ( InvivoGen ) . Vaginal tissues were fixated in 10% formalin , embedded in paraffin , and cut at 4 μm . The tissues sections submitted to antigen retrieval in a citrate buffer and blocked in 5% BSA . The sections were incubated overnight with primary antibody for HSV2 ( B0116; Dako ) and viperin ( MABF106; EMD Millipore ) , and the stain was visualized with appropriate secondary antibodies from Molecular Probes . DAPI was used to stain cell nuclei . 37 , 500 Vero cells were seeded into each well of a 96-well plate . The next day , sample was added in 8 replicates and serially diluted . The result was read at 48 h p . i . , using light microscopy , and the 50% tissue culture infective dose ( TCID50 ) was calculated , using the Reed-Muench method [79] . Data represented as scatter plots with the geometric mean ± SD . Protein was isolated with RIPA buffer supplemented with protease inhibitors , and the protein concentration was measured with a Bradford protein assay . The samples were run on a reduced gel , and the membrane was blocked in either 2 . 5% BSA ( for phosphorylated proteins ) or 5% skim milk . The following primary antibodies were used: anti-vinculin ( V9131; Sigma ) , anti-viperin ( MABF106; EMD Millipore ) , anti-STING ( AF6516; R&D ) , anti-VP5 ( Ab6508; Abcam ) , anti-pSTAT1 ( 7649S; Cell Signaling ) , ISG15 ( 2743S; Cell Signaling ) . Appropriate secondary antibodies from Jackson ImmunoResearch were used . The blots were visualized with an ImageQuant LAS 4000 Mini Luminescent Image Analyzer ( GE Healthcare ) . RNA from non-CNS tissues was isolated with the RNeasy Mini kit with on-column DNase digestion ( Qiagen ) . RNA from CNS tissues was isolated with TRIzol ( Ambion ) and digested with DNase I ( InvitroGen ) . Gene expression was determined by reverse transcriptase quantitative PCR ( RT-qPCR ) , using the SYBR Green ( Agilent Technologies ) and TaqMan ( Applied Biosystems ) systems . 50 ng of RNA was used for each reaction . Expression levels were quantified relative to the expression of GAPDH , using the 2-ΔΔCT method [80] , and normalized to the control group as a fold change . Data are represented as scatter plots with the mean ± SD of biological replicates . The following TaqMan primers were used: mIfnb ( Mm00439552_s1; Applied Biosystems ) , mMx1 ( Mm00487796_m1; Applied Biosystems ) . The following SYBR Green primers were used: mGapdh ( FW: 5’-CAA TGT GTC CGT CGT GGA-3’; RW: 5’-GAT GCC TGC TTC ACC ACC-3’ ) , mA20 ( FW: 5’-TGC AAT GAA GTG CAG GAG TC-3’; RW: 5’-TGG GCT CTG CTG TAG TCC TT-3’ ) , mIl6 ( FW: 5’-GAA AAT CTG CTC TGG TCT TCT GG; RW: 5’-TTT TCTG ACC ACA GTG AGG AAT G ) , mTnfa ( FW 5’-CAC AGC CTT CCT CAC AGA GC; RW: 5’-GGA GGC AAC AAG GTA GAG AGG ) . THP1-Dual Cells ( thpd-nfis ) reporter cell line was obtained from InvivoGen and enable detection of the ISG54 and the NFκB pathways . Briefly , this cell line expressed a SEAP reporter gene under control of a promoter that is responsive for the NF-κB or AP-1 pathways and a Lucia reporter gene under the control of a promoter that comprises five IFN-stimulated response elements ( ISREs ) fused to a minimal promoter of the human ISG54 gene . 24h stimulation of this cell line with CDN analogs subsequently induces production of Lucia and SEAP , which are measured using Quanti-Luc and Quanti-Blue respectively . All statistical analyses were performed with GraphPad Prism version 7 . 03 for Windows , GraphPad Software , La Jolla California USA , www . graphpad . com . Statistical significance of p<0 . 05 was marked with an asterisk ( * ) . For details about the statistical procedures , please refer to the respective figure legends . In general , all analyses were initially sought to be carried out with parametric procedures ( one-way ANOVA , two-way ANOVA ) and adequate post-hoc tests for multiple comparisons ( Šidák correction , Dunnett’s test ) . Virus titers were log10-transformed prior to analysis [81] . When the underlying assumptions of the parametric test were not met , e . g . the data were not normally distributed or the SDs were significantly different as tested by Prism , we used non-parametric tests ( Mann-Whitney-Wilcoxon U test , Kruskal-Wallis test ) and adequate tests for multiple comparisons ( Dunn’s test ) . Comparison of overall survival was done with the log-rank test and the post-hoc Holm-Bonferroni correction . Animal studies were reviewed and approved by Dyreforsøgstilsynet under the Danish Ministry for Veterinary and Food Administration ( permission#: 2015-15-0201-00686 ) . The study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals , EEC Council Directive 2010/63/EU . | Herpes simplex virus ( HSV ) -2 is the leading cause of genital ulcers , and HSV-2 infection has also been reported to amplify HIV-transmission . So far , all attempts at making an effective anti-HSV2 vaccine have failed . In recent years , there has been an increasing interest in immunomodulatory therapy as a means to treat infections . Although the TLR7 agonist imiquimod has been shown to have antiviral activity in individual patients , no significant effects were observed in clinical trials , and the compound also exhibited significant side effects including local inflammation . Type I interferon ( IFN ) s are key players in antiviral defense , and it is now known that the DNA sensor cyclic GMP-AMP synthase produces the cyclic di-nucleotide ( CDN ) 2’3’-cyclic GMP-AMP ( cGAMP ) , which activates the adaptor protein STING to induce IFN expression . In this work we show that natural and non-natural CDNs activate strong type I IFN responses in vivo without stimulating significant expression of genes driven by the transcription factor NF-κB , which induces inflammation . Application of CDNs at epithelial surfaces gave rise to local IFN activity , but only limited systemic responses . Importantly , all tested treatment regimens , strongly reduced replication of HSV-2 in a model for genital herpes , and significantly reduced development of disease . Finally , when comparing to TLR agonists , CDNs showed the best profile with strong IFN response specifically in the epithelial cells and limited induction of inflammation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods",
"and",
"materials"
] | [
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"epithelial",
"cells",
"animal",
"models",
"model",
"organisms",
"signs",
"and",
"symptoms",
"experimental",
"organism",
"systems",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"inflammation",
"animal",
"cells",
"proteins",
"gene",
"expression",
"biological",
"tissue",
"mouse",
"models",
"vagina",
"immune",
"response",
"toll-like",
"receptors",
"biochemistry",
"signal",
"transduction",
"diagnostic",
"medicine",
"anatomy",
"cell",
"biology",
"interferons",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"immune",
"receptors",
"genetics",
"genital",
"anatomy"
] | 2018 | STING agonists enable antiviral cross-talk between human cells and confer protection against genital herpes in mice |
Cross-species comparison of great ape gesturing has so far been limited to the physical form of gestures in the repertoire , without questioning whether gestures share the same meanings . Researchers have recently catalogued the meanings of chimpanzee gestures , but little is known about the gesture meanings of our other closest living relative , the bonobo . The bonobo gestural repertoire overlaps by approximately 90% with that of the chimpanzee , but such overlap might not extend to meanings . Here , we first determine the meanings of bonobo gestures by analysing the outcomes of gesturing that apparently satisfy the signaller . Around half of bonobo gestures have a single meaning , while half are more ambiguous . Moreover , all but 1 gesture type have distinct meanings , achieving a different distribution of intended meanings to the average distribution for all gesture types . We then employ a randomisation procedure in a novel way to test the likelihood that the observed between-species overlap in the assignment of meanings to gestures would arise by chance under a set of different constraints . We compare a matrix of the meanings of bonobo gestures with a matrix for those of chimpanzees against 10 , 000 randomised iterations of matrices constrained to the original data at 4 different levels . We find that the similarity between the 2 species is much greater than would be expected by chance . Bonobos and chimpanzees share not only the physical form of the gestures but also many gesture meanings .
In a series of well-known children’s books , Doctor Dolittle was able to talk to nonhuman animals , but in reality , deciphering meaning in nonhuman communication presents a much bigger challenge . First , there is the question of whether animal signals can be said to have ‘meanings’ or merely ‘functions’ . Functions are known for many animal signals: for example , various species are able to decode complex information from their conspecifics’ calls on the location or class of food or predators [1–4] , level of risk [4 , 5] , and size of predator [6] . However , for meaning , a signal needs to be produced intentionally—the signaller must aim to change the behaviour ( first-order intentional ) or the mental state ( at least second-order intentional ) of the recipient [7–9] . Mounting evidence shows that , unlike most nonhuman animals [10] , great apes habitually engage in first-order intentional communication: great apes routinely direct their gestures towards a specific recipient; monitor that recipient’s attentional state and choose gestures appropriate to it; wait for the recipient to respond; and , if the recipient does not respond , they persist and elaborate with further gestures [11–17] . These criteria demonstrate that the signaller has a specific outcome in mind and uses gestures to achieve that outcome [18] . It has also been argued that to have meaning , communication needs to be ostensive , drawing attention to the fact that it is being used to communicate [19] . In developmental psychology , eye gaze is taken as an ostensive cue; the audience checking performed by great apes before gesturing serves the same ostensive function [20] . Because great apes deploy gestures intentionally , it is appropriate to go beyond simply describing their function and enquire about the intended meaning that a signaller aims to achieve by gesturing [20] . Although we focus on gestural communication , it should be noted that great apes also appear to deploy some vocal signals intentionally [21–23] . Moreover , we focus on a Gricean approach to meaning , rather than a semantic approach [24 , 25] , given that few great ape gestures appear to be referential ( but see [26] ) . The second challenge is that gesture meanings must be deduced indirectly . Past studies have tackled the issue of meaning by looking at the context in which gestures occur [16 , 27] , thereby showing that the same gesture may occur in several contexts . We have taken a different approach . By using the reaction that each gesture elicits , but only in cases where the signaller’s behaviour indicates that this reaction was their intended aim , we hope to pin down the signaller’s intended meaning for each specific gesture . The meaning of a gesture can thus be defined by the ‘Apparently Satisfactory Outcome’ ( ASO ) , the reaction of the recipient that satisfies the signaller as shown by cessation of gesturing [28] . This method indicates the individual signaller’s intended meaning in each instance , and across many instances and individuals , one can examine the gesture’s general meaning ( s ) in a population . Aggregating the meanings represents population level patterns of meaning but does not infer that meanings are conventionalised nor indeed does it imply any particular ontogeny for gesture meanings . Defining meaning by ASOs , wild chimpanzees use their gestures to achieve at least 19 ASOs; that is , their gestures achieve 19 types of behavioural response from the recipient [28] . Each gesture type has a distinct ( set of ) meaning ( s ) that is calculated by comparing the distribution of meanings for a gesture type to the distribution of meanings across all gesture types [28] . Using ASOs to define the meaning of gestures is a relatively new approach , so gesture meanings have not yet been defined for our other closest living relative—the bonobo ( or ‘bilia’ , as the species is locally known ) [29] . Our study is the first to investigate meaning in the natural gestural repertoire of wild bonobos . Once the meanings of bonobo gestures are defined , we can examine the gestural overlap of bonobos and chimpanzees . All species of nonhuman great ape share the majority of their gestural repertoire in terms of the gestures’ physical forms . The overlap for chimpanzees and bonobos is 88%–96% [30]; for chimpanzees and gorillas , 60% [31]; and for chimpanzees and orangutans , 80% [31] . But simply using the same actions does not mean that chimpanzees and bonobos share a communication system ( that is , that a chimpanzee and bonobo would in principle be able to understand one another ) . Only if bonobo and chimpanzee gestures share the same meanings can they be said to share the same system of communication . Deciding that issue is not straightforward . Ape gestural repertoires are large , with over 70 distinct gestures in the chimpanzee and bonobo catalogues . In captivity , large quantities of gestural data can be collected very quickly , but the majority of it occurs during play [32 , 33] . Data from the wild are needed to examine the full breadth of meaning expressed in nonplayful ape communication . Previous studies have used traditional analyses of variance or goodness of fit tests , demonstrating that different individuals within a chimpanzee group use the same gesture to achieve the same outcome [28] . However , despite data sets containing thousands of gesture cases , large repertoires and the regular use of only a subset of these gestures [34] limits the number of gesture types that can be examined in this way . Furthermore , those tests are not suited to data sets in which many of the possible outcomes never occur for each signal type , as we would expect in a system of communication in which specific signals are employed for specific outcomes . We have therefore adapted methods from numerical ecology to compare the similarity between the meanings of bonobo and chimpanzee gestures . In doing so , we offer the first analysis that examines whether the overlap in the physical form of bonobo and chimpanzee gestures extends to their meaning .
We analysed 2 , 321 intentional gesture instances ( occasions on which a gesture was used ) that successfully achieved an ASO . These instances concerned 33 gesture types ( categories of gestures that share the same physical form ) [30 , 31] ( S1 Table ) and 14 different ASOs: ‘Acquire object/food’ , ‘Climb on me’ , ‘Climb on you’ , ‘Contact’ , ‘Follow me’ , ‘Initiate grooming’ , ‘Mount me’ , ‘Move closer’ , ‘Reposition’ , ‘Initiate copulation’ , ‘Initiate genito-genital rubbing ( GG-rubbing ) ’ , ‘Travel with me’ , ‘Move away’ , and ‘Stop behaviour’ . The first 12 of these ASOs served to initiate or develop an activity , and the last 2 served to stop an activity . Of the 33 gesture types , 17 had only a single ASO , 6 had 2 ASOs , and 10 had >2 ASOs ( Fig 1 ) . The mean number of ASOs per gesture type was 2 . 27 ± 1 . 84 ( median = 2 , range 1–8 ) . Next , in accordance with [28] , we used a series of ANOVAs to analyse whether the distribution of ASOs for a given gesture type differed from the average distribution of ASOs across all gesture types ( Figs 2–5 ) . Fifteen gesture types were suitable for analysis , having been used by at least 3 individuals at least 3 times to achieve an ASO or ASOs ( see Materials and methods for more information ) . If gestures were achieving outcomes at random , we would expect no difference between the distribution of a given gesture type and the average distribution across all gesture types . All but 1 gesture type ( Object shake ) showed significant deviation from the average distribution . Bonobo gesture types , like chimpanzee gesture types [28] , do have distinct ( sets of ) meanings . Using a randomisation procedure , we tested the null hypothesis that the similarity between the 2 species ( Fig 6 ) would be the same under a random assignment of gestures to ASOs for each species ( see Materials and methods ) . We compared 4 different methods of matrix permutation ( R code in S1 Data ) , generating gesture-to-ASO assignment matrices with ( a ) no constraints ( least conservative ) , ( b ) constraints on the column sums , ( c ) constraints on the row sums , and ( d ) constraints on both column and row sums ( most conservative ) , none of which produced a pair of matrices that were more similar than the original data ( Fig 7 ) . When constraining the column or row sums , the total number of ASOs a gesture was assigned to ( row sum preserve ) or gestures an ASO was assigned to ( col . sum preserve ) in a permutation was constrained to that of the original chimpanzee and bonobo matrices , though the actual assignment is random . For example , under the row sum preserve method , the row “Object Shake” would have exactly 7 1s for any permutation of the chimpanzee matrix and a single 1 for the bonobo matrix , as in the original data ( raw data and species matrices in S2 Data ) . We can be confident that the similarity of the gesture matrices for the 2 species is greater than expected by chance assignment of gestures to ASOs , as defined by the randomisation procedure . We further explored the randomisation procedure , in order to find the limits of its application to communication , by repeating the randomisation process on subsets of our data . Specifically , we examined the effects of the available number of observed gestures on the results of our analysis by subsetting our original data to include an incrementally increasing number of gestures , from 4 to the maximum 21 available for use . The probability of generating randomised gesture matrices , using any of the 4 different constraint sets described above , that are more similar than the original gesture matrices remains very low ( <0 . 05 ) as long as at least 8 or more of the gestures are included in the randomisation ( Fig 8 ) . The fact that such a strong signal of similarity between the 2 species’ gesture matrices exists with considerably less data lends weight to the robustness of our main result .
Bonobos intentionally deploy gestures to achieve at least 14 different intended outcomes—12 that initiate or develop an activity and 2 that stop it . They use gestures to request things ( such as food ) and to initiate co-locomotion , grooming , and sex . Because the gestures are intentionally produced , meeting widely accepted criteria for intentional communication [18] , these outcomes are not only the gestures’ ‘functions’—they are their ‘meanings’ [20 , 28] . Moreover , bonobo gesture types have distinct ( sets of ) meanings . Almost all gesture types achieve a different distribution of ASOs to the average distribution , showing distinct aims . Object shake , the 1 exception , may have failed to show a distinctive pattern because of its small sample size and because it is primarily used for sex and grooming—2 behaviours that numerically dominate bonobo gesture instances and thus contribute substantially to the average distribution . Overall , we conclude that bonobos are using gestures to achieve distinct outcomes , as has also been found for chimpanzees [28] . About half of bonobo gestures have only a single meaning , while the others have 2 or more meanings . Words in human language can have a single meaning or polysemous meanings , and this poses no problem for the recipient in deciphering the signaller’s intended meaning . Future research ( with an expansive enough dataset ) can explore how bonobo recipients appear able to correctly interpret the meaning of apparently ambiguous gesture types , perhaps by including analysis of facial expressions , gesture sequences , or local situational context . Having catalogued the meanings of bonobo gestures , we then compared them with the meanings of chimpanzee gestures , finding evidence of their similarity . Across 10 , 000 random permutations of the gesture matrices , we failed to generate a single pair that were more similar than the observed data , implying a negligible probability that such similarity arose by chance . This was the case even under the most conservative constraints , where randomised matrices were generated maintaining the number of assignments in both columns and rows as the original data . Our findings remained robust , with clear similarities found between bonobo and chimpanzee meanings , even with subsets of our main gesture matrices , down to a minimum of just 8 ( of the 21 total ) gesture types . In the future , researchers would ideally be able to use these methods to compare the meaning of gesture repertoires among a range of primate species , determining whether more closely related species have more similar gesture repertoires . The bonobo and chimpanzee gestural repertoires—that is , the physical form of the gestures—overlap by 88%–96% [30] . We now know that there is also a large overlap in the intended outcomes achieved by these shared gesture types in bonobos and chimpanzees . Whilst biological inheritance is one possible explanation for this overlap , we recognise that similar gestures and meanings could emerge through another acquisition mechanism , such as ontogenetic ritualization [35] or a version of imitation [36 , 37] ( but see [17] ) . Bonobos and chimpanzees also experience similar environmental and anatomical constraints that may restrict the available gestures and desired outcomes . More research is needed to explore the precise mechanism behind the overlap of gesture meanings . It is probable that this pattern of gestures and meanings also applied to the last common ancestor we shared with the 2 Pan species . That is , it is likely that the Pan-Homo last common ancestor would have been able to use and understand most of the gestures of modern bonobos and chimpanzees; less likely , but not impossible , the elaborate shared Pan repertoire could have evolved after divergence from the hominin lineage . If we can now discover whether humans also share or understand these great ape gestures , those 2 possibilities can be resolved [38] . Doubtless , gestural communication was an important contributor in the evolution of language [39 , 40]; but it remains to be seen how gesture as it manifests in nonhuman great apes relates to gesture as it manifests in humans alongside thought and language . Understanding this ‘baseline’ of gestural communication may better enable us to predict those new meanings that development of protolanguage offered to human-specific ancestors , ultimately resulting in the evolution of language .
KEG collected data on 2 neighbouring communities of wild bonobos ( E1 group and P group ) at Wamba , Luo Scientific Reserve , Democratic Republic of the Congo ( 00° 10' N , 22° 30' E ) . Habituation began for E1 group ( n = 39 ) in 1976 ( when it was still part of E group ) and for PE group ( n = 30 ) in 2010 ( 24 , 25 ) . At the beginning of this study , in 2014 , the total sample size was 63 individuals , with 28 adults , 12 adolescents , 9 juveniles , and 14 infants . In 2015 , the total sample size was 64 individuals , with 30 adults , 8 adolescents , 10 juveniles , and 16 infants . Bonobo age groups are divided into infant ( <4 years ) , juvenile ( 4–7 years ) , adolescent ( 8–14 years ) , and adult ( 15+ years ) [41] . CH collected data on 1 community of wild chimpanzees ( Sonso community ) at the Budongo Conservation Field Station , Uganda ( 1° 35’–1° 55’ N , 31° 18’–31° 42’ E ) . Habituation began for the Sonso community ( n = 92 ) in 1990 . At the beginning of this study , in 2007 , the total sample size was 81 individuals , with 32 adults , 16 subadults , 15 juveniles , and 18 infants . Chimpanzee age groups are divided into infant ( ≤4 years ) , juvenile ( 5–9 years ) , subadult ( male: 10–15 years , female: 10–14 years ) , and adult ( male: 16+ years , female: 15+ years ) [42] . KEG conducted fieldwork from 4 February 2014 to 28 June 2014 and 19 January 2015 to 13 June 2015 , following bonobos daily from approximately 05:50 to approximately 12:00 , with a weekly schedule of 4 days on and 1 day off . Observation time amounted to 204 days . CH conducted fieldwork from 25 October 2007 to 8 March 2008 , 13 April 2008 to 1 January 2009 , and 5 May 2009 to 8 August 2009 , following chimpanzees daily from approximately 07:30 to approximately 16:30 , with a weekly schedule of 3 days on , 1 day off , 3 days on , 2 days off . Observation time amounted to 266 days . We filmed social interactions using focal behaviour sampling , where the focal behaviour was whenever 2 or more individuals approached within 5 m of each other . We chose this criterion to ensure that any gestures preceding social interactions were recorded . KEG recorded video footage with a Panasonic HDC-SD90 video camera , using the 3-second pre-record feature to increase the likelihood of catching the gestures in time . CH used a MiniDV tape using a Sony Handycam ( DCR-HC-55 ) . We imported video footage each day , labelled it , and entered it into a clip directory in FileMaker Pro . Filemaker Pro was also used for video coding . Each gesture instance ( that is , a single gesture ) was coded in a separate sheet , with the following information: signaller , recipient , signaller age and sex , recipient age and sex , gesture type , part of sequence , audience checking , response waiting , persistence , recipient response , and ASO . The signaller is the individual who produces the gesture , and the recipient is the individual who is the target of the gesture . The gesture type is a category comprising physically similar gesture instances , where gesture instances are grouped by body part and action . A complete list of gesture types is described in [20] , with additional bonobo gesture types found in [19] . A sequence is a series of gesture instances separated by <1 s and produced by 1 individual . Audience checking , response waiting , and persistence are all criteria for intentionality . For audience checking , we reported whether or not the signaller turned to face the recipient; for response waiting , whether or not they paused for >1 s after gesturing; and for persistence , whether or not they continued to gesture . To be included in analyses , we required that each gesture instance meet at least 1 of these criterion for intentionality . Recipient response was categorical: No response , ASO , Gesture ( if the recipient responded with a gesture ) , or Unknown . To analyse meaning , we only used gesture instances where the recipient responded with an ASO . For a gesture to be assigned an ASO , we required that the recipient react to the gesture sequence with an ASO ( that is , a response that satisfies the signaller shown by cessation of gesturing ) . ASO was then the specific outcome by the recipient . KEG coded all the bonobo video footage and , to test interobserver reliability , CH coded 100 gesture instances for several of the aforementioned categories: gesture type , audience checking , persistence , and signaller apparently satisfied . We analysed Cohen’s kappa for interobserver reliability of these variables giving 0 . 87 ( almost perfect ) , 0 . 56 ( moderate ) , 0 . 70 ( substantial ) , and 0 . 63 ( substantial ) , respectively . CH coded all of the chimpanzee video footage; interobserver was conducted in their 2011 paper [31] , with another experienced coder coding 50 gesture instances for directedness , recipient attentional state , and gesture type ( Cohen’s kappa: directedness , κ = 0 . 69 [substantial]; recipient attentional state , κ = 0 . 63 [substantial]; gesture type , κ = 0 . 86 [almost perfect] ) . We recorded 4 , 256 intentionally produced gesture instances for wild bonobos , but we only analysed the 2 , 463 gesture instances ( including those in sequences ) that successfully achieved an ASO . We then excluded gestures used in play ( 231 instances ) , because including cases where gestures were used playfully would risk masking their normal meaning—the very nature of play means that gestures would be used playfully , not necessarily containing the same meaning they would otherwise . All statistical analyses were conducted in R 3 . 2 . 3 . In accordance with Hobaiter & Byrne 2014 [28] , we used a series of ANOVAs to analyse whether the specific distribution of ASOs for a gesture type differed from the average distribution ( the distribution of ASOs across all gesture instances ) . For direct comparability , we set the same parameters in our analyses to those used in their previous study [28] . To be included in parametric analyses , we required that each gesture type achieve an ASO at least 3 times ( per individual ) by at least 3 individuals ( we analysed 1 , 896 gesture instances; 15 gesture types were suitable for this analysis , and 51 individuals contributed data ) . Then , we converted the number of instances a gesture type achieved any 1 ASO into a proportion of the total number of gesture instances in which an individual used that gesture type . We also calculated the average distribution by converting the number of instances in which all gesture instances achieved each ASO into a proportion of the total number of gesture instances . For values of 0 or 1 , we converted them in accordance with Snedecor and Cochran ( 0 → 1/ ( 4N ) and 1 → 1- ( 1/ ( 4N ) ) , where N is the total number of instances for that gesture type ) [43] . Finally , to calculate how the specific distribution deviated from the average distribution , we subtracted the average from the specific distribution . We then conducted the ANOVA with this resulting deviation as the dependent variable , ASO as the independent variable , and signaller identity as a random effect . P values of < 0 . 05 show that the deviation of the specific from the average distribution is significant . The relationships between gestures and ASOs for each species were represented as a matrix in which each row corresponded to a possible gesture and each column corresponded to 1 of the possible ASOs . A ‘1’ in this gesture matrix indicated that the associated gesture was observed to precede the associated ASO in the corresponding species . The criterion for inclusion was that a gesture type must achieve the given ASO at least 2 times and by a minimum of 2 individuals ( that is , ape individual A uses it once and ape individual B uses it once ) . Note that criteria for the previous ANOVA were necessarily strict to meet the requirements for parametric analyses . In comparing the communication of 2 species that differ markedly in social behaviour , it is important not to mistake differences in the frequency of use of signals , which are to be expected , with genuine differences in communication system . To avoid that error , we deliberately adopted a looser criterion so that subtle differences in the bonobo and chimpanzee repertoires could be detected but not confused with spurious differences in usage frequency . A ‘0’ in the matrix indicated that such an association was not observed . In order to deal with the few cases where we had insufficient data for 1 or other species , we defined the possible gestures as the intersection of all gestures used ( n = 22 ) and possible ASOs as the intersection of all ASOs observed across both species ( m = 11 ) . Thus , the dimensions of the gesture matrices were the same for both species ( n × m ) , and each row and column had at least one ‘1’ . We defined the similarity between 2 gesture matrices simply as the sum of all matching corresponding matrix entries , be they 0 or 1 . Using a randomisation procedure , we tested the null hypothesis that the similarity between the 2 species would be the same under a random assignment of gestures to ASOs for each species [44] . To perform the randomisation test , we iteratively generated new gesture matrices for each species by randomly permuting the original gesture matrices and calculating the similarity between the 2 resultant matrices , generating a null distribution for similarity over 10 , 000 iterations . We used 4 different methods of permutation , each imposing different constraints on the possible matrices that could be generated . The simplest method simply randomly shuffled the values in each matrix without any constraints . The row sum preserve method shuffled the entries in each row of a matrix , thus preserving the number of ASOs assigned to each gesture for each species . The column sum preserve method shuffled the entries in each column , thus preserving the number of gestures allocated to each ASO . Finally , the row and column sum preserve method maintained both the number of gestures allocated to ASOs and ASOs allocated to gestures and was performed using the “tswap” algorithm in the vegan package in R , which implements a swap algorithm to generate new matrices that preserve row and column sums whilst sampling the distribution of possible matrices with equal probability [44] . From the null distribution of similarity values generated by each method , we calculated a corresponding P value as the proportion of iterations of the randomisation procedure in which the resultant similarity score was equal to or exceeded that of the original data . We then examined the effects of the available number of observed gestures on the results of our analysis by subsetting our original data to include an incrementally increasing number of gestures , c , from 4 to the maximum 21 available to use . For each gesture count , c , and each iteration , we randomly sampled c gestures ( rows ) from both species’ gesture matrices to form the data subset for that iteration . Using this subset of the original data , randomised matrices were generated and the resultant similarity compared to that of the nonrandomised , subsetted matrices . A probability value was calculated as the proportion of iterations in which the randomised subset matrices were of equal or greater similarity than their nonrandomised subset counterparts . This probability value allowed us to test the null hypothesis that the observed similarity was the same as would be expected by chance , given that only c randomly selected gestures were observed . When c = 21 , the maximum number of gestures available , any subset was the same as the original matrix , so the randomisation test was equivalent to that described in the main text . It should be noted that this randomisation test does not rule out the possibility of another primate species having a more similar gesture–ASO matrix to chimpanzees or bonobos than they have to each other , though , to the best of our knowledge , no such data currently exist in order to test this . The procedure simply compares the similarity of the 2 species to the similarity of hypothetical gesture matrices generated under the above set of constraints . | Bonobos and chimpanzees are closely related members of the great ape family , and both species use gestures to communicate . We are able to deduce the meaning of great ape gestures by looking at the ‘Apparently Satisfactory Outcome’ ( ASO ) , which reflects how the recipient of the gesture reacts and whether their reaction satisfies the signaller; satisfaction is shown by the signaller ceasing to produce more gestures . Here , we use ASOs to define the meaning of bonobo gestures , most of which are used to start or stop social interactions such as grooming , travelling , or sex . We then compare the meanings of bonobo gestures with those of chimpanzees and find that many of the gestures share the same meanings . Bonobos and chimpanzees could , in principle , understand one another’s gestures; however , more research is necessary to determine how such gestures and gesture meanings are acquired . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"acoustics",
"sociology",
"social",
"sciences",
"vertebrates",
"animals",
"mammals",
"animal",
"signaling",
"and",
"communication",
"bonobos",
"primates",
"analysis",
"of",
"variance",
"animal",
"behavior",
"mathematics",
"statistics",
"(mathematics)",
"zoology",
"discrete",
"mathematics",
"combinatorics",
"research",
"and",
"analysis",
"methods",
"behavior",
"mathematical",
"and",
"statistical",
"techniques",
"communications",
"physics",
"eukaryota",
"permutation",
"biology",
"and",
"life",
"sciences",
"apes",
"physical",
"sciences",
"chimpanzees",
"amniotes",
"statistical",
"methods",
"organisms",
"semiotics",
"acoustic",
"signals"
] | 2018 | Bonobo and chimpanzee gestures overlap extensively in meaning |
Gene expression responds to changes in conditions but also stochastically among individuals . In budding yeast , both expression responsiveness across conditions ( “plasticity” ) and cell-to-cell variation ( “noise” ) have been quantified for thousands of genes and found to correlate across genes . It has been argued therefore that noise and plasticity may be strongly coupled and mechanistically linked . This is consistent with some theoretical ideas , but a strong coupling between noise and plasticity also has the potential to introduce cost–benefit conflicts during evolution . For example , if high plasticity is beneficial ( genes need to respond to the environment ) , but noise is detrimental ( fluctuations are harmful ) , then strong coupling should be disfavored . Here , evidence is presented that cost–benefit conflicts do occur and that they constrain the evolution of gene expression and promoter usage . In contrast to recent assertions , coupling between noise and plasticity is not a general property , but one associated with particular mechanisms of transcription initiation . Further , promoter architectures associated with coupling are avoided when noise is most likely to be detrimental , and noise and plasticity are largely independent traits for core cellular components . In contrast , when genes are duplicated noise–plasticity coupling increases , consistent with reduced detrimental affects of expression variation . Noise–plasticity coupling is , therefore , an evolvable trait that may constrain the emergence of highly responsive gene expression and be selected against during evolution . Further , the global quantitative data in yeast suggest that one mechanism that relieves the constraints imposed by noise–plasticity coupling is gene duplication , providing an example of how duplication can facilitate escape from adaptive conflicts .
For cellular adaptation gene expression must respond to changes in conditions . However , expression also varies stochastically among cells in a population . In the budding yeast Saccharomyces cerevisiae both variation across different conditions ( ‘expression plasticity’ , [1] ) and variation among individuals in a constant environment ( ‘expression noise’ , [2] ) have been quantified for thousands of genes . Comparisons across genes have shown that these two levels of expression correlate , suggesting that noise and plasticity may somehow be mechanistically coupled [3]–[7] . Correlations among different levels of expression variance are consistent with some theoretical proposals , which consider that stochastic , environmental , and mutational perturbations are likely to similarly affect biological systems [8]–[11] . The findings are also supported by mechanistic studies where mutations that increase noise have also increased plasticity [12] , [13] . Further , several properties such as initiation from a TATA-box promoter [2] , [4] , [6] and high proximal promoter nucleosome occupancy [3] , [14] , [15] are enriched among genes with both high noise and high plasticity . It should be noted , however , that both properties are also associated with genes with a wide range of noise and plasticity levels [2]–[4] , [6] , [14] , [15] . Theoretical work and intuition do , however , also suggest that noise and plasticity may not always by strongly coupled in this way . For example , by altering the size of transcriptional bursts , the interval time between bursts , or the number of decay steps in the degradation of a protein it should be possible to alter noise independently of plasticity [16] . Therefore it is important to ask whether coupling between expression noise and expression plasticity is , as reported [3] , a general result . Or , rather , is coupling an evolvable trait that can vary among genes ? Further , what is the mechanistic basis of coupling ? Does coupling constrain expression evolution ? How are such constraints relieved ? And has coupling itself been subject to selection ? In some specific situations a strong coupling between noise and plasticity may be disfavored . For example , if high expression plasticity is beneficial , facilitating environmental adaptation , but high noise is detrimental , then strong coupling will cause a fitness cost–benefit conflict . This potential for an adaptive conflict predicts that for many genes strong coupling between noise and plasticity would reduce fitness . In short , when noise is detrimental , noise–plasticity coupling should be disfavored . It is not known if this is the case . Here using global quantitative data from yeast it is shown that noise–plasticity coupling is not a general result , but rather a property of particular promoter architectures , and so is an evolvable trait . Promoter architectures that favor coupling are underrepresented among genes required for viability , and for these genes noise and plasticity are rather independent , consistent with selection against coupling . Following gene duplication , however , when the detrimental effects of expression variation in many cases will be reduced , the constraints on coupling and promoter architectures that favor coupling appear to be relaxed . Thus , noise–plasticity coupling is not a general trait , but one that likely is both constrained by selection and constrains the evolution of gene expression . Further , this constraint may be relieved following gene duplication , providing an example of escape from adaptive conflict [17] , [18] .
Considering all genes in yeast there is a reasonable correlation between levels of gene expression variation in a single condition ( ‘expression noise’ [2] ) and the variability of expression across changing conditions ( ‘expression plasticity’ [1] ) ( Spearman's correlation coefficient , rho = 0 . 30 , p<2 . 2E-16 , n = 2049 [3]–[6] ) . However , considering different classes of genes shows that this is not a general result . In yeast only about 20% of genes are transcribed from promoters containing TATA box elements [19] . For these genes , noise and plasticity are strongly coupled ( rho = 0 . 62 , p<2 . 2E-16 , n = 369 , Figure 1A ) . In contrast , for non-TATA genes coupling is much weaker ( rho = 0 . 16 , p = 1 . 4E-10 , n = 1680 , Figure 1B ) . This indicates that the extent of coupling may relate to the mechanism of transcription initiation . The stronger coupling of TATA genes is confirmed when controlling for possible confounding features such as the absolute level of plasticity ( Table S1 ) , the requirement of a gene for viability ( Figure 1C and 1D ) , gene function ( Table S3 ) , expression level ( Table S4 ) , protein complex membership ( Table S5 ) , the number of upstream regulators ( Table S6 ) , the identity of upstream regulators ( Table S7 ) , histone exchange rates ( Table S8 ) , and nucleosome occupancy ( Table 1 ) . In summary , coupling between noise and plasticity in yeast appears related to the promoter architecture of a gene . In eukaryotes DNA is packaged into chromatin , and this chromatin structure varies across the promoters of different genes [14] , [15] , [20] . Chromatin remodeling is thought to be a major source of transcriptional noise [13] , [21] , [22] . Many genes in yeast contain a DNA-encoded region of low nucleosome occupancy in their proximal promoters , a feature often associated with low expression noise [14] , [15] . Considering the nucleosome occupancy of promoters shows that high noise–plasticity coupling is also associated with high upstream nucleosome occupancy ( Figure S1A and S1B , rho = 0 . 13 , p = 2 . 4E-4 , n = 789 for genes with upstream nucleosome free regions and rho = 0 . 48 , p<2 . 2E-16 , n = 273 for genes with high upstream nucleosome occupancy ) . This result is confirmed when controlling for gene importance ( Figure 2A and 2B ) , the absolute levels of plasticity ( Table S1 ) , and also when only considering non-TATA promoters , even though these promoters show lower overall levels of coupling ( Table 1 , rho = 0 . 08 for non-TATA genes with upstream nucleosome free regions and rho = 0 . 24 for non-TATA genes with high upstream nucleosome occupancy ) . Promoters also differ in their nucleosome dynamics , and the exchange of core histones has been quantified across much of the genome [23]–[25] . High-noise plasticity coupling is also associated with high rates of histone exchange in promoter regions ( Table S8 ) . This is confirmed when controlling for TATA presence ( Table S8 ) , nucleosome occupancy ( Table S9 ) , or when only considering genes with low plasticity ( Table S10 ) . Thus , in addition to the link between noise–plasticity coupling and initiation from a TATA box , stronger coupling is also associated with higher and more dynamic promoter nucleosome occupancy . This strengthens the evidence that noise–plasticity coupling relates to the process of transcription initiation and indicates that coupling relates to chromatin remodeling . It also suggests that the extent of coupling is a trait that has the potential to change during evolution . High expression variation can be detrimental if it results in insufficient protein production , and there is good evidence that this is the case in yeast because genes required for viability have low noise [2] , [5] , [26] , [27] . This predicts that noise–plasticity coupling should also be disfavored for these genes: although genes required for viability still need to respond to external conditions ( for example coupling growth rates to changes in the environment ) , excessive variation in their production would be detrimental . Consistent with this , essential genes , haploinsufficient genes ( genes that reduce fitness when their copy number is reduced by half ) and genes required for growth all show little or no significant coupling between noise and plasticity ( Figure 3 , rho = 0 . 15 p = 0 . 001 n = 492 for essential genes , p = 0 . 19 for haploinsufficient genes , and p = 0 . 43 for genes required for growth ) . Similar results are seen when only considering genes with low absolute levels of plasticity ( Table S1 ) . Consistent with predictions that noise will be detrimental for protein complex subunits [27] , coupling is also lower for these genes ( Table S2 ) . Thus when high noise is likely to be detrimental , noise and plasticity are largely unrelated traits in yeast . These findings are also consistent with differences in promoter architectures . Whereas 24% of non-essential genes use TATA promoters , only 3% of haploinsufficient genes , 9% of essential genes , and 11% of genes required for growth do ( p<10E-14 in all cases , Fisher's exact test ) . Further , genes required for growth or viability usually have nucleosome free regions in their proximal promoters ( 74% compared to 57% for other genes , p<2 . 2E-16 ) and they have low levels of promoter histone exchange ( mean 0 . 9 compared to 1 . 1 for other genes , p = 1 . 2E-11 , Kolmogorov-Smirnov ( KS ) test ) . Thus , for genes encoding core cellular components , promoter architectures associated with high noise–plasticity coupling are avoided . This is consistent with a model in which selection disfavors coupling when it is detrimental . In many cases , therefore , the evolution of highly responsive gene expression from TATA promoters may be constrained by the detrimental effects of high noise . How can genes escape this adaptive conflict and evolve highly plastic TATA-initiating expression ? One event that has been proposed as a general mechanism to facilitate escape from adaptive conflicts is gene duplication [17] , [18] . Here it is argued that one conflict that can be resolved by duplication is the conflict between gene expression noise and plasticity . Following the duplication of a gene , variation in expression can be less detrimental if there is functional compensation between duplicates and a component of the expression variation of the duplicates is independent [28]–[30] . Thus , if the evolution of highly responsive , but noisy , expression is constrained , then this constraint may be relieved by duplication: promoter architectures that favor plasticity ( but that also couple this plasticity to noise ) should be less detrimental . Three sets of observations from yeast are consistent with this proposal . First , duplicates in yeast have high levels of plasticity , and also high levels of noise ( Figure 4A and 4B ) . Even duplicates known to redundantly perform a process required for viability or growth [31] have higher noise than single copy genes ( p = 2 . 8E-7 , KS test ) , showing that variation in their expression is not detrimental . Thus duplicates tolerate higher expression variation than other genes . Second , TATA initiation , which facilitates plasticity but also couples noise and plasticity , is much more frequent among duplicates than among other genes: whereas 9% of genes required for viability or growth initiate from TATA promoters , this rises to 35% of gene duplicates redundantly required for growth or viability ( p<2 . 2E-16 , Fisher's exact test ) . TATA promoters are indeed strongly enriched among duplicates of all ages , and for duplicates arising from both small scale and whole genome duplication events ( Table S11 ) . Similar trends for the enrichment of TATA dependent transcription among duplicates are also seen in other species [32] . Thus , duplicates in general more frequently use promoters that couple noise to plasticity . Third , evidence from the whole genome duplication ( WGD ) also supports this model . Following the WGD ∼100 million years ago , most genes reverted to a single copy but with a substantial number retained as duplicates [33] . This allows a direct comparison between genes retained as duplicates and those reverting to a single copy after a common ancestral event . Considering a set of genes inferred to have non-TATA promoters prior to the WGD ( see Materials and Methods ) , those retained as duplicates are twice as likely to have gained a TATA promoter since the WGD than those reverting to a single copy ( Figure 4C , P = 1 . 24×10−5 , Fischer's exact test , N = 470 and N = 1994 , respectively ) . This shows that not only are duplicates enriched for TATA promoters , but that they also tend to gain TATA promoters post-duplication . Thus , following duplication noise–plasticity coupling likely increased for many genes , which is consistent with this coupling being less detrimental . In summary , the global data in yeast are consistent with duplication relieving a constraint on the evolution of highly plastic but variable gene expression . Thus one benefit of duplication may be that it allows escape from the adaptive conflict [17] , [18] of noise–plasticity coupling .
Previous studies have suggested that each promoter may have a “unique capacity to respond to external signals that can be environmental , genetic or even stochastic” [3] . Here it has been shown that this conclusion is not correct , but that the extent of noise–plasticity coupling relates to the mechanism of transcription initiation , and is confined to a subset of genes in yeast . This means that coupling can be an evolvable trait , as changes in promoter architecture associate with stronger or weaker coupling between expression plasticity and expression noise . This study initiated from the hypothesis that the reported strong coupling between noise and plasticity could be detrimental because of the potential for fitness cost–benefit conflicts . The quantitative data from yeast support this idea , showing that both noise–plasticity coupling and promoter architectures that favor coupling are avoided when coupling is likely to be detrimental . Although only correlative in nature , the data are consistent with noise–plasticity coupling being not just an evolvable trait , but also one that has likely been constrained by selection . It has been shown here that TATA genes show a striking coupling between noise and plasticity . Thus , when high plasticity is adaptive , for TATA genes this will nearly always be accompanied by high noise . This means that although in some instances high noise may be beneficial [34]–[37] , this should not be assumed as the case . Provided that it is not detrimental , high noise may be nothing more than a by-product of high plasticity . For many genes , however , there is good evidence that high noise would be detrimental [2] , [5] , [26] , [27] and for these genes TATA-dependent initiation is disfavored and strong noise–plasticity coupling is not observed . How can genes escape a potential adaptive conflict between the benefits of plasticity and the costs of noise ? One likely mechanism is gene duplication . Whereas prior to duplication the detrimental consequences of noise may limit the evolution of highly responsive expression , following duplication variation may be better tolerated due to functional compensation [28]–[30] . Constraints on the evolution of highly responsive , but noisy expression should therefore be relieved following duplication . The quantitative data from yeast are consistent with this model , showing that noise , plasticity , and the use of TATA promoters all increase among duplicates . Thus one general benefit of duplication may be that it facilitates escape from the adaptive conflict [17] , [18] imposed by coupling between expression noise and expression plasticity , permitting the evolution of responsive and variable expression .
Expression plasticity is defined as the total responsiveness of each gene's expression to environmental change in a large compendium of over 1500 S . cerevisiae expression profiling experiments [1] , as reported in [6] . Values shown in this manuscript , as for noise , are scaled between 0 and 1 . Expression noise is quantified from single cell-profiling measurements of fluorescently tagged proteins , using the ‘DM’ measure of Newman et al . , which accounts for the influence of protein abundance on coefficient of variation measurements [2] . Promoters were classified as ‘nucleosome occupied’ or ‘nucleosome free’ using in vivo nucleosome occupancy data [38] in 100 base pairs upstream of each gene as previously described [15] . A total of 1082 ‘nucleosome occupied’ proximal promoters ( clusters 7 and 8 from Tirosh et al . ) and 1940 ‘nucleosome free’ proximal promoters ( clusters 2 , 3 , and 4 ) are considered . Histone H3 exchange data is from [25] . The average exchange in each promoter is used , with a promoter defined as 500 base pairs upstream of a gene's start site . TATA containing promoters in S . cerevisiae were identified using the classification of Basehoar et al . [19] . Ancestral genes were considered as the set of genes with an ortholog present in each of the closely related pre-WGD species Zygosaccharomyces rouxii , Kluyveromyces lactis , Ashbya gossypii , Saccharomyces kluyveri , Kluyveromyces thermotolerans , and Kluyveromyces waltii [39] . TATA-boxes were identified in promoter regions of these species using the definition of Basehoar et al . by scanning the −70 to −310 region of each gene's promoter for the consensus site TATA ( A/T ) A ( A/T ) ( A/G ) [19] . In the analysis , ancestral non-TATA genes are those inferred from the absence of a consensus TATA-box in any pre-WGD species . Duplicates were identified using the SYNERGY algorithm [40] , which uses gene trees based on sequence similarity and shared gene order across 17 fungal genomes to resolve orthology and paralogy relationships [41] . Whole-genome duplicates ( WGD ) and their orthologs in pre-WGD species were identified using the yeast genome order browser [42] version 3 . 0 [39] . Here conserved synteny and parsimony are used to identify ortholog groups . Genetically redundant genes were compiled from systematic studies and the literature , as described [31] . Here only redundant genes where the single gene deletions do not result in slow growth are considered . Considering all redundant genes or redundant genes arising in the WGD gave very similar results . All statistical tests were performed using R ( www . r-project . org ) . | Gene expression needs to respond to changes in conditions , but also varies stochastically among cells in a homogenous environment . It has been argued that these two levels of expression variation may be coupled , relating to the same underlying molecular mechanisms . However , such a strong coupling between expression “plasticity” and expression “noise” may introduce cost–benefit conflicts during evolution . For example , if plasticity is beneficial , but noise is detrimental , then coupling will be disfavored . In this work , evidence is presented that such cost–benefit conflicts do occur and that they constrain the evolution of gene expression in yeast . In contrast to recent conclusions , it is shown that noise–plasticity coupling is not a general result , but rather one associated with particular mechanisms of transcription initiation . Promoter architectures associated with coupling are avoided when noise is detrimental , and noise and plasticity are not coupled for core cellular components . Noise–plasticity coupling is therefore not a general property of gene expression , but an evolvable trait that may constrain the evolution of gene expression and be selected against during evolution . Further , gene duplication may facilitate escape from the adaptive conflict imposed by coupling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/gene",
"expression",
"genetics",
"and",
"genomics/bioinformatics",
"evolutionary",
"biology/genomics"
] | 2010 | Conflict between Noise and Plasticity in Yeast |
MAP kinases are integral to the mechanisms by which cells respond to a wide variety of environmental stresses . In Caenorhabditis elegans , the KGB-1 JNK signaling pathway regulates the response to heavy metal stress . In this study , we identified FOS-1 , a bZIP transcription factor , as a target of KGB-1-mediated phosphorylation . We further identified two transcriptional targets of the KGB-1 pathway , kreg-1 and kreg-2/lys-3 , which are required for the defense against heavy metal stress . FOS-1 plays a critical role in the transcriptional repression of the kreg-1 gene by recruiting histone deacetylase ( HDAC ) to its promoter . KGB-1 phosphorylation prevents FOS-1 dimerization and promoter binding , resulting in promoter derepression . Thus , HDAC behaves as a co-repressor modulating FOS-1-mediated transcriptional regulation . This study describes the direct link from JNK signaling , Fos phosphorylation , and regulation of kreg gene transcription , which modulates the stress response in C . elegans .
Mitogen-activated protein kinase ( MAPK ) signal transduction pathways are evolutionarily conserved in eukaryotic cells and transduce signals in response to a variety of extracellular stimuli . Each pathway is composed of three classes of protein kinases: MAPK , MAPK kinase ( MAPKK ) and MAPK kinase kinase ( MAPKKK ) [1] , [2] . MAPKKK phosphorylates and activates MAPKK , which in turn activates MAPK . This activation cascade can be reversed by phosphatases . In particular , members of the MAPK phosphatase ( MKP ) family can remove phosphate groups from activated MAPK [1] , [2] . Three subgroups of MAPKs have been identified: extracellular signal-regulated kinase ( ERK ) , c-Jun N-terminal kinase ( JNK ) , and p38 kinases [1] , [2] . JNK and p38 MAPKs function as key mediators of stress and immune signaling in mammals . The MKK4 and MKK7 MAPKKs have been shown to activate JNK , and the MKK3 and MKK6 MAPKKs serve as the major activators of p38 MAPK . The specific MAPKKs are themselves phosphorylated and activated by specific MAPKKKs . Different MKPs display different activities toward ERK , JNK , and p38 . Invertebrate model organisms such as Drosophila melanogaster and Caenorhabditis elegans are useful for understanding the effects and interactions of JNK proteins , especially since they are amenable to the analysis of cytoprotective gene expression and the specific contributions of different tissues [3] , [4] . Recent studies in C . elegans have revealed that the JNK MAPK signaling components are highly conserved between C . elegans and mammals . One such C . elegans JNK pathway is the KGB-1 pathway , composed of an MLK-type MAPKKK MLK-1 , an MKK7-type MAPKK MEK-1 and a JNK-type MAPK KGB-1 [5] . The KGB-1 pathway is required for the protection against heavy metals and protein folding stress [5] , [6] , [7] , and regulates the transcriptional responses to bacterial pore-forming toxins [8] . Another component of this pathway is the MKP VHP-1 , which negatively regulates the KGB-1 pathway by dephosphorylating KGB-1 [5] . However , the components that function downstream of the KGB-1 pathway have yet to be elucidated . Various targets of JNK phosphorylation have been identified in mammalian systems , including members of the basic region leucine zipper ( bZIP ) family of transcription factors such as ATF2 and Jun [9] , [10] . The activating protein 1 complex ( AP-1 ) constitutes an important subset of bZIP transcription factors [9] , [10] . AP-1 component proteins interact as homodimers or heterodimers , bind DNA through conserved bZIP domains , and regulate transcription of their target genes . A large body of research supports a model in which extracellular stimuli trigger AP-1 phosphorylation by JNK , leading to reprogramming of target gene expression [11] , [12] . Given the importance of chromatin dynamics in the control of gene expression , recent work has focused on factors interacting with AP-1 that can mediate chromatin modification and remodeling , notably enzymes that reversibly modify histone tails by acetylation . The histone deacetylase ( HDAC ) complex was thus found to inhibit the JNK pathway [13] , [14] . Gene repression by the HDAC complex is relieved by phosphorylation of Jun , which causes it to dissociate from the promoter [15] , [16] . These findings suggest that chromatin dynamics may play a central role in the cellular response to JNK signaling . To understand the role of KGB-1 signaling in the heavy metal stress response , we screened for proteins that may interact with KGB-1 and identified FOS-1 , a C . elegans homolog of Fos , and showed that it functions downstream of KGB-1 . In addition , we identified two genes whose expression is induced by copper in a KGB-1-dependent manner: kreg-1 and kreg-2 ( KGB-1 regulated genes ) . We found that FOS-1 represses transcription via the recruitment of a Class I histone deacetylase HDA-1 to the promoter . Biochemical assays demonstrated that phosphorylation by KGB-1 inhibits FOS-1 self-association and binding to the kreg-1 promoter . These results suggest that FOS-1 and HDA-1 play an inhibitory role in the response to heavy metal stress , and that the KGB-1 pathway confers tolerance to heavy metals by phosphorylating and thereby negatively regulating FOS-1 .
To identify components that function downstream of KGB-1 , we screened a C . elegans mixed-stage cDNA library by the yeast two-hybrid method to isolate proteins that interact with KGB-1 . Generally , kinase-negative ( KN ) forms of protein kinases constitutively associate with their substrate . Therefore , as bait we used KGB-1 ( K67R ) , a KN form in which Lys-67 in the ATP-binding motif has been mutated to arginine . From this screen , we identified 10 proteins that interact with KGB-1 ( Table S1 ) . One of them is FOS-1 , an ortholog of the mammalian Fos transcription factor [10] , [17] . Because Fos is a known substrate of MAPK in many systems , we considered FOS-1 as a likely substrate of KGB-1 . The FOS-1 protein is similar to other Fos proteins in that it possesses a basic DNA-binding domain , a leucine zipper region , and a carboxyl terminus rich in serine and threonine residues , which are typical sites of phosphorylation ( Figure 1A ) . The fos-1 gene encodes two FOS-1 isoforms , FOS-1A and FOS-1B [17] . As FOS-1A has previously been characterized as a regulator of anchor-cell invasion during nematode development [17] , we focused our investigations on the FOS-1B form ( hereafter referred to as FOS-1 ) . To confirm an interaction between KGB-1 and FOS-1 , we co-expressed HA-tagged KGB-1 KN and T7-tagged FOS-1 in COS-7 cells , immunoprecipitated HA-KGB-1 KN with anti-HA antibodies , and probed for T7-FOS-1 on a Western blot with anti-T7 antibodies . We found that KGB-1 KN co-immunoprecipitated with FOS-1 ( Figure 1B ) , indicating that KGB-1 can physically associate with FOS-1 . The physical association of KGB-1 with FOS-1 suggested that FOS-1 may be a phosphorylation target of KGB-1 . Indeed , in COS-7 cells , co-expression of KGB-1 activated by MEK-1 resulted in the appearance of slower migrating forms of the FOS-1 protein when analyzed by SDS-polyacrylamide gel electrophoresis ( SDS-PAGE ) ( Figure 1C , lane 2 ) . Pre-treatment of extracts with alkali phosphatase reduced the intensities of the band shifts ( Figure 1C , lane 3 ) , which is a typical indicator of dephosphorylation . Expression of MEK-1 in the absence of KGB-1 did not induce any mobility shift ( Figure S1A ) . The FOS-1 protein contains six putative MAPK phosphorylation sites ( S/TP ) : Ser-151 , Thr-263 , Thr-278 , Thr-304 , Thr-316 , and Thr-318 ( Figure 1A ) . We generated a mutant form of FOS-1 , [FOS-1 ( 6A ) ] , in which all 6 Ser/Thr residues had been changed to Ala . When we analyzed extracts from COS-7 cells transfected with FOS-1 ( 6A ) together with active KGB-1 , we observed no slowly migrating bands in SDS-PAGE ( Figure S2A , lane 12 ) . To identify the specific phosphorylated residue ( s ) in FOS-1 , we introduced various combinations of Ala mutations into the six Ser/Thr residues . We observed that the T304A , T316A , T318A triple mutation completely abrogated phosphorylation of FOS-1 ( Figure S2B , lane 9 ) , suggesting that Thr-304 , Thr-316 , and/or Thr-318 are potential phosphorylation sites . We further generated three FOS-1 mutants that individually changed Thr-304 , Thr-316 , or Thr-318 to Ala and found that the FOS-1 ( T304A ) mutation exhibited decreased phosphorylation by KGB-1 ( Figure 1D , line 3 and Figure S2 ) . These results suggest that T304 is a major site of phosphorylation . However , we did also observe a minor slower-migrating band , indicating that there is some residual phosphorylation of FOS-1 ( T304A ) and that Thr-316 and/or Thr-318 may be minor sites of KGB-1 phosphorylation . To confirm that KGB-1 phosphorylates FOS-1 at the Thr-304 residue , we generated anti-phospho-FOS-1 antibodies that specifically recognize FOS-1 phosphorylated at Thr-304 . Transfection with active KGB-1 , but not with the kinase-negative mutant KGB-1 KN , resulted in strong reactivity of FOS-1 with this antibody ( Figure 1D , lanes 1 , 2 ) . In contrast , we found that the FOS-1 ( T304A ) mutated form could not be detected by this antibody ( Figure 1D , lane 3 ) , confirming that it was specific for FOS-1 phosphorylated at Thr-304 . Fos family proteins function as dimers that bind DNA and regulate the transcription of target genes [9] , [10] , [18] . We therefore next investigated whether FOS-1 undergoes homo-dimerization . FOS-1 was fused to both GFP and FLAG and expressed in COS-7 cells together with T7-FOS-1 . We immunoprecipitated the GFP-FLAG-FOS-1 protein with anti-GFP antibodies , and tested for co-precipitation of T7-FOS-1 by blotting with anti-T7 antibodies . We differentiated between GFP-FLAG-FOS-1 and T7-FOS-1 by virtue of their different molecular weights . Indeed , GFP-FLAG-FOS-1 readily co-immunoprecipitated with T7-FOS-1 ( Figure 1E , lanes 1 , 2 ) , indicating that the two proteins oligomerized , presumably as dimers . We next examined whether KGB-1 phosphorylation correlated with the degree of FOS-1 self-association . Co-expression of active but not inactive KGB-1 resulted in reduced co-immunoprecipitation of T7-FOS-1 with GFP-FLAG-FOS-1 ( Figure 1E , lanes 3 , 4 ) . We next examined the phosphorylation state of FOS-1 self-association using anti-phospho-FOS-1 antibodies and observed that the phosphorylated form of T7-FOS-1 was not co-precipitated with GFP-FLAG-FOS-1 ( Figure 1E , lane 3 ) . This indicates that phosphorylation inhibits self-association of FOS-1 . We also generated a mutant intended to mimic FOS-1 phosphorylation by replacing the Thr-304 residue with glutamic acid , with the purpose to examine its self-association potential . However , when expressed in COS-7 cells , FOS-1 ( T304E ) exhibited faster migration on SDS-PAGE compared to wild type FOS-1 ( Figure S1B ) , suggesting that the structure of FOS-1 ( T304E ) is different from that of phosphorylated FOS-1 . Thus , this mutation does not appear to mimic FOS-1 phosphorylation . Since the KGB-1 MAPK pathway regulates the response to heavy metal stress [5] , [6] , [7] , we tested whether FOS-1 also regulates the stress response to heavy metals . Existing fos-1 loss-of-function mutants could not be used to assay for heavy metal toxicity , because they have a sterile phenotype ( data not shown ) . We therefore tested the effect of fos-1 knockdown on the stress response using a feeding RNA interference ( RNAi ) method . Animals were placed on agar plates containing copper ( Cu2+ ) ions , fed a bacteria strain expressing the double-stranded RNA for fos-1 , and their development was monitored for any signs of an altered response to heavy metal stress . As shown Figure 2A , fos-1 RNAi had no effect on the sensitivity to Cu2+ ions . Animals treated with fos-1 RNAi exhibited an everted/protruded vulval phenotype in the adult as observed in fos-1a loss-of-function mutants [17] . This indicates that fos-1 RNAi indeed had caused knockdown of fos-1 . In contrast to the lack of effect in wild-type animals , fos-1 RNAi suppressed the sensitivity to Cu2+ ions in kgb-1 ( km21 ) mutants ( Figure 2A and Figure S3 ) , suggesting that FOS-1 negatively regulates the tolerance to heavy metal stress . The above results raised the possibility that KGB-1-mediated phosphorylation of FOS-1 Thr-304 relieves FOS-1-mediated inhibition in response to stress . To test this possibility , we expressed wild-type FOS-1 or the non-phosphorylatable FOS-1 ( T304A ) mutant from the heat shock promoter ( Phsp-16 ) in wild-type animals . We found that expression of FOS-1 ( T304A ) resulted in sensitivity to Cu2+ ion compared to expression of wild type FOS-1 ( Figure 2B ) . These results suggest that KGB-1 phosphorylation at Thr-304 negatively regulates FOS-1 function . To understand how the KGB-1 pathway modulates gene activity and to define the physiological processes in which the heavy metal stress response may be involved , we examined gene expression changes in wild-type and kgb-1 mutant animals subjected to heavy metal stress by carrying out a microarray analysis ( see Materials and Methods ) ( Figure S4A and Tables S2 , S3 , S4 , S5 , S6 , S7 ) . From this , we identified six kreg ( KGB-1-regulated gene ) genes whose expression was regulated by KGB-1 ( Figure S4B and Table S8 ) . Among these , expression of two of the genes was increased in response to Cu2+ ions ( Figure S4B and Table S8 ) . These were designated kreg-1 and kreg-2 . The protein encoded by kreg-1 ( F53A9 . 2 ) is a novel 83 amino acids protein with polyhistidine streches , while the kreg-2 gene is identical to lys-3 , which encodes a lysozyme . We validated our microarray data by quantitative real-time RT-PCR ( qRT-PCR ) ( Figure 3A and 3B ) . In wild-type animals , Cu2+ induced the expression of both kreg-1 and kreg-2 , but in kgb-1 ( km21 ) mutants induction of both genes was considerably reduced . To determine whether the kreg genes play functionally important roles in the resistance to heavy metal stress in C . elegans in vivo , we used RNAi to inhibit the expression of kreg-1 or kreg-2 and then examined the stress response . RNAis against either kreg-1 or kreg-2 caused a partial sensitivity to Cu2+ ions ( Figure 3C and Figure S5 ) . The kreg-2/lys-3 gene encodes a secreted lysozyme that is presumably involved in anti-bacterial defense [19] . This raised the possibility that there may be a role for bacteria in the susceptibility to heavy metal stress . To test this possibility , we fed the worms on viable versus heat-killed bacteria and asked if this affected their heavy metal sensitivity . We found that heat treatment of bacteria had no effect on either the heavy metal sensitivity in wild-type animals or the heavy metal sensitive phenotype caused by kgb-1 and lys-3 mutations ( data not shown ) . Thus , bacteria appear to play no role in the susceptibility to heavy metal stress and it remains unclear how LYS-3 may protect against heavy metal stress . To analyze in vivo kreg-1 expression patterns and to develop tools for further analysis , we generated a Pkreg-1::venus reporter , consisting of the kreg-1 promoter driving expression of venus . Wild-type animals harboring the Pkreg-1::venus reporter exhibited weak Venus expression in the absence of Cu2+ ( Figure 3D and 3E ) . However , Pkreg-1::venus expression was robustly induced in the intestine of animals following incubation with Cu2+ ( Figure 3D and 3E ) . To confirm that the Pkreg-1::venus reporter behaves similarly to endogenous kreg-1 mRNA , we tested whether Pkreg-1::venus induction is dependent on the KGB-1 MAPK pathway , which is negatively regulated by the VHP-1 phosphatase [5] . In contrast to the wild-type animals , very little Pkreg-1::venus expression was induced by Cu2+ in kgb-1 ( km21 ) mutants ( Figure 3D and 3E ) . Treatment of animals with vhp-1 RNAi resulted in the constitutive expression of the Pkreg-1::venus transgene in wild-type , but not in kgb-1 ( km21 ) animals ( Figure 3D and 3E ) . Thus , the Pkreg-1::venus reporter is induced in response to heavy metal stress through the activation of the KGB-1 pathway . To understand the role of FOS-1 in the induction of kreg-1 in response to Cu2+ stress , we examined the effect of fos-1 RNAi on Pkreg-1::venus expression in C . elegans . Treatment with fos-1 RNAi markedly increased intestinal Pkreg-1::venus expression even in the absence of Cu2+ ( Figure 4 ) . The effect of fos-1 RNAi on expression of kreg-1 and kreg-2 was further confirmed by qRT-PCR ( Figure S6 ) . These results raised the possibility that FOS-1 functions as a repressor for gene induction activated by the KGB-1 pathway . To test this hypothesis , we carried out epistasis analysis using fos-1 RNAi and kgb-1 ( km21 ) mutants . We observed that while expression of the Pkreg-1::venus reporter gene was diminished in kgb-1 ( km21 ) mutants , treatment with fos-1 RNAi was epistatic to this and resulted in increased kreg-1 reporter activity ( Figure 4 ) . This indicates that FOS-1 functions downstream of KGB-1 as a repressor of kreg-1 induction by Cu2+ . Incubation with Cu2+ induced Pkreg-1::venus expression in the intestine in a manner dependent on the KGB-1 pathway . This observation suggests that activation of the KGB-1 pathway in the intestine is critical in the defense against heavy metal stress . Consistent with this , MEK-1 , a MAPKK in the KGB-1 pathway , is expressed in intestinal cells [6] , [20] . However , we have previously shown that expression of MEK-1 in the epidermis can rescue the Cu2+- sensitive phenotype of mek-1 null mutants [6] . To test whether expression of MEK-1 in the intestine of mek-1 mutants confers resistance to heavy metal stress , we expressed the mek-1 cDNA in the intestine using the elt-2 promoter . The mek-1 ( ks54 ) deletion mutant carrying Pelt-2::mek-1 exhibited resistance to heavy metal stress ( Figure S7 ) . The Pkreg-1::venus reporter may lack the region required for its expression in epidermis . Fos proteins bind to Jun or other bZIP proteins to create an AP-1 dimer complex , which regulates gene expression [9] , [10] , [18] . In fact , similar to mammalian and Drosophila Fos and Jun proteins , C . elegans FOS-1 and JUN-1 form heterodimers [18] , [21] . To examine whether C . elegans jun-1 plays the same role as fos-1 in modulating kreg-1 expression , we treated wild-type animals with jun-1 RNAi , however it failed to increase intestinal Pkreg-1::venus expression ( Figure S8A ) . ATF-7 is a member of the bZIP transcription factor family and functions in innate immunity mediated by the PMK-1 p38 pathway [22] . We therefore tested the effect of atf-7 RNAi on Pkreg-1::venus reporter activity and similarly observed no effect ( Figure S8A ) . Consistent with these results , neither knockdown of jun-1 nor a loss-of-function atf-7 ( qd22 ) mutation resulted in enhanced heavy metal stress sensitivity in wild-type animals or suppression of the stress-sensitive phenotype of kgb-1 mutants ( Figure S8B and S8C ) . Thus , JUN-1 and ATF-7 do not participate in the heavy metal stress response mediated by the KGB-1 pathway . The bZIP domain of Fos binds the consensus sequence , TGA ( C/G ) TCA , called the TPA-responsive element ( TRE ) [23] . The promoter region of the kreg-1 gene contains two TRE binding motifs , termed TRE1 and TRE2 ( Figure 5A ) . To determine whether these TRE motifs are required for FOS-1-mediated repression of Pkreg-1::venus expression , we deleted each motif independently within the Pkreg-1::venus reporter ( Figure 5A ) . Deletion of TRE1 ( Pkreg-1Δtre1::venus ) had no effect on the expression pattern of the transgene ( Figure 5B and 5C ) . In contrast , deletion of TRE2 ( Pkreg-1Δtre2::venus ) resulted in constitutive expression in both wild-type and kgb-1 ( km21 ) mutant animals ( Figure 5B and 5C ) . Furthermore , we found that treatment with fos-1 RNAi did not enhance constitutive expression of the Pkreg-1Δtre2::venus transgene ( Figure 5B and 5C ) . Thus , the TRE2 binding site is required in cis to mediate repression of kreg-1 by FOS-1 . These results support the possibility that FOS-1 negatively regulates kreg-1 expression through the TRE2 site in the promoter . To examine whether FOS-1 binds directly to the kreg-1 promoter via TRE2 , we conducted chromatin immunoprecipitation ( ChIP ) assays . Human embryonic kidney ( HEK ) 293 cells were co-transfected with the Pkreg-1::venus reporter together with either T7-FOS-1 or the negative control T7-hGrhl2 . Lysates were immunoprecipitated with anti-T7 antibodies , and quantitative PCR analysis was performed to amplify DNA fragments contained in the immunoprecipitated complexes . PCR analysis showed that FOS-1 bound efficiently to the kreg-1 promoter , whereas the negative control human Grhl2 protein did not ( Figure 6A ) . We could detect binding of FOS-1 to the Pkreg-1Δtre1::venus transgene ( data not shown ) , but not to the Pkreg-1Δtre2::venus transgene ( Figure 6A ) , indicating that FOS-1 associates with the kreg-1 promoter via an interaction with the TRE2 motif . As shown above , self-association of FOS-1 is prevented by KGB-1-mediated phosphorylation . We next addressed whether FOS-1 phosphorylation affects its ability to interact with the TRE2 element of the kreg-1 promoter . Cell extracts obtained from COS-7 cells expressing T7-FOS-1 were incubated with probes and analyzed in a gel-retardation assay . We found that FOS-1 was able to associate with a probe containing the optimal TRE2 sequence , but not with a probe in which the core 6 bases of TRE2 were deleted ( Figure 6B , lanes 1 , 2 , 5 ) . To further confirm the interaction of FOS-1 with the TRE2 element , we utilized site-directed mutagenesis to convert the consensus TGAGTCA sequence to AAGCTTA in the TRE2 element . A similar alteration has been shown to inhibit the AP-1-DNA interaction [24] . Indeed , we observed that FOS-1 was not able to bind to the mutated TRE2 probe ( Figure 6B , lane 6 ) . In addition , the protein-DNA complex was supershifted by pre-incubation with anti-T7 antibody ( Figure 6B , lane 3 ) , indicating that T7-FOS-1 is involved in this complex . When MEK-1 and KGB-1 were co-expressed with T7-FOS-1 in COS-7 cells , the association of FOS-1 with the optimal TRE2 probe was decreased ( Figure 6C , lanes 1–3 ) . This reduction was dependent on the kinase activity of KGB-1 ( Figure 6C , lane 4 ) . Thus , FOS-1 phosphorylation by KGB-1 decreases the association of FOS-1 with its target gene promoter . Taken together , these results suggest that the KGB-1 pathway activates transcription of target genes by phosphorylation of FOS-1 , which inhibits FOS-1 self-association and binding to its target promoter . How does FOS-1 repress kreg-1 transcription ? Given the importance of chromatin dynamics in the control of gene expression , recent work has focused on AP-1 interaction partners capable of chromatin remodeling and modification [13]–[16] , [25] , [26] . It has been reported that AP-1 , during the innate immune response , recruits HDAC1 , a member of the Class I histone deacetylase ( HDAC ) family , to the promoter of a gene that encodes an antibacterial protein where it deacetylates promoter-associated histones [26] . Therefore , we examined whether HDACs might affect Pkreg-1::venus expression . C . elegans possesses three HDAC genes , hda-1 , hda-2 and hda-3 , which encode Class I HDAC homologs [27] , [28] . We found that treatment with hda-1 RNAi resulted in constitutive expression of the Pkreg-1::venus reporter in wild-type animals ( Figure 7A and 7B ) . Furthermore , hda-1 knockdown significantly restored loss of intestinal Pkreg-1::venus expression in kgb-1 ( km21 ) mutants ( Figure 7A and 7B ) . We also found that hda-1 RNAi had little effect on the constitutive expression caused by the Δtre2 deletion of the Pkreg-1::venus reporter ( Figure 7A and 7B ) , indicating that negative regulation of kreg-1 expression by HDA-1 requires the TRE2 motif in the promoter . In addition , we observed by qRT-PCR that hda-1 RNAi enhanced expression of the kreg-2 gene ( Figure S9 ) , confirming that this effect is not specific only to kreg-1 . Next we asked whether FOS-1 could interact with HDA-1 . T7-FOS-1 and FLAG-HDA-1 were co-expressed in HEK293 cells . We immunoprecipitated FLAG-HDA-1 with anti-FLAG antibodies , and probed for the T7-FOS-1 on a Western blot with anti-T7 antibodies . We failed to detect an association between FOS1- and HDA-1 ( Figure 7C , lane 1 ) . However , if we transfected in the Pkreg-1::venus reporter along with T7-FOS-1 and FLAG-HDA-1 , we could detect an association between FOS-1 and HDA-1 ( Figure 7C , lane 3 ) . Furthermore , removal of the TRE2 site from the Pkreg-1::venus reporter reduced this interaction ( Figure 7C , lane 4 ) . These results suggest that HDA-1 and FOS-1 can associate on the kreg-1 promoter . Finally , we examined whether HDA-1 contributes to the response to heavy metal stress . Knockdown of hda-1 by RNAi in wild-type animals had no effect on their sensitivity to Cu2+ ions ( Figure 7D ) . In contrast , knockdown of hda-1 by RNAi suppressed the sensitivity to Cu2+ ions in kgb-1 ( km21 ) mutants . Thus , HDA-1 negatively regulates the heavy metal stress response , consistent with the observation that kreg-1 expression is repressed by HDA-1 .
A key step in understanding the KGB-1 JNK pathway is the identification of downstream targets that are activated by KGB-1 and that perform the actual protective function . Analysis of gene expression comparing wild type and kgb-1 mutants has led to the identification of two targets of the KGB-1 pathway , namely kreg-1 and kreg-2/lys-3 . Both targets are transcriptionally induced by stress , both require the KGB-1 pathway for their full induction , and both are required for protection of the animal against heavy metal stress . These data suggest that activation of the KGB-1 pathway leads to increased production of these proteins that this in turn leads to protection and defense against heavy metal stress . The identity of one of these genes is particularly revealing: The protein encoded by kreg-1 contains polyhistidine stretches , which are well known to bind metal ions ( e . g . Ni2+ , Cu2+ , Co2+ and Zn2+ ) and widely used as an affinity tag [29] . A previous study also revealed that Hpn , a 60 amino acids protein with polyhistidine stretches in Helicobacter pylori , preferentially binds Cu2+ ion and is able to confer copper resistance when expressed in Escherichia coli [30] . Thus , we speculate that the KREG-1 protein may confer resistance to Cu2+ stress by chelating this ion through these polyhistidine stretches . In this study , we identified the FOS-1 bZIP transcription factor as a downstream component of the KGB-1 pathway . FOS-1 was isolated as a protein that binds to KGB-1 and we showed that KGB-1 phosphorylates FOS-1 in the C-terminal regulatory region . Fos and Jun of bZIP transcription factors form part of the AP-1 transcription factor complexes [18] , [23] . These transcription factors are homologous within two adjacent domains: a basic region and a leucine zipper motif , which are necessary for DNA binding and factor dimerization , respectively . Indeed , C . elegans FOS-1 acts as an activator of spermathecal-specific plc-1 gene expression by forming heterodimers with JUN-1 [21] . In addition , a genome-wide RNAi screen identified fos-1 and jun-1 as genes important for the KGB-1-mediated defense pathway against pore-forming toxins made by soil bacterium [8] . Thus , it is likely that the JNK-AP-1 pathway has the role in protection against pore-forming toxins by regulating transcriptional responses . However , we found that JUN-1 is not involved in the KGB-1-mediated stress response pathway . We demonstrated that FOS-1 is capable of forming homodimers and acts as a repressor of its target gene expression . Dimerization of FOS-1 most likely serves to enhance its DNA binding affinity to target promoters and it is therefore likely that the C . elegans FOS-1 binding partner determines whether FOS-1 functions as a repressor or activator . It has been proposed that bZIP transcription factors can switch between repressor and activator mode , as illustrated by the transcriptional regulation of C . elegans ATF-7 and the yeast Sko1p resulting from MAPK activation [22] , [31] . Activation of the PMK-1 p38 MAPK pathway in response to pathogen infection results in PMK-1 phosphorylation of ATF-7 , leading to a switch in ATF-7 from a transcriptional repressor to an activator [22] . In yeast , Sko1p is phosphorylated via the Hog1p MAPK pathway in response to osmotic stress , and this converts Sko1p from a repressor to an activator [31] . Here , we found that depletion of FOS-1 suppressed the heavy metal sensitivity of kgb-1 mutants , but had no effect on the heavy metal sensitivity in wild-type animals . These results strongly suggest that FOS-1 simply acts as a transcriptional repressor of the heavy metal stress response mediated by the KGB-1 pathway . Thus , FOS-1 regulation of the heavy metal stress response does not appear to involve switching of its transcriptional regulation activity . Our analysis showing FOS-1 phosphorylation by KGB-1 and its biological consequences has provided some novel molecular insights into the regulation of FOS-1 . We found that phosphorylation blocks FOS-1 dimer formation and that this results in reduced binding to the promoter of target genes . We imagine that dimeric FOS-1 binds DNA with a higher affinity than the monomeric form . Based on these data , we propose that activation of the KGB-1 pathway in response to heavy metal stress results in FOS-1 phosphorylation , leading to a switch of FOS-1 from dimer to monomer and consequent loss of promoter binding ( Figure 8 ) . How does FOS-1 act as a repressor of kreg-1 transcription ? Our results suggest that the HDA-1 histone deacetylase co-operates with FOS-1 to repress transcription of the kreg-1 gene ( Figure 8 ) . Many transcription factors have been shown to recruit protein complexes that locally alter the acetylation of histones . Recruitment of HDAC can lead to transcriptional repression , whereas recruitment of histone acetyltransferase can lead to transcriptional activation . These results suggest that FOS-1 acts as a transcriptional repressor by recruiting HDA-1 to the promoter of the kreg-1 gene . Therefore , it is quite likely that KGB-1 activates kreg-1 expression by derepressing this FOS-1/HDA-1 repressor complex ( Figure 8 ) . In this model , FOS-1 forms homodimers and binds to the TRE2 motif in the kreg-1 promoter . FOS-1 dimerization might also potentiate the recruitment of HDA-1 to the promoter . Thus , the FOS-1/HDA-1 repressor complex may function to prevent inadvertent activation of the kreg genes in the absence of heavy metal stress . When signaled by heavy metal stress , KGB-1 is activated and phosphorylates FOS-1 , which leads to dissociation of the FOS-1 dimer and dissociation of the FOS-1/HDA-1 repressor complex from the kreg-1 promoter , resulting in the activation of kreg-1 expression . The ability of Fos to function as a repressor has also been described in Drosophila [14] . HDAC is recruited to promoters occupied by unphosphorylated DFos and represses transcription of its target genes . JNK-mediated phosphorylation of DFos not only releases the HDAC corepressor complex and leads to activation by derepression but also unmasks the function of histone acetyltransferase and results in increased transcriptional efficiency . However , the mechanism of C . elegans FOS-1 derepression described here represents a unique case where transcription factor phosphorylation leads to reduced dimerization , DNA binding and loss of HDAC association . Comparing Drosophila Fos and C . elegans FOS-1 , we find that significant homology is present only in the adjacent basic and leucine zipper motifs . In addition , the amino acid sequence of the region flanking the phosphorylation sites is not conserved between Drosophila Fos and C . elegans FOS-1 [32] . Nevertheless , the basic mechanisms of JNK-mediated phosphorylation of Fos and its effects on Fos/HDAC repressor complex formation are evolutionally conserved between C . elegans and Drosophila . This finding thus reveals a common underlying mechanism by which the JNK signaling pathway modulates the activities of the Fos family of bZIP transcription factors . In summary , we have described a mechanism of transcriptional regulation whereby KGB-1 activates expression of the stress response genes by promoting the dissociation of a FOS-1/HDA-1 repressor complex . This is a new finding that could provide valuable insights into the stress response in the context of the whole organism . It would greatly enhance our understanding of the stress response mediated by JNK signaling to elucidate how the kreg genes confer tolerance to heavy metals in C . elegans .
The yeast expression vector for the LexA DNA-binding domain ( DBD ) -fused KGB-1 ( K67R ) was constructed by inserting each coding sequence into pBTM116 . The mammalian expression vectors for HA epitope-tagged KGB-1 ( HA-KGB-1 ) and FLAG epitope-tagged MEK-1 ( FLAG-MEK-1 ) were described previously [5] . The cDNA for fos-1 was isolated by the Y . Kohara EST project ( National Institute of Genetics , Mishima , Japan ) . The cDNAs for hda-1 and human Grhl2 were amplified by PCR from C . elegans and human cDNA libraries , respectively , and completely sequenced . The mammalian expression constructs for T7-FOS-1 , GFP-FLAG-FOS-1 , FLAG-HDA-1 and T7-hGrhl2 were constructed by inserting each coding sequence into a vector expressing epitope-tagged protein under the control of the cytomegalovirus ( CMV ) promoter . Each coding sequence was amplified by PCR using primer sets to create restriction sites immediately before the first codon and after the stop codon . Mutated forms of FOS-1 were made by oligonucleotide-directed PCR and the mutations were verified by DNA sequencing . To construct the Phsp-16::t7::fos-1 plasmids , each t7::fos-1 fragment from the mammalian expression vectors for T7-FOS-1 was subcloned into the pPD49 . 78 vector . Gateway cloning technology ( Invitrogen ) was used to construct the Pkreg-1::venus plasmid for expression in animals . The Pkreg-1::venus plasmid was constructed by fusion of the venus coding sequence to a 2 . 8 kbp genomic fragment containing the kreg-1 promoter . Deletions of Pkreg-1::venus were made by oligonucleotide-directed PCR and the deletions were verified by DNA sequencing . The Pelt-2::mek-1::venus plasmid was constructed by fusing three DNA fragments in the following order: a 2 . 9 kbp genomic fragment containing the elt-2 promoter , the mek-1 coding sequence , and the venus coding sequence . The Pmek-1::mek-1::venus , Pttx-3::gfp and sur-5::gfp plasmids were described previously [6] , [33] , [34] . Anti-phospho-FOS-1 rabbit polyclonal antibody was raised against a synthetic phospho-polypeptide , CSNTGL ( P ) TPSGQP [ ( p ) , phosphorylated] , which corresponds to the C-terminal portion of FOS-1 and affinity purified . Anti-HA monoclonal antibody 16B12 ( Covance ) , anti-FLAG monoclonal antibody M2 ( Sigma ) , anti-T7 monoclonal antibody ( Novagen ) and anti-GFP polyclonal antibody ( Clontech ) were used . All strains were maintained on nematode growth medium ( NGM ) plates at 20°C and fed with bacteria of the OP50 strain , as described [35] . The alleles used in this study were N2 Bristol as the wild type , kgb-1 ( km21 ) , mek-1 ( ks54 ) , atf-7 ( qd22 ) , and eri-1 ( mg366 ) . Strains carrying the Phsp-16::t7::fos-1 transgene were generated by injecting this DNA together with the sur-5::gfp plasmid , which expresses GFP in the nuclei of most somatic cells from embryogenesis , into the gonads of young adult N2 animals as described [36] . Strains carrying the Pkreg-1::venus transgene were generated by injecting this DNA together with the Pttx-3::gfp plasmid , which expresses GFP in a pair of AIY interneurons , into the gonads of young adult N2 animals . Assays for the effect of fos-1 transgenes on heavy metal toxicity were carried out as follows . Animals were grown and allowed to lay eggs on NGM plates seeded with bacteria of the OP50 strain . Embryos expressing GFP were transferred to NGM plates containing the indicated concentrations of copper sulfate . After incubation for 1 day at 20°C , the numbers of hatched embryos were determined by counting unhatched embryos . After additional incubation for 3 days either at 20°C or 33°C for 1 hour twice a day , the animals that developed into adulthood were counted . The percentage of adults was calculated by multiplying the number of adults by 100 and dividing by the number of hatched animals . The relative viability was estimated by dividing the percentage of adults in the presence of heavy metals by the percentage of adults in the absence of heavy metals . Assays for the effect of RNAi on heavy metal toxicity were performed as follows . Animals were grown and allowed to lay eggs on NGM plates seeded with bacteria of the OP50 strain . Embryos were transferred to NGM plates containing the indicated concentrations of copper sulfate and seeded with bacteria of the HT115 strain carrying plasmids expressing the respective double-stranded RNAs for fos-1 , kreg-1 , kreg-2 , jun-1 or hda-1 . After incubation for 1 day at 20°C , the numbers of hatched embryos were determined by counting unhatched embryos . The animals that developed into adulthood were counted 4 days after egg laying . The relative viability was estimated as described above . Adult worms of each strain were incubated with H2O or 1 mM copper sulfate for 1 hour . Total RNA was then prepared using Trizol reagent ( Invitrogen ) , followed by DNase I treatment , phenol/chloroform extraction and ethanol precipitation . RNA was dissolved in water and used as a template for a genome-wide microarray analysis and real-time qRT-PCR . Affymetrix GeneChip microarray processing was performed once by Takara Bio Inc . according to the manufacturer's protocol ( Affymetrix ) . Briefly , total RNA was prepared from wild-type and kgb-1 mutant animals subjected to Cu2+ ion exposure or left untreated ( control ) . Biotinylated cRNA was hybridized to Affymetrix Genechips containing probes against 22 , 500 transcripts . qRT-PCR was performed with a 7300 real-time RT-PCR system ( Applied Biosystems ) using SYBR Premix Ex Taq ( Takara ) . A standard curve was generated from diluted RNA derived from wild-type animals , and levels of gene expression were normalized to act-1 expression . The microarray results were used as an initial screen to identify genes whose expression was increased in response to Cu2+ ions and in a manner dependent on KGB-1 . We selected target genes by the following process ( Figure S4A ) . First , transcript expression levels were compared between animals with or without Cu2+ treatment ( Tables S2 and S3 ) . 334 genes were chosen that were up-regulated greater than 2-fold by Cu2+ in wild-type animals ( Table S4 ) . Second , we compared Cu2+-mediated gene induction in wild-type versus kgb-1 mutant animals to identify genes whose induction was affected by kgb-1 . We identified 66 genes whose induction by Cu2+ in kgb-1 mutants was <50% of the induction seen in wild-type animals ( Table S5 ) . Third , we compared basal expression levels between wild-type and kgb-1 mutant animals , since basal activity of KGB-1 can be detected in wild-type animals [5] , [6] . We identified 50 genes whose basal expression was decreased or not changed in kgb-1 mutants versus wild-type animals ( Table S6 ) . Finally , data were manually curated to remove genes no longer predicted to be expressed using data available in Wormbase . From this we chose the top 13 genes whose expression was significantly induced by Cu2+ in wild-type animals ( Table S7 ) . We then re-examined regulation of these genes in a more quantitative manner by qRT-PCR ( Figure S4B ) . From this we obtained a final list of 6 genes whose regulation was reproducibly affected by kgb-1 ( Table S8 ) . Microarray data for the Cu2+-treated/non-treated wild-type animals and Cu2+-treated/non-treated kgb-1 mutant animals have been deposited in NCBI-GEO with the accession numbers GSE42703 . The following links have been created to allow review of records GSE42703: http://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE42703 Wild-type and kgb-1 mutant animals harboring the Pkreg-1::venus transgene as an extrachromosomal array were cultured on plates seeded with a bacteria strain expressing the respective double-stranded RNAs for vhp-1 , fos-1 , jun-1 , atf-7 or hda-1 . At 3 days after hatching , animals were treated with copper sulfate ( 1 mM ) for 1 hour . These animals were then transferred to NGM plates and incubated for 3 hours . The percentages of animals in each expression category are listed . “Weak” refers to animals in which intestinal Venus was present at low levels . “Strong” indicates that Venus was present at high levels in most of the intestine . For phosphatase treatment , cell lysates were incubated with or without calf intestinal alkaline phosphatase ( NEB ) at 36°C for 5 minutes . Immunoprecipitation from COS-7 cells was carried out as described previously [37] . For immunoprecipitation from HEK293 cells , cells were pretreated with 1% paraformaldehyde in PBS for 10 minutes and glycine at a final concentration of 0 . 125 M for 5 minutes and collected . The ChIP assay was performed using ChIP-IT Express Enzymatic Shearing ( Active Motif ) according to the manufacturer's instructions . In brief , the soluble chromatin extracts were prepared from 2×108 HEK293 cells , and immunoprecipitated with anti-T7 monoclonal antibodies and protein G magnetic beads ( VERITAS ) overnight . The immunoprecipitated DNA-histone complexes were incubated overnight at 65°C to reverse cross-linking and then treated with RNase A and protease K . Purified DNA fragments were subjected to quantitative PCR . Transfected COS-7 cells were lysed in lysis buffer containing 20 mM HEPES ( pH 7 . 4 ) , 150 mM NaCl , 12 . 5 mM β-glycerophosphate , 1 . 5 mM MgCl2 , 2 mM EGTA , 10 mM NaF , 2 mM dithiothreitol , 1 mM Na3VO4 , 1 mM phenylmethylsulfonyl fluoride , 100 units/ml aprotinin , 0 . 5% Triton X-100 . Binding reactions were performed at room temperature for 30 minutes by incubating cell extracts and Cy5 . 5-labeled retardation probes in binding buffer containing 25 mM Tris ( pH7 . 9 ) , 250 mM KCl , 1 mM EDTA , 5% glycerol , 1 mM dithiothreitol , 0 . 25 mg/ml BSA , 0 . 1% Triton X-100 and 0 . 1 µg/ml of poly ( dI ) •poly ( dC ) . The samples were analyzed on 3–12% polyacrylamide gels . For supershift experiments , anti-T7 antibodies or normal mouse IgG ( Santa Cruz ) ( 1 µg per lane ) were added in the binding reactions . The sequences of the gel retardation probes are as follows: TRE2 probe , 5′-AATTGCTGAGTCACAGACAT-3′; mutated TRE2 probe , 5′-AATTGCAAGCTTACAGACAT-3′; probe deleting the core 6 bases of TRE2 , 5′-AAATAATTGCCAGACATTAC-3′ . TRE2 and mutated TRE2 are underlined . The LexA DBD-KGB-1 ( K67R ) plasmid was used as bait to screen the Caenorhabditis elegans cDNA library in pACTII [38] . The bait plasmid and the library cDNAs were co-transformed into the Saccharomyces cerevisiae reporter strain L40 [MATa , trp1 , leu2 , his3 , LYS2:: ( lexAop ) 4-HIS3 , URA3:: ( lexAop ) 8-LacZ] . Yeast cells were plated onto a synthetic medium plate lacking histidine and containing 3-amino triazole , and allowed to grow at 30°C . Transformants grown on selective medium plates were then streaked on selective medium plates again . Plasmids were collected from colonies that grew on selective medium plates and subjected to DNA sequencing . | We have investigated the mechanisms by which the soil nematode Caenorhabditis elegans regulates its response to environmental stresses . Previously , we showed that a conserved KGB-1 JNK mitogen-activated protein kinase ( MAPK ) pathway regulates the response to heavy metal stress in C . elegans , illustrating that these key stress signaling pathways in mammals are also conserved in C . elegans . Various proteins have been identified as potential targets of JNK MAPK in mammals; however , the identification of JNK MAPK physiological substrates has proven more challenging in C . elegans . Here we demonstrate that Fos , a bZIP family transcription factor , and HDAC , a Class I histone deacetylase , are crucial components functioning downstream of KGB-1 in the JNK-mediated stress response pathway . We find that FOS-1 functions as a transcriptional repressor by recruiting HDAC to its target promoters . In response to stress , activated KGB-1 relieves this repression via phosphorylation of FOS-1 . Thus , this study shows how JNK signaling induces tolerance to stress at the gene level and describes a novel mechanism of gene regulation by which this is effected . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"model",
"organisms",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2013 | The Caenorhabditis elegans JNK Signaling Pathway Activates Expression of Stress Response Genes by Derepressing the Fos/HDAC Repressor Complex |
The HERC gene family encodes proteins with two characteristic domains: HECT and RCC1-like . Proteins with HECT domains have been described to function as ubiquitin ligases , and those that contain RCC1-like domains have been reported to function as GTPases regulators . These two activities are essential in a number of important cellular processes such as cell cycle , cell signaling , and membrane trafficking . Mutations affecting these domains have been found associated with retinitis pigmentosa , amyotrophic lateral sclerosis , and cancer . In humans , six HERC genes have been reported which encode two subgroups of HERC proteins: large ( HERC1-2 ) and small ( HERC3-6 ) . The giant HERC1 protein was the first to be identified . It has been involved in membrane trafficking and cell proliferation/growth through its interactions with clathrin , M2-pyruvate kinase , and TSC2 proteins . Mutations affecting other members of the HERC family have been found to be associated with sterility and growth retardation . Here , we report the characterization of a recessive mutation named tambaleante , which causes progressive Purkinje cell degeneration leading to severe ataxia with reduced growth and lifespan in homozygous mice aged over two months . We mapped this mutation in mouse chromosome 9 and then performed positional cloning . We found a G⇔A transition at position 1448 , causing a Gly to Glu substitution ( Gly483Glu ) in the highly conserved N-terminal RCC1-like domain of the HERC1 protein . Successful transgenic rescue , with either a mouse BAC containing the normal copy of Herc1 or with the human HERC1 cDNA , validated our findings . Histological and biochemical studies revealed extensive autophagy associated with an increase of the mutant protein level and a decrease of mTOR activity . Our observations concerning this first mutation in the Herc1 gene contribute to the functional annotation of the encoded E3 ubiquitin ligase and underline the crucial and unexpected role of this protein in Purkinje cell physiology .
The cerebellum plays the role of a coordination centre , integrating peripheral sensory information on movement and position of the body parts to fine-tune gait and balance . Structural or functional alterations of this part of the central nervous system result in a complex syndrome , called ataxia , which is characterized by neurological signs that are clinically obvious in most species including the mouse . Many such mutations , either of spontaneous origin or resulting from strategies of genetic engineering performed in vitro , have been studied in detail in this species that , synergistically with human studies , have allowed advancement of our understanding of the developmental mechanisms generating the uniquely complex mature cerebellum . In this publication , we report the positional cloning of an autosomal recessive mouse mutation , called tambaleante ( symbol tbl; meaning staggering in Spanish ) , which is precisely characterized by a severe ataxic syndrome [1] , [2] . Mice homozygous for this mutation ( tbl/tbl ) exhibit an unstable gait , abnormal hindlimb posture and tremor . All these phenotypic characteristics correlate with a progressive degeneration of Purkinje cells ( PCs ) starting by two months of age . tbl mice thus represent a model of recessively inherited ataxia with progressive neurodegeneration of PCs . Using a combination of genetic , histological and biochemical approaches , we have been able to characterize the pathology of this mutation that we could relate to a mutation in the gene encoding the E3 ubiquitin ligase HERC1 .
The tambaleante ( tbl ) mutation arose spontaneously in the DW/JPas inbred substrain , at the Institut Pasteur , and appeared to be inherited as an autosomal recessive condition with complete penetrance . The most remarkable phenotypic feature of homozygous ( tbl/tbl ) mice was an unstable gait , with abnormal hind limb-clasping reflex , which became really obvious from two months of age and worsened with time ( Figure 1A and Video S1 ) . To quantify these observations , we performed rotarod assays with these animals . Figure 1B shows that tbl animals stayed less time on the rotarod without falling . To visualize the progressive degeneration of PC , we performed an analysis of cerebellum sections stained with haematoxylin and eosin ( H&E ) . In Figure 1C–1F , we can observe the great loss of PC between 1–3 months in tbl animals . Immunostaining using anti-calbindin D28-k antibodies ( Figure 1G–1J ) of parasagital sections of mouse cortex of 4 month old shows that tbl mice is almost completely depleted of PC . Compared to their normal littermates , tbl/tbl homozygotes were smaller in size . Growth curves showed that the weight of the mutant animals was significantly and constantly lower than the weight of controls , varying from 15 to 30% according to age and gender ( Figure 2 ) . Mutant animals also showed a lower survival rate since less than 40 percent of the latter survived longer than 40 weeks on the original DW background ( Figure 2 ) . Both sexes appeared to be fertile although poor breeders . Genotyping 30 F2 mutant offspring ( 60 meiotic events ) of an inter-subspecific cross between DW-tbl/tbl males and wild type ( +/+ ) females of the inbred strain MBT/Pas [3] , allowed us to assign the locus for tbl to chromosome 9 , within a 1 . 7 cM interval flanked by markers D9Mit233 and D9Mit165 ( Figure 3A and 3B ) . Although this interval encompassed the locus of the staggerer ( Rorasg ) mutation , which occurred in the gene encoding RAR-related orphan receptor alpha and is also characterized by a severe cell-autonomous defect of Purkinje cell [4] , we could eliminate this gene as causative of the tambaleante phenotype through the finding and characterization of a recombination event between the loci for tbl and the one of Rora ( Figure 3A ) . In addition , a complementation test performed by mating tbl/tbl mice to +/Rorasg mice and yielding exclusively normal offspring confirmed non allelism of the two mutations ( data not shown ) . Finally , the tbl candidate region was reduced to a genomic segment of 0 . 6 cM ( ∼0 . 98 Mb ) between D9Mit233 ( 65 . 97 Mb ) and D9Mit302 ( 66 . 95 Mb ) , which contains eleven known genes as indicated in the Ensembl sequence database ( http://www . ensembl . org/Mus_musculus ) ( Figure 3A and 3B ) . Among the eleven genes that were identified in the interval , three candidates ( Herc1; Usp3 and Rab8b ) appeared top ranked considering their expression profile and the known or putative functions of the encoded proteins ( Figure 3B ) . Among these three candidates , Herc1 seemed to be the most likely one considering its large size and its sequence homology with Herc2 , a locus where mutations that have phenotypes similar to tambaleante have already been reported [5] , [6] . Sequence analysis of the ∼15 kb cDNA corresponding to this gene and comparison with the sequences of the co-isogenic strain DW , allowed to identify a single nucleotide difference ( a G1448A transition at ENSMUST00000042824 ) between the tbl and the normal ( + ) haplotype that resulted in a Gly483Glu substitution ( Figure 3C ) . Sequence comparison with several other unrelated inbred strains confirmed that this sequence alteration was recent and unique to the mutant haplotype . The G⇔A transition in the tambaleante haplotype generated a new restriction site for the MboII enzyme that allowed us to design a PCR assay , helpful for the diagnostic of tbl haplotype by discriminating +/+ from +/tbl or tbl/tbl ( Figure 4 ) . To ascertain that the Gly483Glu substitution identified in Herc1 was really causative of the abnormal phenotype observed in tbl/tbl mice , we decided to attempt the rescue of the tambaleante phenotype by crossing tbl/tbl mice with transgenic mice expressing a normal copy of the Herc1 gene . We used two strategies to generate such transgenic mice: in the first case we used the mouse BAC clone RP23-355L9 ( ∼160 kb - encompassing the Herc1 locus and a gene encoding a leucine-rich repeat protein 22 ( Fbxl22 ) that is not expressed in the brain [7] ( Figure 3B ) ) ; then we used the human cDNA of HERC1 ( 96% amino acid identity with mouse ) [8] . BAC transgenic ( TgRP23-355L9/+ ) mice were crossed with heterozygous ( +/tbl ) mice , then the F1 mice were mated with heterozygous ( +/tbl ) mice to obtain tbl homozygous mice with the transgenic copies ( tbl/tbl; TgRP23-355L9/+ ) . We found that the phenotype of these mice was greatly improved since none of the animals exhibited the phenotypic characteristics of tambaleante mutants during the period they were observed ( as an example see footprint experiments in Figure S1 ) . By 3 months of age , PCs in these transgenic mice remained normal at least in number and size , indicating complete phenotypic rescue ( Figure S1 ) . Complete transgenic rescue was also achieved with the other transgenic strain generated from the human HERC1 cDNA [6] in pCI-neo ( TgHERC1cDNA ) ( Figure S2 ) . The complete phenotypic rescue was also analyzed by weight , rotarod performance , cerebellar staining with H&E and immunohistochemistry with anti-calbindin D28-k antibodies ( Figure 5 ) . No differences in the parameters analyzed were observed between wild-type animals and transgenic mice . From these results we considered that the pathology in tambaleante mice is indeed a direct consequence of the molecular defect in Herc1 . For this reason , we use Herc1tbl as an official symbol for the Herc1 mutant allele . Analysis of the Herc1 gene expression by Northern blotting has been previously reported and revealed ubiquitous expression in mammalian tissues although at very low levels in the liver [8] . We have confirmed these data by RT-PCR in mouse tissues using specific primers for Herc1 ( Figure 6 ) . In situ hybridization of brain sections also confirmed a pattern similar to that shown in the Allen Brain Atlas [9 , and data not shown] . The HERC proteins have a HECT ( Homologous to E6-AP COOH-terminus ) domain and at least one domain related with RCC1 ( Regulator of Chromosome Condensation 1 ) . HECT domains are involved in the transfer of ubiquitin or ubiquitin-like proteins to target substrates . RCC1-like domains ( RLD ) seem more versatile and may have a role in guanine nucleotide exchange on small GTP-binding proteins , in enzyme inhibition and in interaction with proteins and lipids . Proteins containing some of these domains are important regulators of cellular processes such as cell cycle , cell signalling and membrane trafficking . HERC1 protein was the first protein of this family to be identified . It contains one HECT domain , two RLD ( RLD1 and RLD2 ) , seven WD40 repeats and one SPRY domain ( Figure 3D and 3E ) [10]–[12] . The Gly483Glu substitution found in tambaleante mice is located within the highly conserved N-terminal RCC1-like domain ( RLD1 ) of the HERC1 protein and presumably alters its structure and function ( Figure 3D and 3E ) [10] . To check whether Herc1tbl mutation affected the HERC1 protein levels , we performed Western-blot analysis using anti-HERC1 antibodies with samples from brain , cerebellum and kidney of mice older than 3 months . Surprisingly , we observed a significant increase in the amount of HERC1 protein in mutant mice , suggesting a possible increase of its stability ( Figure 7A and 7D ) . No changes were observed with any other protein such as clathrin heavy chain ( CHC ) that was used as loading control ( Figure 7A ) . Similar results were also found in skeletal muscle ( not shown ) . HERC1 has been previously reported to interact with TSC2 protein . TSC1 and TSC2 are tumour-suppressor genes that are mutated in the tumour syndrome TSC ( tuberous sclerosis complex ) . Their gene products form the TSC1-TSC2 complex ( also named hamartin-tuberin complex ) and , through its GAP ( GTPase-activating protein ) activity towards the small G-protein Rheb ( Ras homolog enriched in brain ) , this complex is a negative regulator of mTORC1 ( mammalian target of rapamycin complex 1 ) [13] . Because processes such as growth , autophagy and neuronal plasticity are known to be regulated , at least in part , through the mTOR pathway [14] , [15] and because Herc1tbl/Herc1tbl mice show features related to these processes ( neuronal degeneration and smaller size ) , we hypothesized that a mutant HERC1 protein might affect some of these events through deregulation of the mTOR pathway . With this guess in mind , we first analyzed whether the levels of the TSC1-TSC2 complex and mTOR protein were modified in the brain of Herc1tbl/Herc1tbl mice . We found that the levels of these proteins remained normal in these animals suggesting that the mutant HERC1 protein does not affect their stability ( Figure 7B ) . We then checked whether the mTOR activity was modified , and to achieve this , we analyzed the phosphorylation of a substrate of the mTORC1 activity , the ribosomal protein S6 kinase 1 ( S6K1 ) . We observed a decrease of the phosphorylation of S6K1 at threonine 389 ( P-T389-S6K1 ) in brain of Herc1tbl/Herc1tbl mutant mice ( Figure 7C ) . These data were also confirmed and quantified in kidney where P-T389-S6K1 levels were found higher compared to +/+ mice ( Figure 7C ) . Because it had been previously reported that mTOR negatively regulates autophagy [16] , we thought that the decrease of the mTORC1 kinase activity could correlate with an increase of autophagy in tbl/tbl mice . The conversion of the microtubule-associated protein light chain 3 ( LC3-I ) to its phosphatidylethanolamine-modified form ( LC3-II ) has been used as marker of the accumulation of autophagosomes [17] . We have measured by immunoblot analysis this autophagy marker observing an increase of LC3-II levels in brain and cerebellum of tambaleante mice ( Figure 7B and 7C ) . To check that this increase was due to increases in autophagic activity and not to reduced turnover of autophagosomes [18] , we also measured the steady-state levels of the known substrate for autophagy p62/SQSTM1 [19] . We observed a significant decrease in the steady-state levels of p62/SQSTM1 ( Figure 7C ) indicating that autophagic flux was not blocked . Altogether these data suggest that autophagy is induced in tbl/tbl mice . An attractive hypothesis would be that this activation is the cause of PC death in the tambaleante mice . This however is difficult to ascertain because an increase in autophagy is also a protective mechanism for cells in response to stressing stimuli . To determine whether this was the case or not , we analyzed the occurrence of autophagy at an earlier phase of the PC degeneration; in 2-month-old Herc1tbl/Herc1tbl cerebellum . Double labeling with antibodies against anti-calbindin D28-k and lysosome-associated membrane protein LAMP-1 allowed us to identify lysosomes with large cytoplasmic accumulations in dendrites and somata being particularly numerous in regions of the cerebellar cortex which had lost the calbindin D28-k expression ( Figure 7E ) . Electron microscopy also performed on 2-month-old mice showed that PC bodies contained numerous autophagosomes ( vacuoles with double membrane and filled with cytoplasmic organelles ) and autolysosomes ( vacuoles with a single membrane and filled with larger inclusions ) ( Figure 7F–7H ) . Altogether these data show that activation of autophagy in Herc1tbl/Herc1tbl mutant mice is indeed the earliest pathological process that seems to be involved in the degeneration and death of PCs . This high degree of autophagy is unique to the Herc1tbl mutation and has never been reported before for any other cerebellar mutation [20] .
Genes of the Herc/HERC family are absent in prokaryotes and in eukaryotes such as fungi and plants . In mammalian genomes there are several HERC paralogous copies encoding two subgroups of proteins: large ( HERC1–2 ) and small ( HERC3–6 in human; HERC3–5 in mouse ) . The HERC1 giant protein , which was the first to be identified in a screening looking for human oncogenes , contains several domains ( one HECT , two RLDs , seven WD40 repeats and one SPRY ) involved in protein ubiquitilation , guanine nucleotide exchange and protein-protein or protein-lipid interaction . This structure probably reflects the multiple interactions of this protein with other cellular proteins . HERC1 interacts with phosphoinositides and with several other proteins such as clathrin , ADP-ribosylation factor ( ARF ) , M2-pyruvate kinase and TSC2 and , through these interactions it has been involved in membrane trafficking , cell growth and proliferation [8] , [10]–[12] . The Gly483Glu substitution that we found in tambaleante mice , which is located within the highly conserved RLD1 domain ( Figure 3D–3E ) , presumably alters the structure of the HERC1 protein and very likely impairs its functions as well [10]–[11] . The structural alteration might be causative of an increase in its stability , leading to the observation of an unexpected increased level of this protein in all studied tissues ( Figure 7A , 7B , and 7D ) . Impairment of HERC1 function through the mTOR pathway could explain the neuronal degeneration and the smaller size observed in tambaleante mice . Since mTOR has been reported to be a negative regulator of autophagy [16] , a decrease of its activity would indeed correlate with an increase of autophagy observed in the PCs of tambaleante mice ( Figure 7 ) . Although future studies are required to understand the precise role of HERC1 , we can however conclude that HERC1 has a profound impact in the animal growth and the maintenance of the cerebellum structure . A consensus is emerging among molecular geneticists stressing that a missense mutation , affecting only one site of a multidomain protein , is sometimes of better value for gene annotation than a knockout allele that , in general , suppresses at once the protein . In the case of Herc1tbl the situation may be more complex . If the Gly483Glu amino acid substitution has an effect on the protein structure , then one may expect heterozygous mice to be affected to some extent . This however does not seem to be the case since , as far as we could observe from those Herc1tbl/+ breeders that we kept for more than one year , we never noticed any symptoms in their gait or behaviour that would have been evocative of a pathology of the central nervous system ( CNS ) . We did not conduct any observation at the histological level on these mice but it is not sure that this would have been of great value if we consider that a decrease in PCs number seems to be a common observation in mice heterozygous for most mutations affecting the cerebellum ( nr , Rorasg , Agtpbp1pcd and Relnrl ) [20] . It does not appear that the Herc1tbl mutation has a dominant negative effect on the HERC1 function because transgenic mice could rescue the tambaleante phenotype . Our data seem to indicate that the tambaleante protein is not functional or has acquired a different function to the wild-type protein and that the presence of wild-type protein has a dominant effect . For this reason , heterozygous or rescued mice , where the wild-type HERC1 protein is present , do not exhibit a tambaleante phenotype . Because ataxia is the most apparent feature in tambaleante mice and because this symptom is commonly associated to a cerebellar defect , we focused more on this part of the CNS than on any other in our morphological survey . However , in all cases , the paraffin embedding and serial sectioning after Nissl staining that were achieved on adult mutant CNS , did not allow detection of any obvious lesion outside the cerebellum ( retina was not analyzed ) . Nevertheless , the possibility that some type of alteration could be disclosed using more specialized methods ( immunohistochemistry , electron microscopy ) remains open . In humans , genes encoding proteins with mutations in their RCC1 domains have been found to be involved in several diseases [11] . The best studied is probably the RPGR ( Retinitis pigmentosa GTPase regulator ) gene , which is responsible for 70–80% of the most severe forms of human the X-linked retinitis pigmentosa [16] , and in which more than 200 independent mutations have been identified . Functional studies suggest a role for this protein in microtubule-dependent transport along cilia [21] . Another example is provided by the ALS2 ( amyotrophic lateral sclerosis 2-juvenile ) locus , which encodes for a protein where mutations have been associated to an autosomal recessive form of juvenile-onset amyotrophic lateral sclerosis ( jALS ) [22] , [23] . All mutations found in this gene lead to the production of truncated proteins . Interestingly , truncations affecting its amino-terminus , where the RCC1-like domain is located , lead to jALS with degeneration of upper and lower spinal cord motorneurons , whereas less severe truncations in the protein Alsin lead only to degeneration of the upper motorneurons [24] , [25] . Gene expression of HERC1 has also been reported to be increased in human tumour cell lines [8] and decreased in heroin users with a genetic variation of the opioid receptor [26] . HERC2 , the other member of the large HERC family , has also been found associated with pathologies . Mutations in the mouse Herc2 gene were found to be responsible for the so-called runty , jerky , sterile-syndrome or , in short , rjs-syndrome , also known as jdf-2 ( juvenile development and fertility-2 ) [5] , [6] . The pathogenic mechanisms of this syndrome are not known at the molecular level , but it has been suggested that at least some of its symptoms could be due to pituitary defects . In humans , the HERC2 genomic locus , including several partially duplicated paralogs ( duplicons ) of HERC2 [27] , [28] , corresponds to the chromosomal breakpoint region in deletions that cause the Prader-Willi and Angelman syndromes [29] , [30] although lack of HERC2 protein does not seem to play a role in these syndromes [6] . Recently , it has been reported that a single nucleotide polymorphism in intron 86 of the HERC2 gene determines human blue/brown eye colour by controlling the expression of the neighboring gene OCA2 [31]–[33] . In summary , the present study unambiguously demonstrates that the gene Herc1 is involved in the mutation tambaleante and shows , for the first time , that this gene has a profound impact on growth and maintenance of the cerebellar structure . To our knowledge , no other mutant allele has ever been reported at the Herc1 locus before Herc1tbl . Considering the relative great size of this gene ( 78 exons - with a predicted coding region of 14 , 559 bp ) this is rather surprising and probably means that a majority of the mutations likely to occur at this locus either have no deleterious effects or , most probably , that they are lethal in utero and accordingly remained undetected so far . This is an important difference with the Herc2 locus where at least a dozen mutations have been reported that lead to the rjs/jdf2 syndrome [5] , [6] . This also means that , in spite of ancestral relationships , the two proteins have acquired some specific , non-redundant functions .
The tambaleante mutation is available at the RIKEN BioResource Center , Tsukuba , Japan ( <http://www2 . brc . riken . jp/> - Ref: RBRC00188 ) . The mouse strain transgenic for the mouse BAC clone RP23-355L9 was generated in Kyoto University by direct in ovo injection . The mouse strain transgenic for the full-length human HERC1 cDNA was generated at the Institute Pasteur . The transgene was previously generated in two steps: first , we digested pCIneo vector ( The vector carries the cytomegalovirus ( CMV ) immediate-early enhancer/promoter , from Promega ) with the restriction enzyme BglII and ligated the annealed oligonucleotide GATCTATCGATA generating a new ClaI restriction site in pCIneo vector . Then the full-length human HERC1 cDNA [8] was cloned into this modified pCIneo ( pJLR189 ) . The transgene cassette was linearized by ClaI , purified by agarose gel and microinjected in ovo . All animal experiments were performed following European and Japanese institutional guidelines for animal handling and research . Tail DNA samples were prepared from mice according to [34] and PCR amplification of the exon containing the tambaleante mutation was performed using the primers: 5′-GCTTGTGGTAAAGGCAGCTATGGG-3′ and 5′-CCTCACATGTCCCCACACAC-3′ , yielding a 476 bp product ( PCR settings were: 94°C×5 minutes , 94°C×30 seconds , then annealing at 60°C for 30 seconds , and elongation at 72°C for 30 seconds . Number of cycles: 35 ) . Amplification products were then digested with MboII enzyme and fractionated in a 10% PAGE to distinguish among tbl/+ , tbl/tbl and +/+ mice . Transgenic mice were PCR-genotyped using the primers: 5′-TGGTGGAAATAGTATCCCAC-3′ and 5′-CACGGTCAGTAGTCAGTGTC-3′ , yielding a 588 bp product ( PCR settings: 94°C×5 minutes , 94°C×30 seconds , then annealing at 55°C for 30 seconds , and elongation at 72°C for 45 seconds . Number of cycles: 35 ) . A mouse multiple-tissue and embryo cDNA set ( Mouse MTC Panel I , Clontech ) was used for expression analysis of Herc1 gene . RT-PCR was performed with following primers: 5′-GAAGATGTGGATGCAGCAGA-3′ and 5′- GGTCTGTCCGGTGAAGGATA-3′ for mouse Herc1 cDNA ( 199 bp ) , and Gapdh 5′ and 3′ PCR primers ( 983 bp ) as a control . To assess transgene expression , total RNA was isolated from mouse brains using the Ultraspec RNA Isolation System ( Biotecx ) . 2 µg of total RNA were reverse-transcribed using the cDNA Reverse Transcription kit ( Applied Biosystems ) and random primers . PCR was carried out with primers: 5′-AGTCGACTGGATCCGGTACC-3′ and 5′-AGTCTGGCAACTGTGGTCCT-3′ for the transgene and 5′-ATGGATGACGATATCGCTG-3′ and 5′-ATGAGGTAGTCCGTCAGGA-3′ for the actin control . Mice were perfused transcardially with a fixative containing 4% formaldehyde in 0 . 1 M phosphate buffer after being deeply anesthetized by diethyl ether inhalation . The cerebellum was removed and postfixed in the same fixative for two hours , then embedded in gelatin 8% and subsequently processed for respective histological analyses . Tissue samples were stored at −80°C and cut on a cryostat . Cryosections were stained with H&E or immunostained with rabbit polyclonal anti-calbindin D28k or anti-LAMP-1 antibodies . For immunostaining , the cerebellum was processed according to [35] . The rotarod test was used to assess motor coordination and function ( rotarod apparatus: ROTAROD/RS Panlab; diameter: 3 . 5 cm , length: 5 cm ) . For Figure 5 , three groups were constituted: +/+ ( n = 8 , four males , four females ) , tbl/tbl ( n = 9 , four males , five females ) , tbl/tbl TgHERC1cDNA ( n = 5 , two males , three females ) . Testing began at 4 weeks of age and was conducted until week 12 to follow the progression on each phenotype . In brief , animals were trained a constant speed ( 16 rpm ) for 60 s . The animals were put on the rotarod until the latency to fall off reached the total time of 60 s . Each mouse was placed on the rotarod with its head in the direction of rotation and so had to turn to the opposite direction . We performed three trials per day with 2–6 min intervals , on three consecutive days . During the pauses between the turns , mice were allowed to rest in their home cages . After training , mice were evaluated once at 6 , 8 and 12 weeks of age at a constant speed of 12 rpm until the latency of fall reached 1 min . The percentage ( % ) of animals that stayed for 60s is shown in Figure 5 . For Figure 1 , each mouse had three trials per test with an intertrial interval of 5 minutes . Mice ( n = 4 to 9 ) were placed on the rotating drum at 20 rpm and the time the animal stayed on the rotarod without falling off was measured . The hind limb clasping reflex was assessed by holding the mouse by its tail for 30 seconds . Horseradish peroxidase-coupled secondary antibodies ( Molecular Probes ) ; anti-mTOR and anti-P-T389-S6K1 ( 1A5 ) antibodies ( Cell Signalling Technology ) ; anti-TSC2 ( C-20 ) , anti-S6K1 ( C-18 ) and anti-LAMP-1 ( N-19 ) antibodies ( Santa Cruz Biotechnology , Inc . ) ; anti-CHC antibody ( BD Transduction laboratories ) ; anti-HERC1 antibodies [8]; anti-LC3 antibody ( MBL ) ; anti-calbindin D28k antibody ( Swant , Bellinzona , Switzerland ) ; anti-p62/SQSTM1 antibody ( Abnova ) . Mice ( 3–7 months old ) were euthanized by cervical dislocation . Organs were collected and frozen in liquid nitrogen and stored at −80°C until analysis . Tissues were prepared in lysis buffer ( consisting of 10 mM Tris-HCl , pH 7 . 5 , 100 mM NaCl , 1 . 5 mM MgCl2 , 50 mM NaF , 1 mM sodium vanadate , 1 mM phenylmethylsulfonyl fluoride , 5 mg/ml leupeptin , 5 mg/ml aprotinin , 1 mg/ml pepstatin A , 50 mM β-glycerophosphate , 100 mg/ml benzamidine ) , homogenized in a motor-driven Polytron PT3000 ( Kinematica AG ) incubated in a precooled tube with CHAPS 0 . 3% for 20 minutes , and centrifuged at 13 , 500g for 15 min at 4°C . Total protein levels were measured by BCA ( Pierce ) . Equal amounts of supernatant proteins ( 200µg/lane ) were separated by electrophoresis . To analyze simultaneously in the same SDS/PAGE gel giant proteins such as HERC1 or mTOR and small proteins such as LC3 , lysates were loaded in a combination of SDS/PAGE gels named LAG gel [36] . After running the gel overnight , the proteins were transferred to PVDF membranes and visualized by immunoblotting using specific antibodies as previously described [36] . Band intensities were analyzed with a gel documentation system ( LAS-3000 Fujifilm ) . The protein levels were normalized with respect to CHC , mTOR or S6K1 levels and expressed as percentage of controls . | The cerebellum is a coordination center whose function is to fine-tune vertebrates' gait and balance; and for this reason , alterations or damage affecting this structure result in a complex syndrome , called ataxia , with neurological signs that are easily recognized . In the mouse , many mutations producing ataxia have been identified and characterized . They have contributed to a better understanding of the genetics of cerebellum development , physiology , and pathology . The present study identifies the recessive allele responsible for the progressive and massive degeneration of the Purkinje cells observed in mutant mice previously named tambaleante . The mutation leads to a single amino acid substitution in a highly conserved domain ( RCC1-like ) of the giant protein HERC1 . This protein belongs to the families HECT ( E3 ubiquitin ligases ) and RCC1 ( GTPases regulators ) . While a variety of mutations have been reported in several members of these families , leading to sterility , growth retardation , retinitis pigmentosa , amyotrophic lateral sclerosis , or cancer , no mutation had ever been reported so far in the mouse Herc1 gene . This report demonstrates the crucial and unexpected role of HERC1 in Purkinje cell physiology that could be considered helpful in the development of new therapeutic strategies for neurodegenerative disorders . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/animal",
"genetics",
"biochemistry",
"cell",
"biology/cell",
"signaling",
"neurological",
"disorders/movement",
"disorders",
"genetics",
"and",
"genomics/disease",
"models",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/gene",
"function",
"neuroscience",
"molecular",
"biology",
"cell",
"biology/gene",
"expression"
] | 2009 | Progressive Purkinje Cell Degeneration in tambaleante Mutant Mice Is a Consequence of a Missense Mutation in HERC1 E3 Ubiquitin Ligase |
Interferons ( IFNs ) are a group of cytokines with a well-established antiviral function . They can be induced by viral infection , are secreted and bind to specific receptors on the same or neighbouring cells to activate the expression of hundreds of IFN stimulated genes ( ISGs ) with antiviral function . Type I IFN has been known for more than half a century . However , more recently , type III IFN ( IFNλ , IL-28/29 ) was shown to play a similar role and to be particularly important at epithelial surfaces . Here we show that airway epithelia , the primary target of influenza A virus , produce both IFN I and III upon infection , and that induction of both depends on the RIG-I/MAVS pathway . While IRF3 is generally regarded as the transcription factor required for initiation of IFN transcription and the so-called “priming loop” , we find that IRF3 deficiency has little impact on IFN expression . In contrast , lack of IRF7 reduced IFN production significantly , and only IRF3−/−IRF7−/− double deficiency completely abolished it . The transcriptional response to influenza infection was largely dependent on IFNs , as it was reduced to a few upregulated genes in epithelia lacking receptors for both type I and III IFN ( IFNAR1−/−IL-28Rα−/− ) . Wild-type epithelia and epithelia deficient in either the type I IFN receptor or the type III IFN receptor exhibit similar transcriptional profiles in response to virus , indicating that none of the induced genes depends selectively on only one IFN system . In chimeric mice , the lack of both IFN I and III signalling in the stromal compartment alone significantly increased the susceptibility to influenza infection . In conclusion , virus infection of airway epithelia induces , via a RIG-I/MAVS/IRF7 dependent pathway , both type I and III IFNs which drive two completely overlapping and redundant amplification loops to upregulate ISGs and protect from influenza infection .
The type I interferon family is a group of cytokines encoded by a single IFNβ gene and a tandem cluster of multiple IFNα genes that were first characterized for their ability to interfere with influenza virus replication [1] and are now recognized as powerful inducers of the host response to viral infections . IFN induction by influenza A virus ( IAV ) depends on recognition of viral components by either cytoplasmic receptors or the Toll-like receptor ( TLR ) system , depending on the infected cell type . While plasmacytoid dendritic cells ( pDC ) use TLR7 to recognize influenza virus , in fibroblasts and conventional DCs IFNβ induction requires recognition of RNA viral genomes by the cytoplasmic RNA helicase retinoic acid-induced gene I ( RIG-I ) [2] , [3] . Upon RNA binding , RIG-I interacts with the mitochondrial adaptor protein MAVS ( also known as IPS-1 , CARDIF and VISA ) and initiates a signalling cascade that culminates in the activation of the transcriptional factors AP-1 , NF-κB and IRF3 , and the expression of IFNβ and IFNα4 in mouse ( or IFNα1 in humans ) [4]–[7] . Once secreted , IFNβ and IFNα4 acts in both a paracrine and autocrine way through binding to the ubiquitously expressed heterodimeric IFNα/β receptor ( IFNAR1/2 ) to induce activation of the receptor–associated tyrosine kinases JAK1 and Tyk2 and subsequent phosphorylation of the transcriptional factors STAT1 and STAT2 [8] . Activated STATs then form transcription factor complexes , including STAT1 homodimers and a STAT1/STAT2/IRF9 heterotrimer known as ISGF3 [9] , and mediate the induction of hundreds of IFN-stimulated genes ( ISGs ) , whose expression determines the establishment of an antiviral state inside the cell . Recently , a novel group of IFNs was described and named type III IFNs . This new IFN family has three members: IFNλ1 ( a pseudogene in mouse ) , IFNλ2 and IFNλ3 , alternatively named IL-29 , IL-28A and IL-28B respectively [10] , [11] . IFNλ induction depends on the same triggers and signalling pathways that regulate type I IFN expression [12] , [13] , with the RIG-I/MAVS/TBK1/IRF3 axis being particularly relevant in mouse embryonic fibroblasts ( MEFs ) upon viral infection [14] . IL-28/29 act through a distinct receptor complex consisting of IL-28Rα , specific for type III IFNs , and the IL-10Rβ chain , which is also part of the receptors for IL-10 , IL-22 and IL-26 . Despite activating different receptors , both type I and III IFNs activate the JAK-STAT signalling pathway , that leads to the formation of the ISGF3 complex [15] and the induction of ISGs . The most important determinant of the different biological activities of IFNα/β and IFNλ is the distribution of their receptors . While the receptor for type I IFNs is expressed on all cells , IL-28Rα is found primarily on epithelial cells of both the respiratory and the gastrointestinal tract [16]–[18] . This finding suggests that type III IFNs may act in a cell-type restricted manner and may selectively contribute to the innate immunity of mucosal surfaces , potential entry sites for many pathogenic viruses . The airway epithelium is a pseudostratified , columnar epithelium consisting of ciliated , basal and secretory goblet cells , that lies at the interface between the host and the environment and provides the first line of defence against inhaled microorganisms . Airway epithelial cells represent the target of many respiratory viruses , including Influenza virus , Adenovirus , Rhinovirus and RSV [19] . As epithelial cells express both cell surface and endosomal pattern recognition receptors ( PRRs ) and intracellular viral sensors , they can promptly detect invading microbes and react by producing cytokines , chemokines and antimicrobial peptides , thus initiating inflammatory and immune responses [20] , [21] . While the PRRs and downstream signals required for influenza ISG induction have been mapped out in detail in other cell types , it is less clear which recognition systems are in action in airway epithelia . Moreover , while the importance of type III IFNs in epithelial responses has been documented , it is unclear whether the signatures induced by interferon type I or III overlap and which , if any , ISG subsets are selectively induced by one or the other . To address these issues , we established cultures of primary differentiated murine tracheal epithelial cells ( MTEC ) and , using a genetic approach together with microarray analysis , investigated the mechanisms that lead to IFN induction in response to influenza A infection . We also assessed the relative contribution of type I and III IFNs to the establishment of an antiviral state by comparing the pattern of influenza-induced gene expression in the absence of either IFNα/β signalling , IFNλ signalling or both . These studies help define the biology and nature of the antiviral state induced by different IFNs in primary cells and determine whether and which genes are still induced by influenza infection in the complete absence of both IFN type I and III signalling .
Primary MTEC were grown to confluence and exposed to air for 14 days , leading to formation of a fully differentiated , polarized epithelium containing ciliated and secretory goblet and Clara cells ( Fig . 1A ) , a cellular composition which closely matches that of airway epithelia in vivo . These differentiated cultures were then infected with IAV , at a multiplicity of infection ( moi ) of 0 . 3 . Intracellular FACS analysis of the infected cultures showed that approximately 10% of MTEC were expressing viral nucleoprotein ( NP ) 24 hours post infection ( hpi ) ( Fig . S1 ) . Total RNA was isolated from five replicate cultures 24 hpi and analyzed using microarrays . We first investigated which type of IFN was induced in response to the infection . To this purpose , after normalizing the signal intensity of each probe in each sample to the median intensity of that probe in the control group , and filtering for genes that were expressed above background level , we searched for the probe sets which represent IFNs in the data set . As shown in Fig . 1B , both type I ( IFNβ , α4 and α5 ) and type III IFNs ( IL-28A/B ) were significantly induced in the infected samples , with IL-28A/B being the most strongly upregulated . The induction of all other IFNα genes did not reach statistical significance . Type II IFN ( IFNγ ) did not pass the initial filtering for genes expressed above background and was not induced upon infection ( not shown ) . To identify the pattern recognition receptors responsible for IFN induction in this experimental model , we infected MTEC cultures from wild-type and knock-out mice and analyzed the expression of both IFNβ and IL-28 by quantitative RT-PCR and ELISA . While in some immune cells TLR7 is the major PRR mediating IFN induction by influenza virus [22] , we found that the absence of TLR7 or its downstream adaptor MyD88 had no impact on the induction of IFNs in influenza-infected epithelia . Similarly , the adaptor TRIF used by TLR3 and TLR4 is not required for IFN induction in epithelia . In contrast , MTEC deficient in MAVS were unable to produce IFNs in response to influenza infection , suggesting involvement of the RIG-I pathway ( Fig . 1C , 1D ) . We next sought to determine whether virus replication is required for the induction of IFNs . Treatment of PR8 at 65°C inactivates the viral polymerase and prevents viral replication , without abrogating virus attachment to the cells ( Fig . S2A , S2B ) . This treatment abolished the virus' ability to induce IFNs ( Fig . 1C , 1D ) indicating that IAV virus access to the cytoplasm and subsequent replication are required to initiate an IFN response in MTEC . To identify the transcripts that were differentially expressed in infected MTEC cultures at 24 hpi , we performed a supervised analysis under stringent conditions ( ≥4-fold change relatively to mock infected samples , t-test unpaired p value <0 . 01 , Benjamini-Hochberg multiple statistical correction ) . The differentially expressed genes were then partitioned by K-means clustering ( Fig . 2A ) into 2 groups ( 177 downregulated genes; 234 upregulated genes ) , and the 234 upregulated genes were hierarchically clustered to generate the heat map in Fig . 2B . Ingenuity Pathway Analysis of the list of upregulated genes confirmed “activation of interferon regulated factors by cytosolic PRR” and “interferon signalling” as two overrepresented pathways in the infected samples ( Fig . 2C ) . Moreover , 80 ( 35% ) of the upregulated genes were recognized as interferon stimulated gene by the INTERFEROME database . A major level of control of IFN production depends on transcriptional regulation . The general paradigm for IFNβ induction involving recruitment of the transcription factors IRF3 and p50/RelA NF-κB has recently been shown to apply also to type III IFN induction , at least in MEFs [14] . Unexpectedly , IRF3 deficient MTEC were not impaired in their ability to upregulate both IL-28 and IFNβ in response to infection ( Fig . 3A ) . In contrast , IRF7−/− epithelia showed a marked reduction in the amount of IFN induced; however , the induction of both IL-28 and IFNβ was completely abolished only in doubly deficient IRF3−/−IRF7−/− MTECs , as assessed by qPCR for gene expression and by ELISA for protein secretion ( Fig . 3B , C ) . Previous studies have demonstrated that the entry of enveloped viruses like HSV and VSV into fibroblast cells can lead to the induction of a subset of ISGs in an IFN-independent manner and that either IRF3 [23] , [24] , IRF7 [25] or IRF1 [26] may have functions that are redundant to that of ISGF3 and therefore induce an IFN-like transcriptome in the absence of IFN signalling . For these reasons , we sought to determine whether in airway epithelia , ISGs could be induced in the absence of IFNs . The data shown so far point to two situations , i . e . the MAVS−/− and the IRF3−/−IRF7−/− epithelia , in which no IFN production could be detected , both at the protein and at the RNA level ( Fig . 1C , 1D , 3B , 4A ) . Surprisingly , in both conditions , infection with influenza virus led to the induction of most of the ISGs tested , albeit at lower levels than in wild-type epithelia ( Fig . 4B and S3 ) . This correlated with virus control as infected wild- type , MAVS and IRF3/7 deficient epithelia had similar virus titers over the course of infection ( Figure 4C ) . These results can be interpreted in two ways; first , following IAV infection , a subset of ISGs may be induced through a MAVS and/or IRF3/IRF7 independent pathway that does not require interferons . Alternatively , in MAVS−/− and IRF3−/−IRF7−/− cells , minute , steady state amounts of IFNs could still be produced and be sufficient to induce ISGs in a context of viral infection . To test these alternative hypotheses , we infected epithelia deficient for either type I , type III or both IFN receptors and analysed their transcriptional response to influenza A virus by microarray analysis . RNA from five replicate samples were first normalized to the median of mock infected samples and then filtered on expression ( 20–100th percentile in at least 50% of samples ) . A supervised analysis under stringent conditions ( ≥4-fold change versus wild-type mock in at least one infected group , 2-way ANOVA , p value of <0 . 01 , Benjamini-Hochberg multiple statistical correction ) and k-means clustering led to a list of 136 upregulated genes and a list of 50 downregulated genes ( not shown ) . The induced genes were then hierarchically clustered to generate the heat map in Fig . 5A . Influenza infection of IFNAR1−/−IL-28Rα−/− double knock-out epithelia induced the expression of IFNβ and IL-28A/B at levels comparable to the wild-type controls even at later time points during an infection ( Fig . 5B ) , indicating that these genes are most likely upregulated directly downstream of the RIG-I/MAVS pathway and do not require IFN-driven positive feed-back on themselves . In contrast , IFNAR1−/−IL-28Rα−/− cells have lost the ability to upregulate many of the genes that were induced in the wild-type control ( Fig . 5A ) , including known ISGs such as Rsad2 , Oasl2 and others ( Fig . 5B ) . To analyse more globally the 136 up-regulated genes shown in Fig . 5A , they were further partitioned by K-means clustering into those that were not induced in infected IFNAR1−/−IL-28Rα−/− cells ( 110 “IFN-dependent” genes , Fig . S4C ) and those that were still induced ( 26 “IFN-independent” genes , Fig . S4A , B ) : analysis by the INTERFEROME database scored 58 ( 53% ) of the “IFN-dependent” genes as ISGs . The “IFN-independent” genes comprise a smaller group of genes , including many chemokines ( Fig . S4B ) . Although 5 ( 23% ) of these genes ( CXCL1 , CXCL3 , CSF-2 , CXCL5 and CD274 ) were identified as ISGs by the INTERFEROME database , it has been described elsewhere that their expression can be also induced independently of IFNs , most likely through regulation by transcription factors like NF-kB , PPARγ and GATA-1 [27] . The mechanism by which they are induced by IAV in our system is currently under investigation . The transcriptional signatures obtained for the single IFNAR1 and IL-28Rα knock-outs were very similar to the one for cells of wild-type origin ( Fig . 5A ) . To address directly whether induction of some ISGs specifically depends on one type of IFN , we filtered the list of 136 infection-induced genes in Fig . 5A for genes that differ between either the wild-type and single knock-outs or between the two single knock-outs ( Fig . S5 ) . Indeed , only 11 genes were found , and in most cases , genes were induced in all three genotypes although at lower intensity in the single knock-outs . Overall , our results indicate that in airway epithelia , the induction of an antiviral state depends on either type I or type III IFN signalling . Both types of IFN independently drive parallel , completely redundant amplification loops , each leading to the induction of the same set of genes . In the absence of both receptors , the IFN signature disappears almost entirely , indicating that no other mechanism can replace the IFN loop for the induction of ISGs . Importantly , the lack of ISG induction in IFNAR1/IL28Rα deficient epithelia has biologic consequences as it leads to significantly higher virus titers at later points during infection ( Figure 5C ) . No significant differences in viral titers were seen between wild-type and either IFNAR1 or IL-28Rα single knock-out . To test in vivo whether lack of IFN responsiveness in lung epithelia impacts the disease course of influenza infection , we generated bone marrow chimeras where B6 wild-type bone marrow was grafted into either wild-type or IFNAR1/IL28Rα deficient hosts . These two groups have both fully functional immune cells but differ in the ability of radioresistant cells including lung epithelia to respond to type I and III IFNs . We confirmed successful immune cell reconstitution by staining blood cells for IFNαβR ( Fig . S6 ) and by testing Sca-1 upregulation on blood cells in response to IFNβ ( not shown ) . Both experiments confirmed that >85% of immune cells in these chimeras have wt phenotype . When these chimeras were infected with the PR8 strain , high susceptibility and mortality was found only in the group lacking IFN receptors on stromal cells and this correlated with higher viral titers ( Fig . 6A , B ) . These results suggest that the ability to respond to IFNs in infected airway epithelia is crucial for successful elimination of the virus from infected animals . Infected MAVS−/− and IRF3−/−IRF7−/− epithelia show ISG induction in the absence of any detectable IFN upregulation , while IFNAR1−/−IL-28Rα−/− cells , which produce but cannot respond to IFNs , did not express ISGs in response to infection . Although type I IFN genes are tightly regulated in response to viral infection , many tissues constitutively secrete low amounts of type I IFN even in the absence of infection ( reviewed in [28] ) . It has been proposed that these constitutive levels of IFNs are required to maintain basal expression of IFN-inducible signalling intermediates ( STAT1/2 , IRF7/9/5 ) and to modulate the relative expression of STAT proteins , therefore “priming” cells for future responses . For these reasons , we sought to determine the basal level of different IFN-signalling intermediates at steady state and upon infection in IFNAR1−/−IL-28Rα−/− cells . Wild-type and double knock-out cells were infected with influenza A and the level of different STAT and IRF molecules analyzed by qPCR . Some of these molecules ( IRF7 , IRF9 , STAT1 , STAT2 ) are known ISGs and were upregulated in wild-type but not in IFNAR1−/−IL-28Rα−/− epithelia at 24 hours post infection . However , the levels of all these transcripts measured relatively to HPRT at steady state were comparable in the two genotypes ( Fig . 7A , 7B ) and in MAVS−/− and IRF3−/−IRF7−/− cells ( not shown ) . To test directly whether residual IFN production is responsible for ISG induction in MAVS and IRF3/7 deficient epithelia , we compared ISG induction in infected MAVS deficient epithelia in the presence or absence of an antibody cocktail blocking both IFNαβ and IFNλ signalling . As shown in Figure 8A , antibody-treated epithelium had further reduced ISG expression compared to the untreated one . To confirm independently the requirement of autocrine signalling by soluble factors for ISG induction , we infected wild-type epithelia in the presence or absence of brefeldin A ( BFA ) . BFA treatment left IFN gene induction unaffected but abolished ISG induction , indicating that soluble factors which include IFNs are required for ISG induction ( Fig . 8B ) . Collectively , these results indicate that the different responses to infection in MAVS−/− and IRF3−/−IRF7−/− compared to IFNAR1−/−IL-28Rα−/− cells can not be ascribed to a lack of priming in the latter , but are most likely due to a residual production of IFN in MAVS−/− and IRF3−/−IRF7−/− cells , that , in the context of infection , is sufficient to ensure ISG induction .
Here , we delineate the influenza-triggered pathways leading to the induction of an antiviral state in primary airway epithelia , the first and most important target tissue of the virus in an infected organism . We show that TLR7 or other TLRs relying on the adaptor molecules MyD88 and TRIF are not involved in the induction of interferons , while the RIG-I/MAVS pathway is crucial for this process . We also show that between the two transcription factors IRF3 and IRF7 implied in IFN induction , IRF3 is of less importance than IRF7 , but complete abolition of influenza-triggered IFN expression is seen only in the absence of both molecules . Most importantly , we show that , upon influenza infection , IFN type I and III independently mediate parallel amplification loops leading to the induction of a completely overlapping set of ISGs , and that this induction is abolished only when none of the two amplification loops are active . The general paradigm for type I IFN induction involves recruitment of transcription factors that are activated by phosphorylation in response to signalling cascades stimulated during viral infection . The IFNβ promoter contains four positive regulatory domains ( PRDI-IV ) , which are occupied by different transcription factors . PRDI and III are binding sites for IRF3 ( early during infection , due to its constitutive expression ) and IRF7 ( with delayed kinetics , due to its inducible expression through an IFN-dependent positive feedback loop ) , while PRDIV and PRDII bind the ATF-2/c-Jun AP-1 and the p50/RelA NF-κB complexes respectively ( reviewed in [8] ) . This initial model of a positive feedback loop , in which IRF3 is primarily responsible for the early induction of IFNβ while IRF7 is required later in the response [29] , was subsequently modified when a study performed on IRF7−/− MEFs revealed that both the early and the late production of type I IFN induced by VSV or EMCV is abolished in the absence of IRF7 [30] . Our results are in line with these observations and identify IRF7 as the major regulator of both type I and type III IFN responses in epithelial cells . Indeed , our data in fig . 6A suggests that the expression level of IRF7 was higher than that of IRF3 at steady state , which would support the notion that IRF7 could act directly downstream of viral recognition to induce IFNs at the earliest stage of infection . Moreover , our data indicates that , even later during an infection , the expression of IFNβ1 and IL-28A/B can be sustained independently of the IFN-driven positive amplification loop ( Fig . 5B ) . IFNλ is preferentially induced by influenza A virus , both in vivo and in vitro [31] . We extend these findings to primary airway epithelia and demonstrate that epithelial-derived IFN type I or type III are sufficient to fuel their respective amplification loop and that no extrinsic IFNs from other cells , for instance immune cells , is required to induce an epithelial IFN signature . The importance of type III IFNs in epithelial responses has been well documented . Several studies have shown that IFNλ protects the epithelium of lung , intestine and vagina from viral infections and that IFNAR1−/−IL-28Rα−/− double deficient mice are more susceptible to viral infection than each single knock-out strain [17] , [18] , [32] . Using chimeric mice with a wild type immune system and either wild type or IFNAR1−/−IL-28Rα−/− double deficient stroma , we show here that IFN unresponsiveness in the stromal cell compartment is sufficient to render mice more susceptible to influenza infection . Previous studies have shown that the promoters of many ISGs have a simple structure and can be easily turned on directly by IRF proteins independently of interferon . In these studies , alternative pathways of direct ISG induction were suggested that rely on either IRF3 [23] , IRF7 [25] , IRF1 [26] or peroxisomal MAVS/IRF1/IRF3 [33] . More recently , the cytosolic exonuclease Trex1 has also been identified as a negative regulator of a novel pathway involving STING , TBK1 , IRF3 and IRF7 that can lead to interferon-independent activation of ISGs [34] . The finding that a subset of ISGs could still be induced in both MAVS- and IRF3/IRF7-deficient epithelia suggested to us that IFN-independent ISG induction may take place here . However , the nearly complete absence of ISG induction in IFNAR1−/−IL-28Rα−/− epithelia led us to conclude that , at least in our experimental model , no other mechanism can efficiently replace the IFN loop for the induction of ISGs . Constitutive low-level signalling of IFNβ ( IFN “priming” ) has been suggested to help preserve IFN responsiveness but also to allow IFN-independent ISG induction [28] , by maintaining the expression of STATs and other signalling intermediates . It could be argued that , unlike IFNAR1−/−IL-28Rα−/− epithelia , MAVS−/− and IRF3−/−IRF7−/− cells still possess this sub-threshold signalling which helps maintain STATs and IRFs at sufficient levels to allow for direct ISG induction upon viral trigger , even in the absence of IFNs . We did however not detect steady-state differences in the expression of a range of IRF and STAT molecules between genotypes and therefore have no evidence that differences in IFN priming contribute to the phenomenon described here . Previous studies that assessed IFN independent ISG induction have mostly relied on IFNAR deficient cells to confirm IFN independence . Here we show that care must be taken to evaluate IFN independence . As IFN type III can stand in for IFN type I in inducing an IFN signature , the analysis of each single receptor knock-out epithelium would have wrongly suggested complete independence of ISG induction from the IFN system . Moreover , while the IFN signature was completely abolished in influenza infected IFNAR1−/−IL-28Rα−/− cells , the addition of neutralizing antibodies against secreted type I and type III IFNs , used in combination on wild-type epithelia or for the complementary IFN on IFN receptor single knock-out epithelia , had little effect on ISG induction ( not shown ) , indicating that even minute concentrations of IFN were still able to induce ISG expression in responsive cells . In vivo studies with single-knock-out mice clearly showed that ISG induction by type III IFN translates into less powerful protection against influenza A virus than ISG induction by type I IFN [32] . At present it is unclear whether slight differences in the kinetics of virus-triggered induction of type I and type III IFN may account for this observation . An alternative explanation is that lung macrophages which are productively infected by most influenza A virus strains and which do not respond well to type III IFN may quickly amplify the incoming virus in the respiratory tract of IFNAR1-deficient mice and thus overwhelm the type III-mediated protection of epithelial cells in such mice . Through the induction of ISGs , the IFN system has potent effects not only to directly combat virus , but also on cell physiology and survival and on the immune response . Therefore , it is considered a very tightly controlled system to avoid excessive inflammation , cell death and tissue damage . On this background , we were surprised to find that in MAVS and IRF3/7 deficient epithelia , IFNs at levels below ELISA or qPCR detection threshold still lead to only slightly reduced ISG upregulation . ISG induction disappeared in MAVS deficient epithelia when IFN type I and III signalling was blocked by an antibody cocktail , indicating that a residual IFN production drives ISG induction in these cells . In apparent contrast to these results , when exogenous IFN was titrated onto epithelial cultures , the minimum amount of IFN required to induce ISG expression was clearly detectable by the ELISA assay ( not shown ) . Possible reasons for this discrepancy could be differences in the bioactivity of endogenous versus recombinant IFNs and differential ability by the ELISAs to detect endogenous or recombinant IFNs . Moreover , biological explanations include local concentrations of endogenous IFNs that may be much higher than those measured in the total supernatant , and the possibility that autocrine IFNs bind to their receptor already in intracellular vesicles and are therefore not measured by ELISA . Overall these data suggest that presence or absence , rather than absolute amounts of IFN , determine the response . The biological sense of such a binary switch could be to respond robustly even to small perturbations of the steady state , which may gain the host precious time when infected by fast-replicating viruses . In the uninfected state , absolute IFN shut-down would be required to avoid chronic “Flu-like symptoms” , a scenario that is in contradiction to the proposed priming effect of sub-threshold IFN levels . One hypothesis for how the vast differences in IFN levels are translated into a largely unaltered IFN signature in MAVS−/− and IRF3−/−IRF7−/− epithelia is that Influenza A infection may render cells much more IFN-sensitive by unknown pathways . To test this hypothesis , we titrated exogenous IFN on uninfected or infected wt or MAVS-deficient epithelia and measured ISG induction . No increased IFN responsiveness was found in infected versus uninfected epithelia , suggesting that there is no synergy between these two signals ( not shown ) . An alternative hypothesis is that IFN protein that is prestored [35] and therefore not measurable by qPCR ( not transcriptionally controlled ) or by ELISA ( too low sensitivity ) could be released locally and mediates the observed ISG induction . In conclusion , we show here that airway epithelia rely on two parallel , redundant amplification loops to induce an IFN signature in response to influenza A infection . Only a small fraction of genes , mostly non ISGs , are induced by the virus in the absence of both IFN systems , and no ISG appears to rely specifically on one IFN system only . This complete redundancy may guarantee induction of antiviral responses even if one or the other IFN system is blocked , for instance by specific virally encoded antagonists . In contrast to the two redundant IFN loops in epithelia , the majority of immune cells respond only to IFN type I , thus potentially allowing for differential control of the epithelial antiviral state and the induction of immune responses: while high levels of IFN type I would activate both epithelia and immune cells , high levels of type III would specifically activate epithelial responses but leave immune responses unaffected , which may help limit immune-mediated pathology in the lung and at other mucosal surfaces .
All animal breeding was approved by the local ethical committee of the NIMR and is part of a project approved by the UK Home Office ( licence number 80/2236 ) . Breeding was conducted according to local guidelines and UK Home Office regulations under the Animals Scientific Procedures Act 1986 ( ASPA ) . Influenza A virus strain A/PR/8/34 ( H1N1 ) was grown in day 10 embryonated chicken eggs , and titrated on MDCK by 50% tissue culture infective dose ( TCID50 ) , according to the Spearman-Karber method . All cells used in this study were derived from mice on the C57BL/6 background . MAVS−/− mice [36] and tracheae from TLR7−/− mice [37] were kindly provided by Dr . C . Reis e Sousa; MyD88−/− [38] , IFNAR1−/− [39] and TRIF−/− mice [40] were kindly provided by Dr . A . O'Garra . These and C57BL/6 wt mice were bred in-house under SPF conditions . Tracheae from Irf3−/− [29]; Irf7−/− [29]; Irf3−/−Irf7−/− cells were also used . IL-28Rα−/− [12] , IFNAR1−/− [39] and IL-28Rα−/−IFNAR1−/− cells were obtained from mice on a congenic B6 . A2G-Mx1 background carrying an intact Mx1 gene [41] . To generate chimeric mice , naïve B6 . A2G-Mx1 and B6 . A2G-Mx1 IL-28Rα−/−IFNAR1−/− recipient mice were lethally irradiated with 1000 rad and reconstituted with donor B6 . A2G-Mx1 BM cells ( 7×106 ) by intravenous injection . Chimeric mice were maintained for 7 weeks and chimerism assessed by IFNAR1 ( MAR1-5A3 antibody ) expression on Gr1+ , CD19+ , CD4+ and CD8+ cells in the blood ( Fig . S6 ) . Chimeric mice were then infected intranasally with 105 TCID50 of Influenza A/PR/8/34 in 30 µl PBS after anesthesia . Isolation and culture of primary MTEC were performed as previously described [42] . Briefly , cells isolated by enzymatic treatment were seeded onto 0 . 4 µm pore size clear polyester membrane ( Corning ) coated with a collagen solution . At confluence , media was removed from the upper chamber to establish an air- liquid interface ( ALI ) . Fully differentiated , 10–14 days-old post ALI cultures were routinely used for experiments . In some experiments , cultures were infected in the presence of brefeldin A ( 2 . 5 µg/ml ) or neutralizing anti-IL28A/B , anti-IL28B ( R&D Systems ) and blocking anti-IFNAR1 ( MAR1-5A3 ) antibodies , at a final concentration of 10 µg/ml each . Differentiated , ALI day 14 cultures were fixed in 4% paraformaldehyde and permeabilized with 0 . 1% Triton X-100 . Cells were then incubated with the indicated primary antibodies for 1 hour at room temperature , washed , incubated with fluorochrome-conjugated secondary antibody , washed and finally mounted . Image acquisition and processing information: ( i ) microscope: Olympus IX70; ( II ) magnification: 20×; ( III ) imaging medium: Vectashield with DAPI ( Vector labs ) ; ( IV ) fluorochromes: Alexa Fluor 488 , Alexa Fluor 568 . ( V ) acquisition software: Softworx . Images were processed with Image J . The apical surface of MTEC cultures was washed extensively to remove accumulated mucins before inoculation with IAV ( moi = 0 . 3 ) . After incubation at 37°C for 1 h , the virus inoculum was removed and the cultures were incubated in complete growth medium for 24 hours . Aliquots of the supernatants were collected at different time points and titrated by ELISA as described below . Cells were then lysed to extract RNA or for detection of viral protein by Western blotting . RNA was isolated from MTEC cultures by directly lysing the cells in the transwells , using the Qiagen RNeasy mini kit , according to the manufacturer's instructions . One microgram total RNA was reverse transcribed using the ThermoScript RT-PCR System kit ( Invitrogen ) . The cDNA served as template for the amplification of genes of interest and the housekeeping gene ( Hprt1 ) by real-time PCR , using TaqMan Gene Expression Assays ( Applied Biosystems ) , universal PCR Master Mix ( Applied Biosystems ) and the ABI-PRISM 7900 sequence detection system ( Applied Biosystems ) . The fold increase in mRNA expression was determined using the ΔΔCt method relatively to the values in mock treated samples , after normalization to Hprt1 gene expression . Total RNA harvested from MTEC cultures was hybridized using Affymetrix Mouse Genome 430 2 . 0 microarrays . The raw intensities values for each entity were preprocessed by RMA normalization against the median intensity in mock infected samples . Using GeneSpring 11 . 5 , all transcripts were filtered based on signal values , to select the ones whose level of expression was in the 100–20th percentile , in at least 50% of samples . Student's t test ( infected versus mock infected ) or 2-way ANOVA ( parameters: treatment and genotype ) were performed to identify gene significantly differentially expressed relative to controls ( ≥4-fold change; p<0 . 01 , Benjamini-Hochberg multiple test correction ) . Ingenuity Pathway Analysis ( IPA ) was used to select , annotate and visualize gene by function and pathway . ISGs were identified with the Interferome database ( www . interferome . org/ ) . Microarray data can be accessed at GEO under accession number GSE43710 for the superseries . Cell culture supernatants were harvested from the apical compartments of mock or IAV infected samples . IL-28A/B was measured using the IL-28A/B ELISA Duo kit ( R&D Systems ) , IFNβ with the Verikine IFNβ ELISA kit ( PBL Interferon Source ) . | The response of cells to virus infection depends on Interferons ( IFNs ) , a group of cytokines which activate the expression of hundreds of genes that help control viral replication inside infected cells . While type I IFN was discovered in 1957 , type III IFN ( IFNλ , IL-28/29 ) was characterized recently and is known for its role in the response to hepatitis C virus . Airway epithelia are the primary target of influenza virus , and we studied how infection induces IFNs and which IFN is most important for the epithelial anti-influenza response . We found that infected epithelia detect virus through the cytoplasmic RIG-I/MAVS recognition system , leading to activation of the transcription factor IRF7 and subsequent induction of both type I and III IFNs . All ensuing cellular responses to infection are dependent on the production and secretion of IFNs , as responses are lost in epithelia lacking receptors for both type I and III IFNs . Finally , gene induction is indistinguishable in single receptor-deficient and wild-type cells , indicating that the two IFN systems are completely redundant in epithelia . Thus , influenza infection of airway epithelia induces , via a RIG-I/MAVS/IRF7 dependent pathway , both type I and III IFNs which drive two overlapping and redundant amplification loops to upregulate antiviral genes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Type I and Type III Interferons Drive Redundant Amplification Loops to Induce a Transcriptional Signature in Influenza-Infected Airway Epithelia |
The gap between the number of known protein sequences and structures continues to widen , particularly as a result of sequencing projects for entire genomes . Recently there have been many attempts to generate structural assignments to all genes on sets of completed genomes using fold-recognition methods . We developed a method that detects false positives made by these genome-wide structural assignment experiments by identifying isolated occurrences . The method was tested using two sets of assignments , generated by SUPERFAMILY and PSI-BLAST , on 150 completed genomes . A phylogeny of these genomes was built and a parsimony algorithm was used to identify isolated occurrences by detecting occurrences that cause a gain at leaf level . Isolated occurrences tend to have high e-values , and in both sets of assignments , a sudden increase in isolated occurrences is observed for e-values >10−8 for SUPERFAMILY and >10−4 for PSI-BLAST . Conditions to predict false positives are based on these results . Independent tests confirm that the predicted false positives are indeed more likely to be incorrectly assigned . Evaluation of the predicted false positives also showed that the accuracy of profile-based fold-recognition methods might depend on secondary structure content and sequence length . We show that false positives generated by fold-recognition methods can be identified by considering structural occurrence patterns on completed genomes; occurrences that are isolated within the phylogeny tend to be less reliable . The method provides a new independent way to examine the quality of fold assignments and may be used to improve the output of any genome-wide fold assignment method .
The prediction of protein structures from sequences is becoming increasingly important , particularly as the gap between the number of experimentally determined sequences ( >6 , 000 , 000 ) and structures ( <35 , 000 ) widens . Knowledge of protein structure is essential to the understanding of biochemical processes . In practical terms , knowledge and prediction of protein structures can aid the discovery of new drugs [1] . When predicting the structure for a target sequence , a major step is achieved when an evolutionarily related protein with a known structure is identified . Since structure is more conserved than sequence , it is presumed that the target sequence has a similar fold to the related protein . This process is called fold recognition and has been a major force behind improvement in structure prediction in recent years [2] . At present there are several ways to recognise a fold for a given sequence . If there is close homology between the target sequence and a known structure in the Protein Data Bank ( PDB ) [3] , a simple sequence search , such as BLAST [4] , will be sufficient to identify the fold . To detect more distant homologies we can use position-specific scoring methods such as PSI-BLAST [5] and hidden Markov model ( HMM ) –based methods such as SAM-T98 [6] . These methods are the least expensive forms of fold recognition and are sequence based . There are more computationally expensive methods that take structural information into account; an example of such a technique is THREADER [7] . Recently , many studies have used fold recognition to look at the structural content of entire genomes with aid of PSI-BLAST [8] , HMMs [9 , 10] , or threading procedures [11] . These fold-recognition assignments can produce an occurrence pattern on a set of species for a given family , superfamily , or fold as defined by structural classifications such as SCOP [12] or CATH [13] . Sets of such occurrence patterns have proved to be useful for building a phylogeny of species [14 , 15] , for grouping proteins within a similar pathway [16] , and for estimating the ages of folds [17] . A major challenge for all fold-recognition techniques is to discriminate a true homologue from a false positive ( specificity ) using confidence scores such as e-values . e-Values ( expectation values ) indicate how likely it is that an alignment with the search sequence would occur by chance in a given database ( i . e . , they should reflect the chance of a false positive assignment ) . Previous studies have suggested that analysis of structural assignments on completed genomes may indicate false positives of fold-recognition techniques . Yang and coworkers [15] showed that the number of hits on completed genomes drastically increased above a certain e-value cutoff , which could be explained by a sudden influx of false positives . Furthermore , Winstanley and coworkers [17] showed that occurrence patterns of most superfamilies can explain the phylogeny of the genomes reasonably well , but observed a difference in fitness between patterns from two different fold-recognition techniques . In particular it appeared that , in one set , more assignments were made that occurred on isolated leaves of the species tree . We propose that false positives in fold-recognition assignments might be identified by considering a phylogeny of species . False assignments in such a set might be expected to occur randomly across the genome tree , whereas true positive assignments to a superfamily should be evolutionary related . Hence , we expect that false occurrences have a stronger tendency to be scattered across the tree than true assignments . This study investigates whether isolated occurrences within a phylogenetic occurrence pattern are indeed more likely to be false positives . Using phylogeny to improve homology searches is not a new idea . It has been known for a long time that by considering the phylogeny of related proteins one can improve sequence alignments . An example is progressive multiple sequence alignment , where an approximate phylogeny of the sequences is used to aid the alignment of multiple sequences [18] . Assignment of function can be facilitated by phylogenomics: a set of known homologs is used to create a phylogeny of proteins in which speciation and duplication events are marked . These can be used to subclassify the proteins in the phylogeny into specific functions . Several protocols as well as automated procedures based on phylogenomics have been able to improve functional annotation [19–21] . Recently it has also become clear that confidence in a modelled structure increases when homologues of the target sequence give a similar structure prediction [22 , 23] . Here , precalculated phylogenies of entire genomes are used rather than phylogenies of individual proteins . In this study we will assess genome-wide assignments obtained by PSI-BLAST searches as well as assignments from the SUPERFAMILY database . PSI-BLAST is an iterative version of BLAST , which uses a position-specific scoring matrix ( PSSM ) to include information about homologous sequences . The PSSM is used to identify the amino acids that are most likely to occur at a given position in the sequence . During each run , sequence information from all hits , with an e-value below a threshold , is added to the PSSM . Final e-values between the target and each database sequence are based on sequence similarity to the PSSM , which is subsequently normalised for amino acid composition with respect to the entire database . The SUPERFAMILY database [9] is built with SAM-T99 [6] , a procedure to find distant homologues using HMMs . The SAM-T99 program was finetuned with expert knowledge to recognise superfamilies as defined by SCOP . For each search sequence a profile HMM is created from homologues , which are found by a simpler sequence similarity search . The HMM is a statistical model that describes the evolutionary behaviour for a set of homologous sequences . These models are then used to calculate a score for each sequence in the database . The e-values are normalised by the reversed score of the searching sequence . To assess the assignments , occurrence patterns were obtained for every superfamily from our two fold-recognition sets ( PSI-BLAST and SUPERFAMILY ) . We developed a method that identifies isolated occurrences within such occurrence patterns by considering if an occurrence causes a gain at leaf level in the phylogeny . This study demonstrates that false positives in fold-recognition assignments can indeed be identified by considering a phylogeny of species: isolated occurrences are shown to have higher e-values ( are less reliable ) than other occurrences . We formulated criteria to predict false positives based upon these results . The set of predicted false positives were validated by comparisons to overlapping PSI-BLAST assignments and to assignments that changed between different versions of the SUPERFAMILY database . Both tests confirmed that the predicted false positives are far more likely to be falsely assigned than other occurrences . Analysis of occurrence patterns from genome-wide fold recognition also provides a new way to examine the quality of fold assignments . We show that the frequency of occurrences drastically increases for high e-values ( >10−8 for SUPERFAMILY and >10−4 for PSI-BLAST ) , and that this influx is likely to be caused by false positive assignments . In addition , the accuracy of the fold recognition is demonstrated to differ significantly for the different structural classes as defined by SCOP . In principle , this technique can screen assignments of any existing fold-recognition method for false positives . An extended version of the method is given , which can be applied to assignment sets with a high proportion of false positives . This version might be able to improve the search capacity of any genome-wide fold-recognition technique .
We used fold-recognition methods to find members of a superfamily . Hence , all assignments and occurrences within a pattern should be evolutionarily related . We propose that an isolated occurrence on a species tree might indicate that that occurrence is a false positive . Below we consider what other mechanisms might cause a bad fit within an occurrence pattern for a superfamily . The more occurrences there are in a pattern , the lower the chance that an isolated occurrence can appear . Therefore , if a pattern saturates with occurrences , fewer false positives can be predicted through our gain at leaf level technique . There is a strong anticorrelation ( r2 = 0 . 95 ) between the number of occurrences in a pattern and potential gains at leaf level . A potential gain at leaf level is a leaf without an occurrence , which would cause a gain at leaf level when added to the pattern . Occurrence patterns can become saturated through an increase in either true or false positive assignments ( see Discussion for consequences of additional true assignments ) . An increase of false positive occurrences could be created by the inclusion of very high e-values into the set of assignments . This might be desirable to enlarge the capacity of existing fold-recognition methods . However , if false occurrences start to dominate the pattern , our technique would begin to fail . The problem can be overcome using a slight modification to the technique . The parsimony algorithm is first run on a base pattern that is created from assignments below a strict e-value threshold . Genomes without an occurrence in this base pattern are checked for potential gains at leaf level . Then the set of isolated occurrences becomes the union of all gains at leaf level in the base pattern and all potential gains at leaf level , which have an occurrence with an e-value above the threshold . This modification could cause the algorithm to overpredict the number of false positives if the e-value cutoff is set too low . Figure 5 shows the results of running this procedure with different e-value cutoffs for the base pattern . Only when using extremely low e-value cutoffs ( e . g . , 10−20 ) are a considerable number of isolated occurrences detected for middle-range e-values . Hence , the above method can potentially be used without a huge overprediction of false positives . In fact , comparing the distributions ( coloured lines ) to the consensus data ( striped line ) , a sensible threshold can be determined with an e-value cutoff of 10−8 for SUPERFAMILY and 10−2 for PSI-BLAST . This version of our method can be used on occurrence pattern sets with a high proportion of false positives . At the highest level of the SCOP hierarchy , domains are grouped into seven different classes , predominantly based on secondary structure content . In our analysis , predicted false positives rates for these structural classes differ significantly from one another . For example , the first column in Table 2 shows that the proportion of occurrences with a gain at leaf level is lower for the alpha/beta and multidomain classes than for the all-alpha , all-beta , transmembrane , and small protein classes ( SUPERFAMILY ) . Some of these differences , although not all , might be explained by the average chain length of the proteins . Domains from the alpha/beta and multidomain classes have on average longer chain lengths than domains from the all-alpha and small-protein classes . Karplus and coworkers [6] previously observed that HMM-based methods might be less accurate at estimating correct e-values for small sequences; this could result in a higher false positive rate . Note that the multidomain class is different from the other six , as it contains proteins that cannot yet be split up into separate domains based on SCOP classification rules . This class is included here because it shows some evidence for sequence-length dependence . To investigate in more detail why certain structural classes have higher rates of predicted false positives , we submitted a small number of predicted false-positive gene regions to the meta servers ( http://bioinfo . pl/meta and http://genesilico . pl/meta ) . The servers predicted quite a few of the predicted alpha-class regions to be in coil-like structures . On inspection of the source domain , we observed that the region which generated the assignment often contained a very long alpha helix . This strong helical–helical scoring may explain the slightly higher false positive rate for the all-alpha class . These results must be interpreted with a little care , as the proportions of isolated occurrences might correlate with saturation of the occurrence patterns . As described above , more saturated occurrence patterns result in fewer potential gains at leaf level . Domains from the class of small proteins have on average a lower age than domains from the alpha/beta class [8 , 17] , and will therefore have on average emptier occurrence patterns . To correct for this , the second column of Table 2 shows the proportion of occurrences with a gain at leaf level that have a high e-value ( >10−10 for SUPERFAMILY and >10−4 for PSI-BLAST ) . These proportions indicate that the all-alpha and small-protein classes still display a higher proportion of false positives than the alpha/beta class for both the SUPERFAMILY and PSI-BLAST assignments . The second score seems to deviate more per class for SUPERFAMILY than for PSI-BLAST , while the standard deviations for the sample distributions are very similar . This may indicate that there is a higher e-value dependence on secondary structure content or domain length for the former method . Conventional benchmarking of fold-recognition methods involves blind tests on subsets of protein domains with a known structure [27–29] or on sets of sequences annotated with expert knowledge [30] . Using such a procedure , the ratio of false predictions can be estimated as well as the ratio of false negatives . A similar fashion of benchmarking would be difficult for our method , since the majority of genes have unknown structures , and using all genes on the genomes is essential to the method . We have already shown through e-value distributions that isolated occurrences are less reliable . Below we describe two other independent tests to verify our set of predicted false positives . Predicted false positives are defined as an occurrence with a gain at leaf level , a single copy , and a high e-value ( >10−10 for SUPERFAMILY and >10−4 for PSI-BLAST ) . Although no actual structures are available for the majority of the genes , a set of likely false positive assignments can be obtained by checking for overlap . For a given gene region , more than one assignment may be obtained . If the assignments are to two different superfamilies , the assignment with the worse e-value is likely to be false . Translating this to occurrence patterns , we can say that occurrences , which are solely generated by assignments with a stronger assignment in the same region , are likely to be false . We selected these occurrences and compared them with our sets of predicted false positives . Table 3 shows that the overlapping occurrences are found nine times more often in our set of predicted false positives than in general occurrences . Fisher's exact test confirms that this difference in numbers is highly significant ( p < 2 . 2 × 10−16 ) . Moreover , if the ratio of overlapping regions is measured within a set of occurrences above the same e-value threshold as the set of predicted false positives and is compared with this ratio of the set of predicted false positives , the fraction of overlapping regions remains significantly higher for the predicted false positives ( p = 2 . 7 × 10−5 ) . Unfortunately , this test can only be carried out on PSI-BLAST assignments , because in the SUPERFAMILY database , overlapping regions have already been removed . Instead we can compare two different versions of the database . Our study has been carried out on SUPERFAMILY version 1 . 65 . The more recent version ( 1 . 69 ) is deemed to be more reliable . We assessed assignments that were present in SUPERFAMILY version 1 . 65 , but which had been removed in version 1 . 69 . Again , we selected occurrences for which all assignments on the genome were removed ( Table 4 ) . More than 90% of our predicted false positives in SUPERFAMILY 1 . 65 had been removed in version 1 . 69 . The fraction of removed occurrences is significantly higher in the set of false positives than in general occurrence ( p < 2 . 2 × 10−16 ) and significantly higher than in the set of occurrences with a similar e-value ( p < 2 . 2 × 10−16 ) . These results both confirm that our set of predicted false positives are significantly more likely to be false than general occurrences .
As described above , fewer isolated occurrences can be detected as occurrence patterns become more saturated . Such saturation could be caused by an increase in true positive assignments , when fold-recognition techniques become far more sensitive , or when many more protein structures become available . Previous work [8] has shown that the number of occurrences is not uniformly distributed . The distribution peaks at a low and a high number of occurrences , with a minimum in the middle . In addition , recent work by Yan and Moult [31] shows that the number of known ( super ) families , which occur on very few genomes , is expected to grow . This implies that although some superfamily patterns might become more saturated , the number of superfamilies with emptier patterns will also grow . Hence , this method will remain applicable to a large number of superfamilies . When interested in the fold assignment of a single protein , it is important to keep in mind that an isolated occurrence may appear due to lateral gene transfer or extensive gene loss . Nevertheless , it can be advantageous to visually inspect predicted occurrence patterns of homologous sequences with a weak hit to the protein of interest . Anomalies in such a pattern can give an idea about the reliability of the assignments , and may also indicate false negatives or deletions in cases of a “loss at leaf level . ” When working on a specific protein , it is also important to be aware of the differences in aim and methodology between this method and phylogenomics methods used for functional annotation . The aim of this study is to find true homologues in a set of likely homologues , whereas phylogenomic methods aim to subclassify a set of known homologues . The difference in aim results in a distinctively different methodology . For known ( close ) homologues , it is feasible to calculate a phylogeny of proteins and subsequently assign duplication and speciation events [21] , whereas in this work , the proteins of interest have very distant ( if any ) evolutionary relationships and a precalculated phylogeny of genomes is used instead . Hence , our method is not a substitute for a phylogenomics method , but could perhaps be used as a prefilter in cases where structural assignments are involved ( e . g . , see Sjolander [20] ) . Previously we described how the method could be extended to assess assignment sets with a large proportion of false positives using a base pattern . However , rather than assignments below an e-value threshold , consensus occurrences could be used to create the base pattern . In this study , two profile-based searching methods were analysed . Although a large proportion of occurrences is in agreement ( 84% for SUPERFAMILY and 95% for PSI-BLAST ) , only about 50% of these occurrences were caused by the same number of assignments . The two fold-recognition methods might recognise different homologues on the same genome . Consensus occurrences appear to be more reliable on evaluation by our method: the proportion of isolated occurrences was significantly lower than for occurrences obtained by a single method , even though consensus occurrence patterns are naturally sparser . The usage of consensus occurrences rather than consensus assignments could therefore provide additional information about the reliability of the assignments . This technique might be used as an additional quality check for meta servers , using consensus data of several genome-wide fold-recognition methods . This study shows that false positives assigned by fold-recognition methods on completed genomes can be detected by determining isolated occurrences in a phylogeny . We developed a method to identify these isolated occurrences by applying a parsimony algorithm to a phylogenetic occurrence pattern . An occurrence is said to be isolated if it causes a “gain at leaf level” in the most parsimonious evolutionary scenario . It is shown that in principle , isolated occurrences are less reliable than other assignments: the majority of isolated occurrences have a high e-value ( >10−8 for SUPERFAMILY , >10−4 for PSI-BLAST ) . e-Values are shown to be almost independent of evolutionary distance between the source sequence and the genome of assignment . Deletions and/or false negatives are therefore unlikely to cause the observed high e-values of isolated occurrences . To predict false positives in practise , additional constraints should be imposed . The e-value of the occurrence should be higher than a given e-value ( >10−8 for SUPERFAMILY , >10−4 for PSI-BLAST ) to minimise the number of isolated occurrences caused by lateral gene transfer . In addition , the occurrence ( of a superfamily ) should be caused by a single assignment to the genome . Using this technique , more than 1 , 000 false positives can be predicted for both SUPERFAMILY and PSI-BLAST . Tests with independent means to indicate false positives were performed to validate our predicted false positives; one test was based on overlap of PSI-BLAST assignments , and the other considered changes between different versions of the SUPERFAMILY database . Both tests confirmed that isolated occurrences are more likely to be falsely assigned . The method can be extended to assess sets of assignments with a large proportion of false positives and could be used to enhance the searching power of existing fold-recognition techniques . This technique could therefore provide a way to fundamentally improve the assignment sets of genome-wide fold recognition . Considering occurrence patterns from genome-wide fold recognition also gives a new way to examine the quality of fold assignments . It was observed that the number of assignments and occurrences on genomes drastically increases for high e-values . The most likely explanation for this phenomenon is a sudden increase in false positive assignments above certain e-values , since the sudden increase is not observed in consensus data , with occurrences predicted by both SUPERFAMILY and PSI-BLAST assignments , or for data where isolated occurrences were removed . When examining the rate of predicted false positives for different structural classes , a significant variance was observed . Domain length and secondary structure content might cause this dependency between false positive rate and structural class .
The first set of assignments was taken from SUPERFAMILY database [24] version 1 . 65 and covers 1 , 269 different superfamilies as defined by SCOP [12] on 150 completely sequenced genomes . About 750 , 000 structural assignments were made , with the following restrictions imposed: ( 1 ) all assignments have an e-value lower than 10−4; and ( 2 ) no other assignment on the same region of the gene is made with a lower e-value . Table S1 shows the genomes and their coverage by assignments . The second set consisted of assignments obtained by PSI-BLAST [5] searches on the same 150 genomes . Sequences with less than 95% sequence identity from the ASTRAL [32] database were used to search for structural domains in a nonredundant database created from all genes in the 150-genome set . PSI-BLAST was used with a SEG filter and an e-value cutoff of 1 × 10−5 for inclusion in the PSSM . Assignments with an e-value smaller than 1 . 0 were included after the final run . The e-values for the assignments were taken from the PSI-BLAST run in which the assignment first falls below 1 × 10−5 ( i . e . , when it is not yet included for scoring in the PSSM ) . Note that there was no check for overlap of assignments within a gene . In contrast with SUPERFAMILY , a repeat of a domain from the same superfamily within a gene was counted as a single copy . Phylogenies of the genomes were created using the SUPERFAMILY occurrence data ( the number of copies were not included ) . A neighbour-joining algorithm was used to create a tree . The branch lengths were then normalised so that all leaves were at an equal distance from the root ( 1 . 0 ) , following the method used by Winstanley and coworkers [17] . An occurrence is in this study defined as the occurrence of a superfamily on a genome . An occurrence can therefore be caused by more than one assignment . The e-value of an occurrence is defined as the lowest e-value in the set of assignments covering the occurrence . A parsimony algorithm is used to find isolated occurrences in a pattern given the phylogeny . This minimises the number of loss and gain events for a superfamily in the species tree [17 , 33 , 34]; a detailed description of the algorithm can be found in [33] . Isolated occurrences are identified as occurrences that create a gain at leaf level . The algorithm used a gain penalty that was twice the size of the loss penalty in order to take a high number of false negatives into account . Experimentation with a lower relative gain penalty showed only a small increase in isolated occurrences . This indicates the technique is relatively robust against false negatives . The parsimony algorithm was implemented in Java ( J2SE 5 . 0; http://java . sun . com ) . A potential gain at leaf level is defined as a genome without an occurrence for a given superfamily that would cause a gain at leaf level if an occurrence were added to the existing pattern . To calculate potential gains at leaf level , the parsimony algorithm is run for every genome without an occurrence in the pattern . A base pattern is created from assignments with an e-value below a given threshold . Subsequently , the parsimony algorithm is run on the base pattern , and potential gains at leaf level are predicted . A set of isolated occurrences can be found as the union of the set of ( 1 ) occurrences within the base pattern which cause a gain at leaf level; and ( 2 ) all occurrences caused by assignments above the e-value threshold , with a potential gain at leaf level . The distance to source for an occurrence is the age of the youngest common ancestor between the occurrence and any genome containing the source domain . A genome is said to contain a source sequence for a SCOP superfamily if it covers 80% of its length and has at least 95% sequence identity . Simple BLAST searches were used to identify the source sequences . No source sequences could be identified for a few superfamilies on our set of genomes . The cluster distance reflects how “far away” an occurrence is from any other occurrence within a pattern , given the phylogeny . The cluster distance is calculated as the sum of distances to every occurrence in the tree divided by the distance to every leaf in the tree . This score was subsequently divided by the average score for each occurrence in the pattern so that the average cluster distance of a pattern became 1 . 0 . A cluster score of 1 . 0 indicates average clustering; a score of <1 . 0 indicates tighter clustering . Mediation by the average distance to each leaf was used , since some leaves in the tree lie in tighter clusters and would generally produce lower scores without mediation . A consensus occurrence is an occurrence that is identified by both SUPERFAMILY and PSI-BLAST . Note that the occurrence does not necessarily have to be caused by assignments to the same gene . A simple sampling procedure was used to determine if the proportion of false positives for a structural class deviates significantly from the overall false positive rate in Table 2 . The ratio of predicted false positives was calculated for 500 random samples of the occurrence patterns . For each sample , the number of random genome entries was chosen to match the class size . The resulting distribution of sampled false positives rates was used to determine if the rate for each structural class was significantly lower , falling within the lowest 1% of the sampled distribution , or significantly higher , falling within highest 1% . The number of genome entries for a class depends on the number of superfamilies and ranges from almost 6 , 000 entries ( multidomain ) to just over 50 , 000 entries ( alpha/beta ) . Overlapping regions of assignments were identified as assignments with a stronger assignment to a different superfamily on the same gene region and with at least 50% of the weaker assignment covered by the stronger assignment . Overlapping occurrences are occurrences that would be taken out of the dataset if only the strongest assignment within a region was retained . All figures are plotted using R [35] except Figure 1 , which was created by a modified version of a Java applet ( http://www . stats . ox . ac . uk/∼abeln/howold ) . Linear regression was also performed using R . Figure 2 is plotted using a kernel estimate for the density function; the amplitude is then multiplied by the number of data elements in the set to obtain an approximate frequency . | When predicting the structure for a protein sequence , a major step is achieved when an evolutionarily related protein with a known structure is identified . This process is called fold recognition , and has been a major force behind improvement in structure prediction . Moreover , fold-recognition techniques have become increasingly important in recent years because of the huge numbers of protein sequences with unknown structures available through sequencing projects on completed genomes . However , all fold-recognition methods tend to produce either a large number of false negatives ( at high confidence scores ) or a large number of false positives ( at low confidence scores ) . Here we show that the reliability of a fold-recognition technique can be explored by analysing its predictions across a set of completed genomes . We have developed a method that can indicate false positives in these genome-wide assignment sets . The basic idea behind the method is that a fold assignment on a genome is less reliable if the prediction is not observed on evolutionary related genomes . The ability of the method to discriminate false positives is confirmed by independent tests . The method can be used on the output of any genome-wide fold assignment method . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology",
"none",
"molecular",
"biology",
"computational",
"biology"
] | 2007 | Using Phylogeny to Improve Genome-Wide Distant Homology Recognition |
Conditional deletion of Apc in the murine intestine alters crypt-villus architecture and function . This process is accompanied by multiple changes in gene expression , including upregulation of Cited1 , whose role in colorectal carcinogenesis is unknown . Here we explore the relevance of Cited1 to intestinal tumorigenesis . We crossed Cited1 null mice with ApcMin/+ and AhCre+Apcfl/fl mice and determined the impact of Cited1 deficiency on tumour growth/initiation including tumour multiplicity , cell proliferation , apoptosis and the transcriptome . We show that Cited1 is up-regulated in both human and murine tumours , and that constitutive deficiency of Cited1 increases survival in ApcMin/+ mice from 230 . 5 to 515 days . However , paradoxically , Cited1 deficiency accentuated nearly all aspects of the immediate phenotype 4 days after conditional deletion of Apc , including an increase in cell death and enhanced perturbation of differentiation , including of the stem cell compartment . Transcriptome analysis revealed multiple pathway changes , including p53 , PI3K and Wnt . The activation of Wnt through Cited1 deficiency correlated with increased transcription of β-catenin and increased levels of dephosphorylated β-catenin . Hence , immediately following deletion of Apc , Cited1 normally restrains the Wnt pathway at the level of β-catenin . Thus deficiency of Cited1 leads to hyper-activation of Wnt signaling and an exaggerated Wnt phenotype including elevated cell death . Cited1 deficiency decreases intestinal tumourigenesis in ApcMin/+ mice and impacts upon a number of oncogenic signaling pathways , including Wnt . This restraint imposed by Cited1 is consistent with a requirement for Cited1 to constrain Wnt activity to a level commensurate with optimal adenoma formation and maintenance , and provides one mechanism for tumour repression in the absence of Cited1 .
Inactivation of the APC ( adenomatous polyposis coli ) gene marks one of the earliest events in colorectal tumourigenesis [1] , an observation that has given rise to the concept of Apc as a ‘cellular gatekeeper’ protecting against tumourigenesis [2] . This role in suppressing tumour formation has been closely associated with its ability to regulate the level of β-catenin within cells . Thus , Apc normally forms part of the scaffold of proteins that phosphorylate β-catenin and target it for degradation . In the absence of Apc , β-catenin levels become elevated and translocates to the nucleus , where it drives increased transcription of Wnt target genes associated with cell proliferation and cell death [3] . To investigate the biological consequences of Apc loss and Wnt activation , we and others have previously used a conditional model of Apc loss . In this model , deletion of Apc is achieved through use of an inducible AhCre transgene , which is responsive to exposure to the xenobiotic β-napthoflavone . Following Cre induction and loss of function of Apc , we observe a range of rapid phenotypic changes . These include promiscuous entry of cells into S phase , loss of differentiated cell types , loss of cell polarity and disorganisation of the crypt-villus structure to the point that discrete crypts are no longer discernable . Apc deficiency also reduces the normal migration of cells along the crypt villus axis , leading to the preferential retention of Apc deficient cells . These changes may all be considered pro-tumourigenic , however we also observe a considerable stress signal within Apc deficient cells , most clearly shown by a significant elevation in apoptosis . These phenotypic changes are accompanied by the expected elevation in levels of nuclear β-catenin and marked changes in the transcriptome [3] . One of the changes we observe in the intestinal epithelial cells of AhCre+Apcfl/fl mice is a strong induction of Cited1 , a bi-functional transcriptional cofactor which is able to activate or repress transcription in association with other transcription factors [4] , [5] . We also found this induction to be dependent upon functional c-Myc as Cited1 expression returns to basal levels in the additional absence of c-Myc , which completely rescues the phenotype of Apc deficiency [6] . These observations suggest that elevation of Cited1 may be directly associated with the preneoplastic phenotype . Cited1 was originally identified in a mouse melanoma cell line [7] . During vertebrate development , Cited1 is expressed in progenitors of the heart , limb , axial skeleton , kidney , and placenta [8] , [9] . It is implicated as a key co-ordinator during renal epithelial morphogenesis [5] and is involved in mammary gland development [10] . Cited1 is required for placental development with effect on embryo growth and survival [9] . However , Cited1 null mice that survive the early postnatal period are otherwise grossly phenotypically normal [9] . Cited1 is also able to enhance TGF-β signaling and inhibit Wnt signaling depending on cellular context [5] , [11] . Both activation and inhibition of transcription are dependent on the CBP/p300 binding C-terminal transcription activation domain CR2 , which is conserved throughout the Cited family [5] , [11]–[13] . Deregulation of CITED1 has been implicated in several human cancers , including melanomas , Wilm's tumours and nephroblastomas [7] , [14]–[16] . In the mouse , Cited1 is up-regulated in MMTV-Cre/FloxNeoNeuNT mammary tumours and associates with the transcription factor EGR2 to regulate the expression of the oncogene ErbB2 ( HER2 , Neu ) [17] . Recently it has been shown that Cited1 expression , together with another transcription regulator Six2 , specify self-renewing nephron progenitor cells in kidney development and it is suggested that Cited proteins may contribute to the maintenance of the self-renewing capping mesenchyme in the developing kidney [18]–[20] . Thus , although a body of studies have implicated Cited1 in both embryogenesis and carcinogenesis , its potential role in Wnt-induced intestinal tumourigenesis remains unresolved . Given the data implicating Cited1 as a regulator of the Wnt pathway , we have tested the hypothesis that Cited1 plays a key role in intestinal tumourigenesis . We show that CITED1 is upregulated in human colorectal cancers and that Cited1 deficiency increases the survival of ApcMin/+ mice . When crossed into our acute model of Apc deficiency , we show that loss of Cited1 accentuates nearly all aspects of the Apc deficient phenotype , including the transcription of a range of oncogenic signaling pathways , including the Wnt pathway .
We have previously shown that deletion of Apc in the mouse intestine leads to nuclear β-catenin translocation and up-regulation of Wnt target genes , including Cited1 , as scored by microarray analysis [3] . To confirm this upregulation , we analysed mouse intestinal epithelium from AhCre+Apcfl/fl and AhCre+WT ( WT: wild type ) mice which had been induced by intraperitoneal injection of β-napthoflavone 4 days previously [3] . Quantitative PCR analysis revealed significant upregulation of the Wnt targets c-Myc , Axin2 and Cd44 in the absence of Apc . Similarly , Cited1 showed a 15-fold increase in expression ( p<0 . 05 , Figure 1A ) . To determine if CITED1 was also deregulated in human cancers , we performed a Taqman quantitative PCR on human colorectal tumour tissues . In comparison to paired normal tissues from the same patient , we observed over-expression of the human orthologues of the Wnt target genes c-MYC , AXIN2 , CD44 and CITED1 ( p<0 . 01 , Figure 1B ) . These data demonstrate the potential transferability of our data from the acute Apc deletion mouse model to human colorectal carcinogenesis . We next assessed Cited1 levels in adenomas developing in the ApcMin/+ mouse model of human colorectal cancer , which allows evaluation of the effects of loss of Apc function over the course of polyp development . Again , the level of Cited1 expression was significantly increased in intestinal polyps from ApcMin/+ mice compared to normal tissue from the same mouse ( p<0 . 01 , Figure 1C; Figure S1A ) . Detection of high levels of Cited1/CITED1 expression in both human and murine tumours suggests that Cited1/CITED1 may play a role in intestinal tumourigenesis . We performed in situ hybridization using a Cited1 probe designed against the deleted sequence in the Cited1− mouse ( Figure 1D ) . We observed low levels of staining ( compared to expression of the housekeeping gene Polr2a ) throughout the crypt-villus structure in the AhCre+WT mouse intestine ( WT ) with a trend to higher levels within the crypt . There was no apparent specificity for the stem cell region at the base of the crypt or for any differentiated cell type . Consistent with QPCR data , we observed an increase in the level of staining throughout the crypt-villus structure of the AhCre+Apcfl/fl mouse . As Cited1 is upregulated in colonic tumors we next asked whether deficiency of Cited1 could inhibit intestinal adenoma formation in the ApcMin/+ mouse . To achieve this , we crossed Cited1 null ( Cited1− ) mice onto ApcMin/+ . Given that Cited1 is on the X-chromosome , we aged male cohorts of ApcMin/+Cited1− and ApcMin/+ mice until they displayed symptoms of intestinal neoplasia ( rectal bleeding and paling feet ) . The median lifespan of ApcMin/+ mice was 230 . 5 days , which is increased to 515 days in the ApcMin/+Cited1− mice ( Log-Rank p = 0 . 001 , Figure 2A ) . Cited1− mice had a survival rate that was not significantly different to that of WT mice ( Figure 2A ) . We next counted the number of adenomas in the small and large intestine ( Figure 2B ) . ApcMin/+Cited1− mice developed significantly less tumours compared to ApcMin/+ in both the small intestine ( 5 versus 18 . 5 adenomas p<0 . 05 ) and the large intestine ( 1 versus 6 . 5 adenomas , p<0 . 05 ) . The tumour distribution in the small intestine and the colon was analysed at ill health ( Figure S1C ) . There was no significant difference in the percentage of tumours found in the duodenum or jejunum of the small intestine , or in the large intestine . However , we did observe a significant increase in the percentage of tumours found in the last part of the small intestine which corresponds to the human ileum ( Figure S1C ) . Total tumour burden of ApcMin/+Cited1− mice was not significantly different from that of ApcMin/+ ( Figure 2C ) , and shared the same tubular morphology and degree of invasiveness , as assessed histologically by the frequency of invasion into the submucosa ( 52 . 8% High grade +47 . 16% Low grade in ApcMin/+ vs 42 . 8% High grade +57 . 4% Low grade in ApcMin/+Cited1− , Chi-square x2 = 2 . 1 , DF = 1 , p>0 . 05 ) . These data suggest that mice became symptomatic of disease when they had developed an equivalent tumour burden , but that in the Cited1 mutant background this was significantly later and reflected fewer , but larger lesions at these later time points , hence implicating Cited1 in intestinal tumour initiation . To address the mechanism underlying the reduction of adenoma formation in ApcMin/+Cited1− mice , we crossed Cited1− mice with mice conditionally mutant for Apc . We have previously demonstrated that we can achieve almost 100% recombination of the Apcfl/fl allele in the intestine using the β-napthoflavone inducible AhCre transgene to drive recombination [3] . Thus , AhCre+WT , AhCre+Apcfl/fl , AhCre+Cited1− , and AhCre+Apcfl/flCited1− mice were induced with β-napthoflavone and culled 4 days after the first injection to determine the role of Cited1 immediately following deletion of Apc . To confirm the level of Apcfl/fl recombination we used quantitative RT-PCR and again found that 100% of the PCR products obtained were from the recombined Apc allele ( Figure S1B ) . We also confirmed Cited1 deficiency in Cited1− mice using RT-PCR . We observed a significant 3 . 81 fold difference decrease in Cited1 expression in AhCre+Cited1− compare to AhCre+WT . The small difference observed is most likely due to the low level of expression of Cited1 in the intestine [21] . Due to the increased level of Cited1 expression after loss of Apc , Cited1 deficiency is more noticeable in the intestinal epithelial cells of AhCre+Apcfl/flCited1− mice which showed a 277 . 81 fold decrease compared to AhCre+Apcfl/fl mice ( p<0 . 05 Mann-Whitney U test ) . We have previously shown that the loss of Apc leads to an increase in proliferation and apoptosis and also to a loss of migration [3] . To analyse the effects of Cited1 deficiency after Apc loss , we first counted the number of cells in S phase within the crypt or hyperplastic areas ( formed after Apc loss ) . On day 4 after β-napthoflavone induction , mice were injected with BrdU to label cells in S-phase and culled 2 hrs later ( Figure 3A ) . In AhCre+WT and AhCre+Cited1− mice the number of proliferating cells was not significantly different ( AhCre+WT: 18 . 97 vs AhCre+Cited1−: 20 . 65 BrdU positive cells/Crypt; p>0 . 05 . Figure 3B ) . However , the number of cells in S-phase was significantly increased in the hyperplastic areas of AhCre+Apcfl/flCited1− mice compared to AhCre+Apcfl/fl mice ( AhCre+Apcfl/fl: 74 . 13 vs AhCre+Apcfl/flCited1−: 106 . 4 BrdU positive cells/area , p<0 . 05 , Figure 3B ) suggesting a role for Cited1 in controlling cell proliferation in the context of active Wnt signaling . We next analysed the histology on HE sections of the intestinal tissue from all the genotypes after β-napthoflavone induction . There were no gross changes in the crypt/villus architecture in AhCre+WT compared to AhCre+Cited1− mice , and induced AhCre+Apcfl/flCited1− mice had similar large aberrant crypts to those observed in AhCre+Apcfl/fl mice ( Figure 4D ) . Given our findings of decreased adenoma formation in ApcMin/+ mice , we also examined the extent of the hyperplastic area within the crypts of AhCre+Apcfl/fl mice compared to AhCre+Apcfl/flCited1− mice as determined by the extent of BrdU labelling . Surprisingly , the number of cells in the hyperplastic area was greater in AhCre+Apcfl/flCited1− mice compared to AhCre+Apcfl/fl mice ( AhCre+Apcfl/fl: 78 cells/area vs AhCre+Apcfl/flCited1− :106 cells/area , p<0 . 05 , Figure 4E ) . We next determined migration rates by comparing the position of cells 2 hrs and 24 hrs after BrdU labelling . The difference between the 2 hrs and 24 hrs distributions for a genotype was analysed with the Kolmogorov–Smirnov test . The distribution of Brdu positive cells from 2 hrs to 24 hrs varies significantly for all genotypes ( p = 0 . 01 ) indicating cell migration . Enterocytes in AhCre+WT mice and AhCre+Cited1− mice migrate at the same rate ( 20 cell positions at the 50% cumulative frequency ) whereas cells in both AhCre+Apcfl/fl mice and AhCre+Apcfl/flCited1− mice show greatly reduced migration rates ( Figure 3C ) . Critically , although deletion of Apc in AhCre+Apcfl/fl mice results in strong suppression of migration ( 9 cell position migration ) , some movement of cells was detected in these samples ( Figure 3C ) . By comparison , the absence of Cited1 in AhCre+Apcfl/flCited1− mice , resulted in even less migration ( 2 cell position migration ) than observed for enterocytes in AhCre+Apcfl/fl mice ( Figure 3C ) . Together , these data demonstrate that Cited1 deficiency further exacerbates both the proliferation and migration phenotypes of Apc loss in the intestine , which is surprising given that ApcMin/+Cited1− mice developed significantly less intestinal tumours than ApcMin/+mice . Consistent with these observations that proliferation is increased in AhCre+Apcfl/flCited1− mice , we also observed a decrease in the number of differentiated entoendocrine cells and goblet cells in these mice ( Figure S5A–D ) . We and others have previously demonstrated that the location and the number of paneth cells in the intestinal crypt are regulated by Wnt signalling [6] , [22]–[24] . It is observed by the increased number of paneth cells after loss of Apc ( Figure S5F ) and the loss of positioning at the bottom of the crypt ( Figure S5E–G ) . Consistent with our observations that the phenotype of AhCre+Apcfl/fl mice is enhanced upon deficiency of Cited1 we also observe a change in position of the paneth cells in the hyperplastic areas of the AhCre+Apcfl/flCited1− compared to AhCre+Apcfl/fl ( Figure S5G ) . This is most likely due to the increase in crypt size seen in the double mutant , which gives cells a bigger area to be distributed . The increase in proliferation observed in the intestine following deletion of Apc is also associated with a dramatic increase in apoptosis [3] . We therefore examined if Cited1 was regulating apoptosis by counting apoptotic bodies in H&E sections and also scoring Caspase 3 staining . We observed no significant difference in apoptosis between AhCre+WT and AhCre+Cited1− mice , however , there was a significant increase in the number of apoptotic cells in AhCre+Apcfl/flCited1− mice compared to AhCre+Apcfl/fl mice , which was verified by both methods ( Figure 4A ) . As mentioned above , we observed an increase in the number of cells per hyperplastic area in the AhCre+Apcfl/flCited1− samples . To verify that the increase in cell death was not an artefact of the difference in the number of cells per area , we corrected for this difference between AhCre+Apcfl/fl and AhCre+Apcfl/flCited1− mice . The normalised data confirmed increased cell death in AhCre+Apcfl/flCited1− compared to AhCre+Apcfl/fl ( p<0 . 05 ) after H&E counting ( AhCre+Apcfl/fl : 11 . 53 apoptotic cells/area vs AhCre+Apcfl/flCited1−: 15 . 33 apoptotic cells/area , p<0 . 05 ) and after anti cleaved-Caspase3 staining ( AhCre+Apcfl/fl: 7 . 32 apoptotic cells/area vs AhCre+Apcfl/flCited1−: 9 . 53 apoptotic cells/area , p<0 . 05 ) ( Figure 4B–C ) . These data indicate that the increase in cell death is not proportional to the increase in cell proliferation . Therefore , Cited1 deficiency in a Wnt perturbed background accentuates the apoptotic response . Scoring of Caspase3 positive cells revealed no change in the number of apoptotic cells in the intestinal tumors of ApcMin/+Cited1− mice compared to ApcMin/+ mice at time of death ( Figure 4F ) . Therefore , the increased apoptosis we observe in the absence of Cited1 is only manifested in the context of acute Wnt activation , which underlines the role of Cited1 in restraining tumour initiation , and also implies that in those tumours that do develop in the absence of Cited1 , they have developed alternate mechanisms to restrain the Wnt pathway . We next wished to investigate the mechanism through which Cited1 may be modifying Wnt driven tumorigenesis . One possibility is a direct effect upon Wnt signaling , and in support of this , Cited1 has previously been shown to be able to bind to β-catenin and consequently inhibit Wnt induced transcription during Xenopus development [5] . Two potential TCF-4 sites were identified in the Cited1 promoter region ( ctttgt and cattgaa in the 2 kb prior exon1 ) . This implicates Cited1 in the control of the Wnt pathway , however this is not the only pathway known to be altered by Cited1 . Cited1 has been shown to bind to the p300/CBP coactivators and also to Smad4 , thereby enhancing their transcriptional activity [11] , [25] . To analyse the effects of Cited1 deficiency on various transcriptional pathways we performed a microarray analysis using the Affimetrix Chip 430 2 . 0 and AffylmGUI software [26] . We then submitted our microarray data to ingenuity pathway analysis software ( IPA ) to identify pathways significantly affected by Cited1 deficiency . In the AhCre+WT after additional loss of Cited1 , a number of signaling pathways identified by IPA analysis were found to be affected , amongst them: P53 ( p = 6 . 87×10−6 , ratio = 0 . 146 ) , PI3K/AKT ( p = 6 . 74×10−6 , ratio = 0 . 114 ) , Pten ( p = 7 . 7×10−4 , ratio = 0 . 097 ) ; Wnt ( p = 1 . 08×10−1 , ratio = 0 . 057 ) ; and TGFβ ( p>0 . 1 , ratio = 0 . 034 ) . Several targets were analysed by QPCR including c-Myc , Axin2 , CD44 , Sox4 , p53 , Pten , Akt1 , and Smad4 but none were found to be significantly deregulated ( N = 6 , p>0 . 05; Mann-Whitney ) . Several signaling pathways identified by IPA analysis were affected in AhCre+Apcfl/flCited1− mice compared to AhCre+Apcfl/fl mice , including: P53 ( p = 3 . 5×10−8 , ratio = 0 . 177 ) ; PI3K/AKT ( p = 2 . 62×10−5 , ratio = 0 . 107 ) ; Pten ( p = 7 . 22×10−4 , ratio = 0 . 097 ) , Wnt ( p = 5 . 4×10−2 , ratio = 0 . 063 ) ; and TGFβ ( p = 2 . 22×10−1 , ratio 0 . 056 ) . The validity of the IPA analysis was subsequently verified by QPCR . We analysed several targets from these pathways by QPCR , and found significant upregulation of p53 , Runx1 , Sox4 ( Figure S2C ) and a number of Wnt targets known to be deregulated in the intestines of AhCre+Apcfl/fl mice [3] or listed as Wnt target genes in the Nusse webpage ( http://www . stanford . edu/group/nusselab/cgi-bin/wnt/target_genes ) ( Figure 5A–B ) . 10 Wnt target genes , including c-Myc , Axin2 , and CD44 were confirmed by QPCR to be significantly up-regulated in AhCre+Apcfl/flCited1− mice compared to AhCre+Apcfl/fl mice ( Figure 5A ) . Three additional transcripts were analysed by microarray analysis that have previously been identified as key players in the Wnt pathway ( Nucleophosmin , Nucleolin , and β-catenin respectively: [27]–[29] ) . These were also found to be upregulated in AhCre+Apcfl/flCited1− mice compared to AhCre+Apcfl/fl mice . These data indicate that Cited1 inhibits several signaling pathways , including the Wnt pathway following Apc loss . The Wnt signalling pathway has been shown to play a critical role in intestinal homeostasis which includes stem cells maintenance . Because Cited1 loss leads to a deregulation of the Wnt pathway and due to the potential role of Cited1 in the stem cell niche in the cap mesemchyme in the developing kidney [18] , we analysed the effect of Cited1 deficiency in the intestine . RT-QPCR analysis revealed a significant upregulation of several stem cell markers ( Gpr49 , Ascl2 , Musashi and Olfm4 ) in AhCre+Apcfl/flCited1− tissues compared to AhCre+Apcfl/fl controls ( Figure 5C ) . We also performed ISH for the surrogate marker of Lgr5 , Olfm4 ( Figure 5D ) . In AhCre+Wt and AhCre+Cited1− mice the location of Olfm4 expressing cells is confined to the stem cell niche at the base of the crypts . In AhCre+Apcfl/fl mice , Olfm4 expressing cells were distributed throughout the hyperplastic area . In AhCre+Apcfl/flCited1− mice Olfm4 expressing cells are also mislocalised throughout the aberrant crypts but expression is increased , consistent with our RT-QPCR data of the same tissue ( Figure 5C ) . Loss of Apc has been shown to drive an increase in total β-catenin and more importantly a re-localisation of the active form of β-catenin to the nucleus [3] . To test if Cited1 deficiency modified this phenotype , we analysed the localisation of total β-catenin in the small intestine by immuno-histochemistry ( Figure 6A ) . We observed a normal pattern of localisation in both AhCre+WT and AhCre+Cited1− mice consistent with previous findings [30] . Upon deletion of Apc , we observed nuclear translocation of β-catenin in the aberrant crypts of both AhCre+Apcfl/fl and AhCre+Apcfl/flCited1− mice ( Figure 6A ) indicative of de-regulated Wnt signalling . β-catenin regulates important cellular functions such as transcription and adhesion [31] , and the cellular concentration and phosphorylation status of β-catenin has been shown to impact on these functions [31] , [29] . As we observe an increase in the transcription of several Wnt target genes in AhCre+Apcfl/flCited1− mice we examined the level of total β-catenin , the extent of phosphorylation at multiple sites and the ratio of transcriptionally active free β-catenin in purified intestinal epithelial cells ( Figure 6B ) as previously described [32] . First , we observed a significant increase in total β-catenin accompanied by an increase in the active form of β-catenin ( dephosphorylation at pS33 , pS37 , pT41 sites ) in AhCre+Apcfl/fl and AhCre+Apcfl/flCited1− mice compared to AhCre+WT and very importantly in AhCre+Apcfl/fl mice compared to AhCre+Apcfl/flCited1− mice . These data were confirmed by western blot analysis using an antibody raised against total β-catenin ( Figure 6C ) or against the active form of β-catenin ( dephosphorylated sites pS33 , pS37 , pT41 ) ( Figure 6D ) and were verified with a second antibody against dephosphorylated β-catenin ( Figure S2A–B ) . There was no significant difference in the phosphorylated ( inactive ) form of β-catenin ( phosphorylated β-catenin at pS33 , pS37 , pT41 sites is degraded as a mechanism of regulating Wnt signalling ) between all genotypes ( Figure 6B ) . β-catenin phosphorylated at pS45 ( phosphorylated by casein kinase Iα as part of degradation pathway ) is significantly increased in AhCre+Apcfl/flCited1− compared to AhCre+WT and AhCre+Apcfl/fl . Given that β-catenin can also be phosphorylated by Protein Kinase A ( PKA ) at Ser552 and Ser675 which acts to inhibit ubiquitination and therefore increase levels of active β-catenin [33] , we also analysed levels of pS552 and pS675 and found phosphorylation at S675 significantly increased in AhCre+Apcfl/flCited1− tissues compared to AhCre+WT and AhCre+Apcfl/fl , demonstrating the ability of Cited1 to regulate β-catenin at multiple sites ( Figure 6B ) . We also measured the intracellular free β-catenin ( active β-catenin ) levels by pull-down with a GST-fusion protein of the inhibitor of β-catenin and TCF-4 ( ICAT ) . We observed a significant increase in free β-catenin in AhCre+Apcfl/fl and AhCre+Apcfl/flCited1− mice compared to AhCre+WT and noticeably a significant increase in AhCre+Apcfl/flCited1− compared to AhCre+Apcfl/fl mice ( Figure 6B ) . These data support our findings above which indicate that Cited1 deficiency increases the levels of active dephosphorylated β-catenin . These data demonstrated that the active dephosphorylated form of β-catenin in purified intestinal epithelial cells is markedly increased upon Cited1 deficiency . Although the level of dephosphorylated β-catenin is increased in Cited1− intestinal cells compared to WT mice , it is below that observed in the AhCre+Apcfl/fl intestinal cells ( Figure 6B , 6D ) . As Cited1− mice do not develop any intestinal phenotypes such as hyperproliferation this suggests that the level of Wnt activation in Cited1− mice is below the critical threshold required to induce neoplasia [34] . However , when Apc is deleted in Cited1 deficient mice ( AhCre+Apcfl/flCited1− ) the level of dephosphorylated β-catenin is greater than that observed with Apc loss alone , thus providing an explanation for the increased transcription of Wnt target genes observed in these mice ( Figure 5A ) .
Colorectal cancer is driven by a multiplicity of different biochemical pathways , however , key amongst these is the Wnt pathway , which we and others have previously shown to activate a set of c-Myc dependent genes which are critical for the early stages of colorectal cancer [6] , [35] . One of these genes is Cited1 , which has been found to interact at the protein level with β-catenin and thereby negatively regulate β-catenin transcription [5] . Its relevance to carcinogenesis has already been described as Cited1 up-regulation has been observed in various cancers [14] , [36] , [37] . Here , we have extended those observations and find that CITED1 is significantly up-regulated in colorectal tumours from patients and in intestinal adenomas developing in the ApcMin/+ mouse model [38] . We also previously found Cited1 to be over-expressed in intestinal epithelial cells immediately following deletion of the Wnt regulator gene Apc in AhCre+Apcfl/fl mice in a c-Myc dependent manner [6] . These data establish Cited1 as an immediate Wnt target gene in the intestine . On the basis of these data we hypothesised that Cited1 might control β-catenin activity and thereby modulate Wnt signaling activation and its effects on colorectal tumorigenesis . To investigate this , we used microarray analysis and quantitative PCR studies to show that loss of Cited1 on an Apc deficient background does indeed impact upon a range of oncogenic signaling pathways , including Wnt . Our array data therefore show multiple effects of Cited1 deficiency including negative regulation of the Wnt-pathway . To investigate the requirement of Cited1 during Wnt induced tumourigenesis , we analysed the effects of deletion of Cited1 in two well characterised mouse models of Wnt signaling activation; the ApcMin/+ mouse model of colorectal tumourigenesis and the AhCre+Apcfl/fl mouse , a conditional model of Apc loss in which the immediate phenotypic consequences of Apc deletion can be studied [3] . Surprisingly , we obtained the apparently paradoxical result that although ApcMin/+Cited1− mice developed fewer intestinal tumours ( associated with an increased life-span ) than ApcMin/+ mice , the phenotypes induced upon conditional loss of Apc ( including perturbed cell proliferation , apoptosis , differentiation and migration ) were enhanced , rather than diminished , with additional loss of Cited1 . Of note , we observed reduced capacity to differentiate ( reflected by a reduced number of goblet cells and enteroendocrine cells ) , but no difference in total paneth cell numbers , although we did observe a difference in the positioning of paneth cells , which may well reflect differences in the Wnt signalling environment . Our studies , suggest that a possible explanation for this apparent paradox is that the hyper-activated Wnt phenotype that occurs in the absence of Cited1 includes increased apoptosis . Several studies in cell culture systems already support such a model . For example , it has been reported that overexpression of β-catenin when transfected into cell lines leads to a 3–4 fold increase in cell death [39] . In addition , it has been demonstrated that high levels of c-Myc induce apoptosis in vivo [40] . This is consistent with our observations that c-Myc is overexpressed immediately following deletion of Apc in the intestine and that levels are significantly increased further with additional absence of Cited1 . We interpret our data to indicate that the increase in apoptosis may counteract the increase in proliferation to the extent that the overall effect is reduced development of Wnt transformed cells and consequently inhibition of tumourigenesis . The mechanism underlying such hyper-activation of Wnt signaling appears to be at least in part mediated through increased levels of dephosphorylated β-catenin , which we found to be up-regulated in AhCre+Apcfl/flCited1− tissue compared to comparable AhCre+Apcfl/fl tissue at both the transcriptional and protein levels . Thus , we found increased levels of the dephosphorylated forms of β-catenin ( T41 , S33 , S37 ) . These sites when phosphorylated are involved in the degradation of β-catenin by the proteasome pathway [31] . This is accompanied by an increase in the levels of phosphorylation at serine 675 which has been shown to be phosphorylated by protein kinase A ( PKA ) and which has been shown to lead to inhibiting of ubiquitination of β-catenin causing its accumulation and subsequent Wnt signalling activation . We therefore show that Cited deficiency increases the pool of active β-catenin , consistent with the enhanced Wnt pathway activation we observe . It does however remain possible that Cited1 may in addition be mediating its effects downstream of β-catenin . We cannot rule out the possibility that the other pathway changes we observe are responsible for the reduction in tumourigenesis in ApcMin/+Cited1− mice , as the effects of loss of Cited1 are not exclusive to the Wnt pathway . We also cannot rule out that the effects we observe may be secondary to Cited1 deletion . For example , it is possible that some of the changes we observe may be due directly to the upregulation of c-Myc rather than a direct consequence of Cited1 loss . Functional delineation of the precise relevance of all the changes we observe requires multiple crosses onto the relevant pathways to probe such dependency . Finally , AhCre+WT , AhCre+Apcfl/fl , AhCre+Cited1− , and AhCre+Apcfl/flCited1− mice were maintained on an outbred background , and because the comparisons of genotypes within the same littermates was restricted due to the small number of litter size , we cannot completely rule out the effect of gene modifiers on the Cited1 loss phenotype . Our current primary hypothesis is that Cited1 deficiency mediates its effects upon adenoma formation primarily through the apparently paradoxical derepression of the Wnt pathway . This result is however consistent with a “just right model” wherein a specific level of Wnt signaling activity is required for maximal tumour development and those levels of Wnt signaling above or below this level compromise tumour growth [41] . This model is further supported by recent studies on a novel mutant Apc mouse ( Apc1322T ) , which has reduced Wnt signaling compared to ApcMin/+ littermates , but surprisingly develops significantly more intestinal tumours [42] . Recently it was shown by Leedham et al [43] that in normal mouse intestine , stem cell markers and Wnt target genes are expressed in a physiological gradient compatible with normal intestinal homeostasis . Pathological activation of Wnt activity using the Ctnnb1Δex3 mouse model led to variable gradients in stem cell number and Wnt signalling activity which influenced tumour susceptibility , with regional differences in tumour predisposition throughout the length of the intestinal tract . This data , which supports the just right model hypothesis , may explain the variation we observe in the tumour distribution in the ileum between the ApcMin/+ and ApcMin/+Cited1− mice models . These observations clearly show that there is not a simple linear relationship between Wnt pathway activity and tumour burden . Our data is consistent with another version of this “just right” concept where perturbation of Cited1 leads to increased dephosphorylated β-catenin and hyper-activation of the Wnt pathway to a level that is incompatible with maximum tumour growth ( Figure 7 ) . Notably , this relationship appears specific to the intestine as similar analysis of kidney tumorigenesis in these mice showed no effect of Cited deficiency ( Figure S3 ) . Furthermore , to define the precise relationship between Wnt levels and tumourigenicity will require mouse modelling experiments in which Wnt activity is precisely regulated at numerous levels . Wnt/beta-catenin signalling plays a key role in the homeostasis of the intestinal epithelium and its role in the fate and maintenance of the stem cell compartment have been clearly demonstrated [44] . Our data clearly show that Cited1 is an immediate target of Wnt signalling and is an important regulator of the Wnt pathway . The loss of Cited1 has a direct impact on stem cell status in the small intestine as we have found several stem cell markers to be upregulated including Lgr5 ( Gpr49 ) , Musashi and Olfm4 . These alterations in expression could be a direct consequence of the ‘hyper’ activation of the Wnt pathway we observe after combined loss of Apc and Cited1 . This would implicate Cited1 as an important player in Wnt dependant stem cell maintenance in the small intestine . This has been already suggested in the developing kidney where Cited1 may contribute to the maintenance of the self-renewing capping mesenchyme [18] . By regulating the Wnt pathway , Cited1 may be an important regulator of the self-renewal compartment in the crypt of the small intestine . Our data show that Cited1 deficiency represses tumourigenesis . The consequences of Cited1 deficiency are diverse , but in particular impact upon Wnt pathway activity . We propose a model whereby loss of Cited1 , in the context of deregulated Wnt signaling , hyper-activates the Wnt pathway resulting in apoptosis of Wnt induced transformed cells and thus inhibits tumourigenesis . As Cited1 mice are fertile and viable this suggests that Cited1 represents a possible target for therapeutic intervention , where Cited1 inhibition induces cytotoxic effects due to very high Wnt signalling .
Total RNA samples from patient colorectal tumour tissues were obtained from the Cancer Tissue Bank Research Centre ( CTBRC ) . All colorectal cancer tissues and adjacent uninvolved colonic mucosa were obtained from surgically removed specimens with informed patient consent . Uninvolved colonic mucosa was generally 5–10 cm away from the malignant tissue . All experiments were performed under the UK Home Office guidelines . Mice were obtained and genotyped as follows: Cited1 null ( Cited1− ) [9]; ApcMin/+ [38]; AhCre transgene ( AhCre+ ) [45]; Apc580S allele [46]; β-catfl/fl [47]; ApcMin/+ and ApcMin+Cited1− mice were maintained on an inbred C57BL/6J background and were confirmed as congenic for the C57BL/6 Mom-1 allele via PCR analysis . Mice were sacrificed at ill-health . Intestine were fixed in Methacarn ( methanol-chloroform-glacial acetic acid [4∶2∶1] ) , and the lesion numbers were scored macroscopically . To study the role of Cited1 after the early loss of Apc , AhCre+WT , AhCre+Apcfl/fl , AhCre+Cited1− and AhCre+Apcfl/flCited1− mice were generated and maintained on an outbred background . Cre activity was induced by three intraperitoneal injections of 80 mg/kg β-naphthoflavone within 24 h and mice were taken Day4 or Day5 later . Tissues analysed were from age ( 8–12 weeks ) , sex ( males ) , background and genotype matched animals , however these were not always littermates . Apoptosis was scored from H&E or after anti cleaved-Caspase3 immuno-staining as previously described [3] . For proliferation analysis , mice were injected with 0 . 25 ml of BrdU ( Amersham ) before culling and were taken either 2 hrs ( day4 ) or 24 hrs ( day5 ) after BrdU injection . Staining was performed as previously described [3] . The number of cells in AhCre+Apcfl/fl and AhCre+Apcfl/flCited1− hyperplastic area was scored using the position of the last BrdU positive cells in the hyperplastic area . For each analysis , 25 full crypts or areas were scored from at least 3 mice of each genotype and time point . In situ hybridization of Olfm4 and Cited1 in the small intestine was performed for all genotypes using sections embedded in paraffin sectioned at 5 µm . Olfm4 hybridization was performed as described in Gregorieff et al . , 2005 . [48] . Cited1 hybridization was performed using a probe against the sequence deleted in the Cited1− allele designed by Advanced cell Diagnostics inc ( ACD ) . RNAscope 2 . 0 FFPE Reagent Kit – Brown kit was used according manufacturer instructions . Negative control Probe-DapB was used together with a positive control probe Polr2a from the ACD manufacturer . The DNA microarray were performed from three mice of each genotype using Mouse Genome 430 2 . 0 Affymetrix chips at Liverpool Microarray Facility according to the manufacturer's instructions . The Microarray data were analyzed using AffylmGUI ( Affymetrix linear modeling Graphical User Interface; http://bioinf . wehi . edu . au/affylmGUI/#citation ) [26] . The p values presented have been corrected for multiple testing using the BH method to control the false discovery rate . The B statistic is the log odds that the gene is differentially expressed and is adjusted for multiple testing using the assumption that 1% of genes are expected to be differentially expressed [26] , [49]–[51] . Microarray data were deposited in MIAME format at www . ebi . ac . uk/arrayexpress/ ( Accession Number: E-MEXP-3202 ) QPCR protocols , routine methods and a description of the statistical analyses used are provided in Protocol S1 . List of primers for Taqman RT-QPCR , Sybr green RT-QPCR , and Cited1 semi quantitative RT-PCR are provided in Figure S4 . Ingenuity pathway analysis ( IPA ) software ( www . ingenuity . com ) was used to determine which signaling pathways were affected by the loss of Cited1 in AhCre+WT or AhCre+Apcfl/fl mice . The comparative ( AhCre+WT vs AhCre+Cited1− and AhCre+Apcfl/fl vs AhCre+Apcfl/flCited1− ) data from the microarray analysis were filtered for a p value of less than 0 . 05 and imported into the IPA software . The significance of the association between the data set and the pathway was measured in 2 ways: by the ratio and by a p value . The ratio corresponds to the number of genes from our data set that map to the ingenuity pathway divided by the total number of genes that map to the Ingenuity canonical pathway . The p value is calculated by a right tailed Fischer's exact test . The p-value associated with a pathway is a measure of the likelihood that the association between a set of focus genes in your experiment and a pathway is due to random chance . Analysis of biological function , localization , and posttranslational modification of the different forms of β-catenin were carried out as previously described [32] . Two additional assays were included in the analysis . Anti-dephospho S33/S37 and T41 ( Cell Signalling Technologies ) was used as an additional capture antibody to measure dephosphorylated β-catenin and GST-ICAT was employed as an additional bait protein to study free β-catenin . | Colorectal cancer is the fourth leading cause of cancer related deaths worldwide , and a key genetic change associated with this disease is mutation of the gene APC . APC encodes a protein which plays a regulatory role in the Wnt signalling pathway . To better understand the mechanisms leading to colorectal cancer after APC loss , we have used a mouse model in which we deleted Apc in the bowel and which developed several characteristics of early stage cancers . Here , we show that after Apc loss , the expression of another gene , Cited1 , is increased in mice and human colorectal tumours . To study the role of Cited1 in bowel cancer after loss of Apc , we generated mice mutant for Apc ( Min ) or mutant for Apc and Cited1 ( MinCited1 ) . We observed that MinCited1 mice developed fewer intestinal tumours and lived longer than Min mice suggesting that Cited1 is pro-tumourigenic . However , we also observed that Cited1 deficiency actually increased many of the aspects associated with loss of Apc , including deregulation of the Wnt pathway and cell death . To explain this apparent paradox , we propose a model whereby loss of Cited1 , in the context of deregulated Wnt signalling , ‘over-stimulates’ the Wnt pathway , the net effect of which is to inhibit tumourigenesis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"biology",
"cancer",
"genetics"
] | 2013 | Cited1 Deficiency Suppresses Intestinal Tumorigenesis |
The neurodegenerative disease Friedreich's ataxia ( FRDA ) is the most common autosomal-recessively inherited ataxia and is caused by a GAA triplet repeat expansion in the first intron of the frataxin gene . In this disease , transcription of frataxin , a mitochondrial protein involved in iron homeostasis , is impaired , resulting in a significant reduction in mRNA and protein levels . Global gene expression analysis was performed in peripheral blood samples from FRDA patients as compared to controls , which suggested altered expression patterns pertaining to genotoxic stress . We then confirmed the presence of genotoxic DNA damage by using a gene-specific quantitative PCR assay and discovered an increase in both mitochondrial and nuclear DNA damage in the blood of these patients ( p<0 . 0001 , respectively ) . Additionally , frataxin mRNA levels correlated with age of onset of disease and displayed unique sets of gene alterations involved in immune response , oxidative phosphorylation , and protein synthesis . Many of the key pathways observed by transcription profiling were downregulated , and we believe these data suggest that patients with prolonged frataxin deficiency undergo a systemic survival response to chronic genotoxic stress and consequent DNA damage detectable in blood . In conclusion , our results yield insight into the nature and progression of FRDA , as well as possible therapeutic approaches . Furthermore , the identification of potential biomarkers , including the DNA damage found in peripheral blood , may have predictive value in future clinical trials .
Friedreich's ataxia ( FRDA; OMIM# 229300 ) is the most common autosomal-recessively inherited ataxia beginning in childhood and leading to death in early adulthood . Patients exhibit neurodegeneration of the large sensory neurons and spinocerebellar tracts , along with variable systemic manifestations that include hypertrophic cardiomyopathy , scoliosis , and diabetes mellitus ( see http://www . ncbi . nlm . nih . gov/entrez/dispomim . cgi ? id=229300 ) . FRDA results from the partial loss of frataxin ( FXN; Entrez Gene ID 2395 ) , a small nuclear encoded 18-kDa protein targeted to the mitochondrial matrix [1] . A GAA triplet repeat expansion in the first intron impairs transcription of frataxin , resulting in a significant reduction in mRNA and protein levels [2]–[4] . The exact physiological function of frataxin is still unclear , but it has been shown to bind iron and play a role in iron-sulfur cluster ( ISC ) assembly [5] , [6] . A decrease in frataxin may also increase reactive oxygen species ( ROS ) produced by increases in bioavailable iron [5] , [7]–[10] and the lack of iron detoxification [11] . The conclusions of several studies indicate that a defect in ISC assembly is the primary event in frataxin-deficient cells [5] , [12]–[14] and that ROS production is a secondary event [8] , [15] . Napoli et al . [12] believe the dysfunction of biosynthesis of mitochondrial iron-sulfur clusters , and deficiency of ISC enzyme activity , produces a defect in heme , which in turn causes a loss of cytochrome C . Impairment of electron transport activity results in higher levels of ROS production [14] , and according to Napoli et al . [12] , it is the decrease in cytochrome C that leads to the unchecked increase in production of mitochondrial ROS in Friedreich's ataxia patients . This hypothesis is further supported by studies of yeast strains with reduced frataxin , which accumulate mitochondrial iron and generate reactive hydroxyl radicals that damage membranes , proteins , and mitochondrial DNA ( mtDNA ) , ultimately resulting in the decreased capacity for ATP synthesis through impaired oxidative phosphorylation [15] , [16] . Moreover , evidence consistent with nuclear DNA ( nDNA ) damage is demonstrated by decreasing the levels of frataxin in a RAD52 ( 854976 ) double-strand break repair deficient yeast strain , which results in rapid G2/M cell cycle arrest [16] . In FRDA patients , iron deposition is observed in neuronal and myocardial cells and suggests the potential for free radical damage [17] , [18]; however , we note that the case for oxidative stress has been somewhat controversial . Cell models support sensitivity to oxidative stress , and patient studies have found markers of oxidative stress [7] , [19] , [20] , but a conditional knock-out ( KO ) mouse model did not show oxidative stress , or improvement , when overexpressing superoxide dismutase ( SOD ) [21] . Recent studies have also failed to replicate the previous marker data [22] , [23] . Therefore , it is important to examine other markers of oxidative stress by more sensitive and specific means , such as testing for mtDNA damage in the patient . There is good evidence to suggest that hypertrophic cardiomyopathy , which leads to the death of most FRDA patients , is probably a consequence of iron-catalyzed Fenton chemistry causing damage to mitochondrial macromolecules followed by muscle fiber necrosis and a chronic reactive myocarditis [24] . More work is needed to understand the causes of the pathobiology associated with the progression of FRDA . While genome-wide scans in frataxin-deficient model organisms and mammalian cells have previously been published [15] , [25]–[27] , we report the first study involving transcription profiling of total blood from children with FRDA . These gene expression data were further validated in a second cohort of adults with FRDA , who were compared to an independent group of controls . Importantly , we observed previously unreported signatures of gene expression associated with DNA damage responses . Based on these results , we further analyzed patient mitochondrial and nuclear DNA from peripheral blood and detected high levels of damage as compared to control samples . These results provide insights into the nature of the disease and a working model for frataxin deficiency in humans .
We set out to identify mechanisms involved in the nature and progression of Friedreich's ataxia by analyzing global gene expression changes in blood samples from 28 FRDA children involved in an idebenone clinical trial [22] ( Table S1 ) . Blood samples were collected from the children prior to the administration of idebenone . The protocol only allowed one 8 . 5 ml sample of blood for the RNA isolation , which resulted in a limited amount of RNA for this study . Furthermore , control unaffected children were not included in this clinical trial; therefore , we used the youngest control adults available from an NIEHS sponsored study [28] for the gene expression analysis ( Table S1 ) . Significance Analysis of Microarray ( SAM ) [29] identified 1 , 370 differentially expressed genes at a false discovery rate ( FDR ) less than 0 . 023% ( Dataset S1 ) . A majority of genes , 899 , were downregulated in FRDA compared with control , while 471 genes were upregulated . We further investigated whether these altered transcripts ( FDR<0 . 023% ) were associated to specific gene ontology ( GO ) terms , at p≤0 . 05 , in order to assess the global impact of FRDA on gene expression . This analysis identified significant functional groups that included apoptosis signaling , transcription/RNA processing , cell-cell signaling , cell cycle , ubiquitin cycle , proteolysis/protein catabolism , response to stimuli , and fatty acid beta-oxidation ( Figure 1 ) . Although age-matched control children were not available for this study , we decided not to use data uploaded to GEO by other laboratories . It is important to minimize cross-platform differences , which can give rise to large technical variation that may obscure small biological differences in blood cells . Therefore , the controls we used were gene expression profiles from young adults processed in our same facility , on the same oligonucleotide chip design , by the same operator . However , we did assess what effect age might have on gene expression by using SAM to test for any association . This analysis was performed in the controls , and age was dichotomized by comparison to the median age of the controls . Age was not found to be associated with any gene expression value after multiple testing correction ( min q = 0 . 47 ) ; thus no age-specific gene expression changes were discerned in our control group . As an alternative approach to gene ontology enrichment analysis , the microarray data for children with FRDA were further analyzed by employing Gene Set Analysis ( GSA ) , a tool that uses predefined gene sets to identify significant biological changes in microarray datasets [30] , [31] . We searched for significantly associated gene sets from Molecular Signatures Database subcatalog C2 , a database of 1684 microarray experiment gene sets , pathways , and other groups of genes [31] . The analysis yielded many biologically informative sets ( Dataset S2 ) consisting of genes enriched in brain cortex and heart atria , as well as biological processes such as mitochondrial fatty acid beta-oxidation , and reactive oxygen species . The application of GSA also identified 23 gene sets associated to genotoxic stress response ( Table 1 ) . P53genes_all is composed of transcriptional targets of p53 ( 7157 ) , a regulator of gene expression in response to various signals of genotoxic stress , with genes such as GADD45A ( 1647 ) , PMAIP1 ( 5366 ) , and SESN1 ( 27244 ) displaying repressed expression . Genotoxins_all_24h_reg consists of downregulated genes regulated in mouse lymphocytes at 24 hours by cisplatin , methyl methanesulfonate ( MMS ) , mitomycin C , taxol , hydroxyurea , and etoposide [32] . Other gene sets consist of mostly downregulated genes in response to bleomycin , MMS , ultraviolet B ( UVB ) , and ultraviolet C ( UVC ) radiation , which were also downregulated in the FRDA dataset ( denoted by the negative GSA scores ) ( Table 1 ) . We next asked if there were genes in common across the 23 genotoxic-stress-response gene sets . Transcripts present in at least 25% of the gene lists ( 81 genes total ) were subjected to unsupervised clustering and displayed segregation into two groups of controls and patients ( Figure 2A , right panels , and Dataset S3 ) . Most of these genes fall in the gene ontology categories of transcription , signal transduction , and cell cycle ( data not shown ) . To validate the gene expression changes observed in the child FRDA cohort , we examined an adult FRDA cohort ( n = 14 ) . These patients were compared with a new group of 15 adult controls , obtained from an NIEHS sponsored study [28] ( Table S2 ) . SAM analysis yielded 2 , 874 genes at an FDR of less than 0 . 018% ( Dataset S4 ) . This dataset was also analyzed with GSA , yielding significant gene sets related to genotoxic stress response , DNA repair , insulin response , and apoptosis ( Dataset S5 ) . When we performed unsupervised clustering of the same list of 81 genes found in the children with FRDA , we observed a similar segregation of patients from controls in this independent group of adults ( Figure 2A ) , thus helping to validate this gene set . The FRDA adult and FRDA children's datasets were further compared by limiting to significant SAM genes in common and taking into account the direction of differential expression . As stated earlier , a total of 1370 probesets were found significant in the children's cohort ( FDR<0 . 023% ) , and the median FDR for these probesets in the second cohort was 0 . 34% . These analyses resulted in 228 mostly downregulated genes in common in both gene sets ( FDR<0 . 018%; overlap p = 0 . 007 – based on the probability of the hypergeometric distribution ) ( Figure 2B ) . This list of 228 genes was analyzed for specific GO category enrichment ( Figure 2C ) ( categories similar to those displayed in Figure 1 are not included; for the full list , see Table S3 ) and significantly associated to ubiquitin cycle , protein ubiquitination , and proteolysis/catabolism , as well as cell cycle . Using the Ingenuity Knowledge Base ( see Materials and Methods ) , we further mapped the genes to biological function categories and specific pathways , which included apoptosis signaling and oxidative phosphorylation ( Table S4 ) . Additionally , the Fisher's exact test was used to search the Molecular Signatures Database for gene sets enriched with genes found in the overlap . These results yielded similar genotoxic gene sets as those observed with the children's data as a whole ( Table S5 ) . Having seen a genotoxic stress response in the microarray data of the FRDA patients , we sought to validate these findings and test whether damage to the nuclear and mitochondrial genomes is elevated in patients with FRDA . Moreover , although genotoxic responsive gene sets in the children and adult cohorts had highly significant p-values , some had high false discovery rates , requiring further biological validation . Blood DNA from the same 28 children studied for gene expression was analyzed using a quantitative PCR ( QPCR ) assay to detect DNA damage [33] . In addition to the 28 children evaluated by gene expression profiling , we obtained 19 more DNA samples from affected children in the same clinical study ( Table S6 ) [22] . Only one 8 . 5 ml sample , per child , of whole blood was allotted for DNA isolation; all DNA samples were prepared by one person using the same protocol ( see Materials and Methods ) . Blood from 15 young adults was obtained from an NIH blood bank in Bethesda , MD and used as controls ( Table S6 ) . The QPCR technique we used has successfully identified lesions in nDNA and mtDNA , resulting from oxidative stress , in a number of organisms , including human and rat cell cultures [34]–[36] , yeast [16] , and mice [37] . However , this is the first time the assay has been used for DNA derived from human blood . The approach involves the amplification of a large 8 . 9 kb and 12 . 2 kb fragment for mtDNA and nDNA , respectively . Previous work by our group suggests that damage is distributed evenly throughout both genomes [34] , [38] , [39] . The large mtDNA amplification product constitutes ∼54% of the genome and is the representative of the overall genome . The primers used to amplify the product were designed to specifically avoid the D-loop region – a region that is often single-stranded and could increase DNA damage because of its high mutation frequency . Additionally , amplification of a short , ∼200 bp mtDNA fragment , which due to its small size has less chance of containing a lesion , is used to normalize for mitochondrial copy number in the amplification of the large mtDNA fragment . Oxidative damage induces a spectrum of lesions , such as strand breaks , abasic sites , and some base damage ( i . e . thymine glycol ) , which interfere with the progression of the thermostable polymerase to replicate the template DNA . Thus , during amplification , the presence of an oxidative lesion results in the inability of the polymerase to synthesize the template DNA . The final amount of amplified DNA is inversely proportional to the number of oxidative lesions . The QPCR gene-specific damage assay is based on differential amplification of target genes in affected controls as compared to a control population . Excess DNA damage in the affected group will show as a decrease in amplification as compared to the control group . A significant number of nuclear ( 0 . 53 lesions/10 kb ) and mitochondrial DNA lesions ( 0 . 81 lesions/10 kb ) was observed in the 47 FRDA children compared to 15 young adult controls ( Table S6 ) , with p<0 . 0001 , respectively , by Mann Whitney U-test ( Figure 3 ) . There was also a significantly higher number of mitochondrial lesions than nuclear lesions ( p<0 . 002 , Mann Whitney U ) , and both mtDNA and nDNA lesions were found to be highly correlated by Spearman's rank test ( Rho = 0 . 700; p<0 . 0001 ) . Since this study did not have age-matched control children , for the two DNA damage variables ( nDNA and mtDNA ) , we used a t-test to detect any association with age in the control group . This analysis was performed in the controls after stratifying the subjects as “higher” or “lower” than the median age of the group . Age was not found to be associated with mtDNA damage ( p = 0 . 98 ) or nDNA damage ( p = 0 . 10 ) . Finally , mtDNA and nDNA damage samples were classified as “high damage” or “low damage” if the lesions/10 kb of DNA was >0 . 85 or <0 . 85 ( based on the distribution pattern; data not shown ) , respectively . GSA was then used to test the association of the DNA damage to predefined gene sets . A positive score would indicate enrichment in samples with high DNA lesions/10 kb , and a negative score would point to enrichment in samples with less lesions/10 kb . This analysis showed higher levels of DNA lesions associating to gene sets involving neuronal and synapse formation , as well as several others regarding a genotoxic stress response ( p≤0 . 01 ) ( Dataset S6 ) . Since the bioinformatics analyses in this study were applied to SAM-derived data based on pooling the raw data from all 28 children ( Materials and Methods ) , we tested the hypothesis that a discrete set of genes would be differentially expressed in patients with the lowest levels of frataxin . We next generated a transcription profile comparing patients with the lowest amount of frataxin , analyzed by Real-time PCR , to those with the highest . The samples were stratified into two groups , based on the expression distribution , where the values formed two distinct modes ( see Figure S1 legend ) . A threshold of −2 . 5 was selected to separate these two modes , resulting in six patients considered “high-frataxin expressers” and 21 patients designated “low-frataxin expressers” ( Figure 4A and Figure S1 ) . Significant gene changes were determined using SAM at a cutoff of FDR≤8% ( p≤0 . 05 ) for a total of 973 genes . These genes were analyzed for gene ontology enrichment and mapped to pathways in the Ingenuity Knowledge Base . Top scoring categories and pathways included protein biosynthesis , oxidative phosphorylation , ubiquinone biosynthesis , nucleotide excision repair , and protein ubiquitination , all of which were downregulated in those patients expressing lower levels of frataxin ( Figure 4B ) . Since we had quantified the level of frataxin for each patient , we sought to find an association of these levels to the expression of each gene in our genotoxic signaling list ( Dataset S3 ) . A univariate linear model was constructed to test this association , and no genes in this set were found to be significantly associated with frataxin levels . The minimum p-value was 0 . 002 , which was not significant after multiple testing correction ( q = 0 . 20 ) . We further sought to find correlations of frataxin mRNA levels to all available clinical data for each child , including DNA lesions and disease duration , but only found an association , by univariate linear modeling , with age of onset and number of short GAA repeats . This resulted in p = 0 . 00277 and r2 = 0 . 305 for the relationship of frataxin mRNA levels to age of onset ( Figure 4C ) , and p = 0 . 00131 ( r2 = 0 . 344 ) for the relationship of frataxin mRNA levels with the short GAA repeat length ( Figure S2 ) . These correlations of frataxin mRNA to short GAA repeat length and age of onset are in agreement with other published studies [1] , [40] , [41] . Intrigued by these associations to clinical data , we wondered if we could directly associate the global gene expression data to all clinical data and decided to use a method called EPIG , which extracts microarray gene expression patterns and identifies co-expressed genes ( see Materials and Methods ) [42] . Indeed , not only did we validate the SAM gene lists for the FRDA children , but we extracted patterns that associated the mostly downregulated transcriptional changes to their frataxin levels , age of diagnosis , and International Cooperative Ataxia Rating Scale scores – a scoring method used to discern the level of disability in the FRDA patient ( Table 2 ) [43] . Patterns correlating levels of frataxin to gene expression encompassed 37 responsive genes . These genes were grouped most strongly in the GO categories of immune response and protein biosynthesis . Patient age of diagnosis and ICARS score associated with 98 and 48 genes , respectively , with protein biosynthesis as a highly significant category . Supervised correlation analysis was also performed , where correlation r values and p-values between each gene expression profile and ICARS score were calculated . This analysis generated a list of 144 positively and negatively correlated genes at a p-value threshold of 0 . 01 . Protein biosynthetic function was once again highly significant as a GO category , resulting from the input of these genes ( Table 2 ) . Genes correlated with ICARS are potential biomarker candidates that may help to classify the progression of the disease . EPIG was further utilized to extract patterns and compare between the FRDA children and FRDA adults . We were interested in finding biomarker gene candidates based on duration of disease and representative of severity . Of particular interest were genes that were very significant in both cohorts , regulated in the same direction , but displaying larger differential expression in the adults . A detailed examination of the gene profiles was employed using the signal to noise ratio , ANOVA , Student's t-test , and fold changes , and resulted in fifteen genes that are potential biomarkers of disease progression in need of further testing: SERPINC1 ( 462 ) , DHFRL1 ( 200895 ) , IRX2 ( 153572 ) , SGCE ( 8910 ) , ADAM23 ( 8745 ) , TBX3 ( 6926 ) , SLC5A4 ( 6527 ) , CCAR1 ( 55749 ) , MS4A2 ( 2206 ) , IMPACT ( 55364 ) , NDUFA5 ( 4698 ) , CEACAM6 ( 4680 ) , C10orf88 ( 80007 ) , ITGA4 ( 3676 ) , and CD69 ( 969 ) ( first seven genes are upregulated and the subsequent eight genes are downregulated ) . Entrez IDs for these genes and all genes described in this paper can be found in Table S7 .
Cortopassi and coworkers have previously published two global gene expression analyses of cells from FRDA patients [25] and tissue from frataxin-KO mice ( 26 ) . While direct analyses of these published studies with the work presented here is not possible , since these previous data were not apparently deposited in a public database , several points are worth noting . Tan et al . [25] reported 48 significant frataxin-dependent differentially expressed genes in at least two of the human cell types . In particular , they focused on seven downregulated transcripts belonging to the sulfur amino acid ( SAA ) and iron-sulfur cluster biosynthetic pathways . When they combined data from mouse and human frataxin-deficient cells and tissues , mitochondrial coproporphyrinogen oxidase ( CPOX; 1371 ) , which is involved in the heme pathway , and the homologue of yeast COX23 ( 856516 ) were most consistently downregulated [27] . These transcripts were not found to be significant in our dataset . The authors conclude that frataxin deficiency leads to heme deficiency . While our work with both FRDA patient cohorts did not show the heme pathway as significantly repressed , the downregulated mitochondrial pathways we did observe are easily affected by , or could contribute to , heme deficiency . Ultimately , these changes in heme biosynthesis could cause DNA damage [44] recapitulated in our patients . We wanted to further validate the datasets derived from children and adults with FRDA , so we analyzed ten FRDA lymphoblastoid cell lines , compared with seven age-matched controls , and data from a previous report involving two lymphoblastoid cell lines ( one control and one affected ) ( Table S8 ) [45] . Particularly interesting overlaps with the blood data were observed with GSA analysis , which also yielded significant gene sets related to genotoxic stress response ( for biologically informative sets in common between the FRDA children's data and the lymphoblastoid data , see Dataset S7 and Figure S3 ) . Furthermore , GSA analysis of the lymphoblastoid data also found an association to significant gene sets like electron_transport_chain , mitochondrion , and ubiquinone biosynthesis , which are indicative of the mitochondrial dysfunction expected in frataxin-deficient cells ( Dataset S7 ) . While the QPCR assay used in this study cannot directly identify the type of DNA damage inhibiting the progression of the thermostable polymerase , the increase in mtDNA damage , as compared to nDNA damage , is consistent with a large number of studies from our and other laboratories , indicating that mtDNA is more prone to oxidative stress [16] , [34]–[37] . The increased DNA damage observed in the children suggests oxidant injury in their blood cells , probably caused by an increase in bioavailable iron in the mitochondria . Persistent mtDNA damage in FRDA patients could impair mitochondrial function . Experiments with mammalian cell cultures , treated with hydrogen peroxide , indicate that relentless mtDNA damage decreases oxidative phosphorylation and ATP production ( unpublished observation ) . In vivo evidence of impaired mitochondrial ATP production has , in fact , been seen in the muscle of FRDA patients [46] , [47] and in KO mice [48] . Karthikeyan et al . [16] also demonstrate how yeast strains with reduced frataxin accumulate mitochondrial iron and generate reactive hydroxyl radicals , which damage cell membranes , proteins , and mitochondrial DNA , resulting in the decreased capacity for ATP synthesis through impaired oxidative phosphorylation . The same study further demonstrates how low levels of frataxin in a RAD52 double-strand break repair deficient yeast strain lead to rapid G2/M cell cycle arrest , which is consistent with nuclear damage . Moreover , reports depicting an increased sensitivity to gamma irradiation in FRDA skin fibroblasts , and the induction of chromosomal damage by mutagens in blood lymphocytes , support a hypothesis of increased susceptibility and/or altered DNA repair capacity in these patients [49] , [50] . The eukaryotic cell , in response to DNA damage , employs different strategies for damage recognition and repair in order to maintain the integrity of the genome . DNA damage sensors such as ATM ( 472 ) and p53 are crucial in detecting double-strand breaks and general DNA damage responses , respectively [51] . Although the ATM and p53 genes are not differentially expressed in the datasets presented here , we observed , via GSA and the Ingenuity Knowledge Base ( data not shown ) , that many interacting network partners of these two proteins are significantly altered . The significant increase in persistent DNA damage we found in FRDA patients , as well as the transcription profiling results , indicate altered repair capacity and altered apoptosis signaling events . Furthermore , a significant number of genes included in the Ingenuity biological function category of Cancer ( 104 and 475 genes in Cancer and its subcategories , p≤0 . 05 , for FRDA children and FRDA adults , respectively; data not shown ) may explain the malignant transformation potential of frataxin-deficient cells , both in vitro and in vivo [10] , [48] , [52] . This hypothesis is supported by work done in mice by Thierbach et al . [48] , albeit a mouse model completely deleted for frataxin . When they disrupted expression of frataxin in mouse hepatocytes , lifespan was not only reduced , but the livers had increased oxidative stress and mitochondrial dysfunction . This was paralleled by reduced activity of iron-sulfur cluster containing proteins and the development of multiple hepatic tumors . The authors also reported impaired phosphorylation of the stress-inducible p38 MAP kinase and suggest that frataxin may , in fact , be a mitochondrial tumor suppressor protein . Thus , while reports of cancer in FRDA patients are rare [53]–[56] , the incidence may be underestimated due to premature mortality of these patients in early adulthood . The overall decrease in transcription and DNA damage we observed are likely consequences of dysfunctional ISC biosynthesis and reduced activity of proteins containing iron-sulfur centers . In fact , several damage recognition and DNA repair proteins are iron-sulfur containing and could be directly linked to the DNA damage described [57]–[60] . Currently , we are analyzing protein levels , in frataxin-deficient cell lines , of a panel of iron-sulfur containing proteins important to DNA repair . Some candidate proteins with iron-sulfur centers include the MutY ( 4595 ) homologue ( a glycosylase in base excision repair ) ; the yeast protein , Rad3 ( 856918 ) , which is essential for viability , and its human homologues XPD ( 2068 ) and Fancj ( 83990 ) ( helicases involved in nucleotide excision repair and the Fanconi anemia repair pathway , respectively ) ; and Pri2 ( 853821 ) ( essential to RNA primer synthesis ) [57]–[60] . The likely sequence of events leading to the DNA damage we observed are as follows: 1 ) deficiency of frataxin generates a defect in ISC assembly and biogenesis [5] , [12]–[14]; 2 ) the dysfunction of biosynthesis of mitochondrial iron-sulfur clusters , and deficient ISC enzyme activity , produces a defect in heme and a lack of cytochrome C [12]; 3 ) impairment of electron transport activity , which is dependent on iron-sulfur biogenesis , and the decrease in cytochrome C , results in higher levels of ROS production [12] , [14]; and 4 ) due to lack in antioxidative capacity , which we explain below , eventual DNA damage occurs . Therefore , we believe the DNA damage in the blood of the Friedreich's ataxia patient is a secondary event to the primary one of frataxin depletion and neurodegeneration . However , the secondary event of cellular oxidative stress and DNA damage is a significant component to the underlying pathology of the disease . We further propose that the downregulation of many key pathways ( Figure S4A and Figure 4C ) and GO categories ( Figure 1 , Figure 2C , and Figure 4B ) , in this study , may suggest a systemic survival response to chronic genotoxic stress and consequent DNA damage . Chronic genotoxic stress in FRDA probably results from iron accumulation in the mitochondria , and it might be expected that cellular redox homeostasis , such as that regulated by NRF2 ( 4780 ) , would protect the cell from excessive reactive oxygen metabolites . However , we observed the downregulation of the NRF2-mediated oxidative stress pathway ( Figure S4B ) , strengthening published reports suggesting a disabled antioxidant defense response in FRDA [9] , [10] , [61] , including a recent study by Paupe et al . [62] showing that cultured fibroblasts from patients with FRDA exhibit hypersensitivity to oxidative stress because of an impaired NRF2 signaling pathway . Furthermore , Chantrel-Groussard et al . [9] found that reduced frataxin does not induce superoxide dismutases nor the import iron machinery by endogenous oxidative stress in FRDA fibroblasts compared to controls . Superoxide dismutase activity is also not induced in the heart of conditional knock-out mice [21] . Conversely , overexpression of human frataxin in murine cells increases antioxidant defense via activation of glutathione peroxidase and elevation of reduced thiols , and reduces the incidence of ROS-induced malignant transformation [10] . Sturm et al . [61] reported data strongly indicating that a reduction in frataxin does not affect the mitochondrial labile iron pool in human cell lines and suggests that these cells have a decreased antioxidative capacity . Overall , these studies support a mechanism by which iron-sulfur proteins are lost [13] and there are increased amounts of ROS and a disabled antioxidant defense system . Based on our blood analysis of FDRA patients showing chronic genotoxic stress responses and chronic DNA damage , we believe these stressors cause a genetic reprogramming of fundamental biological pathways as a protective survival response . A similar idea was reported by Niedernhofer et al . , [63] who analyzed a case of XPF/ERCC1 ( 2072/2067 ) progeroid syndrome and a knockout mouse model of this disease . They concluded that chronic DNA damage causes cells to deemphasize growth activities in order to ensure organismal preservation and maximal lifespan , despite an increase in cellular senescence and apoptosis . The level of injury in the cells of these patients is not only exacerbated by the loss of antioxidative defense , but also by the downregulation of oxidative phosphorylation and the shutdown of protein synthesis and translation , as was observed in the gene expression analysis of patients with lower levels of frataxin as compared to patients with higher levels . EPIG analysis further demonstrated a marked decrease in genes involved in protein synthesis , and genes encoding ribosomal proteins , correlating with frataxin levels , age of diagnosis , and ICARS scores . Many of the significant genes involved in the category of protein synthesis include the repression of several initiation factors . Paschen et al . [64] discuss how such events suggest the relationship between the shutdown of translation and induction of neuronal cell death . It is our hypothesis that such global responses are triggered by chronic stress . We also found interesting the effect frataxin deficiency has on ubiquitin cycle and protein degradation in both FRDA children and FRDA adults . Modifications to the function of ubiquitinating enzymes by oxidative stress have been reported [4] . Degradation of damaged proteins by the ubiquitin-proteasome system ( UPS ) is one of the most important processes in the cell , and a decreased capacity for protein degradation is related to several neurodegenerative diseases and pathologies of the inflammatory immune response [65] , [66] . In summary , this study provides the first evidence of increased mitochondrial and nuclear DNA damage , as well as gene expression patterns consistent with DNA damage , in peripheral blood cells of patients with FRDA . Analyses of clinical features and gene expression patterns correlate with age of onset and frataxin mRNA levels , as well as altered protein synthesis with frataxin levels , ICARS score , and age of diagnosis . Future studies with Friedreich's ataxia patients will help better define these gene sets and DNA damage as candidate biomarkers of disease severity and progression . Additionally , biomarkers are vital to the development of therapeutic approaches , and our study points to possible drug interventions , like modulating the ubiquitin-proteasome system or upregulating molecular chaperone activity , which may be as useful for FRDA as they are in other neurodegenerative diseases . However , the development of effective therapeutic approaches also depends on an enhanced understanding of signaling pathways and other cellular responses to chronic genotoxic stress .
Peripheral blood samples were collected from 48 children with FRDA participating in a randomized , placebo-controlled clinical trial for idebenone [registered with ClinicalTrials . gov , number NCT00229632 , and approved by the NIH Institutional Review Board at the National Institute of Neurological Disorders and Stroke ( NINDS ) , protocol # 05-N-0245] . Samples from 14 anonymous FRDA adults were collected in the reference center clinic dedicated to cerebellar ataxias and aspartic paraplegias at the University Salpêtrière Hospital in Paris; samples were exempted by the NIH Institutional Review Board at the National Institute of Environmental Health Sciences ( NIEHS ) , exempt # 3984 . All controls used for transcriptional profiling were young healthy adults from an acetaminophen study [28] , approved by the Institutional Review Board at the University of North Carolina , Chapel Hill , protocol # GCRC-2265 . Controls used for the DNA damage assay were obtained from an NIH blood bank . Blood and/or apheresis samples were obtained from healthy volunteer donors who gave signed consent to participate in an IRB-approved protocol for use of their blood in laboratory research studies; these samples were approved by the Institutional Review Board at the National Cancer Institute ( NCI ) , protocol # 99-CC-0168 . Peripheral blood samples were collected from 48 children with FRDA participating in a randomized , placebo-controlled clinical trial . All whole blood samples in this study were collected before administration of idebenone . A detailed description of all subjects and clinical endpoints was recently published [22] . Due to other endpoints , this study only allotted one 8 . 5 ml sample of blood from each patient for RNA isolation . RNA was isolated by one person , utilizing the PAXgene blood RNA isolation kit ( PreAnalytiX/QIAGEN , Hilden , Germany ) according to the manufacturer's protocol , including the optional on-column DNase digestion , except that the centrifugation time after proteinase K digestion was increased from 3 to 20 minutes in order to obtain a tighter debris pellet . RNA quality was assessed with an Agilent Bioanalyzer ( Palo Alto , CA ) to ensure that samples with intact 18S and 28S ribosomal RNA peaks were used for microarray analysis . Of the 48 patients , twenty samples were lost during the isolation procedures , leaving 28 high-quality RNA samples remaining . The demographics for these subjects are detailed in Table S1 . RNA was also isolated , using the same methods already described , from 14 anonymous FRDA adults ( Table S2 ) . All controls used were young healthy adults ( see demographics data in Table S1 and Table S2 ) from an acetaminophen study [28] . Two independent sets of control populations were used separately to compare to the children and the adult validation cohort . Gene expression profiling was conducted using Agilent Human 1A ( V2 ) Oligo arrays with ∼20 , 000 genes represented ( Agilent Technologies , Palo Alto , CA ) . Each sample was hybridized against a human universal RNA control ( Stratagene , La Jolla , CA ) . 500 ng of total RNA was amplified and labeled using the Agilent Low RNA Input Fluorescent Linear Amplification Kit , according to manufacturer's protocol . For each two color comparison , 750 ng of each Cy3- ( universal control ) and Cy5-labeled ( sample ) cRNA were mixed and fragmented using the Agilent In Situ Hybridization Kit protocol . Hybridizations were performed for 17 hours in a rotating hybridization oven according to the Agilent 60-mer oligo microarray processing protocol prior to washing and scanning with an Agilent Scanner ( Agilent Technologies , Wilmington , DE ) . The data were obtained with the Agilent Feature Extraction software ( v9 . 1 ) , using defaults for all parameters . The Feature Extraction Software performs error modeling before data are loaded into a database system . Images and GEML files , including error and p-values , were exported from the Agilent Feature Extraction software and deposited into Rosetta Resolver ( version 5 . 0 , build 5 . 0 . 0 . 2 . 48 ) ( Rosetta Biosoftware , Kirkland , WA ) . All gene expression data have been deposited in the public Gene Expression Omnibus ( GEO ) database and are available under the series ID GSE11204 . Supervised analysis to find genes associated with case versus control or low frataxin expression versus high expression was performed using Significance Analysis of Microarrays ( SAM ) after pooling the raw data [29] . The two-class unpaired SAM algorithm was used and the false discovery rate was set to less than or equal to 1% for all analyses . Gene Set Analysis ( GSA ) [30] was also performed for these comparisons to test the association of gene sets instead of individual genes . The database of gene sets used for GSA was obtained from the Molecular Signatures Database ( MSigDb ) [30] . Gene sets demonstrating a p-value less than 0 . 01 were considered significant . Biologically relevant themes in the lists of significant genes from SAM were analyzed with gene ontology tools , GoMiner and DAVID ( Database for Annotation , Visualization and Integrated Discovery ) [67] , [68] . GO terms with p≤0 . 05 for upregulated , downregulated , and/or combined direction of change were selected for analysis . Both tools group genes according to the GO categories of biological process , cellular component , and molecular function , based on ranking by a hypergeometric test p-value . These data were also uploaded into Ingenuity Pathway Analysis ( IPA ) software v 5 . 5 . 1 ( Ingenuity Systems , Redwood City , CA ) , a program that categorizes genes into biological functions but also enables visualization of biologically relevant networks and canonical pathways ( “canonical” implies “established” ) . Go to www . Ingenuity . com for specifics regarding the application . Unsupervised clustering and heat-map generation were carried out with Cluster and Treeview programs [69] . The levels of DNA damage were analyzed by Mann-Whitney U test or Spearman's Rank Correlation because the data is not normally distributed or homoskedastic . In order to test association between gene expression and age , we used SAM . For the two DNA damage variables we used a Student's t-test to detect association with age . All these analyses were performed in the controls , and age was dichotomized by comparison to the median age of the controls . A univariate linear model was constructed to test the association of each gene in the genotoxic gene set list ( genes found in 25% of gene lists from significant gene sets ) to patient frataxin levels ( determined by Real-time PCR ) . Correlations of frataxin mRNA levels to all available clinical data for each child were also performed by univariate linear modeling . In extracting gene expression patterns , EPIG [42] uses a filtering process where all profiles initially are considered as pattern candidates . Briefly , using all pair-wise correlations , any candidate profile , whose local cluster size is less than a predefined size Mt or its correlation with another profile is higher ( >Rt ) but has a lower local cluster size M , is removed from pattern construction consideration . Among the remaining profiles , EPIG then creates representative profiles for the corresponding local clusters and removes those profiles with a signal-to-noise ratio or magnitude less than given thresholds . After this filtering processing , the remaining profiles consist of the extracted patterns , which are used to be the representatives to each of the local clusters . Subsequently , EPIG categorizes each significant gene to a pattern , for which it has the highest correlation with the gene profile . A gene not assigned to any extracted pattern is considered an “orphan” if its highest correlation r-value is lower than the given threshold R . The frataxin probe on the Agilent chip was observed to lack sensitivity for both the individual lymphoblastoid cell lines from affected people ( data not shown ) and the whole blood from FRDA patients . We , therefore , decided to obtain relative gene expression levels of frataxin by TaqMan Real-time PCR [70] . The sequence information of the probe used for TaqMan is the proprietary information of PE Applied Biosystems , but we know it is 75 nucleotides that span across the exon 1 and exon 2 boundary . On the other hand , the probe on the microarray chip is mostly from the 3′ UTR of frataxin . One microgram of total RNA from 27 patients and 10 controls was used for reverse transcription with TaqMan Reverse Transcription Reagents ( PE Applied Biosystems ) . To determine relative frataxin mRNA levels , Real-time PCR was carried out using the ABI Prism 7900HT sequence detection system . Primer and probe sets for frataxin and glucuronidase-beta were purchased as pre-developed assays from PE Applied Biosystems . Relative quantification was obtained using the threshold cycle method after verification of primer performance , following the manufacturer's guidelines . The levels of frataxin obtained are relative to the average ΔCT from 10 controls . Total DNA from whole blood was successfully isolated from 47 children enrolled in the study and 15 adult controls obtained from an NIH blood bank in Bethesda , MD , ( Table S6 ) using the PAXgene blood DNA isolation kit ( PreAnalytiX/QIAGEN , Hilden , Germany ) according to the instructions of the manufacturer . Briefly , one 8 . 5 ml sample of whole blood , per child , was collected in PAXgene blood DNA tubes for DNA isolation; each blood sample was transferred to a processing tube containing a lysing solution . Lysed red and white blood cells were centrifuged , and the resulting pellet of nuclei and mitochondria was washed and resuspended . After digestion with protease , DNA was precipitated by addition of isopropanol and dissolved in water . All DNA samples were prepared by one person . DNA lesion frequencies were calculated as described previously [33] . Briefly , the amplification of patient samples ( Apatient ) was compared to the amplification of non-damaged controls ( Acontrol ) resulting in a relative amplification ratio . Assuming a random distribution of lesions and using the Poisson equation [f ( x ) = . e ˜−λ λx/x ! , where λ is the average lesion frequency for the nondamaged template ( i . e . , the zero class; x = 0 ) ] , the average lesion per DNA strand was determined by the following equation: . λ = . −ln APatient/Acontrol . Amplification of the large mitochondrial target was normalized to mitochondrial copy number by examination of a short mitochondrial target , which due to its short size , should be free of damage . | Friedreich's ataxia is an inherited disease that causes progressive damage to the nervous system and affects the muscles and heart . The disease is caused by a defect in the frataxin gene , which is involved in iron homeostasis and likely protects against reactive oxygen species . In order to identify mechanisms involved in the nature and progression of the disease , we performed transcriptional profiling and measurements of mitochondrial and nuclear DNA damage on blood cells from FRDA patients . Transcriptional profiling was performed on blood samples from a cohort of 28 children compared to a control group . These data were then validated with a cohort of 14 adults with FRDA compared to a second independent control group . DNA damage was assessed on the blood samples from the 28 FRDA children , plus an additional 19 affected children , by quantitative PCR ( QPCR ) . Transcriptional profiling revealed changes in gene expression consistent with the presence of genotoxic stress in FRDA patients . This finding was supported by the direct evidence that FRDA patients accumulated significantly higher levels of mitochondrial and nuclear DNA damage as compared to controls . The identification of potential biomarkers , including the DNA damage found in peripheral blood , may help identify therapeutic approaches for this devastating disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/gene",
"expression",
"molecular",
"biology/dna",
"repair",
"neurological",
"disorders",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"biochemistry/replication",
"and",
"repair",
"genetics",
"and",
"genomics/bioinformatics",
"cardiovascular",
"disorders/heart",
"failure"
] | 2010 | Altered Gene Expression and DNA Damage in Peripheral Blood Cells from Friedreich's Ataxia Patients: Cellular Model of Pathology |
Given the importance that the evolution of cooperation bears in evolutionary biology and the social sciences , extensive theoretical work has focused on identifying conditions that promote cooperation among individuals . In insects , cooperative or altruistic interactions typically occur amongst social insects and are thus explained by kin selection . Here we provide evidence that in Lutzomia longipalpis , a small biting fly and an important vector of leishmaniasis in the New World , cooperative blood-feeding in groups of non-kin individuals results in a strong decrease in saliva expenditure . Feeding in groups also strongly affected the time taken to initiate a bloodmeal and its duration and ultimately resulted in greater fecundity . The benefits of feeding aggregations were particularly strong when flies fed on older hosts pre-exposed to sand fly bites , suggesting that flies feeding in groups may be better able to overcome their stronger immune response . These results demonstrate that , in L . longipalpis , feeding cooperatively maximizes the effects of salivary components injected into hosts to facilitate blood intake and to counteract the host immune defences . As a result , cooperating sand flies enjoy enormous fitness gains . This constitutes , to our knowledge , the first functional explanation for feeding aggregations in this species and potentially in other hematophagous insects and a rare example of cooperation amongst individuals of a non-social insects species . The evolution of cooperative group feeding in sand flies may have important implications for the epidemiology of leishmaniasis .
Phlebotomine sand flies ( Diptera: Psychodidae ) are important vectors of leishmaniasis , a parasitic disease affecting an estimated 1–1 . 5 millions people and causing over 50'000 deaths worldwide each year [1] . Medical entomologists have long noticed that some sand fly species tend to arrive in waves to bite mammal hosts , that they feed in aggregations and that new-arriving flies promptly join existing groups [2] . Feeding aggregations have also been observed in the laboratory , including in the New World sand fly Lutzomyia longipalpis which is commonly reared for research on drugs , vaccines and chemical attractants ( Fig . 1 ) , The functional explanation of this ‘invitational effect’ [3] , which in L . longipalpis has been shown to be mediated by a pheromone produced on the female maxillary palps [4] so far has eluded scientists . Because sand flies are tiny ( 2–4 mm ) and produce large amounts of potent vasodilators [5] , [6] to facilitate blood intake , we hypothesized that individual flies might benefit from inviting other flies to feed in their vicinity by sparing costly saliva . The ‘cooperative feeding hypothesis’ would provide a simple explanation for the invitational effect observed in group-feeding sand fly species and , possibly , in a number of other group-feeding hematophagous dipteran species [4] , [7] , [8] , [9] . This hypothesis is also intriguing from an evolutionary point of view because cooperative interactions in insects are usually seen in social insects where it occurs among related individuals and can thus be explained through kin selection [10] , [11] . Given that sand flies , as most Diptera , do not provide parental care to their offspring [12] and are known to disperse over substantial distances in search of hosts [13] , [14] , cooperative feeding occurs amongst unrelated flies and would constitute a rare example of non-kin cooperation in insects . Finally , understanding the dynamics and function of feeding aggregations in sand flies and its consequences in terms of saliva usage bears special significance as salivary components have been implicated in the successful development of Leishmania parasites in the host [6] , [15] . In Old and New world sand fly genera Phlebomomus and Lutzomiya , co-injecting salivary gland extracts with Leishmania parasite enhances infection , which translates in increased lesion size at the site of the bite and higher parasite burden within those lesions [16] , [17] , [18] , [19] , [20] , [21] . This suggests that a behavioural trait such as cooperative feeding , which results in saliva being injected by multiple flies at a single biting site on the host , could potentially also play an important role in the dynamics and severity of Leishmania infections . We tested the cooperative feeding hypothesis in L . longipalpis , an important New-World vector of visceral leishmaniasis . Sand flies were blood-fed on a host either in groups or singly and the effect of group-feeding on their salivary use , feeding duration and fecundity was recorded . The results show that L . longipalpis individuals prefer to feed in aggregations and that by doing so greatly increase feeding efficiency , bloomeal profitability and fecundity . The implications of these findings for our understanding of the evolution and maintenance of cooperation in non-social organisms and for the epidemiology of leishmaniasis are discussed .
All animal handling and anaesthetizing procedures used in these experiments were approved by the Keele University sub-ethics committee and carried on under the conditions specified by the UK Home Office Animals Scientific Procedures Act . Experimental designs aimed at minimizing the number of hamsters used while achieving adequate statistical power . Two groups of golden hamsters , respectively 3-months and 12-months old were used in this study . Three-months old hamsters were naïve to sand fly bites whilst the 12-months old ones had been exposed to Lutzomyia sand fly bites on average once every 1–1 . 5 months prior to our experiments . At the time of the experiment they had not been exposed to bites for a period of 30 days . The L . longipalpis s . l . colony used in all experiments ( except some preliminary observations – see below ) was established in the 1980's from a large number of wild-caught females collected in Jacobina ( 11° 11′S , 40° 30′W ) , which is located in Bahia State , Brazil . This population is characterized by the 3M∝H , sesquiterpene 3-methyl-∝-himachalene male sex pheromone type as well as a characteristic mating song and may thus represent a distinct sub-species of L . longipalpis [22] , [23] . All flies were used 4 d after emergence and given sugar water until blood-feeding . Flies were picked at random but males were excluded from all experiments in order to avoid confounding effects of male-produced courtship pheromones on female feeding aggregations . Before initiating the study , we established that sand flies kept in our laboratory colonies showed the same inclination for feeding in aggregations as what has been reported for their wild counterparts . To do so , forty female flies were released in a cage with a sedated hamster covered with a slotted paper-towel so as to expose its 4 paws . The distribution of feeding sand flies on the paws was recorded at 1 min intervals . For the time interval at which the highest number of flies was feeding and in 3 replicates , a chi-squared test was conducted on the number of flies feeding per paw to reveal significant deviations from random distribution . Significant aggregation was observed in two of the tests and flies tended to aggregate in the third replicate ( Likelihood-ratio: X2 = 5 . 7 , n = 17 , P = 0 . 002 ) , ( X2 = 6 . 3 , n = 17 , P = 0 . 097 ) and ( X2 = 13 . 6 , n = 18 , P = 0 . 004 ) , yielding a combined p-value<0 . 001 . Feeding aggregation was also observed amongst a mixture of 20 individuals from two L . longipalpis colonies from Marajo , district of Pará , Brazil and El Callejon in Colombia . These populations are characterized by the cembrene-1 and 9MGB , sesquiterpene ( S ) 9-methyl-germacrene B pheromonal types and contrasting mating songs suggesting that they may also represent distinct sibling species of L . longipalpis [22] , [23] . Groups of 20 female flies were fed on a single exposed hamster paw . For single flies , 5 flies were introduced in the cage and when the first fly initiated feeding we carefully aspirated the remaining flies . Five replicates of group feeding alternating with single-fly feeds were conducted on 5 young hamsters ( 3 months-old ) . After each experiment 6 randomly picked group-fed flies and all single-fed flies were dissected in a saline solution and a digital picture of the salivary gland pair was taken . The surface of the salivary gland pictures was measured on screen using the programme ImageJ1 . 38 available at http://rsb . info . nih . gov/nih-image/ . All measurements were conducted blind with regard to the flies' original groups . The data was analyzed non-parametrically to account for non-normality . A design similar to that described above was used but in group-fed flies we first powder-dyed 10 of the 20 flies to facilitate behavioural observations . Group and single-fed flies were either fed on a young naïve or an older exposed hamster ( see details above ) to study the potential effect of an enhanced immune response due to previous exposures to sand fly bites . As above and to avoid biases , 4 hamsters ( replicates ) were used alternately to feed groups of flies or single flies . The time at which flies initiated and finished blood intake was recorded and the first 6–10 individuals that initiated feeding where captured with an aspirator as they completed their blood meal and left the host . Group and single-fed flies were set to lay eggs singly in a vial containing a damp filter paper folded to produce a concertina effect . All data were checked for deviation from normality , unequal variances and heteroscedasticity and analyzed using JMP6 . 0 available at http://www . jmp . com/ .
In line with the cooperation hypothesis , single flies presented a host were so reluctant to feed that we had to give them temporary companion flies to initiate blood-feeding . In contrast , group-feeding flies started feeding significantly faster ( Mann-Whitney: Z = 4 . 56 , n = 45 , P<0 . 001 ) ( Table 1 ) . The variance in the time at which blood-feeding started was also an order of magnitude smaller in group-feeding flies than in single-feeding flies indicating that they fed synchronously ( Bartlett: F1 , 44 = 21 . 5 , P<0 . 001 ) ( Fig . 2 ) . Since older hosts that have been more exposed to bites may exhibit a stronger immune response to bites or tougher skin , we also compared the time it took to initiate the blood meal in relation to host age . Given the unequal variances this was done non-parametrically across both experimental groups . There was no effect of host age on the time taken by flies to initiate their bloodmeal ( Mann-Whitney: Z = 0 . 09 , n = 45 , p = 0 . 927 ) . Group feeding strongly decreased the time taken to acquire a bloodmeal . Overall single-feeding flies spend 70 . 5% longer acquiring their bloodmeal than group-feeding ones . Host age also had a very strong effect on feeding duration . Single flies took much longer to feed , but particularly so when feeding on older pre-exposed hosts ( 2-way Anova: aggregation , T1 , 41 = −5 . 2 , p<0 . 001; host age , T1 , 41 = −6 . 7 , P<0 . 001; interaction , T1 , 41 = 5 . 0 , P<0 . 001; r-square = 0 . 674 ) ( Table 1 , Fig . 3 ) . Feeding on an older hamster resulted in a 219% increase in feeding time in this group , whilst in group-feeding flies feeding on young naïve hamsters , the duration increased by 31 . 9% only . Importantly , in group-feeding flies the order in which the flies initiated feeding had a significant effect on the duration of their blood meal . The later a fly joined a group and/or initiated feeding , the longer it took her to acquire a bloodmeal suggesting that there maybe a cost in delaying feeding or waiting for other flies to start feeding in cooperating groups ( regression: F1 , 22 = 7 . 12 , P = 0 . 014 ) ( Fig . 4 ) . Using digital pictures of dissected salivary gland pairs and imaging software ( see methods ) , we compared salivary gland use for a single bloodmeal in single and group-feeding flies as well as in unfed flies . All blood-fed flies were fed on young hamsters in this experiment . After blood-feeding , flies kept in groups or singly had significantly larger right glands than flies that were kept unfed suggesting that flies produced saliva whilst feeding instead of simply emptying existing reserves ( Kruskal-Wallis: X2 = 10 . 2 , n = 56 , P<0 . 006 ) . In contrast , in all groups the left gland was significantly smaller than the right gland ( Wilcoxon Sign-Rank: P<0 . 001 in all cases ) . The difference between the two glands was significantly smaller in unfed flies than in both groups of blood-fed flies suggesting that they produce but also use more saliva ( Mann-Whitney: P<0 . 001 in both cases ) ( Table 2 and Fig . 5 ) . There was no statistical difference between the right gland of flies fed on hamsters in groups or singly . However aggregation had a strong effect on the difference in size between the right and left glands , with single flies having an average 45% less saliva in their left salivary gland than those feeding in groups ( Mann-Whitney: X2 = 20 . 9 , n = 36 , P = 0 . 004 ) ( Table 2 and Fig . 5 ) . This comparison was particularly statistically powerful as it effectively controlled the data for variation in body size ( power equal to 1 ) ; something we could not achieve by controlling gland size by wing length because the two variables did not correlate ( Pearson correlation: P>0 . 05 in all three groups ) . The number of eggs laid by females following a bloodmeal was dramatically reduced in flies feeding singly ( 50 . 5% decrease ) . Fecundity was also significantly affected by host age although to a much lesser extend than feeding duration ( 2-way Anova: aggregation , T1 , 41 = 11 . 3 , P<0 . 001; host age , T1 , 41 = −2 . 6 , P = 0 . 013; interaction , T1 , 41 = 0 . 8 , NS , r-square = 0 . 778 ) ( Fig . 6 ) . Despite a non-significant interaction term in the latter analysis , the data suggest a much stronger effect of host age on fecundity in single-feeding flies than in aggregated flies ( Table 1 ) . Indeed , separate analyses of egg production for each experimental treatment revealed no significant difference between flies feeding in groups on younger and older hosts ( T-test: df = 20 , T = 1 . 2 , P = 0 . 232 ) whilst single flies were strongly affected ( T-test: df = 20 , T = 2 . 7 , P = 0 . 014 ) . In terms of profitability , single flies that fed on older exposed hosts produced on average 54 . 4% eggs per unit of time spent feeding than those feeding on young naive hosts ( Kruskal-Wallis: n = 21 , X2 = 10 . 1 , P = 0 . 002 ) ( Table 1 ) . Compared to that of group-feeding flies , the profitability of singly-feeding flies on old hosts was 72 . 9% lower than that of group feeding flies on older hosts and 82 . 2% lower than those feeding on young hosts ( Table 1 ) . The latter two groups did not differ significantly ( Kruskal-Wallis: n = 23 , X2 = 1 . 2 , P = 0 . 268 ) . There was a strong linear relationship between feeding duration and the number of eggs laid in both experimental groups ( GLM: aggregation , T1 , 42 = 12 . 5 , P<0 . 001; feeding duration , T1 , 42 = 3 . 2 , P<0 . 001; r-square = 0 . 790 ) ( Fig . 7 ) . Importantly , in group-feeding flies the order in which the flies initiated feeding had a no significant effect on fecundity suggesting that there is no fecundity advantage in delaying feeding or waiting for other flies to start feeding in cooperating groups ( regression: F1 , 22 = 0 . 02 , P = 0 . 903 ) .
Our results show unambiguously that the invitational effect observed in L . longipalpis is driven by the enormous benefits of feeding in aggregations over feeding singly . Sand flies are tiny flies and whereas the majority of other blood-feeding insects species rely on long piercing mouthparts for reaching into capillary vessels , sand flies use instead chisel-like mouthparts to cut and pierce the epidermis and dermis of their host in order to create a small hemorrhagic pool upon which they feed [24] . The injection of large amounts of their saliva , rich in anticoagulants , antiplatelets , vasodilators and immunomodulators in the wound , enables them to maintain a constant blood flow throughout their bloodmeal [5] , [25] . Here we show that , in L . longipalpis , feeding in group not only significantly reduces the amount of saliva used per individual during blood-feeding but also drastically decreases the amount of time taken to acquire their bloodmeal and sharply increases their fecundity . At present the costs of producing pheromones such as that responsible for sand fly feeding aggregations [7] is not known but given our results we can only assume that they are small or largely out-weighed by benefits . Our observations and the results of our experiments suggest that several mechanisms could lead to the fitness gains observed in group-feeding flies . Firstly , individual flies injecting saliva into the host do not only facilitate their own bloodmeal but also that of flies feeding next to them . This result may not be that surprising given that in humans and other mammals L . longipalpis bites often cause large erythemas suggesting that the effects of saliva are not narrowly localized [26] , [27] . The maxadilan protein alone , an extremely potent vasodilator and an important component of their salivary repertoire , has been shown to elicit such large erythemas both in humans and rabbits [27] , [28] . Thus , because the effects of saliva are not narrowly localized , group-feeding flies benefit from the combined anticoagulant , antiplatelet , vasodilatory and immunomodulatory effects of multiple bites and thus produce and spend less saliva acquiring their bloodmeal . As a consequence , they should be able to replenish their salivary glands at a faster rate and could invest more resources in egg production or body maintenance . At present we do not know why flies used saliva from their left gland preferentially . That a paired organ such as salivary glands is used asymmetrically seems counter-intuitive . In contrast to laboratory flies that are fed once , lay eggs , and usually die within their egg-laying pot , wild flies may have more opportunities to feed several times before ovipositing and may use the content of the right gland on those occasions . Another important source of fitness gain stems from the fact that shortened feeding duration may translate in increased survival . It is generally accepted that host behavioural defences are an important determinant of survival or feeding success in blood-feeding arthropods [29] , [30] , [31] , [32] . Thus , acquiring a bloodmeal faster could decrease the likelihood of eliciting potentially life-threatening host behavioural defences . Albeit feeding in groups may itself entail risks due to density-dependent host defences [30] , [33] , [34] , the much higher blood intake observed in group-feeding flies may overall still translate into better survival , hence higher reproductive value . Finally , group-feeding L . longipalpis females have much higher fecundity despite shortened feeding durations , which suggests that in addition to sparing saliva and feeding more efficiently , they could acquire larger bloodmeals and/or bloodmeals that are easier to digest . Group-feeding flies might not only overwhelm vasoconstriction factors but may also better counteract blood coagulation and immunity factors that would otherwise interfere with bloodmeal acquisition and digestion . Indeed , the much higher feeding durations and lower profitability observed in single flies feeding on older hosts that had been regularly exposed to bites further suggest that single flies may be adversely affected by increased host immune defences . The fact that the overall profitability of group-feeding flies was much higher than that of singly-feeding flies but that those fed on exposed hosts tended to produce less eggs per unit of time spent feeding than those fed on naïve hosts lends further support to that hypothesis . Milleron et al . [35] showed that mice sensitised by sand fly bites or by injection with the maxadilan protein , produced antibodies that reduced vasodilation and negatively affected egg production in L . longipalpis . Blood meal size was found to decrease by 10 . 16% in flies feeding on pre-exposed hosts whilst the number of eggs they produced diminished by 39 . 5% [35] . Taken together these results emphasize the impact of host immune defence on blood digestion and the fact that flies feeding cooperatively may be better able to counteract them . In addition to being able to locally inject a combined amount of saliva that is much larger than that of single flies , groups of flies probably inject saliva with a much higher antigenic diversity , and there is some evidence that this could be maximizing its effectiveness [35] , [36] , [37] . The maxadilan protein is highly polymorphic within and among sand fly populations and between sibling species [36] . Rabbits immunized with different maxadilan variants have been shown to develop antibodies specific to these variants [37] . The same variants bound to the antibodies contained in the serum of individual pigs and humans exposed to sand fly bites in a specific manner [37] . Thus , group-feeding flies may be better able to swamp the host antibody repertoire thereby increasing feeding efficiency and bloodmeal profitability . An abundant body of theoretical work in evolutionary ecology has focused on delineating conditions required for the evolution of altruism and cooperation amongst organisms [10] , [11] . When kin selection is not implicated , cooperation can evolve in the forms of reciprocity amongst individual that can recognize each other in repeated interactions , or alternatively , as what is know as weak altruism i . e . cooperative interactions amongst individuals that directly benefit from them [10] , [11] . Pheromone-mediated aggregations are not uncommon in insects , but only in rare cases have been directly associated with group-feeding benefits in the form of increased resource acquisition [38] . The cooperative interactions observed amongst sand flies are unique in that they do not involve kin selection and their dynamism and high efficiency is reminiscent of those observed in social animals and complex animal societies . Despite the common occurrence of feeding aggregations in nature , which we assume to be consisting largely of unrelated individuals , we were cautious about working on potentially inbred colony material . Our random picking of individuals from large pools of individuals insured that we were working with non-kin individuals but some relatedness would be expected in such an old colony . Consequently , in a preliminary experiment , we tested whether flies belonging to two different cryptic taxa of L . longipalpis would feed in ‘mixed’ aggregations . This was observed , proving that cooperative interactions can occur among totally unrelated flies in the laboratory and that they are likely to be found across taxa in mixed sand fly populations in the field . Cooperative behaviour can only evolve if it is ‘resistant’ to potential cheating individuals . Given that group-feeding flies decrease their salivary expenditure whilst nevertheless feeding faster , it is probable that the increase in feeding efficiency derived from co-injecting larger amounts of saliva into the host levels off at some point . Thus a cheating fly joining a group of flies that already injected their saliva into the host could potentially inject less saliva than other group members and nevertheless benefit from increased blood intake . Although we do not have data on the exact amount of saliva produced and injected by flies in relation to the order in which they initiate feeding in a group , the positive relationship found between feeding order and bloodmeal duration suggest that late-feeding flies may be disadvantaged . Furthermore , early-feeding flies were as fecund as those feeding later thereby confirming the apparent lack of benefits for potentially cheating flies . Thus the complex mode of action of salivary components , their importance for bloodmeal ingestion and their potential role in bloodmeal digestion may prevent cheating , making cooperative feeding an evolutionary stable strategy . Feeding aggregations have also been observed in other blood-feeding Dipteran species . McCall and Lemon [9] reported an invitational effect in black flies Simulium damnosum and the same patterns has been observed in the mosquitoes Aedes sierrensis [7] , Ae , aegypti [7] , Ae . cantans and others [8] . It is unclear at this point whether the invitational effect observed in these other blood-feeding Dipteran species is symptomatic of similar cooperative processes or if group feeding has a different function in those species . Salivary components are key to the development of Leismania parasites inside the host [6] , [15] . In both L . longipalpis and L . whitmani , another important vector species , co-injecting salivary gland extracts with Leishmania parasites enhanced the infection , which translated in increased lesion size at the site of the bite and higher parasite burden within those lesions [16] , [17] , [18] , [19] , [21] . Understanding how sand fly salivary components , the Leishmania parasite and the host immune response interact to determine the course of infections is crucial for the development of vaccines and/or drugs against the disease . It is probably also key to understanding why some human infections develop into the more severe visceral form and why that form is more abundant in certain geographical areas than others [39] , [40] . It has also been shown that Leishmania actively manipulates sand fly feeding behaviour through the secretion of a promastigote secretory gel ( PSG ) rich in filamentous proteophosphoglycan ( fPPG ) that is regurgitated alongside parasites when feeding [41] , [42] . Importantly , the PSG and particularly its fPPG component are strong parasite virulence factors [41] , [42] . Given the known importance of saliva and PSG for parasite development and the large density-dependent effects we report here on sand fly fitness , there is a possibility that fly density at the bite site could have an impact on the course of an infection . This work therefore begs for follow up studies aimed at exploring the interactive effects of salivary component polymorphism , PSG , and sand fly species composition and density , on disease transmission and pathology . | Understanding the processes that promote cooperation amongst animals in nature is a fundamental question in evolutionary biology with ramifications in the social sciences . In insects , cooperative or altruistic interactions are usually observed amongst genetically related social insects ( kin selection ) . Here we provide evidence that in Lutzomia longipalpis , a small biting fly and an important vector of disease in the New World , cooperative blood-feeding occurs in groups of non-kin individuals . Groups of 20 flies and single flies were fed on hamster hosts and we compared their salivary gland usage as well as the time taken to initiate a bloodmeal , its duration , and the number of eggs they produced . Our results show that flies feeding in aggregations benefit from decreased saliva expenditure and greatly enhanced blood intake and egg production . These effects were particularly strong on older hamsters pre-exposed to sand fly bites , suggesting that group-feeding flies may better overcome their stronger immune response . These experiments demonstrate that , in L . longipalpis , feeding cooperatively maximizes the effects of saliva injected into hosts to facilitate blood intake and to counteract the host immune defences , resulting in much increased fecundity . This constitutes the first explanation for the function of feeding aggregations in hematophagous insects and a fascinating example of cooperation amongst individuals in a non-social organism . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/protozoal",
"infections",
"ecology/population",
"ecology",
"ecology/evolutionary",
"ecology",
"ecology/behavioral",
"ecology"
] | 2009 | Cooperative Blood-feeding and the Function and Implications of Feeding Aggregations in the Sand Fly, Lutzomyia longipalpis (Diptera: Psychodidae) |
The protein kinases are a large family of enzymes that play fundamental roles in propagating signals within the cell . Because of the high degree of binding site similarity shared among protein kinases , designing drug compounds with high specificity among the kinases has proven difficult . However , computational approaches to comparing the 3-dimensional geometry and physicochemical properties of key binding site residue positions have been shown to be informative of inhibitor selectivity . The Combinatorial Clustering Of Residue Position Subsets ( ccorps ) method , introduced here , provides a semi-supervised learning approach for identifying structural features that are correlated with a given set of annotation labels . Here , ccorps is applied to the problem of identifying structural features of the kinase atp binding site that are informative of inhibitor binding . ccorps is demonstrated to make perfect or near-perfect predictions for the binding affinity profile of 8 of the 38 kinase inhibitors studied , while only having overall poor predictive ability for 1 of the 38 compounds . Additionally , ccorps is shown to identify shared structural features across phylogenetically diverse groups of kinases that are correlated with binding affinity for particular inhibitors; such instances of structural similarity among phylogenetically diverse kinases are also shown to not be rare among kinases . Finally , these function-specific structural features may serve as potential starting points for the development of highly specific kinase inhibitors .
The protein kinases constitute the largest enzyme family encoded by the human genome , with currently 518 known sequences , making up 1 . 7% of all human genes [1] , [2] . Because these protein kinases are intimately involved in cellular communication and regulation networks , the loss of normal kinase regulation has been implicated in a wide variety of pathological conditions . The large number of disease states found to be associated with kinase dysregulation has motivated the development of kinase-specific inhibitor compounds and research to discover protein kinase inhibitors has come to constitute 20–30% of the drug development programs at many companies [1] . The bulk of this effort has been directed at identifying inhibitors that bind at the atp binding site . However , due to the large number of existing protein kinase domains and the high degree of ( atp ) binding site similarity among them , designing highly selective inhibitors has proven difficult . For example , type I kinase inhibitors that only target the atp site have typically been found to have low selectivity across the kinome [3] . To increase inhibitor selectivity , type II inhibitors bind both the atp site and the immediately adjacent allosteric site . By also binding to the allosteric site , type II inhibitors are able to make additional highly specific interactions , thereby allowing them to be more selective [3] . Identifying highly specific structural features that can be uniquely targeted by inhibitors can be facilitated by comparative analysis of multiple kinase structures [4] . Comparative analysis of multiple structures allows for the identification of kinase structural features that are available for inhibitor targeting as well as insight into the effect of activation conformation dynamics , such as structural features that are only available for targeting in the inactive , DFG-out conformation [3]–[6] . Furthermore , combining structure and sequence is important when analyzing the kinases holistically due to the large degree of sequence divergence among the protein kinases [7] . A specific example of the insight derived from the comparative analysis of kinase structural features follows . Many of the effective inhibitor selectivity strategies involve exploiting the differences in the size of the atp binding site and targeting residue variability at a few key positions [3] , [8] . These structure-based comparison approaches have proven more useful than sequence-only measures of overall kinase similarity in evaluating the potential selectivity profile of inhibitors [8] . For example , the size of the gatekeeper residue directly moderates the availability of a hydrophobic pocket . Inhibitors having larger functional groups that bind this hydrophobic pocket may be able to select for the roughly 20% of protein kinases that have a relatively small gatekeeper residue ( e . g . , Gly , Val , Ala or Thr ) . This is because kinases with a larger gatekeeper residue ( e . g . , Phe ) do not have a large enough hydrophobic pocket to accommodate the inhibitor [8] . However , in order to select for an even more specific subset of the human kinome , it has proven necessary to take advantage of multiple structural features of the kinase binding site ( both atp and allosteric sites ) simultaneously [3] , [8] . A review of related work is given below . Recent work has illustrated that local structural similarity exists among phylogenetically diverse groups of kinases [5] , [9] and has highlighted the importance of large-scale , multiple-structure analysis of structure-affinity relationships among the kinases [9] , [10] . The PharmMap method [10] incorporates an aligned set of receptor-ligand co-crystals in order to identify pharmacophores common to a set of inhibitors . It has been developed to identify kinase inhibitor pharmacophores useful for selecting molecules for kinase screening panels . Huang et al . have utilized a knowledge-based approach to constructing a minimal binding site “fingerprint” that captures only a pre-specified set of well-studied , structurally selective features , such as the size and hydrogen-bonding ability of the gatekeeper residue [8] . The per-kinase fingerprint utilizes nine binding site features ( e . g . , residue type at gatekeeper position ) that have been shown to encode for selectivity among type I inhibitors . Anecdotally , kinases with similar fingerprints were shown to also have similar inhibitor selectivity profiles [8] , illustrating the utility of structural features in predicting and understanding kinase selectivity . Rather than relying upon pre-specified structural features , the recently developed Pocketfeature method decomposes a binding site into all possible “micro-environments” [11] . Pairs of kinase binding sites with highly similar sets of micro-environments were anecdotally shown to share a common inhibitor in 9 out of the top 50 most similar ( as calculated by Pocketfeature ) kinase binding site pairs . The CavBase [12] cavity matching approach has been used to cluster kinase atp binding cavities from multiple families across the kinome [5] . The kinase binding cavity clusterings derived from CavBase have been shown to generally agree with the sequence-derived kinase families and sub-families [5] , demonstrating that the overall kinase cavity is well-conserved within families . Recent work by Jackson et al . demonstrated a related structural clustering approach to predicting kinase inhibitor binding affinities [9] . Their geometric hashing approach to whole-site comparison of the atp binding pocket was demonstrated to be effective at identifying possible instances of inhibitor cross-reactivity and further emphasized the importance of taking into account subtle conformational changes in the binding site . However , despite the successes of existing approaches , several outstanding problems to identifying structural features of the kinase binding site that are predictive of inhibitor selectivity remain . The reliance upon a single , representative structure precludes the ability of existing methods to identify features common only to active conformations if an inactive structure is chosen as representative ( and vice versa ) . Additionally , choosing one representative structure disregards the role that binding site flexibility and plasticity may play in inhibitor interactions . Furthermore , the availability of multiple structures for individual kinases , exhibiting a variety of binding site conformations and bound ligands , provides a vast quantity of structure data that remains unexploited by existing approaches . Much of the difficulty in incorporating multiple conformations per individual kinase sequence into existing analyses stems from the non-uniform distribution of available kinase structures , with kinases such as CDK2 having more than a hundred available crystallographic structures while other kinases have only a single ( or no ) available structure . Finally , the availability of multiple kinase structures known to bind a given inhibitor and other kinase binding sites known not to bind that same inhibitor provides a rich set of structural examples and counter-examples beyond a single instance of pairwise similarity . Existing receptor-based methods focus on identifying meaningful pairwise similarity to a characterized kinase known to bind a given inhibitor . These methods currently do not account for the similarity of a given kinase binding site to other kinase sites that have been characterized to not bind the inhibitor in question . To this end we have developed the Combinatorial Clustering Of Residue Position Subsets ( ccorps ) method . ccorps solves the following problem: given a set of sequence-aligned kinase domains ( each having available PDB structures ) and a per-sequence inhibitor binding label ( either binds , does-not-bind or unknown ) , predict whether a given kinase domain binds the given inhibitor . Taking a set of kinase binding site residue positions as input , ccorps identifies clusters of kinases that share structurally and chemically similar subsets of residue positions . Given a particular kinase with unknown ability to bind a particular inhibitor , ccorps identifies kinase binding sites that share similar residue positions that are both known to bind and not to bind the inhibitor ( i . e . , it finds evidence both for and against binding a particular inhibitor ) . Finally , ccorps aggregates the residue position subset similarities for all possible -position subsets of the kinase binding site and predicts whether or not the given inhibitor will bind the given uncharacterized kinase binding site . In addition to the development of ccorps , three major results from the analysis of the human kinome are presented here . First , the identification of structural features within the kinase atp binding site that are correlated with the ability of certain kinases to bind specific inhibitors is demonstrated . Second , the existence of affinity-correlated structural features that are shared among kinases from distinct families of the kinome are enumerated , shown to be not rare and also to differ depending upon the inhibitor being analyzed . Third , the ability of ccorps to predict the affinity for kinases lacking affinity annotations is quantified and compared to a recent structural binding site analysis approach [9] . ccorps is demonstrated to make perfect or near-perfect predictions for the binding ability of 8 of the 38 kinase inhibitors studied , while only having overall poor predictive ability for 1 of the 38 compounds . The performance of ccorps for predicting inhibitor binding is compared to the method of Jackson et al . [9] and shown to meet or exceed the predictive ability for the subset of the 38 inhibitors also tested by Jackson et al . We also compare ccorps against a sequence-based approach and show that they have complementary strengths . Finally , ccorps is shown to identify shared structural features across phylogenetically diverse groups of kinases that are correlated with binding affinity for particular inhibitors; such instances of structural similarity among phylogenetically diverse kinases are also shown to not be rare among kinases . These function-specific structural features may serve as potential starting points for the development of highly specific kinase inhibitors and provide a basis for understanding patterns of inhibition by compounds such as sunitinib that target multiple kinases [13] . In contrast to existing pairwise binding site comparison approaches , ccorps provides an automated way to incorporate the similarity of an uncharacterized binding site to all characterized binding site structures , without the need to manually select a reference binding site . ccorps also accounts for the similarity of an uncharacterized binding site to both kinases that bind and those that do not bind a particular inhibitor . The high degree of atp binding site similarity shared across the protein kinases has made them a difficult target for which to design highly selective inhibitors . However , by identifying the patterns of local structural similarity among binding sites at the kinome scale , potential off-target interactions may be identifiable at earlier stages of pharmaceutical development and compensated for by further inhibitor modification . This would allow researchers to make predictions of binding affinity for a given ligand across the kinome with less experimental data . Furthermore , the emergence of kinase inhibitor resistance due to binding site position mutations may be better understood through the identification of kinases having similar structural features at the mutated positions . Structural features that are found to be unique to one or a small number of chosen kinases may provide the starting point for designing highly specific inhibitor interactions and therefore highly selective protein kinase inhibitors .
In order to identify locally similar features among substructures , all -sized combinations of the residue positions ( i . e . , combinations ) are generated . For example , given and , all 3-position subsets ( 1140 subsets ) are generated . Then , each of these position subsets are examined one-by-one . Continuing the example , given the position subset , all the protein structures are compared by examining the pairwise similarity of only positions 7 , 13 , and 14 in isolation ( i . e . , disregarding the other 17 positions ) . 3-position subsets are used in this work because they allow for a unique 3-dimensional lrmsd superposition and are more computationally tractable than subsets of size , while still allowing for binding site position decomposition . The dissimilarity between a pair of substructures is quantified by a combination of structural distance and chemical feature dissimilarity introduced in [26] . Specifically , the distance between any two substructures and is expressed as:The term is the pairwise-aligned side chain centroid lrmsd between the substructures . The remaining terms account for differences in the amino acid properties between the substructures and as quantified by the pharmacophore feature dissimilarity matrix as defined in [26] . For a given set of residue positions , we can calculate a matrix of pairwise distances between substructures using the distance measure defined above . Each row can be thought of as a feature vector that represents how different a protein is with respect to all other proteins in terms of the selected residues . The distance matrix is highly redundant . We use Principal Component Analysis ( pca ) to obtain a low-dimensional embedding . Our previous work [27] showed that this dimensionality reduction typically results in negligible information loss . Some technical details on how we correct for overrepresentation are described in Text S3 . The dimensionality-reduced feature vectors are then clustered to identify sub-groups that share strong structural similarity . The number of clusters is not known beforehand and the number of clusters will vary depending on the set of positions being compared . The Gaussian Mixture Model ( gmm ) clustering method implemented in the mclust package [28] was used to identify both the number of clusters present and the cluster memberships for each of the feature vectors . The above feature vector computation and clustering steps are repeated for each possible 3-position subset in order to compare all possible local structural features across all proteins . Structural variation in most subsets is not expected to be informative , either because no significant variation is present , or because spurious patterns can occur due to chance . However , functionally relevant structural variation can be detected with many different subsets and therefore distinguished from random patterns , as will be shown below . A cluster that is dominated by one annotation label can be used to predict the label for other structures in that cluster whose annotation is unknown . We therefore call such clusters “highly predictive” ( HPCs ) . Identification of HPCs is performed by selecting a minimum threshold for the label purity of clusters , and then selecting all clusters with equal or greater label purity than this minimum as HPCs; we used the strictest purity threshold possible ( 1 . 0 or 100% purity ) in this work ( see Fig . 1 ) . In general , purity is calculated for a multiset of labels , , as where is the multiplicity function of a label within the multiset and is the most frequent label within . As with the dimensionality reduction mentioned above , we need to correct for overrepresentation bias , the details of which are described in Text S3 . Purity alone does not account for the distinctness of the proteins in the cluster relative to the remainder of the dataset . For example , an hpc for label that partially overlaps a second hpc for label is less likely to be informative than an cluster greatly separated from the remainder of the dataset . The “degree of separation” or “distinctness” of a cluster was quantified by calculating the cluster silhouette score [29] . The mean silhouette score for a cluster was then used as a further selection criteria for identifying HPCs by removing potential HPCs with negative average silhouette scores ( malformed clusters ) . Each time an unlabeled protein falls within an hpc , that protein receives a single vote in favor of the majority label associated with the hpc . Because a protein can be a member of at most one cluster per -position subset , the maximum number of votes any protein can receive is equal to the number of possible -position subsets . For any given -position subset , it is possible that all clusters are HPCs or that no clusters are HPCs , depending on how the labels are distributed among the clusters . It is also possible that a protein may never fall within any hpc and therefore would receive zero votes for any label; such proteins are excluded from further analysis after the voting step . In the experiments described below this case rarely occurred . After tallying the label votes across all -position subsets , the label predicted for a given structure is determined by an SVM-derived decision boundary as described below . Given a set of label votes that have been determined for an unlabeled structure , the threshold ( s ) used to decide which of the two or more label classes to assign to the structure requires the definition of a decision boundary procedure . For example , given a set of annotation labels containing the label classes ( e . g . , indicating whether a kinase is known to bind to a given ligand ) , a simple decision rule may be that given a structure with true vote , predict the true label for that structure . However , determining a single threshold for deciding the number of label votes required to classify a structure into one of several classes is difficult to generalize . Because ccorps is a semi-supervised approach , the labels for the training structures are known and can be used to empirically estimate a vote count decision boundary . For example , given structure with known label , the number of times that appeared in a false hpc or a true hpc , across all -position subsets , can be calculated using the same approach as for unlabeled structures . The structure is then represented by an -dimensional vote vector , where each of the dimensions corresponds to the number of votes received for label ( for the case of kinase binding affinity , since we only have false and true labels ) . Application of this procedure to all labeled structures in the dataset provides an empirical basis for calculating a decision boundary in the vote space given the vote distribution for labeled structures . For example , the blue and red points shown in the scatter plot of Fig . 2 denote the vote vectors for training set substructures with known true and false labels , respectively . Given the vote vectors calculated for all labeled training set substructures in the dataset , it is then possible to train any number of classifiers in order to determine a decision boundary . To compute a decision boundary in the vote space for classifying unlabeled proteins , svms were selected in this paper . First , an svm ( linear kernel ) is trained using the vote vectors of labeled training set substructures . For example , the decision boundary determined by training an svm on vote vectors is shown in Fig . 2 as the bold , black line . Next , for an unlabeled substructure with a given vote vector , the label for the substructure can be predicted by determining which side of the svm decision hyperplane the unlabeled structure falls within . As illustrated in Fig . 2 , test vote vectors falling within the blue region will be predicted as having the true label and those falling within the red region , the false label . For training svms and calculating the -values of predictions made by those svms , libSVM [30] was used in this work . To validate the predictive ability of the structural features identified by ccorps an extensive dataset of 48 families was automatically constructed using the Pfam database [17] as a source of well-curated protein alignments . The annotation labels analyzed in the validation set were per-structure Enzyme Commission ( EC ) number classifications . Cross-validation was performed in order to evaluate the predictive power of ccorps and the utility of the distinguishing structural features identified . The overall classification accuracy of ccorps ( Text S2 ) when applied to the validation dataset demonstrates the ability of ccorps to identify structural features that distinguish functionally different protein homologs and the ability of ccorps to generalize to non-kinase protein families .
In order to enable the kinome-scale analysis of the protein kinase atp binding site presented here , a dataset of protein kinase binding site structures was assembled and then mapped to the affinity dataset of Karaman et al . [31] . Karaman et al . studied the affinity of 38 kinase inhibitors across 317 kinases and was one of the most comprehensive studies of kinase inhibitor selectivity at that time . Mapping a structure-affinity-phylogeny dataset by further incorporating the kinome family labeling of Manning et al . [2] has enabled the incorporation of all available crystallographic structures of the atp binding site and the analysis of shared structural features between major kinase families that is presented later in this paper . The process by which ccorps recognizes structural features that are associated with kinase binding affinity is through the identification of Highly Predictive Clusters ( HPCs ) . Given the 27-position binding site ( Fig . 3 ) , ccorps computes a clustering for each of the unique 3-position subsets . For example , consider the 3-residue substructure shown in Fig . 4A . The 3 residues shown correspond to 3 positions in the full kinome alignment and the corresponding residues for each structure in the kinome dataset are structurally compared to compute the substructure clustering shown in Fig . 4B . Each of the 1958 substructures within the kinase structure dataset is shown in Fig . 4B as a single point . The color of each point in Fig . 4B corresponds to the cluster assignment as computed by ccorps . Several informative observations regarding kinase structural diversity and its association to inhibitor binding affinity can be made by further examination of the substructure clustering shown in Fig . 4B . Immediately upon examination of the substructure clustering it can be noted that multiple distinct clusters of kinases exist . This observation alone indicates that the 3-position substructure that resulted in this clustering is highly diverse among kinase binding sites . Conversely , the presence of a single large cluster would indicate that the 3-position substructure was structurally conserved , exhibiting little variance across the kinome; indeed instances of clusterings with a single dominating cluster were also observed for some 3-position subsets . As demonstrated in Fig . 4C , where one randomly selected representative substructure is shown for each of the 21 clusters identified by ccorps , both the geometry and residue types vary significantly for this 3-position subset . By incorporating the affinity annotation labels for a particular inhibitor , further observations can be made about the association between the 3-position substructure shown in Fig . 4A and the kinases capable of binding that inhibitor . For example , mapping the affinity annotation labels for the inhibitor flavopiridol onto the substructure clustering ( Fig . 4D ) reveals that some of the clusters consist of only a single annotation label while others are a mixture of labels . In Fig . 4D , kinases capable of binding flavopiridol are colored red ( true label ) , kinases incapable of binding flavopiridol are colored black ( false label ) and kinases lacking affinity annotation are colored white ( undefined label ) . As shown in Fig . 4D , multiple clusters of purely true labels exist and are considered to be HPCs by ccorps . The existence of true-only clusters indicates that the 3-positions shown in Fig . 4A are a distinguishing structural feature for identifying kinases that bind flavopiridol . More interestingly , however , is the fact that multiple , structurally distinct versions of the same 3-position substructure exist for different kinases that all are capable of binding flavopiridol . This result is significant because it indicates that across the kinome there are different structural motifs that are associated with binding flavopiridol , as opposed to a single , shared structural motif across all flavopiridol-binding kinases . The ability to identify multiple structural motifs that can each be associated with inhibitor binding is a strength of ccorps . Furthermore , the existence of clusters containing only kinases incapable of binding flavopiridol can also be observed in Fig . 4D . These HPCs are also informative because they identify particular structural versions of the 3-position substructure in Fig . 4A that are all incapable of binding flavopiridol . Finally , clusters consisting of a mixture of kinases that are both capable and incapable of binding flavopiridol can be identified in Fig . 4D . For kinases in these clusters , the 3-position substructure is not a distinguishing feature of flavopiridol-binding ability . Finally , while flavopiridol is discussed in detail here for illustration , the same analysis was computed by ccorps for each of the 38 different inhibitors within the affinity dataset . For each of the inhibitors , the affinity labels can be mapped separately onto the same substructure clustering as shown in Fig . 5 . However , it should be noted that no information is shared between the results for different inhibitors in this work; that is , each inhibitor is computed in a fully separate ccorps computation ( the substructure clusterings do not vary , just the annotation labels ) . Examination of the affinity-annotated substructure clusterings shown in Fig . 5 reveals that the set of clusters which are HPCs varies greatly depending on the inhibitor considered . While the flavopiridol-annotated substructure clustering contains multiple HPCs for both true and false labels , the correspondingly annotated clustering for other inhibitors , such as VX-745 , PI-103 and imatinib , contain only false HPCs . This result demonstrates that the substructures that are informative of inhibitor binding are inherently inhibitor-specific . That is , a subset of binding site positions that are predictive for one inhibitor are not necessarily predictive for another inhibitor . It is important to note that Fig . 4 and Fig . 5 represent the same clustering for just one 3-residue substructure . However , all 2925 clusterings are computed and all HPCs detected in these clusterings are used to predict binding affinity . The particular three-residue subset shown in Fig . 4A was chosen because the resulting clustering exhibits a number of illustrative features . First , the clustering is relative “clean” with well-separated clusters . Second , it contains highly predictive clusters for both binding and not binding to flavopiridol ( the ocher cluster in the top-left and the red cluster in the bottom right of figure Fig . 4B , respectively ) . None of these features are essential for predicting binding affinity; all automatically selected HPCs in all clusterings are used to predict affinity , each casting one “vote . ” Numerous instances of cross-family affinity for both type I and II kinase inhibitors have been identified , as was clearly illustrated by the kinome affinity maps created by Karaman et al . [31] . It is important to identify structural features shared among phylogenetically diverse kinases that share affinity for a particular inhibitor , because they provide a basis for reasoning about inhibitor cross-reactivity when overall sequence similarity will be low . Furthermore , by identifying these shared structural features , it may be possible to rationally re-engineer the specificity of inhibitors by avoiding the targeting of these features , since they are not unique to the intended kinase target . In order to identify the number of instances of cross-family structural features that can be associated with specific inhibitor binding , the distribution of substructure clusters across all 3-position subsets was analyzed . Each individual cluster , across all 2925 clusterings and all 38 inhibitors , was evaluated to calculate the purity of both affinity labels and family-level phylogenetic labels . For example , a cluster containing 3 distinct kinase sequences with affinity labels and family labels {agc , camk , tk} would have an affinity purity of and a phylogenetic purity of 0 . 33 . By plotting the affinity and phylogenetic purity scores of each cluster ( separately for each inhibitor ) as shown in Fig . 6 , the distribution of clusters across the spectrum of possible scores can be evaluated . Note that only the clusters having a true label majority are plotted in Fig . 7 ( i . e . , a true label majority is purity in the true label ) . Additionally , Table 2 lists per inhibitor statistics for cluster distributions shown in Fig . 7 . In order to build intuition for interpreting the cluster distributions , the cluster distribution for VX-680 ( Fig . 7 ) is examined in more detail because it is representative of the distribution for many of the other inhibitors . As listed in Table 2 , 23 , 495 clusters were identified by ccorps that have purity in the true label for VX-680 ( hereafter referred to as true-majority clusters ) . Only these true-majority clusters are plotted in the cluster distribution shown in Fig . 7 , meaning the minimum “affinity purity” displayed in Fig . 7 is 0 . 5 by definition ( because only 2 different affinity labels exist , true and false ) . As can be seen in Fig . 7 , the vast majority of clusters identified by ccorps have low affinity purity as well as low phylogenetic purity . This is to be expected because highly conserved portions of the kinase atp binding site are known to exist . Structural features that consist of conserved residue positions will be common to many kinases from different families due to the fact that these positions are so heavily conserved , which explains the low phylogenetic purity of these clusters . Furthermore , these conserved features are unlikely to be correlated with the affinity for a particular inhibitor because most inhibitors have been engineered to not have broad cross-reactivity across the kinome . Staurosporine is an exception as it is a very non-selective inhibitor due to its interaction with highly conserved binding site features; the cluster distribution corresponding to staurosporine ( Fig . 6 ) is markedly different from the other inhibitors with most clusters having high affinity purity across a range of phylogenetic purity values . Examination of the extremes of the VX-680 cluster distribution reveals further insights into the frequency of structural similar features among kinases with different degrees of sequence similarity . Clusters having a phylogenetic purity of 1 . 0 ( i . e . , all proteins belong to the same family ) but having low affinity purity exist , and for VX-680 276 such clusters were identified by ccorps . This observation is interesting because it illustrates that kinases sharing sequence similarity ( relative to kinases outside the family ) have multiple common structural features that are not informative of the ability of these kinases to bind VX-680 and are therefore unlikely to be good features to target in design studies . Because ccorps only incorporates clusters with high affinity purity ( i . e . , HPCs ) , these conserved structural features that are not indicative of VX-680 binding are ignored by ccorps when predicting affinity for unannotated kinases . This observation can also be made for each of the other inhibitors as shown in Fig . 6 . Another interesting extreme of the VX-680 cluster distribution to examine is the existence of HPCs that are phylogenetically diverse . The HPCs selected by ccorps correspond to the right-most column of points in Fig . 7; these clusters all have an affinity purity of 1 . 0 for VX-680 and therefore contain only structures with known VX-680 affinity . As can be noted in Fig . 7 , HPCs exist at a range of phylogenetic purity values . ccorps identified a total of 2707 HPCs for VX-680 , and 1786 ( 66% ) of these HPCs contain proteins belonging to two or more distinct kinase families . This result demonstrates that ccorps is capable of identifying cross-family structural features that are associated with VX-680 binding . Furthermore , this result is not unique to VX-680 . As shown in Fig . 6 and tabulated in Table 2 , cross-family structural features associated with inhibitor binding were identified for all of the inhibitors tested with the exception of GW-2580 , for which no true-majority clusters were identified . Examination of the cluster distributions across each of the inhibitors reveals a wide range of observations . While many inhibitors have a cluster distribution similar to that of VX-680 , for some inhibitors ccorps identified relatively fewer true-majority clusters . For example , only 133 clusters with affinity purity 0 . 5 were identified by ccorps for SB-431542 and all of these happen to be HPCs . However , even among this relatively low number of HPCs , 69 ( 52% ) of the clusters contain kinases from two or more families . As demonstrated by the corresponding distributions for all 38 inhibitors in Fig . 6 , such shared structural similarity is not rare . The approach used by ccorps to classify an unlabeled kinase is to identify the cluster to which the unlabeled kinase belongs . If the associated cluster is an hpc , the label for the hpc is transferred to the unlabeled kinase . Non-informative clusters containing a mix of labels ( non-HPCs ) do not contribute to the label prediction process . This “co-clustering” analysis approach is repeated for all of the 2925 substructure clusterings and the final label prediction for an unlabeled kinase is then selected as detailed in Methods . The ability of ccorps to predict the binding of each inhibitor for proteins within the annotated structural dataset was assessed using the cross-fold validation approach described in the following section . For each of the 38 inhibitor annotation label sets , an independent evaluation of ccorps was performed . No information was shared among the evaluations in order to validate the predictive ability of ccorps to identify structural features predictive of the binding ability of each inhibitor independently .
Identifying structural features of the kinase binding site that directly or indirectly mediate the binding ability of inhibitors is a significant component in developing and optimizing kinase inhibitors . Given the increasingly large number of available kinase structures , kinome-wide comparative binding site analysis is now possible as has been demonstrated here . By combining available structure data with large-scale inhibitor affinity data , it becomes possible to automatically learn the features of the kinase binding site that predict the binding ability of a given inhibitor . This is useful for predicting whether kinases whose binding affinity is unknown will bind to a given drug , but , perhaps more importantly , knowing the structural basis for binding to a particular drug can be exploited in the design of analogs that bind more strongly and have fewer off-target interactions . This information could further improve well-established structure-based computer-aided drug design methods , where it is challenging to develop reliable models for the contributions of individual interactions or groups of interactions between inhibitor and protein to binding affinity . ccorps has been demonstrated here to be capable of learning the features of the kinase binding site that are informative of inhibitor binding across a set of 38 inhibitors . Furthermore , the binding site features selected by ccorps as informative of inhibitor activity/inactivity have been shown to be interesting in and of themselves , for example , the existence of residue triad clusters that are unique in kinases capable of binding a given inhibitor but that exist within kinases from different major branches of the kinase family tree . The identification of such shared binding site features among sequence-diverse kinases is an important contribution for structure-based methods because of the relative difficulty of identifying small subsets of sequence non-contiguous but spatially compact positions that are correlated with a given indicator , such as inhibitor binding ability . The complete set of 41 , 964 true-majority HPCs that contain kinases from two or more of the kinome families as defined by Manning et al . [2] is provided as Dataset S1 to facilitate further analysis of these phylogenetically diverse structural features that distinguish kinases binding each of the 38 inhibitors . As was demonstrated here , ccorps is capable of incorporating all of the available protein kinase structure data , so as to operate at the kinome scale , and then using this data to construct highly accurate predictors of kinase affinity for a variety of different small molecule inhibitors . While ccorps relies upon the aggregation of structural similarity that coincides with affinity similarity to build predictors , the individual instances may be informative in and of themselves . Further analysis of the vast number of structurally similar features shared among phylogenetically distant kinases may provide additional insights into the structural mechanisms of inhibitor recognition occurring across the kinome . The existence of affinity datasets containing structurally similar inhibitors , that differ by only one or a small number of chemical substitutions , provides the opportunity to associate specific structural features identified by ccorps with specific inhibitor pharmacophores . A recent approach by Milletti and Hermann [6] has been demonstrated to identify specific chemical transformations that can be associated with selectivity differences . In future work we will seek to further incorporate this cross-inhibitor level of analysis and broaden the scale of the structure dataset by further incorporating newly available kinase crystallographic structures . Several potential optimizations of ccorps may increase its inhibitor binding prediction performance on broad spectrum inhibitors . For the 38 inhibitor dataset analyzed in this paper , the number of HPCs identified was well correlated with the number of kinases inhibited ( ) . That is , ccorps tended to perform less well on inhibitors for which large numbers of HPCs were identified . Developing an approach to weighting and ranking the large number of HPCs generated by broad spectrum inhibitors may aid in increasing the predictive performance of ccorps for these inhibitors . For example , ranking HPCs by the mean within-cluster affinity ( ) would more heavily weight structural features correlated with strong binders and decrease the impact of structural features only correlated weak binders . Such an approach would help to increase the signal-to-noise ratio of HPCs when the number of HPCs identified grows large . As our results showed , there are cases where ccorps significantly outperforms a sequence-based method , but there also cases where the reverse is true . While this paper focused on quantifying the extent at which structure alone can be used to predict binding affinity , for practical usage we envision that structure- and sequence-based methods are used in tandem . A major advantage of the work presented is the generality of ccorps to detect structurally distinguishing features for a wide variety of applications beyond the kinase inhibitor affinity analysis presented here . No assumptions regarding the nature of the annotation labels nor of the alignment type are made at any point by ccorps . ccorps provides a general framework for automatically learning structural features that distinguish proteins having different annotation labels . This allows the incorporation of purely structure-based alignments , such as those available in databases like homstrad [33] or even local structure alignments such as those identified by motif/template search algorithms ( e . g . , soippa [34] , and LabelHash , [35] ) . Other sources of annotation labels , including Gene Ontology ( GO , [14] ) terms , binding affinity for a given molecule and ligation state can be incorporated as-is with ccorps without modification to the method . | The kinases are a group of essential signaling proteins within the cell and are the largest family of enzymes encoded by the human genome . The high degree of binding site similarity shared across the protein kinases has made them difficult targets for which to design highly selective inhibitors , but kinome-wide binding site analysis can help predict unintended off-target inhibitions . Given the increasingly large number of available kinase structures , kinome-wide comparative analysis of binding sites is now possible . In this paper , the Combinatorial Clustering Of Residue Position Subsets ( ccorps ) method is introduced and used to synthesize kinome-wide structure datasets with a kinome-wide inhibitor affinity screening dataset consisting of 38 kinase inhibitors . ccorps identifies structural features of the kinase binding site that are correlated with an inhibitor binding and uses these features to predict if this inhibitor will be capable of binding to uncharacterized kinases . This paper demonstrates the ability of ccorps to accurately predict inhibitor binding and identify features of the kinase binding site that are unique to kinases capable of binding a given inhibitor . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biomacromolecule-ligand",
"interactions",
"algorithms",
"computer",
"science",
"protein",
"structure",
"biology",
"computational",
"biology",
"biophysics",
"macromolecular",
"structure",
"analysis"
] | 2013 | Combinatorial Clustering of Residue Position Subsets Predicts Inhibitor Affinity across the Human Kinome |
Since 2010 , WHO has recommended oral cholera vaccines as an additional strategy for cholera control . During a cholera episode , pregnant women are at high risk of complications , and the risk of fetal death has been reported to be 2–36% . Due to a lack of safety data , pregnant women have been excluded from most cholera vaccination campaigns . In 2012 , reactive campaigns using the bivalent killed whole-cell oral cholera vaccine ( BivWC ) , included all people living in the targeted areas aged ≥1 year regardless of pregnancy status , were implemented in Guinea . We aimed to determine whether there was a difference in pregnancy outcomes between vaccinated and non-vaccinated pregnant women . From 11 November to 4 December 2013 , we conducted a retrospective cohort study in Boffa prefecture among women who were pregnant in 2012 during or after the vaccination campaign . The primary outcome was pregnancy loss , as reported by the mother , and fetal malformations , after clinical examination . Primary exposure was the intake of the BivWC vaccine ( Shanchol ) during pregnancy , as determined by a vaccination card or oral history . We compared the risk of pregnancy loss between vaccinated and non-vaccinated women through binomial regression analysis . A total of 2 , 494 pregnancies were included in the analysis . The crude incidence of pregnancy loss was 3 . 7% ( 95%CI 2 . 7–4 . 8 ) for fetuses exposed to BivWC vaccine and 2 . 6% ( 0 . 7–4 . 5 ) for non-exposed fetuses . The incidence of malformation was 0 . 6% ( 0 . 1–1 . 0 ) and 1 . 2% ( 0 . 0–2 . 5 ) in BivWC-exposed and non-exposed fetuses , respectively . In both crude and adjusted analyses , fetal exposure to BivWC was not significantly associated with pregnancy loss ( adjusted risk ratio ( aRR = 1 . 09 [95%CI: 0 . 5–2 . 25] , p = 0 . 818 ) or malformations ( aRR = 0 . 50 [95%CI: 0 . 13–1 . 91] , p = 0 . 314 ) . In this large retrospective cohort study , we found no association between fetal exposure to BivWC and risk of pregnancy loss or malformation . Despite the weaknesses of a retrospective design , we can conclude that if a risk exists , it is very low . Additional prospective studies are warranted to add to the evidence base on OCV use during pregnancy . Pregnant women are particularly vulnerable during cholera episodes and should be included in vaccination campaigns when the risk of cholera is high , such as during outbreaks .
Cholera represents a risk of complications for pregnant women and their fetus . Published literature reports fetal loss rates during cholera episodes of between 2% and 36% [1–7] . However , comparison of pregnancy outcomes among different reports is difficult , due to differences in inclusion criteria , treatment provided , and access to care . Although the exact cause of fetal death during a cholera episode has not yet been identified , several studies suggest an association between fetal loss and the degree of dehydration and hypovolemia [2 , 4–7] . In cholera-endemic countries , the World Health Organization ( WHO ) recommends vaccination “for groups that are especially vulnerable to severe disease and for which the vaccines are not contraindicated , such as pregnant women and HIV-infected individuals” [8] . WHO has prequalified two oral cholera vaccines ( OCV ) , both consist of killed whole-cells of V . cholerae . One consists of several strains of V . cholerae O1 and a recombinant B subunit of the cholera toxin ( WC-rBS , marketed as Dukoral ) ; the other contains strains from both serogroups O1 and O139 , but no component of the cholera toxin ( BivWC , marketed as Shanchol ) [8] . According to the package inserts , neither vaccine is contraindicated in pregnant women , but only recommended when the potential benefits are considered higher than the risk . Inactivated OCVs are unlikely to have a harmful effect on fetal development as the killed bacteria in the vaccine do not replicate , the vaccine antigens act locally in the gastrointestinal mucosa , are not absorbed and do not enter the maternal or fetal circulation . In addition , the vaccines do not trigger systemic reactions ( e . g . fever ) associated with miscarriage in early pregnancy [9] . Pre-licensure studies and post-marketing surveillance suggest that Dukoral has a good safety profile when used during pregnancy [4] and inadvertent vaccination of pregnant women with the vaccine during a mass vaccination campaign in Zanzibar in 2009 was not associated with any harmful effects [9] . However , pregnant women have been excluded systematically from most other cholera vaccination campaigns because of the weak data on safety during pregnancy for Dukoral and the absence of safety data during pregnancy for Shanchol [10] . Shanchol has several advantages compared with Dukoral for public health use . The vaccine is cheaper , has a lower storage volume and does not require water for administration . Thus , understanding the safety of BivWC during pregnancy will provide essential information for its future use throughout the cholera-endemic world . The Ministry of Health and Public Hygiene ( MHPH ) of Guinea , with the support of Médecins Sans Frontières ( MSF ) , carried out mass OCV campaigns using BivWC in 2012 in Boffa and Forécariah Prefectures as part of a comprehensive response to a cholera epidemic that was spreading in remote rural areas with limited access to health facilities [11 , 12] . These campaigns targeted all people aged one year and above living in the target areas [11 , 12] . Pregnant women were not excluded from the target population . In order to assess whether there was a difference in pregnancy outcomes between women who exposed their fetus to OCV and those who did not , we report the results of a retrospective cohort study , which compared the incidence of pregnancy losses ( miscarriages and stillbirths ) and malformations between these two groups .
Women were included in the study if they were residents of the Koba and Boffa subprefectures , were 15 to 49 years old , were pregnant in 2012 ( i . e . , conception and/or birth occurred that year ) and if they ( or their guardians for minor participants ) provided informed consent . Exclusion criteria were non-residence in Boffa prefecture at the time of the vaccination campaign , absence from the home after two visits , lack of knowledge of their vaccination status , and refusal to participate . Based on published literature [13–16] , we assumed a 10% incidence of pregnancy loss , an unexposed/exposed ratio of 0 . 3 ( based on 77% of pregnant women vaccinated in the vaccination coverage survey ) , an alpha error of 0 . 05 , and a statistical power of 0 . 8 . Thus , 1 , 200 vaccinated pregnant women and 360 non-vaccinated pregnant women were necessary to estimate a 1 . 5 increase in the risk of pregnancy loss among vaccinated women . All interviewers and supervisors were recruited locally and received theoretical and practical training . They visited all households ( defined as a group of individuals living under the same roof and regularly sharing the same meals ) . Interviewers revisited households later in the day where no one was at home . If there was no response the second time , the household was skipped . Interviewers asked the head of household for the number of women between 16 and 50 years old living in the household , and the number of women who were pregnant in 2012 , irrespective of pregnancy outcome . They obtained written informed consent from the women who were pregnant in 2012 and conducted face-to-face interviews in the local language . A standardized pre-tested questionnaire was used to collect inclusion criteria , socio-demographic data , information about the pregnancy , pregnancy history and other risk factors for pregnancy loss . Vaccination status was assessed at the end of the questionnaire . Interviewers also completed a questionnaire to determine the health condition of live-born babies . Mothers and children were referred to a pediatrician if the questionnaire elicited concerns . The pediatrician completed a clinical examination and determined if the child was ill or presented any malformation . The medical team was also in charge of patient management ( i . e . ambulatory treatment or transfer to hospital ) , if needed . The primary outcome of the study was the incidence of pregnancy loss , defined as any loss of a product of conception after the woman recognizes she is pregnant . Secondary outcomes included the incidence of miscarriage , stillbirth and malformation in live children . A miscarriage was defined as a loss of a clinically recognized pregnancy before the end of the fifth month of gestation and a stillbirth as the delivery of a dead fetus ( without pulse ) after the end of the fifth month of gestation . These outcomes were reported orally by the mother and verified by documentation when possible . A malformation was defined as a physical defect in a live infant that was identified by the study pediatrician . Primary exposure was defined as the intake of OCV during pregnancy . Participants were asked whether they had been vaccinated and , if so , to show their vaccination cards . A fetus was considered exposed if the mother was pregnant during the campaign , received at least one dose of OCV ( card-confirmed or reported orally ) , and at least one dose was received after the estimated date of conception and before the date of birth or fetal loss . Date of birth was reported orally and verified by documentation when possible . The date of conception was calculated by subtracting the duration of the pregnancy ( reported orally or confirmed by documentation ) from the date of birth or fetal loss . When date of birth or fetal loss was unknown , the mother was asked if she was pregnant during the vaccination campaign . The primary data analysis included women who were pregnant during the mass vaccination campaign . Descriptive analysis of these women was stratified by their vaccination status . Qualitative and quantitative variables were compared , respectively , through Fisher and Wilcoxon tests . The fetus was then considered as the unit of analysis since some women had multiple pregnancies . We calculated crude cumulated incidence of pregnancy loss as the number of pregnancy losses divided by the number of conceived fetuses . We compared the risk of pregnancy loss through a binomial regression . Possible confounders were variables for which p-values were less than 0 . 20 in the bivariate analysis . We obtained an adjusted estimate of relative risk ( aRR ) of pregnancy loss and its 95% confidence interval ( 95%IC ) according to OCV exposure using a forward stepwise procedure . The interaction between trimester of the pregnancy on 18 April 2012 and primary exposure was tested . All covariates significantly associated with the risk of a pregnancy loss ( p-value <0 . 05 ) or those improving model fit ( based on Bayesian Information Criterion ) were retained in the final model . Women with missing data were excluded from the analysis . In a secondary analysis , the same procedure was applied to other negative outcomes ( miscarriages , stillbirths and malformations ) . Fetuses born to mothers who had been pregnant for more than five months on 18 April 2012 were excluded from the analysis of the risk of miscarriage . Fetuses who did not complete five months of gestation were excluded from the analysis of the risk of stillbirth . Children who were not alive at the time of the survey ( fetal or perinatal deaths ) were excluded from the analysis of the risk of malformations . A bias-indicator analysis of fetuses conceived in 2012 after the second vaccination round was conducted to assess bias from possible misclassification of the women vaccination status or fetal outcome . This analysis again compared pregnancy outcomes of woman who had been vaccinated during the campaign with women who did not receive the vaccine . Since OCV intake before conception is not supposed to have an effect on pregnancy outcome , this analysis provides information about possible information bias . Since the exact dates of vaccination , conception and fetal lost were mainly estimates , we conducted sensitivity analyses by excluding all fetuses born or lost within seven days of the first round of the vaccination campaign and those whose estimated date of conception was within two weeks following the first round of the campaign . Data entry was performed using EpiData 3 . 1 ( EpiData Association , Denmark ) and data analysis was performed using Stata 12 . 0 ( College Station , USA ) . This study was conducted according to the ethical principles for research on human subjects , described in the Helsinki Declaration , and in accordance with international principles and guidelines for biomedical research involving human subjects , published by the Council for International Organizations of Medical Sciences . The study protocol was approved by National Ethics Committee of the Republic of Guinea and the Médecins Sans Frontières Ethics Committee . Each woman ( or her legal representative ) received the information on the methods and potential risks and benefits of the study . The participant or her representative signed an informed consent form after being informed that participation in the study was voluntary and that she could withdraw from the study at any time . Anonymity and confidentiality of collected data were ensured throughout the study . If there was any suspected illness in the live-born babies , they were referred to the pediatrician , were treated or referred and hospitalized , if needed . All treatment was provided free of charge .
Overall , 84 . 8% [95%CI: 83 . 0–86 . 6%] of the women pregnant during the campaign received at least one dose of OCV and could therefore have exposed their fetus to the vaccine . Vaccine coverage was significantly higher among women who were pregnant during the vaccination campaign ( primary analysis ) than those who became pregnant after the campaign ( bias-indicator analysis ) , both for the first round ( 81 . 1% [95%CI: 79 . 2–83 . 1%] vs 76 . 1% [95%CI: 73 . 4–78 . 8%] , p-value = 0 . 003 ) and the second round ( 64 . 0% [95%CI: 61 . 6–66 . 4%] vs 55 . 5% [95%CI: 52 . 4–58 . 7%] , p-value<0 . 001 ) . Vaccination status was confirmed by vaccination card in 24% of the cases . Women vaccinated during their pregnancy were not significantly different from those not vaccinated in terms of socio-demographic variables , pregnancy history , pregnancy status and practices , aside from owning a television ( p = 0 . 033 ) and an oven ( p<0 . 001 ) ( Table 1 ) . Vaccinated and non-vaccinated women included in the bias-indicator analysis were also similar in their baseline characteristics ( Table A in S1 Appendix ) . Most ( 84 . 3% ) of the women pregnant during the vaccination campaign presented a child health record booklet . The percentage of women who received antenatal care services and who delivered in a health facility was higher among those who received the vaccine during their pregnancy than those who did not , though the differences were not statistically significant ( Table 1 ) . A total of 1 , 584 fetuses whose mother was pregnant during the campaign were included in the primary analysis; 1 , 312 ( 82 . 8% ) were exposed to the vaccine ( Table 2 ) . A total of 56 fetuses were classified as lost . There was no difference in the crude cumulative incidence of pregnancy loss between fetuses exposed to the vaccine and those who were not ( p = 0 . 350 ) . The adjusted risk ratio for pregnancy loss ( aRR ) was 1 . 13 [95%CI: 0 . 54–2 . 38 , p-value = 0 . 738] ( Table 2 ) . The risk of pregnancy loss was found to be higher among fetuses of mothers who reported a cholera episode in 2012 than those who did not in the adjusted analysis ( aRR = 3 . 18 [95%CI: 1 . 56–6 . 48] , p-value = 0 . 002 ) ( Tables B-D in S1 Appendix ) . The interaction between the trimester of pregnancy on April 18 , 2012 and the primary exposure was not significant ( p = 0 . 465 ) ( Table G in S1 Appendix ) . In the bias-indicator analysis , the risk of pregnancy loss was not associated with the vaccination status ( aRR = 1 . 19 [95%CI: 0 . 47–3 . 00] , p-value = 0 . 717 ) . A total of 1 , 263 fetuses exposed to the vaccine and 265 non-exposed fetuses were born alive . Among them , 18 exposed ( 1 . 4% [95%CI: 0 . 7–2 . 1%] ) and five non-exposed ( 1 . 9% [95%CI: 0 . 2–3 . 5%] ) babies died before the survey . This difference was not statistically significant ( p-value = 0 . 577 ) . In addition , 133 children ( 8 . 8% ) were referred to the study pediatrician among those screened in the primary analysis , as were 87 children ( 9 . 4% ) in the bias-indicator study . After the pediatrician’s clinical examination , seven vaccine-exposed children and three non-exposed children were considered to have a malformation ( Table 2 ) . Malformations were mainly from limbs ( five from lower limbs and two from hands ) ( Table E in S1 Appendix ) . There was no statistically significant increase in the risk of malformation for fetuses exposed to OCV in the primary analysis ( p-value = 0 . 314 ) ( Table 2 ) . After adjusting for other factors , the risk of malformation was significantly associated with the mother’s profession ( p-value = 0 . 008 ) ( Table F in S1 Appendix ) . In the bias-indicator analysis , the risk of malformation was not associated with vaccination status ( aRR = 0 . 51 [95%CI: 0 . 13–2 . 02] , p-value = 0 . 341 ) .
These are the first estimates of the risk of pregnancy loss following vaccination of pregnant women with the bivalent , whole-cell only oral cholera vaccine . Exposure of the fetus to this vaccine was not significantly associated with the risk of pregnancy loss and malformation in this study . Vaccine coverage among pregnant women was high ( 83% ) and similar to the overall vaccination coverage of the campaign [11] . This suggests that pregnant women who were offered OCV during the campaign chose to participate rather than forego vaccination . Vaccination coverage was higher among women who were pregnant during the campaign than among those who become pregnant after the campaign . Pregnant women may have been better informed about the vaccination campaign , less occupied by outside activities on the day of vaccination , and more willing to follow the advice of the Ministry of Health to get the vaccination than non-pregnant women . Overall , vaccinated and non-vaccinated women had similar baseline characteristics , both in the primary and in the bias-indicator analyses . Vaccinated pregnant women included in the primary analysis were more likely to attend antenatal care services and delivered more frequently in health facilities than those not vaccinated , which could be the result of a greater interest and awareness of preventive activities during pregnancy . The lack of association between the exposure of the fetus to OCV and pregnancy loss in both the crude and the adjusted primary analysis is consistent with the findings with Dukoral in Zanzibar ( aRR = 1 . 62 [0 . 76–3 . 43] , p-value = 0 . 21 ) [9] . In the present study , the exposure of the fetus to OCV was not significantly associated with miscarriage or stillbirth . In the Zanzibar study , analysis of pregnancy loss was not broken down by miscarriage or stillbirth , although the crude incidence of stillbirths was slightly higher among vaccinated women ( 4 . 6% versus 2 . 1% ) [9] . Another key finding in this study is that women who reported having had cholera in 2012 while they were pregnant were at six times higher risk of miscarriage and three times higher risk of having a stillborn child than women who did not report having had cholera . Although consistent with the literature [1–7] , biological confirmation of cholera cases and determination of the date of onset of the illness would have strengthened the causal link between cholera episodes and pregnancy loss . The number of reported cholera episodes was lower among vaccinated versus non-vaccinated women who were pregnant during the campaign . This is in line with the vaccine effectiveness ( 86% ) reported following the campaigns in Guinea [17] . The main reason newborns were referred to the pediatrician for clinical examination was illness rather than malformation . Malformations were detected mainly on upper and lower limbs . After adjusting on other factors , exposure to OCV was not statistically associated with malformation . There are several important limitations of note in this study . First , the incidence of pregnancy loss was lower than expected both in vaccinated and non-vaccinated women , especially in the first trimester . Pregnant women may not have reported , or been aware of , pregnancy losses during the study period . Conversely , some women could have falsely reported pregnancies or loss of pregnancies , since few pregnancy losses could be verified on official documentation . Since the number of pregnancy losses is low , this possible information bias could affect our point estimates , though it is difficult to determine in which direction . Second , less than 25% of the women could present a vaccination card , leading to potential misclassification of their vaccination status . In order to minimize this potential bias , we reminded participants about the way the vaccination campaigns were organized and the route of administration . To understand further the potential presence of information bias , we conducted a bias-indicator analysis to estimate the risk of pregnancy loss among women who were pregnant after the vaccination campaign . As in the primary analysis , the risk of pregnancy loss in the bias-indicator analysis was slightly but not significantly higher among vaccinated women . Another possible bias influencing our results is the presence of a seasonal component in pregnancies and pregnancy losses ( Fig A in S1 Appendix ) . When comparing non-vaccinated women , the incidence of pregnancy loss was higher among women who were pregnant during the campaign than among women who become pregnant afterwards . We could therefore not consider fetuses conceived after the vaccination campaign as controls in the primary analysis , reducing the power of our study . Lastly , as previously discussed , the number of negative events was lower than expected and the vaccine coverage was higher than expected , leading to a low number of non-exposed fetuses with negative events . This reduced the power of our analysis to detect statistical differences . In conclusion , we found no association between fetal exposure to OCV and risk of pregnancy loss or malformation . Despite the weaknesses of a retrospective design and a decreased statistical power due to the low number of fetuses not exposed to the vaccine , we can conclude that if there is a risk of poor pregnancy outcomes from taking OCV during pregnancy , it is likely to be very small . Further studies are needed to confirm these results and provide further evidence about the risks and benefits of OCV for pregnant women and their fetus . As far as possible , these studies should be prospective cohort studies to reduce the likelihood of misclassifying negative pregnancy outcomes or exposure to the vaccine . It is also important to note that any small potential risk of pregnancy loss could be offset by the possible benefit of vaccination . During preventive campaigns in non-epidemic periods , if the risk of infection is low , vaccination of pregnant women could be delayed , notably for women who have other risk factors for pregnancy loss . However , during epidemics , when the risk of cholera infection is high , vaccination should be offered to all pregnant women , since they are at particularly high risk of losing their fetus if they become ill with cholera . | Pregnant women are at high risk of complications and fetal deaths when ill with cholera . However , they have been excluded in most cholera vaccination campaigns because of the lack of safety data on oral cholera vaccines during pregnancy . This study aimed to determine if the risk of pregnancy loss changed after the administration of the oral cholera vaccine in Guinea in 2012 . We visited all households in Boffa and Koba sub-prefectures , where the vaccination campaign took place , and enrolled a total of 2 , 493 women in the study . In this large retrospective cohort , we found no association between fetal exposure to the cholera vaccine and the risk of pregnancy loss or malformation . Pregnant women are particularly vulnerable during a cholera episode and should be included in vaccination campaigns when the risk of cholera is high , such as during the outbreaks . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Pregnancy Outcomes after a Mass Vaccination Campaign with an Oral Cholera Vaccine in Guinea: A Retrospective Cohort Study |
The factors determining the functional avidity and its relationship with the broad heterogeneity of antiviral T cell responses remain partially understood . We investigated HIV-specific CD8 T cell responses in 85 patients with primary HIV infection ( PHI ) or chronic ( progressive and non-progressive ) infection . The functional avidity of HIV-specific CD8 T cells was not different between patients with progressive and non-progressive chronic infection . However , it was significantly lower in PHI patients at the time of diagnosis of acute infection and after control of virus replication following one year of successful antiretroviral therapy . High-avidity HIV-specific CD8 T cells expressed lower levels of CD27 and CD28 and were enriched in cells with an exhausted phenotype , i . e . co-expressing PD-1/2B4/CD160 . Of note , a significant increase in the functional avidity of HIV-specific CD8 T cells occurred in early-treated PHI patients experiencing a virus rebound after spontaneous treatment interruption . This increase in functional avidity was associated with the accumulation of PD-1/2B4/CD160 positive cells , loss of polyfunctionality and increased TCR renewal . The increased TCR renewal may provide the mechanistic basis for the generation of high-avidity HIV-specific CD8 T cells . These results provide insights on the relationships between functional avidity , viremia , T-cell exhaustion and TCR renewal of antiviral CD8 T cell responses .
CD8 T cells play a critical role in antiviral immunity and a large number of studies in both human and murine models indicate that virus-specific CD8 T cells are directly involved in the control of virus replication and disease progression [1] , [2] , [3] , [4] , [5] , [6] , [7] . Functional avidity of T cells , also defined as antigen ( Ag ) sensitivity , is thought to be a critical component of antiviral immunity . Functional avidity reflects the ability of T cells to respond to a low Ag dose and is determined by the threshold of Ag responsiveness . There is a general consensus that high functional avidity CD8 T-cell responses are of higher efficacy against cancers [8] and acute virus infections [9] . However , their relevance in chronic persistent virus infections and established tumors [10] , [11] , [12] remains to be determined since conflicting results were obtained in these contexts [13] , [14] as well as in HIV infection [15] , [16] , [17] , [18] , [19] . HIV-specific CD8 T-cell responses in non-progressive infection were associated with high avidity and superior variants recognition [11] , [12] , [20] , [21] , whereas other studies indicated similar avidity between patients with progressive and non-progressive chronic infection [16] , [18] , [19] , [22] , [23] . In this regard , we have previously shown that polyfunctional virus-specific CD8 T-cell responses during chronic virus infections were predominantly of low functional avidity [24] . Furthermore , it is also well established that high functional avidity T-cell responses preferentially led to viral escape and T-cell clonal exhaustion [10] , [24] , [25] , [26] . However , the factors determining the level of T-cell functional avidity and its relationship with the phenotypic and functional heterogeneity of T-cell responses are only partially understood [15] , [16] , [17] , [18] , [19] , [22] . Functional avidity is based on the ability of T cells to respond following stimulation with a cognate Ag and it is well established that responding CD8 T cells are clonally heterogeneous ( i . e . oligoclonal ) [27] , [28] , [29] , [30] . Therefore , the clonotypic composition of the responding T-cell population ( and its TCR diversity ) can influence functional avidity [27] , [28] . Indeed , we and others reported that HIV-specific CD8 T cells responding to various epitopes harbor a diverse TCR repertoire in chronically-infected patients [31] , [32] , [33] . HIV-specific CD8 T cells in primary HIV infection ( PHI ) are temporally associated with the initial control of viremia [1] . Lichterfeld and colleagues suggested that high-avidity HIV-specific CD8 T-cell responses are present during early infection ( defined as HIV seroconversion within 6 months ) and are then preferentially lost overtime [33] . In the present study , we have performed a comprehensive cross-sectional characterization of HIV-specific CD8 T-cell responses in patients with PHI or chronic ( progressive and non-progressive ) HIV infection in both steady-state conditions as well as following virus rebound . The primary observations of the present study indicate that a ) the functional avidity of HIV-specific CD8 T cells is not different between patients with progressive and non-progressive chronic infection , b ) the functional avidity of HIV-specific CD8 T cells is significantly lower in PHI patients as compared to patients with chronic infections , c ) increased functional avidity is associated with T-cell exhaustion and lack of expression of markers of co-stimulation , and d ) great increase in functional avidity is observed after virus rebound following spontaneous interruption of antiretroviral therapy and is associated with increased TCR renewal .
We recruited 85 HIV-infected patients and performed a cross-sectional analysis of the functional avidity of HIV-specific CD8 T-cell responses . The distinct groups included a ) 37 patients with very early stage of acute infection ( i . e . prior to seroconversion and incomplete western blot; hereafter referred to as PHI ) , 39 patients with progressive chronic infection ( i . e . typical progressors; hereafter referred to as CP ) and 9 patients with non-progressive chronic infection ( i . e . LTNP ) ( Table S1 ) . We first investigated the 115 HIV-specific CD8 T-cell responses obtained in 26 untreated PHI ( PHI-B ) , 19 untreated CP ( CP-B ) patients and 9 LTNP ( Fig . 1A–B ) . As described in the Methods , blood mononuclear cells were stimulated with decreasing concentrations of the cognate peptides and the peptide dose able to induce half of the maximal response ( i . e . effect concentration 50%; EC50 ) was determined ( Fig . 1A ) . The results of this analysis indicated that the functional avidity of HIV-specific CD8 T cells was lower in PHI-B as compared to CP-B or LTNP patients ( both P<0 . 0001; Fig . 1A–B ) while there was no difference between CP-B and LTNP ( Fig . 1A–B ) . However , there were no significant difference in the magnitudes of HIV-specific CD8 T-cell responses among all groups ( Fig . 1C ) and no significant association between the functional avidity and the magnitude of HIV-specific CD8 T-cell responses ( Fig . S1A ) . Furthermore , the differences in functional avidity of HIV-specific CD8 T cells between the different cohorts were not influenced by distinct peptides/MHC class I associations , since these differences remained significant when common epitopes ( i . e . epitopes recognized by patients from distinct cohorts ) were analyzed ( Fig . 1D ) . Of note , the B*2705-KRWIILGLNK ( i . e . KK10 ) epitope has been previously reported as a protective epitope [28] , [34] , [35] and it was one of the common epitopes recognized by CP-B and LTNP patients . While the functional avidity of B*2705-KK10-specific CD8 T-cell responses in CP-B and LTNP patients was almost identical , it was rather low as compared to the other HIV-specific CD8 T-cell responses from both groups ( Fig . 1D ) . Taken together , these results indicate a lack of association between the functional avidity of HIV-specific CD8 T cells and virus control consistently with the recent study from Chen and colleagues [22] . Furthermore , the HIV-specific CD8 T-cell responses in acute infection have lower functional avidity than in chronic infection . It has been previously reported that high functional avidity HIV-specific CD8 T-cell responses are selectively deleted early after acute HIV infection [33] . We addressed this issue by repeating the epitopes mapping in patients with acute infection after one year of antiretroviral therapy ( ART ) ( PHI-T1Y ) . Forty-five HIV-specific CD8 T-cell responses were identified in PHI-B patients using ICS . Among these 45 responses , 38 ( 85% ) remained detectable after one year of ART whereas 7 ( 15% ) became undetectable . Interestingly , at the time of acute infection , these 7 responses were already of lower magnitude as compared to the 38 responses which remained detectable ( P = 0 . 03; Fig . 2A ) . Furthermore , the functional avidity of HIV-specific CD8 T-cell responses at the time of acute infection was not different between the 7 lost and the 38 remaining responses ( P>0 . 05; Fig . 2A ) . These results indicate that the minor proportion of HIV-specific CD8 T-cell responses selectively lost after acute infection did not have higher functional avidity . It has been suggested that Ag load may influence the responsiveness of HIV-specific CD8 T cells [36] . To address this issue , we assessed whether the functional avidity of HIV-specific CD8 T-cell responses would change after control of virus replication , i . e . after 1 year of successful ART , in patients with acute or chronic infection . Of note , 46 additional HIV-specific CD8 T-cell responses were considered; 17 responses were identified in the initial 26 PHI patients following re-mapping after 1 year of ART and 29 responses were identified in 11 additional PHI patients only mapped after 1 year of ART . Both magnitude and functional avidity of HIV-specific CD8 T-cell responses generated during ART were similar to those measured at baseline ( Fig . 2A ) . Furthermore , PHI patients were treated either with ART alone or with ART+CyclosporinA ( CsA ) but CSA treatment had no significant impact on the magnitude or the functional avidity of HIV-specific CD8 T-cell responses ( Fig . S2 ) . The functional avidity of the same HIV-specific CD8 T-cell responses measured longitudinally either prior to ART or after 1 year of ART remained stable in both PHI and CP patients ( both P>0 . 05; Fig . 2B ) . Furthermore , the lack of significant effect of ART on the functional avidity of HIV-specific CD8 T cells was also confirmed in non-longitudinal , independent , T-cell responses from PHI or CP patients ( both P>0 . 05; Fig . 2C ) . Therefore , HIV-specific CD8 T-cell responses remained of lower avidity ( P = 0 . 0003 ) in PHI-T1Y as compared to CP-T1Y patients ( Fig . 2D ) . Consistently with the above-mentioned analyses performed in the untreated groups , differences in functional avidity of HIV-specific CD8 T cells between PHI-T1Y and CP-T1Y were not related to distinct peptide-MHC class I associations since the differences remained significant also when common epitopes were considered ( P = 0 . 0003; Fig . 2E ) . All together , these observations indicate that even after control of virus replication HIV-specific CD8 T-cell responses from patients with acute HIV infection remain of lower avidity as compared to patients with chronic infection . We then assessed the functional profile of HIV-specific CD8 T cells from PHI-B , CP-B and LTNP patients . Although , the magnitudes of HIV-specific CD8 T-cell responses were not significantly different between PHI-B , CP-B and LTNP ( Fig . 1C ) , perforin expression was significantly ( P≤0 . 001 ) higher in HIV-specific CD8 T cells from PHI-B patients as compared to CP-B or LTNP ( Fig . S1B and 3A ) . As previously shown [37] , [38] , HIV-specific CD8 T cells from LTNP contained more IL-2 , whereas those from CP-B patients were mostly composed of single IFN-γ-producing cells ( both P<0 . 0001; Fig . S1B and 3A ) . We then performed a phenotypic characterization of HIV-specific CD8 T-cell responses and monitored CD27 and CD28 expression to assess co-stimulation and PD-1 , 2B4 and CD160 expression to assess T-cell activation and exhaustion . For these analyses , only HIV-specific CD8 T cells detectable using cognate peptide-MHC class I multimers ( Table S2 ) were taken into consideration . Regarding T-cell co-stimulation , HIV-specific CD8 T cells from PHI-B expressed a higher proportion of CD27+CD28+ cells than those from CP-B ( P = 0 . 02 ) or LTNP ( P = 0 . 003 ) patients ( Fig . S1C and 3B ) . Also , analyses of the expression of co-inhibitory receptors indicated that HIV-specific CD8 T cells from CP-B and LTNP were both composed of significantly higher proportions of PD-1+2B4+CD160+ ( P<0 . 006 and P<0 . 025 , respectively ) or PD-1−2B4+CD160+ ( P<0 . 025 and P<0 . 0001 , respectively ) as compared to PHI-B ( Fig . S1D and 3C ) . HIV-specific CD8 T cells from PHI-B mostly ( about 70% ) lacked all three markers or expressed 2B4 alone ( all P<0 . 006; Fig . 3C ) and expressed lower frequency and intensity of PD-1 as compared to CP-B ( both P<0 . 002; data not shown ) . These data suggest that HIV-specific CD8 T cells from patients with acute and chronic infection are functionally and phenotypically distinct . We then assessed the association between functional avidity and the expression of co-stimulatory or co-inhibitory receptors . The functional avidity of HIV-specific CD8 T cells was negatively correlated to the proportion of CD27+CD28+ cells ( P = 0 . 01; Fig . 4A ) and directly correlated to the proportion of cells co-expressing PD-1/2B4/CD160 ( P = 0 . 005; Fig . 4B ) . Furthermore , we performed correlations and rank correlation's matrix to explore the partial associations of variables and to assess the dependency and potential hidden effect of confounding variables in pairs associations . These analyses indicated that the proportions of CD27+CD28+ and of PD-1+2B4+CD160+ HIV-specific CD8 T cells were not significantly dependent on each other . This allowed us to perform a regression model analysis and to postulate that the functional avidity of HIV-specific CD8 T cells may be a linear function of the two aforementioned explained variables ( after log10 transformation ) . Interestingly , this regression analysis indicated that about 28% of the functional avidity of HIV-specific CD8 T cells was explained by a combination of the proportion of CD27+CD28+ and of PD-1+2B4+CD160+ CD8 T cells ( P = 0 . 0013; data not shown ) . We did not , in contrast , observe any significant correlation between the functional avidity of HIV-specific CD8 T cells and their functional profile . Overall , these observations indicate that high-avidity HIV-specific CD8 T-cell responses are preferentially composed of cells lacking the expression of co-stimulatory molecules but co-expressing high levels of co-inhibitory receptors . However , the functional avidity can only be partially predicted from the expression of co-stimulatory or co-inhibitory molecules . We then performed a longitudinal analysis to investigate the effect of changes in viremia levels on HIV-specific CD8 T cells . To address this issue , we longitudinally monitored HIV-specific CD8 T cells in two distinct models: a ) in conditions of viremia below the limit of detection , i . e . viremia <50 HIV RNA copies/ml of plasma ( in patients successfully treated by ART ) and b ) in conditions of rapid and major changes in viremia occurring in patients experiencing virus rebound following spontaneous treatment interruption ( TI ) ( Fig . 5A ) . In particular , we evaluated HIV-specific CD8 T-cell responses in PHI-T1Y and compared them with those after 5 years ( PHI-T5Y ) of uninterrupted successful ART or after TI ( PHI-ATI ) ( Fig . 5A ) . Nine out of the 37 patients identified during acute infection spontaneously interrupted ART . These patients were treated since PHI for ≥1 year ( mean±SE 131±15 weeks ) and all had undetectable viremia ( <50 HIV RNA copies/ml ) at the time of TI . After TI , all patients experienced a virus rebound with an average plasma viremia of 5 . 18 log10 HIV RNA copies/ml . The functional profile of HIV-specific CD8 T-cell responses at the time of TI was different from that of baseline . HIV-specific CD8 T cells in PHI-T1Y were mostly polyfunctional ( associated to a large fraction of IL-2-producing cells and little perforin ) ( P<0 . 0001; Fig . 5B–C ) as compared to the typical effector profile ( Fig . 3A ) observed in PHI-B . In patients remaining on ART , HIV-specific CD8 T cells became more polyfunctional ( i . e . further shifted toward IL-2 production ) after 5 years of ART as compared to 1 year of ART ( Fig . 5B–C ) . Conversely , in patients interrupting ART , as shown for patient #1023 who interrupted ART after two years of treatment and experienced a virus rebound of 122'000 HIV RNA copies/ml , the proportion of HIV-specific CD8 T cells co-producing IFN-γ and IL-2 decreased ( Fig . 5A–B ) . Cumulative analyses confirmed the significant ( P<0 . 01 ) decrease in polyfunctionality of HIV-specific CD8 T-cell responses after TI and the significant ( P = 0 . 03 ) increase in polyfunctionality of HIV-specific CD8 T-cell responses from patients who remained on ART ( Fig . 5C ) . Then we determined PD-1 ( as well as 2B4 and CD160 ) expression in a subset of PHI-ATI and PHI-T5Y patients with known HIV-specific CD8 T-cell responses using cognate peptide-MHC class I multimers ( Table S2 ) . As shown in the representative flow cytometry profiles from patients #1017 and #1023 , PD-1 expression increased in patient #1023 who interrupted ART but not in patient #1017 who remained on ART ( Fig . 5D ) . Along the same line , the proportion of triple PD-1+2B4+CD160+ HIV-specific CD8 T cells also increased in patient #1023 but not in patient #1017 ( Fig . 5F ) . Cumulative analyses confirmed that PD-1 expression as well as the proportion of cells co-expressing PD-1/2B4/CD160 in HIV-specific CD8 T cells were significantly increased in patients who interrupted ART ( both P = 0 . 03; Fig . 5E and 5G ) . No increase in PD-1 expression or in the co-expression of PD-1/2B4/CD160 , however , was observed in the patients who remained on ART ( both P>0 . 05; Fig . 5E and G ) . Finally , consistently with the differences in the functional profile of HIV-specific CD8 T-cell responses between patients who did or did not interrupt ART ( Fig . 5B–C ) , we observed that the proportion of dual IFN-γ/IL-2-producing HIV-specific CD8 T cells was negatively correlated with the proportion of cells co-expressing PD-1/2B4/CD160 ( P = 0 . 009; data not shown ) . These data indicate that major changes of viremia levels in TI patients caused reduction of polyfunctional HIV-specific CD8 T cells and were associated with an increased level of exhaustion . We then analyzed the effects of virus rebound following TI on the functional avidity of HIV-specific CD8 T cells . As shown for patient #1023 , the functional avidity of B*0701-GPGHKARVL- and A*0301-RLRPGGKKK-specific CD8 T-cell responses significantly increased after TI ( ATI ) as compared to pre-TI ( PTI ) ( Fig . 6A ) . Furthermore , an additional HIV-specific CD8 T-cell response against B*0701-IPRRIRQGL , which was below the detection level PTI , was observed ATI ( Fig . 6A ) . Cumulative analyses confirmed the increase in functional avidity of HIV-specific CD8 T cells occurring ATI ( P = 0 . 007; Fig . 6B ) and also indicated that new responses generated following virus rebound were of high avidity ( P = 0 . 04; Fig . 6B ) . Consistently , no significant ( P>0 . 05 ) differences in functional avidity were observed during the same period when a similar analysis was performed in HIV-specific CD8 T-cell responses from patients who did not interrupt ART ( i . e . for an average of 4 years; Fig . 6B ) . Furthermore , the functional avidity of HIV-specific CD8 T-cell responses ATI was in the same range as compared to those observed in CP patients ( data not shown ) . Of note , consistently with the increase in the co-expression of co-inhibitory molecules occurring in patients experiencing virus rebound ( Fig . 5F–G ) , we observed a positive association ( P = 0 . 02 ) between the fold increase in functional avidity of HIV-specific CD8 T cells and the fold increase in the proportion of PD-1+2B4+CD160+ HIV-specific CD8 T cells ( Fig . 6C ) . Of interest , we performed a comprehensive statistical modeling of the changes in functional avidity of HIV-specific CD8 T cells and used mixed-effect linear models [39] , [40] to assess the evolution of functional avidity as a function of time and virus rebound . For this analysis , all longitudinal measures ( n = 231 ) of functional avidity were included . The statistical model revealed that the interaction between avidity and time was not significant in steady-state conditions , i . e . neither in patients on ART ( Fig . 6D; red line ) , nor in patients off ART ( Fig . 6D; green line ) . In both conditions , an increase of 0 . 013 units per month was determined but did not reach statistical significance ( P = 0 . 05; Fig . 6D ) , thus indicating that functional avidity does not significantly change under steady-state circumstances . However , we found a significant ( P = 0 . 013 ) interaction between functional avidity and virus rebound . An immediate increase of functional avidity of HIV-specific CD8 T cells of about 1 order of magnitude ( 0 . 95 units ) occurred directly after TI ( Fig . 6D , grey dashed lines ) and was not related to the duration of ART prior to TI . These observations indicated that the functional avidity of HIV-specific CD8 T cells is stable overtime in steady-state conditions regardless of viremia levels , but does increase after rapid increase in viremia levels associated with virus rebound . We recently demonstrated that the global CD8 TCR repertoire of virus-specific CD8 T cells was diverse and subjected to continuous renewal [32] . We then evaluated the TCR repertoire in PHI patients experiencing a virus rebound following TI . For this purpose , we measured CDR3 diversity and the percentage of renewal of HIV-specific CD8 T cells and compared those to the changes in functional avidity of HIV-specific CD8 T cells occurring before and after virus rebound . As shown for patient #1023 ( Fig . 5A ) , TRBV usage and CDR3 size pattern were analyzed for B*0701-GPGHKARVL-specific CD8 T cells at week ( W ) 18 , W96 and W125 ( Fig . S3A–B ) . Using our previously-described model to determine CDR3 renewal [32] , we calculated a renewal of 76% between W18 and W96 ( i . e . on ART ) and a renewal of 82% between W96 and W125 ( i . e . after TI ) . Cumulative analyses confirmed a significantly ( P = 0 . 008 ) higher CDR3 renewal of HIV-specific CD8 T cells after virus rebound than in steady-state condition , i . e . during treatment ( Fig . 7A ) . Interestingly , the level of CDR3 renewal was directly associated ( P = 0 . 036 ) with the extent of increase in functional avidity of HIV-specific CD8 T cells ( Fig . 7B ) . Taken together , these observations suggest that increase in CDR3 renewal may contribute to the increase in functional avidity of HIV-specific CD8 T cells occurring after virus rebound .
T-cell functional avidity reflects the ability of T cells to respond to various concentrations of Ag and may be assessed ex vivo through a quantification of a biological function such as IFN-γ production , cytotoxic activity or proliferation capacity . Several parameters concur to determine the threshold of T-cell responsiveness . These include: a ) the affinity of the TCR for the peptide-MHC ( pMHC ) molecule , i . e . the strength of the interaction between the TCR and pMHC [41] , [42] , b ) the density of pMHC-TCR interactions ( reflecting both the amount of Ag and the ability of Ag presenting cells ( APC ) to present Ags ) [43] , [44] , [45] , [46] , c ) the expression of co-stimulatory and co-inhibitory molecules by T cells and APC [47] , and d ) the T-cell distribution and composition of signaling molecules [44] , [48] . However , the factors determining functional avidity and the relationship between functional avidity and the heterogeneity of T-cell responses are not well understood . In the present study , we comprehensively investigated the functional avidity of HIV-specific CD8 T-cell responses in a cross-sectional study of different cohorts of HIV-infected patients . The evaluation of the functional avidity of HIV-specific CD8 T cells was based on optimal epitopes , i . e . epitopes not necessarily corresponding to the autologous virus sequences . Since pMHC/TCR affinity is one of the parameters potentially influencing the functional avidity [49] , [50] , a mismatch between the epitope sequences and the TCR or the MHC may impact the determination of avidity . However , the same strategy was used throughout all cohorts of HIV-infected patients , thus minimizing the potential biases in our observations . HIV-specific CD8 T cells generated during acute infection were of lower functional avidity as compared to those from patients with chronic progressive or non-progressive infection . These differences were not biased by distinct peptide-HLA associations and remained significant after ART-induced control of virus replication . In addition , a preferential deletion of HIV-specific CD8 T-cell responses of higher avidity was not observed , as previously described in a cohort of early HIV infection [33] . The discrepancy between our and the previous study may be explained by differences in the individual cohorts as well as by the fact that all the 37 patients received ART at the time of diagnosis of PHI in our study , whereas only 5 of the 10 patients in Lichterfeld's study received ART [33] . Our results also indicated that the minor proportion of HIV-specific CD8 T-cell responses lost after acute infection had an initial lower magnitude rather than a higher avidity . Furthermore , consistently with previous studies [16] , [18] , [19] , [22] , [23] , [51] , there were no significant differences in the functional avidity of HIV-specific CD8 T-cell responses from chronic progressive and non-progressive infection . These observations suggest that T-cell functional avidity does not represent a correlate of virus control , at least in the context of chronic and persistent virus infections . Along the same line , HIV-specific CD8 T-cell responses commonly associated with virus control [52] , i . e . HLA-B*27- , B*57- or B*5801-restricted T-cell responses , were consistently found in the lower range of functional avidity ( data not shown ) . These observations do not support previous studies showing a relationship between higher avidity T-cell responses and better virus control [11] , [12] , [20] , [21] , [53] , [54] . One potential explanation is that in most of these studies , specific T-cell epitopes ( e . g . TW10 or KK10 ) were considered predominantly in individuals with non-progressive infection . Of note , consistently with our study , Chen and colleagues recently demonstrated that KK10-specific CD8 T-cell responses in elite controllers showed better virus control and broader viral recognition but similar functional avidity as compared to progressors [22] . They also confirmed the overall lack of difference in the functional avidity of HIV-specific CD8 T cells between patients with progressive and non-progressive infection [22] . Taken together , these observations suggest an association between higher avidity T-cell responses and chronic HIV infection . Of note , we also assessed the relationship between T-cell functional avidity and the expression of markers of exhaustion . It is important to underscore that HIV-specific CD8 T cells which had higher avidity in chronic infection expressed also higher levels of exhaustion markers . Therefore , these results further indicate that higher functional avidity does not correlate with better virus control but rather with the status of cells activation/exhaustion . We also determined the impact of the expression of costimulation ( i . e . CD27 and CD28 ) and exhaustion markers ( i . e . PD-1 , CD160 and 2B4 ) on the levels of functional avidity in HIV-specific CD8 T cells using a regression model . The regression model indicated that the expression of the above markers only partially accounts for the establishment of the functional avidity of HIV-specific CD8 T cells , thus indicating that additional factors may contribute to determine the levels of functional avidity . The lower avidity of HIV-specific CD8 T cells in PHI patients may also be potentially explained by the fact that patients were identified very early in the course of infection and received ART within 24 hours . Therefore , one cannot exclude the possibility that this early control of virus replication blunted the natural evolution and maturation of the immune response , as previously shown for T- and B-cell responses [55] , [56] , [57] . Consistently , it was also shown in mice that functional avidity of antiviral CD8 T cells continuously increased ( avidity maturation ) during the first month of infection [58] . In this regard , when patients treated during PHI experienced a virus rebound , the functional avidity of HIV-specific CD8 T-cell responses significantly increased . The mixed-effect linear model we used indicated a punctual increase of about one order of magnitude following virus rebound . However , there was no quantitative correlation between either the peak or the steady-state of the virus rebound and the increase in avidity . Several mechanisms were proposed to modulate T-cell functional avidity maturation including: 1 ) the formation of clusters comprising several TCRs and other molecules able to reinforce the immunological synapses [59] , [60] , [61] , 2 ) the optimization of the signal transduction machinery such as an increase in the amount of and in the basal phosphorylation levels of signaling molecules [62] , [63] and 3 ) a selective expansion of high TCR avidity clones and/or the loss of clones with low TCR avidity [33] , [64] , [65] , [66] , [67] . We cannot exclude that the same mechanisms may also contribute to explain the increase in avidity observed following treatment interruption and virus rebound . Interestingly , we showed that TCR renewal was also significantly higher following virus rebound and associated with an increase in T-cell functional avidity . Therefore , our data indicate a potential role of TCR renewal in the modulation of the levels of functional avidity . However , our results do not distinguish between the recruitment of new clones , selective expansion of pre-existing high-avidity clones or depletion of low-avidity clones since the study was performed at the population level . Taken together , these results support the following model . HIV-specific CD8 T cells of lower functional avidity are generated during primary immune responses; then , persistence of detectable viremia drives an increase in functional avidity as supported by the major increase in functional avidity associated with the sudden increment in viremia levels; the increase in viremia levels is also associated with massive TCR renewal which , in turn , causes the generation/selection of T-cell clones with higher functional avidity . These results provide insights on the relationships between functional avidity , viremia , T-cell exhaustion and TCR renewal of antiviral CD8 T-cell responses .
These studies were approved by the Institutional Review Board of the Centre Hospitalier Universitaire Vaudois and all subjects gave written informed consent . Seventy-six patients with primary ( PHI ) or progressive chronic ( CP ) HIV infection were enrolled . Diagnosis of PHI included the presence of an acute clinical syndrome , a negative HIV antibody test , a positive test for HIV RNA in plasma , and ≤3 positive bands in a Western blot . All PHI patients started ART alone or ART+CsA within 72 h as described [68] and were followed for up to 10 years . Patients with chronic progressive ( CP ) HIV infection were infected for more than a year , were ART-naïve at the time of inclusion , had ≥400 CD4 T-cells/µl , ≥5000 plasma HIV RNA copies/ml and were directly treated with ART upon diagnosis as described [69] , [70] . Four CP patients were investigated both prior to ( BSL ) and then after 1 year of ART ( T1Y ) . Furthermore , 9 additional HIV-infected patients with non-progressive disease , i . e . LTNP , as defined by documented HIV infection since >10 years , stable CD4 T-cell counts >500 cells/µl , and plasma viremia <500 HIV RNA copies/ml were also included . Clinical and virological characteristics of the different cohorts are detailed in Table S1 . Epitope mapping was performed using a panel of 192 HPLC-purified ( >80% purity ) previously-described optimal epitopes [71] . Confirmation of specificity was achieved based on the HLA class I genotype of the patients and ICS assays . Peptide-MHC class I multimers ( listed in Table S2 ) were purchased from ProImmune ( Oxford , UK ) except HLA-B*0801-RAKFKQLL , HLA-B*0702-RPPIFIRRL , HLA-B*0702-TPRVTGGGAM and HLA-B*5701-TSTLQEQIGW ( Table S2 ) which were produced as described [72] . The following antibodies were used in different combinations . CD8-PB , CD8-APCH7 , CD3-APCH7 , CD45RA-PECy5 , PD-1-FITC , IFN-γ-APC , TNF-α-PECy7 , and IL-2-PE were purchased from Becton Dickinson ( BD , San Diego , CA ) , CD4-ECD , CD3-ECD , CD28-ECD , CD27-APC from Beckman Coulter ( Fullerton , CA , USA ) , Perforin-FITC ( clone B-D48 ) from Diaclone ( Besançon , France ) , CCR7-FITC from R&D Systems ( Minneapolis , MN , USA ) , 2B4-PECy5 . 5 and CD160-APC from Biolegend ( San Diego , CA , USA ) . ELISPOT assays were performed as per the manufacturer's instructions ( BD Biosciences ) . In brief , 2×105 cryo-preserved blood mononuclear cells were stimulated with 1 µg of single peptide or peptide pools in triplicate conditions as described [73] . Media only and staphylococcal enterotoxin B ( SEB ) were used as negative and positive controls , respectively . Thresholds for assay validation and positivity were determined as described [73] . Results are expressed as the mean number of SFU/106 cells from triplicate assays . Only cell samples with >80% viability after thawing were analyzed , and only assays with <50 spot forming unit ( SFU ) /106 cells for the negative control and >500 SFU/106 cells after SEB stimulation were considered valid . An ELISpot result was defined as positive if the number of SFUs was ≥55 SFU/106 cells and ≥4-fold the negative control . Cryo-preserved blood mononuclear cells ( 1–2×106 ) were stained for dead cells ( 4°C for 20′; Aqua LIVE/DEAD , Invitrogen ) and then stained with appropriately tittered peptide-MHC class I tetramer complexes at 4°C for 30′ in Ca2+-free media as described [74] . Cells were then washed and directly stained at 4°C for 20′ with the following Abs in various combinations: CD3 , CD8 , CD28 , CD27 , PD-1 , 2B4 , CD160 . Finally , cells were fixed ( CellFix , BD ) , acquired on an LSRII SORP ( 4 lasers ) and analyzed using FlowJo 8 . 8 . 2 ( Tree star Inc , USA ) . Analysis and presentation of distributions were performed using SPICE version 5 . 1 , downloaded from http://exon . niaid . nih . gov/spice/ [75] . The number of lymphocyte-gated events ranged between 0 . 6–1×106 in the flow cytometry experiments . Cryo-preserved blood mononuclear cells ( 1–2×106 ) were stimulated for 6 h or overnight in 1 ml of complete media ( RPMI ( Invitrogen ) , 10% fetal bovine serum ( FBS; Invitrogen ) , 100 µg/ml penicillin , 100 units/ml streptomycin ( BioConcept ) ) in the presence of Golgiplug ( 1 µl/ml , BD ) , anti-CD28 ( 0 . 5 µg/ml , BD ) and 1 µg/ml of peptide as described [74] . Staphylococcus enterotoxin B ( SEB; Sigma ) stimulation ( 100 ng/ml ) served as positive control . At the end of the stimulation period , cells were stained for dead cells ( 4°C for 20′; Aqua LIVE/DEAD , Invitrogen ) , permeabilized ( RT for 20′; Cytofix/Cytoperm , BD ) and then stained at RT for 20′ with CD4 , CD8 , CD3 , IFN-γ , IL-2 , TNF-α and perforin ( clone B-D48 ) . Cells were then fixed ( CellFix , BD ) , acquired on an LSRII SORP and analyzed using FlowJo 8 . 8 . 2 . Analysis and presentation of distributions were performed using SPICE version 5 . 1 , downloaded from http://exon . niaid . nih . gov/spice/ [75] . The number of lymphocyte-gated events ranged between 0 . 6–1×106 . With regard to the criteria of positivity of the ICS , the background in the unstimulated controls never exceeded 0 . 03% . An ICS to be considered positive had to have >0 . 03% of cytokine-positive cells after subtraction of the background ( media alone ) and to be >5 fold higher that the background . The analysis of the CDR3 diversity and renewal was performed as described [32] . CDR3 renewal corresponds to the percentage of TCR sequences specific for a given epitope that changed between two time points . Briefly , blood mononuclear cells were stained with cognate multimers and anti-CD3 , anti-CD8 , anti-CD45RA , and anti-CCR7 mAbs ( BD Biosciences ) . CD45RA+ CCR7+ naïve and Ag-specific ( multimer+ ) CD8 T cells were directly sorted ( FACSAria , BD Biosciences ) in RLT lysis buffer ( Qiagen , Hilden , Germany ) containing 20 ng RNA carrier ( Roche Diagnostics , Rotkreuz , Switzerland ) and RNA extracted ( Qiagen ) . Then , cDNA preparation and amplification were performed by using the SuperSMART PCR cDNA Synthesis Kit according to the manufacturer's instructions ( Clontech Laboratories , Saint-Germain-en-Laye , France ) . Amplified cDNA was subjected to TRBV–TCR-b–chain C region ( TRBC ) PCR reactions as described [32] . For spectratyping , aliquots of positive samples were mixed with Genescan-500 ROX size standards and run on an ABI 3130 capillary sequencer ( Applied Biosystems , Foster City , CA ) . The CDR3 junction ( length ) was analyzed using the IMGT system as described [32] . Peptide stimulations were performed as described above . Functional avidity of T-cell responses was assessed by performing limiting peptide dilutions ( ranging from 2 µg/ml to 1 pg/ml ) in in vitro assays as described [24] . The peptide concentration required to achieve a half-maximal IFN-γ response ( EC50 ) was determined . Four-digit HLA class I genotyping was performed by direct sequencing methods as described [76] . The data were analyzed and alleles were assigned using Assign-SBT version 3 . 5 ( Conexio Genomics , Applecross , Australia ) . Mann-Whitney and Wilcoxon-matched paired tests were performed using GraphPad Prism version 6 . 00 ( San Diego , CA ) . Analyses of the functional avidity of CD8 T-cell responses were performed on log10-transformed data using non-parametric tests . Associations among variables were performed by Spearman test . Rank correlations matrix and linear regression analysis were performed after log10 transformation of variables using R software . Bonferroni corrections for multiple analyses were applied . Regarding SPICE analyses of the flow-cytometry data , comparison of distributions was performed using a Student's t-test and a partial permutation test as described [75] . Furthermore , mixed-effect linear models were used to assess the evolution of functional avidity as a function of the time and virus rebound , as described [39] , [40] . In brief , let Y_ij be the measured avidity for subject i at time j ( time_ij ) and rebound_ij , the covariate coded as 1 if a patient i is an on-therapy at time j and coded as 0 if not on therapy ( off-therapy ) . We fitted the following mixed effect linear model: Y_ij = ( β_0+r_i ) + ( β_1 ) time_ij+ ( β_2 ) rebound_ij+ε_ij where β_0 is the global mean , β_1 the effect of the time on avidity , β_2 the effect of the virus rebound on avidity , r_i the random effect which represents the individual deviation from the global intercept and ε ij are independent measurement errors with mean zero . The interaction between time_ij and rebound_ij was tested . | CD8 T cells directed against virus are complex and functionally heterogeneous . One relevant component of CD8 T cells is their functional avidity which reflects their sensitivity to cognate antigens , i . e . how prone T cells are to respond when they encounter low doses of antigens . In patients with chronic and established HIV infection , we observed that the sensitivity of HIV-specific CD8 T cells was not different between patients with progressive or non-progressive disease . In contrast , the sensitivity of HIV-specific CD8 T cells was significantly lower in patients with early and recent HIV infection . Furthermore , CD8 T cells of high avidity were preferentially associated with a state of functional impairment known as exhaustion . Of interest , some patients treated with antiretroviral therapy during acute infection spontaneously interrupted their treatment and experienced a rebound of virus . In these patients , the avidity of HIV-specific CD8 T cells increased and this increase was associated to stronger cell exhaustion and greater renewal of the population of antiviral CD8 T cells , thus potentially providing the mechanistic basis for the generation of high-avidity CD8 T cells . Overall , our data suggest that rapid perturbation in viremia levels drove increases in the functional avidity of HIV-specific CD8 T cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"immunodeficiency",
"viruses",
"medicine",
"infectious",
"diseases",
"adaptive",
"immunity",
"immune",
"cells",
"immunity",
"virology",
"t",
"cells",
"immunity",
"to",
"infections",
"immunology",
"host-pathogen",
"interaction",
"biology",
"microbiology",
"infectious",
"disease",
"control",
"immune",
"response"
] | 2013 | Rapid Perturbation in Viremia Levels Drives Increases in Functional Avidity of HIV-specific CD8 T Cells |
A clear understanding of the knowledge , attitudes and practices ( KAP ) of a particular community is necessary in order to improve control of human African trypanosomiasis ( HAT ) . New screening and diagnostic tools and strategies were introduced into South Sudan , as part of integrated delivery of primary healthcare . Knowledge and awareness on HAT , its new/improved screening and diagnostic tools , the places and processes of getting a confirmatory diagnosis and treatment are crucial to the success of this strategy . A KAP survey was carried out in Yei County , South Sudan , to identify gaps in community KAP and determine the preferred channels and sources of information on the disease . The cross-sectional KAP survey utilized questionnaires , complemented with key informant interviews and a focus group discussion to elicit communal as well as individual KAP on HAT . Most ( 90% ) of the respondents had general knowledge on HAT . Lower levels of education , gender and geographic locations without a history of HAT interventions were associated with incorrect knowledge and/or negative perceptions about the treatability of HAT . Symptoms appearing in the late stage were best known . A majority ( 97 . 2% ) would seek treatment for HAT only in a health centre . However , qualitative data indicates that existing myths circulating in the popular imagination could influence people’s practices . Seventy-one percent of the respondents said they would offer social support to patients with HAT but qualitative data highlights that stigma still exists . Misconceptions and stigma can negatively influence the health seeking behaviour of HAT cases . In relation to communication , the top preferred and effective source of communication was radio ( 24% ) . Gaps in relation to KAP on HAT still exist in the community . Perceptions on HAT , specifically myths and stigma , were key gaps that need to be bridged through effective education and communication strategies for HAT control alongside other interventions .
Human African trypanosomiasis ( HAT ) , also known as sleeping sickness , is a parasitic neglected tropical disease of public health significance that mostly afflicts poor populations in endemic areas of rural Africa [1] . It is caused by two species of Trypanosoma: T . brucei rhodesiense , mostly found in eastern and southern Africa and causes the acute form of the disease , and T . b . gambiense , found in west and central Africa , and causes the chronic form of the disease . The disease occurs in two stages , early and late , and if not diagnosed early , affects the central nervous system causing severe neurological disorders , which can subsequently lead to death [2] . During the past decade , the number of HAT cases reported to the World Health Organization ( WHO ) per year has been falling progressively , arising from a concerted campaign to control the disease at international and country levels[3] . South Sudan is one of the countries that is still reporting cases of HAT [1] . According to [4] , the disease is endemic in the southern and southwestern regions of South Sudan , near the borders with Uganda , the Democratic Republic of the Congo ( DRC ) and the Central African Republic ( CAR ) . In the years 2011 and 2012 , the number of cases reported in South Sudan was 272 and 317 respectively , which was the third highest behind the DRC and CAR [5] . Recurrent outbreaks of the disease in this country have been attributed to reduced control measures and/or socio-political crisis [6] . About 250 , 000 km2in South Sudan is infested with the tsetse fly vector , with about 1–2 million people at risk of HAT . Glossina fuscipes fuscipes is the sole vector for HAT in the three Equatoria States; Eastern Equatoria State , Central Equatoria State and Western Equatoria State . Yei county is one of the 10counties in Central Equatoria State and among the nine that are endemic for HAT[6] . Over the years , HAT activities in South Sudan were mainly carried out by NGOs but after 2006 , when the numbers of new cases begun to decrease , most NGOs stopped or reduced their HAT related activities . This was due to difficulties in securing funding for what was no longer perceived as a serious problem [6 , 7] . Currently a project in South Sudan has brought together the Government of South Sudan , Malteser International and Foundation for Innovative New Diagnostics ( FIND ) , to intensify surveillance and control of HAT in the country in a sustainable manner , through the introduction and use of new screening and diagnostic tools and strategies as part of an integrated delivery of primary healthcare[8] . The HAT rapid diagnostic test ( RDT ) is part of the new screening and diagnostic tools for HAT . If one tests HAT RDT positive , they are referred to the nearest facility for parasitological confirmation , including microscopy . If found positive by microscopy , the patient is staged and treated for HAT [9] . Besides availability of new screening and diagnostic tools , well–trained staff and well-equipped health centers , accurate diagnosis and early identification of cases also depends on other community aspects , including knowledge , attitudes and practices ( KAP ) . A study in north western Uganda recommended that the success of community-based interventions against tsetse depend on early engagements with communities and carefully designed sensitization campaigns that reach all communities [10] . An understanding of community KAP therefore has relevance for disease control programmes and has been instrumental in revealing the level of knowledge , and misconceptions or misunderstandings , that could present obstacles to control intervention activities and potential barriers to behaviour change[11] . They help to suggest intervention strategies that reflect specific local circumstances and the cultural factors that influence them , and plan activities that are suited to the respective population involved [12] . Knowledge , attitudes and practices surveys are frequently used in health-seeking behaviour research with knowledge being assessed in order to see how far community knowledge corresponds to biomedical concepts[12] . A limitation in these kinds of studies is that other types of knowledge tend to be highly neglected with very little information being sought on knowledge about the health system [13] . Further , attitudes are never easy to obtain from such a survey while questions on practices are usually hypothetical , hence hardly permitting statements about actual practices or taking into consideration underlying contextual factors , hence affecting their reliability[14] . Besides the limitations of KAP surveys , they are very useful for assessing distribution of community knowledge in large-scale and for evaluating changes in knowledge after education and media campaigns . They permit rapid assessments , yielding quantitative data , and are therefore a cheap way to gain quick insights into main knowledge data[13] . Combining both qualitative and quantitative methods can address some of the limitations of each method ( eg qualitative methods can help identify potential respondent bias present in KAP and inappropriate framing of questions , while quantitative surveys can collect data at scale ) . Both qualitative and quantitative studies play a role in highlighting information on community knowledge , perceptions and acceptance of HAT related control interventions . Qualitative studies on HAT in Uganda , South Sudan and the DRC have been useful in highlighting the importance of understanding community attitudes and perceptions in effective engagement in HAT interventions [10 , 15 , 16] . Similarly , a number of surveys on community’s KAP on HAT have been carried out in various endemic countries , including Tanzania [17] , the DRC [18] and Nigeria[19] . These studies highlight people’s knowledge about cause , symptoms and prevention of HAT , people’s beliefs about the disease and where they would seek for treatment . They also highlight the notion that knowledge about tsetse and HAT can help communities understand and support tsetse and HAT control interventions . Given that there is a dearth of KAP studies in South Sudan , this survey sought to establish gaps in KAP of the community in Yei county in relation to HAT , and to identify the preferred channels and sources of information for HAT . The aim was to gather relevant information that would be used to develop an effective communication strategy and community awareness programme , through information education and communication ( IEC ) campaigns to run alongside the passive screening and diagnosis strategy recently introduced in the country .
Ethical clearance for this study was granted by the Ministry of Health , Republic of South Sudan before commencement of activities . Prior to the interviews and discussions , all the eligible participants were consented individually . At that time the objectives , procedures , benefits and discomforts of the study were explained and they were assured of confidentiality . Their voluntary participation was recorded through a thumb print or signature before the interviews were conducted . Yei county is located southwest of Juba , South Sudan and lies close to the border of Uganda and the Democratic Republic of Congo ( DRC ) . It has an area of 6 , 730 km2 , is administratively divided into five payams , including Yei town , Otogo , Lasu , Mugwo and Tore . It is further divided into 22 Bomas and 100 villages , and has 31 health facilities and 74 educational facilities . The 2015–2020 projections indicate the human population of Yei county to be about 267 , 656 , with 33 , 393 households [20] . The majority socio-linguistic groups are the Kakwas followed by Bari . Others include Avukaya , Pojulu , Kuku , Mundu and Keliko . The area receives adequate rainfall all year round and is suitable for food and cash crop farming . The main occupations of the inhabitants include farming , livestock farming and fishing . This social survey was carried out from 17th November to 9th December , 2015 . The baseline KAP survey ( hereafter referred to as survey ) reported here involved a cross-sectional descriptive research design utilising mixed methods , whose unit of analysis was the individual participant . Each of the five payams ( districts ) in Yei county were included in the sample . Smaller units , that is , the boma/village in each of the payams , formed the units from which systematic sampling of households took place . According to Raosoft 2004 , a sample size of 560 respondents was envisaged , given a confidence level of 95% , an error margin of 4 . 08% and a response distribution of 50% . However , the sample size was increased to 610 to ensure that all the bomas were included . The questionnaires were administered proportionately in five main payams , including Lasu111 ( 18 . 4% ) , Yei town 166 ( 27 . 5% ) , Mugwo79 ( 13 . 1% ) , Otogo156 ( 25 . 9% ) and Tore 91 ( 15 . 1% ) respectively . A total of 610 respondents across the payams were randomly sampled and interviewed using the structured questionnaires which were proportionately distributed to the various payams then to the bomas and villages . The sampling ensured that all the villages were represented . Households were selected from a random starting point and household members , either male or female , who consented to be interviewed , were selected . Sampling of males and females were alternated to ensure an almost equal representation of both . Both male and female respondents , aged between 18 and 90 years , participated in the study . To complement and contextualize the KAP survey , eight in-depth interviews were held with key informants in charge of HAT related activities from different health facilities , who also happened to have participated in past sleeping sickness programmes . Out of the eight , three of them were former sleeping sickness patients and one had a spouse who had recently been diagnosed with HAT . The purpose of interviewing them was to explore their perspective of community KAP and their individual opinions and experiences ( past and present ) with HAT . One focus group discussion ( FGD ) was held with nine male discussants from one of the payams ( Otogo ) . The participants included those who had tested positive with the HAT rapid diagnostic test ( RDT ) , but had not gone for confirmatory testing . This category of discussants was chosen to try and get their perceptions about HAT and the current diagnostic and treatment interventions as well as the challenges they face in completing the referral process for HAT diagnosis . After the interviews , the participants of the FGD were linked with the head of the Yei county health department to ensure they underwent the HAT confirmatory tests . The aim of the FGD was to explore participant’s opinions and experiences with HAT and its past and current interventions . We had envisaged to have more than one FGD but it became a challenge to get the category of people we were keen on having discussions with , coupled with the long distances on rough terrain , and hence we focused more on the key informants who were easier to access . Before initiation of the study , the study team had a meeting with the Yei county administration team to introduce the study and plan how to conduct the baseline surveys , including mobilizing and sensitizing the community about the study . Research supervisors and assistants were recruited from the community and taken through a one-day training on the basics of conducting the research , including sampling and how to administer the questionnaires . Role plays were carried out to help the research assistants familiarize with the questions in the tool . The questions on attitude using the likert scale gave some research assistants challenges but we managed to put a lot of emphasis in the training on these questions , as well as reviewed all the filled-in questionnaires to clarify any anomalies . Structured questionnaires , composed mainly of closed ended questions focusing on the tsetse and sleeping sickness situation in the study area , and on participants’ awareness , attitudes and practices about symptoms of sleeping sickness , its transmission , diagnosis and management , were administered to all participants . The questionnaires were in English but were translated to the respondents verbally during the interviews . Respondents were asked questions from the questionnaire without having the choices read to them and the research assistants recorded the answers given based on the listed choices . The questionnaire was structured to allow for ease of administration and analysis . However , it had an option of recording anything that did not fall in the given choices , under the category “other” including also a few open ended questions . The questionnaire was based on four main themes , including ( 1 ) level of knowledge and awareness about HAT , ( 2 ) attitudes and perceptions about HAT , ( 3 ) health seeking behavior in relation to HAT , and ( 4 ) sources of information about HAT . The theme on health seeking behavior was a hypothetical question given that not all respondents would have experienced HAT . However their responses gave indications of community health behaviour . Given that communication plays a vital role in health campaigns to create awareness , increase knowledge and skills about health conditions and initiate the uptake of health services , it was therefore necessary to also establish the preferred sources and effective channels of communication in the community . The FGDs/key informant interviews were conducted by the first author who is trained in qualitative methodologies while the second author , with training in community public health and work experience in the study area , recorded the notes . The key informant interviews and FGD were conducted in English but a translator helped with translation when necessary . The key themes in the FGD and key informant interviews included: historical perspectives of HAT including myths , common signs and symptoms of the disease , perceptions about HAT , its patients and preferred channels of information dissemination to the community . Seven out of 610 questionnaires were excluded from the analysis of quantitative data because they were incomplete . Analysis was run on 603 questionnaires which were initially sorted and cleaned and entered in the statistical package for social sciences ( SPSS ) . Descriptive statistics was used to analyse the data using frequencies and percentages . In the analysis , responses related to options that were true of HAT were considered as correct responses , hence indicative of good knowledge of HAT , while responses that were not true of HAT were considered incorrect hence indicative of a gap in knowledge about that specific component . Qualitative data from the FGD and the key informant interviews were analysed using a deductive approach based on already identified themes . Notes from the qualitative data were organized and summarized according to the identified themes and presented in the form of quotes to contextualize the quantitative data .
Out of 603 respondents who participated in the study , 332 ( 55% ) were male and 271 ( 45% ) female , 419 ( 70% ) were between 21 to 50 years old , most ( 79 . 6% ) had attained a primary level education and below and were mainly crop farmers ( 75% ) ( Table 1 ) . Detailed aspects of knowledge about HAT among the communities in Yei are as shown in Table 2 . Human African trypanosomiasis in this community , depending on one’s ethnic background , is known as “ponge” in Kakwa language and “ayanlototo” or “gilolototo” in Bari language . These literally translate to “the disease of sleeping” or “sleeping sickness” . Most ( 99% ) of the 603 respondents had heard about HAT . Although more than three quarters ( 85 . 7% ) reported correctly that the disease is caused by tsetse flies , others mentioned incorrect causes such as mosquitoes ( 3 . 2% ) , shaking hands ( 0 . 2% ) , changes in weather ( 0 . 2% ) , and others ( 1 . 8% ) such as sugar , dirty environment , dirty water and some unknown fly as the causes of HAT . Further analysis in terms of differences in knowledge across the payams , Yei had the most ( 93% ) respondents who gave correct answers , followed closely by Lasu ( 91% ) , Otogo ( 85% ) , Tore ( 75% ) and Mugwo ( 72% ) . The reverse , that Mugwo also had the highest percentage of respondents who gave incorrect answers for causes of HAT , also applies . In terms of gender , more women ( 43% ) than men ( 25% ) gave incorrect responses to causes of HAT while in terms of education , those with at most a primary level of education had higher percentage of people ( 15% ) giving incorrect answers as opposed to those ( 4% ) who had attained at least a secondary level of education . The respondents were asked about the common signs and symptoms of HAT , and all the symptoms they mentioned were ticked against the choices . Given the many varied symptoms mentioned a number of times , the percentages under each symptom are low . However , they still illustrate the common signs and symptoms of HAT as known by the community . The general signs and symptoms of HAT , according to the respondents , included abnormal sleep ( 18% ) , weight gain ( 15% ) and mental problems ( 14% ) . Erectile dysfunction and memory loss were classified in the “others” ( 3% ) category of signs and symptoms . When asked about the three main symptoms of HAT , 21% of the respondents mentioned abnormal sleeping , 20% mentioned severe headache , while 16% mentioned weight gain . However , discussants in the FGDs were in consensus that abnormal sleeping and abnormal itching/rashes were the key signs of HAT . This was also echoed by some key informants . The main correct activities that respondents mentioned as pre-disposing people to get infected with HAT included fishing ( 42 . 8% ) , herding ( 17 . 2% ) and farming ( 16 . 9% ) . A few ( 1 . 5% ) respondents mentioned other incorrect factors , such going to dirty places , sleeping a lot , using dirty water , and walking in the sun . More men ( 43% ) than women ( 35% ) mentioned these incorrect factors . Bathing and washing along the rivers , burning charcoal , walking in the forest , and walking outdoors early in the morning were also mentioned under the ‘other’ category , and are more or less related to correct responses . Correct potential high risk areas for contracting HAT were along rivers ( 50 . 5% ) and bushes/forests ( 34 . 5% ) . Others ( 1% ) included incorrect information such as cold places , dirty places , near dirty water , and in houses where mosquito nets are not used . Tore payam topped the list with the most respondents ( 16% ) citing incorrect responses in comparison to the other payams , as relates to high risk areas or HAT . Mugwo payam followed by 8% , Otogo ( 7% ) , Yei ( 4% ) and Lasu ( 3% ) . Key informants consider rivers in the villages as high risk areas for contracting HAT . The main rivers considered as places with a lot of tsetse flies were River Kobo , River Kiju , River Boko and River Tore . For prevention of HAT , respondents mentioned clearing of bushes around the homestead ( 48% ) , wearing light coloured clothes ( 16% ) , sleeping under bed nets ( 15% ) , vaccination ( 14% ) , and applying insect repellants ( 7% ) . Most ( 68 . 3% ) respondents reported that men are more prone to sleeping sickness , while 15 . 8% cited women as most prone . Other respondents ( 14 . 4% ) did not know which gender is most prone to sleeping sickness . To establish whether aspects of stigma existed , respondents were asked how people perceive those suffering from HAT and their responses were matched to the choices listed in the questionnaire . Most respondents ( 71% ) reported that HAT patients would get community support and help , 13% reported that they would be avoided , 9% reported that they would be rejected , and 7% mentioned others , including being secluded to eat alone , or being kept in their own rooms . Contrary to the aspect of community support , this quote exemplifies the aspect of possible stigma towards HAT cases . “When you test positive for sleeping sickness , people talk bad of you , that you have this sickness so and so… . ” ( FGD ) . One way of gauging the perception of respondents about sleeping sickness included reading several statements to the respondents , and asking them to indicate their agreement or disagreement on a likert scale . Results indicate that a majority ( 69 . 5% ) of the respondents perceive that HAT affects only adults , close to half ( 41 . 8% ) of them perceive it to be a contagious ( person to person ) disease , while others believe that it can be treated by herbalists and witch doctors ( 11 . 7% ) , and that HAT patients can survive without treatment ( 10% ) . More women ( 34% ) than men ( 21% ) perceived that HAT has no cure . Other perceptions were that HAT is a killer disease ( 63% ) , can be treated in any health facility ( 36 . 6% ) , and that its treatment is expensive ( 45 . 5% ) .
Given that not all respondents for this survey were necessarily former HAT patients , some of the questions related to practice were hypothetical and hence their responses to these were not based on reality . Additionally , results from KAP surveys are highly descriptive without explaining why people behave the way they do hence needs to be interpreted with caution [13 , 14] . In this case , ethnographic studies are more applicable to in-depth studies of health seeking behavior , attitudes and practices while placing them in context [14] . KAP surveys have an underlying assumption that there is a direct relationship between knowledge and action . In line with this , the discussions and conclusions should be viewed with this context in mind . However these responses are still valid because they give general indications of what may happen in a real situation of infection with HAT and thus can be used to tailor health education messages . This survey revealed a range of important socio-demographic characteristics of the community in Yei that should be taken into account when developing health education campaigns . Previous studies have shown that socio-economic and demographic factors such as education , age , gender , and main occupation among others , play a critical role in disease epidemiology , and need to be taken into consideration when designing health interventions [21 , 22] . Most of the respondents in the present survey had low levels of education or none at all and are also among the group that still had misconceptions about HAT . The low level of education could present a barrier to effective reading and comprehension of written IEC materials . This is similar to findings by [18] in a KAP survey of HAT in the DRC , who concluded that education among other factors was significant in the acquisition of knowledge . Developing IEC materials for such communities requires the use of simple illustrative methods of communication that are easy to conceptualize . Given that more women than men gave incorrect responses in relation to HAT , communication interventions need to take the gender factors into consideration to ensure that public health interventions do not only target either of the genders by virtue of their position in the household or community , but ensures that both men and women are targeted appropriately to ensure inclusivity of all the household members . Some previous interventions on reproductive health have registered low uptake because spouses of the women enrolled in the programs were not included in the study , and hence they stopped their wives from participating [23] , while on the other hand , some agricultural projects have left out women , given that they only deal with household heads , who are mostly the owners of the farms[24] . On the other hand , given that the main economic activity in the community in this survey was crop farming , the timing of public health campaigns should take into consideration this socio-economic context to avoid scheduling them during peak seasons when activities such as planting , weeding and harvesting are going on . Findings by [25] showed that engagement in farming activities was one of the barriers to community participation in active screening for HAT , as people were reluctant to interrupt their activities to go for screening . Findings from this survey show that some misconceptions still exist among a small percentage of people in relation to the cause of HAT based on gender , educational levels and geographic location ( payams ) . These variables also played a significant role as determinants for acquisition of knowledge of HAT in the DRC [18] . Some of the misconceptions about the cause of HAT in the South Sudan study included mosquitoes , shaking hands and sugar . A study by [10] also noted similar routes of transmission of HAT including food , animal blood and water among others in Uganda . On the contrary , in the DRC , Kenya and Uganda , traditional beliefs , sorcery and witchcraft were perceived to cause the disease [26 , 27 , 28] . Further , Brown ( 1999 ) reported that the Azandes , a Central-African people also found in South Sudan , associated the chronic conditions of HAT as those caused by witchcraft . The solution to getting cured was appeasing the spirit of the person who had caused the disease . However how these beliefs/practices could influence conceptualization of HAT may not have been adequately captured in this KAP survey , potentially because of limitations in the study design . However , from inference , if a community’s causal explanation for disease differs from those of a biomedical perspective , there is a greater chance of areas of conflict in preventive practices[29] . Tore and Mugwo payams seemed to have higher percentage of respondents who did not know key information about HAT as opposed to those respondents from Yei municipality and Lasu payams which had more people knowledgeable about HAT . Similar findings were observed by [10] where acceptance of tsetse control interventions varied in relation to duration of experience with previous tsetse control programs . These could relate to previous experience and exposure with HAT and HAT related activities . Field reports from Malteser International BMZ indicate that over the years ( 2002–2007 ) the number of HAT cases had been higher in Yei followed by Otogo , Lasu , Tore and lastly Mugwo payams . Presence of HAT cases could signify more HAT related activities in that particular area hence more people getting exposed to HAT related information . Differential geographic knowledge might also emerge from learning about the disease from experiences of correctly diagnosed patients . This carries important implications in designing an effective communication strategy for HAT . Many times in the past , interventions have focused more where cases have been reported . However , findings from this survey show the need to also focus on areas with few cases especially in the face of intensifying surveillance and control of HAT in a sustainable manner , because the few undiagnosed cases who may be infected and not treated act as reservoirs of the disease , hence maintaining the transmission cycle [6] The community in the present survey had good knowledge about signs of HAT , with sleep reported as the key sign . Abnormal sleep , the characteristic symptom for HAT , has also been noted in several studies [17 , 26 , 28 , 30] . However , using sleep as a key sign of HAT can have a negative effect on passive screening and early detection of the disease , given that the sleep signs manifest in the second or late stage of the disease , and hence one may not present to a health facility for HAT screening until this sign is manifested . Furthermore , as much as the community members indicated that they would seek treatment as soon as they have symptoms , a problem arises in the fact that the few symptoms experienced at the onset of HAT can easily be confounded with other common diseases for which people might not seek treatment for . This contributes to delayed diagnosis and presents a much bigger problem , not only for the individual , but also for the community as the infected person becomes a source of infection for others . Acknowledging that irregular sleep patterns is a classic symptom of HAT that is commonly used in IEC materials and algorithm promoted by national programmes , it is equally important to note that relying on the ‘sleep’ symptom , which was also noted by the respondents as one of the key symptoms of HAT , in the era of elimination , may slow down the progress towards achievement of this goal . Furthermore , relying on this classical symptom may make those who are asymptomatic to remain undiagnosed in the community . This calls for including and stressing the other non-classical symptoms of HAT in community education and sensitization as well as undertaking screening ( both active and passive ) on a regular basis , given that active surveillance and case treatment have been found to be extremely effective in reducing disease transmission [31] . Fishing , herding and farming as mentioned in this survey to be the key activities that mainly expose people to HAT are in tandem with other reports on the same [32 , 33] . Other pre-disposing activities mentioned by the respondents related to activities carried out in conducive tsetse habitats . Despite the fact that most respondents knew where they were highly likely to contract sleeping sickness , with places along rivers being the high risk areas , about 10% still held on to various misconceptions such as eating together , using dirty water and walking in the sun . Education and communication interventions need to address such inaccurate information about HAT . Almost a quarter of the respondents had misconceptions about prevention of HAT , such as sleeping under bed nets and vaccination . This presents a gap of knowledge on how to prevent HAT , given that there is currently no vaccine for the disease , while tsetse flies mostly bite people outside of their houses when they are carrying out activities in tsetse habitats . Myths and misconceptions still exist as shown in the results , however this study only relied on one FGD , hence a limitation that needs to be noted while situating the study in context . Myths such as one’s male/reproductive organs ‘dying’ because they have been diagnosed with HAT , or that if they go to health facilities they will die , still exist among some of the community members as was also noted in the DRC[16] . The historical perspectives presented in this study relate to the 1990s when MSF–France responded to an outbreak of HAT and set up a treatment centre in Omugo . Patients from South Sudan were normally transported to Omugo hospital for treatment and due to the long distances back home , they were kept for six months until they completed their first follow up . Similar observations were also noted in a study on perceptions of sleeping sickness in Uganda [34] . They further also note the pressure that was put on those who refused to go for testing as well as the perception that the whites were coming to eat the community members or poison them in hospital . The aspect of poison could relate to the toxic drug melarsoprol that was previously used to treat the late stage of the disease [33] . Such perceptions still exist and may hinder people from going for diagnosis of sleeping sickness . Collective memory influences community response to disease control programs [34] . There is therefore a need to address such myths and misconceptions , given that misconceptions act as barriers to the adoption of control strategies . They limit people’s ability to change their behavior , and can spread and negatively influence the rest of the community [35] . Just as [34] indicates , it is important to consider whether memories of the past could interfere with current interventions . Given that some of the myths were brought up by community members who had defaulted in completing the diagnostic process after testing positive with RDTs , it is important to take such perceptions into account when working on changing community health/treatment seeking behavior , given that new interventions may trigger some historical myths . Information on the new diagnostic and treatment regimes are therefore key in shaping and dealing with existing myths and perceptions on HAT . As [34] noted , knowledge on the new diagnostic procedures and new treatment approaches which have shorter durations of hospitalization , limited cases of relapses and few side-effects and death , can improve on community response to testing . The perception that sleeping sickness only affects adults as held by 31% of the respondents is contrary to findings in Tanzania and Sudan , which showed that HAT affects children and adults , as well [1 , 18 , 36] . Although sleeping sickness occurs more in adults due to their higher exposure to tsetse fly bites , the disease can equally affect children , as long as they are exposed to tsetse fly bites , hence the importance of having IEC materials that show that everyone , the young , the old , both men and women are equally at risk of HAT infections by nature of their activities . This implies also mobilizing all members of the community and not just adults for the screening activities , whether active or passive . The education and communication strategies need to also target the health workers to prompt them in considering children in the differential diagnosis of HAT . Despite the finding in this survey that a majority ( 89% ) of the respondents perceive HAT as serious , there is still need to communicate the seriousness of the disease to the community so that the few who perceive it as less serious get to appreciate it as a disease of public health importance that leads to death if not treated . When a disease is perceived as less serious , then minimal or no measures at all are taken against it , which can contribute to more infections and subsequent deaths related to the same . As such , clear communication about the seriousness of HAT should be made to enable the community put in place appropriate preventive measures and timely response in case of suspicion of HAT infection . Most of the respondents were sure that HAT can be cured , but a small percentage ( 6% ) did not know or were not sure whether it can be cured . This has implications in health seeking behavior , given that this category of people in the community may not bother to seek medical intervention when HAT is suspected . A majority of people surveyed said they would offer social support to patients with the disease . However , from both the quantitative and qualitative study , aspects of stigma were still present although not widespread . This is an issue that needs to be dealt with when a communication strategy is developed . However , as [37] notes , much as stigma is a powerful element in determining health behavior and a key factor in social exclusion , it is not the only cause of social exclusion and should not be the driver of sensitization campaigns if it is not relevant . Fear , feeling stigmatized , embarrassment , shame and sadness , or hopelessness when one is diagnosed with HAT , are negative reactions that could hinder community members from seeking health services . As such , patients and the community should be encouraged to have courage to go to the hospital in case they realize that they may have contracted HAT , given that drugs for treatment are available and the prognosis is much better if detected early , besides the new/improved testing and treatment approaches in place [2] . Confirmatory diagnosis of HAT can be undertaken in selected peripheral health facilities , while treatment is only undertaken in specific health facilities . Findings from this survey indicate that when people do not know what is causing their illness or HAT related symptoms , they would seek health services from other sources such as chemists , traditional healers , family members and even prayers . This is similar to findings from Kenya and Uganda , which showed that HAT patients utilized other options besides health facilities in the process of seeking health care , leading to delays in accurate diagnosis , and thus presentation in the late stage of the disease[26] . It is therefore important to take into consideration the few people who , in case of a HAT infection , are most likely not to get a correct diagnosis , since the diagnosis and treatment of sleeping sickness cannot be fully accomplished outside the designated health facility settings . Sometimes seeking of alternative sources for treatment may be due to barriers that hinder access to quality health services . Due to the scope and limitations of this study design , in-depth focus on the alternative/ethnomedical perspectives of the community was not adequately captured . However , in the face of integrated strategies to eliminate HAT , it is important to incorporate the alternative sources where people seek treatment , as part of the HAT control strategies by sensitizing them on HAT and making them key stakeholders in community–based referral programmes for HAT control . Costs of services were identified as key barriers to seeking treatment in health facilities . Similar findings have been reported by [25] where payment for HAT in the DRC was perceived as unpredictable , excessive and unfair . While HAT is diagnosed and treated free of charge in public hospitals , only 64% of the respondents were aware of the same . Such information should be widely communicated to the public to create awareness on the same . Some respondents gave monetary value to the cost of treating HAT . However , treatment of HAT is carried out for free , courtesy of donation of drugs by manufacturers , and supply to endemic countries by WHO at no cost [1] . Hence the cost mentioned by the respondents could have been indirect costs , or just perceptions of cost . With the current effort to eliminate HAT , there is need to make people aware that diagnosis and treatment for HAT is free , so as not to hinder the community from seeking diagnosis and treatment when they suspect HAT . Distance was reported as a barrier to seeking services from health facilities if HAT is suspected . In South Sudan , access to designated HAT screening health facilities was reported to be difficult , given that most villages were located more than an hour’s walk from the health facilities , in addition to poor and unreliable road transport [7 , 30] . In the face of intensifying surveillance and control of HAT using the new screening and diagnostic tools and strategies as part of the integrated delivery of primary healthcare , distance to the designated health facilities for HAT need to be taken into consideration . Discordance of the health facilities’ hours of operation with the working hours of some respondents also presented a barrier to utilization of such health facilities by patients . This finding conforms to findings in the DRC[26] . Lack of trust and negative attitude of health workers was reported as a barrier to utilization of health facilities by suspected HAT patients . Issues of trust between health workers and patients have also been raised as barriers to seeking health care or participating in HAT campaigns in the DRC and Uganda [26 , 38] . Health workers therefore need to be trained on how to handle such patients in a friendly manner , in order to create trust between them and the patients . The preferred sources of information were radio , health workers and village elders , while the most effective places of information dissemination were market places , churches , funeral places , and at health facilities . Similar finding were observed by [38] especially as related to the church and health facilities as key places where information on HAT interventions were conveyed . However , they noted varying responses to this information by ethnic groups hence the importance of taking ethnic variations into consideration . This survey also noted some gender variations in preferred sources of information which are critical to consider in enabling effective dissemination of information . The radio has been hailed as a preferred channel of dissemination of information due to its wide coverage in most areas , rural communities included[39 , 40] . It is therefore important that communication strategy for South Sudan take into consideration the preferred sources and identified channels of communication dissemination . Failure to use trusted channels and sources of information may hinder communities from adopting the messages disseminated [41] . Village elders have also been recognized as important sources of information , given that they are the gatekeepers of the community and sometimes custodians of community culture , hence the need to be made key stakeholders in community-based disease control programs [34] . This survey has highlighted the need to develop a communication package that is capable of reaching the community , for purposes of awareness creation , and for informing the health seeking behavior of respondents . Providing health care providers and community members in HAT endemic areas with better information about treatment-related side effects would also be beneficial for increasing the uptake of HAT control efforts [15] . However one needs to also take into consideration that change in knowledge does not necessarily lead to change in behavior , given that many other factors are equally important in influencing health/treatment seeking behavior [12] . Gender , education level , geographic location , previous experience and exposure to HAT among others , are key aspects to consider and incorporate in developing an effective communication strategy , besides incorporating the gaps in knowledge and practices highlighted in this survey . In addition , health education campaigns should include information that explains the development of HAT , the new/improved testing and treatment approaches and how a late treatment can affect the whole community . The findings of this survey have highlighted useful information in developing an effective communication strategy and community awareness program that can run hand in hand with the passive and active screening interventions by helping to mobilize and sensitize communities on HAT as part of the bigger goal of elimination of the disease .
A limitation inherent in most KAP studies that equally applies to this survey is that the findings relate to the reported versus observed practice . In relation to attitude questions , respondents may give answers they believe to be generally acceptable . This needs to be taken into consideration while interpreting the results . The authors worked with local research assistants who were trained on the use of the tool among other basic research concepts . Careful planning and pre-testing and training of the research assistants was carried out to minimize on the cultural gap and help place any unclear issues into context . This KAP survey has a limitation in that one is not able to probe further on the depth of the responses given however , it helped to elicit general information about HAT . The study included the use of a few qualitative methods to try and explore some of the issues in-depth however , the sample used was small hence may not represent a broad range of community perceptions on HAT , much as it gives some insights on the same . We also did not manage to conduct more than one FGD due to constraints of access , hence the qualitative information in this survey was from 8 key informants and only one FGD . However , the findings provide useful data that future researchers can build upon for further studies . The findings of this survey are specific to Yei county , hence cannot be generalized to other areas with varying contexts , although the survey provides useful information on knowledge , perceptions and practices in relation to HAT that can guide future HAT control interventions . This survey was carried out in 2015 , and since then , the country has undergone some security upheavals . These have caused many members of the community to move to neighbouring countries , mainly Uganda . The findings of this study especially as relates to health/treatment seeking behaviour and the access to health facilities , therefore need to be situated in the current context . However , the information is still useful in improving interventions on control and management of HAT as well as development of a communication strategy which is still applicable to the South Sudanese communities who are still in the country or those who may be refugees in neighbouring countries as well as in populations with similar settings . | Misconceptions about sleeping sickness , a neglected tropical disease transmitted by tsetse flies , can be a hindrance to effective implementation of control interventions especially in the face of accelerating work to eliminate the disease . Understanding community knowledge , attitudes and practices about sleeping sickness is important in developing appropriate material for educating and sensitizing communities at risk of the disease . We conducted a study to establish community knowledge , attitudes and practices , including preferred channels of disseminating sleeping sickness information . Despite the fact that the community in Yei County knew about the disease , existing myths and stigma have the potential of influencing their health seeking behaviour . The radio , community health workers and village elders were the most preferred sources of sharing information with the community . There is need to develop education and awareness material to address issues of existing myths , potential stigma , treat ability of HAT , importance of testing and treatment , as well as provide information on the new/improved testing and treatment approaches for HAT . In addition , this should be provided through use of preferred and trusted sources of information dissemination , which is critical in uptake of HAT control , management and prevention activities . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusions"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"behavioral",
"and",
"social",
"aspects",
"of",
"health",
"african",
"trypanosomiasis",
"tropical",
"diseases",
"geographical",
"locations",
"social",
"sciences",
"south",
"sudan",
"parasitic",
"diseases",
"animals",
"health",
"care",
"glossina",
"research",
"design",
"neglected",
"tropical",
"diseases",
"surveys",
"tsetse",
"fly",
"insect",
"vectors",
"africa",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"zoonoses",
"behavior",
"protozoan",
"infections",
"trypanosomiasis",
"insects",
"disease",
"vectors",
"arthropoda",
"people",
"and",
"places",
"psychology",
"eukaryota",
"survey",
"research",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"health",
"education",
"and",
"awareness",
"organisms"
] | 2018 | Knowledge, attitudes and practices about human African trypanosomiasis and their implications in designing intervention strategies for Yei county, South Sudan |
The dynamics of neuronal excitability determine the neuron’s response to stimuli , its synchronization and resonance properties and , ultimately , the computations it performs in the brain . We investigated the dynamical mechanisms underlying the excitability type of dopamine ( DA ) neurons , using a conductance-based biophysical model , and its regulation by intrinsic and synaptic currents . Calibrating the model to reproduce low frequency tonic firing results in N-methyl-D-aspartate ( NMDA ) excitation balanced by γ-Aminobutyric acid ( GABA ) -mediated inhibition and leads to type I excitable behavior characterized by a continuous decrease in firing frequency in response to hyperpolarizing currents . Furthermore , we analyzed how excitability type of the DA neuron model is influenced by changes in the intrinsic current composition . A subthreshold sodium current is necessary for a continuous frequency decrease during application of a negative current , and the low-frequency “balanced” state during simultaneous activation of NMDA and GABA receptors . Blocking this current switches the neuron to type II characterized by the abrupt onset of repetitive firing . Enhancing the anomalous rectifier Ih current also switches the excitability to type II . Key characteristics of synaptic conductances that may be observed in vivo also change the type of excitability: a depolarized γ-Aminobutyric acid receptor ( GABAR ) reversal potential or co-activation of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors ( AMPARs ) leads to an abrupt frequency drop to zero , which is typical for type II excitability . Coactivation of N-methyl-D-aspartate receptors ( NMDARs ) together with AMPARs and GABARs shifts the type I/II boundary toward more hyperpolarized GABAR reversal potentials . To better understand how altering each of the aforementioned currents leads to changes in excitability profile of DA neuron , we provide a thorough dynamical analysis . Collectively , these results imply that type I excitability in dopamine neurons might be important for low firing rates and fine-tuning basal dopamine levels , while switching excitability to type II during NMDAR and AMPAR activation may facilitate a transient increase in dopamine concentration , as type II neurons are more amenable to synchronization by mutual excitation .
Midbrain dopamine ( DA ) neurons predominantly fire in a low frequency , metronomic manner ( i . e . tonic ) and display occasional , yet functionally important , high frequency , burst-like episodes [1 , 2] . While regular tonic firing is observed in isolated preparations ( i . e . slices ) , tonic firing pattern in vivo is somewhat more variable due to active synaptic inputs [3 , 4] . Such tonic activity is important for maintaining a constant basal level of dopamine in projection areas . Accordingly , abnormal basal DA levels are linked to psychiatric disorders from depression to schizophrenia [5 , 6] . While the maintenance of basal DA levels seem to be critical for normal brain function , a consistent picture has not yet emerged regarding how changes in firing patterns of the DA neuron facilitates this important biological function . Background activity of the DA neuron appears to rely on the intrinsic pacemaking mechanism that generates tonic firing . The current composition producing low-frequency pacemaking in DA neurons is of vibrant debate among researchers in the field . A number of experimental [7–19] studies suggests that the maintenance of tonic firing in at least a subpopulation of DA neurons relies on the interactions of the voltage gated calcium ( Ca2+ ) and SK-type Ca2+-dependent potassium ( K+ ) currents Slow pacemaking in our model relies on a subthreshold Ca2+-K+ oscillatory mechanism , similar to a number of well-established models [4 , 20–24] . Interaction between Ca2+ and Ca2+-dependent K+ currents periodically brings the neuron to the spike threshold and generates a metronomic firing pattern . In our model , spike-producing currents ( fast sodium and the delayed rectifier potassium ) play a mostly subordinate role in this dynamic , adding a spike on top of the oscillations without significant changes to the period or shape of voltage and calcium oscillations , as in the study by Wilson and Callaway 2000 [4] . A number of studies suggest an additional Ca2+-independent oscillatory mechanism [25–27] . In particular , they emphasize the contribution of sodium currents to pacemaking . We review the literature on the mechanisms of DA neuron pacemaking in the discussion section in more detail . The specific composition of currents contributing to oscillations determines the response of the DA neurons to stimuli , their synchronization properties and , ultimately , the computations they perform . In this paper , we use recent experiments to calibrate the dynamical properties of the DA neuron and determine its excitability type . A standard method to classify neuronal excitability is via characterizing the frequency-to-input relationship , or F-I curve . Two major types of excitability can be determined based on how the onset of tonic firing occurs as the applied current increases and the neuron is released from quiescence at the hyperpolarized rest state [28] . A type I-excitable neuron can fire at an arbitrary low frequency near the onset of firing , whereas a type II neuron shows a discontinuous jump to a minimal frequency above a certain current threshold and fires only in a limited range of frequencies [21] . Generally , the onset of repetitive firing occurs through one of two mechanisms: 1 ) a saddle-node bifurcation on invariant circle or SNIC ( type I excitability ) or 2 ) an Andronov-Hopf bifurcation ( type II excitability ) [29] . Type II neurons , such as fast-spiking inhibitory interneurons in the cortex , display precise spike timing even in the presence of noise and are therefore suitable for the implementation of spike time coding [30 , 31] . A type I neuron , such as a weakly adapting cortical pyramidal neuron , was shown to relay the stimulus rate by modulating its own frequency , and , therefore , displayed rate coding [31] . Further , type II neurons display resonance and controlled synchronization in networks [31–33] . For neurons that are tonically active without any injected current , such as DA neurons , the transition to the non-spiking rest state occurs when a sufficiently strong hyperpolarizing current is injected . For these neurons , the excitability type would be defined by the transition from tonically firing to quiescent/excitable: again type I would show a smooth frequency decrease to zero , while type II should show an abrupt transition to quiescence . At this point , there is no direct evidence defining to what type of excitability DA neurons belong , and how the different intrinsic conductances and the different synaptic inputs influence their type . Determining the type of excitability will allow us to predict the behavior of the DA neuron during application/blockade of different currents and better understand computations it performs in different input conditions ( e . g . rate coding vs . resonance at a particular input frequency ) . A number of experimental studies provide indirect evidence of excitability type of the neuron in control conditions and during activation of synaptic inputs . It has been shown that the firing rate of DA neurons increases linearly in response to a ramping depolarizing current until it goes into depolarization block ( e . g . [3 , 34] ) . Further , injection of a tonic hyperpolarizing current to the regularly firing DA neuron in vitro increases its interspike intervals [7] . The firing properties of the neuron in response to a combination of tonic inhibitory and excitatory synaptic conductances were investigated by Lobb and colleagues [35 , 36] . Using the dynamic clamp technique , they injected inhibitory γ-Aminobutyric acid ( GABA ) and excitatory N-methyl-D-aspartate ( NMDA ) receptor conductances in SNc DA neurons . Injection of tonic GABAR conductance decreased the firing rate of the neuron several-fold . Furthermore , the neuron fired at low frequencies when NMDAR and GABAR conductances balanced each other . Thus , NMDAR activation , which strongly increases the firing frequency [37–40] , can be effectively compensated by GABAR activation . Such compensation would be impossible if the inhibition produced an abrupt transition to quiescence and the neuron jumped from a high frequency to zero . The experimentally observed compensation suggests , again , a smooth frequency decrease upon GABAR activation rather than an abrupt transition to the resting state at hyperpolarized potentials . Together , these data resemble the tonic firing/quiescence transition in type I neurons with two distinctions . First , the transition parameter is not an injected current , but an ohmic GABAR conductance . In experiments , a conductance has already been used instead of an injected current to determine the neuronal excitability [41] . Second , the co-activation of the NMDA receptor introduces an additional parameter ( its maximal conductance ) . Both of these extend the definition of excitability into the space of synaptic conductances . Formally , the excitability type is an intrinsic property of a neuron , yet viewing synaptic inputs as changing excitability of a neuron is a powerful concept used to understand neuron dynamics in vivo [31 , 42] . We used the experiments described above to parameterize a model of the DA neuron , determine its type of excitability , and determine how intrinsic and synaptic currents shape the excitability type and , therefore , the computational properties of the neuron . These experiments suggest that the DA neuron exhibits type I excitability in isolation from synaptic inputs and under the balanced influence of excitatory and inhibitory synaptic conductances . However , the excitability type has been shown to vary depending on the intrinsic currents and network connectivity [42 , 43] . For example , modeling results suggest that changes in the intrinsic currents , e . g . L-type Ca2+ current , can switch the excitability type of the DA neuron [44 , 45] . Here we address the variability in the excitability type under different conditions by studying the contribution of intrinsic and synaptic currents to regulation of the low-frequency DA neuron firing .
We investigated the behavior of a simulated dopamine neuron in response to irregular asynchronous GABA and glutamate ( Glu ) inputs to mimic temporal structure of neural firing in in vivo conditions . The Glu input was produced by Poisson distributed spike trains and GABA inputs was explicitly modeled as activity of a population of GABA neurons ( detailed description of the inputs and equations are given in the methods section ) . We quantified changes in the firing rate and the regularity of DA neuron firing in response to synaptic inputs of different strengths ( Fig 1A1 and 1A2 ) . We identified a parameter region where the excitatory and inhibitory inputs balance to produce low frequency DA neuron firing at rates similar to background firing ( Fig 1A1 , between the black lines ) . This happens because asynchronous GABA and Glu inputs ( see rasters in Fig 1B1 and 1B2 ) activate GABARs and NMDARs nearly tonically ( Fig 1B3 and 1B4 ) and provide quasi-constant levels of inhibition and excitation to the DA neuron respectively . Under the influence of these two inputs , the DA neuron fires similarly to the in vitro-like conditions ( tonic inputs ) , but with less regularity , which is typical of the background firing in vivo . An example voltage trace of the DA neuron in response to synaptic inputs formed by asynchronous Glu and GABA populations is shown in Fig 1B5 . Fluctuations in the firing of neural populations innervating the DA neuron can produce irregular spiking as observed in in vivo experiments . Considering that asynchronous inputs produce nearly constant receptor activation ( see Fig 1B3 and 1B4 ) , for further analysis we substituted these Glu and GABA inputs by tonic currents . Moreover , tonic synaptic currents mimic long-lasting injection of the conductances in dynamic clamp experiments [46 , 35 , 36] , or iontophoresis of the agonists [39 , 40] , or bath application of the agonists . Transition from asynchronous inputs to tonic currents is described in the methods section . Our next goal was to reproduce the experimentally-observed compensatory influence of tonic NMDAR and GABAR conductances [35] . Using the dynamic clamp technique , it was shown in vitro that a balanced injection of GABAR and NMDAR conductances leads to DA neurons firing at frequencies comparable with background frequencies ( 1–5 Hz ) . Removal of inhibition in such conditions evokes a classical disinhibition burst ( the disinhibition model of burst generation is well known and described in e . g . [35 , 47–49] ) . Fig 2A reproduces the voltage traces obtained in the experiments by Lobb et al . 2010 [35] . In this example , the simulated DA neuron is tonically active at 1 . 5 Hz during tonic co-activation of NMDA and GABA receptors ( gNMDA = 16 . 9 m S/cm2 , gGABA = 5mS/cm2 ) . Removal of the GABAR conductance produces an episode of high-frequency firing ( Fig 2A ) . Removal of the NMDAR conductance produces a pause in firing ( Fig 2A ) . We explored the range of NMDAR and GABAR conductances that produce tonic firing in the DA neuron model ( Fig 2B ) . Compensation of NMDAR and GABAR activation can be readily achieved near the upper boundary of the firing region ( Fig 2B , blue ) . When both receptors are activated , low frequency tonic activity is observed ( Fig 2A ) . The dot labeled as A on the heat plot indicates conductances taken for this simulation . As in the experiments [35] , the balanced region is stretched linearly on the conductance plane with NMDA/GABA slope around 3 . 4 . Moving to the left on the diagram corresponds to deactivation of the NMDAR current and blocks DA neuron firing due to the remaining GABAR activation ( Fig 2A ) . A pause may also be produced by stronger activation of the GABAR ( Fig 2C ) . Conversely , moving down on the diagram corresponds to deactivation of the GABAR and evokes high-frequency firing ( Fig 2A ) . The firing frequency also increases by moving from the upper boundary of the firing region to the right ( increasing NMDAR conductance; Fig 2D ) . These two directions correspond to two ways of eliciting a DA neuron burst: strengthening NMDA excitation or removing inhibition , respectively . Excessive tonic NMDAR activation leads to a depolarization block , as shown in Fig 2B and 2E at high NMDA and low GABA receptor conductances . Interestingly , application of a tonic GABAR conductance in combination with an excessive NMDAR conductance may rescue high-frequency firing in the model ( Fig 2E ) . Thus , the compensatory influence of GABAR activation removes depolarization block induced by an excessive NMDAR activation and restores the intrinsic oscillatory mechanism required for tonic firing . The smooth frequency decrease to zero as the neuron transition to quiescence when GABAR conductance increases suggests type I excitability for the DA neuron both in in vivo and in vitro like cases ( Figs 1A1 and 2B ) . However , excitability is classically defined by the structure of the transition between spiking and hyperpolarized rest state induced by an injected current , as opposed to a synaptic conductance . We show that the DA neuron exhibits type I excitability by standard definition with a continuous F-I curve and place it in Supporting Information ( S1 Fig ) as this case has less physiological significance than the influence of synaptic currents . We further investigate the influence of intrinsic and synaptic currents on the excitability type of the DA neuron . In vivo , the type of excitability may change due to tonic synaptic inputs [42] , and next we explore how this change occurs in the DA neurons . AMPA receptor may co-activate together with NMDA and GABA receptors in vivo . By contrast to NMDAR , conductance of which is voltage-dependent , AMPA and GABA receptor currents are purely ohmic . Their combination is also an ohmic current: IAMPA+IGABA=Ieff=geff ( Eeff−v ) , Where geff = gAMPA + gGABA is a combined synaptic conductance , and Eeff=gGABAEGABA+gAMPAEAMPAgGABA+gAMPA|EAMPA=0=gGABAgGABA+gAMPAEGABA is a synaptic reversal potential . Fig 8 shows the frequency distribution and the type of bifurcation at the transition to the rest state on the plane of these two parameters: conductance and the reversal potential of the ohmic synaptic current . For instance , if the AMPA receptor is blocked , the reversal potential coincides with the GABAR reversal potential , which is in the range of -90 mV to -70 mV [55] . In this range , an increase in the conductance Eeff leads to a gradual decrease in the frequency to zero and a transition to the rest state via a SNIC bifurcation . By contrast , at higher reversal potentials , the frequency drops to zero abruptly and the transition corresponds to an Andronov-Hopf bifurcation . This suggests a transition to type II excitability for the DA neuron . Thus , elevated GABAR reversal potential or tonic activation of AMPAR leads to a switch in the excitability to type II . AMPAR-mediated input induces depolarization block in the DA neurons and firing cannot be restored . This is consistent with the previous modeling studies showing that NMDA can elicit bursting [56–58] or burst envelope [59] , while AMPA abolishes high frequency firing . The dynamical explanation is that AMPAR activation shifts the minimum of the voltage nullcline across the Ca2+ nullcline , so that for high AMPAR conductance values ( as well as positive applied currents ) , voltage oscillations decrease in amplitude and depolarization block occurs ( S3B Fig ) . Thus , DA neuron firing does not exceed the frequency of ~10 Hz when driven with AMPAR activation , similarly to the experimental results ( see e . g . [37 , 38 , 40] ) , Application of GABAR-mediated input shifts the voltage nullcline even further to the right and makes the stable equilibrium more robust . Therefore , the region of parameters for which spiking is produced is much smaller for combined application of AMPA and GABA than for combined NMDA and GABA activation ( compare Figs 2B and 9A ) . Therefore , the prediction of our model is that a disinhibition burst can be supported by tonic background activation of NMDA but not AMPA receptors . Please note that nullclines can be produced only for the model without the fast sodium current ( see S3 Fig for the illustration of depolarization block in the model with spike-producing currents ) . Together with AMPA and GABA receptors , the NMDAR may be also co-activated , since glutamate binds to both AMPA and NMDA receptors . To make the analysis of the excitability type possible in the parameter space of all three synaptic currents , we further perform 2-dimensional bifurcation analysis . The point marked Bogdanov-Takens bifurcation in Fig 8 is a good predictor of the type of excitability at the boundary between spiking and the rest state . Mathematically , it is defined as a junction of the SNIC bifurcation and the Andronov-Hopf bifurcation , as it appears in the figure . In Fig 10 , we plot this point as a function of the NMDA receptor conductance . As in the previous figure , the transition to quiescence occurs as the combined conductance of the ohmic synapses geff grows . The information about the specific value of the conductance is omitted in Fig 10 because the transition occurs in a dimension orthogonal to the plane of the figure . For example , the transition in Fig 8 is represented by one line at gNMDA = 0 . The diagram in Fig 10 shows that the separation between the types of excitability shifts to lower values of the combined reversal potential for the AMPAR and GABAR currents as the NMDAR conductance increases . However , this shift quickly saturates and is restricted to the range of GABAR reversal potentials . Thus , similar to Ih , NMDAR activation may switch the type of excitability of the DA neuron from type I to type II in a certain window of other synaptic currents received by the neuron . To illustrate the importance of changing DA neuron type of excitability , we simulated heterogeneous populations of DA neurons under two conditions: 1 ) in control ( in the absence of the tonic synaptic inputs ) , and 2 ) during the tonic influence of AMPA and GABA inputs . The DA population is electrophysiologically heterogeneous [60–62] , and its uncoordinated activity produces a homogeneous low-level DA concentration . In order to have similar firing rates and basal DA levels in both cases , we balanced the increase in firing rate produced by the application of AMPAR conductance with GABAR conductance ( note that GABA can balance AMPA only for a very limited range of values ) . In both cases , DA neurons received correlated fluctuating NMDA inputs ( Fig 11A , see methods for the detailed description of the inputs ) . Our simulations show that the population of DA neurons that receive the background synaptic tone produces higher DA levels in response to bursty correlated NMDA input than a population without the synaptic tone ( Fig 11D ) . As described above , DA neurons are type I excitable in the absence of synaptic tone , while AMPAR activation switches DA neuron excitability to type II . Thus , the transition from type I to type II excitability of the DA neurons is accompanied by higher dopamine release in response to a correlated synaptic input . The higher responsiveness is partially due to a greater synchronization of the DA neurons receiving the synaptic tone , as evident by the higher number of peaks in their summed activity in Fig 11C . Type II neurons display more robust synchrony when they receive a common input , even in the presence of independent noise [33 , 63] . Thus , in vivo background synaptic tone might be important not only for regulating basal DA levels , but also for the responsiveness of the DA neurons , so that they are more ready to produce coincident bursts in response to correlated synaptic inputs .
The type of excitability for a neuron determines the neurons’ responses to stimuli and their dynamics in a population ( phase response curve , synchronization , resonators vs . integrators ) . The type of bifurcation determines the type of excitability: neural oscillations that arise via an Andronov-Hopf bifurcation have type II excitability , while those appearing via SNIC have type I excitability [52 , 65] . Based on the bifurcation analysis and frequency responses to hyperpolarizing inputs ( negative injected current and GABAR conductance ) , we have shown that in control conditions , the simulated DA neuron is type I-excitable . It’s known that the type of excitability in vivo may be different from in vitro [42] . In vivo , DA neurons display irregular low-frequency activity occasionally interrupted by high-frequency bursts . This low-frequency regime may reflect the balance of inhibitory and excitatory inputs . We found that , in the most prominent low-frequency regime , the DA neuron is type I excitable , in either high or low synaptic conductance states . The baseline level of dopamine is important for the normal function of the brain . The level is determined by the background activity of the DA neurons . This activity is intrinsically generated by the neuron and controlled by its synaptic inputs [66] , reviewed in [67] . Thus , the capacity of the DA neuron to adjust its firing rate according to the inputs is vital . The graded response curve of a type I-excitable neuron , as opposed to an on-off response of a type II neuron , provides this capacity . Accordingly , at every level of excitation provided by NMDAR input , inhibitory GABA synaptic conductance can balance it and bring the frequency down to an arbitrary low value . A similar hyperpolarizing current activated by dopamine D2 receptors on the DA neuron has also been shown to slow down its firing rather than abruptly block it all together [68] . These are very important autoregulatory functions of the DA system that allow it to adjust basal DA levels in target areas . Persistent sodium current is known to amplify subthreshold oscillations [69] and increases neural excitability of DA neurons by contributing to spontaneous depolarization in between the spikes [25] . Consistent with experimental observations , the subthreshold sodium current increases the firing rate of the DA neuron in the model . Additionally , we found that the current is necessary for achieving gradual frequency decrease upon application of hyperpolarizing current , thus , maintaining type I excitability of the DA neuron . The type of excitability is determined by the internal properties of the currents contributing to pacemaking in the DA neuron . L-type Ca2+ and SK-type Ca2+-dependent K+ currents are the core currents that traditionally constitute this mechanism [7–18] ( but see the section on the mechanisms of DA neuron pacemaking ) . However , our model shows that the mechanism results in type II excitability , in which a hyperpolarizing current blocks the voltage oscillations without restoring a low frequency . Our explanation is that , without the contribution of further currents , the steep part of the Ca2+ nullcline is very close to the minimum of the voltage nullcline ( Fig 4A ) , because they reflect the same event: opening of the Ca2+ channel . This positions the system close to the Andronov-Hopf bifurcation , which occurs whenever the minimum of the voltage nullcline moves across the Ca2+ nullcline . When the subthreshold sodium current is included into the mechanism , the minimum of the voltage nullcline reflects opening of this current , and it is shifted away from the steep part of the Ca2+ nullcline ( Fig 6B ) . This shifts the system away from the Andronov-Hopf bifurcation . Now , a downward shift of the voltage nullcline following inhibitory inputs moves its minimum across the flat part of the Ca2+ nullcline and produces a SNIC bifurcation . In this transition , the firing frequency gradually reduces to zero , and this allows a balance between NMDAR and GABAR conductances and restores the background firing frequency . The influence of NDMA and GABA receptor conductances on the DA neuron have been studied in several papers [23 , 22 , 70] . The compensatory influence of NDMA and GABA receptor activation on the firing frequency has been predicted in modeling studies by Komendantov et al . [23] . Lobb et al . [70] modified our previous model [24] to capture the balance of the inhibition and excitation and disinhibition bursts . In these models , a high maximal frequency ( >10 Hz ) can be achieved by tonic activation of the AMPAR or in response to depolarizing current injection , which contradicts experimental observations [7 , 34] . To the contrary , a number of experimental studies suggest that stimulation of NMDA receptors evokes a burst of high-frequency firing , whereas AMPA receptor activation evokes modest increases in firing [37–40] ( but see [61 , 71] ) . This is an important distinction , which impacts the excitability of the neuron . Further , in Komendantov et al . [23] and Canavier & Landry [22] , the NMDAR conductances were restricted to dendrites , whereas GABAR conductance was somatic . The mechanism of frequency rise during dendritic application of NMDA is different from the mechanism of response to somatic NMDAR stimulation [56 , 24] . Somatic NMDAR stimulation has been shown to elicit high-frequency firing in earlier experiments [46 , 40] and used to achieve the NMDA-GABA balance [35] . Here , we base a new model on our previous model [56] that presented a mechanism for somatically-induced high-frequency firing in a reconstructed morphology first and reduced it to a single compartment . In the current model , we have integrated the mechanism for high-frequency firing together with the balance of NMDAR and GABAR activation . The mechanism of low frequency pacemaking in the DA neurons has been extensively studied . However , it is still a matter of on-going debate in the literature since different experimental results lead to contradicting conclusions , proposing that different currents are critical for the DA neuron spontaneous firing . In a number of experimental and modeling studies it has been shown that spontaneous tonic firing relies on the interactions between the voltage gated calcium ( Ca2+ ) and SK-type Ca2+ -dependent potassium ( K+ ) currents [7–19] . Wilson and Callaway [4] and later Chan et al . [19] showed that calcium-driven slow oscillatory potentials ( SOPs ) drive the spiking rate of the SNc DA neurons . Chan et al 2007 [19] also showed that dependence of pacemaking on Ca2+ oscillations changes with the age . Particularly , TTX blocks slow oscillations in juvenile neurons , but not in adult neurons , which is related to the change in density of Ca2+ channels with age . Ping and Shepard 1996 showed that the frequency of SOPs after the application of TTX is approximately the same as the frequency of spiking . In contrast , in a more recent study , Guzman et al . [26] demonstrated that , in a number of DA neurons , SOPs and spiking frequencies are weakly correlated , and that TTX inhibits spontaneous oscillatory potentials pointing to the importance of sodium currents for pacemaking . A number of other studies also suggest that sodium channels are highly involved in controlling spontaneous DA neuron frequency [26 , 25] , especially in the VTA DA neurons [27] . The sources of the apparent discrepancy in the experimental results were investigated by Drion and colleagues [21] . Based on the combination of experimental and modeling approaches , authors suggested that calcium and sodium currents likely cooperate to produce pacemaking and prevailing mechanism depends on the density of the ion channels in the neuron . Further , authors showed that the lack of correlation between spikes and SOPs does not lead to a conclusion that generating mechanisms are different . The complementary role of the two mechanisms , notably if they co-exist in the same cell or represent pacemaking in distinct populations , is a matter of on-going research in the field . For the sodium-based pacemaking mechanism , it is still unclear what hyperpolarizing current provides the long interspike interval when the SK current is not functional . This renders the Ca2+-independent oscillatory mechanism incomplete and is a matter of a future investigation . Control of the firing by the SK current has been shown to be stronger in the DA neurons positioned more laterally in the midbrain [15] . Thus , we focus on the subpopulation of DA neurons that are more abundant in the substantia nigra pars compacta ( SNc ) than the ventral tegmental area ( VTA ) . Changes in intrinsic currents can affect the excitability type and , thus , computational properties of the DA neuron . For instance , we observed that potentiation of Ih current promotes type II excitability of the simulated DA neuron ( Fig 7 ) . Further , we show that Ih current can induce bistability in the DA neuron , and the bistability region increases with the increase in Ih conductance ( Fig 7B ) . This can affect the behavior of the neuron near the boundary between spiking/resting states , particularly , if the current reached the value necessary to induce spiking onset , a small perturbation in the current will not silence the neuron . This makes spiking more robust near the threshold . In addition to the contribution of Ih current to pacemaker activity , has been shown in DA neurons [72] , as well as in the other neuronal types that Ih induces intrinsic subthreshold resonance [73–75] . Thus , augmentation of Ih current increases oscillatory behavior of the DA neurons , as well as their synchronization in response to excitatory pulses . However , low-frequency tonic firing could not be maintained at high conductances of Ih current , likely affecting background DA levels . Further , the influence of tonic synaptic inputs can also change the transition to the rest state and , therefore , be described as altered excitability . Tonic activation of AMPA receptors or an elevated reversal potential of the GABAR conductance may make the low-frequency balanced state unreachable . The reason for that is a transition to type II excitability: firing is blocked at higher frequencies . A model prediction that follows from this result is that tonic AMPAR activation induces depolarization block and firing cannot be rescued by application of GABA . Our explanation is that shunting is so strong that opening of the subthreshold sodium current cannot sustain the voltage growth . In other words , these changes unfold the voltage nullcline and bring its minimum close to the steep part of the Ca2+ nullcline ( Fig 9B ) . This primes the system for the Andronov-Hopf bifurcation responsible for type II excitability . Further , we found that NMDAR activation also biases the neuron towards type II excitability ( Fig 10 ) . Although the type may change as parameters shift away from the boundaries of the firing region [76] , together , these results suggest that in high-frequency regimes the DA neuron displays type II excitability . This switch in excitability type may play a role in a transient increase in DA concentration in response to salient stimuli as it is easier to synchronize type II neurons by an excitatory input . Fig 11 supports this hypothesis by showing higher transient DA release produced by heterogeneous population of DA neurons receiving synaptic AMPA and GABA tones than in the absence of synaptic tone . Thus , correlated excitatory synaptic inputs are more likely to evoke robust coincidence DA release when DA neurons display type II excitability . A growing body of literature links the type of excitability to neural coding [29 , 31 , 41 , 65 , 77–79] . For instance , Prescott et al . [80] suggested that type I neurons are best suited for coding stimulus intensity . Hence , the DA neuron is designed for encoding the intensity of the tonic depolarizing and hyperpolarizing inputs by its smooth frequency dependence . This further supports and augments a recently found unique computational role for the DA neuron: it performs subtraction of inhibitory and excitatory inputs [81] . The operation is optimal to calculate unpredicted value of an event but rarely observed and hard to implement in the brain . The first type of DA neuron excitability is necessary to quantitatively encode the level of input by the firing rate and perform the subtraction . Several studies , for example Eshel et al . 2016 [82] and Tobler et al . 2005 [83] , showed that activation of DA neurons gradually increases with the increase in the reward value . We attempted to reproduce this experimental result in our model . For the simulations , we assumed that the reward value is proportional to the strength of NMDA input coming to the DA neurons ( gNMDA ) . GABA inputs do not seem to be a plausible candidate because Eshel et al . 2016 [81] showed that firing of the GABA neurons , as opposed to the DA neurons , does not vary consistently with the reward value . We calculated firing frequency dependence of the type I/ type II DA neurons on the input strength ( Fig 12 ) . Type II DA neurons were modeled by applying tonic AMPA along with NMDA conductance ( additional excitation was compensated by increasing GABA conductance ) . Type II DA neurons are unable to encode low reward values as their firing abruptly drops to zero with the decrease in the input strength . In contrast , frequency dependence of type I DA neurons resembles the experimental curve of the DA neuron frequency dependence on the reward value shown in [82] . The difference can also be seen in the raster ( Fig 12A and 12C ) and frequency responses ( Fig 12B and 12D ) to a transient increase in the input strength , representing the reward value . First peak in the frequency response represents salience , and was simulated by applying constant level of NMDA for all the values , while second peak represents the reward value , and was simulated by scaling NMDA input accordingly . There is a gap in a frequency response to low reward values of type II DA neurons ( Fig 12D ) . Thus , type I DA neurons best encode the value of an event , i . e . the difference between predicted and received reward . On the other hand , spike timing of type I neurons in response to weak or noisy transient inputs is not reliable [84 , 85] , while temporal precision of type II neurons is much higher . The ability of the DA neuron to switch excitability type from type I to type II under certain synaptic inputs might play a significant role in producing enhanced transient DA release , since it likely relies on the precise coordinated activity of the DA neurons . Multiple drugs of abuse , including EtOH ( e . g . [86] ) evoke transient increases in the DA concentration in nucleus accumbens . Using our model , we show that EtOH shifts excitability of the DA neurons to type II and induces higher DA release . EtOH acts on multiple intrinsic and synaptic currents . Mainly , it enhances Ih current [87 , 88] and increases AMPA/NMDA ratio [89] . Moreover , it increases GABA release onto DA neurons [90] . Thus , we modeled EtOH action by increasing Ih , AMPA and GABA receptor currents . Ih and AMPAR currents switch DA neuron excitability to type II and , therefore , promote synchronization in the population of DA neurons in response to noisy excitatory inputs ( Fig 13 ) . This is one of the mechanisms that can produce higher DA transients . By switching DA neuron to type II , EtOH can increase the motivational potential of a stimulus because the same excitatory input produces enhanced DA signal under EtOH . In other words , neutral stimulus can become salient after EtOH exposure . Our modeling prediction regarding the excitability switch after EtOH could be tested in-vitro by studying how the shape of F-I curve changes after EtOH . In in-vivo conditions it could be checked by stimulating DA neurons before and after EtOH exposure with a chirp pattern signal in order to check for spiking resonance . To test whether DA neurons better synchronize after EtOH , a phase response curve ( PRC ) or a spike triggered average ( STA ) could be calculated . Presence of the negative PRC component and narrow STA indicates that neurons are more amenable to synchronization by common synaptic noise . In conclusion , our results predict that DA neurons can exhibit traits of both integrators and resonators and these traits are modulated by intrinsic and synaptic conductances . Depending on the current constitution , DA neurons can perform rate coding by integrating slow variations in the inputs and adjust basal DA concentration or they can detect transient coherent changes in the inputs and synchronize for producing robust DA transients .
The main currents of the model that produce pacemaking activity of DA neuron are an L-type voltage-dependent calcium current ( ICa ) and an SK-type calcium-dependent potassium current ( IKCa ) . Gating of the calcium current is instantaneous and described by the function: gCa=gCa¯⋅αc4 ( v ) αc4 ( v ) +βc4 ( v ) ( 2 ) Calibration of the calcium gating function reflects an activation threshold of an L-type current , which is significantly lower in DA cells than in other neurons ( ~ -50mV; [4] ) . Calcium enters the cell predominantly via the L-type calcium channel . Contribution due to the NMDA channel is minor [92] . Thus , calcium concentration varies according to the second equation of the system ( 1 ) . It represents balance between Ca2+ entry via the L channel and a Ca2+ component of the leak current , and Ca2+ removal via a pump . In the calcium equation , β is the calcium buffering coefficient , i . e . the ratio of free to total calcium , r is the radius of the compartment , z is the valence of calcium , and F is Faraday’s constant . Pca represents the maximum rate of calcium removal through the pump . A large influx of Ca2+ leads to activation of the SK current , which contributes to repolarization as well as afterhyperpolarization of the DA cell . Dependence of the SK current ( IKCa ) on calcium concentration is modeled as follows: gK , Ca=gK , Ca¯⋅[Ca2+]4[Ca2+]4+[K+]4 ( 3 ) The neuron is repolarized by the activation of a large family of voltage-gated potassium channels . In addition to the already described potassium current , the model contains voltage-dependent K current ( IK ) . Conductance of this current is given by a Boltzmann function: gK=gK¯⋅11+exp ( − ( v+10 ) 7 ) ( 4 ) The DA neuron expresses voltage-gated sodium channels that carry a large transient current during action potentials ( a spike-producing sodium current ) and a non-inactivating current present at subthreshold voltages ( a subthreshold sodium current ) . Even though the persistent subthreshold sodium current is much smaller than the transient spike-producing current , it influences the firing pattern and the frequency of the DA neuron by contributing to depolarization below the spike threshold [25] . We modeled the voltage dependence of the subthreshold sodium current as follows: gsNa=g¯sNa11+exp ( − ( v+50 ) 5 ) ( 5 ) The kinetics and the voltage dependence of the subthreshold sodium current were taken from [93] . The majority of DA neurons express a hyperpolarization-activated nucleotide-gated ( HCN ) inward cation current ( Ih ) . The HCN current contributes to spontaneous firing of subpopulations of DA neurons [94] . The activation variable of Ih is governed by a first-order ordinary differential equation ( the third equation of the system ( 1 ) ) . The maximal activation of Ih current is described by the following voltage-dependent equation [88] q∞=11+exp ( v+958 ) ( 6 ) The voltage-dependent time constant is described by τq=625⋅exp ( 0 . 075 ( v+112 ) ) 1+exp ( 0 . 083 ( v+112 ) ) ( 7 ) The leak current ( Ileak ) in the model has the reversal potential of -35 mV , which is higher than in the majority of neuron types . In DA neurons , several types of depolarizing , nonselective cation currents are expressed , which likely contribute to depolarization during interspike intervals . DA neurons receive glutamatergic ( Glu ) excitatory drive through AMPA and NMDA receptors and inhibitory drive through GABA receptors . Changes in the membrane potential induced by synaptic conductances are described by the following equation Isyn=gNMDA ( v ) sig ( sNMDAact ) ( ENMDA−v ) ︷INMDA+gAMPAsig ( sAMPAactsAMPAdes ) ( EAMPA−v ) ︷IAMPA+gGABAsGABA ( EGABA−v ) ︷IGABA ( 11 ) where gNMDA ( v ) , gAMPA , gGABA are the maximal conductances of NMDA , AMPA and GABA receptor currents accordingly , sNMDA , sAMPA , sGABA are gating variables that depend on the input spike trains . The AMPA and GABA conductances are voltage-independent , but the NMDA conductance has voltage sensitivity as in [95] gNMDA ( v ) =g¯NMDA1+0 . 1[Mg2+]e−mev ( 12 ) where [Mg2+] denotes the amount of magnesium , taken to be 0 . 5μM . The low slope of the voltage dependence ( me = 0 . 062 ) is critical for the increase in the frequency of spikes or subthreshold oscillations during NMDA application [56] . A multidimensional system can be analyzed with two-dimensional nullcline methods only after its reduction to a two-dimensional system . The description of this method is provided by Strogatz [104] and Rinzel and Ermentrout [29] . The reduction was done by eliminating the spike-producing currents , which do not significantly change the firing frequency of the neuron [4] . Nullclines are the curves where either dvdt=0 or d[Ca2+]dt=0 . Accordingly , nullclines of our system were obtained by numerically solving the following equations: dvdt=1cmgCa ( v ) ( ECa−v ) + ( gKCa ( [Ca2+] ) +gK ( v ) ) ( EK−v ) +gsNa ( v ) ( ENa−v ) +gl ( El−v ) =0;d[Ca2+]dt=2βr ( ( gCa ( v ) +0 . 1gl ) zF ( ECa−v ) −PCa[Ca2+] ) =0 . ( 16 ) The type of bifurcation was determined by systematically varying parameters of the system ( in our case gGABA ) until the behavior of the system qualitatively changes . The model of DA release is adopted from Wightman and Zimmerman ( 1990 ) [105] and is described by the following equation d[DA]dt=[DA]maxδ ( t−tspike ) −Vmax[DA]Km+[DA] ( 17 ) The first term describes the release from spiking activity of the DA neuron . Dirac delta function δ ( t − tspike ) represents the release at time of a spike . Maximum amount of DA released per spike is [DA]max = 0 . 1μM . The second term represents DA uptake described by Michaelis-Menten equation , where Vmax = 0 . 004μM/ms is the maximal rate of uptake by a transporter and Km = 0 . 2μM is the affinity of the transporter for dopamine . Heterogeneity in the population of DA neurons was putatively introduced by varying the leak conductance . Further , neurons received correlated fluctuating NMDA inputs . NMDAR conductance to each DA neuron was given by a linear summation of Ornstein-Uhlenbeck ( OU ) processes [106] described as following: gNMDA ( t ) =μ+σ ( 1−cxi ( t ) +cxc ( t ) ) ( 18 ) where μ = 1 . 5mS/cm2 and σ = 0 . 5mS/cm2 are the values of the mean and the standard deviation of the NMDAR conductance used for the simulations . xc ( t ) is the common component of the NMDA input that was applied to all of the DA neurons , whereas xi ( t ) is the independent component , which was generated individually for each neuron . A shared fraction of the input is determined by the input correlation c and was set to c = 0 . 5 . Each OU process was formed by the following equation: dx=−xτdt+Nτxdt ( 19 ) where x ( t ) is Gaussian white noise with zero mean and unit variance . Nτ = ( 2/τ ) 1/2 is a normalization constant that makes x ( t ) have unit variance . A correlation time τ = 5ms was used [33 , 64] . In the model with fast sodium and the delayed rectifier potassium spike-producing currents , a spike was registered whenever voltage oscillation reached the threshold of 0 mV . In the reduced model ( without spike-producing currents ) , a spike was registered every time voltage oscillations crossed the threshold of -40 mV , as experimentally it was shown that a DA neuron action potential is triggered when the voltage is depolarized to approximately -40 mV [3] . Voltage oscillations that were below these thresholds in the models with and without the spike-producing currents respectively were not counted as spikes and did not contribute to the firing frequency . To analyze firing pattern of simulated DA neuron in the presence of different synaptic currents , we quantified its firing rate and bursting . Mean firing rate of the simulated DA neuron was calculated as an inverse of the mean interspike interval ( ISI ) . To calculate bursting we used ISI coefficient of variation ( CV ) , calculated as the SD/mean of 200 ISIs . | Dopamine neurons play a central role in guiding motivated behaviors . However , complete understanding of computations these neurons perform to encode rewarding and salient stimuli is still forthcoming . Network connectivity influences neural responses to stimuli , but so do intrinsic excitability properties of individual neurons , as such properties define synchronization qualities and neural coding strategy . We investigated the excitability type of the DA neuron and found that , depending on the synaptic and intrinsic current composition , DA neurons can switch from type I to type II excitability . In short , without synaptic inputs or under balanced excitatory and inhibitory inputs DA neurons exhibits type I excitability , while excitatory AMPAR inputs can switch the neuron to type II . Type I neurons are best suited for coding the stimulus intensity due to their ability to decrease the firing rate smoothly . Type I excitability might be important for achieving low a basal DA concentration necessary for normal brain functioning . Switching to type II excitability further enables robust transient DA release of heterogeneous DA neuron population in response to correlated inputs , partially due to evoked population synchrony . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"action",
"potentials",
"neurochemistry",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"engineering",
"and",
"technology",
"electronics",
"depolarization",
"membrane",
"potential",
"electrophysiology",
"neuroscience",
"organic",
"compounds",
"hormones",
"receptor",
"potentials",
"amines",
"neurotransmitters",
"catecholamines",
"dopamine",
"rectifiers",
"animal",
"cells",
"chemistry",
"gamma-aminobutyric",
"acid",
"biochemistry",
"cellular",
"neuroscience",
"cell",
"biology",
"organic",
"chemistry",
"physiology",
"neurons",
"biogenic",
"amines",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"neurophysiology"
] | 2016 | Dopamine Neurons Change the Type of Excitability in Response to Stimuli |
Australia is underprepared for a rabies incursion due to a lack of information about how a rabies outbreak would spread within the susceptible canine populations and which control strategies would be best to control it . The aim of this study was to collect information to parameterize a recently developed dog rabies spread model as well as use this information to gauge how the community would accept potential control strategies . Such information–together with model outputs–would be used to inform decision makers on the best control strategies and improve Australia’s preparedness against a canine rabies incursion . The parameters this study focussed on were detection time , vaccination rates and dog-culling and dog movement restriction compliance . A cross-sectional survey of 31 dog-owners , using a questionnaire , was undertaken in the five communities of the Northern Peninsular Area ( NPA ) in northern Australia regarding community dog movements , veterinary visits , reporting systems , perceptions of sick dogs and potential human behaviours during hypothetical rabies outbreaks . It highlighted the significant shortfalls in veterinary care that would need to be vastly improved during an outbreak , who educational programs should be targeted towards and which dog movements should be restricted . The results indicate that men were significantly more likely than women to allow their dogs to roam and to move their dogs . The current low vaccination rate of 12% highlighted the limited veterinary services that would need to be substantially increased to achieve effective rabies control . Participation in mass vaccination was accepted by 100% of the respondents . There was lower acceptance for other possible rabies control strategies with 10–20% of the respondents stating a resistance to both a mass culling program and a ban on dog movements . Consequently , movement bans and mass dog culling would have limited effectiveness as a control strategy in the NPA community . More than half of the respondents said that they would report their sick dogs within a week . This would lead to a much more optimistic rabies detection time than observed in other regions with recent dog rabies outbreaks . Findings from this study can be used to parameterize a recently developed dog rabies spread model as well as to develop informed policies for managing a future rabies incursion , thus improving Australia’s preparedness against a canine rabies incursion .
Rabies is an acute viral zoonosis that causes approximately 60 , 000 human deaths annually , despite being preventable [1] . The disease occurs worldwide , with half of the annual deaths occurring in Asia [1 , 2] . Although Australia is one of the few countries free of canine rabies [3] , the increasing number of islands becoming infected in Indonesia − including Bali , Flores , Ambon and Yamdena − has brought rabies within 300km of Australia ( Fig 1 ) and the risk of an incursion is escalating [4–6] . The Northern Peninsula Area ( NPA ) − located in far north Queensland and adjacent to the Torres Strait − is the most likely location for a rabies incursion [7 , 9] . Not only is it close to Indonesia , the NPA has similar characteristics to Bali , which experienced an incursion in 2008 [4 , 7] . These include large populations of free roaming domestic dogs that can facilitate disease spread , inefficient surveillance and vaccination , plus limited access for dog owners to health resources [6 , 8] . It is vital to understand how a rabies outbreak would spread in such high-risk areas of northern Australia in order to develop response plans [9 , 10] . It is also necessary to understand community attitudes and perceptions towards control strategies to anticipate potential barriers to implementation . However , such information is lacking because − except for one isolated incursion in the 1860s − there has never been an outbreak of rabies in Australia . Consequently , Australia is underprepared for a potential rabies incursion [7] . Epidemiological models are powerful tools that provide insight on disease spread and impacts [11–15] . Rabies epidemiological models have mainly been used in rabies endemic regions to refine and evaluate control strategies . Such models are also invaluable in rabies free areas to simulate outbreaks and evaluate potential control strategies . However , models for use in rabies free regions are scarce . There is only one recently developed model able to simulate a rabies outbreak in Australia and how different control strategies would influence its spread [9] . This model was based on limited data and many assumptions from northern Australia ( the NPA and East Arnhem shire ) and as a result , some key parameters for detection time and the main control strategies ( rates for dog vaccination , dog culling and movement restrictions of dogs ) are not based on extensive empirical data [9] . This study aimed to provide these data and to enable more accurate parameterization of the model to increase its predictive power . In addition , the study assessed community attitudes towards potential control strategies , and thus their efficacy in the event of a rabies incursion into the NPA .
A household survey was conducted in the NPA , which is a local government area located at the northern tip of Cape York ( Fig 1 ) . The NPA has a land area of 105 , 691 ha and consists of five communities: Bamaga , Umagico , Injinoo , New Mapoon and Seisia with the distances between individual communities ranging from 2 to 5 km . In 2011 there were 811 households in the NPA and a population of 2 , 298 [16] . A census of dogs in the NPA , conducted in 2009 , estimated the total dog population to be 437 within 276 households , with 1 . 6 dogs per household [17] . A questionnaire was designed ( S1 Questionnaire ) with a combination of closed- and open-ended questions . The majority were closed questions with yes/no options . Some hypothetical questions were included to gauge what respondents would do if there were a rabies outbreak in community dogs . The questionnaire had four sections: 1 . human and dog demographics; 2 . dog movements within the community; 3 . current dog vaccination rates; and 4 . estimated detection time for rabies . Section 1 sought information on human demographics and the age , sex and number of dogs owned in that household . In Section 2 , questions were asked about dog movements , either free roaming movements or human mediated movements , and whether the respondents ever restricted their dogs in the preceding 12 months . Some questions also focussed on whether the owners would hypothetically change their dog restriction practices if there were a rabies outbreak . In Section 3 , the current level of vaccination ( such as for canine parvovirus , distemper and infectious hepatitis ) in the dog population was estimated based on when the dogs were last taken to a veterinarian , when their last vaccination was and what it was given for . Respondents were also asked whether they would hypothetically vaccinate their dogs if a rabies outbreak occurred in their community as well as if they would euthanize their dogs during an outbreak . Section 4 focused on estimating a potential detection time of a disease outbreak by asking questions about how respondents would describe sick dogs , how long they would wait until seeking advice about their sick dog and where they would go for veterinary services if they had been bitten by a dog in the last 12 months , if their dog had been bitten by another dog in the last 12 months and if they had reported any of these incidents . The survey was conducted from June 15 to 18 , 2015 . The questionnaire was administered with the assistance of the local NPA Animal Management Worker ( AMW ) , who identified eligible and willing participants , explained the reasons for the survey and provided information regarding confidentiality and confirmed verbal consent to participate prior to the interview . The AMW is a community member and local government employee who has undergone basic training in animal management and handling and is usually the first point of contact for community members for animal related problems . They are also responsible for distribution of animal related information within the community . All questionnaires were administered in person as face-to-face interviews by the first author . Only community residents who were dog owners were selected . For ethical reasons , an age restriction of 18 years or older was also applied for the respondent selection . All questions were asked and answers were given in English , with interpretation help from the AMW as needed . Questionnaires were generally conducted at the dog owner’s place of residence , with their dogs usually being present . A small number of interviews were conducted at the respondents’ workplace . The survey was approved by The University of Sydney’s Human Research Ethics Committee ( number 2013/757 ) . The data were compiled in Microsoft Excel ( S1 Dataset ) . Responses to open-ended questions were categorised to allow for easier interpretation and analysis . For example , the answers for how respondents would describe a sick dog were condensed into categories such as behaviour , skin condition , physical ailments and body condition . Descriptive analyses were carried out for each question and included count and percentage for categorical responses ( e . g . “Did you move your dog ( s ) in the previous 12 months ? ” ) , and median and range for continuous responses ( e . g . “How many dogs did you own in the previous 12 months ? ” ) . The demographic variables were condensed into two variables per category for statistical analysis: male and female for human gender , young ( 20–39 years ) and old ( >-40 years ) for human age; and ≥3 versus ≤2 dogs per house for number of dogs per house . Fisher’s exact test was used to explore the associations and due to the relatively small sample size , a liberal significance level of P = 0 . 1 was used to identify significant associations . The analyses were carried out using the R statistical program ( Version 0 . 98 . 1091 ) [18] .
Thirty-one dog owners were interviewed in the survey with each person representing one household . Based on the 2009 survey , which estimated 276 dog-owning households , this represents approximately 11% of the dog owning households in the NPA [17] . All five communities were represented in the survey . Umagico and Seisia were both overrepresented: the percentage of dog owning houses surveyed ( 14% and 29% , respectively ) was greater than the overall survey percentage . Bamaga and New Mapoon were under represented ( each approximately 7% ) and Injinoo was substantially underrepresented ( 3% ) . The age of most respondents was between 40−49 ( 42% ) years and most ( 81% ) were male . This differs from the overall population in 2011 , when the median age was 22 years and there was nearly a 1:1 male to female ratio ( Table 1 ) [16] . The difference in population ages between the study and the census data was a result of the ethics restrictions allowing for only owners over the age of 18 . This skewed the study age towards the older end of the scale compared to the census data that includes all ages . A total of 74 dogs were owned by the 31 respondents , an average of 2 . 4 dogs per household ( range 1−5 ) . Most dogs in these households were 1−4 years of age ( Table 2 ) . However , a large proportion ( 34% ) of dogs were of unknown age ( Fig 2 ) . The dog sex ratio was equally distributed ( 51% males , 49% females ) and a small proportion were de-sexed ( 18% ) . More than half of the interviewed community members ( 58% ) allowed their dogs to roam within their community of residence and many dogs had been moved by owners during the preceding 12 months , mostly for pig hunting and camping ( Table 3 ) . Most respondents had imposed some sort of restriction on the movement of their dogs in the last 12 months , by closing the gate , chaining the dogs or keeping them inside the house . In the event of a disease outbreak the proportion of respondents who would restrict their dogs' movements increased by 6% from 87% to 93% ( Table 3 ) . Of the respondents that would restrict their dogs during a disease outbreak , most would impose more restrictive conditions than in normal circumstances . For instance , if the gate was closed in normal circumstances , the owner would chain or keep the dog inside during a disease outbreak . Likewise , if the dog were chained normally , during an outbreak the owner would keep the dog inside . The minority of respondents visited a veterinarian in the preceding 12 months; most had visited a veterinarian in 2013 or earlier ( Table 4 ) . More than half of the respondents said their dogs have had a “needle” in their lifetime , with 40% of these respondents reporting them as vaccinations . As the term “needle”was substituted in the questionnaire for any injections given to ease communication , the rest of the needle administrations ( 60% ) were for reasons such as mange and worming treatments and arthritis alleviation . However , of needles given , only a small proportion had been during the preceding 12 months and not necessarily administered by a veterinarian . All respondents were willing to vaccinate their dogs in the event of a disease outbreak . Only a small number of respondents ( 10% ) were opposed to or unsure of euthanizing their dogs during a disease outbreak if they were prompted by the AMW or by local influences . However , the number of respondents opposed to euthanasia increased by 6% if they were forcibly told to euthanize their dogs by a non-local government official . The most common signs respondents used to determine if a dog was sick were physical ailments ( scratches , limping , pus ) and skin conditions ( mange and hair loss ) . Other signs identified were body condition ( skinny ) , behavioural changes ( not coming when called or very lethargic ) and gastrointestinal signs ( diarrhoea and vomiting ) ( Table 5 ) . Most respondents would wait a few days before reporting their sick dog and the majority would report it to the animal management worker . A small number of respondents would report a sick dog to the veterinarian ( 5 ) , ranger ( 1 ) , hospital ( 1 ) and manager of the local abattoir ( 1 ) . Half ( 50% ) of the respondents who were bitten by dogs reported their injury and the event . Conversely , only 10% of respondents reported that their dogs had been bitten by other dogs . There were few significant associations between human demographic and behavioural and knowledge variables ( Tables 6 and 7 ) . Men were 9 . 8 times more likely to allow their dogs to roam than women and only men took their dogs outside of the community , mainly for pig hunting ( P = 0 . 007 and P = 0 . 059 respectively; Table 6 ) . Having two or less dogs in a house meant there was significantly lower chance of one of the dogs being sick in the last 12 months compared to houses with three or more dogs ( P = 0 . 008; Table 7 ) , whilst younger dog owners were 5 . 01 times more likely to have had more dogs sick in the last 12 months than older dog owners ( P = 0 . 058; Table 6 ) . Although not statistically significant , there was a trend between having >2 dogs in the house and having more dog-dog bites ( P = 0 . 15; Table 7 ) as well as more human mediated movements ( P = 0 . 15; Table 7 ) .
This study successfully collected information on dog health management in remote , northern Australian indigenous communities to better parameterize a rabies epidemiological model . It revealed potential flaws in a dog movement ban , as the compliance of dog owners was not 100% , and emphasised significant shortfalls in veterinary care that would need to be vastly improved during an outbreak to reach the 70% coverage recommended to control rabies . The detection time was optimistic compared to the current model estimation and other rabies detection times seen in previous foreign outbreaks . The study also provided useful information on how the control programs of dog vaccination , culling and movement bans would be accepted by the dog-owning community and highlighted issues to be targeted by educational programs and potential barriers to implementation ( such as potential decreased compliance when non-local government officials are involved ) . Both types of information could be used to better inform decision makers on best practice for containing a potential rabies outbreak in this high-risk region and therefore improve preparedness against a rabies incursion . However , more detailed information is needed to understand potential barriers to implementation of control strategies , and the impacts of rabies control strategies on the wider community . | Australia is underprepared for a rabies incursion due to limited information about how a rabies outbreak would behave and which control strategies would be best to control it . A disease spread model of rabies has been developed to help policy-makers decide on the best response to a rabies incursion . However , data to inform this model are lacking . Therefore , the aim of this study was to gather information to parameterize the existing rabies spread model and to gauge how the community would accept potential control strategies . A survey of dog-owners , using a questionnaire , was undertaken in five remote , indigenous communities in northern Australia . We found that compared to women , men were more likely to allow their dogs to roam and to move their dogs . The current vaccination rates in these dog populations are low due to limited veterinary services . This would make delivery of vaccine in the event of a rabies incursion potentially challenging . However , compliance of dog owners with mass vaccination campaigns would be high . However , compliance with movement control of dogs might be problematic , as would the mass culling of dogs , although , rabies detection following an incursion could optimistically occur within a week . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"demography",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"australia",
"dogs",
"animals",
"mammals",
"preventive",
"medicine",
"rabies",
"animal",
"behavior",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"zoology",
"veterinary",
"science",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"behavior",
"people",
"and",
"places",
"oceania",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"amniotes",
"organisms"
] | 2016 | A Survey of Dog Owners in Remote Northern Australian Indigenous Communities to Inform Rabies Incursion Planning |
B cells and antibodies are involved not only in controlling the spread of blood circulating Trypanosoma cruzi , but also in the autoreactive manifestations observed in Chagas disease . Acute infection results in polyclonal B cell activation associated with hypergammaglobulinemia , delayed specific humoral immunity and high levels of non-parasite specific antibodies . Since TNF superfamily B lymphocyte Stimulator ( BAFF ) mediates polyclonal B cell response in vitro triggered by T . cruzi antigens , and BAFF-Tg mice show similar signs to T . cruzi infected mice , we hypothesized that BAFF can mediate polyclonal B cell response in experimental Chagas disease . BAFF is produced early and persists throughout the infection . To analyze BAFF role in experimental Chagas disease , Balb/c infected mice were injected with BR3:Fc , a soluble receptor of BAFF , to block BAFF activity . By BAFF blockade we observed that this cytokine mediates the mature B cell response and the production of non-parasite specific IgM and IgG . BAFF also influences the development of antinuclear IgG and parasite-specific IgM response , not affecting T . cruzi-specific IgG and parasitemia . Interestingly , BAFF inhibition favors the parasitism in heart . Our results demonstrate , for the first time , an active role for BAFF in shaping the mature B cell repertoire in a parasite infection .
Chagas disease is a chronic disease caused by infection with Trypanosoma cruzi . Initially , the disease goes through an acute episode which is characterized by circulating parasites and immunological disturbances that , at the level of B cell compartment , are mature and immature B cell apoptosis [1] , [2] as well as massive B cell response [3] . This polyclonal response [4]–[6] , which results in hipergammaglobulinemia [7] and delayed parasite specific humoral response [8] , [9] , may be the original cause for the autoimmune phenomena that have been described in chronic phases of infection [10] , [11] . It is well known that polyclonal activation of B lymphocytes is invariably accompanied by autoantibody production [12] , [13] . The polyclonal B cell activation that occurs during infection , induced by the host response ( T cell response , cytokines ) or parasite antigens , may disrupt normal immune regulatory mechanisms and cause autoimmunity [14] It has been reported that polyclonal B cell response in T . cruzi infected mice is predominantly helper T-cell dependent [15] . However , Ig-secreting plaque-forming cells are recorded in athymic ( nude ) mice after T . cruzi infection [5] suggesting that T-independent mechanisms can also mediate polyclonal B cell response . Several parasite-encoded proteins have been identified as B cell mitogens [13] , [16]–[18] and some of these T . cruzi antigens trigger in vitro polyclonal B cell activation and differentiation in a T-independent way [16] , [17] . We have reported that macrophages from normal mice cultured with T . cruzi glutamate dehydrogenase , a T-independent type II polyclonal B cell activator , secrete high level of BAFF that mediates B cell polyclonal activation [17] , suggesting that BAFF may mediate the polyclonal B cell response in vivo during T . cruzi infection . BAFF is a crucial factor for the survival of peripheral B cells [19]–[21] . But , in excess , BAFF leads to the development of autoimmune disorders in animal models . It has been described that BAFF transgenic mice show clear signs of B cell hyperplasia and hyperglobulinemia . These mice have enlarged spleen , Peyer's patches and lymph nodes , circulating immune complexes , rheumatoid factors , and anti-DNA Abs [22] . In addition , high levels of BAFF have been detected in the serum of patients with various autoimmune disorders [23] , [24] . Based on the fact that BAFF transgenic and T . cruzi infected mice share many immunological features like polyclonal activation , autoantibody production and autoimmunity , we hypothesized that BAFF can participate in the polyclonal B cell response observed in experimental Chagas disease . In the present study , we quantified the levels of BAFF and analyzed the participation of BAFF on B cell response by blocking its activity with a soluble BAFF-receptor in T . cruzi infected mice .
BALB/c mice were originally obtained from School of Veterinary , La Plata National University ( La Plata , Argentina ) and housed in our animal facility where all experiments were performed in compliance with the Institutional Review Board and Ethical Committee of the School of Chemical Sciences , National University of Cordoba . BALB/c mice 6–8 wk old were intraperitoneally ( i . p . ) infected with 500 trypomastigotes from T . cruzi ( Tulahuén strain ) diluted in physiological solution , as previously described [2] , [25] . Non-infected normal littermates were injected i . p . with physiological solution and processed in parallel . For BAFF activity blocking , one day after infection , mice were injected i . p . with 150 ug of BR3:Fc ( Genentech Inc . , South San Francisco , CA , USA ) three times per week . As control , infected mice were injected with 150 ug of IgG2a or physiological solution . Non-infected normal littermates were injected i . p . with physiological solution and injected i . p . with 150 ug of BR3:Fc or 150 ug of IgG2a or physiological solution with the same schedule described above and processed in parallel . At 15 days after infection , mice ( number indicated in each figure ) were killed by cervical dislocation , blood was collected and lymphoid organs were removed . BR3:Fc efficacy of BAFF neutralization was tested in vivo evaluating the reduction of splenic B cell subsets according to Lin et al [26] . Also , BR3:Fc neutralizing BAFF activity was evaluated in an in vitro assay measuring IgA concentration in the supernatant of peritoneal B cells cultured with CpG plus recombinant BAFF [27] , [28] in presence or in absence of BR3:Fc ( data not shown ) . Blood was collected by retro-orbital bleeding , erythrocytes were lysed in a 0 . 87% ammonium chloride buffer , and viable trypomastigotes counted in a Neubauer counting chamber [2] . Spleen and inguinal lymph nodes were obtained and homogenized through a tissue strainer . Peritoneal cells were obtained by peritoneal washouts and bone marrow cells were isolated by flushing femurs and tibias of mice with RPMI 1640 . When it was necessary , red blood cells were lysed for 5 min in Tris-ammonium chloride buffer . Viable mononuclear cell numbers were determined by trypan blue exclusion using a Neubauer counting chamber . Cell suspensions were processed for Flow cytometry studies or culture as indicated below . To obtain B cells , T cells , dendritic cells and F 4/80+ macrophages , splenic cells from infected mice were stained with anti-B220 APC , anti-CD3 FITC , anti-CD11c PE , anti-F4/80 Biotin followed by Streptavidin Per-CP purchased from BD , and sorted by positive selection with FACSAria Cell Sorter ( Becton Dickinson ) to enrich populations to 98% for B and T cells and 88% for CD11c+ and F 4/80+ . Cells were incubated with TRIzol reagents ( Life Technologies ) and RNA was extracted according to the manufacturer's recommendation and stored at −70°C . RNA was reverse transcribed using Moloney murine leukemia virus reverse transcriptase ( Invitrogen , USA ) at 42°C for 60 min . One microgram of RNA was used to generate first cDNA strain . Real Time PCR reactions for mouse BAFF and β-Actin detection were performed using the following primers pairs: BAFF ( Mm00446347_m1 , Applied Biosystems ) and HPRT ( HPRT-F: 5′-AAGCTTGCTGGTGAAAAGGA-3′; and HPRT R: 5′-TCCAACAAAGTCTGGCCTGT-3′ ) . The reaction mixtures contained: TaqMan Universal PCR Master Mix in the case of BAFF or 2X SYBR Green PCR Master Mix , 800 nM of HPRT primers , and 50 ng of cDNA . All reactions were performed in triplicate and were cycled as follows: 95°C for 10 min , 1 cycle; 95°C 15 s , 60°C 1 min , 40 cycles; 95°C 15 s , 1 cycle , 60°C 1 min , 1 cycle , followed by a melting curve rising from 60 to 95°C incrementally using in a 7500 System apparatus ( Applied Biosystems ) . HPRT was used as standard to normalize cDNA loading [29] . Cell suspensions were washed twice in ice-cold FACS buffer ( Physiological solution with 2% fetal bovine serum ( FBS , Gibco ) ) and preincubated with anti-mouse CD32/CD16 mAb ( Fc block ) for 30 min at 4°C . The cells were then incubated with each PE- , FITC- , or biotinylated Ab ( e-Bioscience , San Diego , USA ) for 30 min at 4°C and washed with FACS buffer: PE-anti B220 , FICT-anti IgM and biotin-anti IgD to identify mature B cells; PE-anti CD138 and FICT-anti B220 to identify plasma cells; PE-anti B220 , FICT-anti CD24 and biotin anti IgD to identify mature B cells in bone marrow , and biotin anti B220 to identify peritoneal B cells . Data were acquired on a FACSCanto II cytometer ( Becton Dickinson ) and analyzed using Flow Jo ( Tree Star ) software . BAFF concentration ( ng/ml ) was determined in sera and culture supernatant of mononuclear cells from lymphoid organs from normal or T . cruzi infected mice by ELISA following manufacturer's instructions ( Axxora , USA ) . Mononuclear cells from lymphoid organs from normal or T . cruzi infected mice treated with physiological solution or BR3:Fc or IgG2a control obtained at day 15 post infection ( p . i . ) were cultured with media for 30 h . IgM and IgG concentrations ( ng/ml ) were determined by ELISA as previously described [17] . In brief , plates were coated with 2 . 5 ug/ml of the type-specific goat anti-mouse Ab ( IgM , IgG; Sigma-Aldrich Chemical Co ) overnight at 4°C , and blocked with 1% Bovine serum albumin ( BSA ) . Culture supernatants were incubated overnight at 4°C . Peroxidase-conjugated anti-mouse IgG or anti-mouse IgM ( 2 . 5 µg/ml ) were added and incubated for 1 h at 37°C . The reaction was developed with TMB Substrate Reagent ( BD OptEIA™ ) . The concentration was measured with reference to standard curves using known amounts of the respective murine Ig ( Sigma-Aldrich Chemical Co . ) . The titers of T . cruzi specific seric IgM and IgG were determined by ELISA [30] , [31] using T . cruzi trypomastigotes recombinant Ags following manufacturer instruction ( Wiener lab , Argentina ) . ELISA test sera were considered positive if the mean OD value was two standard deviations above the mean value for control sera assayed in parallel . Amount of antinuclear specific IgG Ab ( ng/ml ) and IgG1 , IgG2a and IgG3 ( OD ) was determined by ELISA using a ANA ELISA kit ( Alpha Diagnostic , San Antonio , TX , USA ) following the manufacturer's instructions in sera from T . cruzi infected mice treated with physiological solution or BR3:Fc or IgG2a control . Hearts were fixed in formaldehyde and embedded in paraffin blocks , after which 5 to 20-µm-thick transverse sections were mounted on slides and subsequently stained with hematoxylin and eosin . Photographs were taken using a Nikon Eclipse TE 2000 U equipped with a digital video camera . The grade of heart parasitism was evaluated analyzing the number of nests containing amastigotes per section . Statistical significance of comparisons of mean values was assessed by a two-tailed Student's t test or nonparametric Mann-Whitney U test using Graph pad software . In the case of T . cruzi specific Abs , mean values was assessed by a one way non parametric Kruskal-Wallis test . p≤0 . 05 was considered significant .
To investigate systemic BAFF concentration during T . cruzi infection , serum samples were collected at different times p . i . and analyzed for levels of circulating BAFF by ELISA . During an ongoing T . cruzi infection , a significant increase in the seric levels of BAFF was detected at 11 days p . i . BAFF concentration presented the highest value at the peak of parasitemia ( day 11–15 p . i , [2] ) and persisted elevated until the last day analyzed ( day 32 p . i ) ( Fig 1A ) . Additionally , cells from the spleen and bone marrow , but not from the lymph nodes and peritoneal cavity , obtained from 15-day infected mice and cultured in the absence of stimuli , released higher concentrations of BAFF than cells from non-infected mice ( Fig 1B ) . To identify which mononuclear cells produced BAFF in T . cruzi infected mice , CD3+cells , CD11c+cells , F4/80+ cells and B220+ cells were obtained by cell sorting from the spleen of infected mice . By real-time PCR we determined that F4/80+ cells and CD11c+ cells showed high levels of mRNA coding for BAFF . The transcript for BAFF was present at low levels in T cells and undetectable in B cells . mRNA from total spleen cells from T . cruzi infected mice was used as positive control ( Fig 1C ) . In order to analyze the potential role of BAFF in the massive B cell response in experimental Chagas disease , BAFF activity was blocked by injecting BR3:Fc , a soluble BAFF receptor , into T . cruzi infected mice . Considering that the different B cell compartments are dissimilarly affected by T . cruzi infection [2] , [32] , [33] , we evaluated how the BAFF blockade affected each B cell compartment individually . We observed that treatment with BR3:Fc resulted in a significant reduction of splenic and lymph nodes B220+ cell number in non-infected and in T . cruzi infected mice ( Fig 2 ) . We determined that the reduction affected mature B220+IgD+IgM+ spleen cells but did not induce significant changes in the number of B220+IgD−IgM+ spleen cells ( phenotype compatible with immature , marginal zone and activated B cell ) ( Fig S1 A ) . BAFF inhibition reduced the already low number of B cells present in the bone marrow of T . cruzi infected but did not affect the number of B cells in the bone marrow of normal mice ( Fig 2 ) . This result could be explained by the fact that BAFF blockade affected mature B220+ IgD+CD24+ but not immature B220+ IgD− CD24hi B cells from bone marrow ( Fig S1 B ) . Immature B cells are the major B cell population in the bone marrow of normal mice but are almost absent in T . cruzi infected mice that present only mature B cells [2] . In addition , we determined that BAFF blockade did not change the number of peritoneal B220+ cells observed in T . cruzi infected mice ( Fig 2 ) that were reduced as a consequence of infection [34] . Mice injected with control isotype non-blocking Ab presented no significant difference in B cell number in comparison to non-treated non infected ( NI ) or infected ( I ) mice ( Fig 2 ) . Next , we analyzed if B cell reduction observed in lymphoid organs as a consequence of BAFF blockade also affected plasma cell number . B220+CD138+ cell number declined markedly in the spleen but remained unchanged in lymph nodes of BR3:Fc treated-T . cruzi infected mice ( Fig 3 ) . T . cruzi infection per se induced a severe reduction of plasma cell number in bone marrow ( Fig 3 ) and no further reduction was observed after BAFF blockade . The reduction of plasma cell number in T . cruzi infected mice treated with BR3:Fc correlated with a significant reduction in IgM and IgG concentration in culture supernatant of splenic cells ( Fig 4 ) . Interestingly , while the concentration of IgG was maintained , a reduction of IgM concentration was detected in the culture supernatant of cells from peritoneum of T . cruzi infected mice with blocked BAFF activity . According to plasma cell number results , no differences were detected in IgM and IgG concentrations produced by cells from lymph nodes and bone marrow from treated and untreated infected mice ( Fig 4 ) . To analyze the biological significance of the reduced Ab response observed in BR3:Fc treated T . cruzi infected mice , we evaluated the levels of parasite specific Abs in sera . We observed that preventing BAFF binding drastically diminished trypomastigotes T . cruzi-specific IgM titers while T . cruzi-specific IgG titers did not change ( Fig 5 ) . Parasite-specific Abs were practically undetectable in the culture supernatants of the lymphoid organs obtained from infected mice indicating a low frequency of T . cruzi antigen-specific B cells as reported [8] , [9] and a high frequency of non-parasite specific Abs ( data not shown ) . To address the role of BAFF in the autoreactive humoral response , we evaluated autoreactive Abs in the sera of treated and control infected mice at 15 days p . i . . It has been described that autoreactive Abs such as anti-actin , anti-myosin , anti-myoglobin and antinuclear Abs ( ANA ) among others are present in the acute and chronic phase of the pathology [35]–[37] . However , we were unable to detect autoreactive Abs others than ANA at the acute phase of the infection ( day 15 p . i . ) . Therefore , ANA were tested as markers of the effect of BR3:Fc treatment on the autoreactive B cell population during acute phase of infection . We observed that BAFF blockade prevented the production of ANA of IgG isotype in T . cruzi infected mice . The isotype of IgG involved in this reaction was IgG3 but not IgG2a or IgG1 ( Fig 6 ) . To analyze if B cell reduction and the decrease of IgM parasite-specific Abs affected the parasite replication , we measured the number of circulating trypomastigotes in blood of infected mice and the grade of tissue parasitism by evaluating amastigote niches in heart . Parasitemia was similar in T . cruzi infected mice treated with physiological solution , BR3:Fc or IgG2a control ( Fig 7A ) , while cardiac parasitism was increased in BR3:Fc treated infected mice in comparison to untreated infected mice ( Fig 7B , C ) . Thus , the hearts of infected mice in which BAFF activity was blocked had higher number of nest of amastigotes in the myocardial fibers of the auricle than non-treated infected mice ( Fig 7B , C ) .
Most of the information about BAFF in disease is linked to autoimmune pathologies [38] , but little information is available about BAFF in infectious diseases [39] , [40] . Previously , we reported that BAFF is involved in the polyclonal B cell activation triggered in vitro by a T . cruzi antigen [17] . Here , we extended these data showing that BAFF increased early in T . cruzi infected mice and persisted at high levels throughout the infection . Our data show that cells from immune system as macrophages and dendritic cells which are susceptible to be infected by T . cruzi and in which the parasite can replicate intracellularly [41] , [42] are an important source of BAFF . It has been reported that BAFF is mainly produced by innate immune cells such as neutrophils , macrophages , monocytes , dendritic cells ( DCs ) and follicular DCs [43] stimulated by cytokines often produced during inflammation and infections [43] , [44] , as well as by the Toll-like receptors ligands [45] . Then , it is possible that IL-10 and/or IFNγ , which increase during infection [46]–[48] , as well as TLR ligands expressed by T . cruzi [49] , [50] or parasite antigens [17] trigger BAFF secretion . To analyze the role of BAFF in T . cruzi infection , infected mice were treated with a soluble BAFF-R , BR3:Fc , to block BAFF activity . The use of BAFF-R Fc considerably reduced mature peripheral B cell numbers in T . cruzi infected mice . As it was previously reported for normal mice [21] , [51] , [52] , we observed that mature B cells from T . cruzi infected mice have different BAFF requirements than peritoneal mature B cells and/or immature B cells for their survival . Reduction in mature B cells was also observed in the bone marrow of infected mice , probably as a consequence of the diminution of peripheral mature B cells . Following B cell depletion , BAFF blockade resulted in a reduction of splenic B220+CD138+ plasma cells . When we analyzed the proportion of plasma cells with respect to the number of total B cells we observed similar values in BR3:Fc treated or untreated infected mice ( data not shown ) . These results suggest that during T . cruzi infection , BAFF apparently controls plasma cell numbers by reducing the pool of mature B cells . Concomitantly with plasma cell diminution , a reduction of total IgM and IgG production by splenic cells was observed in BR3:Fc treated infected mice . Since most of the Abs produced by splenic B cells are non-parasite specific [8] , our results suggest that BAFF is regulating polyclonally activated B cells rather than antigen-specific activated B cells . In agreement , parasite-specific IgG titers were not affected by BAFF blockade while parasite-specific IgM almost disappeared in BR3:Fc treated infected mice . Similar behaviour was observed in three different independent experiments analyzed at 15 and 22 days p . i . Our results confirm and complete previously reported findings on the role of BAFF-BAFF-R signalling in the survival and maintenance of the mature B cell compartments [reviewed in 53] , and that BAFF inhibition had a markedly small effect on IgG+ B cells and long-lived plasma cells . Scholz et al [54] reported that IgM-bearing memory cells are sensitive to BAFF depletion whereas IgG-bearing memory cells are not . Chagas disease pathology is associated to autoimmunity [55]–[57] . There are several mechanisms to explain autoimmunity induced by infectious agents [58] . All are based on the observation that an immunocompetent host possesses circulating autoreactive T and B cells that are normally tolerant to self antigens [59] . As a consequence of the favorable proinflammatory environment induced by microorganisms , an unspecific activation may occur [60] . Excess BAFF may lower the threshold for BCR signaling and maintain survival when a normal autoreactive B cell would undergo death [61] . In agreement with reports showing evidences of BAFF participation in autoreactive response [62]–[67] , BAFF blockade decreases the ANA IgG in T . cruzi infected mice . According to our results , it has been previously reported that a 4-week course of BAFF-R–Ig prevents the emergence of IgG anti-DNA antibodies in NZB/W mice [68] . The presence of anti-self antibodies was reported at the acute and chronic phase of T . cruzi infection [35] , [37] . However , in our infection model , probably as consequence of different experimental conditions and the parasite and mouse strain used , we were unable to detect , by ELISA , anti-myosin , anti- myoglobin and anti-skeletal muscle antibodies in sera of acutely-infected mice ( day 15 p . i . ) . Consequently , we ignore whether BAFF blockade affects the production of other autoreactive Abs different from ANA . Our data analyzed together indicated that autoreactive IgG3+ B cells activated during the infection and the parasite specific IgG+ B cells induced by T . cruzi show a differential requirement of BAFF to proliferate and/or differentiate and/or survive . This difference is probably related to the nature of B cell response: T-independent , extrafolicular or from germinal center [53] . Probably IgG3 ANA are produced during a pre-germinal center response [69]–[71] while anti-T . cruzi IgG Abs are produced in the course of a germinal center response ( Bermejo et al . unpublished observation ) . Importantly , this differential requirement of BAFF may become an important target of manipulation to control a possible pathological autoreactive response without dampening the protective parasite specific response . Interestingly , blocking BAFF is apparently not affecting the Abs involved in the control of circulating trypomastigotes . It seems that the conserved levels of IgG are sufficient to control parasite replication , or that , other populations different from B cells and not affected by BAFF inhibition , may be controlling parasite spreading . However , BAFF blockade does impact in T . cruzi replication in the heart , favoring the appearance of T . cruzi pseudocyts . The increase in the parasite replication in target tissues observed in BR3:Fc treated mice could be consequence of the markedly reduced mature B cell numbers that affect not only the production of Abs but also the development of protective cellular responses [72] . Strikingly , in spite of the high number of T . cruzi pseudocyts observed in the hearts of infected mice , preliminary data showed that BR3:Fc treated mice survive longer than non-treated infected mice ( data not shown ) . Our findings enlighten a new role of BAFF-BAFF-R signalling in a parasite infection where it controls mature B cell numbers , polyclonal B cell activation and self-reactive response but does not affect protective anti-parasite IgG response . | Chagas disease , caused by the protozoan Trypanosoma cruzi , is endemic in Central and South America . It affects 20 million people and about 100 million people are at risk of infection in endemic areas . Some cases have been identified in non-endemic countries as a consequence of blood transfusion and organ transplantation . Chagas disease presents three stages of infection . The acute phase appears one to two weeks after infection and includes fever , swelling around the bite site , enlarged lymph glands and spleen , and fatigue . This stage is characterized by circulating parasites and many immunological disturbances including a massive B cell response . In general , the acute episode self-resolves in about 2 months and is followed by a clinically silent indeterminate phase characterized by absence of circulating parasites . In about one-third of the cases , the indeterminate phase evolves into a chronic phase with clinically defined cardiac or digestive disturbances . Current knowledge suggests that the persistence of parasites coupled with an unbalanced immune response sustain inflammatory response in the chronic stage . We believe that an effective treatment for chronic Chagas disease should combine antiparasitic drugs with immunomodulators aimed at reducing inflammation and autoreactive response . Our findings enlighten a new role of BAFF-BAFF-R signaling in parasite infection that partially controls polyclonal B cell response but not parasitespecific class-switched primary effectors B cells . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"immunology/immunity",
"to",
"infections"
] | 2010 | BAFF Mediates Splenic B Cell Response and Antibody Production in Experimental Chagas Disease |
Many viruses , including adenovirus , exhibit bidirectional transport along microtubules following cell entry . Cytoplasmic dynein is responsible for microtubule minus end transport of adenovirus capsids after endosomal escape . However , the identity and roles of the opposing plus end-directed motor ( s ) remain unknown . We performed an RNAi screen of 38 kinesins , which implicated Kif5B ( kinesin-1 family ) and additional minor kinesins in adenovirus 5 ( Ad5 ) capsid translocation . Kif5B RNAi markedly increased centrosome accumulation of incoming Ad5 capsids in human A549 pulmonary epithelial cells within the first 30 min post infection , an effect dramatically enhanced by blocking Ad5 nuclear pore targeting using leptomycin B . The Kif5B RNAi phenotype was rescued by expression of RNAi-resistant Kif5A , B , or C , and Kif4A . Kif5B RNAi also inhibited a novel form of microtubule-based “assisted-diffusion” behavior which was apparent between 30 and 60 min p . i . We found the major capsid protein penton base ( PB ) to recruit kinesin-1 , distinct from the hexon role we previously identified for cytoplasmic dynein binding . We propose that adenovirus uses independently recruited kinesin and dynein for directed transport and for a more random microtubule-based assisted diffusion behavior to fully explore the cytoplasm before docking at the nucleus , a mechanism of potential importance for physiological cargoes as well .
Viruses generally depend on active transport inside host cells as the crowded cytoplasm restricts their ability to diffuse [1 , 2] . Viruses have evolved mechanisms to hijack microtubule motor proteins for this purpose during cell entry and egress [3] . Adenovirus [4 , 5] , herpesvirus [6] , vaccinia [7] , adeno-associated virus [8] and HIV-1 [9 , 10] each exhibit bidirectional movements along microtubules ( MTs ) , consistent with possible use of both minus- and plus- directed microtubule motors . Cytoplasmic dynein , in particular , has been implicated in microtubule minus end-directed transport for several viruses , but less is known about the contributions of kinesins [11] . The human adenoviruses are non-enveloped , dsDNA-containing particles , consisting of more than 57 serotypes grouped into seven species [12] . Adenovirus infections are usually self-limiting , but can have fatal outcomes in immunocompromised patients . However , engineered versions are preferred vehicles for vaccine delivery and therapeutic gene transfer [13] . Adenovirus enters target cells by receptor-mediated endocytosis [14] . Following endosomal escape , Ad particles travel along microtubules ( MT ) , and then dock at nuclear pore complexes ( NPC ) to deliver their DNA genome into the nucleus [4 , 15 , 16] . In enucleated cells or those treated with the nuclear export inhibitor leptomycin B , capsids bypass the nucleus and accumulate in the vicinity of the centrosome [17 , 18] . Transport to the cell center involves the MT minus end-directed motor protein cytoplasmic dynein , and can be inhibited by microtubule-destabilizing agents or blocking dynein or dynactin function using dominant negative cDNAs , RNAi , or acutely injected function-blocking antibodies and inhibitory fragments directed at dynein subunits [4 , 5 , 15 , 19 , 20] . We have found that cytoplasmic dynein is directly recruited to adenovirus by its major capsid subunit hexon , via the dynein intermediate and light intermediate chains ( ICs; LICs ) and without the need for adaptor proteins used by physiological forms of dynein cargo [5] . We found in addition that exposure of hexon to reduced pH , as would happen during passage through early endosomes , triggers a reversible change in the hexon hypervariable region 1 ( HVR1 ) , to activate dynein recruitment [21] . This function is also markedly stimulated by PKA , which phosphorylates LIC1 and enhances dynein binding to the virus , while releasing the motor protein from endogenous cargo sites associated with lysosomes [19] . The role of kinesins in adenovirus transport is less well understood . A form of kinesin-1 , Kif5C , has been reported to function in disruption of Ad5 capsids and associated nuclear pore complexes to facilitate genome transfer into the nucleus [22] . However , kinesins responsible for adenovirus plus-end directed transport along MTs were not evaluated and remain to be identified . In the current study , we performed an RNAi screen of all 38 plus-end directed transport kinesins in the human genome . We identified Kif5B as the major motor responsible for plus end-directed adenovirus transport , along with several minor kinesins which may contribute to this role . Kif5B interacted with the capsid subunit penton base ( PB ) , independent of hexon or low pH exposure . Kif5B RNAi caused a marked increase in microtubule minus end-directed transport toward the cell center , and a resulting accumulation of Ad5 particles in the pericentrosomal region . Following this period in normal or kinesin-knockdown cells , adenovirus transport along microtubules was more randomly directed , though the magnitude of this cytoplasmic exploration behavior was also reduced by Kif5B knockdown . These results suggest distinct evolution of dynein and kinesin recruitment to adenovirus and a coordinated requirement for the two motors during virus entry .
As a first step toward identifying microtubule plus end-directed kinesins for Ad5 transport , we quantified virus redistribution to the nucleus and the pericentrosomal region in human lung epithelial A549 cells at 0 , 30 , 60 and 90 min post infection ( p . i . ) . Ad5 particles could be seen to accumulate in the pericentrosomal region of a small subset of cells by 30min p . i . ( Fig 1A ) , well before maximal targeting to the nuclear envelope , which usually occurs by 60 min [5] . We reasoned that depletion of an Ad5 transport kinesin might , in principle , lead to an increase in dynein-driven Ad5 accumulation in the pericentrosomal region . We tested this possibility with an siRNA library targeting the heavy chains of each of the known 38 plus end-directed human kinesins , covering all 14 families . Three days following siRNA treatment , we infected the cells with Ad5 and determined the percentage of cells showing pericentrosomal Ad5 capsid accumulation at 30 min p . i . ( Fig 1B and 1C ) . Reduction of the level of Kif5B , a well-known member of the kinesin-1 subfamily , had the most pronounced effect on virus redistribution , with a 5-fold increase in cells exhibiting clear pericentrosomal virus accumulation ( Fig 1C ) . The level of Kif5B was reduced by ~70% ( Fig 1D ) , as tested with an isoform specific antibody ( S1A Fig ) . Incomplete reduction , however , may still be responsible for some of the residual pericentrosomal targeting we observed ( Fig 1E ) . RNAi directed against the two other kinesin-1 heavy chain isoforms , Kif5A and Kif5C , also showed significant , though weaker , effects on virus redistribution to this region of the cell ( Fig 1C ) . In addition , we observed increased pericentrosomal accumulation with knockdown of kinesin-4 subfamily heavy chain genes , Kif4A and Kif21A . Although the nucleus is the physiological destination for in-coming capsids [23] , we reasoned that the attachment of the virus to the nucleus terminates motility and complicates analysis of underlying transport mechanisms . We , therefore , treated cells with the nuclear export inhibitor leptomycin B ( LMB ) , which blocks adenovirus binding to nucleoporins [18] . Upon LMB treatment , the percentage of cells exhibiting nuclear accumulation markedly decreased ( Fig 2A ) . However , the number of cells exhibiting pericentrosomal virus accumulation at 30 min p . i . increased modestly ( from 6 . 1% to 9 . 1% ) ( Figs 1A and 2B ) . Importantly , the combination of LMB treatment with Kif5B knockdown led to a ~10-fold increased fraction of cells showing pericentrosomal virus accumulation ( 9 . 1% to 90 . 1% ) ( Fig 2B and 2C ) . We also quantified the concentration of individual virus particles reaching the MT organizing center ( MTOC ) by measuring the capsid density in the pericentrosomal region ( see Materials and Methods ) . Although there was considerable variation between cells , this analysis again revealed strong enhancement of Ad5 redistribution to the pericentrosomal region upon Kif5B RNAi in LMB treated cells ( Fig 2D ) . These results imply that Kif5B contributes substantially to Ad5 transport , normally serving in virus transport away from the cell center . Physiological cargo recognition by kinesin-1 is often mediated through its light chains ( KLCs ) . Indeed , KLCs were previously reported to interact with Ad protein IX ( pIX ) , as part of the mechanism mediating nuclear pore binding and disruption [22] . We noted that Kif5B knock down also reduced KLC1 and KLC2 protein levels ( Fig 1D ) , suggesting the HC-LC interaction stabilizes the LCs . However , to test more directly for an LC contribution to Ad5 behavior , we also performed RNAi for the individual KLCs . Knockdown of KLC1 also depleted KLC2 , despite the sequence specificity of the siRNAs used , whereas KLC2 siRNA was specific for this isoform alone ( Fig 1D ) . Knockdown in neither case had any detectable effect on pericentrosomal virus accumulation in the presence of LMB ( Figs 2E and S1B ) . Thus , our results support a direct role for the heavy chain in kinesin-1 recruitment to the virus capsid and subsequent transport , whereas KLC1 and KLC2 are dispensable for these functions . We also tested the combined effect of LMB treatment with RNAi directed at additional kinesin heavy chain candidates . RNAi against Kif5A , Kif5C and Kif4A each increased the number of infected cells showing pericentrosomal Ad5 accumulation by 2-3-fold in the presence of LMB compared to a scrambled RNAi control ( Figs 3A and S1C ) . We note that Kif5A and C are expressed at much lower levels than Kif5B in A549 cells ( Source: Protein Abundance Database , http://pax-db . org/ ) , suggesting that all three isoforms might contribute to Ad5 transport , but , perhaps , in proportion to their abundance . To test the functional relationship between the isoforms and as a control for potential off-target effects of Kif5B RNAi , we determined the ability of each of the Kif5 isoforms to compensate for Kif5B knockdown . We observed clear rescue in each case , suggesting that the three kinesin-1 isoforms contribute to a common function , though with the most abundant form , Kif5B , playing by far the most substantial role in Ad5 transport ( Figs 3B , 3C and S1D ) . Interestingly , Kif4A overexpression partially rescued the effects of Kif5B knockdown on virus redistribution ( Fig 3B and 3C ) . Kif21A rescue was not tested in light of its relatively minor contribution . To test the role of Kif5B in adenovirus behavior more directly , we performed continuous imaging of Alexa-546-labeled Ad5 in live , LMB-treated A549 cells from 15 to 60 min p . i . at 30 second intervals . Net particle movement towards the pericentrosomal region was observed in 13 out of 23 Kif5B siRNA-treated cells , compared to 2 out of 24 scrambled siRNA- treated cells ( Fig 4A , S1 and S2 Movies ) . The net inward movement was most prominent from 15–30 min p . i . ( Fig 4A and S2 Movie ) , supporting a normal role for Kif5B in MT plus end-directed Ad5 transport during the early stages of infection . We then examined individual virus capsid motility at higher temporal resolution ( 23 frames/second , fps ) over 2 min intervals from 15–30 min p . i . , a period when capsids undergo the greatest redistribution toward the centrosome in LMB-treated cells . Tracings for stationary particles or those exhibiting small-scale Brownian motion are filtered out as part of the particle-tracking analysis [5] . In Kif5B knockdown cells , we observed a pronounced increase in inward minus-end directed run length ( Fig 4B ) . Small , though statistically significant , increases in outward , MT plus-end directed run length and in the relative frequency of plus- vs . minus-end runs and pauses could also , however , be detected ( Fig 4C ) . Virus transport velocity was unaffected ( Fig 4D ) . The persistence of outward runs in Kif5B knockdown cells ( Fig 4B , 4C and 4D ) despite the global effects of capsid translocation towards the centrosomal region , is consistent with incomplete Kif5b knockdown ( Fig 1D ) . A minimal change in run length has previously been observed from kinesin inhibition in Drosophila embryonic lipid droplet motility ( Shubeita et al . , 2008 ) , and may reflect compensatory effects on dynein behavior and the small number of motor protein molecules involved in transport of individual cargo ( see Discussion ) . Using a fluorescence focus assay [24] , we also observed an ~20% inhibitory effect of Kif5b knockdown on Ad5 infectivity ( Fig 4E ) . This , too , is a relatively small effect , but consistent with the limited impact of microtubule motor-driven transport on measures of whole cell physiological changes ( see below ) . We also monitored the effects of Kif5B RNAi on virus behavior later in infection ( 30–60 min ) , when directed movements are still seen , but the distribution of virus in the LMB-treated cells has reached a steady state , with no clear bias in transport toward or away from the cell center . To analyze this behavior we employed a different imaging regimen , examining Ad5 capsids at 0 . 5 fps for 10min between 30 and 60 min p . i . ( Fig 4F ) . At this time scale , we observed numerous capsids exhibiting non-directional “wandering” behavior within the cytoplasm ( Fig 4F ) . Local virus movements appeared to be linear , but random over-all with regard to direction and distance . This was confirmed by analysis of mean squared displacement ( MSD ) , which we found to increase linearly with time ( Fig 4G and 4H ) . This behavior is characteristic of random motility , such as Brownian movement , but the scale of over-all virus transport was greater than expected for simple diffusion . Knockdown of Kif5B resulted in a clear decrease in MSD values ( Fig 4I and 4J ) , indicating that capsids explore a smaller region of the cell under this condition . To test the dependence of the stochastic virus movements on MTs further we exposed cells to the MT depolymerizing drug nocodazole , which dramatically reduced virus wandering behavior . The apparent diffusion coefficient for the virus particles obtained from the slope of MSD plot was more than 4-fold lower than in control cells ( Fig 4I and 4J ) , and the maximum displacement ( distance between starting and ending positions ) was halved . Thus , together , our data support at least two active MT-based modes of for Ad5 capsid movement: directed transport and assisted diffusion . We previously found that cytoplasmic dynein is recruited to Ad5 by the major capsid subunit hexon , primed for binding by exposure to low pH in the endosome [5 , 21] . To define the mechanism for Kif5B recruitment we tested for a physical interaction with capsid subunits . We used an anti-hexon monoclonal antibody to immuno-purify intact Ad5 capsids or hexon alone [21] , with and without transient low pH exposure . We then tested for co-immunoprecipitation of kinesins from a GTP-release fraction of MT binding proteins isolated from rat brain cytosolic extracts [25] . Kinesin-1 could be readily seen to co-immunoprecipitate with Ad5 capsids , but not with hexon . In contrast to dynein [5 , 21] , the kinesin-1-virus interaction was unaffected by prior exposure of capsids to reduced pH ( Fig 5A ) . In reciprocal experiments , we also found Ad5 capsids to co-immunoprecipitate with kinesin-1 ( Fig 5B ) . The absence of a hexon-kinesin-1 interaction suggested the involvement of a different capsid protein . Potential candidates should , in principle , be exposed at the capsid surface during cytoplasmic transport , and would include the major capsid subunit penton base and the minor capsid protein IX , which has been reported to interact with KLC1 [22] . Notably , the third major capsid component , fiber , is lost during initial endocytosis of the virus [26] . Absolute levels of penton base co-immunoprecipitating with adenovirus virions have been found to decline with time p . i . However , relative levels appeared less affected . In fact , this subunit was still readily detected at > 1 hr p . i . by this approach [27] and by immunocytochemistry of infected cells [5 , 28 , 29] . Based on these considerations we first tested a penton base-containing subassembly for kinesin-1 binding . We isolated “penton dodecahedron” complexes ( Pt-Dd ) found in Ad-infected cells following lysis , which are composed of an ordered assembly of penton base and fiber subunits [30] ( Fig 5C ) and organized in a manner closely resembling that in the complete Ad capsid [31] . We clearly detected Pt-Dd in kinesin-1 , but not in cytoplasmic dynein , pull-downs ( Fig 5D ) . Given the unlikely role for fiber in kinesin-1 recruitment , these results suggest that Ad5 interacts with kinesin-1 through penton base . We also expressed individual capsid subunits in 293A cells and found that kinesin-1 co-immunoprecipitated with HA-tagged penton base , but not hexon or pIX ( Fig 5E ) . To test the ability of the three Kif5 isoforms to bind to penton base , we co-expressed myc-tagged penton base with fluorescent protein-tagged versions of Kif5A , B , or C in 293A cells . After anti-myc IPs of the cell lysates , we detected similar amounts of each of the three isoforms , but not GFP alone ( Fig 5F ) , in the pellet fraction , consistent with the ability of each of the three kinesin-1 isoforms to transport adenovirus ( Fig 3B and 3C ) . We also co-expressed myc-tagged penton base together with the kinesin light chains GFP-KLC1 and -KLC2 . We saw no detectable interaction between penton base the kinesin light chains ( Fig 5G ) , consistent with our KLC RNAi analysis ( Fig 2E ) . Each of the kinesin-1 heavy chain polypeptides consists of an N-terminal motor domain , followed by a neck-linker region , and highly elongated coiled-coil α-helical stalk and tail domains ( Fig 5H ) . The stalk consists of two coiled-coil regions separated by a predicted hinge domain . The tail mediates physiological cargo interactions either directly or through the kinesin light chains [32–34] . To define the penton base binding site within the kinesin-1 heavy chain , we generated GFP-tagged fragments of the Kif5B heavy chain ( Fig 5H ) . Only constructs spanning the hinge region showed a detectable interaction with expressed penton base ( Fig 5I ) . These results indicate that the interaction is mediated by the central portion of the Kif5B stalk , possibly including the hinge region . We note that the penton base binding region partially overlaps with that for the physiological kinesin recruitment factor JIP1and may be involved in kinesin-1 auto-inhibition [35 , 36] .
A number of viruses have been found to rely on kinesin-mediated transport , but most studies have focused on a potential role for this class of motor proteins in virus egress [3 , 37] . We have identified a role for Kif5B in Ad5 transport prior to capsid binding to the nucleus . For our analysis , we focused on the period between endosomal escape and nuclear envelope binding , when Ad5 transport is known to depend on motor proteins [16 , 38 , 39] . Using an RNAi screen of 38 MT plus end-directed kinesin heavy chains we have found Kif5B to be the predominant transport kinesin for Ad5 . Our evidence also suggests a role for additional kinesins , including the other kinesin-1 family members Kif5A and Kif5C . Importantly , expression of each of the kinesin-1 isoforms rescued the effects of Kif5B RNAi , suggesting that Kif5A and C are fully capable of adenovirus transport , though they are expressed at too low concentrations to make an important quantitative contribution to this behavior . The kinesin heavy chain Kif5C together with KLC1 and 2 has been implicated in another study in Ad2 capsid disassembly at the nuclear pore complex , involving a KLC1/2 interaction with the capsid subunit protein IX [22] . We find that knockdown of KLC1 and/or KLC2 had no detectable effect on the redistribution and transport of Ad5 capsids ( Fig 2E ) , and see no evidence for an involvement of protein IX ( Fig 5E ) . However , we find the major capsid protein penton base to mediate direct recruitment of the heavy chain of all three kinesin-1 isoforms ( Fig 5F–5I ) . Based on findings by us and others [5 , 27–29] , penton base remains associated with Ad capsids throughout the cytoplasmic phase of capsid entry , and therefore , represents a plausible kinesin binding partner . These data suggest that we have identified a novel kinesin role in capsid transport distinct from that for nuclear entry . Two kinesin-4 family members , Kif4A and Kif21A , were also implicated by our analysis in plus-end directed adenovirus transport . Both Kif4A and Kif21A have previously been reported to regulate microtubule dynamics and stabilization [40 , 41] . Interestingly , Kif4A has been suggested to contribute to transport of HIV Gap proteins [42] , and Kif4-mediated microtubule stabilization enhanced early infection of HIV-1 [43] . Whether these kinesins affect in-coming Ad5 redistribution ( Fig 3A ) through changes in MT stability or direct effects on virus transport is uncertain . Of note , Ad replicates and assembles within the nucleus and induces lytic cell death to release and spread its progeny virions [44] . Thus , there is no apparent late function for kinesins in this system . The role we identify for Kif5B and the other kinesins found in our screen in early infection seems to be the predominant one in this system . Previously we reported that low pH-primed Ad5 directly recruits the cytoplasmic dynein IC and LIC1 subunits through the hexon hypervariable region HVR1 [19 , 21] . Here , we identify a distinct kinesin-1 recruitment mechanism , which is independent of pH and hexon , but , instead , involves another major Ad5 capsid protein , penton base . We found this subunit to interact with the Kif5A , B , and C heavy chains , and mapped the penton base-binding site to the kinesin stalk . Although this region is not well conserved among the Kif5 isoforms , we suspect there must be sufficient structural conservation to support the penton base interaction in each case . Several cellular proteins have been found to interact directly with the kinesin heavy chain ( see S1 Table ) , and not through the kinesin light chains [32] . Whereas most kinesin-1 heavy chain interactions occur through its tail domain , JIP1 can also bind within the stalk . Kinesin-1 can fold via its central hinge to allow the tail to bind to and inhibit the motor domain [35] . The region of the kinesin-1 stalk implicated by our analysis in penton base binding raises the possibility that Ad5 may have evolved to select for , or induce , the unfolded , activated kinesin-1 state . From extensive live imaging we have noted that only about 10% of Ad5 capsids are motile within any given 5~10 second time window , though almost all virus particles show directed motility within a 10 min window , suggesting that kinesin recruitment or activity may be intermittent and/or limited . This possibility is supported by the short plus end-directed run lengths for adenovirus we observe in vivo similar to published run-lengths for purified individual kinesin-1 molecules ( ~1 um ) . Thus , we suspect that most viruses are transported by no more than a single kinesin , in contrast to physiological cargoes . We were unable to detect colocalization of endogenous or expressed kinesin with individual adenovirus particles by immunocytochemistry , perhaps consistent with the few motile particles at a given instant , and hampered by the high levels of free kinesin-1 in the cytoplasm . We do not observe the juxta-nuclear accumulations of kinesin-positive capsid structures previously reported [22] . Under conditions of reduced Kif5B expression we observed a significant increase in inward virus run length , though , curiously , a modest increase in outward run length as well . The latter observation is consistent with a previous report on the behavior of Drosophila embryo lipid droplets , which displayed slight , but statistically significant increases in plus-end run length , though a more significant increase in minus-end run length after kinesin-1 inhibition [45] . These results may reflect roles for opposite-directed motor proteins in run termination and in “tug-of-war-mediated motor activation” [46] . We proposed two temporally distinct transport roles for Kif5B during entry . Up to ~30 min p . i . , Kif5B-mediated transport delays virus redistribution to the nucleus and retards infection . After ~30 min p . i . , Kif5B-dependent assisted diffusion enables more efficient cytoplasmic exploration by the virus and stimulates infection . The first role is supported by a dramatic centrifugal redistribution effect of Kif5B knock down ( Fig 2 ) by 30 min p . i . , using LMB to block nuclear pore binding . The second role is supported by the virus motility analysis between 30–60 min p . i . ( Fig 4F–4K ) . Adenovirus recruitment of cytoplasmic dynein would seem to provide the evolutionary advantage of aiding virus transport toward the nucleus , and this possibility is supported by a reduction in infectivity in dynein-inhibited cells [19] . We find that Kif5B knock-down also inhibits infectivity , though modestly . The basis for this outcome and its implications for the evolution of kinesin recruitment by Ad5 are uncertain . One hypothesis is that kinesin-mediated virus transport evolved as a means to defend the cell from viral attack , keeping virus particles from reaching the cell center . This seems an unlikely model , as viruses defective in kinesin binding would have an evolutionary advantage , and should quickly have out-competed a kinesin-binding strain . A second hypothesis is that kinesins are needed for full dynein activity . This possibility has been raised for peroxisome transport in Drosophila S2 cells [46] and the inward transport of HIV-1 [10] . However , we find that Kif5B knockdown actually increases centripetal Ad5 movement , arguing against this hypothesis . A third hypothesis is that kinesins might be needed to counteract the tendency of cytoplasmic dynein alone to accumulate virus particles in the pericentrosomal region , a behavior only slightly increased in LMB-treated cells ( Figs 1A and 2B ) . However , the effect of LMB treatment is much more pronounced upon additional kinesin knockdown , indicating that kinesins do normally contribute to a more dispersed virus distribution . Analysis of Ad5 capsid distribution during the early stages of control cell infection , however , revealed little tendency to congregate at the centrosome prior to reaching the nuclear envelope ( Fig 1A ) , arguing against a requirement for the virus to visit the pericentrosomal region before transport to the nucleus [47] . A final hypothesis raised in this study is that dynein/kinesin-mediated bidirectional Ad5 movement allows capsids to explore the cytoplasm more fully ( Fig 6B ) . Despite a basic role for the motor proteins in linear cargo transport along MTs , we find that they can contribute to less ordered behavior as well , the magnitude of which depends on microtubules as well as Kif5B . We propose , therefore , that a combined role for MT plus end- and minus end-directed MT motor proteins is to permit broad and relatively rapid random-walk exploration of the cytoplasm . We reason that centrosomes , where MT minus ends are concentrated , serve as a cul-de-sac for the virus , which the kinesins normally help to avoid . We observed “exploratory” movements throughout the cytoplasm , consistent with a broader role for a microtubule-based assisted diffusion mechanism in helping virus particles search the entire cytoplasmic space . According to this reasoning , kinesin recruitment should help maintain virus particles in an exploratory state until they interact irreversibly with nuclear pores , thereby facilitating infection . Kif5B knockdown did , in fact , reduce Ad5 infectivity , supporting a role for kinesin-1 in aiding the virus in the early stages of infection . Another form of exploratory behavior has been observed for vaccinia and baculovirus , which take advantage of the actin polymerization machinery to move within cells , either to reach the nucleus [48] or to spread to neighboring cells [49 , 50] . Our data suggest that adenovirus , and perhaps other pathogens may have evolved a distinct microtubule-based mechanism to explore the cytoplasm by recruiting both plus-end and minus-end motors . We also note that diverse forms of vesicular motor protein cargo exhibit relatively random-seeming MT motor-driven movements in nonneuronal cells . Conceivably , this represents another manifestation of cytoplasmic assisted diffusion to enhance the chance of cargoes , such as Golgi vesicles , to encounter sites of docking and fusion with other membranous structures . Kif5B depletion reduced adenovirus infectivity by only ~20% , as opposed to the effects of nocodazole ( >75% ) [51] or cytoplasmic dynein light intermediate chain knock down ( 45% ) [19] . We note in this context that the dual effect of kinesin inhibition on Ad motility may minimize the consequences for infectivity . The increased centripetal Ad transport during early infection of Kif5b knockdown cells should increase infectivity , whereas the decrease in assisted diffusion later in infection may inhibit infectivity , with kinesin inhibition suppressing virus exploration of the centrosomal region “in search of” the nucleus . We note that incomplete knockdown of Kif5B and/or the contributions of other kinesin isoforms may also affect the relative magnitude of the two modes of virus transport . Another complicating factor in this analysis is the radial subcellular organization of MTs in the A549 cells used in these studies , and the relatively short distances involved in virus transport to the nuclear surface , especially for viruses that enter the cell at nearby sites within the plasma membrane . In fact , we expect Kif5B and the other kinesins identified in this study to play a more critical role in differentiated cells ( Fig 6C ) . Polarized lung epithelial cells , similar to other epithelia , have a distinct arrangement of MTs and organelles , with centrosomes located apically and the nucleus basolaterally [52] . MTs tend to have a more columnar apical-basal arrangement [53] , as well as a disordered subapical array distinct from the more radial organization in the A549 cells used commonly for adenovirus analysis . We reason that in polarized epithelia cytoplasmic dynein and kinesin-1 may facilitate navigation of Ad along the apico-basal axis of the cell , and , more randomly , through the apical MT meshwork , to allow capsids to find the nucleus efficiently . Further work will be needed to test the role of cytoplasmic dynein and kinesins in adenovirus infection of differentiated cells .
Human A549 ( alveolar basal epithelial; ATCC , CCL-185 ) and 293A ( embryonic kidney; Thermo Fisher , R70507 ) cells were grown in DMEM supplemented with 10% FBS . Amplification , purification , and labeling of replication-deficient Ad5 were engineered for late GFP expression ( courtesy of Dr . Hamish Young , Columbia University , New York , NY ) as previously described [5] . Virus infectivity was assessed by fluorescent-focus assays for AdV5-GFP [24] . Antibodies used include mouse monoclonal anti-hexon ( Novocastra ) , anti- γ-tubulin ( ab27074 from Abcam ) , anti-dynein intermediate chain ( MAB1618 , Millipore ) , anti-KHC ( H1 and H2; MAB1613 and MAB1614 , Millipore ) , rabbit anti-Ad5 ( ab6982 from Abcam ) , anti-GFP ( A-11122 from Invitrogen ) , anti-β-tubulin ( ab6046 ) , goat anti-γ-tubulin ( sc-7396 from Santa Cruz ) , anti-Kif5B ( sc13356 from Santa Cruz ) . Mammalian expression constructs used in this work included RFP-tagged Kif5C as well as GFP-tagged full length Kif5A , Kif5B , and KLC1 and KLC2 ( all obtained from Kristin Verhey ) . Plasmids encoding the Ad proteins hexon , 100K and HA-tagged penton base were previously described [5] . Plasmid encoding myc-tagged penton base or protein IX tagged with FLAG and GFP ( pIX-flag-GFP ) were cloned from purified Ad DNA . Viral DNA was separated from capsid proteins by boiling purified virus particles at 100°C for 5min in 1%SDS and subsequent MiniPrep ( Qiagen ) . Protein IX was cloned into the pEGFP-C1 vector ( Clontech , Mountain View , CA ) by introducing 5’ EcoRI and 3’ AgeI restriction sites by PCR . The additional C-terminal FLAG-tag was introduced during PCR . Mutations in the sequence and in-frame cloning was tested for by 5’ and 3’ sequencing . Transient transfections were performed using either Lipofectamine 2000 ( Invitrogen ) or Effectene ( Qiagen ) . An siGENOME siRNA smartpool custom library targeting 38 plus-end directed human kinesins was designed and purchased from Dharmacon . Other siRNA pools used include Dharmacon siGENOMETM Control Pool and Kif5A , Kif5B , Kif5C , KLC1 , KLC2 , Kif4A , Kif21A smartpools . siRNA oligonucleotides were transfected into A549 cells using Lipofectamine 2000 ( QIAGEN ) . RNAi-resistant Kif5B was generated using QuickChange II site-directed mutagenesis kit ( Agilent technologies ) . Adenovirus infections were all performed in a low volume of DMEM containing 2% FBS at 4°C for 30 min to allow virus attachment . The cells were washed three times in cold PBS and incubated in fresh DMEM/2% FBS for 60 min at 37°C , unless stated otherwise , to allow internalization and intracellular transport . Leptomycin B ( Sigma ) was dissolved in dimethyl sulfoxide ( DMSO ) and kept at -20°C until use . Cells were treated with 20 nM leptomycin B for 60 min prior to adenovirus attachment and during adenovirus infection . Cells were grown on glass coverslips and fixed in methanol at -20°C for 5 min . Coverslips were blocked for 30 min in 0 . 5% donkey serum/PBS; incubated in primary antibody at 37°C for 1 hr , washed , and incubated for 1 hr at 37°C in Cy2- , Cy3- , or Cy5-conjugated secondary antibody; then stained with DAPI for 10 min to visualize DNA . Coverslips were mounted using ProLong Gold antifade mounting medium ( Invitrogen ) and imaged using either a Leica DM IRB/E inverted microscope equipped with a CCD camera ( ORCA 100; Hamamatsu ) or an IX80 laser scanning confocal microscope ( Olympus FV100 Spectral Confocal System ) . For all live cell imaging experiments , cells were grown on coverslips in coverslip-bottomed dishes ( MatTek Corp . ; Ashland , MA ) and infected with Alexa-546 labeled adenovirus [5] . Movies were acquired 15–75 min p . i . using a 63X oil immersion objective ( actual pixel size ~256 nm/pixel ) and a CCD camera ( model C9100-12; Hamamatsu ) attached to an inverted microscope ( IX80; Olympus; Center Valley , PA ) , or 60X oil immersion objective ( actual pixel size ~109 nm/pixel ) and an EMCCD camera ( Andor iXon ULTRA 897BV ) on a spinning disk confocal microscope ( Yokogawa; Olympus ) . Cells were typically imaged every 2 seconds or 30seconds , as noted . Z-stacks of images covering the entire cell thickness were acquired for the latter analysis , and the total intensity projection is shown . For high temporal resolution particle tracking analysis , imaging was at 23 frames/sec , continuously for 3000 frames . Cell outline and nucleus position in movies was identified from bright field images . Cells were judged to exhibit “nuclear accumulation” if they showed > 50% of virus particles associated with nucleus . Cells were judge to exhibit “pericentrosomal accumulation” if virus particles were sufficiently concentrated in this region to colocalize with a centrosomal marker or if centrosome position could be clearly inferred from local adenovirus concentration . To measure the concentration of virus particles in the pericentrosomal area quantitatively , we estimated average adenovirus number within a square of 5 . 8x5 . 8 μm2 around the centrosome . Relative Number Density ( RND ) , the ratio of virus number density per unit area within the “box” divided by average density throughout the rest of the cytoplasm is calculated as a measure of pericentrosomal virus accumulation . A custom-tracking algorithm was used to extract the position of adenovirus as a function of time [5] . Cells infected with Alexa-546 labeled viruses were imaged at a temporal resolution of 43ms/frame and Ad5 run length , velocity , and frequencies of movements were determined as previously described [5] . To separate diffusive and active transport events , we fit mean squared displacement ( MSD ) vs . time linearly and quadratically for each track , and used the Schwarz Bayesian information criterion for model selection . Only active transport events , whose MSD increases quadratically with time , were used for run length , velocity , and frequency analysis . A similar particle tracking program was used to identify Ad5 tracks in movies taken at 2s/frame for 10min ( Fig 4F ) . Most of the tracks in this case had a linearly increasing MSD with time . These tracks , in particular were used to determine apparent diffusion constant ( Fig 4J ) . Values for Ad5 MSD , apparent diffusion constant , and maximum displacement for each track each exhibited an exponential decay distribution , from which the mean and confidence intervals for these parameters were determined . Sprague Dawley rat ( Taconic ) brain tissue was used to prepare rat brain lysate and purified rat cytoplasmic dynein in phosphate-glutamate buffer ( pH 7 . 0 ) as previously described [25] . Kinesin was enriched for using 5–20% sucrose gradient density centrifugation of the microtubule GTP release fraction . The kinesin-1 fractions were devoid of cytoplasmic dynein and dynactin , and were used for further Ad5 and capsid protein binding analysis . Adenovirus hexon was either recovered from the virus-depleted supernatant from post-lytic cells by immunoprecipitation with a monoclonal anti-hexon antibody [5] or by anion exchange chromatography [54] . Initial steps of penton dodecahedra ( Pt-Dd ) purification followed the protocol described in [54] . Additionally , the first eluting peak from the anion exchange step was pooled , concentrated using spin columns ( Millipore ) , and applied to a Superose 6 10/300 GL column equilibrated with PBS . Pt-Dd-containing fractions eluted close to the void volume , and were pooled and concentrated . Purified proteins were flash frozen in liquid nitrogen and stored at -80°C . Mammalian cultured cell lysates were prepared in RIPA buffer ( 50mM Trizma-maleate , 100 mM NaCl , 1 mM EGTA , [pH 7 . 4] ) containing protease inhibitor cocktail ( Sigma ) and 1% NP40 . Insoluble debris was removed by centrifugation . For immunoprecipitations from cell lysate , antigen was recovered with anti-tag antibodies and protein A sepharose beads ( GE Bioscience ) , washed extensively in RIPA buffer , and analyzed for interactions by immunoblotting . Adenovirus capsid and hexon binding assays were described previously [5] . Briefly , virus or hexon was immunoprecipitated with anti-hexon antibody and protein A sepharose beads from purified virus stock or virus-depleted infected 293A cell lysate , respectively . The beads were washed and incubated for 30 min in Tris-maleate buffer ( 50 mM Trizma-maleate , 10 mM NaCl , 1 mM EDTA , and 0 . 1% Tween 20 , pH 4 . 4 or pH 7 . 4 ) , washed in the same buffer at pH 7 . 4 , and then incubated with purified kinesins at 4°C for 1 . 5 hr . Following washing , the beads were analyzed for the presence of kinesin by immunoblotting . In reciprocal experiments , anti-kinesin or anti-dynein IC antibodies were used to immunopurify motor proteins , which was subsequently incubated with purified virus capsid stock , Pt-Dd , or 293A cell lysate expressing viral capsid components . Two-sample comparisons were performed via either Student’s t test unless otherwise specified . F test was used for the comparison of mean run length ( Fig 4B ) , velocity ( Fig 4D ) , apparent diffusion coefficient ( Fig 4J ) , and maximum displacement ( Fig 4K ) , which are assumed to be exponentially distributed . Most error bars shows standard error of the mean , except in Fig 4B , 4C , 4I , 4J and 4K , where a 95% confidence interval is shown . Statistical significance was inferred for P < 0 . 05 ( denoted by * ) , P<0 . 01 ( denoted by ** ) , or P<0 . 001 ( denoted by *** ) for both tests . All statistical tests were performed using MATLAB ( MathWorks ) , Excel ( Microsoft ) or R software . | The role of plus-end directed microtubule motors in virus entry into host cells is a long-standing question . In this study , the authors identify the kinesins responsible for adenovirus plus end-directed transport along microtubules , the mechanism for kinesin recruitment , and both directed and motor-based exploratory movements involved in adenovirus’ search for the nucleus . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"kinesins",
"rna",
"interference",
"pathology",
"and",
"laboratory",
"medicine",
"microtubules",
"gene",
"regulation",
"pathogens",
"microbiology",
"viral",
"structure",
"viruses",
"dyneins",
"molecular",
"motors",
"dna",
"viruses",
"epigenetics",
"cellular",
"structures",
"and",
"organelles",
"motor",
"proteins",
"cytoskeleton",
"small",
"interfering",
"rnas",
"genetic",
"interference",
"proteins",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"biochemistry",
"cytoskeletal",
"proteins",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"virology",
"viral",
"pathogens",
"adenoviruses",
"genetics",
"biology",
"and",
"life",
"sciences",
"capsids",
"non-coding",
"rna",
"microtubule",
"motors",
"organisms"
] | 2018 | Role of kinesins in directed adenovirus transport and cytoplasmic exploration |
Tsetse flies ( Glossina spp . ) are the cyclical vectors of Trypanosoma spp . , which are unicellular parasites responsible for multiple diseases , including nagana in livestock and sleeping sickness in humans in Africa . Glossina species , including Glossina morsitans morsitans ( Gmm ) , for which the Whole Genome Sequence ( WGS ) is now available , have established symbiotic associations with three endosymbionts: Wigglesworthia glossinidia , Sodalis glossinidius and Wolbachia pipientis ( Wolbachia ) . The presence of Wolbachia in both natural and laboratory populations of Glossina species , including the presence of horizontal gene transfer ( HGT ) events in a laboratory colony of Gmm , has already been shown . We herein report on the draft genome sequence of the cytoplasmic Wolbachia endosymbiont ( cytWol ) associated with Gmm . By in silico and molecular and cytogenetic analysis , we discovered and validated the presence of multiple insertions of Wolbachia ( chrWol ) in the host Gmm genome . We identified at least two large insertions of chrWol , 527 , 507 and 484 , 123 bp in size , from Gmm WGS data . Southern hybridizations confirmed the presence of Wolbachia insertions in Gmm genome , and FISH revealed multiple insertions located on the two sex chromosomes ( X and Y ) , as well as on the supernumerary B-chromosomes . We compare the chrWol insertions to the cytWol draft genome in an attempt to clarify the evolutionary history of the HGT events . We discuss our findings in light of the evolution of Wolbachia infections in the tsetse fly and their potential impacts on the control of tsetse populations and trypanosomiasis .
The genus Wolbachia encompasses intracellular maternally inherited Gram-negative bacteria estimated to infect over 40% of insect species , in addition to filarial nematodes , crustaceans , and arachnids [1] , [2] . Wolbachia interactions with its host can have diverse outcomes that range from mutualistic to pathogenic or reproductive parasitism [3] . In arthropods , Wolbachia alterations to host reproduction include parthenogenesis induction , male killing , feminization of genetic males , and cytoplasmic incompatibility ( CI ) [1] , [4] . In its simplest form , CI occurs when a Wolbachia infected male mates with an uninfected female , causing developmental arrest of the embryo . In contrast , Wolbachia infected females can mate with either an uninfected male or a male infected with the same Wolbachia strain , and produce viable Wolbachia infected offspring . It has been suggested that the reproductive advantage afforded by the Wolbachia induced CI mechanism may permit the rapid spread of desirable host phenotypes into natural populations as a novel disease control approach [4]–[7] . A number of Wolbachia whole genome sequence ( WGS ) data are available to date and at least ten more genomes are currently being sequenced from a diverse set of hosts [8]–[15] . The majority of the Wolbachia strains have genomes that range from 1 . 08 to 1 . 7Mb in size [12] . Although most Rickettsiales have small genomes , Wolbachia sets a different pace by carrying an extremely high number of mobile and repetitive elements [4] , [16] , [17] . In addition , a number of Ecdysozoan genomes have been reported to contain chromosomal insertions originating from Wolbachia , including the mosquito Aedes aegypti [18] , [19] , the longhorn beetle Monochamus alternatus [20] , filarial nematodes of the genera Onchocerca , Brugia , and Dirofilaria [21] , [22] , parasitoid wasps of the genus Nasonia [22] , the fruit fly Drosophila ananassae [22] , the pea aphid Acythosiphon pisum [23] , and the bean beetle Callosobruchus chinensis [23] , [24] . Horizontal gene transfer ( HGT ) events in prokaryotes are rather common , and represent a way for bacteria to acquire novel features that enable them to adapt to different environments and to reorganize their genome [25]–[27] . In unicellular eukaryotes , gene transfer events are also relatively common [28] . Since many unicellular eukaryotes are phagotrophic on bacteria and other micro-organisms , they are constantly exposed to prokaryotic DNA , which may predispose them to incorporate foreign genetic material into their genomes [29] . By contrast , in multi-cellular organisms HGTs are rare [30] . It is likely that the localization of Wolbachia within the host germ-line cells [31] may have enabled the transfer of its genetic material to the host chromosomes . Tsetse flies are the exclusive vectors of Human African Trypanosomes ( HAT ) , also known as sleeping sickness , and of the livestock disease Nagana in sub-Saharan Africa . These diseases are caused by different members of the kinetoplastid protozoan parasites , Trypanosoma spp . The World Health Organization ( WHO ) has estimated that 60 million people in Africa live in tsetse infested areas , and are at risk of contracting sleeping sickness [32] . Disease control in the mammalian host is complicated due to the lack of vaccines , cheap and effective therapeutic treatments , and simple accurate diagnostic tools [33] , [34] . Tsetse flies also harbor multiple symbiotic microbes , which display different levels of integration with their host . The obligate mutualist genus Wigglesworthia provides dietary supplements to support host fecundity and is also necessary during larval development for the adult immune maturation processes [35]–[38] . The facultative symbiont genus Sodalis is present in some individuals in natural populations and may play a role in tsetse's trypanosome transmission ability ( vector competence ) [38] , [39] . The ability to cultivate Sodalis in vitro and transform and repopulate tsetse with modified Sodalis has led to a potential paratransgenic control strategy to modify tsetse's vector competence by expressing trypanocidal molecules in recSodalis [40]–[44] . Natural populations of many tsetse species also harbor a third symbiont , which belongs to the genus Wolbachia . Recent surveys indicate that Wolbachia infection prevalence in natural populations of different tsetse species can vary considerably , with some populations having near 100% infection prevalence [41] , [45] . We recently demonstrated that Wolbachia infections in Glossina morsitans morsitans ( Gmm ) induce CI in the laboratory and confer a reproductive advantage to infected females [41] . Further modeling of CI demonstrated the potential use of Wolbachia to drive a desirable host phenotype into a natural tsetse population [41] , [46] . Thus , it is suggested that tsetse carrying modified Sodalis expressing antiparasitic molecules in their midgut can be used to replace their wild parasite-susceptible counterparts through Wolbachia-mediated CI . One population control method that has been successful for tsetse , and currently being implemented in Africa , is the sterile insect technique ( SIT ) , where males rendered sterile through irradiation are released to mate with wild females and suppress their fecundity [41] , [47] . A promising alternative/complementary approach to SIT could be the use of the incompatible insect technique ( IIT ) , which relies on Wolbachia-induced sterility in the released males instead of irradiation [48] , [49] . In this paper , which is being submitted as a satellite to the manuscript describing the WGS of the tsetse species Gmm , we report on the draft genome sequence of its associated cytoplasmic Wolbachia endosymbiont ( cytWol ) . Moreover , we mined the WGS of Gmm and report on the presence of multiple extensive chromosomal insertions of Wolbachia ( chrWol ) in the host genome . These results confirmed our previous PCR-amplification based data suggesting the presence of HGT event ( s ) between Wolbachia and Gmm [45] . The HGT events were validated by Southern blot and Fluorescent in situ Hybridization ( FISH ) analyses on Gmm chromosomes . We compared the chrWol insertions discovered in the assembled Gmm genome to cytWol to understand the evolution of HGT events , and discuss our findings in light of the evolution of Wolbachia infections in tsetse . Finally , we analyzed the presence of Wolbachia HGT events in several Gmm natural populations , and discuss the potential to harness Wolbachia effects for the control of tsetse-transmitted diseases .
For the genome sequencing of the naturally infected Wolbachia strain of G . m . morsitans ( wGmm ) , approximately 250 ovaries were dissected from adult females from the Gmm colony maintained in the Yale University insectary . DNA was prepared using Qiagen DNeasy kit ( Qiagen , Inc . , Valencia , CA ) . The complete genome sequence was determined using whole-genome shotgun pyrosequencing using the Roche 454 GS sequencer FLX Titanium system ( 454 Life Sciences , Branford , CT , USA ) . In order to improve the wGmm draft genome , Illumina read libraries from the tsetse genome assembly were used . These were obtained from: ( a ) a pool of five tsetse flies . and ( b ) the first larval progeny of tetracycline-treated female . Two sets of Illumina reads were used: a PCR-free small fragment ( ∼300 bp ) library and Hi-Seq mate-pair libraries with an insert of approximately 1 . 6 kb . The tsetse ovary DNA used for wGmm sequencing contained a mixture of host genetic material , as well as cytoplasmic ( cyt ) and chromosomal ( chr ) Wolbachia DNA . A customized informatics pipeline was developed to computationally distinguish between sequence reads . An initial assembly was performed using MIRA [50] . First , all host sequences were removed by mapping the 454 reads to the Wolbachia reference genomes ( wMel , wRi , wPip and wBm ) . The filtered sequence reads contained chromosomal and cytoplasmic reads . The chromosomal reads were further removed using MIRA by mapping the filtered sequences to the chromosomal Wolbachia contigs ( 99% cut-off ) . The same procedure was followed for the Illumina data . The resulting 454 and Illumina reads were de novo assembled using MIRA . This initial assembly was subsequently improved using approaches described in the PAGIT protocol [51] . In brief , the contigs were aligned to the wMel genome using ABACAS [52] , creating one large scaffold that consisted of the contigs successfully mapped to the wMel genome and a set of contigs that did not map . An attempt was made to close the gaps in the large scaffold using IMAGE [53] with the PCR-free small fragment library . After gap closing , the large scaffold was reduced once more to a set of contigs by breaking it around any of the unclosed gaps . This is because there are usually many genome rearrangements between different Wolbachia strains , and we would therefore expect a number of rearrangements to exist between the wMel and wGmm genomes . Breaking the scaffold makes allowance for these gaps . Finally , scaffolding was then performed on this reduced set of contigs using SCARPA [54] with the Hi-Seq mate-pair libraries . The statistics for the assembly at each stage of the process are given in Table S1 . The genome was annotated with XBASE and RAST [55] , [56] , followed by manual curation . Putative protein-encoding genes were identified using GLIMMER [57] and tRNA by tRNAscan-SE [58] . Predicted proteins were examined to detect frame-shifts or premature stop codons to identify pseudogenes using ARTEMIS [59] . Those for which the frame-shift or premature stops were of high quality by examining re-mapped reads in these regions were annotated as “authentic” mutations . This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession AWUH00000000 . The version described in this paper is version AWUH01000000 . The Sanger and 454 reads used in the tsetse genome assembly were obtained from flies treated with tetracycline as described previously [41]; therefore , these reads did not contain cytWol sequences . As mentioned above , Wolbachia specific sequences were filtered out from WGS reads of each sequencing technology with MIRA [50] using the complete genomes of wMel ( AE017196 ) , wRi ( CP001391 ) , and wBm ( AE017321 ) as reference sequences . We obtained 5 , 306 ( Sanger ) , and 10 , 978 ( 454 ) Wolbachia-specific sequences respectively . All the filtered putative Wolbachia-specific sequences were further examined using blast and a custom made Wolbachia database . ChrWol-specific sequences were assembled with MIRA and AMOS [50] , [60] using as a reference sequence the wGmm draft genome . The statistics for the two chrWol assemblies are as follows: N50 2970 , mean contig length 1261 . 97 , longest contig 15053 , total length 527 , 504 bp for insertion A , while for insertion B N50 2791 , mean contig length 1092 . 82 , and total length 484 , 123 . Genes were identified with Glimmer [61] , followed by a round of manual curation using Blastn [62] and MegaBlast [62] against the non-redundant and custom made Wolbachia databases . The predicted CDSs were translated and used to search the NCBI non-redundant database , KEGG , and COG databases . The tRNAScan-SE tool [58] was used to identify tRNA genes . Phylogenetic analyses were performed using Maximum-Likelihood ( ML ) and Neighbor-Joining ( NJ ) estimation for a concatenated set of six phage genes from wGmm , wRi , wMel and wPip . The genes used for the phylogeny included HK97 family phage major capsid protein ( wGmm_0882 , WD_0458 , WRi_002750 , WP0102 ) , phage integrase family site-specific recombinase ( wGmm_0004 , WD_1148 , WRi_009900 , WP0980 ) , phage SPO1 DNA polymerase-related protein ( wGmm_0674 , WD_0164 , WRi_000900 , WP0922 ) , prophage LambdaW5 baseplate assembly protein W ( wGmm_0971 , WD_0640 , WRi_005480 , WP0303 ) , prophage LambdaW1 , baseplate assembly protein J ( wGmm_0970 , WD_0639 , WRi_010130 , WP0302 ) and a prophage LambdaW1 , site-specific recombinase resolvase protein ( wGmm_0960 , WD_0634 , WRi_005400 , WP0342 ) . In addition , a concatenated set of ten genes ( DNA-directed RNA polymerase , DNA polymerase III ( alpha subunit ) , DNA gyrase B , translation elongation factor G , aspartyl-tRNA synthetase , CTP synthase , glutamyl-tRNA ( Gln ) amidotransferase B , GTP-binding protein , cell division protein FtsZ , fructose-bisphosphate aldolase ) from the identified Gmm chromosomal insertions , wGmm , wRi , wMel , wPip , and wBm were used . All sequences were aligned using MUSCLE [63] and ClustalW [64] as implemented in Geneious 5 . 4 [65] , and adjusted manually . ML and NJ trees were constructed using MEGA 5 . 0 [66] with gamma distributed rates with 1000 bootstrap replications and the method of Tamura-Nei as genetic distance model [67] . To determine the number of chromosomal insertions , genomic DNA from tetracycline-treated Gmm females and normal Gmm individuals were restricted with HindIII endonuclease , electrophoresed on 1% agarose gel in 1× TBE buffer , and transferred to a positively charged nylon membrane according to Southern protocol [68] . The membrane was hybridized at 55°C with 350 ng of a 569 bp probe corresponding to part of the wsp gene labeled with the Gene Images Alkphos Direct labeling system ( GE Healthcare , Little Chalfont , UK ) using the random primer method following manufacturer protocols . Signal detection was performed using CDP-star followed by exposure to autoradiographic film ( X-OMAT AR , Kodak ) . The absence of cytWol from the tetracycline-treated Gmm DNA was confirmed by a PCR assay , which resulted in only a single 16S rRNA amplification product originating from the chromosomal insertions [45] . Mitotic chromosome spreads were obtained from freshly deposited larvae from the Slovakia Academy of Sciences Institute of Zoology tsetse laboratory Gmm strain . Briefly , larval nerve ganglia were incubated on a slide in 100 µl 1% sodium citrate for 10 min at room temperature , and sodium citrate was replaced with methanol-acetic acid ( 3∶1 solution ) for 4 min . The tissue was disrupted by pipetting in 100 µl 60% acetic acid for fixation and dropped onto clean slides heated on a hot plate at 70°C until acetic acid evaporation . After dehydration in 80% ethanol , slides were stored at −20°C for at least 2 weeks . For in situ hybridization experiments , multiple probes specific for Wolbachia 16S rRNA , fbpA and wsp genes were amplified from the Slovakian strain DNA [45] , [69] . To generate the labeled probes , 1 µg of DNA resuspended in 16 µl ddH2O was denatured by boiling for 10 min . 4 µl of labeling mix ( Biotin High Prime kit; Roche , Basel , Switzerland ) were added and the reaction was incubated overnight at 37°C . After the reaction was stopped , ddH2O ( 5 µl ) , 20×SSC buffer ( 25 µl ) and formamide ( 50 µl ) were added and 25 µl of denatured probe was placed on each pre-treated slide . The hybridization was performed at 37°C overnight in a humid chamber and detection of hybridization signals was performed using the Vectastain ABC elite kit ( Vector Laboratories , Burlingame , CA , USA ) and Alexa Fluor 594 Tyramide ( Invitrogen ) . Chromosomes were DAPI stained and the slides were mounted using the VECTASHIELD mounting medium ( Vector Laboratories ) . Chromosomes were screened under an epifluorescence Zeiss Axioplan microscope and images were captured using an Olympus DP70 digital camera . For the localization of signals on mitotic chromosomes the karyotype description of Willhoeft [69] , [70] was adopted . Natural samples of Gmm used to examine HGT fragments originated from four populations collected in Zambia , Zimbabwe and Tanzania ( Table 1 ) . DNA was isolated from adult flies stored in EtOH using the Qiagen DNeasy kit ( Qiagen , Valencia , CA ) following the manufacturers' instructions and stored at −20°C . The aposymbiotic ( Wolbachia-free ) Gmm line [41] was used as a control . For detection of Wolbachia , a PCR assay that amplified a 438 bp 16S rRNA fragment was used with the specific primer set wspecF and wspecR [71] . For input DNA control , a 377 bp fragment of the mitochondrial 12S rRNA gene was amplified with the primer set 12SCFR and 12SCRR [72] . The PCR amplification protocol was 10 min at 95°C , 35 cycles of 30 sec at 95°C , 30 sec at 54°C and 1 min at 72°C , and 10 min at 72°C . The identification of the Wolbachia strain infections was based on MLST ( gatB , coxA , hcpA , fbpA and ftsZ ) and wsp-based genotyping approaches [45] , [69] . PCR reactions were performed using the following program: 5 min of denaturation at 95°C , followed by 35 cycles of 30 sec at 95°C , 30 sec at the appropriate temperature for each primer pair ( 52°C for ftsZ , 54°C for gatB , 55°C for coxA , 56°C for hcpA , 58°C for fbpA and wsp ) and 1 min at 72°C . All reactions were followed by a final extension step of 10 min at 72°C . Both strands of the products were sequenced using the respective primers . In addition , PCR products of 16S rRNA , wsp and MLST genes from the Gmm populations analyzed were cloned in pGEM-T Easy Vector System , and PCR products from several clones generated by the primers T7 and SP6 were sequenced in both directions using the BigDye Terminator v3 . 1 Cycle Sequencing Kit ( PE Applied Biosystems ) and were analysed using an ABI PRISM 310 Genetic Analyzer ( PE Applied Biosystems ) . All Wolbachia gene sequences were manually edited with SeqManII by DNAStar and aligned using MUSCLE [63] , as implemented in Geneious 5 . 4 [65] , and adjusted manually . To determine if genes from the chromosomal insertions were potentially expressed in locations other than the gonotrophic tissues , we utilized mapping of Illumina datasets from other studies , that included transcriptome reads from somatic tissues [73]–[75] . Reads were mapped to the chromosomal insertions using CLC Genomics Workbench ( CLC Bio , Cambridge , MA ) allowing no mismatches per reads , a maximum of 10 hits per read and 80% of the gene must match at 95% . Predicted open reading frames ( ORF ) from the insertions were extracted and the following criteria were utilized to determine the possibility of expression: 1 ) at least 25 reads were recovered from the ORF and 2 ) those represented had coverage of over 85% of the ORF . This filtering approach excluded genes with a high number of mapped reads that were only present in small limited sections of the ORFs . These sections with high read numbers mapping but low coverage could be where sequence similarity between Gmm , Wigglesworthia or Sodalis is high enough to yield mapping to the chromosomal insertions .
The draft genome of cytoplasmic wGmm contains 201 contigs of 1 , 019 , 687 bp , comprised of 800 putative functional coding sequences ( CDS ) and 16 pseudogenes ( Figure 1 and Table 2 ) . The GC content of wGmm is 35 . 2% , in the range observed for the other sequenced Wolbachia genomes ( Table 2 ) . Although , the wGmm genome is not complete , based on comparison of the identified contigs , it is most similar to the two Wolbachia strains associated with Drosophila melanogaster and D . simulans , wMel and wRi , respectively ( Table S2 ) . It is more distantly related to the genomes of the Wolbachia strains associated with Culex pipiens and Brugia malayi , wPip and wBm , respectively ( Table S2 ) . The majority of the regions and genes missing from the wGmm genome relative to the wMel and wRi genomes encode phage , ankyrin and hypothetical proteins ( Tables S3 and S4 ) . Both PCR-based evidence from Wolbachia infected tsetse flies , and analysis of the Gmm annotated genome data indicated the presence of Wolbachia gene fragments inserted in the host genome . We mined the final assembly of the Gmm host genome and were able to identify 261 contigs that carried chrWol DNA sequences . Based on nucleotide diversity , close examination of the 261 contigs indicated that these represented at least three different events , which we refer to as insertions A , B and C . Manual editing and implementation of the AMOS snps script enabled the separation of the contigs into different insertions , with insertions A and B being the largest in size . Figure 2 shows the mapping of these two insertions on the wGmm reference genome . The observed pattern suggests that at least two large Wolbachia genome segments of 527 , 507 and 484 , 123 bps have been integrated into the Gmm chromosomes indicating that at least 51 . 7% and 47 . 5 . % of the draft Wolbachia genome were transferred to the host nuclear genome . Sequence analysis of insertion A predicted 197 putative functional coding sequences , 148 pseudogenes , and 15 tRNAs . Remnants of 163 pseudogenes were discovered that are greater than 100 bp in size and that have either partially been integrated into the host genome , or only represent part of the pseudogene . For insertion B , sequencing analysis revealed the presence of 159 putative functional coding sequences , 148 pseudogenes and 13 tRNAs . In insertion B , 157 remnants of pseudogenes were also identified . Thus , on average more than 60% of the genes transferred to the tsetse nuclear genome have been pseudogenized . The average length of the putative functional coding sequences is slightly smaller than wMel , wRi and the cytoplasmic wGmm at 690 bp for insertion A and 677 bp for insertion B ( Table S4 ) . The GC% content for insertion A and B is 35 . 1% . Comparison between the chromosomal insertions A and B and the wGmm draft genome using Blastn and lastz indicated that: ( a ) the two insertions are very similar to each other ( Figure S4 ) and ( b ) at least four genes , three hypothetical proteins and hemK are present in the chromosomal insertions but not in the cytoplasmic Wolbachia genome . The sequence identity between chromosomal and cytoplasmic genes and phylogenetic analysis based on ten concatenated genes clearly suggests that the chromosomal insertions A and B are closely related to the cytoplasmic wGmm genome ( Table 5 and Figure S3 ) . In more detail , comparison of the sequence identity in eleven chromosomal genes indicates that the majority of them exhibit a high sequence identity with the wGmm sequences ( Table 5 ) . The third Wolbachia HGT segment , insertion C , is only 2 , 089 bp in size and sequence analysis predicted the presence of only six pseudogenes . A number of different types of mutations were identified in insertions A and B present in the host nuclear genome , and these shed light on the pseudogenization process . Our analysis suggests that more than 80% of the mutations that accumulated in the putative functional coding sequences represent single nucleotide polymorphisms ( SNPs ) ( Figure 3 ) . The majority of the genes that have been pseudogenized accumulated mutations that consist of nucleotide polymorphisms with deletions ( NPD ) and NPs . In both insertions , genes that have been pseudogenized contain mutations that combine NPs and deletions ( NPDs ) are more than those pseudogenized by NPs ( Figure 3 ) . In addition , we identified two additional types of mutations , NPs with insertions and NPs with deletions and insertions , associated with both chromosomal Wolbachia insertions but to a much lesser degree . A list of partial and full genes corresponding to the chrWol insertions is available in Tables S6 , S7 and S8 . Based on our results , there were very few ORFs that met our criteria for expression from chromosomal insertions . In general , there were multiple ORFs that had high number of mapped reads ( >100 ) , but in nearly all cases the coverage of the mapping was below 30% indicating that these may represent reads from another symbiont or tsetse transcripts . Results were similar for the three transcriptomes analyzed from heads , salivary glands and the bacteriome . However , three putative ORFs satisfied our criteria: serB , ccmB and a degenerate transposase located at both insertions ( 102636-102894 for insertion A and 97255-97523 for insertion B ) . These analyses suggest that most of the genes present in the chromosomal insertions are likely not expressed , but the few specific genes we identified may have low levels of expression . Further studies will be necessary to validate their expression . Hybridization of the wsp probe to Gmm female DNA restricted with the HindIII enzyme produced five bands of about 1200 , 1600 , 2150 , 2600 and 2700 bp ( Figure 4 , lanes 1 and 3 ) . DNA from tetracycline-treated females ( cytWol-free ) had a similar profile , except that the 2700 bp band , corresponding to the expected cytWol wsp fragment , was absent ( lane 2 ) . Untreated male DNA displayed an additional band of 1500 bp , indicating the presence of insertions on the Y chromosome ( Lane 4 ) . This banding pattern suggests the presence of at least five independent wsp chromosomal insertions , including one on the Y chromosome , supporting the in silico analyses . To determine the location of Wolbachia insertions on Gmm chromosomes , we performed FISH analyses on mitotic spreads using wsp , 16S rRNA and fbpA specific probes . The Gmm mitotic complement , comprising the supernumerary dispensable chromosomes ( B chr ) [78] is depicted in Figure 5 , where the AT-rich heterochromatic nature of Y and B chromosomes is indicated by the strong DAPI-staining . The two autosomes , L1 and L2 , as well as the X chromosome , appear to contain heterochromatic regions on both sides of the centromere . FISH results indicate that the Wolbachia genes 16S , fbpA , and wsp consistently display a biased location on the distal part of the X , Y and B chromosomal arm . Although tyramide labeling generates strong and site-specific signals , it is difficult to detect the presence of multiple insertions on one chromosome if these events are localized in close proximity . The 16S rRNA signal detected on the short arm of the X chromosome appears to be particularly strong and diffused , and may thus represent more than one insertion event in that region . Our previous characterization of the laboratory Gmm strain by Wolbachia-specific 16S rRNA-based PCR screening , the wsp-based and the MLST typing system revealed several HGT events [45] . Our results presented above indicate that these transfer events are in fact more extensive than previously considered . We next investigated the presence of HGT events in natural populations of Gmm originating from Zambia , Tanzania and Zimbabwe . We detected the pseudogenized fragment of the 16S rRNA gene carrying a deletion of 142 bp ( Figure 6 ) , similar to that we described in Gmm colony DNA prepared from the tetracycline-treated ( cytWol-free ) samples [41] , [45] . We observed a similar phenomenon for fbpA , where a pseudogenized gene fragment could be amplified containing two deletions of 47 and 9 bp from the same four natural populations , as well as from the cytWol-free Gmm laboratory strain DNA sample . Finally , the HGT event of the Wolbachia wsp gene , which has been pseudogenized through a deletion of 7 bp , was also detected in two natural samples ( Figure 6 ) . Unlike the laboratory line of Gmm , in which all individuals analyzed carried the cytWol strain ( 100% infected ) , the prevalence of Wolbachia varied in the different populations and was not fixed ( Table 1 ) .
wMel genome ( AE017196 ) , wRi genome ( CP001391 ) , wPip genome ( AM999887 ) , wBm genome ( AE017321 ) , DNA-directed RNA polymerase ( WD_0024/GI 42409679 , WRi_000230/GI:225591874 , WP0554/GI:190357240 , Wbm0647/GI:58419220 ) , DNA polymerase III ( alpha subunit ) ( WD_0780/GI:42410358 , WRi_006220/GI:225592374 , WP0658/GI:190357336 , Wbm0499/GI:58419072 ) , DNA gyrase B ( WD_0112/GI:42409755 , WRi_001420/GI:225591969 , WP1103/GI:190357759 , Wbm0764/GI:58419337 ) , translation elongation factor G ( WD_0016/GI:42409671 , WRi_000140/GI:225591866 , WP0562/GI:190357248 , Wbm0344/GI:58418918 ) , aspartyl-tRNA synthetase ( WD_0413/GI:42410026 , WRi_003280/GI:225592127 , WP0387/GI:190357090 , Wbm0012/GI:58418589 ) , CTP synthase ( WD_0468/GI:42410077 , WRi_002850/GI:225592086 , WP1235/GI:190357884 , Wbm0169/GI:58418745 ) , glutamyl-tRNA ( Gln ) amidotransferase B ( WD_0146/GI:42409786 , WRi_003090/GI:225592108 , WP0087/GI:190356829 , Wbm0445/GI:58419018 ) , GTP-binding protein ( WD_1098/GI:42410645 , WRi_012740/GI:225592933 , WP0891/GI:190357560 , Wbm0032/GI:58418609 ) , cell division protein FtsZ ( WD_0723/GI:42410305 , WRi_007520/GI:225592482 , WP0577/GI:190357261 , Wbm0602/GI:58419175 ) , fructose-bisphosphate aldolase ( WD_1238 GI:42410776 , WRi_012130/GI:225592879 , WP1081/GI:190357738 , Wbm0097/GI:58418674 ) , HK97 family phage major capsid protein ( WD_0458/GI:42410067 , WRi_002750/GI:225592077 , WP0102/GI:190356840 ) , phage integrase family site-specific recombinase ( WD_1148/GI:42410690 , WRi_009900/GI:225592678 , WP0980/GI:190357644 ) , phage SPO1 DNA polymerase-related protein ( WD_0164/GI:42409803 , WRi_000900/GI:225591926 , WP0922/GI:190357589 ) , prophage LambdaW5 baseplate assembly protein W ( WD_0640/GI:42410229 , WRi_005480/GI:225592306 , WP0303/GI:190357018 ) , prophage LambdaW1 , baseplate assembly protein J ( WD_0639/GI:42410228 , WRi_010130/GI:225592699 , WP0302/GI:190357017 ) and a prophage LambdaW1 site-specific recombinase resolvase protein ( WD_0634/GI:42410223 , WRi_005400/GI:225592300 , WP0342/GI:190357056 ) | African trypanosomes are transmitted to man and animals by tsetse fly , a blood sucking insect . Tsetse flies include all Glossina species with the genome of Glossina morsitans morsitans ( Gmm ) being sequenced under the International Glossina Genome Initiative . The endosymbionts Wigglesworthia glossinidia , Sodalis glossinidius and Wolbachia pipientis ( Wolbachia ) have been found to establish symbiotic associations with Gmm . Wolbachia is known to be present in natural and laboratory populations of Glossina species . In this study we report the genome sequence of the Wolbachia strain that is associated with Gmm . With the aid of in silico and molecular and cytogenetic analyses , multiple insertions of the Wolbachia genome were revealed and confirmed in Gmm chromosome . Comparison of the cytoplasmic Wolbachia draft genome and the chromosomal insertions enabled us to infer the evolutionary history of the Wolbachia horizontal transfer events . These findings are discussed in relation to their impact on the development of Wolbachia-based strategies for the control of tsetse flies and trypanosomiasis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"invertebrates",
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"microbiology",
"animals",
"glossina",
"genome",
"sequencing",
"microbial",
"evolution",
"molecular",
"biology",
"techniques",
"tsetse",
"fly",
"epidemiology",
"molecular",
"biology",
"disease",
"vectors",
"insects",
"arthropoda",
"vector",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"organisms"
] | 2014 | Presence of Extensive Wolbachia Symbiont Insertions Discovered in the Genome of Its Host Glossina morsitans morsitans |
Epithelial cells in the colon are arranged in cylindrical structures called crypts in which cellular proliferation and migration are tightly regulated . We hypothesized that the proliferation patterns of cells may determine the stability of crypts as well as the rates of somatic evolution towards colorectal tumorigenesis . Here , we propose a linear process model of colonic epithelial cells that explicitly takes into account the proliferation kinetics of cells as a function of cell position within the crypt . Our results indicate that proliferation kinetics has significant influence on the speed of cell movement , kinetics of mutation propagation , and sensitivity of the system to selective effects of mutated cells . We found that , of all proliferation curves tested , those with mitotic activities concentrated near the stem cell , including the actual proliferation kinetics determined in in vivo labeling experiments , have a greater ability of delaying the rate of mutation accumulation in colonic stem cells compared to hypothetical proliferation curves with mitotic activities focused near the top of the crypt column . Our model can be used to investigate the dynamics of proliferation and mutation accumulation in spatially arranged tissues .
Colorectal cancer is the third most prevalent cancer type for both men and women in the United States , accounting for 9% of all cancer deaths [1] . This large incidence can be partially attributed to the rapid cell divisions that continuously replenish the colonic epithelium , as this large amount of cell turnover increases the risk of accumulating the genetic changes leading to colorectal tumorigenesis [2] . The identity and order of genetic alterations leading to colorectal cancer have been extensively studied [3] . The gene most frequently altered in colorectal cancer is adenomatous polyposis coli ( APC ) , with more than 85% of all colorectal cancer cases harboring mutations in this gene [4] . APC , a tumor suppressor , is a negative regulator of the β-catenin oncoprotein [5] , and mutations in APC lead to elevated levels of β-catenin in the cytoplasm , which in turn induce changes in proliferation , differentiation , migration , adhesion , and apoptosis [6] . Germline APC mutation results in the familial adenomatous polyposis ( FAP ) syndrome , which is characterized by an early onset of colorectal cancer in almost all afflicted individuals [7] . Other frequently altered genes in colorectal cancer include KRAS [8] , the SMAD genes [9] , TP53 [10] , and MYC [11] , [12] . In addition to alterations of oncogenes and tumor suppressor genes , colorectal tumors often display a mutator phenotype , which has been broadly categorized as microsatellite instability ( MIN ) [13] or chromosomal instability ( CIN ) [14] . About 15% of sporadic colorectal cancers display MIN , caused by a loss of DNA mismatch repair gene function [13]; the remaining 85% have CIN , characterized by an excessive rate of gaining or losing whole chromosomes or parts of chromosomes , at a rate of up to 10−2 per chromosome per cell division [14] . An important feature associated with tumors harboring CIN is the accelerated rate of loss of heterozygosity ( LOH ) , which increases the rate of tumor suppressor gene inactivation . It is not unusual for more than half of the genes in colorectal tumor cells to display LOH [15] . More than one hundred genes associated with CIN have been identified in yeast , many of which have human homologs [16] , [17] . In addition to the genetic sequence leading to colorectal cancer , the physical architecture and proliferation kinetics of epithelial cells have also been the topic of many investigations . Epithelial cells in the colon are arranged in cylindrical compartments called crypts [18] . Each crypt contains on average 2 , 000 cells , with about 40 cells in circumference and 80 cells in height [19] . A small number of stem cells ( 4–6 ) are located at the bottom of the crypt [20] , [21] . These cells divide to produce the differentiated progenies populating the crypt . The latter cells divide and migrate upward with limited lateral movement and are eventually shed off into the lumen of the large intestine [22] . The proliferation kinetics of cells follows a complex and spatially specific pattern , with proliferating cells concentrated at the bottom half of the crypt , near the stem cells , and the upper half of the crypt consisting of non-dividing migrating cells [19] , [23] . This proliferation pattern is tightly controlled , and changes in this pattern have been shown to be associated with the progression towards colorectal cancer [24] . Quantitative measurements in animal models demonstrated that the speed of the upward migration increases from 0 . 02 cell positions per hour per crypt column at the bottom of the crypt to approximately 1 . 0 cell positions per hour near the top [25] . Under normal circumstances , the entire crypt is regenerated every 2 to 7 days [2] . Overall , the human colon contains about 107 crypts , thus bringing the total number of epithelial cells in the colon to about 2×1010 cells . As experiments involving colonic epithelial cells remain technically challenging or infeasible in humans , several mathematical and computational models were developed to enhance our understanding of crypt kinetics and the somatic evolution leading to colorectal cancer . Early work has led to the postulation that colorectal cancer is the result of a sequential accumulation of mutations [26] . Since then , many mathematical models have been proposed to describe the accumulation of mutations leading to colorectal cancer . For instance , some investigators have addressed the effects of tissue architecture on the rate of mutation accumulation in colonic tissues . These studies include the spatially explicit models proposed by Komarova and Wang [27] to investigate the location within the crypt at which APC mutations arise , by Michor et al . to elucidate the time during tumorigenesis at which CIN arises [28] , and by Buske et al . to investigate the changes in tissue dynamics resulting from gains or losses of specific gene functions using an agent-based model [29] . In addition , Nowak et al . proposed a linear process model to study the speed of somatic evolution in colonic crypts [30] . In this model , N cells within a crypt column are projected onto a one-dimensional grid . During each time step , a cell is selected for reproduction . A cellular division yields two daughter cells , with one daughter occupying the original position and the other daughter residing in the position immediately to the right of the original cell . All cells on the right of the dividing cell move to the right by one position in the grid , and the last cell is shed off into the lumen of the intestine . During each cell division , a mutation may occur with a certain probability; each daughter cell has a chance of 1/2 of inheriting the mutation . Compared to a well-mixed population of cells , this linear process was shown to slow down the speed of somatic evolution and to conceal the selective effects of advantageous mutants [30] . This observation suggests that the cellular architecture of multicellular tissues has the potential to delay the onset of cancer . Several other models were designed to specifically investigate crypt kinetics . Two excellent reviews by van Leewen et al . [31] and de Matteis et al . [32] provide an in-depth discussion of these studies: a two-dimensional lattice-based model [33] , [34] , a one-dimensional lattice-based model with an intraepithelial growth factor gradient [35] , a two-dimensional lattice-free model based on Voronoi tessellation [36] , and a cellular Potts model ( CPM ) [37] . Recently , Mirams and Fletcher presented an integrated model incorporating both proliferation kinetics and tissue architecture for investigating mutation fixation within a crypt [38] . Using the number of proliferating cells as a proxy for proliferation kinetics , they showed that the dynamics of cell division have a significant effect on the spread of mutated cells within the population . Despite these forays , several open questions remain regarding the effects of proliferation kinetics on the rate of mutation accumulation towards colorectal cancer . To address these issues , we developed a spatially arranged stochastic model of the colonic crypt . We investigated several different proliferation kinetics curves , including one quantitatively measured using labeling indices in the normal colonic epithelium [19] , [23] , and their effects on the rate of somatic evolution towards colorectal tumorigenesis . This model contributes to a quantitative understanding of the initiation and progression of colorectal cancer and can be used to investigate the effects of spatial patterns on mutation accumulation .
In order to investigate the effects of proliferation kinetics on the rate of somatic evolution toward colorectal tumorigenesis , we designed a spatial model capturing the essential features of tissue architecture and cellular movement in colonic crypts . Each colonic crypt is modeled by a representative column of cells , which is projected onto a linear lattice ( Fig . 1A . ) . The total number of cells per column is given by , as determined by in vivo measurements with a measured mean of 81 . 9 cells ( ±9 . 7 cells ) [19] . Position 1 on the left end of the lattice represents the stem cell and position on the far right represents the apex of the crypt , close to the gut lumen . During every elementary time step of this stochastic process , a cell at position i is selected to divide according to a probability weight , wi , defined by a specific proliferation kinetic curve for , zero elsewhere . The two daughter cells are then placed into positions i and i+1 , causing cells that previously resided in positions i+1 to to shift by one position to the right . The last cell is shed into the gut lumen . During each division , a mutation may occur with probability u . If a mutation arises , then each of the daughter cells has a chance of 1/2 of inheriting the mutation . This flexible model then allows us to investigate the effects of different proliferation curves on the rate of somatic evolution . Our model closely resembles the one originally proposed by Nowak et al . [30] , with the difference of incorporating specific proliferation kinetics . The dynamics of proliferation is a function of the cell position in the crypt column . The proliferation kinetic curve assigns a mitotic probability to each position in the crypt column: a more proliferative position in the crypt column is represented by a larger mitotic probability . We examined five proliferation kinetic curves ( Fig . 1C . ) . Position 1 , the stem cell position , has the same mitotic probability in all curves , such that is identical for all curves . Weights for positions 2 to 80 are assigned by different proliferation curves . Proliferation curve 1 represents the measured kinetic curve , which is extrapolated from in vivo bromodeoxyuridine ( BrdU ) labeling experiments [19] , [23] . The measured distribution was approximated using a normal distribution with mean 18 and standard deviation 15 to best match the 90th percentile interval of the measured curve . The probability weights for choosing a cell at each position in the crypt column , i = 2 to 80 , are specified by the probability density function , . Curve 2 represents the logistic proliferation curve generated from ; this curve is used to investigate the effects of spreading the proliferation activities upwards in the crypt column . Curve 3 represents a uniform curve . Curve 4 represents the mirror image of curve 2 , with a vertical plane of reflection between positions 40 and 41 . Finally , curve 5 represents the mirror image of curve 1 , with a vertical plane of reflection between positions 40 and 41 . Curves 4 and 5 were selected to examine the effects of proliferating activities concentrated far away from the stem cell . In addition to the normal shedding of the last cell in the crypt column , accidental premature cell death may also occur ( Fig . 1B . ) . The rates of apoptosis have been measured using terminal deoxynucleotidyl transferase dUTP nick end labeling ( TUNEL ) in normal cells , hyperplastic polyps , adenomas and carcinomas; in these cases , the percentage of cells being labeled ranged from 1% to 4% [39] . These observations suggest that multiple cell deaths may occur during each round of cell division . In the context of our model , such cell death may occur after each normal cell division event . The number of dying cells , m , follows a Poisson distribution with mean λ , which denotes the mean number of cell deaths per round of cell division . To incorporate the possibility of having multiple cell deaths , λ can vary between 0 and 3 . 2 . The upper bound of 3 . 2 corresponds to 4% of the length of the crypt column , which is the observed percentage of cells undergoing apoptosis using Tunel labeling [39] . Cells are selected for death according to a particular apoptosis curve specifying the likelihood of cell death for each position in the crypt column . Due to a lack of quantitatively measured apoptosis curves as a function of cell position within the crypt column , we used the five proliferation curves discussed above as proxies for apoptosis curves and tested the resulting twenty five combinations of proliferation and apoptosis curves for their effects on the dynamics of the system . The vacancy resulting from a cell death at position j can be filled by an additional cell division occurring at position k , where k≤j+δ and δ indicates the size of the interval in which more differentiated cells can replace the dead cell . All following numerical examples are calculated based on δ = 5 . Position k is selected for a replenishing cell division according to the reweighted proliferation kinetics , , where indicates the existence of a viable cell at position . If multiple cells die simultaneously , the replenishing cell divisions occur sequentially , in the order of j ( 1 ) , j ( 2 ) … j ( m ) , where j ( 1 ) , j ( 2 ) … j ( m ) are ordered positions for the m dying cells . The m positions for the replenishing cell divisions are again selected according to the sequentially reweighted proliferation kinetic curve . This design was chosen to ensure that the dead cell ( s ) can only be replaced by cells of a similar differentiation stage ( j+1≤k≤j+δ ) or a less differentiated stage ( k<j ) . In this model , the differentiation hierarchy is fully specified by each cell's position in the crypt column , with position 1 representing the least differentiated cell and position representing the most differentiated cell . In addition , this replacement rule captures the essential features of two biological observations governing cellular repopulation of crypt columns: ( i ) replacement mitotic activities are concentrated near the stem cell [40] , [41]; and ( ii ) newly divided cells migrate upward in the crypt column [25] . The accumulation of mutations is described by specific transition probabilities between different mutational states . The possibility of back mutations , which reconstitute a less mutated state , is neglected . The mutation rates are assumed to be constant with respect to cell positions and with respect to time . Table 1 provides the range of plausible values used for the individual mutation rates . In addition to mutations arising during cell division , we also incorporated random mutations not linked to replication into our framework . Due to the lack of data on the positional dependency of this type of random mutation , such mutations are assumed to be equally likely to occur in any position in the crypt column . The mutation rate for the cell division-independent mutation events is considered to be the same as that of mutation events linked to cell division . Time is measured in rounds of cell shedding . At each elementary time step , the cell at position 80 is shed off into the lumen of the gut as the result of cell division and the associated cell movement . Cell deaths and replenishing cell divisions that replace dead cells are not counted as extra increments in time . The rationale for this design is to distinguish between the regenerative cell divisions under normal circumstances and the compensatory cell divisions for apoptotic losses . The instantaneous time scale for replenishing cell divisions is extrapolated from the observation that under strong external stimuli , the rate of progression through the cell cycle is accelerated; for instance , ionizing radiation was shown to induce proliferative activity [42] and shorten the duration of the cell cycle [43] . In addition , thermal injury [44] and starvation-induced stress [45] also tend to increase mitotic activity . In this study , we adopted the assumption of instantaneous replenishing cell divisions regardless of whether strong external stimuli are present or not . Relative fitness is defined as the ratio of the proliferation rate of a mutant cell to that of a wild type cell at position i in the crypt column , . A relative fitness greater than 1 indicates that a mutant cell as position i is more likely to be selected to undergo cell division than a wild type cell at the same position . A relative fitness value of less than 1 represents a fitness disadvantage and thus decreases the probability of cell divisions at that position . A relative fitness value of 1 signifies a neutral mutation . We investigated the effects of relative fitness values between 0 . 5 ( representing a 50% fitness disadvantage of a mutant cell ) and 2 . 0 ( representing a 100% fitness advantage of a mutant cell ) [46] .
Using this spatially arranged stochastic process model , we first investigated the effects of different proliferation curves on cell movement . Five different proliferation curves were selected to illustrate their effects on cell movement ( Fig . 1C . ) . Curve 1 represents the measured labeling indices ( LI ) from in vivo experiments [23] , [42]; curve 2 was selected to investigate the effects of spreading proliferating activities upwards in a crypt column; curve 3 represents the uniform kinetics in which cell proliferation is equally likely to occur in any cell along the crypt column; and curves 4 and 5 represent the mirror images of curves 2 and 1 , respectively . The selection probabilities for position 1 in all curves are identical; furthermore the area under each curve is normalized to the same total in order to make the effects of each curve comparable . The rates of cellular movement , measured by the number of mitoses required for a cell at a particular position in the column to reach the top of the crypt ( position 80 ) , depends on the shape of the proliferation curve ( Fig . 2A Left . ) . As expected for all kinetic curves investigated , cells located near the bottom of the crypt require a larger number of cell divisions to reach the top . For kinetic curves with mitotic activities concentrated near the base of the crypt column ( curves 1 and 2 ) , substantially fewer cell divisions are required to push out a cell located in the bottom half of the column . The numbers of cell divisions required to accomplish this task are very similar for curves 1 and 2 . The uniform curve ( curve 3 ) requires slightly more rounds of cell divisions ( Fig . 2A Right . ) . In contrast , curves 4 and 5 , which have mitotic activities concentrated near the top of crypt column , require a much larger number of cell divisions than the other curves . In addition , we observed significant variations among individual runs of the stochastic simulation in the number of cell divisions needed to push a cell out of the crypt when considering the five different curves . At each position in the crypt , curves 4 and 5 show larger variations among simulation runs than curves 1 , 2 , and 3 . Interestingly , cells proliferating according to kinetic curves 4 and 5 show a stronger positional dependency in the amount of variation as compared to curves 1 , 2 and 3 . For instance , for curves 4 and 5 , the number of cell divisions needed to push out a cell located near the top of the crypt column is less variable than that for a cell located near the bottom ( Fig . 2A . ) . In addition to the qualitative descriptions of cellular movement for different growth kinetic curves , we also derived a mathematical representation of the cellular behavior . The dynamics of cellular movement in the crypt column can be represented by a Markov chain . After each round of cell shedding , the transition probability for a cell in position m to remain in position m is given by , and transition probability from position m to m+1 is given by for m ranging from 1… , zero elsewhere . The absorbing state for this transition matrix represents the event of a cell exiting the crypt column by reaching its top . The expected number of cell divisions needed for a cell to exit the crypt column ( E ) and the variance of the number of cell divisions ( V ) are given as follows [47]: ( 1 ) where is the identity matrix of size , is the transition matrix , and is a vector of length and ( 2 ) where denotes the Hadamard product of the expected number of cell divisions [48] . The concordance between simulation results and results from the Markov chain is shown in Fig . S1 . The presence of cell death results in changes in the transition matrix . The new transition matrix is then given by products of the transition matrix for normal shedding , and the transition matrix for replenishing cell divisions , . The transition probabilities in the matrix are all zero , except for , and , which are given by:where for k≤j+δ and k≠j , zero otherwise; this quantity denotes the probability of apoptosis occurring at position j and a replacement cell division occurring at position k; and I denotes an indicator variable . The overall transition probability matrix is then given by ( 4 ) This matrix contains the average transition probabilities in the linear system for a given death rate λ . The expected number of cell division needed for a cell to exit the crypt column ( E ) and the variance for the number of cell divisions ( V ) can be calculated from equations 1 and 2 . The concordance between the simulation and analytical results calculated using equations 1 and 2 are shown in Fig . S2 . The most important effect of cell death is its ability to accelerate cellular movement in the crypt column ( Fig . 2B . ) . As the mean number of deaths , λ , per cell division increases , fewer rounds of cell division are required to push a cell out of the crypt column ( Fig . 2B . ) . These acceleration effects on cellular movement were observed even when λ was as low as 0 . 1 and were more prominent in crypts proliferating according to curves 4 and 5 than those proliferating according to the other curves . Regression analysis indicated that the median number of cell divisions needed to push a cell at position 2 out of the crypt was not significantly different among crypt columns proliferating according to curves 1 , 2 and 3 . Also , interactions between birth and death curves were not significantly correlated with the speed of cellular movement ( see Table S1 ) . In addition to causing this acceleration effect , cell deaths also increase the mitotic burden on the stem cell ( Fig . 2C . ) . The mitotic burden of a cell is defined as the mean number of times a cell is selected to undergo cell divisions per time step . These divisions could either be cell divisions occurring under normal circumstances or replenishing replications that replace dead cells . Note that , in the absence of cell death ( λ = 0 ) , the average number of times the stem cell is selected to divide is identical for all curves . However , in the presence of cell death , λ>0 , the number of times of selecting the stem cell to undergo replenishing cell divisions depends on the shapes of the proliferation kinetic and death curves as well as the death rate . Under the uniform death curve , when λ>0 , proliferative curves 1 and 2 show the smallest increases in the mitotic burden of the stem cell , due to the positioning of the proliferating cells close to the stem cell . These proliferating cells may undergo replenishing cell divisions to replace dead cells arising further up in the crypt column , which help lessen the mitotic burden on the stem cell . In contrast , the stem cell's mitotic burden increases drastically for proliferative curves 4 and 5 due to the long distance between the stem cell and the other proliferating cells . In these cases , cell deaths occurring in the middle of the crypt column can only be replaced by stem cell divisions due to the shape of the mitotic curves and the replenishing rules imposed; this effect thus increases the mitotic burden on the stem cell . The mitotic burden for other cells is less affected by cell death as compared to that of the stem cell and the overall shape of the kinetic curves remains similar to the original shape in the absence of cell death ( see Fig . S3 . ) . One important consequence of over-using the stem cell to replace dead cells is that it accelerates the rate of mutation accumulation in the stem cell , as explored in the following section . We then incorporated a simple mutation model into the linear process . This addition captures for instance the accumulation of an inactivating mutation in one APC allele , thus transforming an APC+/+ cell into an APC+/− cell ( Fig . 3A . ) . As a representative example , the dynamics of the somatic evolution process within a single crypt column is displayed in Fig . 3B . In the absence of cell death , all mutations arising in non-stem cells within the crypt column are eventually flushed out of the crypt; only mutations arising in the stem cell have the ability to reach fixation , i . e . reach 100% frequency within a crypt column . This spatial restriction highlights the fundamental difference between the linear process and a stochastic process model describing a well-mixed population , for instance the Moran process [28] or the Wright-Fisher process . In the latter models , any cell within the population has an equal chance of taking over the population if all fitness values ( i . e . growth and death rates ) are equal . In the linear process , in contrast , only a mutation arising in the stem cell has the ability to take over the entire population of cells . Somatic evolution in the crypt column in the absence of cell death can be split into two disjoint events: ( i ) a mutation arising in the stem cell , and ( ii ) mutation propagation through the crypt column ( Fig . 3B . ) . Since the mutation rate at which APC is inactivated per allele , u0 , is low ( estimated to be on the order of 10−7 per cell division [49] ) , the rate-limiting event of this process is the time it takes until the mutation arises in the stem cell; mutations arising in any other cell are neglected here since they cannot reach fixation in the crypt . The number of times the stem cell divides is specified by the proliferation kinetic curve . In the absence of cell death , the probability of selecting the stem cell for cell division is identical for all kinetic curves; thus the time , measured by the number of mitoses occurring , until all cells in the crypt column are in the APC+/− state is identical for all kinetic curves ( Fig . 3C . ) . The distribution of mitotic activities along the crypt column plays no role in modifying the speed of mutation accumulation in the stem cell , since the rate of stem cell divisions are equal among all curves investigated ( see Methods ) . The probability of fixation by time t in the linear process is determined by the probability that a mutation has arisen in the stem cell by time t , ( 5 ) For a small u0 , this expression can be approximated by ( 6 ) Here denotes the probability of selecting the stem cell position for replication and the factor of ½ arises from the two possible arrangements of cells after cell divisions , only one of which can lead to a mutant cell residing in the stem cell position . Once the stem cell becomes mutated , fixation of its offspring in the crypt column quickly ensues . However , the rate of propagation of mutated cells throughout the crypt is heavily influenced by the kinetic curves ( Fig . 3D . ) . Mutated cells in crypts proliferating according to curves 1 and 2 reach fixation faster than in those proliferating according to curves 4 and 5 . Curve 3 , the uniform curve , leads to fixation of mutant cells on a slightly slower time scale than curves 1 and 2 , but still faster than curves 4 and 5 . In addition , the amount of variability in the number of cell divisions required for the APC+/− mutation to propagate through the crypt column also depends on the shapes of the kinetic curves . As the mitotic activity becomes more concentrated near the top of the crypt , towards the gut lumen , the amount of variability among individual simulation runs in the fixation time increases ( Fig . 3D . ) . In contrast , in the presence of non-stem cell death , the interactions between birth and death curves determine the speed of somatic evolution ( Fig . 3E . ) . We derived an analytical approximation to show the accelerating effects of non-stem cell death on the rate of mutation accumulation in the stem cell , in the absence of relative fitness differences between mutated and normal cells and in the absence of stem cell death . The probability of selecting the stem cell to undergo one round of additional cell division , and thus to shift all downstream cells along the crypt column to fill the vacancy created by a cell death at position j , is given by , where δ specifies the size of the interval ( i . e . number of cells ) in which more differentiated cells can replace the dead cell . Thus , the probability of a mutation arising in the stem cell during this additional round of cell division is given by . Since the expected number of cell deaths occurring during a time interval t is , the probability that a mutation occurs in the stem cell during additional cell divisions is given by ( 7 ) where indicates the position at which the rth cell death event occurs . For small , this expression can be approximated by . If >> , it can further be approximated by , where denotes the probability of cell death occurring at a particular position in the crypt column , and . In addition to division-linked mutations , cell division-independent mutation also contributes to the rate of mutation accumulation in the stem cell . Given the assumption that cell division-independent mutation may occur anywhere in the crypt column , the additional contribution to the overall rate of mutation accumulation in this scenario can be modeled by including the factor . Thus the overall fixation rate in the crypt column of a new mutant cell is given by ( 8 ) The interactions between birth kinetics and death positions determine the probability of selecting the stem cell for replenishing cell divisions after apoptosis , as stated by the term in equation ( 8 ) . A comparison between the analytical approximation using equation ( 8 ) and simulations is shown in Fig . S4 . For the five curves examined , the interactions between birth kinetics and death selection are shown in Table 2 . Curve 1 has the lowest probability of selecting the stem cell for replenishing cell divisions for all death curves . In addition , the rates of selecting the stem cell are less variable for curve 1 across different death selection curves ( row 1 in Table 2 ) than for other proliferation kinetic curves . Deaths occurring near the top of the crypt column , such as in death curve 5 , lower the mitotic burden on the stem cell for all birth kinetics curves . Of all twenty-five combinations examined , the interaction between birth kinetics curve 1 and death selection curve 5 results in the lowest amount of mitotic burden on the stem cell . In contrast , birth kinetics curve 5 , with mitotic probabilities concentrated near the top of the crypt , and death selection curve 1 , with apoptosis occurring near the stem cell , result in substantial increases in the probability of selecting the stem cell for replenishing cell divisions . In addition to the interaction between the location of proliferating cells and cells undergoing apoptosis , the rate of cell death , λ , also controls the rate of mutation accumulation in the stem cell . As λ increases , the rate of mutation accumulation in the stem cell is accelerated . Lastly , cell division-independent mutation increases the overall mutation rate , as shown by the last term in equation ( 8 ) . We also investigated the effects of fitness changes of mutant cells on the speed of somatic evolution . Relative fitness is defined as the ratio of selection probabilities for proliferation of a mutant cell to that of a wild type cell at the same position in the crypt column ( see Methods ) . Changes in the relative fitness can potentially affect the speed of both the rate of mutation accumulation in the stem cell and the rate of mutant propagation through the crypt column . We investigated the effects of relative fitness values between 0 . 5 ( representing a 50% fitness disadvantage of a mutant cell ) and 2 . 0 ( representing a 100% fitness advantage of a mutant cell ) [46] . The speed of somatic evolution was not significantly affected by changes in fitness within this range ( Fig . 3F . ) . This effect arises because the rate-limiting event is represented by the generation of the first mutation in the stem cell , which – unlike its propagation throughout the crypt column – is not dependent on fitness . In the case of an extremely advantageous mutation with a relative fitness value greater than 10 , fixation of its offspring in the crypt column can be substantially delayed for the cases in which the initial mutation arises in a non-stem cell . In these cases , cell divisions in non-stem mutants are driving the cellular movement and tissue regeneration in the crypt column , thus reducing the probability of the stem cell undergoing cell divisions . When conditioning on the event that the stem cell is already mutated , we found that the relative fitness has a significant effect on the fixation time for some kinetic curves ( Fig . 3G . ) . For instance , curves 4 and 5 are more sensitive to changes in relative fitness; as the relative fitness of mutants decreases , more cell divisions are required for a mutant cell to reach fixation . In contrast , curves 1 and 2 are less sensitive to changes in the relative fitness . The variation in the fixation time also depends on the kinetic curves and on the relative fitness of the mutant cells . Again , curves 1 and 2 lead to less variation compared to curves 4 and 5 . Furthermore , at low relative fitness values , the number of cell divisions necessary for fixation is more variable than at high relative fitness values for curves 4 and 5 . The uniform curve is more sensitive to fitness variations compared to curves 1 and 2 ( Fig . 3G . ) . We then modified the basic mathematical framework to allow for the special case of stem cell death . In such situations , a mutant stem cell may be replaced by a more differentiated cell . Apoptosis of the stem cell delays the rate of mutation accumulation in the stem cell and mutant fixation . Regression analysis indicates that the number of times a mutant stem cell is replaced by a wild type cell as the result of apoptosis in the stem cell position is significantly related to the birth kinetics , death rate , relative fitness , and specific interactions between birth and death curves ( see Table S2 ) . Kinetics curves with mitotic probabilities concentrated near the top of the crypt ( curves 3 , 4 and 5 ) are less likely to lead to a loss of a mutation that had arisen in the stem cell position . Mutant stem cells with a low relative fitness are more likely to be replaced by wild type cells , since such mutant cells have a slower rate of propagation . Furthermore , large death rates increase the mean number of times a mutant stem cell is replaced by a wild type cell through apoptosis and dedifferentiation . Since the death selection function is normalized such that the stem position has an equal probability of being selected for all curves , death curves have no effect on the rate at which mutant stem cells are replaced by wild type cells . The interactions between birth and death curves only have weak effects on how often the mutant stem cell loses a mutation . Overall , these observations suggest that apoptosis in the stem cell has the ability to delay mutant fixation , whereas apoptosis in the differentiated cells accelerates mutation accumulation by increasing stem cell divisions . Consistent with this observation , recent studies suggest that a high rate of stem cell apoptosis in the small intestine is partially responsible for the low incidence of small intestine cancer compared to colorectal cancer; in the latter , the stem cell is protected from undergoing apoptosis by the expression of bcl-2 , an anti-apoptosis protein [50] . We then investigated a two mutation model within the linear process . This scenario captures , for instance , the mutations inactivating both alleles of the APC gene ( Fig . 4A . ) . Similarly to the single mutation model , the rate of emergence of double mutant cells ( i . e . APC−/− cells ) is driven by the rates per cell division at which the two mutations arise , for all growth kinetic curves investigated . We first studied the dynamics of the accumulation of these two mutations by investigating the effects of proliferation kinetics on the probability that an APC+/− cell acquires an additional mutation in the second APC allele before being flushed out of the crypt . We investigated these dynamics in the absence of cell death . Fig . 4B . shows a representative simulation run . There are two scenarios: the first APC mutation may arise in the stem cell , or alternatively it could arise in a non-stem cell . Once the stem cell harbors the first APC mutation , in the absence of cell death , this mutation is permanently maintained in the crypt and eventually , the second APC mutation arises . In contrast , if the first APC mutation arises in a non-stem cell , the probability that this cell and its progeny gain an additional mutation before exiting the crypt depends on four factors: the mutation rate for inactivating the second APC allele , u1 , the location of the APC+/− cell , the kinetic curve , and the relative fitness of the APC+/− cell . As expected , an increase in u1 enhances the probability of accumulating an additional mutation in a non-stem mutant cell ( Fig . 4C–F . ) . This observation agrees with the findings by Komarova and Wang [27] that for a large u1 , APC+/− cells are likely to arise among differentiated cells . The position of the APC+/− cell also has a significant influence on the probability of accumulating the additional mutation . An APC+/− cell residing near the stem cell has a higher probability of acquiring a second mutation before being “washed” out of a crypt than a cell residing near the top of crypt ( Fig . 4C–F . ) . Finally , different birth kinetic curves also have effects on the probability of acquiring further mutations . For an APC+/− cell at a particular position between 3 and 80 , curves with mitotic activities concentrated near the stem cell ( curves 1 , 2 and 3 ) have small probabilities of acquiring the second APC mutation before the cell exits the lattice ( Fig . 4C–F . ) . This effect is more prominent for a large than for a small u1 . The probability of accumulating the second mutation depends on the sum of mitotic probability weights for each cell: w1 +…+wi-1 , where i denotes the position of the first APC+/− cell; this sum is inversely related to the probability of accumulating the second APC mutation . For instance , if the APC+/− cell resides at position 8 , then the curve with the smallest sum of mitotic probability weights for positions 1 to 7 leads to the largest chance that this clone accumulates an additional mutation before exiting the lattice . Since all curves have the same mitotic probability for position 1 , the chance that an APC+/− cell at position 2 accumulates a new mutation is the same for all kinetic curves . This effect arises only in the absence of cell death . Therefore , conditional to the event that an APC+/− cell resides between positions 3 and 80 , those kinetic curves with mitotic probability weights concentrated near the stem cell confer a protective effect against acquiring new mutations . Increasing the relative fitness of mutant cells increases the likelihood that an APC+/− cell acquires a new mutation by enhancing the likelihood of this cell and its progeny to be selected for cell divisions . These considerations hold under the assumption of no cell death . The presence of cell death , in contrast , reduces the probability for a mutant cell to gain an additional mutation; increasing the rate of cell death enhances the likelihood for a mutant cell and its progeny to either die or be flushed out the crypt before they accumulate a new mutation . Finally , we incorporated the effects of chromosomal instability ( CIN ) into our model . CIN arises due to the accumulation of a specific mutation at rate u2 and leads to a large mutation rate , u3 , at which the second APC allele is inactivated during cell divisions [15] ( Fig . 5A . ) . Considering this additional mutation event , we then set out to investigate the effects of CIN on the dynamics of the system . For large u3 , we observed a phenomenon that has previously been termed “stochastic tunneling” [51] . Tunneling refers to the process in which a crypt column moves from a homogeneous state in which all cells harbor ρ mutations to a homogeneous state in which all cells harbor ρ+2 mutations , without ever transiting through a state in which all cells harbor ρ+1 mutations . For instance , in our model , tunneling occurs when cells in the crypt column move from an APC+/− state directly to an APC−/−CIN state , without reaching fixation in the APC+/− CIN state , as illustrated in Fig . 5B . Another possible tunneling scenario is APC+/+ CIN to APC+/− CIN to APC−/− CIN . Tunneling between other states is less likely to occur given the small second mutation rates . Unlike prior investigations concerning the tunneling rate for a well-mixed population of cells [52] , cells in this model remain constrained to a one-dimensional lattice . The tunneling rate depends on the mutation rate , u3 , the death rate , the birth kinetic curve , the death selection curve , and the relative fitness of mutated cells ( see Table S3 ) . Consider the uniform death curve ( curve 3 ) as an illustration: for this curve , as u3 increases , a larger proportion of all simulation runs display tunneling for all five proliferation kinetic curves ( Fig . 5C . ) . Even though all curves lead to this phenomenon , we found that crypts proliferating according to curves 4 and 5 are more likely to display tunneling than those proliferating according to curves 1 , 2 and 3 . For curves 4 and 5 , more stem cell divisions are required to reach fixation of APC+/− CIN cells , starting from an APC+/− CIN stem cell , than for the other curves . Because of the large number of stem cell divisions in crypts proliferating according to curves 4 and 5 , there is a large probability for the stem cell to lose the second copy of the APC gene before all cells in the column become APC+/− CIN cells . For a given u3 , as the mean number of cell deaths , λ , increases , a larger proportion of simulation runs reach fixation of APC−/−CIN cells via tunneling . Furthermore , we found that increases in λ result in larger increases in the tunneling probability for cells proliferating according to curves 4 and 5 than for those proliferating according to curves 1 , 2 and 3 . Finally , in the presence of cell death , high relative fitness values of APC+/− CIN cells reduce the tunneling probability , whereas high relative fitness values of APC−/− CIN cells promote tunneling . Since the relative fitness values of APC+/− CIN and APC−/− CIN cells are correlated , the ratio of relative fitness values of APC+/− CIN and APC−/− CIN cells determines the tunneling rate . This ratio affects the selection of replacing cell divisions and hence the tunneling probability . Higher-order interaction terms involving birth curves , death curves and death rate are also important determinants of tunneling rates ( see Table S3 ) . Finally , we derived a general solution for the tunneling probability under the assumption of no cell death . The general solution comprises the product of three terms: ( 1 ) the limiting probability of the initial APC−/− CIN cell at position k to reach either of the absorbing states: that an APC−/− CIN stem cell arises , leading to tunneling , or that no further mutation arises in the stem cell position before the APC−/− CIN clone is removed from the crypt column , signifying the fixation of APC+/− CIN cells; ( 2 ) the probability that the initial APC−/− CIN cell arises at position k conditional to the event that an APC+/− CIN to APC−/− CIN mutation occurred while APC+/− CIN cells populating the crypt column reach position l; ( 3 ) the probability that an APC+/− CIN to APC−/− CIN mutation occurs as APC+/− CIN cells populating the crypt column reach position l . The transition matrix for calculating the limiting probabilities is given by ( 9 ) The transition matrix specifies the movement of the left-most APC−/− CIN cell in the crypt column . Under the assumption of no cell death , the movement of the left-most APC−/− CIN cell completely determines the tunneling probabilities . The two absorbing states are state 1 , when the left-most APC−/− CIN cell arises in the stem cell position , resulting in tunneling , and state +1 , when the left-most APC−/− CIN cell is out of the crypt column , resulting in fixation of APC+/− CIN cells . The tunneling probability vector of length −1 can be calculated using the fundamental matrix of an absorbing Markov chain [47] , [48] , ( 10 ) Thus , the overall tunneling vector of length , denoted by [P ( T|k ) ] , is given by ( 11 ) This expression arises since tunneling occurs with probability 1 if the APC+/− CIN to APC−/− CIN mutation arises in the stem cell . Under the assumption of no cell death , the mutation conferring CIN must occur in the stem cell position , since otherwise , it would be flushed out of the crypt column . Thus , the probability for the initial APC−/−CIN cell to arise at position k depends on the position of the right-most APC+/−CIN cell as the latter populate the crypt column: ( 12 ) Thus , the probability that the initial APC−/−CIN cell occurs at position k conditional to the event that the right-most APC+/−CIN cell is located at position l and a mutation occurs can be arranged in matrix form , denoted by [P ( k|l ) ] , with columns representing k and rows representing l . Lastly , the probability that an APC+/− CIN to APC−/− CIN mutation occurs as the right-most APC+/−CI cell moves into position l follows a geometric distribution: ( 13 ) Thus , a vector of length , denoted by [P ( l ) ] , completely specifies the probabilities of a mutation occurring as the right-most APC+/−CIN cell moves from l-1 position to l . Therefore , the overall tunneling probability can be calculated as ( 14 ) We then compared this analytical result with exact computer simulations and found good agreement ( Fig . 5D . ) . Furthermore , using this analytical result , we calculated the tunneling rates for the different proliferation curves at various values of the APC+/− CIN to APC−/− CIN mutation rate u3 ( Fig . 5E . ) . For a given u3 , those proliferation curves with mitotic probabilities concentrated near the stem cell have lower tunneling probabilities than those with mitotic probabilities near the top of the crypt . In addition , in the absence of cell death , a variation in fitness values does not affect the tunneling rates ( Fig . S5 . ) ; in contrast , fitness variation is significantly associated with the tunneling rate when cell death is present ( Table S3 ) .
We designed a spatially arranged computational model of intestinal epithelial cells to investigate the effects of proliferation kinetics on the dynamics of cell movement and mutation accumulation in the colonic crypt . The model considers a single cell column within a colonic crypt , in which cells are arranged onto a one-dimensional lattice . One end of the lattice represents the bottom of the crypt , where the stem cell resides , and the other end represents the orifice of the crypt ( Fig . 1A . ) . Mitotic activities cause cell movements towards the upper end of the crypt and push the last cell off the lattice . During each cell division , a mutation may occur , which might increase the proliferative activity of the resulting mutant cell and can represent a step towards colorectal tumorigenesis . We then compared the effects of five different proliferation curves on the speed of somatic evolution and cell movement ( Fig . 1C . ) . We used these proliferative curves to demonstrate that proliferation kinetics is an important criterion that needs to be considered in modeling the dynamics of somatic evolution in spatially arranged tissues . In addition to the proliferation kinetics , we introduced a differentiation hierarchy to this linear process model . In our model , only cells at similar or less differentiated stages can replace dead cells . Finally , we used a discrete time scale such that under normal circumstances , only cell divisions contribute to the measurement of time while cell death and replenishing cell division events are assumed to be instantaneous in time . Compared to previously published models [33] , [34] , [36] , [37] , in our model the crypt structure is greatly simplified to a one-dimensional lattice; nonetheless , the essential features of cell movement are captured by our simple model , in that cells move upward towards the gut lumen with limited lateral movement [22] . This simplified design allows for the investigation of the effects of proliferation kinetics on the rate of mutation accumulation . In addition , unlike previously published models , our simple design enables us to derive analytical solutions for several quantities of interest such as the rate of mutation accumulation and the tunneling probability . In contrast to compartmental models [53]–[55] , the linear model has the advantage of retaining the spatial structure dictating colonic epithelial cell behavior . In addition , our model contains a gradual differentiation hierarchy , which is represented by the cell positions in the crypt , instead of being characterized by discrete compartmental boundaries . Using this model , we demonstrated that spatially explicit proliferation kinetics have a significant impact on the stability and the dynamics of the crypt column in terms of the speed of cell movement and mutation propagation as well as sensitivities to apoptosis and selective effects of mutant cells . Comparing the proliferation kinetic curves we investigated , we identified three advantages of a spatial architecture in which the proliferative potential of cells is located close to the stem cell: 1 ) this type of proliferation architecture increases the stability of the linear system in terms of providing a less variable rate of cell movement; 2 ) in the presence of cell death , this architecture delays the rate of mutation accumulation in the stem cell; and 3 ) it provides protection against tumorigenesis by reducing the probability of acquiring further mutations in the absence of cell death . These results suggest that the kinetic curve identified using labeling index studies in the human colon [19] , [23] best delays the rate of somatic evolution towards colorectal tumorigenesis when compared to the other curves investigated here . Prior work has demonstrated that both spatial organization and cellular hierarchy need to be considered in modeling somatic evolution [56] . Our findings highlight the importance of proliferation patterns , in addition to spatial arrangements and cellular hierarchies , in studying tissue and mutation dynamics . This area has not been explored in-depth prior to our investigation . Despite the highly simplified nature of our model , we have demonstrated that proliferation curves with mitotic activities concentrated near the stem cell confer an advantage to the colon crypt by increasing the stability of the linear system and by delaying the rate of mutation accumulation . Normal proliferation kinetics , in addition to the linear tissue architecture , can suppress the rate of evolution towards colorectal cancer . A departure from such proliferation kinetics accelerates the rate of mutation accumulation in the colonic crypt and destabilizes natural cell flow , thus representing a step towards cancer . | Mathematical and computational models have a long and rich history in enhancing our understanding of intestinal epithelial cells . A plethora of models have been proposed to describe different aspects of cellular behavior , including cell proliferation , migration , differentiation , and mutation accumulation . Here , we present a novel approach to examine the effects of proliferation kinetics on the rate of somatic evolution in a spatially arranged model of the colon . Based on our simulation results , we demonstrate that spatially determined proliferation kinetics has the ability to delay the rate of somatic evolution , and changes in proliferation patterns can significantly affect the speed of mutation accumulation . Our work highlights the importance of considering proliferation kinetics as well as the spatial organization of tissues when investigating the dynamics of cancer initiation . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"oncology",
"medicine",
"computer",
"science",
"computer",
"modeling",
"mathematics",
"applied",
"mathematics"
] | 2013 | Patterns of Proliferative Activity in the Colonic Crypt Determine Crypt Stability and Rates of Somatic Evolution |
In both fission yeast and humans , the shelterin complex plays central roles in regulation of telomerase recruitment , protection of telomeres against DNA damage response factors , and formation of heterochromatin at telomeres . While shelterin is essential for limiting activation of the DNA damage checkpoint kinases ATR and ATM at telomeres , these kinases are required for stable maintenance of telomeres . In fission yeast , Rad3ATR and Tel1ATM kinases are redundantly required for telomerase recruitment , since Rad3ATR/Tel1ATM-dependent phosphorylation of the shelterin subunit Ccq1 at Thr93 promotes interaction between Ccq1 and the telomerase subunit Est1 . However , it remained unclear how protein-protein interactions within the shelterin complex ( consisting of Taz1 , Rap1 , Poz1 , Tpz1 , Pot1 and Ccq1 ) contribute to the regulation of Ccq1 Thr93 phosphorylation and telomerase recruitment . In this study , we identify domains and amino acid residues that are critical for mediating Tpz1-Ccq1 and Tpz1-Poz1 interaction within the fission yeast shelterin complex . Using separation of function Tpz1 mutants that maintain Tpz1-Pot1 interaction but specifically disrupt either Tpz1-Ccq1 or Tpz1-Poz1 interaction , we then establish that Tpz1-Ccq1 interaction promotes Ccq1 Thr93 phosphorylation , telomerase recruitment , checkpoint inhibition and telomeric heterochromatin formation . Furthermore , we demonstrate that Tpz1-Poz1 interaction promotes telomere association of Poz1 , and loss of Poz1 from telomeres leads to increases in Ccq1 Thr93 phosphorylation and telomerase recruitment , and telomeric heterochromatin formation defect . In addition , our studies establish that Tpz1-Poz1 and Tpz1-Ccq1 interactions redundantly fulfill the essential telomere protection function of the shelterin complex , since simultaneous loss of both interactions caused immediate loss of cell viability for the majority of cells and generation of survivors with circular chromosomes . Based on these findings , we suggest that the negative regulatory function of Tpz1-Poz1 interaction works upstream of Rad3ATR kinase , while Tpz1-Ccq1 interaction works downstream of Rad3ATR kinase to facilitate Ccq1 Thr93 phosphorylation and telomerase recruitment .
Telomeres are protective nucleoprotein structures found at the natural ends of eukaryotic linear chromosomes [1] . In most eukaryotes , telomeric DNA consists of GT-rich repeat sequences , primarily double-stranded but terminating in a single-stranded 3′ overhang , known as G-tail [1] . Telomerase , a specialized reverse transcriptase , solves the “end-replication problem” by de novo addition of GT-rich repeats to chromosome ends [2] , [3] . While its catalytic subunit TERT ( Trt1 in fission yeast Schizosaccharomyces pombe , Est2 in budding yeast Saccharomyces cerevisiae , and hTERT in humans ) and telomerase RNA ( TER1 in S . pombe , TLC1 in S . cerevisiae , and hTR in humans ) [4]–[7] are sufficient for generating telomerase activity that can be monitored in vitro [8] , [9] , additional regulatory subunits are necessary for telomere maintenance in vivo . For example , a regulatory subunit Est1 is required for telomere maintenance in both budding and fission yeasts , as it contributes to efficient recruitment and/or activation of telomerase in late S-phase [10]–[12] . In mammalian cells , the shelterin complex ( consisting of TRF1 , TRF2 , RAP1 , TIN2 , TPP1 and POT1 ) ensures stable maintenance of telomeres [1] . While TRF1 and TRF2 specifically recognize and bind to the double-stranded telomeric repeats , POT1 specifically binds to the G-tail [1] . TIN2 and TPP1 are essential for assembly of the shelterin complex , since these two proteins , via TRF1-TIN2 , TRF2-TIN2 , TIN2-TPP1 , and TPP-POT1 interactions , connect the double-stranded DNA ( dsDNA ) binding proteins TRF1/TRF2 to the single-stranded DNA ( ssDNA ) binding protein POT1 [13] . On the other hand , a fully assembled shelterin complex might occur only transiently during cell cycle , since TRF1 , TRF2 and POT1 show distinct cell cycle-regulated telomere binding patterns [14] , and distinct sub-complexes of shelterin components have been identified in cell extracts [15] . In addition , a careful quantitative western blot analysis indicated that protein expression levels of TPP1 and POT1 are significantly lower than TRF1 , TRF2 , RAP1 and TIN2 , suggesting that a majority of shelterin subunits may be assembled only as the TRF1-TRF2-RAP1-TIN2 sub-complex [16] . Previous studies have found that TRF1 , along with TIN2 , TPP1 and POT1 , function as negative regulators of telomerase-dependent telomere elongation [17]–[20] , but TIN2 , POT1 and TPP1 also play roles in promoting telomere extension by facilitating telomerase recruitment [21]–[25] . The shelterin complex is also important for preventing telomeres from being recognized as broken DNA ends , which can undergo chromosome rearrangements and fusions by various DNA repair proteins , and cause cell cycle arrest mediated by the DNA damage checkpoint kinases ATM and ATR [26]–[28] . Mutations in telomerase and shelterin subunits have been linked to genomic instability and human diseases , highlighting the importance of understanding how the shelterin complex regulates telomerase and DNA damage response factors [29] , [30] . Fission yeast Schizosaccharomyces pombe serves as an attractive model for understanding how cells regulate telomere maintenance , since its shelterin complex ( composed of Taz1 , Rap1 , Poz1 , Tpz1 , Pot1 and Ccq1 ) ( Figure 1A ) shares many conserved features with the mammalian shelterin complex [31] , [32] , and fission yeast cells are highly amenable to genetic and biochemical analyses . Pot1 , the ortholog of mammalian POT1 , binds directly to the G-tail and protects telomeres against chromosome fusions and Rad3ATR-dependent checkpoint activation [33] , [34] . Due to loss of telomere protection , deletions of Pot1 or the Pot1-interacting protein Tpz1 ( TPP1 ortholog ) lead to immediate cell death for the majority of cells , while rare survivor cells carrying circularized chromosomes can be recovered [31] , [33] . Poz1 , proposed to be a functional analog of TIN2 , forms a bridge between the ssDNA binding protein Pot1 and dsDNA binding protein Taz1 through its interactions with both Tpz1 and the Taz1-interacting protein Rap1 [31] . Much like in mammalian cells , different subunits of the fission yeast shelterin complex also show distinct cell cycle-regulated telomere association patterns [35] , [36] , suggesting that a commonly drawn fully connected shelterin complex ( Figure 1A ) might never be formed or formed only transiently during the cell cycle . Taz1 , Rap1 , and Poz1 are all important for the negative regulation of telomerase-dependent telomere elongation , as deletion of any of these three proteins causes telomerase-dependent massive elongation of telomeres [31] , [37]–[39] . We have previously shown that Taz1 , Rap1 and Poz1 are necessary to properly coordinate leading and lagging strand synthesis and to limit accumulation of the DNA damage checkpoint kinase Rad3ATR at telomeres [36] . Since Rad3ATR/Tel1ATM-dependent phosphorylation of Ccq1 at Thr93 is recognized by the 14-3-3-like domain of Est1 to promote Est1-Ccq1 interaction and telomerase recruitment [12] , increased accumulation of Rad3ATR in taz1Δ , rap1Δ and poz1Δ cells leads to increased Ccq1 Thr93 phosphorylation and telomerase recruitment [12] , [36] . Ccq1 interacts not only with Tpz1 and Est1 , but also with the Clr3 subunit of SHREC ( Snf2/HDAC-Containing Repressor Complex ) , a mediator of heterochromatic gene silencing , to facilitate heterochromatin formation at telomeres [31] , [40] . While yeast two-hybrid ( Y2H ) assays have detected Tpz1-Ccq1 and Ccq1-Clr3 interactions and co-immunoprecipitation ( co-IP ) assays have detected association of Pot1 with Ccq1 and association of Clr3 with Ccq1 , co-IP assays failed to detect association of Pot1 with Clr3 [31] , suggesting that shelterin and SHREC likely represent two distinct complexes with the shared subunit Ccq1 . In addition to its roles as telomerase recruiter and promoter of heterochromatin formation , Ccq1 also acts to repress checkpoint activation and recombination at telomeres [31] , [41] . However , previous studies have utilized ccq1Δ strains to examine its roles at telomeres , and thus did not address which of these previously described Ccq1 functions is dependent on the proper formation of Tpz1-Ccq1 interaction . Poz1 , besides functioning as a negative regulator of telomerase recruitment , is also required for heterochromatin formation at telomeres [32] , and disruption of Poz1-Rap1 interaction leads to massively elongated telomeres and loss of heterochromatin [42] . However , it remained unclear if Poz1-Tpz1 interaction might play any functional role in this negative regulation of telomerase recruitment or heterochromatin formation at telomeres , as previous studies have utilized only poz1Δ strains [12] , [31] . Simultaneous deletion of ccq1 and poz1 genes leads to severe telomere de-protection phenotypes reminiscent of tpz1Δ and pot1Δ cells , suggesting that Ccq1 and Poz1 play redundant roles in allowing Pot1-Tpz1 to protect telomeres against DNA repair and checkpoint proteins [31] . However , it had not been established if Tpz1-Poz1 and Tpz1-Ccq1 interactions are necessary for Poz1 and Ccq1 to protect telomeres , or the presence of these proteins per se are necessary for telomere capping . Therefore , we have decided to map the domain ( s ) of Tpz1 that are responsible for dictating Tpz1-Poz1 and Tpz1-Ccq1 interaction , and to generate a series of tpz1 mutations that are specifically defective in Tpz1-Poz1 and/or Tpz1-Ccq1 interaction while retaining Tpz1-Pot1 interaction . Using these separation-of-function mutants of Tpz1 , we will establish that Tpz1-Ccq1 interaction is essential for Ccq1 Thr93 phosphorylation and telomerase recruitment , heterochromatin formation , and checkpoint suppression at telomeres , while Tpz1-Poz1 interaction is required for efficient localization of Poz1 at telomeres , heterochromatin formation , and inhibition of Ccq1 Thr93 phosphorylation and telomerase recruitment . Furthermore , we will establish that Tpz1-Poz1 and Tpz1-Ccq1 interactions are redundantly required for telomere protection . These conclusions are in conflict with a recent study that concluded that Tpz1-Ccq1 interaction is not necessary for telomerase recruitment and Tpz1-Poz1 interaction does not contribute to Poz1 recruitment to telomeres [43] . Therefore , we will also discuss contrasting observations between our current study and the previous findings .
A previous study that discovered the fission yeast shelterin complex also identified amino acids 379–508 of Tpz1 as the domain that mediates both Tpz1-Ccq1 and Tpz1-Poz1 interactions by Y2H assay [31] ( Figure 1 ) . In an attempt to identify smaller regions sufficient for mediating either Tpz1-Ccq1 or Tpz1-Poz1 interaction , we further truncated this fragment and found that amino acid residues 421–485 are sufficient for Tpz1-Ccq1 interaction while amino acid residues 486–508 are sufficient for Tpz1-Poz1 interaction by Y2H assay ( Figure 2A ) . Analysis of Tpz1 sequences from four Schizosaccharomyces species [44] identified highly conserved amino acid residues within the minimum Tpz1-Ccq1 and Tpz1-Poz1 interaction domains ( Figures 1C and S1A ) . Therefore , we decided to mutate them to determine if these conserved residues play any role in promoting Tpz1-Ccq1 or Tpz1-Poz1 interaction . For charged residues ( Lys , Arg , Asp and Glu ) , we introduced either a charge swap ( Lys/Arg to Glu ) or Alanine mutation ( Lys/Arg/Asp/Glu to Ala ) , but failed to find any mutation that disrupt Tpz1-Poz1 or Tpz1-Ccq1 Y2H interaction ( Figure S1B ) . On the other hand , we were able to identify mutations that specifically disrupt either Tpz1-Ccq1 or Tpz1-Poz1 Y2H interaction by changing hydrophobic residues ( Tyr/Trp/Leu/Ile ) to Arginine ( Figure 2B–C ) . Within its Ccq1 interaction domain , simultaneous mutations of Tyr439 and Leu445 or a single mutation of Leu449 to Arg disrupted the Tpz1-Ccq1 Y2H interaction without affecting Tpz1-Poz1 and Tpz1-Pot1 interactions ( Figure 2B ) . Within its Poz1 interaction domain , simultaneous mutations of Trp498 and Ile501 to Arg significantly weakened the Tpz1-Poz1 Y2H interaction for full-length Tpz1 , and completely eliminated the interaction when introduced into the minimal Poz1-binding domain construct ( residues 486–508 ) ( Figure 2C ) . When mutated to Ala , tpz1-W498A , I501A only mildly affected Tpz1-Poz1 Y2H interaction , and tpz1-Y439A , I445A failed to disrupt Tpz1-Ccq1 Y2H interaction ( Figure S2B ) . While tpz1-L449A mutation also caused disruption of Tpz1-Ccq1 Y2H interaction ( Figure S2B ) , the disruption was more severe for tpz1-L449R than tpz1-L449A ( Figure S2C ) . Since we were interested in generating mutants that fully disrupt Tpz1-Ccq1 or Tpz1-Poz1 interaction , we then decided to focus our efforts on characterizing hydrophobic to Arg mutants rather than Ala mutants , except for tpz1-L449A . Based on our results from Y2H analysis , we integrated mutant tpz1 alleles into their endogenous locus in fission yeast , and characterized their effects on Tpz1-Ccq1 , Tpz1-Poz1 and Tpz1-Pot1 interactions by co-IP ( Figures 3 and S3A ) . As predicted by Y2H assays , we found that Tpz1-L449R , Tpz1-Y439R , L445R and Tpz1-Y439R , L445R , L449R mutant proteins specifically lose Tpz1-Ccq1 interaction but retain Tpz1-Poz1 and Tpz1-Pot1 interactions in fission yeast cells ( Figure 3B–D ) . In addition , co-IP analysis revealed that both L449R and L449A mutations both completely disrupt Tpz1-Ccq1 interaction even when lower salt ( 60 mM NaCl ) than our standard salt ( 150 mM NaCl ) lysis and wash conditions were used ( Figure S3A ) . Furthermore , we found that both Tpz1-W498R , I501R and the C-terminally truncated Tpz1-[1–485] ( Figure 3A ) specifically disrupt Tpz1-Poz1 interaction without affecting Tpz1-Ccq1 and Tpz1-Pot1 interactions ( Figure 3E–G ) . While all tpz1 mutant alleles showed comparable expression levels of Tpz1 protein and did not affect protein levels of Pot1 or Ccq1 ( Figure 3B–G ) , we found that Poz1 expression levels are reduced in cells expressing Tpz1-W498R , I501R or Tpz1-[1–485] , suggesting that Poz1 stability partially depends on its interaction with Tpz1 ( Figure 3E ) . Taken together , our results indicated that two distinct C-terminal regions of Tpz1 mediate its interaction with either Ccq1 or Poz1 . Identification of Tpz1 mutants that specifically disrupt either Tpz1-Ccq1 or Tpz1-Poz1 interaction allowed us to investigate the functional significance of these interactions for fission yeast telomere maintenance . For all four Tpz1-Ccq1 interaction mutants tested ( tpz1-L449R , tpz1-L449A , tpz1-Y439R , L445R and tpz1-Y439R , L445R , L449R ) , Southern blot analysis indicated that newly generated haploid tpz1 mutant cells ( derived from heterozygous diploid cells ) show progressive telomere shortening when restreaked on agar plates , very much like ccq1Δ cells ( Figure 4A lanes 6 , 7 , 9 , 10 , 12 and 13 , and Figure S3B lanes 2–9 ) [31] , [41] . However , colony sizes of tpz1 mutant cells on plates were highly variable , and when smaller colonies were selectively picked to monitor telomere structures , even early generation colonies ( restreaked twice ) showed very strong signals for a band corresponding to intra-chromosome fusion ( I+L ) , and faint signals for bands corresponding to inter-chromosome fusions ( I+M and L+M ) on a pulsed-field gel ( Figures 4B and S3C ) . Conversely , when faster growing colonies were chosen for successive restreaks on plates , they maintained short telomeres even after restreaked 11 times ( estimated to be 220–275 cell divisions , assuming 20–25 cell divisions per restreak on agar plates ) ( Figure S4 ) , suggesting that those cells have adapted to the loss of Tpz1-Ccq1 interaction and managed to stably maintain short telomeres . Furthermore , tpz1-L449R or tpz1-L449A cells derived from larger colonies also showed no telomere fusions when analyzed on a pulsed field gel ( Figure 4D lane 5 , and Figure S3E lanes 2 , 3 and 7 ) . Taken together , we thus concluded that the loss of Tpz1-Ccq1 interaction causes heterogeneous phenotypes with either immediate loss of cell viability due to elevated telomere fusions or frequent emergence of cells that manage to maintain short telomeres for many generations . In liquid culture , tpz1-L449R and ccq1Δ cells showed similar changes in cell growth rate . The slowest cell growth occurred on days 6–8 after haploid colonies derived from heterozygous diploid cells were first inoculated and subsequently serially diluted every 24 hours ( Figure S5 ) . Progressive telomere shortening of tpz1-L449R and ccq1Δ cells during days 1–7 correlated with progressive loss of growth , but both cultures eventually generated survivors that recover in growth rate and carry highly elongated and rearranged telomeres ( Figure S5C–D ) . Intensities of telomere signals for tpz1-L449R and ccq1Δ survivor cells were much more intense than for trt1Δ survivor cells ( Figure S5B–D ) , suggesting that elimination of Tpz1-Ccq1 interaction or Ccq1 protein may allow for more efficient telomere elongation by telomere-telomere recombination in survivor cells [41] , [45] . The appearance of survivor cells with highly elongated telomeres in liquid culture is reminiscent of the situation in budding yeast , where rare but faster growing budding yeast Type II recombination survivors with long telomeres predominate in liquid culture , and more common but slower growing Type I survivors with short telomeres predominate when selected on agar plates [46] , [47] . Double mutant tpz1-L449R ccq1Δ cells grew comparably to tpz1-L449R and ccq1Δ single mutant cells , and Southern blot analysis revealed that tpz1-L449R ccq1Δ double mutant cells exhibit a similar extent of telomere shortening as tpz1-L449R and ccq1Δ single mutant cells ( Figure 4C lanes 3 , 5 and 6 ) . By contrast , the majority of tpz1-L449R poz1Δ and tpz1-L449A poz1Δ double mutant cells died immediately after they were generated by dissection of spores derived from heterozygous diploid cells , and rare survivor cells had lost their telomeres ( Figures 4C and S3D ) and carried circular chromosomes ( Figures 4D and S3E ) , much like ccq1Δ poz1Δ double mutant cells . These data supported the notion that disruption of Tpz1-Ccq1 interaction mainly affects the Ccq1-dependent pathway of telomere maintenance . Much like ccq1Δ cells [41] , Tpz1-Ccq1 interaction disruption mutants immediately activated the G2 DNA damage checkpoint , based on the appearance of highly elongated cells and a slow mobility band corresponding to hyper-phosphorylated Chk1 on SDS PAGE ( Figure S6 ) . In addition , tpz1-L449R and tpz1-Y439R , L445R cells , much like ccq1Δ cells , failed to repress the his3+ gene inserted adjacent to telomere repeats , suggesting that Tpz1-Ccq1 interaction is essential for heterochromatin formation at telomere/sub-telomere regions ( Figure S7A ) [48] . Thus , disruption of Tpz1-Ccq1 interaction recapitulated all phenotypes of ccq1Δ cells we have examined , highlighting the importance of this interaction not only for telomerase regulation and telomere protection [12] , [31] , [41] , [49] but also for the SHREC-dependent role of Ccq1 in heterochromatin formation at sub-telomeres [40] . Next , we performed chromatin immunoprecipitation ( ChIP ) assays to understand how disruption of Tpz1-Ccq1 interaction affects the association of Tpz1 , Ccq1 and Trt1TERT with telomeres . We utilized early generation mutant strains , and the presence of telomeres was confirmed by Southern blot analysis ( Figure S8A–C ) . The amount of Tpz1 bound to telomeres was reduced in tpz1 mutant cells which disrupt Tpz1-Ccq1 interaction ( tpz1-L449R , tpz1-Y439R , L445R and tpz1-Y439R , L445R , L449R ) , much like in ccq1Δ cells ( Figure 5A ) . Likewise , Ccq1 association with telomeres was reduced in these mutants , but retained substantial binding ( Figure 5B ) . Furthermore , we found that tpz1-L449A and tpz1-L449R cause comparable reduction in Ccq1 binding at telomeres ( Figure S9A ) . Thus , even though Tpz1 is the only shelterin subunit known to directly interact with Ccq1 [31] , we concluded that Ccq1 could still be recruited to telomeres in the absence of detectable Tpz1-Ccq1 interaction , strongly implicating the existence of additional mechanism ( s ) that allow recruitment of Ccq1 to telomeres . For both Tpz1 and Ccq1 , the reduction in telomere binding we observed could be a consequence of having fewer telomeric repeats . On the other hand , previous studies have found that ccq1-T93A cells , despite having short telomeres , do not show reduced Ccq1 association with telomeres [12] , [50] , suggesting that Tpz1-Ccq1 does indeed contribute to the efficient accumulation of Ccq1 at telomeres . Ccq1 is required for recruitment of telomerase to telomeres [41] , [49] . Thus , we next examined the association of telomerase with telomeres by performing ChIP assays for the catalytic subunit of telomerase Trt1TERT . Despite the fact that substantial amounts of Ccq1 remained bound to telomeres , we found that Trt1TERT binding to telomeres is nearly eliminated in all Tpz1-Ccq1 interaction disruption mutants , much like in ccq1Δ cells ( Figures 5C and S9B ) . Furthermore , we found that Tpz1 fails to co-IP with telomerase RNA TER1 in the absence of Tpz1-Ccq1 interaction ( Figure S8D ) . Because Rad3ATR/Tel1ATM-dependent Ccq1 Thr93 phosphorylation is required for recruitment of telomerase to telomeres [12] , [50] , we hypothesized that the loss of telomerase recruitment may be explained by loss of Ccq1 Thr93 phosphorylation . We examined Ccq1 phosphorylation in both rap1+ and rap1Δ backgrounds , since elimination of Rap1 strongly induces Rad3ATR/Tel1ATM-dependent hyper-phosphorylation of Ccq1 at multiple sites including Thr93 [12] , allowing us to more robustly determine the effect of disrupting Tpz1-Ccq1 interaction on Ccq1 phosphorylation . Based on the appearance of a phosphatase sensitive slow mobility band on SDS PAGE , we found that disruption of Tpz1-Ccq1 interaction alone , much like trt1Δ , is sufficient to induce hyper-phosphorylation of Ccq1 , due to telomere shortening [12] ( Figure 5D bottom panel ) . Moreover , Ccq1 was still hyper-phosphorylated when Tpz1-Ccq1 interaction was disrupted in rap1Δ cells ( Figure 5E bottom panel ) . By contrast , disruption of Tpz1-Ccq1 interaction completely eliminated Ccq1 Thr93 phosphorylation in both rap1+ and rap1Δ backgrounds ( Figure 5D–E top panels ) . Taken together , we thus concluded that Tpz1-Ccq1 interaction plays an essential role in telomerase recruitment by facilitating Rad3ATR/Tel1ATM-dependent phosphorylation of Ccq1 Thr93 . Furthermore , our data indicated that Ccq1 Thr93 phosphorylation is differentially regulated from phosphorylation of other Ccq1 sites and much more dependent on Tpz1-Ccq1 interaction . When various truncation mutants of Tpz1 , which all expressed well in fission yeast based on western blot analysis ( Figure S10A–B ) , were tested for their effects on telomere maintenance , we found that deletion of the internal Tpz1-Ccq1 interaction domain alone ( tpz1-[Δ421–485] ) or deletion of both Tpz1-Ccq1 and Tpz1-Poz1 interaction domains ( tpz1-[1–420] ) result in immediate telomere loss and chromosome circularization ( Figure S10C–D ) . By contrast , deletion of the Tpz1-Poz1 interaction domain alone ( tpz1-[1–485] ) allowed cells to maintain highly elongated telomeres , much like in poz1Δ cells ( Figure 6A lanes 7 and 8 , and Figure S10C lane 6 ) . Tpz1 point mutations that disrupted Tpz1-Poz1 interaction ( tpz1-W498R , I501R ) ( Figure 3E ) likewise caused telomere elongation comparable to poz1Δ , and telomeres did not show any further elongation in tpz1-W498R , I501R poz1Δ cells ( Figure 6A lanes 7 , 9 and 10 ) . Furthermore , tpz1-W498R , I501R ccq1Δ cells immediately lost telomeres , as soon as they were germinated from spores derived from heterozygous diploid ( tpz1+/tpz1-W498R , I501R ccq1+/ccq1Δ ) cells , and survived by circularizing their chromosomes , very much like in ccq1Δ poz1Δ cells ( Figure 6A lanes 11 and 12 , and Figure 6B lanes 4 and 5 ) . We also observed that cells carrying tpz1 mutants that incorporate disruption mutations for both Tpz1-Ccq1 and Tpz1-Poz1 interactions ( tpz1-[1–485]-L449R and tpz1-L449R , W498R , I501R ) fail to protect telomeres against fusions , immediately lose viability for the majority of cells , and exclusively generate survivors with circular chromosomes ( Figure 6C lanes 5 and 7 , and Figure 6D lanes 3 and 5 ) . Taken together , we thus concluded that telomere length deregulation caused by disrupting Tpz1-Poz1 interaction specifically inactivates Poz1's ability to prevent uncontrolled telomere elongation . Furthermore , we concluded that Tpz1-Poz1 and Tpz1-Ccq1 interactions redundantly provide essential telomere protection functions of Tpz1 [31] . While it remains to be established why Ccq1 and Poz1 are redundantly required to prevent telomere fusions , we can rule out the possibility that they are redundantly required for assembly of the Tpz1-Pot1 complex , since Tpz1-Pot1 interaction detected by co-IP remain intact in ccq1Δ poz1Δ cells ( Figure S11D ) . Furthermore , in a Pot1-dependent in vitro pull down assay for Tpz1 utilizing a magnetic-bead coupled telomeric G-tail oligo , wild-type Tpz1 could still be detected in ccq1Δ poz1Δ cells , and both Tpz1-[1–485]-L449R and Tpz1-L449R , W498R , I501R mutant proteins , which interact with neither Ccq1 nor Poz1 , were also detected ( Figure S11A–C ) . Disruption of Tpz1-Poz1 interaction also allowed expression of the his3+ gene inserted adjacent to telomere repeats ( Figure S7B ) , much like poz1Δ cells [49] , suggesting that heterochromatin formation at telomeres also requires Tpz1-Poz1 interaction . However , both tpz1-W498 , I501R and tpz1-[1–485] cells grew slower than poz1Δ cells on selective media lacking histidine , suggesting that Poz1 , even in the absence of Tpz1-Poz1 interaction , weakly contributes to the formation of heterochromatin at telomeres . In order to gain insight into how the disruption of Tpz1-Poz1 interaction affects the association of shelterin subunits and telomerase with telomeres , we next carried out ChIP assays for Tpz1 , Ccq1 , Poz1 and Trt1TERT . It was necessary to utilize dot blot-based ChIP assays , rather than quantitative real-time PCR-based ChIP assays , since tpz1-W498R , I501R caused massive elongation of telomere repeats ( Figures 6A and S12 ) and thus putting the sub-telomeric annealing sites for our PCR primers too far away from actual telomeric ends [12] , [36] . By quantifying hybridization intensities of precipitated and input DNA to a telomeric repeat DNA probe , we first established % DNA that was precipitated by ChIP relative to input ( raw % precipitated DNA ) ( Figure S13 ) . Changes in raw % precipitated DNA values more closely reflect changes in density of a given protein on telomeric repeats , rather than changes in total amount of protein associated per chromosome end . Thus , it became necessary to correct raw % precipitated DNA values for telomere length to better represent changes in amount of protein bound per chromosome end , especially for cells carrying highly elongated telomeres . To account for changes in telomere length , we first established “telomere length correction factors” for individual strains by measuring changes in telomere/rDNA hybridization intensity ratios compared to wild-type cells ( Table S1 ) [36] . We then established “telomere length corrected” ChIP values by multiplying background subtracted % precipitated DNA values ( raw % precipitated DNA from epitope tagged strain – no tag control % precipitated DNA ) with the telomere length correction factors , and normalizing them to wild-type ChIP values ( plotted as “relative ChIP signal” ) [36] . Although not perfect , this adjustment for variations in telomere length allowed us to better estimate changes in amount of protein localized per chromosome end . Analysis of ChIP data revealed that tpz1-W498R , I501R , poz1Δ and tpz1-W498R , I501R poz1Δ cells show comparable increases in amount of Tpz1 and Ccq1 per chromosome end over wild-type cells when corrected for telomere elongation in these mutant cells ( Figure 7A–B ) . Since single and double mutants for tpz1-W498R , I501R and poz1Δ showed comparable changes in Tpz1 and Ccq1 association with telomeres , these ChIP data further confirmed that the loss of Tpz1-Poz1 interaction solely disrupts Poz1 function at telomeres . Further analysis of Poz1 ChIP data indicated that Tpz1-Poz1 interaction is crucial for efficient accumulation of Poz1 at telomeres , as tpz1-W498R , I501R or tpz1-W498R , I501R rap1Δ cells showed a substantial reduction in Poz1 association with telomeres compared to wild-type cells ( Figure 7C ) . By contrast , rap1Δ cells showed an increase in Poz1 association ( Figure 7C ) [36] . We have previously shown that rap1Δ also causes a comparable increase in telomere association for Tpz1 and Ccq1 [36] . Taken together , we concluded that Poz1 association with telomeres is primarily facilitated by Tpz1-Poz1 interaction , and that Poz1-Rap1 interaction does not play a significant role in association of Poz1 with telomeres . On the other hand , it should be noted that Poz1 association , although significantly decreased , is not completely eliminated even in tpz1-W498R , I501R rap1Δ cells ( Figures 7C and S13C ) . As noted earlier , we found that the presence of Poz1 protein appears to contribute weakly to the transcriptional repression of the his3+ marker in tpz1-W498R , I501R cells ( Figure S7B ) . Therefore , residual Rap1- and Tpz1-independent association of Poz1 with telomeres may also be functionally important . Alternatively , since Tpz1-W498R , I501R protein still showed residual interaction with Poz1 in Y2H assay ( Figure 2C ) , it might also retain a residual weak interaction in vivo ( not detected by co-IP ) that is responsible for its residual localization to telomeres . Since introduction of tpz1-W498R , I501R or tpz1-[1–485] caused telomere extensions comparable to poz1Δ ( Figure 6A ) , we expected that loss of Tpz1-Poz1 interaction would cause increases in both telomerase association with telomeres and Ccq1 Thr93 phosphorylation , as previously established for poz1Δ cells [12] , [36] . Indeed , ChIP assays for the telomerase catalytic subunit Trt1TERT revealed that tpz1-W498R , I501R causes a comparable increase in Trt1TERT binding to telomeres as poz1Δ cells ( Figures 7D and S13D ) . In addition , we found that both tpz1-W498R , I501R and tpz1-[1–485] mutations cause hyper-phosphorylation at Thr93 and other sites of Ccq1 ( Figure 7E ) . Thus , we concluded that Tpz1-Poz1 interaction-dependent recruitment of Poz1 is essential for enforcing a negative regulation on Ccq1 Thr93 phosphorylation-dependent recruitment of telomerase .
In this study , we determined amino acid residues within two distinct C-terminal domains of Tpz1 that are responsible for mediating either Tpz1-Ccq1 or Tpz1-Poz1 interaction , and characterized how these interactions individually or in combination affect the ability of the shelterin complex to ensure telomere maintenance and protection in fission yeast . ( Key findings are summarized in Figure 8 ) . Our results indicated that disruption of Tpz1-Ccq1 interaction causes telomere phenotypes that are essentially identical to those of ccq1Δ cells ( Figures 4 , S3 , and S5 ) . Cells lacking Tpz1-Ccq1 interaction fail to efficiently recruit telomerase to telomeres , due to loss of Rad3ATR/Tel1ATM kinase-dependent Ccq1 Thr93 phosphorylation ( Figure 5C–E ) , which is essential for promoting Est1-Ccq1 interaction and telomerase recruitment [12] ( Figure 8 ) . Although Ccq1 association with telomeres was reduced , significant amounts were still detectable in the absence of Tpz1-Ccq1 interaction ( Figure 5B ) , implicating the existence of an alternative mechanism that allows recruitment of Ccq1 to telomeres . Ccq1 also interacts with the SHREC complex that facilitates heterochromatin formation at telomeres [40] and heterochromatin-dependent recruitment of Ccq1 has been proposed as a mechanism to allow recruitment of Pot1 to protect chromosome ends in HATTI survivor cells that lack telomere repeats at chromosome ends [51] . Thus , it is possible that the SHREC complex is responsible for allowing Ccq1 localization at telomeres in the absence of Tpz1-Ccq1 interaction . On the other hand , we cannot completely rule out the possibility that a weak residual Tpz1-Ccq1 interaction , as detected under less stringent ( −His ) Y2H assay conditions ( Figure S2C ) but not by co-IP experiments ( Figure 3C ) , may still contribute to Ccq1 binding at telomeres . In any case , our results established that localization of Ccq1 at telomeres alone is not sufficient for telomerase recruitment to telomeres , due to the fact that Tpz1-Ccq1 interaction is essential for Rad3ATR/Tel1ATM-dependent phosphorylation of Ccq1 Thr93 . Furthermore , we have shown that disruption of Tpz1-Poz1 interaction causes a dramatic reduction in Poz1 association with telomeres ( Figure 7C ) , and results in phenotypes that are essentially identical to those of poz1Δ cells , including strong induction of Ccq1 Thr93 phosphorylation ( Figure 7E ) , enhanced telomerase recruitment ( Figure 7D ) and massive elongation of telomeres ( Figure 6 ) . Since we have previously shown that poz1Δ cells accumulate more RPA ( Replication Protein A ) and Rad3ATR-Rad26ATRIP kinase complex at telomeres [36] , it is likely that increased Ccq1 Thr93 phosphorylation in Tpz1-Poz1 interaction disruption mutant cells is also caused by a failure to limit the accumulation of Rad3ATR kinase . Consistent with the notion that Tpz1-Ccq1 disruption and Tpz1-Poz1 disruption cause telomere defects nearly identical to ccq1Δ and poz1Δ respectively , we also found that ( 1 ) simultaneous disruption of both Tpz1-Ccq1 and Tpz1-Poz1 interaction , ( 2 ) combining the Tpz1-Ccq1 disruption mutation with poz1Δ , and ( 3 ) combining the Tpz1-Poz1 disruption mutation with ccq1Δ , all cause a strong synergistic loss of telomere protection and immediate telomere fusion phenotype ( Figures 4 , 6 , and S3D–E ) , much like in ccq1Δ poz1Δ cells [31] . Thus , our current study establishes that Tpz1-Poz1 and Tpz1-Ccq1 interactions redundantly fulfill the essential telomere protection function of the shelterin complex . Taken together , our data from current and previous studies [12] , [36] suggest that the negative regulatory function of Tpz1-Poz1 interaction works upstream of Rad3ATR kinase to limit Rad3ATR/Tel1ATM-dependent phosphorylation of Ccq1 Thr93 , while Tpz1-Ccq1 interaction works downstream of Rad3ATR kinase to facilitate Ccq1 Thr93 phosphorylation and telomerase recruitment ( Figure 8 ) . While our studies were in progress , another study , which also identified distinct Tpz1 C-terminal domains critical for mediating either Tpz1-Ccq1 or Tpz1-Poz1 interactions , was published [43] . While their results agreed well with our current study for the domains and amino acid residues that promote Tpz1-Ccq1 or Tpz1-Poz1 interaction , our conclusions differ significantly with regard to the functional significance of these interactions for fission yeast telomere regulation . While their ChIP analysis of a Tpz1-Ccq1 disruption mutant ( tpz1-L449A ) showed no effect on Ccq1 or Trt1TERT association with telomeres [43] , our current findings clearly showed that all Tpz1-Ccq1 interaction mutants ( including tpz1-L449A ) reduce Ccq1 binding and almost entirely eliminate Trt1TERT recruitment at telomeres ( Figure 5B–D ) . Even more strikingly , they reported that tpz1-L449A poz1Δ cells carry highly elongated telomeres [43] , unlike our finding that tpz1-L449A poz1Δ and tpz1-L449R poz1Δ cells immediately lose telomeres and survive by circularizing their chromosomes ( Figures 4C–D and S3D–E ) . Furthermore , their ChIP analysis also indicated that disruption of Tpz1-Poz1 does not affect localization of Poz1 ( for tpz1-I501R and tpz1-I501A , R505E ) or Trt1TERT ( for tpz1-I501R ) at telomeres [43] , in contrast to our current study , which found that Tpz1-Poz1 interaction promotes Poz1 localization and prevent telomerase recruitment by limiting Ccq1 Thr93 phosphorylation . Unfortunately , we are not entirely sure why our findings are so different from the recent study , especially with regard to tpz1-L449A where both studies have in theory analyzed the effect of the same single amino acid mutation . A direct comparison of the Tpz1-Poz1 results is more complicated due to the fact that mutant alleles analyzed in two studies are not identical . However , we do note that our Tpz1-Poz1 disruption mutants ( tpz1-[1–485] and tpz1-W498R , I501R ) appears to destabilize Poz1 ( Figure 3E ) , and behave essentially identical to poz1Δ cells . In contrast , their mutants showed significantly less telomere elongation than poz1Δ cells and did not affect Poz1 stability [43] , raising the possibility that their mutants have retained residual Tpz1-Poz1 interaction not detected by their co-IP analysis . Another potential weakness of the previous study was that their ChIP data for Trt1TERT localization was quantified with real-time PCR primers that anneal to the sub-telomeric sequence adjacent to telomeric repeats , even for tpz1-I501A cells , which carry long telomeres [43] . In contrast , we performed dot blot-based Trt1TERT ChIP analysis and corrected for telomere length for tpz1-W498R , I501R cells . In any case , our results are incompatible with the model proposed by Jun et al . [43] , which suggested that Tpz1-Ccq1 interaction works “upstream” ( rather than downstream as we propose here ) of Tpz1-Poz1 interaction to overcome a “non-extendible” shelterin complex status that is defined by the fully connected Taz1-Rap1-Poz1-Tpz1-Pot1 linkage , based primarily on their observation that tpz1-L449A poz1Δ and tpz1-L449A-I501R cells carry highly elongated telomeres in their hand [43] . Furthermore , it should be noted that their proposed model did not even attempt to explain how the shelterin complex enforces late S-phase specific recruitment of telomerase to telomeres [35] , or how it regulates Rad3ATR/Tel1ATM-dependent Ccq1 Thr93 phosphorylation [12] , [50] to allow preferential recruitment of telomerase to shorter telomeres ( Figure S9B ) [36] . In contrast , our current model ( Figure 8 ) [36] provides an explanation for all previous observations [12] , [31] , [36] , [41] , [49] with regard to how telomerase association and telomere extensions are controlled by the shelterin complex and Rad3ATR/Tel1ATM kinases in fission yeast . Furthermore , since our detailed cell cycle ChIP analyses have recently identified Poz1 as a critical regulatory factor that promotes the timely arrival of the lagging strand DNA polymerase α at telomeres to limit accumulation of ssDNA and Rad3ATR kinase in late S-phase [36] , we suggest that Tpz1-Poz1 interaction-dependent localization of Poz1 to telomeres is required to negatively regulate telomere extension during late-S phase by ensuring proper coordination of leading and lagging strand synthesis at telomeres to limit Ccq1 Thr93 phosphorylation and telomerase recruitment ( Figure 8 ) [36] . While the shelterin complex in mammalian cells has been found to negatively regulate the DNA damage checkpoint kinases ATM and ATR [26] , [27] , ATM and ATR kinases also contribute to telomere protection and maintenance [52]–[54] . Furthermore , much like in fission yeast , ATM and ATR specifically associate with telomeres during S/G2-phases [55] , and lagging strand synthesis at mammalian telomeres is significantly delayed compared to leading strand synthesis [56]–[58] . In addition , mammalian TPP1 ( Tpz1 ortholog ) , in collaboration with TRF1 , TIN2 and POT1 , has been shown to play both positive and negative roles in extension of telomeres [18] , [19] , [22]– , much like Tpz1 in fission yeast . Studies have also found that TIN2 plays a major role in POT1-dependent inhibition of ATR activation at telomeres [60] . Therefore , TIN2-TPP1 interaction ( much like Poz1-Tpz1 interaction ) is also likely to function upstream of ATR kinase to ensure telomere stability . Since the mammalian shelterin complex subunit TRF1 ( much like Taz1 ) has been found to promote semi-conservative replication of telomeres by replicative DNA polymerases [36] , [61] , [62] , TRF1-TIN2-TPP1 interaction might control accumulation of ATM and ATR kinases at telomeres by regulating ssDNA accumulation at telomeres by coordinating leading and lagging strand synthesis . Thus , our current findings , which established roles of Tpz1-Poz1 and Tpz1-Ccq1 interactions in regulating Rad3ATR/Tel1ATM-dependent Ccq1 Thr93 phosphorylation and telomerase recruitment , may be relevant in understanding how the mammalian shelterin complex can collaborate with DNA damage checkpoint kinases to ensure telomere maintenance .
Fission yeast and budding yeast strains were generated and cultured using standard techniques and protocols [63] , [64] . Growth analysis of serially diluted liquid cultures was performed as previously described [45] with cells diluted to 4×104 cells/ml every 24 hours in YES ( Yeast Extract Supplemented ) media [63] . Effects of various tpz1 mutations on telomere heterochromatin formation were monitored by utilizing his3+::tel ( 1L ) strains that carry an integrated his3+ marker adjacent to telomere repeats of chromosome I left arm , as previously described [48] . Fission yeast strains used in this study are listed in Table S2 . For ccq1Δ::hphMX , poz1Δ::natMX6 , rap1Δ::ura4+ , and trt1Δ::his3+ , original deletion strains were described previously [4] , [38] , [49] , [65] . For pot1Δ::natMX , the drug resistance marker of pot1Δ::kanMX strain [33] was changed by transformation of a natMX fragment . For tpz1-myc , trt1-myc , poz1-myc , ccq1-myc , ccq1-FLAG , pot1-FLAG and chk1-myc , original tagged strains were described previously [31] , [35] , [49] , [66] , [67] . For poz1-FLAG , an epitope tag was introduced by PCR-based method [68] . Various myc-tagged tpz1 mutant alleles were integrated at the endogenous tpz1+ locus using a BsgI-BglI tpz1-myc-kanMX fragment , excised from mutagenized pBS-tpz1-13myc-kanMX plasmids listed in Table S3 . Point mutations were introduced by Phusion site-directed mutagenesis ( NEB , F-541 ) or QuikChange Lightning site-directed mutagenesis ( Agilent , 210513 ) . Strains carrying various tpz1 mutations without epitope tags were subsequently generated by transforming tpz1-myc mutant strains with a C-terminal tpz1 fragment marked with hphMX within the 3′ untranslated region of the tpz1+ gene . Yeast two-hybrid ( Y2H ) assays were performed by mating Saccharomyces cerevisiae MATa ( Y2HGold: MATa trp1-901 leu2-3 , -112 ura3-52 his3-200 LYS2::GAL1 ( UAS ) -GAL1 ( TATA ) -HIS3 GAL2 ( UAS ) -GAL2 ( TATA ) -ADE2 gal4Δ gal80Δ URA3::MEL1 ( UAS ) -MEL1 ( TATA ) -AUR1-C MEL1 ) strains harboring GAL4-DBD ( DNA-binding domain ) plasmids with MATα ( Y187: MATα trp1-901 leu2-3 , -112 ura3-52 his3-200 ade2-101 gal4Δ gal80Δ met− URA3::GAL1 ( UAS ) -GAL1 ( TATA ) -LacZ MEL1 ) strains harboring GAL4-AD ( activation domain ) plasmids , as described in the MATCHMAKER system manual ( Clontech ) . Plasmids used in Y2H assays are listed in Table S4 . Positive Y2H interactions were identified by spotting mated cells onto SD−HTL ( −His ) or SD−HTLA ( −His −Ade ) plates . To increase selection stringency , 1 mM or 5 mM 3-amino-1 , 2 , 4-triazole ( 3AT ) was added to SD−HTL plates . Co-IP experiments were performed as previously described [12] , [49] . Briefly , cell extracts were prepared in lysis buffer [50 mM Tris pH 8 . 0 , 150 mM NaCl , 10% glycerol , 5 mM EDTA , 0 . 5% NP40 , 50 mM NaF , 1 mM DTT , 1 mM PMSF , 0 . 2 mM APMSF , 1 mM Na3VO4 , ‘Complete’ protease inhibitor cocktail] using glass beads . For low salt co-IP experiment shown in Figure S3 , lysis buffer contained 60 mM NaCl instead of 150 mM . Proteins were immunoprecipitated using either monoclonal anti-myc antibody ( 9B11 , Cell Signaling ) or monoclonal anti-FLAG antibody ( M2-F1804 , Sigma ) , and Dynabeads Protein G ( Life Technologies ) , and washed 3× in lysis buffer . Immunoprecipitated proteins were analyzed by western blot analysis using monoclonal anti-FLAG or anti-myc as primary antibodies . In addition , anti-Cdc2 ( y100 . 4 , Abcam ) was used to detect Cdc2 in whole cell extract ( WCE ) to serve as loading control . To detect Ccq1 Thr93 phosphorylation by anti-phospho ( S/T ) Q ( Phospho- ( Ser/Thr ) ATM/ATR substrate antibody ( 2851 , Cell Signaling ) , Ccq1 was first immuno-purified from whole cell extract using anti-FLAG antibody as previously described [12] . Either horseradish peroxidase ( HRP ) -conjugated ( goat ) anti-mouse ( Pierce , 31430 ) or HRP-conjugated ( goat ) anti-rabbit ( Pierce , 31460 ) was used as the secondary antibody . Pulsed-field gel electrophoresis of NotI-digested chromosomal DNA was separated on a 1% agarose gel as previously described [69] in 0 . 5xTAE buffer at 14°C , using the CHEF-DR III system ( BioRad ) at 6 V/cm ( 200 V ) and a pulse time of 60 to 120 sec for 24 hours . For telomere length analysis by Southern blotting , EcoRI-digested genomic DNA was separated on a 1% agarose gel , transferred to nylon membrane , and hybridized to a P32-labeled telomeric DNA probe as previously described [69] , [70] . ChIP assays were carried out as previously described [36] , [49] , [70] . Fission yeast cells were crosslinked with 1% formaldehyde for 20 min at room temperature , and then incubated with 125 mM glycine for 5 min at room temperature . Cell extracts were then prepared in lysis buffer [50 mM Hepes-KOH pH 7 . 5 , 140 mM NaCl , 1 mM EDTA , 1% ( v/v ) Triton X-100 , 0 . 1% ( w/v ) sodium deoxycholate , 1 mM PMSF , ‘Complete’ protease inhibitor cocktail] using glass beads , and sonicated in an ice bath ( Misonix Sonicator 3000 with cup horn device ) to obtain ∼500 bp DNA fragments . 1–2 mg lysate was prepared in 200–250 µl lysis buffer and 5 µl was set aside for input . 1 µg of monoclonal anti-Myc ( 9B11 ) antibody was added to the ChIP sample and incubated for two hours at 4°C , then 30 µl of Dynabeads Protein G ( Life Technologies ) were added and incubated for one hour at 4°C . IPs were washed 3× with 1 ml lysis buffer , and DNA was recovered from ChIP and input samples using Chelex 100 as previously described [36] , [49] , [70] . Quantitative real-time PCR was performed with primers annealing to a region adjacent to the telomeric repeat ( jk380 5′-TAT TTC TTT ATT CAA CTT ACC GCA CTT C-3′ and jk381 5′-CAG TAG TGC AGT GTA TTA TGA TAA TTA AAA TGG-3′ ) [71] utilizing SYBR green to quantify PCR products . % Precipitated DNA values were calculated based on ΔCt between Input and IP samples with the following formula: [% Precipitated DNA] = [100*EΔCt]/[D*R] , where E = amplification efficiency of primer pairs , ΔCt = [Ct Input]−[Ct ChIP] , D = [dilution factor of ChIP]/[dilution factor of Input] and R = [Sample volume used in IP]/[Sample volume set aside for Input control] . For dot blot-based analysis of ChIP samples , DNA was denatured in 0 . 4 M NaOH , 10 mM EDTA for 10 min at 100°C , then rapidly cooled on ice , and subsequently blotted onto Hybond XL membrane ( GE Healthcare ) using a Bio-Dot Microfiltration Apparatus ( Bio-Rad ) , and hybridized to a P32-labeled telomeric DNA probe as previously described [36] , [49] , [70] . For cells carrying highly elongated telomeres , data were corrected for changes in telomere length as previously described [36] by using correction factors for telomere length , established by measuring the hybridization signal intensity of telomere versus rDNA repeats ( telomere/rDNA ) ( Table S1 ) . “Telomere length corrected” ChIP values were then calculated by multiplying the background subtracted % precipitated DNA values ( raw % precipitated DNA – no tag control % precipitated DNA ) with the correction factors , and normalized to values from wild-type cells . SEM of telomere length corrected ChIP ( SEMQ ) was calculated as previously described [36] . Biotin-DNA primer coated Streptavidin beads were prepared by mixing 30 µl Dynabeads M-280 Streptavidin ( Invitrogen , 10 mg/ml ) with 100 pmol of biotinylated oligonucleotide primers ( C-oligo: 5′ Biotin-CGT AAC CGT AAC CCT GTA ACC TGT AAC CTG TAA CCG TGT AAC C 3′; G-oligo: 5′ Biotin-GGT TAC ACG GTT ACA GGT TAC AGG TTA CAG GGT TAC GGT TAC G 3′ ) in Binding Buffer BW [5 mM Tris , pH 7 . 5 , 0 . 5 mM EDTA , 1M NaCl] for 15 min at RT . Beads were then washed 2× with 1 ml BW and resuspended in 30 µl lysis buffer LB [50 mM Tris pH 8 . 0 , 150 mM NaCl , 10% glycerol , 5 mM EDTA , 0 . 5% NP40 , 50 mM NaF , 1 mM DTT , 1 mM PMSF , 0 . 2 mM APMSF , 1 mM Na3VO4 , ‘Complete’ protein inhibitor cocktail] . 2 mg of fission yeast whole cell extract , prepared in LB , was then incubated with 30 µl Biotin-primer coated beads for 2 h at 4°C , washed 3× with 1 ml LB , and finally resuspended in 25 µl 2× Laemmli sample buffer for subsequent western blot analysis . | Proper maintenance of telomeres is essential for maintaining genomic stability , and genomic instability caused by dysfunctional telomeres could lead to accumulation of mutations that may drive tumor formation . Telomere dysfunction has also been linked to premature aging caused by depletion of stem cells . Therefore , it is important to understand how cells ensure proper maintenance of telomeres . Mammalian cells and fission yeast cells utilize an evolutionarily conserved multi-subunit telomere protection complex called shelterin to ensure protection against telomere fusions by DNA repair factors and cell cycle arrest by DNA damage checkpoint kinases . However , previous studies have not yet fully established how protein-protein interactions within the shelterin complex contribute to the regulation of DNA damage checkpoint signaling and telomerase recruitment . By utilizing separation of function mutations that specifically disrupt either Tpz1-Ccq1 or Tpz1-Poz1 interaction within the fission yeast shelterin , we establish that Tpz1-Ccq1 interaction is essential for phosphorylation of Ccq1 by the DNA damage checkpoint kinases Rad3ATR and Tel1ATM that is needed for telomerase recruitment to telomeres , while Tpz1-Poz1 interaction prevents Ccq1 phosphorylation by promoting Poz1 association with telomeres . These findings thus establish for the first time how protein-protein interactions within the shelterin complex modulate checkpoint kinase-dependent phosphorylation essential for telomerase recruitment . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"biochemistry",
"biology",
"and",
"life",
"sciences",
"molecular",
"biology"
] | 2014 | Tpz1-Ccq1 and Tpz1-Poz1 Interactions within Fission Yeast Shelterin Modulate Ccq1 Thr93 Phosphorylation and Telomerase Recruitment |
The African trypanosome Trypanosoma brucei , which persists within the bloodstream of the mammalian host , has evolved potent mechanisms for immune evasion . Specifically , antigenic variation of the variant-specific surface glycoprotein ( VSG ) and a highly active endocytosis and recycling of the surface coat efficiently delay killing mediated by anti-VSG antibodies . Consequently , conventional VSG-specific intact immunoglobulins are non-trypanocidal in the absence of complement . In sharp contrast , monovalent antigen-binding fragments , including 15 kDa nanobodies ( Nb ) derived from camelid heavy-chain antibodies ( HCAbs ) recognizing variant-specific VSG epitopes , efficiently lyse trypanosomes both in vitro and in vivo . This Nb-mediated lysis is preceded by very rapid immobilisation of the parasites , massive enlargement of the flagellar pocket and major blockade of endocytosis . This is accompanied by severe metabolic perturbations reflected by reduced intracellular ATP-levels and loss of mitochondrial membrane potential , culminating in cell death . Modification of anti-VSG Nbs through site-directed mutagenesis and by reconstitution into HCAbs , combined with unveiling of trypanolytic activity from intact immunoglobulins by papain proteolysis , demonstrates that the trypanolytic activity of Nbs and Fabs requires low molecular weight , monovalency and high affinity . We propose that the generation of low molecular weight VSG-specific trypanolytic nanobodies that impede endocytosis offers a new opportunity for developing novel trypanosomiasis therapeutics . In addition , these data suggest that the antigen-binding domain of an anti-microbial antibody harbours biological functionality that is latent in the intact immunoglobulin and is revealed only upon release of the antigen-binding fragment .
Trypanosomatid protozoan parasites cause many important diseases , including African sleeping sickness in humans and Nagana in domestic livestock in sub-Saharan Africa [1] , [2] . These organisms , like many other successful pathogens , have evolved sophisticated mechanisms for immune evasion [3] . A prominent strategy among African trypanosomes , facilitating chronic persistence in the host bloodstream and lymphatic system , relies on antigenic variation [4] . The major trypanosome surface antigen is the immunogenic variant-specific surface glycoprotein ( VSG ) present at ∼107 copies per cell and representing ∼90% of the total cell surface proteins [5] . This dense VSG coat is envisaged as functioning as a physical barrier , impeding antibody recognition of invariant surface epitopes . By repeatedly switching the VSG coat , antibodies that would recognise trypanosomes , leading to their elimination , are evaded [4] , [6] . Further , trypanosomes can reverse antibody-mediated agglutination in a protein synthesis-dependent manner [7] , and also defend themselves by efficient internalisation of antibody-VSG complexes [8] , [9] , [10] , [11] , delaying elimination by antibody-dependent complement lysis [12] , [13] . Furthermore , several groups have reported that antibody-induced VSG shedding may contribute to protection against antibody-mediated removal [7] , [14] , [15] . Trypanosoma brucei , in common with other trypanosomatids , restricts membrane exchange between the surface and endomembrane compartments to an invagination of the plasma membrane , the flagellar pocket ( FP ) , which is contiguous with the pellicular and flagellar membranes [13] , [16] . The FP comprises ∼5% of the total cellular surface and lacks the subpellicular microtubules [17] , [18] . The lumen of the FP contains an electron dense carbohydrate rich matrix and is bounded by a hemidesmosome-like zone around the neck of the pocket . Solution macromolecules such as antibodies have to transit the hemidesmosomal zone and the matrix to enter the FP . The VSG density is similar at the luminal face of the FP membrane and the bulk plasma membrane , but many other proteins such as macromolecular receptors are enriched within the FP membrane and virtually absent from the cell surface ( Field et al . [13] ) . In the bloodstream stage of the trypanosome , the cell surface turnover is exceptionally high [12]; exceeding rates reported for macrophages and fibroblasts [19] and is sufficient to cycle the entire surface in under 15 minutes [20] , [21] . While the biological significance of this high and developmental-stage specific activity is unclear , it likely contributes to a mechanism for recovering VSG and/or eliminating anti-VSG immunoglobulins bound to the surface of living parasites [12] . The difference in recycling efficiency of VSG and other surface proteins is due in part to differential trafficking through the endocytic pathway . For instance , transferrin is liberated from the parasite transferrin receptor in an acidic compartment , possibly the late endosome or the lysosome , whereas VSG is recycled via early and recycling endosomes [8] , [20] . Following clathrin-dependent endocytosis at the FP , VSG is separated from bound antibodies in sorting endosomes and recycled to the parasite surface while the antibody is directed to a distinct pathway for degradation [8] , [20] , [22] , [23] , [24] . These distinct endosomal populations have been classified depending on the presence of several key components of the vesicle transport system i . e . Rab GTPases and other markers [25] . Specifically , early/sorting endosomes play a role in fluid-phase and transporter-mediated endocytosis and contain Rab5A or 5B [26] , whereas the recycling endosomes mainly contain Rab11 [24] . The dense packing of VSG at the parasites' surface prohibits recognition of conserved membrane proximal VSG epitopes by antibodies . To potentially circumvent this we introduced camelid IgG-derived 15 kDa nanobodies ( Nb ) , representing the intact antigen-binding domains of the unique camelid IgG2 or IgG3 90 kDa heavy-chain antibodies that are devoid of light chains ( HCAb ) [27] . The monomeric Nbs have dimensions of ∼4×2 . 2 nm and offer several advantages over antigen-binding fragments derived from classical antibodies [28] . High-affinity antigen-specific Nbs can be readily obtained from an immunised camelid and selected by phage display . They are also very robust and can be engineered efficiently into larger constructs to confer novel functionality and broaden their utility [29] . Moreover , several VSG-specific Nbs , directed towards distinct regions of the VSG molecule have already been identified , of which one targets a conserved VSG epitope that is on live trypanosomes inaccessible for larger antibodies [30] , [31] . We now report that Nbs recognizing VSG isotype-specific epitopes are competent in lysing parasites both in vitro and in vivo . These trypanolytic Nbs rapidly arrest cell motility , block endocytosis , cause FP swelling , collapse mitochondrial membrane potential and exhaust ATP , ultimately leading to parasite death . However , the Nbs become non-lytic when reconstituted into HCAbs in the absence of complement . Further , polyclonal antibodies directed against VSG recognise live trypanosomes without any detectable toxicity in the absence of complement , whereas proteolytically derived monovalent Fab or Nb antigen-binding fragments are trypanolytic . These data suggest that the antigen-binding fragment of an antibody harbours biological functions which remain latent in the intact immunoglobulin .
Previously several monoclonal Nbs against Trypanosoma brucei AnTat1 . 1 were isolated by panning of a phage-displayed nanobody ( Nb ) library from lymphocytes of a VSG-immunised camelid [30] . Some of these Nbs appear to be highly specific for the AnTat1 . 1 VSG , while others exhibit cross-reactivity towards a wide variety of distinct VSGs . Surprisingly , the AnTat1 . 1 VSG-specific Nbs ( i . e . Nb_An05 , Nb_An06 and Nb_An46 ) provoke efficient lysis of AnTat1 . 1 parasites within five hours ( Fig . 1A ) . The specificity of this Nb-mediated trypanolysis is further confirmed , firstly by demonstrating that Trypanosoma brucei strains expressing MiTat1 . 1 , MiTat1 . 2 , MiTat1 . 5 and MiTat1 . 6 VSGs are not lysed ( Fig . 1B ) and secondly , via inhibiting trypanolytic activity by pre-incubation with a three-fold molar excess of purified soluble AnTat1 . 1 VSG prior to parasite challenge ( Fig . 1C ) . Furthermore , using the most potent trypanolytic Nb , i . e . Nb_An05 , it is noted that this lysis is dose-dependent ( Fig . 1D ) . Similar observations were obtained for Nb_An06 and Nb_An46 , but with different kinetics . Under identical assay conditions , Nb_An33 , which cross-reacts with multiple T . brucei VSGs , had no significant effect on parasite viability . To evaluate the therapeutic potential for Nbs in vivo , mice were infected with virulent monomorphic AnTat1 . 1A parasites and treated with lytic or non-lytic Nbs at daily intervals , starting at day one and progressing to day four post-infection . Untreated mice and those treated with the non-lytic Nb_An33 reached extremely high levels of parasitaemia within four days ( Fig . 1E ) . In contrast , mice treated with trypanolytic Nb_An05 , 06 or 46 had no detectable parasites during the entire treatment period . However , upon interruption of the Nb treatment , parasites reappeared in the blood and proliferated to ∼2×108 parasites/ml by day seven post infection ( Fig . 1E ) . The progress of the Nb-mediated trypanolysis was followed by immuno-fluorescence . Addition of ALEXA-labelled Nb_An05 to AnTat1 . 1 trypanosomes , maintained at 4°C , stained the parasites over their entire surface ( Fig . 2A upper left panel ) , whereas at 37°C , the stain concentrated rapidly in the flagellar pocket ( FP ) ( Fig . 2A , upper middle panel ) . Monitoring trypanosomes at 37°C revealed that the parasites were rapidly hampered in their mobility , within minutes following the addition of lytic Nbs , and finally became immobile . Subsequently , morphological abnormalities became noticeable whereby a gradual swelling takes place until a globular shape is adopted was attained . These parasites exhibited progressively weaker Nb_An05 surface staining ( Fig . 2A ) , and eventually lysed . Although the kinetics are slightly distinct , highly similar observations were obtained for Nb_An06 and Nb_An46 . Transmission electron microscopy on ultrathin sections was used to investigate these morphological changes further ( Fig . 2B ) . The early and prominent feature was emergence of a large vacuole , determined to be the FP on account of morphology , presence of a flagellum , and kinetoplast , matrix material in the lumen and position within the cell [13] , [32] . No FP enlargement was observed in the absence of Nbs ( Fig . 2B ) . The FP surface area after one hour incubation was significantly increased in the presence of trypanolytic Nb_An05 as compared to control cells either treated with non-lytic Nb_An33 or no Nb ( Fig . 2C ) . Interestingly , a small but significant increase in FP surface area was observed between parasites incubated with non-lytic Nb_An33 compared to untreated parasites . Nb_An05 and Nb_An46 stain the FP more intensively when compared to the non-lytic Nb_An33 ( Fig . 3 ) . Furthermore , in the presence of lytic Nbs , especially Nb_An05 , the FP is greatly enlarged as compared to the control Nb_An33 where the FP size remains essentially unaltered . Both , enlargement of the FP and failure to traffic the Nbs into internal compartments suggest that endocytosis is blocked at the FP . A flow cytometry-based pulse-chase experiment indicated greatly impaired clearance of Nb-VSG complexes when parasites are incubated with trypanolytic Nbs compared to non-lytic or conventional anti-VSG antibodies ( Fig . 4A ) . Interestingly , clearance of the non-lytic Nb_An33 was slower than conventional anti-VSG IgGs . Since motility plays a key role in clearance of antibody-bound VSG , we evaluated the effect of trypanolytic Nbs on parasite motility . Within 10–60 minutes of exposure to lytic Nbs greatly reduced parasite motility occurs , which precedes cell lysis ( Fig . 4B and C ) . Endocytosis is temperature dependent [20] , [33] , and specifically at 4°C is fully arrested [34] . The Nb-mediated lysis gradually decreased at lowered temperature , reaching a minimum at 4°C ( Fig . 4D ) . We next investigated the possible influence of trypanolytic Nbs with transporter-mediated and fluid-phase endocytosis using FITC-labelled transferrin and dextran , respectively ( Fig . 4E & F ) . Clearly , both transporter-mediated and fluid-phase uptake were reduced rapidly following the addition of lytic Nbs and this suggests that the presence of these Nbs in some manner obstructs endocytosis . Furthermore , Nb_An05 or Nb_An46 greatly reduced 2-deoxy-D-[3H]glucose uptake , which relies primarily on facilitated diffusion through glucose transporters , rather than endocytic activity [35] , [36] ( Fig . 4G ) . The non-lytic Nb_An33 did not have this effect . This reduced accumulation of glucose , which provides the major carbon source for glycolysis , may underlie the decline in cellular ATP at later time points when parasites are incubated with trypanolytic Nbs . The ATP levels were unaffected by non-lytic Nb_An33 ( Fig . 4H ) . The arrest of facilitated diffusion and endocytosis occurs within a time span of ∼10 minutes , whereas the energetic crisis through ATP depletion and the loss of mitochondrial membrane potential as assayed with the cationic dye MitoPT JC-1 ( Fig . 4I ) are clearly a secondary effect of the presence of the trypanolytic Nbs . Internalisation of surface VSG and fluid-phase uptake are both clathrin-mediated [23] , [32] , while endocytosis and recycling is regulated by Rab5A and Rab11 [8] , [37] . The localization and expression of these endocytic markers was determined during Nb-induced trypanolysis . ALEXA-labelled Nb_An05 and Nb_An46 stained the entire parasite surface and accumulate in the FP , but no obvious co-localization with clathrin or Rab11 occurs after 30 or 60 minutes ( Fig . 5A ) . Surprisingly , there is also no co-localization observed for the ALEXA-labelled control Nb_An33 with clathrin and Rab11 , indicating that none of these Nbs are internalized to a detectable level . Next , protein levels of clathrin , Rab5A and Rab11 were determined after zero , one and two hours incubation with Nb_An05 , Nb_An46 or control Nb_An33 by Western blotting ( Fig . 5B ) . The protein levels of clathrin , Rab5A and Rab11 declined after incubation with Nb_An05 and Nb_An46 , but remain unaffected with Nb_An33 . While reduction of Rab5A , Rab11 and clathrin correlates with swelling of the FP and is in accordance with earlier data [38] , it is unlikely that this represents the direct mechanism whereby endocytosis is compromised , and may rather reflect macromolecular leakage from the cells during the lysis period . The observation of a decrease of detectable Rab5 and Rab11 , while BiP levels are unaffected may be due to the requirement of an extensive cell lysis; Rab5 and Rab11 are cytosolic proteins , but BiP is located within the lumen of the endoplasmic reticulum ( ER ) . Our experiments with murine infections revealed that trypanolytic Nbs are able to control trypanosome levels ( see Fig . 1E ) . However , camelids that produce anti-VSG HCAbs do suffer from trypanosomiasis , suggesting that the presence of anti-VSG antibody alone is insufficient for parasite control [39] . To resolve this potential contradiction we reconstituted the Nbs into a monoclonal anti-VSG HCAb molecule . The coding sequence for the AnTat1 . 1-specific and trypanolytic Nb_An05 was fused to the Fc-domain , including the hinge , of human IgG1 , and this construct was transfected into NSO cells . The transfectants secrete 90 kDa Nb-Fc homodimers that lack both the CH1 domain and the light chain , and are similar to naturally occurring camelid HCAbs ( see Fig . 6A upper panel ( 2 ) ) . Surprisingly , addition of the purified chimeric Nb_An05-Fc HCAb to AnTat1 . 1 trypanosomes fails to induce lysis ( Fig . 6B ) . Nevertheless , the synthetic HCAb was perfectly functional in terms of antigen binding , bivalency and effector function as ( i ) Nb_An05-Fc HCAb recognised the VSG antigen by ELISA and surface plasmon resonance , ( ii ) addition of larger amounts of HCAb led to parasite aggregation , and ( iii ) addition of guinea pig complement to the trypanosomes exposed to the HCAb elicited complement-mediated parasite lysis ( Fig . 6B ) . To confirm that the presence of the Fc-domain in the reconstituted HCAbs abolished the Nb trypanolytic activity , the monoclonal chimeric Nb_An05-Fc HCAb protein was digested with pepsin and papain to release ( Nb ) ′2 and Nb , respectively ( see Fig . 6A upper panel ( 3 ) and ( 4 ) ) . Remarkably , these proteolytic fragments regained the lytic activity towards AnTat1 . 1 trypanosomes , especially the monovalent Nb obtained by papain digestion ( Fig . 6C ) . The data above suggested that intact immunoglobulins may possess latent functions that become apparent once the Fc and antigen-binding domains are separated . Therefore , we tested the trypanolytic activity of intact camelid serum antibodies from animals immunised with AnTat1 . 1 sVSG and from which the cloned Nbs were derived . Camelid serum contains two classes of IgG [27]; conventional 150 kDa antibodies consisting of light and heavy chains , and 90 kDa HCAb consisting of heavy chains only . The conventional subclass i . e . IgG1 and the HCAb subclasses , i . e . IgG2 and IgG3 , of the immunised camelid were purified by differential adsorption on Protein-A and Protein-G . Antibodies in these fractions recognize purified AnTat1 . 1 VSG in ELISA and Western blot and also stain living T . brucei parasites by flow cytometry and immunofluorescence [30] . Addition of the purified camelid IgG1 fraction to trypanosomes , in absence of complement , did not lyse parasites ( Fig . 7A ) . However , proteolysis of the camelid IgG1 by pepsin and papain resulting in 100 kDa Fab′2 and 50 kDa Fab fragments respectively , of which only the latter demonstrated trypanolytic activity ( inset Fig . 7A , right panel ) . Similarly , protease digestion of camelid polyclonal HCAb IgG2 and IgG3 yields bivalent 35 kDa Nb′2 and monovalent 15 kDa Nb antigen-binding fragments ( inset Fig . 7B and C , respectively ) . While incubation of AnTat1 . 1 trypanosomes with camelid HCAbs did not result in lysis , the ( Nb ) ′2 fragments elicited moderate lysis following prolonged incubation periods , while the corresponding Nb fragments provoke significant lysis ( Fig . 7B and C ) . These results are consistent with previous observations for bivalent Nb′2 and monovalent Nbs derived from reconstituted Nb-Fc HCAbs ( Fig . 6B ) . To assess whether trypanolysis could be achieved by non-camelid antibodies , we immunized a rabbit with AnTat1 . 1 sVSG . The IgG fraction contained antibodies that recognized purified VSG by ELISA and Western blot ( see Fig . 1 in [30] ) , and stained the entire surface of parasites expressing AnTat1 . 1 VSG ( data not shown ) . Similarly , pepsin and papain digestions of the rabbit IgG were performed , generating Fab′2 and Fab , respectively ( Fig . 7D , inset ) . The effect of the pools of polyclonal rabbit IgG , Fab′2 and Fab fragments in absence of complement was tested on trypanosomes in vitro . Only the Fab fragments lysed parasites significantly over a four hour incubation period ( Fig . 7D ) . Collectively , these data demonstrate that generation of bivalent antigen-binding fragments suppresses the trypanolytic property of a monomeric Nb or Fab . Besides the molecular weight and monovalency , we considered that the antigen binding properties may contribute to the trypanolytic activity . Therefore , the affinity and epitope specificity of the Nbs were analysed by surface plasmon resonance ( SPR ) and flow cytometry . The competitive or cumulative binding of Nbs to the AnTat1 . 1 antigen on intact parasites ( Fig . 8A ) revealed that Nb_An05 and Nb_An06 share overlapping VSG epitopes , which are distinct from the epitopes recognized by Nb_An46 and Nb_An33 . The binding of the two distinct lytic Nbs , Nb_An05 and Nb_An46 , to immobilised VSG occurs with comparable kinetic on-rates of 3 . 5 and 7 . 4×105 M−1 s−1 respectively and more distinct off-rates of 2 . 3×10−3 and 3 . 25×10−2 s−1 . Equilibrium dissociation constants ( KD = koff/kon ) of 6 . 6 , 18 and 44 nM were calculated for Nb_An05 , Nb_An46 and Nb_An33 ( Fig . 9 and Table 1 ) . Interestingly , the trypanolytic Nbs exhibited smaller koff values than the non-lytic Nb_An33 . The SPR measurements suggest that the Nb-mediated trypanolysis may only occur above a critical threshold koff value ( Fig . 9 ) . To test this , the trypanolytic Nb_An05 was subjected to randomization of select tyrosine residues in its complementarity determining regions 1 and 3 ( Fig . 10A ) . This resulted in the identification of several paratopic variants that retained trypanosome-binding , but with greatly reduced affinity ranging from 120 nM to 2 µM as compared to the 6 . 6 nM affinity for the wild type Nb_An05 ( Fig . 10B , Table 1 ) . Relative to the wild type Nb_An05 , these variants had kon rates reduced by 2 to ∼50-fold , whereas the koff rates are ∼1 . 5 to ∼25-fold faster . Therefore the variants exhibit a wide diversity in KD , kon and koff constants . When tested against live trypanosomes , two Nb_An05 mutants , Nb_An05-04 and Nb_An05-12 , retained some lytic activity against the parasite ( Table 1 ) . Remarkably , these two Nbs had the slowest koff rates of all the variants . Overall , this experiment indicates that a koff rate slower than 10−2 s−1 is required to detect trypanosome lysis under our current in vitro test conditions . The correlation between KD or kon and trypanolysis is less clear .
Trypanosoma brucei has evolved very efficient systems for immune evasion , which include antigenic variation and mechanisms for removal of antibody-VSG complexes from the surface by endocytosis and proteolysis of the immunoglobulin . Hereby , the VSG is efficiently recycled [8] . It is conceivable that uptake of antibody-VSG complexes and subsequent trafficking is influenced by the antibody valency and molecular weight . Exposing trypanosomes to the antigen binding domain ( Fab or Nb ) of an immunoglobulin alone represents a non-physiological circumstance , and the evidence presented here suggests that this presents a challenge which the parasite may be unable to circumvent . We show that small , monovalent VSG-specific antibody fragments , Fabs or Nbs , efficiently lyse trypanosomes both in vitro and in vivo . Hereby , the monovalency of these fragments is pivotal for trypanolysis as bivalent Nb′2 are significantly less lytic than monomeric forms . Reconstitution of monovalent Nbs into an HCAb ( increasing valency , molecular weight and incorporating an Fc-domain ) abolished the trypanolytic activity in vitro , whereas remarkably , releasing the Nb domain via proteolysis of the recombinant HCAb restored the trypanolytic activity . This suggests that intact , bivalent monoclonal or polyclonal immunoglobulins , including rabbit and camelid classical antibodies and camelid HCAbs , are essentially harmless to trypanosomes in the absence of complement or any other bystander effector . In contrast , polyclonal Fabs or Nbs derived from the serum antibodies and deprived of classical effector Fc-domains , acquire trypanolytic activity . It should be emphasized that the trypanolytic potency of recombinant Nbs , Nbs from polyclonal HCAbs ( IgG2 or IgG3 ) , and Fabs from polyclonal IgG are not directly comparable as the exact titre of VSG-specific antigen-binding fragments within the polyclonal pool is unknown and the efficiency of lysis is clearly concentration dependent ( Fig . 1D ) . Immunoglobulins evolved with the antigen binding site ( Fab ) at one pole and with Fc effector functions that trigger complement-mediated killing and receptor-mediated phagocytosis at the other . Normally these effector functions are exerted only following antigen binding and are mediated by the Fc-domain . Nevertheless , intrinsic activities within antigen-binding domains may be present . For example Nbs with competitive enzyme inhibiting capacity [40] , [41] or Fabs with catalytic activity , ‘Abzymes’ , have been described [42] , [43] . In addition , Fabs can exhibit an intrinsic ability to convert molecular oxygen into hydrogen peroxide , which may contribute to destruction of the bound antigen [44] , [45] . This latter activity cannot be at the origin of the trypanolytic activity described here , as hydrogen peroxide formation occurs in the hydrophobic cavity between the VH and VL domains and reduction of singlet oxygen is catalysed by VH residues Trp-36 and Trp-47 [46] . These conditions are absent in Nbs , which lack a VL-domain , while Trp47 is also substituted . Moreover , we were unable to detect hydrogen peroxide formation during trypanolysis ( data not shown ) . Despite the highly similar phenotypes following trypanolytic Nb exposure and RNAi of specific endocytic factors [32] , [38] , [47] , two very important differences suggest distinct mechanism . Firstly , the Nbs elicit FP enlargement much more rapidly than RNAi , and too quickly for this to be possible via turnover of critical proteins . Secondly , the Nb is an exogenous agent . Other small exogenously delivered molecules , including aptamers [48] , cathelicidins [49] , neuropeptides [50] and a modified bovine host defence peptide ( BMAP-18 ) [51] can also elicit trypanolysis in the absence of any bystander or toxin , but their modes of action are clearly distinct from the Nbs . For example cathelicidins disrupt surface membrane integrity , which is preceded by immobilisation and rapid swelling of the parasite . Although immobilisation and swelling of the parasites also occurs with Nbs , the outer membrane integrity is not significantly affected as evidenced flow-cytometrically by the lack of leakage of FITC-labelled Dextran ( 4 kDa ) when parasites are incubated with trypanolytic Nbs . Therefore , with Nbs it seems that reduction in ATP-levels , with effects on motility , endocytosis and morphology , rather than direct surface membrane disruption , is a crucial step in parasite killing . The rapid and very severe block to endocytosis is remarkable and multiple lines of evidence demonstrate this; including accumulation of lytic Nbs in the FP and impairment to removal of the VSG-bound Nb from the parasite surface compared with non-lytic Nbs and IgG . The absence of intracellular staining or co-localisation with clathrin or Rab11 with any trypanolytic Nb strongly suggests that Nbs are not internalized to any significant degree , but temperature dependence suggests that this is an active process . It is possible that the protection accorded by lower temperature is due to inhibition of membrane transport , so preventing the FP enlargement . Drastic swelling of the FP is indicative of a block to bulk membrane endocytosis , in the presence of ongoing exocytosis , and has been reported previously in energy-depleted cells [52] . Furthermore , MitoPT™ JC-1 staining detected depolarization of the mitochondrial membrane at later times after trypanolytic Nbs exposure , but given an absence of intracellular Nbs , no direct effect on the mitochondrial membrane can be assumed . It is difficult to pinpoint the critical parameter responsible for the intrinsic destructive capacity of the monovalent , antigen-binding fragments and where a bivalent character or the presence of the Fc-domain is counterproductive for trypanolysis . Several factors are likely important , although they probably act synergistically in attaining lytic activity . Firstly , to be trypanolytic Nbs must bind VSG with high affinity . Mutagenesis-derived trypanolytic Nb_An05 variants that recognize the same epitope with modified binding kinetics indicate that toxicity requires slow release kinetics ( low koff ) , suggesting that prolonged interaction with VSG is beneficial to lysis . However , as monovalency dominates the binding parameters it is not possible to increase trypanolytic potency with bivalent constructs . Secondly , the Nb-VSG complex , unlike the IgG-VSG complex where the IgG potentially cross-links two VSG dimers , is not internalized and therefore remains at the surface . Engstler et al [12] found that smaller antibody fragments have reduced clearance from the parasite surface compared to intact antibodies and our results indeed confirm that there is greatly reduced VSG-Nb elimination from the surface; however in the case here we also find a failure to be internalized into the parasite cell . Third , the precise VSG epitope targeted by the Nb is likely important . Interestingly , some epitopes including the conserved N-glycan present on various VSG serotypes and recognized by Nb_An33 [30] , [31] failed to induce lysis , whereas Nb_An46 and Nb_An05 , targeting different epitopes , induced potent lysis . Remarkably , the competition binding experiments suggest that the most potent trypanolytic Nb has a binding site furthest from the membrane and may even occlude access of molecules to underlying epitopes ( see Fig . 8B ) . Fourth , the observation that parasites in presence of trypanolytic Nbs have reduced ATP levels suggests a correlation between energy-depletion and reduced endocytosis . Fifth , the observation that parasites in presence of trypanolytic Nbs exhibit a loss in mitochondrial membrane potential ( Δψm ) likely contributes to the observed reduced ATP levels . Sixth , the impaired flagellar motility observed very rapidly and only in presence of trypanolytic Nbs might be a crucial initiation step in the trypanolysis process [53] . Of the different mechanisms by which trypanolytic Nbs could cause lysis , the model we favour is that high affinity binding ( mediated by a low koff ) of trypanolytic Nbs to VSG impairs recycling of the surface , and within minutes this translates into impaired cellular motility . This rapidly blocks formation and/or budding of clathrin-coated pits , i . e . endocytosis . The swelling of the FP is likely to lead rapidly to cell lysis , which was also observed using clathrin RNAi . It is however unlikely that this process itself leads to decreased ATP , as ATP levels decrease more slowly than the onset of cellular defects . This slow loss of low molecular weight ATP ( 509 Da ) also effectively eliminates the possibility of rapid generation of pores or disruptions in the plasma membrane , although we cannot rule out possible smaller disruptions to the lipid bilayer that could result in ionic imbalance for example . Further , we also observed very rapid loss of glucose accumulation , but as glucose is mainly accumulated through GLUT channels in the bulk plasma membrane , it is unlikely that this is directly related to decreased endocytosis . One possible explanation for the decreased glucose transport across the plasma membrane is that lower endocytic activity and motility reduce the draw on ATP and hence glucose utilization , decreasing the concentration gradient for glucose transport into the cell , and hence lowering glucose uptake . It may also be that this reduced consumption masks an otherwise more prominent change to intracellular ATP levels . Therefore trypanolytic Nbs in some manner are able to compromise cellular energetics , but the connection between binding VSG , compromised endocytosis and lower cellular energy remains unclear . In contrast , Nb_An33 binds to a sugar epitope and therefore does not cause the above described phenotype . Furthermore , given that Nb_An33 has a higher koff-value means that it dissociates faster from the coat so that it can not exert a trypanolytic effect . In conclusion , the present work demonstrates firstly that high affinity antigen-binding antibody fragments can exert a direct biological “cytotoxic” function in the absence of the effector Fc-domain , and which is latent in intact immunoglobulins . Secondly , targeting the trypanosome surface with such small high affinity antigen-binding fragments is sufficient to efficiently kill the parasite . In addition , Nbs targeting various epitopes on the surface coat of trypanosomes offers possibilities for novel treatments for trypanosomiasis by developing small trypanotoxic compounds that compromise cell viability . Indeed , since the lytic Nb_An05 and Nb_An46 recognize distinct AnTat 1 . 1-specific peptidic epitopes , it seems that multiple sites on VSG could serve to target therapeutics . Moreover , the observation that Nb_An05 and Nb_An06 have overlapping epitopes that are not necessarily identical ( these Nbs have different CDR sequences ) suggests that the therapeutic epitope could be reduced in size to a small footprint of only a few hundred Å2 that might be conserved among VSGs of various serotypes . Although the specificity for a particular VSG , as is the case for the trypanolytic Nbs here imposes a limitation on therapeutic value , our data indicate that a cross-reactive therapeutic Nb recognizing many or all VSGs would require binding at a conserved VSG epitope with very high affinity . Under this assumption it might become feasible to design or select small organic compounds that would bind with high affinity to a VSG epitope , leading to trypanosome clearance , and as such be used as a novel therapeutical approach .
The experiments , maintenance and care of mice complied with the guidelines of the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes ( CETS n° 123 ) . The experiments for this study were approved by the Ethical Committee for Animal Experiments of the Vrije Universiteit Brussel , VUB , Brussels , Belgium ( Permit Number: 08-220-8 ) . Purification of Trypanosoma b . brucei ( AnTat1 . 1 , MiTat1 . 1 , MiTat1 . 2 , MiTat1 . 5 and MiTat1 . 6 ) bloodstream parasites , their soluble VSGs and the immunization of a camelid with AnTat1 . 1 sVSG was as described [30] . Camelid serum IgG fractionation , cloning , selection and purification of Nbs was according to published methods [30] , [54] , and reconstitution of HCAbs by fusing the Nb_An05 or Nb_An33 to the human Fc of IgG1 was as explained in [55] . Purified IgG was digested with 1% Hg-papain ( Sigma ) or porcine pepsin ( EC 3 . 4 . 23 . 1 , Sigma ) following the manufacturer's instructions . The digest was passed over Protein-G Sepharose and gel-permeation Superdex-200 ( 10/30 ) ( GE Healthcare ) in PBS ( pH 7 . 4 ) to purify Fab , Fab′2 , ( Nb ) ′2 or Nb . The protein concentration was assessed spectrophotometrically . From each antibody , or digested material , 130 pmole was loaded onto a 12% SDS-polyacrylamide gel ( under non-reducing conditions ) , and transferred to nitrocellulose . After blocking with 1% ( w/v ) bovine serum albumin , the membrane was incubated sequentially with a rabbit polyclonal anti-VHH IgG and a goat anti-rabbit-IgG antibody conjugated to horseradish peroxidase ( Sigma ) . In between the successive two hour incubations was a PBS-0 . 1% Tween 20 wash . Thirty minutes after adding chromogenic substrate ( methanol/4-chloro-1-nafthol in PBS/H2O2 ) the reaction was stopped by rinsing the membrane with water . The determination of the protein expression levels of Clathrin , Rab5A , Rab11 and BiP during the trypanolysis assay was performed as described elsewhere [56] . Chemiluminescense detection was by exposure to X-ray film ( Kodak BioMax MR ) , and ImageJ software used for quantification . For the affinity determination with Biacore 3000 , different concentrations , ranging from 500 nM to 7 . 5 nM , of Nb_An05 , Nb_An06 , Nb_An46 or Nb_An33 were added to a CM5 chip to which 500 RU of AnTat1 . 1 VSG had been coupled [57] . Sensograms were fitted for a 1∶1 binding model using the BIA-evaluation software version 4 . 1 ( GE Healthcare ) , resulting in kon , koff and KD values as output . The affinities of the Nb_An05 mutants were measured by surface plasmon resonance on a Biacore T100 system . Between 1000 and 1500 RU of soluble AnTat1 . 1 VSG was coupled onto a CM5 chip ( GE Healthcare ) via amine groups according to the manufacturer's descriptions using EDC and NHS as cross-linking agents and ethanolamine to block free esters . For the affinity determination , Nb concentrations ranging from 500 to 7 . 5 nM were added to the antigen-coated chip at a flow-rate of 30 µl/min in HBS buffer [10 mM Hepes ( pH 7 . 5 ) , 150 mM NaCl , 3 . 5 mM EDTA and 0 . 005% ( v/v ) Tween-20 ) ] . Bound Nbs were eluted with 10 mM glycine-HCl ( pH 2 . 5 ) . Sensograms were fitted for a 1∶1 binding model using the Biacore T100 Evaluation Software 2 . 0 . 2 ( GE Healthcare ) , calculating kon , koff and KD values . The different Nb clones were evaluated on live , bloodstream form AnTat1 . 1 trypanosomes through flow cytometry following a direct or three-step labeling procedure . The direct labeling required conjugation of Nbs with ALEXA Fluor 488 according to the manufacturer's instructions ( Molecular Probes ) . Hereby , parasites ( 2×105 in 100 µl PBS/10% FCS ) were cooled in an ice-bath ( 30 minutes ) before adding Nbs . After 10 minutes incubation with ALEXA-labelled Nbs ( 1 µg ) , cells were washed with ice-cold PBS/10% FCS and analyzed . The three-step labeling procedure relied on the detection of the surface-bound Nbs with a mouse anti-6⋅His IgG and a phycoerythrin-labeled rat anti-mouse IgG . Flow cytometry analyses were performed on a FACS Canto II and histograms were prepared using the FlowJo software ( Becton Dickinson , San Jose , CA ) . To evaluate the antibody-clearance rate by trypanosomes , a pulse-chase experiment was performed . This consisted of incubation of 2×105 parasites with 1 µg ALEXA-labelled Nbs ( Nb_An05 , Nb_An46 , Nb_An33 or irrelevant Nb ) or 10 µg rabbit polyclonal IgGs against VSG for 10 minutes on ice in HMI-9 medium/5% FCS . Next , the free antibodies were washed away by washing the parasites 2 times with ice cold HMI-9 medium . The parasites were resuspended at 5×106/ml in HMI-9/5% FCS in separate tubes and brought at 37°C . At different time-points ( 0-0 . 5-1-1 . 5-2 . 5-5-7 . 5-10-30-60-120 minutes ) , aliquots were washed with 2 ml HMI-9 to remove free antibodies . The parasites were resuspended in 100 µl HMI-9 followed by addition of 100 µl 4% paraformaldehyde/PBS to stop the metabolic activity . Following a 30 minutes fixation step , the cells were washed with ice-cold PBS/10% FCS and analyzed as described above . The mean-fluorescence intensity of parasites incubated with the antibody at time 0 was taken as the 100 percent signal . Nbs were labelled with ALEXA-488 ( Molecular Probes ) according to the manufacturer . Aliquots of 106 parasites were incubated with 10% normal rabbit serum in PBS for 30 min in an ice-bath before adding different ALEXA-labelled Nbs ( 1 µg ) , ALEXA-labelled rabbit polyclonal anti-VSG IgG ( 5 µg ) , camelid polyclonal anti-VSG IgG ( 5 µg ) or Nb_An-Fc chimer ( 6 µg ) . After 30 minutes the parasites were pelleted , washed with 10% normal rabbit serum in PBS , and analysed by fluorescence microscopy ( Nikon ECLIPSE E600 with phase contrast , 500×–1250× magnification ) . To assess the role of the membrane fluidity in uptake of Nbs , parasites were pre-incubated for 1 hour at 4°C or 37°C before adding ALEXA-labelled Nb_An05 or control Nb . After 30 minutes , parasites were washed 3 times with PBS/5% FCS and analysed by immuno-fluorescence microscopy . Individual samples were taken every 30 minutes and visualised by fluorescence microscopy to study the kinetics of Nb clearance by parasites . The co-localization experiments were performed as described [56] , [58] . Images were obtained using a Nikon ECLIPSE E600 epifluorescence microscope fitted with optically matched filter blocks and a Hamamatsu charge-coupled-device camera or a Leica confocal laser-scanning microscope . Images were false-coloured and assembled using Adobe Photoshop . In vitro: short term ex vivo trypanolysis was performed using 200 µl DEAE52-purified parasites ( stock: 106 parasites/ml HMI-9 medium/5% FCS ) which were incubated at 37°C at 5% CO2 in a humidified atmosphere with different antibodies ( rabbit polyclonal anti-VSG , Fab , Fab′2 , fractionated polyclonal camelid IgG , Nbs , ( Nb ) ′2 or Nb_An-Fc ) at a maximum concentration of 0 . 067 nmole . The surviving parasites were counted at regular intervals over a time period up to 5 hours using a Bürker hematocytometer . For the inhibition of the Nb-mediated trypanolytic activity , the Nbs were pre-incubated for 30 minutes with a 3-times molar excess of purified AnTat1 . 1 soluble VSG , prior to addition to the parasites . The percentage lysis was calculated relative to the condition with a non-trypanosome specific Nb or without Nb . For the site-directed mutagenized Nb_An05 trypanolysis experiments , 2×105 parasites in 200 µl HMI-9 medium supplemented with 10% decomplemented fetal bovine serum ( FBS ) were incubated with 1 µg Nb , followed by incubation at 37°C in a conditioned atmosphere with 5% CO2 for 5 hours . Lysis was quantified by parasite counting using a Bürker hematocytometer and the percentage lysis of the different paratope variants was calculated relative to that of wild type Nb_An05 ( i . e . 100 percent ) . In vivo: Eight-weeks old F1-mice were injected intra-peritoneally ( i . p . ) with 5000 virulent monomorphic AnTat1 . 1A parasites per mouse . Starting from day 1 till day 4 post infection , 100 µg Nb was i . p . injected . The parasitemia in 2 . 5 µl blood ( obtained from the tail vein of infected mice ) diluted in 500 µl PBS was monitored microscopically , and the survival of the mice was recorded . Parasites ( 2×105 in 200 µl HMI-9 medium/5% FCS ) were incubated for 1 hour at 37°C , 25°C , 15°C and 4°C to reduce or stop the membrane fluidity , before adding Nbs ( 1 µg ) and to monitor their survival over a 4–5 hour period . The interference from Nbs on the specific or non-specific uptake of nutrients by trypanosomes was assessed by adding FITC-labelled transferrin or dextran ( Sigma ) , respectively . After a total of 10 minutes incubation , parasites were pelleted , washed 3 times with HMI-9 medium/5% FCS , suspended in PBS and FITC-labelled nutrient uptake monitored by fluorescence readings ( Cytofluor II , PerSeptive Biosystems ) . The 2-deoxy-D-[1-3H]glucose ( 1 mM , 1 µCi , Perkin Elmer ) uptake by 2×106 parasites in presence of 2 µg Nbs was as described in [59] . Cells were lysed and the glucose concentration determined by triplicate measurements using a liquid scintillation beta-counter ( Perkin-Elmer Rackbeta , Boston USA ) . The total protein concentration was determined as described in [60] . The data are expressed as percentage of glucose uptake relative to the glucose-level of parasites incubated for the same time period without Nbs . To determine the effect of endocytosis disturbance by Nbs ( 10 µg ) on the internal ATP concentration , 107 parasites ( at 37°C ) were lysed after different time intervals by three freeze-thawing cycles . The ATP concentration of triplicate samples was quantified by the ATP-assay ( Molecular Probes ) . The Mitochondrial Permeability Potential ( Δψm ) was determined using the cationic dye MitoPT™ JC-1 ( Immunohistochemistry Technologies , Bloomington , MN ) , which exhibits potential-dependent accumulation in mitochondria . At low membrane potentials , JC-1 continues to exist as a monomer and produces a green fluorescence ( emission at 527 nm ) . At high membrane potentials or concentrations , JC-1 forms J aggregates ( emission at 590 nm ) and produces a red fluorescence . The staining procedure was as recommended by the suppliers and the trypanolysis assay was performed as described above . Briefly , after different time points of the trypanolysis assay a 1∶1 ratio of the MitoPT JC-1 staining solution was added and the cells incubated for an additional 15 minutes in a CO2 incubator at 37°C . Next , the cells were pelleted and washed twice with 2 ml assay buffer warmed at 37°C . Finally , the fluorescent signals were measured by flow cytometry . As positive control , the cells were incubated with a final concentration of 50 µM Carbonylcyanide m-chlorophenylhydrazone ( CCCP ) for 60 minutes in a CO2 incubator at 37°C . For electron microscopy cells were prepared as described elsewhere [38] . Observations were made on a Tecnai 10 electron microscope and images were captured with a MegaView II camera and processed with AnalySIS and Adobe Photoshop software . The co-localization experiments were performed as described in [56] . All manipulations were conducted using HMI-9/5% FCS . Parasites ( 107/ml ) were incubated with ALEXA-labelled Nbs ( 10 µg/ml ) at 37°C for 0–30 or 60 minutes followed by two washes with PBS and fixed with 4% paraformaldehyde ( PFA ) in ice-cold PBS . Immunofluorescence was performed as described in [58] with a few modifications . Using an ImmEdge pen ( Vector Laboratories , Burlingame , CA ) , compartments were drawn on a poly-lysine slide ( Polysine; VWR International , Leuven , Belgium ) and 200 µl of 4% PFA-fixed cells was placed in each compartment . The slides were incubated in a moist chamber , and the cells were allowed to settle on the slide followed by a permeabilisation with 0 . 1% Triton X-100 . Staining was performed as described previously [58] . The trypanosomal Golgi complex was stained using dapi ( Vectashield ) . Cells were observed either on a Nikon Microphot-FX epifluorescent microscope attached to a Photometrics CH350-CCD camera or with a Laser Scanning Microscope 510 ( Zeiss ) . Images were false-coloured and assembled using Adobe PhotoShop . To follow the kinetics of ALEXA-labeled Nb uptake , parasites were incubated ( 37°C ) in presence of ALEXA-labelled Nbs over a 2 hours time period . Next , parasites were washed twice with PBS/5% FCS , fixed with PFA in ice-cold PBS and processed as described above . Variants of Nb_An05 were generated by site-directed mutagenesis using degenerate primers ( 5′CCGGCCATGGCCGATGTGCAGCTGGTGGAGTCTGGGGGAGGCTCGGTA CTAACTGGAGGGTCTCTGAGACTCTCCTGTGCAGCCCCTGGAATCACCNHTCGTATGTACTGCATGGCC-3′ and 5′-GGAGACGGTGACCTGGGTCCCCCGGCCCCAGTAA GCAAACTCAGTTCCCTGAGCGGAGCCTGAADNGCAGCCADNGTCTCTTGCCG-3′ ) which allow replacement of tyrosine residues in complementarity determining regions ( CDR ) 1 and CDR3 that are anticipated to contribute in the interaction with the VSG-antigen . The PCR amplicons were subsequently cloned into pMES using PstI and BstEII restriction sites and individual mutant clones were screened for their functionality in a VSG-specific ELISA , using a peroxidase conjugated anti-6×His detection IgG ( Serotec ) . Nb_An05-02 ( HQ680967 ) , Nb_An05-04 ( HQ680968 ) , Nb_An05-05 ( HQ680969 ) , Nb_An05-06 ( HQ680970 ) , Nb_An05-12 ( HQ680971 ) , Nb_An05-17 ( HQ680972 ) , Nb_An05-19 ( HQ680973 ) . | Haemoparasites , such as African trypanosomes , have developed potent immune evasion mechanisms to avoid antibody-mediated elimination . Consequently , trypanosome surface antigen-specific immunoglobulins in the absence of complement are non-trypanocidal . In contrast , certain monovalent nanobodies ( Nb ) , monomeric antigen-binding domains derived from camelid Heavy-Chain Antibodies ( HCAb ) and which have a much lower molecular weight ( 15 kDa ) than classical antibodies ( 150 kDa ) , efficiently lyse trypanosomes both in vitro and in vivo . This is surprising as classically immunoglobulin effector functions are mediated via the Fc-domain , which is absent from the Nb . We demonstrate that the Nb-mediated trypanolysis depends on the low molecular weight , monovalency and high affinity and is associated with loss of motility , a major block to endocytosis , energy depletion and cell death . Overall , targeting the parasite surface with low molecular weight , high affinity Nbs is sufficient to exert a direct therapeutic action . Therefore , the exploitation of Nbs against African trypanosomiasis represents a novel therapeutic strategy . Furthermore , demonstration that a high affinity antigen-binding Nb or Fab fragment lacking an effector domain ( i . e . , Fc-domain or an attached toxin ) can exert a direct biological function , suggests that intact antibodies likely harbour latent functionality which only become revealed upon removal of the Fc-domain . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biotechnology",
"immunology",
"biology",
"molecular",
"cell",
"biology"
] | 2011 | High Affinity Nanobodies against the Trypanosome brucei VSG Are Potent Trypanolytic Agents that Block Endocytosis |
The initial step in target cell infection by human , and the closely related simian immunodeficiency viruses ( HIV and SIV , respectively ) occurs with the binding of trimeric envelope glycoproteins ( Env ) , composed of heterodimers of the viral transmembrane glycoprotein ( gp41 ) and surface glycoprotein ( gp120 ) to target T-cells . Knowledge of the molecular structure of trimeric Env on intact viruses is important both for understanding the molecular mechanisms underlying virus-cell interactions and for the design of effective immunogen-based vaccines to combat HIV/AIDS . Previous analyses of intact HIV-1 BaL virions have already resulted in structures of trimeric Env in unliganded and CD4-liganded states at ∼20 Å resolution . Here , we show that the molecular architectures of trimeric Env from SIVmneE11S , SIVmac239 and HIV-1 R3A strains are closely comparable to that previously determined for HIV-1 BaL , with the V1 and V2 variable loops located at the apex of the spike , close to the contact zone between virus and cell . The location of the V1/V2 loops in trimeric Env was definitively confirmed by structural analysis of HIV-1 R3A virions engineered to express Env with deletion of these loops . Strikingly , in SIV CP-MAC , a CD4-independent strain , trimeric Env is in a constitutively “open” conformation with gp120 trimers splayed out in a conformation similar to that seen for HIV-1 BaL Env when it is complexed with sCD4 and the CD4i antibody 17b . Our findings suggest a structural explanation for the molecular mechanism of CD4-independent viral entry and further establish that cryo-electron tomography can be used to discover distinct , functionally relevant quaternary structures of Env displayed on intact viruses .
About 2 . 5 million individuals are newly infected with human immunodeficiency virus ( HIV ) each year , and over 2 million deaths result annually from HIV/AIDS ( http://www . unaids . org ) . HIV-1 and the closely related simian immunodeficiency virus ( SIV ) bind to target cells by the interaction of trimeric envelope glycoprotein spikes ( Env ) , a heterodimer of a transmembrane glycoprotein ( gp41 ) and a surface glycoprotein ( gp120 ) [1] , with CD4 and a co-receptor ( CCR5 or CXCR4 ) [2] . Understanding of the molecular structure of trimeric Env on infectious virus particles before and after contact with the T-cell surface is fundamental to the informed design of immunogens to elicit broadly neutralizing antibodies and for deciphering the detailed molecular mechanisms underlying HIV infection [3] , [4] . Presently , no atomic resolution structures are available for trimeric Env in any conformational state , although there are several sets of coordinates available from X-ray crystallography for the truncated core of monomeric gp120 in unliganded [5] and liganded forms [6] , [7] , [8] . Recent advances in cryo-electron tomography to obtain 3D density maps from pleiomorphic biological structures provide new methods to tackle the challenge of describing the structure of trimeric Env as displayed on infectious viruses under near-native conditions [9] . Starting from a series of tilted projection images of plunge-frozen viruses , tomograms that capture the distribution of density on the surface and interior of the virus can be determined . In general , the resolution that can be obtained in a tomogram of a single virus is barely enough to discern molecular shapes because images are recorded at the lowest possible electron doses in order to minimize damage from electron irradiation of the sample . However , by extracting subvolumes corresponding to each trimeric spike , and accounting properly for the missing wedge [10] of data that is inherent to electron tomography , 3D classification and averaging can be used to obtain density maps at signal-to-noise ratios that are sufficiently high for molecular interpretation [9] , [10] , [11] , [12] . Cryo-electron tomographic studies have been recently used to obtain density maps for trimeric Env on both HIV-1 and SIV [13] , [14] , [15] , [16] ( see also Figure S1 ) . Analysis of HIV-1 virions resulted in structures of trimeric Env in unliganded , b12-antibody neutralized and sCD4/17b-antibody liganded states at ∼20 Å resolution and in a working model for structural changes in trimeric Env that occur upon engagement of the CD4 receptor on a target cell [15] . The validity of the experimental and computational procedures used to obtain the density maps of trimeric HIV-1 gp120 were first established [10] by determining structures of protein complexes with known structures at atomic resolution and of gp120-antibody complexes . Previous tomographic studies of trimeric SIV Env have only focused on unliganded Env and have been controversial . Density maps and derived molecular models reported for trimeric Env displayed on SIVmac239 by Zhu et al . [13] and SIVmneE11S by Zanetti et al . [14] differ significantly from each other and from that reported in our earlier studies of unliganded HIV-1 BaL [10] . It has remained unclear whether the disagreements in the density maps reflect genuine differences in structures of trimeric Env between the different viral isolates studied or originate largely from variations in the strategy and methods used for structure determination . The choice of gp120 atomic coordinates chosen to interpret the results from tomography is also controversial , as the available atomic coordinates of monomeric SIV gp120 in unliganded [5] , and monomeric HIV-1 gp120 in various liganded and antibody-bound states [6] , [7] , [8] display significant differences , especially in the inner domain . To address this controversy , we have carried out electron tomographic studies of SIVmneE11S and SIVmac239 from the same purified virus preparations used in the previous conflicting reports [13] , [14] , using the computational procedures established in our previous work [10] , [15] on trimeric HIV-1 BaL Env . Here , we present both biochemical and computational evidence to validate the fits of gp120 to the density maps derived by cryo-electron tomography and extend the analysis to a CD4-independent SIV strain where we show that trimeric Env is displayed in a different quaternary conformation .
A low-dose projection image recorded at zero tilt from a specimen of plunge-frozen SIVmneE11S is shown in Figure 1a . The viral membranes are visible , but the surface spikes can be barely discerned . Much more detail is evident in a tomogram constructed from a series of images spanning a tilt range of ±65° where the viral envelope and individual spikes can be easily detected ( Video S1 , S2 ) . Inspection of slices through the tomogram ( Figure 1b ) provides an indication of the improvement in image quality compared to the single projection image in Figure 1a . The tomographic volumes can be visualized either as an image stack ( Video S1 , S2 ) , or as a segmented rendering ( Figure 1c ) . To obtain a density map for each trimeric Env , typically 3000–4000 spikes present on the surface of reconstructed virions were selected using automated procedures and subjected to 3D alignment , classification and 3D averaging ( Figure 1d ) as described previously [10] to obtain profiles of 3D density distribution ( see Figure S2 for expanded version of Figure 1d ) . The maps can also be visualized as a cross-sectional slice through the density to view the overall structural profile of Env ( Figure 1e ) . The highest densities and highest signal-to-noise ratios are observed in the gp120 region and at the putative gp120/gp41 interface , indicating that these are the best-defined regions of the structure . There is lower density and a lower signal-to-noise ratio in the region immediately above the membrane , corresponding to the gp41 stalk , implying that there is less scattering power ( i . e . protein mass ) and/or possibly greater disorder in this region relative to the gp120 region . The methods that we have used for image classification and alignment and for missing wedge correction minimize the possibility that these less defined regions of the map lead to artifacts in the final averaged map . The maps for SIVmneE11S and SIVmac239 strains , shown as isosurface representations ( Figures 1f and 1g , respectively ) establish that the two spikes are similar in shape , each being composed of a propeller-shaped region with three gp120 blades connected both at the apex and at the base , corresponding to the gp120/gp41 interface . This architecture closely resembles that reported previously for unliganded HIV-1 Env at ∼20 Å resolution ( Figure 1h; [15] ) . Our finding that the gp41 stalk connecting the gp120/gp41 interface to the membrane is compact is in agreement with this feature of the SIVmneE11S Env map reported by Zanetti et al [14] and our earlier study of HIV-1 BaL Env , but at variance with the reports by Zhu et al . for SIVmac239 [13] and HIV-1 BaL Env [16] where the stalk region was separated into the legs of a tripod that are 70 Å apart ( Figure S1 ) . In addition to discrepancies in the overall shape of the maps , there are significant differences in the molecular interpretation of the spike architectures , derived by fitting X-ray crystallographic atomic coordinates for monomeric gp120 into the density maps obtained using cryo-electron tomography . There are multiple X-ray coordinates for truncated gp120 monomers , in unliganded , as well as sCD4-liganded or antibody-bound forms [6] , [7] , [8] . Of these , the coordinates for monomeric , unliganded SIV gp120 ( PDB ID: 2BF1; [5] ) might be considered a natural choice to obtain a molecular interpretation of the unliganded HIV-1 and SIV Env trimers since these are the only published coordinates for unliganded SIV or HIV-1 gp120 . The two previously reported density maps for trimeric SIV Env were interpreted in terms of these coordinates using manual fitting . Zhu et al . [13] oriented the monomeric gp120 coordinates in the map so that the V1 and V2 loops are located at the base of the spike . Zanetti et al . [14] presented two alternate interpretations in which the V1 and V2 loops also localized to the base of the spike , but with substantial differences in the predicted positions of the V3 loop . Each of these models also differed from the theoretical model proposed by Chen et al . [5] which also had the V1/V2 loop at the base of the spike ( Figure S3c , S3d ) , but with the V3 loops of the three gp120 monomers located at the apex of the spike . Thus , in addition to the differences in overall shape of the density maps , these earlier models derived by electron tomography also provided radically different interpretations for the molecular structure of trimeric SIV Env . The differences in the final outcome can only arise from four possible sources: ( i ) procedures and instrumentation used for data collection , ( ii ) the computational procedures used to average individual tomographic spike subvolumes to arrive at the final density map , ( iii ) the choice of gp120 coordinates used to interpret the maps and ( iv ) the procedures used to fit the coordinates into the density map . Of these , the first source can be excluded because the differences in methods for data collection between the two previously published tomographic analyses [13] , [14] and the present work are not expected to alter the profile of the density maps , with only minor differences such as pixel size ( 5 . 5 Å vs . 4 . 1 Å ) and underfocus values ( 4 to 6 µm vs . 1 . 5 to 2 . 5 µm ) . The remaining three possibilities and a description of the procedure for choice and fitting of coordinates into the density maps are discussed below ( see also Figures S2 , S3 , S4 , S5 , S6 ) . The determination of density maps for the complex of trimeric Env with b12 and sCD4/17b was a critical element of our earlier analysis with antibody-bound HIV-1 spikes because the X-ray structures available for the binary [8] and ternary [7] complex , respectively , provided unambiguous tests of map quality and interpretation [15] . Since there are no obvious constraints such as the presence of complexed b12 or sCD4 for obtaining a molecular interpretation of unliganded SIV/HIV-1 Env density maps , the fitting is based on the fit with the gp120 coordinates alone . All of the available gp120 coordinates include only ∼60% of gp120 polypeptide mass , and none of the gp41 ectodomain ( ∼20 kD ) , which makes the fit of these coordinates into the density map challenging . Placing any of these gp120 coordinates into the map by manual fitting as in the work of Zhu et al . [13] or Zanetti et al . [14] is non-quantitative , since the outcome is based on subjective evaluation of the map . The automated fitting procedures we use employ correlation coefficient-based maximization , providing an objective way of obtaining the fits . Further details about the fitting procedures are also available publicly through the website of the publicly distributed software package UCSF Chimera ( http://www . cgl . ucsf . edu/chimera/hiv2009/ ) . The three possible sets of atomic coordinates that are available as starting points to carry out the automated fitting are the structure reported for unliganded SIV gp120 core ( 2BF1 ) , the structure of HIV-1 gp120 from the complex with b12 ( 2NY7 ) , or the structure of HIV-1 gp120 from the complex with sCD4/17b ( 1GC1 ) . The automated fitting procedures , identical to those used in our previous work with HIV-1 , result in a single solution when the HIV-1 gp120 coordinates are used , and no clear solution and lower correlation coefficients ( indicative of a poor fit ) when the 2BF1 SIV gp120 coordinates are used in any orientation . Comparison of the fit obtained from our work using 1GC1 coordinates ( Figures S3a , S3b ) with the theoretical model ( Figures S3c , S3d ) proposed by Chen et al . [5] based on their crystallographic structure of unliganded , truncated SIV gp120 shows that the two differ dramatically in terms of the inferred locations of the V1/V2 loop regions . In the preferred fits with 1GC1 coordinates , the V1/V2 loop regions are at the apex of the spike , while the model proposed by Chen et al . [5] places the V1/V2 loop regions closer to the base of the spike ( Figures S3c , S3d ) . These differences can be visualized more clearly in Figure S4 . The first column in Figure S4 shows density maps calculated from each of these gp120 coordinates to a resolution of 20 Å , which illustrates the appearance of each coordinates' density at this resolution when rendered as an isosurface . The gp120 density profiles obtained from the coordinates of the sCD4/17b complex ( Figure S4a ) and the b12 complex ( Figure S4b ) are rather similar when viewed at ∼20 Å resolution , even though there are important differences in their atomic resolution structures . However , both of these HIV-1 gp120 structures differ significantly from the conformation reported for monomeric , unliganded , SIV gp120 ( Figure S4c ) . When each of these coordinates are fitted into the SIVmneE11S map we present here , it is the HIV-1 gp120 coordinates determined for the complexes with sCD4/17b ( Figures S4d , S4g ) or b12 ( Figures S4e , S4h ) that show a visually better fit rather than the coordinates reported for the unliganded SIV gp120 core ( Figures S4f , S4i ) . The estimated position of the V1/V2 loops , based on the location of the truncated loop in the coordinates indicated by red spheres , also varies between the 3 sets of gp120 coordinates . The fits of both sets of HIV-1 gp120 coordinates situate the V1/V2 loop region with excellent correspondence to a region of unassigned density at the apex of the spike ( Figure S4d , S4e , S4g , S4h ) . In contrast , the estimated location of this loop in the unliganded SIV gp120 coordinates ( Figure S4f , S4i ) is not consistent with the observed architecture of the spike , falling in a region where there is no unassigned density . All three sets of coordinates also have significant deletions in the N and C-terminal regions that reside at the base of the spike . In order to further compare the different fits , we extended our analysis to use a stringent “geometric fitting” criterion [15] that evaluates the fit over a range of density thresholds ( Figure S5 ) . In cryo-electron microscopy , choice of the “right threshold” to represent density maps generally depends on the size of the protein , and on an effective “temperature-factor” , which describes the resolution-dependent fall-off in image amplitudes . The reason why the use of correlation coefficients alone can lead to false confidence in fit quality is because the density map is influenced by both parameters , and there is no a priori way to know these parameters precisely . Analyzing the map over a complete range of thresholds is thus a thorough and transparent way to illustrate map quality , and provides a better way to assess quantitative fitting coefficients . For this purpose , the fit of the monomeric HIV-1 and SIV gp120 coordinates as shown in Figures S4b , S4e , or S4f was kept fixed , and the threshold values for map visualization was progressively varied ( Figure S5 ) . At each threshold value , the proportion of atoms that fell outside the map contour was calculated . We note that the coordinate fits are to the density map and not the threshold , and therefore are identical across all thresholds . Compared to the fits obtained using 1GC1 or 2NY7 coordinates , a higher proportion of atoms are distributed outside the map contour at every threshold when 2BF1 coordinates are used to carry out the fit in the orientation roughly corresponding to that suggested by Chen et al . [5] . These results provide further validation regarding the choice of coordinates used for fitting by giving a quantitative estimate of the number of atoms that are included within the map over a range of density thresholds . These calculations do not include contributions from the V1/V2 loop which are absent in the coordinates; if this loop is included , the improved fit of the HIV-1 coordinates to the map would be further accentuated . Inclusion of the stump of the loop present in the 1GC1 coordinates also makes no difference to the fits . In summary , the molecular model for trimeric SIV and HIV-1 envelope glycoproteins derived from published coordinates for liganded monomeric HIV-1 gp120 , but not SIV gp120 is consistent with the density maps we have obtained from native virions . The most likely explanation is that the three-dimensional crystals used to determine the structure of the truncated SIV gp120 core may have captured a conformation of gp120 that is different from the conformation in the native trimer . The V1 and V2 loop regions together represent ∼15–20% of the overall polypeptide mass of gp120 . The earlier studies with b12-complexed HIV-1 BaL [15] , as well as the detailed fitting exercises described in Figure S4 provide strong evidence that the V1 and V2 loop regions are located at the apex of the spike . As a further stringent test of this assignment , we took advantage of the availability of a previously reported V1/V2 loop-deleted variant of the HIV-1 R3A strain [17] . For this purpose , a matched pair of full-length and loop-deleted viruses of the R3A strain were prepared in parallel using the same cell system , purified in parallel under the same conditions , and analyzed by cryo-electron tomography . The quaternary structure of Env from the R3A strain ( Figure 2a ) is essentially the same as that previously reported in our earlier studies of HIV-1 BaL [15] . Comparison of trimeric Env structures from V1/V2 loop-deleted HIV-1 R3A ( Figure 2b ) with the wild-type R3A shows missing density in the raw map of the deletion mutant at the predicted location and is confirmed in the difference map ( Figures 2c–d ) . These results establish unequivocally that the V1/V2 loops must be located at the top of the trimeric spike . Further , automated fitting of the density maps with 1GC1 HIV-1 gp120 coordinates results in molecular structures for trimeric Env in wild-type and V1/V2 loop-deleted variants of HIV-1 R3A that are closely comparable to each other ( Figures 2e–g ) , and to that reported for HIV-BaL [15] . Using the principles for fitting gp120 coordinates described above , we obtained a molecular model for trimeric Env from SIVmneE11S and SIVmac239 ( Figures 3a and 3b , respectively ) . Visualization of the fits of the gp120 coordinates to the map identifies the locations of the V1/V2 loop region absent in the X-ray coordinates , and the gp41 ectodomain . The fitted gp120 trimers for SIVmneE11S and SIVmac239 display similar architectures , with differences in orientations and total displacement of the fitted gp120 molecules that are within ∼10° and ∼10 Å , respectively . In turn , both are similar to the structure of trimeric HIV-1 BaL [15] . We extended the investigation of trimeric SIV spike structure to a third strain , SIV CP-MAC , a cytopathic SIV variant derived by serial passaging of SIVmac BK28 on CCR5-negative SupT1 cells [18] , [19] , [28] . SIV CP-MAC is of interest because its Env has been reported to exhibit CD4-independent utilization of rhesus CCR5 on Env-pseudotyped reporter particles [20] , in contrast to viruses such as SIVmac239 where infection is CD4-dependent . Using BC7 , a CD4-negative variant of the T-cell line SupT1 [19] , [30] that was engineered to stably express human or rhesus CCR5 , we confirmed that CP-MAC , but not SIVmac239 , could efficiently infect these cells in the absence of CD4 ( Figures 4a , 4b , S8 ) . Density maps for trimeric Env from SIV CP-MAC show that Env on this virus is present in a structurally distinct “open” conformation ( Figure 4c ) in which the three blades of the propeller are splayed outwards at the apex of the spike in striking contrast to SIVmneE11S and SIVmac239 that , display Env in a “closed” conformation ( Figure 1f , 1g respectively ) with the gp120 propeller blades pointed inwards towards the apex . Comparison of the fits of gp120 to maps of the closed ( Figures 5a , 5c ) and open states ( Figure 4d , 5b , 5d ) of trimeric Env reveal that relative to the fits obtained for the closed conformation , each of the gp120 monomers in the open SIV CP-MAC Env displays an in-plane rotation of ∼50° coupled with an out-of-plane rotation of ∼20° and upward displacement of ∼10 Å ( Figure 5 ) . In this open conformation , the V1/V2 loop is rotated away from the apex of the spike towards the periphery , the CD4 binding site is also similarly rotated , and the V3 loop is closer to the apex of the spike , poised for binding the co-receptor . Notably , the quaternary arrangement of gp120 in trimeric SIV CP-MAC Env is in essentially the same “open” conformation previously observed for trimeric gp120 on CD4/17b complexed HIV-1 BaL ( PDB ID: 3DNO; [15] ) , which represents a conformational state achieved prior to viral entry but after binding to cellular CD4 . Several residues , including those that are highly conserved across most SIV and HIV sequences , which are buried in the closed state ( Figure 5c ) become exposed in the constitutively open state ( Figure 5d ) . It is of particular interest that a swath of residues which becomes exposed upon opening at the apex of the spike include those in the vicinity of the V3 loop region , spanning regions that may potentially represent co-receptor-interactive sites . The discovery of the constitutively open conformation in SIV CP-MAC suggests that the formation of this open conformation is the underlying molecular mechanism by which this CD4-independent virus enters cells lacking CD4 , and further investigation is necessary to determine whether other immunodeficiency viruses with similar entry requirements share this molecular architecture . The finding that SIV CP-MAC Env displays a quaternary conformation that is similar to that achieved by HIV-1 subsequent to CD4 binding suggests that this should be testable by the structural analysis of complexes formed between SIV CP-MAC and antibodies that recognize potential co-receptor binding site regions . To test the functional relevance of the “open” SIV CP-MAC conformation , we therefore carried out structural analysis of the complex formed with the 7D3 neutralizing antibody , which has been proposed to target co-receptor binding sites on SIV CP-MAC gp120 in a manner similar to that of the CD4i antibody 17b [22] , [18] . Whole virion binding studies establish that 7D3 binds much more efficiently to purified SIV CP-MAC than to SIVmac239 under the conditions of the tomography experiments ( Figure 6 ) . Cryo-electron tomographic analyses of 7D3 bound viruses show that density for bound antibody is clearly visible on trimeric CP-MAC Env ( Figures 7a , 7b ) and fitting atomic coordinates for gp120 into 7D3-bound CP-MAC Env density maps shows that 7D3 binds over the base of the V3-loop ( Figures 7c , 7d ) . The overall conformation of the trimeric gp120 itself is unchanged in the complex with 7D3 , thus validating the conclusion that SIV CP-MAC Env is in a constitutively open conformation that does not undergo further significant changes upon binding of a co-receptor binding site antibody .
Knowledge of the structure and functionally relevant conformational changes of trimeric Env remains an important challenge in HIV structural biology . While there are X-ray structures reported for selected monomeric gp120 constructs , mostly in ternary complexes with sCD4 and bound Fab fragments , there is a large gap in structural information on the structures of the trimeric gp120 and gp41 as they are displayed in the surface of infectious viruses . Earlier reports of the structure of trimeric SIV Env [13] , [14] spawned a controversy [23] that appears attributable to technical problems in data analysis and lack of unambiguous fitting constraints . The results from cryo-electron tomographic analyses of trimeric SIV Env presented here ( using wedge-corrected , reference-free image classification methods [10] ) address these concerns since the strains used here are the same as those in the previous studies with SIVmac239 [13] , and with SIVmneE11S [14] . Comparative analyses of the V1/V2 loop-deleted viral strain further validates the methodology we have employed for structure determination . Our analysis of SIV Env using cryo-electron tomography combined with wedge-corrected 3D averaging and classification methods reveals that SIVmneE11S and SIVmac239 have molecular architectures similar to each other , and to that previously determined for HIV-1 BaL [15] . In contrast , SIV CP-MAC , a laboratory-adapted CD4-independent variant of SIV shows an open trimer resembling that previously reported for HIV-1 BaL Env complexed with sCD4 and 17b . An important technical result from our analyses is the demonstration that cryo-electron tomography can be used to distinguish the conformational variability of native Env trimers on intact viruses , establishing the utility of this approach to study Env structural variation in different strains at ∼20 Å resolution ( Figure S6 ) . The discovery of the open state in the CD4-independent strain SIV CP-MAC strain immediately suggests how CD4-independent viruses [21] may gain entry into target cells . A schematic summary of the proposed difference between the molecular mechanisms of cellular entry of CD4-dependent and CD4-independent viruses is presented in Figure 8 . The formation of the open state results in exposure of buried regions of gp120 that would normally only be exposed or formed transiently during virus-cell contact . Prior studies with a CD4-independent variant of HIV-1 IIIB demonstrated that in the absence of CD4 , gp120 displayed preformed epitopes that overlapped with co-receptor binding sites that are not normally exposed or formed in gp120 from other virus strains until after CD4 binding occurs [24] . Interestingly , CD4-independent strains tend to be neutralization-sensitive [18] , [25] and have been isolated from SIV-infected animals with advanced immunodeficiency [26] , from immune-privileged areas such as the CNS [27] , or upon passage under selection in cell culture [20] , [28] , [19] , [24] . Further , antibodies that react with the exposed epitopes on SIV CP-MAC Env ( Figure 5d ) that include highly conserved residues such as Asn 262 and Gly 263 have already been identified [18] . 7D3 , a potent , neutralizing antibody elicited against CP-MAC Env , binds a conformational epitope thought to reside near the CCR5 binding site [18] . Automated fits of gp120 coordinates to the 7D3-bound CP-MAC Env confirm that the footprint of 7D3 resides over the base of the V3 loop , which is known to interact with co-receptor . These density maps provide a structural foundation for the findings [18] , [24] that there is stable exposure of the co-receptor binding site in CD4-independent immunodeficiency viruses . Our results are also consistent with the idea [19] , [24] , [25] that the joint requirement of CD4 and CCR5 for HIV entry represents an evolutionary adaptation of CCR5-dependent HIV and SIV to use CD4-dependent spike opening as a mechanism to cloak conserved antigenic sites on gp120 that represent targets for neutralizing antibodies until contact occurs on the target cell , where access to antibodies is likely kinetically and sterically restricted . Successful efforts at HIV vaccine design are likely to require a good understanding of Env architecture not only at the level of the structure of a particular trimer or trimer-ligand complex , but of a range of Env-ligand complexes from a variety of SIV , HIV-1 and HIV-2 viral isolates in unliganded , sCD4-liganded and antibody-neutralized states in the physiologically relevant context of an infectious virus . Continuing technological advances in high-resolution cryo-electron microscopy in combination with advanced methods for 3D image processing [9] promise to provide increasingly sophisticated tools to address this challenge .
SIV and HIV-1 viruses were produced by infection of SupT1 cells , treated with 2 , 2′dithiodipyridine ( Aldrithiol-2 , AT-2 ) and purified by sucrose gradient centrifugation to obtain concentrated preparations ( typically ∼1011 virions/mL ) that retained functionally intact Env . AT-2 treatment renders viruses non-infectious by preferential covalent modification of internal virion proteins required for infectivity , but under conditions known to maintain the conformational and functional integrity of the viral envelope glycoprotein , including full competence for receptor dependent target cell membrane binding and fusion and viral entry [29] . The SIVmneE11S and SIVmac239 virus preparations , provided by the Biological Products Core of the AIDS and Cancer Virus Program , SAIC Frederick , Inc . , were identical to those used in the previous conflicting reports of SIV Env structure for SIVmac239 by Zhu et al . [13] and SIVmneE11S by Zanetti et al . [14] . HIV-1 R3A containing a deletion of all but the first and last amino acid of the V1/V2 loop , plus a Gly-Ala-Gly linker , has been previously described [17] . Holey carbon-coated 200 mesh grids for electron microscopy were purchased from Quantifoil GmbH ( Jena , Germany ) and glow-discharged immediately prior to specimen preparation . 7D3 antibody was purified from mouse ascites fluid with Pierce NAb Spin Kits using Protein G resin , and added to CP-MAC at a concentration of 4 µM and incubated on ice for 30 minutes . Samples were deposited on the grids at room temperature and transferred to the chamber of a Mark III Vitrobot ( FEI Company , OR ) maintained at 25°C and 100% humidity . Grids were then blotted for 6 sec and plunged into liquid ethane cooled by liquid nitrogen . The total length of time that the specimens were handled at room temperature before plunge-freezing was ∼3–4 minutes . CD4-independent replication of SIVmac239 and CP-MAC was determined by adding virus ( 50 ng of p27gag ) to SupT1 cells or BC7 ( a CD4-negative variant of SupT1 ) , that were engineered to stably express human or rhesus CCR5 [30] . Cells were washed after 18 hours to remove exogenous virus and reverse transcriptase levels in culture supernatant determined over time . Infection was also monitored directly by immunofluorescence microscopy on methanol/acetone fixed cells using an anti-p27gag monoclonal antibody ( Figure S8 ) . Based on estimated Env concentrations , whole virions were equally coated onto a 96-well plate , sealed and incubated overnight at 4°C . All subsequent steps were performed at 4°C and washes were performed with cold TNE buffer ( 100 mM Tris , 150 mM NaCl and 1 mM EDTA pH = 7 . 5 ) . Wells were washed once , blocked with 1% BSA for 1 hour , and washed three more times . Primary 7D3antibody ( 1 µg/mL ) was incubated for 1 hour and washed three times . Goat α-Mouse IgG , H and L chain alkaline phosphatase-conjugated secondary antibody ( Calbiochem EMD , 1∶5000 dilution ) was incubated for 1 hour and three washes were performed . 4-nitrophenyl phosphate ( Sigma-Aldrich , 1 mg/mL ) in alkaline phosphatase substrate buffer ( 50 mM NaHCO3 , 1 mM MgCl2 , pH = 9 . 8 ) was added at 25°C and absorbance was measured at 405 nm at various time points until saturation . Controls against non-specific binding of primary and secondary antibodies were performed and data were corrected accordingly . Data sets were collected on an energy-filtered ( Gatan ) Tecnai G2 Polara transmission electron microsocope ( FEI , Netherlands ) equipped with a 2K×2K post-energy filter CCD camera , operated at 200 kV and with the specimen maintained at −193°C . Data collection was carried out over a tilt range spanning ± 65° with tilt increments ranging from 1–2° and a defocus of 2 . 5 µm , with pixel sizes of 4 . 1 Å at the level of the specimen . Doses used for each image were between 1–2 el/Å2 . Tilt series were aligned using manual fiducial-based alignment as implemented in IMOD [31] . Protein A gold fiducials ( 10 nm ) were selected , tracked and positions refined . Tomograms were then reconstructed using R-weighted back projection . Virion centroids were identified manually and subtomograms ( 480×480×480 voxels ) containing only the virions were selected . The final number of tilt series used for reconstruction for the SIVmneE11S , SIVmac239 , SIV CP-MAC and SIV CP-MAC/7D3 maps were 21 , 54 , 57 and 37 , respectively , consisting of 136 , 305 , 257 and 208 virions respectively , yielding 3800 , 4576 , 3797 and 1600 putative spikes , respectively . For purposes of selection of Env spikes , virion subtomograms were down-sampled by a factor of 4 , denoised using edge-enhancing anisotropic diffusion as implemented in IMOD and subjected to unsupervised membrane segmentation using an energy-based three-dimensional approach [32] . In order to identify the location of spikes in an automated manner , a scalar value was associated to every point on the segmented virion surface corresponding to the cross-correlation value between an external 3D template and the image data immediately outside the membrane . Spikes were identified at the locations corresponding to the local maxima of this function that were above a given threshold . A cylindrically symmetric phantom was used as a template for the search; the same template was used for all maps . Subvolumes ( 100×100×100 voxels ) corresponding to reconstructions of individual spikes ( without denoising or binning ) were cut from the virion subtomograms at the automatically extracted positions . The orientations of the long axis of the spike were determined using the normal to the automatically segmented membrane at the location of each spike , providing initial estimates for two of the three Euler angles . The remaining in-plane rotation was initially randomized to prevent any possible bias in subsequent alignments . After application of the Euler angles , sub-volumes were translationally aligned to their cylindrically averaged global average to ensure they all shared the same center of mass . The 10% of subvolumes that correlated most poorly with the updated global average were left out of the analysis . Spike volumes were aligned and classified without using external references and with proper accounting of the missing wedge using the framework described in Bartesaghi et al . [10] . Subvolume alignments were progressively refined at each iteration and spike volumes repeatedly clustered into 10 classes . Early stages of classification clearly showed classes with inherent 3-fold symmetry , and typically at the fourth iteration , 3-fold symmetry was imposed . At each round , the classes that showed the most clearly delineated features in all regions of the spike ( typically ∼50–60% ) were selected and combined to be used as reference for the next round . Typically ∼4000 spikes were selected for each dataset . Final maps were obtained after ∼5–12 refinement rounds and included contributions from ∼50% of sub-volumes in each dataset . The computational procedures we have used to obtain 3D density maps by averaging tomographic subvolumes were extensively tested first using simulated phantom objects to develop robust alignment and classification routines that take into account the missing wedge data and faithfully recover 3D structures from a series of heterogeneous 3D objects [10] . Experimental assessment of the computational developments were followed by obtaining 3D tomographic models of purified GroEL , a multimeric molecular complex whose structure is known at high resolution from X-ray crystallographic analysis , thereby enabling quantitative evaluation of each map . This computational approach yielded a 26 Å GroEL structure from as few as ∼300 volumes and without imposition of 7-fold symmetry inherent to the GroEL complex [10] . Application of these methods to ∼3000–4000 tomographic volumes of trimeric Env from SIV/HIV-1 strains resulted in maps at a resolution of ∼20 Å ( Figure S6 ) . The classes are relatively homogeneous at the end of the refinement as illustrated by inspecting sections through the maps corresponding to each class average ( Figure S7 ) . Notably , the computational procedures used by Zhu et al . [13] did not employ procedures to correct for the missing wedge , did not use either reference-free methods for 3D alignment of the individual spike subvolumes , or reference-free methods for image classification . The procedures used by Zanetti et al . [14] used missing wedge correction , but did not use image classification within the iterative alignment procedure . Steepest-ascent local optimization , as implemented in UCSF Chimera was utilized for fitting coordinates into density maps [33] . Coordinates were initially placed in random orientations and local maxima of the sum of pointwise products between the coordinates and the map were determined . Fitting was performed to convergence by performing multiples of 100 steepest ascent steps . Atomic coordinates were fit by generating a map simulated from the atomic coordinates at 20 Å . The fits shown in Figure 2–5 and Figure 7 were carried out using 1GC1 coordinates . These coordinates contain a truncated version ( residues 119–129 , and 194–202 ) of the V1/V2 loop region ( spanning residues 119–202 ) . To eliminate any bias in the fits from the inclusion of the partial loop residues present in the coordinates , they were excluded for purposes of coordinate fitting; even with their inclusion there were no significant changes in the fits . To fit the 7D3 density , the atomic coordinates for 17b were utilized and originally placed in difference density calculated from subtracting the CP-MAC density map from the 7D3-bound CP-MAC map and the 7D3 position was further refined to convergence . | HIV and SIV contact and infect target T-cells following the binding of trimeric Env spikes displayed on the viral membrane with cellular receptors . The conformational changes in trimeric Env that are triggered by the interaction between trimeric Env and cell surface receptors lead ultimately to fusion of the viral and cell membranes and delivery of the viral core into infected cells . Knowledge of the molecular structures of trimeric Env at different stages of virus-cell contact is therefore of fundamental interest for defining viral entry mechanisms and vaccine design . Cryo-electron tomography is a powerful structural tool to determine the structures of viral spikes when they are present on the surface of intact virions . Using this approach , we have determined the molecular structures of several SIV and HIV-1 strains , including an SIV strain that does not require cell surface receptor CD4 for entry and infection . Our results represent the first experimental demonstration that strain differences can result in distinct unliganded spike conformation as displayed on the surface of intact virions . The differences in structure between the different strains correlate with functional differences displayed by the viruses , and suggest a novel molecular explanation for the mechanism of CD4-independent viral entry . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/macromolecular",
"structure",
"analysis",
"virology/vaccines",
"virology/virion",
"structure,",
"assembly,",
"and",
"egress",
"virology/immunodeficiency",
"viruses",
"infectious",
"diseases/hiv",
"infection",
"and",
"aids",
"infectious",
"diseases/viral",
"infections",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines",
"virology/immune",
"evasion"
] | 2010 | Molecular Architectures of Trimeric SIV and HIV-1 Envelope Glycoproteins on Intact Viruses: Strain-Dependent Variation in Quaternary Structure |
Dengue fever ( DEN ) is the most common arboviral disease in the world and dengue virus ( DENV ) causes 390 million annual infections around the world , of which 240 million are inapparent and 96 million are symptomatic . During the past decade a changing epidemiological pattern has been observed in Africa , with DEN outbreaks reported in all regions . In Senegal , all DENV serotypes have been reported . These important changes in the epidemiological profile of DEN are occurring in a context where there is no qualified vaccine against DEN . Further there is significant gap of knowledge on the vector bionomics and transmission dynamics in the African region to effectively prevent and control epidemics . Except for DENV-2 , few studies have been performed with serotypes 1 , 3 , and 4 , so this study was undertaken to fill out this gap . We assessed the vector competence of Aedes ( Diceromyia ) furcifer , Ae . ( Diceromyia ) taylori , Ae . ( Stegomyia ) luteocephalus , sylvatic and urban Ae . ( Stegomyia ) aegypti populations from Senegal for DENV-1 , DENV-3 and DENV-4 using experimental oral infection . Whole bodies and wings/legs were tested for DENV presence by cell culture assays and saliva samples were tested by real time RT-PCR to estimate infection , disseminated infection and transmission rates . Our results revealed a low capacity of sylvatic and urban Aedes mosquitoes from Senegal to transmit DENV-1 , DENV-3 and DENV-4 and an impact of infection on their mortality . The highest potential transmission rate was 20% despite the high susceptibility and disseminated infection rates up to 93 . 7% for the 3 Ae . aegypti populations tested , and 84 . 6% for the sylvatic vectors Ae . furcifer , Ae . taylori and Ae . luteocephalus .
Dengue fever ( DEN ) is the most common arboviral disease in the world and is caused by four genetically distinct serotypes of virus ( DENV-1 , DENV-2 , DENV-3 , DENV-4 ) belonging to the genus Flavivirus of the family Flaviviridae . Among the 390 million annual infections estimated around the world , 240 million are inapparent and only 96 million are symptomatic [1] . Dengue fever causes a wide clinical spectrum similar for the four serotypes . The different clinical manifestations of DENV infection range from asymptomatic to several symptomatic forms ranging in severity from classical dengue fever , to Dengue Hemorrhagic Fever ( DHF ) and Dengue Shock Syndrome ( DSS ) . Dengue viruses are transmitted to humans by mosquitoes of the genus Aedes , mainly by the peridomestic mosquito Aedes aegypti aegypti and secondarily by Ae . albopictus and other anthropophilic Aedes mosquitoes . In Africa , the sylvatic circulation of DENV-2 appears to be predominant [2] in contrast to Asia and South America where endemic/epidemic DENV strains circulating in peridomestic cycles are most common , and a sylvatic , nonhuman primate-amplified enzootic cycle has not been identified except for in Malaysia . The first isolations of DENV-2 from naturally infected mosquitoes in Africa date to 1969 when two strains were isolated from Ibadan and Jos in Nigeria [3] . Thereafter , several epizootics of DENV-2 were reported through the periodic amplifications of the sylvatic cycle involving wild populations of mosquitoes and monkeys in several West African countries [4] . However , despite these frequent epizooties and the presence of the epidemic vector Ae . aegypti in all bioclimatic areas , only sporadic DEN cases were recorded in West Africa . This could be explained by the presence of Aedes aegypti formosus , the ancestral African sylvatic and zoophilic form that uses tree holes as its larval habitat . Indeed , both sub-species exist in Africa but the presence of Aedes aegypti aegypti ( the domestic , highly anthropophilic and primarily endophilic subspecies ) in West Africa remains debatable mainly because of the lack of reliable methods to distinguish the two subspecies . The first documented outbreak caused by DENV-2 in West Africa occurred in Burkina Faso in 1982 and was suspected to be triggered by an introduction from the east of an epidemic Seychelles strain [2] . Most African DEN outbreaks caused by DENV-2 have occurred in East Africa . The others DENV serotypes ( 1 , 3 and 4 ) are only known from endemic-epidemic cycles in Africa with no evidence of enzootic circulation . Only DENV-1 has been found associated with Ae . aegypti . During the last century , DENV-1 epidemics were notified in South Africa in 1926–27 , Sudan in 1984 , and Nigeria in 1964 and 1975 while the unique DENV-3 outbreaks occurred in Mozambique in 1985 [5 , 6] . Serotype 4 was only reported in Senegal in contexts which still remains enigmatic [7] . Amarasinghe et al . 2011 [6] have presented an exhaustive review on dengue situation in Africa . Over the last 2 decades a changing epidemiological pattern has been observed in Africa , with outbreaks of DEN reported in all regions and several cases exported to Europe [8] . DENV-2 , responsible for several epidemics in East Africa ( Somalia , Djibouti , Kenya and Tanzania ) and usually circulating in a sylvatic cycle ( between Aedes mosquitoes and non human primates ) in West Africa , spilled over into urban areas in 2014–2015 in Senegal and Mauritania , Gabon in 2007 , Angola in 2013 and Burkina Faso in 2016 . Serotype 3 ( DENV-3 ) , never reported in Africa after its first emergence in 1985 in Mozambique , caused a major urban outbreak in 2009 in Cape Verde , Cote d’Ivoire , Gabon and Senegal . Since September 2017 , Burkina Faso and Senegal face up to major urban outbreaks Ouagadougou and Louga respectively ( S1 Table ) [9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23] . In Senegal all DENV serotypes have been reported ( S2 Table ) . These important changes summarized above in the epidemiological African profile of DEN are occurring in a context where there is no vaccine against DENV recommended for all populations . Furthermore , there is a significant gap of knowledge on DENV vector bionomics and transmission dynamics in Africa to effectively prevent and control epidemics . The vector competence of mosquitoes associated with DENV in nature is poorly characterized . Except for DENV-2 [24 , 25] few studies have been performed with serotypes 1 , 3 , and 4 [26] . Following the 2009 Dakar DENV-3 epidemic , we initiated a vector competence study to evaluate the ability of Ae . aegypti populations from Dakar and Kedougou to transmit DENV-1 and -3 , for which there is no evidence of enzootic , sylvatic circulation in Africa [27]; these two serotypes appear to circulate only in an endemic/epidemic cycle with peridomestic human amplification . Our prior results showed low susceptibility to DENV-3 but high infection and dissemination rates with DENV-1 . However , the oral DENV doses used were low and transmission potential was not tested . Furthermore , only Ae . aegypti was tested and vector competence data for sylvatic vectors were generated for DENV-1 , -3 and -4 . Thereby the present study assessed the vector competence of Senegalese Ae . aegypti , Ae . furcifer , Ae . taylori and Ae . luteocephalus for DENV-1 , -3 and -4 .
The University of Texas Medical Branch ( UTMB ) Institutional Animal Care and Use Committee approved all experiments involving animal-derived cells/tissues/sera/samples under protocol 02-09-068 . UTMB complies with all applicable regulatory provisions of the U . S . Department of Agriculture ( USDA ) -Animal Welfare Act; the National Institutes of Health ( NIH ) , Office of Laboratory Animal Welfare-Public Health Service ( PHS ) Policy on Humane Care and Use of Laboratory Animals; the U . S Government Principles for the Utilization and Care of Vertebrate Animals Used in Research , Teaching , and Testing developed by the Interagency Research Animal Committee ( IRAC ) , and other federal statutes and state regulations relating to animal research . The animal care and use program at UTMB conducts reviews involving animals in accordance with the Guide for the Care and Use of Laboratory Animals ( 2011 ) published by the National Research Council . Mosquito species used in this study were collected from three Senegalese localities: Dakar , Saint Louis and Kedougou ( Fig 1 ) . The Table 1 describes the characteristics and geographic origins of Ae . aegypti , Ae . furcifer , Ae . taylori , and Ae . luteocephalus populations tested . The sylvatic Ae . aegypti population from Kedougou breeding in tree holes represented Ae . aegypti formosus morphologically characterised by the lack of pales scales on the first abdominal tergite and the urban populations from Dakar and Saint Louis breeding in artificial containers were consistent with Ae . aegypti aegypti contrariwise characterised by the presence of pales scales . These species were chosen based on their abundance , anthrophophilic behavior and association with DENV in nature . For each population , several larval habitats were sampled and immature stages were collected and reared in the laboratory . For Ae . furcifer , Ae . taylori and Ae . luteocephalus adult females were caught in a gallery forest at Kedougou and reared in the laboratory . Progeny of these populations were considered as the F1 generation that we used for experimental infections . Adult mosquitoes were maintained with a 10% sucrose solution at 27 °C , 75–80% relative humidity ( RH ) , 12:12 h ( Light:Dark ) photoperiod . Hosts origin , year of collection and passage histories of the virus strains used in this study are presented in Table 2 . DENV-1 , DENV-3 and DENV-4 strains obtained from the World Reference Center for Emerging Viruses and Arboviruses at the University of Texas Medical Branch in Galveston , Texas . For the Ae . furcifer experiment we used the DENV-4 strain from Haiti ( Haiti 73 ) , and DENV-3 strain from Barbados in the Caribbean region of North America ( Carec 01–11828 ) . For other mosquito species we used the following African strains: DENV-1 ( SH 29177 ) ; DENV-3 strain ( S-162 TvP-3622 ) , and DENV-4 strain SH 38549 . An additional passage on C6/36 cells was performed for each strain to obtain the viral stock used to infect mosquitoes . Cell lines were provided by the American Type Culture Collection ( Manassas , Va . ) , and cultured in Gibco DMEM ( Dulbecco’s Modified Eagle Medium ) , High glucose ( Gibco Cat . No . 11965–092 ) supplemented with 10% fetal bovine serum ( FBS; Atlanta Biologicals Cat . No . S11150 ) heat-inactivated in 56° C water bath for 60 min , penicillin-streptomycin ( Gibco Cat . No . 15140–122 ) and 10% of Bacto Tryptose Phosphate Broth ( Becton , USA ) . Virus in cell culture supernatants was concentrated using Millipore UFC910024 Amicon Ultra-15 Centrifugal Filter Concentrator with Ultracel 100 Regenerated Cellulose Membrane . Concentrated viruses were collected , aliquoted and frozen at -80°C , and used as viral stocks for mosquito infection . Virus stocks were titrated using the method focus forming assays and immunostaining described below . Virus titers for stocks and infectious blood meals after 1 hour of exposure to mosquitoes were determined by focus forming assays and immunostaining as described previously [28] . Briefly , Ten-fold serial dilutions of virus in MEM supplemented with 2% FBS and antibiotics ( Invitrogen , Carlsbad , CA ) , were added in duplicate to confluent C6/36 cell monolayers attached to 24-well Costar ( Corning , NY ) plates , and incubated for 1 h with periodic gentle rocking to facilitate virus adsorption at 28 °C . Wells were then overlaid with 1 ml of 0 . 8% methylcellulose ( Sigma-Aldrich , St . Louis ) diluted in warm Optimem ( Invitrogen ) supplemented with 2% FBS , antibiotics and 1% ( w/v ) L-glutamine and incubated undisturbed for 4 days at 28 °C . Methylcellulose overlay was aspirated and cell monolayer rinsed once with phosphate buffered saline ( PBS ) , pH 7 . 4 ( Invitrogen ) followed by fixation with a mixture of ice-cold acetone and methanol ( 1:1 ) solution and allowed to incubate for 30 min at room temperature ( RT ) . Fixation solution was aspirated and plates were allowed to air dry . Plates were washed thrice with PBS supplemented with 3% FBS , followed by hour-long incubation with a dengue-specific hyperimmune mouse ascitic fluid . Mouse hyperimmune sera ( MIAF ) to DENV were prepared in adult mice; using 10% crude homogenates of DENV- infected newborn mouse brain in phosphate-buffered saline as the immunogen . The immunization schedule consisted of four intraperitoneal injections of antigen mixed with Freund’s adjuvant , given at weekly intervals . After the final immunization , mice were inoculated with sarcoma 180 cells , and the resulting immune ascitic fluids were collected . All animal work was done at UTMB under an IACUC approved animal use protocol ( number 9505045 ) . Plates were washed thrice followed by hour-long incubation with a secondary antibody , goat anti-mouse conjugated to horseradish peroxidase ( HRP ) ( KPL , Gaithersburg , MD ) . Detection proceeded with the addition of aminoethylcarbazole ( AEC ) substrate ( ENZO Life sciences , Farmingdale , CT ) prepared according to vendor instructions . Three- to 5-day-old F1 female mosquitoes were placed into 500 mL cardboard containers and sucrose-starved for 48 hours before being exposed to an infectious artificial blood meal ( Hemotek Ltd , UK ) using BALB/c mouse skins obtained from the University of Texas Medical Branch Animal Resource Center , as membranes . The blood meal contained a 33% volume of washed sheep erythrocytes and a 33% volume of a cell culture-derived virus stock supplemented with 21% FBS , and adenosine triphosphate ( ATP ) to a final concentration of 0 . 005 M as a phagostimulant , and sucrose at a final concentration of 10% . After feeding for up to 60 minutes , the remaining blood meal was kept at– 80 °C for virus titration using plaque assay then mosquitoes were cold-anaesthetized and fully engorged specimens were incubated with 10% sucrose at 27°±1°C , a relative humidity of 70–75% and 12:12 h ( Light:Dark ) photoperiod . At 7 or 15 days post bloodmeal ( dpbm ) , mosquitoes were cold-anaesthetized and their legs and wings were removed . The proboscis of each mosquito was then inserted into a capillary tube containing 1–2 μL of FBS for salivation for up to 30 min then expectorated saliva was collected into a tube containing 100 μL of DMEM supplemented with 5% FBS . Detection of DENV in the mosquito body but not the wings/legs indicated a non-disseminated infection ( limited to the midgut ) , whereas the presence of virus in both the body and wings/legs indicated dissemination into the hemocoel . Mosquito bodies as well as wings/legs of infected bodies were tested for DENV after homogenization in 400 μl of MEM containing 5% of FBS , and centrifugation for 2 min at 11 , 500 x g at 4 °C to separate virus supernatant and debris . For each sample , 100 μl of supernatant were cultured in 24-well plates containing Vero cell monolayers and DENV was detected by focus forming assays and immunostaining described above , but without the ten-fold serial dilutions . So detection was limited to presence/absence revelation . Saliva of infected wings/legs were tested to detect DENV presence by real-time RT-PCR using an internal control of 10 no-infected mosquito saliva pooled together; 100 μl of each sample was used for RNA extraction using the QIAamp Viral RNA Extraction Kit ( QIAgen , Heiden , Germany ) , according to the manufacturer’s protocol . Dengue virus RNAs extracted from mosquito saliva were amplified using Bio-Rad iTaq universal probes one-step kit ( Cat#172–5141 ) following Manufacturer’s protocol . For detecting DENV-1 and DENV-3 , forward primer ( 5’ATTAGAGAGCAGATCTCTG 3’ ) , reverse primer ( 5’TGACACGCGGTTTC 3’ ) , and Probe 5’/56-FAM/TCAATATGCTGAAACGCG/3BHQ_1/-3’ were used; for DENV-4 , forward primer 5’AAT AGA GAG CAG ATC TCTG 3’ was used . The RT‐PCR was performed by Quant Studio 6 Flex instrument made from applied BioSystems by life technologies . The cycling conditions were RT step at 50 . 0 °C for 10 min , at 95 . 0 °C for 3 min , and 43 cycles of 15 s at 94 . 0 °C and 1 min at 55 °C . During our experiment with Ae . furcifer , we observed 5 days after oral DENV exposure a high mortality rate . Based on this observation , we planned subsequent experiments to include a negative control cohort exposed to uninfected blood meals to assess the effect of DENV on mortality . The uninfected blood meals used as the negative control contained a 33% volume of washed sheep erythrocytes and 33% volume of cell culture media ( Gibco DMEM , High glucose supplemented with 10% fetal bovine serum , penicillin-streptomycin and 10% of Bacto Tryptose Phosphate Broth ) supplemented with 21% FBS , and adenosine triphosphate ( ATP ) to a final concentration of 0 . 005 M as a phagostimulant , and sucrose at a final concentration of 10% . The Table 3 showed the sample size for each virus strain and for each mosquito populations . These mosquitoes were monitored twice daily for mortality until 15 dpbm for Ae . taylori , Ae . aegypti from Kedougou and St . Louis and 20 dpbm for Ae . aegypti from Dakar , then surviving mosquitoes were tested for DENV infection as described above . Infection ( number of positive bodies/total number of engorged mosquitoes incubated and tested ) , disseminated infection ( number of mosquitoes with positive wings-legs/ total number of engorged mosquitoes incubated and tested ) and transmission ( number of mosquitoes with infected saliva/ total number of engorged mosquitoes incubated and tested ) rates were calculated for each species and each dpbm . The rates obtained were compared using Fisher’s exact test . For Ae . aegypti populations potential impact of the virus serotype , incubation and mosquito origin were estimated using beta regression model . A Wilcoxon test was performed to compare differences between survivals among groups . For all tests , differences were considered statistically significant at p < 0 . 05 using R v . 2 . 15 . 1 ( R Foundation for Statistical Computing , Vienna , Austria ) [29] .
The titers of DENV stocks used ranged between 107 and 108 PFU/ml and the Table 4 presented the titers of the blood meals prepared from these stocks after 1-hour exposure at 37±1 °C for mosquitoes feeding in different days . These titers ranged between 1 . 2 x 106 and 4 . 7 x 107 PFU/ml . A total of 606 Ae . aegytpi ( 240 from Dakar , 206 from St . Louis and 160 from Kedougou ) , 86 Ae . taylori , 71 Ae . furcifer and 22 Ae . luteocephalus was tested after DENV exposure and incubation for 7 or 15 days . For Ae . aegypti , the minimum and maximum values of infection rates were 87 . 5–92 . 5% and 90–95% for the population from Dakar , 62 . 5–71 . 42% and 88 . 23–100% for the St . Louis population , and 70–80% and 87 . 5–100% for the Kedougou population , respectively , at 7 and 15 dpbm ( Fig 2 ) . Disseminated infection rates were 57 . 5–67 . 5% and 60–72 . 5% for the population from Dakar , 50–62 . 85% and 86 . 66–93 . 75% for the population from St . Louis and 50–56 . 66% and 66 . 66–93 . 33% for population from Kedougou respectively at 7 and 15 dpbm . While the infection and dissemination rates were high , the potential transmission ( saliva infection ) rates were globally low ( 0–20% ) , 0–5% for Dakar , 0–2 . 85% and 0–5 . 88% for St . Louis , and 0% and 0–3 . 33% for Kedougou , respectively at 7 and 15 dpbm . Results showed that all species were susceptible to disseminated infection with DENV-1 , -3 and -4 ( Fig 2 and S1 Fig ) . Ae . aegytpi population from Dakar showed higher infection rates ( IR ) than populations from St . Louis and Kedougou for all 3 dengue serotypes at 7 dpbm . However , differences were significant only between the Dakar and St . Louis populations for DENV-1 ( Fisher’s exact test: p = 0 . 01 ) and DENV-3 ( Fisher’s exact test: p = 0 . 003 ) . Infection rates of Ae . aegypti populations increased significantly between 7 and 15 dpbm for all 3 serotypes except for the population from Dakar . At 15 dpbm , IRs of the 3 populations did not differ significantly ( Fisher’s exact test: p > 0 . 05 ) . For all Ae . aegypti populations , IRs with DENV-3 were higher than those obtained with DENV-1 and DENV-4 . However , the difference was statistically significant only for Ae . aegypti from Kedougou when we compare IR obtained with DENV-3 versus DENV-1 ( Fisher’s exact test: p = 0 . 04 ) . The minimum and maximum values of disseminated infection rates of the 3 populations of Ae . aegypti were 50–65% for DENV-1 , 50–57 . 5% for DENV-3 and 56 . 66–67 . 5% for DENV-4 and were statistically comparable at 7 dpbm ( Fisher’s exact test: p > 0 . 05 ) , while at 15 dpbm Ae . aegypti from St . Louis showed significantly higher DIR than populations from Dakar ( Fisher’s exact test: p = 0 . 001 ) and Kedougou ( Fisher’s exact test: p = 0 . 005 ) for DENV-4 . With Ae . aegypti populations from Kedougou and St . Louis the IRs and DIRs increased between 7 and 15 dpbm , however the population from Dakar were susceptible to infection and developed disseminated infection with same rates at 7 and 15 dpbm . Among the sylvatic vectors , Ae . furcifer showed the highest IRs for DENV-3 and DENV-4 but differences were not significant ( Fisher’s exact test: p > 0 . 05 ) . No significant difference was observed for DENV-4 infection and dissemination among the three species despite the higher IR with Ae . furcifer and lower with Ae . luteocephalus . The Fig 3 shows titers of the infected saliva . Globally we observed mainly for Ae . aegypti a decreasing of titer between 7 dpbm and 15 dpbm . The highest titer ( 29 PFU/ml ) was observed with Ae . taylori . For Ae . aegypti highest titers were 15 and 16 PFU/ml for Dakar and Kedougou mosquito strains at 7 and 15 dpbm respectively . The regression model did not reveal an effect of the virus serotype on the infection rates of Ae . aegypti populations ( Table 5 ) . However , odds ratio of mosquito strain from Saint-Louis versus mosquito strains from Dakar , decreases significantly by a factor of 0 . 4 ( p<0 . 001 ) while relative proportion of infected mosquitoes increases by a factor of 3 . 5 at 15 dpbm compared to 7 dpbm ( p<0 . 001 ) . For the dissemination rates , no effects of the virus serotype and mosquito origin were observed . The incubation period was the unique parameter affecting the dissemination with an odds ratio at 15 dpbm increasing by a factor of 2 . 59 compared to 7 dpbm . No statistically significant relationship between transmission rate and mosquito origin , virus strains and dpbm . When we compared survival of Ae . aegypti mosquitoes from Dakar exposed to DENV-1 , -3 and -4 with infection rates of 92 . 68 , 93 . 02 and 91 . 66% , respectively , to that of the unexposed control group , globally the difference was statistically significant ( Wilcoxon test: p <0 . 001 ) ( Fig 4A ) . For Ae . aegypti mosquitoes from St . Louis exposed to DENV-1 , -3 and -4 with infection rates of 88 . 23 , 100 and 93 . 75% respectively compared to unexposed group ( Fig 4B ) , significantly higher mortality was observed ( Wilcoxon test: p = 2 . 57x10-12 ) . For Ae . aegypti from Kedougou ( Fig 4C ) exposed to DENV-1 , -3 , -4 with infection rates of 87 . 5 , 100 , 91 . 66% , respectively , mortality was also significantly higher than that of the negative controls ( Wilcoxon test: p = 0 . 0009 ) . The difference was also overall significant for Ae . taylori ( Wilcoxon test: p = 0 . 01 ) ( Fig 4D ) , which showed infection rates of 68 , 76 . 66 , and 83 . 87% for DENV-1 , DENV-3 and DENV-4 respectively . For Ae . aegypti from Kedougou and Ae . taylori , our analysis did not reveal any effect between DENV serotypes and the mosquito population survival rate ( p = 0 . 344 and p = 0 . 378 respectively ) . But survivals were significantly different between DENV serotypes for Ae . aegypti Dakar ( p <0 . 001 ) and Ae . aegypti Saint Louis ( p = 0 . 002 ) . For Ae . aegypti from Dakar the DENV-1 induced the highest mortality ( Wilcoxon test: p<0 . 001 ) . Kaplan–Meier survival curves are shown that mortality was higher for Ae . aegypti from Dakar exposed to DENV versus unexposed . Also , survival of this population was more affected by DENV-1 than by DENV-4 and DENV-3 . Our results showed that DENV infection also affected the survival of Ae . aegypti from St . Louis . Survival of mosquitoes exposed to all three DENV was reduced compared to negative controls from the 6th dpbm . Ae . aegypti from Kedougou and exposed to DENV-4 survived better until 9 dpbm , then mortality increased highly compared to controls from the 11th dpbm . Mosquitoes exposed to DENV-1 had reduced survival early compared to DENV-3 and DENV-4 . From the 11th dpbm , survival of Ae . aegypti from Kedougou was significantly affected by all three DENV serotypes . Survival of Ae . taylori exposed to DENV-4 or -3 was significantly lower than controls at all dpbm , but there was no significant difference between DENV serotypes . While mosquitoes exposed to DENV-1 showed reduced survival for the first 7 dpbm , no significant difference was observed later .
Our study provides important information on the vector competence of both sylvatic and domestic populations of Ae . aegypti and three sylvatic species of Aedes while some aspect could be considered as minor limitations . First , due to limited number of specimens available we were not able to test Ae . furcifer with the African DENV strains after an early experiment performed with DENV-4 and DENV-3 strains respectively from Haiti and Barbados . For the same reason we also limited the experiment with Ae . furcifer , Ae . taylori and Ae . luteocephalus at 15 dpbm only . Furthermore , only the RT-PCR was used to detect DENV genomes whether infectious particles or not for 2 reasons: i ) the purpose of this article is to show the competence of the vector and we have been focused on the detection of DENV in the different compartments of the mosquitoes and in the saliva . As we have shown that the virus reached the saliva , it implies that the vector is competent ii ) In our experience with other viruses , ( West Nile , Usutu ) , we have noticed that RT-PCR and infectious viral particles are generally very consistent and concordant in their conclusions and trends [30 , 31] . Such a trend has also been confirmed on C6/36 cells for other virus [32] . Despite the relevance of the capillary feeding method for virus transmission assessment there is no proof of salivation activity for each tested individual . However , as the use of animals has many limitations , it is currently the best alternative technique [33 , 34] , successfully used over years with different media like defibrinated blood [35] , mineral or immersion oils [36 , 37] , foetal bovine serum [38] to measure virus transmission . One way to assess the presence of saliva could be the detection of saliva components like protein or carbohydrates . Such an approach will require , however , a larger volume of media for saliva collection to achieve the different analysis without compromising virus detection . During our experiments , no DENV-susceptible laboratory mosquito strain was used as control . In our knowledge , there is no unique mosquito strain that can serve as a single control for each of the DENV serotypes and genotypes . Beyond the current scope of the study , future experiments would take into account this aspect by integrating at least one laboratory susceptible mosquito strain for each DENV serotype . Our results showed high infection and disseminated infection rates with DENV serotypes 1 , 3 and 4 , both for Ae . aegypti populations and for the sylvatic mosquito vectors Ae . furcifer , Ae . taylori and Ae . luteocephalus . IRs with Ae . aegypti population from Dakar reached their maximum values as early as 7 dpi , while for Ae . aegypti from Kedougou and St . Louis , IRs increased between 7 dpbm and 15 dpbm to reach their maximum values later . Ae . aegypti mosquitoes from Dakar develop DENV infection earlier than populations from Kedougou and St . Louis and this could be explained by highest blood meal titer for Ae . aegypti aegypti Dakar than others populations . However differences were not significant between populations from Dakar and Kedougou for DENV-1 and -3 , as observed earlier [27] , but IRs and DIRs were higher in the present study . This could be explained by difference of virus strains and especially by high oral virus doses ( 106–107 PFU/ml ) compared to those used previously ( 103–104 MID50/ml ( Mice Infectious Dose 50 ) ) with different Ae . aegypti populations from Thailand [39] . Moreover Ae . aegypti populations from western ( Burkina Faso: Koro , Bobo and Kari mosquito strains ) and eastern ( Kenya: Rabai and Shimba Hills mosquito strains ) Africa showed lower rates despite the same viral titers ( 107 . 3–108 . 1 MID50/ml ) . Similarly in many DEN-endemic countries infection and dissemination rates obtained [40] were lower than those showed by this study . Comparisons between serotypes show that Senegalese Ae . aegypti were more susceptibility to DENV-3 than to other serotypes . Also , DENV-3 was detected in the saliva of all three populations; this could explain recent DENV-3 outbreaks in many African countries ( S1 Table ) . The St . Louis population showed transmission potential for all three DENVs even if TRs were low , suggesting a salivary gland barrier within some Ae . aegypti populations . However , the population from Dakar showed higher TRs than other populations and at 7 dpbm these rates for DENV-4 reached 20% of total engorged mosquitoes and 7 . 5% for DENV-3 . These transmission rates observed just a week after mosquito infection may explain the Dakar DENV-3 epidemic like the Cape Verde Ae . aegypti population transmitting during the 2009 outbreak while it showed infection rates of 0% at 7 dpbm and until 10 dpbm the transmission rates did not exceed 20% of only mosquitoes which disseminated the infection [41] . The potential transmission rates obtained with DENV-4 show that even if a large outbreak has not been reported , the risk of a DENV-4 epidemic is present in Dakar and St . Louis because , as noted before , even with low transmission rates a vector can cause epidemics based on its abundance , density , survival and human feeding frequency [42] . Regarding enzootic , sylvatic vectors , infection and dissemination rates were relatively high for the different serotypes tested ( Fig 2 and S1 Fig ) . The same population of Ae . furcifer , with different DENV-2 strains , showed similar IRs ( 26 to 97% ) and DIRs ( 17 to 75% ) to the rates we obtained [24] . But in our study we did not detect transmission potential by Ae . furcifer . The decreasing of viral titers in mosquito saliva observed at least in Aedes aegypti between 7 and 15 dpbm may explain the low transmission rates obtained . Vectorial capacity is the efficiency of a vector in the transmission of a pathogen due to the combined effects of many factors , both intrinsic and extrinsic . The mortality rate is an important component [43 , 44 , 45] . Even if vectors become infected after taking an infectious blood meal , if they fail to survive to bite another host , the potential for transmission by this population is low . As a result , changes in the mosquito mortality rate would directly affect transmission of the pathogen . In our study we found that the exposure of Ae . aegypti from Dakar , Kedougou and St . Louis to the 3 DENV serotypes significantly increases mortality compared to negative control cohorts exposed to uninfected blood meals . Adverse effects on the fitness of Ae . aegypti due to DENV infection were also reported previously [46] . We also found increased mortality of infected Ae . taylori mosquitoes . Several other studies showed effects of other arboviruses on the survival of mosquito vectors [47 , 48] . Our results showed that for all 3 populations of Ae . aegypti , DENV-1 exposure affects mosquito survival . However , for the Ae . taylori population , after 7 dpbm we no longer detected an effect on survival of mosquitoes with DENV-1 infection . This absence of effect of DENV-1 infection on Ae . taylori is surprising because this species is not normally adapted to this DENV serotype , which is not known to circulate in the forest galleries frequented by Ae . taylori in Africa [49] . Indeed , only DENV-2 has been shown to circulate regularly in a sylvatic cycle in southeastern Senegal in the Kedougou region [7 , 50] . In summary , our results indicate that DENV-4 exposure did not affect survival of Ae . aegypti from Kedougou before 9 dpbm but it affected early the survival of the Ae . aegypti populations from Dakar and Saint Louis and Ae . taylori . DENV-3 caused high mortality in all mosquito populations tested , mainly in Ae . taylori and Ae . aegypti from Dakar . Survival was been most affected by DENV-1 which showed the highest potential transmission rates . | Dengue fever remains a major public health problem in all tropical regions of the world and causes 390 million infections each year . In Africa , while dengue outbreaks were rare during the last century , recurrent urban epidemic have been reported in all regions the last decade . Serotype 3 , never reported in West Africa , caused major outbreaks in 2009 in several capital cities while serotype 2 , usually confined to the forest cycle , spilled over into urban areas in Senegal and Mauritania in 2014–2015 . These changes are occurring in a context where vector control remains the only effective approach to prevent and control epidemics . However , the design and the implementation of efficient vector control require an accurate knowledge of the vector bionomics and competence while such data are lacking in the African region . To fill out this gap we experimentally infected domestic and wild mosquitoes from Senegal to assess their vector competence for dengue serotypes 1 , 3 and 4 . Finally both domestic and wild Senegalese mosquitoes showed a low ability to transmit dengue viruses . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"invertebrates",
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"geographical",
"locations",
"microbiology",
"tropical",
"diseases",
"saliva",
"animals",
"viruses",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"insect",
"vectors",
"africa",
"senegal",
"infectious",
"diseases",
"aedes",
"aegypti",
"medical",
"microbiology",
"dengue",
"fever",
"microbial",
"pathogens",
"disease",
"vectors",
"insects",
"arthropoda",
"people",
"and",
"places",
"mosquitoes",
"eukaryota",
"blood",
"anatomy",
"flaviviruses",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"viral",
"diseases",
"organisms"
] | 2019 | Potential for sylvatic and urban Aedes mosquitoes from Senegal to transmit the new emerging dengue serotypes 1, 3 and 4 in West Africa |
Drosophila melanogaster has played a pivotal role in the development of modern population genetics . However , many basic questions regarding the demographic and adaptive history of this species remain unresolved . We report the genome sequencing of 139 wild-derived strains of D . melanogaster , representing 22 population samples from the sub-Saharan ancestral range of this species , along with one European population . Most genomes were sequenced above 25X depth from haploid embryos . Results indicated a pervasive influence of non-African admixture in many African populations , motivating the development and application of a novel admixture detection method . Admixture proportions varied among populations , with greater admixture in urban locations . Admixture levels also varied across the genome , with localized peaks and valleys suggestive of a non-neutral introgression process . Genomes from the same location differed starkly in ancestry , suggesting that isolation mechanisms may exist within African populations . After removing putatively admixed genomic segments , the greatest genetic diversity was observed in southern Africa ( e . g . Zambia ) , while diversity in other populations was largely consistent with a geographic expansion from this potentially ancestral region . The European population showed different levels of diversity reduction on each chromosome arm , and some African populations displayed chromosome arm-specific diversity reductions . Inversions in the European sample were associated with strong elevations in diversity across chromosome arms . Genomic scans were conducted to identify loci that may represent targets of positive selection within an African population , between African populations , and between European and African populations . A disproportionate number of candidate selective sweep regions were located near genes with varied roles in gene regulation . Outliers for Europe-Africa FST were found to be enriched in genomic regions of locally elevated cosmopolitan admixture , possibly reflecting a role for some of these loci in driving the introgression of non-African alleles into African populations .
Drosophila melanogaster has a well known history and ongoing role as a model organism in classical and molecular genetics . Its well-annotated genome [1] , [2] and genetic toolkit have also made it an important model organism in the field of population genetics , in many cases motivating the development of broadly applicable theoretical models and statistical methods . Prior to the advent of DNA sequencing , studies of inversions and allozymes in D . pseudoobscura [3] , [4] , and later D . melanogaster [5] , [6] , provided some of the field's first glimpses of genetic polymorphisms within and between populations , often providing evidence for geographic clines consistent with local adaptation . The analysis of DNA sequence data from the Drosophila Adh gene motivated the development of methods that compare polymorphism and divergence at different gene regions [7] or functional categories of sites [8] , and offered examples of non-neutral evolution . Sequence polymorphism data from additional D . melanogaster genes revealed that recombination rate is strongly correlated with nucleotide diversity but not between-species divergence in D . melanogaster [9] . This result suggested that genetic hitchhiking [10] could be an important force in molding diversity across the Drosophila genome , but it also motivated the suggestion that background selection against linked deleterious variants [11] should likewise reduce diversity in low recombination regions of the genome . Larger multi-locus data sets initially came from studies of microsatellites and short sequenced loci . Several of these studies compared variation between ancestral range populations from sub-Saharan Africa and more recently founded temperate populations from Europe , finding that non-African variation is far more strongly reduced on the X chromosome than on the autosomes [12] , [13] . Sequence data also allowed larger-scale comparisons of polymorphism and divergence , leading to suggestions that significant fractions of substitutions at nonsynonymous sites [14] and non-coding sites [15] were driven by positive selection . Although previous studies have found considerable evidence for a genome-wide influence of natural selection , a thorough and confident identification of recent selective sweeps in the genome requires an appropriate neutral null model that incorporates population history . Both biogeography [16] and genetic variation [17] , [18] indicate that D . melanogaster originated within sub-Saharan Africa . Even within Africa , D . melanogaster has only been collected from human-associated habitats , and so its original habitat and ecology , along with the details of its transition to a human commensal species , remain unknown [19] . A few studies have found populations from eastern and southern Africa to be the most genetically diverse [18] , [20] , [21] , suggesting that the species' ancestral range may lie within these regions . Small but significant levels of genetic structure are present within sub-Saharan Africa [18] , which could reflect either long-term restricted migration or short-term effects of bottlenecks associated with geographic expansions within Africa . On the order of 10 , 000 years ago [22]–[24] , D . melanogaster is thought to have first expanded beyond sub-Saharan Africa , perhaps by traversing formerly wetter parts of the Sahara [16] or the Nile Valley [18] . This expansion involved a significant loss of genetic variation [12] , [13] , brought D . melanogaster into the palearctic region ( northern Africa , Asia , and Europe ) , and largely gave rise to the “cosmopolitan” populations that live outside sub-Saharan Africa today . American populations were founded only within the past few hundred years [25] , and their complex demography appears to involve admixture between European and African source populations [26] . Recent advances in DNA sequencing technology have allowed genetic variation to be studied on the whole-genome scale . The sequencing of six D . simulans genomes [27] provided the first comprehensive look at fluctuations of polymorphism and divergence across the genome and their potential causes , including potential targets of adaptive evolution . More recently , larger samples of D . melanogaster genomes have been sequenced , yielding further insight into the potential impact of natural selection on diversity across the Drosophila genome [28] , [29] and connections between genetic and phenotypic variation [29] . However , a large majority of the sequenced genomes are of North American origin , and before we can clearly understand the demographic history of that population , we must investigate genomic variation in its African and European antecedents . Here , we use whole genome sequencing and population genomic analysis to examine genetic variation in wild-derived population samples of D . melanogaster . We use a new method to detect pervasive admixture from cosmopolitan into sub-Saharan populations . We use geographic patterns of genetic diversity and structure to investigate the history of D . melanogaster within Africa . Finally , we identify loci with unusual patterns of allele frequencies within or between populations , which may represent targets of recent directional selection .
The sequenced genomes vary significantly in mean sequencing depth ( average number of reads at a given bp ) present in the assemblies . Among genomes with relevant data from all five target chromosome arms , mean depth ranges from 18X to 47X ( Table S2 ) . Depth was found to have a substantial influence on pairwise genetic distances . Mean depth showed positive , non-linear relationships with distance from the D . melanogaster reference genome , and with average distances to other African samples , such as Zambia-Siavonga ( Figure 2A ) . The relationship between depth and genetic distance from Zambia is especially strong ( Spearman ρ = 0 . 63; P<0 . 00001 ) , suggesting that population ancestry has little influence on this quantity ( a property of the ZI sample further discussed below ) . This correlation is especially pronounced for genomes with depth below 25X , while only a modest slope is present above this threshold . Mean depth was also correlated with genomic coverage – the portion of the genome with a called base at the quality threshold ( Figure 2B; Spearman ρ = 0 . 62; P<0 . 00001 ) . The lowest depth genomes were found to have ∼2% lower coverage than a typical genome with average depth . Some correlation of depth with genetic distance and genomic coverage might be expected if genomes with higher depth were more successful in mapping reads across genomic regions with high levels of substitutional ( and perhaps structural ) variation . Additionally , a consensus-calling bias in favor of the reference allele , such that higher depth genomes were more likely to have adequate statistical evidence favoring a non-reference allele , might contribute to the reduced genetic distances and genomic coverage exhibited by the lowest depth genomes . The influence of depth on genetic distance has the potential to bias most population genetic analyses . We found that strict sample coverage thresholds ( only analyzing sites covered in most or all assemblies ) could ameliorate the depth-distance correlation , but at the cost of excluding most variation and introducing a substantial reference sequence bias ( Figure S2 ) . Instead , we addressed the depth-distance issue by focusing most analyses on genomes with >25X depth and made additional corrections when needed , as described below . Assemblies derived from haploid embryos with >25X depth were defined as the “primary core” data set ( Table S1 ) . Haploid embryo genomes with <25X depth were denoted as “secondary core” . Genomes not derived from haploid embryos were labeled “non-core” , and were not analyzed further in this study . Previous work has suggested that introgression from cosmopolitan sources ( i . e . populations outside sub-Saharan Africa ) may be an important component of genetic variation for at least some African populations of D . melanogaster [18] , [33] , [34] . Preliminary examination of this data set revealed a number of sub-Saharan genomes with unusually low genetic distances to cosmopolitan genomes ( the latter represented by the European FR sample and the North American reference genome ) . Undetected admixture could undermine the demographic assumptions of many population genetic methods , altering genetic diversity and population differentiation , and creating long-range linkage disequilibrium . Hence , we attempted to identify specific chromosome intervals that have non-African ancestry , so that they could be filtered from downstream analyses when appropriate . We developed a Hidden Markov Model ( HMM ) method to identify chromosome segments from sub-Saharan genomes that have cosmopolitan ancestry , as described under Materials and Methods . The method utilized a “European panel” ( the FR sample ) and an “African panel” ( the RG sample ) which may contain some admixture . Because of the diversity-reducing out-of-Africa bottleneck , non-African genomes should be more closely related to each other than they are to African genomes . Therefore , if we examine genomic windows of sufficient length , genetic distances between two FR genomes should be consistently lower than between an RG and an FR genome ( Figure S3 ) . To take advantage of this contrast , we constructed chromosome arm-wide emissions distributions by evaluating two locally rescaled quantities in ∼50 kb windows . One distribution , representing African ancestry , was formed from genetic distances of each RG genome to the FR panel . The other distribution , representing non-African ancestry , was formed from genetic distances between each FR genome and the remainder of the FR panel . Individual African genomes were then compared to the FR panel to determine the likelihood of African or non-African ancestry in each window ( essentially , using the emissions distributions to determine whether we are truly making an Africa-Europe genetic comparison , or if we are actually comparing two non-African alleles in the case of an admixed African genome ) . The HMM was then applied to convert likelihoods to admixture probabilities for each genome in each window . This approach was validated using simulations ( see Materials and Methods; Figure S4 ) . For the empirical data , the above approach was applied iteratively to the RG sample to eliminate non-African intervals from the “African panel” used to create emissions distributions . Emissions distributions generated using the FR and RG samples were also used to calculate admixture probabilities for the other sub-Saharan primary core genomes . Simple correction factors were applied to account for the effects of sequencing depth and other quality factors for each genome ( Materials and Methods ) . When applied to the RG primary core genomes , the admixture detection method produced generally sharp peaks along chromosome arms , with only 3 . 3% of window admixture probabilities between 0 . 05 and 0 . 95 ( Figure S5; Table S5 ) . When primary core genomes from other population samples were analyzed , results still appeared to be of reasonable quality , with 8 . 3% “intermediate” admixture probabilities as defined above ( Figure S5; Table S5 ) . However , inferences for the secondary core and non-core genomes appeared less reliable , with 22 . 5% intermediate admixture probabilities and more admixture predicted in general ( Figure S5; Table S6 ) . Hence , the influence of lower sequencing depth may have added significant “noise” into the admixture analysis . Below , we focus on admixture inferences from the primary core genomes only . The estimated proportion of cosmopolitan admixture varied dramatically among the twenty sub-Saharan population samples represented in the primary core data set ( Figure 3A , Table S7 ) . In general , populations with substantial admixture were observed across sub-Saharan Africa , but admixture proportion varied substantially within geographic regions . At the extremes , one Zambia sample ( ZI ) had 1 . 4% inferred admixture among four genomes , while another Zambia sample ( ZL ) had 84% inferred admixture from the single genome sequenced . A Kruskal-Wallis test for the 14 populations with n≥3 primary core genomes supported a significant effect of population on admixture proportion ( P<0 . 0001 ) . Testing whether admixture might be related to anthropogenic activity , we found that human population size of the collection locality had a strong positive correlation with admixture proportion ( Spearman ρ = 0 . 60; one-tailed P = 0 . 003; Figure S6 ) . For the seven collection sites with population sizes below 20 , 000 , all but one population sample had an admixture proportion below 7% ( the exception , KR , may reflect a higher regional effect of admixture in Kenya ) . In contrast , for the eight cities with a population above 39 , 000 , admixture proportion was always above 15% . These results mirror previous findings that urban African flies are genetically intermediate between rural African flies and European flies , when population samples from the Republic of Congo [33] , [35] and Zimbabwe [34] were examined . Our results suggest that African invasion by cosmopolitan D . melanogaster is not limited to the largest African cities , and has occurred in moderately sized towns and cities across sub-Saharan Africa . In theory , higher admixture levels in urban African locations could result from either neutral or adaptive processes . If larger cities are more connected to international trade , then selectively neutral immigration would affect urban populations first . However , the large size of admixture tracts ( e . g . a mean admixture tract length of 4 . 8 centiMorgans or 3 . 8 Mb for the RG sample; Table S7 ) suggests an unusually rapid spread of cosmopolitan alleles into Africa , which may not be compatible with plausible levels of passive gene flow . We used the method of Pool and Nielsen [36] to estimate , for the RG sample , the parameters of a two epoch migration rate change model . This method found the highest likelihood for a change in migration 59 generations before present , with near-zero migration before this time ( point estimate 1 . 2×10−8 ) , and an unscaled migration rate of 0 . 0010 since the change . It is not clear whether a neutral model invoking thousands of immigrants per generation should be viewed as realistic . Note that the rate of admixture would have to be higher yet if this small Rwandan town was not the point of African introduction for cosmopolitan immigrants . Alternatively , cosmopolitan admixture into sub-Saharan D . melanogaster could be a primarily adaptive phenomenon . Certain cosmopolitan alleles might provide a selective advantage in modern urban environments and may now be favored in modernizing African cities , but may be neutral or deleterious in rural African environments . Or , some cosmopolitan genotypes ( such as those conferring insecticide resistance [37] ) may now be advantageous in both urban and rural African environments , but have thus far spread primarily into urban areas . In either scenario , there is still a role for demography ( i . e . migration rates within Africa ) in governing the geographic spread of cosmopolitan alleles into African environments in which they are adaptive . Perhaps more striking than the between-population pattern of admixture are the stark differences in ancestry observed within populations . This individual variability is well-illustrated by the RG sample ( Figure 3B ) , but similar patterns are also observed in other populations ( Table S8 ) . Among the 22 RG primary core genomes , nine have no inferred admixture at all , eight others have less than 3% admixture , while the other five genomes contain 20–76% admixture . Based on forward simulations with recombination and migration [36] , the observed variance among genomes in cosmopolitan admixture proportion for the RG sample was found to be unlikely under the point estimates of demographic parameters reported above ( one-tailed P = 0 . 02 ) . The unexpectedly high variance in admixture proportion may require a combination of biological explanations . Inversion frequency differences between African and introgressing chromosomes would reduce the rate of recombination , potentially keeping admixture in longer blocks . However , the genome-wide prevalence of long admixture tracts ( including in regions that do not overlap common inversions ) makes this explanation incomplete at best . Alternatively , African populations may be subject to local heterogeneity for any number of environmental factors , and cosmopolitan alleles may confer a greater preference for and/or fitness in specific microhabitats . Such differences might provide a degree of spatial isolation between flies with higher and lower levels of admixture . However , the RG sample was collected from a handful of markets , restaurants , and bars in the center of the relatively small town of Gikongoro , Rwanda ( an area less than 200 m across ) , and it's not clear whether any meaningful isolation could exist on this scale . Finally , sexual selection may play a role in generating this pattern . African strains of D . melanogaster are known to display varying degrees of “Z-like” mating behavior , in which females discriminate against males from “M-like” strains , which include cosmopolitan populations [38] , [39] . Hence , one would expect many African females to avoid mating with males carrying the cosmopolitan alleles responsible for the M phenotype . And indeed , mating choice experiments [33] found that matings between rural Brazzaville females and urban ( apparently admixed ) Brazzaville males were much less frequent than homogamic pairings . This phenomenon might help to explain the prevalence of admixture in only a subset of genomes in RG and other samples . Further empirical and predictive studies will be needed to assess the ability of these and other hypotheses to explain the inferred patterns of cosmopolitan admixture among sub-Saharan genomes . If cosmopolitan admixture is partly due to adaptive processes , it may be worthwhile to examine variability in admixture proportion across the genome . Figure 4 shows the number of primary core genomes with admixture probability above 50% for each window analyzed by the admixture HMM . By including admixture tracts from 95 sub-Saharan genomes across all populations , we may lose some population-specific signals , but we gain resolution that would not exist within small samples . Clear differences were observed between chromosomes in admixture levels . Averaging across all windows , arms 3L and 3R had the highest number of admixed genomes ( 18 . 1 and 18 . 0 , respectively ) , while 2L and 2R were somewhat lower ( both averaged 14 . 7 ) . Both autosomes , however , were considerably more admixed than the X chromosome , which averaged just 9 . 3 admixed genomes per window . A qualitatively consistent pattern has been reported [34] in which cosmopolitan admixture was detected on the third chromosome but not the X chromosome in a sample from Harare , Zimbabwe . A lesser contribution of the X chromosome to cosmopolitan admixture might be expected if males contributed disproportionately to introgression . However , the mating preferences described above might be expected to yield the opposite result , suppressing genetic contributions from cosmopolitan males into African populations . Additionally , the loci responsible for the M/Z behavioral polymorphism are thought to reside primarily on the autosomes [38] , which should impede autosomal introgression rather than X-linked introgression . Another explanation for the deficiency of X-linked admixture is more efficient selection due to the X chromosome's hemizygosity [40] . The X chromosome might have experienced a higher rate of “out-of-Africa” selective sweeps [12] , and even though some cosmopolitan adaptations may now be favored in Africa , it is conceivable that the X chromosome contains a greater density of cosmopolitan alleles that are still deleterious in sub-Saharan Africa and limit X-linked introgression . Even if cosmopolitan alleles that remain deleterious in Africa occur at similar rates on the X chromosome and autosomes , selection might be more effective against introgressing X chromosomes . Alternatively , the X chromosome's higher recombination rate may lead advantageous X-linked cosmopolitan alleles to introgress within smaller chromosomal blocks . The recombination rate difference predicted by mapping crosses [28] will be magnified by the lack of recombination in males , and perhaps also by the autosomes' generally higher levels of inversion polymorphism [41] , which should decrease autosomal recombination rates in nature and increase X chromosome recombination ( due to the interchromosomal effect [42] ) . Considerable variation in the proportion of admixed individuals was also apparent within chromosomes . For example , the X chromosome's dearth of admixture was most dramatic for the telomere-proximal half of its windows ( average 6 . 9 admixed genomes ) and less severe for the centromere-proximal half ( average 11 . 7 ) . On a finer scale , the proportion of admixed genomes showed relatively narrow genomic peaks and valleys ( Figure 4 ) , with the most extreme admixture levels often limited to intervals on the order of 100 kb . If the adaptive hypothesis of cosmopolitan admixture is correct , genomic peaks and valleys of admixture could include cosmopolitan loci that are advantageous and deleterious , respectively , in sub-Saharan Africa . We return to the specific content of these intervals later , in the context of out-of-Africa sweeps . In order to evaluate the effectiveness of admixture identification and to examine geographic gradients of genetic variation , principal Components Analysis ( PCA ) [43] was applied to admixture-filtered and unfiltered data . In both cases , the first principal component clearly reflected cosmopolitan versus African ancestry . Comparison of these results suggested that our admixture detection method had successfully filtered most , but not all , cosmopolitan admixture from sub-Saharan genomes ( Figure 5A ) . For example , RG35 was by far the most admixed genome in its Rwanda population sample , with a pre-filtering -PC1 of 0 . 153 . After filtering , its PC1 dropped to −0 . 001 – a considerable improvement , although slightly higher than the population average of −0 . 049 . Hence , a minority of admixture may remain undetected , and for analyses that may be especially sensitive to low levels of admixture , users of the data could opt to exclude genomes with higher levels of detected admixture . Focusing on PCA from admixture-filtered sub-Saharan data , PC1 separated southern African populations from western African and Ethiopian populations , with eastern African samples having intermediate values ( Figure 5B ) . PC2 mainly distinguished Ethiopian samples from all others , while subsequent principal components lacked obvious geographic patterns ( Table S9 ) . Nucleotide diversity was evaluated for windows and for full chromosome arms , in terms of both absolute and relative π ( the latter based on comparison with the RG sample ) . The use of relative π allowed unbiased comparisons of diversity involving populations with incomplete genomic coverage of admixture-filtered data , and it enabled populations lacking two or more genomes with >25X depth to be considered by using RG genomes with similar depth for comparison ( see Materials and Methods ) . Under simple demographic scenarios of geographic expansion , populations with the highest genetic diversity are the most likely to reflect the geographic origin of all extant populations . Hypotheses for the ancestral range of D . melanogaster within sub-Saharan Africa have ranged from western and central Africa ( based on biogeography [16] ) to eastern and southern Africa ( based on a smaller sequence data set [18] ) . Among 19 African populations , the greatest diversity was found in the ZI sample collected from Siavonga , Zambia ( Figure 6; Table 1; Table S10 ) , followed by the geographically proximate ZS and ZO samples . The inferred nucleotide diversity of ZI ( 0 . 70%; 0 . 83% for higher recombination regions ) is lower than estimates for geographically similar samples based on multilocus Sanger sequencing [13] , and slightly lower than a recent population genomic analysis [28] , but higher than an earlier population genomic estimate of θ [44] . While differences in the genomic coverage of these data sets may help to explain some differences , mapping and consensus-calling biases against non-reference reads may also play a role . Such factors are not expected to have a dramatic impact on comparisons of diversity levels between African populations . Hence , based on the samples represented in our study , southern-central Africa appears to contain the center of genetic diversity for D . melanogaster . Although this hypothesis requires further confirmation , these results are consistent with a southern African origin for D . melanogaster . Much of Zambia and Zimbabwe is characterized by a subtropical climate and seasonally dry Miombo and Mopane woodland . Whether this landscape might reflect the original environment of D . melanogaster is unclear , because the species has never been collected from a completely wild environment [16] and the details of its transition to an obligately human-commensal species are unknown [19] . Compared with related species , African strains of D . melanogaster have superior resistance to desiccation [45] and temperature extremes [46] . These characteristics would be predicted by an evolutionary origin in subtropical southern Africa , as opposed to humid equatorial forests . Most populations from eastern Africa ( including Kenya , Rwanda , and Uganda ) had modestly lower diversity compared to Zambia and Zimbabwe , while western populations ( including Cameroon , Guinea , and Nigeria ) showed an additional slight reduction . The two Ethiopian samples showed the lowest variation among African populations , with roughly three quarters the diversity of ZI , potentially indicating a bottleneck during or since the species' occupation of Ethiopia . The distinctness of Ethiopian samples was also indicated by an analysis of mitochondrial and Wolbachia genomes from these same genomes [47] . Otherwise , only two population samples had reduced diversity relative to the overall geographic pattern described above , and one of these ( CK ) had limited pairwise comparisons in the admixture-filtered data . The other sample with locally reduced variation , TZ , displays an unusual pattern of diversity loss on chromosome arms X and 2L specifically ( Table 1 ) , associated with all three sampled genomes carrying inversions In ( 1 ) A and In ( 2L ) t [48] . Similarly , two of the three Kenya samples ( KN and KR ) show reduced diversity on arm 2L only , also apparently in association with In ( 2L ) t [48] . These results suggest the possibility that selection on polymorphic inversions , which are common in sub-Saharan populations [41] , can be an important determinant of genome-scale diversity levels . Although this hypothesis is contrary to some theoretical predictions [49] and empirical findings [50] that would lead one to expect the effects of inversions to be mainly restricted to breakpoint regions , it is supported by an analysis of inversion polymorphism and linked variation in the genomes studied here [48] . Consistent with previous work [12] , [13] , [51] , variation for the cosmopolitan sample ( FR ) is much more strongly reduced on the X chromosome relative to the autosomes ( Table 1 ) . However , further genomic patterns in the ratio of πFR to πRG can be observed ( Figure 7 ) . This diversity ratio ranges from below 0 . 2 ( at the X telomere ) to above 1 ( for a window on arm 3R ) , with similar patterns observed if πFR is instead compared against πZI ( Figure S7 ) . Diversity ratios on autosomal arms showed distinct differences: FR retains 59% of the RG diversity level on arm 2R but 84% on arm 3R ( Table 1 ) . Based on the inversions identified for each genome [48] , we examined the influence of inversions on chromosome arm-wide diversity by recalculating πFR and πRG using standard chromosome arms only . For the RG sample , the exclusion of inversion-carrying arms had negligible influence on diversity , except that the inclusion of In ( 3R ) P ( present in four of 22 genomes ) increased πRG on arm 3R by 4% ( Table S10 ) . More dramatic contrasts were observed for the FR sample , in which inversions were found to result in arm-wide diversity increases of 10% on arm 2L ( due to one of eight FR genomes carrying In ( 2L ) t ) and 18% on arm 3L ( due to a pair of In ( 3L ) P chromosomes ) . As further detailed in a separate analysis [50] , arm 3R was even more strongly affected , with a 29% diversity increase due to the presence of In ( 3R ) P ( in three of eight genomes ) , In ( 3R ) K and In ( 3R ) Mo ( one genome each ) . Although the French sample only contains inversions on these three arms , they contribute to a 12% genome-wide increase in nucleotide diversity . In light of the above observations , it is possible that inversions have had important effects both in reducing chromosome arm-wide diversity ( for the Tanzania and Kenya populations ) and also in elevating it ( for non-African autosomes ) . As further suggested in a separate analysis [48] , the spatial scale of increased diversity associated with inversions in the France sample ( Figure 7 ) may indicate a recent arrival of inverted chromosomes from one or more genetically differentiated populations . Given that similar levels of gene flow are not indicated by polymorphism on chromosome arms lacking inversions , the spread of genetically divergent inverted chromosomes into France may have been primarily driven by natural selection . In light of their powerful elevation of πFR , inverted chromosomes in this sample may have originated from a more genetically diverse African or African-admixed population . Similarly , the more modest elevation of πRG associated with In ( 3R ) P might indicate the recent introgression of these inverted chromosomes from a genetically differentiated population . However , the nature of selective pressures acting on inversions in natural populations of D . melanogaster remains largely unknown . Without inversions , relative πFR for autosomal arms ranged from 0 . 58 to 0 . 63 , with chromosome 3 showing higher values than chromosome 2 . In light of the above hypothesis to account for the presence of divergent inverted chromosomes in the France sample , some of the remaining differences in relative πFR among inversion-free chromosomes might stem from recombination between standard chromosomes and earlier waves of introgressing inverted chromosomes . Alternatively , given that D . melanogaster autosomes frequently carry recessive deleterious mutations [52] , associative overdominance during the out-of-Africa bottleneck might have favored intermediate inversion frequencies [53]–[55] . This hypothesis is mainly plausible in small populations [56] , which may have existed due to strong founder events during the out-of-Africa expansion [57] . Given the opportunity for recombination between standard and inverted chromosomes since that time , past associative overdominance related to inversions ( or centromeric regions ) might contribute to the modest difference in relative πFR between inversion-free second and third chromosomes , as well as the larger gap between both autosomes and the X chromosome . The ratio of relative πFR for the X chromosome versus the inversion-free autosomes appears consistent with some previously explored founder event models [57] if chromosomes X and 2 are compared ( ratio = 0 . 692 , compared to a minimum of 0 . 669 in the cited study ) , but not if chromosome 3 is examined instead ( ratio = 0 . 646 ) . Some studies have concluded that the difference between X-linked and autosomal diversity reductions in cosmopolitan D . melanogaster exceeds the predictions of demographic models involving population bottlenecks and/or a shift in sex-specific variance in reproductive success [12] , [13] . Instead , the X chromosome's disproportionate diversity reduction might result from more efficient positive selection on this chromosome ( due to male hemizygosity [40] ) during the adaptation of cosmopolitan populations to temperate environments . However , it appears relevant that the above studies examined autosomal loci on chromosome 3 , but not chromosome 2 . Further theoretical , simulation , and inferential studies to elucidate the relative influence of selection , demography , and inversions on the X chromosome and autosomes is needed before their relative contribution to diversity in cosmopolitan D . melanogaster can be clearly understood . Levels of genetic differentiation between populations were evaluated in terms of Dxy and FST [58] for each chromosome arm . In order to minimize the effects of any residual admixture in the filtered data , only genomes with admixture proportion below 15% were included . Populations with sufficient data for this analysis included CO , ED , FR , GA , GU , KR , NG , RG , TZ , UG , ZI , and ZS . Within Africa , FST values on the order of 0 . 05 were typical ( Table 2 ) . Geographically proximate population pairs often had lower FST ( at minimum , a value of 0 . 009 between ZI and ZS ) . Comparisons involving the ED sample gave uniformly higher FST than other African comparisons ( median 0 . 147 ) , consistent with the loss of diversity observed for Ethiopian samples . As expected , comparisons of African samples with the European FR sample yielded the highest FST values ( median 0 . 208 ) . Genetic differentiation at putatively unconstrained short intron sites [59] , [60] showed similar patterns ( Table S11 ) , but as expected , magnitudes of Dxy and π were more than twice as high as for all non-centromeric , non-telomeric sites ( for ZI , short intron π = 0 . 0194 ) . In order to assess the compatibility our data with a model of geographic expansion from southern Africa , we examined the ratio of each population's DZI ( average pairwise genetic distance , or Dxy , between this population and the ZI sample ) and πZI . This ratio will be near 1 if a population's genomes are no more divergent from ZI genomes than ZI genomes are from each other , consistent with the recent sampling of this population's diversity from a ZI-like ancestral population . In contrast , ratios exceeding 1 indicate that a population contains unique genetic diversity not present in ZI . Populations from eastern Africa ( KR , RG , TZ , UG ) and Europe ( FR ) had ratios compatible with a recent ZI-like origin ( Table 2 ) . However , populations from western Africa ( CO , GA , GU , NG ) and Ethiopia ( ED ) showed modest levels of unique variation . The highest ratio , for Guinea ( GU ) , indicated a 2 . 9% excess of DZI over πZI . Elevated ratios could indicate a relatively ancient occupation of at least some of the above regions ( perhaps on the order of tens of thousands of years ) . Alternatively , under the hypothesis of an expansion from southern Africa , these regions may have received a genetic contribution from a different part of a structured southern African ancestral range ( e . g . migration into Gabon and western Africa from Angola , which also contains Miombo woodlands but has not been sampled ) . Examination of genomewide genetic differentiation may also shed light on the sub-Saharan origins of cosmopolitan D . melanogaster . Geographic hypotheses for expansion of D . melanogaster from sub-Saharan Africa have ranged from a Nile route starting from the equatorial rift zone [18] to a more western crossing of the Sahara via formerly wetter areas of “Paleochad” [16] . A simple prediction is that the sub-Saharan samples most closely related to the cosmopolitan source population should show the lowest values of Dxy and FST relative to the cosmopolitan FR sample . However , even low levels of undetected cosmopolitan admixture in sub-Saharan genomes could obscure this signal , and so only genomes with <15% detected admixture were considered below . Among the eleven African populations analyzed ( see above ) , the Kenyan KR sample showed the lowest genome-wide DFR ( Table 2 ) and would have had the lowest FR FST if not for its anomalous pattern of variation for arm 2L ( Table S11 ) . However , KR is the sample with the highest proportion of detected cosmopolitan admixture , which clouds the interpretation of these results . After KR , the lowest DFR values come from the western group of samples ( NG , CO , GA , and GU ) , of which two ( CO and GU ) had relatively low levels of detected admixture . Despite its northeast sub-Saharan location , the Ethiopian ED sample does not appear to represent a genetic intermediate between cosmopolitan and other sub-Saharan populations , and may instead represent a separate branch of this species' geographic expansion . Further sampling and analysis may be needed to obtain compelling evidence regarding the geographic origin of cosmopolitan D . melanogaster . One scenario for the sub-Saharan expansion of D . melanogaster is illustrated by the geographic fit of a simple neighbor-joining population tree based on Dxy values ( Figure 8; Figure S8 ) . This tree is consistent with the hypothesis of a southern Africa origin for D . melanogaster , with an initial expansion into eastern Africa , followed by offshoots reaching Ethiopia , the palearctic ( northern Africa and beyond ) , and western Africa . Of course , even after the filtering of cosmopolitan admixture , a tree-like topology is not likely to fully describe the history of sub-Saharan D . melanogaster populations . However , the history described above seems consistent with levels and patterns of population diversity ( Figure 6; Table 2 ) , and may capture some important general features of the species' history . Even if the general expansion history described above ultimately proves to be accurate , many historical details await clarification . Diversity differences among African populations could indicate population bottlenecks during a sub-Saharan range expansion , and population growth during such an expansion is also possible . Further analysis of population genomic data is also needed to establish whether ancestral range populations have also been affected by population growth [23] or a bottleneck [21] . Lastly , although migration within Africa has not erased the observed diversity differences and genetic structure , the historical and present magnitudes of such gene flow are not clear . The quantitative estimation of historical parameters may be addressed by detailed follow-up studies . However , for a species like D . melanogaster , in which very large population sizes may allow relatively high rates of advantageous mutation and efficient positive selection , one concern is that the effects of recurrent hitchhiking may be important on a genomewide scale [27] , [28] , [61] , [62] . Hence , the application of standard demographic inference methods to random portions of the D . melanogaster genome ( or even putatively unconstrained sites ) may yield estimates that are biased by violations of the assumption of selective neutrality . Under the assumption of demographic equilibrium , Jensen et al . [62] estimated a ∼50% reduction in diversity due to positive selection for Zimbabwe D . melanogaster , and selective sweeps may have similarly important influences on the means and variances of other population genetic statistics as well . Hence , further methodological development may be needed before accurate demographic estimates can be obtained for species in which large population sizes facilitate efficient natural selection . Focusing on our largest population sample ( 22 primary core RG genomes ) , we investigated relationships between genetic diversity and mapping-based recombination rate estimates [28] . To minimize the effects of direct selective constraint on the sites examined , we focused on the middles of short introns ( bp 8 to 30 of introns ≤65 bp in length ) , which are among the most polymorphic and divergent sites observed in the Drosophila genome [59] , [60] . Since each 23 bp intronic locus is too small to be considered individually , we show broad-scale patterns of diversity from all relevant sites within a given cytological band . Consistent with previous findings [9] , strong relationships between recombination and variation were observed for all chromosome arms ( Figure 9 ) , with Pearson's r ranging from 0 . 68 ( for 3L ) to 0 . 95 ( for 2R ) , with P = 0 . 0005 or lower for all arms ( Table S12 ) . Curiously , bp position along the chromosome arm was a stronger predictor of diversity than estimated recombination rate for arms 3L and 3R ( Table S12 ) , which could reflect imprecision in recombination rate estimates for chromosome 3 , or the influence of polymorphic inversions on recombination in nature . Across all autosomal arms , the strongest correlation between recombination and diversity was for low rates of crossing-over ( adjusted rate below 1 cM/Mb , equivalent to an unadjusted 2 cM/Mb rate , Pearson r = 0 . 56 and P = 0 . 0002 ) . However , a strong correlation persisted above this threshold as well ( Pearson r = 0 . 44 , P = 0 . 002 ) . Correlations within these categories were not significant for the X chromosome , potentially due to smaller numbers of chromosome bands , especially for the low recombination category ( n = 4 ) . Overall , the above results are consistent with the well-supported role for natural selection in reducing variation in regions of low recombination . However , the relative contributions of specific selection models such as hitchhiking [10] and background selection [11] to this pattern have not been quantitatively estimated . Examining the RG sample's allele frequencies at short intron sites , we observed an excess of singleton polymorphisms ( sites with a minor allele count of 1 ) for all chromosome arms relative to the predictions of selective neutrality and demographic equilibrium ( Figure 10A ) . The degree of this excess varied among chromosome arms: compared to a null expectation of 31% singleton variants , the autosomal arms ranged from 33% to 37% , while the X chromosome had 44% singletons . The general excess of rare alleles could reflect population growth , as suggested for a Zimbabwe population sample [23] , and growth has some potential to influence X-linked and autosomal variation differently [63] . Recurrent hitchhiking may contribute to the genomewide excess of rare alleles [64] . Under this hypothesis , the difference in singleton excess between the X chromosome could reflect more efficient X-linked selection due to hemizygosity [40] . Without a difference in the rate of X-linked and autosomal adaptation , this contrast could instead result from a greater fraction of X-linked selective sweeps acting on new beneficial mutations , with relatively more autosomal sweeps via selection on standing variation . The autosomes may have more potential to harbor recessive and previously deleterious functional variants , and sweeps from standing variation do not strongly influence the allele frequency spectrum [65] . We also used short intron allele frequencies to conduct a preliminary analysis of the relationship between recombination and rare alleles . Specifically , we tested whether the proportion of singletons among variable sites differed between low recombination regions ( defined here as <1 cM/Mb ) and moderate to high recombination regions ( >1 cM/Mb ) . No clear relationship between recombination and allele frequency was observed above this cutoff ( results not shown ) . For the 1 cM/Mb threshold , the X chromosome showed an elevated proportion of singletons in the low recombination category ( 53% vs . 43%; Pearson χ2 P = 0 . 032; Figure 10B ) . Data from the autosomes are inconclusive: while three arms show non-significant trends toward more rare alleles in low recombination regions ( Figure 10B ) , arm 2L showed a significant pattern in the opposite direction ( 30% vs . 36%; P = 0 . 025 ) , possibly reflecting specific evolutionary dynamics of the 2L centromere-proximal region . The X chromosome result is qualitatively consistent with the predictions of the recurrent hitchhiking model [64] and some ( but not all ) previous findings from D . melanogaster [51] , [66] , [67] . Under this hypothesis , the lack of a comparable autosomal pattern might indicate a lesser influence of classic selective sweeps on the autosomes relative to the X chromosome , or a greater effect of inversion-related selection on the autosomes obscuring predictions of the recurrent hitchhiking model . Background selection may also increase the proportion of singletons [68] , [69] , although a greater X-linked effect of background selection has not been suggested . Further study is needed to quantify the influence of positive and negative selection at linked sites on nucleotide diversity and allele frequencies in the D . melanogaster genome . Linkage disequilibrium ( LD ) was examined using a standard correlation coefficient ( r2 ) between single nucleotide polymorphism ( SNP ) pairs , and also via the directional LD metric rω [70] , [71] . The rω statistic is positive when minor frequency alleles at two sites tend to occur on the same haplotype , negative if they tend to be on different haplotypes . Although we lack a comprehensive understanding of the evolutionary forces capable of influencing rω , it is known that hitchhiking strengthens positive rω ( since recombination near a selective sweep leaves groups of positively linked SNPs [72] ) , while negative rω may result from epistatic interactions among beneficial or deleterious alleles [70] . Empirical data from the RG sample was compared against neutral simulations with equilibrium demographic history . Importantly , equilibrium may not accurately reflect the history of the RG sample: recent population growth may have occurred , and the RG sample's modest diversity reduction compared to the ZI sample may imply a mild population bottleneck . Although the full effects of demography can not be eliminated by any simple procedure , we can reduce the influence of growth or other forces responsible for this population's excess of singleton polymorphisms by excluding singletons from the empirical and simulated data . In general , an excess of LD was observed over neutral , equilibrium predictions for all chromosome arms ( Figure 11A ) . The X chromosome's lower LD is consistent with its higher average recombination rate ( 54% higher for the regions examined [28] ) . The RG pattern contrasts with data from a North American population , which showed elevated X-linked LD [28] , [29] that likely reflects a stronger influence of demography and possibly selection on the X chromosome during the species' out-of-Africa expansion . For the RG sample , the X chromosome's LD excess was largely confined to the 10–100 bp scale . In contrast , autosomal arms showed an excess of LD at all scales 10 bp and above ( Figure 11A ) . Since the simulations account for differences in average ( inversion-free ) recombination rate between the X and autosomes , the autosomes' more pronounced LD excess could result from a stronger influence of inversions on these arms . As noted above , the autosomes' higher inversion polymorphism should reduce autosomal recombination rates in nature and increase X chromosome recombination rates . Arm 3R contains the largest number of common inversions in Africa [41] , and LD for this arm is by far the highest . Arm 3R's somewhat lower average recombination rate ( 7–27% lower than other autosomal arms for the analyzed regions ) may contribute to this pattern as well . The above observations regarding LD are concordant with estimates of the population recombination rate for the RG sample , which are elevated for the X chromosome ( in spite of its potentially lower population size ) and reduced for 3R [73] . Notably , the observed LD excess is driven entirely by SNP pairs with positive rω ( Figure 11B ) , while negative SNP pairs show no departure from equilibrium expectations ( Figure 11C ) . Although cosmopolitan admixture has been largely removed from the analyzed data set , it remains possible that demographic events of this nature might inflate positive LD specifically . Inversions may well play a role in boosting positive LD , since inversion-associated polymorphisms may often be present at similar frequencies on the same haplotypes . However , given the excess of LD on all chromosome arms and on relatively short spatial scales , it is not yet clear whether inversions are a sufficient explanation . Recurrent hitchhiking may also contribute to the genome-wide excess of positive LD [72] . Further studies will be needed to evaluate the compatibility of specific hypotheses with genome-wide LD patterns . Identifying the genes and mutations underlying Darwinian selection is an important aspect of evolutionary biology , and of population genomics in particular . The lack of a precise demographic model limits our ability to formally reject the null hypothesis of neutral evolution for specific loci , since certain demographic models can mimic the effects of selective sweeps [74] . However , we have still sought to learn about general patterns of directional selection in the genome by conducting a series of local outlier analyses to detect unusual patterns of allele frequencies within and between populations that are consistent with recent adaptive evolution . These outlier analyses necessarily involve a strong assumption about the proportion of the genome affected by selection . However , the enrichment analyses we perform on these outliers should be robust to some level of random false positives within the outliers , and should still be informative if not all adaptive loci are detected . We searched for putative signals of selective sweeps in the RG sample using a modified version of the SweepFinder program [75] , [76] that looks for both allele frequency spectra and diversity reductions consistent with recent selective sweeps . As further described in the Materials and Methods section , we analyzed the RG data in windows and used the Λmax statistic in an outlier framework , rather than making an explicit assumption regarding the appropriate demographic null model – as would have been required for typical simulations defining statistical significance . Here , we focus on the most extreme 5% of windows from each chromosome arm . After merging neighboring outlier windows , a total of 343 outlier regions were obtained ( Table S13 ) . For each outlier region , the gene with the closest exon to the Λmax peak was recorded . Genes within extreme outlier regions included Ankyrin 2 ( cytoskeleton , axon extension ) , Girdin ( actin filament organization , regulation of cell size ) , Laminin A ( behavior , development , meiosis ) , narrow abdomen ( ion channel , circadian rhythm ) , Odorant receptor 22a [77] , and ribosomal proteins S2 and S14b ( separate regions ) . Several strong outliers corresponded to genes also implicated in a recent genome scan based on outliers for low polymorphism relative to divergence [28] , including bendless ( axonogenesis , flight behavior ) CENP-meta ( mitotic spindle organization , neurogenesis ) , female sterile ( 1 ) homeotic ( regulation of transcription ) , Heterogeneous nuclear ribonucleoprotein at 27C ( regulation of splicing ) , loquacious ( RNA interference , nervous system development , germ-line stem cell division ) , and no distributive junction ( meiotic chromosome segregation ) . Despite a similar number of outlier regions as the FST analyses described below , the Λmax scan yielded a much larger number of significantly enriched gene ontology categories: 115 categories had P<0 . 05 based on random permutation of target windows within chromosomal arms ( Table 3; Table S14 ) . Consistent with previous results from a population genomic outlier analysis of diversity and divergence [28] , numerous biological processes related to gene regulation were observed , including positive and negative regulation of transcription , positive regulation of translation , regulation of alternative splicing , mRNA cleavage , chromatin organization , regulation of chromatin silencing , and gene silencing . Many enriched cellular components ( e . g . nucleus , precatalytic spliceosome , mRNA cleavage and polyadenylation complex , ribonucleoprotein complex , heterochromatin , and euchromatin ) and molecular activities ( e . g . DNA binding , mRNA binding and especially mRNA 3′-UTR binding ) were also consistent with a broad importance for regulators of gene expression in recent adaptive evolution . A number of the GO terms listed in Table 3 were also reported from the above-mentioned genome scan [28] , including negative regulation of transcription , positive regulation of translation , ribonucleoprotein complex , precatalytic spliceosome , protein ubiquitination , nuclear pore , lipid particle , and spermatogenesis . Other enriched biological processes included oogenesis , neurogenesis , male meiosis and female meiosis chromosome segregation , regulation of mitosis and apoptosis , and phagocytosis . Additional cellular components included microtubule-associated complex , kinetochore , and fusome while enriched molecular activities also included ATP binding and voltage-gated calcium channel activity . Nine African population samples with larger sample sizes after admixture filtering were included in an analysis of local genetic differentiation . FST was evaluated for each pair of populations , and the mean FST for each window was noted . Examination of the 2 . 5% highest mean FST values for each chromosome arm and the merging of neighboring outlier windows resulted in 294 outlier regions ( Table S15 ) . For each outlier region , the gene with the closest exon to the center of the most extreme window was noted . Genes associated with unusually strong FST outlier regions included Odorant receptor 22b ( tandem paralog of the above-mentioned Or22a ) , Cuticular protein 65Au , Dystrophin , P-element somatic inhibitor , and CG15696 ( predicted homeobox transcription factor ) . Of course , many of the strongest putative signals of adaptive differentiation are wide , and further investigation will be needed to confirm specific targets of selection . Permutation of putative target windows indicated that genes from 34 GO categories were significantly over-represented among our outliers at the P = 0 . 05 level ( Table 4; Table S16 ) . These GO categories included biological processes ( e . g . oocyte cytoskeleton organization , regulation of alternative splicing , regulation of adult cuticle pigmentation ) , cellular components ( e . g . mitochondrial matrix , dendrite ) , and molecular functions ( e . g . olfactory receptor activity , mRNA binding ) . A windowed FST outlier approach was also applied to detect loci that may contain adaptive differences between sub-Saharan ( RG ) and European ( FR ) populations . Some of these loci might have had adaptive importance during the expansion of D . melanogaster into temperate environments , but others could reflect recent selection within Africa . A total of 346 outlier regions resulted from analyzing the upper 2 . 5% tail of Rwanda-France FST ( Table S17 ) . Genes associated with strong FST outliers included Or22a ( which may be under selection in Africa , see above ) , CHKov1 ( insecticide and viral resistance [78] , [79] ) , ACXC ( spermatogenesis ) , and Jonah 98Ciii ( digestion ) , plus a number of genes involved in morphological and/or nervous system development ( e . g . Bar-H1 , Death-associated protein kinase related , Enhancer of split , hemipterous , highwire , mastermind , rictor , sevenless , Serendipity δ , and wing blister ) . Other genes at the center of strong outlier regions were also detected by a genome-wide analysis of diversity ratio between U . S . and Malawi populations [28] , including dpr13 ( predicted chemosensory function ) , Neuropeptide Y receptor-like , rugose ( eye development ) , and Sno oncogene ( growth factor signaling , neuron development ) . The genes identified in this analysis still yielded 31 significantly enriched GO categories ( Table 5; Table S18 ) . Biological processes among these GO categories included chromosome segregation , locomotion , female germ-line cyst formation , histone phosphorylation , and alcohol metabolism . Cellular components included basal lamina and polytene chromosome interband , while molecular activities included transcription coactivators and neuropeptide receptors . The detected GO categories were essentially distinct from those obtained from the diversity ratio analysis of Langley et al . [28] . The lack of overlap may stem at least partially from differences in the statistics and populations used in each analysis . The well-known challenges of identifying positive selection in the presence of bottlenecks [74] , along with uncertainty regarding the portion of the genome affected by adaptive population differences , may also contribute to these findings . Both analyses , however , should motivate new adaptive hypotheses to be tested via detailed population genetic analyses and experimental approaches . If the rapid introgression of non-African genotypes into African populations documented above is driven by natural selection , then sharp peaks and valleys of admixture along the genome ( Figure 4 ) should contain functional differences between sub-Saharan and cosmopolitan populations . Such differences may have been driven by natural selection after these populations diverged , and hence may be detectable by the Africa-Europe FST outlier scan presented above . Given that the scale of these FST outliers ( on the order of 10 kb ) is narrower than our admixture peaks and valleys ( on the order of 100 kb ) , population genetic signals of elevated differentiation may be helpful in localizing genes responsible for driving or opposing non-African gene flow into African populations . We selected eight clear genomic peaks of admixture within the higher recombination regions analyzed for FST . These peaks were delimited by windows containing the local maximum number of admixed genomes , and identified FST outlier regions that either overlapped them or were within 100 kb . Valleys of admixture were more difficult to clearly distinguish from gaps between peaks and minor fluctuations ( Figure 4 ) – three were identified , one of which overlapped several FST outlier regions ( Table S19 ) . For peaks of admixture , seven of these eight regions were associated with FST outlier regions ( Table S19 ) , exceeding random expectations ( permutation P = 0 . 017 ) . Stronger outlier regions associated with admixture peaks included the genes Bar-H1 , Enhancer of split , Neuropeptide Y receptor-like , and sevenless . Further studies will be needed to evaluate the possibility that cosmopolitan alleles at one or more of these loci may now confer a fitness advantage in urban African environments .
Consensus sequences with reduced reference bias are now available from http://www . dpgp . org/dpgp2/dpgp2 . html
Genomes reported here are derived from the population samples listed in Table S1 and depicted in Figure 1 . The collection methods for samples collected in 2004 or later correspond to a published protocol [80] . Information about individual fly stocks is presented in Table S2 . Most of the relevant stocks are isofemale lines , each founded from a single wild-caught female . In some cases , intentional inbreeding was conducted by sib-mating for five generations; such lines have an ‘N’ appended to the isofemale line label . Although not a focus of our analysis , we have also released genomic data from a small number of chromosome extraction lines , created using balancer stocks . Except for the three ZK genomes , DNA for all inbred and isofemale lines was obtained from haploid embryos [30] . Briefly , a female fly from the stock of interest was mated to a male homozygous for the ms ( 3 ) K811 allele [81] . This mating produces some eggs which are fertilized but fail to develop because the clastogenic paternal genome . Rarely , such eggs bypass apparent checkpoints and develop as haploid embryos . Eggs with partially developed first instars were visually identified under a microscope . DNA was isolated from haploid embryos and genome-amplified as previously described [30] . For the ZK genomes and chromosome extraction lines , DNA was isolated from 30 adult flies ( generally females; mixed sexes in the case of autosomal extraction lines ) . For all samples , library preparation for sequencing ( ligation of paired end adapters , selection of ∼400 bp fragments , and PCR enrichment ) was conducted as previously described [30] . In some cases , bar code tags ( 6 bp ) were added to allow multiplexing of two or more genomes in one flow cell lane . Sequencing was performed using standard protocols for the Illumina Genome Analyzer IIx . Initial data processing and quality analysis was performed using the standard Illumina pipeline . Sequence reads were deposited in the NIH Short Read Archive as project SRP005599 . Alignments to the D . melanogaster reference genome ( BDGP release 5 ) using BWA version 0 . 59 [31] with default settings and the “-I” flag . Program defaults included a 32 bp seed length; reads could therefore map to the reference only if two or fewer reference differences were present within a seed . Although read lengths varied from 76 bp to 146 bp within this data set , only the first 76 bp of longer reads was used for the assemblies reported here . In order to exclude ambiguously mapping reads , those with a BWA mapping quality score less than 20 were eliminated from the assemblies . Consensus sequences for each assembly were obtained using the SAMtools ( version 0 . 1 . 16 ) pileup module [32] . These diploid consensus sequences generally included a few thousand heterozygous calls , scattered across the genome . Such sites are not expected to represent genuine heterozyosity in these haploid/homozygous samples ( with the exception of ZK , in which large-scale heterozygosity was observed , presumbaly due to incomplete inbreeding ) . All putatively heterozygous sites were masked to ‘N’ . Sites within 5 bp of a consensus indel were also masked to ‘N’ – this criterion was found to reduce errors associated with indel alignment; no appreciable benefit was observed if 10 bp was masked instead ( data not shown ) . Data were only considered for “target” chromosome arms , as defined in Table S1 . These are chromosome arms expected to derive from the population sample of interest ( as opposed to originating from laboratory balancer stocks ) , and observed to be free of heterozygous intervals . Chromosome arms were further defined as “focal” ( the genomic regions analyzed here , namely the euchromatic portions of X , 2L , 2R , 3L , and 3R ) or “non-focal” ( the mitochondria and heterochromatin , including chromosomes 4 and Y ) . The assemblies analyzed here were defined as “release 2” data and are available for download at http://www . dpgp . org/dpgp2/DPGP2 . html . Assemblies of mitochondrial and bacterial symbiont genomes are reported and analyzed separately [47] . Although the above assemblies provide nominal quality scores , we performed a separate evaluation of statistical confidence in the accuracy of assemblies . This analysis utilized five haploid embryo , reference strain ( y1 cn1 bw1 sp1 ) genomes resequenced with comparable depth and read characteristics as the non-reference genomes reported here ( Table S2 ) . In order to simulate the effects of genetic variation , the maq fakemut program [82] was used to introduce artificial substitutions and indels into the resequenced reference genomes . Substitutions were introduced at rate 0 . 012/bp , while 1 bp indels were introduced at rate 0 . 0024/bp . Alignment and consensus sequence generation was then performed as described above . The artificially mutagenized reference genomes allowed us to examine the tradeoff between minimizing error rates and maximizing genomic coverage . Based on the joint pattern of these quantities for various nominal quality scores ( Figure S1 ) , we selected a nominal quality threshold of Q31 as the basis for downstream analyses . The observed consensus sequence error rate for the nominal Q31 cutoff suggested was equivalent to an average Phred score of Q48 ( roughly one error per 100 kb ) . Long tracts of identity-by-descent ( IBD ) between genomes may result from the sampling of related individuals . Because such relatedness violates the assumptions of many population genetic models , we sought to identify and mask instances of IBD caused by relatedness . Target chromosomes from all possible pairs of genomes were compared to search for long intervals of identity-by-descent ( IBD ) that may result from close relatedness . Following Langley et al . [28] , windows 500 kb in length were moved in 100 kb increments across the genome , and sequence identity was defined as less than 0 . 0005 pairwise differences per site . A large number of pairwise intervals fit this criterion ( Table S3 ) . Some chromosomal intervals , including centromeres and telomeres , had recurrent IBD in between-population comparisons ( Table S4 ) . Cross-population IBD occurred at scales up to 4 Mb within these manually delimited “recurrent IBD regions” , and its occurrence between different populations suggests that processes other than close relatedness are responsible . Such intervals were not masked from the data . We identified clear instances of “relatedness IBD” between two genomes when within-population IBD exceeded the scale observed between populations: when more than 5 Mb of summed genome-wide IBD tracts occurred outside recurrent IBD regions , or when tracts greater than 5 Mb overlapped recurrent IBD regions . Only nine pairs of genomes met one or both of these criteria ( Table S4 ) , and two of these pairs were expected based on the common origin of isofemale and chromosome extraction lines ( Table S2 ) . For these pairs , one of the two genomes was chosen for filtering , and all identified IBD intervals from this pairwise comparison were masked to ‘N’ for most subsequent analyses . Relevant for the inference of non-African admixture is a panel of eight primary core genomes from France ( the “FR” sample ) . D . melanogaster populations from outside sub-Saharan Africa show reduced genetic diversity and are more closely related to each other than to sub-Saharan populations [22] , [26] . Hence , whether admixture came from Europe or elsewhere in the diaspora , FR should represent an adequate “reference population” for the source of non-African admixture . However , we lack an African population that is known to be free of admixture . And while a variety of statistical methods exist for the detection of admixture , options for detecting unidirectional admixture using a single reference population are more limited . We therefore developed a new method to detect admixture in this data set . We constructed a windowed Hidden Markov Model ( HMM ) machine learning approach based on a given haplotype's average pairwise divergence from the non-African reference population ( DFR ) . The admixed state is based on comparisons of individual FR haplotypes to the remainder of the FR sample . The non-admixed state is based on comparisons of haplotypes from a provisional “African panel” to the FR sample . Here , 22 genomes from the Rwanda “RG” sample are used as the African panel . We allow for the possibility of admixture within the African panel as described below . Formally , the emissions distribution for the non-admixed state was constructed as follows . For each window , each RG haplotype was evaluated for average pairwise divergence from the FR sample ( DRG , FR ) . Each of these values was rescaled in terms of standard deviations of DRG , FR from the window mean DRG , FR . Standardized values were added to the emissions distribution in bins of 0 . 1 standard deviations , and these bins were ultimately rescaled to sum to 1 . Hence , the emissions distribution reflects the genome-wide pattern of DRG , FR , accounting for local patterns of diversity . The emissions distribution for the admixed state was constructed similarly . For each window , each FR haplotype was evaluated for average pairwise divergence from the remainder of the FR sample ( DFR , FR ) . However , these DFR , FR values were still rescaled by the window mean and standard deviation of DRG , FR . An alternative version of the method in which the admixed state's emissions distribution was instead rescaled by the local mean and standard deviation of DFR , FR was slightly less accurate when applied to simulated data . Given these genome-wide emissions distributions , we can examine DRG , FR for each African allele for each window , and obtain its likelihood if we are truly making an “Africa-Europe comparison” with this DRG , FR ( non-admixed state ) or if we are actually making a “Europe-Europe comparison” ( admixed state ) . These likelihoods form the input for the HMM process , which was performed using an implementation [83] of the forward-backward algorithm . A minimum admixture likelihood of 0 . 005 was applied to HMM input , in order to reduce the influence of a single unusual window . Admixed intervals were defined as windows with >50% posterior probability for the admixed state . For the purpose of masking admixed genomic intervals for downstream analyses , one window on each side of admixed intervals was added ( to account for uncertainty in the precise boundaries of admixture tracts ) . The admixture detection method was tested using simulated data containing known admixture tracts . Population samples of sequences 10 Mb in length were simulated using MaCS [84] , which can approximate coalescent genealogies across long stretches of recombining sequence . Demographic parameters were based on a published model for autosomal loci [13] , [23] . The command line used was “ . /macs04 63 10000000 -s 12345 -i 1 -h 1000 -t 0 . 0376 -r 0 . 171 -c 5 86 . 5 -I 2 27 36 0 -en 0 2 0 . 183 -en 0 . 0037281 2 0 . 000377 -en 0 . 00381 2 1 -ej 0 . 00381000001 2 1 -eN 0 . 0145 0 . 2” , specifying simulations with present population mutation rate 0 . 0376 and population recombination rate 0 . 171 , gene conversion parameters based on a weighted average of loci from Yin et al . [85] , and historic tree retention parameter h = 1000 [84] . The above simulations generate population samples that may resemble data from sub-Saharan and cosmopolitan populations of D . melanogaster , but they do not involve any admixture . If admixture was specified with the command line , then without modifications to the simulation program , there would not be an output record of admixture tract locations . Instead , extra “non-African” haplotypes were simulated ( one for each African haplotype ) , and these “donor alleles” became the source for admixture tracts which were spliced into the African population's data after MaCS simulation was completed . The locations and lengths of admixture tracts were determined by a separate simulation process . The forward simulation program developed by Pool and Nielsen [36] accounts for drift , recombination , and migration , recording intervals with migrant history . By using this program to simulate a region symmetric to the African MaCS data , we identified intervals that should contain admixture tracts after g generations of admixture . These intervals were then spliced from the non-African donor alleles into African haplotypes from the MaCS simulated polymorphism data . The simulated data with admixture was then analyzed using the admixture HMM method described above . In this case , windows of 10 kb were analyzed . Times since the onset of admixture ( g ) of 100 , 1000 , and 10000 generations were examined . Migration rates were specified to approximate a total admixture proportion of 10% ( hence testing the robustness of the method to this level of admixture in the “African panel” ) . As indicated by representative simulation results shown in Figure S8 , the admixture detection method was highly accurate for g = 100 and g = 1000 , and moderately accurate for g = 10000 . Based on preliminary observations from the data , we suspected that much of the admixture in our data set was on the order of g = 100 or less . The admixture HMM was initially applied to the RG sample alone . Compared with the simulated data , the empirical data showed more overlap between the admixed and non-admixed emissions distributions . This contrast could result from demographic differences between the African population used here ( from Rwanda ) and the one from which demographic parameter estimates were obtained ( from Zimbabwe ) , and/or an effect of positive selection making Africa-Europe diversity comparisons more locally heterogeneous than expected under neutrality . We responded by expanding the window size used in the empirical data analysis . Windows were based on numbers of non-singleton polymorphic sites among the 22 RG primary core genomes . We chose a window size of 1000 such SNPs , which corresponds to a median window size close to 50 kb . Smaller windows led to noisier likelihoods ( results not shown ) , while larger windows might exclude short admixture tracts without an appreciable gain in accuracy . Another concern regarding the empirical data was the effect of sequencing depth on pairwise divergence values . After restricting the admixture analysis to genomes with >25X mean depth , we still observed a minor degree of “wavering” in admixture probabilities for genomes with the lowest depth . We therefore applied a simple correction factor to approximate each genome's quality effects on divergence metrics . In theory , we wish to know the effect of depth and other aspects of quality on DFR . In practice , however , genomes differ in DFR in part based on their level of admixture . Instead , DRG ( average pairwise divergence from the rest of the Rwanda sample ) was used as a proxy . For each chromosome arm , a genome's DRG was compared to the RG population average . Each genome's DFR was then multiplied by the correction factor . Following this correction , no effect of depth on admixture inferences was observed within the primary core data set . Although simulations suggested that our admixture method is robust to ∼10% admixture in the African panel , we sought to maximize the method's accuracy by applying it iteratively to the RG sample . Identical-by-descent regions ( as defined above ) were masked during the creation of emissions distributions , but likelihoods were then evaluated for full RG chromosome arms . After one full “round” of the method ( emissions , likelihoods , and HMM ) , admixture tracts were masked from the RG sample . This masked RG sample became the revised African panel for a second round of analysis , this one with a more accurate emissions distribution for the non-admixed state ( since it contains more true “Africa-Europe” comparisons , and is presumably less influenced by admixture ) . Admixture masking for RG was redone based on round 2 admixture intervals , and the re-masked RG data was used to create a third and final set of emissions distributions . The round 3 emissions distributions were used to generate final admixture calls not only for the RG sample , but also for the other African genomes in the primary core data set . The use of RG as an “African panel” when examining admixture in other African populations is not without concern . Fortunately , in addition to being the largest African sample , RG also occupies a genetically intermediate position within Africa ( see results section ) , which reduces the potential impact of genetic structure on the accuracy of admixture inferences for non-RG genomes . It also appears that aside from the effects of admixture , no other African sample has a much closer relationship to FR than RG does ( see results section ) , thus mitigating a potential source of bias . Standard linear regression was used to investigate the possible relationship between cosmopolitan admixture proportion ( for a population sample ) and the human population size of the collection locality ( city , town , or village population size ) . Census-based population estimates were obtained from online sources for 15 of 20 population samples . For the remainder , satellite-based estimates were obtained from fallingrain . com ( Table S1 ) . While a set of uniform and perfectly accurate population figures is not available for these locations , the estimates used here may still allow a significant effect of human population size on cosmopolitan admixture proportion to be detected . The centiMorgan length of each admixture interval was calculated based on recombination rates inferred from smoothed genetic map data [28] . The extra buffer windows added to each side of conservative admixture tract delimitations described above were not included in these length estimates . CentiMorgan tract lengths were then used with a method [36] that estimates three parameters of a migration rate change model: the current migration rate , the previous migration rate , and the time of migration rate change . A minimum detectable tract length of 0 . 5 cM was chosen , corresponding to roughly 200 kb or 4 windows on average . Forward simulations [36] including recombination , migration , and drift were performed under the estimated demographic model . Simulated data were compared to empirical data , to test how often simulated variance in cosmopolitan admixture proportion exceeded that observed in the RG sample . Regions of lower recombination proximal to centromeres and telomeres were excluded from most analyses , except where indicated below . Recombination rates were taken from mapping-based estimates [28] , and the threshold between “low” and “high” recombination rates was defined as 2×10−8 cross-overs per bp per generation . In most cases , a single transition point was apparent where a chromosome arm transitioned from low to high recombination , moving away from a centromere or telomere . A few narrow “valleys” of recombination rate estimates slightly below this threshold within broader high recombination regions , along with one peak of recombination rate slightly above this threshold close to the 3L centromere , were ignored in the definition of centromere-proximal and telomere-proximal boundaries . “Mid-chromosomal intervals” reflecting the higher recombination intervals used in this analysis for each chromosome arm were: X:2 , 222 , 391–20 , 054 , 556 , 2L:464 , 654–15 , 063839 , 2R:9 , 551 , 429–20 , 635 , 011 , 3L:1 , 979 , 673–12 , 286 , 842 , 3R:12 , 949 , 344–25 , 978 , 664 . Principal components analysis ( PCA ) was conducted using the method of Patterson et al . [43] . Mid-chromosomal data from all primary core genomes were included . The analysis was run twice , on data sets with and without admixture filtering . Applying additional filters ( excluding sites with >5% missing data or <2 . 5% minor allele frequency ) had little effect on results . Nucleotide diversity ( π ) was initially calculated in 100 kb windows , and weighted values for each population sample ( based on the number of sites in each window with data from at least two genomes ) were then averaged to obtain a population's mean absolute π for each chromosome arm . Relative π was calculated by obtaining the ratio of window π from a given population versus that for the RG population ( the largest African sample ) , and window ratios were weighted by the number of sites with data from two or more RG genomes . Relative π values should therefore be robust to cases where a population has large blocks of masked data in a genomic region with especially high or low diversity ( since π in each window is standardized by that observed for the RG sample ) , which could bias estimates of absolute π . Genome-wide relative π was calculated as the unweighted average value of the five major chromosome arms . Three samples ( CK , RC , SP ) had only one primary core genome , but one or more secondary core genomes . Relative π for these samples was calculated based on comparisons between primary and secondary core genomes , both for the target sample and for RG ( which also contains primary and secondary core genomes ) . A similar re-estimation of relative π for the CO sample yielded genome-wide relative π of 0 . 914 from primary-secondary comparisons , versus 0 . 927 from primary core genomes only . Dxy , the average rate of nucleotide differences between populations , was calculated for a subset of populations with high levels of genomic coverage in the admixture-filtered data ( CO , ED , FR , GA , GU , KR , NG , RG , TZ , UG , ZI , ZS ) . FST was calculated using the method of Hudson et al . [58] , with equal population weightings regardless of their sample sizes . Arm-wide and genome-wide estimates of both statistics were calculated as described above for relative π . Using the above summary statistics , we calculated the ratio of a population's DZI ( genetic distance from the four Zambia ZI genomes ) to πZI . Here , the intention was to test which populations contained unique genetic diversity not observed in the maximally diverse ZI population , leading to ratios greater than one . The significance of ratios greater than one was assessed via a bootstrapping approach . Windows 100 kb in length were sampled with replacement until 667 were drawn , to match the number present in non-centromeric , non-telomeric regions of the empirical data . One million such replicates were conducted for each population , and the proportion of replicates with a ratio less than one became the bootstrapping P value . The use of windows much larger than the scale of linkage disequilibrium implies a conservative test . For each population's genome-wide relative π ( Figure 6 ) , and for the DZI to πZI ratio ( Table 2; described ) , we applied a correction factor to reduce the predicted influence of sequencing depth on these quantities . From a linear regression of primary core genomes' sequencing depth versus DZI ( Figure 2 ) , the slope and y intercept of this relationship were obtained . Based on population mean sequencing depth , a population's predicted DZI was compared to the predicted DZI of the reference population ( RG for π , ZI for the ratio analysis ) . Observed summary statistic were multiplied by the ratio of these predicted values to obtain a corrected estimate . For both statistics , this adjustment led to changes of ∼1% or less . In addition to the standard correlation coefficient ( r2 ) of linkage disequilibrium ( LD ) , we also examined directional LD via the rω statistic [70] , [71] . Here , LD is defined as positive if minor alleles preferentially occur on the same haplotype , and otherwise LD is negative . Empirical LD patterns were compared to data simulated under neutral evolution and equilibrium demography using ms [86] . In these simulations , the population mutation rate was taken from observed π . The population recombination rate was then inferred from the ratio of empirical estimates of recombination rates ( the average rate from Langley et al . [28] for the analyzed X-linked and autosomal regions , simulated separately ) and mutation rate [87] . Estimates for the rate of gene conversion relative to crossover events ( 5x ) and the average gene conversion tract length ( 86 . 5 bp ) were taken from a weighted average of the locus-specific estimates obtained by Yin et al . [85] . The Λmax statistic of Sweepfinder [75] uses allele frequencies to evaluate the relative likelihood of a selective sweep versus neutral evolution . To add information regard diversity reductions , we implemented the approach of Pavlidis et al . [76] to include a fraction of the invariant sites . One invariant site was added to the input for every 10 invariant sites that had <50% missing data . Likelihoods were evaluated for 1000 positions from each window . The folded allele frequency spectrum from short intron sites ( see below ) was used for background allele frequencies , assumed by the method to represent neutral evolution . Local outliers for Λmax and FST were examined in overlapping windows of 100 RG non-singleton SNPs ( roughly 5 kb on average ) . For FST , overlapping windows were offset by increments of 20 RG non-singleton SNPs , in order to identify outlier loci that could result from adaptive population differentiation . Outlier windows were defined by the upper 2 . 5% ( FST ) or 5% ( Λmax ) quantile for each chromosome arm . The lower threshold for FST avoids an excessive number of outliers due to the greater number of ( overlapping ) windows , compared to the non-overlapping windows for Λmax . Outliers with up to two non-overlapping non-outlier windows between them were considered as part of the same “outlier region” , since they might reflect a single evolutionary signal . For FST , the center of an outlier region was defined as the midpoint of its most extreme window . The nearest gene to an outlier region was calculated based on the closest exon ( protein-coding or untranslated ) to the above location , based on D . melanogaster genome release 5 . 43 coordinates obtained from Flybase . Two FST outlier analyses were conducted . One , with the aim of identifying loci that may have contributed to the adaptive difference between African and cosmopolitan populations , focused on FST between the FR and RG population samples . The other scan was intended to search for potential adaptive differences among African populations . The nine population samples with a mean post-filtering sample size above 3 . 75 were included ( CO , ED , GA , GU , NG , RG , UG , ZI , ZS ) . The mean FST from all pairwise population comparisons was evaluated for each window , and outlier regions for this overall FST were obtained . Each population was also analyzed separately , in terms of the mean FST from eight pairwise population comparisons . Here , outliers were analyzed separately for each African population , but the lists of population-specific outliers were also combined for more statistically powerful enrichment tests . The enrichment of gene ontology ( GO ) categories among sets of outliers was evaluated . For each GO category , the number of unique genes that were the closest to an outlier region center ( see above ) was noted . A P value was then calculated , representing the probability of observing as many ( or more ) outlier genes from that category under the null hypothesis of a random distribution of outlier region centers across all windows . Calculating null probabilities based on windows , rather than treating each gene identically , accounts for the fact that genes vary greatly in length , and hence in the number of windows that they are associated with . P values were obtained from a permutation approach in which all outlier region center windows were randomly reassigned 10 , 000 times ( results not shown ) . | Improvements in DNA sequencing technology have allowed genetic variation to be studied at the level of fully sequenced genomes . We have sequenced more than 100 D . melanogaster genomes originating from sub-Saharan Africa , which is thought to contain the ancestral range of this model organism . We found evidence for recent and substantial non-African gene flow into African populations , which may be driven by natural selection . The data also helped to refine our understanding of the species' history , which may have involved a geographic expansion from southern central Africa ( e . g . Zambia ) . Lastly , we identified a large number of genes and functions that may have experienced recent adaptive evolution in one or more populations . An understanding of genomic variation in ancestral range populations of D . melanogaster will improve our ability to make population genetic inferences for worldwide populations . The results presented here should motivate statistical , mathematical , and computational studies to identify evolutionary models that are most compatible with observed data . Finally , the potential signals of natural selection identified here should facilitate detailed follow-up studies on the genetic basis of adaptive evolutionary change . | [
"Abstract",
"Introduction",
"Results",
"Conclusions",
"and",
"Prospects",
"Materials",
"and",
"Methods"
] | [
"genome",
"evolution",
"population",
"genetics",
"genome",
"sequencing",
"genome",
"databases",
"model",
"organisms",
"genetic",
"polymorphism",
"comparative",
"genomics",
"biology",
"evolutionary",
"genetics",
"haplotypes",
"natural",
"selection",
"genetics",
"gene",
"flow",
"genomics",
"evolutionary",
"biology",
"genomic",
"evolution",
"computational",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Population Genomics of Sub-Saharan Drosophila melanogaster: African Diversity and Non-African Admixture |
Mutations in leucine-rich repeat kinase 2 ( LRRK2 ) are the most common cause of familial Parkinson disease . Genetics and neuropathology link Parkinson disease with the microtubule-binding protein tau , but the mechanism of action of LRRK2 mutations and the molecular connection between tau and Parkinson disease are unclear . Here , we investigate the interaction of LRRK and tau in Drosophila and mouse models of tauopathy . We find that either increasing or decreasing the level of fly Lrrk enhances tau neurotoxicity , which is further exacerbated by expressing Lrrk with dominantly acting Parkinson disease—associated mutations . At the cellular level , altering Lrrk expression promotes tau neurotoxicity via excess stabilization of filamentous actin ( F-actin ) and subsequent mislocalization of the critical mitochondrial fission protein dynamin-1-like protein ( Drp1 ) . Biochemically , monomeric LRRK2 exhibits actin-severing activity , which is reduced as increasing concentrations of wild-type LRRK2 , or expression of mutant forms of LRRK2 promote oligomerization of the protein . Overall , our findings provide a potential mechanistic basis for a dominant negative mechanism in LRRK2-mediated Parkinson disease , suggest a common molecular pathway with other familial forms of Parkinson disease linked to abnormalities of mitochondrial dynamics and quality control , and raise the possibility of new therapeutic approaches to Parkinson disease and related disorders .
Parkinson disease is the second most common neurodegenerative disorder , following Alzheimer disease . Parkinson disease has a prevalence of approximately 1% at age 65 , which rises to nearly 5% by age 85 [1 , 2] . Particularly given the increasing age of the United States population , Parkinson disease represents a significant economic burden to the healthcare system and to patients and their caregivers . There are currently no treatments that alter the course of this progressive and debilitating disorder . Parkinson disease has classically been defined as a movement disorder with loss of dopaminergic neurons from the substantia nigra and the pathological finding of aggregated α-synuclein in Lewy bodies within affected neurons . More generally , Parkinson disease belongs to a larger group of neurodegenerative parkinsonian syndromes . These parkinsonian syndromes include progressive supranuclear palsy and corticobasal degeneration , degenerative disorders characterized pathologically by abnormal aggregation and deposition of the microtubule-binding protein tau into neurofibrillary inclusions in neurons , and glia [3 , 4] . Diseases manifesting pathologically by neurofibrillary tau pathology are termed “tauopathies . ” While parkinsonian disorders are typically regarded in terms of motor dysfunction , more recently , there has been increased recognition of nonmotor features , including psychiatric , cognitive , and autonomic dysfunction , which reflect system degenerations outside of nigrostriatial pathways and contribute significantly to patient morbidity [4–8] . Most Parkinson disease is apparently sporadic , but penetrant single-gene mutations can give rise to the disorder , and analysis of the function of encoded proteins has significantly advanced our understanding of the molecular pathogenesis of the disease [9 , 10] . Mutations in leucine-rich repeat kinase 2 ( LRRK2 ) , a large multidomain protein kinase of as yet incompletely understood function , represent the most common genetic cause of Parkinson disease and predispose to sporadic Parkinson disease as well [11 , 12] . Clinically , patients can present as Parkinson disease , Parkinson disease dementia , or dementia with Lewy bodies . Pathologically , LRRK2-associated disorders can be accompanied by a variety of neuropathologies , including aggregation and deposition of the microtubule-binding protein tau [13 , 14] . Mutations in the locus encoding tau have been linked to Parkinson disease [15–17] , further connecting the biology of tau and sporadic Parkinson disease . We have previously demonstrated that binding and abnormal stabilization of actin by tau is an important mediator of neurotoxicity in vivo [18] . In tauopathy models , excess stabilization of actin leads to altered mitochondrial dynamics through mislocalization of the critical mitochondrial fission protein dynamin-1-like protein ( Drp1 ) . Subsequent increased accumulation of oxidative free radicals promotes premature neuronal cell death . While the function of LRRK2 in normal physiology and disease states remains incompletely characterized , proteomic , biochemical [19 , 20] , and cell biological [21–25] studies have suggested a role in regulating actin dynamics , perhaps through direct binding to actin [19] . Motivated by the genetic , neuropathological , and cell biological connections between Parkinson disease and tau , we performed studies in our well-characterized Drosophila model of tauopathy , which demonstrate enhancement of tau neurotoxicity in vivo by either knockdown or overexpression of Lrrk , the single fly homolog of mammalian LRRK proteins . Tau neurotoxicity is further enhanced by expression of Lrrk carrying mutations homologous to Parkinson disease mutations in human LRRK2 . These findings raise the intriguing possibility that LRRK2 mutations may act at least partially through a loss of function mechanism , perhaps via dominant negative effects in the context of autosomal dominant human mutations . In support of a possible dominant negative effect , we demonstrate inhibition of LRRK2-mediated actin severing by mutant LRRK2 in biochemical mixing experiments . In vivo , manipulation of Lrrk levels enhances tau neurotoxicity by stabilizing the actin cytoskeleton and promoting mislocalization of Drp1 , leading to mitochondrial dysfunction .
Unlike vertebrates in which the presence of two highly related family members , LRRK1 and LRRK2 , may complicate genetic analysis , Drosophila contains a single LRRK protein , Lrrk [26] . To explore a possible interaction between tau and Lrrk , we first determined if either loss of endogenous fly Lrrk or overexpression of Lrrk could influence neurotoxicity in our genetically accessible transgenic Drosophila model of tauopathy . Our model is based on expression of either human wild-type or the familial neurodegenerative tauopathy , frontotemporal dementia with parkinsonism linked to Chromosome 17 ( FTDP-17 ) –associated mutant tau using the GAL4/UAS bipartite expression system with the panneuronal elav-GAL4 driver [27] . In the current Drosophila studies , we expressed human tau carrying the R406W mutation , which we refer to here as “tau” for simplicity . Neuronal expression of R406W mutant tau results in a level of neurotoxicity that is easily manipulated experimentally , with good conservation of the underlying mechanisms of toxicity with the wild-type human tau [28–31] . When we reduced Lrrk levels using either a homozygous protein-null mutant Lrrk allele or transgenic RNAi [32] , we found significant enhancement of tau neurotoxicity , as assessed by the cleavage of a transgenic caspase reporter ( Fig 1A ) [30 , 33] . Interestingly , expression of wild-type fly Lrrk also enhanced tau neurotoxicity , with a further enhancement in neurotoxicity when Lrrk harboring mutations homologous to Parkinson disease—associated mutations in human LRRK2 were expressed [32] . No significant toxicity was observed when Lrrk levels were manipulated in the absence of transgenic human tau ( Fig 1A ) . Caspase activation within neurons was confirmed by costaining with the neuronal marker elav ( Fig 1B ) . Gain or loss of Lrrk function did not modulate toxicity by simply altering levels of transgenic tau , as determined by western blot analysis ( Fig 1C ) . We have previously shown that cell death occurs through inappropriate activation of the cell cycle in our tauopathy model . Cell cycle activation can be monitored by expression of proliferating cell nuclear antigen ( PCNA ) [28] . Aberrant cell cycle activation in tau transgenic flies with Lrrk manipulation paralleled neurotoxicity , as monitored by caspase activation , supporting a role for cell cycle activation downstream of Lrrk-mediated tau neurotoxicity ( Fig 1D ) . To ensure that the enhancement in toxicity with the Lrrk mutants was not due to greater levels of overexpression , we monitored the expression levels of wild-type and mutant Lrrk and found that wild-type and mutant Lrrk proteins were expressed at similar levels ( Fig 1E ) . To examine the specificity of the interaction between Lrrk and tau , we examined an unrelated model of age-dependent neurodegeneration in Drosophila . We expressed mutant SCA3 , a polyglutamine-expanded protein linked to spinocerebellar ataxia type 3 ( Machado—Joseph disease ) , using the panneuronal driver elav-GAL4 and assessed Kenyon neuron degeneration , as we have described previously [29 , 34] . There was no alteration of mutant SCA3 neurotoxicity with manipulation of Lrrk levels , supporting a specific effect on the pathways mediating tau neurotoxicity ( S1 Fig ) . To investigate the role of Lrrk kinase activity in enhancement of tau neurotoxicity , we expressed a mutant form of Drosophila Lrrk carrying three point mutations ( K1781M , D1882A , and D1912A , analogous to K1906M , D1994A , and D2017A in human LRRK2; Lrrk-3KD ) , which are predicted to disrupt ATP binding [32 , 35] . Expression of Lrrk-3KD was less effective at enhancing tau neurotoxicity compared to wild-type Lrrk ( S1 Fig ) , consistent with a contribution of kinase activity to the toxic effects of Lrrk in tau transgenic flies . We next addressed the cellular mechanism by which Lrrk enhances tau neurotoxicity . We have previously shown that tau binds to and stabilizes actin and that actin stabilization is critical for tau neurotoxicity using genetic analyses [36] . LRKK2 has also been implicated in regulation of the actin cytoskeleton [19–24] . We thus assessed the effect of Lrrk on actin dynamics . To monitor stabilization of the actin cytoskeleton in multiple genotypes in parallel , we performed filamentous actin ( F-actin ) enzyme-linked immunosorbent assays ( ELISA ) on brain lysates . We examined brains of tau transgenic flies homozygous for the Lrrk mutant allele or expressing either wild-type Lrrk or Lrrk carrying the G1914S mutation ( Lrrk-GS ) , the mutation homologous to the most common Parkinson disease—associated mutation in human LRRK2 ( G2019S ) . There was a robust increase in the levels of F-actin in the brains of tau transgenic flies with Lrrk manipulation ( Fig 2A and 2B ) . We have previously demonstrated the presence of actin-rich rods similar to Hirano bodies in the brains of tau transgenic flies [36] . We found a significant increase in the number of actin rods in the brains of tau transgenic flies with Lrrk manipulation ( Fig 2C and 2D ) . To ensure that the effects of overexpressing Lrrk reflected the function of the endogenous protein , we expressed wild-type Lrrk along with tau in a Lrrk mutant background . Total levels of Lrrk were normalized by removing endogenous Lrrk ( Fig 2E ) . Accordingly , the enhancement of caspase activation ( Fig 2F ) , cell cycle activation ( Fig 2G ) , and the number of actin rods ( Fig 2H ) were all rescued by reducing Lrrk levels . Note that manipulating Lrrk expression did not alter expression of transgenic tau ( Fig 1C ) . The enhancement in neurotoxicity and the number of actin rods seen in the Lrrk mutant animals were also rescued by mutant Lrrk-GS , indicating that the disease-linked mutant retains the activity of wild-type Lrrk ( S2 Fig ) . We also assessed the ability of human LRRK2 to enhance tau neurotoxicity in our Drosophila model . Expressing human LRRK2 with tau enhanced the neurotoxicity of tau , as assessed by caspase activation ( S2 Fig ) . Cell cycle activation was also increased with LRRK2 coexpression , as were the number of actin rods ( S2 Fig ) , suggesting that the same pathways of toxicity are activated with expression of human LRRK2 as are activated with expression of Drosophila Lrrk . We have shown using genetic and biochemical methods that abnormal stabilization of actin by tau promotes mislocalization of the critical mitochondrial fission protein Drp1 , leading to elongation of mitochondria [18] . To determine if there is Drp1 mislocalization and altered mitochondrial morphology in tau transgenic flies in Lrrk-modified backgrounds , we expressed two additional transgenes . To visualize mitochondria , we expressed mitochondrially directed green fluorescent protein ( mito-GFP ) . To visualize Drp1 , we used a 9 . 35-kb genomic rescue strain that has an in-frame FLAG-FIAsH-hemagglutinin ( HA ) tag after the start codon of Drp1 , thus expressing tagged Drp1 under its endogenous promoter [18 , 37] . Manipulation of Lrrk promoted additional mitochondrial elongation in the brains of tau transgenic flies ( Fig 3A and 3B ) . There was also a significant loss of localization of Drp1 to mitochondria in tau transgenic flies with manipulated Lrrk backgrounds ( Fig 3A and 3C ) . The colocalization analysis was confirmed by computing the Pearson’s correlation coefficient . The average value of the Pearson’s coefficient for the cells analyzed decreased in tau transgenic flies compared to control , with an even further decrease in tau transgenic flies with Lrrk manipulations ( Fig 3D ) , indicating an enhancement in the loss of Drp1 localization to the mitochondria . Utilizing a computational method to assess mitochondrial interconnectedness [38] , we observed an increase in the degree of mitochondrial interconnectedness in tau transgenic flies with Lrrk manipulation ( Fig 3E ) . Drp1 protein levels remained unchanged in the heads of flies with Lrrk genetic manipulation ( Fig 3F ) , demonstrating that the loss of Drp1 localization to the mitochondria was not due to a reduction in Drp1 levels . Since loss of dopaminergic neurons is clinically important in Parkinson disease , we examined these monoaminergic neurons in our model . Examination of tyrosine hydroxylase ( TH ) immunoreactive neurons in the anterior medulla of flies expressing human tau [39] revealed loss of TH-positive neurons . Loss of TH immunoreactive neurons was enhanced when Lrrk levels were reduced or if either wild-type or mutant Lrrk was expressed in tau transgenic flies ( S3 Fig ) . Dopaminergic pathology has previously been reported with expression of human LRRK2 in transgenic Drosophila [40 , 41] , although at later time points than were examined in the current study . Drp1 mislocalization and mitochondrial elongation are accompanied by increased levels of reactive oxygen species ( ROS ) production in the brains of tau transgenic flies [18] . To determine if changes in mitochondrial dynamics by the manipulation of Lrrk correlate with an increase in markers of mitochondrial stress , we first used MitoSOX on freshly dissected whole brains . The MitoSOX dye permeates live cells , in which it is targeted to the mitochondria and becomes oxidized by superoxide , resulting in strong red fluorescence , which can be detected by microscopy [42] . Fluorescence in the brains of flies with manipulated Lrrk along with tau expression was greater than that in control flies ( Fig 3G and 3H ) , consistent with elevated levels of mitochondrial superoxide . Next , we used a genetic reporter , MitoTimer , which is targeted to the mitochondria , and the fluorescence shifts from green to red on oxidation . The ratio of red to green fluorescence gives a measure of mitochondrial stress [43] . We observed an increase in oxidized MitoTimer protein in the brains of tau transgenic flies with altered Lrrk function ( Fig 3I ) . To measure intracellular ROS , we used ROSstar 650 , a hydrocyanine-based probe designed to detect specifically superoxide and hydroxyl radicals [44] . There was an increase in ROS levels in the brains of flies expressing tau with Lrrk manipulation ( Fig 3J ) . To explore the translational potential of our findings , we treated tau transgenic flies with actin-depolymerizing drugs . Latrunculin A binds actin monomers and thereby prevents their polymerization [45] . Consistent with our prior genetic data [18 , 36 , 46] , feeding flies 5 or 10 μM latrunculin A produced a robust , dose-dependent rescue of tau neurotoxicity ( Fig 4A ) . We also observed a dose-dependent decline in cell cycle activation with oral administration of latrunculin A ( Fig 4B ) . Cytochalasins are a widely studied family of compounds , which disrupt actin polymers [47] . We found that cytochalasin B feeding also robustly rescued neurotoxicity in tau transgenic flies; cytochalasin D was moderately effective ( Fig 4C and 4E ) . We observed rescue of cell cycle activation by the two compounds , with cytochalasin B having greater efficacy than cytochalasin D ( Fig 4D and 4F ) . All three compounds were also able to rescue the enhanced neurotoxicity of tau expressed in a Lrrk mutant background or with increased Lrrk expression ( S4 Fig ) . Latrunculin A , cytochalasin B , or cytochalasin D did not alter the levels of tau ( Fig 4G ) . Actin-targeting compounds also rescued the mitochondrial dysfunction in tau transgenic flies , as indicated by a reduction in the oxidation of the MitoTimer protein ( Fig 4H ) and decreased MitoSOX fluorescence ( Fig 4I and 4J ) . To investigate the interaction of LRRK2 and tau in a mammalian system , we used transgenic mice expressing FTDP-17-associated P301L mutant human tau in the forebrain via a calcium/calmodulin-dependent protein kinase II reverse transactivator ( CAMKIIα-tTA ) transgene [48] crossed to wild-type LRRK2 bacterial artificial chromosome ( BAC ) transgenic mice . These tau transgenic mice have robust neurodegeneration , including in the cornu ammonis 1 ( CA1 ) region of the hippocampus , reflecting expression of the tau transgene [49] in an anatomic area involved in common tauopathies and α-synucleinopathies [50–52] . The LRRK2 BAC mice have no overt phenotypic or pathological abnormalities at baseline [53] . We observed increased neuronal loss in the CA1 region of transgenic mice expressing both tau and LRRK2 ( Fig 5A–5C ) . These human tau transgenic mice also have actin-rich rods similar to Hirano bodies in their brains [36] . The number of actin rods in the brains of mice expressing both tau and LRRK2 was increased ( Fig 5D and 5E ) . Rods were most abundant in the deep gray nuclei , including the basal ganglia and thalamus , consistent with prior observations [36] . We next assessed Drp1 localization and mitochondrial morphology in the murine model . We have previously demonstrated mislocalization of Drp1 and elongated mitochondria in these tau transgenic mice [18] . The failure of Drp1 to localize to the mitochondria was enhanced in transgenic mice expressing human tau together with human LRRK2 ( Fig 5F and 5G ) . The average length and the interconnectedness of the mitochondria in hippocampal pyramidal neuronal cell bodies of mice expressing both tau and LRRK2 were increased ( Fig 5H and 5J , S5 Fig ) . We have previously shown that the expression levels of transgenic tau are not affected by the expression of LRRK2 and that LRRK2 expression from the BAC transgenic construct remains robust at the 5 . 5 month time point examined in the current study [48] . To explore the mechanism by which Lrrk influences actin dynamics , we performed in vitro depolymerization assays using pyrene-labeled actin and recombinant human wild-type LRRK2 and LRRK2 with the G2019S mutation ( LRRK2-GS ) . Since LRRK2 mutations causing Parkinson disease can act in an autosomal dominant fashion , with both wild-type and mutant protein being present , we also included a mixture of wild-type and mutant recombinant proteins in a 1:1 ratio ( mix ) in our experiments . The LRRK2 proteins were used at 1 nM , a concentration that is comparable to estimated endogenous levels of LRRK2 in brain [54] . Wild-type LRRK2 and LRRK2-GS as well as the mix enhanced the depolymerization of actin filaments to varying degrees ( Fig 6A ) . The initial depolymerization rate , measured by computing the slope of the depolymerization curve in the first 10 minutes , was similar for the wild type and the mix ( Fig 6B and 6C ) . However , the fold increase in depolymerization at the end of 1 hour was greater for the wild type and LRRK2-GS compared to the mix , indicating a misregulation of actin dynamics in the presence of both wild type and LRRK2-GS ( Fig 6D ) . To investigate the mechanism underlying altered behavior of the mixture biochemically , we examined the oligomerization state of the proteins . We incubated the recombinant LRRK2 proteins at room temperature for 1 hour and visualized oligomeric species on a native gel by silver staining . We observed an increase in the ratio of oligomeric to monomeric species over time , with greater oligomerization in the mix compared to the wild type or LRRK2-GS alone ( Fig 6E and 6F ) . To assess Lrrk oligomerization in vivo , we used a LrrkHA knockin line in which HA-tagged Lrrk is expressed from its endogenous promotor [55] . We expressed transgenic wild-type or mutant Lrrk-GS in flies also carrying one LrrkHA allele and used head lysates to perform native polyacrylamide gel electrophoresis ( PAGE ) . Probing the resultant western blots with an antibody directed to HA revealed dimerization as well as higher-order oligomerization in flies overexpressing wild-type Lrrk protein , which was enhanced in the flies that express Lrrk-GS ( S5 Fig ) . To determine how LRRK2 promotes depolymerization of actin filaments , we assayed actin filament—severing activity . We observed that wild-type LRRK2 , LRRK2-GS , as well as the mix could sever polymerized F-actin filaments , with the severing activity of the mix being intermediate between that of the wild type and LRRK2-GS ( Fig 6G and 6H ) . The severing experiment in Fig 6G and 6H was performed with freshly prepared LRRK2 and actin , thus reflecting the initial phase of the depolymerization studies ( Fig 6B and 6C ) , in which we expect more monomeric species to be present in all the samples ( Fig 6E ) . When we repeated the actin severing assay after incubating LRRK2 , LRRK2-GS , and the mix at room temperature for 1 hour , thus allowing oligomer formation , we saw a significant reduction in the actin-severing activity of the mix ( Fig 6I and 6J ) . The reduction in the severing activity in the mixture of wild-type and mutant LRRK2 is consistent with increased oligomer formation in the mix ( Fig 6E and 6F ) , suggesting a link between the actin-severing activity and the oligomeric state of LRRK2 . To confirm that Lrrk and F-actin interact in vivo , we precipitated F-actin biochemically using biotinylated phalloidin . We observed coprecipitation of Lrrk with F-actin in head homogenates from flies ( Fig 6K ) . Similarly , Lrrk and F-actin colocalized in sections from fly brains ( Fig 6L and 6M ) . Consistent with a close interaction between F-actin and mitochondria [18 , 56–58] , Lrrk also colocalized with mitochondria in fly brains ( S6 Fig ) . The model in Fig 6N predicts that strong loss of Lrrk function should promote F-actin stabilization , even in the absence of transgenic human tau expression . While we did not observe clear alterations in the actin cytoskeleton ( Fig 2 ) or mitochondria ( Fig 3 ) in Lrrk mutant animals at the 10-day time point we typically use to assess tau neurotoxicity in our model ( Fig 1 ) [18 , 28 , 36] , we did observe modest but statistically significant abnormalities in older animals . When we examined actin stabilization in flies with loss of endogenous Lrrk aged to 20 days , we observed actin stabilization as assessed by phalloidin staining and precipitation of F-actin using biotinylated phalloidin ( S7 Fig ) . Increased numbers of actin-rich rods were also present in older Lrrk mutant animals ( S7 Fig ) . As predicted ( Fig 6N ) , Drp1 mislocalization and mitochondrial elongation were also observed in the neurons of these older mutant flies ( S7 Fig ) . Note , however , that the changes in the actin cytoskeleton and mitochondria were significantly less in magnitude than those present in tau transgenic flies ( compare S7D Fig and Fig 2D; S7F Fig and Fig 3B ) , consistent with a lack of overt neurodegeneration in the Lrrk mutant flies ( Fig 1 ) .
While Parkinson disease and the tauopathies , including Alzheimer disease , share some clinical and pathological similarities , the disorders have generally been seen as distinct entities . However , the observation of tau aggregation and deposition as the primary neuropathology in some Parkinson disease patients with LRRK2 mutations and the implication of tau in sporadic Parkinson disease patients through genome-wide association studies has raised the intriguing possibility that the disorders share a common pathogenesis . Here , we describe a plausible molecular basis for the interaction between Lrrk and tau . We have previously demonstrated , using genetic and biochemical approaches , that tau exerts neurotoxicity by binding to and stabilizing F-actin [36] , which promotes mislocalization of the critical mitochondrial fission protein Drp1 and subsequent mitochondrial dysfunction [18 , 57–61] . We now show that the single Drosophila LRRK family homolog Lrrk acts in the same pathway . Manipulating Lrrk expression in tau transgenic flies further increases F-actin levels , decreases Drp1 localization to mitochondria , enhances mitochondrial dysfunction , and promotes neurodegeneration . Fibroblasts from patients with LRRK2-GS mutations show increased numbers of F-actin bundles [21] , consistent with our evidence for abnormal stabilization of F-actin following Lrrk manipulation ( Fig 2A , 2C and 2D ) . To probe the molecular mechanisms underlying in vivo stabilization of F-actin in Lrrk mutant animals , we performed a series of biochemical studies with purified LRRK2 and actin . Addition of LRRK2 to purified actin in vitro has previously been associated with a shift in actin from F-actin to G-actin , as monitored by cosedimentation [19] . Consistent with these findings , we demonstrate here that LRRK2 promotes depolymerization of actin ( Fig 6A–6D ) by directly severing actin filaments ( Fig 6G–6J ) . Importantly , severing activity appears preferentially associated with the monomeric state of LRRK2 , while oligomerization is correlated with loss of LRRK2-mediated severing . Further , we find that disease-associated LRRK2 has an increased propensity to oligomerize with wild-type LRRK2 and , under these conditions , reduced ability to sever actin ( Fig 6I and 6J ) . These biochemical findings have implications for the mechanism of LRRK2 action in disease . Both loss of function [62–64] and gain of function [65] mechanisms have previously been proposed in LRRK2-associated Parkinson disease . Our biochemical data suggest that mutant LRRK2 may have a dominant negative effect . These data fit well with our genetic findings that loss of Lrrk function using either a genetic mutation or transgenic RNAi strongly enhances the toxicity of human tau ( Fig 1A ) . Further , our biochemical data demonstrating that under conditions favoring oligomerization of wild-type LRRK2 actin-severing activity is decreased fits with our observation that increasing expression of wild-type Lrrk enhances tau neurotoxicity ( Fig 1A ) . We thus propose that at lower levels , Lrrk acts to sever actin and maintain normal actin dynamics and downstream mitochondrial dynamics . Disease-associated mutations in LRRK2 promote oligomerization of the protein , reduce actin-severing activity , and result in overstabilization of the actin cytoskeleton ( Fig 6N ) . In the case of the disease-associated mutation G2019S , increased kinase activity may favor formation of dimeric LRRK2 [66] , with a resultant decrease in active , actin-severing monomeric LRRK2 ( Fig 6 , S1 Fig ) . However , we note that expression of a form of Lrrk with mutations in three residues key for kinase activity can still promote tau toxicity to a limited extent ( S1 Fig ) . Thus , Lrrk kinase activity may not be absolutely required for toxicity , or there may be multiple pathways mediating Lrrk toxicity . These findings are consistent with a potential therapeutic benefit to LRRK2 kinase inhibition [67] , particularly in the context of increased LRRK2 expression or activity . The dosage sensitivity to wild-type Lrrk that we demonstrate in vivo ( Fig 1 ) and model biochemically ( Fig 6 ) is consistent with recovery of noncoding Parkinson disease risk variants at the LRRK2 locus , which presumably predispose to disease by modulating LRRK2 expression [11] . Indeed , elevated levels of LRRK2 have been reported in patients with Parkinson disease [68 , 69] . Our findings are also consistent with prior cell culture data demonstrating a strong dosage dependence of LRRK2 toxicity [70] . However , we cannot exclude the possibility that loss of Lrrk function enhances tau neurotoxicity through a different mechanism than overexpression of Lrrk . Indeed , depolymerization [71] as well as excess stabilization of actin inhibits Drp1 localization and disrupts mitochondrial dynamics , consistent with a requirement for proper actin dynamics in localizing Drp1 to mitochondria . Thus , increasing expression of an actin-severing protein ( Fig 6 ) might disrupt mitochondrial Drp1 localization ( Figs 3 and 5 ) by destabilizing actin . However , our data demonstrating increased F-actin and numbers of actin-rich rods when Lrrk is overexpressed argues against this possibility ( Fig 2 ) , and we thus rather favor a dominant negative mechanism for enhancement of tau neurotoxicity by Lrrk manipulation in neurons ( Fig 6 ) . Nonetheless , we note that knockout of LRRK2 is associated with lung and kidney pathology not seen in animals with knockin of Parkinson disease—associated LRRK2 mutations [72 , 73] . Similarly , while actin dynamics in myeloid cells is altered in a Wiskott—Aldrich syndrome protein family member 2 ( WAVE2 ) -dependent fashion in both LRR2 knockout and LRRK-G2019S knockin mice , F-actin levels are decreased in knockout microglia and increased in knockin microglia [25] . Further work will be required to investigate fully the mechanisms underlying LRRK2-mediated dysfunction in various tissues and cell types and their contribution to human disease phenotypes . The multiple , seemingly disparate cellular pathways influenced by LRRK2 in various experimental organisms and systems has been a puzzling aspect of LRRK2 biology [11 , 74–76] . Our data suggest a possible explanation for these diverse results . A primary effect of LRRK on actin stabilization in disease states is consistent with the cellular pathologies previously linked to LRRK dysfunction , including altered vesicle trafficking [72 , 77–83] , miRNA and translational regulation [32 , 84] , and nucleoskeletal changes [85] , because these processes are all regulated by the actin cytoskeleton [46 , 86–89] . Alternatively , LRRK2 could promote disease pathogenesis through multiple pathways , including by influencing phosphorylation and transmission [48 , 90–92] of tau . Consistent with our pharmacological rescue of tau neurotoxicity and mitochondrial dysfunction in vivo ( Fig 4 ) , chemical depolymerization of actin with latrunculin A reverses actin cytoskeletal abnormalities in patient-derived cells [21] . Manipulation of the actin cytoskeleton may thus represent a new target for therapy development in Parkinson disease as well as in other parkinsonian disorders and in the larger group of tauopathies . Although manipulation of the microtubule cytoskeleton for therapeutic purposes has a long history and multiple effective agents in clinical use for diverse disorders ranging from cancer to gout [93] , targeting of the actin cytoskeleton represents a relatively unexplored therapeutic arena [47] . While caution certainly is warranted given the essential roles of the actin cytoskeleton in normal biology , manipulation of actin polymerization by targeting actin-binding proteins has shown promise in preclinical studies in cancer [94] and renal disease [95] . Similarly , in the current studies , we observed rescue of mitochondrial functional defects and neurotoxicity without excessive toxicity following systemic delivery of the pharmacological agents latrunculin A or the cytochalasins ( Fig 4 ) . These findings complement our prior work demonstrating that genetic destabilization of the actin cytoskeleton is protective in tauopathies [18 , 46] . The mitochondrial pathology we observe is consistent with prior reports of elongated , fusiform mitochondria in LRRK2 G2019S knockin mice [96] . In contrast , a separate study did not observe enhanced neuropathology in mice expressing P301S mutant human tau under the control of the mouse prion promoter and coexpressing mutant human LRRK2 R1441G from a bacterial artificial chromosome [97] . The reasons for the differences between our study and the work of Mikhail and colleagues [97] are not clear; however , the divergent results may reflect the forms of tau and LRRK2 expressed and levels and cellular expression patterns of the transgenes . Additional genetic analyses in LRRK2 knockin and knockout rodent models with concomitant expression of wild-type and disease-linked mutant versions of human tau may provide important information regarding the in vivo interactions of these two proteins . Knockout of both LRRK1 and LRRK2 in the brain may be required in these studies [64] . Our current findings are particularly intriguing in the context of Parkinson disease because two other proteins implicated in familial Parkinson disease , parkin and phosphatase and tensin homolog induced putative kinase 1 ( PINK1 ) , play important roles in mitochondrial dynamics and mitophagy [11 , 27 , 98] . Of note , these recessive disorders fit well within the category of parkinsonian neurodegeneration because they have clinical features atypical for sporadic Parkinson disease and are only infrequently characterized by Lewy body formation [7 , 9] . Thus , interference with the molecular machinery controlling mitochondrial dynamics , quality control , and function , rather than a specific neuropathology , may best define the clinicopathological entity that is parkinsonian neurodegeneration . In summary , we have outlined here a molecular pathway that plausibly connects the pathogenesis of the two most common , seemingly disparate neurodegenerative disorders: Parkinson disease and the tauopathies . Our findings correlate with prior implication of mitochondrial dysfunction in recessive forms of Parkinson disease and have important implications for therapy . Our data raise the possibility of a dominant negative , loss-of-function mechanism in LRRK2-related Parkinson disease and thus suggest caution when considering strong or complete inhibition of LRRK2 in therapeutic approaches to the disorder . Therapies that target excess stabilization of F-actin ( Fig 4 ) or mitochondrial dysfunction may be promising avenues for further investigation .
Mice were housed and treated in accordance with the National Institutes of Health ( NIH ) Guide for the Care and Use of Laboratory Animals . All animal procedures were approved and performed in accordance with the Mayo Clinic Institutional Animal Care and Use committee and the University of Florida Institutional Animal Care and Use committee . Mice were maintained in a pathogen-free facility on a 12-hour light/dark cycle with water and food provided ad libitum . All mice were euthanized by cervical dislocation to maintain the brain biochemistry by avoiding anesthesia-induced tau changes . All Drosophila crosses and aging were performed at 25 °C . Assays were performed on 10-day-old flies , with the exception of experiments in Fig 6 , S5 and S6 Figs ( 1–3 days of age ) and S7 Fig ( 20 days of age ) . The panneuronal driver elav-GAL4 was used for all experiments . The human UAS-tauR406W line has been described previously [27] . The following Drosophila stocks were kindly provided by the indicated investigators: Lrrke03680 , UAS-Lrrk RNAi , UAS-Lrrkwt , UAS-LrrkI1915T , and UAS-LrrkY1383C by Dr . Bingwei Lu; UAS-LrrkG1914S by Dr . Ming Guo; LrrkHA knockin animals by Dr . Patrick Verstreken; UAS-CD8-PARP-Venus transgenic caspase reporter by Dr . Darren Williams; FLAG-FlAsH-HA-Drp1 by Dr . Hugo Bellen; UAS-mito-GFP by Dr . Thomas Schwarz; and UAS-SCA3 ( MJD ) -78 by Dr . Nancy Bonini . elav-GAL4 , UAS-EGFP , and UAS-MitoTimer were obtained from the Bloomington Drosophila Stock Center . All mouse experiments for the tau transgenic strains rTg4510 and rTg4510-LRRK2-WT BAC [48] were performed at 5 . 5 months of age . Controls were age- and gender-matched nontransgenic animals . Mice were housed and treated in accordance with the NIH Guide for the Care and Use of Laboratory Animals . All animal procedures were approved and performed in accordance with the Mayo Clinic Institutional Animal Care and Use committee and the University of Florida Institutional Animal Care and Use committee . Mice were maintained in a pathogen-free facility on a 12-hour light/dark cycle with water and food provided ad libitum . The parental TauP301L responder line and parental tTA activator line were generated and maintained on an FVB and 129/S6 background , respectively . The parental bacterial artificial chromosome ( BAC ) -LRRK2 mice were maintained on an FVB background . Mice from the TauP301L responder line were crossed with mice from the BAC-LRRK2 mouse line for one generation to obtain LRRK2 . TauP301L responder mice on an FVB background . LRRK2 . TauP301L responder mice were then crossed with mice from the tTA activator line to obtain the resultant F1 LRRK2/TauP301L mice on a 50% FVB , 50% 129S background [48] . For all Drosophila and mouse experiments , equivalent numbers of male and female animals were used . A polyclonal antibody against Lrrk was generated in rabbit using Lrrk 1–336 amino acids as the epitope using the services of Covance . This antibody was used at a dilution of 1:1000 in 5% BSA for western blotting . The following antibodies were used at the indicated concentrations: mouse monoclonal actin ( JLA20 ) , 1:500 , Developmental Studies Hybridoma Bank; rabbit polyclonal actin ( A2066 ) , 1:1000 , Sigma; HA-11 , 1:200 , Covance; GAPDH , 1:200 , 000 , Abcam; tau ( tau5 ) , 1:20 , 000 , Developmental Studies Hybridoma Bank; elav , 1:50 , Developmental Studies Hybridoma Bank; cleaved PARP ( E51 ) , 1:200 , 000 , Abcam; Drp1 , 1:200 , SantaCruz Biotechnology; GFP ( N86/8 ) , 1:1000 , NeuroMab; ATPVa , 1:500 , Novex; and TH , 1:200 , Immunostar . Drosophila brains were homogenized in 2X sample buffer and analyzed by 4%–12% SDS-PAGE and immunoblotted according to standard protocols . Each blot was repeated at least three times with similar results and quantified using ImageJ . An image of a representative blot is shown in the figures . For native PAGE and silver staining , recombinant wild-type human LRRK2 or G2019S mutant human LRRK2 protein samples ( 5 nM ) were prepared in 2X native gel sample buffer ( 62 . 5 mM Tris-HCL , pH 6 . 8 , 25% glycerol , and 1% bromophenol blue ) . The samples were run on 4%–20% prescast tris gels ( Lonza 58517 ) in running buffer without SDS ( 25 mM Tris , 192 mM Glycine ) . Silver staining was performed on the gel using the Thermo Scientific Pierce Silver Stain Kit ( Cat . #24162 ) following the manufacturer’s protocol . Quantification of the number of neurons with caspase activation and PCNA staining was performed on 4 μm paraffin-embedded tissue sections , and positive cells throughout the entire brain were counted . Mitochondrial length was assessed by imaging Drosophila Kenyon neurons and murine hippocampal pyramidal neurons using laser scanning confocal microscope . Individual mitochondrial length was measured by freehand line length using ImageJ ( http://rsbweb . nih . gov/ij ) . All the mitochondria quantified for length were also scored for colocalization . Mitochondrial interconnectivity was assessed using the ImageJ macro “mitochondrial morphology” publicly available at http://imagejdocu . tudor . lu/doku . php . The average area/perimeter normalized to the average circularity was taken as the measure of mitochondrial interconnectivity . Pearson’s coefficient was calculated by circling individual cells and using the ImageJ macro Coloc2 . Fifty cells were analyzed for each genotype for Pearson’s coefficient calculations . Percent colocalization of Lrrk and F-actin or mitochondria was also assessed using Coloc2 . The number of actin-rich rods was determined by counting all rod-shaped or round structures over 3 microns in size that stained for actin . Actin rods were counted throughout the entire Drosophila brain . All the actin rods experiments in mice were done blind , and the counts were performed without the knowledge of the genotypes . For assessment of fluorescence in Drosophila brains , the samples were processed simultaneously using the same acquisition parameters . For the quantification of fluorescence , average pixel density from two-dimensional projections of z-stacks for the entire brain was computed using ImageJ . The density of Kenyon cells per μm2 was determined on H&E stained paraffin sections . The density of TH-positive cells per μm2 was determined on paraffin sections of the anterior medulla , as described [39] . Freshly dissected brains from 10-day-old flies were homogenized in 50 μl actin stabilization buffer from the G-actin/F-actin in vivo assay kit from Cytoskeleton Inc . ( Cat . #BK037 ) . Ten μl of the homogenate was used in the F-actin ELISA assay from MyBioSource Inc . ( Cat . #MBS702018 ) , and the assay was performed according to the manufacturer’s protocol . The remainder of the homogenate was used in western blotting to ensure equal levels of total actin and the input . Each experiment was performed with two technical replicates . Flies were collected on the day of eclosion and were fed 0 , 5 , or 10 μM of drug dissolved in ethanol and mixed into Drosophila culture medium , as described [99 , 100] . In each vial , 10–12 flies were kept , and the vials were changed every third day and analyzed at 10 days . Sagittal brain sections were stained with H&E to visualize structure and align slides . Matching brain sections for each animal were scanned into the ScanScope XT scanner and visualized through ImageScope version 11 . 2 . 0 . 780 software ( Aperio ) . An individual ( MJH ) who was blinded to genotype and sex of animals outlined the regions of interest . He then performed a direct count of all cells within the CA1 , placing a mark on each cell that was counted . If the matched section was ripped within the region to be counted , that animal was excluded from analysis . Actin depolymerization assays were performed using the fluorescent form of the Actin Polymerization Biochem Kit from Cytoskeleton Inc . ( Cat . #BK003 ) , as described by the manufacturer . Pyrene-labeled actin was polymerized at room temperature for 1 hour in actin polymerization buffer ( containing 2 mM MgCl2 and 2 mM ATP ) and then mixed either with buffer or with human recombinant LRRK2 , LRRK2-G2019S , or a 1:1 mixture of the two . Actin was used at a final concentration of 4 μM and the LRRK2 proteins at 1 nM . The samples were read every 30 seconds in a plate reader at the excitation/emission wavelengths of 350/410 for 1 hour . The experiment was repeated three times with two technical replicate each time per sample . The values were normalized to 0 at the starting time . Each data point represents the mean of three separate experiments . Actin ( 4 μM ) was polymerized in actin polymerization buffer ( containing 2 mM MgCl2 and 2 mM ATP ) at room temperature for 1 hour . Polymerized actin was incubated with LRRK2 , LRRK2-GS , or mix ( 1 nM final concentration ) for 2 minutes . Fluorescently labeled phalloidin ( Acti-stain 555 ) was added to a final concentration of 2 μM , and the samples were diluted 50-fold with PBS . Of each sample , 2 μl was adsorbed on coverslips coated with 0 . 01% poly-L-lysine and imaged using a fluorescence microscope . The filament lengths were quantified using ImageJ software by freehand drawing tool . Three different randomly selected areas were quantified for each sample , with at least 10 filaments measured per area . The experiment was repeated three times . Brains from 20-day-old flies were dissected in PBS and fixed in 4% PFA on ice for 30 minutes . After a 10-minute incubation in 0 . 3% Triton-X , the brains were stained with Acti-stain 555 phalloidin ( Cytoskeleton Inc . , Cat . #PHDH1 ) at a concentration of 14 nM for 30 minutes in the dark . Brains were then washed three times in PBS for 60 minutes each , mounted , and imaged using confocal microscopy . Forty fly heads were homogenized and centrifuged at 800 x g to pellet debris . The supernatant was incubated with 0 . 3 units biotinylated phalloidin ( Molecular Probes ) , followed by precipitation with streptavidin-coupled Dynabeads ( Invitrogen ) . Flies were 10 days old in all coprecipitations . All reported n values are biological replicates . The sample sizes used were similar to the ones reported in previous publications [18 , 28] . For Drosophila immunostaining and genetic reporter- and dye-labeling experiments , the sample size was six per genotype and time point . Western blot quantifications were performed on at least three independent blots . The ELISA experiment was repeated three times , with two technical replicates per sample . The sample sizes for mouse experiments were determined by power analyses . For the histological studies , we determined that a sample size of five is sufficient to detect 20% difference between control and experimental genotypes with a power of 80% . Exact sample size for each experiment is provided in the figure legends . Data collection and analysis in mouse experiments were performed blinded to the conditions of the experiment . Experiments in flies were not performed blinded . Statistical analysis was performed using one-way ANOVA , and multiple comparisons among the data sets were performed . Variance was similar between groups compared . | Parkinson disease is a common and debilitating neurodegenerative disorder . Family and larger population-based studies have revealed a number of genes important in the development and progression of the disease , but specific mechanisms linking these seemingly unrelated genes and proteins with specific cellular pathways within affected neurons have often remained elusive . Here , we connect the pathobiology of two proteins associated with Parkinson disease through genetics and neuropathology in patients , leucine-rich repeat kinase 2 ( LRRK2 ) and tau . Using a combination of biochemistry , cell biology , and genetic manipulation in fruit fly and mouse models , we show that LRRK2 and tau interact through the actin cytoskeleton and thereby control mitochondrial dynamics . Specifically , we find that LRRK2 can directly sever actin filaments in purified protein preparations in vitro . Consistent with our biochemical results , we see that altering the levels of LRRK changes the amount of filamentous or polymerized actin within neurons in the brain . We explore the therapeutic implications of our findings by demonstrating that decreasing polymerized actin with drugs can rescue neuronal death in animal models of disease . Our findings also link the biology of LRRK2 and tau with less common forms of familial Parkinson disease caused by mutations in the genes encoding parkin and phosphatase and tensin homolog induced putative kinase 1 ( PINK1 ) , proteins , which have previously been implicated in mitochondrial dynamics and function . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"neurodegenerative",
"diseases",
"cell",
"processes",
"neuroscience",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"drosophila",
"research",
"and",
"analysis",
"methods",
"contractile",
"proteins",
"polymer",
"chemistry",
"actins",
"actin",
"polymerization",
"animal",
"cells",
"proteins",
"animal",
"studies",
"chemistry",
"mouse",
"models",
"insects",
"arthropoda",
"movement",
"disorders",
"biochemistry",
"cytoskeletal",
"proteins",
"cellular",
"neuroscience",
"eukaryota",
"cell",
"biology",
"neurology",
"parkinson",
"disease",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"energy-producing",
"organelles",
"depolymerization",
"organisms"
] | 2018 | Lrrk promotes tau neurotoxicity through dysregulation of actin and mitochondrial dynamics |
Human T Lymphotropic virus ( HTLV ) infection can persist in individuals resulting , at least in part , from viral escape of the innate immunity , including inhibition of type I interferon response in infected T-cells . Plasmacytoid dendritic cells ( pDCs ) are known to bypass viral escape by their robust type I interferon production . Here , we demonstrated that pDCs produce type I interferons upon physical cell contact with HTLV-infected cells , yet pDC activation inversely correlates with the ability of the HTLV-producing cells to transmit infection . We show that pDCs sense surface associated-HTLV present with glycan-rich structure referred to as biofilm-like structure , which thus represents a newly described viral structure triggering the antiviral response by pDCs . Consistently , heparan sulfate proteoglycans and especially the cell surface pattern of terminal β-galactoside glycosylation , modulate the transmission of the immunostimulatory RNA to pDCs . Altogether , our results uncover a function of virus-containing cell surface-associated glycosylated structures in the activation of innate immunity .
Human T-Lymphotropic Virus type 1 ( HTLV-1 ) infects over an estimation of 5–10 million people . HTLV-1 is mainly present in Japan , central Africa , Caribbean and South America [1 , 2] . After a long period of clinical latency , HTLV-1 infection leads , in a fraction of infected individuals , either to Adult T-cell Leukemia/Lymphoma ( ATL ) [3] an uncontrolled CD4+ T–cell proliferation of very poor prognosis , or to an inflammatory disorder named HTLV-1 Associated Myelopathy / Tropical Spastic Paraparesis ( HAM/TSP ) [4] . In chronically infected individuals , HTLV-1 provirus is mainly found in CD4+ T-cells , yet infected dendritic cells ( DCs ) are also detected [5 , 6] . Their function is subsequently altered in vivo [6–8] , thereby most likely contributing to viral pathogenesis . Viral persistence leading to chronic infection and its associated diseases implies that innate and adaptive immune responses fail to eliminate HTLV-1 infected cells , possibly because HTLV-1 has evolved efficient strategies to escape immune pathways [9] . Type-I interferons ( referred herein to as IFN-I , i . e . , IFNα and β ) are key mediators of innate immunity . They induce the expression of IFN-stimulated genes ( ISGs ) that suppress viral spread at different stages of the viral cycle , and stimulate the onset of adaptive immune responses . The IFN-I response is initiated via the recognition of pathogen-associated molecular patterns ( PAMPs ) by pattern-recognition receptors ( PRRs ) , including the Toll like receptors ( TLRs ) [10] . Like virtually all viruses [11] , HTLV-1 inhibits several steps of the PRR-induced pathways [12–14] , and as a consequence , blunts IFN-I induction and signaling [15 , 16] , leading to very limited production of IFN-I by infected cells . Because the acute phase of the infection is asymptomatic , very little is known regarding host innate responses in HTLV-1-infected individuals . Nonetheless , indirect evidence infers that the IFN-I response exerts an antiviral action against HTLV-1 . First , while not easily detectable in vivo [17] , viral proteins expression is induced in T-lymphocytes isolated from infected patients when cultured ex vivo [18] , likely as a result of the relief from in vivo repression . Consistently , culture of HTLV-1-infected cells with IFN-β-expressing stromal cells represses viral protein expressions [18] . Second , exogenous IFN-I decreases viral protein translation in vitro , and protects lymphocytes from de novo infection [19] . Thus , IFN-I-mediated antiviral control of HTLV-1 infection is likely to occur in vivo . Nonetheless , the cell type that produces IFN-I during infection remains enigmatic . Plasmacytoid dendritic cells ( pDCs ) act as sentinels of viral infection , as they are the major IFN-I producers in vivo [20] , being 1000-fold more potent for IFN-I production as compared to other cell types [20] . They predominantly recognize viral nucleic acids , i . e . single-stranded RNA and non-methylated CpG-containing DNA , by TLR7 and TLR9 , respectively [21] . Cell-free HTLV-1 particles , when added at high concentration , were shown to induce IFN-I production by pDCs in vitro , in a TLR7-dependent manner [22] . Nonetheless , cell-free viruses are undetectable in the plasma of HTLV-1-infected individuals , which leaves open the question of the modality of pDC activation in vivo . Importantly , we and others recently revealed that cell contacts are required for efficient pDC activation by evolutionary divergent RNA viruses belonging to distinct families , such as Flaviviridae , Picornaviridae , Arenaviridae , Retroviridae , and Togoviridae [23–31] . Transfer of immunostimulatory viral RNAs from infected cells to pDCs was further shown to involve carriers in the form of non-infectious and/or non-canonical viral particles , including exosomes [25 , 27 , 29] and immature virus particles [24] . Interestingly , cell-cell transmission of viral material is reminiscent of HTLV cell-cell transmission [32] , which is the only efficient way to infect new target cells . HTLV-1 viral transmission occurs through the transfer of neo-synthesized HTLV-1 virions via a virological synapse formed at the cell contact [33] , and/or infectious viral particles embedded at the surface of infected cells within an extracellular matrix components ( ECM ) -rich structure [34] . The latter is referred to as the HTLV-1 biofilm-like structure [34] . This HTLV-1 biofilm-like structure has been further defined as the minimal infectious structure allowing viral transmission [35] . Importantly , the role of the cell surface associated virus within biofilm-like structure in the activation of the innate immune response is still unknown . Here , we demonstrate that the pDC-mediated IFN-I response requires physical contacts with HTLV-infected cells . Moreover , we show that HTLV-1 biofilm-like structure represents the minimal virally induced-structure able to trigger an IFN-I response by pDC , and thus recapitulating pDC activation induced by contact with infected cells . Further , comparison of a panel of HTLV1/2 infected cells reveals that pDC-mediated IFN-I response inversely correlates with the ability of the HTLV-infected cells to transmit infectivity and with their surface glycosylation pattern . Indeed , we show that the density of terminal β-galactoside glycosylation at the surface of infected cells regulates IFN-I production by pDC . Altogether , our results highlight an unforeseen function of virus-containing cell surface-associated structures in the activation of pDCs by cell contacts , as well as its fine-tuning by the glycosylation pattern at the surface of the sensed infected cells .
We first determined the production of type I IFN ( referred to as IFN-I ) by PBMCs and pDCs upon recognition of infected cells as compared to cell-free virions present in supernatant ( SN ) of infected cell lines ( Fig 1 ) . pDCs , representing 0 . 2–0 . 5% of total PBMCs , were isolated from healthy blood donors with >91% of purity ( Fig 1A , middle panel ) , consistently with our previous reports [24 , 25 , 30] . PBMCs or purified pDCs ( referred to as responders ) were co-cultured with HTLV-1 chronically infected cells , i . e . , C91-PL cell line [36] , ( referred to as inducer ) . These HTLV-1-infected cells induced a potent IFN-I response by both PBMCs and purified pDCs , when in physical contact ( Fig 1B ) . In sharp contrast , cell-free viruses present in the supernatant from HTLV-1-infected cells ( approximately 10–25 ng/mL of the HTLV-1 capsid p19gag , i . e . , representing the viral concentration reached in the supernatant of inducer cells at the time of coculture ) failed to induce very low , or undetectable levels of IFN-I production ( Fig 1B , around 5 U/mL when detected , or below the detection limit ) . Next , we tested the contribution of pDCs relative to other PBMC cell types in the IFN-I response to HTLV-1 infected cells . Depletion of pDCs from PBMCs ( Fig 1A , lower panel ) abrogated the response to HTLV-1 infected cells ( Fig 1B ) . We controlled that pDC-depleted PBMCs and PBMCs produced comparable levels of IL-6 after LPS stimulation , confirming that pDC depletion did not impair PBMC responsiveness ( Fig 1C ) . pDCs obtained from 27 donors , reproducibly demonstrated robust IFN-I responses to HTLV-1 infected cells ( Fig 1D; median value of 13 400 U/mL ) , albeit with some donor-to-donor variations . Of note , HTLV-1 infected cells alone did not produce IFN-I ( Fig 1D ) . Together , these results indicate that pDCs are the main , if not exclusive , IFN-I producers among PBMCs in response to the contact with HTLV-1-infected cells . Next , we tested whether pDC sensing of HTLV-1 infected cells involves TLR7 , a sensor of single-stranded RNA . Inhibition of TLR7 recognition using a competitive inhibitor significantly decreased the IFN-I response to HTLV-1-infected cells ( Fig 1E ) . The specificity of TLR7 inhibitor was validated by the inhibition of IFN-I production triggered by a TLR7 agonist but not by a TLR9 agonist , as expected ( Fig 1E ) . These results suggested that pDCs sense HTLV-1-infected cells via TLR7 , implying that HTLV-1 viral RNA is likely the immunostimulatory signal . Since IFN-I production by pDCs following incubation with cell-free viruses was not or barely detectable , we hypothesized that cell contacts are required for pDC activation . We thus measured IFN-I production when pDCs were physically separated from HTLV-1-infected cells by a 0 . 4μm permeable membrane ( Fig 1F , TW ) . The absence of physical contact between inducer and responder cells abrogated IFN-I production ( Fig 1F ) . We controlled that pDC responsiveness was maintained in this experimental setting , as pDCs produced similar amounts of IFN-I upon TLR7 agonist stimulation when cultured in transwell chambers or not ( Fig 1F ) . This demonstrated that pDC contact with infected cells is required to trigger IFN-I production . Exosomes have been involved in the transfer of immunostimulatory RNAs to pDCs for other viruses [25 , 27 , 29 , 37] and HTLV-1 infected cells are known to produce exosomes [38] . To test whether exosomes are involved in the transfer of the HTLV-1 immunostimulatory signals , we used the C8166 HTLV-1 cell line , which is impaired for expression of the structural proteins Gag and Env and thus do not produce infectious viral particles [39] , as confirmed by absence of infectivity transmission to Jurkat-LTR-Luc reporter cell line ( Fig 1G ) . While C8166 HTLV-1 cells retain the capacity to produce the Tax regulatory protein , and exosomes that contain several viral mRNAs [38] , they failed to induce IFN-I production by co-cultured pDCs ( Fig 1G ) . This inferred that the transmission of activating signal to pDCs likely requires Env gp46 and/or Gag mediated extracellular export of viral RNA , rather than exosomal export of viral RNAs . To address the importance of Env gp46 in pDC IFN-I response , we tested pDC activation upon co-culture with Jurkat cells transfected with the WT HTLV-1 molecular clone ( i . e . , pACH ) or with the counterpart molecular clone lacking the envelope glycoprotein ( i . e . , pACH-ΔEnv ) . As expected , Env gp46 was not expressed when Jurkat cells were transfected with the ΔEnv molecular clone , while p19gag levels were similar ( Figs 1 and S1A ) . Cells harboring WT but not ΔEnv molecular clone or only Tax expressing vector induced a robust IFN-I production by co-cultured pDCs ( Fig 1H ) . Next , we tested whether primary HTLV-1 infected cells from HAM/TSP patients were also able to induce IFN-I production by pDCs . As HTLV-1 infected cells isolated from the blood of patients do not express HTLV-1 [18] , PBMCs from 3 HAM/TSP patients were first cultured in presence of IL-2 and PHA to induce viral re-expression . This was controlled by p19gag detection ( S1B Fig ) . Viral re-expression was observed in all patient samples , with some donor-to-donor variation as expected ( S1B Fig ) . These cells were then co-cultured with pDCs . PBMCs from the 3 independent HAM/TSP patients significantly induced pDC IFN-I production ( Fig 1I ) , as opposed to the absence of response to PBMCs from healthy donors used as controls . We then aimed at determining whether pDCs are susceptible to HTLV-1 infection as previously reported [5] , in our experimental setup leading to IFN-I production ( i . e . , within 24h-incubation with HTLV-1 infected cells ) . The productive infection of pDCs at the end of co-culture with HTLV-1 infected cells was assessed by the detection of Tax , as we previously reported [32] . In contrast to monocytes-derived dendritic cells ( MDDCs ) , Tax expression by pDCs was not readily detected 24h after co-culture with HTLV-1 infected cells ( S2A Fig ) . Thus , this suggests that pDC IFN-I response to HTLV-1 infected cells does not involved a productive infection . Altogether , our results demonstrated that pDCs sense HTLV-1 infected cells by Env gp46-mediated transmission of pDC-activating signal by cell contact leading to robust IFN-I response via TLR7-induced signaling . The capture of HTLV-1 cell-free virus by target cells involved binding of Env gp46 to NRP-1/BDCA-4 in cooperation with HSPG [40] and then to Glut-1 [41] . The latter also serves as the receptor mediating fusion of HTLV envelope with the cellular membrane [42] . NRP-1/BDCA-4 , Glut-1 and HSPG are all readily expressed at the pDC surface ( Fig 2A ) . We thus sought to determine the contribution of these receptors in the transfer of the activating signal from the infected cells to the pDCs . Previous reports showed that the binding of HTLV-1 Env gp46 to its receptors is mediated by the receptor binding domain ( RBD; the first 215 amino acids of gp46 ) , and can thus be out-competed by recombinant RBD [41] . Competition with recombinant RBD significantly reduced IFN-I production by pDCs ( Fig 2B ) , viral binding to pDCs ( Fig 2C and S2B Fig ) and viral transmission to reporter cells ( Fig 2D ) . This suggests that pDC sensing requires HTLV-1 Env binding to its receptor ( s ) . RBD comprises residues that have been specifically involved in NRP-1/BDCA4 ( i . e . , at the position 90-to-94 ) [40] and the 94-to-101 stretch known to be pivotal for Glut-1 binding and subsequent viral fusion [43] . Thus , it does not allow to discriminate between binding to NRP-1/BDCA-4 versus Glut-1 . Nonetheless , binding of Env gp46 to NRP-1/BDCA4 can be prevented by addition of recombinant VEGF165 , a known ligand of NRP-1/BDCA4 , that interacts directly through a peptide stretch similar to the 90–94 sequence found in Env gp46 but also using an HSPG dependent manner [40] . Thus VEGF165 does not allow discriminating binding to NRP-1 versus HSPG . Competition with recombinant VEGF165 did not prevent the IFN-I production by pDCs ( Fig 2B ) , viral binding to pDCs ( Fig 2C ) nor cell-cell viral transmission to reporter cells ( Fig 2D ) , suggesting that NRP-1/BDCA-4 HSPG-mediated and/or direct binding may not be involved in HTLV-1 transfer by cell-cell contact . The effectiveness of VEGF165 competition was confirmed by the expected reduction of the binding of cell-free HTLV-1 virion to reporter cells measured by flow cytometry detection of p19gag ( Fig 2E ) , consistent with a previous report [40] , whereas no competition by VEGF165 was observed in co-culture experiments with HTLV-1 infected cells ( C91-PL ) ( Fig 2E and S2C Fig ) . VEGF165 and RBD treatment did not impair the pDC IFN-I response upon stimulation with a TLR7 agonist , thus ruling out non-specific effects of recombinant RBD and VEGF165 on pDC responsiveness ( Fig 2B ) . Altogether , these results suggested that both transmission of infection to target cells after cell-cell contact as well as HTLV-1 sensing by pDCs require Env gp46 interaction with at least Glut-1 . Previous reports showed that HTLV-1 virions are present at the cell surface embedded within carbohydrate-rich elements , referred to as a viral biofilm-like structure [34] , and involved in the infectivity transmission [34 , 35] . Since pDCs respond to HTLV-1-infected cells upon physical contact , we first determined whether HTLV-1 biofilm-like structure was present at the contact site between pDCs and HTLV-1-infected cells . Confocal microscopy analyses of pDC in contact with C91-PL cells revealed that HTLV-1 Env gp46 accumulates at the pDC/infected cell interface , together with carbohydrate-rich elements , known to be present in the viral biofilm-like structure [34] , as revealed here by WGA lectin staining ( Fig 3A ) . Of note , Env gp46 and WGA clusters were co-localized at the contact site for most of the analyzed pDC/infected cell contacts ( Fig 3B , approx . 85% ) , suggesting that cell contacts are preferentially oriented toward these specific biofilm-like structures , or inversely that the biofilm-like structures are preferentially positioned at the contact site . Next , we determined whether HTLV-1 biofilm-like structures could trigger pDC response . pDCs were cultured in the presence of HTLV-1 biofilm-like structures , isolated from infected cells as previously described [35] . Virus concentration in isolated biofilm-like structures ranged from 23 . 5 to 31 . 6 ng/mL of p19gag , a concentration similar to that found in the supernatant of infected cells . While cell-free virus-containing supernatants failed to induce IFN-I production by pDCs ( Fig 1B ) , isolated biofilm-like structures significantly activated pDC IFN-I response ( Fig 3C ) . This was specific to HTLV-1 infected cells , since similar isolation procedure from uninfected cells failed to activate the pDCs ( Fig 3C , cont . ) , ruling out a putative non-specific activation by the experimental process ( e . g . , cellular debris ) . As expected , isolated biofilm-like structures from HTLV-1-infected cells transmitted infectious virions to Jurkat target cells , ( Fig 3D ) . To further confirm that pDCs sense HTLV-1 biofilm-like structures , HTLV-1 biofilm-like structure was depleted using metalloprotease that digest the extracellular matrix [44] as previously described [34] . The metalloprotease treatment of HTLV-1-infected cells decreased the levels of surface envelope gp46 compared to untreated cells ( Fig 3E ) , in association with a reduction in both IFN-I production by co-cultured pDCs and viral transmission to Jurkat-LTR-Luc reporter target cells ( Fig 3F and 3G , and S3A and S3B Fig ) . This is shown for a representative experiment ( i . e . , pDCs from one blood donor co-cultured with biofilm-depleted HTLV-1 infected cells in Fig 3F and 3G ) and for the means of independent experiments using pDCs from 3 blood donors ( S3A and S3B Fig ) . Altogether , these results show that the HTLV-1 biofilm-like structure contains the immunostimulatory signal that triggers IFN-I production by pDCs . Since HTLV-1 embedded in biofilm-like structure but not cell-free virons induced pDC IFN-I response , we next sought to examine the elements present in the biofilm-like structure that contribute to viral transmission and subsequent pDC activation . HTLV-1 biofilm-like structures contain ECM components and linkers , including collagen , agrin and heparan sulfate proteoglycans ( HSPGs ) [34] . As HSPGs are involved in cell-cell and cell-ECM interactions [45] , we hypothesized that HSPGs present in HTLV-1 biofilm-like structure and/or in association with HSPGs at the pDC surface ( Fig 2A , right panel ) could favor cell-cell adhesion . To test this hypothesis , we used heparin , a polyanionic glycosaminoglycan that mimics the sulfate groups of HSPGs and that could thus act as a bridge to increase the pDC/infected cell contacts via the HTLV-1 biofilm-like structure . The impact of heparin on the frequency of cell conjugates formed between HTLV-1 infected cells and pDCs was analyzed by imaging flow cytometry ( Image Stream X technology ) ( S4A Fig ) , as we previously established [24] . This quantitative analysis revealed that heparin significantly increased the frequency of pDC-HTLV-1-infected cell conjugates ( Fig 4A ) . Consistently , heparin augmented pDC IFN-I production induced by HTLV-1 infected cells ( Fig 4B ) . Importantly , heparin increased as well pDC activation induced by isolated HTLV-1 biofilm-like structures ( Fig 4B ) . Similar results were obtained using blood samples from different donors ( S4B Fig ) . HSPGs are known to act as attachment factors via an interaction with Env gp46 [46] , thus heparin could compete for HTLV-1 binding and subsequent HTLV-1 infection . Nonetheless , similar heparin addition had no impact on the viral transmission to Jurkat reporter cells using either isolated HTLV-1 biofilm-like structures or HTLV-1-infected cells ( Fig 4C ) . This contrasts with a previously reported impact of heparin in the context of distinct experimental procedure that showed that heparin , when added during biofilm isolation , reduced the ability of the isolated biofilm to infect Jurkat reporter cells [34] . The presence of heparin during biofilm isolation might have loosened the biofilm structure allowing a better exposure of the viral envelop to heparin competition . The lack of heparin competition both on infected cells and isolated biofilm ( Fig 4C ) demonstrated that under our experimental conditions , heparin did not compete with HSPGs for HTLV capture when HTLV-1 is embedded in an intact viral biofilm . Altogether , these experiments suggested that heparin increases pDC/infected cell contact as well as pDC transfer of the immunostimulatory signal from the isolated HTLV-1 biofilm-like structures and , likely as a consequence the potentiation of pDC IFN-I response , albeit additional effect can contribute as well . Importantly , the absence of modulation by these heparin treatments of infectivity transmission to target cells highlighted that the transfer of the immunostimulatory signal to pDCs features distinct regulatory mechanism ( s ) as compared to the infectivity transmission to other cell types . Next , we sought to define the viral determinants and other cellular component ( s ) modulating the level of pDC response to infected cells , including the amounts of viral RNAs , cell contact efficiency and ability to transmit viral infectivity . To address these questions , we compared three HTLV-1 chronically infected cell lines ( C91-PL , MT-2 and Hut102 ) , known to produce different amounts of viral proteins [47] and two HTLV-2-infected cell lines ( MO and C19 ) . All HTLV-cell lines triggered IFN-I production by co-cultured pDCs , albeit at different levels ranking from lower to higher inducer cell lines , at the optimal pDC/infected cells ratios ( S5D Fig ) , as follows: C19 , MO , C91 , Hut102 and MT-2 cell lines ( Fig 5A ) . Neither differences in the amount of intracellular genomic RNA nor viral RNA released in the supernatant of infected cells was correlated to the observed differences in the induction of pDC IFN-I response ( Fig 5B and 5C and S5A and S5B Fig ) . This suggests that the amount of RNA produced by the infected cells is not rate limiting for activation of pDCs . Thus , we next assessed whether frequency of pDCs engaged in contacts with the different HTLV-infected cell lines regulated the intensity of pDC activation . The frequency of pDC conjugates with the different HTLV-infected cell lines was similar , except higher level for the HTLV-infected MO cell line ( Fig 5D ) . Nonetheless , this higher frequency of cell conjugates with the HTLV-infected MO cell line did not translate into higher levels of IFN-I production ( Figs 5D and S5C ) , suggesting that additional factor ( s ) , other than pDC ability to establish contact with HTLV-infected cells , govern ( s ) the levels of pDC IFN-I response to HTLV-infected cells . We thus tested whether variations of pDC induction by HTLV-infected cell lines might be explained by distinct mechanisms for viral capture by pDCs . To address this , we first evaluated viral binding and internalization in the pDCs upon co-culture with the different HTLV-1/2 cells lines by detection of intracellular p19gag in the CD123+ pDCs population by FACS ( Fig 5E ) . Except pDCs co-cultured with C91-PL , we observed no differences in virus binding on pDCs . This suggests that the reduced IFN response induced by C19 cells does not result from diminished HTLV-2 capture by pDCs . Furthermore , consistent with results obtained using cell-free virus [22] , HTLV-1 infected cells induced TRAIL expression by co-cultured pDCs ( Fig 5F ) , as did HTLV-2 infected cells ( Fig 5F ) . This suggests that the reduced pDC IFN response to C19 cells is not associated with other impairment link in their ability to respond to virus . Furthermore , using C19 cells that induced the lowest pDC IFN-I production ( Fig 5A ) , we showed that both the pDC response and viral transmission were significantly out-competed by recombinant RBD , but not by recombinant VEGF165 ( S6A–S6D Fig ) , suggesting that different HTLV viral receptor usages are not likely responsible for the difference in pDC IFN-I response to the various cell lines . Additionally , pDC sensing of C19 cells was specifically inhibited by TLR7 inhibitor ( S6E Fig ) . This rules out the involvement of other PRR that would induced lower IFN-I induction upon HTLV-2 sensing , as suggested for other viruses [28] . As we showed that pDCs are activated by cell-cell contacts with infected cells and via viral biofilm-like structures , we then compared the viral accumulation at the surface of the panel of HTLV-infected cell lines ( S6F Fig ) . The p19gag proteins were detected as patch/cluster at the surface of all infected cells , suggesting that the pDC activation is not directly linked to an absence of virus accumulation at the surface of the different HTLV-infected cell lines . We next asked whether the level of pDC activation by the HTLV-infected cell lines correlate with their ability to transfer infectious virions to target cells . Regressive exponential correlation analysis revealed that IFN-I production by pDCs was inversely correlated with the ability of infected cells to transfer infectious virions ( Fig 5G and 5H , p-value = 0 , 011 ) . Altogether these results indicate that the sensing of HTLV-infected cells by pDCs is not strain-specific , and , importantly , inversely correlated to infectivity transmission to alternative target cells . It thus implies that pDC activation is likely modulated by other features of the infected cells . Our results using heparin suggested that glycosylated proteins , including HSPGs are involved in the tethering of HTLV-1 stimulating signals to the pDC surface and/or its transfer , resulting in increased pDC IFN-I production . The density of surface glycosylation , including HSPGs [48] , is known to be cell type specific [49] . We thus quantified cell surface glycosylation using various lectins known to bind different terminal glycosylation patterns ( S7A Fig ) . The staining by Peanut agglutinin lectin ( PNA ) , which bind to oligosaccharide structures with terminal β-galactose residues on the different HTLV-infected inducer cells ( S7B Fig ) revealed that the amount of this type of surface glycosylation was inversely correlated to the magnitude of IFN-I production by co-cultured pDCs ( Fig 6A and 6B ) . As opposed , the levels of PNA lectin staining at the surface of infected cells positively correlated with their ability to transmit viral infection to target cells ( Fig 6C ) . Consistent observations were obtained by confocal microscopy analysis of PNA lectin , displaying very weak PNA staining for MT2 and Hut102 , the highest IFN-inducer cell lines ( S7C Fig ) . Similar trend was observed for stained SBA-lectin ( S7B Fig ) , thought to detect α- or β-linked N-acetylgalactosamine residues , albeit to with lower magnitudes of difference ( S7B Fig ) , and without statistical correlation with IFN-I production ( S7D Fig ) . In contrast , binding of the lectins UEIA , WGA and ConA , that recognized other glycosylated residues , did not demonstrate difference between the different cell lines ( S7B Fig ) . Altogether , these results suggest that the composition of the terminal oligosaccharide residues , especially dense terminal β-galactose glycosylation , present at the surface of HTLV-1-infected cell lines might inversely govern both IFN-I response by co-cultured pDCs and viral transmission . To further study the role of terminal β-galactose residues in IFN-I induction , we performed assay to mask such surface glycosylation by pretreating C19 cells with PNA-lectin prior contact with pDC . We controlled that the PNA concentrations used in the co-culture did not affect cell viability ( S8A Fig ) . The presence of PNA-lectin at the surface of C19 cells significantly increased pDC-induced IFN-I production ( Fig 6D ) . Conversely , the removal of sialic acid using neuraminidase treatment resulted in augmented exposure of terminal β-galactose at the surface of treated C91 cells ( i . e . , at levels similar to that of C19 cells , Fig 6E ) , and in significant decreased of pDC-induced IFN-I production ( Fig 6D ) . Of note , the limited impact of neuraminidase treatment on pDC IFN-I production likely result from its short timeframe of impact on the exposure of β-galactoside residues , as revealed by the reduction overtime of PNA staining of neuraminidase-treated cells ( S8B Fig ) . Thus , to strengthen the role of β-galactoside residues in the regulation of pDC IFN-I production , we determined whether viral expression in PBMCs from HAM/TSP patients was also associated with a modulation of PNA staining . In vitro culture of primary PBMCs from both healthy donors and HAM/TSP patients is enough to expose β-galactoside residues ( S8C and S8D Fig ) . Consequently , β-galactoside was not preferentially/exclusively induced in PBMCs that expressed HTLV-1 ( S8E Fig ) . However , we observed a higher proportion of infected PBMCs that expressed β-galactoside residues in the PBMCs from patient #1620 compared to the two others ( S8E Fig ) . Interestingly , although the level of virus re-expression in PBMCs from patient #1620 was the highest ( i . e . 70% of positive PBMCs compare to 40% for #1668 and 5% for #1485 , see S1 Fig ) , this was not associated with a higher induction of type-I IFN by co-cultured with pDCs ( Fig 1I , compare patients #1620 to #1668 ) . Altogether , our results show that β-galactoside glycosylations at the surface of infected cells likely negatively regulate pDC activation by HTLV-infected cells .
Evidences suggested that IFN-I response is likely pivotal to repress HTLV-1 replication [18 , 19 , 50 , 51] . Yet , the persistent HTLV infection is thought to result from escape viral mechanisms and consequent failure of the immune detection and clearance of infection [52] . Along this line , HTLV-1 inhibits IFN-I induction and signaling [12–15] [16] , leading to very limited production of IFN-I by infected cells . In this study , we elucidated an alternative sensing pathway mediated by the recognition of infected cells by pDCs , a sentinel cell type known to be a potent producer of IFN-I . We demonstrated that pDCs preferentially sense HTLV-1 infected cells via physical contact rather than HTLV-1 cell-free virions . This sensing pathway is thus congruent with the absence of , or very low , detection of cell-free virus in the blood of infected patients [53] , a consequence of active repression of viral expression [54] . Recent report suggested that viral latency observed in vivo might be transiently relieved under changes in nutrients availability in the extra-cellular environment of the infected cells [55] , thus potentially leading to sporadic viral expression in privileged sites such as lymphoid organs . Thus localized IFN-I response by pDCs upon contact with transiently reactivated infected cells would be in agreement with the absence of detection of IFN-I at the systemic level in infected carriers [56] . HTLV-1 viral transmission through physical contact likely compensates for the low levels of cell-free viruses found in the patients and/or their poor infectivity [35 , 53 , 57] . Upon cell contact , the virus is transmitted either through a virological synapse , in which virus assembly and budding are polarized toward the cell contact [58] , or through the transfer of viral biofilm-like structures , an extracellular accumulation of viruses embedded in the infected-cell extracellular matrix ( ECM ) [34] , both mechanisms being likely not mutually exclusive . Here , we showed that isolated viral biofilm-like structure is sufficient to trigger IFN-I production by pDCs . We further propose that the increased potential of biofilm-like structures for pDC activation , as compared to cell-free virus , could be due to components of these structures favoring the transmission of the pDC-activating signal , possibly by tethering the immunostimulatory RNA carrier to the pDC surface . Although pDCs are largely refractory to most viral infection [23] , Jones et al . [5] observed viral production by pDCs exposed to cell-free HTLV-1 for at least 3 days . This timeframe is longer than our experimental setting ( i . e . , 24 hours , in accordance with previously reported pDC half-life limited to couple of days [59 , 60] ) , and during which we failed to detect productive HTLV-1 infection of pDCs . Thus , productive infection of pDCs by HTLV-1 is very unlikely needed for their rapid IFN-I production induced by contact with infected cells . Our results suggested that pDC sensing of HTLV-infected cells is mediated via the HTLV entry receptor Glut-1 , and NRP-1/BDCA-4 receptor seems dispensable . Previous report showed that infection of myeloid DC by cell-free HTLV-1 particles is independent of NRP-1/BDCA4 , viral binding being ensured by the DC-SIGN lectin [61] . It is conceivable that pDC-expressed glycosylated surface factors , including lectins and HSPGs known as capture molecules for HTLV and other viruses , could act as cofactor for HTLV capture at the pDC surface via glycan-mediated interactions with HTLV before its delivery to Glut-1 . Alternatively , virus delivery through cell-cell contact could bypass the need for attachment factors and virus concentration at the surface of target cells before virus interaction with its cognate receptor . Our results showed that heparin , an HSPG mimic , increased the frequency of pDC contacts with HTLV-1 infected cells and IFN-I production by pDCs in response to HTLV-1-infected cells , implying a putative function of HSPG in the stabilization of the pDC-infected cells interface required for the capture of the immunostimulatory RNA carrier present in the HTLV-1 biofilm-like structures by pDCs . Our results uncovered differences of surface glycan pattern of HLTV-infected cells , including the composition of the terminal oligosaccharide residues at the cell surface ( i . e . , plasma membrane ) , with densification of certain residues being inversely correlated to the level of pDC IFN-I response . Further , enzymatic and pharmacological inhibition/modulation of the cell surface glycans impacted pDC response to infected cells . These observations support the proposition that the extracellular matrix and/or glycosylated proteins expressed at the plasma membrane of the infected cells likely govern both IFN-I production by co-cultured pDCs and viral transmission . Importantly , our results suggest that these two processes are inversely correlated . By analogy to previously reported abilities of several viruses to modify ECM composition of the host cells to favor their own dissemination and/or immune escape [62–64] , one might speculate that the chronic infection by HTLV modulates the cell surface glycan pattern [65] . We showed that terminal β-galactoside glycosylation density is inversely correlated with the ability of infected cells to promote contact-dependent pDC IFN-I production . Together with the absence of correlation between the levels of pDC IFN production , the amount of viral RNA production or capture of HTLV by pDCs , shown for different HTLV-infected cells , our results suggest that composition of the extracellular matrix and/or cell surface expression of glycosylated proteins that embed cell surface-attached viruses of the different HTLV-infected cells might regulate pDC activation . The results obtained using PBMCs from HAM/TSP patients suggest that β-galactoside residues induction at the surface of infected cells might not however be regulated by viral expression . In addition , we could not determine whether these residues are specific components of the viral biofilm , or whether β-galactoside-containing proteins surrounding the viral biofilm at the plasma membrane of the infected cells are enough to regulate IFN-I production . Previous reports showed that pDC response to viral infections can be modulated by different cell surface factors including the regulatory receptors ILT7 , BDCA2 or DCIR [66 , 67] . For example , ILT7 binds BST-2 , an IFN-induced gene , initially described as an HIV restriction factor that impedes viral release from infected cells [68] . Since HIV Vpu protein counteracts viral tethering by BST2 , virus-free surface BST-2 can readily interact with ILT7 , and thereby inhibits pDC IFN-I production [69] . While BST-2 is also expressed upon HTLV-1 infection , as opposed to the negative regulation of HIV release , it participates in efficient HTLV-1 cell-cell transmission [70] . Other negative regulatory receptors e . g . , BDCA2 and DCIR bind complex-type sugars chains ( terminal β-galactoside containing complex sugars [71 , 72] , and mannose/fucose containing complex sugars [73 , 74] , respectively ) . Since pDC IFN-I response to HTLV is modulated by the available terminal sugar composition at the surface of infected cells , one might speculate that the density of specific glycans , via the interaction with negative regulators , e . g . , BDCA2 might regulate the levels of pDC IFN-I production induced by HTLV-infected cells as previously shown in other contexts [75–77] . Interestingly , these dense specific glycans might not be part of the biofilm , but still might be engaged after contact with the pDC . Thus , along with the interaction of the viral envelop protein with the Glut-1 receptor , other cell surface factors , including heparan sulfate-containing proteins , and terminal galactoside-conjugated proteoglycans , at the pDC/infected-cell interface , could regulate the strength of pDC activation . Altogether our results provided an original illustration of the regulation of pDC IFN-I response by the surface glycan pattern of infected cells .
Jurkat cells ( from ATCC , ref ACC 282 ) stably transfected with a plasmid encoding the luciferase ( Luc ) gene under the control of the HTLV-1 long terminal repeat ( LTR ) promoter and the HTLV-1 Tax-transactivator ( Jurkat-LTR-Luc ) [34] were maintained under hygromycin selection ( 450 μg/mL , Sigma ) in culture RPMI medium: RPMI 1640 medium supplemented with 10% fetal calf serum ( FCS; Gibco Life Technologies ) , L-Glutamine ( 2 mM , Gibco Life technologies ) and penicillin-streptomycin ( 100 U/mL and 100 μg/mL respectively; Gibco Life Technologies ) . C91-PL ( HTLV-1 infected T-cell line , Cellosaurus , ref CVCL_0197 ) , MT-2 ( HTLV-1 infected T-cell line , NIH , ref 237 and [78] ) , Hut102 ( HTLV-1 infected T-cell line , Cellosaurus , ref CVCL_3526 and [79] ) and C8166 ( HTLV-1 infected T-cell line which does not produce infectious virus [39] , ECACC ref 88051601 ) were maintained in culture RPMI medium . PBMCs from healthy blood donors or from HAM/TSP patients were cultured 18h in RPMI medium supplemented with 20% FCS supplemented with IL-2 ( 150 U/mL ) and PHA ( 1μg/mL ) . C19 ( HTLV-2 infected cell line , obtained from [80] ) and MO ( HTLV-2 infected cell line , ATCC ref CRL-8066 ) were maintained in culture RPMI medium supplemented with 20% FCS . The human fibrosarcoma cell line containing a plasmid encoding the luciferase gene under the control of the immediate early IFN-I inducible 6–16 promoter ( HL116 ) ( a kind gift from S . Pelligrini , Institut Pasteur , France ) [81] was maintained under HAT selection in DMEM medium supplemented with 10% FCS and penicillin-streptomycin ( 100 U/mL and 100 μg/mL respectively ) . All cells were grown at 37°C in 5% CO2 . Ficoll-Hypaque ( GE Healthcare Life Sciences ) ; LPS , TLR7 agonist ( R848 ) and TLR9 agonist ( ODN2216 ) ( Invivogen ) ; TLR7 antagonist , IRS661 ( 5’-TGCTTGCAAGCTTGCAAGCA-3’ ) synthesized on a phosphorothionate backbone ( MWG Biotech ) ; Fc Blocking solution ( MACS Miltenyi Biotec ) ; BDCA-4-magnetic beads for selective isolation of pDCs ( MACS Miltenyi Biotec ) ; IL-6 ELISA kit ( Affymetrix , eBioscience ) ; Lipofectamine 2000 ( Life Technologies ) ; 96-well format transwell chambers ( Corning ) ; LabTek II Chamber Slide System , 96-Well Optical-Bottom Plates and Nunc UpCell 96F Microwell Plate ( Thermo Fisher Scientific ) ; Vibrant cell-labeling solution ( CM-DiI , Life Technologies ) ; rat anti-HSPG antibody ( clone A7L6 , Upstate Biotechnology ) Hoescht and anti-mouse AlexaFluor 647-conjugated secondary antibody ( Life Technologies ) ; anti-mouse DyLight 488-conjugated secondary antibody ( Vector ) ; anti-rat APC-conjugated secondary antibody ( SouthernBiotech ) ; High Capacity cDNA reverse transcription kit ( Applied Biosystems ) ; Powerup Sybr Green Master Mix ( Applied Biosystems ) ; pDC specific markers: mouse PE or APC-conjugated anti-CD123 ( clone AC145 , Miltenyi ) , mouse APC-conjugated anti-BDCA-2 ( AC144; Miltenyi ) ; mouse PE-conjugated anti-TRAIL ( ThermoFisher ) ; Metalloproteinase 9 ( Enzo Life Sciences ) ; FITC-conjugated Peanut Agglutinin ( PNA ) ( Sigma Aldrich ) ; Alexa Fluor 680-conjugated Wheat Germ Agglutinin ( WGA ) ( Thermo Fisher ) ; FITC-conjugated WGA , Soy bean Agglutinin ( SBA ) , Ulex europaeus agglutinin I ( UEA-I ) and Concanavalin A ( ConA ) were from Vectors; poly-L-lysine ( Sigma , P4832 ) , anti-HTLV-1 p19gag antibody ( 1:1000 , Zeptometrix ) ; recombinant VEGF165 protein ( R&D ) ; Glut-1 . RBD . GFP ( Metafora biosystem ) ; recombinant IFN-α 2a ( TEBU BIO PBL ) ; anti-HTLV-1 Env gp46 antibody ( 1:1000 , Zeptometrix ) , luciferase reporter activity assay ( Promega ) ; paraformaldehyde 20% ( PFA; Electron Microscopy Sciences ) ; saponin ( Sigma ) ; Heparin ( Sigma ) . HTLV-1 molecular clone ( pACH ) and HTLV-1 molecular clone lacking the expression of the envelope protein ( pACH ΔEnv , [82] ) were provided by Dr . Pique ( Institut Cochin , France ) . The Alexa Fluor 488-conjugated anti-Tax antibody ( LT-4 ) was provided by Pr Tanaka ( University of Ryukyus , Japan ) . The pDCs and PBMCs were isolated from blood or cytapheresis units from healthy adult human volunteers which was obtained according to procedures approved by the “Etablissement Français du sang” ( EFS ) Committee . All donors provided informed consent to EFS . PBMCs from HAM/TSP patients were obtained in the context of a Biomedical Research Program approved by the Committee for the Protection of Persons , Ile-de-France II , Paris ( 2012-10-04 SC ) . All individuals gave informed consent . PBMCs were isolated using Ficoll-Hypaque density centrifugation . pDCs were positively selected from PBMCs using BDCA-4-magnetic beads ( MACS Miltenyi Biotec ) and pDCs were depleted from PBMCs , as previously described [24 , 25] . The typical yields of PBMCs and pDCs were 800x106 and 2x106 cells , respectively , with a typical purity of >91% pDCs , as we previously reported [24 , 25] . After isolation , pDCs ( 2x104 ) were platted in 96-well round bottom plates and cultured at 37°C with HTLV-1 or HTLV-2 infected cell lines ( 2x104 or other count when indicated ) , or with PBMCs from healthy donors or from HAM/TSP patients ( 2x104 ) , or with Jurkat cells microporated with the HTLV-1 molecular clone pACH or pACH ΔEnv ( 2x104 ) , or with Jurkat cells ( 2x104 ) as negative control , or with isolated HTLV-1 biofilm-like structures ( 100μL ) , or with HTLV-1 biofilm-like structures depleted cells ( 2x104 ) . When indicated , HTLV-2 infected cells ( C19 , 106 cells in RPMI culture medium ) or HTLV-1 infected cells ( C91 , 106 in PBS ) were treated with PNA ( 10 μg/ml , SIGMA ) for 30 min at 4°C or neuraminidase ( 0 , 1U/ml , SIGMA ) for 1h at 37°C respectively . Treated cells were then washed twice in RPMI culture medium prior to co-culture ( 2×104 ) with pDC ( 2×104 ) . Culture with isolated pDCs or PBMCs were maintained in RPMI 1640 medium ( Life Technologies ) supplemented with 10% FCS , 10 mM HEPES , 100 units/mL penicillin , 100 μg/mL streptomycin , 2 mM L-glutamine , non-essential amino acids and 1 mM sodium pyruvate at 37°C/5% CO2 . The supernatants were collected at 20-24h after co-culture . When indicated , infected cells or uninfected cells were co-cultured with pDCs in 96-well format transwell chambers separated by a 0 . 4 mm membrane ( Corning ) , as previously [24 , 25] . HL116 cells were seeded at 2 . 104 cells/well in 96-well plate 24 h prior the assay , and incubated for 17 h with supernatant collected from pDC co-cultures ( 100 μL ) or serial dilution of recombinant human IFN-α 2a ( PBL Interferon Source ) , used for standard curve titration . Cells were then lysed and luciferase activity assayed . IFN-I levels were expressed as equivalent of IFN-α 2a concentration , in Unit/mL . The detection of IL-6 by ELISA was performed as previously[24] using kit ( Affymetrix , eBioscience ) and according to the manufacturer instructions . Jurkat cells ( 8x104 cells ) were transfected with 3 μg of pACH or pACH ΔEnv together with 1 μg of pSG5M-Tax1 [83] using the Neon Transfection System ( ThermoFischer Scientific ) following manufacter’s instructions . Cells were cultured 48h at 37°C before co-culture with pDCs . HTLV-1 viral biofilm-like structure was prepared with a method that is slightly different from the original one [34] and as previously described [35] . Briefly C91-PL cells were platted ( 3x105 cells/mL ) and cultured for 4 days . HTLV-1–infected cells were washed twice in RPMI-1640 serum-free medium and incubated at 1x106 cells/ml for 1 h at 37°C , with gentle agitation every 10 minutes . Then , FCS ( 10% final ) and penicillin-streptomycin ( 100 μg/mL final ) were added , and cells centrifuged . Supernatant containing biofilm-like structures preparation was collected and supplemented with Hepes ( 10 mM ) , non-essential amino acid ( 2 . 5 mM ) , sodium pyruvate ( 1 mM ) , β-mercaptoethanol ( 0 . 05 mM ) before immediate use . Cell-free viruses were also obtained from C91-PL cells ( 106 cells/mL ) cultured for 24h at 37°C 5% CO2 . Supernatant were clarified by centrifugation ( 5 minutes at 800g ) and filtrated through a 0 . 45 μm-diameter pore filter ( Millipore , MA ) to eliminate cell debris . Virions were purified by ultracentrifugation through a 20% ( wt/vol ) sucrose cushion at 100 , 000g in SW32 ( Beckman ) for 1h30 at 4°C and stored at -80°C before use . Virus concentration was determined using Retrotek HTLV-1/2 p19gag Antigen ELISA kit ( Zeptometrix ) following manufacturer’s instructions and as previously described [19] . C91-PL cells ( 106 cells /mL ) were treated with Metalloproteinase 9 ( 20 nM ) in RPMI serum-free medium for 1h at 37°C 5% CO2 . Cells were washed twice with culture RPMI medium , and immediately used . The efficacy of HTLV-1 viral biofilm-like structures shedding was controlled by analyzing gp46 viral envelope level by FACS and viral transmission to T-cells using Jurkat LTR-Luc reporter cells . HTLV-1 ( C91-PL , Hut102 , MT-2 ) , HTLV-2 ( C19 , MO ) or uninfected ( Jurkat ) cell lines ( 103 , 104 or 105 ) were co-cultured with Jurkat LTR-Luc cells ( 104 ) . Different ratio of infected cells/target cells ( 1/10; 1/1 or 10/1 ) were incubated for 24 hours in round-bottom 96-wells plates at 37°C . Cells were washed once with PBS and stored at -80°C as dry pellets until assayed for luciferase reporter activity using manufacturer’s instructions ( Promega ) . Luciferase results were normalized according to the amount of proteins determined by Bradford ( Biorad ) . Jurkat LTR-Luc cells ( 2x105 ) were incubated in culture RPMI medium with VEGF165 ( 80–100 ng/mL ) or 10μL Glut-1 . RBD . GFP at 4°C during 30 minutes before co-culture with C91-PL cells ( 2x104 ) or with cell-free viruses ( 50 ng/mL of p19gag equivalent as measured by ELISA ) for 2 hours at 37°C . Cells were then harvested , washed with PBS , fixed in 4% PFA , permeabilized in PBS / 1% BSA / 0 . 05% Saponin and stained with an anti-p19gag antibody ( 1:1000 ) followed by FITC or Alexa Fluor 549-conjugated anti-mouse antibody . Fluorescence was acquired on at least 10 000 events with a FACSCanto II cytometer ( BD Biosciences ) and data analyzed on FlowJo software ( Tree Star , Inc . Ashland , OR ) . HTLV-1 ( C91-PL , Hut102 , MT-2 ) , HTLV-2 ( C19 , MO ) or uninfected ( Jurkat ) cell lines ( 2x105 ) were fixed with 4% PFA and stained with FITC-conjugated lectins ( 10 μg/ml ) for 30 minutes at 4°C . The level of Env gp46 surface expression was determined on unfixed C91-PL cells or on Jurkat transfected cells using anti-HTLV-1 Env gp46 antibody ( 1:1000 in PBS-1% BSA ) for 1h at 4°C followed by Alexa488-coupled anti-mouse antibody for 30 min at 4°C . pDC were surface-stained with mouse PE or APC-conjugated anti-CD123 and mouse APC-conjugated anti-BDCA-2 , with Glu1 . RBD . GFP protein ( 5μl/ 1x105 cells , Metafora ) , or with anti-NRP-1 ( clone 12C2 , Biolegend ) . Alternatively , pDCs were fixed in 4% PFA and stained with anti HSPG antibody ( 1:100 ) for 30 min at 4°C followed by APC-conjugated anti rat antibody ( 1:100 ) for 30 min at 4°C . Cells were then washed with PBS and fluorescence acquired using 20 000 events on a FACSCanto II cytometer , and analyzed with FlowJo software ( Tree Star , Inc . Ashland , OR ) . Transfected Jurkat cells or PBMCs from HAM/TSP patients were cultured 18h in presence or not of IL-2 and PHA and fixed with 4% PFA , permeabilized in PBS / 1% BSA / 0 . 05% Saponin , and stained with anti-p19gag antibody ( 1:1000 ) for 30 min at 4°C followed by DyLight488-conjugated anti-mouse antibody ( 1:1000 ) . Cells were then washed and fluorescence acquired using at least 10 000 events on a FACSCanto II cytometer ( BD Biosciences ) , and analyzed with FlowJo software . After 24h co-culture with HTLV-1/2 infected cells , pDCs were collected , washed and stained with PE-conjugated anti-TRAIL and APC-conjugated anti-BDCA-2 antibodies . Cells were then washed and fixed in 4% PFA . Fluorescence was acquired using at least 10 000 events with a FACSCanto II cytometer ( BD Biosciences ) and analyzed with FlowJo . pDCs ( 105 ) were co-cultured with C91-PL ( HTLV-1 infected cells , 105 ) in the presence or not of Glu1 . RBD . GFP ( 10μl ) for 4h . Cells were then washed in PBS . For subsequent viral binding analyses , cells were surface-stained with anti-gp46 antibody ( 1:1000 ) followed by Alexa Fluor 647-conjugated anti-mouse antibody ( 1:1000 ) . For viral capture analyses , cells were fixed in 4% PFA , permeabilized in PBS / 1% BSA / 0 . 05% Saponin , and stained with anti-p19gag antibody ( 1:1000 ) followed by Alexa Fluor 647-conjugated anti-mouse antibody ( 1:1000 ) . After washing , pDCs were surface-stained with anti-CD123-Vioblue-conjugated antibody and fixed in 4% PFA . pDCs ( 105 ) or MDDCs ( 2 . 5x105 ) were co-culture with C91-PL ( 105 ) for 24h or 72h respectively . For pDCs infection analysis , cells were washed , surface-stained with Vioblue-conjugated anti-CD123 antibody , fixed and permeabilized according to the manufacturer’s instructions ( eBiosciences ) . For MDDCs infection analysis , cells were washed in PBS and in normal goat serum ( 7% , Sigma ) , fixed and permeabilized according to the manufacturer’s instructions ( eBiosciences ) . pDCs or MDDCs were stained with biotin-coupled anti-Tax antibody ( LT4 ) followed by streptavidin labeled with PE-Cy7 ( BioLegend , Ozyme ) . After extensive washes , MDDCs were finally surface-stained with a V450-coupled anti-CD11c antibody . Fluorescence was acquired using at least 10 000 events with a FACSCanto II cytometer and data analyzed on FlowJo software . RNAs were isolated from samples harvested in guanidinium thiocyanate citrate buffer ( GTC; Sigma-Aldrich ) by phenol/chloroform extraction procedure as previously described [25] . Reverse transcription was performed using the random hexamer-primed High Capacity cDNA reverse transcription kit ( Applied Biosystems ) and quantitative PCR was carried out using the Powerup SYBR Green Master Mix ( Applied Biosystems ) . The absolute numbers of HTLV-1 transcripts were normalized to the total amount of RNA . For supernatant samples , qRT-PCR was controlled by the addition of exogenous carrier RNAs encoding xef1α ( xenopus transcription factor 1α ) in supernatant diluted in GTC buffer , as previously described [24 , 25] . For quantification of viral genomic RNA the following primers were used: HTLV-1 Forward ( AAAGCGTGGAGACAGTTCAGG ) , HTLV-1 Reverse ( CAAAGGCCCGGTCTCGAC ) , HTLV-2 Forward ( CCTTGGGGATCCATCCTCTC ) , HTLV-2 Reverse ( TCTCTAAAGACCCTCGGGGAG ) . For quantification of viral RNA present in the supernatant of infected cells , the following primers were used: Tax 1 Forward ( GGATACCCAGTCTACGTGTTTGG ) , Tax 2 Forward ( GGATACCCCGTCTACGTGTTTGG ) , Tax 1/2 Reverse ( GGGGTAAGGACCTTGAGGGT ) . HTLV-1 ( C91-PL , Hut102 , MT-2 ) , HTLV-2 ( C19 , MO ) cell lines and control uninfected Jurkat cells ( 5x105 ) were transduced in 24 well plate with lentiviral-based vector pseudotyped with VSV glycoprotein to stably express GFP . Briefly , 105 GFP-expressing HTLV infected cells and control cells were co-cultured with 4x104 pDCs in low-adherence micro-plate designed for cell harvesting by temperature reduction ( Nunc UpCell 96F Microwell Plate from Thermo Scientific ) for 4–5 hours at 37°C , and , as indicated , in presence or not of heparin . The co-cultured cells were detached by 20 minute-incubation at room temperature , harvested and fixed in 4% PFA . Cells were then washed twice with staining buffer ( PBS 2% FBS ) , and pDCs were stained by the pDC-specific CD123 marker . Co-cultured cells were analyzed by Image Stream X technology ( Amnis ) at magnification x40 using IDEAS software , as previously described [24 , 25] . The cell population defined as pDC/HTLV-infected or uninfected cell conjugates comprises conjugates of at least one CD123 APC positive cell and at least one GFP positive cell among the total of CD123 APC-positive cells , GFP-positive cells and conjugates . As shown by representative images ( S4A Fig ) , the population gated as pDC-HTLV-1-infected cell conjugates corresponded to GFP positive/CD123 positive cell conjugates , as expected , with 80–95% purity . The cell populations were sorted using masks ( IDEAS software ) to eliminate cells out of focus and/or with saturating fluorescent signal , and then selected based on cell size of the positive cells ( i . e . , fluorescent signal area ) . For lectin/HTLV-1/2 virus localization analysis , HTLV-1 ( C91-PL , Hut102 , MT-2 ) , HTLV-2 ( C19 , MO ) or uninfected ( Jurkat ) cell lines cultured on Lab-tek chamber slides ( Nunc ) previously treated with 0 . 01% poly-L-lysine ( Sigma , P4832 ) were surface-stained with FITC-conjugated PNA or WGA ( 10μg/ml ) , fixed in 4% PFA , then permeabilized and stained with antibodies against HTLV-1/2 matrix protein p19gag ( 1:1000 ) followed by Alexa Fluor549-conjugated anti mouse antibodies . Cells were counterstained with DAPI-Fluoromount G before analysis on Zeiss LSM 800 microscope . Images were acquired on ImageJ . For pDC/HTLV-infected cells conjugates analysis , 4x104 pDCs were stained using 0 . 5 μM Vibrant cell-labeling solution as previously[24] . Labeled pDCs were washed twice with PBS and then co-cultured with 2x104 HTLV infected cells for 4–5 hours at 37°C , in a 96-well optical-bottom plate pre-coated with 8 μg/mL poly-L-lysin for 30 minutes at 37°C . Cells were then fixed in 4% PFA , washed with PBS and PBS-3%BSA , and stained with anti-HTLV-1 Env gp46 antibody ( 1:1000 in PBS—3%BSA ) for one hour at room temperature . Prior antibody staining , cells were stained by WGA lectin coupled to Alexa 680 ( Molecular Probes , ref W32465 ) diluted at 10μg/mL in HBSS , for 10 minutes at room temperature , then washed three times with HBSS . After three washes with PBS , cells were incubated with Alexa488-conjugated-anti-mouse antibody in 3% BSA-PBS and added to the cells along with Hoechst diluted at 1:500 ( Molecular Probes ) for 1 hour at room temperature . After three washes with PBS , cells were observed with a Zeiss LSM 710 laser scanning confocal microscope . The quantification of the phenotypes defined to as clusters at the contact were performed using Image J software package . Statistical analysis was performed using PRISM v7 . 03 software ( Graphpad ) . One-way analysis of variance ( ANOVA ) with Sidak’s multiple comparison test was used to determine statistically significant differences . Paired two-tail t-test was used to compare two groups from the same donor . Differences were considered significant if the p-value was < 0 . 05 . | Human T Lymphotropic virus type ( HTLV ) establishes persistent infections , leading to adult T-cell Leukemia , a life-threatening cancer in chronically-infected individuals . Viral persistence likely results from a failure of immune responses to eradicate viral replication , a least in part , by viral escape from innate immunity , and notably via decreased production of type I interferons ( IFN-I ) by infected cells . Plasmacytoid dendritic cells ( pDCs ) are known as robust producers of IFN-I in response to virus stimulation , thus bypassing the viral mechanisms to evade pathogen-sensing pathways in infected cells . However , HTLV particles are not detected in biological fluids of infected individuals , raising the question of the pDC-activating signal . Here , we demonstrate that pDCs produce IFN-I upon physical contacts with HTLV-infected cells . We show that pDCs sense surface associated-HTLV present with glycan-rich structure , referred to as HTLV-biofilm-like structure . Importantly , the sensing of infected cells by pDCs is modulated by the glycosylation pattern at the surface of infected cells . This newly ascribed regulation of innate immunity activation by cell surface-associated glycans might contribute to the differential activation levels of antiviral response to infected cells when their glycosylation profile is modified , such as for chronically infected cells or tumor cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"binding",
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"drugs",
"microbiology",
"viral",
"structure",
"retroviruses",
"viruses",
"rna",
"viruses",
"pharmacology",
"glycosylation",
"heparin",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"proteins",
"medical",
"microbiology",
"htlv-1",
"microbial",
"pathogens",
"biochemistry",
"cell",
"staining",
"lectins",
"cell",
"biology",
"post-translational",
"modification",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"glycobiology",
"organisms"
] | 2019 | Sensing of cell-associated HTLV by plasmacytoid dendritic cells is regulated by dense β-galactoside glycosylation |
Most mucosal surfaces of the mammalian body are colonized by microbial communities ( “microbiota” ) . A high density of commensal microbiota inhabits the intestine and shields from infection ( “colonization resistance” ) . The virulence strategies allowing enteropathogenic bacteria to successfully compete with the microbiota and overcome colonization resistance are poorly understood . Here , we investigated manipulation of the intestinal microbiota by the enteropathogenic bacterium Salmonella enterica subspecies 1 serovar Typhimurium ( S . Tm ) in a mouse colitis model: we found that inflammatory host responses induced by S . Tm changed microbiota composition and suppressed its growth . In contrast to wild-type S . Tm , an avirulent invGsseD mutant failing to trigger colitis was outcompeted by the microbiota . This competitive defect was reverted if inflammation was provided concomitantly by mixed infection with wild-type S . Tm or in mice ( IL10−/− , VILLIN-HACL4-CD8 ) with inflammatory bowel disease . Thus , inflammation is necessary and sufficient for overcoming colonization resistance . This reveals a new concept in infectious disease: in contrast to current thinking , inflammation is not always detrimental for the pathogen . Triggering the host's immune defence can shift the balance between the protective microbiota and the pathogen in favour of the pathogen .
The evolution of pathogenic microorganisms has been shaped to a great extent by their interaction with cognate host species . Colonization is the first step of any infection . For enteropathogenic bacteria , this poses a formidable task as the target host organ is already colonized by a dense microbial community , the microflora , or “microbiota” . Intestinal colonization by microbiota begins immediately after birth and lasts for life . In a healthy intestine , the microbiota is quite stable , and its gross composition at higher taxonomic levels is similar between individuals , and even between humans and mice [1] . The intestinal ecosystem is shaped by symbiotic interactions between the host and the microbiota . Microbiota composition is influenced by nutrient availability , local pH , and possibly also by the host's immune system [2] . Conversely , the microbiota optimizes nutrient utilization [3 , 4] , and boosts maturation of intestinal tissues and the intestinal immune system [5–7] . In addition , the microbiota provides an efficient barrier against infections ( “colonization resistance” ) , which must be overcome by enteropathogenic bacteria . It is poorly understood how enteropathogens can achieve that task . Here , we used Salmonella enterica subspecies 1 serovar Typhimurium ( S . Tm ) and a mouse colitis model to study strategies by which enteropathogenic bacteria break colonization resistance . S . Tm infects a broad range of animal species and is a frequent cause of intestinal infections in the human population . The normal murine microbiota provides colonization resistance and prevents intestinal colonization upon oral S . Tm infection . Oral treatment with the antibiotic streptomycin ( 20 mg of streptomycin intragastric [i . g . ] ) transiently reduces the microbiota by >80% and disrupts colonization resistance for a period of 24 h [8 , 9] . The residual microbiota re-grows within 2–3 d , and colonization resistance is re-established ( [9]; unpublished data ) . These studies have provided the basis for a “streptomycin mouse model” for Salmonella enterocolitis [10]: 1 d after streptomycin treatment , oral infection with S . Tm leads to efficient colonization of the murine intestine , especially the cecum and the colon ( approximately 109 colony-forming units [CFU]/gram; Figures 1A and S1 ) [8 , 9 , 11] . Wild-type S . Tm ( S . Tmwt ) triggers pronounced intestinal inflammation ( colitis ) and colonizes the intestinal lumen at high densities over extended periods of time [8 , 10–12] . This “streptomycin mouse model” can be used to study bacterial virulence factors required for colonization and triggering of intestinal inflammation . For example , S . Tm strains lacking the two virulence-associated type III secretion systems ( e . g . , S . Tm ΔinvG sseD::aphT [S . Tmavir] [13] ) cannot trigger colitis . In addition , these mutants were found to colonize the murine intestine only transiently [11 , 13] . The reason for this colonization defect has remained elusive . To explore this , we analyzed microbiota compostition in S . Tmwt– and S . Tmavir–infected mice and the role of inflammation for Salmonella colonization and competition against the intrinsic microbiota . We found that inflammation shifts the balance between the protective microbiota and the pathogen S . Tm in favour of the pathogen . This principle might apply to various other pathogens and therefore constitute a novel paradigm in infectious biology .
First , we confirmed the differential colonization efficiency of S . Tmwt and S . Tmavir in the streptomycin mouse model . Unlike S . Tmwt , intestinal S . Tmavir colonization levels decreased significantly by day 4 post-infection ( p . i . ) in a highly reproducible fashion ( Figure 1B ) . This coincided with re-growth of the microbiota as revealed by immunofluorescence microscopy ( Figure 1C–1H ) . By anaerobic culture , DNA isolation , and 16S rRNA gene sequencing , high densities of characteristic members of the intestinal microbiota ( Clostridium spp . , Bacteroides spp . , and Lactobacillus spp . [14] ) were found in S . Tmavir–infected , but not in S . Tmwt–infected , animals at day 4 p . i . ( Table 1 ) . Both the S . Tm/microbiota ratio and the composition of the microbiota itself differed between mice infected with S . Tmavir and S . Tmwt . These data demonstrated that residual microbiota surviving the streptomycin treatment can re-grow , outcompete S . Tmavir , and thereby re-establish colonization resistance . In contrast , S . Tmwt can suppress re-growth of the residual microbiota . Therefore , the streptomycin mouse model allows study of the principal mechanisms by which enteropathogens manipulate the intestinal ecosystem . To better characterize the effect of S . Tm on microbiota composition , we employed 16S rRNA gene sequencing ( see Materials and Methods ) . This method allows a quantitative comparison of microbial communities , including bacterial species that cannot be cultivated in vitro . The analysis comprised five different groups of mice and addressed the effect of the streptomycin pretreatment per se as well as the effect of S . Tmavir and S . Tmwt infection on microbiota composition ( Figure 2 ) . In line with published data , a large fraction of the murine microbiota in unmanipulated mice belonged to either the Firmicutes ( including Clostridium spp . and Lactobacillus spp . ; 39% ± 10% ) or the Bacteroidales ( 53% ± 13%; Figure 2 ) [1 , 15–17] . Streptomycin treatment reduced the global density of the microbiota by approximately 90% ( Figure 2; see also Figure 1C and 1D ) and changed its relative composition ( Figure 2A and 2B; Table 2 ) . The composition of the remaining microbiota varied substantially between individual members of this group ( Figure 2B ) . Most likely , this is attributable to the unstable situation created by the antibiotic and may arise from slight animal-to-animal variations in the timing or speed of the gut passage of the antibiotic and/or from species-specific differences in antibiotic susceptibility and rate of re-growth . Five days after the antibiotic treatment , the microbiota had re-grown to normal density and microbiota composition , at least at the phylum level ( Figure 2A and 2B; Table 2; p = 0 . 35078 ) . Infection with S . Tmavir did not interfere detectably with re-growth of the normal microbiota in the streptomycin-pretreated mouse model ( Figure 2B; Table 2 ) . In contrast , S . Tmwt significantly altered the cecal microbiota composition ( Figure 2A and 2B; Table 2; p < 0 . 00001 ) . Proteobacterial 16S rRNA gene sequences represented >90% of all sequences , and Salmonella spp . generally represented the most prominent ( up to 100% ) proteobacterial species in the S . Tmwt–infected animals . These observations were confirmed by fluorescence in situ hybridization ( FISH ) of fixed cecal content ( Figure S2 ) . This demonstrates that S . Tmwt interferes with microbiota re-growth and represents the predominant species at day 4 p . i . It should be noted that other proteobacterial species ( e . g . , Escherichia coli ) were also present in significant numbers in the cecum of most S . Tmwt–infected animals ( Figure 2A ) . These proteobacterial strains are low abundance members of the normal gut microbiota of our mouse colony ( <107 CFU/g of cecal content ) . In many mice the proportion of these commensal proteobacterial species increased concomitant with the S . Tmwt infection . This suggests that other bacterial species closely related to S . Tm may also be able to benefit from the S . Tmwt–triggered inflammation . Further work will be required to address this issue . The observed changes in microbiota growth in S . Tmwt–infected mice were verified in a competitive infection experiment with a specific member of the microbiota . For this purpose we selected a rifampicin-resistant variant of Lactobacillus reuteri strain RR ( L . reuteri RRRif ) . This strain was isolated as a commensal from our mouse colony . Streptomycin-treated mice were infected i . g . with either S . Tmwt or S . Tmavir ( 5 × 107 CFU i . g . ) and gavaged 1 d p . i . with L . reuteri RRRif ( 8 × 106 CFU i . g . ) . L . reuteri RRRif colonized the S . Tmavir–infected mice at levels of 105–106 CFU/g of intestinal content . In S . Tmwt–infected mice , similar L . reuteri RRRif colonization levels were observed at day 2 p . i . , but colonization levels declined below the detection limit by day 4 p . i . ( p = 0 . 008; Figure 3 ) . Thus , alteration of microbiota composition by S . Tmwt can be demonstrated at the level of a single bacterial strain . The above findings prompted us to investigate whether there is a cause-and-effect relationship between triggering of inflammation and enhanced colonization by S . Tm . In this case one would predict that S . Tmavir ( which cannot trigger inflammation ) competes successfully with the microbiota if inflammation is triggered by other means . Three different experimental approaches lent evidence for this hypothesis: First , we analyzed whether inflammation induced by S . Tmwt improved S . Tmavir colonization efficiency . Earlier experiments had shown that infections with 1:1 mixtures of S . Tmwt and attenuated mutants led to full-blown colitis ( Figure 4A and data not shown ) . Thus , streptomycin-treated mice were infected with a 1:1 mixture of S . Tmwt and S . Tmavir ( a total of 5 × 107 CFU i . g . ) . Control groups were infected with S . Tmwt or S . Tmavir only ( 5 × 107 CFU i . g . ; Figure 4A ) . Pronounced colitis was observed in all animals infected with S . Tmwt and the S . Tmwt–S . Tmavir mixture , but not in animals infected with S . Tmavir alone . Furthermore , S . Tmavir was severely defective at colonizing lymph nodes and spleen in single and mixed infections . Despite its non-pathogenic phenotype , S . Tmavir colonized the cecal lumen up to wild-type levels in mixed infections with S . Tmwt . Thus , concomitant colitis created favourable conditions in the intestinal lumen that suppressed microbiota regrowth and rescued S . Tmavir colonization in tandem . This was confirmed in long-term infection experiments using 129Sv/Ev mice , which develop a chronic form of colitis ( Figures 4B and S3 ) [12] . Next , we studied whether cecal inflammation per se ( in absence of S . Tmwt ) could enhance S . Tmavir colonization . For this purpose we employed knockout mouse models lacking the key anti-inflammatory cytokine IL10 . Depending on the exact genetic background and the composition of the microbiota , these animals develop colitis spontaneously earlier ( week 6; C3H/HeJBirIL10−/− model [18] ) or later in life ( week 30–50; C57BL/6IL10−/− model [19] ) . To test the effect of pre-existing colitis on S . Tmavir colonization , groups of 8-wk-old C3H/HeJBirIL10−/− mice and C3H/He control mice were infected ( 5 × 107 CFU of S . Tmavir i . g . ; no streptomycin treatment ) . Fecal shedding ( day 1 p . i . ) , colonization , and colitis ( day 2 p . i . ) were analyzed . Colonization of the intestinal lumen by S . Tmavir was significantly enhanced in mice displaying colitis ( day 2 p . i . , p = 0 . 016; Figures 5A , S3 , and S4 ) . Similar observations were made using the C57BL/6IL10−/− model . In C57BL/6IL10−/− mice , the onset of colitis is quite random and varies anywhere from 30 to 50 wk even between littermates . Accordingly , we infected C57BL/6IL10−/− littermates 30–50 wk of age ( 5 × 107 CFU of S . Tmavir i . g . ; no streptomycin treatment ) . Again , colonization of the intestinal lumen by S . Tmavir was enhanced in littermates displaying colitis ( day 1 p . i . , p = 0 . 016; Figures 5B , S3 , and S4 ) . This suggested that inflammation per se can enhance S . Tmavir colonization . To verify this hypothesis we employed the alternative , recently developed VILLIN-HACL4-CD8 mouse model for T cell–induced colitis [20] . This model employs VILLIN-HA transgenic mice expressing the HA epitope in the gut epithelium and T cells ( CD8+; HA-directed α/β T cell receptor; from CL4-TCR transgenic mice ) recognizing the HA epitope . Adoptive transfer of these T cells into VILLIN-HA transgenic mice results in severe inflammation of the small and the large intestine at 4–5 d post-transfer ( Figure 6A ) [20] . This model was of particular interest because intestinal inflammation develops quickly , occurs in the majority of animals , and does not involve i . g . treatment with chemicals that might themselves influence the microbiota–pathogen competition . To study the impact of inflammation on S . Tmavir colonization we infected VILLIN-HA transgenic mice receiving CL4-CD8 T cells and unmanipulated VILLIN-HA control mice . In the unmanipulated VILLIN-HA mice ( no T cells transferred ) , no intestinal inflammation was apparent and S . Tmavir colonization efficiency was low ( Figure 6B ) . In contrast , the animals receiving 4 × 106 CL4-CD8 T cells ( VILLIN-HACL4-CD8 mice ) developed intestinal inflammation 4 or 5 d after T cell transfer , and S . Tmavir efficiently colonized the intestine of these animals ( p = 0 . 01; Figure 6B ) . It should be noted that the initial colonization by S . Tmavir was poor ( fecal samples at days 2 and 3 after T cell transfer ) and that the onset of efficient S . Tmavir colonization closely correlated with the onset of the intestinal inflammation ( day 4–5 after T cell transfer [20] ) . At this stage , “Salmonella” sequences represented 26%–46% of all bacterial 16S rRNA genes recovered from the cecal contents ( Figure 6C ) . This confirmed that colitis per se creates conditions in the gut skewing the competition between Salmonella spp . and the microbiota in favour of the pathogen . As additional controls , we analyzed the fecal microbiota composition of unmanipulated VILLIN-HA transgenic mice ( n = 4 ) and non-infected VILLIN-HA transgenic mice ( n = 2 ) at day 4 after CL4-CD8 T cell transfer ( Figures 6C and S6; Table S2 ) . The latter two animals showed intestinal inflammation comparable to that in mice that received CL4-CD8 T cells and S . Tmavir ( data not shown ) . At the phylum level , we did not detect any significant differences between the microbiota recovered from the feces of the unmanipulated mice ( no gut inflammation ) , the VILLIN-HA transgenic mice that had received CL4-CD8 T cells ( gut inflammation ) , and the S . Tmavir–infected VILLIN-HA transgenic mice that had not received CL4-CD8 T cells ( no gut inflammation ) . These data suggest that inflammation per se does not drastically alter the gross gut flora composition ( at least not in the short term ) . Further work is required to determine whether the loss of colonization resistance in the inflamed VILLIN-HA transgenic mice is attributable to suppression of some particular , low abundance member ( s ) of the microbiota . Finally , our data show that S . Tmavir colonization efficiency in the murine intestine is restricted by the intestinal microbiota . In the absence of microbiota , S . Tmavir should colonize efficiently . This was confirmed in germ-free mice that lack microbiota in the first place . S . Tmavir colonized the large intestine of germ-free mice at wild-type levels up to day 4 p . i . ( approximately 109 CFU/g ) but did not cause colitis ( Figure S5 ) . Thus , S . Tmavir efficiently colonizes the murine intestine as long as competing microbiota is lacking . Furthermore , inflammation is not required for colonizing the intestinal lumen in the absence of microbiota . However , it should be noted that germ-free mice represent a useful but highly artificial tool . In natural habitats , Salmonella spp . always encounters a dense intestinal microbiota , and intestinal colonization will be enhanced by the triggering of inflammation .
Based on these data we propose a three-way microbiota–pathogen–host interaction model for murine Salmonella colitis ( Figure 7 ) . The resident microbiota and the incoming pathogen compete for growth . In a “healthy” intestine the normal microflora is shaped and stabilized by mutually beneficial interactions with the intestinal mucosa . It effectively excludes S . Tmwt and S . Tmavir from the intestinal lumen . Colonization resistance can be transiently alleviated by streptomycin treatment . Inflammatory host responses—triggered by specific S . Tm virulence factors ( TTSS-1 and TTSS-2 ) , by genetic pre-disposition ( IL10−/− ) , or by T cell–inflicted damage ( VILLIN-HACL4-CD8 model ) —alter conditions in the intestinal lumen and shift the competition in favour of the incoming pathogen . Suppression of the microbiota or enhanced pathogen growth may be involved ( Figure 7 ) . In either case , S . Tmwt can enhance intestinal colonization via an indirect mechanism—by triggering the host's immune defence . Thus , S . Tmwt infection involves two different steps: triggering inflammation , and surviving in and profiting from the altered ecological niche . The avirulent mutant S . Tmavir is unable to trigger colitis but it is still capable of taking advantage of the ecological niche opened by inflammation and thus successfully competes with the microbiota if inflammation is induced by other means . How does intestinal inflammation subvert colonization resistance ? The inflammation involves increased secretion of antibacterial peptides and lectins [21 , 22] and mucins ( B . Stecher and W . Hardt , unpublished data ) , phagocyte infiltration/transmigration , and release of oxygen and nitrogen radicals . Potentially , there are a number of different ways this may subvert colonization resistance . ( 1 ) Released antibacterial factors may kill or retard growth of specific members of the microbiota that would normally inhibit S . Tm growth in the healthy intestine . ( 2 ) There may be “commensal network disruption” , i . e . , loss of one or more specific microbiota species that might be required for efficient growth of the microbiota species that slow pathogen growth in the normal , healthy intestine . These protecting species and their integration into microbiota growth networks have not been identified . ( 3 ) There may be differential defence susceptibility . Microbiota species conferring colonization resistance might be susceptible to antibacterial defences that S . Tm can resist . This would be in line with the discovery of numerous S . Tm genes that function to enhance antimicrobial peptide resistance and radical detoxification [23–25] . ( 4 ) There may be enhanced pathogen growth . The altered nutrient mix available in the inflamed gut might foster efficient pathogen replication . Under these conditions , microbiota may simply grow slower and are thus overgrown by the pathogen . The model is summarized in Figure 7 . Future work will have to address which of these mechanisms contribute to subversion of gut inflammation by S . Tm . Inflammation induced by S . Tm , self-reactive T cells , or IL-10 deficiency enhances colonization by the pathogen and reduces growth of the commensal microbiota . Other proteobacteria closely related to S . Tm may also benefit from inflammation ( e . g . , E . coli; see Figure 2 ) . Thus , this principle may also apply to other enteric infections . For example , in calves , which are naturally susceptible to Salmonella enterocolitis , defects of Salmonella TTSS-2 mutants in triggering inflammation are associated with attenuation of intestinal colonization [26 , 27] . Similar observations were made with Shigella flexneri , Vibrio cholerae , and Citrobacter rodentium , the causative agents of bacillary dysentery , cholera , and transmissible murine colonic hyperplasia: ablation of colitis by disrupting the hosts' innate immune response or specific bacterial virulence factors coincided with reduced intestinal colonization [28–31] . Thus , intestinal inflammation and efficient colonization may be linked in a broad range of enteropathogenic infections . Some data are available for human Salmonella enterocolitis . In line with findings in the murine system , antibiotics are known to reduce human colonization resistance , and altered microbiota composition is commonly observed in patients with inflammatory bowel disease ( IBD ) [32–34] . Furthermore , some studies suggest an increased incidence of Salmonella colonization in IBD patients [35–40] . Microbiota composition in IBD patients significantly differs from that in healthy controls . Currently , an imbalance in normal gut microbiota is regarded as one possible factor triggering the inflammation in Crohn disease and ulcerative colitis [41–43] . Our data suggest that the altered gut flora might not be the cause , but rather one of the many symptoms , of intestinal inflammation in IBD patients . Further investigation into this idea will be of importance for basic research exploring the aetiology and pathogenesis of Crohn disease and ulcerative colitis . The outcome of any infection is determined through competition between the bacterial virulence factors ( enhancing pathogen replication/persistence ) and the host's immune defences ( eliminating the pathogen ) . In the case of enteropathogens , which target a niche colonized by the microbiota , the virulence factors can serve an additional function that has remained unrecognized: they allow triggering of intestinal inflammation that subverts the host's immune defences for undermining colonization resistance . This may represent a common virulence strategy of enteropathogenic bacteria including Clostridium difficile , which is a frequent cause of antibiotic-associated colitis . In fact , inflammation may promote pathogen competitiveness at any colonized site of the human body , and pathogens infecting the respiratory tract , the uro-genital system , and the skin might also use this strategy . Molecular analysis of the complex three-way pathogen–host–microbiota interactions poses a great technological challenge for future research and promises to reveal novel avenues for determining prevention strategies and cures for infectious disease .
All aspects of animal procedures were approved by local authorities and performed according to the legal requirements . Sex- and age-matched specified pathogen free ( SPF ) C57Bl/6 ( Elévage Janvier; http://www . janvier-breedingcenter . com/ ) , 129Sv/Ev , C3H/He ( Charles River Laboratories , http://www . criver . com/ ) , C57BL/6IL10−/− [19] , and C3H/HeJBirIL10−/− [18] mice were held under barrier conditions at the Rodent Centre , Swiss Institute of Technology Zurich , Zurich , Switzerland , and the Biologisches Zentrallabor , University of Zurich , Zurich , Switzerland . VILLIN-HA [44] and CL4-TCR [45] transgenic mice were raised under SPF barrier conditions at the Helmholtz Centre for Infection Research , Braunschweig , Germany , and transferred to the Rodent Centre 1 wk before the infection experiment . Germ-free C57BL/6 mice were bred and infected in the germ-free facility of the Biologisches Zentrallabor . 129Sv/Ev mice used for long-term infection experiments ( Figure 4 ) were transferred to C57BL/6 foster mice at the day of birth , and raised and weaned as usual . In the streptomycin mouse model , mice were treated with streptomycin ( 20 mg i . g . ) [13] and infected 24 h later with S . Tm strains ( 5 × 107 CFU i . g . ) as indicated . For super-infection , L . reuteri RRRif ( 8 × 106 CFU i . g . ) was administered 24 h after S . Tm infection . No streptomycin treatment was performed in spontaneous colitis models and germ-free mice ( Figure 3C and 3D ) . For induction of acute colitis , CD8+ T cells from CL4-TCR transgenic mice that express an α/βT cell receptor recognizing an epitope of the HA protein presented by MHC class I ( the H-2Kd:HA512–520 complex ) were adoptively transferred into VILLIN-HA mice that express the A/PR8/34 HA epitope from influenza virus A under control of the enterocyte-specific villin promoter [20] . Single-cell suspensions were prepared from the spleen of CL4-TCR transgenic mice . Cell suspensions were depleted of CD4+ , CD11b+ , CD45R+ , DX5+ , and Ter-119+ cells by using the MACS CD8 T cell isolation kit ( Miltenyi Biotec , http://www . miltenyibiotec . com/ ) . CL4-TCR T cells were purified by negative selection according to the manufacturer's instructions . Isolated CD8+ T cells were washed once in PBS and resuspended ( 4 × 107 cells/ml of PBS ) . Then 4 × 106 purified CL4-TCR transgenic T cells were injected intravenously into VILLIN-HA transgenic mice . Disease symptoms ( weight loss and diarrhoea ) were observed 4–5 d after adoptive transfer . The streptomycin-resistant wild-type strain S . Tmwt ( SL1344 wild-type [46] ) and the isogenic mutant S . Tmavir ( ΔinvG sseD::aphT; kanR [13] ) were grown in LB 0 . 3 M NaCl as described [13] . Colonization was defined by plating on MacConkey agar plates ( Oxoid , http://www . oxoid . com/; 100 μg/ml streptomycin ) . Co-infections with S . Tmavir were evaluated by replica-plating on medium containing kanamycin ( 50 μg/ml ) . Culturable intestinal microbiota were grown on Wilkins Chalgren agar supplemented with 5% defibrillated sheep blood ( Oxoid ) for 3–5 d in an atmosphere of 7% H2 , 10% CO2 , and 83% N2 at 37 °C in anaerobic jars . 16S rRNA gene sequencing was performed as described below . L . reuteri RRRif was selected on MRS medium ( 100 μg/ml of rifampicin; Laboratoire Labo'Life , http://www . labolife . com/ ) and grown anaerobically . Fresh fecal pellets collected from individual mice and cecum content were resuspended in PBS . Mesenteric lymph nodes ( mLN ) , spleen , and liver were removed aseptically and homogenized in cold PBS ( 0 . 5% tergitol and 0 . 5% BSA ) . Bacteria were enumerated by plating on appropriate medium . Colonies were isolated and purified twice on Wilkins Chalgren agar ( 5% sheep blood ) . DNA was recovered by lysis ( Tris/EDTA; 0 . 5% SDS and 0 . 1 mg/ml of proteinase K; 37 °C; 1 h ) , CTAB treatment ( 1%; 62 . 5 mM NaCl; 65 °C; 10 min ) , phenol-chloroform extraction , and 2-propanol precipitation . Broad-range bacterial primers fD1 ( 5′-AGA GTT TGA TCC TGG CTC AG-3′ ) and rP1 ( 5′-ACG GTT ACC TTG TTA GCA CTT-3′ ) [47] were used for 16S rRNA gene PCR amplification ( 94 °C , 5 min; 35 cycles of 94 °C , 1 min; 43 °C , 1 min; 72 °C , 2 min; and 7-min final extension at 72 °C ) . The PCR product was purified and sequenced with primer rP1 . First , bacteria were grouped according to colony morphology . Then , representative colonies were typed by 16S rRNA gene sequencing and comparison to the Ribosomal Database Project II [48] . This allowed a rough estimation of the abundance of the respective bacterial species ( Table 1 ) . Two mice were analyzed per condition ( S . Tmwt and S . Tmavir infection day 4 p . i . ) . Six colony morphological groups were assigned for S . Tmwt infection , and ten for S . Tmavir infection . Tissues were cryo-embedded in Tissue Tek OCT Compound ( Sysmex , http://www . sysmex-europe . com/ ) , 5-μm cryosections were stained with hematoxylin and eosin ( HE ) , and cecum pathology was evaluated using a histopathological scoring scheme as previously described [49 , 50] ( see Figure S1 ) . Evaluation scored submucosal edema ( score 0–3 ) , polymorphonuclear leukocyte infiltration into the lamina propria ( score 0–4 ) , loss of goblet cells ( score 0–3 ) , and epithelial damage ( score 0–3 ) . The combined pathological score for each tissue sample was determined as the sum of these averaged scores: 0–3 , no to minimal signs of inflammation that are not sign of a disease ( this is frequently found in the cecum of SPF mice ) ; 4–8 , moderate inflammation; and 9–13 , profound inflammation . Cecal tissues were fixed in PBS ( 4% paraformaldehyde [pH 7 . 4]; 4 °C; 12 h ) , washed in PBS , equilibrated in PBS ( 20% sucrose and 0 . 02% NaN3; 4 °C; 12 h ) and cryo-embedded in OCT . Cryosections ( 7 μm ) were mounted on glass slides , air-dried ( 21 °C; 2 h ) , fixed in PBS ( 4% paraformaldehyde , 5 min ) , washed , and blocked with 10% ( w/v ) normal goat serum in PBS ( 1 h ) . S . Tm was stained with polyclonal rabbit anti–Salmonella O antigen group B serum ( factors 1 , 4 , 5 , and 12 , Brunschwig , http://www . brunschwig-ch . com/; 1:500 in PBS , 10% goat serum ) and a Cy3-conjugated goat anti-rabbit antibody ( Milan; 1:300 in PBS , 10% goat serum ) . The specificity of the anti–Samonella O ( 1 , 4 , 5 , and 12 ) antiserum was checked extensively by immunofluorescence microscopy . This was done by analyzing cecum tissue sections from uninfected mice ( negative ) , S . Tm–infected mice ( positive ) , S . enterica serovar Enteritidis–infected mice ( negative; the LPS of this serovar does not react with this antiserum ) , and mice with >10 different commensal species , including commensal E . coli strains from our mouse colony , grown in vitro ( all negative ) . DNA was stained with Sytox green ( 0 . 1 μg/ml; Sigma-Aldrich , http://www . sigmaaldrich . com/ ) and F-Actin with Alexa-647-phalloidin ( Molecular Probes , http://probes . invitrogen . com/ ) . Sections were mounted with Vectashield hard set ( Vector Laboratories , http://www . vectorlabs . com/ ) and sealed with nail polish . Images were recorded using a PerkinElmer ( http://www . perkinelmer . com/ ) Ultraview confocal imaging system and a Zeiss ( http://www . zeiss . com/ ) Axiovert 200 microscope . For quantification of total bacterial numbers , cecal contents were weighed , fixed in 4% paraformaldehyde , and stained with Sytox green ( 0 . 1 μg/ml ) . Bacteria were counted in a Neubauer's counting chamber using an upright fluorescence microscope ( Zeiss ) . Total DNA was extracted from cecal contents using a QIAmp DNA stool mini kit ( Qiagen , http://www1 . qiagen . com/ ) and a Tissuelyzer device ( Qiagen ) . 16S rRNA genes were amplified by PCR using primers Bact-7F ( 5′-AGA GTT TGA TYM TGG CTC AG-3′ ) and Bact-1510R ( 5′-ACG GYT ACC TTG TTA CGA CTT-3′ ) and the following cycling conditions: 95 °C , 5 min; 22 cycles of 95 °C , 30 s; 58 °C , 30 s; 72 °C , 2 min; followed by 72 °C , 8 min; 4 °C , ∞ . Reaction conditions ( 100 μl ) were as follows: 50 mM KCl , 10 mM Tris-HCl ( pH 8 . 3 ) , 1 . 5 mM Mg2+ , 0 . 2 mM dNTPs , 40 pmol of each primer , and 5 U of Taq DNA polymerase ( Eppendorf , http://www . eppendorf . com/ ) . Fragments were purified by gel electrophoresis , excised , recovered using the gene clean kit ( Qbiogene; http://www . qbiogene . com/ ) and dried . The PCR products were suspended in 10 μl of sterile distilled water and between 2 and 5 μl was ligated into pGEM-T Easy Vectors ( Promega , http://www . promega . com/ ) . The ligated vectors were transformed into high-efficiency competent JM109 E . coli cells ( Promega ) , plated on LB-carbenicillin agar , and subjected to blue-white screening of colonies . White colonies were picked into 96-well boxes containing 500 μl of Circlegrow medium ( Qbiogene , http://www . qbiogene . com/ ) per well and grown overnight at 37 °C , and the plasmid DNA was then prepped using a modified semi-automated alkaline lysis method . Sequencing was carried out using Applied Biosystems ( http://www . appliedbiosystems . com/ ) BigDye terminators ( version 3 . 1 ) and run on Applied Biosystems 3730 sequencers . The 16S rRNA gene inserts were sequenced using two primers targeted towards the vector end sequences , M13r ( 5′-CAGGAAACAGCTATGACC-3′ ) and T7f ( 5′-TAATACGACTCACTATAGGG-3′ ) , and one towards an internal region of the gene , 926r ( 5′-CCGTCAATTC[A/C]TTT[A/G]AGT-3′ ) , in order to bridge any gaps between the sequences generated from the two end primers . Contigs were built from each three-primer set of sequences using the GAP4 software package [51] and converted to “sense” orientation using OrientationChecker software [52] . These files were then aligned using MUSCLE [53] , and the alignments were manually inspected and corrected using the sequence editor function in the ARB package [54] . The files were then tested for the presence of chimeric sequences using Mallard [52] and Bellerophon [55] , and putative chimeras were checked using Pintail [56] and BLAST [57] . Positively identified chimeras were removed , and the remaining sequences were examined with the Classifier function at the Ribosomal Database Project II Web site [48] in order to give a broad classification at the phylum level . To obtain more detailed taxonomic information the sequences were divided into phylotypes by generating distance matrices in ARB ( with Olsen correction ) , which were then entered into the DOTUR program [58] set to the furthest neighbour and 99% similarity settings . The resulting phylotypes were then assigned similarities to nearest neighbours using BLAST . Statistical analyses of viable CFU and pathological scores were performed using the exact Mann-Whitney U Test and the SPSS version 14 . 0 software , as described before [8] . Values of p < 0 . 05 were considered statistically significant . Box-plots were created using GraphPad Prism 4 version 4 . 03 ( GraphPad Software , http://www . graphpad . com/ ) . Differences in the phylogenetic compositions of samples were assessed by first assigning the detected 16S rRNA gene sequences to their respective phyla , and then computing the normalized Euclidean distance between the phyla counts . The observed differences were judged for their statistical significance by performing Monte Carlo randomizations: 16S rRNA gene sequences were shuffled between two samples , such that overall sample sizes and total counts for each phylum were maintained . Euclidean distances were then re-computed , and the fraction of distances larger than or equal to the observed distances determined the p-values . Bonferroni correction for multiple testing means that p-values below 0 . 005 indicate statistical significance in Figures 2 and 6 and Table 2 .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/ ) accession numbers for the 16S RNA gene sequences shown in Figure 2 are EF604903–EF605247 , and for those shown in Figure 6C are EF604904–EF605247 and EU006095–EU006496 . | A dense microbial community colonizes the intestinal tract of mammals , contributing to health and nutrition and conferring efficient protection against most pathogenic intruders . Intestinal pathogens can overcome this colonization resistance and cause disease; however , the mechanisms used to do this are still elusive . In this study we analyzed intestinal infection by the model pathogen Salmonella enterica subspecies 1 serovar Typhimurium ( S . Tm ) . We show that the virulent wild-type pathogen overcomes colonization resistance by inducing the host's inflammatory immune response and exploiting it for its purpose . In contrast , an avirulent Salmonella mutant defective in triggering inflammation was unable to overcome colonization resistance by itself . Colonization by this mutant was restored if inflammation was provided concomitantly , in mice with inflammatory bowel disease ( genetic and inducible ) or by co-infection with wild-type S . Tm . These findings reveal a previously unrecognized strategy by which pathogenic bacteria overcome colonization resistance: abusing the host's inflammatory immune response to gain an edge against the normal microbial community of the gut . This represents a first step towards unravelling the molecular mechanisms underlying this three-way interaction of host , microbiota , and pathogens . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"infectious",
"diseases",
"gastroenterology",
"and",
"hepatology",
"immunology",
"microbiology",
"mus",
"(mouse)"
] | 2007 | Salmonella enterica Serovar Typhimurium Exploits Inflammation to Compete with the Intestinal Microbiota |
The genes encoding ribosomal RNA are the most abundant in the eukaryotic genome . They reside in tandem repetitive clusters , in some cases totaling hundreds of copies . Due to their repetitive structure , ribosomal RNA genes ( rDNA ) are easily lost by recombination events within the repeated cluster . We previously identified a unique gene amplification system driven by unequal sister-chromatid recombination during DNA replication . The system compensates for such copy number losses , thus maintaining proper copy number . Here , through a genome-wide screen for genes regulating rDNA copy number , we found that the rtt109 mutant exhibited a hyper-amplification phenotype ( ∼3 times greater than the wild-type level ) . RTT109 encodes an acetyl transferase that acetylates lysine 56 of histone H3 and which functions in replication-coupled nucleosome assembly . Relative to unequal sister-chromatid recombination-based amplification ( ∼1 copy/cell division ) , the rate of the hyper-amplification in the rtt109 mutant was extremely high ( >100 copies/cell division ) . Cohesin dissociation that promotes unequal sister-chromatid recombination was not observed in this mutant . During hyper-amplification , production level of extra-chromosomal rDNA circles ( ERC ) by intra-chromosomal recombination in the rDNA was reduced . Interestingly , during amplification , a plasmid containing an rDNA unit integrated into the rDNA as a tandem array . These results support the idea that tandem DNA arrays are produced and incorporated through rolling-circle-type replication . We propose that , in the rtt109 mutant , rDNA hyper-amplification is caused by uncontrolled rolling-circle-type replication .
The ribosome is an abundant macromolecular protein-RNA complex that translates mRNA into protein . In Saccharomyces cerevisiae , ribosomal proteins ( RP ) account for approximately 50% of total protein and ribosomal RNA ( RNA ) represents approximately 80% of total RNA [1] . While the amount of RP increases during translation from mRNA , the amount of rRNA is dependent on transcription level . To meet this huge biosynthetic demand for rRNA in eukaryotic cells , the rRNA genes are present in hundreds of copies and are transcribed by a highly efficient RNA polymerase , RNA pol I [1] . We recently reported that about half the rDNA genes are transcriptionally silent , but serve as a “foothold” for repair enzymes to maintain the integrity of the rDNA region [2] . Nonetheless , the highly repetitive structure of the rDNA makes it fragile and vulnerable to loss of copies following homologous recombination events within the repeat . As an apparent adaptation for a requirement for many copies of the rRNA genes and their simultaneous genetic instability , an rDNA-specific amplification system has evolved in eukaryotic cells . In Saccharomyces cerevisiae , the approximately 150 tandemly-repeated rDNA genes are located on chomosome XII ( Figure 1A ) [3] . One unit of rDNA is 9 . 1 kb and is composed of a region encoding a pre-35S rRNA , a small 5S rRNA and two intergenic spacer ( IGS ) . The IGS region encodes the key elements involved in the amplification system that maintains rDNA gene copy number [4] . During S phase , progress of the DNA replication fork is blocked at the replication fork barrier sequence ( RFB ) through the action of the fork-blocking protein , Fob1 [5] . This inhibition induces DNA double-strand breaks ( DSBs ) and subsequent repair by unequal sister-chromatid recombination which increases rDNA copy number through re-replication of neighbor copies [6] . Moreover , this amplification is promoted by RNA pol II dependent transcription from E-pro , a non-coding bi-directional promoter near the RFB sequence [7] . Transcription from E-pro is proposed to clear cohesin proteins from nearby , which in turn may promote unequal sister-chromatid recombination , because cohesin proteins suppress unequal sister chromatid recombination [7] . In a wild type strain with ∼150 rDNA copies , E-pro transcription is repressed by a histone deacetylase , Sir2 . Once rDNA copy number is reduced , Sir2 repression is relieved and amplification by unequal sisiter-chromatid recombination is enhanced . This amplification gradually increases copy number at a rate of ∼ one per cell division until a maximum of about 150 copies is reached [3] . To elucidate the mechanism by which rDNA copy number is regulated , we sought and found mutants in which copy number increased abnormally by screening a genome-wide deletion library consisting of ∼4 , 800 mutants ( OpenBiosystems ) . Among these mutants , the rtt109Δ strain exhibited a prominent phenotype in which rDNA copy number is ∼400 in contrast to ∼150 in the wild type . RTT109 encodes an acetyltransferase that is stimulated by the histone chaperone Asf1 [8]–[11] to acetylate histone H3 on residue K56 . The acetylated histones are incorporated into newly-synthesized DNA to promote replication-coupled nucleosome assembly by chromatin assembly factor 1 ( CAF-1 ) and histone chaperone Rtt106 [12] . It is known that the defect in acetylation affects DNA damage sensitivity , stability of the replication fork , and recombination between sister-chromatids [8] , [13]–[16] . Recently , Houseley and Tollervey also independently found this phenotype in the asf1 and rtt109 mutants [17] . We analyzed the relationship between histone acetylation and rDNA expansion in depth . We found that mutation of K56 of histone H3 also induced “hyper-amplification” of rDNA . Moreover , the histone deacetylase double mutant , hst3 hst4 , had an increased rDNA copy number as well . Interestingly , the rate of the hyper-amplification in the rtt109 mutant was much faster ( ∼100 copies per cell division ) , than the ∼1 copy per cell division in wild-type . During hyper-amplification , extra chromosomal rDNA circles ( ERC ) generated from the rDNA by Fob1-dependent recombination did not accumulate . Instead , integration of plasmids with rDNA sequences was detected . They replicated sequentially and repeatedly during the integration step . These observations suggest that intra-chromosomal recombinational repair of Fob1-driven DSB that produces ERC does not occur . Instead , the data suggest that break-induced replication ( BIR ) takes place between the DSB end and ERC , resulting in rolling circle replication and hyper-amplification . We conclude that Rtt109 prevents such unusual replication and facilitates completion of recombinational repair with a sister-chromatid after a DNA double-stranded break occurs at the blocking site .
To elucidate the mechanism which regulates rDNA copy number , we measured the size of rDNA in the ∼4 , 800 mutants that comprise the yeast deletion library ( S288c genetic background ) by pulsed-field gel electrophoresis ( CHEF , Figure S1 ) . We identified eight mutants with higher-than-usual copy numbers ( ∼150 ) , and presumed that they were defective in copy number regulation ( Table 1 ) . The rtt109 mutant was among those with the highest rDNA copy number . Chromosome XII of the rtt109 mutant is bigger than the biggest size maker ( 3 . 13 Mb ) and sharper than other chromosome XIIs ( lane 6 in Figure S1 ) . This indicates that the chromosome XII is too big to be analyzed under these electrophoretic conditions and failed to migrate . To confirm the “hyper-copy” rDNA phenotype in the rtt109 mutant , we deleted RTT109 in a different genetic background ( W303 ) and tested the size of chromosome XII in this mutant by CHEF . As shown in Figure 2A , chromosome XII in the rtt109 strain was much longer than in the wild type strain . Moreover , we measured rDNA copy number in the rtt109 mutant based on signal intensities of Southern blots after BglII digestion ( Figure 2B ) . The copy number was ∼2 . 5 times greater ( ∼400 copies ) than in the wild type strain ( ∼150 copies , Figure 2B ) . To determine whether hyper-amplification of the rDNA was a direct result of the deletion of RTT109 , we transformed the rtt109 mutant with plasmid-encoded RTT109 and monitored the size of chromosome XII by CHEF . Chromosome XII remained longer up to ∼24 generations following the transformation ( Figure 2C , lane 6–10 ) , but decreased in size to that observed in the wild-type strain after ∼90 generations ( Figure 2D , lane 5–6 in the W303 background , lane 13–16 in the S288c background ) , indicating that the hyper-copy rDNA phenotype was due to deletion of RTT109 . Rtt109 acetylates the K56 residue of newly-synthesized histone H3 and plays an important role in chromatin remodeling during DNA replication , repair , and recombination [8] , [12] , [14] , [15] . Acetylation of K56 no longer occurs in the rtt109 mutant [8]–[10] . To determine whether the acetylation activity is related to the hyper-amplification of rDNA , we measured rDNA copy number in the histone H3 mutants , H3K56R ( non-acetylated ) and H3K56Q ( acetylated mimic ) . As expected , the H3K56R mutant had a longer chromosome XII similar to the rtt109 mutant ( Figure 3A , lane 2 ) , indicating that the hyper-amplification of rDNA occurred by inhibition of acetylation of the histone H3K56 residue . However , unexpectedly , the H3K56Q mutant also had a longer chromosome XII ( Figure 3A , lane 3 ) , although the rDNA copy number did not reach the level observed in the histone deacetylase mutant . Moreover , in the W303 genetic background , the double H3K56 histone deacetylase ( HDAC ) mutant , hst3 hst4 , in which almost all H3 K56 residues were acetylated , exhibited the hyper-amplification phenotype as well ( Figure 3B ) [18] . These results indicate that deacetylation of K56 on histone H3 on the sister chromatin also prevents hyper-amplification and maintains proper copy number . We have previously shown that rDNA amplification from low copy number to wild type levels is dependent on the replication fork blocking activity of Fob1 and subsequent recombination [3] . To determine whether hyper-amplification starting from wild type copy number levels also requires this blocking activity , we deleted RTT109 in the fob1 mutant . As shown in Figure 4A ( right , lanes 9–10 ) , hyper-amplification in the double mutant was not observed . However , after more than 50 generations , some rtt109Δ fob1 clones had a longer chromosome XII than the single fob1 mutant , although they were still shorter than chromosome XII in the rtt109 single mutant ( Figure S2 , lanes 15–16 ) . To confirm the contribution of Fob1 to the hyper-amplification phenotype , we monitored amplification rate in the rtt109 fob1 GAL-FOB1 double mutant over a time course after induction of FOB1 transcription . Amplification was observed 6 hrs post-induction ( Figure 4B , asterisks ) . Because the doubling time of the mutant we measured was 3 . 4 hrs , the hyper-amplification phenotype was detectable within ∼2 generations . On the other hand , in cells grown on glucose where FOB1 was not induced , chromosome XII appeared slightly smeared in the gel , but the longer chromosome XII variant was not detected even after 32 hrs ( Figure 4B ) . Taken together , these observations indicate that Fob1 plays an essential role in the hyper-amplification phenotype . We also tested whether hyper-amplification occurred in mutants belonging to the RTT109 epistasis group . A histone chaperone , Asf1 , functions with Rtt109 to acetylate histone H3K56 [19] , [20] . In an asf1 mutant , the rDNA was also hyper-amplified , although the length of chromosome XII was much shorter than in the rtt109 mutant ( Figure 4A , lane 4–5 ) . Deletion of the other histone chaperone Vps75/Rtt109 complex that acetylates H3 N-terminal residues did not promote a hyper-amplification phenotype ( Figure 4A , lane 6–7 ) [21]–[23] . A similar result was reported by Houseley & Tollervey ( 2011 ) [17] . In the double fob1 asf1 mutant , amplification was substantially repressed , while in the vps75 mutant , no hyper-amplification was observed ( Figure 4A , lane 11–14 ) . This is consistent with a previous report that showed that an Rtt109-Asf1 , but not an Rtt109-Vps75 complex , acetylated the histone H3/H4 dimer , which is a key event required for genome stability [9] . Because histone H3K56 acetylation is important for replisome integrity at a stalled replication fork under replication stress [8] , [10] , we monitored stability of a replication fork arrested at the RFB site in the rtt109 mutant , in which FOB1 is induced by galactose as used in Figure 4B , by two-dimensional gel electrophoresis during hyper-amplification . In both the wild type and rtt109 mutant , the RFB spot on the Y-arc appeared only when FOB1 was induced ( cells grown on galactose ) . Because the signal intensities were similar ( Figure 5A ) , the fork stability at the RFB appears not to be affected by RTT109 function . In addition , the amount of recombination intermediates was not increased in the rtt109 mutant ( Figure 5A ) . Non-coding transcription from the IGS region ( E-pro ) is up-regulated during amplification from low copy number rDNA [7] . The transcription reduces association of the cohesin complex , resulting in unequal sister-chromatid recombination [7] . Therefore , we tested whether transcription increased in the rtt109 mutant . By northern analysis , the RNA increased in the rtt109 mutant ( Figure 5B ) . However , high level E-pro expression was not observed in the rtt109 fob1 double mutant ( Figure 5B ) in which rDNA copy number is maintained at wild type levels ( Figure 4A , lanes 9–10 ) . This suggests that the increased rDNA copy number causes the high level of E-pro transcription rather than deletion of RTT109 . To address this possibility , we measured the level of E-pro in the wild type strain with a higher rDNA copy number constructed by introduction of a plasmid-encoded RTT109 into the rtt109 mutant shown in Figure 2C . As the result ( Figure S3 ) , the strain with the longer rDNA had a high level of E-pro expression as well as the rtt109 mutant . Therefore , we conclude that Rtt109 is not involved in repression of E-pro transcription , but that the abnormally increased rDNA copy number leads to a failure to repress transcription . Moreover , we examined the effect of deletion of RTT109 on the association of cohesin with the IGS region of rDNA by chromatin immunoprecipitation ( ChIP ) assay . To exclude the indirect effect of increaed rDNA copy number , we compared cohesin ( Mcd1 ) association level between the rtt109 mutant and the wild type in a FOB1 defective background . As shown in Figure 5D , we confirmed cohesin association in the IGS in the mutant . Because the association represses unequal sister-chromatid recombination [7] , the hyper-amplification in the rtt109 mutant is most likely not induced by the recombination . It appears to be difficult for break-induced unequal sister-chromatid amplification at the standard level to achieve the high rate of amplification observed in the rtt109 mutant ( >100 copies/cell division ) . As mentioned above , we detected cohesin association that repressed the frequency of unequal sister-chromatid recombination . Rather , rolling circle type amplification appears to occur in such a manner that a large increase in copy number occurs in just a few cell divisions . In fact , the rolling circle structure was observed in the fraction remaining in the agrose gel plug subjected to pulsed-field gel electrophoresis ( Figure 6 ) . To initiate rolling circle replication , the broken end at the RFB must recombine with the intra-sister-chromatid ( Figure 7B ) or with popped-out circular molecules ( ERC , extra chromosomal rDNA circle , Figure 7C ) . In the former case , the recombination usually produces an ERC [4] . Therefore , we monitored ERC formation during hyper-amplification by 2D gel electrophoresis in which the non-linear DNA molecules are separated from genomic DNA . In the wild type control , ERCs were detected in a Fob1-dependent manner as previously described [24] , [25] . In contrast , in the rtt109 mutant , more ERCs were detected than in the wild type strain before FOB1 induction , while after induction , little increase in ERC level was observed ( Figure 8A and 8B ) . This suggests that Fob1-induced DSB does not change the ERC level , but promotes an alternative pathway , perhaps rolling circle replication . To test the possibility of a rolling circle amplification for which the ERC serves as a template ( Figure 7C ) , we transformed a URA3-plasmid containing an intact rDNA unit ( ERC-like plasmid ) into the rtt109 fob1 double mutant and induced FOB1 . As shown in Figure 8C by Southern hybridization with a URA3 probe , a signal was detected at the position of the hyper-amplified chromosome XII ( ChrXII hyp . ) in the rtt109 mutant . This indicates that the ERC ( in this case ERC-like plasmid ) was integrated into the rDNA during hyper-amplification in the mutant and the integration may explain little increase of ERC level after Fob1 induction ( Figure 8A and 8B ) . To monitor the status of integration , we performed 2D gel analysis as shown in Figure 8A . For this purpose , the rtt109 mutant was transformed with another URA3-plasmid ( IGS-plasmid ) containing only the IGS region , after which the DNA was analyzed by digestion with BglII . The emzyme cuts only chromosomal rDNA units because the IGS region lacks a BglII recognition site . If rolling circle integration were to occur with the plasmid , tandem arrays of the plasmid sequence would be expected to form , generating long fragments following BglII digestion ( Figure 8D , upper ) . In contrast , if the integration did not accompany rolling circle replication , a few integrated plasmid sequences would be expected . As shown in Figure 8D lower , signals for the plasmid-specific probe ( amp ) were observed both at positions corresponding to circular ( IGS plasmid ) and linear DNAs . Apparently , the copy number for the linear position in the rtt109 mutant is much higher than in the wild type strain ( Figure 8D ) . This indicates that in the rtt109 mutant , rolling circle replication occurs more efficiently than in wild type and contributes to hyper-amplification . It has been reported that some mutants missing transcription factors for rRNA genes ( Rrn9 , Rrn5 , Rrn6 , Uaf30 ) , acquired more copies of rDNA to compensate for less efficient transcription [26] , [27] . Therefore , we examined whether deletion of RTT109 decreased transcription of rRNA genes , resulting in compensatory hyper-amplification of rDNA . To exclude indirect effects leading to copy amplification , we compared nascent rRNA in the rtt109 mutant and in wild type in a FOB1-defective background , where hyper-amplification is repressed temporarily . As shown in Figure 9A and 9B , transcript abundance did not decrease in the rtt109 mutant . Therefore , the amplification is not due to transcriptional compensation . Further , we examined the transcription status of the rRNA genes in the rtt109 mutant by psoralen crosslinking [28] . In this assay , because the transcribed ( active ) rDNA copies have fewer nucleosomes , more psoralen can intercalate into the DNA and retard electrophoretic mobility . As shown in Figure 9C , the intensity of the lower band ( silent copies ) increased ∼3-fold in the rtt109 mutant , while that of the upper band ( active copies ) did not change significantly . This indicates that the hyper-amplification does not increase the number of copies with an open chromatin conformation .
The ribosomal RNA gene is the most abundant and unstable gene in cells . Abundance is maintained in spite of the instability through several mechanisms . Gene amplification after deleterious recombination is one . We found that modification of a histone represses rolling circle replication and maintains wild type rDNA copy number . Study of the relationship between histone modification , as a phenotype of the sir2 mutant , and rDNA stability was initiated by Gottlieb and Esposito ( 1989 ) [29] . They found that rDNA recombination was greatly enhanced in a sir2 mutant . Fritze et al . ( 1997 ) found that the mutation affected chromatin structure and speculated that it prevented recombination ( instability ) in the repeat [30] . Later , Sir2 was identified as a histone deacetylase [31] . Subsequently , we found that such alterations of chromatin structure repressed the noncoding promoter ( E-pro ) whose transcription reduces cohesion association with rDNA , and induces recombination between unequal sister-chromatids , because cohesin proteins hold sister-chromatids together and suppress unequal sister-chromatid recombination [7] . Therefore , one putative mechanism for rDNA maintenance by histone modification is through E-pro regulation . As reported here , however , deletion of RTT109 alone did not activate E-pro expression when copy number was normal ( Figure 5B ) . Instead , as the result of the hyper-amplification , E-pro transcription was elevated in the rtt109 mutant . This may be because the copy number is too high to be silenced transcriptionally . It is known that the shorter rDNA also induces a high level of E-pro and induces rDNA amplification [7] . In this case , the activation can be explained by self-regulation of SIR2 as follows . When rDNA copy number is reduced , a large amount of Sir2 is released from the rDNA and may silence other parts of the genome . In fact , Michell et al . reported that telomere silencing is enhanced in a low rDNA copy strain ( 2005 ) [32] . They also showed that SIR2 itself is a target for repression . Therefore , we propose that autoregulated less SIR2 expression induces more E-pro transcription to amplify rDNA in a low rDNA copy strain . As copy number increases , more Sir2 is consumed by interaction with rDNA and repression of SIR2 is reduced . When copy number reaches wild type levels , SIR2 expression is maximal which represses E-pro resulting in a halt to further amplification . Therefore , in the case of hyper-amplified rDNA , Sir2 levels may be too low to interact with all rDNA copies resulting in a net partial loss of E-pro repression . Hyper-amplification was observed in the mms22 mutant which has been shown to interact genetically with RTT109 ( Table 1 ) [33] . In both rtt109 and mms22 mutants , cells become more sensitive to DNA damage during S phase than in wild type . Both Rtt109 and Mms22 are also involved in regulation of homologous recombination induced by replication inhibition [14]–[16] , [34] . Our results show that the hyper amplification is accelerated by replication fork-blocking activity in the RFB . Therefore , one explanation is that the blocked fork induces DNA double stranded breaks and that recombinational repair of the broken sister-chromatid can be modulated by these proteins through histone modification . However , as shown in our 2D gels , no significant abnormality was observed in amount of recombination intermediates in the rtt109 mutant ( Figure 5A ) , e . g . , the early steps of recombination ( DSB , homologous annealing , strand invasion and formation of Holliday structure ) . This suggests that Rtt109 and Mms22 function in later steps of the recombination reaction . For example , after formation of the Holliday structure , the donor strand may not separate from the template strand , or , alternatively , replication , such as break-induced replication ( BIR ) may occur [35] . In case of repetitive sequences such as rDNA , this BIR initiates rolling circle replication which generates an enormous increase in copy number during the cell cycle ( Figure 7B ) . The pattern of integration of the IGS plasmid into the chromosomal region indicates that rolling circle replication occurs in the rtt109 mutant ( Figure 8D ) . However , integration of the ERC-like plasmid does not seem to be the sole or main contribution to rDNA amplification . The total signal intensities of both an integrated ERC-like plasmid into chromosome XII [Figure 8C , Chr XII ( hyp ) ] and intermediates of on-going rolling circle replication in the well were ∼15 times greater than that of single copy URA3 on Chromosome V ( Figure 8C; lanes 7–8 in the right pannel , Chr . V ) . Of course , because both the ERC-like plasmid and native ERC become templates for rolling circle replication , the actual extent of replication is expected to be greater . Intra-chromosomal recombination may also contribute to promote an additional pathway of rolling circle replication . As shown in Figure 7B , the broken end selects the intra-sister chromatid as a recombination template , instead of ERC , and converts into a Break-Induced Replication ( BIR ) event . Thus , we think that such template switch from sister-chromatid to ERC ( results in ERC integration ) and BIR conversion ( results in less ERC production ) reduce Fob1-induced ERC level during hyper-amlification in the rtt109 mutant ( Figure 8A and 8B ) . Our rolling circle replication model is supported by other reports concerning Rtt109 function . Wurtele et al . ( 2012 ) showed that the histone H3K56R mutation reduced marker loss rate in rDNA [36] . As we showed in Figure 2A , the mutation induced rDNA hyper-amplification which should have reduced loss of rDNA copies . Another study reported that mutated RTT109 reduced the frequency of unequal recombination between sister chromatids in non-rDNA regions [16] . This is consistent with our observation that cohesin associates with the IGS region of rDNA . This suggests that rDNA amplification by unequal sister-chromatid recombination ( Figure 7A ) does not seem elevated in the mutant . Houseley and Tollervey ( 2011 ) recently reported that hyper-amplification occurred in an asf1 mutant and proposed a model based on unequal sister chromatid interaction [17] . Because the rate of amplification is much lower in the asf1 mutant than in the rtt109 mutant , their model appears plausible . But in the case of the rtt109 mutant , models other than one based on rolling circle replication would not seem compatible with such a high amplification rate . In addition to the rtt109 mutant and the histone H3K56R mutant , we found that the double deletion mutant of histone deacetylase that acts on the K56 residue of histone H3 , Hst3 Hst4 , and the H3K56Q mutant all exhibited a hyper-amplification phenotype . These mutations also increased sensitivity to DNA damage relative to wild type cells [18] . Either an excess or an absence of acetylation on H3K56 of sister chromatids during recombination at the replication fork blocking site leads to rolling-circle amplification of rDNA . While we currently lack a satisfying explanation , the downstream event of the unusual H3K56 modification is important for the completion of recombination at the RFB in order to inhibit BIR-based rolling circle replication . Rolling circle replication is known to play important roles in cell physiology . The most well known example may be rDNA amplification during amphibian oogenesis [37] . In this stage , extra-chromosomal rolling circle amplification increases copy number up to 2–3 orders of magnitude resulting in a huge increase in rRNAs ( and ribosomes ) needed for early development [38] , [39] . This mechanism of replication is not limited to rDNA but occurs elsewhere in the genome and in some cases , helps cells adapt to various stresses ( for review , see [40] ) . In cancer cells , gene amplification is frequently observed and may be mediated in part by rolling circle replication [41] . Abnormal histone modification has also been reported in malignant tissue [42] consistent with the possibility that gene amplification plays a role in acceleratating tumorigenesis .
Yeasts strains used in this study were derived from NOY408-1b in the W303 genetic background and from strain S288c ( OpenBiosystems ) and are listed in Table S1 . The strains were grown at 30°C in YPD medium except for the hst3 hst4 double mutant that was cultured at 23°C . To monitor rDNA hyper-amplification , the rtt109 mutant with the plasmid containing galactose-inducible FOB1 ( YCplac22-Gal-FOB1 ) was precultured in synthetic complete medium minus tryptophan with raffinose ( SC-trp+raf ) . The strain was then transferred to SC-trp with galactose instead of raffinose ( at time zero ) and harvested at the indicated times ( 0 , 6 , 12 , 18 , 24 , 36 hr ) to prepare genomic DNA in agrose plugs [43] . As a control , the RTT109 strain was monitored in parallel . Raffinose ( 1/10 volume of a 10X raffinose stock ) and galactose were added just before use . Yeast strains with amino acid substitutions at histone K56 were generated by plasmid shuffling [44] . Genomic DNA was purified from ∼4 , 800 yeast mutants from the genome-wide deletion library ( OpenBiosystems ) as described [43] . They were then subjected to pulsed-field gel electrophoresis as described below . Gels were stained with ethidium bromide ( EtBr ) . The mutant strains with a longer chromosome XII than the 3 . 13 Mb size marker ( H . wingei ) were classified as hyper-amplified rDNA mutants ( Table 1 ) . We used a wide gel with a long comb ( 45 teeth ) to perform the electrophoresis . Three thin gels were stacked on one another in the electrophoresis tank permitting simultaneous running of ∼130 samples . Pulsed-field gel electrophoresis ( CHEF ) using CHEF Mapper XA ( BioRad ) and Southern hybridization were performed as described [25] , [43] . The conditions were a 300–900 sec pulse time and 100 V for 68 hrs at 14°C in a 0 . 8% agarose gel . To detect replication and recombination intermediates by 2D gel electrophoresis during hyper-amplification , DNA was prepared from cells ( YSI204 , YSI205 ) growing in SC-trp+Gal ( 9 hrs ) and SC-trp+Raf , and was digested with NheI in agarose plugs [43] . rDNA was detected with an rDNA-specific probe . DNA was isolated and digested in agarose plugs . To detect ERC during hyper-amplification by 2D gel electrophoresis , DNA was isolated from the cells before and 9 hours after release into SC–trp +Gal . To separate large DNAs between 5 and 50 kb by 1D gel electrophoresis , a 0 . 3% SeaKem Gold Agarose gel ( Lonza ) was used in TAE at 1 V/cm for 20 hours . After excising each lane above 5 kb from the 1D gel , 2D gel electrophoresis was performed in 1 . 0% SeaKem LE Agarose gel ( Lonza ) with EtBr at 6 V/cm for 5 hours at 4°C . The gel was blotted and part of the membrane containing DNAs longer than 7 kb was hybridized with an rDNA-specific probe . The other part of the membrane containing DNAs shorter than 7 kb was probed with URA3 . The ERC signal was normalized to the URA3 signal in the each blot . The relative amounts of ERC were determined from three independent experiments as shown in Figure 8B . To detect rolling circle amplification ( Figure 8D ) by 2D gel electrophoresis , cells with the IGS plasmid , YEplac195 containing the IGS region of rDNA were used . DNA was isolated from cells 48 hrs after release into SC–trp +Gal . Conditions used for 2D gel electrophoresis and induction of hyper-amplification were similar to those described above , except that DNA was digested with BglII before electrophoresis . BglII does not cut the plasmid but cuts the chromosomal rDNA , such that only plasmids integrated on the chromosome form linear structures following BglII digestion . The plasmid sequence was detected using a specific probe ( amp ) . Chromatin immunoprecipitation was performed as described [6] . DNA fraction remaining in the agrose gel plug after pulsed-field gel electrophoresis was extracted by crushing . AFM observation was performed as described [45] . During the sample preparation , DNA was sheared into smaller fragments ( <∼50 kb ) . Among these fragments ( ∼1 , 500 ) , we looked for the circle with the proper size ( 8–11 kb ) . Imaging was carried out with a Didital Instrument MultiMode AFM in the tapping mode . Image J ( NIH ) was used for measurement of DNA length . | Gene amplification is one of the major strategies used by cells to increase the abundance of gene products . We have been studying amplification of the ribosomal RNA genes cluster , also known as rDNA ( ribosomal DNA ) , in yeast and found that unequal sister-chromatid recombination increased copy number following accidental deleterious recombination events among the repeats . This amplification is highly regulated and ceases when the copy number reaches ∼150 . We isolated mutants , including rtt109 , which have abnormally high copy numbers of rDNA . RTT109 encodes an acetyl transferase that affects chromatin structure . In the mutant , rolling circle-type amplification that is observed in the early developmental stage in amphibians , oogenesis occured . We speculate that RTT109 plays a key role in regulating the mode of rDNA amplification . Variation in gene copy number ( amplification ) has been widely observed in a variety of organisms , contributing to both beneficial adaptation and pathology ( e . g . , cancer ) . Our results shed new light on molecular mechanisms of gene amplification . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbiology",
"dna",
"replication",
"dna",
"recombination",
"dna",
"chromosome",
"biology",
"dna",
"amplification",
"biology",
"molecular",
"biology",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"genomics",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2013 | Rtt109 Prevents Hyper-Amplification of Ribosomal RNA Genes through Histone Modification in Budding Yeast |
To investigate the DNA damage response , we undertook a genome-wide study in Saccharomyces cerevisiae and identified 86 gene deletions that lead to increased levels of spontaneous Rad52 foci in proliferating diploid cells . More than half of the genes are conserved across species ranging from yeast to humans . Along with genes involved in DNA replication , repair , and chromatin remodeling , we found 22 previously uncharacterized open reading frames . Analysis of recombination rates and synthetic genetic interactions with rad52Δ suggests that multiple mechanisms are responsible for elevated levels of spontaneous Rad52 foci , including increased production of recombinogenic lesions , sister chromatid recombination defects , and improper focus assembly/disassembly . Our cell biological approach demonstrates the diversity of processes that converge on homologous recombination , protect against spontaneous DNA damage , and facilitate efficient repair .
Homologous recombination ( HR ) , a repair mechanism that depends on DNA sequence homology , underlies a number of important DNA processes that act to both stabilize and diversify a genome . In mitotic cells , HR functions to maintain the integrity of the genome through such processes as the repair of DNA double-strand breaks ( DSBs ) , the maintenance of rDNA copy number , and the rescue of collapsed replication forks . HR is entirely conservative when it occurs following DNA replication where a sister chromatid is available as a template . However , utilization of sequences on a homologous chromosome can lead to crossovers and potential loss of heterozygosity ( LOH ) , while recombination at ectopic or repeated sequences may lead to genomic rearrangements such as deletions , duplications , and translocations ( reviewed in [1] ) . In Saccharomyces cerevisiae , Rad52 is the defining member of an epistasis group that includes: RecA homologs Rad51 , Rad55 , Rad57 , and Dmc1; putative SWI/SNF family ATPase Rad54; Rad52 homolog Rad59 and Mre11 , Xrs2 , and Rad50 . The Rad52 epistasis group is essential for HR . Rad52 binds single-stranded DNA in vitro and has been shown to stimulate DNA annealing and to enhance Rad51-catalyzed strand invasion [2–5] . In response to DNA damage , proteins involved in HR relocalize into discrete subnuclear foci . Fluorescently tagged repair and checkpoint proteins have been used to explore the composition and dynamics of these foci , which are giga-dalton-sized assemblies of proteins [6] . Repair foci colocalize with fluorescently tagged inducible DSB sites , regions of single-stranded DNA , and sites of unscheduled DNA synthesis [7–9] . Multiple DSBs often colocalize at a single focus showing that foci reflect recombination centers capable of the simultaneous repair of more than one lesion . The assembly of proteins into repair foci is a coordinated process beginning with detection of damage by the Mre11/Rad50/Xrs2 complex . Next , checkpoint proteins are bound and activated to arrest cell cycle progression until completion of repair . The lesion is repaired through HR performed by the Rad52 epistasis group proteins and finally the repair apparatus is disassembled [10] . From a cell biology perspective , Rad52 focus formation is an excellent marker for HR , since it is required for the recruitment of all other HR proteins into repair foci . While exogenous DNA damage greatly stimulates the formation of Rad52 foci , foci also form spontaneously in S phase cells , likely reflecting the repair of spontaneous DNA lesions such as DSBs , nicks , and single-stranded gaps [6 , 11] . Time-lapse microscopy indicates that foci form in approximately 50% of cells during S phase and most spontaneous foci persist for less than 10 min [7] . Since spontaneous foci generally last for only a fraction of S phase , they are observed in 20% of S phase cells in a population of logarithmically growing cells ( 5% of the total population ) . Mutants defective in various aspects of DNA metabolism , including damage checkpoints ( mec1 sml1 ) , HR ( rad51Δ ) , and DNA replication ( pol12–100 ) exhibit elevated levels of spontaneous foci [6] . This elevation may be the consequence of an increased incidence of focus formation reflecting the generation of more DNA lesions , or the consequence of foci that persist over time resulting from an alteration in the dynamics of focus assembly/disassembly . Here we report the results of a genome-wide screen designed to identify gene deletions that significantly alter levels of spontaneous Rad52 foci . The set of gene deletions identified includes many known genes involved in DNA and chromatin processes such as replication , repair , silencing , and chromosome segregation , as well as a number of other processes with no reported link to HR . In addition , 22 previously uncharacterized ORFs designated IRC2–11 , 13–16 , 18–25 ( Increased Recombination Centers ) were identified . Measurement of HR between sister chromatids and between homologous chromosomes in these focus mutants established four different classes demonstrating that several distinct mechanisms are involved in precipitating increased Rad52 foci . Furthermore , several IRC genes exhibit synthetic interactions with a rad52Δ allele suggesting a direct role in the maintenance of genomic integrity .
Rad52 , a central recombination protein , relocalizes to form sub-nuclear foci in response to spontaneous and induced DNA damage . Therefore , Rad52 focus levels can be used as a sensitive indicator for processes that impinge on the genome . An initial screen to identify gene deletions affecting the levels of spontaneous Rad52 foci was performed by transforming a plasmid containing a Rad52-YFP fusion gene directly into library haploid strains . However , this approach yielded an excessive number of false-positive results caused by additional recessive factors in the individual library strains . We therefore developed a method that permits the systematic creation of hybrid diploids that are homozygous for the gene deletion from each library strain ( Figure 1A ) , while simultaneously facilitating the introduction of plasmid or chromosomal constructs into the gene deletion strains [12] . Using the systematic hybrid LOH method , we screened 4 , 805 nonessential gene deletions for levels of spontaneous foci ( Figure 1B ) . Twenty gene deletions could not be constructed as hybrid homozygous diploids , including 11 that are deficient in mating . The distribution of focus levels is shown in Figure 2 . To evaluate the reliability of the mutant screen , we partitioned the genes into four sub-sets shown ( A , B , C , and D ) . Sub-set B consists of the strains that exhibit focus levels within the range of variation seen in a wild-type strain . These account for approximately 90% of the deletions screened . For 233 gene deletions ( sub-set A ) , no foci were observed . Upon retest , most of these strains formed foci; however , the distribution was shifted slightly to the left compared to the entire deletion library ( Figures 2 and S1A ) . Furthermore , for the 96 strains that showed the lowest levels upon retest ( less than 1% foci ) , all gave rise to foci after ionizing radiation , indicating that no fundamental aspect of Rad52 focus formation was disrupted in these deletion strains ( Figure S1B ) . Further study of this group of genes may shed some light on processes that generate spontaneous DNA damage . We next examined those mutants that exhibit foci in more than 20% of all cells ( sub-set D ) , a 4-fold elevation over average wild-type levels . These 108 gene deletions ( 2 . 2% of the library ) fall into a number of broad functional groups involved in various aspects of DNA and chromatin metabolism . Following two independent retests , 80 mutants ( 75% of the 108 ) consistently exhibited elevated focus levels including deletions of 16 previously uncharacterized ORFs ( Tables 1 and S1D ) . Subsequently , we examined the 144 deletions from the genome-wide screen that exhibited foci in 15%–20% of cells ( sub-set C ) . We suspected that interesting candidate genes would be found in this sub-set since it includes deletions of several genes directly associated with genomic integrity , including replication ( CTF18 , RAD27 ) , homologous repair ( MRE11 , RAD50 ) , chromatin remodeling ( HPA1 , SET2 , SNT1 , SWI6 ) , and cell cycle control ( CLB5 , RAD17 ) . To decide how much further to study this sub-set , we retested all 41 uncharacterized ORFs contained within it . We found that only nine ( 22% ) consistently exhibited elevated focus levels upon retesting ( Table S1C ) . Since only 22% of uncharacterized ORFs in this 15%–20% focus range maintained an elevated focus phenotype compared to 75% in the greater than 20% range , we concluded that the likelihood of identifying additional new genes in this sub-set was dramatically decreased . Therefore , we did not delve deeper into our candidate pool , and the nine new genes along with the 80 deletions exhibiting the highest levels of foci were selected for further analysis ( Table 1 ) . Lastly , we recognize that the screen produces some false negatives since several gene deletions expected to exhibit elevated focus levels ( e . g . , mre11Δ , xrs2Δ , rad50Δ , rad27Δ ) failed the cutoff for significance ( 16% , 13% , 9% , and 16% , respectively ) . In addition , the screen identified some gene deletions ( e . g . , ddc1Δ , rad57Δ , slx8Δ , mms1Δ , and rtt109Δ ) , but failed to identify other members of the same complexes ( e . g . , mec3Δ and rad17Δ , rad55Δ , slx5Δ , asf1Δ , and mms22Δ ) that were expected to be phenotypically equivalent . We then retested the other members individually and they consistently demonstrated elevated levels of spontaneous foci ( Table S1E ) indicating that the focus phenotype was shared among all members of each complex and that the mutants not identified in the initial screen were false negatives . We did not include these mutants in our subsequent analysis , as we studied only those gene deletions identified by our original screening criteria . Among the 89 deletions that exhibit elevated levels of spontaneous Rad52 foci were 25 previously uncharacterized ORFs . Eight of these IRC genes are listed as dubious ORFs on the Saccharomyces Genome Database due to the small size of the coding region . In addition , three IRC ORFs overlap other genes identified in the screen ( IRC1 and BDF2 , IRC15 and CTF19 , IRC17 and RTT103 ) , while several others overlap genes not identified in the screen . To verify that the focus phenotype observed in each IRC mutant was the result of disruption of that ORF , we performed a complementation test . Twenty-two of the 25 ircΔs were complemented by their corresponding ORF . However , a wild-type copy of the IRC ORF was unable to complement irc1Δ ( ydl071cΔ ) , irc12Δ ( yor024wΔ ) , and irc17Δ ( ydr290wΔ ) ( Figure S2A ) . Of these three , irc1Δ and irc17Δ remove sequences from overlapping genes that were also identified in the screen and their focus phenotype was complemented by a plasmid containing the neighboring gene ( Figures S2B and S2D ) . For irc12Δ the adjacent non-overlapping gene HST3 complemented the phenotype ( Figure S2C ) . We conclude that the elevated focus phenotype in these three strains was likely a consequence of mutation of the neighboring genes or regulatory sequences , thus reducing the number of mutant strains to 86 . Increased levels of Rad52 foci may indicate a dependency on RAD52 gene function . In fact , 14 of the 86 mutants previously showed a synthetic fitness defect or lethality when combined with a rad52Δ allele [13–19] . Synthetic genetic interactions between all 86 gene deletions and rad52Δ were assayed through a comprehensive quantitative analysis ( see Materials and Methods ) . The analysis revealed 27 synergistic interactions , 15 of which had not been previously described . All interactions observed resulted in synthetic growth defects , with no double mutants exhibiting improved growth compared to the single mutants . One gene deletion , nup60Δ , exhibited synthetic lethality with the rad52Δ allele . To distinguish between strong and weaker interactions , the remaining 26 were parsed into synergistic and additive growth defects ( see Materials and Methods ) . Growth differences revealed 17 synergistic and 10 additive interactions ( Table 2 ) . Finally , we are unable to reproduce the previously reported synthetic interactions of mus81Δ and rtt107Δ with the rad52Δ allele [18] . The 27 interactions described here suggest that the absence of these genes intensifies the requirement for HR in cell survival and growth . The mutants were parsed into functional classes based on their effect on homologous recombination . The increase in spontaneous Rad52 foci observed in the mutant strains may correspond to altered levels of spontaneous recombination . This hypothesis is underscored by the observation that a number of these mutants have been previously demonstrated to exhibit either elevated ( hpr1Δ ) or reduced ( rad51Δ ) mitotic recombination [20 , 21] . Each focus mutant was subjected to two heteroallelic recombination assays to sort them into functional classes corresponding to the mechanisms that lead to the accumulation of Rad52 foci ( Figure 3 ) . The direct repeat assay measures the rate of sister chromatid gene conversion and intrachromosomal single-strand annealing ( SSA ) events , while the interhomolog assay measures recombination between alleles on homologous chromosomes in diploids , specifically those recombination events that do not utilize the sister chromatid , which is the preferred template for HR [22] . Interhomolog recombination between heteroalleles is one underlying cause of LOH . The recombination rates for the two assays divide the set of mutants into four classes ( Tables 1 , 3 , and 4 ) . The three Class I mutants are in the RAD52 epistasis group and exhibit significantly reduced levels of gene conversion between both sister chromatids and homologous chromosomes but wild-type levels of SSA , consistent with previous studies [21 , 23] . The 14 gene deletions that exhibit increased rates of both direct repeat and interhomolog recombination define Class II . The 32 mutants in Class III exhibit increased rates of recombination specifically between homologous chromosomes , while sister chromatid recombination ( SCR ) in the direct repeat assay is unaffected . Surprisingly , 37 Class IV gene deletions fail to demonstrate any alteration in the rate of recombination with either assay . It is possible that these mutants affect recombination only at specific genomic regions . To begin to examine this notion , we tested 33 of the Class IV deletion strains using an assay that measures recombination in the multiple tandem rDNA array [24] . Elevated frequencies of recombination in the rDNA array compared to wild type ( 1 . 7 × 10−2 ± 0 . 7 ) are observed in four gene deletion strains ( bck1Δ 18 ×10−2 ± 2 . 2 , bdf1Δ 7 . 4 ×10−2 ± 2 . 9 , rtt107Δ 5 . 8 × 10−2 ± 2 . 3 , and trf4Δ 9 . 3 × 10−2 ± 3 . 7 ) , demonstrating that the focus phenotype may be triggered by events at specific loci .
Three members of the RAD52 epistasis group , rad51Δ , rad54Δ , and rad57Δ display elevated levels of Rad52-YFP foci and decreased levels of Rad51-dependent recombination ( this study and [25] ) . Rad51 , Rad54 , and Rad57 are recruited to a DNA lesion subsequent to Rad52 focus formation and their recruitment is dependent on Rad52 [25] . Deletion of genes that encode proteins that function downstream of Rad52 in HR leads to a failure to complete the recombination process . Thus , the increased focus levels in these mutants likely reflect the persistence of Rad52-YFP foci . Since SSA is dependent upon Rad52 , but not Rad51 or other later HR proteins , the rate of that process is not affected by these gene deletions [21 , 26] . Class I gene deletions do not exhibit synthetic interactions with rad52Δ , since they are in the same epistasis group . It is noteworthy that none of the newly defined IRC gene deletions fall into this class , indicating that these genes are not required for Rad51-dependent HR . Class II mutants exhibit elevated levels of recombination between sister chromatids and between homologous chromosomes likely reflecting an increase in the generation of spontaneous DNA lesions requiring repair via HR . Thus , the Class II focus phenotype reflects an increase in the overall frequency of formation of Rad52-YFP foci in these mutants . This class includes a number of genes with well-characterized roles in the maintenance of genomic integrity and the suppression of spontaneous DNA damage ( SGS1 , RMI1 , ELG1 , MLH1 , HPR1 , and SLX8 ) , as well as all three nuclear pore genes identified in the study ( NUP60 , NUP133 , and POM152 ) [27–35] . The Class II recombination phenotype is exhibited by two mitochondrial genes ( COX16 and MRPI17 ) , which suggests that an increase in oxidative damage ( e . g . , reactive oxygen species ) stimulates spontaneous DNA lesions in these mutants [36–38] . Among the Class II mutants , ten of 14 , including the previously uncharacterized gene IRC8 , sensitize cells to the absence of a functional RAD52 allele . Such a synthetic defect is consistent with an increased requirement for Rad52-mediated HR in these gene deletion strains as a result of the increased generation of spontaneous lesions . Class III mutants exhibit elevated levels of recombination between homologous chromosomes in diploid cells , but wild-type levels of SCR in haploids . It is unlikely that the Rad52 focus phenotype observed in this class reflects a general increase in endogenous DNA damage , but rather that spontaneous lesions that do arise are processed differently in these mutants . In diploid cells , lesions are preferentially repaired using the sister chromatid as template , effectively preventing LOH that can result from repair from the homologous chromosome . In Class III mutants , the 2- to 20-fold elevation in interhomolog recombination occurs without a concomitant increase in SCR . We suggest that the increased utilization of the homolog for repair reflects a defect in the efficiency of SCR . However , reduced SCR in these mutants is not a result of a defect in sister chromatid cohesion since only one Class III mutant , ctf4Δ , exhibited precocious sister chromatid separation ( [39] and unpublished data ) . The increase in interhomolog recombination observed in Class III mutants may reflect a defect in the recombination center itself . Null mutants of members of the Mre11-Rad50-Xrs2 complex and certain separation-of-function alleles of rad52 exhibit elevated levels of spontaneous Rad52-YFP foci and increased interhomolog recombination , identical to the phenotype of our Class III [11 , 40] . Class III includes three genes ( DDC1 , RAD59 , and SAE2 ) that encode proteins that themselves localize to repair foci in response to DNA damage [25] . In addition , Mus81 protein functions downstream of Rad52 in the resolution of recombination intermediates [41 , 42] . Similar to Class I , these four mutants do not exhibit synthetic growth interactions with rad52Δ as they function in the same pathway . Class III also includes deletions of a large number of genes with roles in chromatin modification and remodeling ( AHC1 , ESC2 , LRS4 , RRM3 , VPS71-VPS72 , MMS1-RTT101-RTT109 , and HST3 ) [20 , 43–49] . Modified chromatin at the site of damage may itself function as a scaffold for the recruitment and assembly of the repair machinery . Remodeling tightly condensed chromatin is critical to allow repair proteins access to both the damaged DNA and the homologous template [50 , 51] . Perhaps chromatin defects delay use of the sister chromatid as the repair template . The delay , observed here as persistent Rad52 foci , may increase the use of the homologous chromosome for repair . Since the majority of the previously characterized Class III genes function in DNA replication , repair , and chromatin dynamics , it is likely that the six newly identified IRC genes in this class ( IRC4 , IRC5 , IRC7 , IRC9 , IRC14 , and IRC19 ) are also involved in DNA metabolism . Like six of the 12 Class III genes implicated in chromatin metabolism described above , irc5Δ , irc14Δ , and irc19Δ exhibit synthetic growth defects with rad52Δ , suggesting related roles for these genes . The irc19Δ mutant also exhibits synthetic defects with other genes associated with the maintenance of genome integrity including genes involved in replication and HR [18] . IRC5 encodes a putative Snf2 family DNA helicase with homology to the mammalian lymphoid-specific helicase HELLS , supporting a role in chromatin remodeling [46 , 52] . Genetic interactions of the irc5Δ mutant correlate strongly with those of replicative proteins ( rad27Δ , elg1Δ , rnh201Δ , pol30–79 , rfc4-DAMP , rfc5-DAMP ) , suggesting that Irc5 protein remodeling activity may be involved in DNA replication ( N . Krogan , personal communication ) . Class IV gene deletions do not affect recombination specifically at the LEU2 locus . Since it is possible that recombination is affected in other regions of the genome , we measured recombination rates within the rDNA multiple tandem array . We chose to examine this array next because it is a highly organized region where recombination is tightly regulated [53] . We identified four Class IV mutants ( bck1Δ , bdf1Δ , rtt107Δ , and trf4Δ ) that exhibit rDNA hyper-recombination . Similar to most of the mutants in Class II , three of the four ( bck1Δ , bdf1Δ , and trf4Δ ) show synthetic interactions with rad52Δ that is consistent with their potential roles in the suppression of genetic instability within this array . It is tempting to speculate that IRC13 , BUB2 , and RTT103 , the three other Class IV genes that show synthetic interactions with rad52Δ may also be involved in the suppression of region-specific damage . Other regions of the genome that require specific factors for maintenance are telomeres . For example , defects in telomere capping may lead to increased recognition of telomeres as DSB ends that would recruit Rad52 . Interestingly , irc6Δ may have such a defect ( W . Zhang and D . Durocher , personal communication ) . Alternatively , increased levels of Rad52 foci may occur without measurable effects on the products of recombination . For example , the focus phenotype may be due to slowing of the HR process or delaying the disassembly of foci without changing the outcome . In addition , a gene deletion may generate unrepairable lesions that could result in the accumulation of repair proteins into foci but lead to cell lethality . For the spindle assembly checkpoint mutants , mad1Δ , mad2Δ , mad3Δ , and bub2Δ , an increase in the number of cells with foci in G1 was observed ( unpublished data ) . Mitotic division before resolution of a repair focus formed during the previous round of DNA replication could result in two G1 cells with potentially unresolvable foci [54 , 55] . The unique phenotype of the Class IV mutants underscores the utility of taking a cell biological approach to investigate HR . Examination of intermediate steps in the process permits the identification of genes that would not have been found using assays requiring a measurable recombination product . The large number of chromatin remodeling , mitochondrial , and spindle checkpoint genes found in Class IV suggests that some of the 12 IRC genes in this class may be involved in these processes . For example , IRC3 encodes a putative DEAD/DEAH box helicase that localizes to mitochondria and exhibits synthetic lethal interactions with deletions of spindle assembly checkpoint proteins , histones , and proteins associated with sister chromatin cohesion . Furthermore , IRC20 encodes a putative Snf2/Swi2 family helicase , which localizes to nuclei and mitochondria and is implicated in transcriptional regulation , while IRC21 is predicted to function in chromatin remodeling [14 , 38 , 46 , 56–58] ( G . Prelich , personal communication ) . Overall , the 86 genes identified in the Rad52-YFP focus screen are largely conserved throughout eukaryotic evolution , with 49 having homologs in nearly every sequenced eukaryotic species ( Table S4 ) . Homologs for another 15 have been found in evolutionarily divergent yeast species including Schizosaccharomyces pombe and Candida albicans , but not in mammals , Drosophila melanogaster , or Caenorhabditis elegans . The remaining 22 are found in the closely related sensu stricto yeast species . Seven IRC genes have homologs identified across eukaryota . This includes three putative helicases , IRC3 , IRC5 , and IRC20 . Other IRC homologs have been linked to human diseases , including the IRC7 homolog CTH , which encodes a cystathionase implicated in premature births and cancers , and the IRC24 homolog SPR , which contains polymorphisms associated with Parkinson disease [59–61] . IRC21 has homology to NADPH cytochrome B5 oxidoreductase , which is linked to insulin-dependent diabetes in mice , while IRC15 resembles mammalian mitochondrial dihydrolipoamide dehydrogenase , associated with a number of human diseases including Alzheimer disease [62 , 63] . The Rad52-YFP focus screen described here applies systematic hybrid LOH in a genome-wide cell biological search to identify proteins involved in diverse pathways contributing to genomic integrity . This assay permitted the identification of a large set of gene deletions that affect the incidence and dynamics of HR foci in living cells regardless of the genetic outcome . In particular , we uncovered 22 previously uncharacterized ORFs , many having only subtle effects on the process of recombination , which prevented their identification in other screens . Since recombination is a multi-step process , it will be of interest to examine the dynamics of other HR factors within our mutant set and the genome as a whole . Additionally , it will be important to show that the conserved genes in other organisms play similar roles in this process .
Individual deletions of nonessential genes made in the BY4742 and BY4739 MATα strains were obtained from the Saccharomyces Gene Deletion Project [64] . The conditional chromosome strains used to create the hybrid LOH strains and the method of inducing LOH in these strains have been described [12] . Briefly , the systematic hybrid LOH method utilizes a set of 16 strains , each containing a conditional centromere construct on one of the 16 yeast chromosomes . Mating each gene deletion strain to the appropriate conditional centromere strain creates hybrid diploids that are transiently heterozygous for the gene deletion . After chromosome loss is induced for the conditional chromosome , homozygosis of the marked gene deletion occurs more than 95% of the time as a result of endoduplication of the monosomic chromosome . This method allows the introduction of any plasmid or chromosomal construct into individual mutants of the gene deletion library through mating rather than individual transformations . Strain BY4742 was used as the wild-type control for the library background . Gene deletions from the library were backcrossed into the W303 background to create congenic strains . Gene deletion strains assayed for direct repeat recombination are congenic ( minimum of four backcrosses ) to W5880-3A ( MATα ADE2 leu2-ΔEcoRI::URA3::leu2-ΔBstEII TRP1 lys2Δ RAD5 ) . Homozygous diploid deletion strains tested for interhomolog recombination are congenic ( minimum of five backcrosses ) to W5309 ( MATa/α ADE2/ADE2 TRP1/trp1–1 LYS2/lys2Δ leu2-ΔEcoRI/leu2-ΔBstEII ) . Deletion strains assayed for marker loss in the rDNA array are congenic ( minimum of six backcrosses ) to W6921-5A ( MATα ade2–1 can1–100 TRP1 LYS2 rDNA:ADE2-CAN1 ) . The pWJ1314 plasmid , which expresses Rad52-YFP from the native RAD52 promoter , was used for all focus measurements performed using library gene deletion strains [12] . Plasmids used for complementation analysis of irc mutants were built utilizing plasmid pWJ1250 , designed to facilitate cloning via HR in yeast . To construct plasmid pWJ1250 , overlapping primers C-D-TOP ( AGCGAGGTCGACTAGGGATAACAGGGTAATCCGCTGCTAGGCGCGCCGTGGTTAACGTCAG ) and C-D-BOTTOM ( CAGCTGGAGCTCTAGGGATAACAGGGTAATGCAGGGATGCGGCCGCTGACGTTAACCAC ) were fused and amplified by PCR with primers C-D-Forward ( AGCGAGGTCGACTAGG ) and C-D-Reverse ( CAGCTGGAGCTCTAGG ) . This reaction resulted in a synthetic DNA containing common adaptamer sequences C and D separated by the restriction site for the blunt-cutting HpaI enzyme . The C-D cassette is flanked by 17 base pair recognition sites for the I-SceI meganuclease , and then flanked by SacI and SalI sites . The amplified DNA was digested with SacI and SalI for cloning into the yeast shuttle vector pRS416 [65] . Thus , DNA amplified with C- and D-containing adaptamers can be cloned into the HpaI-linearized pWJ1250 by transformation into yeast . Cloned cassettes can be recovered by I-SceI digestion of the resulting plasmid . For each IRC and adjacent gene cloned for complementation analysis , C- and D-containing adaptamer primers were selected 200–300 bp upstream and downstream of the ORF . PCR products for each gene were amplified from a wild-type yeast strain and cotransformed into yeast with HpaI-linearized pWJ1250 . The resulting gap-repaired plasmids were recovered from Escherichia coli and transformed into the gene deletion strains congenic ( minimum of six backcrosses ) to W3749-14C ( MATa ADE2 bar1::LEU2 trp1–1 LYS2 RAD52-YFP ) along with plasmid pWJ1250 as a control . Complementation was performed by comparing levels of spontaneous Rad52-YFP foci in each gene deletion strain when transformed with the vector containing the deleted gene and with the empty vector as a control ( Figure S2 ) . In the focus screen described , 108 mutants were initially identified with Rad52-YFP foci in greater than 20% of cells examined . Of these 108 , 80 gene deletions maintained the focus phenotype following repeat experiments and were selected for further analysis . After retesting all uncharacterized ORFs with focus levels between 15%–20% ( 41 ) , nine consistently exhibited elevated foci and were added to the set of deletions . Three hypothetical ORFs were removed from the set after expression of the wild-type gene failed to complement the focus phenotype in the gene deletion strain . The remaining 86 gene deletions were prepared for the additional assays described below . Examination of Rad52-YFP focus levels by microscopy was performed as previously described [66] . Briefly , cells were grown overnight in SC-Leu media at 23 °C and exponentially growing cultures were prepared for microscopy . A single DIC image and 11 YFP images obtained at 0 . 3-μm intervals along the z-axis were captured for each frame , and Rad52-YFP foci were counted by inspecting all focal planes intersecting each cell . For each gene deletion strain in the screen , 200–400 cells were scored for Rad52-YFP foci . All Rad52-YFP focus data are presented in Figures 2 , S1 , and S2 , and Table S1 . We took advantage of the existing Gene Ontology ( GO ) annotations to determine whether our set of 86 mutants was enriched for any particular categories of genes . We utilized the Fisher's Exact Test to compute statistically significant enrichment of GO categories within the 86 mutants compared to those among the complete set of 4 , 805 mutants assayed for Rad52-YFP foci . The set of focus mutants exhibited enrichment in the GO component category for proteins localized to the nucleus ( p-value = 3 . 2 × 10−6 ) , as well as the GO biological process categories for DNA metabolism and response to stress ( p-value = 1 . 3 × 10−4 and 6 . 0 ×10−4 ) . The Bonferroni corrected threshold for significance among GO components was 2 . 0 × 10−3 and among GO processes was 1 . 5 × 10−3 . Synthetic interactions between library gene deletions and rad52Δ were determined on the basis of spore colony size following tetrad dissection . In the initial analysis , all 86 library gene deletion strains were mated to W303 background strain W3777-17A ( MATa ADE2 bar1::LEU2 trp1–1 LYS2 rad52::HIS5 RAD5 ) and sporulated . Twenty-four tetrads were dissected for each cross , and dissection plates were scanned using Adobe Photoshop ( Adobe Systems ) . Colony size was measured for each cross using a macro ( Y . Deng , unpublished data ) written for ImageJ software ( W . Rasband , National Institutes of Health ) , followed by genotyping for segregating alleles . Average colony size and standard deviation were derived in each cross for the wild type , gene deletion , rad52Δ , and the double mutants , and normalized to the mean value for the wild-type segregants in the cross . For this study , we defined two different classes of synthetic interactions for growth . Additive interactions are defined when the average colony size for the double mutant is significantly less than the colony size of either single mutant . Synergistic interactions are defined when the average colony size of the double mutant is significantly less than the product of the normalized colony size values of either single mutant . The synthetic interactions reported here were verified in a more closely related W303 genetic background by performing a second trial with congenic deletion strains ( minimum four backcrosses ) and W3777-17A ( rad52Δ ) . All data for synthetic interactions are presented in Table S2 . Spontaneous mitotic recombination between leu2-ΔEcoRI and leu2-ΔBstEII heteroalleles was measured between sister chromatids in haploid strains and between homologous chromosomes in diploid strains as previously described [67 , 68] ( Figure 3 ) . Rates of mitotic recombination were calculated as described by Lea and Coulson [69] . For each gene deletion mutant , eight independent trials were performed for each assay . Two-tailed t tests were applied to determine significant changes in recombination rates in gene deletion strains . Recombination rates measured for all mutants in both assays are presented in Table S3 . Frequencies of marker loss in the rDNA array was determined using a modification of a described method [24] . | Homologous recombination ( HR ) is a cellular process that permits efficient repair of both endogenous and exogenous DNA damage . Although the principal players in HR have been well characterized , the interplay of diverse processes with the HR pathway remains mysterious . Traditionally , genetic screens investigating HR have utilized genetic assays , such as survival following exposure to DNA damaging agents or alterations in the rate of the generation of recombinant products . In this work , we instead utilize a cell biology phenotype , the relocalization of the central HR protein Rad52 into subnuclear foci reflecting repair centers actively engaged in HR . This approach allows us to identify mutants that affect the kinetics of HR repair center assembly and disassembly regardless of the outcome of recombination . We identified 86 gene deletions that lead to increases in the levels of spontaneous foci in proliferating diploid cells , 22 of which were deletions of previously uncharacterized ORFs ( designated IRC2–11 , 13–16 , 18–25 ) . Genetic characterization of the mutants revealed a diversity of mechanisms that underlie the focus phenotype . These include increasing the generation of DNA lesions , blocking the completion of HR , and altering the kinetics of genetic recombination and the assembly/disassembly of the HR protein complexes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology",
"genetics",
"and",
"genomics",
"molecular",
"biology",
"saccharomyces"
] | 2007 | Genome-Wide Analysis of Rad52 Foci Reveals Diverse Mechanisms Impacting Recombination |
The transcriptome , as the pool of all transcribed elements in a given cell , is regulated by the interaction between different molecular levels , involving epigenetic , transcriptional , and post-transcriptional mechanisms . However , many previous studies investigated each of these levels individually , and little is known about their interdependency . We present a systems biology study integrating mRNA profiles with DNA–binding events of key cardiac transcription factors ( Gata4 , Mef2a , Nkx2 . 5 , and Srf ) , activating histone modifications ( H3ac , H4ac , H3K4me2 , and H3K4me3 ) , and microRNA profiles obtained in wild-type and RNAi–mediated knockdown . Finally , we confirmed conclusions primarily obtained in cardiomyocyte cell culture in a time-course of cardiac maturation in mouse around birth . We provide insights into the combinatorial regulation by cardiac transcription factors and show that they can partially compensate each other's function . Genes regulated by multiple transcription factors are less likely differentially expressed in RNAi knockdown of one respective factor . In addition to the analysis of the individual transcription factors , we found that histone 3 acetylation correlates with Srf- and Gata4-dependent gene expression and is complementarily reduced in cardiac Srf knockdown . Further , we found that altered microRNA expression in Srf knockdown potentially explains up to 45% of indirect mRNA targets . Considering all three levels of regulation , we present an Srf-centered transcription network providing on a single-gene level insights into the regulatory circuits establishing respective mRNA profiles . In summary , we show the combinatorial contribution of four DNA–binding transcription factors in regulating the cardiac transcriptome and provide evidence that histone modifications and microRNAs modulate their functional consequence . This opens a new perspective to understand heart development and the complexity cardiovascular disorders .
It is long known that an evolutionary conserved orchestra of transcription factors controls cardiac development and function . More recently the contribution of epigenetic and post-transcriptional mechanisms has been identified . Many successful studies have focused on these different aspects independently , but it is still an open question , how these molecular regulatory mechanisms interact . The ability of transcription factor binding to DNA is highly influenced by the chromatin status and epigenetic mechanisms play an important role in establishing and maintaining transcriptional programs . This layer of control comprises posttranslational modification of histones , DNA methylation and chromatin remodeling . To understand networks directing gene expression , the interplay between different transcription and epigenetic factors has to be considered . Furthermore , recent studies have started to unveil powerful roles for microRNAs ( miRNAs ) in regulating and fine-tuning mRNA profiles via either translational repression or mRNA degradation . It should be noted that gene expression profiles as obtained by standard microarrays or next-generation sequencing reflect mRNA profiles , which depend on the gene transcription as well as the decay of mRNA . In line with this , we present a study integrating mRNA profiles with transcription factor-DNA interaction data , histone modification marks and posttranscriptional regulation by miRNAs . The DNA-binding transcription factors Gata4 , Mef2a , Nkx2 . 5 and Srf play pivotal roles for the differentiation , maturation and homeostasis of cardiomyocytes . Mice lacking Gata4 die at E8 . 5 with failure of ventral morphogenesis and heart tube formation [1] , [2] . Targeted disruption of Nkx2 . 5 leads to abnormal heart morphogenesis with lethality at E9 . 5 [3] . Mef2a knockout mice die within the first postnatal week and exhibit myofibril fragmentation and impaired myocyte differentiation [4] . Srf-null mice show severe defects in the contractile apparatus of cardiomyocytes and die at the gastrulation stage [5] , [6] . However , Gata4 , Mef2a , Nkx2 . 5 and Srf not only display independently a central role for cardiac development and function , they also regulate each other's expression [7]–[10] . Despite their impact , we still have limited understanding of the global cardiac transcription networks driven by these factors in a direct and indirect manner . Moreover , we lack knowledge to which extent epigenetic marks such as histone modifications interfere with the regulation of downstream targets . The N-terminal histone tails serve as targets for a variety of reversible posttranslational modifications including acetylations or methylations . Both have a main impact on chromatin structure and represent binding sites for transcriptional regulators [11]–[13] , thus promoting or inhibiting transcription . They are put in place by specific enzyme families and are removed by others [14] , [15] . Hence histone acetylation is a dynamic process and for instance , mice lacking histone deacetylases 5 and 9 ( HDAC5 and 9 ) show cardiac defects typical for abnormalities in growth and maturation of cardiomyocytes [16] . A superactivation of Mef2 based on its interaction with HDACs is proposed [17] . However , our understanding of the underlying molecular mechanisms is still premature . It was reported that only 5–15% of differentially expressed genes in short interference RNA ( siRNA ) knockdown experiments are also direct transcription factor targets identified by chromatin immunoprecipitation ( ChIP ) and vice versa [18]–[23] . There is considerable evidence , that cascades of transcriptional regulators form networks . Loss of one factor will directly affect a few genes , but among those genes are other transcriptional regulators whose function is now altered , affecting a further set of downstream genes . In addition , it has to be considered that a significant proportion of downstream effects mediated by DNA-binding transcription factors is promoted via downstream miRNAs or other regulators . For example Srf regulates the transcription of miRNAs such as the smooth muscle relevant miR-143 and miR-145 [24] . Feedback loops between Srf/Mef2 and muscle-specific miR-133/miR-1 have been described , and both miRNAs are expressed throughout heart development and play important roles in muscle proliferation and differentiation [25]–[28] . Furthermore , miR-1 promotes myogenesis by targeting HDAC4 [26] , a transcriptional repressor of muscle gene expression , and thus represents an interface to histone acetylation . Taken the above , we investigated the transcription network driven by Gata4 , Mef2a , Nkx2 . 5 and Srf in cardiomyocytes in a genome-wide approach . First , we focused on the direct downstream targets by evaluating in-vivo DNA-binding sites of the respective factors and correlated these binding events with the expression level of related genes . Second , we investigated the functional consequence of the proposed regulation in knockdown experiments and built respective transcription networks . Third , we analyzed if co-occurrence with activating histone modifications could impact on gene expression levels of direct targets . Fourth , we studied the modulation of mRNA levels by miRNA alterations seen in knockdown experiments . Finally , we integrated the three levels regulating mRNA profiles and generated a comprehensive transcription network centered on Srf . Based on our analysis we argue that transcription networks have a comparable dependency on transcription factor binding , modulation by histone modifications as well as regulation by miRNAs . In addition to the global perspective , our networks provide distinct information on the regulation of individual genes especially with regard to cardiac function .
We developed a custom two-array set ( 2×385K ) with NimbleGen using a tilling approach for promoter and enhancer regions ( 10kb upstream ) , and first exon and intron sequences of 12 , 625 transcripts ( 65% of all RefSeq promoters ) . This enabled the analysis of sequence regions beyond standard promoter arrays . We selected 89Mbp of the mouse genome related to transcripts of 13 data sources ( Table S1 ) , which included all known expressed skeletal , smooth and cardiac muscle genes in human and mouse . Using ChIP-chip analyses we identified several hundreds of transcription factor binding sites ( TFBS ) for Gata4 ( 447 ) , Mef2a ( 999 ) , Nkx2 . 5 ( 383 ) and Srf ( 1 , 335 ) in mouse HL-1 cardiomyocytes , which were related to 345 Gata4 , 701 Mef2a , 276 Nkx2 . 5 and 1 , 150 Srf target genes ( Table S15 ) . Figure 1B shows the distribution of observed binding sites relative to the transcriptional start site ( TSS ) . An average of 24% of TFBS were localized in potential enhancer regions with a distance between 2 . 5kb to 10kb upstream from any transcriptional start site . The respective target genes of the studied transcription factors included 42 known targets ( Table S2 ) , substantiating the reliability of the system . In addition , we found several genes previously shown to be deregulated in mutants as direct targets of the respective transcription factor . For example , Gata4 and Nkx2 . 5 levels are decreased in cells depleted of Mef2 [8] and both genes show binding of Mef2a at their promoters in our data . To gain further insights into the transcription factor functionality , we investigated which Gene Ontology ( GO ) terms were significantly overrepresented among target genes of each factor when compared to all genes represented on the array . The significant GO terms ( p<0 . 001 ) show a stronger than expected association with heart development and function and are highly related to the phenotypes reported for the respective transcription factor ( Tables S3 , S4 , S5 , S6 ) . For example , the GO terms ‘muscle cell differentiation’ and ‘heart looping’ are significantly overrepresented among Mef2a and Nkx2 . 5 targets , respectively , and both are key features of corresponding knockout mouse models [3] , [4] . We investigated the sequences underlying the transcription factor binding sites in more detail and searched for TRANSFAC [29] motifs within the presumably bound sequences . The TRANSFAC matrices used for motif search are listed in Table S7 . For Gata4 , Nkx2 . 5 and Mef2a 84–94% of all ChIP binding events harbored respective binding motifs . For Srf , the fraction of ChIP binding events with predicted motifs was very small ( 169 out of 1 , 335 binding sites ) . However , Srf is well-known to bind the CArG-box CC ( A/T ) 6GG [30] , which is only partially represented by TRANSFAC motifs . Using a pattern matching approach we found the CArG-box in 1 , 063 ( approximately 80% ) Srf binding events . Furthermore , more than every second binding event of the studied factors occurred at sequence sites containing at least two times the respective transcription factor motif or pattern . We studied the cross-species conservation of binding motifs and found in ∼10% complete sequence conservation between human and mouse . In 27% the binding sites were localized in regions conserved across 18 vertebrate species based on PhastCons elements [31] . Thus , by focusing only on conserved sequence regions a priori more than two-third of the binding sites would be missed . The investigated transcription factors are known to co-regulate targets and pairwise physical interaction has been described between several of these factors [32] , [33] . Nevertheless , it is unknown how frequently this co-binding occurs in-vivo . Consequently , we investigated the assignment of Gata4 , Nkx2 . 5 , Mef2a and Srf to the same gene . We observed frequent co-regulation by more than one transcription factor , where Gata4 and Nkx2 . 5 shared 143 targets ( 41% and 52% , respectively ) and Mef2a and Srf shared 320 target genes ( 46% and 28% , respectively ) ( Figure 1C and 1D ) . For 91 genes co-binding of all four transcription factors was observed and in 85 cases the binding was observed at close proximity within a 500bp window . These data underline the complexity and cooperative regulation of gene regulation shown in our model of four DNA-binding transcription factors . We investigated whether the transcription factors act mainly as activators or repressors in a wildtype situation . We carried out genome-wide expression array analysis of the contracting HL-1 cardiomyocytes and classified all transcripts as expressed or non-expressed . We found that for each of the four transcription factors approximately 80% of the target genes were expressed and their expression levels were significantly enhanced compared to non-targets ( p<0 . 005 ) . Considering the cooperative co-binding of the investigated transcription factors , we were interested in the functional consequence of significantly reducing the quantity of each of the factors . Therefore , we used siRNA technique to reduce the protein levels of investigated transcription factors by more than 70% and studied its consequence for gene transcription . The reduction at mRNA and protein level was monitored by quantitative PCR and Western Blot analysis ( Figure S1 and Figure 2A ) and the genome-wide effects on transcript levels were measured by expression array analysis ( Table S17 ) . All data were based on a total of 4 replicate experiments using duplicates of two different siRNAs per transcription factor . The majority of deregulated transcripts were downregulated in the siRNA treated samples , confirming a primarily activating function of the transcription factors . Performing Annexin assays and Tryptophan Blue staining , we observed an increased apoptosis and cell death in particular when Gata4 or Srf were knocked down ( Figure S2 ) , which is in line with previous data [34]–[36] . Analogous to our analyses of common downstream targets based on transcription factor binding events , we studied common differentially expressed genes . The different factors share a comparable proportion of differentially expressed genes when knocked down ( Figure 2B ) . An opposing effect for targets regulated by several transcription factors was only observed in two cases: Myocd and Tpm1 . Figure 2C shows the combinatorial regulation of a selection of heart and muscle relevant , directly bound and differentially expressed genes . This includes genes coding for structural proteins like Actc1 , Actn2 , Tnnt2 , Mybpc3 or Myh6; growth factors like Igf1 or apoptosis factors like Casp3 . The transcription factor Tbx20 represents an example for a gene that is bound and regulated by all four factors . A broad panel of differentially expressed genes was further confirmed by quantitative real-time PCR ( Table S8 ) . Finally , we compared the differentially expressed genes in siRNA knockdown experiments to the direct target genes identified by ChIP . Analyzing the overlap focusing on the functional role of the respective genes , we found that both datasets share the Gene Ontology terms reflecting heart and muscle development and function . For example , muscle cell differentiation , muscle contraction and heart development are the main functional roles for direct targets of Srf as well as the respective differentially expressed genes . However , only a small fraction of direct target genes ( ∼10% ) was also differentially expressed , pointing to the combinatorial nature of gene regulation . In accordance , we found that genes bound by multiple transcription factors were significantly less likely differentially expressed ( χ2-test p<0 . 001 ) . Likewise transcription factors having a high number of common binding targets share only a small number of co-regulated genes in RNAi knockdown ( correlation shown in Figure 2B is inverse to the correlation in Figure 1D ) . In addition , binding in a poised state or buffering by epigenetic mechanisms such as histone modifications which interfere with the accessibility of the DNA should be considered . It has to be kept in mind that transcription factor binding depends on binding affinity and accessibility of binding sites . The regulatory potential of several factors has been reported to be strongly dosage dependent ( e . g . Tbx5 [37] and Gata4 [38] ) . Furthermore , a significant proportion of differentially expressed genes in RNAi are likely to be regulated in an indirect manner . Recent studies show the powerful roles for miRNAs in controlling mRNA profiles largely by silencing target genes , via either translational repression or mRNA degradation . To explore the influence of histone modifications as an epigenetic mechanism to modulate gene expression , we analyzed our transcription factor binding data in the context of co-occurring histone marks . In a previous study we investigated the localization of four histone modifications , which are known to promote an open chromatin state ( H3K9K14ac , H4K5K8K12K16ac , H3K4me2 and H3K4me3 ) [39] . We found that ∼80% of the respective transcription factor binding events are marked by one or more of these histone modifications , whereas in a randomized simulation only 23% are expected to co-occur ( Figure S3 ) . We consequently investigated whether the presence of any of these marks correlates with higher expression levels of direct target genes and found a significant impact for histone 3 acetylation ( H3ac ) only ( Figure 3A and Figure S4 ) . For Nkx2 . 5 and Mef2a the expression levels of direct targets were significantly higher than the reference group , independent of whether H3ac was present or not . Genes showing neither transcription factor binding nor H3ac were used as a reference . In case of Gata4 and Srf the expression levels of direct targets were only significantly increased when binding sites were additionally marked by H3ac . The enhanced expression levels depending on H3ac co-occurrence is further depicted in Figure 3B , which shows confirmation experiments of nine genes using quantitative PCR . In conclusion , our data provide evidence that acetylation of histone 3 supports the activating function of Gata4 and Srf , which might be mediated via p300 . The histone acetyl transferase p300 not only acetylates lysine residues on histone 3 but also on Gata4 , thereby enhancing the DNA-binding and activating potential of this transcription factor [40] . The Srf cofactor Myocardin has been reported to recruit p300 to Srf binding sites whereby histone 3 acetylation is induced and gene expression enhanced [41] . Finally , we studied the change of H3ac marks as a consequence of Srf knockdown using ChIP followed by qPCR . Strikingly , we found complementary alterations of H3ac in a panel of relevant promoter regions ( Figure 3C and 3D ) . To validate and further investigate the correlation of H3ac with Srf target gene expression , we performed genome-wide ChIP-seq experiments in HL-1 cardiomyocytes ( Table S16 ) . We found a synergistic effect of H3ac and Srf binding when compared to non-bound genes or genes solely bound by either of both ( Figure 3E ) . The influence of H3 acetylation marks was further substantiated by RNAi knockdown of Srf in HL-1 cells ( Figure 3F ) . In accordance to its mainly activating function , we found a significant decrease in expression levels of genes bound by Srf . However , this decrease was significantly reduced in genes additionally marked by H3ac in the wild-type . In a further attempt to confirm our results gathered in cell culture , we studied Srf and H3ac binding and their influence on gene expression in mouse hearts in a time-series during cardiac maturation at three developmental stages E18 . 5 , P0 . 5 and P4 . 5 around birth . From the fetal to the postnatal stage , the heart adapts to the body circulation and cardiomyocytes mature . During this process the heart increases in size ( Figure 4A ) , the cells elongate , myofibrils align and cell-cell contacts become bipolar . Immunostaining of α-Actinin-1 and Connexin 43 illustrates hypertrophy of cardiomyocytes , assembly of the sarcomeric z-discs and development of gap junction [42] . Based on ChIP-chip/seq results for Srf and H3ac in HL-1 cells , we analyzed promoter binding regions of genes and miRNAs relevant for this process using ChIP followed by quantitative real-time PCR . The selection comprises ( Figure 4B ) : Dmpk ( kinase of myogenin ) , Slmap ( sarcomeric protein ) , Picalm ( clathrin assembly protein ) , miR-133a ( cardiac and muscle-specific miRNA ) , the growth factor Igf1 and its receptor Igf1r , Pitx2c ( cardiac transcription factor ) , and Nrp2 ( interactor of Vegf ) . We found a high correlation between the changes of Srf and H3ac binding and the gene expression levels over time . In case of Pitx2c and Nrp2 we identified multiple binding events in HL-1 cells by ChIP-chip/seq of which their functionality could be confirmed by common changes over time in the mouse model ( Figure 4B ) . Taken together , these data support an important role for the co-occurrence of Srf and H3ac in the regulation of the cardiac maturation process and underline the influence of histone modifications . Considering that only a small proportion of differentially expressed genes in loss-of-function experiments are direct targets of the respective transcription factors , we were interested in studying the impact of miRNAs as secondary effectors ( Figure 5A , 5B ) . Again we focused on the transcription factor Srf , which is known to regulate cardiac relevant miRNAs like miR-1 and miR-133 [26] , [43] . We investigated Srf binding using ChIP-seq technology to map Srf binding sites potentially regulating miRNAs . We found 22 miRNAs from the miRNA database miRBase with Srf binding within a region of 10kb . This includes the previously described miR-208 Srf binding site , as well as other well-known muscle relevant miRNAs like miR-1 , miR-125b , miR-133 , miR-143 and miR145 ( Table S9 ) . Second , we performed Srf knockdown using two different siRNAs and quantified the miRNA expression levels by miRNA-seq ( Table S10 and Table S18 ) . We observed 42 miRNAs ( 49 loci ) to be differentially expressed in both siRNA experiments , including miR-208 , miR-125b and miR-21 . The analysis revealed that most of the miRNAs were downregulated ( 78% ) supporting the role of Srf as an miRNA activator ( Figure 5A ) . To explore the potential effect of the differentially expressed miRNAs on the Srf network , we assigned confirmed and predicted targets to each miRNA . We found 192 miRNA targets to be also differentially expressed in Srf knockdown , with a higher fraction of upregulated genes ( 57% of all upregulated genes ) compared to downregulated genes ( 44% of all downregulated genes , Figure 5A ) . The majority of these dysregulated target genes had 3′UTR target sequences for a panel of our differentially expressed miRNAs ( median of 3 ) . The differential expression of miRNAs potentially impacts up to 45% of all differentially expressed genes by Srf knockdown , a higher proportion of genes than expected ( Fisher's exact test , p = 1 . 77e-5 ) , and provides a feasible explanation for the observed consequences on the transcriptional portrait ( Figure 5B ) . A representative example is shown in Figure 5C . It comprises the three genes Igfbp5 ( insulin-like growth factor binding protein 5 ) , Nfic ( nuclear factor I/C ) and Ctnnal1 ( catenin alpha-like 1 ) . None of these factors has a direct Srf binding site in ChIP-chip/seq but all are found to be upregulated in the Srf siRNA knockdown experiment . Using miRNA target prediction a number of downregulated miRNAs were found that provide a possible explanation for this indirect regulation ( see Table S11 ) . We confirmed a panel of observed transcription factor binding sites by qPCR ( Figure S5 ) . Using luciferase reporter gene assays , we validated an Srf binding site in the regulatory region of mouse miR-125b-1 as well as an Nkx2 . 5 binding element in the core promoter region of human DPF3 . Mmu-miR-125b-1 is known to be deregulated in heart diseases [44] and was found to be differentially expressed in Srf siRNA knockdown . Figure 6A shows the Srf binding motif and respective Srf ChIP-seq peak within the regulatory region of miR-125b-1 . Luciferase reporter gene assays with wildtype and mutated fusion constructs confirm its functionality . Mutation of the potential Srf binding sequence ( CAGCCAAC→CATAGTAC ) significantly reduced the transcriptional activity of the reporter gene . DPF3 is a novel epigenetic regulator of heart and skeletal muscle development [13] . Within the 1 . 2kbp promoter region we found three Mef2 matrices and one Nkx2 . 5 matrix using TRANSFAC MATCH [45] . In case of Mef2a , all three potential binding sites can drive reporter gene expression as reported [13] . Figure 6B shows the binding of Nkx2 . 5 to the human DPF3 core promoter . Subsequently , co-transfection of reporter construct and increasing amounts of Nkx2 . 5 expression vector revealed a dose-dependent transcriptional activation by Nkx2 . 5 . In line with this , deletion of the potential Nkx2 . 5 binding element ( TCCACTTTCC ) showed that transcriptional activity was indeed mediated through this motif , as activation was lost in the mutated construct . In addition to a genome-wide perspective , our analysis also provides useful information on the level of individual genes . We conducted an extensive literature search and built an Srf centered cardiac transcription network , where we subsequently integrated our findings from the Srf and histone 3 acetylation ChIP and Srf siRNA-mediated knockdown experiments ( Figure 7 ) . Thus our data add regulatory content to the nodes , which are connected by referenced interactions . The network depicts common regulation by Srf and H3ac as well as the impact of the posttranscriptional modulation of expression levels by miRNAs . Target genes important in the cardiovascular context are arranged to their biological roles like regulation in muscle contractility or cardiac growth and conduction . As an example the apoptotic machinery is regulated at all three levels ( Srf , H3ac and miRNAs ) through several pathways involving pro-apoptotic ( Casp3 , miR-320 , Hsp20/a8/a5 , Bax ) as well as anti-apoptotic ( miR-21 , Bcl2 , Mcl1 ) genes .
We present a systematic in-vivo analysis of three levels regulating cardiac mRNA profiles , namely regulation of gene transcription by epigenetic and genetic factors and posttranscriptional regulation by short noncoding RNAs . We performed genome-wide profiling of the DNA occupancy of four key cardiac transcription factors ( Gata4 , Nkx2 . 5 , Mef2a and Srf ) and studied their co-occurrence with four activating histone modifications ( H3ac , H4ac , H3K4me2 and H3K4me3 ) as well as the potential regulatory impact of miRNAs . We combined these data with mRNA expression profiles in wildtype and RNAi mediated knockdown cells and finally confirmed key conclusions in a time-course of cardiac maturation in mouse around birth . In human and mouse ∼2 , 000 transcription factors , more than 100 different modifications of histone residues and ∼700 miRNAs modulate the mRNA profiles corresponding to ∼23 , 000 genes . Major insights have been gained into the regulation of the transcription process by DNA-binding transcription factors [46]–[48] . The role of histone modifications in establishing and maintaining the chromatin status and their function as protein interaction partners has been discovered [12] , [13] , [49] . More recently , the high impact of miRNAs on mRNA profiles and their function as inhibitors of the translation process has emerged [43] , [50]–[52] . However , we lack data showing the interaction between these three levels of regulation . The initial insights were obtained by focusing on the different levels independently , and it was long thought that transcription factors are the main driving force . We feel that it is a fine-tuned balance and our data favor a comparable impact for all three levels with a high degree of interdependency . Our data indicate that histone 3 acetylation is involved in the regulation of Srf as well as Gata4 dependent cardiac genes and moreover potentially compensates the loss of transcriptional activation in Srf knockdown . Vice versa histone modifying enzymes represent an important group of direct downstream targets of Srf ( e . g . histone demethylases containing a Jumonji domain such as Jmjd1c , Jmjd2b , Jmjd3 , Jmjd4 and Jmjd5 , see Figure 7 ) . A similar picture emerges for the relationship of miRNAs and Srf such that the Argonaute proteins Eif2c2 ( Ago2 ) and Eif2c3 ( Ago3 ) , which are direct Srf targets , play a key role for miRNA mediated-mRNA cleavage via the RISC complex [53] . In line with this , we found a panel of miRNAs deregulated in Srf knockdown , explaining three times more differentially expressed genes than Srf binding events alone could do . We are convinced that these data reflect the high degree of interdependency between the different levels . In addition , our data underline the high potency of compensatory regulation between DNA-binding transcription factors . We show that genes regulated by multiple transcription factors were significantly less likely differentially expressed in RNAi knockdown of one respective factor . So far , it had been postulated that members of a gene family ( e . g . Mef transcription factors [4] , [54] , [55] ) or factors with redundant paralogs could buffer each others dysfunction [18] . Our data extend these findings to primarily unrelated transcription factors , which share common targets . The observed correlation of histone 3 acetylation with Srf and Gata4 target gene activation underlines the beneficial effects seen for HDAC inhibitors for a variety of disease states [17] . Further , we favor the view that modulation of the histone modification status might be a plausible explanation for incomplete penetrance or phenotypic diversity as frequently observed in mouse models with identical genetic background or in human disease such as congenital heart disease . Here , a distinct gene mutation can lead to a broad portfolio of phenotypes , such as mutations in Cited-2 [56] , [57] . Environmental factors are potentially causative for these observation and recent reports show a link between environment and alterations of histone modifications . Thus , the change of the phosphorylation status and thereof the activity of histone modifying enzymes mediated for example via the calcium/calmodulin-dependent protein kinase II ( CaMKII ) could represent a mechanistic explanation [58]–[60] . In accordance with others , we found that the overwhelming proportion of differentially expressed genes in our RNAi experiments were indirect targets of the respective transcription factor . Computational studies suggest that up to 30% of all human genes are regulated by miRNAs , while each miRNA may control hundreds of gene targets [61] , [62] . Our in-vivo data highlight the global impact of miRNAs on expression profile alterations seen in transcription factor loss-of-function studies . Differentially expressed miRNAs in Srf knockdown potentially explain up to 45% of the altered mRNA profile in our study . In summary , our data indicate that the different levels regulating mRNA profiles have a high degree of interdependency . The different nodes of the regulatory network have the potential to modulate each other and should therefore be viewed in context . Further functional tests will be required to evaluate regulatory circuits on a single gene basis . It will be of interest to study how the interdependency of the different factors stabilizes the overall function of given networks , and how it contributes to the resistance to external disturbances as well as to the impact of novel therapeutic tools such as HDAC inhibitors or antagomirs .
Human cardiac tissue was obtained from the German Heart Center with ethical approval by the responsible institutional review committee ( Charité 129/2000 ) and informed consent of patients [63] . HL-1 cells were provided by Prof . William C . Claycomb ( Departments of Biochemistry and Molecular Biology and Cell Biology and Anatomy , Louisiana State University Medical Center , New Orleans , LA 70112 ) and cultured as described [64] . The cells were used for experiments at their maximum contraction . HEK293T cells were cultivated according to standard protocols . Mouse hearts at the indicated stages of CD1 and C57/Bl6 strain were dissected in cold PBS from the rest of the body . For subsequent RNA isolation heart samples were directly snap frozen in liquid nitrogen and stored at −80°C . For ChIP or histology experiments heart samples were fixed with formaldehyde or paraformaldehyde , respectively . For RNAi knockdown HL-1 cells were transfected with two different siRNAs ( Qiagen ) per transcription factor ( Table S12 ) using two biological replicates each ( 4 replicates in total ) . As a control , the cells were transfected with an unspecific siRNA ( siNon ) . Cells were grown to 70–80% confluence for at least two days without addition of antibiotics . 3×105 cells were seeded into 6-well plates with 2ml media resulting in 70–80% confluence after 4h . The mixture of 9µl ( 20µM ) siRNA in 270µl of DMEM media and 16µl Lipofectamine 2000 ( Invitrogen ) in 470µl DMEM media was incubated for 20min at room temperature and added drop wise to the cells . The cell culture media was changed after 24h and cells were harvested for protein extraction or RNA preparation after 48h . For reporter gene assays HEK293T cells were transfected with Transfast ( Promega ) according to manufacturer's instruction . Total RNA of cultured cells and heart tissues was isolated using TRIzol reagent ( Invitrogen ) followed by DNase digest ( Promega ) and ethanol precipitation according to standard protocols . Reverse transcription reactions were carried out via AMV-RT ( Promega ) with random hexamers ( Amersham Pharmacia Biotech ) . Illumina array analysis was performed by Integragen ( France ) . For each set of experiments two biological and two technical replicates were analyzed using Illumina Mouse-6 v1 . 1 genome-wide microarrays . To verify transcript expression levels of HL-1 cells and mouse hearts , quantitative real-time PCR measurements were performed using SYBR Green PCR Master Mix ( ABgene ) and the ABI PRISM 7900HT Sequence Detection System . Gene expression was calculated using the ΔCT method with normalization to the housekeeping gene Hprt . Primer sequences and additional results are given in Tables S8 and S13 and Figure S1 . The raw and transformed data of the Illumina expression microarrays ( Mouse-6 v1 . 1 genome wide arrays ) were deposited in the ArrayExpress database at the EBI ( accession code E-TABM-376 ) . Probe intensities were obtained from Integragen ( France ) . Probes were filtered according to the detection score given by the Illumina array analysis software BeadStudio . Only probes with a detection score greater or equal to 0 . 95 in at least one experiment were retained . Probe intensities were qspline normalized and probes assigned to one transcript ( Ensembl v46 , mm8 ) were normalized using the median polish procedure . Differential expression was determined using the limma package [65] of Bioconductor 2 . 0 [66] and p-values were corrected for multiple testing according to Benjamin and Yekutieli [67] . Only transcripts with p-value smaller or equal to 0 . 05 in both siRNA-mediated knockdowns when compared to siNon-treated cells were considered to be significantly differentially expressed . Small RNAs were isolated from total RNA of HL-1 cells and prepared for miRNA sequencing using Illumina Kit FG-102-1009 according to manufacturer's protocol . For quantification of miR-133a-1 in mouse hearts stem-loop qPCR ( primer sequence: GTTGGCTCTGGTGCAGGGTCCGAGGTATTCGCACCAGAGCCAACCAGCTG ) and TaqMan qPCR ( forward primer: 5′ATTAATTTGGTC CCCTTCAAC , reverse primer:5′GTGCAGGGTCCGAGGT , TaqMan probe 21 ( Roche ) ) were performed as described elsewhere [68] , [69] . Small RNAs were sequenced by Illumina/Solexa next-generation ( single-end ) sequencing technology . The small RNA-seq data were deposited in the Gene Expression Omnibus ( GEO ) database at the NCBI ( accession code GSE26397 ) . In two independent siRNA-mediated knockdowns of Srf ( Srf-si1 & Srf-si2 ) and an unspecific siRNA we retrieved 14 , 911 , 499 ( Srf-si1 ) , 14 , 518 , 157 ( Srf-si2 ) and 14 , 742 , 382 ( siNon ) unfiltered 36bp reads , which yielded 5 , 634 , 650 ( Srf-si1 ) , 5 , 503 , 661 ( Srf-si2 ) and 5 , 674 , 429 ( siNon ) unique ( i . e . non-redundant ) read sequences . These reads were mapped to the mouse reference genome ( NCBI v37 , mm9 ) using MicroRazerS [70] allowing at most 20 equally-best hits for each read and using a seed length of 16 bases with at most one mismatch . For Srf-si1 96 . 7% , for Srf-si2 96 . 2% and for siNon 96 . 5% of all unique sequences could be mapped to the mouse genome . In total 402 miRNAs were identified , corresponding to 450 different loci . To annotate the aligned sequence reads with miRNAs , we checked for overlaps with miRNA positions ( http://www . miRBase . org/ , release 14 . 0 ) . We tested for differential expression between the Srf-si1/2 and siNon libraries using Fisher's exact test with FDR correction for multiple testing ( p≤0 . 05 ) . For all miRNAs identified as significantly differentially expressed in at least one siRNA knockdown of Srf ( but both either up- or downregulated ) compared to negative control we did target gene predictions using the miRanda v3 . 0 algorithm [71] . Finally , the target prediction revealed 192 of 429 differentially expressed genes . Using a fisher exact test we found the number to be statistical significant when compared to a prediction based on all versus differentially expressed genes ( p = 1 . 77e-5 ) . The number of unfiltered 36bp reads for wildtype HL-1 was 14 , 440 , 535 , while the human heart samples produced 14 , 475 , 968 ( NH-1 , normal heart ( NH ) ) , 16 , 270 , 049 ( NH-2 ) , 12 , 940 , 172 ( NH-3 ) and 14 , 890 , 970 ( NH-4 ) unfiltered reads , respectively , yielding 5 , 541 , 954 ( HL-1 ) , 5 , 176 , 852 ( NH-1 ) , 7 , 189 , 852 ( NH-2 ) , 3 , 397 , 365 ( NH-3 ) and 5 , 075 , 129 ( NH-4 ) unique reads . Mapping to mouse and human reference genomes ( mm9 , NCBI v37 and hg18 , NCBI v36 ) was performed by MicroRazerS [70] allowing at most one mismatch and 20 equally-best hits per read with a seed length of 16 ( mouse ) and 18 ( human ) , respectively . 97% ( HL-1 ) and 87–91% ( NH-1-4 ) of all unique sequences could be mapped to the corresponding reference genome . In total 196 common , 107 mouse-specific and 180 human-specific miRNA families were found ( http://www . miRBase . org , release 14 . 0 ) . To account for cross-species differences in the specific-expression levels , we used rank-transformed miRNA expression levels for comparison . ChIP experiments with HL-1 cells and mouse hearts were carried out as previously described [72] with minor modifications . The antibodies used are given in the Table S14 . ChIP-chip experiments of HL-1 cells were performed on NimbleGen custom made microarrays with two biological duplicates ( containing two pooled technical replicates each ) . Samples were labeled and hybridized according to NimbleGen standard procedure . Sample preparation for ChIP-seq of HL-1 cells was performed according the Illumina library preparation procedure . Two pooled biological replicates for Srf and H3ac were sequenced using Illumina/Solexa next-generation ( single-end ) sequencing technology . ChIP-chip and ChIP-seq data were confirmed by quantitative real-time PCR using the SyberGreen I PCR Master Mix ( Abgene ) and the ABI PRISM 7900HT Sequence Detection System or using the RealTime ready DNA Probes Master with the Universal ProbeLibrary and the LightCycler 1536 ( Roche ) . Results of ChIP-qPCR experiments are given in Figure S5 and primer sequences for verifications in Table S13 . ChIP after siRNA knockdown of Srf in Hl-1 cells was performed using the LowCell ChIP Protein A Kit from Diagenode according to the manufacture's instructions . The raw and transformed data of the ChIP-chip experiments and the array design were deposited in the ArrayExpress database ( accession code E-TABM-378 and A-MEXP-893 ) . We designed a set of two 385k NimbleGen arrays to represent enhancer and promoter regions of 12 , 625 transcriptional start sites based on a broad panel of muscle relevant data source ( Table S1 ) [13] . The arrays represented 89Mbp of the mouse genome build mm8 and contained 740 , 000 probes with a tiling of 110bp ( 50–60bp gap between probes ) . This included conserved regions ( based on PhastCons [31] score thresholds of 0 . 2 ) within 10kb upstream , the full sequence within 2kb upstream and the first exon and intron of the corresponding transcript . The array intensities of each channel were normalized and log-transformed using VSN [73] . Log-ratio enrichment levels for each probe were calculated by subtraction of log Cy3 ( Input ) from log Cy5 ( ChIP sample ) . The signal of transcription factors were smoothed by calculating a median over the probes inside a sliding window of 600bp . To distinguish enriched probes a z-score and empirical p-value for each probe on the null hypothesis that these z-scores have a symmetric distribution with mean zero was calculated . Significant probe positions ( corrected for multiple testing [74] , FDR<0 . 1 ) , with a distance less than 210bp were combined into transcription factor binding sites . The histone binding sites were identified as described previously [39] . The ChIP-seq data were deposited in the GEO database ( accession code GSE26397 ) . Of the initial 6 , 967 , 318 and 8 , 364 , 328 sequence reads obtained in the Srf and H3ac ChIP-seq experiment , respectively , 4 , 543 , 634 ( 65 . 2% ) for Srf and 6 , 141 , 144 ( 73 . 4% ) for H3ac could be mapped to the mouse reference genome ( NCBI v37 , mm9 ) using the read mapping tool RazerS [75] . Only uniquely mapped 36nt reads with at most two mismatches were retained . To identify enriched regions we used the CisGenome software [76] . For Srf we used a window size of 100bp , a step size of 25bp and a read count level of 10 ( FDR = 1 . 6% ) . For H3ac we applied a window size of 250bp , a step size of 50bp and a read count level of 10 ( FDR = 4 . 7% ) . After peak localization the found Srf and H3ac peaks were filtered ( see Text S1 ) identifying 2 , 190 Srf and 10 , 486 H3ac ChIP-seq peaks . The occurrence of transcription factor binding motifs within observed peaks was analyzed with TRANSFAC MATCH program [45] by using ±250bp sequence surrounding the peak center and position weight matrices corresponding to the studied factors obtained from TRANSFAC [29] . The TRANSFAC matrices used for motif search are listed in Table S7 . Presence of the CArG-box was determined by searching the pattern CC ( A/T ) 6GG with two errors at most . The degree of conservation of respective motifs was studied using PhastCons conserved elements [31] . In addition 100% conservation between human and mouse was defined using 100bp windows . Identified ChIP-chip or ChIP-seq binding sites were assigned to transcriptional start sites if located within 10kb upstream or in the transcribed region . General co-regulation of a gene by two or more transcription factors was defined irrespective of the distance between the binding sites . In addition , co-occurrence between transcription factors or histone modifications was defined if the centers of the peaks had a distance below ±500bp . The association of gene groups to Gene Ontology ( GO ) terms [77] was assessed as described previously [78] ( conditional hypergeometric test , p<0 . 001 ) . To analyze the association of differentially expressed transcripts with GO categories , transcripts were mapped to genes . Overrepresentation was tested against all genes represented on the ChIP and siRNA array , respectively . Standard bioinformatic analysis was carried out using R and Bioconductor packages [66] as well as Perl and its BioPerl modules . If not mentioned otherwise , p-values given are based on Student's t-test . Specific or non-specific siRNA treated HL-1 cardiomyocytes were used for Western Blot analysis to monitor the knockdown efficiency at protein level . HL-1 cells were treated with lysis buffer ( 20mM Tris-HCl pH 7 . 4 , 150nM NaCl , 1mM EDTA , 1% Triton , 1mM DTT , 0 . 1mM PMSF , 1× Protease Inhibitor Cocktail , 1mM NaVO4 ) for protein extraction . Western Blot was performed according to standard protocols . All antibodies with their respective dilution are given in Table S14 . Reporter constructs were made by cloning the 385bp long human DPF3 minimal promoter ( chr14:72 . 430 . 563–72 . 430 . 943 , NCBI36/hg18 ) and the 485bp long regulatory region downstream of mmu-miR-125b-1 ( chr9:41 . 390 . 238–41 . 390 . 700 , NCBI37/mm9 ) into the pGL3 basic vector ( Promega ) . Transient co-transfections were carried out in triplicates in 96-well plates in HEK293T cells by transfecting 50ng of reporter vector , 5ng of Firefly luciferase vector for internal normalization of transfection efficiency and 50–150ng of the respective expression vectors . Activity was measured by Dual-Luciferase assay ( Promega ) after 48 hours . Site-directed mutagenesis of DNA was carried out using the QuikChange site-directed mutagenesis kit ( Stratagene ) according to manufacturer's instructions . Oligonucleotides for mutagenesis were designed to introduce deletions or mutations of the potential Nkx2 . 5 or Srf binding sites . Mutagenesis was confirmed by plasmid sequencing carried out at MWG Biotech . For immunofluorescence analyses , the heart tissues were fixed over night with 4% paraformaldehyde , dehydrated and embedded in paraffin . Subsequently sections of 8µm were de-paraffinized , rehydrated and antigen retrieval was performed in 10mM citric acid buffer ( pH 6 ) . Blocking was carried out in 5% normal goat serum in PBS for 1h at room temperature . Primary and secondary antibodies were applied in the same buffer for 2h at room temperature , each followed by three washes in PBS and a DAPI counterstaining . Antibodies with their respective dilution are listed in Table S14 . The sections were mounted in Flouromount G ( Electron Microscopy Science ) and examined on a Zeiss LSM 510 META confocal microscope ( Carl Zeiss ) . | An evolutionary conserved orchestra of transcription factors controls cardiac development and function . More recently the contributions of epigenetic and post-transcriptional mechanisms like histone modifications and microRNAs have been identified . The interplay between these regulatory mechanisms is still an open question . However , perturbations of the cardiac transcriptome , triggered by all three levels of regulation , are underlying cardiovascular disease such as congenital heart malformations . Here , we show the impact of the interdependencies of four key transcription factors ( Gata4 , Mef2a , Nkx2 . 5 , and Srf ) and the contribution of activating histone modifications and microRNAs on the cardiac transcriptome . We found that even these non-paralogous transcription factors can partially compensate each other's function . Our data show that histone 3 acetylation correlates with Srf- and Gata4- dependent gene activation . Moreover , we predict a large proportion of indirect Srf targets to be regulated by Srf-dependent microRNAs , which thus might represent an important intermediate layer of regulation . Taken together , we suggest that the different levels regulating cardiac mRNA profiles have a high degree of interdependency and the potential to buffer each other , which presents a starting point to understand the phenotypic variability typically seen in complex cardiovascular disorders . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"biology/post-translational",
"regulation",
"of",
"gene",
"expression",
"molecular",
"biology/histone",
"modification",
"molecular",
"biology/transcription",
"initiation",
"and",
"activation",
"genetics",
"and",
"genomics/gene",
"expression",
"physiology/cardiovascular",
"physiology",
"and",
"circulation",
"molecular",
"biology/bioinformatics",
"genetics",
"and",
"genomics/epigenetics",
"computational",
"biology/systems",
"biology",
"cardiovascular",
"disorders"
] | 2011 | The Cardiac Transcription Network Modulated by Gata4, Mef2a, Nkx2.5, Srf, Histone Modifications, and MicroRNAs |
It is widely recognized that representing a protein as a single static conformation is inadequate to describe the dynamics essential to the performance of its biological function . We contrast the amino acid displacements below and above the protein dynamical transition temperature , TD∼215K , of hen egg white lysozyme using X-ray crystallography ensembles that are analyzed by molecular dynamics simulations as a function of temperature . We show that measuring structural variations across an ensemble of X-ray derived models captures the activation of conformational states that are of functional importance just above TD , and they remain virtually identical to structural motions measured at 300K . Our results highlight the ability to observe functional structural variations across an ensemble of X-ray crystallographic data , and that residue fluctuations measured in MD simulations at room temperature are in quantitative agreement with the experimental observable .
It has been suggested that at temperatures below the protein dynamical transition temperature , TD , there is a dominant native basin in which a protein's dynamics is largely controlled by harmonic motions [1] . Above this temperature , a sudden activation of new anharmonic protein motions that are thought to be dependent on a more fluid solvent environment [2] , correlates with a rapid enhancement of enzymatic function in most cases . The importance of dynamics in mediating protein function is widely recognized [3] , and experimental techniques such as nuclear magnetic resonance ( NMR ) , quasi-elastic neutron scattering , dielectric relaxation , Mossbauer and terahertz time domain spectroscopies have been used to explore the dynamical transition behavior of proteins with temperature and water solvent [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] . Much evidence supports the idea that activation of solvent dynamics must proceed first in order to initiate the dynamical transition to a functional protein , emphasizing that the protein itself plays a more passive role in the concept known as “solvent slaving” . [13] Because the majority of X-ray crystal structures of proteins are modeled as single conformations [14] in crystalline environments , their dynamical information is limited . Dynamics is often indirectly addressed by theoretical estimates of uncertainty in atomic positions using Luzzati or Read plots [15] , [16] and isotropic or anisotropic B-factors that measure primarily molecular disorder in the crystal , and possibly other errors , in addition to thermal motion [17] , [18] . More recently , an increasing number of high-resolution data sets have permitted the anisotropic refinement of disordered regions and the modeling of alternate backbone and side-chain conformations [19] , [20] , although most protein crystals diffract to too low resolution for the modeling of disorder in general [21] . Some limited information on conformational mobility can be determined from multi-start simulated annealing refinement [22] , [23] , multi-copy refinement [24] , or time-averaging with multiple refinements [25] , [26] , [27] , producing different structures because each instance fits the structure factor data slightly differently . Furthermore , given the high amount of automation in modern day protein crystallography , the entire procedure of model building , density modification and refinement can yield an ensemble of structures compatible with a given X-ray data set . [28] In this work we define several experimental ensembles with respect to the X-ray crystallography derived hen egg white lysozyme ( HEWL ) structure 3LZT [29] which are analyzed by comparing them to MD generated ensembles at different temperatures . HEWL is unique among almost all proteins in the PDB because ( 1 ) there is at least one high resolution structure that serves as the reference ( here 3LZT ) , ( 2 ) it contains no prosthetic groups or metals , ( 3 ) it is a protein that has been solved in multiple space groups , and ( 4 ) there are ∼80 independent solved structures to generate an experimental ensemble . Surprisingly virtually no other protein allows us to do the same analysis under these criteria . Based on this data , the experimental ensembles include: ( 1 ) the X-ray ensemble generated from multi-start simulated annealing of 3LZT ( 3LZT-MSSA ) , ( 2 ) the X-ray ensemble of HEWL structures that crystallize in the P1 space group like 3LZT , and ( 3 ) the X-ray ensemble of HEWL structures that crystallizes into alternative space groups to P1 , which we refer to as the non-P1 ensemble . These experimental ensembles , whose structure factor data was generated over the period between 1974 and 2010 ( see supplementary material ) , are compared to MD simulations of HEWL in water , employing 3LZT as an initial configuration , and simulated at temperatures of 200K , 210K , 220K , 230K , and 300K .
The HEWL protein ( 3LZT ) was simulated in the AMBER9 [30] molecular mechanics package using the AMBER99SB ( protein ) [31] and TIP4P-Ew ( water ) [32] force fields . The HEWL protein was immersed in a box of 5736 water molecules and 9 Cl- and equilibrated by first restraining the protein atoms with a 10 kcal/mol/Å2 restraint while the system was heated from 0 to 200K , 210K , 220K , 230K or 300K using the Andersen thermostat under constant volume conditions . After equilibration , the system underwent 10ns of NPT molecular dynamics , sufficient to generate a stable protein within the MD model whose RMSD does not vary after the first 1ns , consistent with previous studies [33] , [34] . The equations of motion are integrated with 1fs timesteps , the long-range electrostatic interactions are calculated using Particle Mesh Ewald method , and a cutoff of 10 . 0Å is used for real space electrostatics and Lennard-Jones interactions . All bonds involving hydrogens were constrained using the SHAKE algorithm . The system was then equilibrated under constant pressure using the AMBER default Berendsen barostat parameters at 1atm for 1 ns . The molecular dynamics ensembles at each temperature were derived from 100 snapshots separated by 0 . 01ns over each 10ns trajectory for each temperature . All crystallographic analysis was performed using the program PHENIX [35] . For consistency the original HEWL model ( 3LZT ) used throughout this study was re-refined in phenix . refine [36] at 1 . 10Å resolution , yielding a high quality structure with an R-factor = 0 . 0990 and R-free = 0 . 1308 . [37] We then performed multi-start simulated annealing refinement against the high resolution data set of 3LZT to generate the 3LZT_MSSA reference ensemble . All X-ray ensembles and molecular dynamics structures at different temperatures were superimposed against the refined 3LZT model using phenix . superpose_pdbs , and using phenix . model_vs_data , and all resultant R-factors and RMSDs recorded ( see selection reported in Table 1 ) . Snapshots sampled from the MD trajectory at each temperature were least square fit to the experimental real space data of 3LZT and B-factors set to an average value . Structure factor data was then calculated but without atomic relaxation of atomic positions , in order to compare the structural deviations of the MD model from the 3LZT X-ray structure reference . We quantify the structural variations on a per residue basis among the ensembles by calculating the local density correlation ( LDC ) coefficients of electron density values between X-ray and MD ensembles computed from the model maps around individual amino-acid residues [38] , [39] , [40] ( 1 ) where the ρ's are electron density values at grid points , and Eref and Ealt refer to densities of the members of the reference 3LZT-MSSA ensemble and the alternative ensemble to be compared , respectively . Computing a LDC requires two maps , and each map can be computed from one single model or an ensemble of models ( for example , a PDB file containing multiple HEWL models in the P1 space group split by MODEL-ENDMDL records ) . Residue level LDCs between the ensembles of structures were computed using phenix . real_space_correlation tool . Computing a LDC for a residue ( for example ) requires defining a region around a residue in both maps and the grid points in those defined regions are then used in the LDC calculation . The region around a residue can be defined assuming that each atom has radius of 1 . 5–2 . 0Å . Since the LDC does not depend on the scale the occupancies of atoms in ensemble containing N models do not need to be divided by N . We categorize LDC values between structural ensembles greater than 0 . 7 as having a strong correlation [41] , [42] .
Figure 1 shows a comparison of the LDCs between the P1 and non-P1 X-ray ensembles for HEWL against the MSSA-3LZT reference ensemble . First it is noteworthy that >90% of residues of the P1 ensemble are well correlated with the 3LZT-MSSA reference , although there are deviations in LDC<0 . 7 in a few isolated regions . More interesting is the greater dissimilarity between the alternative crystal space groups of the non-P1 ensemble , which shows many regions that are poorly correlated with the 3LZT-MSSA and P1 ensembles . It is important to emphasize that if a member of the non-P1 HEWL ensemble had been chosen as a MSSA reference , the same regions of difference would be found for the P1 ensembles since the LDC analysis is symmetric between opposite definition of the reference ensemble . The LDC deviations seen between the experimental X-ray ensembles show remarkable correlation with NMR S2 order parameters for backbone amide groups measured in 15N relaxation experiments [43] , [44] . Although NMR order parameters have been compared to atomic B-factors of HEWL X-ray structures previously [43] , [45] , we have used the LDC and a far larger experimental ensemble of ∼80 different HEWL structure that shows far better quantitative agreement than previously described . Regions of S2<0 . 8 for HEWL correspond to residues 16–19 , 45–50 , 67–70 , 116–119 , while even lower S2 values were measured for residues 85–86 ( loop preceding the C-helix ) , 102–106 ( the loop connecting the C-helix and D-helix in the α-domain ) , and residues 127–129 in the C-terminus[43] . Previous normal mode analysis of HEWL has shown that the lowest frequency mode [45] , a strong mechanistic indicator of protein function [46] , corresponds to activation of the β-turn connecting the first two strands of the β-sheet ( residues 44–50 ) and the central portion of long loop ( residues 67–73 ) in the β-domain , and the enhanced fluctuations in the N-terminus ( 1–39 ) and C-terminus ( 116–129 ) in the α-domain , around a central hinge [47] . It is evident that the X-ray ensembles exhibiting regions of LDC<0 . 7 , captures the NMR disorder of an aqueous thermal environment and normal mode analysis relevant for HEW lysozyme function quite well . Thus the experimental X-ray ensemble can measure the activation of functional motions of the protein at a residue-by-residue level as we compare to LDCs of molecular dynamics simulations below and above TD∼215K . Figure 2 shows the time progression of the root-mean-square deviation ( RMSD ) of the molecular dynamics trajectory from the 3LZT start state at each temperature . It is evident that the simulation model shows that a structural transition has occurred over the temperature range of 210K–230K . Figure 3 shows the LDC's for the experimental ensemble against the averaged MD ensembles at 200K and 210K ( which were found to be very similar to each other and hence we averaged their LDC ensemble data ) . Below the transition temperature of ∼215K , a majority of residues ( 92 out of 129 ) are highly similar ( LDC>0 . 7 ) to the 3LZT reference , while 19 of the 37 residues with an LDC<0 . 7 are within experimental deviations permitted under different crystallization conditions . For the remaining 18 residues outside of experimental differences , ∼11 residues have slightly degraded LDC values ranging from 0 . 6 to 0 . 7 , with the remaining larger differences outside of experiment isolated to residues 80–86 . Nonetheless , the overall dynamical motions of the aqueous solution of HEW lysozyme below 215K are not activated in any of the highly flexible or global motion regions that signify the active state of the protein . Figure 4 shows the LDC's for the experimental ensemble against the averaged MD ensembles at 220K and 230K ( which were found to be very similar to each other and hence we averaged their ensemble data ) . We note in the region around residues 97 to 105 the X-ray ensemble shows very low LDC values , while this LDC minimum is broader over the residue range from 97 to 114 for the MD ensembles . This is because the triclinic and tetragonal crystal forms that dominate our experimental X-ray ensemble have a large number of atomic crystal contacts in the region of 105 to 114 , suppressing their fluctuations . This suggests that our X-ray ensemble is incomplete , and we predict that a different crystal form of HEWL that relieves those contacts would bear out the MD fluctuations in this small region . Nonetheless , above the transition temperature of ∼215K , a majority of residues ( 90 out of 129 ) are now dissimilar ( LDC<0 . 7 ) to the 3LZT reference , with 40 of the 90 dissimilar residues yielding LDC values less than 0 . 5 in the same regions as the overall X-ray ensemble . This is due to activation of global motions of the α- and β-domains about the central hinge , signifying that fluctuations of an active protein are now populated . Figure 3 also shows that the MD ensemble just above TD is measuring structural deviations that are mostly identical to the MD ensemble at 300K , thereby showing that the functional dynamical signatures are largely complete just past the protein dynamical transition temperature .
It is widely recognized that representing a protein as a single static conformation is inadequate to describe the dynamics essential to the performance of its biological function . X-ray crystal structures have historically relied on atomic displacement parameters and similar metrics to provide information on local flexibility and disorder [16] , [17] , but more recently have included multiple models consistent with a given set of structure factor data to better represent the dynamical ensemble [14] , [28] . However the possibility of generating structure factor data for a given protein in different crystal forms and solvent conditions could generate an ensemble of structures that reveal the functionally relevant protein conformational states that are populated under physiological conditions . | There is a well-recognized gap between the dynamical motions of proteins required to execute function and the experimental techniques capable of capturing that motion at the atomic level . We show that much experimental detail of dynamical motion is already present in X-ray crystallographic data , which arises from being solved by different research groups using different methodologies under different crystallization conditions , which then capture an ensemble of structures whose variations can be quantified on a residue-by-residue level using local density correlations . We contrast the amino acid displacements below and above the protein dynamical transition temperature , TD∼215K , of hen egg white lysozyme by comparing the X-ray ensemble to MD ensembles as a function of temperature . We show that measuring structural variations across an ensemble of X-ray derived models captures the activation of conformational states that are of functional importance just above TD and they remain virtually identical to structural motions measured at 300K . It provides a novel analysis of large X-ray ensemble data that is useful for the broader structural biology community . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biophysics/experimental",
"biophysical",
"methods",
"biophysics/theory",
"and",
"simulation"
] | 2010 | Evidence of Functional Protein Dynamics from X-Ray Crystallographic Ensembles |
Most of the natural variation in wheat vernalization response is determined by allelic differences in the MADS-box transcription factor VERNALIZATION1 ( VRN1 ) . Extended exposures to low temperatures during the winter ( vernalization ) induce VRN1 expression and promote the transition of the apical meristem to the reproductive phase . In contrast to its Arabidopsis homolog ( APETALA1 ) , which is mainly expressed in the apical meristem , VRN1 is also expressed at high levels in the leaves , but its function in this tissue is not well understood . Using tetraploid wheat lines with truncation mutations in the two homoeologous copies of VRN1 ( henceforth vrn1-null mutants ) , we demonstrate that a central role of VRN1 in the leaves is to maintain low transcript levels of the VRN2 flowering repressor after vernalization . Transcript levels of VRN2 were gradually down-regulated during vernalization in both mutant and wild-type genotypes , but were up-regulated after vernalization only in the vrn1-null mutants . The up-regulation of VRN2 delayed flowering by repressing the transcription of FT , a flowering-integrator gene that encodes a mobile protein that is transported from the leaves to the apical meristem to induce flowering . The role of VRN2 in the delayed flowering of the vrn1-null mutant was confirmed using double vrn1-vrn2-null mutants , which flowered two months earlier than the vrn1-null mutants . Both mutants produced normal flowers and seeds demonstrating that VRN1 is not essential for wheat flowering , which contradicts current flowering models . This result does not diminish the importance of VRN1 in the seasonal regulation of wheat flowering . The up-regulation of VRN1 during winter is required to maintain low transcript levels of VRN2 , accelerate the induction of FT in the leaves , and regulate a timely flowering in the spring . Our results also demonstrate the existence of redundant wheat flowering genes that may provide new targets for engineering wheat varieties better adapted to changing environments .
The temperate grasses , which include economically important species such as wheat , barley , rye , and oats , are well-adapted to cold winters . Most of these species require a prolonged period of cold treatment for timely flowering , a process referred to as vernalization . This requirement delays the initiation of the reproductive phase and protects the sensitive floral meristems from frost damage during the winter . It also contributes to the precise adjustment of flowering time to seasonal changes , which is important to maximize seed production . Therefore , a better understanding of the mechanisms involved in the regulation of wheat flowering can contribute to the engineering of high yielding varieties adapted to changing environments . The cloning of the three main wheat vernalization genes , VRN1 [1]–[3] , VRN2 [4] and VRN3 [5] , and the characterization of their natural allelic variation [6]–[11] provided an important first step in our understanding of this regulatory pathway . However , the mechanisms involved in the interactions among these genes are still controversial [12]–[14] . To facilitate the discussion of these complex interactions , the information available for these three regulatory genes is presented first . The VRN3 gene is the main integrator of the vernalization and photoperiod signals in the temperate grasses [15] ( Figure 1 ) . This gene encodes a RAF kinase inhibitor–like protein with high similarity to Arabidopsis protein FLOWERING LOCUS T ( FT ) [5] and will therefore , be designated as FT hereafter . In Arabidopsis , FT transcription is induced by long days in the leaves and the encoded protein travels through the phloem to the stem apical meristem [16] . There , FT interacts with the bZIP transcription factor FD and up-regulates the expression of the meristem identity gene APETALA1 ( AP1 ) , which leads to the transition of the stem apical meristem from the vegetative to the reproductive phase [17] . A similar interaction has been observed in wheat , where the homologous FT protein interacts with an FD-like protein ( FDL2 ) that has the ability to bind in vitro the promoter of VRN1 , the wheat homolog of AP1 [18] ( Figure 1 ) . The insertion of a repetitive element in the FT promoter in the wheat variety Hope results in the overexpression of FT and early flowering . Transformation of a winter wheat with this FT allele results in accelerated flowering even in the absence of vernalization , which suggests that high FT transcript levels are sufficient to overcome the vernalization requirement [5] . Furthermore , transcript levels of different wheat and barley FT alleles correlate well with flowering time , which suggests that the amount of FT transcript in the leaves is critical for the regulation of flowering time in the temperate cereals [5] . FT and upstream genes of the photoperiod pathway are well conserved between Arabidopsis and the temperate cereals , but the vernalization genes in these species are very different . The main Arabidopsis vernalization genes FLOWERING LOCUS C ( FLC ) and FRIGIDA ( FRI ) have not been detected in the temperate cereals , and similarly , the central flowering repressor VRN2 has not been detected in Arabidopsis [4] . Despite the fact that they belong to different classes of proteins , VRN2 and FLC both repress FT and prevent flowering until the plants are vernalized ( Figure 1 ) [5] , [19] . Recent studies suggest that the negative regulation of FT transcription by VRN2 in the temperate cereals is mediated by the competition between VRN2 and the photoperiod protein CONSTANS ( CO , a promoter of FT expression ) for binding with a common set of NF-Y transcription factors [15] ( Figure 1 ) . NF-Y transcription factors have been shown to be involved in FT activation in Arabidopsis [20] , [21] . The VRN2 locus includes two tandemly duplicated CCT domain ( CONSTANS , CO-like , and TOC1 ) genes , ZCCT1 and ZCCT2 , which function as long day flowering repressors [4] . Simultaneous deletions or non-functional mutations in all ZCCT genes result in a spring growth habit in both barley and wheat [4] , [11] . In the commercial tetraploid wheat varieties studied thus far ( including the variety ‘Kronos’ used in this study ) , ZCCT-A1 , ZCCT-A2 , and ZCCT-B1 all have natural deleterious mutations in the CCT domain and the only functional copies are the two similar ZCCT-B2 genes present in the VRN-B2 locus [11] . Therefore , a natural deletion including both ZCCT-B2 genes was sufficient to generate a tetraploid wheat with spring growth habit [11] . Indirect evidence suggests that VRN2 transcription is repressed in the spring by the up-regulation of VRN1 , closing a positive feedback regulatory loop that is central for the precise regulation of flowering time in the temperate cereals ( Figure 1 ) [22] . The wheat VRN1 gene encodes a MADS-box transcription factor closely related to the three paralogous Arabidopsis meristem identity genes AP1 , CAULIFLOWER ( CAL ) and FRUITFULL ( FUL ) [1] , [3] . VRN1 transcripts are significantly up-regulated during vernalization , both under long and short days . Since FT and VRN2 transcript levels are undetectable under short days [23] , [24] , it was concluded that VRN1 is a direct target of vernalization . This conclusion agrees with the observation that vernalization promotes an active chromatin state by increasing levels of histone 3 lysine 4 trimethylation ( H3K4me3 ) and decreasing H3K27me3 in VRN1 regulatory regions but not in those of VRN2 or FT [25] . In the temperate cereals there are two additional MADs-box genes similar to VRN1 designated as FUL2 ( = HvMADS8 = OsMADS15 ) and FUL3 ( = HvMADS3 = OsMADS18 ) [26] . The duplications that gave rise to these three paralogous genes in wheat are independent from the duplications in Arabidopsis since they occurred after the monocot-dicot divergence [1] . Therefore , the sub-functionalization of the duplicated meristem identity genes was also independent in these two lineages . In Arabidopsis , AP1 and CAL transcripts are mostly confined to the developing flowers [27] whereas FUL transcripts are detected both in apices and leaves but at low levels [28] . In contrast , high levels of VRN1 have been observed in the leaves of the temperate grasses before the emergence of spikelet primordia , which suggests that VRN1 is part of an early signal involved in the transition from the vegetative to reproductive stages [1] , [29] , [30] . In Arabidopsis , all three paralogs have retained meristem identity functions , and only the triple ap1-cal-ful mutant is unable to form flowers [31] . In rice , in addition to the AP1/FUL2/FUL3 homologs a fourth MADS-box gene ( OsMADS34 = PAP2 ) needs to be deleted to abolish the transition of the shoot apical meristem to the reproductive stage [32] . In contrast , diploid wheat radiation mutants lacking the VRN1 gene are unable to flower , which suggests that VRN1 is essential for wheat flowering [33] . These mutants , designated maintained vegetative phase ( mvp ) showed an unexpected down-regulation of FT transcript levels [13] , that was not predicted by current flowering models ( Figure 1 ) . However , a more recent study showed that the radiation deletions that eliminated VRN1 were much larger than originally described and included the wheat ortholog of PAP2 ( AGLG1 ) , the PHYTOCHROME C ( PHYC ) and several other linked genes . Therefore , the mvp mutant results are open to alternative interpretations that are at the center of conflicting flowering models in the temperate grasses [13] , [14] . To determine the specific role of VRN1 in the induction of flowering and in the regulation of down-stream flowering genes we developed truncation mutants for the A and B genome copies of VRN1 in tetraploid wheat ( henceforth , Δvrn-A1 and Δvrn-B1 ) and combined them to generate two sets of mutants with no functional copies of VRN1 ( henceforth , Δvrn1-null ) . Using these mutants we demonstrate that functional VRN1 proteins are not essential for wheat flowering or for the up-regulation of FT . We also show that VRN1 expression in the leaves is important for maintaining low transcript levels of VRN2 in the leaves after vernalization , which is critical for a timely flowering in the spring .
To understand better the large differences in flowering time observed in the VRN1 and VRN2 mutants we studied the effect of these mutations on the transcription profiles of VRN1 , FT and VRN2 by quantitative RT-PCR ( qRT-PCR ) . Since FT and VRN1 transcription profiles are similar , they are described together . To study the role of VRN2 in the absence of functional VRN1 proteins , we compared the transcriptional profiles of the Δvrn1-null and Δvrn1-Δvrn2-null mutants . These two mutants have the same VRN1 mutations but differ in the presence or absence of functional VRN2 genes ( Figure 4 ) . Since these two mutants lack any functional VRN1 genes , only the transcript profiles of VRN2 ( ZCCT2 ) and FT are presented in Figure 4 . Even in the absence of functional VRN1 proteins , vernalization accelerated flowering of the Δvrn1-null and Δvrn1-Δvrn2-null mutants ( Figure 2C ) , which indicates the existence of additional vernalization responsive genes . To test if VRN1's closest paralogs FUL2 and FUL3 were responsive to vernalization , we characterized their expression profiles using the same cDNA samples obtained from the vernalization experiments described in Figure 3 .
The dramatic non-flowering phenotype of the wheat mvp mutants suggested initially that VRN1 was an essential flowering gene [33] . However , a later study showed that the deletions in the mvp mutants including VRN1 were larger than initially proposed , and encompassed several genes including PHYC , an important light receptor and AGLG1 , the wheat ortholog of rice PAP2 [14] . Phytochromes affect flowering time in Arabidopsis [35] and rice [36] and PAP2 affects flowering time and reproductive development in rice [32] and therefore cannot be ruled out as an alternative cause of the non-flowering phenotype of the mvp mutants . The Δvrn1-null mutants developed in this study allowed us to separate the effect of VRN1 from the effect of the other genes included in the mvp deletions . The production of normal flowers and seeds in the Δvrn1-null and Δvrn1-vrn2-null mutants demonstrates that VRN1 is not essential for wheat flowering , and that the non-flowering phenotype of the mvp mutants is not solely determined by the deletion of VRN1 . In addition , the early flowering time of the Δvrn1-vrn2-null mutant indicates the existence of redundant flowering genes that are capable of rapidly inducing flowering in wheat in the absence of functional VRN1 and VRN2 proteins . Vernalization accelerated flowering time by 84–133 days in the two Δvrn-A1 mutants ( functional vernalization responsive VRN-B1 allele ) , but only by 31–37 days in the Δvrn1-null mutants relative to the unvernalized controls . This three- to four- fold reduction in the acceleration of flowering by vernalization in the mutants with truncated copies of VRN1 confirms the important role this gene plays in the vernalization response in wheat . However , the fact that a significant acceleration of flowering by vernalization was still detected in the Δvrn1-null mutants indicates the existence of unidentified genes with the ability to respond to vernalization . This significant response to vernalization in the absence of VRN1 does not seem to be dependent on the presence of VRN2 , because the Δvrn1-vrn2-null mutants also showed a significant acceleration of flowering by vernalization . Loukoianov and collaborators [22] observed a negative correlation between the transcript levels of VRN1 and the transcript levels of VRN2 in the leaves of isogenic hexaploid wheat lines carrying different VRN1 spring alleles , and hypothesized that VRN1 functions as a negative transcriptional regulator of VRN2 . This negative interaction became an integral part of a feedback regulatory loop placed at the center of current flowering models ( Figure 1 ) , but has never been conclusively demonstrated before this study . Results from the Δvrn1-null mutants confirmed the validity of the previous hypothesis . Only the mutant lines with no functional copies of VRN1 showed an up-regulation of VRN2 transcript to pre-vernalization levels when the plants were removed from the vernalization treatment and were transferred back to room temperature ( Figure 3F and Figure S2F ) . This result , observed in the two independent sets of Δvrn1-null mutants , confirmed that one specific role of VRN1 is to maintain low transcript levels of VRN2 in the leaves following winter , when the longer day length and warmer weather conditions favour the induction of VRN2 transcription . The transcriptional activation of VRN1 in the leaves during vernalization provides an important regulatory signal that controls the transition from vegetative growth in the fall to reproductive development in the spring . In the fall , low VRN1 transcript levels result in high VRN2 expression , the repression of FT , and the maintenance of vegetative growth . During the winter VRN1 is up-regulated and VRN2 is down-regulated . The presence of VRN1 in the leaves prevents VRN2 from being up-regulated during the longer and warmer days of spring , which facilitates the transcriptional activation of FT and the induction of flowering . In summary , the negative regulation of VRN2 by VRN1 in the leaves is a key regulatory step in the seasonal responses of winter wheat . It remains to be determined whether the VRN1 protein interacts directly with the VRN2 gene or whether other genes mediate this regulatory interaction . Previous vernalization experiments conducted under short-day conditions demonstrated that VRN1 transcript levels were up-regulated by vernalization in the absence of detectable levels of VRN2 and FT transcripts [23] , [24] . Based on these results it was assumed that vernalization operated mainly on VRN1 and that the downregulation of VRN2 during vernalization was likely an indirect effect of the up-regulation of VRN1 [12] , [37] . This hypothesis was reinforced by the observation that vernalization promotes an active chromatin state in VRN1 regulatory regions but not in those of VRN2 or FT [25] . However , in this study we observed down-regulation of VRN2 during vernalization in both sets of Δvrn1-null mutants , which encode for truncated VRN1 proteins ( Figure 3F and Figure S2F ) . This result demonstrates that the down-regulation of VRN2 during vernalization does not require the presence of VRN1 . Similarly , a developmental down-regulation of VRN2 was observed in both sets of unvernalized Δvrn1-null plants ( Figure 3E and Figure S2E , dotted lines ) . This result suggests that , in addition to VRN1 , there are other negative regulators of VRN2 that are developmentally regulated . We are currently developing a triple vrn1-ful2-ful3 mutant to test if the induction of FUL2 and/or FUL3 transcripts during vernalization ( Figure 5 and Figure S4A–S4B ) and development ( Figure S3A and S3C ) play a role in the down-regulation of VRN2 . In rice , the simultaneous down-regulation of closely related FT paralogs Hd3a and RFT1 by RNAi results in non-flowering plants [38] . Therefore , it is tempting to speculate that the permanent down-regulation of FT in the mvp mutants could account for their non-flowering phenotype . To explain the low levels of FT transcripts detected in the mvp mutants , Shimada and collaborators ( 2009 ) proposed an alternative flowering model for the temperate cereals in which VRN1 promotes FT transcription independently of VRN2 [13] ( Figure 6A ) . In this model , the low FT transcript levels of the mvp mutants are caused by the VRN1 deletion . This model also proposes that FT acts as a transcriptional repressor of VRN2 , and VRN2 represses VRN1 [13] ( Figure 6A ) . Shimada's model includes a regulatory feedback loop including the same three genes present in the model in Figure 1 but operating in the opposite direction , and will be referred hereafter as the ‘reverse’ model ( Figure 6A ) . The simplified model from Figure 1 , included for comparison in Figure 6B , will be referred to as the ‘original’ model for the following discussion . The reverse model assumes that the deletion of VRN1 is the cause of the low transcript levels of FT in the mvp mutants . However , our results indicate that FT can reach very high transcript levels in the absence of functional VRN1 proteins both in the Δvrn1-null and Δvrn1-vrn2-null mutants ( Figure 3C–3D , Figure S2C–S2D , Figure S5B ) . Although , we cannot completely rule out the very unlikely possibility that the truncated VRN1 proteins in the Δvrn1-null mutants are able to up-regulate FT , we favor the hypothesis that the permanent down-regulation of FT in the mvp mutants is caused by the deletion of additional genes closely linked to VRN1 and included in the same deletion . The PHYC is a potential candidate to explain these differences since mutations in the PHYC gene are known to affect flowering time in both rice [36] and Arabidopsis [35] . However , the effects described in these model species are not as large as those observed in the mvp mutants , suggesting the possibility that the non-flowering phenotype of the mvp mutants is caused by other genes in the deletion or by the combination of deletions of more than one gene . We are currently developing a wheat phyC-null mutant to test the contribution of this gene to the regulation of flowering time in wheat . Another potential candidate among the genes deleted in the mvp mutants is the MADS-box gene AGLG1 . The rice homolog of AGLG1 ( PAP2 ) is expressed in the shoot apical meristem during the initiation of the reproductive stage at the same time as the meristem identity genes ( OsMADS14 = VRN1 , OsMADS15 = FUL2 and OsMADS18 = FUL3 ) . A null allele of the PAP2 locus combined with reduced expression of the three meristem identity genes by RNA interference down-regulates FT expression in the leaves and inhibits the transition of the apical meristem from the vegetative to reproductive phase [32] . In wheat , AGLG1 transcripts are detected in the developing spike later than VRN1 and are not detected in the leaves [1] . These results suggest that these homologous genes play different roles in the regulation of flowering initiation in rice and wheat . The transcription profiles of FT and VRN2 in the Δvrn1-null and Δvrn1-vrn2-null mutants provide additional evidence that contradicts the predictions of the ‘reverse’ flowering model . This comparison shows that even in the absence of functional copies of VRN1 , the deletion of VRN2 results in the up-regulation of FT ( Figure 4C–4D ) . This result is difficult to explain by the reverse model , which proposes that the negative regulation of FT by VRN2 requires the presence of functional VRN1 proteins ( Figure 6A ) . In addition to the genetic evidence discussed above , biochemical interactions described in previous studies support the interactions proposed in the ‘original’ model . The proposed regulation of VRN1 by FT ( Figure 6B ) in wheat is consistent with the known interactions between FT and AP1 ( VRN1 homolog ) in Arabidopsis . In this species , the FT protein is synthesized in the leaves and is transported through the phloem to the stem apical meristem where it forms a complex with the bZIP transcription factor FD and binds to the promoter of AP1 [17] . Similarly , the wheat FT protein interacts with the bZIP transcription factor FDL2 which interacts in vitro with the VRN1 promoter [18] . In rice , the FT homolog ( Hd3a ) interacts with OsFD1 and a 14-3-3 protein to form a complex that binds the promoter of the rice homolog of FUL2 ( OsMADS15 ) [39] . A biochemical mechanism has been proposed also for the repression of FT by VRN2 [15] . Using a yeast-three-hybrid system Li et al . ( 2011 ) showed that VRN2 and CONSTANS proteins compete for binding with a common set of NF-Y transcription factors , which have been shown to be involved in the regulation of FT in Arabidopsis [20] , [21] . In summary , available genetic and biochemical evidence support the ‘original’ model of flowering for the temperate cereals [12] , [37] . The fact that both sets of Δvrn1-null null mutants were able to flower and set normal seeds , demonstrated the existence of redundant flowering genes with meristem identity functions overlapping those of VRN1 . The MADS-box genes FUL2 and FUL3 are the closest paralogs of VRN1 [26] and therefore , the most parsimonious hypothesis is that their encoded proteins have retained meristem identity functions that can compensate for the lack of a functional VRN1 protein . A similar redundant system exists in Arabidopsis , in which only simultaneous mutations in the three duplicated meristem identity genes AP1-CAL–FUL results in mutants that are unable to form flowers under all tested environmental conditions [31] . The existence of functional redundancy in the meristem identity genes is also evident in rice , where the down-regulation of individual meristem identity genes OsMADS14 ( = VRN1 ) , OsMADS15 ( = FUL2 ) or OsMADS18 ( FUL3 ) has limited effect on flowering time [32] , [40] . Only the simultaneous downregulation of the three rice meristem identity genes in combination with pap2 mutants resulted in extensive delays in stem elongation and severely perturbed inflorescence development [32] . In wheat and barley , previous studies have provided some indirect evidence that support the hypothesis that FUL2 and FUL3 have retained overlapping functions with VRN1 . First , in situ hybridization studies have shown similar spatial and temporal transcription profiles of these three genes during the initial stages of spike development [29] , [30] , [41] , [42] . In addition , wheat FUL2 and rice FUL3 genes were shown to induce flowering when over expressed in transgenic Arabidopsis and rice plants [30] , [40] . The experiments described in this study showed additional similarities between FUL2 , FUL3 and VRN1 genes . First , the transcript levels of FUL2 and FUL3 were significantly up-regulated during vernalization independently of VRN1 and this up-regulation was proportional to the duration of the cold treatment , as described before for VRN1 ( Figure 5 and Figure S4A–S4B ) . Second , the transcript levels of FUL2 and FUL3 were negatively regulated by VRN2 ( Figure 5 versus Figure S5A–S5B ) and positively regulated by FT in transgenic wheat plants overexpressing FT ( Figure S6B–S6D ) . Finally , when Δvrn1-null mutant plants were moved from the cold to room temperature , FUL2 and FUL3 transcripts returned to pre-vernalization levels in the mature leaves but not in the actively dividing apices ( Figure 5 ) . Since cell division is required to establish epigenetic changes , the different transcriptional profiles in these two tissues may reflect the epigenetic regulation of FUL2 and FUL3 by vernalization , as shown before for VRN1 [25] . Studies of the chromatin changes in the regulatory regions of FUL2 and FUL3 will be required to rule out alternative explanations . In summary , the similar patterns of transcriptional regulation , together with the similar in situ hybridization profiles in early spike development and the early flowering of FUL2 and FUL3 transgenic plants , support the existence of some overlapping functions between FUL2 , FUL3 and VRN1 . We have initiated the development of a triple vrn1-ful2-ful3 mutant in tetraploid wheat to test the roles of FUL2 and FUL3 in flowering initiation in wheat . In addition to the above similarities , a critical difference was observed in the ability of these three meristem identity genes to regulate VRN2 expression in the leaves . In the absence of functional VRN1 genes , VRN2 was up-regulated after vernalization in the Δvrn1-null mutants , in spite of the presence of functional FUL2 and FUL3 genes . This result suggests that FUL2 and FUL3 are unable to maintain the repression of VRN2 after vernalization . In summary , this study demonstrates that a central role of VRN1 in the leaves is to maintain the repression of VRN2 after the winter to promote timely flowering in the spring . In spite of its important role in the seasonal regulation of flowering , our results demonstrate that VRN1 is not essential for flowering and raise new questions regarding the roles of FUL2 , FUL3 , AGLG1 and PHYC in the regulation of flowering initiation in wheat . Flowering time is a key component of wheat adaptation and productivity and therefore , a precise understanding of the regulatory mechanisms involved in the initiation of flowering will be beneficial to engineer wheat cultivars better adapted to a changing environment .
A TILLING ( for Targeting Local Lesions IN Genomes ) population of 1 , 368 lines of the tetraploid wheat cultivar ‘Kronos’ mutagenized with ethyl methane sulphonate ( EMS ) [34] was screened for mutations in the A and B genome copies of VRN1 using the genome specific primers described in Table S1 . Two VRN1 genomic regions , one including exon one and the other including exons three to six , were targeted for mutant detection using the CelI assay described before [34] . The VRN2 natural mutants have been described before [14] . RNA samples were extracted from leaves using the Spectrum Plant Total RNA Kit ( Sigma-Aldrich ) . Purified RNA samples were checked for RNA integrity by running 0 . 5 µg RNA on a 1% agarose gel . All samples showed clear 18S and 25S ribosomal RNA bands indicating lack of RNA degradation . Melting curves showed a single peak , which confirmed amplification of a single product . Standard curves were constructed to calculate amplification efficiency for the SYBR Green systems developed for FUL2 and FUL3 ( Table S3 ) . SYBR Green systems for VRN1 , ZCCT2 and FT were developed in previous studies [5] , [11] . All primers used for qRT-PCR are conserved between the A- and B- genome copies of their respective targets . Quantitative PCR was performed using SYBR Green and a 7500 Fast Real-Time PCR system ( Applied Biosystems ) . ACTIN was used as an endogenous control using primers described before [43] . Transcript levels for all genes and experiments presented in this study are expressed as linearized fold-ACTIN levels calculated by the formula 2 ( ACTIN CT – TARGET CT ) . The resulting number indicates the ratio between the initial number of molecules of the target gene and the number of molecules of ACTIN and therefore , the Y scales are comparable across genes and experiments . Primer efficiencies were all higher than 95% ( Table S3 ) and therefore , were not included in the calculation of the linearized values . In the Δvrn1-null set 2 and the derived Δvrn1-vrn2-null mutants the VRN-B1 transcripts were not detected in the qRT-PCR experiments . The flowering experiments for the VRN1 mutants were performed in CONVIRON growth chambers . Long day photoperiod experiments were performed with 16 h of light ( 6 am to 10 pm , including 1 hour ramp at the beginning and end of the cycle ) and 8 h of dark . Intensity of the sodium halide lights measured at plant head height was 260 µM s−1 . For the unvernalized plants day temperatures were set at 22°C and night temperatures at 17°C respectively . Relative humidity in the chambers was maintained at 60% throughout the entire experiment . All the vernalization experiments in this study were performed under long days to separate the effect of low temperature from the overlapping effects of short days . Day and night temperatures were 4°C whereas light intensity and relative humidity were identical to the unvernalized experiments . Unvernalized control plants were planted 5 weeks after the sowing of the vernalized plants to allow both groups of plants to reach a similar number of leaves by the end of the vernalization treatment . Even though the developmental stages of the two groups were coordinated , the vernalization treatment can still affect the subsequent growth rates after vernalization . To eliminate this potential difference , spring lines with no vernalization requirement were included as controls in both vernalized and unvernalized treatments . The differences in heading time between the vernalized and unvernalized vernalization-insensitive spring lines were used to adjust the heading times of the other vernalized plants . The Δvrn-B1 mutants were used as controls in Figure 2A and the Δvrn-B1-Δvrn2-null mutant was used as a control in Figure 2C . Both control lines carry the functional vernalization-insensitive VRN-A1 allele . The Δvrn-B1-Δvrn2-null mutant has the additional deletion of all functional copies of VRN2 and differs from the Δvrn1-Δvrn2-null lines in the presence of the functional VRN-A1 allele . The experiment comparing the Δvrn1- null and Δvrn1-vrn2-null mutants ( Figure 2C ) was performed in the greenhouse ( both mutants were grown under the same conditions ) ; with an average day time and night time temperatures of 25°C and 17°C , respectively . Day length was extended to 16 h with high pressure sodium lights . Tissues for RNA extraction and qRT-PCR were collected at the same time ( 10:30 am ) in all experiments to avoid potential confounding effects of circadian rhythms . Vernalization in this experiment was performed as described above . To study gene expression in the apices ( Figure 5B and 5D ) , we collected apices from Δvrn1-null mutants set 1 grown under long days in the greenhouse . Samples were collected before vernalization , six weeks after vernalization ( 4°C ) and two weeks after the plants were moved from the cold to room temperature . At each sampling point , 90 apices were harvested and pooled into three replicates including 30 apices per replicate . GenBank accession numbers are JX020745 to JX020760 . | Crop yields are strongly associated with flowering time , therefore a precise understanding of the mechanisms involved in the regulation of flowering is required to engineer varieties adapted to new or changing environments . In wheat , most of the natural variation in flowering time is determined by VERNALIZATION1 ( VRN1 ) , a gene responsible for the transition of the apical meristem from the vegetative to the reproductive phase . Extended exposures to low temperatures during winter ( vernalization ) induce VRN1 expression , which promotes flowering in the spring . VRN1 is expressed in the apices and in the leaves , but its role in the leaves is not well understood . Using two sets of VRN1 knock-out mutants , we demonstrate that a central role of VRN1 in the leaves is to maintain low transcript levels of the VRN2 flowering repressor , which allows the production of the mobile FT protein ( florigen ) required to initiate flowering . Both sets of VRN1 knock-out mutants flowered very late but , eventually , produced normal flowers and seeds , which demonstrates that VRN1 is not essential for wheat flowering . This last result also demonstrates the existence of redundant flowering genes that could provide new targets for engineering flowering time in wheat . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"cereals",
"crops",
"wheats",
"plant",
"growth",
"and",
"development",
"genetics",
"biology",
"plant",
"genetics",
"crop",
"genetics",
"wheat",
"plant",
"biology",
"genetics",
"and",
"genomics",
"agriculture"
] | 2012 | Wheat TILLING Mutants Show That the Vernalization Gene VRN1 Down-Regulates the Flowering Repressor VRN2 in Leaves but Is Not Essential for Flowering |
Vibrio parahaemolyticus is an emerging bacterial pathogen which colonizes the gastrointestinal tract and can cause severe enteritis and bacteraemia . During infection , V . parahaemolyticus primarily attaches to the small intestine , where it causes extensive tissue damage and compromises epithelial barrier integrity . We have previously described that Multivalent Adhesion Molecule ( MAM ) 7 contributes to initial attachment of V . parahaemolyticus to epithelial cells . Here we show that the bacterial adhesin , through multivalent interactions between surface-induced adhesin clusters and phosphatidic acid lipids in the host cell membrane , induces activation of the small GTPase RhoA and actin rearrangements in host cells . In infection studies with V . parahaemolyticus we further demonstrate that adhesin-triggered activation of the ROCK/LIMK signaling axis is sufficient to redistribute tight junction proteins , leading to a loss of epithelial barrier function . Taken together , these findings show an unprecedented mechanism by which an adhesin acts as assembly platform for a host cellular signaling pathway , which ultimately facilitates breaching of the epithelial barrier by a bacterial pathogen .
Vibrio parahaemolyticus is an emerging food- and waterborne bacterial pathogen . It is predominantly associated with gastroenteritis but occasionally manifests as wound infection [1]–[3] . Although infections are self-limiting in immunocompetent patients , in rare cases , usually occurring in patients with an underlying primary disease , V . parahaemolyticus can rapidly disseminate into the blood stream and cause septicemia , a life-threatening condition [4] , [5] . During gastrointestinal disease , the pathogen predominantly colonizes the distal small intestine , where it causes fluid accumulation , extensive tissue damage , a reduction in epithelial barrier function and inflammation [6] . V . parahaemolyticus virulence has so far mainly been attributed to secreted haemolysins ( TDH and TRH ) as well as a range of effector proteins secreted into the host cell cytoplasm via two type III secretion systems ( T3SS1 and T3SS2 ) [7] , [8] . V . parahaemolyticus-mediated cellular toxicity has been attributed to the effects of T3SS1 secreted proteins: VopS , VopQ and VPA0450 contribute to cell rounding , disruption of autophagic turnover and cell lysis , respectively [9]–[11] . T3SS2 has been implicated in intestinal colonization , cellular invasion and enterotoxicity [6] , [12] . However , the increase in epithelial permeability seen during infection in vivo as well as in tissue culture models of infection has not been attributed to any particular virulence factor [6] , [13] . Recently , we have shown that Multivalent Adhesion Molecule ( MAM ) 7 , a constitutively expressed surface protein , contributes to pathogen attachment to host cells during the early stages of infection [14] . V . parahaemolyticus MAM7 recognizes two host surface receptors: it binds host membrane phosphatidic acid ( PA ) lipids with high affinity and uses the extracellular matrix protein fibronectin as a co-receptor . MAM7 contains seven mammalian cell entry ( mce ) domains and each individual domain is capable of binding PA , while a stretch of at least five repeats is required to interact with fibronectin . While PA binding is essential for attachment , binding to fibronectin is dispensable for the interaction but increases the on-rate of binding [15] . Both PA and fibronectin are important signaling molecules in their own right and are implicated in key cellular pathways . PAs make up an average 1–4% of a cell's total phospholipid content [16] and are important as precursors for the biogenesis of other phospholipids , in determining membrane curvature and as signaling molecules [17]–[19] . Several PA-binding proteins are known , including Raf-1 , mTOR and SHP-1 [20]–[22] . As such , PAs are involved in the regulation of a diverse set of cellular functions , ranging from metabolism and trafficking to proliferation . Thus far , studies on PAs have focused on pathways involving PA localized in the inner leaflet of the plasma membrane and cellular organelles , such as the ER . Although PA can also be found in the outer leaflet of the plasma membrane , it is not characterized how this pool is generated or how it is linked to cellular functions [23] , [24] . It has also been reported that PA generation in cells is localized to specific regions within the membrane , but the consequences of this compartmentalization are not well understood [25] . In this study , we found that the clustering of MAM7 molecules on the bacterial surface and subsequent binding of these clusters to phosphatidic acid lipids in the host membrane , causes downstream activation of the small GTPase RhoA . RhoA activation drives actin rearrangements which ultimately lead to redistribution of tight junction proteins and a disruption of epithelial integrity . This breach in the epithelial barrier allows V . parahaemolyticus to translocate across polarized epithelial layers . Thus , we report for the first time that a bacterial adhesin , through direct interactions with host lipid receptors , induces cellular signaling pathways facilitating epithelial barrier breaching by a bacterial pathogen .
Multivalent Adhesion Molecule ( MAM ) 7 present on the outer membrane of V . parahaemolyticus mediates attachment of bacteria to host cells [14] . We used V . parahaemolyticus strain CAB4 to study the infection phenotype in Hela cells . CAB4 is derived from the well characterized , pathogenic RIMD2210633 strain [26] , but lacks both thermostable hemolysins ( ΔtdhA ΔtdhS ) and does not express the two type III secretion systems ( ΔexsA ΔvtrA ) . Despite lacking known virulence factors , infection with V . parahaemolyticus CAB4 strain caused pronounced cytoskeletal changes , with thick strands of filamentous actin forming ( Fig . 1A ) . The appearance of F-actin fibers was observed almost immediately upon infection and persisted throughout the course of the experiment ( Fig . 1C ) . In contrast , no changes in the actin phenotype were observed in cells infected with CAB4Δvp1611 lacking MAM7 ( Fig . 1B ) . As such , MAM7 is necessary to trigger the observed actin rearrangements upon infection with V . parahaemolyticus CAB4 . Next , we investigated if MAM7 is sufficient to cause actin stress fiber formation in Hela cells . Heterologous surface-expression of V . parahaemolyticus MAM7 in otherwise non-adherent Escherichia coli is sufficient to mediate their attachment to a wide range of host cells [14] . Infection of cells with this recombinant , attaching E . coli strain recapitulated the same sustained actin rearrangements seen upon infection with CAB4 ( Fig . 1D , F ) . In contrast , expression of translocation-deficient MAM7 ( MAM7ΔN1–44 ) in E . coli lead to only low levels of attachment and did not trigger actin rearrangements ( Fig . 1E ) . This demonstrates that V . parahaemolyticus MAM7 is necessary and sufficient to convey upon non-pathogenic bacteria the ability to attach to host cells and trigger actin rearrangements . Next , chemical cross-linking was used to directionally couple purified MAM7 protein to the surface of fluorescent polymer beads , thereby mimicking exposure of the adhesin on the bacterial surface . We used this “bacteriomimetic” system to study the effect of MAM7 on host cells independent of other bacterial molecules . Beads directionally coupled to the N-terminus of a protein containing all seven mammalian cell entry ( mce ) domains of V . parahaemolyticus MAM7 ( GST-MAM7 ) attach to host cells and trigger sustained actin rearrangements , mimicking the phenotype seen upon infection with CAB4 ( Fig . 1G , I ) . In contrast , beads coupled to GST alone did not significantly bind to host cells and caused no actin rearrangements ( Fig . 1H ) . Beads coupled to protein containing only a single mce domain ( MAM1 ) also failed to be recruited to the host cell surface in high numbers and did not cause changes in cytoskeletal organization ( Fig . 2A , B ) . Free , soluble , uncoupled MAM7 or free GST also did not cause any cytoskeletal reorganization ( Fig . 2C–E ) . The visually observed changes in actin phenotype were also recapitulated using quantitative analysis of cellular G-actin and F-actin contents by fractionation of lysates , Western Blotting and densitometry ( Fig . 1J and 2F ) . We conclude that V . parahaemolyticus MAM7 , through multivalent binding of host receptors and when clustered on the host cell surface , causes sustained rearrangements in the actin cytoskeleton , visible as bundles of F-actin . Actin rearrangements are generally mediated by activation of small GTPases RhoA , Rac and/or Cdc42 . We tested the activation levels of all three GTPases by studying the fraction of GTP-bound proteins over time , following binding of MAM7-beads to host cells ( Fig . 3 ) . We observed a sustained activation of RhoA , but not Rac or Cdc42 , which persisted over several hours in the presence of cell-bound MAM7 beads ( Fig . 3A–D ) . To analyze if actin rearrangements following MAM7 attachment would be dependent on RhoA , Rac or Cdc42 , we treated cells with Clostridium difficile toxin B ( TcdB ) or C . botulinum C3 transferase . TcdB irreversibly deactivates Rho GTPases by glycosylation of the catalytic threonine residue . C3 selectively inactivates RhoA , B and C but not Rac or Cdc42 by ADP-ribosylation of asparagine 41 in the effector region [27] . While untreated cells displayed stress fibers when incubated with fluorescent MAM7 beads , no actin rearrangements where observed in cells pre-treated with either TcdB or C3 transferase ( Fig . 3E–H ) . The observed change in actin phenotype was also confirmed by quantification of cellular G-actin and F-actin ( Fig . 3I ) . We also studied the effect of MAM7 binding on cells overexpressing either dominant negative RhoA , Rac or Cdc42 . Expression of RhoAT19N-GFP abolished actin rearrangements , while expression of either RacT17N-GFP or Cdc42T17N-GFP had no effect ( Fig . 3J–M ) . We conclude that binding of multivalent , surface-coupled MAM7 to the host cell membrane specifically activates RhoA , which in turn triggers the observed actin rearrangements . Several cellular pathways are involved in relaying signaling between activated RhoA and the actin cytoskeleton and the observed actin rearrangements could be a result of either increased stress fiber formation or a decrease in actin depolymerization [28] . We tested if the MAM-induced RhoA activation and ultimately actin rearrangements proceed via the Rho-associated serine/threonine kinase ROCK , a downstream effector of RhoA , by treating cells with the ROCK inhibitor Y-27632 [29] . Cells incubated with control beads showed no perturbation in the actin cytoskeleton , either in the presence or absence of Y-27632 ( Fig . 4A , C , E ) . In contrast , cells incubated with bead-coupled MAM displayed stress fibers but this phenotype was almost completely abolished in Y-27632 treated cells ( Fig . 4B , D , E ) . These findings were recapitulated when we quantified the cellular G-actin and F-actin content under identical experimental conditions ( Fig . 4F ) . Next , we tested whether MAM-induced ROCK activation takes place upstream or downstream of RhoA . We analyzed RhoA activation in the presence and absence of MAM beads , either on untreated or Y-27632 treated cells . These data show that when ROCK is inhibited , even though MAM-induced stress fiber formation is abolished , RhoA activation levels remain high ( Fig . 4G ) . We thus conclude that MAM-induced ROCK activation occurs downstream of RhoA . Next , we tested the activation of LIM kinase ( LIMK ) and cofilin , two key signaling proteins downstream of ROCK . A significant fraction of LIMK was phosphorylated in the presence of MAM-beads , but the p-LIMK level was much reduced if cells were pre-treated with Y-27632 prior to MAM7 bead adhesion . Incubation with either control beads alone or in combination with Y-27632 treatment did not cause significant LIMK phosphorylation ( Fig . 4H ) . LIMK activation causes phosphorylation of cofilin and thus inhibition of its actin depolymerization activity . We observed an increase in p-cofilin in cells with attached MAM-beads , which was abolished by Y-27632 treatment prior to attachment . Incubation with either control beads alone or following Y-27632 treatment did not cause significant changes in p-cofilin levels ( Fig . 4I ) . In addition , treatment of cells with LIMK inhibitor prior to MAM adhesion lead to a loss of the actin phenotype and concurrent loss of increased F-actin contents ( Fig . 4F ) . We conclude that MAM-induced actin rearrangements proceed via the RhoA/ROCK/LIM-K/cofilin pathway and are the result of abrogated actin depolymerization rather than de novo polymerization . We have previously shown that MAM7 interacts with two types of receptors in the host cell membrane . Each of the seven mce domains within MAM7 is capable of interacting with a phosphatidic acid phospholipid molecule , thereby mediating high affinity binding of bacteria to host cells . Recognition of fibronectin is achieved by a repeat of at least five mce domains and while this interaction is dispensable for attachment , it increases the on-rate of bacterial binding to host cells [15] . We asked if the observed actin rearrangements are a result of MAM binding to fibronectin or phosphatidic acid receptors on host cells , or both . We made MAM attachment to host cells independent of binding to fibronectin by blocking the MAM binding epitope on fibronectin with an antibody [15] . This way , binding of MAM7 to host cells was only mediated by phosphatidic acid receptors . Cells either pre-treated with α-Fn antibodies or non-specific control antibodies were incubated with MAM7 beads or control beads . Following incubation with MAM7 beads , stress fibers were observed in both cells treated with control antibodies ( +Fn ) , or α-Fn antibodies ( −Fn ) , ( Fig . 5A–C ) . In contrast , no actin changes were observed in cells following treatment with either antibody followed by control beads ( Fig . 5C ) . As previously described , uncoupling MAM7 binding from its co-receptor fibronectin did not change the overall number of beads bound per cell if sufficient time was allowed for attachment ( Fig . 5D ) . The interaction between fibronectin and MAM has been mapped to the N-terminal region of fibronectin , which is an epitope commonly exploited by bacterial adhesins for binding [30] . Both Staphylococcus aureus fibronectin binding protein A ( FnBPA ) and Streptococcus pyogenes protein F1 bind the N-terminal part of fibronectin with high affinity [31] , [32] . Thus , we tested whether portions of these two adhesins sharing the same binding epitope with MAM would cause similar actin rearrangements to those observed with MAM7 . We incubated cells with beads coupled to the fibronectin-binding region of either FnBPA ( FnBR1-11 ) or F1 ( FUD ) , as previously described [33] . Although both preparations bound to cells with high efficiency , neither caused stress fiber formation ( Fig . 5E–H ) . Taken together , these findings strongly suggest that fibronectin is not involved in the observed signaling pathway between MAM7 , RhoA and actin . To see whether changes in the membrane lipid composition would impact MAM's ability to trigger RhoA activation , we treated cells with phospholipase C ( PLC ) . MAM7 beads were added to cells either immediately or up to five hours following PLC treatment and subsequent enzyme removal , and levels of beads per cell as well as RhoA activation were measured . In untreated cells , approximately 23 beads were bound per cell ( Fig . 5I , red bar ) . No bead binding was observed if cells were continuously exposed to PLC , since the interaction with fibronectin alone is insufficient to mediate binding ( Fig . 5I , blue bars ) . If PLC was removed , the interaction between lipid receptors and MAM7 , and thus bead binding , was initially completely abolished but was gradually recovered until normal binding levels were regained after four hours of recovery ( Fig . 5I , black bars ) . A similar time course was established for RhoA activation upon MAM bead attachment , with full GTPase activation recovered four hours after removal of PLC ( Fig . 5J ) . We conclude that the MAM7 co-receptor fibronectin is dispensable not just for MAM7 binding , but also for the subsequent activation of RhoA and actin rearrangements caused by adhesion . The observed signaling cascade thus depends on the interaction of multivalent , surface-clustered MAM7 adhesins with phosphatidic acid lipids in the host cell membrane . Vibrio parahaemolyticus mostly causes gastroenteritis and on rare occasions it can lead to systemic disease in immunocompromised patients . To better reflect the in vivo situation , we studied the effect of MAM on polarized intestinal epithelial ( Caco-2 ) cells . Differentiated Caco-2 monolayers are a good model of the epithelium in the small intestine , the main site of V . parahaemolyticus infection . When grown on permeable supports , Caco-2 cells form monolayers with well differentiated brush border microvilli and properties resembling those of the small intestinal epithelium [34] . First , we studied the localization of MAM7 on polarized cell layers . MAM7 exclusively bound to the apical side of the epithelial layer , with the protein being enriched at cellular junctions ( Fig . 6A ) . No binding was observed when MAM protein was added to the basolateral side ( Fig . 6B ) . Similar to the effects seen in Hela cells , MAM-coupled beads and V . parahaemolyticus CAB4 , but not a MAM deletion strain ( CAB4ΔMAM ) , caused a significant increase in RhoA activation compared to untreated cells ( Fig . 6C ) . Because MAM7 was enriched at cell junctions and RhoA activation is capable of affecting the distribution of tight junction proteins , we studied the localization of tight junction markers during infection with V . parahaemolyticus . Apical infection with CAB4 caused a re-distribution of the tight junction markers occludin and zonula occludens protein 1 ( ZO-1 ) ( Fig . 6D , G ) . In contrast , the distribution of both tight junction proteins remained unchanged when cells were infected with CAB4 from the basolateral side ( Fig . 6E , H ) or apically with the MAM knockout strain CAB4ΔMAM ( Fig . 6F , I ) . Next , we asked if re-distribution of tight junction proteins during infection would affect epithelial barrier function . When CAB4 was added to the apical surface of a differentiated layer , a marked decrease in transepithelial electrical resistance ( TER ) was observed approximately three hours post infection . This change was mediated via ROCK/LIMK activation , since treatment of cells with either Y-27632 or LIMK inhibitor abolished the CAB4-mediated decrease in TER ( Fig . 6J ) . Similarly , no significant decrease in TER was observed up to seven hours post infection with either CAB4ΔMAM added apically or CAB4 added to the basolateral side of the epithelium ( Fig . 6M ) . We also investigated whether the disruption of cell-cell junctions was sufficient to allow for bacterial transmigration . Polarized cells were infected with either CAB4 or CAB4ΔMAM and bacterial titers in the opposing compartment were determined either immediately or up to eight hours post infection . When either CAB4 or CAB4ΔMAM were added to the basolateral side , no bacteria were recovered on the apical side . In contrast , CAB4 was recovered from the basolateral side following infection from the apical side . Bacterial numbers on the basolateral side increased significantly 2 . 5 hours post infection and continued to increase until 6 . 5 hours post infection , reaching approximately 1% of the initial infecting population . In epithelial layers apically infected with CAB4ΔMAM , no bacteria were detected on the basolateral side ( Fig . 6K ) . The loss of MAM could be compensated either by the expression of MAM in trans or by treatment of cells with bead-bound MAM , but not with control beads ( Fig . 6L ) . We concluded that MAM selectively binds to the apical side of polarized intestinal epithelial cells , causing a re-distribution of tight junction proteins , disruption of barrier integrity and bacterial transmigration . Finally , we asked if the epithelial disruption caused by MAM-mediated adhesion would contribute to infection in a virulent strain . Polarized intestinal epithelium was infected with the virulent strain POR1 or POR1ΔMAM from the apical or basolateral side . Infection with POR1 apically lead to cytotoxicity and rapid cell lysis , with almost complete cell death five hours post infection ( Fig . 7A ) . The cytotoxicity profile was significantly delayed upon infection with POR1ΔMAM and cell death reached only approximately 70% even seven hours post infection . When cells were infected with either POR1 or POR1ΔMAM from the basolateral side , no significant increase in cytotoxicity was observed over the course of the experiment ( up to seven hours post infection ) . POR1 contains the T3SS effector VopS , which causes RhoA inhibition by irreversible AMPylation of a threonine residue in the switch I region [9] . Thus , we investigated the contribution of MAM to the overall RhoA activation levels in polarized Caco-2 cells infected with the virulent strain . At 2 hours post infection , prior to the onset of cell lysis , RhoA activity was completely abolished in POR1 infected cells . In contrast , RhoA was highly activated in POR1ΔVopS . An intermediate level of RhoA activation was observed in cells infected with POR1ΔMAM ( Fig . 7B ) . We also analyzed the G-actin and F-actin content of polarized Caco-2 cells 2 hours post infection . Within 2 hours , POR1 infection lead to a drop in F-actin content compared to untreated cells , which was mediated by the activity of VopS . In the absence of MAM , or in the presence of ROCK- or LIMK inhibitors , the F-actin content was higher compared to POR1 infected cells ( Fig . 7C ) . Finally , we measured the transepithelial resistance in Caco-2 monolayers infected with the virulent strain . POR1 caused a rapid decrease of TER , which was markedly slowed by treatment of cells with Y-27632 or LIMK inhibitor . Similarly , both POR1ΔMAM and POR1ΔVopS showed a slight delay in depolarization ( Fig . 7D ) .
Previously , we reported that V . parahaemolyticus Multivalent Adhesion Molecule ( MAM ) 7 and several of its homologs from other Gram-negative enteric pathogens mediate initial attachment of bacteria to host cells [14] . In this study , we demonstrated that clusters of multivalent MAM molecules , by binding to the host cell membrane , facilitate activation of the host small GTPase RhoA , which in turn leads to actin rearrangements . Clustering of MAMs is achieved by nature , through display of multiple adhesion molecules on the bacterial outer membrane [14] , but can be mimicked by coupling recombinant MAM molecules to a polymer bead with roughly the same dimensions as a bacterium . Soluble MAM failed to achieve the same effect on host cell signaling . MAMs interact with host cells via two cellular receptors , the protein fibronectin and the phosphatidic acid ( PA ) phospholipids . While the former is a well-characterized pathogen receptor [30] , [35] , [36] , direct binding of a bacterial adhesin to a host cell lipid is a new paradigm of host-pathogen interaction . Over recent years , manipulation of cellular lipids by pathogens has been an emerging field of study , and it has become evident that host cellular lipids are often a primary target of bacterial virulence factors [11] , [37] , [38] . Herein , we showed that MAM's impact on RhoA activation is mediated through its interaction with phosphatidic acid lipids in the host membrane and that its co-receptor fibronectin is dispensable for its function as a signaling effector . Taken together , these findings suggest a mechanism whereby the interaction of clustered MAM adhesins with host membrane lipids causes rearrangements of the latter and that this acts as a signal leading to RhoA activation . However , direct observation of such hypothesized rearrangements of phosphatidic acid molecules within the host membrane on the nanoscale is not within the scope of our studies but is an intriguing possibility and something we are currently investigating . We have elucidated the signaling pathway downstream of RhoA and show the MAM-triggered signal is relayed from activated RhoA , via the Rho-associated serine/threonine kinase ROCK and LIM kinase , to result in phosphorylation of cofilin . Cofilin is an actin-binding protein which mediates actin depolymerization [39] . Its interaction with actin and thus its depolymerization activity is disrupted by phosphorylation , resulting in a net stabilizing effect on filamentous actin and apparent increase in actin stress fibers . Although a large part of our experiments was performed on Hela cells because changes in the actin phenotype following serum starvation are visually easier discernible in this cell type , we show that the MAM-mediated effect on actin also proceeds via ROCK and LIMK activation in polarized intestinal epithelial cells , a more relevant system for studies on V . parahaemolyticus . Since we observe MAM-induced RhoA activation also in polarized epithelial cells , we hypothesize that this RhoA activation facilitates subsequent activation of the ROCK/LIMK/cofilin signaling axis , however we cannot show whether RhoA activation is required in this model , since RhoA inactivation itself leads to increased transepithelial permeability [40] . In the polarized epithelial system , MAM7 selectively attached to the apical side of the layer and attachment caused a marked redistribution of tight junction proteins . A similar phenotype has been described to occur following infection of epithelial cells with other pathogens , such as enteropathogenic E . coli ( EPEC ) or the protozoan parasite Giardia lamblia . With EPEC infection , paracellular permeability also resulted from a redistribution of tight junction proteins upon RhoA activation , although in that case RhoA activation has been largely attributed to the activities of type III system-secreted effectors [41] , [42] . In G . lamblia , barrier failure was attributed to apoptosis of enterocyes [43] . Activation of RhoA through the establishment of a signaling complex consisting of bacterial adhesin clusters and host membrane lipids on the host cell surface is , to our knowledge , a previously unrecognized strategy to achieve epithelial barrier disruption . We demonstrated that the action of MAM7 causes epithelial barrier disruption , as evidenced both by a decrease in transepithelial resistance and the ability of bacteria to transmigrate to the basolateral side of the epithelium . It has previously been shown that CAB4 is unable to invade epithelial cells [12] , so this is likely the result of bacteria moving through compromised cell-cell junctions . It has been shown previously that epithelial integrity is compromised following V . parahaemolyticus infection , both in cultured polarized epithelial cells and in vivo . Animal infection models have shown increased transepithelial permeability using fluorescent dextran as a tracer , but the effect was not attributed to any particular virulence factor [6] . Earlier experiments on polarized Caco-2 cells demonstrated a similar effect on epithelial integrity and ruled out TDH and TRH toxins as the culprit [13] . A comparison between V . parahaemolyticus clinical isolates and environmental strains implicated T3SS2 in transepithelial permeability . However , no whole genome sequences are available for the strains used in this study and we therefore do not know if they encode for a MAM homolog and if so , to what extent it would share sequence similarity to RIMD2210633 MAM7 ( vp1611 ) [44] . More recent studies on Caco-2 and mixed M cell-like co-cultures demonstrated that T3SS1 does not significantly contribute to translocation , while T3SS2 is dispensable but has a moderately enhancing effect on translocation in a RIMD2210633 background [45] . Herein we show that MAM7 is sufficient to cause barrier disruption in cultured polarized epithelium . In the context of a T3SS-competent , virulent strain , MAM induces transepithelial permeability and depolarization of the epithelium early during infection . Since MAM is constitutively expressed and present at the early stages of infection , its effect takes hold almost immediately and RhoA activation is detectable as early as 30 minutes post infection ( the earliest time point measured here ) . The resulting depolarization and disruption of cell-cell junctions leads to an increase in host cell surface available for translocation of type III secreted bacterial effectors . Overall , this mechanism accelerates effector-mediated functional changes in host cells , such as VopS-mediated irreversible RhoA inactivation and concomitant actin depolymerization , thus speeding up infection . These findings strongly indicate experiments comparing the effect of wild type and MAM knockout strains in an animal model and this should be the next step to show if indeed MAM contributes to transepithelial permeability and infection in vivo . Overall , the study we present here demonstrated that the contribution of Vibrio parahaemolyticus MAM7 to the pathogen's infection profile is not limited to its function in early bacterial attachment . By establishing signaling complexes consisting of clustered MAM adhesins and host membrane lipid receptors on the host cell surface , it additionally acts as an effector of host cellular GTPase signaling and its action culminates in breaching of the epithelial barrier . This is , to our knowledge , a previously unrecognized strategy by which a bacterial pathogen disrupts intestinal epithelial function and the detailed molecular mechanism of how this is achieved certainly deserves our further investigation .
The construction of BL21-MAM7 , BL21-MAMΔN1–44 , CAB4 , POR1 , POR1ΔMAM ( POR1Δvp1611 ) and POR1ΔVopS has been described elsewhere [9] , [12] , [14] . The V . parahaemolyticus MAM deletion strain CAB4Δvp1611 was constructed using the same method and same vector construct ( pDM4 containing regions 1 kb up- and downstream of vp1611 ) described in these references . Strains were grown on MLB ( V . parahaemolyticus ) or LB agar ( E . coli ) , with 100 µg/ml of kanamycin or ampicillin added for selection where required . HeLa and Caco-2 epithelial cell lines were cultured at 37°C and under 5% CO2 in Dulbecco's Modified Eagle Medium ( DMEM ) containing 10% heat-inactivated fetal bovine serum , 4500 mg/L glucose , 0 . 5 mM L-glutamine , 100 units/ml penicillin and 20 µg/ml streptomycin . For GTPase activation and microscopy assays , cells were serum-starved for 40 hours prior to treatment . For infection experiments , DMEM with no added antibiotics was used . For experiments on polarized Caco-2 cells , cells were seeded on polycarbonate 3 . 0 µm pore size transwell filters ( Costar ) at 200000 cells/ml . Cells reached confluency after approximately 5–6 days , at which point several transepithelial resistance ( TER ) measurements were taken to check the integrity of the layer and establish baseline measurements . TER measurements before and during infection experiments were taken with a Millicell-ERS resistance apparatus ( Millipore ) . Expression and purification procedures for recombinant proteins have been described in detail elsewhere ( see [14] for GST-MAM7 [15] , for GST-mce1 and [33] for GST-FnBPA FnBR1-11 and F1 FUD constructs ) . Purified proteins were immobilized on amine modified fluorescent blue polystyrene beads with a mean diameter of 2 µm ( Sigma ) using Sulfo-SMPB ( sulfosuccinimidyl 4-[p-maleimidophenyl]butyrate ) cross-linking under reducing conditions , as outlined in the manufacturer's protocol ( Pierce ) . Bead-coupled proteins were added to experiments to give a final concentration of 500 nM immobilized protein and a surface density of 1 . 5×105 molecules per bead ( giving a spacing of approximately 57 nm ) . Tissue culture cells were washed with PBS ( phosphate-buffered saline ) prior to the addition of bacteria in tissue culture medium without antibiotics . Bacteria were added to give a multiplicity of infection ( MOI ) of 100 , except for POR1 and derivatives , where an MOI of 10 was used . Plates were centrifuged ( 1000×g , 22°C , 5 minutes ) prior to incubation at 37°C for 30 minutes to eight hours , depending on the experiment . To uncouple MAM binding from fibronectin or phosphatidic acid , respectively , cultured cells were incubated with anti-Fn antibody ( Sigma , 50 µg/ml in PBS ) or treated with 50 µg/ml phospholipase C ( Sigma ) in PBS for 15 min prior to infection , as previously described [15] . For enumeration of bacteria , samples were removed at time points as indicated and were serially diluted , plated on agar plates , incubated at 37°C for sixteen hours and CFU counts determined the following day . For cytoxicity assays , 200 µl of culture supernatant was removed in triplicate from each well at timepoints as indicated , centrifuged ( 1000×g , 22°C , 5 minutes ) , and 100 µl of the supernatant transferred to a fresh 96-well plate for assays . To quantitate cell lysis , the amount of lactate dehydrogenase ( LDH ) released into the culture medium was determined using the LDH cytotoxicity detection kit ( Takara ) according to the manufacturer's instructions . Results are presented in % lysis , relative to negative ( uninfected ) and positive ( Triton X-100 lysed cells ) controls . Cells were transfected with pcDNA3 containing either EGFP , EGFP-RhoAT19N , EGFP-RacAT17N or EGFP-Cdc42T17N using Fugene HD ( Roche ) transfection reagent according to the manufacturer's protocol . For microscopy , cells were fixed with 3 . 2% formaldehyde , permeabilized with 0 . 1% Triton X-100 and stained with rhodamine-phalloidin to visualize F-actin and SYTO-13 to visualize DNA as indicated . For immunofluorescence microscopy , we used α-GST , α-occludin and α-ZO-1 antibodies ( Sigma ) diluted 1∶500 , followed by FITC-labeled α-rabbit antibody ( Sigma ) at a 1∶1000 dilution . Images were taken either on a Zeiss LSM 510 scanning confocal microscope or a Nikon Eclipse Ti fluorescence microscope and images were prepared using ImageJ and Corel Draw X5 . For quantification of the F-actin phenotype , the total number of cells as well as number of cells containing stress fibers , were enumerated . Some fields contained cells displaying cortical actin , however this phenotype was observed across experiments and was independent of MAM adhesion . Thus , these cells were not counted as positive . Data shown are means ± standard deviation from twelve images ( four frames from triplicate experiments , representing at least 100 cells/experimental condition ) . Proteins were separated by SDS-PAGE and transferred onto nitrocellulose membrane . Membranes were blocked with 5% BSA in TBS-T ( Tris-buffered saline containing 0 . 05% Tween 20 ) for 1 hour at 22°C . Membranes were probed with primary antibodies ( against LIMK , p-LIMK , cofilin , or p-cofilin , all Santa Cruz Biotechnology ) diluted 1∶1000 into blocking buffer for 1 hour at 22°C . After three washes with TBS-T , membranes were incubated with anti-mouse HRP ( horseradish peroxidase ) -conjugated secondary antibody ( GE Healthcare ) diluted 1∶5000 into blocking buffer for 1 hour at 22°C . Membranes were washed three more times with TBS-T and proteins were detected using the ECL plus detection system ( GE Healthcare ) and a Gel Doc XR imager . Bio Rad Quantity One software was used for densitometry . Ratios of globular ( G-actin ) to filamentous ( F-actin ) in cultured , serum-starved cells were determined using the G-actin/F-actin In Vivo Assay Kit ( Cytoskeleton Inc . ) as described in the manufacturer's protocol . Serum-starved , untreated cells ( negative control ) and cells treated with F-actin enhancing solution ( positive control ) were analyzed alongside experimental samples ( MAM-treated and controls , as described in the figure legends ) . G-actin and F-actin levels were determined by Western Blotting and were quantified by densitometry . Results shown are means ± s . e . m . from two independent experiments . Following infection or incubation with beads , cells were washed and collected by scraping into GTPase lysis buffer ( 20 mM Tris HCl pH 7 . 5 , 10 mM MgCl2 , 150 mM NaCl , 1% Triton X-100 . Lysates were homogenized and cleared by centrifugation ( 13000 rpm , 20 min ) . 500 µg of cleared lysates were added to 30 µg of GST-PAK PBD bound to glutathione agarose beads and incubated for 1 hour at 4°C . Samples were separated by SDS-PAGE and immunoblotted with α-Cdc42 or α-Rac antibodies ( Sigma ) and compared to total GTPase levels detected in cell lysates . Activated RhoA was pulled down with the use of a RhoA activation kit ( Cytoskeleton ) according to the manufacturer's instructions . Total and GTP-bound RhoA was detected following SDS-PAGE separation and Western Blotting using α-RhoA antibody ( Sigma ) . To study cellular phenotypes independent of GTPase activation , cells were treated with either Clostridium difficile toxin B ( TcdB ) or C3 transferase to irreversibly inactivate either RhoA , Rac and Cdc42 or RhoA , respectively . Cells were treated wither with 200 ng/ml TcdB ( List Biologicals ) or 1 µg/ml cell-permeable C3 ( Cytoskeleton ) for 4 hours . Attachment experiments were carried out immediately after toxin treatment . | Vibrio parahaemolyticus is a bacterial pathogen which occurs in marine and estuarine environments . It is a main cause of gastrointestinal illness following the consumption of raw or undercooked seafood . In immunocompromised people , the bacteria can sometimes enter the bloodstream and cause septicemia , a serious and often fatal condition . V . parahaemolyticus attaches to host tissues using adhesive proteins . Multivalent Adhesion Molecule ( MAM ) 7 is an adhesin which helps the bacteria to hold onto the host cells early on during infection . It does so by binding two different molecules on the host , a protein ( fibronectin ) and phospholipids called phosphatidic acids . We show that MAM7 does not only play a role in sticking to host cells . By forming adhesin clusters on the host surface and binding to host lipids , it triggers signaling processes in the host . These include activation of RhoA , an important mediator of cytoskeletal dynamics . By doing so , MAM7 perturbs proteins at cellular junctions , which normally maintain the cells in the gut as a tightly sealed layer protective of environmental influences . When bacteria use MAM7 to attach to the intestine , the seals between cells break , permitting bacteria to cross the barrier and cause infection of underlying tissues . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biomacromolecule-ligand",
"interactions",
"biochemistry",
"signal",
"transduction",
"cell",
"biology",
"medical",
"microbiology",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"microbiology"
] | 2014 | Multivalent Adhesion Molecule 7 Clusters Act as Signaling Platform for Host Cellular GTPase Activation and Facilitate Epithelial Barrier Dysfunction |
The Anaplastic Lymphoma Kinase ( Alk ) receptor tyrosine kinase ( RTK ) plays a critical role in the specification of founder cells ( FCs ) in the Drosophila visceral mesoderm ( VM ) during embryogenesis . Reporter gene and CRISPR/Cas9 deletion analysis reveals enhancer regions in and upstream of the Alk locus that influence tissue-specific expression in the amnioserosa ( AS ) , the VM and the epidermis . By performing high throughput yeast one-hybrid screens ( Y1H ) with a library of Drosophila transcription factors ( TFs ) we identify Odd-paired ( Opa ) , the Drosophila homologue of the vertebrate Zic family of TFs , as a novel regulator of embryonic Alk expression . Further characterization identifies evolutionarily conserved Opa-binding cis-regulatory motifs in one of the Alk associated enhancer elements . Employing Alk reporter lines as well as CRISPR/Cas9-mediated removal of regulatory elements in the Alk locus , we show modulation of Alk expression by Opa in the embryonic AS , epidermis and VM . In addition , we identify enhancer elements that integrate input from additional TFs , such as Binou ( Bin ) and Bagpipe ( Bap ) , to regulate VM expression of Alk in a combinatorial manner . Taken together , our data show that the Opa zinc finger TF is a novel regulator of embryonic Alk expression .
During embryogenesis , the Anaplastic Lymphoma Kinase ( Alk ) receptor tyrosine kinase ( RTK ) is dynamically expressed predominantly in the primordia of the visceral mesoderm ( VM ) , the developing CNS , the amnioserosa ( AS ) and in a restricted manner in the epidermis [1] . Alk plays a critical role during VM development , where it is activated in response to the secreted ligand Jelly Belly ( Jeb ) driving the Ras/MAPK/ERK pathway [2–5] . This leads to expression of founder cell ( FC ) specific transcription factors ( TFs ) such as Hand [6] , optomotor-blind related-1 ( org-1 ) [4] and factors important in the muscle cell fusion process like dumbfounded/kin of irre ( duf/kirre ) [3–5] . Jeb/Alk signaling also leads to downregulation of fusion competent myoblast ( FCM ) -specific factors such as sticks and stones ( sns ) [7] and Verprolin 1 ( vrp1 ) [8–10] . In addition , Alk signaling in the VM modulates the subcellular localization of the Gli-family TF Lame duck ( Lmd ) , resulting in Lmd translocation from the nucleus to the cytoplasm [11] . Thus , signaling regulated by Jeb/Alk is critical for embryonic FC-specification and the subsequent fusion with FCMs to form a functional larval midgut muscle [2–5] . While we and others have previously identified and characterized several important components and targets of the Alk RTK signaling pathway , little is currently understood about the molecular mechanisms regulating the spatial and temporal expression of the Alk receptor itself . Development of the early VM requires the activity of the NK4/msh-2-like homeobox TF Tinman ( Tin ) for dorsal mesoderm differentiation , as well as the NK3 and FoxF orthologues Bagpipe ( Bap ) and Biniou ( Bin ) [12–15] . Interestingly , the expression patterns of bap and bin in the VM primordia are similar to that of Alk [15] . In addition , ChIP-on-chip studies have shown the region upstream of Alk gene to be occupied by several mesodermally expressed TFs , such as Bin , Bap , Twist ( Twi ) , Tin and Myocyte enhancer factor 2 ( Mef2 ) at different time points during embryogenesis [16 , 17] . While binding of these factors has been documented , their importance in the regulation of Alk transcription in the VM has only been initially characterized in case of Tin [16 , 17] . Here we address regulation of Alk expression during embryogenesis . We have employed a combination of in vitro and in vivo approaches to identify and characterize Alk-specific enhancer elements , including high throughput yeast one-hybrid screening ( Y1H ) with a library of Drosophila TFs [18] . This Y1H screen identified the zinc finger TF Odd-paired ( Opa ) as binding to an evolutionary conserved cis-regulatory module ( CRM ) within one of the Alk-associated enhancer regions . In agreement with these findings , opa mutants displayed a complete loss of Alk expression in the epidermis and reduced levels of Alk in the VM . Furthermore , CRISPR/Cas9-mediated deletion of the Opa binding site containing region in the Alk locus resulted in a reduction of VM Alk protein together with loss of Alk expression in both the AS and embryonic epidermis , indicating that Opa plays an important role in tissue-specific Alk expression during embryogenesis . We have also identified additional enhancer regions regulated by the Bin and Bap TFs , likely together with additional TFs , that work with the Opa binding CRM to regulate Alk expression in the VM in a combinatorial manner .
To study Alk expression during embryogenesis , we employed transgenic GAL4-lines containing overlapping DNA sequences corresponding to Alk 5-prime upstream regions ( Fig 1A , S1 Fig ) , aiming to identify regulatory elements with activity in the visceral mesoderm ( VM ) . AlkEI6 . 5-GAL4 was previously described [1] as driving expression in the trunk VM with stronger expression in founder cells ( FCs ) ( Fig 1B , stage 11 , arrowhead ) . We also noted that the AlkEI6 . 5-GAL4 driver was expressed in the amnioserosa ( AS ) , in keeping with earlier observations that Alk mRNA is expressed in the dorsal-most region of the embryo corresponding to the presumptive AS at the early gastrulation stage ( S2A and S2B Fig ) [1] . We next analyzed AlkE4-GAL4 , which contains 2 . 4 kb of the AlkEI6 . 5-GAL4 region and an additional 1 . 6 kb upstream ( 4 . 0 kb in total ) . This GAL4-driver promotes expression in a similar pattern to AlkEI6 . 5-GAL4 , suggesting this DNA region also contains regulatory elements involved in Alk transcriptional regulation ( Fig 1C ) . In addition , AlkE2 . 7-GAL4 , covering a shorter sequence within AlkEI6 . 5 and AlkE4 , displays activity in the entire trunk VM , being considerably stronger in FCs ( Fig 1D , arrowhead ) . To ensure the specificity of our transgenic lines for the Alk locus flanking genes we performed in situ hybridization on both neighboring genes namely CG5065 ( upstream ) and gprs ( downstream ) ( Fig 1A , S3 Fig ) . Neither CG5065 nor gprs is expressed in a pattern similar to that of Alk in the VM , suggesting that any VM expressing region identified flanking the Alk locus may be involved in the regulation of Alk transcription . The elevated level of expression of AlkEI6 . 5-GAL4 and AlkE2 . 7-GAL4 in FCs compared with other cells of the developing VM suggests Alk may respond to its own signaling . Since signaling in the FCs is driven by activation of Alk by its ligand Jelly Belly ( Jeb ) , we examined expression of AlkE2 . 7-GAL4 in either the absence of Alk activity ( Alk1/Alk10 ) , or upon activation of Alk by overexpression of Jeb in the VM . AlkE2 . 7-GAL4 expression in the FCs was reduced in Alk1/Alk10 mutants ( Fig 1F; arrowhead ) . In contrast , overexpression of Jeb resulted in robust expression of AlkE2 . 7-GAL4 in all cells of the VM ( Fig 1G; arrowhead ) . These results suggest that Alk expression in the VM is positively regulated by Alk signaling , representing a positive feedback loop . Thus , we have identified CRMs in the 5’ region of the Alk locus that promote Alk expression in the presumptive amnioserosa and developing VM . Additionally , our preliminary GAL4 analysis suggests the presence of inhibitory modules within this region that likely contribute to the overall regulation of Alk expression . ChIP experiments performed by the Furlong laboratory have identified a 547 bp CRM ( MesoCRM-880 ) overlapping the AlkE2 . 7 fragment that binds Bin , Bap , Mef , Tin and Twi TFs [16] ( shown schematically in Fig 1A , S1 Fig ) . Later analysis by the Frasch group identified a 1 , 984 bp region ( AlkE301 ) in a genome wide Tin ChIP analysis that drives expression in the VM [17] ( shown schematically in Fig 1A , S1 Fig ) . Together with our GAL4 analyses these results suggest that the Alk-RB promoter may be important for the VM expression of Alk . To functionally address the role of Alk-RB we generated deletion mutants targeting the Alk-RB isoform with CRISPR/Cas9 [19–21] , employing two independent single guide RNA ( sgRNA ) combinations . This resulted in genomic deletions of 1053 bp ( represented by AlkΔRB_1 . 22 . 2 ) or 1325 bp ( represented by AlkΔRB_15 . 16 . 2 ) in the region of the Alk-RB 5’UTR ( Figs 1A and 2A; S1 Fig; S1 Table ) . Both homozygous mutants were embryonic lethal . We further examined the visceral morphology of homozygous AlkΔRB_1 . 22 . 2 mutant embryos and control siblings using Fasciclin III ( FasIII ) as marker for differentiated VM . In control embryos FasIII was expressed in the visceral musculature surrounding the entire midgut , which at later stages of embryogenesis is subdivided into four chambers ( Fig 2B; arrowhead ) . In AlkΔRB_1 . 22 . 2 embryos , FasIII-positive midgut muscles were absent while FasIII-expression could still be detected in the embryonic foregut and hindgut respectively ( Fig 2C ) , resembling the Alk mutant phenotype [2] . In agreement with their mutant phenotype , AlkΔRB_1 . 22 . 2 mutants lacked detectable Alk mRNA and protein in the VM ( Fig 2K , S4 Fig ) compared to wild-type animals ( Fig 2E and 2H , S4 Fig ) , while Alk expression levels in the CNS were similar to those observed in control embryos ( Fig 2I and 2L; S4 Fig ) . Alk expression was also lost in the AS and epidermis of AlkΔRB_1 . 22 . 2 mutants ( Fig 2J and 2K , asterisks ) . Therefore , expression from the Alk-RB promotor drives Alk expression in the embryonic VM , AS and epidermis and is critical for proper formation of the midgut musculature . A 3 . 6 kb genomic region that covered the putative VM and epidermal Alk enhancer regions identified in our initial experiments ( Fig 1A ) was subjected in parallel to high throughput yeast one-hybrid ( Y1H ) and more detailed reporter gene analyses . Six fragments ( denoted AlkEB6 –AlkEB11; S1A Fig; S2 Table ) of approximately 700 bp in length , including a ~100 bp overlap between neighboring fragments , were analyzed . Embryonic lacZ reporter activity was observed with only two of the DNA fragments studied , namely AlkEB8 and AlkEB9 ( Fig 3A–3E ) . AlkEB8 displayed weaker activity in the VM than that observed with AlkEB9 ( Fig 3B , 3B” , 3D and 3D” , arrowheads; quantified in S5 Fig ) . In addition to VM expression , AlkEB9 was also expressed in the AS and epidermis where it overlapped with Alk protein ( Fig 3E , asterisks; S2D Fig ) . No expression in the AS and epidermis was observed in AlkEB8 ( Fig 3C , asterisks; S2C Fig ) . To further confirm that AlkEB9 contains important enhancer elements for Alk , we performed rescue experiments using AlkE9-GAL4 ( Fig 3F–3H ) . Ectopic over-expression of Alk ( AlkEB9-GAL4>UAS-Alk ) in an Alk1/Alk10 mutant background resulted in a rescue of the embryonic gut phenotype ( Fig 3H ) . Therefore , the AlkEB9 genomic region contains sufficient regulatory information to allow rescue of the embryonic Alk VM expression . High throughput Y1H was carried out on the same six fragments employing a library of Drosophila TFs fused to the yeast GAL4 activation domain [18] ( Fig 1A , S1 Fig ) . Based on our reporter gene analysis we focused on the AlkEB9 DNA bait Y1H data set aiming to functionally characterize novel transcriptional regulators of Alk . A set of TFs was identified to bind to the AlkEB9 DNA bait by Y1H screening ( Fig 4A ) . Among these , Odd-paired ( Opa ) ( Fig 4B ) , Pointed ( Pnt ) , Side and CG14655 bound to the AlkEB9 DNA bait and promoted growth in selective media in all biological replicates performed . We further investigated a role for TFs binding AlkEB9 in Alk transcriptional regulation in vivo , employing paired ( prd ) -GAL4 , which drives expression in alternating parasegments and offers internal control of Alk expression levels in the epidermis . In this assay both Opa and Pnt were identified as potential regulators of Alk , with Opa inducing and Pnt repressing Alk expression ( S6 Fig ) . Of the TFs tested in this study , Opa was the only one that resulted in an increase in Alk protein . We also overexpressed opa with the engrailed ( en ) -GAL4 driver which resulted in an increase in AlkEB9-lacZ reporter activity as well as Alk protein levels in the epidermis ( Fig 4C–4D’ ) , indicating that Opa is sufficient to promote Alk expression . Therefore we focused on a more detailed investigation of the role of Opa Alk transcriptional regulation . Employing the JASPAR online prediction tool [22] , we were able to identify a potential Opa binding site ( BS ) in the AlkEB9 sequence , JASPAR_OpaBS ( GACCTCCGGCTG ) ( Fig 5A and 5B ) . In addition , we identified another Opa BS similar to the Opa consensus motif previously reported by [23] and therefore referred to as SELEX_OpaBS ( GCGGGGATG ) ( Fig 5A and 5B ) . Employing the phastCons database , which identifies evolutionarily conserved elements in a multiple alignment , to analyze this sequence , we found that both binding sites are conserved among Drosophila species ( Fig 5B; conservation score in green; Opa BS highlighted in yellow ) [24 , 25] . We next assessed the ability of Opa to specifically bind these predicted sites by electrophoresis mobility shift assay ( EMSA ) . EMSA was performed on the SELEX_OpaBS and JASPAR_OpaBS sequences , incubating probes with cell lysates from Opa-expressing HEK293 cells in the presence of poly ( dI-dC ) to prevent non-specific binding . Addition of Opa lysate to the binding reaction resulted in a shift of both SELEX_OpaBS and JASPAR_OpaBS probes and was reversed by addition of 100 fold molar excess of non-labelled probe ( Fig 5C and 5D ) . In contrast , addition of cold probes that were mutated within the SELEX and JASPAR binding sites , based on published data [23] , was unable to compete the shift generated upon addition of Opa to the labelled wild-type probe . Furthermore , labelled mutated SELEX_OpaBS and JASPAR_OpaBS probes did not exhibit a mobility shift upon incubation with Opa ( Fig 5C and 5D ) . The above observations led us to characterize the interactions of the Opa with the Alk locus by chromatin immunoprecipitation ( ChIP ) . Consistent with Y1H and EMSA analyses , Opa association is detected with a region upstream of the Alk promoter that spans both the SELEX_OpaBS and JASPAR_OpaBS sequences in chromatin from wild-type embryos ( Fig 5E ) . To address the importance of the JASPAR_ and SELEX_OpaBS for in vivo Alk transcription we first attempted to identify a minimal region within the AlkEB9 region that could drive VM expression . This analysis led to the identification of a 154 bp fragment including both SELEX and JASPAR Opa binding sites ( AlkEB9_OpaBS; schematically shown in S1 Fig ) that drives strong VM and epidermal expression , similar to that observed with the 700 bp AlkEB9 fragment ( Fig 6A–6F’ ) . Quantification revealed that VM expression from AlkEB9_OpaBS was weaker than that of the 700 bp AlkEB9-lacZ reporter ( Fig 6B’ and 6E’; S7 Fig ) , while expression in the epidermis appeared similar in both the 154 bp and 700 bp fragments ( Fig 6C’ and 6F’; S7 Fig ) . In order to examine the role of the predicted Opa binding sites , we introduced the same mutations as in our earlier EMSA analysis within the 154 bp AlkEB9_OpaBS minimal region to create AlkEB9_OpaKO-lacZ . Mutation of these binding sites led to a loss of lacZ expression in both the VM and epidermis ( Fig 6G–6I’ , quantified in S7 Fig ) , implying that the predicted Opa binding sites in AlkEB9 indeed contribute to expression from this element . While mutation of Opa binding sites led to a significant reduction of reporter gene expression in the VM ( Fig 6H’; quantified in S7 Fig ) this was not complete , in contrast to a complete loss of detectable lacZ activity in the epidermis ( Fig 6I’; quantified in S7 Fig ) . Taken together , these data show that the AlkEB9 genomic region contains sequence-specific binding sites for Opa that regulate expression from Alk enhancer elements . To further dissect the potential role of Opa as a regulator of Alk expression , we examined opa expression during embryogenesis [26] . opa mRNA can be detected at stage 5 in the ectoderm and mesoderm progenitors spanning the presumptive segmented region of the embryo . At stage 9 opa expression decreases slightly and appears in the neuroectoderm persisting until late embryo stages . In the VM , opa mRNA is observed in a dynamic pattern , where it is expressed in a clustered fashion in PS 3–5 and PS 9–12 ( S8 Fig ) . We next examined the reporter expression of AlkEB9-lacZ in opa loss-of-function mutants ( opa1/opa8 ) . While AlkEB9-lacZ is activated in the entire VM and epidermis in wild-type embryos ( Fig 7A and 7A’ ) , opa1/opa8 mutant embryos display only weak reporter activity during embryogenesis ( Fig 7B and 7B’; quantified in Fig 7C ) . The severe developmental defects observed in opa1/opa8 mutants make analysis difficult , however we noted lower levels of Alk protein in the VM and a complete loss of detectable Alk in the epidermis of opa mutant animals , in agreement with the loss of AlkEB9-lacZ activity ( Fig 7B ) . These observations were supported by analysis of RNAi-induced Opa knockdown in the developing mesoderm employing 2xPE-GAL4 ( Fig 7D–7F ) . We observed that embryos expressing opa RNAi ( 2xPE-GAL4>UAS-opaRNAi ) displayed a reduction of AlkEB9-lacZ in the VM at later stages when compared with controls ( Fig 7E’ , quantified in Fig 7F ) . Since Opa has been reported to be required for proper midgut formation , with opa mutants exhibiting an interrupted VM that fails to form midgut constrictions during embryogenesis [26] , we also examined Alk signaling in the VM of opa mutants . opa1/opa8 mutants , examined with the FC-marker Org-1 , exhibited Org-1 positive VM FCs , however , the level of Org-1 protein observed was less than in control embryos ( S8B and S8C Fig ) . Since reductions in both Alk and Org-1 protein were seen in opa1/opa8 mutants , we asked whether Opa overexpression was sufficient to drive Alk signaling . As expected , bap3-GAL4 driven expression of Jeb in the VM resulted in an increased expression of the HandC-GFP FC marker reflecting activation of Alk signaling ( S8D–S8D” Fig ) . In contrast , bap3-GAL4 driven expression of Opa did not increase HandC-GFP levels ( S8E–S8E” Fig ) . Thus , while Alk signaling may be reduced in opa mutants , Opa is not sufficient to influence FC specification driven by Alk signaling in the embryonic VM . As a complement to our analysis of opa mutants , we employed the Opa4opt-lacZ transgene as readout for Opa activity , focusing on the epidermis . Opa4opt-lacZ contains four tandem copies of the SELEX determined Opa-BS [23] . In parallel we analyzed the opa3D246 lacZ enhancer trap which reflects opa expression [26] . We observed expression of both opa3D246 and Opa4opt-lacZ in the embryonic epidermis , coinciding with Alk protein ( Fig 7G and 7H ) , suggesting that Opa is both expressed and active in these cells . Furthermore , a mutant Opa4opt-lacZ transgene , called Opa4opt-KO-lacZ , in which the Opa binding sites are mutated , no longer displayed expression overlapping with Alk in the embryonic epidermis ( Fig 7I ) . Taken together , this data supports an important role for Opa in driving embryonic Alk transcription , particularly in the epidermis , through the AlkEB9 regulatory region . However , in agreement with our earlier analyses , Alk expression in the VM does not depend only on Opa activity , since Alk protein is still observed in the VM of opa1/opa8 loss of function animals ( Fig 7B ) . Given the presence of Opa binding sites proximal to the Alk-RB isoform promoter , together with the loss of reporter gene activity after deletion of these sites , we next addressed their in vivo relevance for Alk transcriptional regulation . CRISPR/Cas9 genome editing was again employed to delete the identified Opa binding sites ( Opa-BS ) in the AlkEB9 enhancer region of the Alk locus ( Fig 8A; S1 Fig ) . This resulted in isolation of two viable AlkΔOpaBS mutants: AlkΔOpaBS_10 . 28 . 3 and AlkΔOpaBS_10 . 36 . 1 ( Fig 8A and 8F–8I; S1 Fig ) . Loss of 151 bp containing the Opa binding sites in AlkΔOpaBS_10 . 28 . 3 mutants led to a complete loss of detectable Alk protein in the amnioserosa and epidermis ( Fig 8H; S1 Table ) , indicating this region is essential for Alk expression in these tissues . We also observed reduced Alk protein levels in the VM when compared to control embryos at the same stage ( Fig 8H , compare with Fig 8B; quantified in Fig 8N ) . In close proximity to the Opa binding sites we also observed a cluster of highly scoring JASPAR-predicted binding sites for mesodermal TFs ( Bap , Sna and Tin ) in the AlkEB9 genomic region , here designated as meso-BS ( Fig 8A; S1 Fig ) . Deletion of this meso-BS region alone , in AlkΔmesoBS embryos , does not appear to affect either Alk protein levels or the formation of a fully developed gut ( S1 and S9 Figs; S1 Table ) . Interestingly , AlkΔOpaBS_10 . 36 . 1 removes 178 bp including both the Opa- and the meso-BS sites allowing us to functionally address the contribution of the meso-BS region relative to the Opa binding sites . Deletion of both the meso-BS and the Opa-BS regions ( AlkΔOpaBS_10 . 36 . 1 ) results in viable animals , albeit with reduced Alk protein levels when compared to those in control embryos ( Fig 8F , S1 Table ) . Reduction of Alk protein levels in the VM was noticeably stronger in AlkΔOpaBS_10 . 36 . 1 when compared with AlkΔOpaBS_10 . 28 . 3 mutants ( Fig 8F and 8H; quantified in Fig 8N ) . However , the reduced Alk protein levels observed in AlkΔOpaBS_10 . 28 . 3 and AlkΔOpaBS_10 . 36 . 1 were still sufficient to drive Jeb/Alk signaling in the VM as measured with HandC-GFP reporter expression ( Fig 8G and 8I insets ) , and form a functional gut as visualized by FasIII staining ( Fig 8G and 8I ) . Since we detected VM expression activity in the overlapping Alk proximal AlkEB8-lacZ reporter ( Fig 3B ) , we explored the contribution of the corresponding region in the Alk locus to regulation of Alk VM expression . To do this we employed CRISPR/Cas9 genomic editing to remove 808 bp covering part of AlkEB9 ( 312 bp ) and the majority of AlkEB8 ( 647 bp ) ( represented by AlkΔEB8 ) ( Fig 8A; S1 Fig , S1 Table ) . These mutants were homozygous viable , with a wild-type VM morphology ( Fig 8K; S1 Fig ) . Investigation of Alk protein levels in AlkΔEB8 mutants revealed a decrease , but not complete loss , of Alk in the VM ( Fig 8J; quantified in Fig 8N ) , suggesting that CRM ( s ) within the AlkEB8 region are not essential but contribute to VM expression of Alk . Expression of Alk in the epidermis was not affected , in agreement with a sole epidermal CRM including the Opa binding sites within the AlkEB9 region . To further exclude the possibility that an essential CRM might be located in the overlap between AlkEB8 and AlkEB9 , we generated a series of overlapping reporter constructs in this area S1 Fig . We did not observe any VM expression activity in this reporter series ( S10 Fig ) , suggesting that two CRMs , one in the region of AlkEB8 and one in AlkEB9 function together drive VM expression of Alk . To test the contribution of additional CRMs to VM expression of Alk we extended the AlkΔEB8 deletion to include the Opa binding sites within AlkEB9 . This deletion was denoted AlkΔOpaBS+EB8 ( Fig 8A , 8L and 8M; S1 Fig , S1 Table ) . AlkΔOpaBS+EB8 mutants failed to express Alk protein in the VM , epidermis or AS ( Fig 8L ) , and were homozygous lethal due to lack of FC specification ( Fig 8M; inset ) , supporting our hypothesis of several independent CRMs within this area that are critical for Alk expression in the VM . Taken together , our analysis identifies a CRM proximal to the Alk-RB isoform promotor that contains Opa binding sites as critical for Alk expression in the embryonic AS and epidermis . This region also contributes to Alk expression in the VM . Further deletion analysis reveals additional CRM ( s ) located within the AlkEB8 fragment that contribute to regulation of Alk VM expression . Previous studies identified CRMs binding Bin , Bap , Twi , Tin and Mef2 in the Alk locus [16 , 17] . In particular the ChIP and reporter gene analyses performed by Jin et al . ( 2013 ) suggested Tin binding to be important if not essential for Alk expression . We studied expression of Alk protein and the AlkEB9-lacZ reporter in tin346/ED6058 mutant embryos ( S11 Fig ) . Both Alk protein and reporter gene expression , could be observed in the dorsal epidermis and amnioserosa at stage 10/11 and in the epidermis at stage 14 ( S11 Fig ) , indicating that regulation by Tin is not critical for Alk expression outside the VM . In contrast , Alk and AlkEB9-lacZ reporter gene expression as observed in the VM of control embryos ( S11 Fig ) was not observed . However , this analysis was inconclusive since it is difficult to address if , and to which extent , VM formation proceeds in tin mutant embryos . Bin and Bap TFs are known to have a critical function during Drosophila VM development [12 , 15] . In our initial experiments we were unable to see any effect on Alk expression on ectopic expression of either Bin or Bap alone in the epidermis employing en-GAL4 as driver ( S12 Fig ) , however this may reflect a lack of tissue competence in our experimental approach . Therefore , we analyzed both Alk protein and AlkEB9-lacZ expression in bin and bap mutants focusing on VM expression . Although VM development does not proceed normally in either bin or bap mutants , we could observe Alk protein and AlkEB9-lacZ expression in the VM in both cases ( S13B–S13C’ Fig ) . We next investigated AlkEB8-lacZ expression , which was reduced in both bin mutants and bap mutants ( S13E–S13F’ Fig ) . On closer inspection of the AlkEB8 region we identified four putative Bap binding sites , which we deleted to create AlkEB8ΔBapBS-lacZ . AlkEB8ΔBapBS-lacZ failed to exhibit reporter expression suggesting that Bap may be involved in Alk expression in the VM through binding sites within the AlkEB8 region ( Fig 9A and 9B ) . Based on these findings , we analyzed Alk protein levels in AlkΔOpaBS_10 . 28 . 3;bin1/BSC374 and AlkΔOpaBS_10 . 28 . 3;bap208/ED6058 double mutant backgrounds to test whether Alk expression was affected in a combinatorial manner . We observed a strong reduction of Alk expression in the VM of both AlkΔOpaBS_10 . 28 . 3; bin1/BSC374 mutants ( Fig 9C and 9D; quantified in Fig 9G ) and AlkΔOpaBS_10 . 28 . 3;bap208/ED6058 mutants ( Fig 9E and 9F; quantified in Fig 9H ) , with loss of Bap appearing to have a stronger impact . These results suggest that additional factors , including Bin and Bap , contribute to regulate Alk expression in the VM through the AlkEB8 region of the Alk locus ( Fig 9I ) .
The importance of Jeb/Alk signaling in vivo in the embryonic VM for FC specification is well established [2–5] . From this earlier work we know that activated Alk in the VM triggers not only transcriptional activation but also post-translational modifications that promote the specification of the FC fate [3–6 , 11] . In contrast , very little is known about factors that mediate Alk transcriptional regulation . In this study we aimed to identify CRMs and TFs important for Alk transcription . The Alk-RA and Alk-RB transcripts encode the same protein , but differ in their 5’ non-coding regions which employ alternative promoters [1] . This potentially allows differential expression of the Alk-RA and Alk-RB mRNA isoforms both temporally and spatially . Such regulation has been described previously for genes such as the Drosophila DOA kinase [27] and the BBG PDZ-protein [28] , among others . Embryos in which the promoter of the Alk-RB isoform has been disrupted fail to express detectable Alk protein in the VM , AS and epidermis , and exhibit an Alk loss of function phenotype , revealing that this promoter is critical for Alk expression in these embryonic tissues . However , expression of Alk in the embryonic CNS is not compromised by the removal of the Alk-RB promoter and upstream sequences , suggesting that CNS expression of Alk is independent of the VM , AS and epidermal enhancers identified here . Taken together , our results indicate a critical requirement for Alk-RB expression to ensure sufficient Alk protein levels in the VM for signaling and founder cell specification , as well as for Alk expression in the AS and epidermis where the function of Alk is currently uncharacterized . Previous reports have studied sequences within the Alk locus either by reporter activity assays [1 , 17] or ChIP-on-chip analyses [16 , 17] . Our analysis of reporter activity has identified regions upstream of Alk that are active in the AS , VM and epidermis . These coincide temporally with Alk protein expression , allowing us to define Alk VM , AS and epidermal enhancers located proximal to the Alk-RB promoter . High-throughput Y1H screens performed in this study identified a number of TFs that potentially bind to and regulate these regions of the Alk locus . In addition , a genome-wide ChIP-on-chip screen for mesodermal TFs occupancy identified a CRM upstream of the Alk locus that is active during mesoderm development [16] . This CRM maps to 2R:16 , 639 , 969 . . 16 , 640 , 341 ( relative to Dmel_Release_6 sequence assembly ) and was described to be bound by mesodermal TFs including Bin , Bap , Mef2 and Tin and Twi [12 , 15 , 16 , 29 , 30] . However , none of these factors were found in our Y1H analysis . This may reflect additional requirements for binding of some TFs , which would preclude their identification by Y1H , such as heterodimerization with co-factors or post-translational modifications . Interestingly , homozygous mutants for bin and bap still express Alk protein in the VM , suggesting that while they may be involved in the modulation of Alk expression , additional factors are also important in the regulation of Alk expression in the VM . One such factor could be the NK-4/msh-2 TF Tinman ( Tin ) which has been previously reported to bind CRMs at the Alk locus [16 , 17] . Indeed , expression of Alk in the VM is affected in tin mutant embryos ( [17] , this study ) however it is not clear if this occurs due to direct regulation of Alk expression by Tin or a general lack of induction of the VM lineage . Moreover , our analysis of AlkΔOpaBS;bap double mutants uncovers a severe decrease in Alk protein in the VM suggesting only a minor direct contribution of Tin . Interestingly , opa has been reported to be directly regulated by Tin during heart development [17 , 31] and Tin is critical for the expression of two key VM TFs bin and bap [4 , 15] . Therefore it is likely that the importance of Tin for Alk expression relies on its activating potential for these Alk-regulating TFs . Interestingly , loss of tin does not affect Alk expression in the epidermis . Reporter gene expression analyses suggest the VM Alk enhancer is located upstream of the Alk-RB isoform , in agreement with previously reported AlkE301-lacZ reporter spanning 1 , 984 bp ( Fig 1A; S1A Fig ) [17] and the 547 bp MesoCRM-880 [16] that both cover the AlkEB9 region . Our data suggests that Alk-RB expression can be activated through an upstream enhancer that is bound by Opa located within AlkEB9 . We were also able to identify additional nearby enhancer elements in AlkEB8 that integrate information from factors such as Bin and Bap that are critical to ensure precise and robust VM expression of Alk . Taken together with the earlier ChIP analyses from the Furlong and Frasch groups , our data suggest that Opa , along with mesodermal TFs such as Bin , Bap , Mef2 and Tin and Twi function in a combinatorial manner to drive robust expression of Alk in the VM ( Fig 9I ) . Our efforts to identify novel TFs involved in Alk transcriptional control by in vitro Y1H assay resulted in a cluster of TFs potentially binding the AlkEB9 sequence . Of those TF hits for which UAS-transgenes were available to test , only Opa was observed to induce cell autonomous expression of Alk when ectopically expressed . opa is a pair-rule gene [32] that encodes a zinc finger protein important during embryonic segmentation and midgut formation [26 , 33 , 34] , as well as adult head morphogenesis by direct regulation of decapentaplegic ( dpp ) transcription [23 , 35] . opa transcript is expressed in a spatially and temporally dynamic pattern , starting from stage 5 in a broad expression domain and from stage 11 onwards in two discrete domains in the VM corresponding to the first and third midgut constrictions [26 , 33] . While Opa plays a role in the differentiating midgut musculature , with opa mutants exhibiting an interrupted VM unable to form midgut constrictions during embryogenesis [26] , its role during segment formation presents a challenge when attempting to decipher the contribution of this TF more precisely . One component of this may be the regulation of Alk by Opa shown here . While we observed that opa mutants display lower levels of Alk protein in the VM , Jeb/Alk signaling is not abrogated , suggesting that while reduced , Alk protein levels are not reduced to levels under the threshold critical to drive Alk signaling . The lack of a critical role for Opa in the VM expression of Alk may reflect the importance of Alk signaling in this tissue for survival of the fly , where a more complex network of TFs may be employed to ensure rigorous Alk expression . Additional VM enhancer elements 5’ of AlkEB9 in the Alk locus are regulated in part by Bin and Bap , two TFs that are critical for VM development . Thus multiple partially redundant enhancer regions are employed to safeguard VM expression of Alk , a phenomenon that has been observed in numerous genes expressed in the Drosophila embryonic muscle [36] . Moreover , while we have tested the role of Opa and the Opa binding sites in the AlkEB9 region of the Alk locus in this work , we have done so under standard laboratory conditions , and as a result have not tested whether either Opa itself , AlkEB9 or AlkEB8 VM enhancers may play an increasingly critical role in Alk expression in more demanding environmental conditions , as it has been described for some Drosophila loci [37] . Although Alk is expressed in bin and bap mutants , our experiments combining deletion of the Opa binding region in Alk in a bin or bap mutant background suggest a combinatorial role for Bin , Bap and Opa driving VM expression of Alk [12–15] . Opa , Bin and Bap potentially act in combination with other TFs to control Alk transcription in the VM , as has been described for sloppy paired-1 ( slp1 ) activation in the somatic blastoderm in response to Opa and Runt [38] . In addition to direct regulation of Alk expression , Opa may also impact Alk expression via indirect mechanisms during embryogenesis . Further complexity arises when the regulation of opa itself in the VM is considered . It is known that Dpp signaling restricts the VM spatial expression pattern of opa to PS6-8 , with dpp mutants showing continuous opa expression throughout the VM [26] . Opa is also known to regulate dpp expression during adult head development [23] . In addition , opa is broadly expressed in the mesoderm at stage 6 potentially driving Dpp signaling . The Dpp mesodermal response consists of up-regulation of tin and bap , important regulatory genes in the dorsal mesoderm that essentially contribute to the specification of the VM [12 , 39] . Similarly , Alk activity , the FoxF forkhead domain TF Bin and the Tbx1 Org-1 , are also critical factors for expression of dpp in the VM and subsequent activation of Mad signaling in the midgut endoderm [40 , 41] . Moreover , loss of org-1 , whose expression is maintained by Alk signaling in the VM , results in decreased opa VM expression [4 , 41] , revealing a complex interplay of regulation where both Alk and Opa control each other’s expression in a spatially and temporally regulated manner . Surprisingly , in addition to a non-essential role for Opa in the regulation of Alk transcription in the VM , in this work we have been able to identify a critical role for Opa in Alk expression in the AS and epidermis . Here , in contrast to the VM , Opa appears to be required and sufficient to drive Alk expression , although the functional significance of Alk in these tissues remains uncharacterized . Expression of the AlkEB9-lacZ reporter and derivatives in which the Opa binding sites have been mutated indicate that Opa has an important function in Alk transcription through the predicted Opa BS . This is supported by the absence of detectable Alk protein in the AS and epidermis of AlkΔOpaBS mutants , where the Opa binding sites within the AlkEB9 enhancer have been deleted . Given that AlkΔOpaBS mutants are viable , it may be that Alk signaling is employed in a small population of non-essential cells that remain to be identified . Further work will be required to characterize the role of Alk in this context . We have focused here on the regulation of Alk expression during embryonic development , however , Alk is also observed in larval and adult stages . Although Alk signaling does not seem to be critical for viability post-embryogenesis , a number of important roles in the nervous system have been described [42–47] . While we have not investigated the role of Opa , Bin or Bap in Alk expression at these other stages , nor in the CNS in this study , this would certainly be of interest to address in future experiments .
Standard Drosophila husbandry procedures were employed . Drosophila strains and crosses were maintained on a potato-meal based diet . Crosses were performed at controlled 60% humidity and 25°C conditions . Fly lines used in this study are: UAS-Alk [1] , UAS-GFP ( Bloomington 4775 ) , UAS-bap . ORF . 3xHA ( FlyORF #F000006 ) , UAS-bin . ORF . 3xHA ( FlyORF #F000281 ) , UAS-jeb [6] , UAS-lacZ ( Bloomington 1776 ) , UAS-opa [35] , UAS-opaRNAi ( VDRC KK108975 ) , UAS-pnt . P1 ( Bloomington 869 ) , UAS-side ( Bloomington 9679 ) , Alk1 [2] , Alk10 [2] , bap208 [12] , Df ( 3R ) ED6058 ( Bloomington 24140 ) , bin1 ( Bloomington 1438 ) , Df ( 3L ) BSC374 ( Bloomington 24398 ) , opa1 ( Bloomington 3312 and 3222 ) , opa8 ( Bloomington 5335 ) , tin346 [12] , AlkEI6 . 5-GAL4 [1] , bap3-GAL4 [15] , en2 . 4-GAL4 ( Bloomington 30564 ) , prd-GAL4 ( Bloomington 1947 ) , twi . 2xPE-GAL4 ( Bloomington 2517 ) , HandC-GFP [48] , opa3D246 [26] , Opa4opt-lacZ and Opa4opt-KO-lacZ [23] . Alk alleles generated in this study are summarized in S1 Table . Transgenic flies generated in this study: AlkE4-GAL4 , AlkE2 . 7-GAL4 , AlkEB9-GAL4 , eve . p:empty-lacZ , AlkEB6-lacZ , AlkEB7-lacZ , AlkEB8-lacZ , AlkEB9-lacZ , AlkEB10-lacZ , AlkEB11-lacZ , AlkEB9_OpaBS-lacZ , AlkEB9_OpaKO-lacZ , AlkEB8ΔBapBS-lacZ , AlkEB8∩EB9-lacZ , AlkEB8∩EB9+50flank-lacZ and AlkEB8∩EB9+100flank-lacZ . Molecular details of the regions covered by these fragments are described in S2 Table . Genomic coordinates refer to the Dmel_Release_6 sequence assembly [49] . Embryos were stained as described [1] . Primary antibodies used were: guinea pig anti-Alk ( 1:1000 [3] ) , rabbit anti-β-galactosidase ( 1:150; Cappel 0855976 ) , chicken anti-β-galactosidase ( 1:200; Abcam ab9361 ) , mouse anti-Fasciclin III ( 1:50; DSHB 7G10 ) , rabbit anti-GFP ( 1:500; Abcam ab290 ) , chicken anti-GFP ( 1:300; Abcam ab13970 ) , mouse 16B12 anti-HA . 11 ( 1:500; Covance #MMS-101P ) , rabbit anti-Org-1 ( 1:1000 , this work ) , sheep anti-digoxygenin-AP fab fragment 1:4000 ( Roche ) . Alexa Fluor®-conjugated secondary antibodies were from Jackson Immuno Research . Embryos were dehydrated in an ascending ethanol series before clearing and mounting in methylsalicylate . Images were acquired with a Zeiss LSM800 confocal microscope or Axiocam 503 camera , processed and analyzed employing Zeiss ZEN2 ( Blue Edition ) imaging software . For analysis of protein levels , the laser , pinhole and PMT settings were adjusted on control siblings subsequently employed for imaging of mutant embryos . Fluorescence intensity measurements were quantified using Zeiss ZEN2 ( Blue Edition ) . In brief: mean fluorescence values were acquired from regions of interest ( ROI ) , corresponding to the VM or epidermis ( Alk staining ) selected in confocal sections of stage 11 embryos . This mean fluorescent intensity was corrected using a background ROI chosen from a non-stained area . Measurements were taken from 10 embryos per sample analyzed . For statistical analysis we performed a one-way ANOVA using GraphPad Prism 6 software , where n . s . stands for non-significant , ***p≤0 . 001 and ****p≤0 . 0001 . All plots are visualized as mean ±S . D . Recombinant N-terminal Org-1 protein was produced from pET30a—Org-1-N as generated by [50] was purified by His affinity chromatography and injected into rabbits for antibody generation ( Genscript ) . For in situ hybridization , fragments of Alk , gprs , CG5065 and opa were amplified from genomic DNA with the primer combinations shown in S3 Table . PCR products were cloned into the dual promoter PCRII TOPO vector ( Invitrogen ) and used as template to generate DIG-labeled in situ probes with SP6/T7 polymerases ( Roche ) . In situ hybridization of antisense probes to embryos was carried out as previously described [51] . Samples were mounted in in situ mounting media ( Electron Microscopy Sciences ) . pMW2-vectors containing the different Alk putative CRMs were generated by regular cloning techniques ( primer combinations shown in S4 Table ) and integrated into the yeast genome as described [18] . Each DNA bait yeast strain was then transformed with a library of 647 Drosophila TFs fused to GAL4 . Interaction was assessed by growing transformant yeast strains on selective plates followed by data analysis as previously described [18] . Briefly , selective growth of diploid yeast colonies was analyzed by the Matlab-based image-analysis program TIDY which quantifies bright spots , representing yeast colonies to the dark background . For every biological replicate in the screen , each bait-TF interaction was analyzed in four technical replicates resulting in quandrants of yeast colonies as shown for AlkEB9 DNA bait in the results . For generation of lacZ reporter flies , DNA sequences of for AlkEB6 to AlkEB11 were PCR amplified ( S4 Table ) and cloned into the eve . p-lacZ . attB vector [52] . In addition , the AlkEB9 DNA bait was cloned into pPT-GAL vector ( 1225 , DGRC ) to generate the AlkEB9-GAL4 construct . DNA sequences for AlkEB8ΔBapBS-lacZ , AlkEB9_OpaBS-lacZ and AlkEB9_OpaKO-lacZ were assembled by Genscript and cloned into eve . p-lacZ . attB vector for further PhiC31 directed genome integration . For generation of AlkE4-GAL4 and AlkE2 . 7-GAL4 constructs , DNA genomic regions covering 2R:16 , 638 , 503 . . 16 , 642 , 495 and 2R:16 , 638 , 510 . . 16 , 640 , 834 , respectively , were cloned into pCaSpeR-DEST6 ( 1032 , DGRC ) by the Gateway system ( primer combinations in S4 Table ) . Constructs were sequenced ( GATC Biotech ) and injected into w1118 flies , except for attB constructs which were injected into Bloomington 24482 and 24485 , for PhiC31 directed integration at 51C and 68E respectively ( BestGene Inc . ) . Deletions within the Drosophila Alk enhancer region were generated with CRISPR/Cas9 [53] . The sgRNA targeting sequences used ( listed in S1 Table ) were cloned into pBFv-U6 . 2 expression vector ( Genome Engineering Production Group at Harvard Medical School ) . Constructs expressing sgRNA were injected into vasa ( vas ) -Cas9 ( Bloomington 51323 ) embryos by BestGene Inc . Screening of deletion events was performed by PCR and further sequencing ( GATC Biotech ) . For additional complementation tests we employed balanced Alk10 or Df ( 2R ) Exel7144 flies . DNA coding sequence of opa was synthesized ( Genscript ) in frame with carboxy-terminal OLLAS and 6xHis tags and cloned into the pcDNA3 . 1 ( + ) mammalian expression vector . Binding of Opa to the AlkEB9 was analyzed by a DNA binding assay on dsDNA oligonucleotides with cell lysates from HEK-293F cells expressing Opa-OLLAS . Binding reactions were performed as described in [54] containing 10 mM Tris-HCl ( pH 8 . 0 ) , 25 mM KCl and 1 mM DTT , 1 μg poly-dIdC ( Sigma-Aldrich ) , 2 . 5% glycerol , 0 . 05% Triton X-100 , 0 . 2 mM MgCl2 and the indicated 3’-end biotin labelled probe . After 20 min incubation at room temperature , reactions were separated on a 6% native TBE-PAGE in 0 . 5x TBE buffer at 100V . DNA was transferred to nylon+ membranes ( Amersham ) , UV cross-linked to the membrane and detected by Chemoluminiscence Nucleic Acid Detection Module ( Pierce ) according to manufacturer’s indications . Competition assay was performed by addition of 100 fold molar excess of unlabeled competitor DNA to the reaction mix . Wild-type probes used for band shift experiments were Opa_SELEX and Opa_JASPAR . Mutated version were made according to for Opa_SELEX mutant [23] , and in a similar manner for Opa_JASPAR mutant . All four EMSA probe sequences are shown in S3 Table . Chromatin was prepared from approximately 100 mg of pooled collections of fixed 3–4 hour embryos . The embryos were homogenized for 1 min in 10 mM EDTA and 50 mM Tris ( pH 8 . 1 ) . After addition of SDS to a final concentration of 1% and incubation on ice for 10 min , glass beads ( 150–200 μm ) were added and the homogenates were sonicated to give sheared chromatin preparations with an average DNA size of 300–400 bp . Chromatin immuno-precipitation was performed largely as described previously [55] using an affinity-purified anti-Opa antibody raised against a truncated recombinant protein spanning from amino acids 125–507 of Opa , a region containing the DNA-binding zinc-fingers at a concentration of 0 . 5 μg/ml with 100 μg of chromatin in 1 ml of 0 . 01% SDS , 1% TritonX-100 , 1 mM EDTA , 20 mM Tris , pH 8 , 150 mM NaCl and 1x Protease Inhibitor Cocktail ( Roche ) . After overnight incubation of the chromatin and antibody at 4°C , the mixture was incubated with Protein-A Agarose ( Millipore ) for 2 hours at room temperature , followed by low-salt , high salt and LiCl washes as used in the Chromatin Immunoprecipitation Assay Kit ( Upstate Biotechnology ) . After heat reversal of protein-DNA crosslinks , protein digestion , phenol chloroform extraction and purification of the nucleic acids by ethanol precipitation the amount of recovered DNA was quantified using qPCR and a standard curve generated for each primer pair with a sample of nucleic acid purified from the input chromatin . The control primer pair produces a 115 bp amplicon located 12 . 4 kb upstream of odorant receptor 42b , a region devoid of modEncode hallmarks of cis-regulatory DNA sequences . The DESE-Opa primer pair produces a 209 bp amplicon from a central region of the slp1 DESE enhancer that requires Opa for expression [56] . The Alk primer pair produces a 140 bp amplicon that extends from 21 bp downstream of the SELEX_OpaBS to 53 bp upstream of the JASPAR_OpaBS . The ChIP values that are reported are percent precipitation relative to input DNA with error bars representing the mean ± S . D . from three technical replicates of the qPCR . The sequences of the primers are summarized in S3 Table and are as follows: Or42b forward: 5’ TCAAGCCGAACCCTCTAAAAT 3’ , Or42b reverse: 5’ AACGCCAACAAACAGAAAATG 3’ , DESE-Opa forward: 5’ TGCCGTTCGAGTCCTTTATT 3’ , DESE-Opa reverse: 5’ CGGAGATCGGAAGGTTAGTG 3” , Alk-OpaBS forward: 5’ TTGTGCGTTTCACCAATCG 3’ , Alk-OpaBS reverse: 5’ CGGACTAGCCACATCGAAC 3’ . | The Alk receptor tyrosine kinase is employed repeatedly during Drosophila development to drive signaling events in a variety of tissues . The spatial and temporal expression pattern of the Alk gene is tightly regulated . Identifying factors that influence the expression of Alk is important to better understand how Alk signaling is controlled . In this paper we characterize cis-regulatory sequences in the Alk locus and the transcription factors that bind them to govern Alk expression in the Drosophila embryo . Using a robotic protein-DNA interaction assay , we identified the Zic family transcription factor Odd-paired as a factor that binds to regulatory elements in the Alk locus . Binding of Odd-paired to Alk cis-regulatory elements varies spatially , revealing a requirement for additional transcription factors such as the NK3 and FoxF orthologues Bagpipe and Biniou in a subset of Alk-expressing tissues . Our findings provide new insight into the dynamics underlying temporal and spatial regulation of the Alk receptor during embryogenesis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"skin",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"integumentary",
"system",
"electrophoretic",
"mobility",
"shift",
"assay",
"gene",
"regulation",
"animals",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"epidermis",
"molecular",
"biology",
"techniques",
"embryos",
"drosophila",
"research",
"and",
"analysis",
"methods",
"embryology",
"gene",
"mapping",
"gene",
"expression",
"molecular",
"biology",
"insects",
"genetic",
"loci",
"arthropoda",
"anatomy",
"embryogenesis",
"genetics",
"biology",
"and",
"life",
"sciences",
"restriction",
"fragment",
"mapping",
"organisms"
] | 2017 | The Zic family homologue Odd-paired regulates Alk expression in Drosophila |
Amyloidosis describes a group of protein folding diseases in which amyloid proteins are abnormally deposited in organs and/or tissues as fine fibrils . Mouse senile amyloidosis is a disorder in which apolipoprotein A-II ( apoA-II ) deposits as amyloid fibrils ( AApoAII ) and can be transmitted from one animal to another both by the feces and milk excreted by mice with amyloidosis . Thus , mouse AApoAII amyloidosis has been demonstrated to be a “transmissible disease” . In this study , to further characterize the transmissibility of amyloidosis , AApoAII amyloid fibrils were injected into transgenic Apoa2cTg+/− and normal R1 . P1-Apoa2c mice to induce AApoAII systemic amyloidosis . Two months later , AApoAII amyloid deposits were found in the skeletal muscles of amyloid-affected mice , primarily in the blood vessels and in the interstitial tissues surrounding muscle fibers . When amyloid fibrils extracted from the skeletal muscles were subjected to Western blot analysis , apoA-II was detected . Amyloid fibril fractions isolated from the muscles not only demonstrated the structure of amyloid fibrils but could also induce amyloidosis in young mice depending on its fibril conformation . These findings present a possible pathogenesis of amyloidosis: transmission of amyloid fibril conformation through muscle , and shed new light on the etiology involved in amyloid disorders .
Amyloidosis refers to a group of protein folding disorders . Various proteins that are harmless and soluble under normal physiological conditions can undergo marked conformational changes and subsequent self-assembly outside of the cell into highly stable , insoluble amyloid fibrils with a high content of ß-sheet structures . Currently , twenty-eight different kinds of human proteins , in intact or fragmented forms , have been found to be amyloidogenic in vivo and to be associated with pathological disorders such as prion diseases , Alzheimer's disease , type II diabetes , dialysis-related amyloidosis , and familial , systemic , and sporadic amyloidosis [1] , [2] . Many factors , such as aging and epigenetic factors , including lifestyle and types of food ingested , may influence fibril formation and deposition in organs and/or tissues . Transmission of amyloid fibrils might act as an important etiological factor of amyloidosis . Exogenous amyloid fibrils could act as a seed ( nuclei ) and change the conformation of endogenous amyloid protein into that of fibrils , such as PrPSc , reactive ( AA ) and AApoAII amyloid fibrils [3] , [4] , [5] . Transmissible spongiform encephalopathies ( TSEs ) comprise a group of infectious neurodegenerative diseases that affect humans and other animals and are characterized by accumulation of the misfolded , protease-resistant prion protein PrPSc in the central nervous system [6] , [7] . It is hypothesized that TSEs are transmitted from one species to another through ingestion of urine , saliva , and/or infected meat [7] . Apolipoprotein A-II ( apoA-II ) is present in the plasma of humans , mice , rats , and fish [8] , [9] . In mice , apoA-II is the second most abundant apoprotein in serum high density lipoprotein ( HDL ) , and accumulates to form amyloid fibrils ( AApoAII ) in many organs , leading to senile amyloidosis [10] . In laboratory mice , three major alleles ( Apoa2a , Apoa2b and Apoa2c ) of the apoA-II gene encode three variants of the apoA-II protein [11] , [12] . Several genetic analyses have indicated that the Apoa2c allele markedly accelerates age-associated deposition of AApoAII [13] . Mouse AApoAII amyloidosis has been demonstrated to be a transmissible disease by a prion-like infectious process occurring through a seeding-nucleation mechanism [4] , [14] . Our group found that a single intravenous injection of a very small amount of AApoAII amyloid fibrils ( ∼10−13g ) led to systemic deposition of amyloid in young mice [15] , [16] . AApoAII amyloidosis can also be transmitted by the feces [17] and milk [18] excreted by mice with AApoAII amyloidosis . Furthermore , transmission of AApoAII amyloidosis shows a ‘strain phenomenon’ analogous to the prion strains [10] . Thus , the fibrillar nuclei or amyloid fibrils formed by the aggregation of misfolded protein monomers ( rich in ß-sheet structures ) act as seeds to induce and stabilize conversion of the native monomeric protein [19] , [20] . This mechanism provides a plausible explanation for the transmissible nature of AApoAII amyloidosis . In the present study , we found AApoAII amyloid fibrils in the skeletal muscles of AApoAII amyloid-affected mice . Unexpectedly , amyloid fibrils isolated from the muscles were demonstrated to be sufficient for the transmission of amyloidosis . These findings provide important implications for assessing the potential risk of consuming amyloid-deposited skeletal muscles in the transmission of amyloidosis .
To confirm whether apoA-II mRNA is expressed in mouse skeletal muscle , total RNA was extracted from the triceps brachii muscles in the forelimbs , the femoral quadriceps muscles in the hindlimb , the longissimus thoracis muscle in the back and the greater pectoral muscles from the breast of Apoa2cTg+/− mice and R1 . P1-Apoa2c mice . ApoA-II mRNAs were detected in several muscles obtained from these mice . Interestingly , expression levels in muscle tissues were lower than those seen in liver ( Figure 1A ) . Quantitative real time PCR analysis revealed that the expression levels of apoA-II mRNA in the muscles were about one tenth and one thirtieth of the expression levels observed in the livers from Apoa2cTg+/− and R1 . P1-Apoa2c mice , respectively , and expression levels were significantly higher in Apoa2cTg+/− mice ( Figure 1B ) . To determine whether amyloid fibrils exist in the muscles of AApoAII-deposited mice , we intravenously injected 1 µg of isolated AApoAII fibrils into six 2-month-old female Apoa2cTg+/−mice . Two months later , amyloid deposition was detected by the presence of green birefringence in Congo Red-stained tissue from four muscles of Apoa2cTg+/− mice displaying heavy amyloid deposits throughout the body . Histological examination revealed that muscles of all Apoa2cTg+/− were deposited with amyloid ( 6/6; Table 1 ) . In amyloid fibril-injected normal R1 . P1-Apoa2c mice , obvious amyloid deposition was observed only in one of three mice at two months after injection ( 1/3 ) , and all three had amyloid deposits at four months after injection ( 3/3 ) ; that is , in skeletal muscles , the deposition of AApoAII amyloidosis increased with age . Amyloid deposits were found mainly in the blood vessels of muscle tissues , but were also found in connective tissues around muscle fibers ( endomysium ) both in Apoa2cTg+/− mice ( Figure 2A , B ) and R1 . P1-Apoa2c mice ( Figure 2E , F ) . AApoAII amyloid deposition , which was observed in muscles , was further confirmed with anti-apoA-II staining ( Figure 2C , D , G , and H ) . However , no deposits of AApoAII amyloid fibrils were found in muscle tissues of the R1 . P1-Apoa2c mice ( 0/3 ) without induction by AApoAII fibrils . Amyloid fibril fractions were isolated from various muscles and apoA-II protein was detected by Western blot analysis . ApoA-II proteins were detected in amyloid fibril fractions of femoral quadriceps muscles in the hindlimb of Apoa2cTg+/− mice with AApoAII-deposition but similar results were not observed in the muscles of R1 . P1-Apoa2c control mice lacking AApoAII-deposition ( Figure 3A ) . Moreover , apoA-II was also detected in all four kinds of muscle of Apoa2cTg+/− mice ( Figure 3B ) . In R1 . P1-Apoa2c mice , two months after injection of amyloid fibrils , apoA-II was detected in greater pectoral muscles in the breast of all three mice . Four months after injection , apoA-II deposition expanded to other muscles and amounts of apoA-II increased ( Figure 3C ) . The amount of deposition was different among different muscles: greater pectoral muscles from the breast > longissimus thoracis muscle in the back > triceps brachii muscles in the fore-limbs > femoral quadriceps muscles in the pelvic-limb . To further confirm the existence of amyloid fibrils in muscle tissues , the amyloid fibril fractions of muscle tissues from the AApoAII-deposited Apoa2cTg+/− mice and R1 . P1-Apoa2c mice without AApoAII-deposition were observed by transmission electron microscopy . We found amyloid fibrils extracted from muscles only in the fractions of mice with AApoAII-deposition ( Figure 4A ) , but not in the fractions of mice without deposition . Ultrastructural analysis of cross sections of AApoAII-deposited muscle was performed by transmission electron microscopy . Bundles of amyloid fibrils deposited in endomysiums and capillary walls of skeletal muscles were observed ( Figure 4B and C ) . To elucidate whether AApoAII amyloid transmissibility existed in skeletal muscle , amyloid fibril fractions were isolated from femoral quadriceps muscles of Apoa2cTg+/− mice with AApoAII deposition ( Figure 5 ) and were injected into 2-month-old female R1 . P1-Apoa2c mice . Two months following injection , amyloid deposits were observed in the tongue ( 6/6 ) , lungs ( 6/6 ) , stomach ( 6/6 ) , heart ( 6/6 ) , intestine ( 4/6 ) , and skin ( 1/6 ) of six injected mice ( Table 2 ) . The mean AI was 1 . 19 . In contrast , significantly less amyloid deposition was observed when the mice were injected with fractions isolated from R1 . P1-Apoa2c mice without AApoAII-deposition; the mean AI was 0 . 26 ( p = 0 . 0062 ) . Additionally , we found that injection of fibril fractions extracted from no AApoAII deposited muscles from R1 . P1-Apoa2c mice induced a small amount of amyloid deposition ( 2/5 mice had AApoAII deposits and mean AI = 0 . 10 ) . To confirm this , amyloid fibril fractions were extracted from muscles of young ( 2-month-old ) female R1 . P1-Apoa2c mice . Although these fractions induced secondary transmission in mice , neither apoA-II nor AA could be detected in the fraction by Western blot analysis ( Figure 5 ) . Unexpectedly , slight amyloid depositions were observed in the tongue ( 5/7 ) and stomach ( 3/7 ) of seven injected mice two months after injection; mean AI = 0 . 20 . Amyloid fractions denatured by guanidine hydrochloride were injected into six 2-month-old female R1 . P1-Apoa2c mice . No amyloid deposition was detected in any of these mice two months later ( 0/6 ) ( Table 2 ) .
Our previous work in senescence-accelerated mice determined that AApoAII amyloid fibrils are deposited throughout the body , including in the liver , spleen , stomach , intestine , heart , kidneys , lungs , tongue , skin , gonads , adrenal glands , salivary and thyroid glands , thymus , mesenteric lymph nodes , epineurium of the sciatic nerve and blood vessels in various tissues; however , no evidence of amyloid deposition was found in brain parenchyma or bone marrow in the vertebral body of the lumbar spine [21] . Although amyloid fibrils were also detected in the musculoskeletal systems , there were no detailed descriptions nor further studies on AApoAII amyloid deposition in the skeletal muscles [21] . The current study demonstrates that skeletal muscle tissue is capable of propagating AApoAII amyloidosis in mice . We first detected AApoAII amyloid fibrils in four skeletal muscles in different body regions using immunohistochemistry , Western blot analysis and electron microscopy . Amyloid deposits were observed in blood vessels and interstitial tissues surrounding muscle fibers . However , no AApoAII was observed in muscle cells or in the nerve fibers in which prion proteins were previously detected [22] . Although apoA-II mRNA was detected in muscle tissues ( Figure 1 ) , it is unclear whether apoA-II protein in AApoAII fibrils around muscle fibers originates from muscle cells or blood . Interestingly , the amounts of amyloid deposition in the skeletal muscles differed among body regions; that is , breast > back > forelimb > hindlimb . This order is different from that observed in mice inoculated with prion protein [23] . First , we observed amyloid deposition in Apoa2cTg+/− mice in which apoA-II protein is overexpressed in various tissues under the control of a ubiquitous promoter , and found that serum levels of apoA-II increased the susceptibility to induction of AApoAII [24] . As a result , AApoAII deposits in muscles were found in Apoa2cTg+/− mice . Second , we also observed AApoAII deposits in the muscles of R1 . P1-Apoa2c mice in which apoA-II was expressed under an endogenous promoter/enhancer . According to the above results , it was found that AApoAII is deposited in the skeletal muscles as part of a universal phenomenon . Next , we demonstrated that intravenous injection of amyloid fibrils extracted from muscle tissues could transmit amyloidosis depending on fibril-conformation . That is , transmissibility was lost following denaturation with 6 mol/L guanidine hydrochloride . In previous studies , AApoAII amyloid fibril was extremely efficient in inducing amyloidosis following doses of less than 1 pg; moreover , amyloidosis could be initiated after oral ingestion of AApoAII fibrils [16] , [17] , [25] . Thus , the infectious ability of skeletal muscle raised the possibility that mouse AApoAII amyloidosis may result , in part , from dietary exposure to amyloid fibrils through consumption of muscle/meat containing amyloid materials . Animal muscles , an important food component for most humans , have been examined in several studies for the presence of TSE transmissibility [23] , [26] . Recently , it was reported that high prion titers and the disease-causing isoform of the prion protein PrPSc appear in the skeletal muscles of mice , hamsters and sheep inoculated with prion agents [27] , [28] , [29] , [30] , and in deer infected with chronic wasting disease [31] . Furthermore , PrPSc is also present in skeletal muscle samples of sporadic Creutzfeldt-Jacob disease ( CJD ) in humans [32] , and has been demonstrated to be present in the nerve fibers of skeletal muscles tissue [23] . Although some findings are contradictory to the above reports [33] , elucidation of the contribution of muscle tissues to transmission is important for the prevention of prion-related disorders . Prion-like transmission has been reported in mouse inflammation-associated amyloid A ( AA ) amyloidosis [34] . Dietary supply of amyloid fibrils might also be a trigger in the development of AA amyloidosis , especially for a susceptible population . Notably , it was reported that 71 . 4% of skeletal muscles from cows with systemic AA amyloidosis stained positive with anti-AA antibody [35] . Although an unexpectedly high incidence of visceral AA-amyloidosis in aged slaughtered cattle in Japan was reported , and isolated AA amyloid fibrils exhibited amyloid-enhancing factor activity , amyloid deposition in the skeletal muscles was rare [36] , [37] . Thus , these studies support the idea of the transmissibility of systemic AApoAII and AA amyloidosis from skeletal muscle . Additionally , we found that injection of fibril fractions extracted from either control non-AApoAII injected or , no AApoAII deposited muscles of young R1 . P1-Apoa2c mice induced a small amount of amyloid deposition , although these fractions contained neither apoA-II nor AA detectable by Western blot analysis . It is possible that extracts contain trace amounts of AApoAII amyloid fibrils or oligomers that could not be detected by available techniques and these undetectable AApoAII peptides might be transmissible like PrPres [38] . Alternatively , components other than AApoAII amyloid fibrils might induce amyloid deposition . In R1 . P1-Apoa2c mice , AApoAII amyloid can be seeded by various heterogeneous amyloid fibrils ( cross-seeding ) [39] and unexpectedly many kinds of proteins have been reported to form amyloid fibril-like structures [40] , [41] . For example , collagen fibrils and glycosaminoglycans , supportive structures in skeletal muscle , appear to be actively involved in the induction of ß2-microglobulin amyloid fibril formation [42] . Elevated levels of prion protein in muscle lead to myopathy and neurogenic muscle atrophy in affected patients [43] , [44] , [45] . Accumulated amyloid proteins have been found in inclusion-body myositis and are toxic to skeletal myoblasts [46] . Although we observed amyloid deposits around skeletal muscle fibers after inducing amyloidosis , further studies will be necessary to examine possible myopathy and/or toxicity of these deposits . In summary , apoA-II , which is present in the plasma of humans , mice , rats , and fish [6] , [7] , has been demonstrated in the form of amyloid fibrils with transmissibility in mouse muscle . AApoAII amyloid fibrils were detected in various skeletal muscles , especially in the pectoral muscles . The verification of this transmission pathway is valuable for understanding the pathogenesis and etiology of amyloidosis .
All experimental procedures were pre-approved by Division of Laboratory Animal Research of Shinshu University and were performed according to the guidelines of Division of Laboratory Animal Research of Shinshu University . R1 . P1-Apoa2c is a congenic strain of mice with the amyloidogenic Apoa2c allele from the SAMP1 strain in the genetic background of SAMR1 [13] . Apoa2c transgenic mice ( Apoa2cTg+/− ) were established in the genetic background of R1 . P1-Apoa2c [24] . These strains were maintained by sister-brother mating in the Division of Laboratory Animal Research , Research Center for Human and Environmental Science , Shinshu University . Mice were raised under specific pathogen-free ( SPF ) conditions at 24±2°C with a light-controlled regimen ( 12-hour light/dark cycle ) . A commercial diet ( MF; Oriental Yeast , Tokyo , Japan ) and tap water were provided ad libitum . In this study , only female mice were used to avoid AA amyloidosis and/or other adverse impacts caused by fighting or other behaviors among mice reared in the same cage . All experiment procedures were carried out in accordance with the Regulations for Animal Experimentation of Shinshu University . Total RNAs were extracted from the skeletal muscles and liver ( as control ) using RNeasy Mini Kit ( Qiagen , Hilden , Germany ) according to the manufacturer's instructions . First-strand cDNA was synthesized from 1 µg total RNA of each muscle tissue ( First-strand cDNA Synthesis Kit; Amersham Pharmacia Biotech , Piscataway , NJ ) and subjected to PCR amplification with Taq DNA polymerase ( Promega; Madison , WI ) . The reverse transcriptase-polymerase chain reaction ( RT-PCR ) amplification was carried out in a 50-µl reaction mixture containing 200 µM each dNTP , 1× buffer containing 1 . 5 mM MgCl2 , 0 . 1 µM each primer , and 1 . 25 U of Taq DNA polymerase . The cycling parameters for RT-PCR were initial denaturation for 1 minute at 94°C followed by 23 cycles of 30 sec at 94°C , 30 sec at 55°C , and 1 min at 72°C . A 5-µl aliquot of the PCR product was subjected to 3% agarose ( Takara Bio , Otsu , Japan ) gel electrophoresis . SYBR Premix Ex Taq II ( Takara Bio ) was used as the fluorescent marker to monitor DNA accumulation in quantitative real time PCR analysis with the 7500 Real-Time PCR System ( Applied Biosystems Life Technologies , Tokyo Japan ) . The primers used for PCR amplification of apoA-II mRNA were as follows: A2/acc-F ( 5′-AAGAGACAGGCGGACGGACA-3′ ) and A2/acc-R ( 5′-GAGGTCTTGGCCTTCTCCAC-3′ ) . The AApoAII amyloid fibril fraction was isolated as a water suspension from the livers of a 20-month-old R1 . P1-Apoa2c mouse as described previously [10] . Purified amyloid fibrils were re-suspended at a concentration of 1 . 0 mg/ml in distilled water ( DW ) . One milliliter of this solution was put into an Eppendorf tube and sonicated on ice for 30 sec with an ultrasonic homogenizer VP-5S ( Tietech Co . , Ltd . , Tokyo , Japan ) at power level 4 . This procedure was repeated five times at 30 sec intervals . Sonicated amyloid samples were then immediately injected into the caudal vein of mice to induce AApoAII amyloidosis . Six 2-month-old female Apoa2cTg+/− mice were injected intravenously with 1 µg of AApoAII fibrils to induce AApoAII amyloidosis . Two months later the mice were sacrificed and the triceps brachii muscles in the forelimbs , the femoral quadriceps muscles in the hindlimb , the longissimus thoracis muscles in the back , and the greater pectoral muscles from the breast were dissected . Half of the tissue was kept at −80°C and the other half was fixed in 10% neutral buffered formalin , embedded in paraffin , and cut into 4-µm sections . Six 2-month-old female R1 . P1-Apoa2c mice were injected intravenously with 100 µg of AApoAII fibrils; three were sacrificed after two months and the other three were sacrificed after four months . Muscle tissue was dissected and either stored or fixed and embedded in the same fashion as the Apoa2cTg+/− transgenic mice . Three female R1 . P1-Apoa2c littermates not injected with fibrils , were sacrificed at two months as controls . AApoAII amyloid fibril fractions were isolated from the muscle of amyloid fibril-injected mice by Pras' method [47] . Thawed muscles ( 0 . 1 g ) were sonicated twice for 30 sec with a 30-second rest interval in 1 . 0 ml of 0 . 15 M NaCl on ice using an ultrasonic homogenizer VP-5S ( Tietech Co . , LTD , Tokyo , Japan ) at power level 4 . The homogenate was centrifuged at 40 , 000×g for 20 min at 4 . 0°C , after which the supernatant was discarded and the pellet was re-suspended in 1 . 0 ml 0 . 15 M NaCl . The sonication and centrifugation were repeated two more times , and the pellet was suspended in 1 . 0 ml deionized DW ( DDW ) and centrifuged after sonication once more . The pellet was re-suspended with DDW and sonicated . Following centrifugation at 30 , 000×g for 20 min at 4 . 0°C , the supernatant containing amyloid fibrils was collected and used for Western blotting , transmission electron microscopy analysis and for the secondary transmission experiment . Deposition of amyloid fibrils was identified by the appearance of green birefringence in Congo Red-stained sections [48] visualized under polarizing microscopy . AApoAII amyloid fibril proteins were identified immunohistochemically using the avidin-biotinylated horseradish peroxidase complex method with specific antiserum against mouse apoA-II ( 1∶3000 ) [49] . Isolated amyloid fibril fractions ( 25 µg ) from the muscles were separated on Tris-Tricine sodium dodecyl sulfate-polyacrylamide ( 16 . 5% [w/v] acrylamide ) electrophoresis ( SDS-PAGE ) gels [50] . Proteins on the gel were electrophoretically transferred to Immuno-Blot polyvinylidene difluoride membrane ( 0 . 2 µm pore size; Bio-Rad , Hercules , CA , USA ) . Proteins on the membrane were detected with rabbit anti-mouse apoA-II antiserum ( 1∶3000 ) , followed by peroxidase-conjugated goat IgG against rabbit immunoglobulin ( 1∶1000; ICN Pharmaceuticals , Inc . , Aurora , OH , USA ) . Immunoreactive proteins were visualized with ECL reagents ( Amersham Biosciences , Buckinghamshire , England ) . The film ( Amersham Biosciences , Buckinghamshire , England ) was exposed for 3 min . Aliquots ( 20 µl; 0 . 5 µg/µl ) of amyloid fibril fractions isolated from the muscles were applied to 400-mesh collodion-coated copper grid ( Nissin EM Co . , Ltd . , Tokyo , Japan ) for 1 min and subjected to negative staining with 1% phosphotungstic acid ( pH 7 . 0 ) for 1 min . The negatively stained samples were observed with a JEOL 1200 EX electron microscope ( JEOL , Tokyo , Japan ) operated at 80 kV . Electron micrographs were taken with a Gatan multiscan camera model 791 with Gatan digital micrograph software version 3 . 6 . 4 ( Gatan , Pleasanton , CA , USA ) . For ultrastructural analysis by electron microscopy , greater pectoral muscles were thinly sliced and placed in 2 . 5% glutaraldehyde at 4°C overnight . The tissue was rinsed twice with 0 . 1 M phosphate-buffered saline ( PBS ) and post-fixed with 1% osmium tetroxide on ice for 1 hour . Then , the tissue underwent a graded ethanol dehydration series and was infiltrated using a mixture of one-half propylene oxide and one-half resin for one hour . One hour later , the tissue was embedded in resin for four hours and then polymerized at 37°C for 10 hours followed by 60°C for 24 hours . One hundred nanometer sections were cut and stained with 4% uranyl acetate for 20 minutes and 0 . 5% lead citrate for five minutes . The sections were observed with a JEM-1400 transmission electron microscope ( JEOL , Tokyo , Japan ) operated at 80 kV . Electron micrographs were taken with a Gatan multiscan camera with Gatan digital micrograph software version 1 . 81 . 78 ( Gatan , Pleasanton , CA , USA ) . Amyloid fibril fractions were isolated from the muscles of AApoAII-deposited Apoa2cTg+/− mice or R1 . P1-Apoa2c mice without AApoAII-deposition . 100 µl ( 2 . 5 µg/µl ) of the amyloid fibril fractions were injected into two-month-old female R1 . P1-Apoa2c mice , and after two months , the mice were sacrificed and the intensity of AApoAII amyloid deposition was determined semi-quantitatively using the amyloid index ( AI ) . The AI was determined by taking the mean value of the scores of amyloid deposition ( graded from 0 to 4 ) in the seven major organs ( liver , spleen , tongue , heart , intestine , stomach , and skin ) stained with Congo Red as described previously [10] . Amyloid fibril fractions were denatured in a solution of 6 mol/L guanidine hydrochloride , 0 . 1 mol/L Tris-HCl ( pH 10 . 0 ) , and 50 mmol/L dithiothreitol ( 1 . 0 mg/mL ) for 24 hours at room temperature with gentle stirring . Denatured amyloid fractions were dialyzed quickly against 10 mmol/L NH4HCO3 solution . The solution was injected into two-month-old female R1 . P1-Apoa2c mice that were treated as described above . We used the StatView software package ( Abacus Concepts , Berkeley , CA , USA ) to perform statistical analyses . Significant differences in the value of AI among the various groups of mice were examined using the Mann-Whitney U-test . | “Amyloidosis” , a group of protein folding diseases characterized by deposition of fine fibrils in tissues , is a common disorder of protein metabolism and can be acquired , inherited and/or age-associated . Recently , prion-like transmission has been found in various amyloidoses . AApoAII amyloid fibrils in mouse senile amyloidosis have exhibited transmissibility . For instance , ingested AApoAII amyloid fibrils , which were excreted from mice and contained in feces or milk , function as seeds for changing apoA-II amyloid precursor protein to the fibrillar form and cause mouse senile amyloidosis . However , transmissibility through other pathways has not yet been established . Here , we induced AApoAII systemic amyloidosis in transgenic Apoa2cTg+/− and normal R1 . P1-Apoa2c mice to analyze the transmissibility of mouse senile amyloidosis through muscle tissues . In this study , we not only detected AApoAII deposited in various skeletal muscles , but also found that it could induce secondary transmission of AApoAII amyloidosis . This is the first evidence of transmission through skeletal muscles in non-prion systemic amyloidosis . This pathway of transmission provides new insight into the potential for food-borne pathogenesis and etiology of systemic amyloidosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"pathology",
"biochemistry/protein",
"folding",
"infectious",
"diseases",
"pathology/histopathology"
] | 2010 | Mouse Senile Amyloid Fibrils Deposited in Skeletal Muscle Exhibit Amyloidosis-Enhancing Activity |
It has been shown recently that modification of peptidoglycan by O-acetylation renders pathogenic staphylococci resistant to the muramidase activity of lysozyme . Here , we show that a Staphylococcus aureus double mutant defective in O-acetyltransferase A ( OatA ) , and the glycopeptide resistance-associated two-component system , GraRS , is much more sensitive to lysozyme than S . aureus with the oatA mutation alone . The graRS single mutant was resistant to the muramidase activity of lysozyme , but was sensitive to cationic antimicrobial peptides ( CAMPs ) such as the human lysozyme-derived peptide 107R-A-W-V-A-W-R-N-R115 ( LP9 ) , polymyxin B , or gallidermin . A comparative transcriptome analysis of wild type and the graRS mutant revealed that GraRS controls 248 genes . It up-regulates global regulators ( rot , sarS , or mgrA ) , various colonization factors , and exotoxin-encoding genes , as well as the ica and dlt operons . A pronounced decrease in the expression of the latter two operons explains why the graRS mutant is also biofilm-negative . The decrease of the dlt transcript in the graRS mutant correlates with a 46 . 7% decrease in the content of esterified d-alanyl groups in teichoic acids . The oatA/dltA double mutant showed the highest sensitivity to lysozyme; this mutant completely lacks teichoic acid–bound d-alanine esters , which are responsible for the increased susceptibility to CAMPs and peptidoglycan O-acetylation . Our results demonstrate that resistance to lysozyme can be dissected into genes mediating resistance to its muramidase activity ( oatA ) and genes mediating resistance to CAMPs ( graRS and dlt ) . The two lysozyme activities act synergistically , as the oatA/dltA or oatA/graRS double mutants are much more susceptible to lysozyme than each of the single mutants .
In humans , lysozyme is found in a wide variety of fluids , such as tears , breast milk , and respiratory and saliva secretions , as well as in cells of the innate immune system , including neutrophils , monocytes , macrophages , and epithelial cells [1 , 2] . Lysozyme is an important protein in the innate defense response against invading microorganisms and acts on bacteria by hydrolyzing the ß-1 , 4 glycosidic bonds between N-acetylmuramic acid ( MurNAc ) and N-acetylglucosamine ( GlucNAc ) , resulting in degradation of peptidoglycan ( PG ) , and subsequent cell lysis [3 , 4] . Most bacterial species are sensitive to lysozyme , but some important human pathogens , such as Staphylococcus aureus , Neisseria gonorrhoeae , and Proteus mirabilis , are resistant . The mechanisms behind the high resistance of S . aureus to lysozyme are unknown , although several studies suggest that O-acetylation at position C-6 of the MurNAc residue contributes to lysozyme resistance [5–9] . Recently , we were able to prove that indeed O-acetyltransferase A ( OatA ) of S . aureus is responsible for O-acetylation of the PG , and this leads to resistance to the muramidase activity of lysozyme [10] . We also showed that the MurNAc was O-acetylated only in pathogenic , lysozyme-resistant staphylococci ( e . g . , S . aureus , S . epidermidis , S . lugdunensis , and others ) . All nonpathogenic species ( e . g . , S . carnosus , S . gallinarum , or S . xylosus ) were lysozyme sensitive and lacked PG-specific O-acetylation . Therefore , OatA can be regarded as a general virulence factor [11] . Although the oatA mutant was less resistant to lysozyme than the wild type ( WT ) S . aureus , it still was more resistant than , for example , Micrococcus luteus , suggesting that other factors , such as a high degree of peptide cross-linking , may also contribute to lysozyme resistance [12] . Recently , we showed that the presence of wall teichoic acid ( WTA ) increased lysozyme resistance [13] . One also has to consider that lysozyme does not only comprise muramidase activity but also antimicrobial peptide activity , as demonstrated by catalytically inactivate lysozyme or peptides isolated from digested lysozyme , and by synthetic lysozyme-derived peptides [14–17] . Here , we show that the extremely high resistance of S . aureus to lysozyme can be genetically dissected as a ) resistance to muramidase activity and b ) resistance to inherent cationic antimicrobial peptide ( CAMP ) activity . Furthermore , we characterized via transcriptome analysis the two-component system ( TCS ) , GraRS , which , in addition to many virulence genes , also controls the dlt operon to mediate resistance to lysozyme and other CAMPs .
In our search for highly susceptible lysozyme mutants in S . aureus , we isolated two Tn917 transposon mutants in SA113oatA::kan that revealed higher sensitivity to lysozyme than the oatA mutation alone . Chromosomal sequencing of the flanking Tn917 insertion sites revealed that Tn917 was inserted in SA0615 [18] . SA0615 and the upstream gene SA0614 have the features of a typical TCS and were recently named GraRS ( glycopeptide resistance-associated ) , because overexpression of GraR ( response regulator ) and GraS ( sensor histidine kinase ) increased vancomycin resistance [19] . To further study the role of TCS in lysozyme resistance , we constructed a deletion mutant by substituting graRS with an erythromycin B cassette to yield SA113 graRS::erm ( Figure 1 ) . In addition , we also constructed an oatA::kan/graRS::erm double knockout . Sequencing and complementation with pTXgraRS , a vector in which the graRS genes are induced into expression by xylose , confirmed the correct replacement . Whereas the oatA/graRS double mutant was highly susceptible to lysozyme , both single mutants were only marginally affected , but were still more sensitive than the WT , which is completely lysozyme resistant ( Figure 2A–2D ) . The oatA/graRS double mutant was much more lysozyme sensitive than each of the single mutants . This hypersensitivity of the double mutant can be explained by dual activities of lysozyme that act in a synergistic way . To study this phenotype in more detail , we investigated whether the graRS single mutant is affected by the muramidase activity of lysozyme . Indeed , the isolated PG from the graRS single mutant was completely resistant to lysozyme hydrolysis , in contrast to the oatA mutant . As expected , PG of the oatA/graRS double mutant was also hydrolysed , although the sensitivity was less pronounced , as in the oatA single mutant ( Figure 3 ) . Therefore , the increased sensitivity of the double mutant likely came from its higher susceptibility to lysozyme's CAMP activity . This was confirmed by the addition of LP9 , polymyxin B , or gallidermin to a growing culture , which caused immediate growth arrest in the graRS mutant , whereas the WT was much less affected ( Figure 4A and 4B ) , and only the lantibiotic gallidermin inhibited the WT . In addition , we demonstrated that heat-inactivated lysozyme exhibits CAMP activity , but no muramidase activity . Heat-inactivated lysozyme showed no activity ( neither lytic nor CAMP activity ) to the oatA mutant or to the isolated PG of oatA , but it was able to inhibit the growth of the oatA/graRS double mutant ( Figures 2B , 2D , and 3 ) . This result suggests that GraRS controls genes involved in CAMP resistance . This effect was not only achieved with hen egg-white , but also with human lysozyme . To find out which genes are responsible for the high susceptibility to CAMPs in the graRS mutant , we carried out a comparative transcriptome analysis of the WT strain and the graRS mutant . We detected 115 genes whose mRNAs were up-regulated ( Table 1 ) and 133 genes whose mRNAs were down-regulated by GraRS ( Table 2 ) . The complete list of up- and down-regulated genes with their National Center for Biotechnology Information PID numbers is presented in Dataset S1 . In order to give an impression of which genes are controlled by GraRS , some examples are mentioned below . In the graRS mutant , genes that are involved in RNA and amino acid synthesis and glycolysis shows highly gene transcription rates . In particular , the urease genes ( ureA-G ) all 12 pur genes were 2- to 32-fold up-regulated as compared to the WT , whereas purR ( repressor ) appeared not to be influenced by GraRS . Interestingly , the amount of oatA transcript increased in the graRS mutant , which could explain the slightly higher resistance of the graRS mutant to the muramidase activity of lysozyme ( Figure 3 ) . A number of genes that were down-regulated included global regulators ( rot , sarS , mgrA ) , cell surface protein encoding genes ( the Ser-Asp rich fibrinogen-binding proteins SdrC and SdrE ) , the major autolysin gene ( atlA ) and an autolysin/adhesin gene ( aaa ) [20] , exoprotein encoding genes ( hlb , hlgA , B , lukM , F , and geh ) , transporter encoding genes ( essA/essC , oppB , and norB ) , capsule encoding genes ( capA , H , I , J , K ) and PIA encoding genes ( icaADBC ) , genes responsible for d-alanyl esterification of teichoic acids ( TAs ) ( dltA , B , D ) , and the alanine dehydrogenase gene ( ald1 ) . The pronounced decrease of expression of the ica [21–23] and dlt operons [24] and atlA [25] explains why the graRS mutant showed a biofilm-negative phenotype on microtiter plates ( unpublished data ) . With a few genes , such as rot , ureC , and dltA , we verified the transcriptome data by reverse transcriptase ( RT ) -PCR ( Table 3 ) . Next , we asked which of the 115 less expressed genes in the graRS mutant were responsible for the increased susceptibility to CAMPs . A most likely candidate was the dlt operon ( encoding enzymes involved in d-alanylation of TAs ) . Its transcript was decreased 2 . 1-fold to 2 . 9-fold as compared to WT , and indeed , the d-alanylation of TAs was decreased 46 . 7% in the graRS mutant compared to WT ( Table 3 ) . It has been previously shown that inactivation of the dlt operon in S . aureus confers sensitivity to defensins , protegrins , and other antimicrobial peptides [26] . The observed decrease of ald1 transcription by a factor of 3 . 5 is in line with the decreased dlt transcription . Ald1 is the alanine dehydrogenase , which is involved in the synthesis of l-alanine . Because the dlt operon is less expressed in the graRS mutant , we investigated lysozyme susceptibility with a dltA deletion mutant , which is well-known to be sensitive to CAMPs [26] . Indeed , the dltA mutant was more sensitive to lysozyme ( Figure 2E ) ; however , this sensitivity was not due to its muramidase activity , as the isolated PG of the dltA mutant was not hydrolyzed by lysozyme ( Figure 3 ) . Furthermore , growth of the dltA mutant was inhibited whether active or heat-inactivated lysozyme was applied ( Figure 2E ) . When the susceptibility of graRS and dltA mutants to LP9 , polymyxin B , and gallidermin were compared , both mutants were similarly more susceptible to these CAMPs ( Figure 4B and 4C ) . However , there were two distinctions: a ) the susceptibility of the dltA mutant was more pronounced than that of the graRS mutant , and b ) even in the presence of gallidermin or polymyxin B , the graRS mutant started to grow after some time and reached the same optical density ( OD ) values after 24 h as the control culture lacking CAMPs . In contrast , the dltA mutant remained sensitive to gallidermin and polymyxin B and was unable to resume growth . In the presence of LP9 , growth resumed after a similar lag period as in the graRS mutant; this can possibly be explained by its proteolytic degradation . Not only the single but also the double mutants oatA/graRS and oatA/dltA were sensitive to the CAMP activity of LP9 , although the susceptibility was less pronounced as with the graRS and dltA single mutants . However , the oatA single mutant was completely resistant to LP9 , indicating that oatA is resistant to CAMPs ( Figure 4D–4F ) . With respect to gallidermin- and polymyxin B–induced cell lysis , it has been observed that CAMPs such as lantibiotics induce autolysis in staphylococci by increasing PG hydrolase activity [27] . We assume that gallidermin and polymyxin B , which are also CAMPs , very likely have a similar effect . We asked whether the increasing insensitivity of the graRS mutant after prolonged growth is some short lasting CAMP-induced adaptation or whether it is based on selection of resistant mutants . To answer this question , we inoculated from a 24-h graRS culture treated with polymyxin B ( Figure 5B ) a new culture and challenged it again with polymyxin B ( Figure 5C ) . The subculture revealed no growth retardation , which suggests that the graRS phenotype is unstable and that polymyxin B–resistant revertants were quickly selected . Since the dltA revealed a stable phenotype , we assume that in the selected revertants dltA expression was increased to WT levels . The highest susceptibility to lysozyme was observed with the oatA/dltA double mutant , which was more than 66-fold and 333-fold more sensitive to lysozyme than the dltA and oatA single mutants , respectively ( Figure 2B , 2E , and 2F; Table 4 ) . The oatA/graRS mutant is not quite as sensitive as the oatA/dltA mutant . Another difference is that the oatA/dltA mutant stays lysozyme sensitive even after 24 h of cultivation ( Figure 2D and 2F ) , indicating that the dltA mutant phenotype cannot easily revert to the WT phenotype . The lower susceptibility of the oatA/graRS double mutant can possibly be explained by the fact that the TA in this mutant still contains 53 . 3% d-alanyl esters , whereas the dltA mutant completely lacks d-alanylation in its TAs ( Table 3 ) . The high susceptibility of the double mutants is based on the dual activities of lysozyme: a ) the oatA mutant is sensitive to the muramidase activity of lysozyme but is insensitive to CAMPs ( Figures 2B , 3 , and 4F ) , and b ) the dltA and graRS mutants are sensitive to CAMPs , but insensitive to the muramidase activity of lysozyme ( Figures 3 , 4B , and 4C ) . The extremely high lysozyme susceptibility of the oatA/dltA double mutant can only be explained by a synergistic effect of the two activities . Mutanolysin is a muramidase that is able to hydrolyze O-acetylated PG [28] but does not normally cause cell lysis in WT S . aureus or its graRS mutant at a concentration of 100 μg/ml . However , when the graRS mutant was treated with mutanolysin in combination with lysozyme or LP9 , the lytic activity ( indicated by decrease in OD ) was strongly increased ( Figure 5A ) . Because the O-acetylated graRS mutant is insensitive to the catalytic activity of lysozyme , we assume that mutanolysin acts through its lytic activity , and LP9 and lysozyme through their CAMP properties . We have not investigated how the stimulating effect of lysozyme and LP9 on cell lysis is accomplished . However , we assume that it is caused by the concerted action of PG hydrolysis by mutanolysin and induced autolysis by lysozyme and LP9 , as mentioned above . The minimal inhibition concentration ( MIC ) values for lysozyme , polymyxin B , and gallidermin in WT and various mutants are summarized in Table 4 . Both the WT and the graRS mutant were completely resistant to lysozyme at a concentration of 50 mg/ml . However , the graRS mutant was 17- and 4-fold more susceptible to polymyxin B or gallidermin . The sensitivity to the CAMPs is very likely due to the aforementioned decrease in expression of the dlt operon , which corresponds with decreased d-alanylation of the TAs . The oatA mutant was more susceptible to lysozyme than the graRS mutant , but , similar to WT , was completely insensitive to heat-inactivated lysozyme or CAMPs , indicating that oatA is only sensitive to the muramidase activity of lysozyme . The oatA/graRS double mutant was almost 17-fold more sensitive to lysozyme than the oatA mutant , which can be explained by the fact that this double mutant is sensitive to both the muramidase and the CAMP activities of lysozyme . The two activities exert a synergistic effect on the double mutant . The dltA single mutant was over 25-fold more sensitive to lysozyme than the WT and 5-fold more sensitive than the oatA single mutant , demonstrating the importance of lysozyme's CAMP activity . Furthermore , the dltA mutant exhibited the highest susceptibility to polymyxin B and gallidermin , but was completely insensitive to lysozyme's muramidase activity ( Figure 3 ) . With a MIC of only 30 μg/ml , the oatA/dltA double mutant revealed the highest susceptibility to lysozyme . Indeed , it has a 20-fold greater sensitivity to lysozyme than the oatA/graRS double mutant . The oatA/dltA double mutant is 333-fold and 66-fold more sensitive than the single oatA or dltA mutants , which illustrates the extremely high synergistic effect of lysozyme when it can exert both muramidase and CAMP activities . Overexpression of graRS in the graRS mutant or the WT by pTXgraRS resulted in an approximately 2-fold increase in polymyxin B resistance , indicating that even in WT cells , CAMP resistance can be further increased .
One of our research aims was to identify genes involved in staphylococcal lysozyme resistance . We have already elucidated two genes and corresponding enzymes that contribute to resistance against the muramidase activity of lysozyme . Since the target of muramidase is PG , it is not surprising that the mechanism of resistance is masking PG by modification . In S . aureus there are two PG modifications that are involved in resistance to lysozyme's muramidase activity . One modification is O-acetylation catalyzed by the PG-specific O-acetyltransferase A , OatA , and we have shown that the oatA mutant is more susceptible to the muramidase activity of lysozyme than the WT [10] . The other modification is WTA [29] that is covalently linked to the same C6 position in MurNAc as in the O-acetyl group . TagO is a specific UDP-N-acetylglucoseamine transferase , which is involved in the first step of WTA synthesis . The tagO deletion mutant completely lacks WTA [30] . Although the tagO mutant still shows high lysozyme resistance , a oatA/tagO double mutant , however , is much more susceptible to lysozyme's muramidase activity than the oatA mutation alone [13] . Here , we show that the high lysozyme resistance of S . aureus is not only based on resistance to the muramidase activity of lysozyme , but also to its inherent CAMP resistance . The described global two-component regulator , GraRS , was identified in an oatA-minus background by increased lysozyme susceptibility in an oatA/graRS double mutant . The graRS mutant was more susceptible to CAMPs than the WT . We assume that the reason for the increased susceptibility of the graRS mutant was a decrease in dlt expression , and consequently , GraRS up-regulates dlt expression . The Dlt enzymes modify TAs by the incorporation of d-alanine esters rendering the cells resistant to CAMPs , very likely by repulsion [26] . We showed that the dltA mutant is even more susceptible to lysozyme-derived LP9 and other CAMPs than the graRS mutant , because in the dltA mutant , d-alanine esters were completely absent in TAs , the mutant was stable , and no revertants were observed . Heat-inactivated lysozyme does not affect either the growth of the oatA or that of the graRS mutant . The latter effect is surprising , as the graRS mutant is sensitive to the other CAMPs ( LP9 , gallidermin , polymyxin B ) . However , the oatA/graRS mutant was sensitive to heat-inactivated lysozyme , suggesting that the bulky molecule has better access to the cell envelope when the PG is de-O-acetylated . Likewise , sensitivity of the dltA mutant to heat-inactivated lysozyme can also be explained by better access to the cell envelope because of the lack of d-alanine esters in TAs . The next interesting question was , how do CAMPs act in the dltA , oatA/graRS , or oatA/dltA mutants ? Killing of Gram-negative bacteria could be demonstrated by lysozyme-derived peptides that were transported through the outer membrane and damaged the inner membrane by pore formation [17] . Several authors assume that lysozyme and CAMPs are not only acting as membrane permeabilization agents , but also activate autolytic wall enzymes of Gram-positive bacteria , thus causing cell lysis [31–33] . It has also been shown that lipoteichoic acids can bind and inhibit autolysins , depending on their degree of d-alanylation [34–36] . Similar results were also obtained in a dlt mutant of Lactococcus lactis , which showed increased autolysis [37] . In line with these observations , the graRS and dltA mutants also showed increased autolysis when treated with Triton X-100 ( unpublished data ) , suggesting that in these mutants , too , CAMPs activate autolytic enzymes . We assume that the observed synergistic effect of lysozyme in the oatA/graRS and oatA/dltA double mutants is caused by the simultaneous activation of autolytic enzymes and the muramidase activity of lysozyme . A similar synergistic effect is seen by treatment with mutanolysin in combination with LP9 ( inducing autolysis ) or lysozyme ( cannot exert its muramidase activity as the PG is O-acetylated ) as shown in the graRS single mutant ( Figure 5A ) . For the first time ( to our knowledge ) , we have traced and dissected genes that were responsive to the dual activities of lysozyme . Until now , little was known about the two-component system GraRS . We became interested in the regulation of GraRS because we wanted to trace the gene ( s ) that caused the increased CAMP susceptibility in the graRS mutant . Comparative transcriptome analysis of SA113 , an 8325-derivative , and its graRS mutant revealed that 115 genes were up-regulated and 133 genes were down-regulated by GraRS ( Tables 1 and 2 ) . Among the down-regulated genes was the vraFG operon , which immediately follows the graRS operon . However , in studying intermediate level of vancomycin resistance in S . aureus , Ambrose Cheung and colleagues found that vraFG is positively controlled by GraRS [38] . This contradictory result can be explained by the genetic organization of our graRS::ermB deletion mutant ( Figure 1 ) . In our mutant , the ermB cassette is in the same orientation as the vraFG genes . Since the ermB transcription terminator is very weak , we assume that there is a transcriptional read-through into the vraFG genes . This explains why in our graRS deletion mutant , the vraFG genes were up-regulated instead of down-regulated . GraRS up-regulates transcription of global regulators such as the SarA homologs Rot , SarS , and MgrA . We compared our GraRS transcriptome results with that of the recently published transcriptome studies of Rot [39] , MgrA [40] , and ArlRS [41] ( Tables 1 and 2; Figure 6 ) . Rot is a repressor of exoproteins but positively regulates cell surface proteins , and SarS is a positive activator of protein A . MgrA appears to be an antagonist to Rot , as it up-regulates exoproteins and down-regulates cell surface proteins , including the regulator SarS . We found that Rot and MgrA regulate some of the GraRS-controlled genes in the same direction . For these few genes we do not know whether their up- or down-regulation is directly affected by GraRS or indirectly via up-regulation of Rot and MgrA , respectively . Moreover , there are some genes that were regulated in opposite directions ( Figure 6 , boxed genes ) . Interestingly , GraRS up-regulates both regulators , Rot 3 . 8- and MgrA 3 . 1-fold . GraRS controls many genes involved in cell wall synthesis and transport ( 57 genes ) . Among the transporters are the EssA and EssC proteins , involved in transport of the virulence factor EsxA , oligopeptide transport system ( OppB ) , or NorB , which encodes the Blt-like protein that is an efflux pump involved in multidrug resistance , all of which are up-regulated by GraRS . Interestingly , smpC , which encodes a membrane-spanning protein with unknown transport functions , is the only gene that is increased by all four regulators ( GraRS , Rot , MgrA , and ArlRS ) . The gene which had the highest ( 23 . 3-fold ) up-regulation by GraRS was SA1793 , which encodes a hypothetical protein with a phage-related function . Many of the down-regulated genes are involved in RNA and amino acid synthesis or glycolysis . lrgA , which encodes a holin-like protein with murein hydrolase activity , is also down-regulated by GraRS but up-regulated by ArlRS and MgrA . Most of the genes are exclusively regulated by GraRS , such as ica , pur , mgrA , sirA , C , atlA , aaa , dnaJ , K , grpE , and vraF , G . These results illustrate that there is a distinct cross-regulation between GraRS , ArlRS , Rot , MgrA , and probably some other global regulators . GraRS is not only important for resistance to glycopeptides , lysozyme , and other CAMPs . Our data suggest that GraRS also has an intermediate role between other global regulators ( Agr , MgrA , Rot , and SarA ) , as GraRS up-regulates both adhesins as well as exoproteins and toxins ( e . g . , hlb , hlgA , B , lukM , F , geh ) . GraRS is possibly involved in the establishment of persistent infections by the up-regulation of colonization factors ( e . g . , ica , atl , aaa , fib , sirA , sirC , sdrC , sdrE ) , factors involved in resistance to CAMPs ( dlt ) , factors involved in intermediary vancomycin resistance ( vraF , G , as mentioned above ) , and factors involved in biofilm formation ( e . g . , dlt , atl , ica ) . It would be interesting to study the graRS mutant in an animal model for chronic infection .
All of the strains and plasmids that were used are listed in Table 5 . Bacteria were grown in Basic Medium ( BM ) ( 1% tryptone; Gibco BRL Life-Technologies , http://www . invitrogen . com/ ) , 0 . 5% yeast extract ( Gibco BRL ) , 0 . 5% NaCl , 0 . 1% K2HPO4 , 0 . 1% glucose , or 0 . 5% xylose ) . Transposon mutagenesis was carried out in the ΔoatA::kan mutant using the temperature-sensitive plasmid pTV1ts and was performed as described by Bera et al . [10] . Was performed as described by Bera et al . [10] . The PCR products , up- and downstream of graRS ( SA0614/15 ) ( U0614/15Kpn: TGATATAGGTACCTAATTGTTTACTAGCCGACG , U0614/15Sma: ATTTGTCCCGGGTTCTAGTAGTATTTGCATCC , D0614/15Sal: GGCCGTGTCGACTTTGTCATTTTAAACATGCG , and D0614/15Nhe: ATTGCTAGCTTGGCATAACTTGCTGCAACAGG ) , were cloned into the polylinker of the pBT2 vector flanking the ermB antibiotic cassette . Complementation of the graRS deletion mutant was obtained by cloning the graRS genes ( 1 , 912 bp ) ( C0614/15Bam: AATGATGGATCCTGGCTTTGAAGTTGACTGCC , and C0614/15Eco: AGCGCGAATTCATTTCCTTTAGGCTTTGGCAC ) into the xylose-inducible vector pTX15 in S . carnosus TM300 . The oatA::kan/dltA::spc double mutant was created by bacteriophage φ11-mediated transduction of the oatA::kan knockout into the dltA::spc deletion mutant . Overnight cultures were diluted to an OD578nm of 0 . 1 in 50 ml of BM and the cultures were incubated with shaking at 37 °C . OD was determined every hour . Ten milliliters of each culture were transferred into a new 100-ml flask when the cultures reached an OD578nm of nearly 1 . 0 . Then , cationic agents , such as hen egg-white lysozyme and human lysozyme ( Sigma-Aldrich , http://www . sigmaaldrich . com/ ) , LP9 ( a lysozyme-derived 9-aa peptide , 107R-A-W-V-A-W-R-N-R115–NH2 ) ( EMC , http://www . microcollections . de/ ) , polymyxin B ( Sigma-Aldrich ) , gallidermin ( Genmedics , http://www . genmedics . com/ ) , or mutanolysin ( Sigma ) , were added . Lysozyme was inactivated by heating for 1 h at 100 °C and placed on ice . The OD578nm of all cultures was measured hourly up to 7–8 h and after 24 h . The overnight cultures were diluted in BM with 0 . 5% xylose to a concentration of 0 . 5 × 106 CFU per ml and aliquoted in 0 . 5-ml samples , and cationic agents in different concentrations were added . The cultures were incubated with shaking at 37 °C for 20–24 h and MIC was determined . An overnight culture was diluted 1:200 in fresh TSB with 0 . 5% glucose , and 200 μl were filled into microtiter plates and incubated for 20–24 h at 37 °C without shaking . The supernatant was removed and the plate was washed two times with PBS ( pH 7 . 4 ) . The plate was dried and the cells were colored with 0 . 1% safranine . One liter of BM was inoculated with an overnight culture of the WT SA113 or the mutants . Strains were grown for 12 h with shaking at 37 °C . Cells were centrifuged , washed two times with cold 0 . 9% NaCl , diluted in 0 . 9% NaCl , and boiled for 20 min . After the cells were chilled on ice , they were again centrifuged and washed twice with 0 . 9% NaCl . The cells were disrupted in a mechanical grinding device using glass beads ∅︀150–212 μm ( Sigma-Aldrich ) at 4 °C , centrifuged and washed two times with cold H2Obidest , boiled for 30 min in 2% SDS to remove noncovalently bound proteins , and washed four times with H2Obidest . The cell wall fragments were diluted in 0 . 1 M Tris/HCl ( pH 6 . 8 ) and incubated with 0 . 5 mg/ml trypsin for 16 h at 37 °C to degrade cell-bound proteins . After centrifugation and washing with water , the PG was lyophilized . For analyzing the susceptibility of PG to lysozyme , we used a modified method turbidometric assay as described by Clarke [42] . The PG of the WT SA113 and the mutants were sonicated and diluted to 0 . 5 mg in 1 ml of 80 mM PBS ( pH 6 . 4 ) . After the addition of 300 μg lysozyme per ml , the decrease in optical density was monitored at the beginning ( 0 h ) and after 4 h at OD660nm and calculated as percentages . S . aureus strains were grown in BM with 0 . 25% glucose overnight , centrifuged , washed three times , and resuspended in ammonium acetate buffer ( 20 mM [pH 6 . 0] ) . The OD600nm was adjusted to 30 . Aliquots ( 1 ml ) were heat-inactivated by incubation at 99 °C for 10 min and centrifuged , and pellets were dried . After incubation at 37 °C for 1 h with 100 μl of 0 . 1 N NaOH , 100 μl of 0 . 1 N HCl were added for neutralization and samples were dried . For derivatization , 100 μl of triethylamine and 100 μl of Marfey's reagent ( 1-fluoro-2 , 4-dinitrophenyl-5-l-alanine amide; Sigma ) ( 10 mM ) were added . After incubation at 40 °C for 1 h , samples were dried and resuspended in DMSO:H2O ( 1:1 ) . Quantification of d-alanine was performed by HPLC as previously described [43] . SA113 and the graRS deletion mutant were cultivated in 50 ml of BM and harvested at mid-exponential growth phase . Before RNA isolation , two volumes of RNAprotect bacteria reagent ( Qiagen , http://www . qiagen . com/ ) were added to 10 ml of culture and centrifuged . The cells were lysed by the addition of 50 μg/ml of lysostaphin ( 0 . 5 mg/ml ) ( Genmedics ) in TE buffer and total RNA was isolated using the RNeasy Mini Kit ( Qiagen ) . Contaminating DNA was degraded with the DNase Kit ( Ambion , http://www . ambion . com/ ) according to the manufacturer's instructions . LightCycler RT-PCR was carried out using the LightCycler RNA amplification Kit SYBR Green I or with the LightCycler RNA amplification kit for hybridization probes ( Roche Biochemicals , http://www . roche . com/ ) . The internal control gyr was quantified using 10-fold serial dilutions ( 104 to 108 copies/μl ) of a specific RNA standard using oligonucleotides specific for gyr ( gyr297F: TTAGTGTGGGAAATTGTCGATAAT and gyr574R: AGTCTTGTGACAATGCGTT TACA ) , dltA ( dltA1: TGGCGTTGAAAGACTAGGC and dltA2: TTACGAACTCAGACTGGCG ) , rot ( rot1: TTCAGCGAGATTGAAAGCG and rot2: GTTGCTCTACTTGCAATGG ) or ureC ( ureC1: GATATCATTGCCGCTGAAGG and ureC2: AAAGCAGATGGTGTTGCACC ) as described [44] . Standard curves for dltA and rot were generated using 5-fold serial dilutions of WT SA113 RNA or for ureC of the graRS mutant RNA . Differences between WT and the graRS mutant were determined by n-fold change and calculated as a percentage of the mRNA product . The specificity of the PCR was verified by size determination of the amplicons by agarose gel electrophoresis . To check for DNA contamination , each sample was subjected to PCR by using the LightCycler DNA amplification kit SYBR Green I ( Roche Biochemicals ) . In none of the cases an amplification product was detectable . Transcriptome analysis was carried out as described by the microarray manufacturer Scienion ( http://www . scienion . de/ ) and Resch et al . [45] . cDNA was synthesized from isolated RNA ( 1 μg ) during mid-exponential growth ( 4 h ) derived from WT SA113 ( labeled in green with Cy3 [532 nm] ) or from the graRS mutant ( labeled in red with Cy5 [635 nm] ) . cDNAs from WT and the graRS mutant were pooled and hybridized on four DNA microarrays . Scienion performed DNA transcriptome analysis by comparing the intensity of each Cy3-labeled gene of the WT with the intensity of each Cy5-labeled gene of the graRS mutant as a ratio of the medians ( 532/635 ) . The threshold was set at a 2-fold difference in gene expression . Genes whose RNA level was higher in WT ( 2 . 0 and more ) were categorized as being positively regulated by GraRS . In contrast , genes that had higher RNA levels ( 2 . 0 and more ) in the graRS mutant were described as being negatively regulated by GraRS . The significance of differences ( n-fold ) in gene expression was calculated by One-Sample t-Test-Benjamini–Hochberg ( Adv ) ; results <0 . 051 are significant , and some genes from Tables 1 and 2 are higher than 0 . 05 . | In humans , lysozyme plays an important role in the suppression of bacterial infections . However , some bacterial pathogens , such as Staphylococcus aureus , are completely resistant to lysozyme . Here we demonstrate that lysozyme acts on S . aureus in two ways: as a muramidase ( cell wall lytic enzyme ) and as a cationic antimicrobial peptide ( CAMP ) . S . aureus has developed resistance mechanisms against both activities by modifying distinct cell wall structures . Modification of the peptidoglycan by O-acetylation ( OatA ) renders the cells resistant to the muramidase activity . Modification of teichoic acids by d-alanine esterification ( Dlt ) renders the cells resistant to lysozyme's CAMPs and other CAMPs . Transcriptome analysis of the glycopeptide resistance-associated ( GraRS ) two-component system revealed that this global regulator controls 248 genes such as other global regulators , colonization factors , or exotoxin-encoding genes . Since GraRS also upregulates the dlt operon , it was not surprising that in the graRS mutant teichoic acid d-alanylation is markedly decreased , which explains its increased sensitivity to CAMPs . By comparative analysis of mutants we were able to dissect genes that were responsive to the dual activities of lysozyme . Here we show how efficiently S . aureus is protected from the human defense system , which enables this pathogen to cause persistent infections . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"pathology",
"in",
"vitro",
"immunology",
"microbiology",
"vertebrates",
"molecular",
"biology",
"genetics",
"and",
"genomics",
"eubacteria"
] | 2007 | Molecular Basis of Resistance to Muramidase and Cationic Antimicrobial Peptide Activity of Lysozyme in Staphylococci |
Deciphering how the regulatory DNA sequence of a gene dictates its expression in response to intra and extracellular cues is one of the leading challenges in modern genomics . The development of novel single-cell sequencing and imaging techniques , as well as a better exploitation of currently available single-molecule imaging techniques , provides an avenue to interrogate the process of transcription and its dynamics in cells by quantifying the number of RNA polymerases engaged in the transcription of a gene ( or equivalently the number of nascent RNAs ) at a given moment in time . In this paper , we propose that measurements of the cell-to-cell variability in the number of nascent RNAs provide a mostly unexplored method for deciphering mechanisms of transcription initiation in cells . We propose a simple kinetic model of transcription initiation and elongation from which we calculate nascent RNA copy-number fluctuations . To demonstrate the usefulness of this approach , we test our theory against published nascent RNA data for twelve constitutively expressed yeast genes . Rather than transcription being initiated through a single rate limiting step , as it had been previously proposed , our single-cell analysis reveals the presence of at least two rate limiting steps . Surprisingly , half of the genes analyzed have nearly identical rates of transcription initiation , suggesting a common mechanism . Our analytical framework can be used to extract quantitative information about dynamics of transcription from single-cell sequencing data , as well as from single-molecule imaging and electron micrographs of fixed cells , and provides the mathematical means to exploit the quantitative power of these technologies .
Transcription is a multi-step process that leads to the production of messenger RNA ( mRNA ) molecules from its DNA template . Genetic experiments on cells have identified the key molecular components of transcription , while biochemical studies with purified components have uncovered the basic mechanisms governing their dynamics and interactions in vitro . Still an important question that remains is whether the same mechanisms are also operational in cells . One approach to unraveling the mechanisms of transcription in cells is to measure the outputs of this process , either the proteins that correspond to the genes being transcribed , or the actual mRNA molecules . This idea has motivated numerous experiments that count protein [1–3] , and mRNA [4–6] molecules in single cells . The measured steady state distribution of these molecules in a clonal cell population can then be used to infer the dynamics of transcription [4 , 5] . For instance , analysis of the steady state distributions of cytoplasmic mRNA in yeast for a number of different genes , have suggested that yeast genes may fall into two different classes: those that are transcribed in random uncorrelated events clearly separated in time and without any transcriptional memory [4] ( this is often referred to as Poissonian transcription ) , and those that are transcribed in bursts caused by the promoter switching slowly between an active state and an inactive state ( this is often referred to as Bursty-transcription [7 , 8] ) . While this approach to deciphering transcriptional dynamics in vivo by counting cytoplasmic RNA in single cells has led to important insights , a key limitation is that processes that are downstream from transcription initiation can mask the signature of transcriptional dynamics in measurements of the cell-to-cell variability of mRNA and protein abundances . A striking example of this is the recent finding that spatial and temporal averaging , i . e . , the process of accumulation and diffusion of mRNA transcripts during nuclear cycles , significantly reduces the variability in mRNA copy number expected from stochastic transcription initiation [9] . In addition , effects such as mRNA transport out of the nucleus , mRNA processing , and nonlinear mRNA degradation [10–14] can also in principle affect the level of variability of cytoplasmic mRNA . All of these non-transcriptional sources of variability may propagate to the protein level as well , affecting the cell-to-cell fluctuations in protein copy number , which is also affected by the stochastic nature of translation . Finally , it has been recently shown that partitioning of both mRNA and protein molecules during cell division [15–17] can generate distributions in their abundances similar to those that would be generated by stochastic transcription and translation . Therefore , the cell-to-cell variability of both protein and cytoplasmic mRNA copy number do not necessarily reflect transcriptional dynamics alone but are determined by a combination of stochastic processes of which transcriptional dynamics is just one component [18] . One alternative to analyzing steady state mRNA and protein distributions , has been to directly image transcription in real time using fluorescently labeled RNA-binding proteins that associate with nascent RNA , which is still in the process of being assembled at the gene by the RNA polymerase [19–23] . When applied to E . coli , Dyctostelium or animal cells this technique revealed widespread transcriptional bursting consistent with the mechanism of transcription initiation where the promoter switches between an active state and an inactive state [7] . In contrast , in experiments on two constitutive and cell-cycle activated genes in S . cerevisiae , Larson et al . [23] found that the transcription initiation process is dominated by one rate limiting step . In spite of the great promises of this approach , it is technically challenging and still remains in its infancy . Lately , a score of experimental papers have reported measurements of distributions across a clonal cell population of nascent RNA transcripts at a single gene , using single-molecule fluorescence in-situ hybridization ( FISH ) [4 , 24–26] ( Fig 1B ) . These experiments reveal the number of RNA polymerases engaged in transcribing a single gene in a single cell at a specific instant in time . This information can also be obtained from so-called Miller spreads ( electron micrographs of intact chromosomes extracted from cells ) which provide images of transcribing polymerases along a gene [27–31] ( Fig 1A ) . Perhaps more importantly , single-cell whole genome RNA sequencing is slowly but steadily being developed and turned into a quantitative technique , one which will be able to provide a snapshot of the number of RNA polymerase molecules engaged in the transcription of every gene in the cell at a given instant in time ( Fig 1C ) [32–36] . Counting nascent RNAs ( or the number of transcribing polymerases ) provides a more direct readout for the transcriptional dynamics at the promoter within the short window of time required for an RNA polymerase molecule to complete elongation ( for a typical gene in yeast the elongation time is of the order of few minutes [4] ) . As such , this experimental approach is not affected by the aforementioned stochastic processes that contribute to cytoplasmic mRNA and protein fluctuations . Indeed , as mentioned above , strong discrepancies between cytoplasmic and nascent mRNA distributions have been recently found in Drosophila embryos [9] . Below we also demonstrate similar discrepancies in yeast by analyzing published data obtained from counting nascent and cytoplasmic mRNA in single cells . It is thus starting to become possible to obtain quantitative measurements of the distribution of nascent RNA or , what is the same , of the number of transcribing polymerases per gene . In spite of its many advantages , the potential of nascent RNA distributions has not been fully exploited ( for a notable exception see [37] ) due to the lack of mathematical formalisms that allows one to connect molecular mechanisms of transcription initiation and elongation with measured nascent RNA distributions . One of the key results that we report here is the development of such formalism . In particular , we show how to compute the mean and variance of the distribution of nascent RNAs for an arbitrary mechanism of transcription initiation and stochastic elongation . The results of these calculations provide the tools to extract information about transcriptional dynamics from experimentally determined nascent RNA distributions . We demonstrate the usefulness of our method by analyzing published nascent RNA distributions for a set of constitutively expressed yeast genes [25] . We find that all of these yeast genes have similar average initiation rates . We also find that initiation of transcription of these yeast genes is a two-step process , where the average durations of the two steps are equal . This is in sharp contrast to the conclusion that was reached for some of these genes by counting cytoplasmic mRNAs , namely that transcription initiation is dominated by one rate limiting step [25] . By analyzing the nascent RNA distribution , we are able to reach a level of kinetic detail , particularly fast processes , which are obscured at the level of cytoplasmic RNA . While the molecular identities of the two steps leading to transcription initiation remain unknown , our results point to the existence of multiple transcription initiation steps in vivo . It is worth emphasizing that multiple initiation steps of similar duration lead to a reduction of fluctuations in the number of nascent RNAs in a cell , when compared to those produced by single-step initiation .
In order to connect mechanisms of transcription initiation with nascent RNA distributions , we consider a model of transcriptional dynamics with an arbitrarily complex initiation mechanism followed by an elongation process . We describe both processes using chemical master equations . This approach is inspired by the work of Kepler et al . [38] who computed the moments of the mRNA distribution for a promoter , where it switches between an active and an inactive state . We have previously developed this method further to compute the moments of mRNA and protein distributions for arbitrarily complex promoters that can switch between multiple states , each state leading to transcript production at a particular rate [8 , 39–42] . Here we implement the same master equation approach to compute the first and second moments of the nascent RNA distribution . A new element in our analysis is the explicit inclusion of the stochastic elongation process , which predicts that the nascent RNA distributions depend on the length of the gene being transcribed , for which we find confirmation in published data . This dependence of the distribution of nascent RNAs on gene length has also been described recently in [37] . Our theory also suggests new experimental approaches to deciphering the dynamics of transcription initiation in vivo , in which the length of the transcribed gene is varied and the effect on the number of nascent RNAs is measured . To describe the transcription initiation process we focus on promoter dynamics . ( Here we use the term promoter to denote the stretch of regulatory DNA that controls the initiation of transcription of a specific gene . ) The promoter switches between different states as different transcription factors bind and fall off their respective binding sites , causing the effective initiation rate to fluctuate . We assume that after initiation , each RNA polymerase ( RNAP ) moves along the gene by stochastically hopping from one to the next base at a constant probability per unit time ( Fig 2A ) . Our model assumes that transcription initiation timescales are much slower than the elongation timescale and hence RNAPs do not interfere with each other while moving along the gene . This approximation is reasonable for all but the strongest promoters characterized by very fast initiation [43 , 44] . We demonstrate this explicitly using numerical simulations [45 , 46] which include a detailed model of transcription elongation that takes into account excluded-volume interaction between adjacent polymerases ( i . e . “traffic” as defined in previous work [43] ) , as well as ubiquitous RNAP pausing [43 , 47] ( please see S1A Fig ) . The agreement between analytical results based on our simple model and the stochastic simulations of the more realistic model that incorporates traffic jams and pausing of RNAPs only starts to break down when the initiation time scales become comparable to the elongation time scales ( please see S1C and S1D Fig ) . We conclude that for typical rates reported for RNAP elongation and pausing the simple model of transcription adopted here reproduces the first two moments of the nascent RNA distribution with deviations from those obtained from the more realistic model that are less than 10% as long as initiation of transcription is slower than 30 initiations/min . All the initiation rates that have been reported so far from in vivo measurements are slower [4 , 19 , 23] , with important exceptions such as the ribosomal promoters [43 , 44] . Our model does not explicitly include the rate of termination at which the RNAP departs the last base of the gene . The genes [25] that we analyze have an average initiation rate of the order of kINI = 0 . 145±0 . 025/min . Hence even for a termination rate of the order of 1/min [23] , the variance and mean of the nascent distribution won’t be affected for these genes . Another simplifying assumption that we make is that we place no restriction on the number of transcribing RNAPs that can occupy a given base ( in reality at any given instant the number is zero or one ) . This is equivalent to assuming that the occupancy of any given base of the gene by a transcribing polymerase is much less than one , which holds when the initiation time scale is much slower than the elongation time scale . Hence , despite its simplicity , the model of transcription initiation and elongation we adopt here should apply to most genes . In order to compute the first two moments of the nascent RNA distribution for an arbitrary transcription initiation mechanism , we consider a promoter that can exist in N possible states . The rate of transition from the s-th to the q-th state is ks , q , and the rate at which RNAP initiates transcription from the s-th promoter state is ks , ini . Following the initiation process , every RNAP moves along the gene ( elongates ) by hopping from one base to the next with a probability per unit time k , which is equal to the average rate of elongation . The number of RNAP molecules , which is the same as the number of nascent RNAs , at the i-th base pair is denoted by mi . Hence the number of nascent RNAs ( M ) along a gene whose length is L bases , is given by , M=∑i=1Lmi . As remarked earlier we do not consider the processes of transcription termination and mRNA release , as they tend to be fast on the time scales set by initiation and elongation . However these can be easily incorporated into the model . ( For the mathematical details please see the S1 Text . ) The state of the combined promoter+RNA system is described by ( L+1 ) stochastic variables: the number of nascent RNAs ( m1 , … , mL ) at every base along the gene , and the label s , characterizing the state of the promoter . Hence , the probability distribution function that characterizes the promoter+RNA system is given by P ( s , m1 , … , mL ) . To stream-line the mathematics we define the following probability vector: P→ ( m1 , … , mL ) = ( P ( 1 , m1 , … , mL ) , P ( 2 , m1 , … , mL ) , … , P ( s , m1 , … , mL ) ) . ( 1 ) The time evolution for this probability vector can be described by a set of chemical master equations , which can be written in compact , matrix form as dP→ ( m1 , … , mL ) dt= ( K^−R^−Γ^∑i=1Lmi ) P→ ( m1 , . , mi , . , mL ) +R^P→ ( m1−1 , … , mL ) +∑i=1L−1k ( mi+1 ) Γ^P→ ( m1 , . , mi+1 , mi+1−1 , . , mL ) +k ( mL+1 ) Γ^P→ ( m1 , … , mL+1 ) . ( 2 ) In Eq ( 2 ) , we define the following matrices: K^ , which describes the transition between different promoter states , and whose elements are Kqs = kq , s if q≠s and Kss=−∑qkq , s . R^ is a matrix that contains the rates of initiation from different promoter states . In the case of one-step initiation it is diagonal with the diagonal elements equal to the rates of initiation from different promoter states . In the case of two-step initiation this matrix is off-diagonal owing to the fact that the promoter state changes after initiation ( for details please see the S1 Text ) . Γ^ is also diagonal and its elements represent the hopping rate for the polymerase from one base pair to the next , i . e . , Γsq = k δs , q . We limit our calculation to the steady state nascent RNA distribution for which the left hand side of Eq ( 2 ) is set to zero . To obtain the first and second moments of the number of nascent RNAs , M=∑i=1Lmi . in steady state we use Eq ( 2 ) to compute the quantities 〈mi〉 and 〈mimj〉 for all i , j ≤ L . Even though the random variables mi for different bases i on the gene are mutually dependent , we end up deriving a set of linear equations for 〈mi〉 and 〈mimj〉 ( Please see the S1 Text . ) We find that these equations for the moments close , in other words they do not depend on any further , higher moments of the mi’s . These linear equations can then be solved to obtain exact expressions for the first two moments of M as a function of all the rates that define the molecular mechanism of initiation under investigation . ( For the mathematical details please see the S1 Text . ) In order to demonstrate how the distribution of nascent RNAs at the transcription site can be used to extract dynamical information about the process of transcription initiation in vivo , we consider the canonical model of transcription shown in Fig 2A [38] . The gene can switch between two states: an active state , from which transcription initiation can occur , and an inactive state from which initiation does not occur . The two states might correspond to a free promoter and one bound by a repressor protein , or a promoter occluded by nucleosomes . In most theoretical studies to date transcription initiation from the active state was assumed to be characterized by a single rate-limiting step . Instead of initiation being a one-step process we consider the possibility that there are two rate-limiting steps involved in transcription initiation from the active state . These could represent the loading of the transcriptional machinery at the promoter [48 , 49] ( in prokaryotes , this would correspond to the formation of open complex by RNAP [50–52] ) , which occurs with a rate kLOAD followed by the RNA polymerase escaping the promoter into an elongation state ( with rate kESC ) . Three different limits of our model correspond to the various scenarios that have been previously explored in the literature [4 , 19 , 53–55] . First we consider the limit when the promoter is always active ( kOFF → 0 in Fig 2A ) and initiation is governed by a single rate-limiting step . This is a situation when one of the two kinetic steps leading up to initiation ( either the assembly of the transcriptional machinery or the escape of RNA polymerase from the promoter ) is much slower than the other . In this case we find that the nascent RNA distribution is characterized by a variance that is equal to the mean . In other words the Fano factor , defined as the variance divided by the mean , to characterize cell-to-cell variability is 1 . The second limit of interest is when the rates of assembly of the transcriptional machinery ( kLOAD ) and promoter escape ( kESC ) have comparable magnitudes , i . e . , transcription initiation is a two-step process . In this limit , transcription initiation events are anti-correlated due to the presence of a “dead-time” or refractory period in between subsequent initiation events . The third limit of interest is the “transcriptional bursting” , when the promoter is not always active , but is slowly switching between the active and inactive states [7] A key prediction of our model of stochastic transcription initiation and elongation , which is described in Fig 2A , is how the cell-to-cell fluctuations of the nascent RNA number depend on the length of the gene being transcribed . This dependence was also explored in [37] where the importance of the elongation rate and gene length in determining nascent RNA distributions was described , and nascent RNA distributions were used to infer the kinetic rates . Gene length is an interesting quantity to consider from the point of view of experiments , both due to the natural variation in gene length , and the ability to synthetically alter the length of the gene being expressed from a promoter of interest by genetic manipulation . Calculations of the Fano factor as a function of gene length ( Fig 2B ) reveal that this quantity easily discriminates between the three models of transcription initiation described above . When the gene length is small the Fano factor is close to one for all three models of initiation . As the gene length increases , the Fano factor increases above one for the “bursting” scenario , due to slow switching of the promoter between an active and an inactive state , but it decreases below one when the promoter is always active and there are two rate limiting steps leading up to elongation . Finally , in the case when initiation is dominated by one rate-limiting step and the promoter is always active , the Fano factor is equal to one , independent of gene length . Qualitatively these results can be understood by recalling that with a single initiation step , the waiting time between initiation events is exponentially distributed . In this case the number of initiation events to occur in a time interval set by the elongation time ( which is roughly equal to the number of nascent transcripts ) is given by a Poisson distribution [56] , for which the Fano factor is one . For two or more rate limiting steps leading to initiation , the waiting time between successive initiation events is gamma distributed [55] . As a result the distribution of nascent RNAs is expected to be narrower than Poisson with a Fano factor less than one . The presence of transcriptionally inactive states on the other hand has the effect of broadening the distribution of nascent RNAs , and should lead to a Fano factor greater than one in the case when initiation from the active state is a one-step process . For bursty promoters that switch between an active and an inactive state ( for example the PDR5 gene in yeast [4] ) the nascent RNA distribution can also be used to discriminate between different mechanisms of regulation . Recent experiments [57 , 58] have suggested that transcriptional regulation may be achieved by either modulation of the burst size ( given by kINI/kOFF , where kINI = kLOAD × kESC/ ( kLOAD + kESC ) is the average rate of initiation ) , or by modulating the burst frequency ( kON ) ; it is also possible that both are tuned [7] . In Fig 2C we show the results of our calculations of the Fano factor for the nascent RNA distribution , using parameters that are characteristic of the PDR5 gene and assuming that transcriptional regulation is achieved either by tuning the burst size or the burst frequency . We see that even though both mechanisms of regulation produce Fano factors larger than one , they make qualitatively different predictions for the functional dependence of the Fano factor on the mean number of nascent RNAs . Nascent RNA distribution is determined by stochastic initiation and elongation . However for a long gene , the elongation process becomes practically deterministic due to the law of large numbers . Assuming as we do in our model that each elongation step is a stochastic process , with the same rate k , the elongation time will be Gaussian distributed with a mean and variance that are proportional to the length of the gene . Therefore the deviation of the elongation time away from the mean compared to the mean will decrease as the square root of the gene length . A recently published paper by Senecal et al . [37] has explored how stochastic initiation and deterministic elongation processes affect nascent RNA distribution . However , and although this is indeed the case for FISH data such as the one analyzed in Fig 3 , in the paper , an important application of our method will be the analysis of single-cell sequencing data , where the positions of every polymerase along the gene can be determined . In addition , and as shown in Fig 1 , we anticipate that our method can be applied to electron micrograph data ( e . g . such as those reported in [29 , 31] ) , a method that also allows one to measure the position of each polymerase along a transcribed gene . Using this technique , the number of polymerases in the first L nucleotides of the gene can also be determined , and statistics ( mean and variance ) can be computed and compared to experimental results . This will allow us to computationally bin a gene into smaller chunks of arbitrary length , and use length as a “data analysis” turning knob that we can tune computationally to investigate how it affects the noise ( Fano factor ) . To be able to do this , a fluctuating elongation rate is essential , since in principle the gene length can be made as short as desired during data analysis . In the limit of a long gene , when the residence time of the RNAP on the gene is practically deterministic , we can use queuing theory to compute closed form expressions for the first and second moments of the nascent RNA distribution [59 , 60] . For the one-step model , the promoter is always active ( kOFF → 0 in Fig 2A ) and there is a single rate-limiting step leading up to initiation . In this case the nascent RNA distribution is characterized by a variance that is equal to the mean , which is what we computed for stochastic elongation as well . The second limit of interest is when the rates of assembly of the transcriptional machinery ( kLOAD ) and promoter escape ( kESC ) have comparable magnitudes , i . e . , transcription initiation is a two-step process . For an elongation time T = L/k ( where L is the number of bases along a gene and k is the average rate of elongation ) the mean and variance are given by ⟨M⟩=kLOADkESCTkLOAD+kESCVariance=⟨M⟩[1−kESCkLOADT ( kESC+kLOAD ) +kESCkLOADT ( kESC+kLOAD ) 2[2 ( 1−exp ( − ( kESC+kLOAD ) T ) ( kESC+kLOAD ) ) −2T+T2 ( kESC+kLOAD ) ]] . The Fano factor can hence be computed very easily by taking the ratio of the variance and mean . It is to be noted that when one of the rates that describe the two steps leading to initiation ( kLOAD and kESC ) becomes much smaller than the other , we are back to the case of one rate-limiting step . In the case of one rate-limiting step the Fano factor becomes one , which is the signature of a Poisson initiation process . However when the rates kLOAD and kESC become comparable , the Fano factor is reduced from 1 and attains a minimum value when kLOAD is equal to kESC . This also follows from the equations for the variance and the mean and can be intuited by noting that the two rates appear in the equations in a symmetrical fashion . The third limit of interest is the ON-OFF model of initiation which is characterized by kOFF ( the rate of promoter switching from the ON to OFF state ) , kON ( the rate of promoter switching from the OFF to ON state ) , kESC ( the rate of escape , which we assume is much higher than the rate of assembly of the transcriptional machinery ) , and time of elongation T . The mean and variance are given respectively by , ⟨M⟩=kONkESCTkON+kOFFVariance=⟨M⟩[1+2kESCkOFF ( kON+kOFF ) 2+2kESCkOFF ( kON+kOFF ) 3 ( exp ( − ( kON+kOFF ) T ) −1T ) ] . As shown in S4 Fig , these formulas give almost identical results to those we obtain when taking into account stochastic elongation , when the gene length is of the order of few thousand bases . The theoretical results described above can be used as a mathematical tool to extract information about transcription initiation dynamics from nascent RNA distributions , which have been measured in a series of recent experiments [4 , 25] . To demonstrate the utility of this approach , we analyze a set of nascent RNA distributions for twelve different constitutively expressed genes in yeast [25] . We find that for six of these twelve genes ( RPB2 , RPB3 , TAF5 , TAF6 , TAF12 , KAP104 ) , the mean number of nascent RNAs scales linearly with the gene length , as shown in Fig 3A . If we assume that all of these genes have comparable elongation rates ( k = 0 . 8 kb/min ( 4 ) ) , then the linear relationship between the mean nascent RNA number and gene length implies that the average initiation rates of these genes are all roughly the same and equal to kINI = 0 . 145±0 . 025/min . In addition to the mean , our model allows us to investigate the behavior of the variance of the nascent mRNA distribution with gene length , and compare it to the predictions from different models of transcription initiation . Given that the Fano factors of the nascent RNA distribution for the six genes , RPB2 , RPB3 , TAF5 , TAF6 , TAF12 , and KAP104 , are all less than one , the simplest model consistent with the data is one where the promoter is always active and transcription initiation is a two-step process ( see Fig 2A ) . This model is parameterized by the rates kLOAD and kESC . The Fano factor of the nascent RNA distribution depends on the ratio of these two rates . Our model makes prediction for how the Fano factor changes with the gene length when the rates kLOAD and kESC are tuned , consistent with the mean initiation rate being kINI = kLOAD × kESC/ ( kLOAD + kESC ) = 0 . 145±0 . 025/min . As shown in Fig 3B , this value of the initiation rate defines a region in the Fano factor–Gene length phase ( light-blue shaded area in Fig 3B ) . This region is bounded on its upper side by the limit when one of the rates ( either kLOAD or kESC ) is much larger than the other one ( which turns initiation into a one-step process with a Fano factor equal to one ) , and on the lower side , by the limit in which the two rates are identical . The limit of identical rates gives the minimum Fano factor attainable when the average initiation rate is 0 . 145±0 . 025/min . Remarkably , we find that the six genes in question have the lowest possible Fano factor . In principle , the six genes shown in Fig 3A could have ended up anywhere within the shaded region in Fig 3B . The fact that they all follow the lower boundary of the allowed region suggests that these genes , which have varying length , have not only the same average initiation rate , but also that they have identical promoter cycling kinetics , with roughly the same values of kLOAD and kESC ( kLOAD = kESC = 0 . 29±0 . 013/min ) . It is to be noted that a multi-step initiation model , where initiation happens in more than two sequential steps ( this can also include bursting kinetics ) can also account for the Fano factor for the six different genes being less than one . However a more complicated model with more than two sequential steps will have many free parameters e . g . for a three step model we will have three sequential steps to initiation characterized by three different rates . Although we cannot rule out such possibilities , the two-step model in spite being the simplest possible scenario explains the data well and provides mechanistic insight into the dynamics of initiation for the six different genes . The key result here is that by analyzing nascent RNA distributions , we can exclude the one-step and ON-OFF models of initiation . The remaining six ( RPB1 , MDN1 , PUP1 , PRE7 , PRE3 , PRP8 ) constitutive genes of the twelve studied [25] initiate at rates that are different than the rate of initiation that we found for the six genes discussed above ( see S2 Fig ) . All but one of these six genes have nascent RNA Fano factors that are less than one , consistent with two or more steps leading up to initiation . This second set of genes thus acts as a control group that , as expected for a set of genes having different gene-specific rates of transcription , occupies the allowed region in the Fano factor-Gene length phase space without clustering at the lower boundary of this region , like we found for the six genes discussed above ( S3 Fig ) .
Direct imaging of transcriptional dynamics in real time [19–23] at the molecular scale and in individual cells still remains challenging . As an alternative , a number of recent studies have tried to decipher the dynamics of transcription initiation using the measured cell-to-cell variability of transcriptional outputs ( cytoplasmic messenger RNA or protein molecules ) at the single cell level [1 , 3 , 4 , 25 , 61] . These measurements of transcriptional cell-to-cell variability have been interpreted in the context of a classification scheme for promoters , which are characterized by either a Poisson or a Gamma distribution of their outputs . These differences have then been taken to indicate a difference in the mechanism of transcription . A Poisson distribution is taken as evidence that the promoter transcribes at a constant rate , i . e . , initiation is a one-step process . The Gamma distribution on the other hand is indicative of bursty promoter dynamics [4 , 19] . In practice , the distribution of cytoplasmic mRNA or proteins obtained from a population of cells is fitted to a mathematical model that incorporates the stochastic kinetics of transcription ( and translation in the case of proteins ) , and the fitting parameters are interpreted as representative of the kinetic properties of stochastic gene expression ( e . g . , burst size , burst frequency , average transcription rate , etc . ) [5] . Even though in some cases this approach has produced kinetic parameters whose values are consistent with direct measurements of the same parameters [1] , the interpretation of the kinetic parameters can be difficult given that the distributions of mRNA and protein may be affected by stochastic processes that occur downstream of transcription initiation . Examples of these processes include the non-linear degradation of mRNA and proteins [14] , maturation time of fluorescent reporters [62] , transport of mRNA out of the nucleus [10 , 11] , mRNA splicing [12 , 13] and small RNA regulation [14 , 63] . Furthermore , recent theoretical results [15 , 16] indicate that fluctuations due to random partitioning of molecules during cell division may yield the same mathematical dependence between variance and mean of protein and mRNA copy number in clonal cell populations , as would a stochastic model of transcription initiation and linear degradation . In order to demonstrate that the distribution of mRNAs can be affected by stochastic processes that occur downstream of transcription , thereby obscuring the signature of transcription initiation dynamics , we compare the nascent RNA and cytoplasmic mRNA distributions for the twelve yeast genes analyzed in Fig 4 . First , we compute the Fano factor of the cytoplasmic mRNA distribution predicted by the initiation mechanism inferred from the measured nascent RNA distribution for all twelve genes studied ( 23 ) . ( See the S1 Text for details of the calculation . ) Then we compare the results of our calculations with the experimentally determined distributions obtained by counting cytoplasmic mRNA . We find that for all of the yeast genes examined the predicted Fano factors for the cytoplasmic mRNA distributions are less than the measured ones , as shown in Fig 4 . In other words the signature of two-step initiation observed in the nascent RNA distribution is washed out at the cytoplasmic mRNA level due to other sources of noise . It remains unclear what processes are responsible for these differences . In a recent study of transcription in fly embryos , it was also found that the variability of nascent and cytoplasmic mRNA could differ more than six fold [9] . In this case , the reason for this difference is spatial and temporal averaging of mRNA by diffusion and accumulation of mRNA transcripts during nuclear cycles . The yeast and fly examples demonstrate that the relationship between nascent and cytoplasmic RNA distributions is complex and context dependent . An alternative to counting cytoplasmic proteins or mRNA is to count the number of transcribing polymerases [27–30] , or nascent RNAs [4 , 25 , 37] on the gene being transcribed , using electron micrographs and fluorescence in situ hybridization , respectively . These measurements are not affected by post-transcriptional processes and are more direct readouts of transcriptional dynamics . To date , these distributions of nascent RNAs have been used mostly in a qualitative manner , due to the lack of mathematical models that connect these distributions with the underlying mechanisms of transcription , apart from the recent paper by Senecal et al . [37] . For instance , distributions of nascent RNAs ( or of transcribing RNA polymerases ) have been recently reported in yeast [4 , 23 , 25] , fly embryos [9 , 49 , 64] , and bacteria [29 , 65] . The model of transcription initiation and elongation developed here offers a way to quantitatively analyze these measured nascent RNA distributions , and connect them to molecular mechanisms of transcription . In particular , when we consider three different models of transcription initiation that incorporate three broad classes of initiation mechanisms , we find that they make qualitatively different predictions for nascent RNA distributions . Analyzing the nascent RNA distributions for twelve constitutively expressed genes in yeast [25] , we find that all but one of these distributions have a Fano factor less than one . This observation is consistent with a simple model in which initiation proceeds in two-steps ( for some of the genes more than two steps are implicated by the data; see S3 Fig ) , which are of similar duration . The two rate limiting steps can arise from a number of different sources . For the genes in yeast considered in the paper the initiation complex is formed by assembly of multiple transcription factors and co-factors [66] . After the formation of the initiation complex , an RNAP molecule initiates transcription by escaping the promoter . The two-step model we consider in the paper would be realized if out of all these steps leading up to initiation any two steps become rate-limiting . The most surprising finding when analyzing these twelve genes was that six of them have not only the same average initiation rate , but also the same rates of loading of the transcriptional machinery , and of promoter escape . We do not have a mechanistic interpretation for this finding , but the data suggests the existence of a common molecular mechanism of initiation for these six genes , and given that they represent half of the genes in the data set we have analyzed here , it is tempting to speculate that other yeast genes may share the same kinetics . More experiments are clearly needed to test this hypothesis , ideally ones where the dynamics of transcription are followed directly [23] . Our findings for the yeast promoters , highlight the utility of our theory for deciphering transcriptional dynamics in vivo from nascent RNA distributions . In addition , counting nascent RNAs , mRNAs and proteins simultaneously will undoubtedly further enhance our understanding of how the central dogma of molecular biology plays out in individual cells .
To compute the first two moments of the nascent RNA distribution for the canonical model of transcriptional regulation shown in Fig 2A we apply the general method of deriving moment equations from the master equation , Eq ( 2 ) . The rate matrices that define the master equation , Eq ( 2 ) , are in this case: K^=[−kONkOFF0kON− ( kOFF+kLOAD ) kESC0kLOAD−kESC] , R^=[00000kESC000] , Γ^=[k000k000k] . Here K^ , is the transition matrix , which describes promoter switching between the three possible states shown in Fig 2A . When an RNA polymerase initiates transcription from the state in which the polymerase is bound to the promoter , the state of the promoter changes to the state in which the promoter does not have a bound polymerase . This accounts for the rate of escape appearing in the transition matrix and also explains why R^ ( the initiation rate matrix ) is not diagonal . Using these matrices in the master equation for the nascent RNA distribution ( Eq ( 1 ) ) we compute analytically the mean and the variance of the distribution as a function of the gene length L . These results were used to make the plots in Fig 2B and 2C . Our model ( S1B Fig ) makes the assumptions that RNAP molecules do not pause and do not collide with other RNAP molecules , while moving along the gene . We also take the size of the RNAP footprint to be one base , and we do not restrict the number of RNAPs at each base along the gene . These assumptions are equivalent to the assumption that the average number of transcribing RNA polymerases is much less than one per base . If we consider a constitutive ( one-step ) promoter with an initiation rate kESC , and the rate at which the RNAP translocates from one base to the next is k , then the number m of RNAP molecules on the first base of the gene would be Poisson distributed [40] and given by , P ( m ) = ( kESCk ) mm ! e−kESCk . As the above equation demonstrates , if the ratio of initiation rate and hopping rate kESC/k is of the order of 0 . 01 ( characteristic of MDN1 promoter [25 ) ] , the probability of finding two or more RNAP molecules at the first base of the gene would be 5×10−5 . This justifies one of the main assumptions of our model , namely that we can ignore the constraint that no base can be occupied by more than one polymerase . The assumption will be valid as long as the initiation time scale is slower than the elongation time scale , and it makes the model analytically tractable . As described in the results section , this assumption leads to simple formulas for the first two moments of the nascent RNA distribution in the large gene-length limit , and to a set of L2 linear equations in the case of stochastic elongation . As shown in the S1 Text , these linear equations are readily solved to obtain the moments of the nascent RNA distribution using standard computing tools such as Mathematica or Matlab . This is important in order to test many parameter sets without having to run a new Gillespie simulation for every set which can be impractical for complex kinetic mechanisms of transcription initiation . As argued above , we expect the approximations made in our model to be reasonable for all but the strongest promoters characterized by very fast initiation [43 , 44] . In order to test this intuition , we compare the analytic predictions of our model with numerical simulations of a more realistic one ( referred to as the traffic model in S1A Fig ) , which properly accounts for the footprint of a transcribing RNAP molecule on the DNA , ubiquitous pausing of the polymerase , and excluded volume interactions between adjacent polymerases along the gene . In particular we compare the mean and the Fano factor of nascent RNA distributions , as predicted by our model of transcription for the case when initiation occurs via a single rate limiting step , with those obtained from numerical simulation of the traffic model obtained using the Gillespie algorithm [45 , 46] . A single time step of the simulation is performed in the following way: one of the set of all possible reactions is chosen at random according to its relative weight , which is proportional to the rate of the reaction , and the state of the system is updated by implementing the change described by the chosen reaction . The time elapsed since the last step is drawn from an exponential distribution , the rate parameter of which equals the sum of all the rates of the possible reactions at that time . This process is repeated for a long enough time such that the number of RNAP molecules along the gene ( which is the same as the number of nascent RNAs ) reaches steady state . We consider four different transcription initiation rates , spanning the typically observed values in E . coli and yeast cells [4 , 19 , 23 , 25] , and we observe in the simulations how the mean and Fano factor of the nascent RNA distribution are affected by RNAP pausing and road blocking ( S1A and S1B Fig ) . We find that for initiation rates slower than 30 initiations/min , both the mean and the Fano factor extracted from the simulations are in good agreement ( less than 10% difference ) with the analytical results ( S1C and S1D Fig ) . In simulations we used the following parameters to describe RNAP elongation: kP- = 4/sec , kP+ = 0 . 01/sec , k = 80 bp/sec , as was reported for ribosomal promoters in E . Coli [43] . We also use a gene of length L = 2000 bases and a polymerase whose DNA footprint is 30 bases . We generated the plots for the Fano factor versus gene length ( Fig 2B ) , for the three limits of the model in Fig 2A using the parameters listed below . For the bursty promoter , where the promoter slowly switches between inactive and inactive states , we use kOFF = 5/min , kON = 0 . 435/min , k = 0 . 8kb/min , kLOAD = 5/min and kESC = 0/min; kOFF , kON , k , and kLOAD are the characteristic rates for the PDR5 promoter , as reported in [4] . For the two-step initiation model , where the promoter does not switch between an active and an inactive state but has two rate limiting steps leading up to initiation , we use kLOAD = 0 . 14/min , kESC = 0 . 14/min , kOFF = 0/min , kON = 0/min , k = 0 . 8kb/min; these are characteristics of yeast genes , such as MDN1 [25] . For the one-step model , there is one rate limiting step leading up to transcription elongation and we choose kLOAD = 0 . 09/min , kESC = 0/min , kOFF = 0/min , kON = 0/min , k = 0 . 8kb/min , which are characteristics of the yeast gene RPB1 [25] . Genes that are transcribed from a promoter that switches between an active and an inactive state can be regulated by changing the rates of switching between these two states , either by modulating the burst size ( given by kINI/kOFF , where kINI = kLOAD kESC/ ( kLOAD + kESC ) is the average rate of initiation ) , or by modulating the burst frequency ( kON ) , ( it is also possible that both are modulated ) [57 , 58] . In order to compute the predictions for the nascent RNA distribution for these two mechanisms of regulation in Fig 2C , we change burst size and burst frequency by changing kOFF and kON . In the first case , we change the burst size by changing kOFF and taking the other parameter values to be , kON = 0 . 435/min , k = 0 . 8kb/min , L = 4436 bps , kINI = 5/min as reported for PDR5[4] . Then we change burst frequency by changing kON , where the other parameters are , kOFF = 5/min , k = 0 . 8kb/min , L = 4436 bps , kINI = 5/min as reported for PDR5 [4] . In Fig 2C the Fano factor of the nascent RNA distribution is plotted as a function of its mean normalized by meanmax , where meanmax is the maximum of the mean number of nascent RNAs which is obtained when there is no transcriptional regulation and the promoter is always active . We analyze the measured nascent RNA distributions for twelve different constitutively expressed yeast genes reported in reference [25] . By applying our theoretical results to the published data , we find that the average initiation rates of six ( KAP104 , TAF5 , TAF6 , TAF12 , RPB2 , RPB3 ) of these twelve genes are all roughly the same , and equal to 0 . 145±0 . 025/min . However the other six genes ( RPB1 , MDN1 , PUP1 , PRE7 , PRE3 , PRP8 ) initiate transcription at different rates . This we conclude from S2 Fig , where the mean number of nascent RNAs is plotted against the gene length for all the twelve genes . When considering experiments that count nascent RNAPs it is important to be mindful of the fact that the number of RNAP molecules along a gene is not necessarily equal to the nascent RNA counts . Transcribing RNAPs have partial nascent transcripts attached to them depending on how far along the gene they have progressed ( as indicated in Fig 1C ) . In a single molecule FISH experiment , the RNA sequence that is targeted by the fluorescent probes determines if these transcripts are detected or not . Probes against the 5’ end detect transcripts early on , while probes against the 3’ end will detect only almost finished transcripts . [9] . However , as long as there is a way to correctly extract the RNAP number distribution from nascent RNA intensity , our model can accurately transform this data into information about the transcriptional dynamics . In addition to the mean , we analyze the Fano factor of the nascent mRNA distributions as well , and compare it with the prediction from our model of transcriptional regulation ( Fig 2A ) . It is to be noted that for 9 of the different genes we consider , the number of nascent RNAs does not exceed 2 . These distributions can be described by the probabilities of having 0 , 1 and 2 nascent RNAs . Still the Fano factor is a useful metric for analyzing these distributions and quantifying how much they differ from a Poisson distribution . We find the Fano factors of the nascent mRNA distribution for all of the twelve genes in the data set to be less than ( or at most equal to ) one . Hence the simplest model consistent with the published data for these twelve yeast genes is one where the promoter is always active and transcription initiation is a two-step process ( see Fig 1A ) parameterized by the kinetic rates kLOAD and kESC . We find that Fano factors for the six genes: KAP104 , TAF5 , TAF6 , TAF12 , RPB2 , RPB3 , which all initiate transcription at the same average rate , follow precisely the trend-line expected when kLOAD and kESC are equal ( kLOAD = kESC = 0 . 29±0 . 013/min ) , as shown in Fig 2B . We also analyze the Fano factor of the nascent mRNA distributions of the other six genes: RPB1 , MDN1 , PUP1 , PRE7 , PRE3 , PRP8 , which initiate transcription at a different mean rate ( S2 Fig ) . Their location in the phase space defined by the gene length and Fano factor ( as shown in S3 Fig ) indicates that they all have at least two rate limiting steps leading up to initiation and that these steps are likely parameterized by different rates , unlike what we observe for the other six genes . It is to be noted that it might be difficult to experimentally tell the difference between a mature RNA or a single nascent RNA , there are 12 different genes for which data is available [25] and our conclusions are based on examining the whole set . As pointed out earlier 9 of these genes have up to 2 nascent RNA molecules . Hence for these genes the distributions of nascent RNA molecules have three bins ( for 0 , 1 and 2 mRNA molecules respectively ) . However our analysis also includes genes for which there are more than 1 or 2 nascent RNAs , such as RPB1 ( up to 3 ) , MDN1 ( up to 5 ) , PRP8 ( up to 3 ) . We find all of these genes to have a Fano factor of less than one indicative of two or more steps leading to initiation . In order to compare the nascent RNA and cytoplasmic mRNA distributions , we compute the Fano factor of the cytoplasmic mRNA distribution , predicted by the two-step mechanism of initiation for the seven genes ( KAP104 , TAF5 , TAF6 , TAF12 , RPB2 , RPB3 , MDN1 ) . In other words an mRNA molecule is produced in two sequential steps , e . g . , by first assembling the transcriptional machinery at the promoter DNA , followed by RNA polymerase escaping the promoter . These two steps are parameterized by the kinetic rates kLOAD and kESC , respectively . For the gene RPB1 initiation is a one-step process , while for others ( PUP1 , PRE3 , PRE7 , PRP8 ) three steps are required to account for the measured nascent RNA distribution . We further assume that mRNA is degraded with a constant probability γ per unit time per molecule . The degradation rates of the twelve genes used in the calculation are those reported in reference [25] . Given these assumptions about mRNA production and degradation , we compute the Fano factor ( ratio of variance and mean ) of the mRNA distribution using the approach developed previously in order to find the moments of mRNA distribution [8 , 38–41] . The computed Fano factor is a measure of the expected cell-to-cell variability in the number of cytoplasmic mRNAs if only the initiation process was contributing to this variability . The fact that we observe a discrepancy between the mRNA variability calculated in this way and the measured mRNA variability ( Fig 4 ) is indicative of the presence of significant sources of noise that are downstream of transcription . | Gene expression starts with transcription , a multi-step process that produces an RNA molecule that is complementary to the gene . Cells often control the amount of gene expression by controlling the amount of RNA produced through interactions between regulatory DNA and proteins involved in transcription . While the identity of the molecules that take part in this regulatory process is known for a number of different genes , their dynamics in cells is still poorly understood . We show theoretically that the cell-to-cell variability in the number of nascent RNA molecules , those still in the process of being synthesized by the RNA polymerase , carries the signature of transcriptional dynamics in cells . We analyze published nascent RNA distributions for a set of yeast genes and show that the data is inconsistent with a single-step model of transcription initiation . Instead we propose a coarse-grained model where initiation happens not in one but in two sequential steps . Our analytical framework can be used to extract quantitative information about the dynamics of transcription from single-cell sequencing data , as well as from single-molecule imaging and electron micrographs of fixed cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Deciphering Transcriptional Dynamics In Vivo by Counting Nascent RNA Molecules |
Heterochromatin contains a significant part of nuclear DNA . Little is known about the mechanisms that govern heterochromatic DNA stability . We show here that in the yeast Saccharomyces cerevisiae ( i ) DNA mismatch repair ( MMR ) is required for the maintenance of heterochromatic DNA stability , ( ii ) MutLα ( Mlh1-Pms1 heterodimer ) , MutSα ( Msh2-Msh6 heterodimer ) , MutSβ ( Msh2-Msh3 heterodimer ) , and Exo1 are involved in MMR at heterochromatin , ( iii ) Exo1-independent MMR at heterochromatin frequently leads to the formation of Pol ζ-dependent mutations , ( iv ) MMR cooperates with the proofreading activity of Pol ε and the histone acetyltransferase Rtt109 in the maintenance of heterochromatic DNA stability , ( v ) repair of base-base mismatches at heterochromatin is less efficient than repair of base-base mismatches at euchromatin , and ( vi ) the efficiency of repair of 1-nt insertion/deletion loops at heterochromatin is similar to the efficiency of repair of 1-nt insertion/deletion loops at euchromatin .
Mutations fuel evolution and are also the basis of numerous diseases including cancer [1] . Base substitutions , 1-bp deletions , and 1-bp insertions are the most common mutations in the cell . Mutations are formed as a result of DNA damage and replication errors . Cells have evolved multiple mechanisms that suppress mutations [1 , 2] . The high-fidelity DNA synthesis and DNA mismatch repair ( MMR ) play major roles in protecting the genome from mutations [3–8] . Replicative DNA polymerases achieve the high-fidelity DNA synthesis by selecting correct dNTPs and by proofreading DNA synthesis errors [4 , 9] . The nucleotide selectivity step is responsible for keeping the error rate of replicative DNA synthesis at the level of 10−4–10−6 , and proofreading further increases the fidelity of replicative DNA synthesis by 10–1 , 000 fold . Nuclear DNA is mainly synthesized by DNA polymerases ( Pols ) α , δ , and ε [10–12] . At the eukaryotic DNA replication fork , Pol ε performs the bulk of leading-strand synthesis , and Pol δ carries out the majority of lagging-strand synthesis [10–15] . Pol δ and Pol ε are the only two nuclear DNA polymerases that have proofreading activities [16 , 17] . Pol δ can proofread both its own errors and those of Pol ε , but Pol ε can only proofread its own mistakes [18] . MMR removes a large fraction of DNA polymerase errors that escape proofreading [17 , 19–21] . As a result , MMR decreases the level of spontaneous mutations in the genome by ~100 fold [22] . MMR efficiency is different at different genomic sites [23] , and it is higher on the lagging than leading strand [24] . Strand breaks in the leading and lagging strands are likely to be the signals that direct eukaryotic MMR to remove mismatches on the daughter strands [25–28] . MutLα- and Exo1-dependent MMR is a major mechanism for correction of DNA polymerase errors in eukaryotic cells [29 , 30] . This mechanism includes mismatch excision and DNA re-synthesis steps , and it involves a mismatch recognition factor ( MutSα or MutSβ ) , the replicative clamp PCNA , the PCNA loader RFC , and Pol δ , in addition to MutLα endonuclease and the 5’→3’ exonuclease Exo1 [29–58] . Loss of Exo1 causes a modest defect in MMR , indicating that MMR is able to occur via Exo1-independent mechanism ( s ) [41 , 45 , 48 , 57 , 59–61] . Exo1-independent MMR is not as well understood as Exo1-dependent MMR [29 , 30 , 62] . A genetic study implicated proofreading activity of Pol δ in Exo1-independent MMR in budding yeast [45] , and biochemical analyses of defined systems provided evidence that MutSα , MutLα endonuclease , PCNA , RFC , and Pol δ-driven strand displacement DNA synthesis are involved in human Exo1-independent MMR [57 , 61] . Like proofreading and MMR , the histone acetyltransferase Rtt109 [63–65] is required for high-fidelity DNA replication [66] . Loss of Rtt109 increases the spontaneous mutation rate [66] . Rtt109 supports DNA replication fidelity by acetylating histone H3 on the K56 residue [66] . Histone H3 K56ac is an abundant histone modification associated with S phase and DNA replication in S . cerevisiae [67 , 68] . The mechanism by which Rtt109-dependent H3 K56ac maintains the replication fidelity is not well understood , but a genetic analysis [66] indicated that it is likely to entail Rad51 and Rad52 , key components of the homologous recombination machinery [69] . Nuclear DNA is packaged into euchromatin and heterochromatin soon after the passage of the DNA replication fork [70 , 71] . Compared to euchromatin , heterochromatin is more condensed . Transcription in heterochromatin is silenced/suppressed whereas it is active in euchromatin . In S . cerevisiae , heterochromatin is present at HMR , HML , subtelomeric regions , and the rDNA locus [70 , 72] . Sir2 , Sir3 , and Sir4 proteins are the structural components of heterochromatin at HMR , HML , and subtelomeric regions [70] , but heterochromatin at rDNA does not include the latter two proteins . A Sir2-Sir3-Sir4-nucleosome complex is the basic unit of heterochromatin at the HMR , HML , and subtelomeric loci [70 , 72 , 73] . In this complex , Sir2-Sir3-Sir4 heterotrimer contacts the nucleosome via Sir3 and Sir4 . In addition to being a structural component of yeast heterochromatin , Sir2 also has a NAD+-dependent histone deacetylase activity that is required for heterochromatin formation [74] . In the process of heterochromatin formation Sir2 deacetylates the N-terminal tails of nucleosomal histones H3 and H4 , facilitating loading of Sir2-Sir3-Sir4 complexes onto the nucleosomes . Previous research has been mainly focused on investigating eukaryotic MMR in the context of naked DNA and euchromatin . Up to date , only two studies have analyzed MMR at heterochromatin [75 , 76] . One of the studies used bioinformatic approaches to investigate distribution of base-base substitutions at 1-Mb resolution in late heterochromatic and early euchromatic regions of cancer genomes [75] . It provided evidence that in cancer cells the MMR system removes base-base mismatches less efficiently at heterochromatin than at euchromatin . The other study revealed that Msh6-dependent correction of small insertion/deletion loops and base-base mismatches in S . pombe is less efficient at heterochromatin than at euchromatin [76] . In this study , we examined MMR at heterochromatin in S . cerevisiae . We determined that MMR at heterochromatin involves MutLα , MutSα , MutSβ , an Exo1 and that MMR occurring at heterochromatin in the absence of Exo1 is an error-prone process . In addition , we determined that MMR cooperates with the Pol ε proofreading activity and Rtt109 to maintain the stability of heterochromatic DNA . In agreement with a previous study [75] , we established that the efficiency of repair of base-base mismatches at heterochromatin is lower than the efficiency of repair of base-base mismatches at euchromatin . However , we found that the efficiency of 1-nt insertion/deletion loop repair at heterochromatin is very similar to the efficiency of 1-nt insertion/deletion loop repair at euchromatin . This finding does not support the model that the efficiency of MMR at heterochromatin is reduced by lower accessibility of MMR proteins to heterochromatic DNA compared to euchromatic DNA [75 , 76] .
We started this work to study the impact of MMR on spontaneous mutation rates at heterochromatic loci in S . cerevisiae . In the majority of our experiments , we utilized a forward mutation assay that took advantage of the URA3 gene . In this assay , yeast cells that acquire loss-of-function mutations in heterochromatic URA3 are selected on a medium containing 5-FOA ( 5-fluoroorotic acid ) and 5 mM nicotinamide ( NAM ) . NAM was included into the selective medium because it switches the heterochromatic URA3 to a euchromatic state , which leads to its expression [77 , 78] . We first confirmed that when the URA3 reporter was inserted at hmr ( Fig 1A ) in a wild-type strain , it was in a heterochromatic state ( Fig 1B ) . This is in a full agreement with a previous work that showed that a similar reporter , K . lactis URA3 , is heterochromatic at hmr [78] . We then established that MSH2 was not required to maintain the heterochromatic status of URA3 at hmr ( Fig 1C ) . Next , we studied how MSH2 deletion in the wild-type strain affected the 5-FOAR mutation rate at heterochromatic hmr::URA3 . As shown in Fig 1D , we found that deletion of MSH2 in the wild-type strain increased the 5-FOAR mutation rate at heterochromatic hmr::URA3 by 8 fold . This finding showed that MMR was involved in the protection of heterochromatic hmr::URA3 from mutations . We also investigated whether loss of MSH2 affected mutation rates at two other heterochromatic loci: hml::URA3 and Chr VII-L::URA3 ( Fig 2 ) . ( The latter locus is near the left telomere of Chr VII [79] . ) The data revealed that MSH2 deletion in the wild-type strains increased the 5-FOAR mutation rates at heterochromatic hml::URA3 and Chr VII-L::URA3 loci by 12 and 6 fold , respectively ( Fig 2 ) . Collectively , the results of these experiments demonstrated that MMR was essential for the maintenance of heterochromatic DNA stability in S . cerevisiae . Yeast MutLα ( Mlh1-Pms1 heterodimer ) , MutSα ( Msh2-Msh6 heterodimer ) , and MutSβ ( Msh2-Msh3 heterodimer ) play important roles in MMR at euchromatic loci [31 , 36 , 80 , 81] . We performed experiments to study whether these proteins contributed to MMR at heterochromatic hmr::URA3 . We determined that the FOAR mutation rate in the mlh1Δ , pms1Δ , or msh3Δ msh6Δ strain was similar to that in the msh2Δ strain ( Fig 1D ) . Additionally , we determined that the FOAR mutation rate for the msh6Δ strain was 3 times higher than that for the wild-type strain and that the msh3Δ strain displayed the same FOAR mutation rate as the wild-type strain ( Fig 1D ) . Collectively , these experiments revealed that ( i ) MutLα , MutSα , and MutSβ were involved in MMR at heterochromatin and ( ii ) MutSα played a more important role in MMR at heterochromatic hmr::URA3 than MutSβ . Previous genetic studies implicated Exo1 in MMR at euchromatic regions [41 , 48 , 59 , 60] . We studied whether Exo1 had a role in MMR at a heterochromatic locus . We established that deletion of EXO1 in a wild-type strain increased the 5-FOAR mutation rate at heterochromatic hmr::URA3 by 4-fold and that msh2Δ was epistatic to exo1Δ for 5-FOAR mutations at heterochromatic hmr::URA3 ( Fig 1D ) . These data suggested that loss of EXO1 caused a strong defect in MMR at heterochromatin . We then determined and analyzed the ura3 mutation spectra at heterochromatic hmr in the wild-type , exo1Δ , msh2Δ , and msh2Δ exo1Δ strains ( Table 1 , Figs 3 and 4 ) . It can be seen that the most common mutations in the ura3 mutation spectra of the wild-type and exo1Δ strains were base substitutions , whereas the most common mutations in the ura3 mutation spectra of the msh2Δ and msh2Δ exo1Δ strains were 1-bp deletions . Further analysis of the data revealed that ~95% of 1-bp deletions in the msh2Δ and msh2Δ exo1Δ spectra were within N≥3 mononucleotide runs ( Fig 4 ) , but only ~60% and ~15% of 1-bp deletions in the wild-type and exo1Δ spectra , respectively , were within such runs ( Fig 3 ) . To determine whether the ura3 mutation spectra of the msh2Δ , msh2Δ exo1Δ , exo1Δ , and wild-type strains were statistically different from each other or not , we performed the pairwise comparisons using χ2 test of independence and adjusted the p values with the Bonferroni correction . The data showed that there was no statistical difference between the ura3 mutation spectra of the msh2Δ and msh2Δ exo1Δ strains , whereas those two spectra were statistically different from the ura3 mutation spectra of the wild-type and exo1Δ strains ( Table 2 ) . In addition , we conducted the pairwise comparisons of the ura3 mutation spectra of the msh2Δ , msh2Δ exo1Δ , and wild-type strains utilizing a Monte Carlo modification of the Pearson χ2 test of spectra homogeneity [82] . For this statistical analysis the spectra were arranged in a way ( S2 Table ) that was different from the one shown in Table 1 . The results of this statistical analysis revealed that the ura3 mutation spectra of the msh2Δ and msh2Δ exo1Δ strains were not statistically different from each other ( χ2 = 7 . 9 , P = 0 . 8945 ) , but were statistically different from the ura3 mutation spectrum of the wild-type strain ( χ2 = 38 . 1 and 39 . 7 , respectively , P < 10−5 , the critical 5% value = 20 . 9 ) . Our findings that the ura3 mutation spectra of the msh2Δ and msh2Δ exo1Δ strains were not statistically different from each other and that msh2Δ was epistatic to exo1Δ with respect to FOAR mutations at heterochromatic hmr::URA3 ( Fig 1D ) demonstrated that Exo1 was involved in MMR at heterochromatin . Moreover , our finding that the ura3 mutation spectrum of an exo1Δ strain was statistically different from the ura3 mutation spectra of the msh2Δ and msh2Δ exo1Δ strains showed that Exo1-independent MMR at heterochromatin produced mutational intermediates . The REV3 gene encodes the catalytic subunit of the error-prone Pol ζ [83–85] . Prior work showed that deletion of REV3 in an exo1Δ strain suppresses the mutation rate at euchromatic CAN1 [81] . In agreement with this , we found that introduction of rev3Δ into an exo1Δ strain suppressed the mutation rate at euchromatic Chr V::URA3 ( Fig 5 ) . To understand the origin of mutational intermediates , which arose at heterochromatic hmr::URA3 as a result of Exo1-independent MMR , we carried out experiments to determine whether deletion of the REV3 gene in the exo1Δ , msh2Δ , and msh2Δ exo1Δ strains affected the FOAR mutation rates . These experiments demonstrated that deletion of REV3 in the exo1Δ strain decreased the FOAR mutation rate to the level observed in the wild-type strain , but deletion of REV3 in the msh2Δ and msh2Δ exo1Δ strains did not change the FOAR mutation rates ( Fig 1D ) . Thus , FOAR mutations produced at heterochromatic hmr::URA3 in the exo1Δ strain were REV3-dependent , whereas FOAR mutations produced at heterochromatic hmr::URA3 in the msh2Δ and msh2Δ exo1Δ strains were REV3-independent . Based on these results , we concluded that Exo1-independent MMR at heterochromatin often produced Rev3-dependent mutational intermediates . Our analysis of the mutation rates in the msh6Δ rev3Δ , msh6Δ exo1Δ rev3Δ , msh3Δ rev3Δ , and msh3Δ exo1Δ rev3Δ strains was consistent with this conclusion ( S3 Table ) . The MMR system removes DNA polymerase errors at euchromatic loci [17 , 19–21 , 45 , 86] . To examine whether the MMR system removes Pol ε errors at a heterochromatic locus , we constructed an hmr::URA3 pol2-4 pms1Δ strain . ( pol2-4 encodes the catalytic subunit of Pol ε , which lacks the proofreading activity [16] . ) In agreement with the previous work [17] , we noticed that the pol2-4 pms1Δ mutant grew poorly and single colonies of this mutant were of different sizes ( small , medium and large ) . Our analysis showed that the relative mutation rate in the pol2-4 pms1Δ double mutant was 11 times higher than the sum of the relative mutation rates in the single mutants ( i . e . there was a strong synergistic relationship between pol2-4 and pms1Δ for FOAR mutations at heterochromatic hmr::URA3 ) ( Fig 1D ) . The presence of the strong synergistic relationship demonstrated that at heterochromatic hmr::URA3 Pol ε errors that were not removed by its proofreading activity were corrected by the MMR system . The MMR system and histone acetyltransferase Rtt109 act in overlapping pathways to preserve the replication fidelity at euchromatic sites [66] . We explored whether a similar cooperation between the MMR system and Rtt109 took place at heterochromatic hmr::URA3 . The experiments revealed that there was a weak synergistic relationship between msh2Δ and rtt109Δ for ura3 mutations at heterochromatic hmr::URA3 ( Fig 1D ) . To better understand the nature of the cooperation , we determined and analyzed the spectra of ura3 mutations at heterochromatic hmr in the rtt109Δ and msh2Δ rtt109Δ strains ( S1 Fig and Table 3 ) . The data indicated that the MMR system and Rtt109 acted in overlapping pathways that increased the replication fidelity by suppressing base substitutions and 1-bp deletions . During the course of this work , we noticed that the mutation rates at heterochromatic hmr::URA3 , hml::URA3 , and Chr VII-L::URA3 in our wild-type strains ( Figs 1 and 2 ) were 2–4 times higher than the mutation rates at euchromatic CAN1 in other wild-type strains [45 , 66 , 80] . To ensure that the observed difference in the mutation rates was not a result of genetic background variations and/or the use of the different mutation reporters , we inserted URA3 at a Chr V euchromatic locus ( where it is normally located ) in a wild-type strain , which was isogenic to the strains carrying the heterochromatic reporters ( Figs 1 and 2 ) , and measured an FOAR mutation rate in this strain . As shown in Fig 6A , the mutation rate at euchromatic Chr V::URA3 of this wild-type strain was ~ 6–11 times lower than a mutation rate at a heterochromatic locus of a similar wild-type strain . This observation was consistent with an idea that in wild-type strains , heterochromatic DNA was less stable than euchromatic DNA . To test this idea we disrupted heterochromatin by introduction of sir2Δ , sir3Δ , sir4Δ , sir2-N345A , or hmr-EΔ [87] mutation into a wild-type strain and measured the mutation rates in the constructed strains ( Fig 6A ) . ( The N345A mutation inactivates the Sir2 histone deacetylase activity , which is required for heterochromatin formation [74] . ) Analysis of the data showed that the mutation rate at hmr::URA3 in sir2Δ , sir3Δ , sir4Δ , sir2-N345A , or hmr-EΔ strain was 3–7 times lower than the mutation rate at heterochromatic hmr::URA3 in the wild-type strain ( Fig 6A ) . These findings provided a strong support for the idea that in wild-type strains , heterochromatic DNA was less stable than euchromatic DNA . Additional support for this idea was obtained in experiments in which we established that the mutation rate at euchromatic hmr::CAN1 in a sir2Δ strain was half that at heterochromatic hmr::CAN1 in a wild-type strain ( Fig 6B ) . A previous elegant study demonstrated that the efficiency of MSH2-dependent repair of small insertion/deletion loops varies across the yeast genome [23] . MMR efficiency variations are likely to have important evolutionary consequences . In light of this information , we calculated MMR efficiencies at the heterochromatic and euchromatic loci . The calculated efficiencies of MMR at heterochromatic hmr::URA3 , hml::URA3 , and Chr VII-L::URA3 were 88% , 92% , and 84% , respectively , and the calculated efficiency of MMR at euchromatic Chr V::URA3 was 97% . Thus , these data suggested that MMR was less efficient at heterochromatin than euchromatin . We next determined that ( i ) the efficiencies of repair of base-base mismatches and 1 nt-insertion/deletion loops at heterochromatic hmr::URA3 were 72% and 97–98% , respectively , and ( 2 ) the efficiencies of repair of base-base mismatches and 1 nt-insertion/deletion loops at euchromatic Chr V::URA3 were 96% and ~98–99% , respectively ( Table 4 ) . Based on these data , we concluded that the efficiency of repair of base-base mismatches at heterochromatic hmr::URA3 was significantly reduced compared to the efficiency of repair of base-base mismatches at euchromatic Chr V::URA3 . We thought that the reduced efficiency of repair of base-base mismatches at heterochromatic hmr::URA3 might be a consequence of the heterochromatic environment . We reasoned that if this idea was correct , then disruption of heterochromatin at hmr::URA3 by deletion of SIR2 should increase the efficiency of repair of base-base mismatches at this locus . Our experiments showed that the efficiency of repair of base-base mismatches at euchromatic hmr::URA3 in a sir2Δ strain was 90% ( Table 4 ) . Thus , disruption of heterochromatin at hmr::URA3 by sir2Δ mutation increased the efficiency of repair of base-base mismatches at this locus from 72% to 90% . This observation suggested that the heterochromatic environment decreased the efficiency of repair of base-base mismatches .
Significant progress has been made in understanding of MMR at euchromatin since the demonstration of its importance for euchromatic DNA stability [31 , 88 , 89] . However , much less is known about MMR at heterochromatin [75 , 76] . In this work , we have found that inactivation of MMR in S . cerevisiae significantly increases the spontaneous mutation rates at heterochromatic hmr::URA3 , hml::URA3 , and Chr VII-L::URA3 loci ( Figs 1 and 2 ) . These findings have demonstrated that MMR is essential for the maintenance of heterochromatic DNA stability in S . cerevisiae . Furthermore , our analysis of the genetic interactions has provided strong evidence that in budding yeast MMR cooperates with Pol ε proofreading and Rtt109 to protect heterochromatic DNA from spontaneous mutations ( Fig 1D ) . Previous studies revealed that at euchromatic CAN1 the mutation rates in msh2Δ strains are 27–40 times as high as those in wild-type strains [45 , 66 , 80] . Consistent with those studies , we have established that at a different euchromatic locus , Chr V::URA3 , the mutation rate in an msh2Δ strain is 37 times higher than that in a wild-type strain ( S4 Table ) . However , at each of the three heterochromatic loci , the mutation rate in the msh2Δ strain is only 6–12 times that of the wild-type strain ( Figs 1 and 2 ) . Thus , the budding yeast data substantiate the view that MMR is more important for euchromatic DNA stability than for heterochromatic DNA stability [75 , 76] . The experiments , in which we have determined that at euchromatic hmr::URA3 the mutation rate in an msh2Δ sir2Δ strain is 23 times that in an sir2Δ strain , have provided a direct support for such a view ( Fig 6A ) . Our data have implicated MutLα , MutSα , and MutSβ in MMR at heterochromatin ( Fig 1D ) . MutLα , MutSα , and MutSβ are also involved in MMR at euchromatin [31 , 41 , 44 , 48 , 80 , 90] . Thus , it appears that the roles of MutLα , MutSα , and MutSβ at heterochromatin are not very different from those at euchromatin [29 , 30 , 57] . Biochemical studies with cell-free extracts and reconstituted systems demonstrated the importance of the 5’-3’ exonuclease Exo1 for the mismatch excision step in the process that repairs base-base mismatches and 1-nt insertion/deletion loops [46–49 , 91] . Such a role for Exo1 in MMR is in full agreement with genetic analyses of this process at euchromatic sites in yeast and mice [35 , 41 , 45 , 48] . However , the importance of Exo1 for MMR at euchromatin was brought into question by the finding that the mutator phenotype of a yeast exo1Δ strain was not consistent with an MMR defect [81] . We have conducted experiments to investigate whether Exo1 plays a role in MMR at heterochromatin . In these experiments we have found that ( i ) at heterochromatic hmr::URA3 the mutation rate in an exo1Δ strain is 4 times that in the wild-type strain and half that in the msh2Δ strain; ( ii ) deletion of MSH2 is epistatic to deletion of EXO1 for spontaneous FOAR mutations at heterochromatic hmr::URA3 ( Fig 1D ) ; and ( iii ) the ura3 mutation spectrum at heterochromatic hmr in an msh2Δ exo1Δ strain is not statistically different from the ura3 mutation spectrum at the same locus in an msh2Δ strain ( Table 2 and text in Results section ) . Collectively , these findings have shown that Exo1 plays a major role in MMR at heterochromatin . We have determined that ( i ) the mutation rate at heterochromatic hmr::URA3 in an exo1Δ strain is REV3-dependent whereas the mutation rate at heterochromatic hmr::URA3 in an msh2Δ strain is REV3-independent ( Fig 1D ) and ( ii ) the mutation spectrum at heterochromatic hmr::URA3 in an exo1Δ strain is statistically different from the mutation spectra at heterochromatic hmr::URA3 in the msh2Δ and msh2Δ exo1Δ strains ( Table 2 ) . Furthermore , it can be seen that the majority of 1-bp deletions in the spectra of the msh2Δ and msh2Δ exo1Δ strains are within the N≥5 mononucleotide runs ( Fig 4 ) , whereas not a single 1-bp deletion in the mutation spectrum of the exo1Δ strain is in any of these runs . These findings have indicated that Exo1-independent MMR at heterochromatin is an error-prone process that leads to the formation of Pol ζ-dependent mutations ( Fig 7 ) . Because Exo1 plays more important role in MMR on the lagging strand [92 , 93] , this error-prone process is likely to preferentially occur on the lagging than leading strand . We would like to note that it has been suggested that the mutator phenotype of exo1Δ strains reveals the participation of Exo1 in both MMR and an MMR-unrelated mutation avoidance pathway [81] . However , if Exo1 participated in an MMR-unrelated mutation avoidance pathway functioning across the genome , then a mutation spectrum at a heterochromatic site in an msh2Δ exo1Δ strain should have been different from a mutation spectrum at the same site in an msh2Δ strain . In contrast , we determined that the mutation spectrum at heterochromatic hmr::URA3 in an msh2Δ exo1Δ strain is not statistically different from the mutation spectrum at the same locus in an msh2Δ strain ( Table 2 ) . We propose that error-prone Exo1-independent MMR at heterochromatin consists of three principal steps ( Fig 7 ) . In the initial step , Exo1-independent MMR at heterochromatin leads to the formation of an excessive number of MutLα endonuclease-dependent strand breaks in the discontinuous daughter strand . This happens because mismatch removal step in Exo1-independent MMR is slowed down by the absence of Exo1 , which permits MutLα to produce additional strand breaks in the discontinuous daughter strand . Next , one of these strand breaks is used by Pol ζ ( REV3-REV7-Pol31-Pol32 complex [94 , 95] ) to introduce a mismatch , and the original mismatch is corrected . Finally , the Pol ζ-produced mismatch escapes correction , perhaps due to the presence of an MMR impediment , and then is fixed as a mutation in the next round of DNA replication . We believe that this model also provides a satisfactory explanation for the formation of REV3-dependent mutations at euchromatin in exo1Δ strains [81] ( Fig 5 ) . Although we do not know what could block removal of the Pol ζ-produced mismatch by the Exo1-lacking MMR system , past work revealed that the nucleosome is able to function as an MMR impediment [96 , 97] . Our experiments have shown that the mutator phenotypes of the msh3Δ exo1Δ and msh6Δ exo1Δ strains are REV3-dependent ( S3 Table ) and the mutator phenotype of the msh2Δ exo1Δ strain is REV3-independent ( Fig 1D ) . Taken together , these data have demonstrated that Exo1-independent MutSα-dependent MMR at heterochromatin in msh3Δ exo1Δ cells and Exo1-independent MutSβ-dependent MMR at heterochromatin in msh6Δ exo1Δ cells often causes the formation of REV3-dependent mutations . Wei et al . ( 2003 ) found that loss of Exo1 predisposes mice to the development of lymphomas [48] . That finding allowed the authors to propose that EXO1 mutations may predispose humans to cancer [48] . However , strong evidence to support this proposal is still missing [98 , 99] . In yeast , EXO1 deficiency increases the mutation rates in the forward mutation assays to the levels that are 70–140% of those caused by MSH6 deficiency [45] ( Fig 1D ) . Based on these data , we envision that cancers triggered by EXO1 mutations may be nearly as common as those initiated by MSH6 mutations [100] . Previous work and our mutation spectrum data suggest that cancers triggered by EXO1 deficiency will rarely display microsatellite instability [41 , 45 , 60] ( Figs 3 and 4 ) , which is a hallmark of MMR deficiency caused by MSH2 , MLH1 , or PMS2 inactivation . Thus , to better understand the relationship between EXO1 and cancer , it may be necessary to analyze microsatellite-stable , but not microsatellite-unstable , cancers that display increased mutation rates . We have measured mutation rates at several genomic sites; one of the sites is euchromatic and the others are heterochromatic . We have found that the mutation rate at euchromatic Chr V::URA3 in a wild-type strain is ~9 , ~6 , and ~11 times lower than the mutation rates at heterochromatic hmr::URA3 , hml::URA3 , and Chr VIIL::URA3 , respectively , in similar wild-type strains ( Fig 6A ) . Moreover , we have found that disruption of heterochromatin in a wild-type strain by sir2Δ , sir3Δ , sir4Δ , sir2-N345A , or hmr-EΔ [87] ( S1 Table ) decreases the mutation rate at hmr 2–7 fold ( Fig 6 ) . Together , these data have shown that in S . cerevisiae the heterochromatic DNA is less stable than the euchromatic DNA , which supports the idea that the chromatin environment is a key factor that affects the stability of DNA [23 , 101] . Sun et al . ( 2016 ) have recently described that in an S . pombe msh6Δ strain mutation rate in heterochomatin is ~50% higher that in euchromatin [76] . Consistent with this , we have found that the mutation rate at heterochromatic hmr::URA3 in the msh2Δ strain is 1 . 5–2 times higher than those at euchromatic hmr::URA3 in the msh2Δ hmr-EΔ and msh2Δ sir2Δ strains ( Fig 6A ) . Together , these findings suggest that the heterochromatic environment modestly increases the level of DNA replication errors at heterochromatic sites . Our data ( Figs 1–3 and Table 4 ) corroborate the view that MMR efficiency varies from one locus to another and is an important factor that contributes to locus-specific mutation rates [23 , 75] . Surprisingly , at heterochromatic hmr::URA3 the efficiency of repair of base-base mismatches is only 72% but the efficiency of repair of 1-nt insertion/deletion loops is 97–98% ( Table 4 ) . Thus , the heterochromatic environment decreases the efficiency of repair of base-base mismatches but has a little of influence on the efficiency of repair of 1-nt deletion loops ( Table 4 ) . This finding argues against the model that the efficiency of MMR at heterochromatin is reduced by lower accessibility of MMR proteins to heterochromatic DNA compared to euchromatic DNA [75 , 76] . We do not know how the heterochromatic environment reduces the efficiency of repair of base-base mismatches . It is likely that the overall efficiency of repair of base-base mismatches at heterochromatin is decreased because many base-base mismatches in newly replicated heterochromatic DNA are poor substrates for the MMR reaction . We envision that these base-base mismatches are poor substrates for the MMR reaction because they contain damaged bases , which arise as a result of low level of base excision repair at heterochromatin . In summary , we have performed a detailed analysis of MMR at heterochromatin . Our research has demonstrated that MMR involves MutLα , MutSα , MutSβ , and Exo1 to maintain heterochromatic DNA stability . Surprisingly , it has also revealed that Exo1-independent MMR at heterochromatin is an error-prone process and that the repair of 1-nt insertion/deletion loops at heterochromatin is nearly as efficient as the repair of 1-nt insertion/deletion loops at euchromatin .
The yeast S . cerevisiae strains are derivatives of BY4742 ( MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 ) . The wild-type strains are BKDY155 ( MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 hmr::URA3 hml::HphMX ) , BKDY157 ( MATα his3Δ1 leu2Δ0 lys2Δ0 Chr V::URA3 hml::HphMX ) , BKDY438 ( MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 hml::URA3 ) , BKDY541 ( MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 Chr VII-L::URA3 ) , BKDY834 ( MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 can1::LEU2 hmr::CAN1 ) , and FKY1292 ( MATα his3Δ1 leu2Δ0 lys2Δ0 Chr V::URA3 ) . In BKDY157 and FKY1292 , a DNA sequence between nucleotides 115 , 949 and 117 , 045 of Chr V is replaced with URA3 . Each of the mutant strains is isogenic to one of the wild-type strains . In the hmr-EΔ strain , the 56-bp HMR-E region ( Chr III 292 , 674–292 , 729 ) [87] was replaced with a LEU2 cassette . To create the gene deletions , PCR-amplified disruption cassettes were introduced into yeast cells by lithium acetate/PEG4000/DMSO transformation . The presence of each gene deletion was confirmed by locus and disruption cassette-specific PCRs . The pol2-4 mutation was introduced into the chromosomal POL2 gene using the integration-excision method , and the sir2-N345A mutation was inserted into the chromosomal SIR2 gene utilizing a previously described technique [102] . The spontaneous mutation rates were measured using a fluctuation test . At least 9–18 cultures , which were started from single colonies of two-four independent isolates of the same genotype , were used to determine the spontaneous mutation rate . The cultures were grown to saturation in 3 ml YPDAU medium ( 1% yeast extract , 2% bacto-peptone , 2% dextrose , 60 mg/L adenine , 60 mg/L uracil ) at 30°C . The saturated cultures were diluted in sterile water , and appropriate dilutions were plated on a synthetic complete ( SC ) medium to determine the total number of cells in the cultures and on a selective medium to determine the total number of the mutant cells in the cultures . Unless noted otherwise , the selective medium for FOAR cells was a SC medium containing 1 g/L 5-FOA and 5 mM NAM ( SC + 5-FOA + NAM ) , and the selective medium for CanR cells was a SC medium that lacked arginine and contained 60 mg/L L-canavanine and 5 mM NAM . The plates were incubated for 3–4 days at 30°C , and the colonies were counted . 5–70% of the FOAR colonies grew on a SC—Ura + 5 mM NAM medium . FOAR cells , which formed these colonies , were excluded from calculations of the mutation rates . To identify FOAR colonies that were Ura+ , FOAR colonies formed on the fluctuation test plates were replica-plated onto the SC—Ura + 5 mM NAM medium , and the plates were incubated for 1 day at 30°C . The mutation rates were calculated from the total numbers of cells and mutants in the cultures using the Drake’s formula μ = ƒ/ln ( Nμ ) [103] , where μ is mutation rate per replication , ƒ is the median mutant frequency , and N is population size . Where indicated , the significance of the observed differences in the mutation rates was assessed with the Mann-Whitney U two-tailed test ( GraphPad Prism 6 software ) , in which the null hypothesis is that there is no difference between the two data sets . To examine the relationship between the nominal variables of spectra and mutation type ( Table 2 ) , categorical variables were summarized with frequencies and percentages , and a χ2 test of independence with a Bonferroni correction for multiple comparisons was utilized . In a different method , a Monte Carlo modification of the Pearson χ2 test of spectra homogeneity [82] was used to compare mutation distributions ( S2 Table ) . The calculations were done using the COLLAPSE program [104] . In order to determine the ura3 mutation spectrum , ~50–100 patches each started from a different single colony were grown on YPDAU plates ( 1% yeast extract , 2% bacto-peptone , 2% dextrose , 60 mg/L adenine , 60 mg/L uracil , 2% agar ) . The patches were next replica-plated on the SC + 5-FOA + 5 mM NAM , followed by incubation of the plates for 1 day at 30°C . The patches that were formed on the SC + 5-FOA + 5 mM NAM plates were replica-plated on fresh SC + 5-FOA + 5 mM NAM plates , and the plates were incubated for 2–3 days at 30°C . A single FOAR colony was randomly selected from each patch , purified on a SC + 5-FOA + 5 mM NAM plate , and propagated on a YPDAU plate . The patches were then replica-plated on SC—Ura + 5 mM NAM plates . Patches that grew on the SC—Ura + 5 mM NAM plates were not analyzed further . Genomic DNAs of the remaining FOAR patches were isolated with a MasterPure Yeast DNA purification kit ( Epicentre ) . Each of these genomic DNAs was used as a template to PCR-amplify a 1 . 4-kb DNA fragment encompassing the entire length of ura3 ORF with primers #1 ( 5’- GAGAATAAGCGCAGGTACTCCTG -3’ ) and #2 ( 5’- CGCCATATACGAAAATGTTGGTG -3’ ) . The amplified DNA fragments were purified with a PCR purification kit ( Thermo Fisher ) and sequenced to determine ura3 mutations . | Eukaryotic mismatch repair is an important intracellular process that defends DNA against mutations . Inactivation of mismatch repair in human cells strongly increases the risk of cancer initiation and development . Although significant progress has been made in understanding mismatch repair at euchromatin , mismatch repair at heterochromatin is not well understood . Baker’s yeast is a key model organism to study mismatch repair . We determined that in baker’s yeast ( 1 ) mismatch repair protects heterochromatic DNA from mutations , ( 2 ) the MutLα , MutSα , MutSβ , and Exo1 proteins play important roles in mismatch repair at heterochromatin , ( 3 ) Exo1-independent mismatch repair at heterochromatin is an error-prone process; ( 4 ) mismatch repair cooperates with two other intracellular processes to protect the stability of heterochromatic DNA; and ( 5 ) the efficiency of repair of base-base mismatches at heterochromatin is lower than the efficiency of repair of base-base mismatches at euchromatin , but the efficiency of 1-nt insertion/deletion loop repair at heterochromatin is similar to the efficiency of 1-nt insertion/deletion loop repair at euchromatin . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"deletion",
"mutation",
"insertion",
"mutation",
"euchromatin",
"mutation",
"substitution",
"mutation",
"fungi",
"model",
"organisms",
"experimental",
"organism",
"systems",
"mismatch",
"repair",
"epigenetics",
"dna",
"chromatin",
"research",
"and",
"analysis",
"methods",
"heterochromatin",
"saccharomyces",
"chromosome",
"biology",
"gene",
"expression",
"genetic",
"loci",
"yeast",
"biochemistry",
"eukaryota",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"repair",
"saccharomyces",
"cerevisiae",
"yeast",
"and",
"fungal",
"models",
"organisms"
] | 2017 | Involvement of DNA mismatch repair in the maintenance of heterochromatic DNA stability in Saccharomyces cerevisiae |
Buruli ulcer is a neglected emerging disease that has recently been reported in some countries as the second most frequent mycobacterial disease in humans after tuberculosis . Cases have been reported from at least 32 countries in Africa ( mainly west ) , Australia , Southeast Asia , China , Central and South America , and the Western Pacific . Large lesions often result in scarring , contractual deformities , amputations , and disabilities , and in Africa , most cases of the disease occur in children between the ages of 4–15 years . This environmental mycobacterium , Mycobacterium ulcerans , is found in communities associated with rivers , swamps , wetlands , and human-linked changes in the aquatic environment , particularly those created as a result of environmental disturbance such as deforestation , dam construction , and agriculture . Buruli ulcer disease is often referred to as the “mysterious disease” because the mode of transmission remains unclear , although several hypotheses have been proposed . The above review reveals that various routes of transmission may occur , varying amongst epidemiological setting and geographic region , and that there may be some role for living agents as reservoirs and as vectors of M . ulcerans , in particular aquatic insects , adult mosquitoes or other biting arthropods . We discuss traditional and non-traditional methods for indicting the roles of living agents as biologically significant reservoirs and/or vectors of pathogens , and suggest an intellectual framework for establishing criteria for transmission . The application of these criteria to the transmission of M . ulcerans presents a significant challenge .
Buruli ulcer ( BU ) is a serious necrotizing cutaneous infection caused by Mycobacterium ulcerans [1]–[7] . Before the causative agent was specifically identified , it was clinically given geographic designations such as Bairnsdale , Searles , and Kumasi ulcer , depending on the country [8]–[11] . BU is a neglected emerging disease that has recently been reported in some countries as the second most frequent mycobacterial disease in humans after tuberculosis ( TB ) [12]–[14] . Large lesions often result in scarring , contractual deformities , amputations , and disabilities [2]–[4] , [7] , [14]–[22] ( Fig . 1 ) . Approximately 80% of the ulcers are located on the limbs , most commonly on the lower extremities yet some variation exists [3] , [13] , [23] , [24] . In Africa , all ages and sexes are affected , but most cases of the disease occur in children between the ages of 4–15 years [5] , [13] , [17] , [25]–[28] . BU is a poorly understood disease that has emerged dramatically since the 1980's , reportedly coupled with rapid environmental change to the landscape including deforestation , eutrophication , dam construction , irrigation , farming ( agricultural and aquaculture ) , mining , and habitat fragmentation [3]–[7] , [29] , [30] . BU is a disease found in rural areas located near wetlands ( ponds , swamps , marshes , impoundments , backwaters ) and slow-moving rivers , especially in areas prone to flooding [3] , [4] , [23] , [27] , [29] , [31]–[36] ( Fig . 2 ) . Cases have been reported from at least 32 countries in Africa ( mainly west ) , Australia , Southeast Asia , China , Central and South America , and the Western Pacific [3] , [6] , [20] , [28] , [37] , [38] ( Fig . 3 ) . A number of cases have been reported in non-endemic areas of North America and Europe as a sequel to international travel [20] , [39]–[42] . Buruli ulcer disease is often referred to as the “mysterious disease” because the mode of transmission remains unclear , although several hypotheses have been proposed . The objectives of this article are to: 1 ) review the current state of knowledge on the ecology and transmission of M . ulcerans , 2 ) discuss traditional and non-traditional methods for investigating transmission , and 3 ) suggest an intellectual framework for establishing criteria for transmission .
Selection of the publications cited was based on the following approaches: 1 ) Direct knowledge of the authors of this manuscript regarding their background in the field of Buruli Ulcer research and knowledge of key papers and unpublished data; 2 ) Online search engines for Buruli Ulcer and Mycobacterium ulcerans ( predominantly PubMed , ISI Web of Knowledge , Web of Science , Centers for Disease Control ( CDC ) ; 3 ) Knowledge in the field of Buruli Ulcer research in that three of the authors ( Merritt , Small , Johnson ) are on the WHO Technical Advisory Committee for Buruli Ulcer in Geneva , Switzerland; 4 ) Review of the following websites: Buruli ulcer disease maintained by WHO in Geneva , Switzerland ( http://www . who . int/buruli/en ) , The Buruli Ulcer Disease Ecology Research Consortium ( BUDERC ) ( https://www . msu . edu/~budiseco/index . html ) ; and UBS Optimus Foundation ( http://www . stopburuli . org ) .
M . ulcerans is a slow-growing environmental mycobacterium that can be isolated from primary lesions after a 5–8 week incubation period , although up to 6 months may be required [43] , [44] . M . ulcerans falls into a group of closely related mycobacterial pathogens which comprise the M . marinum complex . The M . marinum complex contains mycobacterial species pathogenic for aquatic vertebrates and includes M . marinum ( fish ) , M . pseudoschottsii ( fish ) and M . liflandii ( frogs ) [45]–[48] . All of these species are characterized by slow growth rates and low optimal growth temperatures [49] . From a genomic standpoint , the species in the M . marinum complex can be considered a single species based on the fact that they share over 97% identity in the 16sRNA gene sequence [50] . However , practical considerations have led to the establishment of separate names based on differences in host tropism and pathogenesis analogous to other mycobacterial groupings , such as the M . avium and M . tuberculosis complexes . Genomic analysis suggests that M . ulcerans evolved from an M . marinum-like ancestor [21] , [51] through the acquisition of a large virulence plasmid and accumulation of multiple copies of insertion sequences , IS2404 and IS2606 . The genome has undergone considerable reductive evolution through a number of mutational events including transposon insertion . As a result , the genome has accumulated over 700 pseudogenes [21] , [52] . Although it has been reported that micro-aerophilic conditions enhance the growth of M . ulcerans in the BACTEC system [53] , the M . ulcerans genome strain lacks both nitrate and fumarate reductase systems , suggesting that M . ulcerans is considerably handicapped in the ability to grow under low oxygen conditions compared with M . marinum . The reported discrepancy in the oxygen requirements of M . ulcerans may be due to strain differences and requires closer investigation . A mutation in crtI , a key gene in the pathway for carotinoid biosynthesis , is suggested to compromise the ability of M . ulcerans to survive in direct sunlight [52] . A number of genes in ion transport and lipid biosynthesis have been lost and the repertoire of PE , PPE genes are considerably reduced compared with M . tuberculosis or M . marinum . Taken together , these results suggest that M . ulcerans is undergoing adaptation to a different and narrower niche than M . marinum . This idea has recently gained support from experimental work in which Medaka fish were infected with M . marinum and M . ulcerans . In these studies , M . marinum produced a lethal infection in Medaka , whereas M . ulcerans was not pathogenic and declined over a 23-week infection period ( L . Mosi , unpubl . data ) . The most important phenotypic characteristic of M . ulcerans is the low optimal growth temperature and the extremely restricted growth temperature range . M . marinum exhibits growth between 25–35°C , although the optimal growth temperature is 30–35°C [54] , [55] and many M . marinum isolates are capable of growth at 37°C . In contrast , growth of M . ulcerans strains under laboratory conditions is characterized by a remarkably narrow temperature range between 28–34°C and optimal growth of most strains is found between 30–33°C [56] . The restricted growth temperature of M . ulcerans is thought to play a substantial role in the pathogenesis of BU by limiting infection to the skin . The organism has never been isolated from internal organs of human patients or from bone in cases of osteomylelitis , or from the internal organs or blood of experimentally infected animals [51] , [57]–[59] . It has been recently reported that many isolates of M . ulcerans survive at 37°C for 13 days , although numbers decline after the first few days . No one has isolated or derived a strain capable of growth at 37°C [60] . The characteristic pathology of BU is mediated by a polyketide-derived macrolide exotoxin called mycolactone , which is cytotoxic and immunosuppressive [51] , [61] , [62] . Because of the large metabolic cost of producing mycolactone , it is likely that mycolactone plays an important role in the survival and growth of M . ulcerans in its environmental niche . Stringent criteria exist in biomedical research for indicting the roles of living agents as biologically significant reservoirs and/or vectors of pathogens . The application of these criteria to the transmission of M . ulcerans presents a significant challenge . The above review reveals that various routes of transmission may occur , varying amongst epidemiological setting and geographic region , and that there may be some role for living agents as reservoirs and as vectors of M . ulcerans , in particular aquatic insects , adult mosquitoes or other biting arthropods . It is also clear that the exact mode of transmission , if indeed there is a single mode , remains unknown . We briefly discuss the process by which a vector is incriminated to the point of as much certainty as is possible , and then discuss the application of this process to indictment of insect vectors for transmission of M . ulcerans . If Buruli ulcer is a vectored disease , intervention might be designed to reduce the possibility of transmission since there are possibilities other than suppressing vector populations . Vector incrimination traditionally involves satisfying a set of criteria analogous to Koch's postulates , summarized by Barnett [127] as follows: ( 1 ) the vector must be shown to acquire the pathogen from an identified source such as an infected vertebrate host or other reservoir , and thereafter become infected with the pathogen; ( 2 ) the vector must be shown convincingly to have close associations with infected hosts , including humans , in time and space; ( 3 ) individual vectors collected in endemic settings must repeatedly be found infected with the pathogen; and ( 4 ) efficient transmission to competent vertebrate hosts must be demonstrated experimentally , under well controlled conditions , by individual vectors , such as by bite or other means of direct contact . These criteria accommodate mechanical transmission if infection includes recovery of the pathogen from the vector's body , without making any assumptions about replication of the pathogen on or in the vector . Further , they do not preclude the possibility of parallel modes of transmission other than vectors . For example , the causative agent of plague , Yersinia pestis , has a flea vector and during sporadic outbreaks is transmitted by flea bites; but these bacteria also are transmitted during epidemics in aerosols generated by sneezing of pneumonically-infected humans or animals such as cats , which is probably the predominant mode of transmission in epidemics [128] . Similarly , human infection with the causative agent of tularemia , Franciscella tularensis , may occur through direct contact with contaminated water , by aerosols , by contact with blood or infected tissues of animals , or by bites of infected ticks , deer flies , or mosquitoes [129] , [130] . The causative agent of Rift Valley fever , a Phlebovirus in the family Bunyaviridae , is transmitted amongst infected vertebrate reservoirs ( mainly ungulates ) by mosquitoes; however , many human infections occur upon exposure to infected animal blood at the time of slaughter , by aerosolization , as well as by mosquito bites [131] . Another useful illustration is that of Chlamydia trachomatis , the causative agent of trachoma , where the transmission to human eyes has been definitively associated with contact by Musca sorbens flies ( Diptera: Muscidae ) that breed in human feces in various parts of Africa [132] . Despite this observation , other mechanisms of transmission for this disease are known , such as person-to-person contact with contaminated fingers and wash towels [133] , [134] . In two of the above examples ( plague and Rift Valley fever ) , the pathogen has a close biological relationship with , and dependency upon , insect vectors; neither pathogen could persist in nature without infecting their respective vectors . For tularemia and trachoma , vectors are not essential to pathogen persistence in nature , even though fly control in the latter case was shown to reduce incidence of disease in humans [135] . However , it is unlikely in the case of tularemia and trachoma that even highly effective fly control could eliminate human infection in endemic areas owing to other modes of transmission [133] . Therefore , using a critical approach to address the issue of insect vector incrimination for M . ulcerans , one must be cognizant of the relative biological dependency of this bacterium on an insect vector , and the potential for facultative and facilitative relationships between these bacteria and various insect “hosts” to exist which may be ancillary or even spurious to the essential and normal transmission modes . The most thorough examination of the role of an insect vector for transmission of M . ulcerans stems from investigations of aquatic , predaceous Hemiptera ( true bugs ) as reviewed above , which go far in addressing and meeting Barnett's criteria . It is important to recognize that the vast number of studies of M . ulcerans in environmental samples provide qualitative , indirect evidence of M . ulcerans based on very sensitive methods for detecting M . ulcerans DNA . Such studies revealed repeatedly that natural infection by M . ulcerans in field-collected bugs occurred , but it was tempered by detection of M . ulcerans in many other aquatic insects [18] , [67] . Thus , definitive incrimination of a single species or group of closely-related aquatic and semi-aquatic Hemiptera to the exclusion of other insects was not initially established . Other studies suggested natural contamination of the surfaces of these insects with M . ulcerans and suggested that M . ulcerans growth could occur as biofilms on the external appendages of such ‘bugs’ [101] . Thus , although aquatic and semi-aquatic Hemiptera and other insects found to harbor M . ulcerans in nature might provide habitat for the bacteria , along with numerous other living and non-living surfaces where biofilms could form [104] , this is insufficient evidence for indicating an obligatory or even facultative vectorial role to these insects . Although the experiments reported by Marsollier et al . [64] , [66] , [98]–[100] suggested modest bacterial replication in internal tissues of bugs , acquisition of bacterial infection from a live source ( infected fly maggots meant to simulate an infected prey item ) , and transmission to mice , this evidence does not establish natural infection coupled with transmission to humans . Finally , there has been no epidemiological association established between spatial and temporal distribution of contacts with aquatic Hemiptera , or bites by them , and development of Buruli ulcer in humans [68] . As reviewed above , the common understanding of the feeding habitats of aquatic and semi-aquatic Hemiptera does not include feeding on humans . More likely , infection in aquatic insects is associated with exposure to M . ulcerans in detritus and on biofilms formed on submerged materials , leading to a generalized distribution of M . ulcerans and M . ulcerans DNA in aquatic environments . In this particular scenario , despite the body of research on the topic , Barnett's criteria have not yet been fulfilled satisfactorily . The recent research by Wallace et al . [124] , whilst firmly documenting growth of M . ulcerans in mosquito larvae and transtadial infection after the molt , showed that infection did not persist upon metamorphosis to the adult stage . Thus , the link between presence of M . ulcerans in aquatic environments in which larval mosquitoes are found and adult mosquito infection with M . ulcerans , was not confirmed experimentally . However , these studies did show that M . ulcerans DNA could be detected on surface components of some adult mosquitoes . This brings up an important issue regarding experimental design and suggests that interpretation of PCR results obtained from whole insect lysates must be cautiously interpreted . These findings suggest that further research is required to confirm the association between mosquito bites , adult mosquito infection , and incidence of Buruli ulcer in humans in Australia ( reviewed above ) , where a link between mosquito feeding on infected possums and transmission of the agent via the same species of mosquitoes was proposed ( Fyfe et al . [76] ) . An analysis of blood host choice by mosquitoes , documenting blood feeding on both possums and humans in the area where human cases of Buruli ulcer are occurring , would be required as one element of satisfying Barnett's criterion #2 . At best , Barnett's criteria for vector incrimination have not been completely satisfied for a mosquito vector role , but more compelling data may be forthcoming on this matter in the future . A second approach to vector incrimination involves application of the Bradford Hill guidelines for establishing causation of infection and disease in epidemiological/ecological contexts [136] . Rather than rely upon experimental evidence , the Bradford Hill guidelines emphasize epidemiological/ecological association and use of logical inference to build up support and evidence for a strong conclusion of cause and effect , where A represents the “cause” and B the “effect” in the relationships under study [137] . The result is an “evidence hierarchy” that can be used in formal deduction [138] , and represents an interdisciplinary approach to causal investigation in disease ecology . Here , “A” would be contact between an insect vector infected with M . ulcerans , and “B” would be human infection with M . ulcerans . The guidelines are qualitative in nature and do not require the clear endpoints of Barnett's criteria , yet represent a logical approach to the problem of cause and effect under epidemiological circumstances [139] . They are as follows ( Table 2 ) : ( 1 ) Plausibility . The cause and effect association of A and B must be plausible , that is , rational and lacking in speciousness . By this is meant that the association reflects the common understanding of the normal behavior and other attributes of both A and B , bringing the appropriate factors together in such a way that abnormally implausible ( i . e . , irrational ) explanations must be discounted . In formal philosophy , plausibility must be demonstrated by sets of binary outcomes whose relationships are clearly defined propositions which can be resolved by the application of logical discourse [140] . Although plausibility can be formulated axiomatically , it cannot be analyzed statistically . It is important , therefore , not to confuse “plausible” with “probable” as the latter allows for rare and unusual circumstances and events to be explanatory under the right circumstances , whereas the former involves a rigorous , but non-probabilistic analytical process . Put more simply , plausibility addresses qualitatively how likely or unlikely it is that A results in B . A common problem in epidemiological scenarios that confronts plausibility is the issue of clusters of cases of infection ( e . g . , [134] ) , which may or may not have spatial associations with other nearby cases or with the landscape qualities near those cases [136] . In the case of Buruli ulcer and vector transmission of M . ulcerans , it is not implausible that Hemiptera and human cases are associated in time and space , but it is not plausible that there is a direct , causal relationship between the pair except in rare , accidental circumstances . Hence , there is insufficient evidence to conclude that biting hemipterans are a significant vector of M . ulcerans , although they may act as environmental reservoirs . ( 2 ) Temporality . If A results in B , then A must consistently precede B in temporal sequence . For Buruli ulcer , there is no evidence that bites of particular insects consistently precede development of patent M . ulcerans infection in humans , although there is evidence that mosquito bites are associated with increased risk [90] . The problem with this guideline is the prolonged period of time between exposure and development of symptoms in Buruli ulcer disease . However , if bites from true bugs always preceded disease , patients are likely to remember these due to the painful nature of a naucorid or belostomatid bite , in contrast to bites by mosquitoes that often go unnoticed . ( 3 ) Strength . Is the “strength” of the association great ? For example , is there a statistically significant correlation between A and B in space and or time ? The association between contact with water sources and M . ulcerans infection in humans is reasonably strong , but between insect bites and infection it is not for hemipterans , nor yet firmly established for mosquitoes in Australia and virtually non-existent for mosquitoes in Africa . ( 4 ) Biological gradient or dose-response relationship . Infection in B should increase proportionately as A increases . This principle can operate at the dose-response level , as in a toxicological series; or at the population level , as when , e . g . , more dengue virus infected mosquitoes results in more human cases of infection with that virus in space and time . The relationship may not be linear , thus confounding the interpretation of the relationship . There is no evidence that higher infection rate of M . ulcerans in aquatic insects results in higher incidence of infection in humans , although there is evidence that adult mosquitoes caught in highly endemic area in southeastern Australia are more likely to be PCR positive than those caught in areas with lower endemicity [35] . ( 5 ) Consistency . Episodes and research data where A and B show spatial and temporal associations commensurate with the other Bradford Hill guidelines must consistently reveal the association to be a positive one . Consistency could be revealed by meta-analysis of many data sets or through replicated , longitudinal studies across time and space . If scenarios emerge in which B occurs , but A does not in space and time , then doubt emerges regarding the veracity of the association . Although there are vignettes , correlations , and observations regarding insect vectors of M . ulcerans , there is no clear consistency among epidemiological scenarios to currently support the notion that insects are the predominant vector in most geographic regions . Consistent data are lacking for the ubiquitous role of vectors in the M . ulcerans transmission system . ( 6 ) Consideration of alternate explanations and analogous situations . Explanations other than causation due to A must be carefully weighed as alternatives . Causation may be inferred by analogous correspondence with other scenarios . For Buruli ulcer , a wide range of alternate explanations for transmission exists , such as human behavior linkages involving activities that increase direct skin contacts with contaminated water and inoculation with infective doses of M . ulcerans through lesions . However , as we have seen , several diseases with insect vector associations have alternative transmission modes , such as tularemia , plague , Rift Valley fever , and trachoma . Thus , it is plausible that there are multiple modes of transmission in Buruli ulcer , with certain modes more likely given specific environmental and socio-cultural contexts . ( 7 ) Experimentation . If experimental manipulations are feasible and can be structured realistically , then outcomes of the treatment regime conferred upon B ( such as exposure to the effects of A ) must reflect the association in a positive way . Often , however , Bradford Hill guidelines are utilized because experiments are either not possible , or not sufficiently rigorous or realistic . Experimental data on insect-M . ulcerans relationships have been reviewed above . There seems to be a sufficient body of work with sufficient variation in outcomes that the treatment manipulations do not lead to easily generalized conclusions on the association . Furthermore , it is often difficult to find true replication for large-scale experiments ( e . g . , treating replicate ponds with a specific chemical agent to test of changes in M . ulcerans ) , making it difficult to rigorously evaluate and experimentally test complex dynamics related to multiple modes of transmission of M . ulcerans within the environment . ( 8 ) Specificity . In this guideline , B follows A , but B does not follow when other plausible explanatory factors and events occur in temporal or spatial association . It is one of the most difficult of the guidelines to satisfy and comes closest to a strict criterion , usually because of incomplete information , multiple causes of B , random effects , and systematic errors of measurement . The review of the literature on cause and effect between insects and Buruli ulcer cases indicates a paucity of data to prove specificity . Furthermore , there are few studies relating disease incidence and insect abundance in time and space especially in Africa , and none of the alternate explanations for transmission reviewed above , such as through aerosols ( 9 ) , have been discounted . The current available data points to a multiple transmission model for Buruli ulcer , indicating that the Buruli ulcer disease system lacks specificity with regard to vector insects , with the possible exception of southeastern Australia . Therefore , more complete and rigorous qualitative assessments of data are critical to provide evidence for consistency and specificity with regard to the role of vectors and reservoirs in transmission of M . ulcerans . ( 9 ) Coherence . The association of B with A must cohere to knowledge of similar relationships in other similar associations . For M . ulcerans , insect transmission is quite unusual , as the remainder of the M . marinum group does not depend upon invertebrate vectors for transmission and infection in fish hosts . Furthermore , there is no scientific precedent for transmission of any disease agent from the direct bites of hemipteran bugs , nor is there precedent for biological transmission of any bacterial pathogen by mosquitoes known . Thus , coherence is overall not strong . However , although closely related to M . marinum , M . ulcerans is a distinct species with a genomic signature indicating it has diverged from its free-living ancestor and now occupies a specialized niche environment . Either a vertebrate gastrointestinal tract ( e . g . possums ) or insects may provide this unknown microenvironment . In summary , neither the application of Barnett's strict criteria nor the Bradford Hill guidelines support conclusively that bites by M . ulcerans-infected insects' result in human infection with M . ulcerans . However , further research will reveal if any associations might result in higher risk of infection under certain circumstances . Infection with anthrax bacteria , Bacillus anthracis , provides a useful comparison , not as a directly transferable model , but rather as a model for conceptualization of how insects , like mosquitoes , may have ancillary roles in bacterial transmission when other transmission modes also exist [141] . In that system , infection occurs in animals endemically and sporadically . When they are stressed ( as in a drought ) , they become susceptible to low dosages of bacterial spores in soil . As animals die , colonization of necrophilic flies during decomposition results in infection locally and increased bacterial sporulation and more animal cases occur as a result ( the so-called “case multipliers” effect of insects ) . As more animals become infected , an insect-mediated dispersal of bacteria occurs by biting flies such as deer flies and horse flies , whose mouthparts can become contaminated with bacteria during blood feeding ( the so-called “space multiplier” effect of insects ) . The role of flies in both modes furthers epizootics of anthrax . Although these two processes are unlikely to occur for Buruli ulcer , which appears to be mainly an endemic disease , the scenario for anthrax establishes a model by which insects might be envisioned to have ancillary roles in transmission for M . ulcerans as well . | Buruli ulcer ( BU ) is a serious necrotizing cutaneous infection caused by Mycobacterium ulcerans . It is a neglected emerging disease that has recently been reported in some countries as the second most frequent mycobacterial disease in humans after tuberculosis ( TB ) . Cases have been reported from at least 32 countries in Africa ( mainly west ) , Australia , Southeast Asia , China , Central and South America , and the Western Pacific . BU is a disease found in rural areas located near wetlands ( ponds , swamps , marshes , impoundments , backwaters ) and slow-moving rivers , especially in areas prone to human-made disturbance and flooding . Despite considerable research on this disease in recent years , the mode of transmission remains unclear , although several hypotheses have been proposed . In this article we review the current state of knowledge on the ecology and transmission of M . ulcerans in Africa and Australia , discuss traditional and non-traditional methods for investigating transmission , and suggest an intellectual framework for establishing criteria for transmission . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] | [
"ecology/plant-environment",
"interactions",
"microbiology/environmental",
"microbiology",
"ecology/evolutionary",
"ecology",
"ecology/behavioral",
"ecology",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"ecology",
"ecology/environmental",
"microbiology",
"ecology/spatial",
"and",
"landscape",
"ecology",
"infectious",
"diseases/skin",
"infections",
"infectious",
"diseases/bacterial",
"infections",
"microbiology",
"microbiology/parasitology",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"microbiology/medical",
"microbiology",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] | 2010 | Ecology and Transmission of Buruli Ulcer Disease: A Systematic Review |
Chikungunya virus ( CHIKV ) is an emerging arboviral infection with a global distribution and may cause fetal and neonatal infections after maternal CHIKV-infections during gestation . We performed a systematic review to evaluate the risk for: a ) mother-to-child transmission ( MTCT ) , b ) antepartum fetal deaths ( APFD ) , c ) symptomatic neonatal disease , and d ) neonatal deaths from maternal CHIKV-infections during gestation . We also recorded the neonatal clinical manifestations after such maternal infections ( qualitative data synthesis ) . We searched PubMed ( last search 3/2017 ) for articles , of any study design , with any of the above outcomes . We calculated the overall risk of MTCT , APFDs and risk of symptomatic neonatal disease by simple pooling . For endpoints with ≥5 events in more than one study , we also synthesized the data by random-effect-model ( REM ) meta-analysis . Among 563 identified articles , 13 articles from 8 cohorts were included in the quantitative data synthesis and 33 articles in the qualitative data synthesis . Most cohorts reported data only on symptomatic rather than on all neonatal infections . By extrapolation also of these data , the overall pooled-MTCT-risk across cohorts was at least 15 . 5% ( 206/1331 ) , ( 12 . 6% by REMs ) . The pooled APFD-risk was 1 . 7% ( 20/1203 ) ; while the risk of CHIKV-confirmed-APFDs was 0 . 3% ( 3/1203 ) . Overall , the pooled-risk of symptomatic neonatal disease was 15 . 3% ( 203/1331 ) , ( 11 . 9% by REMs ) . The pooled risk of symptomatic disease was 50 . 0% ( 23/46 ) among intrapartum vs 0% ( 0/712 ) among antepartum/peripartum maternal infections . Infected newborns , from maternal infections during gestation were either asymptomatic or presented within their first week of life , but not at birth , with fever , irritability , hyperalgesia , diffuse limb edema , rashes and occasionally sepsis-like illness and meningoencephalitis . The pooled-risk of neonatal death was 0 . 6% ( 5/832 ) among maternal infections and 2 . 8% ( 5/182 ) among neonatal infections; long-term neurodevelopmental delays occurred in 50% of symptomatic neonatal infections . Published cohorts with data on the risk to the fetus and/or newborn from maternal CHIKV-infections during gestation were sparse compared to the number of recently reported CHIKV-infection outbreaks worldwide; however perinatal infections do occur , at high rates during intrapartum period , and can be related to neonatal death and long-term disabilities .
Chikungunya virus ( CHIKV ) is a remerging arbovirus[1–5] in the family of Togaviridae , genus Alphavirus that is transmitted by the Aedes spp . mosquitos A . aegyptii and A . albopictus[6] causing a crippling musculoskeletal inflammatory disease in humans characterized by fever , polyarthralgia , myalgia , rash , and headache . [7] It was first identified in Tanzania in 1953[8] , and the name comes from a Makonde word that means “that which bends up” due to the position taken by patients suffering from the severe joint pain . [9 , 10] Since then , it has caused outbreaks in Africa[5 , 11 , 12] , Indian Ocean islands , South East Asia [13–15] , Central and South America[16–18] , US territories[19] and Europe . [20–23] CHIKV has now been identified in 94 countries worldwide . [6 , 24] CHIKV infections reemerged in India after a gap of 32 years with an estimated 1 . 38 million people been infected by the end of 2006; the outbreak subsequently declined and by 2009 there were only a few thousand cases reported yearly . [25 , 26]La Reunion Island , a French territory in the Indian Ocean , had the best-studied epidemic , and over one third of the inhabitants of the island were affected in the 2005–2006 outbreak . [27] In other outbreaks , such as in India and Malaysia much higher post-outbreak seropositivity rates were reported ( 62%-68%[28 , 29] and 56%[30] respectively ) . The first case of CHIKV infection in the Western Hemisphere was reported in 2013[18] , and it has now rapidly spread to 44 countries in the Americas , [19 , 31] including also US territories and the Caribbeans . [32 , 33] In the US , in 2014 there was the first report of local- autochthonous CHIKV transmission in Florida . [34] Among 2 , 799 CHIKV cases reported to ArboNET in 2014 from US states , 12 cases were locally-transmitted ( from Florida ) ; while all the remaining cases were from returning travelers from endemic areas . In contrast , 99% of the 4 , 710 CHIKV cases reported from US territories were locally-transmitted . CHIKV infection became a nationally notifiable condition in 2015 . [4] The number of CHIKV-infections reported in the US , declined after 2015 and in 2017 there were only 36 reported cases from the US , with no locally transmitted cases , and 30 cases from US territories where local transmission continues . [35] The almost global distribution of CHIKV as well as the possibility for autochthonous transmission in the US make CHIKV infections a threat to global health and also to domestic health in the US . According to the CDC predictive model estimates for the US based on climate data , the potential range where the Aedes aegypti and Aedes albopictus mosquitos could potentially live , survive and reproduce in the US is quite extensive . [36–38]These mosquitos are capable of transmitting other arboviral infections as well , except for CHIKV . [39 , 40] Despite the almost global distribution of CHIKV infection , data for the impact of acute CHIKV infections during pregnancy are sparse and uncertainties remain on several important clinical questions . The best studied maternal-fetal cohort for CHIKV infections during gestation is from La Reunion CHIKV outbreak in the Indian Ocean in 2005–2006[27] . Prior reviews on this specific question were not exhaustive on their searches and focused only on very few studies [41 , 42] . The maternal-fetal data from all available cohorts and for all important fetal and neonatal clinical outcomes has not been previously systematically evaluated . We set up to perform a systematic review and when meaningful , synthesize also the data by meta-analysis , to address the following questions: a ) What is the overall risk of Mother-To-Child-Transmission ( MTCT ) from maternal CHIKV infections during gestation . b ) What is the risk for antepartum fetal deaths ( APFDs ) from maternal CHIKV-infections during gestation . c ) How often maternal CHIKV-infections during gestation lead to symptomatic neonatal disease and d ) whether the reported risk-differences ( for MTCT risk and risk of symptomatic disease ) from maternal infections during the intrapartum period vs antepartum or peripartum period are consistent across diverse cohorts . Moreover , we wanted to record the spectrum of clinical manifestations reported in the scientific literature for neonatal CHIKV infections from maternal acute CHIKV-infections during gestation .
For the quantitative data synthesis , we included cohort studies , case series or case control studies that provided data on maternal CHIKV-infections during gestation and the CHIKV-infection status of their fetuses and/or newborns to allow for the calculation of the risk for: a ) mother-to-child transmission ( MTCT ) , b ) antepartum fetal deaths ( APFD ) , c ) symptomatic neonatal disease , and d ) neonatal deaths from maternal CHIKV-infections during gestation . We also calculated the overall combined fetal/neonatal disease impact of maternal CHIKV infection during gestation for a composite outcome of symptomatic neonatal disease plus the APFDs . For the qualitative data synthesis , we considered studies of any study design , including case reports , that reported clinical manifestations of neonates exposed to maternal-CHIKV-infections during gestation . Reports of postnatally acquired CHIKV-infections from mosquito exposure were excluded . For endpoints with ≥5 events in more than one study , we also synthesized data by random-effect- model meta-analyses . [43] From each eligible study for the quantitative data synthesis we extracted the following information: authors , year , locations , year of study , period of recent regional CHIKV-infection outbreak , duration of study , any possible overlap with prior published reports from the same cohort , number of pregnant women infected during gestation , number of neonatal infections from maternal infections during gestation , number of neonatal infections from intrapartum ( -2 ds prior-to-delivery to +2 ds post-delivery ) , peripartum ( -7 ds to -3 ds prior-to-delivery ) and antepartum ( >7 ds prior-to-delivery ) maternal infections , number of antepartum fetal deaths , number of CHIKV-confirmed APFDs , number of symptomatic neonatal infections , number of symptomatic neonatal infections from intrapartum , peripartum and antepartum maternal infections , number of neonatal deaths , methods for ascertainment of maternal and neonatal CHIKV-infections . For the qualitative data synthesis , we extracted information on the clinical manifestations of neonatal infections documented to have occurred from suspected or confirmed maternal infections during gestation . Data were synthesized across cohorts by simple pooling . For each outcome of interest ( MTCT-risk , APFD-risk , CHIKV-confirmed-APFD-risk , Symptomatic neonatal disease-risk; Neonatal death-risk ) we calculated -for each cohort and across all analyzed cohorts- the pooled risk ( and 95% confidence intervals thereof ) among total CHIKV maternal infections during gestation ( N of fetuses/neonates with the outcome of interest/ total N of CHIKV maternal infections during gestation ) . For the neonatal mortality outcome , we also calculated the risk of neonatal deaths among total neonatal infections . As for the majority of the analyzed cohorts , data were reported only for symptomatic neonatal infections rather than for total neonatal infections ( symptomatic plus asymptomatic ) , in our overall data synthesis for the MTCT-risk , we included also studies reporting only symptomatic-neonatal-disease-risk and considered that the MTCT-risk for those studies was at least equal to the risk for symptomatic neonatal disease . We also used random effect models ( REMs ) [43] for the calculation of the above risks to account for the between study variance . For outcomes with events <5 we used only simple pooling as REMs in such cases give unreliable results . We used the i2 test for the calculation of the between study heterogeneity . All proportion meta-analyses were done in STATASE 15 . 0 ( Stata , College Station , TX , USA ) . When there were multiple publications from the same cohorts we considered in our overall data synthesis only the report with the maximum number of events for the outcome ( s ) of interest , per total number of analyzed maternal infections reported from the whole cohort . In our systematic review , we followed the PRISMA ( Preferred Reporting Items for Systematic reviews and Meta-Analyses ) guidelines of reporting ( S1 Table ) . [44]
Of the 563 identified articles , 13 [27 , 45–56] ( from 8 cohorts ) with pertinent data were included in the quantitative data synthesis ( Fig 1 ) . These pertained to data from outbreaks in La Reunion Island ( n = 7 ) , Mayotte Island ( n = 1 ) , Sri Lanka ( n = 1 ) , Thailand ( n = 2 ) and Latin America ( n = 2 ) ( S2 Table ) . Furthermore , 33 articles [27 , 45–51 , 53–55 , 57–77] were included in the qualitative data synthesis of neonatal clinical manifestations from maternal infections during gestation . ( Fig 1 ) . In the majority of articles in the quantitative data synthesis , maternal CHIKV-infections were ascertained by maternal serology ( IgM and IgG ) and /or blood RT-PCR and/or maternal symptoms typical of CHIKV-infections . Only in the recent outbreak from Santo Domingo the diagnosis of maternal CHIKV-infections was based only on clinical criteria ( S2 Table ) . In most of the analyzed maternal/neonatal cohorts only symptomatic neonatal cases were reported among maternal CHIKV infections during gestation . By extrapolation also of these data [27 , 51 , 55] , the overall pooled-MTCT-risk across cohorts was at least 15 . 5% ( 95% CIs: 13 . 57%-17 . 53%; 206/1331 ) ( Table 1; S3 and S4 Tables ) and the risk among maternal infections during the intrapartum period was at least 50 . 0% ( 95% CIs: 34 . 90%-65 . 10%; 23/46 ) vs 0% ( 0/712 ) among antepartum/peripartum maternal infections . The timing of maternal infections was analyzed only in three cohorts [27 , 53 , 56]; 5% of all analyzed maternal infections in these three cohorts occurred during the intrapartum period . Results by REM synthesis of data were similar ( MTCT-overall risk: at least 12 . 6% [95% CIs 4 . 47%-20 . 77%]; MTCT-risk-intrapartum infections: at least 50 . 3% [3 . 75%-96 . 93%] ) . ( Fig 2 , Table 1 ) The pooled-risk of APFDs was 1 . 7% ( 95% CIs: 1 . 02%-2 . 56%; 20/1203 ) among maternal infections ( Table 1; S3 and S4 Tables ) . APFDs occurred with maternal infections in all trimester , including during early gestation . Ascertainment of the CHIKV-infection status of APFDs was very rarely performed and was confirmed in only three cases from La Reunion outbreak , after maternal infections at 12 . 5 weeks , 15 weeks and 15 . 5 weeks of gestation respectively . [57] The pooled-risk of CHIKV-confirmed APFD cases was 0 . 3% ( 95% CIs: 0 . 05%-0 . 73%; 3/1203 ) . Overall , the pooled-risk of symptomatic neonatal infections was 15 . 3% ( 95% CIs: 13 . 36%-17 . 30%; 203/1331 ) among maternal infections during gestation ( Table 1; S3 and S4 Tables ) . However , this risk was 50 . 0% ( 95% CIs: 34 . 90%-65 . 10%; 23/46 ) among intrapartum maternal infections vs 0% ( 0/758 ) among antepartum/peripartum maternal infections . Only three maternal-fetal cohorts ( La Reunion[27] , ShriLanka [53] and Colombian cohort[56] ) analyzed their data according to the timing of maternal infection and the reported risks for symptomatic neonatal disease from intrapartum paternal infections across these three cohorts were 48 . 7% ( 19/39 ) [27] , 100% ( 4/4 ) [53] and 0% ( 0/3 ) [56] respectively . The reported cases of symptomatic neonatal disease were almost exclusively from intrapartum maternal infections . The majority of the cohorts did not provide information on the percentage of pregnant women with infections during the intrapartum period . Results by REM synthesis of data were similar ( Symptomatic Neonatal disease-overall risk: 11 . 9% [95% CIs: 3 . 89%-19 . 95%]; Symptomatic Neonatal Diseases Risk-intrapartum infections: 50 . 3% [95% CIs: 3 . 75%-96 . 93%] ) ( Table 1; Fig 3 ) . Recording of long-term neurodevelopmental outcomes was very limited and was available only from La Reunion cohort , which showed neurodevelopmental delays at ~2 years of age in 50% of symptomatic neonatal infections ( 12 with CHIKV-encephalopathy and 22 with mild/moderate prostration ) ( S3 Table ) . The pooled-risk for neonatal death was 0 . 6% ( 95% CIs: 0 . 20%-1 . 40%; 5/832 ) among all maternal infections and 2 . 8% ( 95% CIs: 0 . 90%-6 . 29%; 5/182 ) among neonatal infections . ( Table 1 ) The pooled combined disease impact to the fetus and newborn ( MTCT and APFD ) was 17 . 0% ( 95% CIs: 15 . 00%-19 . 11%; 226/1331 ) among maternal infections during gestation , considering both the neonatal infections and the APFDs . ( Table 1 ) . Limited data were available on the number of premature births from maternal CHIKV infections during gestation to allow for a meaningful data synthesis; however , the reported rates for premature births were low ( 3–8% ) . [47 , 54 , 56] ( S5 Table )
In this systematic review of published data of the MTCT-risk and risk of symptomatic neonatal infection among maternal CHIKV infections during gestation , the number of identified cohorts , with pertinent data for such analyses , was very small compared to the number of recently reported CHIKV-infection outbreaks and the global distribution of CHIKV . [78] Most cohorts that reported neonatal infections had reported only symptomatic cases . By extrapolation of data also from symptomatic disease cases , the overall pooled-risk of MTCT across the 8 analyzed maternal-fetal cohorts was at least 15 . 5% . The risk of APFDs and CHIKV-confirmed APFDs was small ( <2% and <0 . 5% respectively ) . APFDs and CHIKV-confirmed-APFDs occurred from maternal infections in all trimesters , including also during early gestation . Reporting of data on APFDs was limited across the analyzed cohorts and ascertainment of CHIKV infection status of APFD was reported for only 3 cases from La Reunion outbreak . The overall MTCT risk in our study might have been underestimated as the majority of the analyzed cohorts reported only symptomatic neonatal infections rather than on all neonatal infections . In resource poor settings , where most of the CHIKV outbreaks occurred , asymptomatic neonatal infections might have remained undiagnosed , leading to a possible selection bias among the cases studied . Selective follow-up of the sickest babies may also have skewed the results of several papers . Moreover , many had significant losses to follow-up or relied on neonatal disease incidence to estimate actual neonatal infection rates . The overall pooled risk of symptomatic neonatal disease was 15 . 5% among maternal infections during gestation . However , the risk was 50 . 0% among intrapartum maternal infections vs 0% among antepartum/peripartum maternal infections . Data on the percentage of pregnant women infected during the intrapartum period were limited across the analyzed cohorts . Symptomatic neonatal disease occurred almost exclusively from intrapartum maternal infections . The pooled-risk for neonatal death was 0 . 6% among all maternal infections and 2 . 8% among neonatal infections . Long-term global neurodevelopmental delays have also been reported to occur in 50% of symptomatic neonatal infections during gestation , however this was based on a limited number of 33 such neonatal infections . [79] In our qualitative data synthesis , we generated a compilation list of clinical manifestations reported in CHIKV-infected infants from maternal infections during gestation . Such infants presented with a wide spectrum of clinical manifestations ranging from asymptomatic to severely symptomatic . Symptomatic infected newborns , from maternal infections during gestation usually developed symptoms during their first week of life , but not at the time of birth . Commonly reported symptoms included fever , polyarthralgias , diffuse limb edema , irritability , poor feeding , painful syndrome and rashes; occasionally , also sepsis-like syndrome with multiple organ involvement , meningoencephalitis with brain MRI abnormalities and can also cause long term neurodevelopmental delays and devastating neurologic outcomes such as cerebral palsy . There are anecdotal data for the use of interventions like tocolysis for the prolongation of transplacental transfer of protective maternal antibodies , for maternal infections acquired in the intrapartum period . [56] The average interval of ~6 . 3 +/- 1 . 4 days from the onset of maternal symptoms to delivery might have been enough time for the passive transfer of maternal antibodies to prevent MTCT and symptomatic disease in the newborn . [56] Tocolysis ( as long as there are no obstetric contraindications ) , has been used also in other arboviral maternal infections , such as in dengue virus maternal infections to reduce the risk of vertical transmission [56] and in maternal varicella-zoster-virus infections during the peripartum period . [80] The safety and clinical effectiveness of tocolysis as a preventive measure in such intrapartum maternal infections requires additional systematic evaluation . Moreover , the number of pregnant women infected during the intrapartum period should be reported in maternal-fetal cohorts . In La Reunion , [27] the Sri Lanka[53] and the Colombia cohort[56] only 5% of maternal infections were acquired during the intrapartum period . The role of delivery via caesarean section ( C/S ) was analyzed only in La Reunion cohort [27] and appeared to have no influence on the MTCT risk . In La Reunion cohort[27] this observation may support the notion of transplacental transmission of CHIKV-infection from the mother to the fetus , rather than from exposure in the birth canal . The C/S rate among the 61 pregnant women in this cohort with peripartum/intrapartum infections was elevated compared to the baseline rate ( 43% vs 17% ) ; the majority of those C/S was done due to fetal distress . [27] However , this was not seen in the Thailand cohort were the majority of the infants were born via vaginal delivery . [54] For the interpretation of neonatal serologic test results , pediatricians and neonataologists should be aware that the absence of positive neonatal CHIKV IgG and IgM antibodies at birth in infants born to mothers with acute CHIKV-infections in the peripartum/intrapartum period does not exclude CHIKV neonatal infection . Infected newborns from such late maternal infections may have a delayed development of CHIKV IgG and IgM antibodies , within the first 3–4 weeks of life . [81] Serial serologic monitoring of these infants might be indicated as infected infants , particularly so symptomatic infants , might be at risk for poor long term neurodevelopmental outcomes . Understanding the true impact of acute maternal CHIKV infections in the fetus and newborn requires systematic consideration also of fetal and neonatal mortality as well as ascertainment of long term neurodevelopmental outcomes in addition to the neonatal morbidity . Retrospectively extracted information about clinical signs and symptoms suggestive of acute maternal CHIKV infection during gestation likely underestimates the true incidence of maternal infections , due to recollection bias and non-capturing of mild or asymptomatic maternal infections . Moreover , standardized outcome collection and reporting across maternal-fetal cohorts is mandatory , to allow for prompt identification of the accurate risks to fetuses and newborns from maternal infections during gestation . Focus should be given during study design phase and outcome reporting for congenital/perinatal infections on all of the following: a ) estimated time of maternal infection during gestation , with accurate reporting of the number of intrapartum maternal infections; b ) consideration of APFDs in the overall combined fetal/neonatal disease impact from congenital CHIKV-infections; c ) ascertainment of CHIKV-infections-status of APFDs; d ) ascertainment of CHIKV infection status in all newborns exposed to suspected or confirmed maternal CHIKV-infections during gestation; and prompt documentation of losses to follow-up; e ) serial screening of newborns exposed to late gestation maternal infections for the first month of life , even if they are seronegative at birth , given the likely delayed neonatal IgM and IgG production after late gestation maternal infections and f ) ascertainment of long term neurodevelopmental outcomes for at least the symptomatic neonatal infections . We observed significant variation in the reported rates of MTCT and symptomatic neonatal disease across cohorts . Referral selection bias and confounding by differences in the gestational age during maternal infections across cohorts might have explained the reported differences in the risks of symptomatic neonatal disease across cohorts , as symptomatic neonatal disease occurred almost exclusively from intrapartum maternal infections . [27 , 45 , 46] We were not able to make robust conclusions on the possible role of the implicated CHIKV strain in the observed variation in the MTCT-risks and risks of symptomatic disease across cohorts , given the limited number of cases . There is preliminary evidence that the different CHIKV-strains ( Asian vs Central-East-South Africa [CESA] vs West Africa strain ) [82 , 83] might have different pathogenicity . [84] In outbreaks caused by the CESA CHIKV-strain[83] , such as the La Reunion[50] and Mayotte[51] outbreaks the overall risk of symptomatic neonatal disease among all maternal infections was 6 . 26% ( 37/591 ) and 5 . 5% ( 9/163 ) respectively . In outbreaks where the Asian CHIKV-strains were implicated , the reported rates of symptomatic disease varied even more , with 0% MTCT rates from the Thailand cohorts[52 , 54] and a small Colombian cohort[56]; versus 8% ( 4/50 ) for severely symptomatic neonatal disease from the Shri Lanka cohort , [53] 27 . 7% ( 53/191 ) from the El Salvador cohort[55] and 48% from the Santo Domingo cohort . [55] Nevertheless , there are recent data indicating that in South America CHIKV outbreaks , the African CESA CHIKV strains might also be implicated . [85 , 86] Moreover , differences across cohorts were also noted in the reported neonatal case fatality rates , with 0% for La Reunion cohort [27] versus 5 . 1% for the Santo Domingo cohort . [55] The number of neonatal CHIKV infections could be significantly underestimated using neonatal CHIKV IgG and IgM antibodies at birth . Ramful et al[49] showed that CHIKV infected infants ( even symptomatic ones ) from late-gestation maternal infections during the peripartum/intrapartum period can be seronegative at birth and might have delayed production of CHIKV antibodies; up to 3 weeks after birth for the development of IgMs and up to 4 weeks after birth for the development of positive IgGs . This is known to occur also in other congenital infections ( e . g . congenital Toxoplasmosis[87] ) when maternal infections occur very late in gestation and this might have underestimated the overall rate of mother-to-child transmission of CHIKV-infections in some of the analyzed reports . Continued monitoring of the clinical implications of CHIKV infections during pregnancy is needed as CHIKV outbreaks can reemerge in regions where the virus had already previously circulating or emerge in new regions , where it had not been previously detected . Recently in 2017 a CHIKV outbreak was noted again in Italy , in the Anzio west-coast recreational region close to Rome [88–90] , caused by an CESA strain . This strain was genetically slightly different from the strain implicated in the large 2007 outbreak in the Emilia-Romagna region in North-Eastern Italy . [23] Furthermore , the potential benefit from tocolysis for intrapartum maternal infections is an intervention that needs systematic investigation , and if confirmed in larger scale studies to be effective , routine implementation in pregnant women with intrapartum maternal infection might have important public health implications . This might provide further support for the need for prenatal screening of pregnant women with suspected CHIKV infections during the peripartum/ intrapartum period . Future validation of the diagnostic performance of point-of-care tests for the serologic diagnosis of CHIKV maternal and/or neonatal infection and/or other arboviral infections is urgently needed . Moreover , preventive measures targeting avoidance of mosquito bites in pregnant women close to the expected time of delivery , might be cost-saving and effective strategies , given the high neonatal morbidity associated with intrapartum maternal infections . Furthermore , neonatologist , need to become aware that CHIKV-infected newborns from maternal infections late in gestation would need close clinical and laboratory monitoring of their hematologic parameters during their first week of life , even if they appear asymptomatic at birth , as symptomatic neonatal infections usually develop within 3–7 days after birth . Moreover , transplacentally transferred CHIKV-IgG antibodies on average disappear by 8 months of age in uninfected neonates . [49] However , the time to neonatal seroconversion is inversely related to the time of maternal infection during gestation; with >75% of non-infected neonates being still IgG positive by 12 months of age if maternal infection was in the first trimester vs 30% and <1% if maternal infection was in the second and third trimester respectively . [49] Moreover , it may take as long as 24 months for complete neonatal seroconversion to IgG negative status among uninfected neonates . [49] Some study limitations should be acknowledged: First , in all analyzed cohorts ( even those with serologic and/or molecular confirmation of CHIKV maternal infections ) , it was the presence of maternal symptoms the first indicator that led to subsequent testing of those women for CHIKV-infections . This may have led to overestimation of the MTCT risk and the risk of symptomatic neonatal disease from maternal CHIKV infections , if symptomatic maternal infections have an incrementally higher risk of MTCT , independently of the time of maternal infection . The effect of asymptomatic CHIKV maternal infections during pregnancy remains largely unknown . The majority of the CHIKV infections were originally thought to be symptomatic [91] , nevertheless , recent reports indicate that the number of asymptomatic CHIKV infections might have been higher than it was originally thought . A report from the 2008 Thailand outbreak[92] showed that 50% of all cases were asymptomatic and a more recent report from that region showed that 87 . 5% of pregnant women infected during gestation were asymptomatic . [54] Nevertheless , conflicting results from the same region were also reported , with only 9% reported asymptomatic cases . [93] Additional surveillance studies from recent CHIKV outbreaks in Nicaragua also showed that 65% of cases were asymptomatic , with a symptomatic to asymptomatic ratio of 1:1 . 91 . [94] Recollection bias might also explain some of the observed differences in the reported rates of symptomatic CHIKV-infections . Second , in the majority of the analyzed maternal/neonatal cohorts , only symptomatic neonatal infections were reported which might have underestimated the true MTCT-risk . Third , we could not identify published cohorts with an English abstract with pertinent data for our quantitative data synthesis from the majority of countries with recent CHIKV outbreaks; such as outbreaks in African[5] [12] and Asian countries[14 , 95–97] , including large outbreaks in India[98–104] after the 2005 reemergence of CHIKV in India , outbreaks in the Carribeans [32 , 105 , 106] , Pacific Island[78] and Saudi Arabia . [107] Language-bias is also a possible reason for this phenomenon . Moreover , the published cohort studies with pertinent data from the outbreak in Central and South America since 2013 were very limited[55 , 56] compared to the scale of CHIKV transmission across more than 45 countries throughout the Americas and with >1 . 7 million suspected CHIKV-infection cases reported to the Pan American Health Organization ( PAHO ) . [78] For the majority of these outbreaks , only isolated case reports and small case series were identified , which we included in the qualitative data synthesis on the list of reported clinical manifestations from neonatal CHIKV- infections from maternal infections during gestation . Language-bias is also a possible reason for this phenomenon . It is possible that publications from several of these developing countries where such outbreaks occur are published only in grey literature [108] or in local non-English journals and indexed only in local journal databases , but not in PubMed . Moreover , the lack of financial resources and availability of accurate diagnostic tests[109] in several of the settings where such CHIKV outbreaks occur , contribute to the phenomenon of over-estimation of the true incidence and severity of the disease . In conclusion , CHIKV is an emerging arbovirus with a global distribution that can cause significant morbidity and also death in infected fetuses and newborns after maternal infections during gestation . Neonatal morbidity likely occurs predominantly from intrapartum maternal infections . Improvement is needed in the reporting of clinical important outcomes for congenitally/perinatally acquired fetal and neonatal infections . Data should be collected and reported in a standardized way across maternal-fetal cohorts for all clinically important endpoints to allow for informative meta-analyses and individual patient-level meta-analyses in this field of congenital infections . With increasing climate instability and human migration , additional CKIKV outbreaks may be expected and non-immune pregnant women in developing as well as developed countries are at risk . Additional systematic studies of the impact of the CHIKV maternal infections during gestation to the fetuses and newborns are needed . | Chikungunya virus ( CHIKV ) is an emerging arboviral infection with a global distribution and can cause infections of the fetus and newborn after maternal CHIKV-infections during gestation . In this systematic review , we evaluated the risk for mother-to-child transmission ( MTCT ) , antepartum fetal deaths ( APFD ) and symptomatic neonatal disease from maternal CHIKV-infections during gestation . Whenever meaningful , we also synthesized the data by random-effect-model ( REM ) meta-analysis . We also recorded the list of clinical manifestations of neonatal infections after maternal infections during gestation . Overall , published cohorts with pertinent data to estimate the impact to the fetuses and newborns of maternal CHIKV-infections were sparse compared to the number of recently reported CHIKV-infection outbreaks worldwide . Most cohorts reported data only on symptomatic neonatal infections rather than on all ( symptomatic and asymptomatic ) neonatal infections . By extrapolation also of these data , the pooled MTCT-risk was at least 15 . 5% ( 206/1331 ) , ( 12 . 6% by REMs ) . Symptomatic disease occurred almost exclusively with maternal infections around the time of delivery . Overall , the pooled risk of symptomatic disease was 15 . 3% ( 203/1331 ) , ( 11 . 9% by REMs ) ; however , the risk of symptomatic disease from intrapartum maternal infections was 50 . 0% ( 23/46 ) vs 0% ( 0/712 ) from antepartum/peripartum maternal infections . The pooled APFDs-risk was low ( 1 . 7% ) ; however , APFDs occurred with maternal infections in all trimesters . Infected newborns were either asymptomatic or presented during their first week of life , but not at the time of birth , with manifestations such as fever , irritability , rashes , hyperalgesia syndrome , diffuse limb edema , bullous dermatitis and occasionally also meningoencephalitis . Long-term neurodevelopmental delays occurred in 50% of symptomatic neonatal infections . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"maternal",
"health",
"obstetrics",
"and",
"gynecology",
"pathogens",
"tropical",
"diseases",
"microbiology",
"alphaviruses",
"viruses",
"developmental",
"biology",
"chikungunya",
"virus",
"rna",
"viruses",
"women's",
"health",
"pregnancy",
"neglected",
"tropical",
"diseases",
"symptomatic",
"diseases",
"pregnancy",
"complications",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"arboviral",
"infections",
"fetal",
"death",
"research",
"assessment",
"viral",
"pathogens",
"systematic",
"reviews",
"biology",
"and",
"life",
"sciences",
"neonates",
"viral",
"diseases",
"organisms"
] | 2018 | Mother-to-child transmission of Chikungunya virus: A systematic review and meta-analysis |
Deletions at chromosome 2p25 . 3 are associated with a syndrome consisting of intellectual disability and obesity . The smallest region of overlap for deletions at 2p25 . 3 contains PXDN and MYT1L . MYT1L is expressed only within the brain in humans . We hypothesized that single nucleotide variants ( SNVs ) in MYT1L would cause a phenotype resembling deletion at 2p25 . 3 . To examine this we sought MYT1L SNVs in exome sequencing data from 4 , 296 parent-child trios . Further variants were identified through a genematcher-facilitated collaboration . We report 9 patients with MYT1L SNVs ( 4 loss of function and 5 missense ) . The phenotype of SNV carriers overlapped with that of 2p25 . 3 deletion carriers . To identify the transcriptomic consequences of MYT1L loss of function we used CRISPR-Cas9 to create a knockout cell line . Gene Ontology analysis in knockout cells demonstrated altered expression of genes that regulate gene expression and that are localized to the nucleus . These differentially expressed genes were enriched for OMIM disease ontology terms “mental retardation” . To study the developmental effects of MYT1L loss of function we created a zebrafish knockdown using morpholinos . Knockdown zebrafish manifested loss of oxytocin expression in the preoptic neuroendocrine area . This study demonstrates that MYT1L variants are associated with syndromic obesity in humans . The mechanism is related to dysregulated expression of neurodevelopmental genes and altered development of the neuroendocrine hypothalamus .
Intellectual disability ( ID ) is defined by having a full-scale intelligence quota ( IQ ) of under 70 , causing difficulties with day to day functioning [1] . ID affects 2–3% of people and is a significant public health concern as it is associated with substantial morbidity and mortality [1] . Obesity is defined as a body mass index ( BMI ) of over 30 in adults or greater than the 95th centile in children ( CDC definition ) [2] . Obesity affects around 30% of adults in the United States of America and 10–20% of Europeans [2] . Obesity is associated with cardiovascular disease and certain cancers [2] . Copy number variants ( CNVs ) and single nucleotide variants ( SNVs ) are a well-recognized cause of ID [3] . 10–30% of individuals with ID will have a pathogenic CNV [3] . Exome sequencing can detect pathogenic SNVs in around 30% of people with ID without a CNV [4] . Pathogenic CNVs and SNVs are also found in obesity , usually in association with a syndromic presentation [5] . For example , both CNVs and SNVs of SIM1 are associated with obesity in humans [6 , 7] . In SIM1 deletion heterozygous mice there is impaired development of the paraventricular nucleus of the hypothalamus , with reduced melanocortin-4 receptor and oxytocin expression , in association with hyperphagic obesity [8 , 9] . Deletions at 2p25 . 3 are associated with a syndrome consisting of ID and obesity [10 , 11 , 12 , 13 , 14] . The smallest region of overlap contains the PXDN and myelin transcription factor-1 like ( MYT1L ) genes [10] . Bi-allelic SNVs in PXDN are associated with congenital cataracts [15] . No CNVs containing only PXDN have been reported in DECIPHER in association with ID . Thus PXDN is not a strong candidate gene for the phenotype associated with 2p25 . 3 deletions . SNVs in MYT1L have been reported in 2 children with ID [10] . MYT1L is a member of the myelin transcription factor family , which is defined by the presence of a unique cystein-cystein-histidine-cystein zinc finger domain [16] . MYT1L is a pro-neuronal transcription factor , and , in combination with other transcription factors can re-program fibroblasts into neurons [17] . In vitro studies indicate that MYT1L functions as a transcriptional repressor [16] . The role of MYT1L in brain development is not well understood . However , myelin transcription factor-1 ( MYT1 ) has been shown to repress transcription in neuronal progenitor cells , hence blocking Notch signaling and promoting neuronal differentiation [18] . Based upon its biological function MYT1L is a strong candidate gene for ID . Within certain deletion regions , there are single genes in which SNVs recapitulate the deletion phenotype . For example , we recently demonstrated that 2p25 . 2 deletions and SNVs in SOX11 present with Coffin-Siris syndrome [19] . Here we utilized exome-sequencing data from 4 , 296 parent-child trios in the Deciphering Developmental Disorders ( DDD ) study to demonstrate that SNVs in MYT1L are associated with a phenotype resembling that of 2p25 . 3 deletions with ID and obesity [4] . Through gene expression profiling of an MYT1L null cell line we show that MYT1L regulates a network of transcription factors involved in neurodevelopmental disorders . Knockdown of MYT1L orthologues in zebrafish resulted in altered hypothalamic oxytocin expression , providing a potential mechanism for the obesity phenotype in humans .
We identified 4 individuals with heterozygous de novo MYT1L variants through trio exome sequencing performed as part of the Deciphering Developmental Disorders study and an additional 5 individuals with heterozygous de novo MYT1L variants through a genematcher facilitated collaboration ( https://genematcher . org/ ) [20] . Table 1 and Fig 1 summarize the clinical and genetic findings . Patient 1 ( DECIPHER ID 268494 ) is a 10 year old girl with intellectual disability and autism . She was born at 31 weeks of gestation with bilateral talipes equinovarus and camptodactyly of the ring and middle fingers . Up until 5 years of age she had troublesome gastro-esophageal reflux . She first sat at 12 months old . From age 6 years she could take single steps with support . She requires a wheelchair and has never walked independently . She first said single words at 5–6 years old and at 10 years old speaks in simple sentences . She has dysarthria . She has had surgery for bilateral strabismus . Her parents reported hyperphagia and her BMI was greater than the 99th centile . Comparative genomic hybridization ( CGH ) and SNRPN 15q methylation ( for Prader-Willi syndrome ) were normal . Exome sequencing demonstrated a frameshift variant in MYT1L ( g . 1926242CA>C , p . Leu381fs ) . Patient 2 ( DECIPHER 279017 ) is a 9 year old boy with intellectual disability who attends a special needs school . Pregnancy and birth were unremarkable . He first spoke single words at 15 months . On examination he was noted to have posterior plagiocephaly , 5th finger clinodactyly and an ataxic gait . He was not dysmorphic . His parents described him as having hyperphagia and his BMI was on the 96th centile . At age 9 he wears clothes for an 11–12 year old . CGH was normal and brain magnetic resonance imaging ( MRI ) demonstrated cerebral atrophy . Exome sequencing demonstrated a missense variant in MYT1L ( g . 1915795C>T , p . Arg569Gln ) . Patient 3 ( DECIPHER ID 279061 ) is a 10 year old boy with intellectual disability , Attention Deficit Hyperactivity Disorder ( ADHD ) and verbal dyspraxia . He sat at 18 months and walked first at 2 . 5 years . He first said single words at 4–5 years old . At 10 years old he uses 2 word phrases , but mainly communicates with sign language . He wears glasses for anioshypermetropiic astigmatism . He was described as having hyperphagia and his BMI was on the 85th centile . CGH , fragile X , PWS testing and brain MRI were normal . Exome sequencing demonstrated an MYT1L missense variant ( g . 1915823G>A , p . His560Tyr ) Patient 4 ( DECIPER ID 276823 ) is a 28 year old woman with severe intellectual disability , autism , self-injurious behavior and ADHD . She was born at 42 weeks of gestation with spina bifida . She developed hydrocephalus in the first week of life . She has a ventricular shunt and has required several procedures for shunt blockage . She sat at 18 months . She smiled at 3 weeks . She said single words at 4–5 years . She requires a wheelchair because of spastic paraparesis due to spina bifida . She is not dysmorphic . Her behavior is reported to be challenging , including episodes of biting and pinching , triggered by excessive stimuli such as noise or crowding . She has limited verbal communication , and uses communication aids . She can comprehend short ( 3 key phrases ) sentences . Due to her level of disability she does not request or seek food . However , when given food at mealtimes , she was noted to have a tendency to fill her mouth excessively with food while eating . It was not possible to obtain weight or height . However , she wears UK dress size 22 clothes ( equivalent to dress size 20 in United States of America , dress size 50 in Europe and dress size 24 in Australia ) . CGH was normal . Exome sequencing demonstrated an MYT1L splice donor variant ( g . 1915791C>T ) . Patient 5 is a 17 year old woman with intellectual disability and autism . She was born at 41 weeks gestation with no birth complications . Gross motor , fine motor and speech delay was noted at 2 and 4 years old . She is noted to have dyslexia . She was not dysmorphic . She had complex partial seizures from the age of 11 years old . She was reported to have hyperphagia and BMI was greater than the 99th centile . CGH and brain MRI were normal . Whole genome sequencing demonstrated an MYT1L missense variant ( g . 1921036A>G , p . Leu520Pro ) . Patient 6 is a 15 year old girl with intellectual disability who attends a special needs school . She first walked at 22 months old and had delayed speech . She was not dysmorphic . She did not have hyperphagia and her BMI was on the 74th centile . CGH was normal . Exome sequencing demonstrated an MYT1L frameshift variant ( g . 1895856_1895865del , p . Thr741fs ) . Patient 7 is a 5 year old girl with intellectual disability who required additional help at school . She was born at term with no birth complications . She first sat at 9 months . She walked first at 19 months . She required physiotherapy . She had speech delay , speaking single words after 2 years of age . At 5 years old she had ongoing speech and language delay and was reported to be clumsy . She was not dysmorphic . Eye examination demonstrated hyperopia and strabismus . She had hyperphagia and her BMI was greater than the 99th centile . CGH demonstrated a 20p12 duplication , inherited from her phenotypically normal father . Exome sequencing demonstrated an MYT1L nonsense variant ( g . 1890345G>A , p . Arg839Ter ) . Patient 8 is a 3 year old girl with global developmental delay and autistic behaviour . She was born at term with no pregnancy or birth complications . She had global hypotonia during the first few months of life . She had global developmental delay . She first sat at 9 months of age , walked independently at 2 years of age . At age 3 she had not developed speech , but used sign language . At age 3 she was beginning to learn to run . She could draw a line and a circle . She had absence seizures . Her BMI was on the 75th centile and hyperphagia was not present . MRI brain demonstrated thinning of the corpus callosum . Exome sequencing demonstrated an MYT1L missense variant ( g . 1921025G>T , p . His524Asn ) . Patient 9 is a 13 year old boy with global developmental delay and ADHD . Pregnancy and birth were unremarkable . Hypotonia and poor sleep were noted in the neonatal period . He first sat at 12 months of age and walked at 22 months . He spoke single words at 29 months of age . At 5 years old he knew 20 words , but had pronunciation difficulties . At 7 years old he knew 50 words and was using 2-word sentences . At the age of 13 he was unable to read , write or count . He attends a special educational needs school . His BMI was on the 68th centile . A next generation sequencing gene panel test demonstrated an MYT1L missense variant ( c . 1579G>A , p . Gly527Arg ) . Protein truncating variants ( PTV ) in developmentally crucial genes should occur less frequently than predicted in individuals without developmental disorders . The expected frequency of PTV in human genes has been reported in the ExAC database based upon parameters such as mutation rates for given nucleotide bases [21] . We identified only a single loss of function variant in MYT1L in the ExAC database ( accessed March 2017 ) [21] . This is compared to an expected number of 33 loss of function variants , giving a probability of loss of function intolerance score of 1 . 0 ( a probability of loss of function intolerance score>0 . 9 indicates intolerance to loss of function ) . In addition , 205 missense variants were reported , compared to an expected 402 . 5 missense variants ( Z score = 4 . 81 , indicating constraint on variation ) . The Provean ( median -5 . 2 [interquartile range -6 . 0 - -3 . 85] vs -1 . 1 [interquartile range -2 . 59 - -0 . 43] , Mann-Whitney U-test , u = 344 , z = -2 . 96 , p = 0 . 012 ) and SIFT ( median 0 . 001 [interquartile range 0 . 005–0 . 28] vs 0 . 068 [interquartile range 0 . 0003–0 . 003] , Mann-Whitney U-test , u = 569 , z = -2 . 61 , p = 0 . 025 ) scores for the missense variants in our patients were significantly higher than the scores for ExAC missense variants ( Fig 2A ) . This indicates substantial constraint on both PTV and missense variants . This supports a pathogenic role for PTV and missense variants in the reported phenotype . In addition , an in silico model of the structural effects of the MYT1L missense variants indicated that they were likely to interfere with DNA binding ( Fig 2B ) . To examine the hypothesis that haploinsufficiency for MYT1L drives the 2p25 . 3 deletion syndrome , we compared the phenotype of 2p25 . 3 deletion carriers with those of MYT1L single nucleotide variant ( SNV ) carriers . The phenotypes associated with deletion of 2p25 . 3 were defined by a literature review [10 , 11 , 12 , 13 , and 14] and we report an unpublished case ( patient 10 in Table 1 ) . The smallest region of overlap contains PXDN and MYT1L . No PXDN SNVs were identified in the DDD exome dataset . Using Fischer’s exact test there was no significant difference between the proportions of deletion or SNV patients with the following phenotypes: intellectual disability , gross motor delay , speech delay , autism , overweight/obese or hyperphagia . This supports our hypothesis that MYT1L haploinsufficiency is central to the 2p25 . 3 deletion phenotype . Given the phenotype of intellectual disability and predisposition to overweight/obesity we reasoned that MYT1L should be expressed in relevant neuroanatomical structures . We first confirmed that MYT1L expression is confined to the brain and pituitary in humans using the GTEx Portal ( accessed March 2017 ) [22] . We then utilized the Allen Brain atlas to examine the spatial expression pattern of MYT1L in human brain [23] . In keeping with a role in cognition/intellectual disability , MYT1L is expressed at significantly higher levels in the adult cerebral cortex than in the hippocampus , basal ganglia and hypothalamus ( Mann-Whitney U-test , P<0 . 001 ) ( Fig 3A ) . We could not demonstrate significant expression of MYT1L in hypothalamic structures relevant to appetite and obesity in the adult brain . We then hypothesized that MYT1L might be expressed in hypothalamic structures relevant to appetite and obesity during brain development . Data from the prenatal LMD microarray from the Brainspan atlas of the developing human brain demonstrated that MYT1L was expressed in multiple hypothalamic structures at 15–16 post conception weeks ( pcw ) , with significant reduction in expression at 21 pcw ( Mann-Whitney U-test , p<0 . 001 ) ( Fig 3B ) . This suggests that MYT1L principally plays a role in hypothalamic development rather than postnatal hypothalamic function . To study the transcriptional consequence of loss of MYT1L function we created a Human Embryonic Kidney ( HEK ) null cell line using CRISPR-Cas9 . Sanger sequencing of genomic DNA confirmed creation of homozygous premature stop codons in MYT1L . Gene expression profiling using the Clariom S array identified 955 differentially expressed genes ( 2-fold expression change , false discovery rate 2% ) . Enrichment analysis ( using Enrichr http://amp . pharm . mssm . edu/Enrichr/ ) [24]demonstrated that the differentially expressed gene set was enriched for the Gene Ontology Biological Process 2015 term gene expression ( GO:0010467 , adjusted p-value 0 . 00077 , Z-score -2 . 34 , combined score 16 . 77 ) ( Fig 4A ) and Gene Ontology Cellular Component 2015 terms nucleolus ( GO: 0005730 , adjusted p-value 0 . 0023 , Z-score -2 . 21 , combined score 13 . 43 ) and nucleoplasm ( GO: 0005654 , adjusted p-value 0 . 005853 , Z-score -2 . 08 , combined score 11 . 4 ) ( Fig 4B ) . The gene set was also enriched for the Reactome 2016 pathways Gene Expression_Homo Sapiens_R-HAS-74160 ( adjusted p-value 2 . 2 x 10–7 , Z-score -2 . 16 , combined score 33 ) and Generic Transcription Pathway_Homo Sapiens_R-HAS-212436 ( adjusted p-value 0 . 01586 , Z-score -2 . 26 , combined score 9 . 35 ) ( Fig 4C ) . The differentially expressed genes were also enriched for genes from the OMIM disease ontology term mental retardation ( p-value 0 . 045 , adjusted p-value 0 . 38 , Z-score -1 . 32 , combined score 1 . 27 ) ( Fig 4D ) . This suggests that MYT1L regulates a network of genes that control transcription , and which have themselves been implicated in the etiology of neurodevelopmental disorders . Given the obesity phenotype in patients with MYT1L SNVs we hypothesized that loss of MYT1L function may interfere with development of the neuroendocrine hypothalamus . We sought to explore this by creating a zebrafish knockdown model . MYT1L has two orthologs in zebrafish: myt1la and myt1lb . By using whole mount in situ hybridization ( WISH ) we demonstrate that both orthologs are expressed diffusely within the zebrafish central nervous system , including the hypothalamus , and not within any extra-neuronal tissues ( Fig 5A ) . This resembles the expression pattern in humans . ANRT2 encodes a dimerization partner required for SIM1 function in hypothalamic development . To examine the role of SIM1-ARNT2 in regulating myt1la/mytl1b we performed WISH for myt1la in the previously described homozygous arnt2hi2639Tg null zebrafish [25] . In the mutant zebrafish myt1la was undetectable in the hypothalamus ( Fig 5B ) . This shows a role for SIM1-ARNT2 dimers in regulating myt1la expression in the hypothalamus . These experiments demonstrate that MYT1L lies downstream of SIM1-ARNT2 in the leptin-melanocortin-SIM1 pathway , and , in turn , MYT1L regulates OXT expression in the hypothalamus . Antisense morpholinos ( MO ) targeting myt1la or myt1lb , alone or in combination , were injected into zebrafish embryos at the 1- or 2-cell stage . The amount of MO injected was kept constant . WISH using probes against oxytocin ( OXT ) or arginine vasopression ( AVP ) was performed . MO targeting myt1la or myt1lb resulted in an almost complete loss of OXT in the neuroendocrine pre-optic area ( Fig 5C and 5D ) . MO targeting myt1la or myt1lb also resulted in an almost complete loss of AVP in the neuroendocrine pre-optic area , but not in the ventral hypothalamus ( Fig 5E and 5F ) . This suggests that myt1la and myt1lb may play a specific role in regulating the development of the neuroendocrine hypothalamus .
Here we describe 9 individuals with de novo SNVs in MYT1L . These individuals shared several phenotypic features . All had global developmental delay or intellectual disability . Gross motor delay was present in all , and patient 1 had not walked independently by the age of 10 years old . Patient 4 also required a wheelchair , but this was related to spastic paraparesis secondary to spina bifida . No other individuals with CNVs or SNVs of MYT1L have been reported with a neural tube defect . The spina bifida in patient 4 is likely etiologically unrelated to the MYT1L SNV . Six of the MYT1L SNV carriers were overweight or obese based upon BMI centiles and patient 4 requires a dress size 22 , implying obesity . Five of the patients had hyperphagia , but patient 4 did not have sufficient speech development or motor function to ask for or take food . Three had a diagnosis of an autism spectrum disorder . There was no shared , distinctive facial dysmorphology . Two MYT1L SNV carriers have been reported by De Rocker et al: patient 14 had developmental delay , autism and was described as obese ( BMI not reported ) , patient 15 had developmental delay , autism and BMI >97th centile [10] . An autistic male with developmental delay and febrile seizures in association with an MYT1L nonsense mutation was reported by Wang et al [26] , however BMI was not described in this paper . There was no significant difference in the frequency of phenotypic features between 2p25 . 3 deletion and MYT1L SNV carriers . This suggests that haploinsufficiency for MTY1L drives the 2p25 . 3 deletion phenotype . Our report confirms that CNVs and SNVs of MYT1L are associated with a syndromic presentation consisting of developmental delay/ID , hyperphagia and obesity . The clinical presentation of MYT1L CNV and SNV carriers overlaps with other mendelian causes of obesity . Prader-Willi syndrome ( PWS ) is a well-recognized cause of hyperphagic obesity [27] . The presence of characteristic dysmorphology and hypogenitalism may help differentiate PWS from MYT1L SNV/CNV carriers [27] . Deletions and SNVs of SIM1 are associated with hyperphagic obesity [6 , 7] , with developmental delay associated with 6q14 . 1 deletions [7] . Bi-allelic SNVs of leptin ( LEP ) [28] , leptin receptor ( LEPR ) [29] , and pro-opiomelanocortin ( POMC ) [30] are also associated with hyperphagic obesity . LEP and LEPR can be distinguished from MYT1L due to the association of hypogonadotrophic hypogonadism with LEP and LEPR SNVs . Patients with POMC deficiency present with a range of endocrine problems not reported in association with MYT1L variants . There is also phenotypic overlap between certain microdeletion syndromes and MYT1L . Smith-Magenis syndrome ( 17p11 . 2 deletion ) is associated with developmental delay and variable obesity [31] . Facial dysmorphology , sleep disturbance and self-injurious behavior reported in Smith-Magenis syndrome were not identified in our MYT1L cohort . 22q11 . 2 deletion carriers have increased rates of obesity [32] , and the presence of cleft lip/palate , congenital heart disease or parathyroid disease can permit distinction from MYT1L variant carriers . Our data supports the hypothesis that MYT1L SNVs cause loss of protein function and haploinsufficiency . Data from the ExAC indicated a loss of function intolerance score of one , which indicates that MYT1L is a haploinsufficient gene that will not tolerate heterozygous loss of function variants . Four of the SNVs we report were predicted to be PTV , which would result in loss of protein function . In silico modelling indicates that the 5 missense variants we report would be predicted to interfere with the binding of MYT1L to DNA . Both p . His560Tyr and p . Arg569Gln affect conserved amino acids that directly bind to DNA . These variants are likely to disrupt DNA binding . The p . Leu520Pro variant lies at a protein loop which is crucial for the correct folding of the second zinc finger domain; this missense variant is likely to disrupt protein structure and hence DNA binding . His524 and Gly527 are zinc ligands , and any change will disrupt protein structure . The fact that the phenotypes of 2p25 . 3 deletions and MYT1L SNVs overlap supports haploinsufficient loss of function of MYT1L as the disease mechanism . By GEP in an MYT1L HEK cell line with homozygous MYT1L frameshift variants , we demonstrate altered expression of multiple genes implicated in regulation of gene expression and transcription . Haploinsufficiency for MYT1L has clear potential to disrupt expression of critical genes during brain development and hence cause a neurodevelopmental disorder . The expression pattern of MYT1L in the human brain reflects the clinical features of individuals with 2p25 . 3 deletions and MYT1L SNVs . The widespread expression of MYT1L in brain structures relevant to cognition supports a role for loss of function in the etiology of ID . This is in keeping with the fact that the overwhelming majority of ID and autism genes have widespread expression in the cerebral cortex [33] . The expression pattern of MYT1L in human brain also supports a role for the gene in appetite/obesity . MYT1L was expressed in multiple hypothalamic nuclei at 15–16 pcw , with significant reduction in expression at 21 pcw . This leads us to hypothesise that MYT1L may play a role in the development of the hypothalamus , and that MYT1L loss of function may be associated with obesity by impairing development of hypothalamic nuclei . Similar mechanisms operate for other obesity genes such as SIM1 [8 , 9] . To investigate a role for myt1la/b in development of the neuroendocrine hypothalamus we generated a knockdown model in zebrafish . Injection of MO against myt1la or myt1lb , alone or in combination , resulted in a severe loss of expression of OXT in the neuroendocrine preoptic area . Knockdown of myt1la/b resulted in loss of AVP expression in the neuroendocrine preoptic area but not the ventral hypothalamus . This suggests that myt1la/b may influence the development of the neuroendocrine preoptic area , but not other regions of the hypothalamus . The neuroendocrine preoptic area is the functional equivalent of the paraventricular nucleus in humans [34]; lesions of which cause hyperphagic obesity [7 , 8] . Thus , MYT1L CNVs and SNVs may lead to hyperphagic obesity by impairing hypothalamic development . SIM1 , functioning with its dimerization partner ARNT2 , regulates the development of the paraventricular nucleus [7 , 8] . SIM1 deletion heterozygous mice have hypocellularity of the paraventricular nucleus and hyperphagic obesity [7 , 8] . To examine an interaction of SIM1 with MYT1L we performed WISH for myt1la in the homozygous arnt2hi2639Tgnull zebrafish , which has no functional arnt2 and hence disruption of sim1a/b function . The absence of myt1la expression in the neuroendocrine preoptic area demonstrates that MYT1L lies downstream of SIM1-ARNT2 in hypothalamic development . Our experiments also indicate that OXT is downstream of MYT1L . This suggests that loss of OXT may be a final common pathway in genetic forms of obesity , and represent a treatment target in multiple disorders . In summary , we have identified a series of individuals with MYT1L de novo SNVs who present with a syndrome of ID and obesity . Genes involved in nucleosome remodeling , especially those of the neuron-specific Brg1/hBrm Associated Factor ( nBAF ) complex , have emerged as being central to the pathogenesis of ID [35] . However , MYT1L is not known to play a role in nucleosome remodeling , and GEP did not demonstrate that genetic pathways involved in nucleosome remodeling are dysregulated in MYT1L knockdown cells . This suggests that the ID observed in patients with MYT1L SNVs and CNVs is not related to altered nucleosome remodeling . The mechanism by which MYT1L loss of function results in ID is unclear . Murine studies of MYT1 demonstrate that it promotes neuronal differentiation of neuronal progenitor cells by inhibiting Notch signaling [18] . It seems reasonable to suggest that MYT1L may perform a similar function in the developing brain and that loss of MYT1L function will disrupt this process . The obesity phenotype with MYT1L loss of function is associated with disrupted development of the neuroendocrine hypothalamus in zebrafish , manifested by loss of OXT . This is similar to the effects of SIM1 [8 , 9] and POU3F2 loss of function [24] , both of which are associated with hyperphagic obesity . OXT is emerging as a key neurochemical in both autism and obesity pathogenesis . Polymorphisms in OXT and its receptor are associated with autism risk , and intranasal OXT improves autism symptoms and imaging abnormalities [36 , 37] . OXT treatment reduces food intake in humans and in sim1 mutant mice [38 , 39] . In conclusion , we identify MYT1Lmutations as a cause of syndromic obesity , and establish MYT1L as a member of the leptin-melanocortin-SIM1 pathway , with downstream loss of OXT associated with MYT1L mutations a potential therapeutic target .
For probands and their parents in the DDD study , saliva samples were collected ( Oragene DNA collection kits , DNA Genotek , Kanata , ON , Canada ) and DNA extracted ( QIAsymphony , Qiagen , Venlo , Netherlands ) . Exome sequencing was performed at the Wellcome Trust Sanger Institute with Agilent SureSelect 55MB Exome Plus with Illumina HiSeq to investigate single nucleotide variants ( SNVs ) and small insertion-deletions ( indels ) in coding regions of the genome . An automated variant pipleline was used as previously described [4] . Probands were identified with protein altering SNVs in MYT1L . Blood samples were sent for sequencing at the HudsonAlpha Genomic Services Laboratory ( http://gsl . hudsonalpha . org ) . Genomic DNA was isolated from peripheral blood and WGS was conducted to a mean depth of 35X , with >80% of bases covered at 20X . WGS was done on Illumina HiSeq Xs . Reads were aligned and variants called according to standard protocols [40 , 41] . A robust relationship inference algorithm ( KING ) was used to confirm familial relationships [42] . Using filters related to call quality , allele frequency , and impact predictions , we searched for rare , damaging de novo variation , or inherited X-linked , recessive , or compound heterozygous variation in affected probands . WGS were carried out under a research protocol and were not completed within a CAP/CLIA laboratory . All variants found to be medically relevant and returnable were validated by Sanger sequencing in an independent CLIA laboratory ( Emory Genetics Laboratory ) before being returned to participants , although these validated variant results are not CLIA-compliant as the input DNA was originally isolated in a research laboratory . Ethics approval and consent to participate: Review boards at Western ( 20130675 ) and the University of Alabama at Birmingham ( X130201001 ) approved and monitored this study . Consent for publication: A parent or legal guardian was required to give consent to participate in the study and inclusion of their data for publication , and assent was obtained from those children who were capable . The study has UK Research Ethics Committee approval ( 10/H0305/83 , granted by the Cambridge South REC , and GEN/284/12 granted by the Republic of Ireland REC ) . Written consent taken from all participants and declaration of Helsinki followed . The predicted effect of the MYT1L missense variants was examined using SIFT and PolyPhen . Evolutionary conservation of mutated amino acids was assessed by aligning orthologs in Ensembl ( http://www . ensembl . org/index . html ) . The presence of MYT1L variants in control populations and MYT1L constraint metrics were queried using the ExAC browser ( http://exac . broadinstitute . org/gene/ENSG00000186487 ) . SNVs are reported using MYT1L Isoform-1 ( canonical sequence , 1 186 amino acids ) and ensemble transcript ENST00000428368 ( http://www . ensembl . org/Homo_sapiens/Transcript/Summary ? g=ENSG00000186487;r=2:1791242-2331116;t=ENST00000428368 ) . To investigate the structural effects of the MYT1L missense variants , we generated a 3-dimensional model of the 2nd and 3rd zinc finger domains of MYT1L bound to DNA . This was based upon the structure of MYT1 4th and 5th zinc finger domains . The amino acid sequence of MYT1 was first extracted from the report of Gamsjaeger et al [43] . The 4th and 5th zinc finger domains of MYT1 and 2nd and 3rd zinc finger domains of MYT1L aligned well , indicating that the structure of MYT1 is suitable to model the effects of MYT1L missense variants . A 3-dimensional model of the 4th and 5th zinc finger domains of MYT1 bound to DNA was then generated using Pymol ( https://www . pymol . org/ ) . The MYT1L missense variants were placed at the appropriate residues of this to visualize the structural and hence potential functional , consequences . Human embryonic kidney ( HEK ) cell line HEK-293 ( HD PAR-617 , ATCC CRL-1573 ) was confirmed to be triploid at the MYT1L locus using SNP 6 . 0 arrays . MYT1L transcript ENST00000399161 was targeted . A guide RNA ( gRNA1240 ) was designed to bind at exon 9 of MYT1L , with an adjacent protospacer adjacent motif ( PAM ) site . HEK-293 cultures were transfected with gRNA1240 . Colonies were then genotyped by PCR to identify those with homozygous out of frame variants at the gRNA1240 site . PCR sequencing of genomic DNA demonstrated a 1 base pair insertion of a thymidine base within the gRNA1240 site at allele 1 and 2 . PCR sequencing demonstrated a 10 base pair deletion in allele 3 . Long range PCR did not reveal any larger deletions across the gRNA site . These sequence variants are predicted to cause a downstream STOP codon at amino acid 115 and 170 , respectively . RNA was extracted using trizol and a standard column based system from the isogenic parental HEK-293 line and the knockout line , each in biological triplicate . Whole genome gene expression profiling was performed using the Clariom S array ( affymetrix ) . Differentially expressed genes were defined as those showing a 2 fold or greater change in expression with a false discovery rate of 2% using the Affymetrix Transcriptome analysis console . Enrichment analysis was performed using Enrichr ( http://amp . pharm . mssm . edu/Enrichr/ ) with a crisp gene set ( i . e . no fold change expression assigned to each gene ) [24] . The expression pattern of MYT1L in adult brain was examined using the Allen brain atlas ( microarray data using Agilent 8x60K cDNA chip ) [23] . Expression data ( Z score normalized ) were downloaded for 6 adult brains ( donor id 9861 , 10021 , 12876 , 14380 , 15496 and 15697 ) for the frontal cortex , hippocampus , basal ganglia and hypothalamus ( at least 3 regions from each anatomical site ) . Expression levels were compared between each anatomical site using Mann-Whitney U-tests . The expression pattern of MYT1L in developing human brain was examined using the brainspan atlas . Microarray data ( Z score normalized ) was downloaded from 2 donor brains at 15–16 post conception weeks ( pcw ) ( donor id 12840 and 14751 ) and 2 donor brains at 21 pcw ( donor id 12566 and 12690 ) for the paraventricular nucleus , anterior hypothalamic nucleus , lateral hypothalamic area , dorsomedial hypothalamic area , venteromedial hypothalamic area , posterior hypothalamic nucleus and medial mammillary nucleus . Mann-Whitney U-tests were used to compare median MYT1L expression in each hypothalamic nucleus at 15–16 pcw and 21 pcw . Zebrafish ( Danio rerio ) were raised , maintained and crossed as described [43]] . Development of embryos was at 28°C , and staging was determined by both hours post fertilization ( hpf ) and morphological characteristics [44] . Embryos were genotyped for the arnt2hi2639cTg allele as previously described [25] . All procedures were in accordance with NIH guidelines on the care and use of animals and were approved by the Georgetown University Institutional Animal Care and Use Committee , Protocol 11–008 . Whole-Mount In Situ Hybridization ( WISH ) using DIG-labeled riboprobes was performed as previously described [44] . Zebrafish myt1la and myt1lb templates were generated by PCR amplification from 2 dpf zebrafish cDNA and then cloned into pJET1 . 2 vectors in “backwards” direction . The myt1a exon 5 primers were 1aF:CACCACGACAATTATTCTAGTG , and 1aR:CTTTAGGGTAGTAAGCTC . The myt1b exon four primers were 1bF: AGAGTGACCATATGAATTGCA , and 1bR: CTGCTGCTGGTTATTGTTGAG . The plasmids were linireized with XbaI and probe was synthesized using T7 RNA polymerase and reagents from the DIG labeling kit ( Roche ) . MO injections for sim1 were used as previously published [45] . For the myt1la and myt1lb genes , two MOs for each gene were used to knockdown these proteins . One MO was designed to block translation and the other was designed to block splicing . The following MOs were synthesized by Gene-Tools , LCC: myt1la ATG MO , 5’-ACCTCCATCTGAATGCAGTGGTTGA; myt1la Splice MO , 5’-GGACAGCTGGAGACAAGAGAAATAA; myt1lb ATG MO , 5’-CATCTGCTACATCCACCTCCATCTG; myt1lb Splice MO , 5’-ATATTTGTGCCCTCACCTATTTCAT; and tp53 MO , 5’- GCGCCATTGCTTTGCAAGAATTG . The Standard Control MO from Gene Tools was used as control . Solutions consisting of 4 ng/nl MO plus 0 . 5% tetramethyl rhodamine dextran in dH20 were microinjected into one to four cell stage embryos . Images were acquired using a Zeiss Axioplan2 microscope fitted with an AxioCam camera using AxioVision software , or , with a Zeiss stereoscope fitted with a Canon Oneshot digital camera . Digitized images were imported into PhotoShop CS ( Adobe Systems Inc , San Jose , CA ) , contrast and brightness adjusted as necessary . WISH for myt1la/b expression was quantified using an ordinal scale: 0 = no staining , 1 = dramatically reduced staining , 2 = normal staining . WISH for oxytocin was quantified as follows: ~30 cells = wild type expression , 5–15 cells = reduced , 1–4 = highly reduced , 0 = no expression . Statistical differences in green fluorescent protein ( GFP ) expression in 2 somite embryos was determined using ANOVA , followed by Tukey post-hoc tests for individual groups . Significance of MO induced phenotype categories was evaluated by Ordinal Logistic Regression . The statistical analyses utilized SPSS ( version 22 ) from IBM . Decipher ( https://decipher . sanger . ac . uk/ ) Exome aggregation consortium ( http://exac . broadinstitute . org/ ) Allen Brain Atlas ( http://www . brain-map . org/ ) SIFT ( http://sift . jcvi . org/ ) Provean ( http://provean . jcvi . org/index . php ) Ensembl ( http://www . ensembl . org/index . html ) Enrichr ( http://amp . pharm . mssm . edu/Enrichr/ ) Pymol ( https://www . pymol . org/ ) GTEx portal ( https://www . gtexportal . org/home/ ) | Intellectual disability is defined by having an intelligence quotient of less than 70 points , and it affects about 2–3 people in every 100 . Obesity is defined as having a body mass index of over 30 in adults or over the 95th centile in children . Both of these conditions are major public health concerns in Western countries . Genetic studies have shown that small missing pieces of chromosome ( deletions , which remove many genes ) and changes to the lettering of genes ( which stop the gene from working , mutations ) can cause intellectual disability or obesity . Here we identified 9 children with intellectual disability and obesity who have mutations in a gene called MYT1L . This gene is thought to give an important instruction for brain development . To find out what the effect of loss of MYT1L is on brain development we reduced the levels of MYT1L ( using a special chemical ) in an experimental zebrafish . This showed that loss of MYT1L in zebrafish causes a problem with the development of the hypothalamus , which may explain how MYT1L mutations cause obesity in humans . In the zebrafish there was also reduction of a brain hormone called oxytocin which is involved in thought processes , which may explain why MYT1L mutations cause intellectual disability . We have identified a new genetic condition caused by MYT1L mutations , further study of this gene will help us understand , and treat , intellectual disability and obesity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"body",
"weight",
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"pervasive",
"developmental",
"disorders",
"brain",
"vertebrates",
"autism",
"social",
"sciences",
"developmental",
"psychology",
"animals",
"neuroscience",
"dna-binding",
"proteins",
"hormones",
"animal",
"models",
"osteichthyes",
"model",
"organisms",
"physiological",
"parameters",
"obesity",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"autism",
"spectrum",
"disorder",
"research",
"and",
"analysis",
"methods",
"developmental",
"neuroscience",
"genomics",
"neurodevelopmental",
"disorders",
"fishes",
"proteins",
"neurochemicals",
"gene",
"expression",
"hypothalamus",
"gene",
"ontologies",
"peptide",
"hormones",
"biochemistry",
"zebrafish",
"psychology",
"anatomy",
"oxytocin",
"physiology",
"neurology",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"organisms"
] | 2017 | MYT1L mutations cause intellectual disability and variable obesity by dysregulating gene expression and development of the neuroendocrine hypothalamus |
The small size of RNA virus genomes ( 2-to-32 kb ) has been attributed to high mutation rates during replication , which is thought to lack proof-reading . This paradigm is being revisited owing to the discovery of a 3′-to-5′ exoribonuclease ( ExoN ) in nidoviruses , a monophyletic group of positive-stranded RNA viruses with a conserved genome architecture . ExoN , a homolog of canonical DNA proof-reading enzymes , is exclusively encoded by nidoviruses with genomes larger than 20 kb . All other known non-segmented RNA viruses have smaller genomes . Here we use evolutionary analyses to show that the two- to three-fold expansion of the nidovirus genome was accompanied by a large number of replacements in conserved proteins at a scale comparable to that in the Tree of Life . To unravel common evolutionary patterns in such genetically diverse viruses , we established the relation between genomic regions in nidoviruses in a sequence alignment-free manner . We exploited the conservation of the genome architecture to partition each genome into five non-overlapping regions: 5′ untranslated region ( UTR ) , open reading frame ( ORF ) 1a , ORF1b , 3′ORFs ( encompassing the 3′-proximal ORFs ) , and 3′ UTR . Each region was analyzed for its contribution to genome size change under different models . The non-linear model statistically outperformed the linear one and captured >92% of data variation . Accordingly , nidovirus genomes were concluded to have reached different points on an expansion trajectory dominated by consecutive increases of ORF1b , ORF1a , and 3′ORFs . Our findings indicate a unidirectional hierarchical relation between these genome regions , which are distinguished by their expression mechanism . In contrast , these regions cooperate bi-directionally on a functional level in the virus life cycle , in which they predominantly control genome replication , genome expression , and virus dissemination , respectively . Collectively , our findings suggest that genome architecture and the associated region-specific division of labor leave a footprint on genome expansion and may limit RNA genome size .
Genome size is the net result of evolution driven by the environment , mutation , and the genetics of a given organism [1] , [2] . Particularly mutation rate is a powerful evolutionary factor [3] . The relation between mutation rate and genome size is inversely proportional for a range of life forms from viroids to viruses to bacteria , and it is positive for eukaryotes , suggestive of a causative link [4]–[6] . The genome size of RNA viruses is restricted to a range of ∼2 to 32 kb that corresponds to a very narrow band on the genome size scale ( ranging from 1 kb to 10 Mb ) across which genome size increase is correlated with mutation rate decrease [7] . This restricted genome size range of RNA viruses was believed to be a consequence of the universal lack of proof-reading factors , resulting in a low fidelity of RNA replication [8] , [9] . In the above relation , mutation rate and proof-reading serve as a proxy for replication fidelity and genetic complexity , respectively . Replication fidelity , genome size , and genetic complexity were postulated to lock each other , through a triangular relation [10] , in a low state in primitive self-replicating molecules [11] . This trapping , known as the “Eigen paradox” [12] , was extended to include RNA viruses [13] , providing a conceptual rationale for the small range of genome sizes in these viruses . Recent studies of the order Nidovirales , a large group of RNA viruses that includes those with the largest genomes known to date , provided strong support for the postulated triangular relation [10] , [14] . Unexpectedly , they also revealed how nidoviruses may have solved the Eigen paradox by acquiring a proof-reading enzyme . These advancements implied that the control of genome size may be more complex than previously thought , in RNA viruses in general , and particularly in nidoviruses . The order Nidovirales is comprised of viruses with enveloped virions and non-segmented single-stranded linear RNA genomes of positive polarity ( ssRNA+ ) , whose replication is mediated by a cognate RNA-dependent RNA polymerase ( RdRp ) [15] , [16] . The order includes four families - the Arteriviridae and Coronaviridae ( including vertebrate , mostly mammalian viruses ) , and the Roniviridae and Mesoniviridae ( invertebrate viruses ) . The unusually broad 12 . 7 to 31 . 7 kb genome size range of this monophyletic group of viruses includes the largest known RNA genomes , which are employed by viruses from the families Roniviridae ( ∼26 kb ) [17] and Coronaviridae ( from 26 . 3 to 31 . 7 kb ) [18] , that have collectively been coined “large-sized nidoviruses” [19] . Viruses from the Arteriviridae ( with 12 . 7 to 15 . 7 kb genomes ) [20] and the recently established Mesoniviridae ( 20 . 2 kb ) [21] , [22] are considered small and intermediate-sized nidoviruses , respectively . Nidoviruses share a conserved polycistronic genomic architecture ( known also as “organization” ) in which the open reading frames ( ORFs ) are flanked by two untranslated regions ( UTRs ) [10] , [23]–[26] . The two 5′-proximal ORFs 1a and 1b overlap by up to a few dozen nucleotides and are translated directly from the genomic RNA to produce polyproteins 1a ( pp1a ) and pp1ab , with the synthesis of the latter involving a −1 ribosomal frameshift ( RFS ) event [27]–[29] . The pp1a and pp1ab are autoproteolytically processed into nonstructural proteins ( nsp ) , named nsp1 to nsp12 in arteriviruses and nsp1 to nsp16 in coronaviruses ( reviewed in [30] ) . Most of them are components of the membrane-bound replication-transcription complex ( RTC ) [31]–[33] that mediates genome replication and the synthesis of subgenomic RNAs ( a process known also as “transcription” ) [34] , [35] . ORF1a encodes proteases for the processing of pp1a and pp1ab ( reviewed in [30] ) , trans-membrane domains/proteins ( TM1 , TM2 , and TM3 ) anchoring the RTC [36]–[38] and several poorly characterized proteins . ORF1b encodes the core enzymes of the RTC ( reviewed in [39] , see also below ) . Other ORFs , whose number varies considerably among nidoviruses are located downstream of ORF1b ( hereafter collectively referred to as 3′ORFs ) . They are expressed from 3′-coterminal subgenomic mRNAs [40] , and encode virion and , optionally , so-called “accessory proteins” ( reviewed in [41]–[43] ) . The latter , as well as several domains encoded in ORF1a and ORF1b , were implicated in the control of virus-host interactions [44]–[48] . In addition to comparable genome architectures , nidoviruses share an array ( synteny ) of 6 replicative protein domains . Three of these are most conserved enzymes of nidoviruses: an ORF1a-encoded protease with chymotrypsin-like fold ( 3C-like protease , 3CLpro ) [49]–[51] , an ORF1b-encoded RdRp [49] , [52] , [53] and a superfamily 1 helicase ( HEL1 ) [54]–[57] ( reviewed in [58] ) . For other proteins , relationships have been established only between some nidovirus lineages , mostly due to poor sequence similarity . Two tightly correlated properties separate large- and intermediate-sized nidoviruses from all other ssRNA+ viruses , classified in several dozens of families and hundreds of species: a genome size exceeding 20 kb and the presence of a gene encoding a RNA 3′-to-5′ exoribonuclease ( ExoN ) , which resides in nsp14 in the case of coronaviruses [10] . The latter enzyme is distantly related to a DNA proofreading enzyme , and it is genetically segregated and expressed together with RdRp and HEL1 [14] , [59] . Based on these properties ExoN was implicated in improving the fidelity of replication in large- and intermediate-sized nidoviruses . This hypothesis is strongly supported by the excessive accumulation of mutations in ExoN-defective mutants of two coronaviruses , mouse hepatitis virus [60] and severe acute respiratory syndrome coronavirus ( SARS-CoV ) [61] , the identification of an RNA 3′-end mismatch excision activity in the SARS-CoV nsp10/nsp14 complex [62] , and the high efficacy of a live coronavirus vaccine displaying impaired replication fidelity due to nsp14-knockout [63] ( for review see [64] , [65] ) . Although the molecular mechanisms underlying ExoN's function in fidelity control remain to be elucidated , its acquisition by nidoviruses likely enabled genome expansions beyond the limit observed for other non-segmented ssRNA+ viruses [10] , [19] . Since ExoN-encoding nidoviruses have evolved genomes that may differ by up to ∼12 kb ( from 20 . 2 kb of Nam Dinh virus , NDiV , to 31 . 7 kb of Beluga whale coronavirus SW1 , BWCoV-SW1 ) , there must be other factors in addition to the proof-reading enzyme that control genome size . In this study we sought to characterize the dynamics of nidovirus genome expansion ( NGE ) . The NGE is defined by the entire range of the genome sizes of extant nidoviruses , from 12 . 7 to 31 . 7 kb , and thus concerns both pre- and post-ExoN acquisition events . Our analysis revealed that ExoN acquisition was part of a larger process with non-linear dynamics , during which distinct coding regions of the nidovirus genome were expanded to accommodate both an extremely large number of mutations and virus adaptation to different host species . Our results indicate that genome architecture and the associated region-specific division of labor [1] leave a footprint on the expansion dynamics of RNA virus genomes through controlling replication fidelity and/or other mechanisms . Eventually , these constraints may determine the observed limit on RNA virus genome size .
Nidoviruses have evolved genomes in a size range that accounts for the upper ∼60% of the entire RNA virus genome size scale and include the largest RNA genomes [10] . What did it take to produce this unprecedented innovation in the RNA virus world ? This question could be addressed in two evolutionary dimensions: time and amount of substitutions . Due to both the lack of fossil records and high viral mutation rates , the time scale of distant relations of RNA viruses remains technically difficult to study . Hence , we sought to estimate the amount of accumulated replacements in conserved nidovirus proteins and to place it into a biological perspective by comparing it with that accumulated by proteins of cellular species in the Tree of Life ( ToL ) . To this end , we used a rooted phylogeny for a set of 28 nidovirus representatives ( Table S1 ) , which was based on a multiple alignment of nidovirus-wide conserved protein regions in the 3CLpro , the RdRp and the HEL1 , as described previously [10] . The 28 representatives covered the acknowledged species diversity of nidoviruses with completely sequenced genomes [17] , [18] , [20] , [21] and included two additional viruses . For the arterivirus species Porcine reproductive and respiratory syndrome virus we selected two viruses , representing the European and North American genotypes , respectively , because we observed an unusually high divergence of these lineages; for the ronivirus species Gill-associated virus we selected two viruses representing the genotypes gill-associated virus and yellow head virus , respectively , because these viruses showed a genetic distance comparable to that of some coronavirus species [21] ( CL & AEG , in preparation ) . The nidovirus-wide phylogenetic analysis consistently identified the five major lineages: subfamilies Coronavirinae and Torovirinae , and families Arteriviridae , Roniviridae and Mesoniviridae . The root was placed at the branch leading to arteriviruses ( Fig . 1A ) according to outgroup analyses [10] . Accordingly , arteriviruses with genome sizes of 12 . 7 to 15 . 7 kb are separated in the tree from other nidoviruses with larger genomes ( 20 . 2–31 . 7 kb ) . We compared the evolutionary space explored by nidoviruses , measured in number of substitutions per site in conserved proteins , with that of a single-copy protein dataset representing the ToL [66] ( Fig . 1B ) . Using a common normalized scale of [0 , 1] , comparison of the viral and cellular trees and associated pairwise distance distributions revealed that the distances between cellular proteins ( 0 . 05–0 . 45 range ) cover less than half the scale of those separating nidovirus proteins . ( Fig . S1 ) . Unlike cellular species , nidoviruses are grouped in a few compact clusters , which are very distantly related . The distances between nidovirus proteins are unevenly distributed , reflecting the current status of virus sampling: intragroup distances between nidoviruses forming major lineages are in the 0 . 0–0 . 25 range , while intergroup distances between nidoviruses that belong to different lineages are in the 0 . 55–1 . 0 range . The distances separating the intermediate-sized mesonivirus from other nidoviruses tend to be most equidistant , accounting for ∼15% of all distances in the 0 . 55–0 . 85 range . Consequently , nidovirus evolution involved the accumulation of mutations in the most conserved proteins at a scale comparable to that of the ToL . This observation is instructive in two ways . First , it can be contrasted with the conservation of nidovirus genome architecture [58] , which emerges in this context as truly exceptional by conventional evolutionary considerations . Second , it makes it plausible that other , less conserved proteins might have diverged beyond the level that can be recognized by sequence alignment , thus establishing limits of the applicability of the alignment-based analysis of nidoviruses . We used both these insights to advance our study further ( see below ) . To quantify the relation of genome size change and the accumulation of substitutions , we plotted pairwise evolutionary distances ( PED ) separating the most conserved replicative proteins ( Y axis ) versus genome size differences ( X axis ) for all pairs of nidoviruses in our dataset ( Fig . 2 ) . It should be noted that the observed genome size differences may serve only as a low estimate for the actual genome size change , since it does not account for ( expansion or shrinkage ) events that happened in parallel between two viruses since their divergence . The obtained 378 values are distributed highly unevenly , occupying the upper left triangle of the plot . Using phylogenetic considerations ( Figs . 1A and S1 ) , four clusters could be recognized in the plot . Genetic variation within the four major virus groups with more than one species ( arteri- , corona- , roni- , and toroviruses ) is confined to a compact cluster I in the left bottom corner ( X range: 0 . 033–4 . 521 kb , Y range: 0 . 051–1 . 401 ) . Values quantifying genetic divergence between major lineages are partitioned in three clusters taking into account genome sizes: large-sized vs . large-sized nidoviruses ( cluster II , X: 0 . 002–5 . 433 kb , Y: 3 . 197–4 . 292 ) , intermediate-sized vs . other lineages ( cluster III , X: 4 . 475–11 . 494 kb , Y: 2 . 896–4 . 553 ) , and small-sized vs . large-sized nidoviruses ( cluster IV , X: 10 . 536–18 . 978 kb , Y: 4 . 159–5 . 088 ) . Points in clusters I , III and IV are indicative of a positive proportional relation between genome size change and the accumulation of replacements . The off-diagonal location of cluster II can be reconciled with this interpretation under the ( reasonable ) assumption that the three lineages of large-sized nidoviruses expanded their genomes independently and considerably since diverging from their most recent common ancestor ( MRCA ) . This positive relation is also most strongly supported by the lack of points in the bottom-right corner of the plot ( large difference in genome size; small genetic divergence ) . Overall , this analysis indicates that a considerable change in genome size in nidoviruses could have been accomplished only when accompanied by a large number of substitutions in the most conserved proteins . Next , we asked whether genome size change could be linked to domain gain and loss . We analyzed the phylogenetic distribution of protein domains that were found to be conserved in one or more of the five major nidovirus lineages [10] . Ancestral state parsimonious reconstruction was performed for the following proteins: ORF1b-encoded ExoN , N7-methyltransferase ( NMT ) [67] , nidovirus-specific endoribonuclease ( NendoU ) [68] , [69] , 2′-O-methyltransferase ( OMT ) [70] , [71] , ronivirus-specific domain ( RsD ) ( this study; see legend to Fig . S2 ) , and ORF1a-encoded ADP-ribose-1″-phosphatase ( ADRP ) [72]–[74] . This analysis revealed that domain gain and loss have accompanied NGE ( Fig . S2 and Table S2 ) . Particularly , the genetically segregated ExoN , OMT and NMT domains ( Fig . 3 ) were acquired in a yet-to-be determined order during the critical transition from small-sized to intermediate-sized nidovirus genomes . However , the combined size of these domains [10] accounts for only a fraction ( 49 . 7% ) of the size difference ( 4 , 475 nt ) between the genomes of NDiV ( 20 , 192 nt ) and Simian hemorrhagic fever virus ( SHFV ) , which has the largest known arterivirus genome ( 15 , 717 nt ) . The fraction that could be attributed to these and the three other protein domains is even smaller in other pairs of viruses representing different major nidovirus lineages ( CL & AEG ) . This analysis is also complicated by the uncertainty about the genome sizes of nidovirus ancestors that acquired or lost domains . In order to gain further insight in NGE dynamics , we analyzed large genome areas in which homology signals were not recoverable in the currently available dataset because of both the extreme divergence of distant nidoviruses and the relatively poor virus sampling ( Fig . 1 ) . To address this challenge , we developed an approach that establishes and exploits relationships between nidovirus genomes in an alignment-free manner on grounds other than sequence homology . To this end , we partitioned the nidovirus genome according to functional conservations in the genome architecture , using results for few characterized nidoviruses and bioinformatics-based analysis for most other viruses ( reviewed in [19] ) . With this approach , the genomes of all nidoviruses can be consistently partitioned into five regions in the 5′ to 3′ order: 5′-UTR , ORF1a , ORF1b , 3′ORFs , and 3′-UTR ( Fig . 3 , Table S3 ) . The 5′-UTR and 3′-UTR flank the coding regions and account for <5% of the nidovirus genome size . The borders of the three ORF regions that overlap by few nucleotides in some or all nidoviruses were defined as follows: ORF1a: from the ORF1a initiation codon to the RFS shifty codons; ORF1b: from the RFS signal to the ORF1b termination codon; and 3′ORFs: from the ORF1b termination codon to the termination codon of the ORF immediately upstream of the 3′UTR . As we detail in the Supplementary text ( Text S1 ) , the three ORF regions are of similar size but differ in expression mechanism ( Fig . 3 top ) and principal function . Thus , ORF1a dominates the expression regulation of the entire genome , and ORF1b encodes the principal enzymes for RNA synthesis , while the 3′ORFs control genome dissemination . This region-specific association may be described as a division of labor [1] . We then asked how the different regions contributed to the genome expansion . We initially noted that the intermediate position of the mesonivirus between the two other nidovirus groups is observed only in genome-wide but not in region-specific size comparisons ( Fig . 4 ) . In the latter , the mesonivirus clusters with either small-sized ( ORF1a and 3′ORFs ) or large-sized ( ORF1b ) nidoviruses . This non-uniform position of the mesonivirus relative to other nidoviruses is indicative of a non-linear relationship between the size change of the complete genome and its individual regions during NGE . Accordingly , when fitting weighted linear regressions for the three regions separately to the six datasets formed by nidoviruses with small and large genomes , support for a linear relationship was found only for the 3′ORF dataset of large nidoviruses; for all other regions a linear relationship was not statistically significant ( Fig . S3 ) . These results prompted us to evaluate linear as well as non-linear regression models applied to a dataset including all known nidovirus species ( n = 28 ) ( Fig . 5 ) . Two non-linear models were employed: third order monotone splines and a double-logistic regression . In the monotone splines , two parameters – the number and position of knots – determine the regression fit . We identified values for both parameters that result in the best fit ( Fig . S4 ) . Using weighted r2 values , we observed that the splines model captures 92 . 9–96 . 1% of the data variation for the three ORF regions . This was a 5–22% gain in the fit compared to the linear model ( 75 . 9–90 . 8% ) ( Fig . 5 ) . This gain was considered statistically significant ( α = 0 . 05 ) in two F-tests , a specially designed one and a standard one , as well as in the LV-test for every ORF region ( p = 0 . 019 or better ) and , particularly , their combination ( p = 9 . 1e-6 or better ) ( Table 1 ) . The splines model also significantly outperforms the double-logistic model ( p = 0 . 0014 ) ( Table 1 ) . These results established that the nidovirus genome expanded in a non-linear and region-specific fashion . Like each region , also the entire genome must have expanded non-linearly during NGE . Revealing its dynamic was our next goal . To this end , we analyzed the contribution of each of the five genomic regions to the overall genome size increase under the three models ( Fig . 6 and Fig . S5 ) . The top-ranking splines model ( Table 1 ) predicts a cyclic pattern of overlapping wavelike size increases for the three coding regions ( the 5′ and 3′UTR account only for a negligibly minor increase that is limited to small nidoviruses ) . Each of the three coding regions was found to have increased at different stages during NGE ( Fig . 6 ) . A cycle involves expanding predominantly and consecutively the ORF1b , ORF1a , and 3′ORFs region . One complete cycle flanked by two partial cycles are predicted to have occurred during the NGE from small-sized to large-sized nidoviruses . The complete cycle encompasses almost the entire genome size range of nidoviruses , starting from 12 . 7 kb and ending at 31 . 7 kb . The dominance of an ORF region in the increase of genome size was characterized by two parameters: a genome size range ( X axis in Fig . 6 ) in which the contribution of a region accounts for a >50% share of the total increase , and by the maximal share it attains in the NGE ( Y axis in Fig . 6 ) . For three major regions these numbers are: ORF1b , dominance in the 15 . 8–19 . 3 kb range with 72 . 9% maximal contribution at genome size 17 . 5 kb; ORF1a , 19 . 7–26 . 1 kb and 81 . 3% at 22 . 7 kb; 3′ORFs , 26 . 1–31 . 7 kb and 89 . 6% at 29 . 5 kb ( Fig . 6 ) . Furthermore , the shapes of the three waves differ . The first one ( ORF1b ) is most symmetrical and it starts and ends at almost zero contribution to the genome size change . This indicates that the ORF1b expansion is exceptionally constrained , which is in line with the extremely narrow size ranges of ORF1b in arteri- and coronaviruses ( with mean±s . d . of 4362±86 and 8073±50 nt , respectively; Fig . 4 and Fig . 6 ) . The second wave ( ORF1a ) is tailed at the upper end and is connected to the ORF1a wave from the prior cycle . This ORF seems to have a relatively high baseline contribution ( ∼20% ) to the genome size change up to the range of coronaviruses . The third wave ( 3′ORFs ) is most asymmetrical ( incomplete ) , as it only slightly decreases from its peak toward the largest nidovirus genome size to which this region remains the dominant contributor ( ∼77% ) . One partial cycle , preceding the complete one , is observed inside the genome size range of arteriviruses and involves the consecutive expansions of ORF1a and 3′ORFs , respectively . Also the main , but still very limited contributions of 5′- and 3′-UTRs ( <6% ) are observed here . The start of another incomplete cycle , involving the expansion of ORF1b and overlapping with the complete cycle , is observed within the upper end of coronavirus genome sizes . It must be stressed that nidoviruses occupy different positions on the trajectory that depicts the entire NGE dynamics . For the viruses with large genomes those with smaller genomes represent stages that they have passed during NGE; in this respect the latter may resemble ancestral viruses which have gone extinct . For the smaller genomes those with the larger ones represent stages that they have not reached during NGE . Mesonivirus and roniviruses seem to have been “frozen” after the first ( ORF1b ) and second ( ORF1a ) wave , respectively . The third wave ( 3′ORFs ) was due to the genome expansion of coronaviruses and , to a lesser extent , toroviruses ( compare the genome sizes of these viruses and the third wave position in Fig . 6 ) . These observations reveal that the constraints on genome size due to genome architecture may be modulated in a lineage-dependent manner .
It is broadly acknowledged that high mutation rates and large population sizes allow RNA viruses to explore an enormous evolutionary space and to adapt to their host [76] , [104] . Yet the low fidelity of replication also confines their evolution within a narrow genome size range that must affect their adaptation potential . Above , we present evidence for a new constraint on genome size in RNA viruses . In our analysis of nidoviruses , the conserved genome architecture and associated division of labor emerged as potentially powerful forces that are involved in selecting both new genes and positions of gene insertion during genome expansion . In this respect , the established wavelike dynamics of regional size increase could be seen as the footprint of genome architecture on genome size evolution . Ultimately , these constraints may determine the upper limit of the RNA virus genome size . The reported data point to an important evolutionary asymmetry during genome expansion , which concerns the relation between proteins controlling genome replication , expression , and dissemination , and may certainly be relevant beyond the viruses analyzed here . Importantly , the major diversification of nidoviruses by genome expansion must have started at some early point after the acquisition of ExoN [10] . From that point on , nidoviruses expanded their genomes in parallel in an increasing number of lineages , each of which may have acquired different domains in the same region . Extant representatives of the major lineages have very different genome sizes and essentially offer snapshots of different NGE stages . It seems that the host range may affect the outcome of this process , since the two families that infect invertebrates are on the lower end of the genome size range in the ExoN-encoding nidoviruses . For yet-to-be described nidoviruses , the genome expansion model can predict the sizes of the three coding regions by knowing the genome size only . The mechanistic basis of this fundamental relation can be probed by comparative structure-function analyses , which may also advance the development of nidovirus-based vectors and rational measures for virus control . Thus , the wavelike dynamics model links virus discovery to basic research and its various applications .
A dataset of nidoviruses representing species diversity from the three established and a newly proposed virus family was used ( Table S1 ) . A multiple alignment of nidovirus-wide conserved protein domains ( 28 species , 3 protein families , 604 aa alignment positions , 2 . 95% gap content ) as described previously [10] formed the basis of all phylogenetic analyses . To put the scale of the nidovirus evolution into an independent perspective , we compared it with a cellular dataset previously used to reconstruct the ToL , for which a concatenated alignment of single-copy proteins was used ( 30 species , 56 protein families , 3336 aa alignment positions , 2 . 8% gap content ) [66] . The proteins used in the nidoviral and cellular datasets are the most conserved in their group and , as such , could be considered roughly equivalent and suitable for the purpose of this comparative analysis . Rooted phylogenetic reconstructions by Bayesian posterior probability trees utilizing BEAST [105] under the WAG amino acid substitution matrix [106] and relaxed molecular clock ( lognormal distribution ) [107] were performed as described previously [10] . Evolutionary pairwise distances were calculated from the tree branches . A maximum parsimony reconstruction of the ancestral nidovirus protein domain states at internal nodes of the nidovirus tree was conducted using PAML4 [108] . The quality of ancestral reconstructions was assessed by accuracy values provided by PAML4 . The nidovirus genomic sequences are non-independent due to their phylogenetic relatedness [109] . When calculating the contribution of individual sequences to the total observed genetic diversity the uneven sampling of different phyletic lineages must be accounted for . To correct for the uneven sampling we assigned relative weights to the 28 nidovirus species by using position-based sequence weights [110] that were calculated on the alignment submitted for phylogeny reconstruction . The weights were normalized to sum up to one and were used in regression analyses ( see below ) . The sequence weights varied ∼7 fold from 0 . 017 to 0 . 116 . NDiV , which represents mesoniviruses , showed the largest weight of 0 . 116 that was distantly followed by those of the bafinivirus White bream virus ( WBV; 0 . 075 ) and roniviruses ( 0 . 06 each ) ; coronaviruses , making up the best-sampled clade , were assigned the lowest weights ( 0 . 017 to 0 . 028 each ) . The genome of each nidovirus was consistently partitioned into five genomic regions according to external knowledge ( see Results ) . To model the contribution of each genomic region to the total genome size change , we conducted weighted regression analyses ( size of a genomic region on size of the genome ) using three models – a linear and two non-linear ones . Position-based sequence weights were used and a confidence level of α = 0 . 05 was applied in all analyses . The regressions of the different genomic regions were not fitted separately but were joined to produce a genome-wide analysis . The combined contribution of all genomic regions to the genome size change must obviously sum up to 100% . To satisfy this common constraint , in each analysis , regression functions were fitted simultaneously to sizes of the genomic regions by minimizing the residual sum of squares , thereby constraining the sum of all slopes to be not larger than one . The linear model assumes a constant contribution of each genomic region during evolution which was modeled via linear regions . In the first non-linear model we applied third order monotone splines with equidistant knots [111] . We chose splines because of their flexibility and generality ( we do not rely on a specific regression function ) . The monotonicity constraint was enforced to avoid overfitting which was observed otherwise , and third order functions were chosen to obtain smooth , second-order derivatives . We explored the dependence of the performance of the splines model on variations in two critical parameters , the number of knots and the start position of the first knot . These two parameters define a knot configuration and determine a partitioning of the data into bins . In the first test we evaluated five different configurations generating from three to seven knots . Configurations using eight or more knots resulted in some bins being empty and were therefore not considered . For each number of knots the position of the first knot and the knot distance were determined as resulting in that configuration for which the data points are distributed most uniformly among the resulting bins . The exception was the 3-knot configuration , in which the position of the second knot was selected as the intermediate position in the observed genome size range ( 22 . 2 kb ) . Only configurations with equidistant knots were considered . All probed splines models were evaluated by goodness-of-fit values ( weighted version of the coefficient of determination r2 ) . In the second test we evaluated the model dependence on the position of the first knot by considering all positions that do not result in empty bins for the optimal number of knots determined using the approach described above . As another non-linear model we used a 7-parameter double-logistic regression function that mimics the splines model and more readily allows for biological interpretations . Logistic functions discriminate between two principal states – stationary and growth phases; a double-logistic curve comprises not more than three steady and two growth phases . The “length” of the different phases ( in the dimension of the independent variable; e . g . genome size ) , the steady state values ( in the dimension of the dependent variable , e . g . ORF size ) , and the “strength” of the growth ( e . g . the maximum slope of the curve between two steady states ) are controlled by the parameters of the regression function . Once estimated , the parameter values can be used to infer genome size intervals for which a particular ORF region is in a steady state as well as critical genome and ORF sizes at the transition between two steady states . Since double-logistic regressions did not converge for the 5′- and 3′-UTRs , linear functions were used for these two genome regions instead . Linear ( null hypothesis ) and splines ( alternative hypothesis ) regression models were compared using standard weighted F-statistics and a specially designed permutation test ( see below ) . To exclude overfitting as the cause of support of the more complex models , we utilized a more sophisticated framework ( LV-Test ) for the comparison of non-nested regression models ( linear vs . double-logistic and splines vs . double-logistic ) as detailed in [112] . The test was further modified to include weighted residuals according to virus sequence weights that account for sequence dependence . Since our null hypothesis ( linear model ) is at the boundaries of the parameter space , we developed a permutation test to further compare the linear and splines models . To this end , genome region sizes were transformed to proportions ( region size divided by genome size ) , randomly permuted relative to genome sizes , and transformed back to absolute values . These transformations are compatible with the constraints of the null hypothesis and the requirement that region sizes have to sum to genome sizes . Weights were not permuted . The linear and splines models were fit to the permuted datasets and F-statistics were calculated as for the original dataset . The p-value of the test is the fraction of F-statistics of permuted datasets that are larger than the F of the original dataset . It was calculated using 1 , 000 , 000 permutations that were randomly sampled out of ∼1029 possible permutations . Finally , we analyzed the contribution of each genome region to the total change in genome size under the three regression models . The contribution of each region according to a model was calculated as the ratio of change in region size to change in genome size ( first derivative of the regression function ) along the nidovirus genome size scale . These region-specific contributions were combined in a single plot for visualization purposes . To conduct all statistical analyses and to visualize the results we used the R package [113] . Accession numbers of virus genomes utilized in the study are shown in Table S1 . | RNA viruses include many major pathogens . The adaptation of viruses to their hosts is facilitated by fast mutation and constrained by small genome sizes , which are both due to the extremely high error rate of viral polymerases . Using an innovative computational approach , we now provide evidence for additional forces that may control genome size and , consequently , affect virus adaptation to the host . We analyzed nidoviruses , a monophyletic group of viruses that populate the upper ∼60% of the RNA virus genome size scale . They evolved a conserved genomic architecture , and infect vertebrate and invertebrate species . Those nidoviruses that have the largest known RNA genomes uniquely encode a 3′-to-5′exoribonuclease , a homolog of canonical DNA proof-reading enzymes that improves their replication fidelity . We show that nidoviruses accumulated mutations on par with that observed in the Tree of Life for comparable protein datasets , although the time scale of nidovirus evolution remains unknown . Extant nidovirus genomes of different size reached particular points on a common trajectory of genome expansion . This trajectory may be shaped by the division of labor between open reading frames that predominantly control genome replication , genome expression , and virus dissemination , respectively . Ultimately , genomic architecture may determine the observed genome size limit in contemporary RNA viruses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"viral",
"replication",
"genome",
"complexity",
"functional",
"genomics",
"rna",
"viruses",
"genome",
"evolution",
"virology",
"viral",
"classification",
"viral",
"enzymes",
"comparative",
"genomics",
"biology",
"genomics",
"evolutionary",
"biology",
"genomic",
"evolution",
"microbiology",
"computational",
"biology",
"viral",
"evolution"
] | 2013 | The Footprint of Genome Architecture in the Largest Genome Expansion in RNA Viruses |
A venerable history of classical work on autoassociative memory has significantly shaped our understanding of several features of the hippocampus , and most prominently of its CA3 area , in relation to memory storage and retrieval . However , existing theories of hippocampal memory processing ignore a key biological constraint affecting memory storage in neural circuits: the bounded dynamical range of synapses . Recent treatments based on the notion of metaplasticity provide a powerful model for individual bounded synapses; however , their implications for the ability of the hippocampus to retrieve memories well and the dynamics of neurons associated with that retrieval are both unknown . Here , we develop a theoretical framework for memory storage and recall with bounded synapses . We formulate the recall of a previously stored pattern from a noisy recall cue and limited-capacity ( and therefore lossy ) synapses as a probabilistic inference problem , and derive neural dynamics that implement approximate inference algorithms to solve this problem efficiently . In particular , for binary synapses with metaplastic states , we demonstrate for the first time that memories can be efficiently read out with biologically plausible network dynamics that are completely constrained by the synaptic plasticity rule , and the statistics of the stored patterns and of the recall cue . Our theory organises into a coherent framework a wide range of existing data about the regulation of excitability , feedback inhibition , and network oscillations in area CA3 , and makes novel and directly testable predictions that can guide future experiments .
The hippocampus , together with associated medial temporal-lobe structures , plays a critical role in memory storage and retrieval . A venerable line of classical theoretical work has shaped our understanding of how different hippocampal subfields subserve this function [1] , [2] . At the core of this body of work is the notion that area CA3 of the hippocampus operates as an autoassociator , retrieving previously stored memory traces from noisy or partial cues by the process of pattern completion . Furthermore , the theoretical framework of autoassociative memory networks helped elucidate how recurrently-coupled neural circuits , such as CA3 [3] , are capable of such pattern completion [4]–[10] . In this framework , synaptic plasticity stores memory traces in the efficacies ( or weights ) of the recurrent synapses of the neural circuit , and the recall of memories is achieved by the dynamical evolution of network activity through the synapses that were previously altered [11] . This framework has paved the way for a thorough analysis of the memory capacity of recurrent neural circuits [4] , [8]–[10] , and ensuing experimental results have confirmed many of its qualitative predictions [12] , [13] . However , despite much progress , existing models of auto-associative memories make drastic simplifying assumptions , as we describe below , concerning both the synaptic plasticity rules storing information in the circuit and the dynamics of the network at recall . First , at the level of memory storage , one powerful , yet biologically untenable , simplification made by most existing models [4] , [8]–[10] , [14] is the use of additive learning rules , whereby the cumulative effect of storing multiple memory traces is obtained as the linear sum of the contributions made by storing each individual trace . This simplification makes the analysis of the circuit tractable and suggests high memory capacity , but it also implies that synaptic weights can grow arbitrarily large or even switch sign , thereby violating Dale's principle . The shortcomings of assuming additive learning rules can be partially alleviated by introducing additional mechanisms such as synaptic scaling or metaplasticity , that ensure synapses are maintained in the relevant biological range [15] . Metaplasticity is loosely defined as any mechanism that manipulates or modulates synaptic plasticity; it comes in many forms , from the sliding threshold in BCM-like models [16] to sophisticated cascade models [17] . It is ubiquitous in the neocortex [18]–[20] and the hippocampus [21] , and has long been implicated in endowing synapses with powerful computational properties [16] . Importantly for memory storage , it was shown that one particular form of metaplasticity , the cascade model [17] , enables information to be stored in bounded synapses almost as efficiently as additive learning rules , whereas synapses with the same range of efficacies but without metaplasticity are hopelessly poor [17] . Unfortunately , despite their advantage at storing information , metaplastic synapses were found to perform equally poorly when the amount of recalled information was measured instead [22] , indicating that much of the information laid down in the synapses remained inaccessible for the standard attractor dynamics used at retrieval . Thus , perhaps surprisingly , we still do not know how competent memory recall is possible from more realistic synapses that suffer from a bounded dynamical range . Second , at the level of retrieval , there are also several aspects of hippocampal circuit dynamics of which we lack a theoretical account . For example , experimental work has long shown that synaptic plasticity is accompanied by changes in the excitability of CA3 neurons [23]–[25] , that the activity of pyramidal cells is modulated by several classes of inhibitory neurons [26] , [27] , and that the interaction of excitation and inhibition induces prominent oscillations in multiple frequency bands [26] , [28] . Yet , it is largely unclear whether and how these dynamical motifs contribute to efficient memory recall . Here , we develop a theory that specifically addresses the problem of memory recall from synapses with a limited dynamic range , and thus consider how various neuronal and synaptic biophysical properties of area CA3 contribute to this process . We start by assuming that synaptic efficacies are limited and adopt one particular , oft-studied model of metaplasticity , the cascade model , where synapses make transitions between different states which have the same overt efficacy but differ in their propensity to exhibit further plasticity [17] . In order to understand how memories can be recalled efficiently from such synapses , we derive recurrent network dynamics that are optimal for this purpose . Our approach is based on treating memory recall as a probabilistic inference problem , in which the memory pattern to be recalled needs to be inferred from partial and noisy information in the recall cue and the synaptic weights of the network , and network dynamics act to produce activity patterns that are representative of the resulting posterior distribution . Given the statistical properties of the prior distribution of patterns , the recall cues , and the learning rule , the network dynamics that we derive to be optimal for retrieval are fully specified without free parameters to tune ( except , as we show later , for some parameters affecting the speed of recall ) . The essence of our approach is that there is a tight coupling between the specifics of the learning rule governing memory storage and the dynamics of the circuit during recall . This approach has already helped reveal some basic principles of efficient memory recall in neural circuits [14] , [29]–[31] , but has not yet been applied to bounded metaplastic synapses . While we derived optimal recall dynamics with only minimal a priori regard to biological constraints , we found that approximately optimal retrieval can be achieved in a neural circuit whose basic functional structure resembles the standard , biophysically motivated dynamics used for additive learning rules [4] . Importantly , the solution involves several critical motifs that are not predicted by standard approaches , and yet map onto known features of the dynamical organisation of hippocampal area CA3 . First , precisely balanced feed-back inhibition [32] and pre- and postsynaptic forms of intrinsic plasticity ( IP ) [25] , [33] matched to the form of synaptic plasticity that stores the memory traces , are necessary for ensuring stability during retrieval . Second , oscillations that periodically change the relative contributions of afferent and recurrent synapses to circuit dynamics [34] , [35] can further improve recall performance by helping the network explore representative activity patterns more effectively . In sum , addressing the computational challenges associated with effective retrieval of information from bounded synapses provides novel insights into the dynamics of the hippocampal circuitry implementing this function . Thus , our work extends previous approaches that sought to understand the basic anatomical and physiological organisation of the hippocampus [2] , [36] as functional adaptions towards memory recall by providing a similar functional account of further crucial aspects of hippocampal organisation , involving plasticity and circuit dynamics .
We consider an auto-associative memory task in which a sequence of patterns , , is stored by one-shot learning in the synaptic efficacies ( or weights ) , , of the recurrent collaterals of a neural network . This models the network of pyramidal neurons in hippocampal area CA3 . ( We do not model other cell types or hippocampal subfields explicitly , but do consider their effects on CA3 pyramids , see also below ) . Here , is the activity of neuron in the pattern that was stored time steps prior to recall ( in other words , the age of this pattern is ) , and is the ( overt ) efficacy of the synapse between presynaptic cell and postsynaptic cell at the time of recall ( Fig . 1A–B ) . For tractability , we assume that neural activities are binary; although extensions of the theory to analogue activities are also possible [14] , [31] . Having derived the approximately optimal dynamics for memory recall , we first study its efficiency by numerical simulations , using the simpler Gibbs dynamics ( in light of their formal equivalence to a network of stochastic spiking neurons , see above ) . Specifically , our network dynamics proceed in discrete iterations corresponding to a full network update . In each such iteration , we first sample a random permutation to determine the order in which the neurons are to be updated , and then we update each neuron by applying Eqs . 2–5 . We first consider an example in which we store a specific pattern , followed by a sequence of random other patterns ( Fig . 2B , left ) . The retrieval cue , a noisy version of the original pattern , is used both as an initial condition at the beginning of recall and , as required by Eq . 3 , also as a source of external input biasing the network throughout the retrieval process . The activity of the network is stochastic , asymptotically sampling the corresponding posterior distribution , and the output of the network , , is taken to be the running temporal average of the network activity ( Fig . 2B , middle ) . We measure retrieval performance by root-mean-squared ( r . m . s . ) error , which implies that the optimal response is exactly the posterior mean . Even though sampling-based dynamics may , in general , suffer from slow convergence and mixing , as we will also show below , the particular dynamics here attains its asymptotic performance in only a few time steps ( Fig . 2B , right ) . A useful corollary of these dynamics is that the variability of the responses during recall also represents a computationally relevant quantity: the confidence in the correctness of the output ( the average activity ) . Indeed , as expected from a system with a well-calibrated representation of confidence , variability correlates strongly with the actual errors made by the network ( see Fig . S2 ) . To evaluate the overall retrieval performance of the network more systematically , we repeat the storage and retrieval procedure described above times . The patterns are drawn randomly from , a uniform distribution over binary vectors , and the age of each pattern to be recalled is drawn from , the prior over ( Fig . 1D , left , white curve ) . For each pattern , we simulate the effects of storing other random patterns on the synaptic weights , and then run our network dynamics , by starting it from the recall cue . At the end of each recall trial , lasting 100 time steps , we measure the error ( normalised Euclidean distance ) between the originally stored and the recalled pattern and average the errors across all trials . We compare the average performance of the optimal network to that of a ‘control’ network which is a feed-forward network that retains no information about the particular patterns that have been stored , but does perform optimal inference given the general distribution of patterns and the recall cue ( first two terms in Eq . 1 ) . This should provide an upper bound on recall errors because it simply ignores the information in the recurrent collaterals . While this control may seem trivial , several classical recurrent autoassociative memory networks are , in fact , unable to outperform it [14] , [43] ( see also Fig . S3 ) . The performance of our recurrent network deteriorates as a function of pattern age ( Fig . 2C , see also Text S3 ) , as expected , but the average error across pattern ages reveals that the network performs significantly better than the control ( Fig . 2D ) . In line with previous work that assumed additive synaptic plasticity [4] , [14] , retrieval performance is ultimately determined by the number of synapses per neuron ( Fig . 2D ) . Due to the limited dynamic range of synapses , recall performance is also influenced by the average pattern age , such that a larger network ( with more synapses per neuron ) can recall older patterns more proficiently ( Fig . 2E ) . A similar rescaling of errors is observed when using more biologically plausible sparse patterns [1] instead of the dense patterns we used in other simulations ( Fig . 2E ) . In this case , the amount of information per pattern is reduced and so more patterns can be remembered . The quality of recall of the control ( green dashed line ) also improves , because the prior over patterns also becomes more informative ( specifying a priori that most neurons should be inactive ) . Despite the well-known advantage of the cascade model over simple two-state synapses in storing information [17] , [55] , previous work using heuristically constructed recall dynamics was unable to demonstrate a similar advantage in recall performance [22] . Optimal dynamics confers substantial improvement in recall performance when synapses have multiple metaplastic states ( Fig . 3A ) . Importantly , one of the hallmark benefits of metaplastic synapses is that the time for which they retain information after encoding scales as a power-law of the number of synapses per neuron , instead of the catastrophically poor logarithmic scaling exhibited by deterministic two-state synapses [17] , [56] . The quality of information recall in our network shows the same scaling relationships , thus retaining this crucial advantage of metaplastic synapses ( Fig . 3B ) . Motivated by these findings , we now turn to the question of how these approximately optimal dynamics can be implemented , and further approximated , by neural circuit dynamics . For computational convenience , we will consider moderately sized all-to-all connected networks , dense patterns , and small average pattern ages , but these results generalise to the more realistic case of large sparsely-connected networks recalling sparse memories after longer retention intervals ( as suggested by Fig . 2D–E ) . We will take special care to assess the effects of sparse connectivity in those cases in which the detailed structure of the connectivity matrix can be expected to matter for recall performance . The dynamics defined by equations 2–5 have two appealing properties . First , by construction , they represent an approximately optimal solution to autoassociative recall , with all the parameters of the recall dynamics being derived from those characterising memory storage and the input noise . Second , at the same time , they are in a form that is in a loose agreement with standard reduced models of single neuron dynamics ( thresholded , linear summation of inputs ) . However , several details of the dynamics are unrealistic , and it is therefore necessary to show whether and how these details can be approximated by a neural circuit without severely compromising recall performance . Conversely , neuronal dynamics in cortical areas such as CA3 , that may be involved in the recall of associative memories , exhibit features that are mysterious from the perspective of recall based on conventional , additive , plasticity rules . We consider the possibility that these features might play a role in the approximations . A common theme in the approximations we are to consider is to replace an original , implausible term of the total somatic current by its statistical average . Averages can readily be taken over the activity of the population ( as we will see when we consider the role of the balance between excitation and inhibition ) , or over the statistics of previous patterns ( as we will see when we consider pre- and/or post-synaptic forms of intrinsic plasticity ) . In general we ask two questions about each approximation: The conclusion of the following sections will be that , in fact , several aspects of neural circuit organisation characterising hippocampal area CA3 can be understood as such necessary and efficient approximations . It is important to note that we only explicitly model pyramidal neurons ( the principal cells ) in CA3 , and that all other mechanisms involved in implementing approximately optimal memory recall will be described phenomenologically , in terms of their effects on the total somatic current of pyramidal neurons – which is the only computationally-relevant quantity in our model . Nevertheless , for each of these mechanisms , we will point out ways in which they may be dynamically implemented in the neural substrate and also quantify their effects in a way that allows direct comparisons with experimentally measurable quantities ( see also Discussion ) . Although Gibbs sampling was an attractive starting point for deriving dynamics that both work well in practice and can be related to biologically plausible neural network dynamics [54] , it suffers from a major computational shortcoming: sloth . This means that a very large number of iterations may be necessary before the samples are appropriately distributed ( i . e . , a long burn-in time ) . Further , the alacrity with which Gibbs sampling explores this distribution may be limited ( slow mixing ) . This would mean that consecutive samples are highly correlated , implying that very many of them would be needed to compute reliable expectations under the distribution [64] , [65] thus compounding the error in the output of our network – which is computed as just such an expectation ( Fig . 2B ) . These problems become particularly acute when the posterior distribution that needs to be sampled is multimodal ( when modelling hippocampal flickering , see below ) or itself exhibits strong correlations ( e . g . corresponding to strong coupling in frustrated Ising models ) . In fact , similar problems affect the alternative , deterministic mean-field or MAP dynamics ( discussed in Text S1 ) which suffer from local optima and regions of the objective function that gradient-based methods find hard to traverse .
The recall dynamics in our theory share some of the basic features of standard autoassociative memory networks ( recurrent excitation , linear inhibition [4] , [8] ) , but refine them in several critical ways . First , traditional approaches require considerable fine tuning of parameters for scenarios different from the standard Hopfield network ( storing binary patterns with the additive ‘covariance’ learning rule ) . In our approach , the basic form of the network dynamics during recall is fully specified by the statistical properties of the recall cue and the storage process , with no free parameter left to be tuned . ( Note , though , that tuning the amplitude of population oscillations in the more sophisticated , tempered transition dynamics , might be useful to improve the speed of convergence . ) This allowed us to include the effects of a markedly non-uniform prior over the delay after which a memory needs to be recalled , motivated by human forgetting data [42] , in contrast to traditional autoassociative memory models that assume , mostly implicitly , an ( improper ) uniform distribution over a finite range of recall delays ( see also Text S2 ) . Second , our theory provides an explicit prescription for how the excitability of neurons should be regulated depending on the efficacies of both their incoming and outgoing synapses . Akin to standard approaches , this means that excitability should depend on all previously stored patterns . However , while previous proposals for adjusting neuronal excitability require a somewhat heuristic offline procedure where the full list of stored patterns needs to be known [38] , we were able to show that the appropriate regulation of neural excitability in our system can be well approximated by commonly-assumed online forms of IP [60] resulting in competent recall performance . Third , while standard approaches only consider deterministic dynamics , our dynamics are stochastic . This makes it straightforward to optimise network performance for a squared-error loss , and additionally allows for a simple representation of uncertainty , thereby also naturally accounting for hippocampal flickering phenomena ( see also below ) . Finally , in keeping with other statistical treatments of autoassociative memory , which , however , number only few [30] , the recall cue modulates the network dynamics throughout retrieval ( as an external field ) rather than just being an initial condition . Its relative contribution to the dynamics reflects its quality ( or noisiness ) . One powerful , yet biologically-untenable , simplification made by previous autoassociative memory models [4] , [8]–[10] , [14] is the use of additive learning rules . The memory capacity of such networks is linear in the number of synapses per neuron [1] , [4] , but at the cost of unrealistic synaptic and neural dynamics . At the other extreme , the importance of bounded synaptic plasticity has been investigated using synapses with extremely limited dynamic ranges , including synapses with only two states [7] , [30] . While the capacity of such networks was shown to be disappointingly poor , recent work has shown that metaplastic synapses can store memories far more efficiently than synapses with the same range of efficacies but without metaplasticity [17] . As theoretical work investigating the role of metaplastic synapses in memory has so far concentrated on the benefits for storing information , it has been unclear how these benefits can be translated into recall performance – which is what ultimately matters for the organism ( after all , there is not much point in storing memories if they cannot be recalled ) . Surprisingly , almost no work has considered the quality of memory recall from metaplastic synapses , with the notable exception of Ref . [22] who found only very modest improvements compared with the recall performance of simple two-state synapses . Thus , it had remained unclear if the benefits of metaplasticity in terms of information stored can be translated into recall performance . We have shown that with appropriate recall dynamics , recall performance can in fact be substantially improved using metaplastic synapses ( without explicit optimisation of the synaptic plasticity rule used for storage ) , avoiding the characteristic of simple two-state synapses that they exhibit catastrophically poor logarithmic scaling of memory life time with the number of synapses ( Fig . 3 ) . While our measures of performance are currently based on numerical simulations , it may be possible to apply and extend the analytical approaches originally developed for computing the recall capacity of simpler network dynamics [22] to provide a systematic analysis of the performance of our optimal network dynamics . The performance of our recall dynamics follows the qualitative trends predicted by earlier analyses of metaplastic synapses [17] . However , there remain some quantitative discrepancies: for example , the cascade depth at which stored information is maximised is not the same as that at which recall error is minimised ( Fig . S5 ) . There may be several sources of these discrepancies . First , the degree to which our approximately optimal recall dynamics is able to make use of the information that is stored in the synapses may depend on the parameters of the system . Second , analyses of stored information typically quantify information in single synapses ( using measures such as the signal-to-noise ratio , SNR ) , while recall error ( e . g . fraction of correctly recalled bits for whole patterns ) is a result of the information stored jointly in all synapses . These metrics may themselves only be related to each other in a complex and nonlinear manner . For example , when synaptic weights are correlated , these two information measures will differ in general . In this work , we have side-stepped this issue by using an approximation which treats synapses in the network as independent given a particular pattern has been stored . This is formally incorrect in statistical terms , as we expect dependencies between synapses sharing a pre- and post- synaptic partners . Indeed , weak but significant correlations are observed between such synapses in the cortex [68] . It will be an important next step to explicitly consider these statistical dependencies and their significance for memory retrieval [43] . More importantly , however , these measures implicitly quantify performance on fundamentally different tasks: while SNR is appropriate for measuring recognition performance , i . e . the error in making the relatively simple binary judgement on whether a particular ( and noiseless ) pattern has been stored in the past [17] , our central interest has been recollection performance , i . e . the error on the much more demanding task of recalling the details of a high-dimensional pattern from noisy input [4] , [22] . Our model predicts changes in neuronal excitability that can be traced back to the specifics of CA3 synaptic plasticity ( i . e . the NMDA-receptor dependence of learning ) . In particular , we expect that the excitability of individual neurons should constantly change as a function of the state of the incoming and outgoing connections to and from the neuron . A range of experiments has long confirmed the homeostatic regulation of a neuron's responsiveness to injected current after chronic manipulations of network activity , corresponding to in the model [25] , [33] . More remarkably , recent evidence confirmed that neuronal excitability is also modulated by the strength of a neuron's outgoing connections [57] , [58] , closely matching the predictions of our model for : not only do the shifts in presynaptic neuron excitability follow the apparently anti-homeostatic direction predicted ( increases after LTP , reduction after LTD ) [57] , [58] , but this form of plasticity was also shown to be specific to excitatory-to-excitatory connections [58] , as required by the theory . To our knowledge , we are the first to ascribe a functional role to such presynaptic IP . Furthermore , while homeostatic plasticity has been introduced in some models as a heuristic addition to the network dynamics , meant to enhance stability during learning [59] , [60] , here it is derived from first principles , as a necessity for optimal recall . Our model also offers insights into some of the paradoxical findings surrounding IP . Namely , while homeostatic IP can be robustly expressed in vitro by pharmacological manipulations [25] , as we noted , the changes in excitability reported in vivo after more naturalistic manipulations ( e . g . after learning ) are typically anti-homeostatic [23] , [24] ( see also [25] ) . Our results suggest that , although different experimental manipulations may preferentially expose one or the other mechanism ( see [25] , [57] , [58] ) , both are necessary for circuit function , and that the presynaptic anti-homeostatic component dominates overall ( Fig . 4D ) . This would explain non-homeostatic increases in neural excitability in the hippocampus after hippocampus-dependent learning [23] , [24] . Our model offers an equilibrium theory – we expect constancy of neural excitability over the long-run , irrespective of the details of synaptic plasticity , at least as long as there exists a stationary distribution for the weights . Such a balance is consistent with experimental findings about spatial learning preserving the global firing rate of the network [69] . However , it is not obviously consonant with the observations of net shifts in excitability that have been measured across a population of CA3 neurons following learning [23] , [24] . One possibility is that this comes from a detection bias given sparse population patterns , such as those observed in CA3 [70] . That is , for such populations , we predict that the overall balance in excitability is achieved by large increases in excitability in the small subset of neurons that are active in the pattern , accompanied by small decrements of excitability in the inactive population of neurons . If the larger changes are preferentially detected ( e . g . simply due to signal-to-noise constraints in recordings ) , the changes in excitability that will be evident will be positive but not negative . Indeed , the pattern of experimental reports follows this trend: not all neurons recorded during the course of an experiment show detectable changes in excitability , but when they do , those changes are positive [23] , [25] . The model also makes the novel prediction that differential shifts in excitability should be recorded after separating neurons based on their activity in the pattern being stored . Using indicators of immediate early expression gene ( c-Fos ) expression to generate a lasting tag for the neurons that are active during the encoding of a particular memory ( when the animal is exposed to a novel environment [71] ) should make it possible to probe the excitability of these neurons at various retrieval delays , thus directly testing our prediction for the temporal evolution of excitability following memory storage ( Fig . 4D ) . Another key prediction of our model concerns the structure of the inhibitory circuitry that provides feedback inhibition in CA3 ( most likely by fast-spiking basket cells [72] , [73] ) . In particular , optimal recall dynamics require a form of feedback inhibition that dynamically tracks excitation , but without the need for tonic levels of excitation and inhibition to be tightly balanced . This mode of operation is fundamentally different from previously proposed theories of E/I balance in cortical circuits requiring tonic excitation and inhibition to match [61] , because it is the very difference between tonic excitation and inhibition levels ( on the time scale of a recall trial ) that carries the information about the identity of the pattern that needs to be recalled ( and about the confidence in this pattern ) . The network is therefore operating in a rather different regime from other work on associative memory in balanced spiking networks which has considered additive synaptic plasticity [74] . In our model , when the stored patterns are sparse , this translates into inhibition dominating neural responses , as reported for sensory responses in awake ( but not anesthesized ) mice , at least in V1 [75] . It also predicts a net shift between excitation and inhibition within the same neuron depending on the memory being retrieved , consistent with a shift in the average membrane potential of hippocampal place cells depending on whether the animal is inside or outside their place field [76] . At a finer temporal resolution , we also expect that fluctuations in excitation and inhibition are closely correlated . There is evidence for this in neocortical recordings [75] but a test of this prediction in the hippocampus has yet to be performed . At the level of the underlying hippocampal circuitry , the model predicts a high degree of overlap between a neuron's monosynaptic excitatory and disynaptic inhibitory partners , which could , in principle , be detected anatomically [77] or functionally [78] . Indeed , recordings in behaving rats confirm a close functional coupling between excitatory and inhibitory cell populations [79] . Moreover , as the underlying recurrent connectivity is modified , e . g . during learning , the inhibitory circuitry should be plastic as well , on a time course similar to that of learning at excitatory synapses . One recent experiment demonstrates that , at least in CA1 , such structural plasticity of inhibitory connections does accompany the induction of ( synaptic and structural ) plasticity at the excitatory synapses [80] . At the level of synaptic plasticity , theoretical models of excitatory-inhibitory networks have already predicted the dynamical matching of excitatory and inhibitory inputs in individual excitatory cells [63] . A recent experiment found evidence for this by measuring the profile of inhibition during learning of a new spatial representation [51] . This experiment revealed a reconfiguration of inhibitory activity that mirrored the reorganization of excitatory activity during place map formation , as we would expect from a process actively matching excitation with inhibition . We have shown that a periodic modulation of the relative contribution of external versus recurrent inputs facilitates the exploration of the state space of the network , and hence improves performance when there is limited time to answer a recall query . Such periodic modulation of extrinsic vs . recurrent inputs has been anticipated to be useful in the rather specific context of sequence disambiguation [81] but its general utility for memory recall under time-pressure is a novel aspect of our model . The computational role we ascribe to oscillations leads to a number of predictions that are unique to our theory . First , the main effect of oscillations in CA3 should be on the variability rather than the rates of pyramidal cell responses ( Fig . 7C ) . This points to gamma oscillations as potential biological substrates because they have only weak effects on the firing rates of CA3 pyramidal cells [82] . Second , the transmission of information to read-out areas of CA3 , most prominently to pyramidal cells in hippocampal area CA1 , should also be strongly modulated by the oscillation , because only samples from the target distribution at the peak of the underlying oscillation ( corresponding to temperature = ) are correctly representing the pattern that needs to be recalled . This means that CA3 input into CA1 should be periodically gated such that it impacts CA1 preferentially at this phase of the oscillation , which is consistent with gamma modulation of population rates being stronger in CA1 than in CA3 pyramidal cells [82] . Further evidence for this oscillatory coordination between CA3 and CA1 is that their in-phase synchronization in the lower gamma band is a signature of coordinated memory reactivation across the hippocampal network [83] , [84] , and in particular of the transfer of information between them [84] , [85] . This analysis does not delimit a role for theta oscillations . Another novel prediction of the theory is that response variability across the CA3 pyramidal cell population ( measured , for instance , by the entropy of their responses across trials ) should depend on the phase of gamma oscillations . This can be directly tested using multielectrode hippocampal recordings in awake behaving animals , pooling data across trials in which the same item is being recalled ( e . g . the same spatial position is being traversed ) , and measuring the variability across such trials as a function of the phase of the simultaneously recorded gamma oscillation . Mechanistically , our model of oscillations requires a rhythmic modulation of the different excitatory inputs to pyramidal cells in CA3 , affecting the relative contribution of recurrent versus perforant path inputs . While the specific mechanisms achieving this effect remain unclear , recent evidence suggests that at least two classes of inhibitory neurons – bistratified [26] and oriens-lacunosum moleculare ( OLM ) cells [86] – can rhythmically modulate external versus recurrent inputs to pyramidal cells , as would be required in our model . As OLM cells show strong modulation by gamma [87] , they seem to be ideally placed to play this role . Lastly , our analysis of retrieval performance revealed an inherent tradeoff between the utility of oscillations when exploring complex posterior distributions ( very wide for old patterns or multimodal as in the case of the flickering experiment ) and their detrimental effects when the correct answer is very clear ( the posterior is sharp and unimodal , as for recent patterns ) . It is tempting to speculate that the amplitude of gamma oscillations could be modulated with task difficulty ( estimated by some measure of response confidence , which is readily provided in a sampling based representation ) to optimise retrieval performance . Indirect evidence for this comes from chronic recordings in the human hippocampus showing increased gamma ( and theta ) power for retrieving remote versus recent autobiographical memories [88] . One key aspect of our theory is that the uncertainty about the patterns that are being recalled is represented along with the patterns themselves . This facilitates recall within the network , and it is also essential for downstream functions such as decision-making , for which evidence from recalled memories has to be combined with other , e . g . perceptual , sources of information – weighting each source of information with their respective certainties [45] . The behavioural ability to assess confidence in a retrieved memory trace has been demonstrated in various species , including humans [89] , [90] . We proposed that this is underpinned by a sampling-based neural code for uncertainty in the hippocampus [45] , [91] . Although the neural dynamics considered here are highly simplified , recent theoretical work has shown that the dynamics of more realistic leaky integrate-and-fire neurons can closely approximate those required by Gibbs sampling used here [54] . We showed that a sampling-based representation can explain some puzzling experimental observations revealing transient flickering in population responses following an instantaneous transformation of the spatial context [46] . In order to capture this flickering in traditional attractor dynamics , high levels of input noise would need to be assumed . However , in the actual experiments , special care was taken to make the cues for the individual environments as reliable as possible , so that the animals faced a problem of ambiguity rather than noise . According to our theory , hippocampal flickering is a variant of bistable ‘spatial perception’ , and as such can be viewed as a signature of the dynamics exploring different modes of the posterior , each corresponding to one of the stored memories . Bistability poses a particular challenge to attractor dynamics , which actively eliminate ambiguity by a winner-take-all mechanism . Conversely , sampling-based representations have been used to account for a host of perceptual and neural phenomena surrounding bistable perception [45] , [92]–[95] . If our sampling-based interpretation of flickering is correct , then it should be possible to modulate the degree of flickering , and the distribution of dwell-times for the individual representations , by experimentally manipulating sources of uncertainty ( the reliability of sensory cues , or the prior probabilities of the animal finding itself in any one of the possible environments ) . In sum , our work makes two important contributions . First , it shows for the first time that high-quality recall from metaplastic synapses is at all possible with neurally plausible dynamics . Second , the resulting recall dynamics involve several critical motifs that had not been predicted by standard approaches , and yet map onto known features of hippocampal dynamics . Thus the model provides insights into the computational role of several aspects of hippocampal activity and allows us to make a range of novel , experimentally testable , predictions .
We model a network of neurons , with connectivity defined by matrix , with if there is a synapse from neuron to neuron and , otherwise . To control connectivity ( Figs . 2E , 5E–F , and 6 ) , a randomly selected fraction of elements in was set to and the rest to . The corresponding synaptic efficacies are binary and defined by matrix , which is obtained as the result of storing a sequence of patterns by the cascade learning rule ( see below ) . The patterns are also binary and , consistent with data suggesting that the inputs to the CA3 network are decorrelated by the dentate gyrus [96] , we assume individual bits in a pattern to be independent , such that the distribution of the stored patterns factorizes over neurons ( and also , implicitly , over patterns ) : ( 9 ) where is the pattern density , or coding level . Finally , the recall cue is a noisy version of the original pattern , corrupted by independent noise modelled as a binary symmetric channel: ( 10 ) ( 11 ) with parameter describing the probability of a bit in the original pattern being flipped in the recall cue . Pattern age is assumed to be distributed geometrically with mean : ( 12 ) Learning is stochastic and local , with changes in the state of a synapse being determined only by the activation of the pre- and postsynaptic neurons , and and the current value of . Following the presentation of a pattern with activation and , the synaptic state transitions from the current state to to the new state . In the most general form , the probability of a synapse changing between any two states can be defined through a set of transition matrices , with , which leads to a large number of model parameters . A natural way to reduce this number is to define a transition matrix for potentiating , , and depressing , , events and separately map different neuron activation pairs into such events , possibly with some pairs leading to no change . Here , we assume a postsynaptically-gated rule , where the co-activation of pre- and post- neuron leads to potentiation , while an active postsynaptic neuron causes depression if the presynaptic neuron is silent , i . e . , , , , with denoting the identity matrix . For comparison , we also use the traditionally assumed presynaptically-gated learning rule [22] , [38] , with , , , . We express the two transition matrices using a generalization of Fusi et al . 's 2005 cascade model [17] , parametrized by , and the cascade depth . We index states corresponding to weak and strong synapses with and , respectively ( Fig . 1C ) . We describe the elements of the transition matrix , , as a sum of two terms: describing the probability that a weak ( strong ) synapse in state ( ) will potentiate ( depress ) to become a strong ( weak ) synapse , by occupying the ‘shallowest’ corresponding state in the cascade hierarchy , ( ) ; and describing the probability that a weak ( strong ) synapse in state ( ) , will remain weak ( strong ) , but even more so , by changing to a corresponding state that is one step deeper in the cascade hierarchy , ( ) . The probability of potentiation and depression decays as a geometric progression: for ( for ) , and we set ( ) to compensate for boundary effects . The probability of transitions towards deeper metastates is defined as for ( for ) with the correction parameter ( ) ensuring that different metastates are equally occupied for any pattern sparseness value , as done in the original model [17] . Additionally , the constraint ensures that we have proper transition probabilities for . The two additional parameters are inspired by previous work on simple binary synapses , which showed that , for sparse patterns , it is beneficial to have different transitions probabilities for potentiation and depression [10] . The original Fusi model [17] can be easily recovered by setting , . As is conventional , and plausibly underpinned by neuromodulatory interactions [97] , we assume that network dynamics do not play a role during storage , with stimuli being imposed as static patterns of activity on the neurons; and conversely , that the network does not undergo further plasticity during recall . We construct artificial recall dynamics that perform Gibbs sampling in the space of the joint distribution [49] . Introducing the pattern age as an auxiliary variable in the sampling procedure can be related to other auxiliary variable methods for sampling and is expected to improve sampling efficacy in the case of complex distributions [65] . Formally , the dynamics alternates between sampling an individual neuron's activity , conditioned on everything else ( including the current value to ) , and sampling the pattern age , according to the distribution ( which simplifies to , as is independent of the recall cue after conditioning on the stored pattern ) . This last step makes this sampling procedure biologically unrealistic , as computing the distribution over pattern ages requires knowledge of the full set of recurrent collaterals of the network . Practically , the procedure involves stochastic updates of neuron activities that are very similar to those of the simple Gibbs sampler , with the distinction that now the parameters depend on the pattern age , , and are computed using the distribution , obtained by projecting the distribution over the synaptic states directly into synaptic efficacies , without marginalising out . This means that the expression of the total current Eq . 3 now includes age-dependent parameters . Conceptually , this will result in a modulation of the relative contribution of the recurrent collaterals versus the external input from the cue , such that the recurrent dynamics dominate for recent patterns while the output is driven by the external input when the pattern is deemed to be old , when little or no information about the pattern is available in the weights . Finally , to be able to sample the pattern age , we limit the maximum possible pattern ages resulting in a finite discrete distribution , from which it is easy to sample ( in practice , we assume events with can be treated as equivalent ) . As there is little signal in the tail of the distribution over pattern ages , this does not affect performance for the network sizes considered here . Tempered transitions ( TT ) is a method that can improve sampling efficiency by using annealing , i . e . , systematically increasing and then decreasing a temperature parameter to ensure better exploration of the state space [64] . According to TT , in order to sample from a target distribution we choose a set of intermediate probability distributions , , indexed by the inverse temperature parameter , that are increasingly dissimilar from , but also easier to sample than , as decreases . The target distribution is represented at inverse temperature : . For each intermediate distribution we need a form of stochastic dynamics ( formally , defining a Markov transition operator ) which samples from ( i . e . has as its stationary distribution ) the corresponding . Starting from the current state , which is a sample at , i . e . from , a sampling cycle involves first lowering the inverse temperature in a sequence of steps down to and then increasing it back to . At each temperature level , we run the corresponding stochastic dynamics for a few steps starting from the last sample collected at the previous temperature level . This results in a sequence of intermediate samples , and , for the descending and ascending inverse temperature sequences , respectively ( Fig . S4 ) . Finally , all the intermediate samples produced at inverse temperatures are discarded , and the final sample produced at is accepted or rejected ( in which case network activity would need to return to the state it had at the beginning of the cycle ) with a probability given by the product of pairwise ratios of probabilities of all the intermediate states [64] ( see also Suppl . Info . in Ref . [49] ) . For us , the target distribution is the posterior . Common practice would dictate that we choose the intermediate distributions to be simply exponentiated ( with as the exponent ) versions of the target distribution , which would result in a completely uniform distribution at . However , this would not be efficient as the uniform distribution has no information about the original problem and thus results in unnecessarily wide Markov steps and , as a consequence , in a high rejection rate . Instead , we can use an important insight about the structure of our posterior to construct a better sequence of intermediate distributions . This insight is that the only factor that makes the posterior hard to sample from ( thus motivating the usage of TT in the first place ) is the correlations in it that are solely introduced by the weight-likelihood term , . ( Note that although we approximated this term above as factorized over the elements of , this still does not mean that it also factorizes over the elements of , of which the correlations are of issue here . ) Therefore , we chose only this term to be modulated by temperature , such that ( 27 ) The are two important features of exact TT dynamics that are problematic in the context of our network dynamics: first , the order in which neurons are updated in the ascending phase should be the exact reverse of that used in the descending phase; second , and more critically , an acceptance step is required at the end of each temperature cycle , as we saw above . As both the final acceptance step and the tight control on the ordering of neural updates are biologically unrealistic , the neural network approximates TT dynamics by ignoring sample rejections and by updating neuron activities in a random order during an oscillation cycle [64] . Under these approximations , the network dynamics are essentially identical to those of simple Gibbs ( Eqs . 21–23 ) , with all parameters unchanged , with the only modification that the recurrent currents are multiplicatively modulated by the inverse temperature ( cf . Eq . 21 ) : ( 28 ) At this is equivalent to sampling from a purely feed-forward network which uses no information in the recurrent weights and which is the network that we used throughout the paper as our ‘control’ . In general , the inverse temperature parameter can take values between ( corresponding to control ) and ( the target distribution ) . Here , we took a sequence that linearly interpolated between and a minimum value , with the amplitude of the oscillation being defined as . In all cases the number of neurons updated at each temperature level was chosen such that the total number of neurons updated over a whole cycle was the number of neurons in the network , . Although , due to the approximations we introduced above , the resulting network dynamics is no longer guaranteed to generate samples from exactly the correct posterior distribution , simulation results suggest that this approximation does not significantly alter the estimate of the posterior mean or the average response variability provided that the acceptance probability under the exact dynamics remains high , which we ensure by an appropriate modulation of . We start by defining the general setup and the default parameters used in all simulations , after which we proceed to list the parameter settings specific to each figure , in the order in which they are included in the main text . Unless otherwise specified , we considered a network of fully-connected neurons . The stored patterns were balanced , ( when sparse patterns were used , in Fig . 2 ) ; the recall cue noise was , the average pattern age was and the cascade parameters were , and depth . For measuring retrieval performance , we started from sampling the stationary distribution of the synaptic states , then we sampled from the prior , , one N-dimensional binary pattern which we stored by modifying synaptic states in the network according to the cascade learning rule described above . To separate the effects of synaptic correlations from the correlations among recalled activities , we simulated the effects of the storage of intervening patterns following the storage of by evolving individual synapses independently for steps according to the transition operator , , corresponding to storing a random pattern from the prior . At recall , we sampled a recall cue , which was a noisy version of , according to the noise model . This cue was provided as input to the network throughout retrieval as well as the starting point for the network dynamics . The network was allowed to evolve for 100 steps according to the dynamics we derived above . We took the temporal average across all these samples to be the recalled pattern , and computed the root mean square error between the stored and recalled pattern as . The performance for the control feed-forward network could be computed analytically ( as both the prior over and the recall cue distribution are factorized ) as , which for the case of balanced patterns ( , as in most cases considered here ) reduces to . When plotting mean performance as a function of pattern age , we used 10 trials for estimating the error for each ; for the average performance plots , we repeated the storage-retrieval procedure described above times , with pattern ages drawn randomly from the prior distribution . Average performance was measured as the average error over these independent runs , with all error bars representing the standard error of the mean . For Fig . 2B , we stored the binary pattern shown and used Gibbs dynamics without further approximations to retrieve it at a pattern age . To emphasize the stochastic aspects of the dynamics , we chose to show a subset of neurons whose activity evolution happened to be most variable . When investigating the dependence of the networks' performance on the total number of synapses per neuron in the sparse condition , we fixed network size and varied the connection probability , ; in particular , we used a network size of with one exception: we took for the case of 500 synapses/neuron . To investigate the effects of the prior over pattern ages , we reran the same procedure for different settings of the average pattern age ( adjusting the parameters of the network accordingly and resampling the pattern ages for which retrieval is performed ) . For the memory capacity analysis in Fig . 3 , we used the default parameters for the cascade ( and depth ) and optimized the parameters of the two-state synapse such that the signal decays exponentially with the same time constant as the prior assumed over patterns ( and ) . The memory capacity was defined , in line with classic SNR-based analyses , as the maximum pattern age for which retrieval error ( averaged over 100 trials for each ) was below a predefined threshold . The network evolved by simple Gibbs dynamics ( assuming unknown ) . When optimizing cascade depth ( Fig . S5 ) , we assumed , and estimated for each setting of the average retrieval error under the prior for and the mutual information between a synapse and the activity of its corresponding neurons ( marginalizing over the unknown ; i . e . using Eq . 17 ) . For Fig . 5A , we monitored the excitatory , , and inhibitory , , input to a neuron in a network with the default parameters settings , evolving by Gibbs dynamics , as described above ( the actual values obtained were discrete so we added a small amount of Gaussian jitter to them for visualization purposes ) . These two quantities are plotted against each other for three example neurons , two with low entropy ( red , blue ) and one with high response entropy ( ) . All-to-all connectivity was used for panel B , and sparse connectivity ( ) for panels E and F . Panel E shows the total recurrent current to an example neuron using the exact vs . the approximate expression for computing the inhibitory current , while the dynamics evolve by Gibbs . For Fig . 6 , we used relatively dense connectivity ( ) in order to preserve a relatively high number of synapses per neuron . The parameters for different approximations were steps for the online forms of IP , and we used E/I coherence and oscillation amplitude . For the simulations with oscillations ( Fig . 7 ) , we used different temperature levels , linearly spanning the range , with 5 neurons updated at each temperature step , such that one oscillation cycle corresponded to one full network update ( 50 descending and 50 ascending inverse temperature steps ) , as before . In this case , the posterior mean was computed by averaging over the samples obtained at ( inverse ) temperature . ( We kept the total simulation length constant , which meant that we had a reduced number of samples for estimating the posterior with oscillations , thus slightly favoring simple Gibbs dynamics without oscillations , but we deemed this a fair comparison if the duration of a recall trial is the real constraint ) . For Fig . 7C , we used high amplitude oscillations , , and , for each temperature level ( which defines the phase of the oscillation ) , computed average population firing as and the average response entropy as , with a neuron's response entropy defined as . For the flickering experiment ( Fig . 8 ) , we stored two consecutive patterns , and ( corresponding to the two contexts in the original experiments ) , and simulated the effects of having stored another 8 successive patterns independently across synapses as described above . For creating inputs to the network , cues were sampled independently in each time step from the input distribution conditioned on the pattern ( or ) corresponding to the current context , and hence their statistics changed abruptly at a context switch . For recall , we used oscillatory dynamics ( as in Fig . 7 , with ) with one minor modification: instead of taking a single relatively reliable recall cue as the input , each neuron integrated the evidence from the most recent past of several highly unreliable cues ( 75 cues , each with ) by simply summing them up ( this is optimal in our framework under the assumption that all 75 cues are i . i . d . , which is violated at a context switch ) . For constructing the actual figure , we started the simulations using and switched to at time , marked by the vertical green bar . As the effective recall cue was obtained by integrating over a period of several time steps , there is a corresponding time-window after the switch during which this effective recall cue is ambiguous ( due to the integration of conflicting evidence coming from two different contexts ) , and hence the posterior is determined primarily by the evidence from the weights , which is inherently multimodal . We computed the correlation between the response of the network ( at the peak of the oscillation , corresponding to ) and the two actually stored patterns , and , which are displayed in Fig . 8 . | Memory is central to nervous system function and has been a particular focus for studies of the hippocampus . However , despite many clues , we understand little about how memory storage and retrieval is implemented in neural circuits . In particular , while many previous studies considered the amount of information that can be stored in synaptic connections under biological constraints on the dynamic range of synapses , how much of this information can be successfully recovered by neural dynamics during memory retrieval remains unclear . Here , we use a top-down approach to address this question: we assume memories are laid down in bounded synapses by biologically relevant plasticity rules and then derive from first principles how the neural circuit should behave during recall in order to retrieve these memories most efficiently . We show that the resulting recall dynamics are consistent with a wide variety of properties of hippocampal area CA3 , across a range of biophysical levels – from synapses , through neurons , to circuits . Furthermore , our approach allows us to make novel and experimentally testable predictions about the link between the structure , dynamics , and function of CA3 circuitry . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"neuroscience",
"biology",
"computational",
"neuroscience"
] | 2014 | Optimal Recall from Bounded Metaplastic Synapses: Predicting Functional Adaptations in Hippocampal Area CA3 |
Whole-genome duplications have shaped the genomes of several vertebrate , plant , and fungal lineages . Earlier studies have focused on establishing when these events occurred and on elucidating their functional and evolutionary consequences , but we still lack sufficient understanding of how genome duplications first originated . We used phylogenomics to study the ancient genome duplication occurred in the yeast Saccharomyces cerevisiae lineage and found compelling evidence for the existence of a contemporaneous interspecies hybridization . We propose that the genome doubling was a direct consequence of this hybridization and that it served to provide stability to the recently formed allopolyploid . This scenario provides a mechanism for the origin of this ancient duplication and the lineage that originated from it and brings a new perspective to the interpretation of the origin and consequences of whole-genome duplications .
Ancient whole-genome duplications ( WGDs ) are major evolutionary events that have impacted several eukaryotic lineages , including plants , animals , and fungi [1] . Among plants , ancestral WGDs have been identified in monocots and core eudicots [2] , and more recent events are apparent in many lineages such as Arabidopsis , maize , and soybean [3–5] . In vertebrates , the existence of two ancestral WGDs ( but also more recent ones in teleost fishes and frogs ) has been proposed [2] . Earlier work has focused on establishing the periods at which these events occurred [6 , 7] and on assessing the functional and evolutionary aftermath of the doubling of the entire genetic complement [8] . However , we still do not fully understand what initially triggered these events . Perhaps the best-studied WGD is the one affecting an ancestor of the baker's yeast Saccharomyces cerevisiae , an event supported by the finding of numerous blocks of paralogs with conserved synteny [7 , 9] . It is now established that this event occurred just before the separation of Vanderwaltozyma polyspora from the S . cerevisiae lineage , originating a clade of post-WGD species ( Fig 1A ) [10] . In addition , it has been shown that the genome doubling was followed by extensive genome rearrangements and rampant gene loss that have since shaped these species' genomes , resulting in only a minor fraction of the WGD-derived paralogs ( ohnologs ) being retained [11 , 12] . Based on the high level of synteny found between reconstructed ancestrally duplicated gene blocks , it has been proposed that the yeast WGD has its origin in an autopolyploidization event [11] . This proposition has important implications with respect to the possible initial selective advantages that played a role after the polyploidization event . Polyploidy has been considered to promote evolutionary innovation because it facilitates neo- and subfunctionalization and buffers deleterious mutations . However , these mechanisms only provide an advantage after some time has passed and a number of mutations have accumulated . Conversely , simple increase in ploidy has been considered to put barriers to fast adaptation , as it masks beneficial recessive mutations and avoids rapid purging of deleterious mutations . Furthermore , most experimental work comparing populations of different ploidy generally provides support for the superiority of the normal ploidy versus increased ploidies in a given species [13] . Thus , the nature of the initial evolutionary advantage of the yeast WGD remains an open question . WGDs leave a footprint in the form of cohorts of homologous genes that duplicated in the same period . Phylogenetic analysis of gene families informs on the relative age of duplications [15 , 16] and hence is a powerful tool to study WGDs . When ancestral duplications are inferred from the genes encoded in a genome and their relative dates are mapped to a reference species tree , ancient WGDs are expected to lead to an accumulation of duplications mapped to the lineage in which the event occurred . Earlier analyses have used such approach to detect ancient duplications in vertebrates [17 , 18] and plants [19] . However , despite extensive phylogenetic work [20–22] , no study has assessed the global phylogenetic congruence of gene duplications and the WGD that occurred in the lineage leading to S . cerevisiae . Here , we set out to investigate patterns of past duplications in S . cerevisiae by analysing genome-wide sets of gene phylogenies ( i . e . , phylomes ) .
We based our analyses on a set of 26 completely sequenced genomes , for which we reconstructed a reference species phylogeny based on the alignment concatenation of 516 widespread , single-copy orthologs ( See Fig 1A , Materials and Methods ) . Subsequently , we used the phylomeDB pipeline [23] to reconstruct the evolutionary history of every protein encoded in the S . cerevisiae genome . These gene family trees were used to detect and date well-supported duplication events , using a phylogeny-based method described elsewhere [16] . In brief , the method exploits the temporal information provided by the branching patterns in a given gene tree: a duplication must be older than the lineages diverging subsequent to it and younger than lineages branching earlier . Using this information , we can map duplications to the reference species tree and compute duplication densities per gene and branch ( S1 Fig ) . Unexpectedly , our analyses revealed the largest duplication peak ( 0 . 28 duplications per gene ) at the branch preceding the divergence between Saccharomyces and a clade containing the genera Kluyveromyces , Lachancea , and Eremothecium ( Ashbya gossypii ) [24] , hereafter referred to as KLE ( Fig 1B ) . To assess whether this peak was indeed related to the WGD event , we limited our analysis to those duplications leading to conserved pairs of WGD-ohnologs as defined in the Yeast Gene Order Browser ( YGOB ) database [25] . Note that YGOB uses a synteny criterion which is independent of the specific gene phylogeny . We found that the pre-KLE duplication peak was more apparent in the subset of duplications leading to conserved pairs of ohnologs , which indicates that this ancestral duplication peak is indeed related to the observed WGD paralogous blocks ( Fig 1A and 1B ) . Of note , not all duplications resulting in pairs of conserved , syntenic ohnologs mapped to the pre-KLE peak ( n3 ) . A second accumulation of duplications appeared at the branch preceding the divergence of a clade formed by Zygosaccharomyces rouxii and Torulaspora delbrueckii ( referred to as ZT hereafter ) with the post-WGD species ( n4 ) . A smaller fraction of duplications mapped to the expected WGD location ( n5 ) or subsequent branches . We assessed the degree of divergence between syntenic ohnologs derived from duplications at the pre-KLE peak and those from duplications at the WGD node , as the two more divergent points of interest , and found that the former had significantly larger divergences ( Fig 1C ) . This supports that gene pairs whose duplications are predicted to be more ancestral by a topological approach are also more divergent at the sequence level . It also indicates that the genes in paralogous blocks may be composed of distinct sets of genes , diverged at different times . To discard the possibility that our unexpected result was artifactual and to understand what may have caused the dispersion of the duplication mappings outside the WGD node , we carefully assessed possible methodological and interpretation pitfalls . First of all , given that the pre-KLE branch is among the longest in our species phylogeny , the ancestral peak could simply indicate a higher number of duplications accumulated over a longer period of time . We thus measured the correlation between duplication densities and branch lengths for the whole phylogeny . While a high correlation was indeed observed when considering all the duplications ( r2 = 0 . 92 Pearson ) , this was not the case when the analysis was restricted to only those duplications leading to syntenic ohnologs ( r2 = 0 . 00 ) . We next assessed the effect of using alternative yeast species as a seed in the phylome reconstruction and observed that the use of Candida glabrata or V . polyspora phylomes resulted in similar patterns of duplication densities ( see S2 Fig ) . Another , always contentious point is the use of a reference phylogeny . Although the reconstructed species tree was highly supported and congruent with earlier reconstructions [24] , an alternative branching order for the KLE species had been previously presented [26] . This alternative topology suggested that the Lachancea , Kluyveromyces , and Eremothecium are not monophyletic but rather stem out sequentially from the lineage leading to S . cerevisiae ( see S3 Fig ) . Such organization could potentially affect our results if , for instance , the pre-KLE duplications were found to be partitioned among the new internodes ( i . e . , branches ) created by this topology . To test this , we repeated the analysis using the alternative topology as a reference . Our results show that the underlying topology does not affect the central finding that an apparent duplication peak existed before the divergence of KLE species ( S3 Fig ) . Finally , we tried an alternative method to map the duplication events of ohnologs by using the reconciliation-based algorithm implemented in Notung [27] , which rendered similar results ( S4 Fig ) . Thus , a different species topology and a different duplication detection method do not alter the main result that the majority of ohnologs have apparently diverged before the expected WGD . We next tried to assess the possible effect of stochastic errors or artifacts in the gene trees . We did so by focusing on the trees that contained pairs of conserved ohnologs in S . cerevisiae . Short sequences tend to be less reliable and more prone to stochastic errors . First , we examined the signal present in subsets of sequences of varying lengths ( <500 aa , 500 to 1 , 000 aa , and >1 , 000 aa ) . As seen in ( S5 Fig ) , the three groups of genes consistently provide a very low duplication signal at the WGD , while signals at the two previous branches ( pre-ZT and pre-KLE ) are much larger . Secondly , we assessed the robustness of our main result across a range of different methodological approaches for gene tree reconstruction . We tested three different maximum likelihood programs: PhyML [28] , RAxML [29] , and Fasttree [30]; and one program based on Bayesian inference ( BI ) : Phylobayes [31] . In addition to the best-fitting evolutionary model used in our standard analyses , we used PhyML to test the effect of using two different , more complex models ( C20-CAT [32] and Covarion [33] ) and a different search heuristic , subtree pruning and regrafting ( SPR ) , instead of the default nearest-neighbour interchange ( NNI ) . Finally , we tested two different support methods in RAxML ( rapid bootstrapping and Shimodaira–Hasegawa ( SH ) support ) and in PhyML ( approximate likelihood ratio test [aLRT] and bootstrapping ) . A summary of the different methods can be found in S1 Table . Results of different methods are not directly comparable because different subsets of trees pass the filters for a given procedure ( see S6 Fig ) . However , when a tree passed the filters for any given two methods , the result was highly consistent in most cases ( 86% overall agreement ) . Overall , our main result that duplications are apparently older than the expected WGD remained consistent ( see S7 Fig ) . The fraction of ohnolog duplications mapped to the expected WGD node is minimal ( <15% in all datasets ) , while more ancestral duplications are prominent with >50% of the duplications being mapped to the two nodes preceding the WGD ( pre-ZT and pre-KLE ) , although the balance between these two prominent peaks differed between the methods . These differences notwithstanding , the main conclusion of the duplication density analysis is consistent across methods: the majority of ohnologs have inferred duplication ages that predate the expected time of the WGD . Finally , phylogenetic artifacts such as long-branch attraction ( LBA ) can produce wrong topologies with high support [34] . It is possible that trees containing paralogs diverging at very unequal rates may have been affected by LBA , misplacing duplications closer to the root . In fact , differential rates among paralogs are expected when processes of neofunctionalization are acting . One way to ascertain whether LBA is affecting the topology is reconstructing the tree with and without the out-groups . In the absence of LBA , the in-group topology is expected to remain stable [34] . We applied this test to the trees containing ohnologs and found that the majority of trees ( 85% ) gave consistent mappings of the duplication of the ohnologs , indicating that the effect of LBA is not widespread and does not significantly affect the duplication mapping . We performed a second test to see whether LBA could explain the observed patterns . For this , we devised sequence simulations in which one of the ohnologs was made to evolve 20 times faster than its paralog . Despite the use of such extreme values , the duplication peak at simulations was detectable at the expected location , and artifactual peaks were significantly smaller and not apparent at the pre-KLE lineage ( see Materials and Methods , S8 Fig ) . Gene conversion among duplicates may result in underestimation of duplication ages , possibly accounting for part of the disappearance of the WGD peak , but not for the presence of the pre-KLE peak . Thus , LBA and gene conversion may have blurred the signal of the WGD peak but cannot account for the prominent pre-KLE peak . Our results show compelling evidence that a majority of yeast genes defined as ohnologs have diverged before the expected period of the WGD . This overall result holds even though the exact mapping from individual gene trees may vary across methodologies and datasets . The event under study is very ancient , and genes contain a limited amount of information; thus , degradation of the signal is expected . However , stochastic noise would explain a diffusion of the signal but not the existence of a stronger , more ancient duplication peak . We have also shown that distorting processes such as LBA cannot account for the observed patterns . We thus turned to assess other possible biological explanations for our observation . We further considered possible evolutionary scenarios that could result in the observed patterns of ancestral duplications seen for the ohnologs . We reasoned that an interspecies hybridization would result in phylogenetic patterns reminiscent of duplications that would be mapped to the common ancestor of the two hybridizing species ( Fig 2A ) , providing a possible scenario to explain our puzzling results . The process of hybridization originates a new lineage by bringing together two diverged genomes . Orthologous genes coming from each of the parental species would appear as paralogs in standard analyses , since they are homologous genes encoded in the same genome [35] . A phylogenetic analysis , however , would map the apparent duplication to the time of divergence of the two parental species ( Fig 2B ) . This necessarily predates the time of the formation of the hybrid: that is , the hybridization point does not coincide with the point at which the apparent duplications would be mapped . As we will see below , our hypothesis is that the hybridization may have shortly predated the actual WGD point ( i . e . , occurred at node n5 in Fig 1 ) . In support of this , we calculated the duplication densities on the well-studied yeast interspecies hybrid S . pastorianus [36] . This species is the result of a recent hybridization between S . cerevisiae and S . eubayanus [36] . The sequenced genome of S . bayanus is the closest related genome to S . eubayanus , and therefore we expect the highest duplication peak to appear at the common ancestor between S . cerevisiae and S . bayanus . The duplication density analysis , as predicted , yielded an apparent duplication peak at the common ancestor S . cerevisiae and S . bayanus , but not at the lineage where the hybridization and the doubling of the genome is known to have occurred ( Fig 2C ) . The results found for the S . cerevisiae lineage could thus be readily explained by a past hybridization between lineages diverging just after the observed peak and before the post-WGD species . Considering this and the current genomic sampling , species close to , but not necessarily within the KLE and ZT clades , would be the prime suspects of potential partners in the proposed ancestral hybridization ( Fig 2D ) . To explore this possibility further , we inferred properties of the two putative parental lineages from the current genomic sampling . We did so by inspecting individual S . cerevisiae gene phylogenies in the above-mentioned phylomes ( see Fig 3 as an example ) and by measuring phylogenetic affiliations using phylomes reconstructed with reduced taxonomic sets ( see S2 and S3 Tables ) . Phylogenetic affiliations were measured by scanning the gene tree topologies to examine the species contained in the sister groups ( i . e . , neighbouring clades ) of the sequences from post-WGD species ( see Materials and Methods ) . We categorized them according to one of the two lineages that diverged after the pre-KLE peak and the origin of the post-WGD species: the KLE clade and the ZT clade . From now on , we consider the ZT cluster as the extant clade closest to one of the parents ( parent A ) , while the KLE cluster will be considered as the closest to the other parent ( parent B ) . Although , for simplicity , we refer to ZT and KLE clades as parental lineages , it must be clearly stated that it is our understanding that the actual parents may have been close to , but not necessarily within , these clades . Accordingly , three possible topologies can be considered: two in which the S . cerevisiae seed sequence groups with either parental species ( A or B , respectively ) and a third one in which the S . cerevisiae sequence has the two parental lineages as a sister group ( C ) ( see Fig 4A ) . Our results ( Fig 4A ) indicated that a large majority ( 60%–82% , depending on the choice of species used in the reduced phylome; see S9 Fig ) of the trees showed a topology congruent with the currently accepted phylogeny , i . e . , the post-WGD species grouping with the ZT clade . When only the trees that contain S . cerevisiae proteins with a conserved ohnolog are considered , the results remain very similar ( see Fig 4A ) ( 54%–78% , depending on the choice of species used in the reduced phylome; see S10 Fig ) . This suggests that this or a related lineage would have been involved in the hybridization ( parent A ) and that genes derived from this parental species constitute a majority of the genome in extant post-WGD species . In contrast , a remarkably low fraction of genes showed an affiliation only to the KLE lineage ( 4%–14% ) , whereas a larger percentage ( 14%–28% ) of genes had as a sister group a combination of the two putative parental clades ( C ) . This would suggest that one of the actual parental lineages did not belong to the KLE but rather diverged before . The analysis repeated using different phylogenetic methodologies confirmed these results ( S11 Fig ) . The high percentage of trees supporting the A topology could be the result of total or partial gene conversion , which is common in recent hybrids [37] . We can only clarify this matter by analysing gene trees that contain pairs of conserved ohnologs . Depending on the distribution of the two ohnologous genes when compared to KLE and ZT , we can distinguish between nine different topologies ( see S12 Fig ) . Forty percent of the trees contained a topology in which the two yeast ohnologs grouped together ( topologies A–A 1 , B–B 1 , and C–C 1 ) . This could be due to total or partial gene conversion from one of the parents to the other . The gene conversion events seem to favour genes from parent A , since in 30% of the mentioned cases both retained genes are more closely related to this parent . We performed sequence evolution simulations including different degrees of gene conversion to estimate what levels would be necessary to alter the tree topology ( see Materials and Methods ) . In our settings ( S13 Fig ) , conversion of 25% of the gene sequence was sufficient to lead to a higher probability of the duplication being mapped to a younger node . Thus , gene conversion , which renders duplications to appear younger , has a much larger effect than LBA . These analyses underscore the difficulty of correctly determining the position of parent B . There is a strong signal for the parent B to have diverged just before the KLE clade ( shown by topology A–A 2 and B–B 2 ) , which is present in 33% of the trees . As we will discuss below , we consider that recombination between the two parental subgenomes , including total or partial gene conversion , must have been common in the period following the hybridization , explaining not only the bias in descent among ohnologs and singletons but also the widespread mixture of phylogenetic signals in gene trees that is typical for this clade [20] . The availability of genomes from fungal species in which recent hybridizations or WGDs have been described allows us to assess the patterns of phylogenetic affiliations and compare them with the patterns observed for S . cerevisiae . On the one hand , Rhizopus delemar [38] and Hortaea werneckii [39] are thought to have undergone a recent WGD . On the other hand , S . pastorianus [36] and the wine strain S . cerevisiae x S . kudriavzevii VIN7 [40] are recognized as recent hybrids for which the putative parental species are known . It is important to remark that some of the described WGD species may indeed be as well the result of hybridizations , as it is proposed here for the post-WGD clade , but that the current sampling of species prevents the detection of the alternative parental signals . We reconstructed the phylomes of these four species ( see S2 Table ) and computed phylogenetic affiliations as explained above , but adjusting A and B to the known parents or the corresponding neighbouring clades . Putative WGD species showed a clear dominance of the immediate preceding clade ( Fig 4B ) . The recent hybrids , on the other hand , presented a split topology distribution , with roughly half of the trees supporting the A topology and another half supporting the B topology ( Fig 4C ) . This clearly provides evidence of the dual origin of these species . As negative controls , we examined the phylomes of two species without anomalous ploidy , Candida albicans and Penicillium digitatum ( Fig 4D ) [41] , and the above-mentioned simulated yeast phylome in which one of the ohnologs was evolving at a faster rate ( Fig 4E ) . This analysis shows that hybrids present a clear dual pattern of phylogenetic affiliations when the gene phylogenies are examined in the presence of the two parental lineages . This pattern is clearly distinct from what is in genomes with normal ploidy or in recent WGDs . This dual pattern is also present in the analysis of the yeast genome . Of note , in this case the two alternative phylogenetic affiliations are not equally represented . This difference with respect to recent hybrids can be attributable to the larger period of time since the hybridization and the preferential loss or conversion of genes coming from one of the parental lineages , which necessarily altered the balance between the two phylogenetic affiliations . As mentioned above , inferred ancestral collinearity has been used to favour simpler WGD scenarios involving autopolyploidization [11] . However , such studies indistinctly used KLE and ZT clades to infer ancestral gene arrangements and thus could not inform about differences between the putative parents . Although the position of the parental species cannot be ascertained with confidence , we can take KLE and ZT clades as the two extremes of their possible divergence . We therefore assessed the level of micro- and macrosynteny conservation among the KLE , ZT , and post-WGD clades , by considering them separately . To do this , we reanalysed the information of orthology and syntenic blocks provided by YGOB [42] . We first assessed the differences between ZT and KLE by searching for gene arrangements conserved within ZT and KLE , but different between the two groups . These differences can be considered ancestral to the two groups and thus likely present at the time of the proposed hybridization . Only 32 cases of broken synteny and 11 translocations of a single gene were noted ( S4 Table ) . When searching for these synteny breaks in post-WGD species , we found that they had inherited the arrangement present in either KLE or ZT in similar amounts ( 15 and 17 , respectively ) ( S5 Table ) . Of note , the patterns shared by KLE and post-WGD species could result from lineage-specific rearrangements in the lineage leading ZT clade so we cannot unequivocally impute them to the hybridization . This result is consistent with the absence of disagreements between syntenic ohnologous blocks noted earlier [11] . However , an autopolyploidization scenario would predict a larger number of shared syntenic arrangements between the post-WGD and its closest clade ( ZT ) . Furthermore , the absence of disagreements in such a small number of blocks can be explained by other factors , including gene conversion , so it cannot be considered a definitive proof of autopolyploidization . In addition , we found that the number of conserved pairs of adjacent orthologs between KLE and ZT clade was high , as was the number of conserved pairs between post-WGD species and any of the KLE and ZT clades ( S6 and S7 Tables ) . Finally , we found no differences in terms of the minimal amount of rearrangements [43] between each S . cerevisiae syntenic block [11] and those in either ZT or KLE species ( Fig 5 ) . These results speak for the high collinearity of the two putative parental clades at the proposed time of hybridization ( see S4 and S5 Tables ) , which is congruent with the small divergence time estimated between the two clades at the time when the post-WGD clade originated ( S14 Fig ) . In addition , considering the high level of collinearity between the ZT and KLE clades , and the lack of differences in terms of synteny conservation when compared to S . cerevisiae , the proposed hybridization is as compatible with the observed level of conserved synteny between duplicated blocks as a simpler autopolyploidization scenario . Recent yeast hybrids have been shown to present extensive recombination between parental genomes , including total or partial gene conversion [37 , 44 , 45] , which breaks the initial correlation between phylogenetic origins of neighbouring genes and removes sequence and structural differences between homologous chromosomes . This and extensive differential gene loss and genome rearrangements that have occurred within the post-WGD clade have presumably eroded the few initial differences between the two parents that we can reconstruct . We conclude that , given the similar levels of collinearity implied by both scenarios and the confounding effects of extensive gene loss , homologous recombination , gene conversion , and genome rearrangements , synteny cannot be used in this case to disentangle whether the WGD was triggered by an autopolyploidization or a hybridization event . The proposed hybridization is a very ancient event , and thus , the remaining signal must be necessarily weak . We have shown that gene order differences between the putative parental species involved in the hybridization were extremely low , and we consider that this signal may have been completely eroded , which explains why the hybridization was not evident from earlier analyses based on synteny . Our phylogenetic results , however , do provide clear support for the existence of an ancient interspecies hybridization and are not compatible with a simple autopolyploidization scenario . The observed phylogenetic affiliations in ohnologs and singletons , biased towards one of the putative parental lineages , as well as the absence of synteny disagreements in ohnologous blocks , can be reconciled with the assumption that the proposed hybridization was followed by widespread recombination events between the two parents subgenomes , some of which would have led to partial or total gene conversion . As noted before , this process is common in recent yeast hybrids [37 , 44 , 45] , and it is natural to expect that this would have occurred in an ancient hybridization . Notably , hybridization followed by recombination between parental subgenomes also explains another long-held observation of the post-WGD clade: that there is a variable mixture of disparate phylogenetic signals present across different gene trees [20 , 22] . Our results also indicate that an apparently more ancestral duplication peak occurred in addition to duplications around the expected WGD point . We hypothesize that the occurrence of these two rare events in the same lineage is not the result of coincidence . We propose two possible scenarios that naturally link the two events and explain the observed patterns ( Fig 6 ) . In the simplest scenario , two diploid cells from distinct species form an allotetraploid . Subsequent recombination and massive gene loss would render a lineage in which the number of chromosomes has effectively doubled . In this case , hybridization directly results in the observed WGD , because a fraction of the final gene set is retained as “ohnologous” pairs , either from the same or from different parental species . Alternatively , two haploid cells from different species form an allodiploid . Such hybrids are largely unstable and cannot undergo the sexual cycle , but they can propagate clonally [46] . An additional duplication by autopolyploidization would stabilize the hybrid by enabling meiotic recombination . This mechanism , which also prevents backcross with the parental lineages , has been proposed as a necessary step to stabilize some interspecies hybrids [47] and is a scenario commonly considered in recent plant hybrids [48] . Both scenarios cannot be distinguished with the data at hand , but the ability of haploid cells to fuse through mating provides a possible mechanism for the latter . Further investigation of how these two mechanisms participate in the formation of natural hybrids is necessary [46] . Importantly , some of the steps proposed by our model were also contemplated in models considering autopolyploidization scenarios [10] Our results provide compelling evidence for an ancient hybridization in the yeast lineage and bring about novel implications in our understanding of the evolution of eukaryotic genomes and the origin of functional divergence after WGDs . Remarkably , besides the pattern of ancient duplications , the proposed model provides plausible explanations to other common observations in the post-WGD clade . The phylogenetic relationships within and around the post-WGD clade have always been difficult to resolve , and a great diversity of phylogenetic histories among different genes has been noted [20 , 22] . A chimeric origin of the clade , combined with events of recombination between genes from different parents—as observed in current hybrids [44 , 49]—would readily explain an increased variability in phylogenetic signals recovered from different genes . Such intragenic recombinations , together with full gene conversion and differential gene loss , may as well partially explain the observed dispersion of the phylogenetic mapping of duplications from syntenic ohnologs around the expected WGD point . Furthermore , ohnologs have been shown to present selection pressures intermediate of singleton genes and those from small-scale duplications of a similar age [50] . Finally , notable exceptions to expectations from the gene balance hypothesis , which posits that WGD would favour duplications of entire complexes rather than single subunits , have been noted [51] . Most of these observations have been interpreted in the light of an assumed rapid sequence and functional divergence after duplication . However , under a hybridization scenario , a fraction of the predicted ohnologs originate from distinct species , and thus , sequence and functional differences are expected from the start . In contrast , an autopolyploidization scenario poses the problem of how reproductive isolation was achieved and faces the lack of a clear selective advantage before neo- or subfunctionalization occurs . Interspecies hybridization brings together different physiological properties and isolates sexually the newly formed lineage , hence providing an initial selective advantage to explain observed WGDs in eukaryotes . Considering the widespread presence of hybrids among current species , this scenario should also be considered when interpreting ancient polyploidies . The proposed approach and an increased genome sampling around the relevant lineages will enable testing the possible implication of interspecies hybridization in other eukaryotic WGDs .
Proteomes were downloaded from their original databases ( S8 and S9 Tables ) . The proteomes of S . pastorianus and H . werneckii were not available . We thus downloaded the genomes and predicted their proteomes using Augustus [52] . The final S . pastorianus [36] and H . werneckii [39] proteomes comprised 11 , 460 and 20 , 509 proteins , respectively . Phylomes—complete collections of phylogenetic trees for each gene encoded in a given genome—were reconstructed using the automatic pipeline described in Huerta-Cepas et al . [23] . Briefly , the pipeline starts with a seed genome and proceeds as follows: for each protein encoded in the seed genome , a Smith-Waterman similarity search was performed against a database containing the proteomes listed above . Results were then filtered based on e-value ( <1e-05 ) and sequence overlap ( >50% coverage over the query sequence ) . The query and the selected hits ( homologous sequences ) were then aligned using a sophisticated multiple sequence alignment strategy in which three different alignment programs were used ( Muscle v3 . 8 [53] , Mafft v6 . 712b [54] , and Kalign v2 . 04 [55] ) to align the sequences in forward and reverse orientation . The resulting six alignments were combined into a consensus alignment using M-coffee [56] . This alignment was then trimmed to remove poorly aligned columns with trimAl v1 . 3 [57] using a consistency-score cutoff of 0 . 1667 and a gap-score cutoff of 0 . 9 . Trees were reconstructed using the best-fitting evolutionary model . The selection of the model best fitting each alignment was performed as follows: a neighbour joining ( NJ ) tree was reconstructed as implemented in BioNJ [58]; the likelihood of this topology was computed , allowing branch-length optimization , using seven different models ( JTT , LG , WAG , Blosum62 , MtREV , VT , and Dayhoff ) , as implemented in PhyML v3 . 0 [28]; the two models best fitting the data , as determined by the AIC criterion [59] , were used to derive maximum likelihood ( ML ) trees . Four rate categories were used , and invariant positions were inferred from the data . Branch supports were computed using an aLRT based on a chi-square distribution , as implemented in PhyML [60] . S2 Table lists the complete phylomes reconstructed for this project . Seven complete phylomes were reconstructed using S . cerevisiae , C . glabrata , V . polyspora , S . pastorianus , H . werneckii , the yeast S . cerevisiae VIN7 , and R . delemar as seed species . These phylomes have been deposited in phylomeDB ( http://phylomedb . org [61] ) . A simulated phylome using S . cerevisiae as seed was also reconstructed ( see below ) . In addition , a total of 18 reduced phylomes were reconstructed ( see S3 Table; http://genome . crg . es/~mmarcet/yeast_hybrids/phylome_table . htm ) . In these reduced phylomes , for the seed species , only one sequence was present in the tree; all paralogs for this species were removed to ensure that a clear phylogenetic position could be established . Finally , two previously reconstructed phylomes , stored in phylomeDB , were used for comparative purposes: C . albicans ( phylomeID: 205 ) and P . digitatum ( phylomeID: 150 ) [41] . Phylomes were scanned using ETE v2 . 2 [62] , which implements all the algorithms described here . The reference species tree shown in Fig 1 was reconstructed using a multigene concatenation method . From the S . cerevisiae phylome , we selected 516 protein-coding genes found in single copy across the 26 species considered . Their protein alignments were then concatenated , resulting in a combined alignment of 285 , 507 positions . An ML phylogenetic tree was then reconstructed using PhyML v3 . 0 [28] using the LG model . Four rate categories were used , and invariant positions were inferred from the data . Bootstrap support was calculated based on 100 replicas . All nodes were fully supported ( 100% bootstrap ) . The species tree presented in S3 Fig was reconstructed using the same data as the previous tree , but enforcing the desired topology when reconstructing the tree . For the S . pastorianus tree ( Fig 2C ) , 215 genes were selected from the phylome , and the final alignment contained 117 , 408 amino acids . The same methodology was used to reconstruct the tree . Each tree in a phylome was scanned to detect and date duplications using a phylogeny-based algorithm described earlier [16] . In brief , this algorithm traverses the tree and uses a so-called species-overlap algorithm to detect duplication nodes . Duplication nodes are defined as those nodes where the two daughter branches share at least one species . The relative age of this duplication is assumed to be at the last common ancestor of the species diverged after the duplication ( i . e . , those contained in the two daughter branches ) . Each duplication was then mapped onto the corresponding ancestral lineage in the species tree . The total number of duplications was divided by the total number of trees that were rooted at a deeper branch in the species tree ( i . e . , those that are informative for the evaluated lineage ) . For instance , to estimate the duplication density at the WGD branch , only trees that contain at least one pre-WGD species were considered . S1 Fig shows a schematic representation of the duplication mapping process . This analysis was performed using three different phylomes , in which S . cerevisiae , V . polyspora , and C . glabrata were used as seed , respectively . For each phylome , two different datasets were used . In the first one , all the trees in the phylome were used ( see Fig 1B , green dot , and S2 Fig , lighter dots ) , the second was based on trees in which a pair of retained ohnologs was present , and only the duplication node leading to the two seed ohnologs was used ( see Fig 1B , yellow dot , and S2 Fig , darker dots ) . Ohnologs were obtained from YGOB [42] , which uses a synteny criterion combined with sequence similarity but is not phylogenetically informed . Only trees that contained both ohnologs were considered . This second set ensured that the duplication density was not affected by duplications not related to the WGD event . We plotted the correlation between duplication densities and branch lengths . We mapped the duplication event of the two ohnologs to the species tree and only kept those S . cerevisiae sequences whose duplication point mapped to the WGD node or to the pre-KLE node ( see Fig 1 ) . Only trees that contained at least one ZT sequence , one KLE sequence , and one out-group sequence were considered . Blast scores were normalized by dividing the blast score obtained when searching from a seed yeast protein to the ohnolog pair by the blast score obtained from searching the seed yeast protein to itself . In a separate analysis , pairwise alignments of the conserved ohnologs were reconstructed using Muscle v3 . 8 [53] . The Kimura distance between the two sequences was calculated using protdist as implemented in the phylip package [63] . The frequency of distances of the two different distributions and blast score frequencies were plotted with R [64] . Significance of the difference in distributions was assessed using a two-sample Kolmogorov-Smirnov test ( see Fig 1C ) . The two populations were significantly different , with a p-value for the blast scores of 2e-04 and for the Kimura distance of 2 . 9e-05 . PL-R8s [14] was used to assess the divergence times in the concatenated species tree ( S14 Fig ) . Smoothing parameter was estimated using cross validation . The divergence between S . cerevisiae and C . albicans ( 235 MyA as estimated by Douzery et al . [65] ) was used as calibration point . The same protocol was used in individual trees that contained two ohnologous pairs . Trees were pruned so that they only contained the closest sequence belonging to each ZT-KLE group . The frequencies of ages ( see Fig 1C ) were plotted using R [64] . The two populations were significantly different , with a p-value of 4 . 5e-07 . Notung v2 . 6 [27] was used to reconcile the same set of trees used above to the species tree obtained from the concatenation of 516 proteins ( see above ) . Once the two trees were reconciled , we used the option to estimate upper and lower bounds to obtain the time when the duplication of the two S . cerevisiae ohnologs had taken place . Only estimates that had a definite upper and lower bound that could be mapped to a single branch of the species tree were considered . The number of trees that mapped the duplication onto a given branch was divided by the total number of trees in order to obtain the duplication density . A set of 846 trees were selected from the S . cerevisiae phylome where pairs of conserved ohnologs were found , as predicted by YGOB [42] . The alignments were taken from the phylome reconstruction done previously . Then , for each tree , several additional phylogenetic reconstruction methods were used . Fasttree [30] was used with default values . PhyML [28] was run again three times; in all cases , four rate categories were applied and invariant positions were calculated from the data . The first time the CAT model C20 was used [32] , the second time the Covarion model [33] was used ( —cov_free –cov_ncats = 3 ) , and finally , the same models as in the phylome were used , but instead of using NNI to estimate the tree topologies , SPR was used . For the three methods , the aLRT support was calculated . A fourth run with PhyML was performed using the same method as during the phylome reconstruction , but instead of calculating aLRT support values , bootstrap values based on 100 replicates were computed . RAxML [29] was applied using the PROTGAMMALG model and rapid bootstrapping to obtain the branch support . The SH support as implemented in RAxML was calculated over the same set of trees . A Bayesian approach was also used . Phylobayes [31] was used to reconstruct the trees; for each tree , two chains were run for a minimum of 500 cycles; every 100 cycles , the two chains were automatically compared; and if the discrepancies were lower or equal to 0 . 3 and the effective sizes were larger than 50 , the process was stopped . The majority rule consensus , annotated with posterior probabilities , was obtained for each tree . For each set of trees , the duplication density for the duplication point that led to the diversification of the two S . cerevisiae ohnologs was calculated . Results can be found in S7 Fig . Only nodes in which the support value at the common ancestor of the two ohnologous sequences has an aLRT > 0 . 95 or a bootstrap > 95 or a posterior probability > 95 were considered . The same set of 846 trees was reconstructed with no out-group sequences using the same methodology used for phylome reconstruction ( see above ) . The trees included only the post-WGD sequences and the ZT and KLE sequences . Trees were then checked to see whether the two S . cerevisiae ohnologs had a common ancestor that contained no sequences of the ZT and KLE groups , therefore giving support to the WGD , or if they had sequences of either group in between . Only trees in which the common ancestor of the two S . cerevisiae sequences has a support over 0 . 5 were considered . The same procedure was performed in the same set of trees taken from the phylome . Out-groups in this case were used to root the tree , and then the same analysis was performed . Fifteen percent of the trees gave a different prediction when the two methodologies were performed . For each sequence encoded in the yeast genome that had one-to-one orthologs in all the species considered , alignments obtained during the phylome reconstruction were trimmed to remove all positions with gaps . The number of species considered was reduced to 12 , including S . cerevisiae , all the species belonging to the ZT and KLE clades ( T . delbrueckii , Z . rouxii , Kluyveromyces lactis , A . gossypii , Lachancea kluyveri , L . thermotolerans , and L . waltii ) and four outgroups ( Schizosaccharomyces pombe , Yarrowia lipolytica , C . albicans , and Wickerhamomyces anomalus ) . The species tree ( see Fig 1 ) was pruned to match this set of species . The existing tree branch that contained S . cerevisiae was bifurcated to create two new branches containing simulated yeast paralogs . The first branch contained the original S . cerevisiae leaf , but its branch length was cut in half . The second branch contained a new S . cerevisiae leaf with a branch length ten times longer than the original . Each protein was then made to evolve along this tree using Rose [66] . Tree-puzzle [67] was used to obtain the mutation frequency observed at each site of the alignment . Tree-puzzle was run with the JTT model; the gamma distribution was estimated from the data using 16 rate categories . These mutation frequencies produced very conserved , unrealistic alignments with few mutations; therefore , the frequencies were multiplied by 20 , resulting in more realistic alignments . Indel frequency was set at 0 . 0003 . The resulting sequences were then treated as a newly simulated phylome , which was run through the phylome pipeline . Duplication densities were then mapped onto the species tree ( see S8 Fig ) . Reduced phylomes were reconstructed in such a way that they contained only one species for the post-WGD ( seed species ) , one species for the ZT clade , and one for the KLE clade , in addition to three outgroups ( C . albicans , Y . lipolytica , and S . pombe ) . In addition , for the seed species , only the seed sequence was included; other paralogs in this organism were excluded from the tree . A reduced phylome was reconstructed for each pair of ZT-KLE species . Three post-WGD species were used as seed ( S . cerevisiae , C . glabrata , and V . polyspora ) ( see S3 Table ) . For each seed sequence in the reduced phylomes , the sister branch was analysed . First , trees were excluded if they did not have any homologs in ZT , in KLE , or in any of the out-group species . Then , the support of the clade containing the seed sequence and its most immediate neighbouring clade was evaluated using aLRT values . Only clades with support higher than 0 . 95 were considered . The phylogenetic affiliation of the seed sequence was classified into one of the following groups , according to the species that were present in its neighbouring clade ( i . e . , sister branch ) : A , the species located in the sister branch belonged to the ZT clade formed by Z . rouxii and T . delbrueckii ( putative parent A ) ; B , they belonged to the clade formed by A . gossypii , K . lactis , L . thermotolerans , L . waltii , and S . kluyveri ( KLE clade , putative parent B ) ; and C , they contained a mix of both clades . This was done for the whole phylome ( S9 Fig ) and for the trees in which the seed sequence was part of a conserved ohnologous pair ( S10 Fig ) . Analysis was repeated across several phylogenetic methods ( see above ) ( S11 Fig ) . Topologies of pairs of ohnologs were assessed by reconstructing the trees including the ohnologous pair to those trees that already contained a sequence with a conserved ohnolog . Depending on the relation between the two ohnologs and the chosen KLE and ZT parent sequences , we distinguish between nine possible topologies: A–A 1 , A–A 2 , B–B , B–B 2 , C–C , C–C 2 , A–B , A–C , and B–C ( see S12 Fig ) . For the complete phylomes used ( C . albicans phylome , H . werneckii phylome , S . pastorianus phylome , R . delemar phylome , and S . cerevisiae x S . kudriavzevii VIN7 phylome ) , the two groups of species situated closest to the seed species according to the species tree were used as parental species unless the parental species were known ( see S2 Table ) . Trees were then pruned so that only the seed , the two parents , and out-groups were kept . ETE v2 . 2 . [62] was then used to analyse the sister branch ( i . e . , neighbouring clade ) to the seed sequenced . Sequences were classified as explained above . For the same set of sequences used in the LBA simulation ( see above ) , we used ROSE [66] to make the sequences evolve along a species tree that contained two S . cerevisiae sequences . The branch lengths of the tree were inferred by selecting those genes that had an A topology and a C topology and were consistent across different phylogenetic methods . Two species trees were derived from these two sets of genes , and branch lengths were mapped onto our simulated species tree . Once sequences were reconstructed , sets of genes affected by different levels of gene conversion were reconstructed . For each percentage of gene conversion , one yeast sequence was taken for each set of sequences , and a given percentage of its sequence was replaced by the same fragment of the second yeast sequence . Phylogenetic trees were then inferred in the same way used in the phylome ( see above ) , and duplication densities were calculated ( see S13 Fig ) Orthologous relationships between species and gene order data were obtained from the YGOB . Blocks of conserved synteny between the S . cerevisiae genome and the ancestral genome as predicted by Gordon et al . [11] were considered as conserved syntenic blocks . The genome of L . waltii was not used because of the high fragmentation of the assembly . Genes in the genomes were arranged using Z . rouxii as reference ( see S5 Table ) . Genomes were scanned for the presence of breaks in gene order that were common in the KLE clade and not found in either ZT species . Orthologs of the genes surrounding the breaks were searched in five post-WGD species ( S . cerevisiae , Tetrapisispora blattae , Kazachstania naganishii , Naumovozyma castellii , and C . glabrata ) in order to assess whether they followed the ZT or the KLE clade in their gene order ( see S5 Table ) . For each pair of genes located next to each other in the S . cerevisiae genome , we checked whether the orthologs in each of the ZT-KLE species were also contiguous . The same procedure was repeated in order to compare the ZT and KLE species . For each syntenic block , the orthologs were obtained for each of the seven species in the ZT and KLE clades . MGR [43] was used to compute the number of rearrangements that occurred between each ZT/KLE species and S . cerevisiae . | Genome duplication is a major evolutionary process that has shaped the genomes of several eukaryotic lineages including vertebrates , plants , and fungi . The sequencing of the baker's yeast Saccharomyces cerevisiae in the 1990s revealed the presence of conserved blocks of duplicated genes , indicating an ancestral duplication of the entire genome . Subsequent work has clarified when this event occurred and what genomic rearrangements followed , but the underlying mechanistic origin of such a large-scale event remains poorly understood . Here we used a large-scale phylogenetic approach to examine the individual evolutionary histories of all yeast genes and assessed the time at which each duplication occurred . This survey revealed evidence for an ancient hybridization event between two ancestral species in the lineage in which the whole-genome duplication had occurred . We further characterize this hybridization event and the properties of the putative parental species . We propose that the whole-genome duplication was a direct consequence of this hybridization , providing a means by which the initially sterile hybrid could regain fertility . This scenario provides a mechanistic understanding of the origin of the ancient yeast whole-genome duplication and brings a radically different perspective on the interpretation of the origin and evolutionary consequences of whole-genome duplications in eukaryotic lineages . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Beyond the Whole-Genome Duplication: Phylogenetic Evidence for an Ancient Interspecies Hybridization in the Baker's Yeast Lineage |
Cell mechanics has proven to be important in many biological processes . Although there is a number of experimental techniques which allow us to study mechanical properties of cell , there is still a lack of understanding of the role each sub-cellular component plays during cell deformations . We present a new mesoscopic particle-based eukaryotic cell model which explicitly describes cell membrane , nucleus and cytoskeleton . We employ Dissipative Particle Dynamics ( DPD ) method that provides us with the unified framework for modeling of a cell and its interactions in the flow . Data from micropipette aspiration experiments were used to define model parameters . The model was validated using data from microfluidic experiments . The validated model was then applied to study the impact of the sub-cellular components on the cell viscoelastic response in micropipette aspiration and microfluidic experiments .
Cell mechanics has proved to be a widely used label-free biomarker to discern phenotypes , detect pathologies and more importantly , monitor existence or progression of a disease [1–3] . The most prominent example is the changes in cell biology and morphology when it evolves from a healthy to a cancerous state [1 , 3] . These changes take place at the molecular level affecting properties of individual components of cell internal structure , but eventually leading to alterations in mechanical properties of the whole cell . Eukaryotic cells are composed of multiple components that contribute diversely to cell mechanics . The most important components are cell membrane , internal cytoskeleton , and nucleus . The cell membrane is a viscous fluid-like matter which consists of various lipids , cholesterol , and embedded proteins . It contributes to cell viscosity , bending resistance , and incompressibility . Cytoskeleton , which is a network of interconnected filaments of different types , connects the cell membrane with underlying sub-cellular components . It is believed to be one of the main contributors to cell mechanics [1] . The nucleus is the largest organelle among sub-cellular components , demonstrating solid-elastic behavior [4] , and it is typically stiffer than the cell itself [5] . It is comprised of multiple components including nuclear envelope and chromatin network . Improved understanding of the role that each cell component plays towards cell mechanics may be beneficial for diagnosis and therapy of diseases [2] . One of the novel approaches for studying mechanical properties of cells involves development of custom-designed microfluidic devices where deformability of cells is estimated; this is usually done by measuring the time taken for a cell to pass through a tight straight channel , or its average velocity as it transits through a series of small openings , or by monitoring a cell as it squeezes under hydrodynamic forces [4 , 6–9] . These devices can provide higher-throughput systems than conventional technologies such as atomic force microscopy and micropipette aspiration [5] and can be used as a comparative tool between different subpopulations of cells . They , however , often lack in-depth mechanical analysis ( ex . elasticity , viscosity ) and have little or no regard to the differences in intrinsic properties of these cells . To obtain a more detailed analysis of the cell mechanics with all its major underlying components , researchers have utilized modeling . Computational approaches to model cell deformation through microfluidic devices as complementary of experimental investigations are prominent for multiple reasons . Firstly , such modeling approaches give an insight into how cell components function under stress . Secondly , they can improve our understanding of the changes that occur during disease progression which , in turn , might uncover reasons for corresponding alterations occurring in cell mechanics [10 , 11] . Finally , computational models can be used as predictive tools for the experimental design . Much progress has been made during the last several years in the field of cell modeling . Mature human red blood cell ( RBC ) is perhaps among the simplest cells to model , lacking nucleus and internal cytoskeleton . Indeed , membrane models coupled to flow solvers were able to capture essential biomechanical properties of the RBCs in flow . A popular approach is to model the blood plasma with the Lattice-Boltzmann method ( LB ) , RBC membrane forces with finite element method ( FE ) , and RBC-fluid interactions using immersed boundary method ( IB ) [12–15] . Other models are based on the Finite Volume method [16] , moving particle semi-implicit method [17] , coarse-grained Molecular Dynamics [18 , 19] , and Stochastic Rotation Dynamics [20] . RBC models were successfully applied to simulate the flow in capillaries , bifurcations , and microfluidic devices [14 , 21–25] . Other cells are composed of , in addition to the cell membrane , a nucleus and internal cytoskeleton . We split the models for cells of this type into two groups . Models in the first group do not explicitly describe nucleus and cytoskeleton , and are based on the RBC membrane model with adjusted parameters set . Examples include recent study that employed such a model to describe cell movement in lateral displacement microfluidic device [26] . The results obtained from the modeling of flow of RBCs , white blood cells and circulating tumour cells ( CTCs ) were in a good agreement with experiments in microfluidic devices , which primarily rely on separation of cells by size , thus not requiring to reproduce the mechanical properties of cells very accurately in simulations . Some other models are based on the LB-IB method and were applied to CTC membrane deformations [27] and deformable platelet adhesion [28] . The second group of models takes into account sub-cellular components . A platelet model by Zhang et al . [29] , which consists of 73 thousand particles , was used to simulate dynamic properties of flowing platelets . Ujihara et . al . developed a model to study cell mechanics during tensile test [30]; and Liu et al . studied cytoskeleton deformation during needle injection [31] . The first group of models oversimplifies the cell and does not accurately describe its behavior in situations where the cell undergoes large deformations . At the same time , the computational complexity of the other existing models makes their use in the numerical investigation of cell transport in flow difficult . In this paper , we present a new model that is suitable for modeling of cells with wide range of viscoelastic properties and , at the same time , computationally efficient to be employed to large and complex flow domains . We developed a robust numerical framework employing Dissipative Particle Dynamics . To the best of our knowledge , this is the first study that models significant flow-induced deformations of cells using mesoscale particle-based model that explicitly takes into account cell membrane , nuclues , and internal cytoskeleton . To validate our model experimentally , we have chosen normal breast epithelial cells ( MCF-10A ) . Using micropipette aspiration experiment , we first probed cells’ elastic properties and used these data to set up parameters of the computational model . We then validated the model using data from microfluidic experiments , where MCF-10A cells were flown through three different microfluidic devices with a series of triangular shaped constrictions and transit velocity of these cells was measured . Using our computational model , we related transit velocity to cell elastic and viscous properties and examined the effect of cell model components such as cytoskeleton and nucleus on whole cell mechanics . We envision that the developed model will bring us closer to understanding the role of cell biomechanics in the broad spectrum of phenomena , such as the mechanical consequences of the structural changes that occur as cell evolves from healthy to diseased state .
We developed a new computational model mimicking structure of a cell . The model has three components: cell membrane , nuclear envelope , and internal cytoskeleton , as shown in Fig 1 ( b ) . The cell membrane model describes lipid bilayer together with underlying cortical actin layer . The nuclear membrane with lamins meshwork are defined by the nucleus envelope model . Both membrane models are described in Section Membrane model . The cytoskeleton is represented by a network of cross-linked filaments mimicking the topology of F-actin network , see Section Cytoskeleton model . For the sake of simplicity , the chromatin network inside the nucleus is described by the same model . The random cytoskeleton network formation and its integration with the membranes are explained in Section Cell model formation . To model fluid and fluid-structure interactions , we used the Dissipative Particle Dynamics method [38–40] , described in Section Dissipative Particle Dynamics . Although we explicitly model sub-cellular components , we follow the phenomenological approach , rather than reductionist method , to explain the properties of a system from the properties of its constituents [10] . This is due to our limited knowledge of the cell mechanics on the considered length scale . The aim of our cell model is to describe the correct mechanics of the whole cell , rather than to accurately reproduce the mechanical response of each individual molecular constituent at the microscale . Thus , although we , where it is possible , incorporate the knowledge of the microscale mechanics , we do not state that the model describing every particular constituent is correct . The important implication of this is that many model parameters can not be directly related to experimentally measured properties of individual cell constituents .
In this section , we provide the rational behind the choice of model parameters . We consider the model of medium size MCF-10A cell , which has diameter of D = 16μm with nuclear-cytoplasmic ratio of NC = 0 . 29 . We first focus on parameters defining elastic properties of the cell , while parameters affecting cell viscous properties are considered later . We explain the choice of parameters grouped by sub-cellular components . Cell membrane model describes both lipid bilayer and cortical meshwork . Although some parameters can be directly extracted from experimental data , such as cell and nucleus size , the experimental values of others are unknown . It is widely accepted though that the main contributor to the cell stiffness is the internal cytoskeleton and the impact of the membrane is not that significant . Thus , we model membrane as relatively soft material and we base its parameters on the values previously used for RBC membrane modeling ( persistence length , viscosity , bending stiffness ) . We represent cell surface with triangular mesh with 3500 vertices , so that the average bond equilibrium length is 0 . 518μm . The stiffness of the membrane model can be controlled by the value of the maximum elongation of the WLC link lmax . Considering experimentally observed values of elastic modulus for different cell lines , we found that lmax = 3 . 0μm is a reasonable choice . This rather high value allows us to model even very soft cells , increasing the range of model applications . The cytoskeleton model has more parameters with unknown values than other components . The procedure used to generate the network was explained previously in Section Cytoskeleton model . Parameters of the cytoskeleton model which regulate stiffness of filaments and cross-links , namely spring and angle constants ( see Eq 13 ) , have to be specified . Since every filament in the model represents a bunch of protein filaments of different origin , we cannot relate the model parameters describing them with the molecular-level data . Instead , we have chosen such values which result in the correct estimated elastic modulus for the whole cell model . We used relatively stiff filaments with high bending rigidity , while CLs are one order of magnitude softer which is usually the case for actin networks [63] . Variation of filament-CL torsion stiffness can be used as an additional way to control the overall stiffness of the entire cytoskeleton network . Parameters for the nucleus envelope , which in our model describes both lipid membrane and underlying lamins network , must be defined as well . To minimize the total number of parameters in the cell model , we chose nucleus envelope membrane parameters to be equal to corresponding parameters in the cell membrane model . The only parameter which we vary to control the nucleus stiffness is lmax , which was set to 1 . 2 . To define the values of model parameters mentioned above , we performed a series of simulations of micropipette aspiration experiments ( see Fig 2 ( b ) ) , which allowed us to estimate elastic properties of the whole cell model . Specifically , we tuned model parameters until the desired properties of the cell were obtained . We note here , that the resulting set of parameters is not uniquely defined . It is possible that some of the parameters can be chosen in a more rigorous way . However , since our goal is to match the properties of the entire cell , we are not concerned here with particular values of parameters describing individual constituents of the model . The resulting set of parameters is listed in Table 2 . To verify that this set of parameters gives acceptable results , we performed micropipette simulations with 16 independently generated cells with each cell rotated by 4 different angles . For every case , we examined the dependence between applied aspiration pressure δP and normalized aspirated length Ln . We compared mean Ln among all the simulations for each δP with the corresponding mean values obtained from the experiments and found a good agreement , as shown in Fig 2 ( c ) . The estimation of cell viscous properties usually requires a more complicated analysis than needed for elastic modulus estimation . For the current simulation study , cell viscous properties were determined using the time dependence of micropipette aspiration length at constant applied pressure . The longer it takes for a cell to reach plateau value of aspirated length , the higher is its viscosity . We perform this analysis using an extension of the Theret model proposed by Guevorkian et al . [73] which provides a procedure for more accurate viscosity ( η ) estimation than the original model . In order to control η , we use cutoff radius Rc and dissipative force coefficient γ parameters for interaction between filaments and other particles . We observe that the viscosity positively correlates with both parameters and the highest value of viscosity can be achieved by setting both Rc and γ to the highest possible values , see Fig 2 ( d ) . The proposed model allows to vary viscosity up to 3 times which might be useful for modeling cells of different types . For MCF-10A , experimentally observed values for η are in the range of 6 . 75 − 13 . 75mPa × s [74] . Thus , we chose Rc = 1 . 0 and γ = 65 to have η = 10mPa × s . With the model parameters defined by measuring the whole cell response in micropipette aspiration simulations , we perform validation of the model using experimental data for medium size MCF-10A cells traversal through microfluidic device II described in Section Microfluidic device experiments . In simulations , the geometry of microfluidic device is described by the signed distance function . The no-slip boundary conditions on the solid walls are implemented using the bounce-back reflection coupled with layers of frozen DPD particles inside the wall [75 , 76] . We set up the pressure difference , driving the flow in the working part of microfluidic device in simulations , by matching the average velocity ( 22 . 57mm/s ) of small 3 . 85μm beads which we added to the flow in experiments . We considered the same 16 cell models used in the micropipette aspiration simulations and modeled their passage through microfluidic device . In Fig 3 ( b ) we show snapshots of a typical cell squeezing between the obstacles in an experiment . Corresponding snapshots obtained in simulations are shown in Fig 3 ( c ) and 3 ( d ) . Comparison of mean cell velocity in simulations and experiments is shown in Fig 3 ( e ) . The results are in a good agreement . In the previous section , we validated the cell model using data from microfluidic experiments with medium size cells in device II . In this section , we will consider two other devices , I and III , which were used with small and large size cell populations , respectively . The microfluidic experiments in devices I and III were performed with the same pressure difference driving the flow as in device II . The difference between microfluidic devices I , II and III is the gap size between the obstacles which was chosen based on the average cell size ( see Fig 3 ( a ) ) . Specifically , the ratio between the gap size and the average cell diameter in device III is the same as in device II and is equal to 0 . 75 . Device I , on the contrary , has the ratio between the gap size and the average cell diameter equal to 0 . 83 . The experimental results ( Fig 3 ( e ) ) showed that the average velocity is approximately the same in devices II for medium size cells and device III for large cells . The average velocity of small size cells in device I was found to be much higher than for the medium size cells in device II . We created models for small and large size cells using the same procedure and parameters as used for the medium size cell model . The only difference was the diameters of the cell , which for small and large cell models were 12μm and 20μm , respectively . We employed 16 independently generated cell models for each cell size . The micropipette aspiration setup was used to estimate elastic properties of all cells . Applying the same procedure as for medium cells , we found dependence of normalized aspiration length Ln on pressure δP for small and large cell groups , see Fig 2 ( c ) . We do not observe impact of the cell model size on Ln and , thus , on the elastic properties of cells . Viscosity is also the same for all cell models . In Fig 3 ( e ) we show the results from microfluidics simulations performed with small and large size cell models . Similar to experiment , we observe that the average velocity of small size cells in device I is much larger than the velocity of medium and large size cells in devices II and III . In general , simulation results and experimental data are in a good agreement . Although the interactions of cells with the obstacles in microfluidic devices are complex , some simple considerations may help explain the observed results for three devices . One of the important parameters is the effective size of the opening between the adjacent obstacles in the devices . More specifically , we can define the effective size as a radius of the circle with the same area as the area of the opening . The values which we obtain for three devices then are 9 . 1 , 9 . 93 and 11 . 1 . Another important parameter , is the ratio between the effective size of the openings and the average cell diameter . The values we obtain for three devices are 0 . 76 , 0 . 62 and 0 . 55 . The effective size of the openings in device III is roughly 1 . 1 times larger than in device II . Therefore , we expect the average fluid velocity to be higher in device III comparing to device II . At the same time , the ratio between the opening size and the cell diameter is smaller in device III , and therefore the cells in device III have to squeeze through the opening which has smaller relative size , making it more difficult to pass . These two effects partially cancel each other , and the resulting average cell velocities are approximately the same in the two devices . If we consider devices I and II , the effective size of the opening is smaller in device I comparing to device II , so we expect the average fluid velocity to be smaller . At the same time , the ratio between the opening size and cell diameter is larger in device I , making it easier for cells to pass through device I . The last effect strongly dominates the reduction in the average fluid velocity and , hence , the velocity of the small cells in device I is several times higher . Interactions of cell with the obstacles in all three microfluidic devices make accurate modeling of mechanical properties of cell essential to obtain correct prediction of the average cell velocity . With the help of simulations , we now should be able to analyze and explain such interactions in more details . The cytoskeleton is believed to be one of the main contributors to cell stiffness . During progression of several diseases , changes in cytoskeleton structural properties may lead to significant softening of the cell . Such alternations include reduction of the filaments density as well as decrease in number of cross-links . For instance , it has been shown that to facilitate metastasis , cell undergoes a process called epithelial to mesenchymal transition where its cytoskeleton transforms from well-organized network into fragmented arrangement of filaments [77] . By altering cytoskeletal properties , the present model can accommodate for such processes . In this section we perform simulations to quantify how differences in cell internal structures , such as cytoskeleton and cross-links densities , affect cell mechanical properties . To simulate cytoskeleton density variations , we use filaments number density parameter , Nfil ( model parameters are listed in Table 2 ) . By varying its value between 1 . 25 and 4 , we obtain elastic modulus between 75Pa and 260Pa in micropippete aspiration simulations , demonstrating strong dependence of cell stiffness on the cytoskeletal density . This dependence appears to be similar for all cell sizes as shown in Fig 4 ( a ) –4 ( c ) . Cytoskeletal density also affects the velocity of cells in microfluidic simulations . As expected , the average cell velocity is lower with higher cytoskeleton density for all cell sizes . However , devices II and III appear to be more sensitive , demonstrating faster decrease of cell velocity , in comparison with device I , see Fig 4 ( a ) –4 ( c ) . Smaller ratio of effective opening size to average cell diameter in these devices results in larger relative cell deformations . This suggests that devices II and III are more suitable for studying the effect of cytoskeleton structure variation . Cross-links density , NCL , is another parameter which significantly affects the cytoskeleton properties . In our model , we can vary this parameter directly by changing the number of CLs particles during cytoskeleton network generation . We examined the impact of NCL on elastic modulus for medium size cell model and found that this parameter is as significant as cytoskeleton density . NCL can alter elastic modulus from 110Pa to almost 300Pa , see Fig 4 ( d ) . Dependence of average cell velocity on cross-link density for medium size cells in microfluidic device II obtained in simulations is also shown in the same Figure . The simulation results allow us to predict dependence of cell velocity in microfluidic device on its elastic modulus as shown in Fig 4 ( e ) for medium size cells in device II . Two sets of results are plotted corresponding to two alternative approaches we used to vary elastic modulus of cells , i . e . by changing the cytoskeleton density or by changing the cross-links density . The agreement between two sets of results provides additional support to one of the assumptions we use in our modeling approach , that not all of the structural constituents at the microscale should be resolved explicitly for the purpose of our studies , as long as whole cell properties are captured accurately . The nucleus deformability may be a critical factor in the cells’ ability to pass through small openings . There are two main determinants of nuclear stiffness—nuclear lamina meshwork and the chromatin network inside the nucleus . During the mesenchymal transition , nucleus often becomes bigger and softer . It is known that its softening is primarily due to the chromatin pattern alteration which is the hallmark of malignant nuclei [78] . Altered expression of lamins in a variety of human tumors is also often associated with malignant phenotypes , whether lamins level is upregulated or downregulated depends on the cancer type [79] . Despite current advances in live cell imaging and other biophysical techniques , it is still challenging to study the effect of each component on cells mechanics . In this section , we perform a computational study of the effect of morphological and structural changes of the nucleus . We focus on the medium size cells in microfluidic device II . First , we vary the size of the nucleus to evaluate its influence on cell elastic modulus as well as its velocity in microfluidic device . We varied NC ratio between 0 . 0 ( no nucleus ) and 0 . 7 for medium cells with elastic modulus of around 180Pa , see Fig 4 ( f ) . We have chosen a cell with relatively low elastic modulus because cells with very large nucleus tend to get stuck in microfluidic device . We observed that the nucleus size has a minor impact on results of micropipette aspiration simulations , indicating maximum increase of elastic modulus , E , only by 17% comparing to the cell model without nucleus . Results from microfluidics simulations , on the contrary , suggest that nucleus plays an important role in cell passage as shown in Fig 4 ( f ) . In particular , in the absence of nucleus ( NC = 0 . 0 ) , we observed approximately 2 . 5 times increase in average velocity in comparison to cells with nucleus of normal size ( NC = 0 . 29 ) . From the modeling prospective , it means that it is essential to explicitly model nucleus for the considered type of cells . To study the effect of chromatin concentration , we vary the filaments number density inside the nucleus ( N f i l n u c l ) while keeping the density of the cytoskeleton filaments constant . Our results suggest that the chromatin network has a significant impact on the cell stiffness , see Fig 4 ( g ) . For example , if the network is very sparse ( N f i l n u c l = 1 . 25 ) , the velocity increases significantly , exceeding the velocity of the cell model without nucleus . We note here , that for NC = 0 . 0 , the cell interior is completely filled by cytoskeleton with filament density Nfil . Next , we study the effect of the nuclear lamina on cell traversal in microfluidic device . The density of the nuclear lamina meshwork is modeled in the present study by parameter l m a x n u c l in the nucleus membrane model . By varying l m a x n u c l , we observe that by reducing stiffness of the envelope , we can again increase the average cell velocity significantly , see Fig 4 ( h ) . Our simulation results indicate that the impact of the nucleus on cell traversal through microfluidic device cannot be explained primarily by nucleus size , nuclear lamina or chromatin networks contributions , but rather all components may significantly alter cell dynamics . Cell viscosity is yet another property that may affect cell passage through microfluidic device . In general , dependence of cell velocity in microfluidic device on its viscosity can be non-trivial . In our previous studies with healthy and malaria infected RBCs in microfluidic device of similar design [7] , the device was found to be sensitive mostly to elastic properties of cells . Due to the specific interplay between the time needed for a cell to travel from one row of obstacles to the next and RBC characteristic relaxation time , the average cell velocity was almost independent of its viscosity . The distance between rows of obstacles and the driving pressure gradient were set in experiments so that RBCs did no have enough time to completely recover their shape during passage from one row of obstacles to the next . Cells with higher viscosity required longer time to deform and squeeze between pair of obstacles . However , these cells also required longer time to recover their shape , and therefore approached the next pair of obstacles with the shape making passage through the opening easier . In the devices used in the present study , cells recover their shape almost completely . Therefore , we do not expect similar effects to take place . Indeed , the results of simulations with medium size cells in microfluidic device II show roughly linear decrease of average cell velocity with increasing viscosity , as one would expect ( Fig 4 ( i ) ) . The cell viscosity was varied in simulations by changing cutoff radius Rc and dissipative force coefficient γ parameters for interaction between filaments and other particles . The obtained dependence shows that viscous properties of the cell can have comparable effect on its traversal to cell elastic properties . We developed a new eukaryotic cell model which takes into account cell membrane , cytoskeleton and nucleus . The non-tumorigenic breast epithelial cells ( MCF-10A ) were used in our studies . To estimate the viscoelastic properties of cells and to calibrate our computational model , we performed micropipette aspiration experiments . The model was then validated using data from three microfluidic experiments with devices designed to take into account size variation in MCF-10A cell population . We note here , that the chosen set of model parameters may not be unique and better agreement particularly for small cells in microfluidic device I may be achieved given that there are many parameters in the proposed model . However , we want to emphasize that we did not use any data from microfluidic experiments to set cell model parameters . Taking into account the interplay of average flow velocity and cell interactions with obstacles in microfluidic devices , the fact that the model can predict ( even not perfectly but still within the experimental error bars ) cell velocities is quite remarkable in our opinion . Additional validation and benchmark tests are necessary to tune the model more carefully . Using the model , we probed contributions of sub-cellular components to whole cell mechanics in micropipette aspiration and microfluidics experiments . We obtained that the main contributor to cell stiffness is its cytoskeleton . This finding is in agreement with previous experimental studies [80–82] . Our model showed that both filament and cross-links concentrations play equally important role in defining whole cell mechanics , dominating over the effects due to variation of cell nucleus properties . Simulation results indicate that it is important to model nucleus explicitly in microfluidics simulations . Each of considered nucleus properties , namely nucleus size , stiffness of nuclear lamina and chromatin network , can significantly affect deformability of the cell . The viscous properties of the cell can have comparable effect to cell elastic properties on its traversal through microfluidic device . We believe that the new model will allow to study in silico numerous problems in the context of cell biomechanics in flows in complex domains , such as capillary networks and microfluidic devices . Our ongoing work indicates that the proposed cell model parametrization has the flexibility to be used in simulations of various cell types , including cancer cells with different mechanical properties . With further development , the present model with explicit description of sub-cellular components may be used to study different alterations in cell mechanics caused by diseases or functional changes . | Predictive simulations of cell flow in microfluidic devices and capillary networks may help to quantify the impact of different cell components on its behavior . Cells have complex mechanical properties and can undergo significant deformations , requiring detailed models to give an insight into the cell rheology . We developed computational model for simulations of cells with nucleus and cytoskeleton in flows in complex domains such as capillary networks and microfluidic devices . We validated the model using experimental data and used it to quantify the effects of cell components on its behavior . We envision that the proposed model will allow to study in silico numerous problems related to the cell biomechanics in flows . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"and",
"discussion"
] | [
"engineering",
"and",
"technology",
"biomechanics",
"simulation",
"and",
"modeling",
"viscosity",
"materials",
"science",
"materials",
"physics",
"cell",
"nucleus",
"cell",
"mechanics",
"cellular",
"structures",
"and",
"organelles",
"cytoskeleton",
"physical",
"chemistry",
"microfluidics",
"fluidics",
"research",
"and",
"analysis",
"methods",
"chemical",
"properties",
"nuclear",
"membrane",
"chemistry",
"biophysics",
"cell",
"membranes",
"physics",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] | 2017 | Probing eukaryotic cell mechanics via mesoscopic simulations |
Reduced supply of the amino acid methionine increases longevity across species through an as yet elusive mechanism . Here , we report that methionine restriction ( MetR ) extends yeast chronological lifespan in an autophagy-dependent manner . Single deletion of several genes essential for autophagy ( ATG5 , ATG7 or ATG8 ) fully abolished the longevity-enhancing capacity of MetR . While pharmacological or genetic inhibition of TOR1 increased lifespan in methionine-prototroph yeast , TOR1 suppression failed to extend the longevity of methionine-restricted yeast cells . Notably , vacuole-acidity was specifically enhanced by MetR , a phenotype that essentially required autophagy . Overexpression of vacuolar ATPase components ( Vma1p or Vph2p ) suffices to increase chronological lifespan of methionine-prototrophic yeast . In contrast , lifespan extension upon MetR was prevented by inhibition of vacuolar acidity upon disruption of the vacuolar ATPase . In conclusion , autophagy promotes lifespan extension upon MetR and requires the subsequent stimulation of vacuolar acidification , while it is epistatic to the equally autophagy-dependent anti-aging pathway triggered by TOR1 inhibition or deletion .
Methionine restriction ( MetR ) has been long known to enhance lifespan in various organisms , including mammals [1] , [2] . Nevertheless , the mechanisms underlying this phenomenon are poorly understood . Previous studies have mainly focused on MetR-induced alterations of the function and composition of respiratory chain complexes in mitochondria , although no clear cause-effect relationship between these effects and the beneficial impact on longevity could be established [3]–[5] . Given that MetR represents a regime that limits availability of an amino acid , we wondered if the resulting longevity effect might include the involvement of autophagy [6] , which is known to play a crucial role in cells that are stressed by damage or limited nutrient supply [7] . Only recently , autophagy has been shown to play an important role for lifespan extension by treatment with spermidine , rapamycin , or resveratrol , as well as by depletion of the p53 ortholog from Caenorhabditis elegans , the inhibition of IGF signaling , and the overexpression of sirtuin [8]–[12] . The autophagic process depends on the vacuolar proteolytic activity , which is determined by vacuolar acidification [13] , [14] . Based on these premises , we decided to analyze whether macroautophagy ( hereafter referred to as autophagy ) is also induced under conditions of MetR and if autophagy induction contributes to MetR-induced lifespan extension . In addition , we analyzed the extent of vacuolar acidification , which has been recently shown to be crucial for lifespan extension in replicatively aging cells [15] . To tackle these questions , we decided to use baker's yeast ( Saccharomyces cerevisiae ) as a model system , since ( i ) it constitutes a well-established ( chronological ) aging model [16]; ( ii ) autophagy was discovered and largely uncovered in this model [17]–[19]; and ( iii ) MetR can be reliably controlled by virtue of media supplementation and the deletion of genes involved in its biosynthesis [20] . Here , we report that MetR causes an increase in chronological lifespan ( CLS ) that depends on the enhanced vacuole acidification that follows autophagy stimulation .
To determine whether MetR influences yeast CLS , three different strains were used: ( i ) a MET+ strain that is fully competent in synthesizing methionine , ( ii ) a knockout strain deleted in met15 that suffers only a moderate defect in methionine biosynthesis due to an intact salvage pathway mediated via O-acetyl-homoserine ( the reaction product of Met2p ) , and ( iii ) a met2 deletion strain fully devoid of a de novo methionine synthesis and hence strictly dependent on externally supplied methionine [20] . These three strains are otherwise isogenic ( Table S1 ) . The methionine-prototroph strain ( MET+ ) exhibited a rather short CLS of about 10 to 15 days , while the two methionine-auxotroph strains with limited ( Δmet15 ) and no ( Δmet2 ) endogenous methionine biosynthesis displayed enhanced lifespans of up to 25 days ( Figure 1A ) . Of note , these chronological aging experiments were performed in synthetic complete medium supplemented with 30 mg/l methionine . Using this medium , all three strains showed equivalent cell counts during aging experiments ( Figure S1A ) and cell cycle arrest in G0/G1 was comparable ( Figure S1B ) , therefore excluding potential artifacts secondary to differential growth rates and nutrient consumption . Furthermore , chronological aging experiments of MET+ , Δmet15 and Δmet2 revealed only marginal differences in external media pH ( Figure S1C ) , thus showing that external pH effects do not play a major role in this scenario [21] . Accordingly , it has been recently shown that lifespan-enhancing conditions do not necessarily correlate with changes in pH [22] . We next determined the influence of external methionine availability on the MET+ , Δmet15 , and Δmet2 strains by using media supplemented with varying methionine concentrations . Higher levels of supplemented methionine led to decreased survival in Δmet15 and Δmet2 strains ( Figure 1B and Figure S1D ) whereas a reduction of methionine led to improved longevity - especially of long-term survival - of the semi-auxotrophic strain ( Δmet15 , Figure 1B , cell-counts Figure S1E ) , but not of the prototrophic strain ( MET+ ) ( Figure S1F , cell-counts Figure S1G ) . Of note , lower amounts of cysteine ( a downstream product of methionine biosynthesis ) did not increase CLS of Δmet15 whereas high amounts shorten CLS , potentially via formation of methionine by transsulfuration ( Figure S1H ) [20] . Also note that reduced methionine levels were not used in the case of the auxotrophic Δmet2 strain because cell counts , grown in the presence of 3 mg/l , of this strain are ten times lower compared to 30 mg/l standard conditions ( Figure S1I ) . Thus , to verify effects of low levels of methionine on the fully auxotrophic Δmet2 strain during chronological aging , cells were grown to stationary phase ( 24 hours ) in media with excess methionine ( 30 mg/l ) to support normal growth , and then transferred to media with lower methionine concentrations ( Figure 1C and D ) . Δmet2 cell cultures transferred to media with high methionine concentrations exhibited accelerated aging , while lowering methionine concentrations increased longevity . Optimal cell survival was reached with around 6 mg/l externally supplied methionine ( Figure 1C and D ) . Thus MetR during chronological aging can only be achieved by combining both , the deletion of specific genes involved in its biosynthesis and external methionine supplementation . Moreover , MetR resulted in reduced phosphatidylserine externalization ( a typical sign of apoptosis ) and improved plasma membrane integrity ( which is disrupted in necrosis ) , as determined by AnnexinV/PI-costaining ( Figure S2A ) . The levels of reactive oxygen species ( ROS ) , which are suspected mediators of cellular aging , were determined by monitoring the conversion of dihydroethidium ( DHE ) to fluorescent ethidium ( Eth ) , as driven by superoxide anion radicals . ROS levels were clearly diminished in Δmet15 and Δmet2 strains ( Figure S2B ) . Intriguingly , one genetic difference ( beside the functionality of their LYS2 gene-product and a different mating type ) of the frequently used wild type strains of the EUROSCARF strain collection , BY4741 and BY4742 , also affects their ability to produce methionine . The long-lived BY4741 strain harbors a deletion of MET15 ( and was also used in the above described experiments ) whereas the short-lived BY4742 strain is methionine-prototroph . Deletion of MET15 from BY4742 reestablished a long-lived phenotype in chronological aging experiments , indicating that the short life expectancy of this strain is indeed due to its ability to synthesize methionine ( Figure 1E ) . We conclude that the amount of methionine availability – as determined by de novo synthesis or external supply – dramatically influences the survival of chronologically aging yeast and protects against apoptosis/necrosis . Since autophagy might be one of the major pathways responsible for lifespan extension in various organisms and under diverse circumstances [10] , [23] we determined the rate of autophagy within the cell cultures . During the first days of chronological aging , alkaline phosphatase ( ALP ) activity ( which measures the activity of cytosolic ALP which is delivered to the vacuole exclusively via autophagy ) was significantly higher in Δmet15 and Δmet2 strains compared to the MET+ strain ( Figure 2A ) . Accordingly , vacuolar processing of GFP-tagged Atg8p , which takes place under autophagic conditions for recycling of the protein , was elevated upon MetR , as visualized by immunoblotting ( Figure 2B ) . Furthermore , the localization of GFP-tagged Atg8p , a protein essential for the autophagic process , which is normally evenly distributed in the cytoplasm ( see MET+ strain ) , showed punctuate and/or vacuolar localization in Δmet15 and Δmet2 strains during chronological aging ( shown for day 2 in Figure 2C and D ) . This reflects enhanced autophagosome formation and Atg8p-delivery to the vacuole . Of note , differences in the number of puncta per cell compared to the number of positive vacuoles per cell possibly reflect dynamics of this process and are subject to changes over time . To exclude effects of distinct Atg8p levels deriving from possible differences in synthesis capabilities of the different strain backgrounds , we performed overexpression studies of Atg8p in MET+ . No differences in CLS were observed ( Figure S3A ) . Next , we determined the autophagic flux ( ALP activity ) under different levels of MetR by transferring stationary Δmet2 cell cultures grown in excess of methionine to media supplemented with varying methionine concentrations that showed beneficial effects on longevity ( see above and Figure 1C and D ) . A clear inverse correlation of methionine concentration and autophagy level was visible , reaching a nearly ten-fold difference in early onset of autophagy between a high methionine concentration ( 30 mg/l ) and cells exposed to low concentrations ( 3 to 6 mg/l ) of methionine ( Figure 2E ) . In addition , the MET+ strain showed no altered ALP activity when grown in the presence of 3 mg/l methionine ( Figure 2F ) , as opposed to the met15 deletion strain where ALP activity was strongly up-regulated ( Figure S3B ) . Accordingly , the Δmet2 strain grown in the presence of high levels of external methionine showed decreased autophagy ( Figure S3C ) . We conclude that methionine restriction specifically enhances autophagy during yeast chronological aging . To verify whether autophagy truly impacts rather than only correlating with lifespan extension upon MetR , we subjected several strains harboring single deletions of genes necessary for MetR-triggered autophagy . Deletion of ATG5 , ATG7 or ATG8 , abolished the gain in longevity that was normally conferred by MetR . In both , the Δmet15 and the Δmet2 strains , lifespan was drastically shortened upon ATG gene deletions reaching comparable or even lower survival levels than those of the corresponding MET+ strain ( Figure 3A and S4A , B ) . Of note , an additional RAS2 deletion , a well-established longevity-mediating mutation , in an autophagy-deficient Δmet2 strain ( Δmet2 Δatg5 ) led to increased longevity , clearly indicating that autophagy-deficient strains are not per se unable to survive longer but that other ( non-autophagic ) pro-survival mechanisms are functional ( Fig . S4C ) . Importantly , survival of the MET+ strain deleted for ATG5 , ATG7 or ATG8 , did not alter CLS during the first three days . Only after day 3 , when autophagy started to increase ( Figure 2A and B ) and thus seemed to become a physiological need , CLS was shortened ( Figure S4D ) . Next , we determined whether rapamycin , an established pharmacological inhibitor of TOR and inducer of autophagy , could enhance lifespan of the MET+ strain . Rapamycin treatment indeed extended the longevity of the MET+ strain ( Figure 3B ) . In accordance , the autophagy rate under these conditions was strongly induced ( 2 to 5 times ) as measured via ALP activity ( Figure 3C ) . To minimize its effects on growth behavior , rapamycin was added during mid-log phase ( 8 hours after inoculation of the main culture ) . Therefore , the positive effect on chronological survival is neither mediated by possible growth-delays ( Figure S4E ) [8] nor by an enhancement of respiration in the logarithmic growth phase ( Figure S4F ) [24] . Similar to the pharmacologically mediated inhibition of the TOR pathway ( by rapamycin ) , genetic ablation of TOR1 , the initiator-kinase of the autophagy-repressive TOR pathway , resulted in increased CLS of MET+ cells ( Figure 3D ) . The additional deletion of ATG genes essential for autophagy ( ATG5 , ATG7 , or ATG8 ) prevented the positive effects of TOR1 deletion in the methionine prototrophic strain ( Figure 3E and Figure S4G and H ) , as was already shown for rapamycin mediated lonegvity [8] . We conclude that autophagy is crucial for MetR-induced longevity . Because the TOR pathway is one of the major sensors for ( external ) amino acid availability , we asked whether lifespan extension under MetR conditions could be further enhanced by TOR inhibition . For this purpose , we deleted TOR1 in both the methionine-auxotrophic ( Δmet2 ) and the semi-auxotrophic strains ( Δmet15 ) . Genetic ablation of the TOR pathway had no positive influence on survival ( Figure 4A and B ) . This irresponsiveness seems to be independent from ROS generation since addition of low doses of glutathione did not affect CLS of Δmet2 and Δmet15 strains ( Figure S5A and B ) . Intriguingly , external starvation for methionine in the Δmet2Δtor1 strain led to a small increase of autophagy on day 1 , which became more pronounced with ongoing age ( Figure S5C ) . This possibly shows that initial autophagy induction upon MetR is strongly dependent on TOR1 whereas its maintenance might be additionally supported by TOR1-independent mechanism ( s ) . Pharmacological inhibition of the TOR pathway ( via rapamycin ) under the very same conditions as described for the MET+ strain also failed to increase longevity of the Δmet2 strain and had only rather small positive effects on the Δmet15 strain ( Figure 4C and D ) . Of note , a stronger increase in cell count after day 1 was observed for all strains treated with rapamycin , irrespective of its impact on CLS ( Figure S4E and S5D , E ) . Moreover , rapamycin treatment only marginally increased autophagy rates at day 1 , and had no effects at day 7 ( measured by means of ALP activity ) during Δmet15 and Δmet2 chronological aging ( Figure 4E ) . This was presumably the case because autophagy was already strongly stimulated . In contrast , rapamycin enhanced autophagy rates in the methionine-prototroph strain ( MET+ ) to factors of up to 5 , on day 1 , and 2 to 3 , on day 7 ( see above and compare Figure 3C and 4E ) . We conclude that TOR inhibition and MetR are , at least partly , part of the same anti-aging pathway . The downstream target of autophagy is the vacuole . Thus , we mused if MetR-induced autophagy could enhance the degree of acidic vacuoles within the cell population . For this purpose , we stained chronologically aged MET+ or Δmet2 cells with quinacrine , the most widely used and highly specific stain for acidic cell compartments and counted for cells where only the vacuole was stained . The Δmet2 strain showed a 20% increase in cells harboring acidic vacuoles during chronological aging compared to the MET+ strain ( Figure 5A and B ) . This increase was autophagy-dependent since it was completely abolished in a Δmet2 strain lacking ATG5 and thus autophagy-deficient ( Figure 5A and B ) . Of note , cells showing a very bright quinacrine staining ( older cells from MET+ and Δmet2/Δatg5 , Figure 5A ) represent cells with acidic cytoplasm , which harbor almost no intact vacuoles as demonstrated with quinacrine-stained cells expressing a chromosomal mCherry-tagged version of the vacuolar membrane-located Vph1p ( Figure S6A ) . To further show a direct regulation of vacuolar acidity upon MetR , through autophagy , we shifted Δmet2 or Δmet2/Δatg5 cells to media with different methionine concentrations . The amount of cells with only acidic vacuoles was strictly dependent on the amount of supplemented methionine: ∼80% when shifted to media lacking methionine , ∼55% on 3 mg/l , and ∼35% on 30 mg/l methionine ( Figure 5C ) . This dependency was blocked by an additional ATG5 deletion , which causally links vacuolar acidification to MetR-induced autophagy ( Figure 5C ) . Moreover , the MET+ strain grown in the presence of rapamycin , showed an increased proportion of cells harboring acidic vacuoles during aging ( Figure S6B ) , in line with the beneficial effects of this pharmacological intervention on survival and increased autophagy ( Figure 3B and C ) . Of note , starting at day 3 we could observe that the MET+ strain showed distinctly more vacuolar cargo compared to Δmet2 , which is probably due to limited clearance . We conclude that autophagy is sufficient to significantly enhance the proportion of cells harboring acidic vacuoles during yeast chronological aging . Interestingly , an enhanced proportion of cells bearing an acidic vacuole has been recently shown to be crucial for improved replicative longevity by overexpressing VMA1 or VPH2 and thus increasing vacuolar acidity [15] . To explore whether such causal connection is also present between the enhancement of acidic vacuoles and the extended lifespan by MetR-induced autophagy , we overexpressed Vma1p , a part of the vacuolar ATPase ( v-ATPase ) or Vph2p , essential for v-ATPase assembly [25]–[27] in the MET+ strain . Overexpression of both proteins led to increased CLS ( Figure 5D ) . Moreover a Δmet2 strain deleted for VPH2 showed diminished survival during chronological aging resembling more closely the MET+ strain ( Figure S6C ) . It should be noted that a deletion leading to no detectable acidic vacuoles via quinacrine staining ( Figure S6D ) , causes pleiotropic effects [28] and thus results must be interpreted cautiously . Still and supporting our knockout results , overexpression of Vph2p or Vma1p in the methionine-auxotrophic Δmet2 strain did not improve CLS ( Figure S6E ) . These results support a model , in which MetR-induced autophagy regulates vacuolar acidity which in turn promotes longevity ( Figure 6 ) . This places MetR-induced autophagy at the hub of vacuolar acidification and its positive effects on chronological survival , which are at least partly responsible for the positive effects observed on longevity through MetR ( starvation for methionine ) .
Using S . cerevisiae , the key model organism in which autophagy was first functionally described and genetically dissected [17]–[19] , we demonstrate that methionine restriction ( MetR ) , unlike restriction in other amino acids such as leucine [6] , [29] , promotes clonogenic survival during chronological aging . MetR inhibits the ROS overproduction , as well as the aging-associated mortality by both apoptosis and necrosis . MetR shares analogies to limitations in elemental nutrients such as phosphor and sulfate , which induce a specific cell cycle arrest [30] . However , we could not find any signs of cell cycle blockade and our experiments were performed under conditions that fully supported growth to stationary phase in both methionine auxotrophic ( Δmet2 ) and semi-auxotrophic ( Δmet15 ) strains . Although there were no discernible cell cycle effects , MetR-induced lifespan extension correlated with enhanced autophagy , and the positive effect of MetR on longevity was lost when essential ATG genes were deleted . Accordingly , pharmacological or genetic inhibition of the TOR-pathway ( and thus autophagy induction ) enhanced CLS of a methionine-prototroph strain ( MET+ ) but failed to do so in the methionine-auxotroph strains Δmet2 and Δmet15 . This epistatic analysis fully validates the concept that the beneficial effects of MetR on longevity are mediated by autophagy . Only recently , two new mechanisms for autophagy regulation were described: ( i ) a methionine-related one , involving the protein phosphatase 2a ( PP2A ) , high levels of which were shown to down-regulate autophagy in dependence of methionine availability [31] and ( ii ) a methionine-independent mechanism , where high acetate levels block autophagy induction [32] . In our MetR setup , the first mechanism does not seem to play a major role since the MET+ strain lacking PPM1 ( the methyltransferase of PP2A ) , did not lead to better chronological survival and deletions in PPH21 or PPH22 ( catalytic subunits of the PP2A complex ) had only small positive effects ( Figure S7A ) . Instead , acetate levels in the media of the MET+ strain were about 80% higher compared to those of Δmet15 and Δmet2 strains on day 1 , reaching comparable ( Δmet2 ) or lower levels ( Δmet15 ) on day 2 ( Figure S7B ) of chronological aging . Given the complex regulatory network PP2A is involved in and the metabolic and regulatory implications high acetate levels potentially lead to , contributions to autophagy induction are likely dependent on growth conditions and molecular fine-tuning . However , both pathways potentially influence the TOR pathway or its downstream targets thus supporting our epistasis analysis of MetR and TOR inhibition . Future work will be needed to decipher the specific contributions of these pathways/metabolites under different longevity-mediating regimens . Altogether , methionine as an ubiquitous factor within cell metabolism may impact aging through several mechanisms that share or are independent from the herein described , for instance , a recent study suggests that methionine regulates homeostasis through modulation of tRNA thiolation and thus translation capacity [33] . Furthermore , we could clearly demonstrate that the proportion of cells displaying an acidic vacuole within a population is significantly enhanced via MetR inflicted either by genetic deletion of met2 and external methionine availability , or rapamycin treatment of MET+ . Additionally , we show that this enhancement is strictly dependent on functional autophagy ( as shown by an ATG5 deletion ) . In line , increasing v-ATPase activity ( by overexpression of Vph2p or Vma1p ) , a process already shown to increase vacuolar acidity [15] , is sufficient to increase CLS in a methionine-prototrophic strain . Conclusively , deletion of VPH2 and thus disruption of v-ATPase activity reverses the positive effects of MetR on CLS . Intriguingly , it has been recently demonstrated that an increase in vacuolar pH , specifically during early age , negatively influences the replicative lifespan of yeast [15] . In the same line , a recently published screen for chemical compounds extending CLS in Schizosaccharomyces pombe identified , among others , vacuolar acidification as a key process [34] . Additionally , Hughes and Gottschling showed that decreased vacuolar pH positively impacts mitochondrial function [15] . Interestingly , others have determined that autophagy is required to maintain respiration proficiency under caloric restriction conditions in galactose media [29] , highlighting a protective role of autophagy , especially for mitochondrial function . In the frame of these observations , the decrease in ROS production during MetR may suggest a mechanistic structure that couples MetR-induced autophagy and vacuolar acidification to mitochondrial function . In studies relating longevity to autophagy , doubts can be raised on the interpretation of the negative effects of genetic autophagy defects because the deletion of ATG genes may perturb cell survival per se [35] . However , we find that in our experimental setup , cells deleted for ATG genes , in fact , display normal growth rates and survival during the first days of chronological aging . Furthermore , we clearly demonstrate that an additional RAS2 deletion in an autophagy-deficient Δmet2 strain enhances the mutant's CLS . Nevertheless , the positive effects of a RAS2 deletion during chronological aging under MetR conditions seem to be limited since longevity is only marginally enhanced in a Δmet2 strain . This points towards the concept that autophagy is a process that significantly contributes to enhanced longevity upon RAS2 deletion . Our work clarifies two further , thus far unexplained issues that are of great importance to researchers working on ( yeast ) aging: First , we demonstrate that the heterogeneity in CLS of the EUROSCARF wild-type strains BY4741 and BY4742 can be explained by strain-dependent differences in methionine biosynthesis . Second , we show that treatment with rapamycin or deletion of TOR1 almost only increases the longevity of strains that are methionine-prototrophic . Intriguingly , despite one study that could show enhanced CLS by tor1 deletion in BY4741 [36] , previously published results on extension of CLS via TOR1 inhibition were performed in strains that are prototrophic for methionine [8] , [24] , [37] . We can also show that early phases of MetR-mediated autophagy are largely dependent on TOR1 or impact the same downstream targets . At the same time , there seem to be additional TOR1-independent mechanisms of MetR-induced autophagy later on during CLS , which will need to be addressed in future studies . Taken together , we show that autophagy-mediated vacuolar acidification is essential for the anti-aging effects of MetR , one of the rare lifespan-extending scenarios that is conserved across species .
Experiments were carried out in strains using the EUROSCARF strain collection as basis and are listed in Table S1 . In brief: BY4741 ( MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 ) was used as Δmet15 strain . MET+ was constructed by crossing BY4741 with BY4742 and tetrad selection for being methionine prototroph but otherwise isogenic to BY4741 . Accordingly Δmet2 was generated by crossing BY4741 Δmet2::kanMX ( EUROSCARF ) with BY4742 and tetrad selection for tetra-type ( all four spores are methionine auxotroph ) and further selection for geneticin resistance ( only conferred by Δmet2::kanMX ) . All other deletion strains and chromosomal GFP tagging in these backgrounds were carried out by classical homologous recombination using the pUG and pYM vector systems [38]–[40] , and controlled by PCR . Transformation was done using the lithium acetate method [41] . Notably , at least two different clones were tested for any experiment with these newly transformed strains to rule out clonogenic variation of the observed effects . To generate strains with chromosomal mCherry tags at the C-terminus of the vacuolar membrane protein Vph1 plasmid pFA6a3mcherry-natNT2 or pFA6a3mcherry-hphNT1was used for PCR amplification . After transformation and selection , correct integration was tested via PCR and fluorescent microscopy . For overexpression studies of VPH2 , VMA1 , and ATG8 genes were amplified by PCR and inserted into the pESC-HIS vector ( Stratagene ) . Resulting plasmids were verified by sequencing by eurofins/MWG , transformed via the lithium acetate method and subsequently expression was verified by western blot analysis . All strains were grown on SC medium containing 0 . 17% yeast nitrogen base ( BD Diagnostics; without ammonium sulfate and amino acids ) , 0 . 5% ( NH4 ) 2SO4 , 30 mg/L of all amino acids ( aa ) ( except 80 mg/liter histidine and 200 mg/liter leucine ) , 30 mg/L adenine , and 320 mg/L uracil with 2% glucose ( SCD ) . All amino acids were purchased from Serva ( research grade , ≥98 . 5% ) . For experiments with varying methionine or cysteine concentrations , methionine and cysteine were added at given concentrations . For overexpression with the pESC-HIS system strains were grown in the absence of histidine . When needed , glutathione was added at given concentrations at the time of inoculation . Survival plating was done on YPD agar plates ( 2% peptone , 1% yeast extract , 2% glucose , and 2% agar ) and incubated for 2 to 3 days at 28°C . For chronological aging experiments cells were inoculated to 5×105 cells , or alternatively to an OD600 of 0 . 05 , and grown for the indicated time period at 28°C in SC media . If not stated otherwise standard concentration of methionine ( 30 mg/l ) was used throughout the experiments . Shift aging experiments with the Δmet2 strain were inoculated accordingly grown for 24 hours in excess of methionine ( 30 mg/l ) and subsequently shifted to fresh SC media with indicated amounts of methionine . Of note For overexpression studies strains carrying VPH2 , VMA1 , or ATG8 on a pESC-HIS vector were grown in SCD for 6 hours and subsequently shifted into minimal media containing 0 . 5% galactose and 1 . 5% glucose , or 2% galactose ( ATG8 ) , for induction of expression . At the indicated time points cell survival was determined by clonogenictiy: Cell cultures were counted with a CASY cell counter ( Schärfe System ) and 500 cells were plated on YPD agar plates . Subsequently colony forming units were counted and values were normalized to survival at day one . Alternatively , cell death was measured via propidium iodide staining and subsequent flow cytometry analysis ( BD FACSAria ) . Representative aging analyses are shown with at least three independent cultures aged at the same time . All aging analyses were performed at least twice in total with similar outcome . Experiments involving treatment with rapamycin ( LC laboratories ) were performed as described above . To circumvent growth effects described in other publications rapamycin was added eight hours after inoculation and to lower amounts . Dihydroethidium ( DHE; working concentration: 2 . 5 µg/ml in PBS; ∼1×106 cells; incubation time 5 to 10 min at RT ) staining ( ROS production ) and Annexin V/propidium iodide costaining ( apoptosis/necrosis marker ) were performed and quantified by using a fluorescent plate reader ( Tecan , GeniusPRO ) or by flow cytometry ( BD FACSAria ) as previously described [42] . 30 , 000 cells per sample were evaluated using BD FACSDiva software . Measurement of growth media pH was performed at the indicated time points using a pH-Meter ( Metrohm ) . Oxygen consumption was determined eight hours after addition of rapamycin using an Oxygraph ( Clark-type oxygen electrode connected to an ISO2 recorder; World Precision Instruments ) and subsequent data processing with LabChart ( ADInstruments ) . Quinacrine ( Sigma ) was used to stain for acidic vacuoles following standard protocols [15] . Briefly , ∼2×106 cells were harvested , washed with YPD containing HEPES buffer ( 100 mM , pH7 . 6 ) and then collected and re-susupended in fresh YPD-HEPES containing 200 µM quinacrine After 10 min incubation at 30°C cells were put on ice and washed three times with ice-cold HEPES buffer containing 2% glucose an finally resuspended in the same buffer . All samples were kept on ice until they were viewed under the microscope within 1 hour since sample taking . DNA content was measured as described previously [43] . Briefly , ∼1×107 cells were harvested , re-suspended in cold water and fixed with ice cold ethanol . After ∼16 h cells were harvested , re-suspended in sodium-citrate buffer ( 50 mM , pH7 . 4 ) and sonicated . After treatment with RNase and proteinase K , cells were stained overnight with propidium iodide ( 8 µg/ml ) , and analyzed by flow cytometry ( BD LSRFortessa ) . 30 , 000 cells per sample were evaluated using BD FACSDiva software . Microscopy of Quincarine stained cells as well as GFP-Atg8p expressing cells was performed with a Zeiss Axioskop microscope using a Zeiss Plan-Neofluar objective lens with 63× magnification and 1 . 25 numerical aperture or 40× magnification and 2 . 0 numerical aperture in oil ( using Zeiss Immersol ) at room temperature . Fluorescence microscopic sample images were taken with a Diagnostic Instruments camera ( Model: SPOT 9 . 0 Monochrome-6 ) , acquired and processed ( coloring ) using the Metamorph software ( version 6 . 2r4 , Universal Imaging Corp . ) For creating a statistical analysis 330–600 cells of each GFP-Atg8p expressing strain were evaluated from two independent samples ( Figure 3B ) . For statistical analysis of Quinacrine stained cells , >1000 cells of each strain from 3 to 5 independent samples at each time point were evaluated ( Figure 5B and Figure S6B ) or >500 cells of each condition from two independent samples ( Figure 5C ) . Autophagy was monitored by alkaline phosphatase ( ALP ) activity [44] . Strains were transformed with and selected for stable insertion of pTN9 HindIII fragment ( confirmed by PCR ) . Briefly 1–5×10E7 cells were collected ( and kept on ice from that moment on ) , washed , resuspended in assay buffer ( Tris-HCl , 250 mM; pH = 9; 10 mM magnesium phosphate; 10 µM zinc sulfate ) , disrupted with glass beads , and centrifuged . Protein concentration was determined in the supernatant via a Bradford assay ( BioRad ) following standard protocols and subsequently 1 µg of total protein extract was subjected to the ALP assay . Extracts were incubated with α-naphtyl phosphate ( 55 mM ) for 20 min at 30°C and stopped with 2 M glycine-NaOH ( pH = 11 ) . To correct for intrinsic ( background ) ALP activity , the corresponding strains without pTN9 were simultaneously processed and ALP activity was subtracted . Alternatively , strains were transformed with and selected for pCC5 ( a plasmid carrying the cytoplasmic PHO8Δ60 [45] ) . Relative fluorescence units ( RFU ) were determined by using a fluorescence reader ( Tecan , GeniusPRO ) and applying the same manual gain throughout a series of measurements belonging together . For each transformed strain , two clones were tested . Preparation of cell extracts and immunoblotting were performed as described [46] . Blots were probed with monoclonal mouse anti-GFP antibody ( Roche , Cat . No:11814460001 ) , rabbit polyclonal antibodies against glyceraldehyde-3-phosphate dehydrogenase ( gift from Günther Daum ) and the respective peroxidase-conjugated affinity-purified secondary antibody ( Anti-Mouse IgG-Peroxidase antibody A9044 and Anti-Rabbit IgG-Peroxidase antibody A0545 , Sigma ) . For detection the ECL system was used ( Amersham ) . Error bars ( ± SEM ) are shown for independent experiments/samples . In cases when experiments were performed in parallel , a common overnight culture ( ONC ) for each strain was used . The number of independent data points ( n ) is indicated in the figure legends of the corresponding graphs . Significances were calculated using students t-test ( one-tailed , unpaired ) . For aging experiments , a two-factor ANOVA with strain and time as independent factors was applied and corrected by the Bonferroni post hoc test . Significances: *p<0 . 05 , **p<0 . 01 , ***p<0 . 001 . | Health- or lifespan-prolonging regimes would be beneficial at both the individual and the social level . Nevertheless , up to date only very few experimental settings have been proven to promote longevity in mammals . Among them is the reduction of food intake ( caloric restriction ) or the pharmacological administration of caloric restriction mimetics like rapamycin . The latter one , however , is accompanied by not yet fully estimated and undesirable side effects . In contrast , the limitation of one specific amino acid , namely methionine , which has also been demonstrated to elongate the lifespan of mammals , has the advantage of being a well applicable regime . Therefore , understanding the underlying mechanism of the anti-aging effects of methionine restriction is of crucial importance . With the help of the model organism yeast , we show that limitation in methionine drastically enhances autophagy , a cellular process of self-digestion that is also switched on during caloric restriction . Moreover , we demonstrate that this occurs in causal conjunction with an efficient pH decrease in the organelle responsible for the digestive capacity of the cell ( the vacuole ) . Finally , we prove that autophagy-dependent vacuolar acidification is necessary for methionine restriction-mediated lifespan extension . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"aging",
"cell",
"death",
"developmental",
"biology",
"fungi",
"model",
"organisms",
"organism",
"development",
"cell",
"biology",
"yeast",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"cell",
"processes",
"saccharomyces",
"cerevisiae",
"molecular",
"cell",
"biology",
"saccharomyces",
"population",
"biology",
"organisms",
"research",
"and",
"analysis",
"methods"
] | 2014 | Lifespan Extension by Methionine Restriction Requires Autophagy-Dependent Vacuolar Acidification |
Familial hemiplegic migraine type 2 ( FHM2 ) is an autosomal dominant form of migraine with aura that is caused by mutations of the α2-subunit of the Na , K-ATPase , an isoform almost exclusively expressed in astrocytes in the adult brain . We generated the first FHM2 knock-in mouse model carrying the human W887R mutation in the Atp1a2 orthologous gene . Homozygous Atp1a2R887/R887 mutants died just after birth , while heterozygous Atp1a2+/R887 mice showed no apparent clinical phenotype . The mutant α2 Na , K-ATPase protein was barely detectable in the brain of homozygous mutants and strongly reduced in the brain of heterozygous mutants , likely as a consequence of endoplasmic reticulum retention and subsequent proteasomal degradation , as we demonstrate in transfected cells . In vivo analysis of cortical spreading depression ( CSD ) , the phenomenon underlying migraine aura , revealed a decreased induction threshold and an increased velocity of propagation in the heterozygous FHM2 mouse . Since several lines of evidence involve a specific role of the glial α2 Na , K pump in active reuptake of glutamate from the synaptic cleft , we hypothesize that CSD facilitation in the FHM2 mouse model is sustained by inefficient glutamate clearance by astrocytes and consequent increased cortical excitatory neurotransmission . The demonstration that FHM2 and FHM1 mutations share the ability to facilitate induction and propagation of CSD in mouse models further support the role of CSD as a key migraine trigger .
Migraine is a clinically heterogeneous disorder affecting more than 10% of the general population . It generally occurs with unilateral and pulsating severe headache often accompanied by nausea , photophobia and phonophobia . In approximately one third of migraineurs , the headache attack is preceded by aura , a transient neurological symptom that are most frequently visual but may involve other senses [1] . The migraine attack is triggered by a brain dysfunction that leads to activation and sensitization of the trigeminovascular system , particularly trigeminal nociceptive afferents innervating the meninges and lastly to headache [2] , [3] , [4] . Neuroimaging examination suggests that migraine aura is associated to cortical spreading depression ( CSD ) , a short-lasting , intense wave of neuronal and glial cell depolarization . CSD spreads slowly over the cortex at a rate of approximately 2–5 mm/min and is followed by long lasting depression of neuronal activity [5] , [6] , [7] , [8] . Experimental evidence on patients and animal models supports CSD as both underlying migraine aura [1] , [7] , [8] , [9] and a key triggering event for trigeminal activation [10] , [11] , [12] , although the role of CSD in migraine headache is still debated . As an indirect confirmation , several migraine prophylactic agents cause an increase of CSD initiation threshold [13] . Common migraine has a strong multifactorial genetic component , which is higher in migraine with aura ( MA ) than in migraine without aura ( MO ) [14] , [15] . As for many other multifactorial diseases whose complexity hampers the investigation of the pathogenetic mechanisms , rare monogenic forms that phenocopy most or all the clinical features of the common disease are of great help for describing the complicated events leading to migraine . Familial hemiplegic migraine ( FHM ) is a rare autosomal dominant subtype of MA , whose aura symptoms include hemiparesis . Aura symptoms and headache duration are usually longer in FHM than MA , but all other headache properties are similar . FHM is genetically heterogeneous and is associated to mutations in three different genes . Mutations in CACNA1A [16] , ATP1A2 [17] and SCN1A [18] genes are responsible for Familial hemiplegic migraine type 1 ( FHM1 ) , type 2 ( FHM2 ) , and type 3 ( FHM3 ) , respectively . The CACNA1A and SCN1A genes both encode neuronal voltage-gated ion channels , whereas the ATP1A2 gene encodes the α2 subunit of the Na , K-ATPase , hence suggesting a key role of cation trafficking in the pathophysiology of FHM . Until now , more than 50 FHM2 mutations have been identified and most of these are missense mutations . A small fraction of mutations is represented by microdeletions [19] and a single mutation affecting the stop codon , which causes an extension of the ATP1A2 protein by 27 aminoacid residues [20] . Most of the ATP1A2 mutations are associated with pure FHM without additional clinical symptoms [17] , [19] , [20] , [21] , [22] . However , a number of FHM2 mutations have been associated to complications like cerebellar ataxia [23] , childhood convulsions [24] , epilepsy [25] and mental retardation [26] . Interestingly , ATP1A2 mutations associated with non-hemiplegic migraine phenotypes , such as basilar migraine and even common migraine have been reported [27] , [28] . The Na , K ATPase is a P-type ion pump that utilizes the free energy of ATP hydrolysis to exchange Na+ for K+ and maintains gross cellular homeostasis . The functional pump is a heterodimer , consisting of one α catalytic subunit and one β subunit that is required for protein folding , assembling , membrane-addressing , and modulates substrate affinity [29] . The α subunit exposes both the amino- and carboxy- termini in the cytoplasm and crosses the plasma membrane with ten transmembrane segments ( M1–M10 ) [30] . Four isoforms of α Na , K-ATPase ( α1 , α2 , α3 and α4 ) are present in mammals [29] , [31] . While no pathogenic mutations are known for the ubiquitous α1- and the testis α4-subunits , mutation in both α2 and α3 isoforms cause neurological diseases when mutated , FHM2 and rapid-onset dystonia parkinsonism , respectively [32] . While in the adult brain the α1 isoform is nonspecifically present in both neurons and glial cells and α3 is neuron-specific , the α2 isoform is essentially expressed in astrocytes [33] . Investigation of the functional consequences of FHM2 mutations in heterologous expression systems revealed that these mutations produce partial or complete loss of function of the α2 Na , K pump [34] , [35] , [36] . Here , we report the generation of the first mouse model of FHM type 2 , a knock-in mutant harboring the W887R ATP1A2 mutation . The W887R mutation localizes to the extracellular loop between M7 and M8 , which includes the β subunit binding site [37] and was shown to produce the almost complete loss of pump activity [17] , [38] . Homozygous Atp1a2R887/R887 mutants die just after birth , while heterozygous Atp1a2+/R887 mice are fertile and show no apparent clinical phenotype . However , heterozygous FHM2 mouse displays altered CSD properties , such as decreased threshold and increased velocity of propagation . We hypothesize that inefficient astrocyte-mediated clearance of glutamate from the synaptic cleft is a key event for the enhanced susceptibility to CSD in the FHM2 mouse .
With the aim of investigating the molecular pathogenesis of FHM type 2 , we generated a knock-in mouse model by inserting an FHM2 mutation , the transition T2763C that causes the aminoacid replacement W887R in the Atp1a2 murine gene ( construct details in M&M ) . The amino acid sequence conservation between human and mouse α2 Na , K-ATPase proteins is very high and , in particular , in the extracellular domain between transmembrane domains M7–M8 , where W887R is located [17] . This mutation was one of the first two mutations reported to be associated to typical cases of the disease . Embryonic stem cells harboring the R887 and the neo cassette were injected in C57Bl/6J blastocysts and then transferred to pseudopregnant CD1 females . We obtained three chimeric mice , one of which transmitted the Atp1a2+/R887-neo allele through germline ( Figure 1A ) . Heterozygous Atp1a2+/R887-neo mice were genotyped by Southern blot analysis ( Figure 1C ) , are fertile and display no apparent phenotype . To remove the neo cassette that hampers the natural expression of the mutant allele , we crossed the Atp1a2+/R887-neo mice with transgenic mice expressing the Flippase recombination enzyme ( FLPe ) under the control of the human ACTB promoter ( TgN ( ACTFLPe ) 9205Dym; The Jackson Laboratory ) . Hence , we obtained the heterozygous Atp1a2+/R887 knock-in mice ( Figure 1B ) , which are fertile as well and show no visible clinical phenotype . Contrary to heterozygous mice , homozygous Atp1a2R887/R887 mutants do not survive beyond the first day post partum , thus resembling the neonatal lethal phenotype of the Atp1a2 null mutant [39] , which succumbs for dysfunctional neuronal activity and respiratory distress . Therefore , we addressed our investigation onto the heterozygous knock-in mouse , which shares the Atp1a2 gene asset with FHM2 patient . The general behavior of heterozygous Atp1a2+/R887 mice was tested by a modified SHIRPA protocol [40] that provides comparable quantitative data on animal motor , sensory , autonomic and neuropsychiatric functions . The scored parameters are summarized in Table 1 . No major differences in the sensory-motor functions were observed between heterozygous Atp1a2+/R887 ( n = 8 ) and wild-type ( n = 6 ) mice , except for a higher fear and anxiety of Atp1a2+/R887 at the specific tests of transfer arousal and fear ( p<0 . 05; Table 2 ) . Mutant and wild type Atp1a2 gene expression was evaluated at E19 . 5 in Atp1a2R887/R887 for lethality constrain and at adult age in Atp1a2+/R887 mutants . Semi quantitative reverse transcription PCR ( RT-PCR ) analysis showed that both wild type and mutant Atp1a2R887 alleles express equal amount of transcripts ( Figure 2A , left panel ) . Each RT-PCR experiment was normalized on intra-sample actin transcript level . The nucleotide replacement resulting in the W887R mutation creates a new MspI restriction site that we used to confirm the R887 mutation in the Atp1a2R887 transcript and to quantify the mutant transcript in heterozygous mice as intra-sample control ( Figure 2A , right panel ) . The expression of α2 Na , K-ATPase protein was assessed in embryonic brain . Immunoblot of total lysate and microsomal fractions revealed a markedly reduced amount of α2 Na , K-ATPase in the mutants , which displayed approximately half the level of wild type in the Atp1a2+/R887 mice . In the Atp1a2R887/R887 mutant , the R887 α2 Na , K-ATPase is barely observable ( Figure 2B ) . In order to investigate whether the reduced amount of α2 isoform induces a compensatory increase of expression of the paralogous α1 and α3 isoforms in the adult , when the principal phenotype , CSD , is assessable , we analyzed brain tissues with specific α isoform antibodies . No differences in the level of α1 and α3 isoforms were observed in whole brain of Atp1a2+/R887 mice ( Figure 2C ) compared to wild type ones . On the contrary , the Atp1a2+/R887 model displays an α2 expression level reduced to approximately 50% in the cortex , 35% in cerebellum and 40% in total brain ( Figure 2D ) . The loss of α2 protein in the Atp1a2R887/R887 prompted us to investigate the fate of wild-type and mutant α2 proteins by cell transfections . HeLa cells were co-transfected with pA2-R887 or pA2-wt constructs , which express , respectively , mutant and wild type full length c-myc-tagged ATP1A2 cDNAs , each together with pB2 expressing the β2 subunit ( as described in [17] ) . Immunoblot revealed a decreased amount of R887 mutant ( Figure 3A ) , thus confirming the in vivo results on Atp1a2R887/R887 and Atp1a2+/R887 mutant mice . More important , immunofluorescence staining demonstrated a different subcellular localization . Wild type α2 Na , K ATPase showed a typical plasmamembrane and slightly endoplasmic reticulum staining ( Figure 3B , upper panels ) . Differently , most of R887 mutant protein appeared as punctuated pattern localized in the perinuclear region , which overlapped with the endoplasmic reticulum marker calnexin ( Figure 3B; a colocalization quantification appears on right panels ) . Misfolding of the mutant α2 Na , K ATPase induced by the R887 mutation causes very likely the endoplasmic reticulum retention and the inefficient and delayed secretion process . In fact , by inhibiting the proteasome activity with MG132 , wild-type and more consistently mutant α2 subunits accumulated in transfected cells ( Figure 3C ) . Migraine is a complex phenotype that hampers the objective and quantitative evaluation in animal models . In order to assess the effect of the R887 ATP1A2 mutation on an important component of the migraine attack , cortical spreading depression ( CSD ) , we analyzed this neuronal phenomenon in adult Atp1a2+/R887 mice and in their wild type littermates . CSD was induced by electrical stimulation of the visual cortex using a bipolar electrode and recorded at two sites of somatosensory and motor cortex ( Figure 4A ) . Incremental current stimuli were delivered up to CSD induction and the charge delivered at CSD activation was considered as threshold . Atp1a2+/R887mutants were more susceptible to undergo CSD . Indeed the threshold for induction of CSD in mutant animals was significantly lower than wild type animals ( Atp1a2+/R887 , 13 . 00±1 . 7 µC , n = 20; wild type , 19 . 9±1 . 9 µC , n = 18; t-test p<0 . 01 ) ( Figure 4B , left graph ) . Moreover , CSD propagation rate was altered in the mutant , which showed higher CSD velocity rising from 3 . 85±0 . 35 mm/min ( wild type , n = 18 ) to 5 . 41±0 . 41 mm/min ( n = 20; t-test p<0 . 01 ) in Atp1a2+/R887 mice ( Figure 4B , middle graph ) . No significant difference was observed in CSD duration ( Atp1a2+/R887 40 . 1±3 . 19 sec , n = 20; wild type 41 . 1±3 . 5 sec , n = 18; t-test p = 0 . 83 ) ( Figure 4B , right graph ) . After the first CSD , the trace was monitored for further 90 minutes to reveal repetitive CSDs , a parameter correlated with the phenotype severity of FHM models [41] . Heterozygous R887 mutation did not modify the proportion of mice showing repetitive CSD ( Atp1a2+/R887 4 out of 20 mice , wild type 4 out of 18 mice ) . We conclude that the R887 α2 Na , K-ATPase facilitates CSD induction and propagation , but it neither affects its duration nor promotes the induction of repetitive CSDs . Within groups , no difference in CSD threshold and propagation rate was observed between male and female ( see Methods ) .
Genetic mouse models are essential tools to dissect complex pathogenic mechanisms leading to human diseases . Here , we report data on the generation of the first knock-in mouse carrying the W887R ATP1A2 mutation causing FHM2 . The R887 allele has been associated to a typical form of FHM with hemiparesis and epileptic episodes [17] . Since the effect of this mutation is an almost complete loss of function [17] , [38] , we expected a neonatal lethality of homozygous mutant mice . In fact , Atp1a2R887/R887 mutants die few minutes after birth , closely resembling the knock-out models [39] , [42] , which fail to develop a regular respiratory rhythm [43] , [44] . Heterozygous Atp1a2+/R887 mice are viable and fertile . A general behavior characterization by a modified SHIRPA protocol [40] shows a higher susceptibility to fear and anxiety in Atp1a2+/R887 mice . This result resembles the previous reports by Ikeda et al . [42] and Moseley et al [45] showing similar phenotype in the null allele heterozygous Atp1a2+/− mice by more specific tests for anxiety and conditioned fear . As we demonstrate , the mutant gene is correctly transcribed and translated . However , the mutant R887 protein is ineffectively exported from the endoplasmic reticulum-Golgi system . Mutant R887 isoform is mostly degraded by the proteasomal system as demonstrated by the remarkable accumulation of mutant protein under proteasomal inhibition . This is particularly evident in vivo , where the mutant protein is barely detectable in the homozygous mutant brain . This finding is apparently in contrast with our previous result [17] that showed the mutant R887 subunit localized all over the cytoplasm in COS7 transfected cells and seemingly to plasmamembrane as well . By the recent confocal analysis and employing a transfection system that does not saturate , like the COS7 cells , the cytoplasm of exogenous protein , the mutant protein is shown as mostly endoplasmic reticulum-retained . It is worth noting that Koenderink and coworkers proposed a plasmamembrane localization of the R887 protein by centrifugal fractionation in Xenopus oocytes , probably due to the different cellular system and conditions ( room temperature ) and the indirect method of localization [38] . Infact , this test at room temperature may favor the mislocalization of the α2 ATPase mutant protein , as reported in [36] . Considering the autosomal dominant inheritance of FHM , we have addressed our attention to the phenotype analysis of the heterozygous knock-in mouse . CSD represents an excellent phenotype to be analyzed in animal models of migraine as CSD underlies migraine aura in patients [1] , [7] , [8] , [9] and can activate the meningeal trigeminal nociceptors in animals [12] . Atp1a2+/R887 mutant mice , our FHM2 model , are more susceptible to CSD as shown by the decreased threshold of induction and the increased velocity of propagation of CSD induced by electrical stimulation of the cortex in vivo . Duration of CSD in Atp1a2+/R887 mice is unchanged . The facilitation of CSD in our FHM2 mouse model is thus very similar to that previously described in Cacna1a knock-in mice representing the FHM1 models [41] , [46] . In fact , both homozygous and heterozygous S218L and homozygous R192Q FHM1 models showed a lower threshold for CSD induction and a higher velocity of CSD propagation , whereas CSD duration was not significantly prolonged . Interestingly , the extent of CSD facilitation correlated with the severity of the clinical phenotype of the two FHM1 mutations in humans [41] , [46] , [47] . The demonstration that FHM2 and FHM1 mutations share the ability to facilitate induction and propagation of CSD in mouse models further support the role of CSD as a key migraine trigger . The facilitation of CSD in Atp1a2+/R887 mice could be due to impaired clearance of K+ and/or glutamate by astrocytes during cortical neuronal activity consequent to loss-of-function of the α2 Na , K ATPase pump , as previously suggested [34] , [48] . Pharmacological evidence shows that α3 and/or α2 Na , K pumps participate in the clearance of K+ from the extracellular space during intense neuronal activity , although the relative importance of α3 and α2 Na , K pumps is unclear[49] , [50] . Most models of CSD include local increase of extracellular [K+] above a critical value as a triggering event in the initiation of CSD , hence predicting that a reduced K+ clearance would result in a lower threshold for CSD induction [51] . Indeed , in hippocampal slices the inhibition of α2 and α3 Na , K pumps by local administration of ouabain ( at a concentration which only partially affects the low affinity α1 Na , K pump ) reduced the threshold for CSD induction by local pulses of high [K+] [52] . This reduced CSD threshold was accompanied by a large increase in CSD duration ( and decrease in post-CSD undershoots of membrane potential and external [K+] ) , pointing to the involvement of α3 and/or α2 Na , K pump activity in CSD termination . We speculate that our findings of a lower threshold for CSD induction but unaltered CSD duration in Atp1a2+/R887 mice suggest a relatively minor role of the glial α2 Na , K pump in K+ clearance . This is in agreement with the evidence that the α3 isoform contributes most of the Na , K ATPase activity in mouse brain homogenates [53] and , therefore , we suggest that the reduced CSD threshold in FHM2 knockin mice is not primarily due to impaired K+ clearance by astrocytes . Several lines of evidence indicate a specific role of the α2 Na , K pump in glutamate clearance during synaptic transmission . The α2 Na , K pump is specifically stimulated by glutamate in cultured astrocytes [54] . In the adult somatosensory cortex the α2 Na , K pump shows a specific localization in astrocyte processes surrounding glutamatergic synaptic junctions , which coincides with that of the glial glutamate transporters GLAST and GLT1 [55] , [56] . Also , a physical association and functional coupling between the α2 Na , K pump and glutamate transporters has been demonstrated [56] . We therefore hypothesize that CSD facilitation in the FHM2 mouse model is sustained by inefficient glutamate clearance by astrocytes and consequent enhanced cortical excitatory neurotransmission , particularly the NMDA receptor-mediated transmission during high-frequency action potential trains [57] . This glutamatergic hypothesis finds suggestive echoes in the recent report by Anttila et al . [58] , where MTDH , a modulator of glutamate transporters has been associated to the common form of migraine with aura . In addition , a mutation of the glial excitatory aminoacid transporter type 1 ( EAAT1 ) leads to neuronal hyperexcitability and subsequent seizures , hemiplegia , and episodic ataxia by impaired glutamate uptake [59] . While this scenario remains to be confirmed in the FHM2 mouse model , FHM1 models displayed an enhanced glutamatergic synaptic transmission due to increased Ca2+ influx through the mutant presynaptic CaV2 . 1 channels and increased probability of glutamate release at cortical pyramidal cell synapses [60] . A causative link between gain of function of glutamatergic transmission at recurrent cortical pyramidal cell synapses and facilitation of experimental CSD was demonstrated in the FHM1 mouse model [60] . Both FHM1 and FHM2 mice point to a model of CSD initiation , where the activation of NMDA receptors by glutamate released from recurrent cortical pyramidal cell synapses plays a key role in the positive feedback cycle that provokes CSD [4] . Furthermore , the absence of the α2 Na , K pump from the glial processes surrounding GABAergic terminals [55] suggests that FHM2 mutations fail to affect inhibitory neurotransmission , similarly to the FHM1 model , which showed unaltered inhibitory neurotransmission at synapses between fast-spiking interneurons and pyramidal cells [60] . We therefore propose that episodic disruptions of the excitation-inhibition balance and hyperactivity of cortical circuits due to excessive recurrent excitation underlie the vulnerability to “spontaneous” CSD ignition in both the rare forms of FHM1 and FHM2 and , probably , at least a fraction of common migraine cases .
Commercially available rabbit polyclonal antibody directed against α2 Na , K-ATPase isoform ( cat . AB9094 . Millipore , Billerica , MA , USA ) ; mouse monoclonal antibodies for Na , K-ATPase alpha 1 isoform ( α6F; Developmental Studies Hybridoma Bank , Iowa City , IA , USA ) , for Na , K-ATPase alpha 3 isoform ( cat . MA3-915 , Affinity Bio Reagents Suite 600 Golden , CO , USA ) , anti-bovine α-tubulin , mouse monoclonal antibody ( cat . A11126 , Molecular probes , Inc . 29851 Willow Creek Road , Eugene , OR , USA ) ; GAPDH ( 6C5 ) mouse monoclonal antibody ( sc-32233 , Santa Cruz Biotechnology Inc . , California , USA ) ; Ubiquitin ( P4D1 ) , mouse monoclonal antibody ( sc-8017 , Santa Cruz Biotechnology Inc . CA , USA ) . ECL anti-mouse and anti-rabbit IgG and horseradish peroxidase ( HRP ) -linked species-specific whole antibodies were purchased from GE Healthcare . Polyclonal rabbit anti-goat IgG/HRP was obtained from Dako ( Glostrup , Denmark ) . For immunofluorescence experiments , the following antibodies were used: monoclonal anti c-Myc ( 9E10 ) and rabbit anti- calnexin ( Sigma-Aldrich , Milan , Italy ) . Secondary antibodies were conjugated with Alexa 488 and Alexa 596 ( Invitrogen , Carlsbad , CA , USA ) . Procedures involving animals and their care were conducted in conformity with guidelines of the Institutional Animal Care and Use Committee at San Raffaele Hospital ( Milan , Italy ) in compliance with national ( D . L . No . 116 , G . U . Suppl . 40 . Feb 18 , 1992 , Circolare No . 8 G . U . , 14 Lug . 1994 ) and international ( EEC Council Directive 86/609 , OJ L 358 , 1 DEC . 12 , 1987; National Institutes of Health Guide for the Care and Use of Laboratory Animals , U . S . National Research Council , 1996 ) laws and policies . Animals were housed in Specific Pathogen Free ( SPF ) conditions , maintained on a 12-h light/dark cycle , with free access to food and water . Atp1a2+/R887-neo knock-in mice were generated using homologous recombination in embryonic stem ( ES ) cells to modify the Atp1a2 gene such that the exon 19 contained the human FHM-2 W887R mutation . In the targeting vector , the original TGG triplet codon ( POSITION 2763 , CODON 921 ) was changed into CGG by mutagenesis , creating the W887R mutation . Downstream of exon 19 , a PGK-driven neo cassette flanked by LoxP sites was present . ES cells were electroporated , and clones were selected for homologous recombination by Southern blot analysis . The presence of the W887R mutation was tested by PCR using primers 5′-GGCTTCTTTACCTACTTTGTGATA-3′ and 5′-ATGCCCTGCTGGAACACTGAGTTG-3′ and subsequent sequencing analysis of exon 19 . Targeted ES cells were injected into C57Bl/6J blastocysts and these transferred into pseudopregnant CD1 females to create chimeric animals . Chimeras were backcrossed with wild-type C57Bl/6J to generate F1 progeny and the agouti offsprings were genotyped for transmission of the mutant allele , generating transgenic line Atp1a2+/R887-neo knock-in mice . Heterozygous Atp1a2+/R887-neo knock-in mice were bred with transgenic mice expressing FLPe recombinase under the control of the human ACTB promoter ( TgN ( ACTFLPe ) 9205Dym; The Jackson Laboratory ) to remove the neo cassette . Expression of FLPe recombinase as early as embryonic day 10 . 5 causes the Flippase recognition target ( FRT ) sites recombination and the removal of the neo cassette . Germ line transmission was obtained and transgenic line Atp1a2+/R887-neo was established . Mice were further bred with C57Bl/6J for seven generations , at this stage the background would nevertheless be >90% congenic . Heterozygous Atp1a2+/R88 and Atp1a2+/+ littermates were used for further analysis . Sensory-motor function of mutant mice compared with controls was assessed by a modified version of the SHIRPA protocol primary screening [40] . Briefly , undisturbed behavior of each animal was first observed in its own home cage: body position , spontaneous activity and respiration rate were recorded , assigning a score to each behavior . In addition , manifestations of tremors , bizarre behaviors , stereotypes or convulsions were checked at this stage of the protocol . Thereafter mice were transferred individually to a new arena and were tested for transfer arousal , palpebral closing , piloerection , gait , pelvic and tail elevation , touch escape and positional passivity . There followed a sequence of manipulations using tail suspension and a grid across the width of the arena; animals were scored for trunk curl , limb grasping and grip strength . To complete the assessment , the animals were restrained in a supine position to record autonomic behaviors ( heart rate , skin color , limb and abdominal tone , lacrimation , salivation ) prior to measurement of the righting reflex after flip of the animal . Vocalizations and irritability ( during supine restrain ) were also recorded . Fear was assessed based on reaction to transfer to a new environment . A score was assigned to each behavioral test as described in Table 1 . Total RNA was extracted from embryonic mice ( E19 . 5 ) ( n = 9 , 3 embryos for each genotype ) neuronal ( brain ) tissues by Trizol method ( Invitrogen , Carlsbad , CA , USA ) . RNA was reverse transcribed using random hexamers SuperScript® First-Strand Synthesis System ( Invitrogen , Carlsbad , CA , USA ) according to the manufacturer's instructions . Atp1a2 cDNA was amplified using forward primer on exon 19 ( 5′-GGCTTCTTTACCTACTTTGTGATA-3′ ) and reverse primer on exon 20 ( 5′-ATGCCCTGCTGGAACACTGAGTTG-3′ ) with Hot Master Taq DNA polymerase ( Eppendorf , Hamburg , Germany ) at 94°C for 2 min , 35 cycles at 94°C for 30 s , 58°C for 30 s , 65°C for 30 s , and 65°C for 5 min . This strategy allows amplification of both endogenous wild-type and mutant allele ( PCR product: 254 bp ) . The relative Atp1a2 amount was normalized to the β-actin expression levels ( 610 bp PCR product ) . Since the R887 missense mutation introduces a new restriction site for MspI enzyme , the PCR product was subsequently digested with MspI ( New England Biolabs , Ipswich , MA , USA ) to discriminate the endogenous gene ( uncut , band size: 254 bp band ) and the mutant ( cut , bands size: 178 bp+76 bp ) . PCR products were run on a 2% agarose gel in TAE buffer . To prevent proteolysis during the procedure , all steps were carried out on ice , and all buffers contained protease inhibitor cocktail ( Roche , Mannheim , Germany ) and phenylmethanesulfonyl fluoride ( 1 mM ) . Embryonic brains of the various genotypes ( n = 12 , 4 for each genotype ) were processed simultaneously . For the extraction of membrane proteins , whole brain was homogenized with a glass-Teflon homogenizer in Sucrose solution ( 0 . 32 M Sucrose , 5 mM Hepes pH 7 . 4 , 2 mM EDTA ) . After a short centrifugation ( 5000 rpm , 20′4°C ) the supernatant was centrifuged for 1 hr at 42 , 000 rpm 1 h 4°C ( Beckman , ultraTL100 , rotor TL100 . 3 ) and the pellet resuspended in Sucrose buffer . Protein concentration was measured using the Bio-Rad Protein Assay according to the manufacturer's instructions . The preparation of cells and tissues ( total brain , cortex and cerebellum ) total lysates were performed adding RIPA buffer ( 50 mM Tris-HCl , pH 7 . 4 , 150 mM NaCl , 0 . 5% sodium deoxycholate , 0 . 1% SDS , 2 mM EDTA , and 1% Igepal CA-630 ) to the collected samples and left 30′ on ice than the lysates were centrifuged 13000 g 20′4°C . The protein content of the supernatant was measured using the BCA protein assay with bovine serum albumin as standard . We resuspended equal amounts of proteins ( 15 µg each sample in 20 µl ) in SDS-PAGE buffer ( 100 mM Tris-Glycine pH 6 . 8 , 0 . 56 M mercaptoethanol , 2% SDS , 15% glycerol , and 0 . 1% BFB ) , and separated them for 2 h at 100volt in 8% SDS-polyacrilamide gels . Proteins were electrophoretically transferred to hybond ECL nitrocellulose membranes ( GE Healthcare , Munich , Germany ) and blots were blocked overnight with 5% non-fat milk 0 . 1% Tween-20 in PBS . The blocked blots were incubated for 2 h with subunit specific antibodies , washed three times for 10 minutes each with 0 . 1% Tween-20 in PBS then incubated with the appropriate peroxidase conjugated secondary antibodies . After another series of washes ( three times for 10 minutes each ) peroxidase was detected using a chemiluminescent substrate ( GE Healthcare , Munich , Germany ) . Plasmid constructs were the same as in [17] ) by metafectene ( Biontex , Martinsried/Planegg , Germany ) according to the manufacturer's instructions . We selected the ratio range of Metafectene ( µl ) to plasmids DNA ( µg ) of 5∶1 . 48 h after transfection , we fixed HeLa cells in 4% paraformaldehyde ( PFA ) for 30 min at RT and blocked and permeabilized with 10% donkey serum 0 . 2% Triton-X100 in phosphate-buffered saline solution ( PBS ) for 30 min at RT . Permeabilized cells were then incubated with primary antibodies for 2 hr at RT , than washed ( three times ) in PBS , incubated with appropriate secondary antibodies and washed three times with PBS solution . We placed cells in fluorescent mounting medium ( Dako Cytomation , Glostrup , Denmark ) over microscope slides and confocal microscopy was performed on the Perkin Elmer UltraVIEW . Immunofluorescence colocalization was visualized by confocal microscopy and analyzed by Wright Cell Imaging Facility ( WCIF ) colocalization plug-in of Image J software ( http://www . uhnresearch . ca/facilities/wcif/imagej/colour_analysis . htm ) . The following parameters were measured: Pearson's correlation coefficient ( Rr; 1 , perfect correlation , to −1 , perfect exclusion ) ; Mander's overlap coefficient ( R; 1 , highest , to 0 , random correlation ) ; Ch1∶Ch2 , the red∶green pixel ratio . Proteasome inhibitor MG132 ( carbobenzoxy-L-leucyl-L-leucyl-L-leucinal ) was obtained from Sigma-Aldrich , Milan , Italy ( cat . C2211 ) . MG132 were dissolved in DMSO and applied to cells at the concentration of 10 µM , after 48 hours of transfection for the time periods indicated in the text and . An equivalent volume of DMSO was added to control cells . Anti ubiquitin antibody was used to reveal the increase of ubiquitinated proteins after proteasome inhibition . CSD was recorded as described in Van den Maagdenberg , et al . [46] . Briefly , mice ( 20–30 g ) were anaesthetized with urethane ( 20% in saline; 6 ml/kg i . p . ) . Animals , mounted on a stereotaxic apparatus were continuously monitored for adequate level of anesthesia , temperature , heart rate and nociceptive reflexes . Blood oxygen saturation and flux as well as heart and breathing rates were monitored non-invasively using an oximeter ( Starr , Life Science Corp . ) . Oxygen was supplied to maintain blood oxygenation above 93% for the entire duration of the experiment . Heart rate was between 400–600 beats/min , and breathing rate approximately 200 breaths/min . Animals not meeting these criteria were excluded from our sample . To record CSD three holes were drilled in the skull over the left hemisphere . The first corresponded to the occipital cortex and was used for access of the electrical stimulation electrode ( 0 mm A-P , 2 mm M-L from lambda ) . The second hole , at the parietal cortex ( 1 mm M-L , 1 mm caudal to bregma ) and the third hole , at the frontal cortex ( 1 mm M-L , 1 mm rostral to bregma ) , were used for placement of the CSD recording electrodes . The steady ( DC ) potential was recorded with glass micropipettes filled with NaCl ( 3 M , tip resistance 1–2 MΩ ) inserted 200 µm below the dural surface . An Ag/AgCl reference electrode was placed subcutaneous above the nose . Cortical stimulation was conducted using a copper bipolar electrode ( 0 . 2 mm tip diameter , 0 . 3 mm intertip distance ) placed on the cortex surface after removing the dura . Single pulses of increasing intensity ( 20 , 30 , 40 , 50 , 60 , 80 , 100 , 120 , 140 , 160 , 180 , 200 , 230 , 260 , 290 , 320 , 350 , 380 , 430 , 480 , 530 , 600 , 700 , 800 , 900 , 1000 µA ) were applied for 100 ms at 3-min intervals by using a stimulus isolator/constant current unit ( WPI , USA ) until a CSD event was observed [13] . DC cortical potential was amplified ( 10× ) and low-pass filtered at 200 Hz ( Cyberamp , Axon Instruments , Union City , CA ) . Signals were continuously digitized and recorded using Labview data acquisition and analysis system . The minimal stimulus intensity at which a CSD event was elicited was taken as the CSD threshold . In all mice , when CSD was elicited , recordings were continued for 90 min to detect multiple CSDs . To estimate CSD propagation velocity , the distance between the two recording electrodes was divided by the time elapsed between the CSD onsets at the first and second recording sites . The percentage of mice with multiple CSD events was determined only from the mice that could be recorded for one full 90 min following the first detected event . CSD duration was measured at half-maximal amplitude [13] . Because no difference in CSD threshold and propagation rate was observed between male ( N = 11 wild type and N = 11 mutants ) and female ( N = 7 wild type and N = 9 mutants ) within each genotype ( wild type: threshold male 20 . 7±2 . 1 µC , female 18 . 7±3 . 5 Mann-Whitney test p = 0 . 61; propagation rate male 3 . 9±0 . 42 mm/min , female 3 . 8±0 . 66 mm/min Mann-Whitney test p = 0 . 86; mutants: threshold male 13 . 4±3 . 0 µC , female 12 . 6±1 . 3 t-test p = 0 . 82; propagation rate male 5 . 2±0 . 32 mm/min , female 5 . 6±0 . 84 mm/min Mann-Whitney test p = 0 . 94 ) the results from males and females were pooled . For SHIRPA protocol primary screening , comparisons were performed with the Mann-Whitney nonparametric test . The Student's t-test with one-tail distribution was used for significance calculation in densitometric analysis . Statistical analysis for CSD recordings was performed using Sigma Stat 3 . 1 ( Systat Software , Chicago IL USA ) . Multiple groups were compared by ANOVA followed by post-hoc comparisons applying Bonferroni correction or Holm-Sidak test . When two groups were compared , t-test was applied . Normality and homoschedasticity of the data was checked . Data not normally distributed were compared using nonparametric Kruskal-Wallis ANOVA or Mann-Whitney rank sum test . Significance level was equal to 0 . 05 . Data are reported as average ± SEM . | We previously reported that mutations of the α2 subunit of the Na , K-ATPase cause familial hemiplegic migraine type 2 ( FHM2 ) , a dominant form of migraine with aura . This paper describes the first animal model of FHM2 and represents the further proceeding in this disease investigation . Homozygous knock-in mutant mice die just after birth , while heterozygous mice show no apparent clinical phenotype . However , in vivo analysis revealed a marked facilitation of cortical spreading depression ( CSD ) , the phenomenon underlying migraine aura . Given the evidence for specific functional coupling between the glial α2 Na , K pump and glutamate transporters , we hypothesize that CSD facilitation in the FHM2 mouse model is sustained by inefficient glutamate clearance by astrocytes and consequent increased cortical excitatory neurotransmission . We finally propose this FHM2 mouse as a valuable in vivo model to investigate migraine mechanisms and , possibly , treatments . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"medicine",
"model",
"organisms",
"neurological",
"disorders",
"neurology",
"biology",
"mouse",
"migraine"
] | 2011 | Increased Susceptibility to Cortical Spreading Depression in the Mouse Model of Familial Hemiplegic Migraine Type 2 |
Plant genomes encode large numbers of nucleotide-binding ( NB ) leucine-rich repeat ( LRR ) immune receptors ( NLR ) that mediate effector triggered immunity ( ETI ) and play key roles in protecting crops from diseases caused by devastating pathogens . Fitness costs are associated with plant NLR genes and regulation of NLR genes by micro ( mi ) RNAs and phased small interfering RNAs ( phasiRNA ) is proposed as a mechanism for reducing these fitness costs . However , whether NLR expression and NLR-mediated immunity are regulated during plant growth is unclear . We conducted genome-wide transcriptome analysis and showed that NLR expression gradually increased while expression of their regulatory small RNAs ( sRNA ) gradually decreased as plants matured , indicating that sRNAs could play a role in regulating NLR expression during plant growth . We further tested the role of miRNA in the growth regulation of NLRs using the tobacco mosaic virus ( TMV ) resistance gene N , which was targeted by miR6019 and miR6020 . We showed that N-mediated resistance to TMV effectively restricted this virus to the infected leaves of 6-week old plants , whereas TMV infection was lethal in 1- and 3-week old seedlings due to virus-induced systemic necrosis . We further found that N transcript levels gradually increased while miR6019 levels gradually decreased during seedling maturation that occurs in the weeks after germination . Analyses of reporter genes in transgenic plants showed that growth regulation of N expression was post-transcriptionally mediated by MIR6019/6020 whereas MIR6019/6020 was regulated at the transcriptional level during plant growth . TMV infection of MIR6019/6020 transgenic plants indicated a key role for miR6019-triggered phasiRNA production for regulation of N-mediated immunity . Together our results demonstrate a mechanistic role for miRNAs in regulating innate immunity during plant growth .
Plant nucleotide-binding leucine-rich repeat receptors ( NLR ) recognize specific pathogen effectors and trigger effective defenses against invading pathogens that are usually accompanied by a hypersensitive response ( HR ) in infected tissues [1] . Plant genomes usually encode hundreds of NLRs that are divided into TIR-NB-LRR ( TNL ) and CC-NB-LRR ( CNL ) groups , which contain an N-terminal Toll-like and Interlukin-1 receptor domain ( TIR ) and coiled-coil domain , respectively [2] . In humans , Toll-like receptors ( TLRs ) are innate immune receptors that recognize a variety of ligands from viruses , bacteria , fungi and other types of pathogens [3] . Extensive medical research and clinical observations suggested that TLR-mediated immunity and TLR expression are regulated in a growth-stage specific manner [4] . However , whether plant NLR-mediated immunity responses to infection differ during plant growth is unknown . Micro ( mi ) RNAs are 20- to 24-nucleotide ( nt ) long short RNAs that are processed from hairpin RNA precursors by Dicer-like ( DCL ) enzymes [5 , 6] . They form an RNA induced silencing complex ( RISC ) with the endoribonuclease Argonaute ( AGO ) protein and guide AGO to cleave target mRNAs based on sequence complementarity [7 , 8] . miRNAs play general roles in plant and animal development . For example , the conserved plant miR156 and animal miRNA let-7 control developmental phase changes in plants and animals , respectively [9–11] . Both plant and animal miRNAs can trigger mRNA degradation and translational inhibition in their targets [12] , whereas some plant miRNAs have unique functions in triggering phased siRNA ( phasiRNA ) generation from the cleavage products of their targets [13] . These miRNA precursors usually have an asymmetric bulge in their hairpin structures and produce 22-nt mature miRNAs instead of the more typical 21-nt mature miRNAs [14 , 15] , which confer unique functionality to AGO1 to feed the 3' cleavage product into the RNA dependent RNA Polymerase 6 ( RDR6 ) /DCL4 pathway for phasiRNA production . Plant NLRs are frequent targets of plant miRNAs , many of which are 22-nt in length and can trigger phasiRNA synthesis from NLR target transcripts [16–18] . The tobacco mosaic virus ( TMV ) resistance gene N is regulated by the miRNA cluster miR6019/6020 in tobacco plants and the 22-nt miR6019 cleaves the N transcript at its TIR coding region and triggers phasiRNA production in an RDR6/DCL4-dependent manner [17] . Viral and fungal pathogen infections have been reported to inhibit miRNA function and induce NLR expression , suggesting that miRNA-mediated NLR regulation can be modulated by pathogen infection [18 , 19] . However , whether miRNA-mediated NLR regulation is modulated during plant growth is unclear . Here we showed that during growth plant NLR expression gradually increased , and this increase was accompanied by decreased accumulation of NLR regulatory sRNAs . Using an N-miR6019/6020-TMV trilateral system , we showed that N-mediated immunity strengthened as plants matured , which correlated well with increasing accumulation of N transcripts . Further analysis showed that transcriptional regulation of miR6019/6020 was involved in growth-regulated N expression and function . As NLRs represent rich natural resources for disease resistance , enhanced understanding of NLR regulation mechanisms will facilitate better uses of NLR in breeding programs . Our studies described here provide a mechanism for miRNAs in regulation of plant innate immunity during plant growth .
To test whether sRNAs play a role in the growth regulation of NLRs , we conducted genome wide mRNA and sRNA expression profiling using high-throughput sequencing of RNA samples prepared from tomato and tobacco plants at 1 , 3 and 6 weeks after germination ( WAG ) ( S1 Fig ) . Before quantifying expression of NLR and its sRNA regulators ( referred to as NLR silencer hereafter ) , a complete set of 624 and 177 NLR genes was extracted from tobacco and tomato genomes , respectively , and divided into TNL , CNL and NL ( without N terminal TIR or CC domain ) classes based on the N-terminal structure ( S1 Data and S1 Table ) . Using these NLR cDNAs , small RNA databases , and degradome RNA databases , we identified 210 and 747 NLR silencers from tobacco and tomato genomes , respectively , which were predicted to cleave NLR genes . The predicted cleavages were supported by degradome RNAs mapped to the NLR transcripts ( S2 Data and S2 Table ) . Among the 747 NLR silencers in tomato , 151 , 503 and 61 targeted TNL , CNL and NL , respectively . Furthermore , 13 targeted both TNL and CNL whereas 18 targeted both CNL and NL . One NLR silencer targeted all three classes ( Fig 1A ) . Ten out of 20 TNL , 64 out of 119 CNL , and 13 out of 39 NL were directly targeted by NLR silencers ( Fig 1A ) . A similar situation was observed for tobacco , but fewer NLR silencers were identified and smaller portion of NLRs were directly targeted ( S2A Fig ) . These data suggest that NLR silencers are specific to TNL and CNL classes and have a broad impact on NLR regulation . Since secondary siRNAs are processed from NLR transcripts and play an important role in NLR gene regulation [20] , we also mapped sRNAs to different classes of NLR transcripts to assess NLR secondary siRNAs ( NLR siRNAs ) . In tomato , over 20 , 000 20- to 22-nt siRNAs mapped to CNL transcripts , about 2 , 000 siRNAs mapped to TNL and about 5 , 000 siRNAs mapped to NL transcripts with 100% identity ( Fig 1B ) . In tobacco , over 10 , 000 siRNAs perfectly matched with TNL transcripts , nearly 10 , 000 siRNAs matched with NL transcripts , and about 6 , 000 siRNAs matched with CNL transcripts ( S2B Fig ) . As for the NLR silencers , secondary siRNAs were also specific to TNL and CNL classes and about 98% of tobacco NLR genes and 96% of tomato NLR genes spawned secondary siRNAs ( Fig 1B and S2B Fig ) . These data indicate that secondary siRNAs have an even broader impact on NLR regulation than miRNAs . The expression levels of each NLR silencer were determined in sRNA databases derived from different growth stages in both tobacco and tomato plants ( S2 Table ) . For quality control , the expression profile for the conserved miR156 family members was determined and showed clearly decreasing accumulation in maturing tomato and tobacco plants ( Fig 1C and S2C Fig ) , which was consistent with a previous report for Arabidopsis [21] . The overall expression trend of NLR silencers targeting each class of NLR genes was calculated by combining the total transcripts per million ( TPM ) of all NLR silencers . In tomato , the 22-nt and non-22-nt TNL silencer levels were around 700 and 400 TPM , respectively , at 1 WAG and gradually decreased by the 3 and 6 WAG stages ( Fig 1D , left ) . The CNL and NL silencer levels ranged from 200 to 2 , 200 and 10 to 300 TPM , respectively , and showed a decreasing accumulation pattern during plant growth , except that non-22-nt CNL silencers accumulated to the highest level at 3 WAG ( Fig 1D , middle and right ) . In tobacco , there were around 500 and 50 TPM for 22-nt and non-22-nt TNL silencers , respectively , in 1 WAG plants , and these silencers showed a general decreasing pattern , except for a slight increase in 22-nt silencers at 3 WAG ( S2D Fig , left ) . The CNL and NL silencer counts ranged from 1 , 000 to over 10 , 000 and their levels first increased before then decreased ( S2D Fig , middle and right ) . The NLR siRNAs mapping to different regions of NLR transcripts were also quantified . In tomato , CNL genes spawned more abundant siRNAs than TNL and NL genes ( Fig 1E , middle ) . The CNL and NL siRNAs were enriched in the NBS coding region whereas TNL siRNAs were enriched in the LRR region ( Fig 1E ) . All siRNA levels showed a decreasing pattern during plant growth ( Fig 1E ) . In tobacco , TNL genes spawned more abundant siRNAs compared to CNL and NL genes ( S2E Fig ) . Among the TNL siRNAs , those derived from the TIR coding region represented around 80% of the total and showed a decreasing accumulation pattern during plant growth ( S2E Fig , left ) . In contrast , CNL and NL siRNAs were distributed among different regions at similar levels ( S2E Fig , middle and right ) . We further determined NLR gene expression levels using mRNA sequencing data . Tomato TNL genes were expressed with a dynamic range from 0 to 20 TPM . The dynamic range of tomato CNL transcripts was between 0 and 100 TPM . Most tomato NL transcripts were expressed at low levels ranging from 0 to about 50 TPM , except the level of one RPW8-like transcript ranged from 50 to 350 TPM , which was the highest level among all tomato NLRs ( Fig 1F ) . The majority of the tomato NLR transcript levels showed a gradually increasing pattern during plant growth ( Fig 1F and 1G ) . In tobacco , the TNL , CNL and most NL transcript levels were below 10 TPM and showed no trends ( S2F Fig ) . Interestingly , six RPW8-like transcripts accumulated above 16 TPM at 6 WAG and showed a clear upregulation during growth , which was similar to that seen from tomato ( S2F and S2G Fig , right panels ) . Overall , high-throughput sequencing of sRNA and mRNA samples from different plant growth stages revealed that most NLRs in tomato and tobacco were regulated by sRNAs . Levels of NLR silencers generally decreased and their NLR target levels generally increased as plants matured . Our results thus indicate that sRNAs play an important role in regulating NLR innate immune receptors during plant growth . The N-TMV interaction has served as a classical model system for the study of plant immune responses to pathogens [22] . Since the N gene is regulated by tobacco miR6019/6020 [17] , we chose the N-TMV-miR6019/6020 system to investigate the potential biological significance of NLR regulation during plant growth . For this , we challenged our well-characterized N transgenic TG34 tobacco plants at different growth stages with TMV [22] . We germinated TG34 seeds at consecutive time points , so that plants at 1 , 3 and 6 WAG were inoculated with the TMV U1 strain at the same time ( Fig 2A–2C ) . The results showed that hypersensitive response ( HR ) lesions appeared on inoculated leaves of all plants at 2 days post inoculation ( DPI ) ( Fig 2D–2F , red arrows ) . However , at 7 DPI , plants that were inoculated at 1- and 3 WAG died due to systemic HR ( Fig 2G and 2H ) , whereas plants inoculated at 6 WAG survived ( Fig 2I ) . Furthermore , these plants showed no viral symptoms even at 21 DPI ( Fig 2J ) . The experiment was repeated three times with similar results and average survival rates of TMV-infected and untreated plants clearly showed that TMV U1 infection was lethal to plants infected at 1 and 3 WAG , while plants infected at 6 WAG were fully resistant to TMV ( Fig 2K ) . To rule out the possibility that the death of young TG34 seedlings was due to hypervirulence or rubbing damage , SR1 control plants without the N transgene were also inoculated with the TMV U1 strain at 1- , 3- and 6 WAG ( S3 Fig , upper panel ) . Subsequent symptom observation showed that all SR1 tobacco plants survived at 21 DPI , although plants infected at 1- and 3 WAG showed more severe deformation in terms of leaf morphology ( S3 Fig , lower panel ) . These results indicate that N-mediated resistance against TMV in tobacco is regulated during growth and the resistance response strengthens as plants mature . To investigate the molecular mechanisms underlying regulation of N-mediated immunity during growth , we determined the expression of N by real-time reverse transcription-PCR ( qRT-PCR ) using N-specific primers ( S3 Table ) . We performed this experiment in three different N-expressing Nicotiana species ( N . glutinosa , N . tabacum TG34 and N . tabacum cultivar Samsun NN ) , as well as in the cultivar Petite Havana SR1 ( SR1 ) tobacco lacking the N gene . N . glutinosa is a wild tobacco species from which N has been introgressed into cultivated tobacco [23] . Samsun NN expresses N from a large genomic region introgressed from N . glutinosa . TG34 is a transgenic line that expresses N genomic DNA isolated from Samsun NN [22] . The qRT-PCR experiment detected a similar pattern of N expression in all three N-expressing Nicotiana species , starting at low levels for 1 WAG and gradually increasing for 3 to 6 WAG ( Fig 3A ) . As expected , the N transcript was largely undetectable in SR1 at all three time points ( Fig 3A ) . These results show that N expression is subject to developmental regulation and its expression level increases as plants mature . To further investigate mechanisms of N regulation during plant growth , we constructed a GUS reporter gene under the control of a 4 kilobase ( kb ) N promoter ( Npro::GUS , Fig 3B ) and transformed the construct into SR1 plants . The 4 kb N promoter was previously shown to be sufficient to direct N expression that conferred complete resistance to TMV in tobacco [22] . Reporter gene expression levels in Npro::GUS plants at 1- , 3- and 6 WAG were determined by GUS staining and qRT-PCR . GUS staining showed that the intensity of blue color resulting from GUS activity was similar among plants from the three stages as determined by visual observation ( Fig 3C , left panel ) and quantification of the average gray value ( Fig 3D ) . In contrast , wild type control plants showed no staining ( Fig 3C , right panel ) . The qRT-PCR results showed that the GUS mRNA levels increased by 0 . 5- and 0 . 2-fold in the 3- and 6 WAG stages respectively , relative to the 1 WAG stage , whereas N mRNA levels increased around 5- and 17-fold in TG34 tobacco plants at the corresponding stages ( Fig 3E and 3A ) . These data suggest that N gene transcription does not increase significantly as plants mature and thus the increased level of N mRNA in older plant leaves could be due to a relief from miR6019-mediated post-transcriptional gene silencing . Our previous work showed that N and its TNL homologs were regulated by the miR6019/miR6020 cluster in tobacco plants [17] . Therefore , we tested whether miR6019/miR6020 was involved in regulation of N expression during growth by examining miR6019 levels in TG34 plant leaves at 1- , 3- and 6 WAG by northern blot analysis . The results showed that the relative accumulation levels of nta-miR6019 were 1 . 7 , 1 . 0 and 0 . 6 at 1- , 3- and 6 WAG , respectively ( Fig 4A ) . As expected , the nta-miR156 control showed decreasing relative expression levels of 6 . 2 , 1 . 0 and 0 . 2 ( Fig 4A ) , which was consistent with our sequencing results ( S2C Fig ) and the previous findings in Arabidopsis [21] . Meanwhile , the nta-miR168 levels were 0 . 8 , 1 . 0 and 2 . 1 and nat-miR166 levels were 1 . 1 , 1 . 0 and 1 . 2 ( Fig 4A ) . These results indicate that miR6019/miR6020 expression is regulated and its expression decreases as plants mature . To determine the mechanism by which the MIR6019/6020 cluster is regulated , we next attempted to identify the promoter region of nta-MIR6019/6020a and nta-MIR6019/6020b in tobacco [17] . A panel of nta-MIR6019/6020 genomic clones containing exon 1 and different lengths of upstream sequences were cloned ( S4A Fig ) and transiently expressed in N . benthamiana . Northern blot analysis showed that miR6019 driven by the 3 kb nta-MIR6019/6020a ( a-3k ) and the 2 kb nta-MIR6019/6020b ( b-2k ) promoter showed higher expression levels ( S4B Fig ) . Hence , we generated transgenic SR1 tobacco expressing GUS driven by these promoters ( Fig 4B ) . GUS staining assays performed on 1- , 3- and 6 WAG seedlings showed that the GUS expression was very high at 1 WAG and the expression gradually decreased at 3- and 6 WAG as determined by visual observation ( Fig 4C , left and middle panels ) and quantification of the average gray value ( Fig 4D ) . In agreement with the decreased GUS activity , qRT-PCR analysis showed decreased GUS mRNA levels in both GUS reporter transgenic plants ( Fig 4E ) . These results indicate that MIR6019/6020 transcription decreases as plants mature . The observed opposite expression pattern between miR6019 and the N gene suggest that miR6019/miR6020 could play a role in regulating N expression and N-mediated innate immunity during plant growth . To determine if the upregulation of N gene expression during plant growth is due to downregulation of nta-miR6019/6020 expression , we analyzed the two transgenic reporter lines N-CFPT2T1 and N-CFPt2t1 that we described previously [17] . Both lines express Cyan Fluorescent Protein ( CFP ) under the control of the same N gene promoter and 3' regulatory sequences . However , N-CFPT2T1 has the wild type binding sites for miR6019/6020 whereas N-CFPt2t1 has mutated binding sites ( Fig 5A ) . RNA samples were prepared from 1- , 3- and 6 WAG plants from N-CFPT2T1 and N-CFPt2t1 and the CFP transcript levels were determined by qRT-PCR . The relative CFP levels of 1 , 3 . 5 and 4 . 5 in N-CFPT2T1 plants at 1- , 3- and 6 WAG , respectively ( Fig 5B , left ) , demonstrated a gradual increase in expression during growth . This result is similar to the pattern of N gene expression ( Fig 3A ) . In contrast , the relative CFP transcripts levels in N-CFPt2t1 plants did not change significantly ( Fig 5B , right ) . However , miR6019 accumulation gradually decreased in both N-CFPT2T1 and N-CFPt2t1 plants ( S4C and S4D Fig , top panel ) . These results further support that miR6019/6020 regulate N expression during plant growth . To further investigate whether the decreased miR6019 levels seen during plant growth contribute to an increased level of N-mediated immunity , we generated transgenic SR1 plants that overexpressed wild type and the AXC mutant of the MIR6019/6020 cluster , which produce wild type 22-nt and mutated 21-nt miR6019 , respectively , with wild type miR6020 ( Fig 5C ) [17] , designated miR6019WTOE and miR6019AXCOE . The selected lines were crossed to TG34 expressing N and F2 progeny with homozygous N loci were selected ( miR6019WTOE/NN and miR6019AXCOE/NN ) . The sRNA northern blot analyses confirmed that the 22- and 21-nt miR6019 accumulated to 10-fold greater levels in miR6019WTOE/NN and to around 40-fold greater levels in miR6019AXCOE/NN plants compared to the miR6019 level in SR1 and TG34 plants ( Fig 5D ) . Notably , miR6019WT or miR6019AXC overexpression did not affect the growth phenotype of tobacco plants ( S5 Fig ) . We next analyzed the N transcript level in different miR6019WTOE/NN and miR6019AXCOE/NN lines at 6 WAG . As expected , the N transcript level in miR6019WTOE/NN was reduced to about 36% of the N transcript level in TG34 ( Fig 5E ) , while miR6019AXCOE/NN was reduced to about 61% of its level in TG34 plants ( Fig 5E ) . It is interesting to note that even though the level of 21-nt miR6019 in miR6019AXCOE is much higher than that of the 22-nt miR6019 in miR6019WTOE , downregulation of N expression by miR6019AXC was not as efficient as that by miR6019WT . Since the 21-nt miR6019AXC cannot trigger phasiRNA production as miR6019WT does , these results suggest that nta-miR6019-triggered phasiRNA production significantly potentiate miR6019-mediated repression of N expression . To determine if nta-miR6019 overexpression affects N-mediated resistance to TMV , we infected miR6019WTOE/NN and miR6019AXCOE/NN TG34 and SR1 plants at 6 WAG with TMV U1 . The miR6019WTOE/NN plants displayed strong systemic HR symptoms , whereas the miR6019AXCOE/NN plants showed only mild wrinkling symptoms in systemic leaves ( Fig 5F ) . These results provide further evidence to support a role for both miR6019 and miR6019-triggered N phasiRNAs in the negative regulation of N-mediated resistance to TMV . We extended our study on growth regulation of N-mediated TMV immunity to include our N- expressing transgenic D51 and control VF36 tomato lines [24] . D51 tomato plants at 1- , 3- and 6 WAG were infected with TMV and HR was observed on inoculated leaves of all plants at 7 DPI ( Fig 6A–6F ) . At 21 DPI , strong systemic HR was observed for plants infected at 1- and 3 WAG while complete resistance was observed for plants infected at 6 WAG ( Fig 6G–6I ) . The tomato infection experiment was also repeated three times and the average survival rates for infected and control plants were calculated . The survival rates were about 10% , 50% and 100% at 21 DPI for plants infected with TMV at 1- , 3- and 6 WAG , respectively , which is consistent with increased immunity as the plant matured ( Fig 6J ) . For VF36 plants without N , all plants survived at 21 DPI , although they showed TMV symptoms ( S6 Fig ) . These results indicate that N-mediated immunity is also regulated during tomato growth . We further investigated the mechanism of N regulation in tomato . The qRT-PCR analysis determined that the relative N-transcript levels were 1 , 5 and 15 in D51 tomato plants at 1- , 3- , and 6 WAG , respectively , while the levels remained low in all VF36 tomato plants ( Fig 7A ) . These results show that N-mediated immunity and N-expression are also subject to developmental regulation in tomato . Then slicer detector analysis was performed . Using N transcript sequences and the solanaceae small RNA databases [25] , we identified conserved miR6019/miR6020 in N . benthamiana and miR6020 in S . lycopersicum ( Fig 7B and 7C ) . Unlike its counterpart in tobacco , the sly-miR6020 is 22-nt , which may trigger phasiRNA synthesis similar to that with miR6019 . Small RNA sequencing analysis showed that sly-miR6020 accumulated at around 1 TPM at 1- , 3- and 6 WAG with a slight decreasing pattern ( Fig 7D ) . However , the secondary siRNAs mapped to N transcript sequences accumulated at levels comparable to those in TG34 tobacco plants and showed a clear decreasing pattern ( Fig 7D and 7E ) . To test whether sly-MIR6020 can produce 22-nt miR6020 and trigger phasiRNA production from N transcripts , we cloned the sly-MIR6020 foldback structure with flanking sequence into a binary vector driven by a 35S promoter ( S3 Data ) . As a positive control , we cloned an artificial miRNA using MIR171 as a backbone to express sly-miR6020 ( AMIR6020 , S3 Data ) . We used the N-CFPT2T1 and two miRNA sensor constructs , MS4miR6020 and MS4miR6020E ( S3 Data , Fig 8A and 8B ) , as reporters for N transcripts . N-CFPT2T1 and MS4miR6020 have wild type binding sites for sly-miR6020 , which has a mismatch and G-U wobble pairs that may interfere with target cleavage and triggering phasiRNA production ( Fig 8A and 8B ) . Meanwhile , MS4miR6020E has a mutated binding site that completely matches the sly-miR6020 and served as positive control for sly-miR6020 triggered phasiRNA production ( Fig 8C ) . The sly-miR6020 and reporters were co-expressed in N . benthamiana using Agrobacterium-mediated infiltration ( Fig 8C ) . We successfully detected sly-miR6020 expressed from both sly-MIR6020 and AMIR6020 constructs ( Fig 8C ) and confirmed that sly-MIR6020 generated both 22- and 21-nt miR6020 , whereas AMIR6020 generated only 22-nt miR6020 ( Fig 8C , top panel lane1-3 vs . 4–6 ) . Predicted phasiRNAs were detected in leaves co-expressing miR6020 and MS4miR6020E as expected ( Fig 8C , TAS5-8 , lane 2 and 5 ) . Both the N-CFPT2T1 and MS4miR6020 transcript targeted by sly-miR6020 or amiR6020 also produced predicted phasiRNAs ( Fig 8C , TAS1-4 , lane1 and 4; TAS5-8 , lane 3 and 6 ) . These results indicate that tomato sly-miR6020 may cleave the N transcript and trigger phasiRNA synthesis and thus was functionally similar to nta-miR6019 in tobacco .
In plants , NLR genes function as double-edged swords . Although they can recognize pathogen effectors and mediate effective immune responses to protect plants from disease caused by pathogens , NLR expression in the absence of pathogen pressure is accompanied by a fitness cost [26] and causes intraspecific genetic incompatibility [27] . miRNA-mediated repression of NLR gene expression provides a simple and general mechanism for plants to solve this problem yet maintain a large repertoire of NLR receptors because NLR-targeting miRNA genes can be generated during NLR gene duplication and diversification processes [17 , 28] . Consistent with this idea , many miRNAs were reported to be involved in NLR regulation [16–18] . Using bioinformatic analysis , we showed that about half and a quarter of tomato and tobacco NLR genes , respectively , were directly targeted by NLR silencers ( Fig 1A and S2A Fig ) . Considering that we used strict criteria to select only NLR silencers that could cleave NLR transcripts and had dRNA reads supporting the predicted cleavage site , and that NLR silencers could also direct translational repression of NLR without NLR transcript cleavage , a larger portion of NLR genes are expected to be under direct regulation by NLR silencers . Indeed , our analysis showed that over 98% of tobacco NLR genes and over 96% of tomato NLR genes spawned secondary siRNA synthesis ( Fig 1B and S2B Fig ) . These results support a broad impact for sRNA-mediated NLR regulation . In our previous study , we showed that miR6019 could specifically cleave N and its homologous transcripts , triggered secondary siRNA synthesis from cleaved N transcripts and attenuated N-mediated resistance against TMV in transient expression assays [17] . In this study , we found that N-mediated resistance to TMV in transgenic tobacco and tomato was incomplete during the young seedling stage and reached full resistance as the plant matured ( Fig 2 and Fig 6 ) . This process is accompanied by an increased level of N gene transcripts ( Fig 3 and Fig 7 ) . These findings are in agreement with the dose-dependence of CNL-mediated resistance to TMV among independent transgenic lines [29] . It also suggests that N-mediated resistance can be modulated at the N transcript level during plant growth . GUS staining and qRT-PCR analysis of two reporter genes driven by the N gene promoter revealed only moderate changes in N transcription during plant growth . In contrast , northern blot analysis of miR6019 and analysis of MIR6019 promoter reporter activity in transgenic plants showed a decreasing pattern of MIR6019 transcription and miR6019 accumulation as plants matured , which prompted us to hypothesize that miR6019/6020 could play a role in regulating N expression during plant growth . Comparing the expression pattern of the N reporter gene with or without the miR6019/6020 binding sites provided further evidence that dynamic changes in miR6019 levels could be associated with growth-dependent regulation of N expression . We also showed that overexpression of the 22-nt but not the 21-nt miR6019 significantly impaired N-mediated resistance . Together with the previous finding that the 22-nt but not the 21-nt miR6019 triggers phasiRNA synthesis [17] , these results point to a key role for phasiRNA in regulating N transcript levels and N-mediated innate immunity . Results from our N-TMV interaction experiments in D51 tomato are in agreement with this conclusion despite the possibility that phasiRNA could be triggered by the 22-nt miR6020 or derived from homologous N-like transcripts cleaved by other 22-nt miRNAs . Although pathogen infection has been reported to interfere with miRNA-mediated regulation of NLR [18 , 30 , 31] , whether this regulation is modulated during plant growth is unknown . Our results showed that N targeting miR6019/6020 is transcriptionally regulated during plant growth . We first demonstrated by high-throughput sequencing that accumulation of NLR silencers and NLR secondary siRNAs is downregulated during plant growth and accompanied by upregulation of their NLR targets ( Fig 1 and S2 Fig ) . Our viral infection experiments on susceptible plants showed that younger plants are more sensitive to developmental interferences . As NLR gene overexpression promotes abnormal development [32 , 33] , high levels of sRNAs that target NLR expression during early developmental stages can maintain low levels of NLR expression and in turn minimize the possibility of developmental defects caused by auto-activation of NLR-mediated immunity . Activation of immune responses often results in growth inhibition [32–35] and mechanisms underlying the trade-off between defense and growth are being actively investigated . Resource allocation theory was used to explain the growth-defense tradeoff in an ecological study of plant-pest interaction [36 , 37] . Studies of hormone cross-talk in the context of pathogen-triggered defense and growth inhibition showed that many important plant hormones , including auxin , jasmonate ( JA ) , gibberellins ( GA ) , salicylic acid ( SA ) , brassinosteroids ( BR ) and cytokinin ( CK ) , were all involved in balancing growth and defense during pathogen attack [34 , 38–40] . These studies were mostly done using a fixed time point during plant growth and thus could address how growth and defense were balanced at a certain time point . Our study on miR6019 regulation of N-mediated immunity during growth of tobacco and tomato plants provides a new angle to examine the trade-off between growth and defense and suggests that miRNAs can play a role in balancing growth and defense during plant growth by fine-tuning NLR expression . As reported previously , NLR genes are usually colocalized with various transposons in the plant genome and may be subjected to transcriptional silencing by transposon-derived siRNAs [41] . Moreover , NLR proteins are subject to negative regulation by SKP1-CULLIN1-F-box ( SCF ) complex-mediated stability control [42] . Thus , transcriptional and post-translational regulatory mechanisms may also contribute to regulation of growth and defense during plant maturation .
Tomato line D51 ( Solanum lycopersicum ) and tobacco lines SR1 ( Nicotiana tabacum ) , TG34 ( Nicotiana tabacum ) , Samsun NN ( Nicotiana tabacum ) , Nicotiana glutinosa , Nicotiana benthamiana were described previously [22 , 24] . N-CFPT2T1 and N-CFPt2t1 transgenic plants were described recently [17] . The other transgenic plant materials were generated in this study . All plants were grown in a growth chamber at 22 ± 2°C with a 16-h-light/8-h-dark photoperiod . Total RNA samples were prepared using the TRIzol reagent ( Invitrogen ) from the aerial parts of TG34 and D51 plants at 1- and 3 WAG and fully expanded leaves from plants at 6 WAG that were grown in soil . Paired-end mRNA libraries were prepared from total RNA samples using NEBNext Ultra RNA Library Prep Kit for Illumina ( NEB , USA ) according to the manufacture’s instruction and was sequenced on an Illumina HiSeqX TEN platform using the PE150 sequencing mode . For data analysis , tomato genome annotation ITAG2 . 4 and tobacco genome BX , K326 and TN90 were used [43 , 44] . Fastq data were mapped to the genome using bowtie2 and the length distribution of the library insert was analyzed by Picard ( CollectInsertSizeMetrics ) . mRNA expression values were determined using a perl script in Trinity , which calls bowtie and RSEM to do the mapping and calculation . mRNA expression was further normalized based on the TMM model using Trinity and its R module ( edgeR ) . Reproducibility between biological replicates was estimated by the R program with plot methods ( S1 Fig ) . sRNA sequencing library was prepared using NEBNext Multiplex Small RNA Library Prep Set for Illumina ( NEB , USA ) and was sequenced on an Illumina HiSeq2500 platform using SE50 sequencing mode . Fastq data was processed with an in-house perl script to remove adapter sequences , retrieve sRNA read sequences and read number . Reproducibility between biological replicates was also estimated by the R program with plot methods ( S1 Fig ) . The degradome RNA library was prepared as described earlier [25 , 45] and sequenced on an Illumina Genome Analyzer ( QB3 , UC Berkley ) using SE50 mode . NLR gene was extracted from annotated tobacco and tomato genomes and combined with results from HMM search using NBS , LRR , TIR and CC domain consensus sequences . The number of tomato NLR genes identified is consistent with a previous report on the total number of TNL- , CNL- and NL-type NLRs and their full-length NLR genes [46] . To identify the NLR silencers , sRNAs of 20- to 24-nt and TPM greater than 1 were extracted from each tobacco and tomato sRNA databases ( S4 Table ) . NLR-sRNA pairs were identified using the SlicerDetector Perl program . The degradome RNAs were mapped to the NLR transcripts using the dRNAmapper perl program . The NLR silencer-NLR pair with dRNA read support was obtained using the SmartCompare Perl program [25] . For NLR secondary siRNA detection , all small RNAs of 20- to 22-nt were aligned to NLR transcripts using bowtie with 0 mismatches and counted with an in-house perl program . Subsequent data integration and statistics were carried out using in-house perl and R scripts . All constructs used in this study were described in S3 Data . To investigate their transcriptional regulation , the promoters of MIR6019/6020a , MIR6019/6020b , and N gene were amplified and inserted into the pCambia1381Xb vector ( Cambia ) upstream of the GUS reporter gene . After sequencing , all constructs were introduced into tobacco cv . SR1 by Agrobacterium-mediated leaf disc transformation . After screening on MS medium with 10 mg/L hygromycin , positive transformants were used for subsequent analysis . To analyze the promoter activity of three nta-MIR6019/6020a ( a-1k , a-3k , a-5k ) and two nta-MIR6019/6020b ( b-2k , b-4k ) fragments , each sequence fragment fused with Exon1 of nta-MIR6019/6020a or b were amplified and inserted into the empty pH7Lic14 . 0 vector . After sequencing , all constructs were tested in tobacco N . benthamiana by Agrobacterium-mediated infiltration . Primers used in the experiments are listed in S3 Table . TMV U1 strain was propagated in the TMV susceptible SR1 tobacco . Viral inoculations were performed as described [47] . For plants at the 1 WAG stage , the seedlings were very small . We generated wounds using needles on the cotyledons or leaves , and then covered them with wet gauze immersed in the viral sap . Inoculated plants were placed in the incubator at 22°C . Total RNA was extracted from leaf tissue using the TRIzol reagent ( Invitrogen ) according to the manufacture’s protocol . sRNA northern blot analysis was performed as previously described [41] . For nta-miR6019 northern blot , we used the locked nucleic acid ( LNA ) modified oligonucleotide probe as previously described [48] . Probe sequences are listed in S3 Table . Quantitative real-time PCR was carried out using SYBR Green fluorescence and a Light Cycle 96 machine ( Roche ) . The threshold cycle ( Ct ) value was automatically calculated by the Roche Light Cycle 96 1 . 1 system software and the ΔΔCt method was used to calculate the relative expression levels [49] . GAPDH was used to normalize the expression of genes in various RNA samples . Three independent biological replicates and three technical replicates of each sample were used for quantitative PCR analysis . Primers used in the experiments are listed in S3 Table . All plants were sown on the 1/2MS medium . The histochemical GUS staining was performed as described [50] . After staining , samples were photographed using a NIKON D3300 digital camera . The value of GUS intensity was quantified using the PIL ( Python Imaging Library ) based on the average RGB values of each pixel in each image . Then the average grey value of each image was obtained using the "Color turn Gray" formula: Gray = R*0 . 299 + G*0 . 587 + B*0 . 114 . For this method , the gray value is smaller for the darker color and the gray value of white color is assigned the maximum value of 255 . The GUS intensity value was obtained with the formula: Intensity = 255—Gray value . Raw reads of Illumina RNA-seq libraries generated in this study are available from the Sequence Read Archive ( SRA ) at NCBI ( http://www . ncbi . nlm . nih . gov/sra/ ) under the accession number SRP125463 . These data have also been deposited in the genome sequence archive in the BIG Data Center [51] under accession number CRA000618 that are publicly accessible at http://bigd . big . ac . cn/gsa . | In plants , nucleotide-binding ( NB ) leucine-rich repeat ( LRR ) receptors ( NLR ) mediate pathogen-specific effector triggered immunity and are widely used in breeding to generate pathogen-resistant crops . However , dysregulation of NLR expression can inhibit plant growth and how NLR expression and function are regulated in different stages of plant growth is poorly understood . Using a high-throughput sequencing and bioinformatics approach , we found an overall increase in NLR expression , but expression of NLR-targeting sRNA during plant growth was decreased . We also used resistance to tobacco mosaic virus ( TMV ) mediated by the resistance gene N as a model system to study the biological significance of growth regulation of NLR by miRNAs . We found that N-mediated TMV immunity strengthened and N transcript levels increased during plant maturation . Using genetic analysis , we showed that up-regulation of N was due to transcriptional down-regulation of the N-targeting miR6019/6020 cluster during plant growth . We also showed that sRNA-mediated growth regulation of N expression and function was conserved between tobacco and tomato plants . This study therefore reveals a role for miRNAs in regulating innate immunity during plant growth . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"biotechnology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"plant",
"growth",
"and",
"development",
"gene",
"regulation",
"pathogens",
"microbiology",
"tobamoviruses",
"viruses",
"developmental",
"biology",
"plant",
"science",
"micrornas",
"rna",
"viruses",
"genetically",
"modified",
"plants",
"plant",
"pathology",
"plants",
"flowering",
"plants",
"genetic",
"engineering",
"tobacco",
"mosaic",
"virus",
"small",
"interfering",
"rnas",
"genetically",
"modified",
"organisms",
"nicotiana",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"tomatoes",
"fruits",
"biochemistry",
"rna",
"eukaryota",
"plant",
"viral",
"pathogens",
"nucleic",
"acids",
"viral",
"pathogens",
"plant",
"pathogens",
"genetics",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna",
"plant",
"biotechnology",
"organisms"
] | 2018 | A role for small RNA in regulating innate immunity during plant growth |
Gliomas are the most common form of malignant primary brain tumors in humans and second most common in dogs , occurring with similar frequencies in both species . Dogs are valuable spontaneous models of human complex diseases including cancers and may provide insight into disease susceptibility and oncogenesis . Several brachycephalic breeds such as Boxer , Bulldog and Boston Terrier have an elevated risk of developing glioma , but others , including Pug and Pekingese , are not at higher risk . To identify glioma-associated genetic susceptibility factors , an across-breed genome-wide association study ( GWAS ) was performed on 39 dog glioma cases and 141 controls from 25 dog breeds , identifying a genome-wide significant locus on canine chromosome ( CFA ) 26 ( p = 2 . 8 x 10−8 ) . Targeted re-sequencing of the 3 . 4 Mb candidate region was performed , followed by genotyping of the 56 SNVs that best fit the association pattern between the re-sequenced cases and controls . We identified three candidate genes that were highly associated with glioma susceptibility: CAMKK2 , P2RX7 and DENR . CAMKK2 showed reduced expression in both canine and human brain tumors , and a non-synonymous variant in P2RX7 , previously demonstrated to have a 50% decrease in receptor function , was also associated with disease . Thus , one or more of these genes appear to affect glioma susceptibility .
Gliomas are the most common form of malignant primary brain tumors in humans , characterized by rapid growth and the invasion of neoplastic cells into healthy brain . Despite aggressive therapy , malignant gliomas are rarely curable . In the USA , the yearly mortality rates for primary malignant brain tumors are 5 . 6 and 3 . 7 per 100 , 000 in men and women , respectively [1] . To predict the biological behavior of the neoplasm and to standardize therapy regimes , brain tumors are classified and graded by the World Health Organization ( WHO ) according to location and histopathological appearance [2] . Mortality differs significantly by histology and age [1] , with 2 year survival rates of less than 15% for the most aggressive and most common histological subtype , glioblastoma ( GBM ) [3] . Studies of syndromes and familial aggregation have suggested genetic susceptibility to gliomas , and although rare inherited mutations account for only few cases they are important for identifying pathways for gliomagenesis according to Brain Tumor Epidemiology Consortium [1] Spontaneous gliomas in dogs are usually classified and graded using the human WHO criteria , [2] and have striking similarities to their human tumor counterparts at the biological and imaging levels [4 , 5 , 6 , 7 , 8] . Central nervous system tumors occur in dogs at an incidence of around 15 per 100 , 000 animals or 2–4% of necropsy cases , with gliomas representing approximately 35% of all CNS primary tumors [5 , 6 , 9] . This is similar to human patients where gliomas represent 24% of all primary CNS tumors with an incidence of approximately 20 per 100 , 000 [3] Extensive analysis of gliomas in humans has defined commonly disrupted pathways involving the receptor tyrosine kinases/PI3K/RAS , TP53 and RB1 pathways [10 , 11] . Defining molecular subclasses of glioma will likely guide future therapeutic and prognostic stratification [10 , 12] . Although the frequency of specific glioma subtypes varies between humans and dogs , with humans having more high grade glioblastomas and dogs having more high grade oligodendrogliomas [3 , 4 , 5] , ppreliminary analysis suggests that the same key pathway abnormalities are also present in gliomas in dogs [4 , 13 , 14 , 15 , 16] . Dogs are excellent spontaneous models of human complex diseases including cancers , by sharing both genetic and environmental factors . In addition , the recent breed creation events , resulting in certain diseases becoming overrepresented in specific breeds , have made disease gene mapping easier than in human populations . Genome wide association studies in dogs have been highly successful owing to long-range linkage disequilibrium ( LD ) , available SNP genotyping tools and a robust genome assembly [17 , 18] . There are many examples of successful GWAS in dogs for Mendelian traits [17 , 19 , 20 , 21 , 22 , 23] as well as for more complex traits [24 , 25 , 26 , 27 , 28] . In addition , across-breed mapping studies by us and others have successfully identified biologically relevant loci for several traits and related diseases that are fixed within breeds [29 , 30 , 31 , 32] . Some brachycephalic dog breeds have been reported to have a considerable elevated risk of glioma , such as Boxer ( relative risk ~23 ) , Bulldog and Boston Terrier ( relative risk ~5 ) [6 , 33] . Even though all brachycephalic breeds are likely to share a common major mutation that causes this phenotype [30] all brachycephalic breeds are not reported to be at higher risk of developing glioma . A study using neighbor-joining trees on genome wide SNP-data for dog breeds has shown that Boxer , English Bulldog , Boston Terrier and French Bulldog are closely related [34] and likely to share a recent common ancestor . Other brachycephalic breeds such as Pug and Pekingese are closely related to each other [34 , 35] but not as recently related to the high-risk glioma breeds . We hypothesized that genetic risk factors for glioma have been segregating in an “ancestral bulldog” line . Because of the extraordinary high risk of all types of glioma in the Boxer , we hypothesized that genetic risk factors might be in a nearly fixed region of the Boxer genome . Since it has been reported that the risk for glial tumors in dogs increases with age until 10–14 years [33] , and the possibility of a nearly fixed risk locus , we concluded that the traditional affected—nonaffected within breed association mapping approach would not be suitable . Instead we hypothesized that some genetic risk factors for glioma might be shared between breeds and given the suggested recent relationship for some of the breeds at high risk , an across-breed mapping approach comparing glioma-cases from several breeds to controls from several breeds could identify these risk factors . Given the similarities between human and canine gliomas at the histological and genetic level , we hypothesized that this approach could identify genes or pathways that may also be relevant for human glioma .
During the period of 1987–2013 a total of 228 gliomas were histologically diagnosed at UC Davis . Among these , oligodendrogliomas were overrepresented in English Bulldog , French Bulldog , Boston Terrier , and Boxers; astrocytomas were overrepresented in Boston Terrier , Boxer , Jack Russell Terrier and Pit Bull Terrier . When all gliomas including mixed oligoastrocytomas were evaluated , Mastiffs also had an overrepresentation of tumors ( Table 1 ) . We performed genome-wide association mapping , using 39 glioma cases diagnosed with varying types and grades of glioma and 141 controls comprising 25 different dog breeds and 4 mixed breed dogs ( Table 2 , S1 Table ) . The controls were selected to represent several breeds with a few individuals from each breed to maximize the power of the study , and to decrease bias due to population stratification . Based on the uneven number of cases and controls between breeds , we expected a presence of population structure and applied the genomic control ( GC ) option in PLINK [36] to correct for population structure . After GC correction , a major locus on CFA 26 ( top SNP = 9 , 780 , 187 , CanFam2 . 0 , praw = 5 . 47x 10−38 , pgc = 2 . 77 x 10−8 , Fig 1A ) was identified . Visual inspection of graph Fig 1B limits the associated region to ≈CFA 26: 8 . 5–12 Mb . ( Five most associated SNPs in S2 Table ) . Bonferroni correction was applied to calculate a significance threshold adjusted for multiple comparisons . Since the Boxer breed had the largest number of cases ( n = 9 , 23% ) , a separate association analysis was performed excluding the representatives of this breed from the dataset , in order to check for an excess impact from this population . The same region on CFA 26 was identified ( praw = 9 . 7 x 10−28 ) . After correction by genomic control the association was in fact stronger after removal of the Boxers ( pgc = 9 . 1 x 10−11 , S2 Fig ) , reflecting a lower genomic inflation . A Quantile Quantile ( QQ ) plot showed low remaining inflation , with p-values starting to slightly deviate from the expected curve at–logP of 2 , but with a much sharper deviation at expected significance level of–logP 6 . 5 ( S3 Fig ) . We calculated pairwise identity-by-state genetic distances between all samples using PLINK [36] , and then constructed a phylogenetic tree using the software Phylip . The results ( S1 Fig ) supported previous studies [34 , 35] reporting a close relationship for the four breeds: Boxer , English Bulldog , French Bulldog and Boston Terrier . Reduced heterozygosity in the dog genome can occur because of genetic drift or selection . Homozygous blocks longer than 1 Mb have been proposed to be more likely to have arisen through selection than drift [29] . We examined the minor allele frequency ( MAF ) on CFA 26 to check for signs of selection in the region associated with glioma . The MAF for the four most recently related breeds of the ancestral bulldog line was compared to the MAF for Pugs ( Fig 2 ) . A common completely homozygous region in the ancestral Bulldog line spanning ~ 4 Mb ( Fig 2B ) including the glioma-associated region ( Fig 2A ) was identified . An additional ~ 8 Mb flanking this region showed reduced heterozygosity . This can be interpreted as a likely selective sweep having taken place in the ancestral Bulldog line leaving traces behind in descending modern breeds . Since it is known that brachycephaly is a trait that has been under selection we made a comparison to Pugs to investigate if this was a common sweep in brachycephalic breeds . Pugs showed no comparable homozygosity or signs of selection in this region ( Fig 2C ) , supporting their difference in ancestry . In this study the sweep was identified using dogs with glioma . Other reports support that this region has been under selection in a normal cohort of Boxers [37] and English Bulldogs [29] . This locus was also identified as the second most associated in our previous study identifying a major locus for brachycephaly [30] , which could be explained by its presence in only a subpopulation of brachycephalic breeds . In order to find putative genetic risk factors for glioma we performed targeted re-sequencing using NimbleGen capture followed by Illumina sequencing . In total 3 . 4 Mb were re-sequenced spanning 8 . 5–11 . 9 Mb on CFA26 . The sequencing was performed in two experiments . In the first experiment the most homozygous region on CFA26 ( 8 . 5–9 . 2 Mb and 10 . 8–11 . 9 Mb ) was sequenced in six dogs; three brachycephalic ( one Pug and two Boxers ) and three control breeds ( Dachshund , Welsh Corgi , and Basset Hound ) . The focus was on identifying differences between brachycephalic and non-brachycephalic dogs . In the second experiment the target region on CFA26 was extended . Four dogs were sequenced: three brachycephalic dogs diagnosed with glioma and one Dachshund as a control . The two experiments were analyzed together to identify potential risk variants for glioma . In addition a pool of dog-breeds from a whole genome sequencing project were utilized as controls ( See Materials and Methods ) . In total , 6 , 957 Single nucleotide variants ( SNV ) were identified in the re-sequenced region when mapping the reads against the Boxer reference genome sequence [18] . Out of these , 490 SNVs were located within conserved elements ( +- 5bp ) . In addition , we identified one large structural polymorphism . An insertion ( in the Boxer reference genome sequence ) of ~2 , 200 bp on CFA26: 9 , 550 , 700–9 , 552 , 900 was identified in the glioma-associated region by showing a total lack of coverage for the Dachshund , but normal coverage for the brachycephalic/glioma individuals . Candidate variants were selected from the re-sequencing data for a replicate study to evaluate disease association . Since a mutation in a region that is conserved across species is more likely to have a function , we used SEQscoring [38] in the selection process to score the SNVs by conservation and to rank them according to differences between cases and controls . A selection of 56 candidate SNVs were genotyped in a total of 168 dogs using Sequenom . Association tests were performed in three different ways using PLINK [36] . In the first analysis , all dogs were included with their original status as cases ( n = 34 ) or controls ( n = 134 ) . After the first analysis we calculated the allele frequency for all breeds separately for the two most significant SNVs ( S8 Table ) . We concluded that Boxer , English Bulldog , and Boston Terrier seemed fixed at these positions ( frequency 0 . 93–1 . 00 for risk alleles ) . In a second analysis we removed all samples from these fixed breeds to investigate if the same genes would be the most associated in the remaining 21 cases and 113 controls . In a third analysis we instead assumed that if the samples from the high risk breeds that seemed fixed actually all carried the risk factor then it would make sense to perform a test where we assign all these samples a status as cases ( n = 55 ) In all three tests the most associated SNV was identified at position CFA26:10 , 893 , 462 located in an intron of the gene Calcium/calmodulin-dependent protein kinase kinase 2 ( CAMKK2 , alias CAMKK , CAMKKB ) ( Fig 3A , 3B and 3C ) ( with p-values p = 3 . 67e-09 ( A ) , 1 . 04e-05 ( B ) , 4 . 75e-26 ( C ) . The second most associated SNV was identified at position CFA26:9 , 722 , 698 in the first ( p = 2 . 98e-08 ) and the third test ( p = 5 . 04e-26 ) , located in an intron of the gene Density-regulated protein ( DENR ) . The position was conserved among placental mammals [39] . In the second test the second most associated SNV was identified at position CFA26:10 , 969 , 340 ( p = 1 . 05e-05 ) located in an intron of the gene Purinergic receptor P2X , ligand gated ion channel 7 ( P2RX7 ) . This position was not conserved , but another position located in the third exon of this gene is conserved and causes a non-synonymous codon change ( CFA 26: 10 , 984 , 721 ) . In the third test , the non-synonymous SNV in P2RX7 was more associated ( p = 6 . 70e-22 ) than the intronic SNV ( CFA26:10 , 969 , 340 p = 8 . 89e-18 ) in this gene . This non-synonymous polymorphism causes an amino acid change from phenylalanine ( F ) to leucine ( L ) ( p . Phe103Leu ) . The exchanged amino acid is involved in the extracellular loop of the trans-membrane protein P2RX7 . We selected the three SNVs: CFA26:10 , 893 , 462 , CFA26:9 , 722 , 698 and CFA26:10 , 984 , 721 in the three genes CAMKK , DENR and P2RX7 for further analysis . ( P-values reported in the evaluation of SNPs are obtained from basic Chi-square tests with no further correction . ) Individual genotypes for the candidate SNVs can be found in S5 Table . In addition , a candidate structural variant was evaluated for disease association . The ~2 , 200 bp insertion on CFA26: 9 , 550 , 700–9 , 552 , 900 was genotyped in 147 dogs ( 32 cases and 115 controls ) . The insertion was shown to be fairly common in several breeds , and was much less associated ( p = 0 . 001 , S5 Table ) with glioma than evaluated SNVs in the region . To further investigate if the identified variants were truly associated with glioma , we selected six breeds that were segregating for the identified variants . To better avoid confounding breed effects , we added more healthy controls from the same breeds . In total we genotyped 15 cases and 119 controls ( Table 3 ) for three the selected SNVs , located in CAMKK , DENR and P2RX7 . The odds ratios combined across all six breeds were CAMKK: 8 . 7 , P2RX7: 7 . 8 , DENR: 7 . 1 Individual genotypes for the SNVs can be found in S9 Table . To evaluate the effect of potential candidate variants on the expression of genes in the region of interest we performed quantitative PCR experiments on tumor and matched normal cerebral samples from the same animals . A reduced expression of CAMKK2 was seen in tumor versus normal samples ( n = 6 , p = 0 . 03 ) in dogs with the three SNPs glioma risk haplotype ( GRH ) ( Fig 4A ) . The risk haplotype also appeared to induce a non-significant ~2-fold lower expression in normal tissue in dogs with the risk haplotype ( n = 6 ) versus dogs without the risk haplotype ( n = 4 , p = 0 . 22 , Fig 4B ) . No significant changes were seen for DENR . P2RX7 showed an increased mRNA expression in tumors versus normal tissue ( n = 6 , p = 0 . 04 ) in dogs with the risk haplotype ( Fig 5A ) while there was no significant difference in normal tissue in dogs with the risk haplotype ( n = 6 ) versus dogs without the risk haplotype ( n = 4 , p = 0 . 20 , Fig 5B ) . P2RX7 protein expression was assessed in matched and unmatched normal cerebrum and tumors ( Fig 5C and 5D ) by western blotting and was detected in all canine normal brain samples and in 16/17 glioma samples . Major bands previously reported to represent glycosylated ( ~75kD ) and un-glycosylated ( ~60kD ) protein were seen , as well as a consistent band at around 50kD that was present in all samples . No significant difference in total protein levels was detected . However , normal brain samples consistently expressed only the 60kD band , while the majority of tumor samples expressed the 75kD and other higher molecular weight bands with minimal expression of the 60kD band . Based on the expression changes seen in canine tumors , we examined human high-grade gliomas for similar expression changes . Analysis of an available array dataset [40] revealed a significantly lower mRNA expression of CAMKK2 in infiltrating astrocytic tumors ( n = 5 , p = 0 . 01 ) , ependymoma ( n = 4 , p = 0 . 002 ) , and oligodendroglioma ( n = 5 , p = 0 . 03 ) compared to normal brain tissue ( n = 4 , Fig 4C ) . Using a subset of data from the Cancer Genome Atlas ( TCGA ) [11] consisting of mRNA expression data of surgical specimens from 24 glioblastoma patients , and 10 non-tumor control brains ( epilepsy resections ) , we confirmed the lower mRNA expression of CAMKK2 in high-grade glioma ( Fig 4D ) . We found no significant difference for DENR or P2RX7 expression in these same dataset ( S4 Fig ) . Furthermore , we analyzed the levels of CAMKK2 expression in seven human glioblastoma patient-derived cell lines , maintained in neural stem cell medium under serum-free conditions [41] and found that all tumors showed 20–60% lower level of expression of CAMKK2 protein compared to normal frontal cortex ( Fig 4E ) . Because CAMKK2 expression is associated with a mature cellular phenotype [42] we next treated two of the glioblastoma cell lines with 5% serum for seven days to induce their differentiation . For both cell lines CAMKK2 mRNA expression increased 2–3 fold with differentiation ( Fig 4F ) , suggesting that CAMKK2 expression is reduced in the more stem cell like tumor cells . There was a trend towards lower DENR protein expression in GBM compared to normal brain , but P2RX7 showed no conclusive difference in protein expression .
In this study we successfully identified a locus strongly associated with glioma across several dog breeds . Given that the power of canine disease gene mapping typically results from the enrichment of genetic risk factors for specific diseases within a breed , and that disease mapping tools have been designed for within breed mapping , this means that across-breed mapping studies have been challenging so far . Here we took advantage of the fact that glioma has an increased frequency in several related brachycephalic breeds derived from an ancestral Bulldog , the Boxer , English Bulldog , French Bulldog and the Boston Terrier . The glioma-associated region resides in a region showing a sweep in these brachycephalic dogs , suggesting that the across-breed mapping is made feasible by the long selected haplotype in this region . We hypothesize that genetic risk factors for glioma have been enriched in these breeds related to an ancestral Bulldog and have either hitch-hiked with a desirable trait during selective breeding or are pleiotropic effects resulting from selected genetic variants . The risk allele was also present in six additional breeds affected with glioma ( the Australian Cattle Dogs , West Highland White Terrier , Labrador Retriever , Mastiff , Pit Bull Terrier and the American Staffordshire Terrier ) and conferred a significant risk within those breeds . Re-sequencing and fine mapping of the region on CFA 26 associated with glioma identified three candidate genes; DENR , CAMKK2 and P2RX7 with the strongest associated SNV located within an intron of CAMKK2 . Intriguingly , all three genes are potentially relevant candidate genes for cancer development and contain highly associated SNVs , offering the possibility that multiple variants contribute to disease at this extended locus . The CAMKK2 gene encodes a Ca2+/calmodulin-activated kinase , which is highly expressed in the adult brain [42] . Following an increase in intracellular Ca2+ , CAMKK2 activates CAMKI , CAMKIV , Akt and AMP-activated protein kinase ( AMPK ) in a number of pathways [43 , 44] . It has been shown that elevated intracellular Ca2+ stimulates ERKs with a requirement for CAMKK2 acting through CAMKI and via RAS [45] . In our study , CAMKK2 expression was significantly lower compared to normal brain in both canine and human gliomas , making it an attractive candidate gene for further investigations . CAMKK2 has been shown , together with CAMKIV , to be involved in cerebellar granule precursor migration and differentiation during normal development [42] . Since we found that differentiation of glioblastoma stem cells was correlated with a higher expression of CAMKK2 we suggest that CAMKK2 down regulation may render tumor cells more stem cell-like thereby increasing the aggressiveness of the tumor . It was recently reported that inhibition of CAMKK2 blocks migration of medulloblastoma via CAMKI , and CAMKK2 was proposed as a putative target to limit metastasis in this type of brain tumor [44] . On the other hand , CAMKK2 is a versatile activator of signaling pathways and in non-small cell lung cancer [46] its role in AMPK activation was proposed as a mechanism for tumor regression . Because normal brain , in both dogs and humans , express higher levels of CAMKK2 than glioma , we propose that targeting CAMKK2 should involve its activation rather than inhibition . The DENR gene encodes the Density-regulated protein , which acts together with the oncogene multiple copies in T-cell lymphoma-1 ( MCT-1 ) in translation initiation [47] . Because DENR inactivation in Drosophila is lethal due to impaired histoblast proliferation , the DENR-MCT-1 complex was suggested to regulate translation of specific mRNAs , presumably from “cancer-relevant” genes , i . e . those involved in cell growth [48] . While no significant expression changes were seen in this study , DENR may be involved in a translational control system with a key role in supporting proliferation and tissue growth . P2RX7 is a trimeric ligand-gated cation channel that mediates numerous downstream events following activation by extracellular adenosine 5'-triphosphate ( ATP ) . The non-synonymous canine SNV associated with glioma in this study has recently been shown to result in approximately 50% decrease in P2RX7 function [49] , and non-synonymous SNPs in P2RX7 have been identified in an increasing number of human patients with a variety of conditions including cancer [50] . P2RX7 receptors are expressed in a wide variety of immune cells including microglia , the primary antigen-presenting cell of the central nervous system and P2RX7 function is important for IL-β release and downstream priming of IFNγ producing CD8+ T cells involved in adaptive immunity against tumors [51 , 52 , 53] . Abrogation of P2RX7 has been associated with decreased response to foreign material in graft versus host models , increased metastatic potential in human breast cancer and increased susceptibility to colon and epithelial cancers [51 , 54 , 55 , 56] . However , the role of P2RX7 in glioma is complex and expression has been shown to result in both suppression and an increase in glioma growth in a variety of models , suggesting additional effects beyond immune surveillance [53 , 57 , 58 , 59 , 60] . In addition to the higher P2RX7 mRNA levels that we detected in dog glioma , we report an intriguing pattern of differently sized proteins occurring between tumor and non-tumor dog brain . It remains to be investigated how these relate to a potential glioma susceptibility mechanism . In conclusion , this study identifies a locus associated with canine glioma that has likely been under selection in brachycephalic dog breeds related to the original Bulldog , where highly disease-associated SNVs are found in three neighboring candidate genes . Candidate functional consequences were observed for two of the genes suggesting that the glioma susceptibility may be conferred by multiple variants within this locus .
Blood samples were collected from canine patients of the University of California at Davis William R . Pritchard Veterinary Medical Teaching Hospital ( VMTH ) . In total , 39 glioma cases of differing types and grades ( as determined by a board certified veterinary pathologist ) and 142 controls comprising 25 different dog breeds and 4 mixed breeds were collected for the GWAS ( Table 1 , S1 Table ) . The controls were selected to represent several breeds with a few individuals from each breed to maximize the power of the study , and to decrease bias due to population stratification . Frequency of glioma types within specific breeds was determined for all histologically diagnosed gliomas during the years 1987–2013 . A likelihood ratio chi-square test was used to compare the presence or absence of individual tumor types with breed relative to the VMTH population for each breed . A p-value of <0 . 05 was used to define unusually large or small breed associations based on the distribution of breeds examined . All samples were genotyped on the Illumina 170 K canine SNP array [29] . Association analysis was performed using the software package PLINK [36] calculating single marker chi-square association . Data quality control was performed to ensure a minor allele frequency ( MAF ) > 0 . 05 and call rate > 95% for both SNPs and individuals to be retained . One individual was removed because of low genotyping . After frequency and genotyping pruning , there were 143 , 007 SNPs left . The association calculations were further corrected for stratification by the use of genomic control ( GC ) . PLINK was also used to format data for a QQ plot ( S3 Fig ) and MAF graphs ( Fig 2 ) . Significance level corresponding to p-value 0 . 05 after adjusting for 143 K test with Bonferroni correction was calculated and set to a log p-value of ≈6 . 5 To evaluate the relationship between breeds we calculated pairwise identity-by-state genetic distances between all samples using PLINK [36] , and then constructed a phylogenetic tree using the Neighbor-Joining/UPGMA method of software Phylip ( version 3 . 695 , Joseph Felsenstein , University of Washington , Seattle ) . Targeted massive parallel re-sequencing was performed in two experiments . In the first experiment a total of ≈1 . 8 Mb was targeted ( CFA26: 8 , 534 , 645–9 , 176 , 011 , CFA26:10 , 800 , 000–11 , 900 , 000 ) and in a second experiment the region at CFA26 was extended and a total of ≈3 . 4 Mb was targeted ( CFA26: 8 , 500 , 000–11 , 900 , 000 ) . Fragment libraries were prepared as described in Olsson et al . [61] . Sequence capture was performed using a 385K custom-designed probe array from Roche NimbleGen according to the manufacturer’s instructions . Captured enriched libraries were sequenced using Illumina sequencing technology . In the first experiment , six dogs: three brachycephalic ( 2 Boxer , 1 Pug ) and three control breeds ( 1 Dachshund , 1 Welsh Corgi , 1 Basset Hound ) were sequenced with a read length of 60bp ( single end reads ) , using Genome Analyzer II . In the second experiment , four dogs: three brachycephalic+glioma ( 1 French Bulldog , 1 English Bulldog , 1 Boston Terrier ) and one control ( 1 Dachshund ) were sequenced with a read length of 100bp ( paired end reads ) , using HiSeq 2000 . Information about breed and health status for the individual dogs that were sequenced is shown in S3 Table . Obtained reads were mapped to CanFam 2 . 0 [18] using Burrows-Wheeler Aligner ( BWA ) [62] . SAMtools [63] was used for variant calling using mpileup format . Recommended setting for BWA reads -C50 was used as a coefficient to downgrade mapping quality , and–E to increase sensitivity . For coverage calculation the SEQscoring tool was used with SAMtools pileup format ( S3 Table ) . Before variant calling , PCR duplicates were removed using the tool Picard ( hosted by SAMtools ) . The presence of structural variants was investigated by comparing coverage for cases and controls using SEQscoring [38] and IGV [64] . To select SNVs for genetic validation in additional dogs , we used SEQscoring [38] to score variants according to conservation ( SiPhy constraint elements detected by the alignment of 29 eutherian mammals [39] , phastCons [65] multiple alignment from UCSC of the human ( hg17 ) , mouse ( mm6 ) , rat ( rn3 ) , and the dog ( canFam2 ) and a multiple alignment of 16 amniota vertebrates [66] from Ensembl release 56 ) . SEQscoring [38] was also used to calculate pattern scores in order to rank the SNVs by association to phenotype . Pattern scores are based on pairwise comparison of all individuals for all variants , where variants are scored based on similarities and differences between cases and controls . Genotypes from a whole genome re-sequencing study of six pools of breeds were included as controls . The pools consisted of 12 wolves and 60 dogs from 14 other breeds as described in a study by Axelsson and colleagues [67] . When calculating the pattern score , we used three different sets of individuals . In the first set we evaluated the region that was the most associated with glioma ( CFA26:9 , 176 , 012–10 , 799 , 999 ) with available samples ( three glioma-diagnosed cases , one individual control and six control pools ) . In the second set we accounted for the possibility of a risk factor that might be fixed in breeds related to the ancestral Bulldog , and thus classified the breeds Boxer , French Bulldog , English Bulldog and Boston Terrier as cases . Four of the brachycephalic dogs had been diagnosed with glioma ( S3 Table ) . The entire region was included in the third set ( CFA26:8 . 5–11 . 9 Mb ) . A total of 56 SNVs ( 31 conserved and 25 non-conserved ) were successfully genotyped using iPLEX Sequenom MassARRAY platform in a total of 168 dogs ( 34 cases and 134 controls ) ( S4 and S5 Tables ) . Association between SNV allele and phenotype was evaluated using PLINK [36] chi-square calculations . Primary tumor tissue was obtained at necropsy or from surgical biopsy of clinical cases presented to the VMTH . Necropsy samples were collected within 20 min after death . Control normal cerebral tissue was collected from contralateral cerebral hemispheres in tumor bearing dogs , and from neurologically normal dogs . All samples were snap-frozen in liquid nitrogen for storage . Samples of adjacent tumor tissue were paraffin-embedded and processed for histological analysis . All tumors were histologically classified by a board-certified pathologist according to the WHO classification of human tumors of the central nervous system [2] . All canine samples were obtained with their owner's consent , and in strict accordance with good animal practice , with study protocols approved by the Institutional Animal Care and Use Committee ( IACUC ) at UC Davis . Total RNA was isolated from whole blood of control and affected dogs using QIAamp RNA Blood Mini Kit ( QIAGEN , Valencia , CA ) . The optional on column DNase treatment was carried out to eliminate gDNA contamination . 100ng of total RNA was used for cDNA synthesis with the SuperScript III First-Strand Synthesis System for RT-PCR ( Life technologies , Grand Island , NY ) . Primers for quantitative real-time PCR ( qRT-PCR ) , for PR2X7 , CAMKK2 and DENR were designed using Primer3Plus [68] and are shown in S6 Table . Semi quantitative RT-PCR using AmpliTaq Gold® ( Life Technologies , Grand Island , NY ) was performed to confirm product size and sequence identity was confirmed by sequencing . PCR was performed in triplicates using the Rotor-Gene SYBR Green PCR Kit ( QIAGEN ) . A 2 step cycle protocol ( 35 cycles; 95°C; Annealing- 15 seconds at 60°C; Extension- 90 seconds at 60°C; Final Melt curve ) was carried out on the Rotor Gene Q real-time PCR instrument ( QIAGEN ) . Each replicate containing 0 . 2ng template cDNA . All data were normalized to the housekeeping gene B2M , using published primer sequences [69] . Amplification efficiency and differences in takeoff values between affected and unaffected dogs were analyzed by REST2009 [70] . Box plots of calculated delta-delta CT values were generated using GraphPad Prism version 5 ( GraphPad Software , La Jolla , CA ) . Protein extraction and Western blots were carried out similarly as described before [16] . Briefly , tissues were lysed in RIPA buffer ( Boston BioProducts , Inc . , Worcester , MA ) with 1X Halt protease and phosphatase inhibitors ( Thermo Fisher Scientific , Inc . , Rockford , IL ) and proteins were quantified using a Coomassie protein assay reagent ( Pierce/Thermo Fisher Scientific , Rockford , IL ) 20ug of protein was heat denatured and resolved by SDS PAGE electrophoresis followed by transfer to a nitrocellulose membrane . Blots were blocked for 1 h and then incubated overnight with primary antibodies . Primary rabbit polyclonal antibodies used were anti-P2RX7 antibody recognizing amino acids 576–595 of rat P2X7R ( 1:500 , APR-004 Alamone Labs , Jerusalem , Israel ) and anti-actin antibody ( 1:10 , 000 , A2066 , Sigma-Aldrich , St Louis , MO ) . Blots were washed then incubated for 2 h with HRP-conjugated goat anti-rabbit IgG ( 1:5 , 000 , 12–348 , EMD Millipore , Temecula , CA ) . Blots were visualized with SuperSignal West Femto solution ( Pierce/Thermo Fisher Scientific , Rockford , IL ) and detection was performed with Vision Works LS digital capture software ( UVP , Upland CA ) . One set of glioma and normal brain expression data was obtained from a study by Liu and colleagues [40] and can be found in the ArrayExpress database ( GSE21354 ) . Briefly , the dataset contains five samples of diffusely infiltrating astrocytic gliomas , four samples of ependymomas , five samples of oligodendrogliomas , and four samples of normal brain tissue . All 18 samples were hybridized on the Affymetrix GeneChip Human Genome U133 Plus 2 . 0 . Raw CEL files where downloaded from ArrayExpress and processed in the Affymetrix Expression Console using the RMA algorithm for background correction , quantile normalization and probe summarization . Multiple transcripts/probes for the same gene were collapsed using the mean value . A second dataset of gene expression values in GBM and normal brain tissue was downloaded from TCGA ( http://cancergenome . nih . gov/ ) . Specifically , the data contained a total of 10 epileptic brain normal tissue samples processed in a single batch as well as 24 GBM tissue samples processed within the same batch . For these samples the downloaded data corresponded to level 3 gene expression as measured by the HT Human Genome U133A array . For further processing , the distribution of gene expression values within each sample was standardized using the z-score . One-way ANOVA analyses were employed to test for an overall significant difference between group means for each gene and two-sided Welch t-tests were employed to test for significant differences between the means of any pair of two groups . Human glioblastoma cell cultures were developed as follows: human GBM grade IV biopsies were obtained in accordance with the protocol approved by the Uppsala ethical review board ( 2007/353 ) , and were graded by neuropathologist Irina Alafuzoff , Uppsala University Hospital , according to WHO guidelines . Tumor biopsies were minced ( 1 mm × 1 mm pieces ) and digested by 1:1 ratio of Accutase ( eBioscience , San Diego , CA ) /TrypLE ( Life technologies , Grand Island , NY ) at 37°C for 15 min and triturated through 18g and 21g needles 5 times . Dissociated cells were resuspended in DMEM/F12 Glutamax ( Life technologies , Grand Island , NY ) and Neurobasal medium ( Life technologies , Grand Island , NY ) mixed 1:1 with addition of 1% B27 ( Life technologies , Grand Island , NY ) , 0 . 5% N2 ( Life technologies , Grand Island , NY ) , 1% penicillin/streptomycin ( Sigma-Aldrich , St Louis , MO ) , 10 ng/ml EGF and FGF2 ( Peprotech , Rocky Hill , NJ ) , and plated at 100 , 000 cells/ml . After primary sphere formation , spheres were seeded onto poly-ornithine/laminin-coated dishes and cultured as adherent cells as described in Pollard et al . [71] . Twenty thousand human glioma cells per well were seeded onto polyornithine-laminin coated glass coverslips in 6-well plates in maintenance medium ( see above ) . The next day , medium was changed to differentiation medium , DMEM/F12 Glutamax: Neurobasal ( ratio1:1 ) with 1% B27 , 0 . 5% N2 supplement , 1% penicillin/streptomycin and with the addition of 5% FBS but no EGF or FGF-2 . Medium was changed every 3–4 days . Seven days later , cells were lysed for RT-PCR . Cells were lysed , and total RNA was extracted using RNeasy Mini kit ( Qiagen , Valencia , CA ) according to instructions from the manufacturer . Genomic DNA was digested with RNase-free DNase I ( Qiagen , Valencia , CA ) throughout the RNA extraction process . Five hundred nanograms of RNA was transcribed into cDNA using iScript cDNA synthesis kit ( BIO-RAD , Berkeley , CA ) . Quantitative RT-PCR was performed in triplicates with Ssofast EvaGreen Supermix kit ( BIO-RAD , Berkeley , CA ) . A no-template negative control was included for each primer set and constantly found not to generate any PCR products . The primers used in this experiment are shown in S7 Table . The PCR products were loaded on 2% agarose gel and photographed using Molecular Imager gel-doc XR+imaging system ( BIO-RAD , Berkeley , CA ) and Image Lab software . For relative expression analysis , a comparative cycle threshold method ( ΔΔCT ) was used . Briefly , each gene of interest was first normalized against endogenous housekeeping control ( β-actin ) , and then the normalized values were further normalized using the control sample . Normal human frontal cortex lysate was purchased from Abcam . Human glioma cells were lysed in RIPA buffer ( 150 mM NaCl , 1% NP-40 , 0 . 5% deoxycholate , 0 . 1% SDS , 50 mM Tris , pH7 . 5 ) containing 1% protease inhibitor ( Roche Diagnostics ) . Lysate was centrifuged at 10 , 000g for 30min and supernatant was used for protein estimation using BCA protein assay kit ( Thermo scientific , Rockford , IL ) following the manufacturer’s instruction . Equal amount of protein was loaded onto 10% gel NuPAGE ( Invitrogen ) . The protein was transferred to nitrocellulose membrane through iBlot gel transfer device ( Invitrogen ) . Membrane was blocked in 5% milk in TBST for 1h , followed with overnight incubation with primary antibody in 4°C and incubation with secondary antibody for 1h in room temperature . Primary antibodies used were mouse anti-CAMKK2 antibody ( 1:200 ) from Abnova ( Taipei , Taiwan ) , and mouse anti-beta-actin antibody ( 1:5000 ) from Sigma-Aldrich ( St Louis , MO ) . Secondary antibody used was HRP goat anti-mouse IgG from GE healthcare ( Little Chalfont , UK ) . Blot was visualized using Amersham ECL kit ( GE healthcare , Little Chalfont , UK ) and detection was performed with ImageQuant LAS 4000 ( GE healthcare , Little Chalfont , UK ) . | Gliomas are devastating malignant brain tumors that are very rarely curable . Despite extensive research to define pathways and genes involved in the development of disease , there is still an urgent need to improve therapy . Some dog breeds have a considerable elevated risk of glioma , making the dog a suitable model for locating genes potentially of importance also for development of human glioma . In this study we defined a genomic region strongly associated with glioma in dogs . We also showed that this genomic region had likely been under selection in the dog breeds with the highest risk of developing glioma . Sometimes selection for breed specific traits results in amplification of disease causing mutations together with the variant selected for . We located three candidate genes in the identified region: CAMKK2 , P2RX7 and DENR . We performed further functional studies to evaluate the potential role of these genes in both canine and human glioma . By comparing normal and tumor tissue we could show that two of the genes—CAMKK2 and P2RX7 , were affected at the level of gene expression and protein structure , respectively . We propose that further investigation of all three genes could be of interest with potential benefit to both dog and human . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"animal",
"types",
"medicine",
"and",
"health",
"sciences",
"cancers",
"and",
"neoplasms",
"vertebrates",
"pets",
"and",
"companion",
"animals",
"dogs",
"animals",
"mammals",
"oncology",
"neurological",
"tumors",
"genome",
"analysis",
"mammalian",
"genomics",
"molecular",
"biology",
"techniques",
"zoology",
"genotyping",
"research",
"and",
"analysis",
"methods",
"gene",
"mapping",
"gene",
"expression",
"molecular",
"biology",
"animal",
"genomics",
"neurology",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"glioma",
"amniotes",
"computational",
"biology",
"organisms",
"human",
"genetics"
] | 2016 | Utilizing the Dog Genome in the Search for Novel Candidate Genes Involved in Glioma Development—Genome Wide Association Mapping followed by Targeted Massive Parallel Sequencing Identifies a Strongly Associated Locus |
Aedes albopictus , the Asian tiger mosquito , originates from the tropical and subtropical regions of Southeast Asia . Over the recent decades it has been passively spread across the globe , primarily through the used tyre trade and passive transportation along major traffic routes . A . albopictus is a proven vector for many arboviruses , most notably chikungunya and dengue , with recent outbreaks also in continental Europe . In southern Switzerland , in the Canton of Ticino A . albopictus was spotted for the first time in 2003 . Since then the local authorities have implemented a control programme based on larval source reduction . Despite these efforts , mosquito densities have increased over the last decade , casting doubts on the effectiveness of such larval control programmes . The Italian communities just across the Swiss-Italian border lack a control programme . This motivated us to compare the intervention and the non-intervention areas side by side in an attempt to find evidence for , or against , the effectiveness of larval A . albopictus control . Using ovitraps and a randomised sampling scheme , we examined the seasonal and spatial abundance of A . albopictus in sylvatic and urban environments across the Swiss-Italian border in 2012 and 2013 . In the urban environments of the non-intervention area , egg densities were 2 . 26 times higher as compared to the intervention area . In the sylvatic environments , as compared to the urban environments , egg densities were 36% in the intervention area and 18% in the non-intervention area . Though alternative explanations are also valid , the results support the hypothesis that the Ticino intervention programme does have an impact . At the same time the data also suggest that current larval interventions fall short in gaining full control over the mosquito , calling for the evaluation of additional , or alternative , approaches . Ideally , these should also consider inclusion of the neighbouring Italian communities in the surveillance and control efforts .
Aedes ( Stegomyia ) albopictus ( Skuse , 1894 ) , the Asian tiger mosquito , originates from the tropical and subtropical regions of Southeast Asia . During recent decades this mosquito species has spread to North America , Europe , Latin America and Africa , primarily by the transport of dormant eggs in used tyres [1] and through the importation of Dracaena sanderiana plants , also known as “lucky bamboo” [2] . At the regional level the mosquito is further passively dispersed through adults displaced by vehicles along traffic routes such as motorways [3] . Under laboratory conditions , A . albopictus is a competent vector for at least 26 arboviruses , notably chikungunya , dengue , yellow and West Nile fever [4 , 5] . A . albopictus is also of veterinary significance because it is equally a competent vector for Dirofilaria immitis , a nematode that causes dirofilariosis in dogs [4] . Therefore , the establishment of A . albopictus represents a potential threat for both public and veterinary health . How realistic this threat is also for mainland Europe has been clearly demonstrated by several reports of autochthonous chikungunya and dengue cases over the recent years . In 2007 , an outbreak of chikungunya associated with the establishment of A . albopictus occurred in Ravenna , Italy , with over 200 confirmed cases [6 , 7] . More recently , between August and September 2010 , autochthonous cases of dengue have been reported from Croatia and metropolitan France with A . albopictus deemed responsible for its transmission [8 , 9] . In the same year , two people became also infected with the chikungunya virus in Fréjus , France [10] . Then additional autochthonous dengue cases were reported from southern France in 2013 [11] and again in 2014 , alongside new cases of chikungunya [12] . In Italy , A . albopictus was first detected in Genoa in 1990 from where it spread to many parts of Italy , including the border region south of Switzerland [13] . In response to its presence in northern Italy an A . albopictus surveillance programme was put in place by the local authorities in southern Switzerland in the Canton of Ticino ( in the following simply called Ticino ) in 2000 . Three years later , the first A . albopictus eggs were detected [14] . As increasing egg numbers were detected between 2003 and 2006 , the surveillance effort was gradually intensified and control measures implemented [14] . Control measures entailed removing of potential breeding sites and use of larvi- and adulticides . In the following years the estimated A . albopictus density was still low , suggesting that individual adult mosquitoes had been sporadically introduced from Italy but had not yet established a sustained population in Ticino . Yet , in 2007 the situation changed significantly , when a dramatic increase of positive mosquito traps in Chiasso , right at the Swiss-Italian border , was observed , indicating that a local mosquito population had then been established [14 , 15] . In 2007 the monitoring system consisted of 292 oviposition traps ( ovitraps ) that were regularly controlled , covering a defined area of approx . 4 . 6 km2 . Ovitraps are a widely used tool for the surveillance of container breeding Aedes [14 , 16–19] as they are sensitive , relatively inexpensive and easy to maintain [20][16 , 18] . The ovitrap is a device that consists of a water-filled black bucket with a piece of wood , or styrofoam , onto which female mosquitoes may deposit their eggs . In Ticino , the ovitraps used consist of a flower pot filled with water into which a wooden strip is plunged for the females to lay eggs [14] . The traps are set within communities as well as at lay-bys and service areas along the motorway E35 [21] . The E35 is a south-north European route that runs from Rome ( Italy ) to Amsterdam ( the Netherlands ) . In addition , places with stagnant water that cannot be averted otherwise were treated with Bacillus thuringiensis var . israelensis ( Bti ) , a biological control agent for larval mosquito stages [22] . During the last years , the ovitrap network has been continuously expanded and adapted . Today , over 1 , 000 ovitraps are deployed within the frame of the Ticino surveillance and control programme covering an area of approx . 60 km2 . The traps are inspected biweekly and the number of positive traps serves as an indicator if and where the application of insecticide would be necessary [14] . In addition , information campaigns are carried out to raise public awareness in order to sensitise residents for the occurrence of A . albopictus and to eliminate potential breeding sites from their private properties . Despite these measures A . albopictus densities have still increased in Ticino over the last decade [14] . Larval source reduction by removing water containers that may serve as breeding sites is considered the best method for the control of A . albopictus by several authors [23 , 24] . Studies from North Carolina [25] , Spain [26] and New Jersey [27] reported that source reduction campaigns resulted in a temporary suppression of immature A . albopictus . Indeed , Bartlett-Healy et al . [28] showed that artificial containers on private properties are the most productive sources for the emergence of A . albopictus , highlighting the importance of public involvement in the overall control effort . Awareness campaigns showing the public how to identify and eliminate potential breeding sites from their properties have become an integral component of Aedes mosquito control [20] . Such campaigns go often hand in hand with larvicide treatments and spraying of insecticides targeting adult mosquitoes . Comparing different intervention approaches , Fonseca et al . [27] concluded that careful source reduction by trained personnel , in combination with efforts to educate the public in removing breeding sites , results in a significant decrease in adult A . albopictus numbers . Despite the above evidence there is still much debate as to how effective such larval control measures really are , particularly in areas where mosquitoes are continuously re-introduced such as being the case in southern Switzerland . This motivated us to examine the potential impact of the current surveillance and control programme by comparing relative mosquito densities between Ticino and two neighbouring Italian provinces where ecological parameters are comparable; yet , no intervention programme is in place .
Field surveys were carried out from July to November 2012 and from May to November 2013 . The study area enclosed the southernmost border region of Ticino , the Mendrisiotto district , and the provinces of Varese and Como in Lombardy , Italy ( Fig 1 ) . Hereafter , the part of the study area in Ticino is called the “intervention” area and that of Varese and Como the “non-intervention” area . In total , the study area covered a surface area of 118 km2; 65 km2 on the Italian side and 53 km2 on the Swiss side of the border . The difference in the surface areas were to make up for places that were either inaccessible or covered by the Lake of Como . The landscape of the study area is similar on both sides of the border and dominated by deciduous forests and agriculture . Approximately 20% are covered by buildings or roads . Population densities are almost equal and are 440 and 480 inhabitants per km2 in the Ticino and the Lombardy part , respectively [29 , 30] . The traffic-intense European route E35 runs through the study area , connecting the South of the continent with North-western Europe . On average , on a single work day over 62 , 000 people cross the Swiss-Italian border , mostly by car [31 , 32] . Most of the people crossing the border commute to Switzerland for work . The climate in the study area is continental with relatively mild temperatures , yet distinct annual seasons . Mean annual temperature and rainfall are 11 . 1°C and 1 , 311 mm [33] . Besides its relatively sunny weather , the region is also well known for its heavy thunderstorms during the summer . Using the ArcGIS version 10 . 0 ( ESRI Inc . , USA ) geographic information system ( GIS ) software a grid with 250 m by 250 m cells was virtually superimposed over the study area . From this grid , all grid cells within a lake and those that were inaccessible in the field were excluded from sampling . The remaining grid cells were then stratified into “urban” and “sylvatic” environments . A cell was classified as sylvatic if at least 50% of the surface were covered with trees , and vice versa . For each of the four combinations of area and environment 35 cells were randomly picked from the grid to avoid sampling bias . For this purpose the cells were first numbered through and then the numbers drawn using a random number generator . The total number of cells included in the study was chosen on the basis of a power calculation that used simulation methods described in Johnson et al . [34] . For this exercise we assumed a minimal effect size of 10% difference in egg counts between the two countries and a power of 1-β = 0 . 8 . Relative densities of A . albopictus were estimated using ovitraps . The traps mimic breeding sites , attracting gravid females to deposit their eggs . In the present study , an ovitrap consisted of a 1 . 5 l , black plastic flower pot , filled with 1 . 2 l tap water . Three small holes with a diameter of 5 mm were drilled at equal distances , 2 cm below the rim , to prevent the trap from being flooded by rain . A wooden strip made of untreated beech wood was placed inside the pot so that it was partially submerged and partially sticking out of the water . The strip measured 20 cm x 2 . 5 cm x 0 . 5 cm . In order to prevent the ovitraps from becoming potential breeding sites larvicide granules of Bti ( VectoBac , Valent BioSciences , USA ) were added . The strips , water and Bti were replaced biweekly . When replaced , the traps were cleaned and the wooden strips wrapped in clingfilm for transportation and preservation . Each strip was labelled with the date and a unique code together with additional information related to the trap condition and the presence of larvae . The final trap position within the assigned sampling grid cell was chosen in the field . Traps were placed at shaded , wind protected locations that , in the optimal case , were surrounded by green vegetation as done in previous studies ( e . g . [14 , 27] ) . All traps were geo-referenced with a handheld GPS device ( nüvi 1390 , Garmin , Switzerland ) . In the laboratory , the strips were inspected for the presence of mosquito eggs using a stereo microscope ( EZ4D , Leica Microsystems , Germany ) and , where present , the number of eggs counted . During the first season in 2012 , eggs were identified to species level by morphology . At that time only two container-breeding mosquito species , A . albopictus and A . geniculatus , were known to occur in the region . Both species can easily be distinguished by morphology [20 , 35] . As a quality control measure an additional identification method was introduced for the 2013 mosquito season . Here , for each collection round , eggs from two randomly selected positive traps were also analysed by matrix-assisted laser desorption/ionization mass-spectrometry ( MALDI TOF MS ) [35] . Only eggs were chosen for the analysis that had previously been morphologically determined as being A . albopictus and , where present , were still intact . For MALDI-TOF MS three to five apparently intact eggs were carefully removed using forceps from the ovitrap strips and then transferred to 1 . 5 ml Eppendorf tubes . The samples were sent to Mabritec SA ( Riehen , Switzerland ) where they were prepared and analysed according to the protocol described in Schaffner et al . [35] . The numbers of A . albopictus eggs on each wooden strip were counted and recorded in an Excel data base together with additional information such as the trap location , date , condition of the trap , etc . Data were then imported into the GIS software ArcGIS Version 10 . 1 ( ESRI Inc . , USA ) to produce spatio-temporal density maps . For statistical analysis , data were loaded into the freely available software R , version 3 . 1 . 2 [36] . Relative egg densities per trap were modelled by a zero-inflated negative binomial ( ZINB ) regression model using the R package “glmmADMB” [37 , 38] . The ZINB accounted for an excessive number of zeros in the ovitrap count data . In the ZINB model , the outcome was the biweekly egg count per trap , while the predictors “area” ( non-intervention vs . intervention ) and “environment” ( urban vs . sylvatic ) and their interaction were included as fixed effect terms . To account for the slight bias in altitude towards higher elevations in the intervention area ( Fig 2 ) and the potential relationship between altitude and temperature , a predictive term for “altitude” was also included in the model . Altitude was entered as metres above sea level . As egg counts were repeatedly ( i . e . biweekly ) measured for the same ovitrap , an intercept was included for “trap” as a random term in the ZINB model , accounting for correlations in the number of eggs caught in the same trap . Also included as a random term was an intercept for the week in which the traps were replaced in order to account for seasonal variations . The model was also inspected for signs of spatial correlations in the residuals using the variogram function in the R package “gstat” , version 1 . 0–19 [39] . The statistical graphics were produced with ggplot2 , version 1 . 0 . 0 [40] . The level of significance was set at α = 0 . 05 .
In 2012 , ovitrap collections ran over 20 weeks ( i . e . 10 rounds ) from July to November , while the survey covered 26 weeks ( i . e . 13 rounds ) from May to November in 2013 . The first eggs in the season were found in early June , followed by a steady increase with a peak between 19 and 26 August . In September , egg counts dropped again and eventually ceased in mid-November ( Fig 3 ) . From the potentially 6 , 440 available strips for the analysis ( 280 traps x 23 rounds ) , 357 ( 5 . 5% ) have gone missing ( Table 1 and S1 Table ) ; either they have been taken from the traps or the traps themselves became dysfunctional ( e . g . traps were found turned over or missing completely ) . From the remaining 6 , 083 strips , 2 , 508 ( 41 . 5% ) were positive for A . albopictus , 689 ( 11 . 4% ) for A . geniculatus and 333 ( 5 . 5% ) for both species . While for A . albopictus a total of 224 , 728 eggs were counted , egg numbers were not recorded for A . geniculatus , only whether eggs were present or absent . In 2012 , egg counts per trap ranged from 0 to 1 , 537 in the non-intervention area ( i . e . Lombardy , Italy ) and from 0 to 441 in the intervention area ( i . e . Ticino , Switzerland ) . In 2013 , egg counts ranged in the non-intervention and intervention area from 0 to 1 , 039 and from 0 to 1 , 333 , respectively . Egg counts were generally higher in the non-intervention area ( Table 1 ) . In all ( i . e . 20 ) instances the morphological identification was confirmed by MALDI-TOF MS . Remarkably , A . albopictus eggs were found across the whole altitude range ( Fig 2 ) and were even repeatedly found at higher altitudes up to 781 m above sea level ( S1 Table ) . In the urban environment , the average ratio in egg densities between the non-intervention and the intervention area was 2 . 26 ( 95% confidence interval , CI: 1 . 40–3 . 65; Fig 4 and Table 2 ) . Mosquito eggs were also detected in the sylvatic environment , although , as compared to the urban environment , the counts were much lower . The average ratios between the sylvatic and the urban environments were 0 . 36 and 0 . 18 in the intervention and in the non-intervention area , respectively . In the model , the difference in these ratios is accounted for by the interaction term ( Table 2 ) with an estimated ratio of 0 . 504 ( CI: 0 . 254–0 . 997 ) and graphically illustrated in Fig 4 . In addition , the model improved by adding a term for altitude; an increase of altitude by one meter decreases egg counts by a ratio of 0 . 995 , that is by 0 . 54% ( 95% CI: 0 . 37%– 0 . 71% ) . The model did , however , not improve when adding “year” as a term , indicating that egg counts did not significantly differ between the two years ( χ2 = -2 . 6 , p = 1 ) . Moreover , inspecting the residuals for spatial correlations did not detect violation of independence . When plotting the positive traps on the geographic map , it becomes apparent that not only the numbers of eggs were higher in the non-intervention area but , equally , more traps were positive ( Fig 5 ) . The picture remained the same in both years and in the early ( July ) and late ( September ) mosquito season . Combining egg counts from both seasons , 32 . 4% ( 72 , 869 eggs ) of the A . albopictus eggs were collected alone in the city of Como . In Switzerland the communities of Chiasso and Balerna , which are located at ( i . e . Chiasso ) or very close to ( i . e . Balerna ) the border , had the highest A . albopictus egg counts . Over both sampling periods , total egg numbers in Chiasso and Balerna were 12 , 637 and 12 , 212 , respectively . Together they represent 11% of the total A . albopictus egg count . As traps were , however , distributed randomly to make inference about the whole region these numbers have to be interpreted with caution .
We found that A . albopictus egg densities in the non-intervention area on the Italian side of the Swiss-Italian border were more than twice compared to the intervention area in Ticino . Though other factors might explain the difference in mosquito densities , the present data support the hypothesis that the currently implemented surveillance and control programme in Ticino has a positive impact . Presumably public awareness is a major component in reducing A . albopictus densities . However , it remains to be shown experimentally how big the actual impact of the current interventions really is . | The Asian tiger mosquito ( Aedes albopictus ) has gained increased attention in public health because it is a globally spreading , highly invasive mosquito species that may transmit several viruses . Outside of its original range in Southeast Asia it has been increasingly implicated in local transmission of chikungunya and dengue fever in many places including La Réunion , continental Europe , the Americas and Japan . The Asian tiger mosquito lays eggs that are adapted to desiccation and colder climate . This , together with the mosquito’s ability to breed in almost any small , stagnant water body , makes its control extremely difficult , and there is much debate as to what interventions would be effective . This motivated us to compare the occurrence of the Asian tiger mosquito in southern Switzerland , where a mosquito surveillance and control programme is in place , with its neighbouring Italian districts where no such programme exists . The Swiss programme is based on public awareness campaigns to remove breeding sites and the use of insecticides against larvae . Using specialised traps that collect eggs from egg laying female mosquitoes , we found 2 . 26 times more A . albopictus eggs in the non-intervention area . The results support the hypothesis that targeting larval sources does have a significant impact . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2016 | Surveillance and Control of Aedes albopictus in the Swiss-Italian Border Region: Differences in Egg Densities between Intervention and Non-intervention Areas |
Polymyxin is the last line of defense against severe infections caused by carbapenem-resistant gram-negative pathogens . The emergence of transferable MCR-1/2 polymyxin resistance greatly challenges the renewed interest in colistin ( polymyxin E ) for clinical treatments . Recent studies have suggested that Moraxella species are a putative reservoir for MCR-1/2 genetic determinants . Here , we report the functional definition of ICR-Mo from M . osloensis , a chromosomally encoded determinant of colistin resistance , in close relation to current MCR-1/2 family . ICR-Mo transmembrane protein was prepared and purified to homogeneity . Taken along with an in vitro enzymatic detection , MALDI-TOF mass spectrometry of bacterial lipid A pools determined that the ICR-Mo enzyme might exploit a possible “ping-pong” mechanism to accept the phosphoethanolamine ( PEA ) moiety from its donor phosphatidylethanolamine ( PE ) and then transfer it to the 1 ( or 4’ ) -phosphate position of lipid A via an ICR-Mo-bound PEA adduct . Structural decoration of LPS-lipid A by ICR-Mo renders the recipient strain of E . coli resistant to polymyxin . Domain swapping assays indicate that the two domains of ICR-Mo cannot be functionally-exchanged with its counterparts in MCR-1/2 and EptA , validating its phylogenetic position in a distinct set of MCR-like genes . Structure-guided functional mapping of ICR-Mo reveals a PE lipid substrate recognizing cavity having a role in enzymatic catalysis and the resultant conference of antibiotic resistance . Expression of icr-Mo in E . coli significantly prevents the formation of reactive oxygen species ( ROS ) induced by colistin . Taken together , our results define a member of a group of intrinsic colistin resistance genes phylogenetically close to the MCR-1/2 family , highlighting the evolution of transferable colistin resistance .
Prevalent antibiotic resistance is posing a threat to providing safe and effective health care worldwide . Antimicrobial resistance ( AMR ) is associated with 700 , 000 deaths each year [1 , 2] . It is estimated by O’Neill and his team that AMR would claim as many as 10 , 000 , 000 deaths per year globally by 2050 [3] . Despite this prediction being exaggerated and unreliable [4] , we acknowledge that AMR has been posing an ever-growing burden on clinical therapies and public health , thereby highlighting the urgent need for a nationwide response to alleviate this burden [4 , 5] . Polymyxins , a class of cationic cyclic polypeptide antibiotics , act as a “last-resort” option against infections by carbapenem-resistant gram-negative pathogens [6–9] . However , the emergence and global spread of plasmid-borne mobilized colistin resistance determinants ( mcr-1 ) has greatly threatened the renewed interest of colistin ( polymyxin E ) in clinical therapies [10] . To the best of our knowledge , mcr-1-harboring Enterobacteriaceae have been detected in no less than 40 countries worldwide , spanning 5 of 7 continents [11] , including the United States of America [12 , 13] . MCR-1 is annotated as a transmembrane PEA transferase , which belongs to the “YhjW/YjdB/YijP” alkaline phosphatase super-family [10 , 14] . MCR-1 catalyzes the transfer of phosphoethanolamine ( PEA ) from phosphatidylethanolamine ( PE ) to the 1 ( or 4’ ) -phosphate position of lipid A glucosamine ( GlcN ) moieties ( S1 Fig ) [15–17] . In general , the addition of PEA to lipid A is believed to reduce the net negative charge of the bacterial outer-membrane [6 , 18 , 19] , consequently leading to a phenotypic resistance to polymyxin , a final line of defense against severe infections by pathogens with multi-drug resistance ( MDR ) [7 , 8] . In comparison to the paradigm PEA transferase , the Neisseria EptA protein [16 , 20] , MCR-1 seems to render the recipient E . coli strains more resistant to colistin [14] , implying that it is a more efficient enzyme . In contrast to the more prevalent MCR-1 , its cousin MCR-2 with almost 80% amino acid identity , seems to be rare in that it is only detected in Belgium in 2016 [21] . In particular , it seems likely that both MCR-1 and MCR-2 have an evolutionarily-conserved catalytic motif having essential roles in enzymatic activities and resultant antibiotic resistance [14 , 22] . Along with other transferable determinants of colistin resistance ( MCR-3 [23 , 24] , MCR-4 [25] and MCR-5 [26] ) , a growing body of MCR-1 variants with point substitutions have been reported , such as MCR-1 . 2 [Q3L] [27] , MCR-1 . 3 [I38V] [28] and MCR-1 . 6 [R536H] [29] . More recently , mcr-1 and mcr-2 variant genes have been discovered on chromosomes of Moraxella species [30 , 31] . In brief , MCR-1 . 10 with 98 . 7% amino acid identity to MCR-1 is present in M . porci MSG13-C03 [30] , MCR-2 . 2 with 99% amino acid identity ( only 8 amino acid substitutions of 538 residues ) to MCR-2 is encoded on the chromosome of M . pluranimarium [31] , and MCR-2 . 3 with 87 . 9% amino acid identity to MCR-2 is harbored in a M . pluranimalium-like isolate , MSG47-C17 . Moreover , it seems that MCR-like enzymes with about 60% amino acid identity are prevalent in Moraxella species [32] , suggesting a potential reservoir of mcr-like genes . However , this requires further experimental evidence . In this study , we have integrated multiple lines of approaches to study the structure and mechanism for one such mcr-like gene , AXE82_07515 in M . osloensis ( hereafter designated as icr-Mo [of note: Mo denotes Moraxella osloensis] ) . This will allow it to serve as a representative member of this family for future studies . Consistent with genetic speculations raised by other groups [30–32] , our results formulate a functional definition of icr-Mo and provide an evolutionary relationship between itself , mcr-1/2 variants and non-mcr genes that confer colistin resistance . By probing the biochemical and physiological relevance , our findings highlight the importance of the Moraxella family in understanding the growing body of mcr-1/2-like genetic determinants conferring colistin resistance .
The gene mcr-1 , is a prevalent determinant of plasmid-borne mobilized colistin resistance with a global distribution . In addition to a closely-related but rare mcr-1 homolog , mcr-2 sharing 81% protein identity [21] , a number of genetic variants of mcr-1 ( like mcr-1 . 2 [27] ) with point substitutions have also been elucidated . Earlier this year , Kieffer and coworkers [32] showed genetic evidence that Moraxella species might be a potential reservoir for mcr-like genes . This prediction is further validated by additional discovery of mcr-1 . 10 ( 98 . 7% amino acid identity ) in M . porci MSG13-C03 [30] and mcr-2 . 1 ( 99% amino acid identity ) in M . pluranimarium [31] . In addition to their similarity , certain members of these chromosomally encoded mcr-1/2-like genes also neighbor almost identical mobilizable elements like the insertion sequence ISApl1 , associated with mcr-1/2 dissemination [31 , 32] . To probe the evolutionary relationships between the icr-Mo in Moraxella and other extant mcr genes , we performed phylogenetic analyses ( Fig 1 ) , using their coding nucleotide sequences . As anticipated , MCR-1 and its variants ( namely MCR-1 . 2 , MCR-1 . 3 , … MCR-1 . 9 and MCR-1 . 10 ) cluster together to form a tight group ( Fig 1A and 1B ) . In fact , we have indicated 2 additional variants of MCR-1 with single synonymous point mutations . MCR-2 and its two variants ( MCR-2 . 1 and MCR-2 . 2 ) also form a distinct group ( Fig 1A and 1B ) while still being closely-clustered within a single subclade , comprising MCR-1/2 variants ( Fig 1A and 1B ) . Surprisingly , Neisseria EptA , a chromosomal colistin resistance determinant ( 31 . 5% amino acid identity to ICR-Mo , S3 Fig ) clusters with other putative sulfatases , a large family of enzymes encoding broad functionality that includes PEA transferases . The sulfatase sub-clade along with a cluster comprising MCR-3 and MCR-4 variants are only distantly-related to the MCR-1/2 and MCR-like genes . In fact , Z1140 ( 539 aa ) of E . coli O157:H7 , a putative PEA transferase lacking detectable ability to confer phenotypic resistance to polymyxin ( S2 Fig ) is found within the MCR-3/4 subclade leading us to categorize them as non MCR-like genes ( Fig 1A ) . Given that the identity between ICR-Mo and EptA is only 31 . 5% , it seems likely that mcr-1/2 and icr-Mo genes have diverged from sulfatases much earlier than the recent colistin usage in agriculture and clinical medicine . A much closer ancestor might be found in bacteria that naturally produce polymyxins as secondary metabolites . In contrast , a subset of MCR-like members exclusively from the Moraxella species ( porci , lincolnii and catarrhalis subspecies , in Fig 1 ) with about 60% amino acid identities ( S3 Fig ) are clustered into a neighboring clade that includes ICR-Mo ( Fig 1A and 1B ) . This also includes the recently crystallized ICR-Mc ( Accession No: AIT43666 ) from Moraxella catarrhalis which shares 55 . 2% amino acid identity with ICR-Mo [33] . We are inclined to believe that plasmid-borne MCR-1/2 variants and the intrinsic ICR-Mo might share a common ancestor . Intriguingly , certain species of Moraxella contain chromosomally encoded colistin resistance determinants that are grouped with members of MCR-1/2 family ( Fig 1B ) . Therefore , a systematic evaluation of biochemical and physiological properties of ICR-Mo , is necessary . Prediction with TMHMM server v2 . 0 ( http://www . cbs . dtu . dk/services/TMHMM ) suggests that ICR-Mo is an integral membrane protein possessing five N-terminal helices ( S3A Fig ) , which is almost identical to those of MCR-1 [14] and MCR-2 [22] . As recently described for MCR-1/2 [14 , 22] , the transmembrane protein ICR-Mo was overexpressed and purified to homogeneity with nickel affinity column in the presence of 1% detergent dodecyl-β-D-maltoside ( DDM ) . Following visualization by 12% SDS-PAGE ( S4A and S8A Figs ) , identity of the recombinant ICR-Mo protein was determined with MALDI-TOF MS ( Matrix-Assisted Laser Desorption/Ionization Time of Flight Mass Spectrometry ) with a 84 . 67% fragment sequence ( S10C Fig ) . The structure of full-length EptA ( PDB: 5FGN ) was used as a template [34] for the structural modeling of ICR-Mo via Swiss-Model ( https://swissmodel . expasy . org/interactive ) [35] . Despite possessing only 31 . 5% amino acid identity to EptA ( S2 Fig ) , the modeled structure of ICR-Mo shows an appreciable coverage of 93% ( 24–558 ) , indicating a satisfied prediction ( S4 Fig ) . ICR-Mo also displays a similar topology to EptA ( S4B Fig ) [34] . The TM domain spanning the inner-membrane involves five α-helices ( S4B Fig ) , and the catalytic domain has a hydrolase-fold ( S4C Fig ) comprising 10 α-helices and 7 β-sheets ( S4C Fig ) . The two domains are linked by four short periplasm loops ( PH2 , PH2’ , PH3 and PH4 ) , a bridge helix ( BH ) and a long-coiled loop ( S4B and S4C Fig ) . As described by Anandan et al . [34] for EptA activity , we established an in vitro system for ICR-Mo enzymatic reaction . In this reaction system , the PEA donor substrate used as a fluorescent label , 1-acyl-2-{12-[ ( 7-nitro-2-1 , 3-benzoxadiazol-4-yl ) amino] dodecanoyl}-sn-glycerol-3-phosphoethanolamine ( abbreviated as NBD-glycerol-3-PEA , Fig 2A ) . It is hypothesized that ICR-Mo removes the PEA moiety from the alternative substrate NBD-glycerol-3-PEA , and gives rise to a NBD-glycerol product ( Fig 2A ) . Given that an adduct of PEA-Thr280-EptA has been observed by Vrielink and coworkers [20] , we anticipate that an intermediate of ICR-Mo-bound PEA ( i . e . , PEA-Thr287-ICR-Mo ) is produced in the hydrolytic reaction ( Fig 2A ) . Thereafter , we utilized thin layer chromatography ( TLC ) to separate the mixture of the ICR-Mo reaction ( Fig 2B and 2C and S5 Fig ) . As expected , the product NBD-glycerol is released in the presence of ICR-Mo , which is consistent with observations of EptA and MCR-1/2 ( S5A and S5B Fig ) [36 , 37] . Under blue light ( 455–485 nm ) , TLC-based assay illustrates that the NBD-glycerol product ( Fig 2C ) migrates faster than the substrate NBD-glycerol-3-PEA does ( S5 Fig , Fig 2B and 2C ) . Subsequently , liquid chromatography mass spectrometry ( LC/MS ) confirms the identities of NBD-glycerol-3-PEA ( Fig 2B ) and its resultant product , NBD-glycerol ( Fig 2C ) . This demonstrates clearly that ICR-Mo displays an enzymatic activity of removing the PEA moiety from the lipid substrate in vitro , which is similar to the well-studied EptA [34] , as well as MCR-1/2 ( S5 Fig ) . To further the in vivo transfer of PEA from PE lipids ( S1A Fig and Fig 2D–2G ) , we engineered E . coli MG1655 strains to carry arabinose inducible plasmid pBAD24-borne icr-Mo ( and eptA/mcr-1 ) . We prepared and purified the LPS-lipid A from an array of E . coli strains . Unlike the negative control MG1655 exhibiting a single peak ( m/z of 1796 . 370–1797 . 843 ) for lipid A alone ( Fig 2E–2G ) , MALDI-TOF mass spectrometry unveiled an additional peak associated with PPEA-1 ( or 4’ ) -lipid A , the modified form of lipid A and this peak appeared in the E . coli strain harboring mcr-like genes . We believe that lipid A accepts PEA moiety from the intermediate of ICR-Mo-bound PEA , giving PPEA-4’-lipid A ( Fig 2D ) . In fact , the molecular mass of PPEA-lipid A in different strains exhibits slight variation , namely m/z 1920 . 014 for EptA ( Fig 2E ) , m/z 1920 . 334 for MCR-1 ( Fig 2F ) , and m/z 1919 . 577 for ICR-Mo ( Fig 2G ) . The in vitro and in vivo evidence leads us to believe that ICR-Mo catalysis proceeds by PEA transfer from the donor PE lipid substrate to the receiver lipid A , giving the two final products , PPEA-4’-lipid A and diacyl glycerol ( S1A Fig ) . Taking into account the similarity between the adduct of Thr280-PEA in EptA [20 , 34] and its counterparts in the reactions involving alkaline phosphatase-type phosphate transferase [38] , it can be hypothesized that i ) the ICR-Mo-bound PEA intermediate is released from the PE lipid molecule in the first-half reaction of ICR-Mo catalysis ( S1B Fig ) ; ii ) In the second half-reaction , PEA is transferred from the ICR-Mo-bound PEA adduct to the 1 ( 4’ ) -phosphate position of lipid A GlcN moieties , generating PPEA-4’- lipid A ( S1C Fig ) . It seems likely that ICR-Mo exploits a possible “ping-pong” mechanism for enzymatic catalysis , similar to those proposed for EptA [34] and MCR-1/2 [36 , 37 , 39] . On the basis of the newly-determined complex structure of EptA with the detergent DDM , an analogue of its physiological PE substrate [34] , we applied molecular docking to reanalyze the binding to PE molecule ( Fig 3A ) . As a result , it allows us to propose a 12 residue-constituting cavity for PE entry/binding ( Fig 3 ) . Superposition of the modeled structures of MCR-1 ( ICR-Mo ) to EptA also gives a similar PE-recognizable cavity ( Fig 3C and 3D ) . Amongst them , five amino acids ( E248 , T287 , H397 , D472 and H473 ) are centering with the zinc ion ( Fig 3E–3G and S7A and S7B Fig ) , and remaining 7 residues ( N110 , T114 , E118 , S332 , K335 , H402 and H485 ) are anticipated to be involved in the recognition of its physiological lipid PE substrate ( Fig 3H–3J and S7C and S7D Fig ) . Similar observations have been made earlier for K333 , H395 and H478 in MCR-1 [40–44] . This raises a possibility that it is an evolutionarily-conserved functional cavity shared amongst the family of MCR-like enzymes ( Fig 3 ) . However , this hypothesis needs further experimental validation . We thereafter conducted structure-guided site-directed mutagenesis , generating 12 point-mutants of ICR-Mo ( S2 Fig and Figs 3E–3J and 4A ) . Prior to functional assays in vivo , western blotting demonstrates that all the mutated versions of ICR-Mo express well in E . coli ( Fig 4A ) . Assays of colistin resistance elucidate that ( i ) ICR-Mo permits the recipient E . coli to grow on LBA plates with up to 8 μg/ml colistin [which is equivalent to that of EptA ( S6 Fig ) , but significantly lower than MCR-1 does ( 16 μg/ml ) ] ( Fig 4 and S6 Fig ) ; ( ii ) all the 5 point-mutants of ICR-Mo with defects in zinc-binding residues fail to support the growth of the recipient E . coli strains under the condition with up to 0 . 5 μg/ml colistin ( Fig 4B ) ; ( iii ) Among the 7 ICR-Mo mutants with the alanine substitution in putative PE substrate-interacting residues , 3 of them seem to retain partial activity ( namely N110A [2 μg/ml] , T114A [1 μg/ml] and S332A [1 μg/ml] , Fig 4C ) and the remaining 4 mutants are nonfunctional [0 . 5 μg/ml] ( Fig 4C ) . These results are in agreement with the observations in the colistin MIC measurements ( Fig 4D ) . To gain a mechanistic glimpse of the PE substrate-interactive cavity , we purified the 12 point-mutants of ICR-Mo protein to homogeneity ( S8A Fig ) . First , catalytic activities of the ICR-Mo mutants were examined using the in vitro enzymatic reaction with the fluorescence-labelled substrate , NDB-glycerol-3-PEA ( S5A and S8B Figs ) [34] . As anticipated , our TLC results elucidate that i ) Two mutants of ICR-Mo [N110A and S332A] retain partial activities of hydrolyzing of NBD-glycerol-3-PEA into NBD-glycerol ( S8B Fig ) ; and ii ) The other 10 mutated versions of ICR-Mo have no detectable enzymatic activity [i . e . , an alternative substrate NDB-glycerol-3-PEA is not converted into NDB-glycerol , almost identical to the negative controls] ( S8B Fig ) . In general , the variation in enzymatic activities of ICR-Mo mutants was further verified by structural analyses of LPS-lipid A moieties in a physiological context ( Fig 5 ) . As expected , MALDI-TOF MS analysis reveals that a unique peak of PPEA-1 ( or 4’ ) -lipid A ( m/z , 1920 . 140 ) is present in the E . coli expressing the wild-type ICR-Mo ( Fig 5C ) , relative to the recipient strain MG1655 with a single peak of lipid A [m/z , 1796 . 782~1796 . 946] ( Fig 5A and 5B ) . Consistent with the observations of the TLC experiments , 9 of 12 ICR-Mo point-mutants cannot modify lipid A species in vivo since only a single peak of lipid A is detected ( Fig 5F–5H and 5J–5O ) . In contrast , the remaining 3 mutants seem to retain partial abilities of PEA transfer , in that an extra-peak of PPEA-4’-lipid A [m/z , 1919 . 865~1920 . 457] is produced in addition to the lipid A [1796 . 709~1797 . 230] ( Fig 5D , 5E and 5I ) . Of note , as for the ICR-Mo ( T114A ) mutant retaining partial activity in vivo ( Figs 4C and 5E ) , no enzymatic activity in vitro is detected ( S8B Fig ) . This could be partially attributed to the lack of sufficient sensitivity of TLC assays in this case . Evidently , this highlights the importance of cavity-forming residues in ICR-Mo biochemistry . Phylogenetically , ICR-Mo is distinct from EptA and MCR-1/2 . This prompted us to probe the possible association of the two modules of MCR-like enzymes [transmembrane region ( TM ) and PEA transferase domain] with respect to their functional evolution ( Fig 1 ) . As described with MCR-2 [22] , we adopted the method of domain swapping to engineer a collection of hybrid versions of the icr-Mo gene ( Fig 6A ) . In addition to three parental proteins ( EptA , MCR-1 and ICR-Mo ( AXE82_07515 ) ) , four hybrid proteins engineered here include i ) TM ( EptA ) -AXE82 , a modified ICR-Mo carrying TM region of EptA; ii ) TM ( AXE82 ) -EptA , a derivative of EptA whose TM domain is replaced with its ICR-Mo counterpart; iii ) TM ( MCR-1 ) -AXE82 , a modified version of ICR-Mo containing the TM domain of MCR-1; and iv ) TM ( AXE82 ) -MCR-1 , a hybrid derivative of MCR-1 carrying the TM region of ICR-Mo ( Fig 6A ) . Western blotting analyses prove that all the MCR-like enzymes ( ICR-Mo and its domain-swapped derivatives ) express pretty well in E . coli ( Fig 6B ) . Subsequently , they were subjected to functional evaluation under both in vitro and in vivo conditions ( S9–S11 Figs ) . Following an MS-based confirmation of their polypeptide sequence identity ( S9 and S10 Figs ) , circular dichroism ( CD ) experiments reveal that all the hybrid derivatives of ICR-Mo consistently display typical CD spectra of being rich in α-helices ( S11D–S11G Fig and S3 Table ) , almost identical to that of MCR-1 ( S11A Fig ) , EptA ( S11B Fig ) and ICR-Mo ( S11C Fig ) . As observed with soluble domains of both MCR-1 [40–43 , 45] and MCR-2 [46] , the result of ICP/MS ( inductively coupled plasma mass spectrometry ) reveals that zinc is specifically occupied with ICR-Mo and its derivatives ( S12A Fig ) , and the relative ratio of zinc ion to protein is calculated to be around 1:1 ( S12B Fig ) . Assays of antibiotic resistance on LBA plates show that i ) 3 of 4 chimeric versions are inactive; and the only one variant , TM ( MCR-1 ) -AXE82 , has partial activity to allowing the recipient E . coli to grow on the condition with up to 4 μg/ml colistin ( Fig 6C ) , less than those of its parental versions MCR-1 ( 16 μg/ml colistin ) and ICR-Mo ( 8 μg/ml colistin ) ( Fig 6C ) . Similar scenarios were seen in the measurement of colistin MIC ( Fig 6D ) . In addition , MALDI-TOF MS analyses of lipid A species demonstrated that like the wild-type MCR-like proteins ( Fig 7C–7E ) , TM ( MCR-1 ) -AXE82 is the only one hybrid derivative of ICR-Mo having a role in the transfer of the PEA moiety from PE to the acceptor LPS-lipid A ( Fig 7H ) . In contrast , the other 3 derivatives ( Fig 7G–7I ) , have no detectable activities similar to those of negative controls ( Fig 7A and 7B ) . Evidently , MS-based visualization of altered structures of lipid A moieties agrees with scenarios observed in the both enzymatic tests and colistin resistance assays ( Fig 6 ) . In summary , these data suggest that protein evolution has created a gradient with respect to the acquisition/differentiation of antibiotic resistance amongst EptA , ICR-Mo and MCR-1/2 to some extent . Successful killing and eradication of the Gram-negative bacteria by polymyxins is determined by two critical factors: i ) Efficient binding/entry to the initial target , bacterial LPS-lipid A [9]; ii ) Activation of a downstream hydroxyl radical death pathway [47 , 48] . The aforementioned results have proved that ICR-Mo can modify the lipid A moieties , whereas it is unclear as to whether or not it influences the downstream event . Therefore , we integrated two different methods [confocal microscopy ( Figs 8 and 9 ) and chemical rescue assay ( Fig 10 ) ] to address this question . First , an oxidant-sensitive dye DCFH2_DA ( 2′ , 7′-dichlorodihydrofluorescein diacetate ) was applied in monitoring intra-cellular H2O2 species in E . coli with/without ICR-Mo ( or EptA/MCR-1 ) ( Fig 8 ) . As expected , the treatment with colistin boosts ROS production in the E . coli MG1655 ( Fig 8A and 8B ) . However , expression of icr-Mo alleviates colistin-triggered ROS formation regardless of colistin treatment ( Figs 8G and 9 ) . This is in agreement with observations in E . coli strains expressing either EptA ( Fig 8C and 8D ) or MCR-1 ( Fig 8E and 8F ) . In chemical rescue trials , the presence of a ferric chelator , bipyridine , significantly bypasses cell death of the icr-Mo-negative E . coli with that are subjected to treatment with colistin ( Fig 10A and 10C ) . It seems likely that a Fenton reaction is involved in the formation of free hydroxyl radicals ( Fig 10A ) . Similarly , the presence of the ROS scavenger , L-cysteine alone [or accomplished with bipyridine] also greatly improves bacterial survival of E . coli under conditions of colistin treatment ( Fig 10A and 10C ) . Intriguingly , the presence of icr-Mo ( or eptA/mcr-1 ) prevents the recipient E . coli from entering into the hydroxy radical death pathway in E . coli ( Fig 10B and 10C ) . Of note , this impact is independent of presence of pyridine ( and/or L-cysteine ) . Thus , we believe that ICR-Mo ( EptA/MCR-1 ) modifies the bacterial LPS-covered membrane , prevents efficient entry of polymyxin into cells , quenches/alleviates ROS production in vivo , and consequently bypasses antibiotic killing by colistin ( Fig 10 ) .
Polymyxin is a paradigm for cationic antimicrobial polypeptides ( CAMP ) . Physical basis for CAMP-type resistance is due to structural alteration of bacterial LPS , leading to a reduction of the net negative charge of the cell surface . In total , three types of chemical modifications of lipid A species are involved in surface remodeling of Gram-negative bacterium . Among them , the most prevalent form refers to the modification of the LPS-lipid A with an addition of the cationic sugar 4-amino-4-deoxy-l-arabinose [49–51]; Second , the transfer of PEA to the 1 ( or 4’ ) -phosphate position of lipid A moieties [16 , 17 , 51–54]; Third , the glycine attachment to LPS-lipid A GlcN moieties by a tripartite system [Vc1577 ( AlmG ) , Vc1578 ( AlmF ) , and Vc1579 ( AlmE ) ] in Vibrio cholerae [55] . It is known that rare cases of an intrinsic colistin resistance are generated by spontaneous point-mutations of chromosomal genes , especially two-component systems of PhoB/Q [53 , 56] and PmrA/B [17 , 52–54] . In addition , the EptA enzyme of the Neisseria species , which is a representative member of PEA lipid A transferase family , also contributes to the generation of an intrinsic resistance to colistin [16] . In contrast , the discovery of MCR-like determinants belonging to PEA lipid A transferases constitutes a new mechanism for the transferability of polymyxin resistance [10 , 14 , 21 , 22 , 24–26] . It seems more worrisome that global transmission of mcr-1 genetic determinant on diversified plasmids raises significant challenge to clinical therapy and public health [11 , 57 , 58] . Therefore , it is essential to understand the evolution of mcr-1-like variants . Very recently , Moraxella species have been proposed as a reservoir for mcr-1/2 genetic determinants [30–32] . Our phylogenetic analyses showed that the MCR-1/2 family is much more closely related to the Moraxella family of MCR-like genes than either Sulfatases or other non MCR-like genes . In fact , the pap2 gene originally found next to mcr-1 on plasmids , has also been detected on the chromosome next to mcr-1 ( and its genetic variants ) in certain other species of Moraxella [30] . Given the fact that i ) both M . pluranimalium [59] and M . porci [60] colonize in pigs; and ii ) pigs are major host reservoirs for mcr-1-harboring Enterobacteriaceae [10] , it is ecologically reasonable that genetic exchange of mcr-1/2-like determinants between Moraxella species and Enterobacteriaceae like E . coli occur . This leads us to believe that evaluating the chromosomally-encoded genetic determinants of colistin resistance in Moraxella might further our understanding of evolution in the MCR1/2 family . The data we report here presents a biochemical and physiological understanding of ICR-Mo , found on the chromosome of Moraxella osloensis . Relative to MCR-1 , ICR-Mo confers a lower level of resistance to colistin to a recipient E . coli strain ( S6 Fig ) , which argues the possibility that mcr-like genes are in either a developing or degenerating state [61] . Of note , any direct evidence of ICR-Mc [33] or ICR-Mo conferring resistance to colistin in native Moraxella strains is also lacking . The two domains ( TM region and extra-domain of PEA transferase ) of MCR-1 and MCR-2 are functionally-exchangeable [22] , while those in ICR-Mo and MCR-1 cannot be fully switched ( Fig 6 ) . This indicates a fundamental difference between these two proteins that further validates their phylogenetic placement . Structural and functional characterization of ICR-Mo facilitates the identification of an important PE substrate cavity , which is essential for its enzymatic activity and for the conferring a phenotypic resistance to colistin ( Figs 3 and 4 ) . In light of its similarity to those of EptA and MCR-1 ( Figs 3 and 4 ) , this hints an evolutionarily-conserved mechanism for intrinsic and transferable colistin resistance . In fact , this is in close agreement with observations made from a very recent X-ray crystal structure of an intrinsic colistin resistance gene product , ICR-Mc , in Moraxella catarrhalis [33] . Moreover , ICR-Mo displays similar abilities in both modifying lipid A structure ( Fig 2 ) and quenching the production of ROS in vivo ( Figs 8 and 9 ) , when compared with that of MCR-1 . Intriguingly , it seems likely that a putative “ping-pong” mechanism is shared amongst these MCR-like enzymes ( S1 Fig and Fig 2 ) . It was thought that the ancestral source of colistin resistance might be from proteins containing sulfatase/hydrolase domains . These domains possess the same catalytic architecture as PEA transferases . Further , polymyxin is naturally produced by ( and originally isolated in ) certain members of the Paenibacillus family as a secondary metabolite . However , based on our biochemical and phylogenetic data it is tough to conclude whether the members of the non MCR-like family ( comprising MCR-3/4 ) , which cluster closely with the Sulfatase family , are in a state of development or degeneration of colistin resistance . Thus , we simply propose that ICR-Mo and the other members of the Moraxella family have a same ancestor as the MCR-1/2 family does , providing a current source of genetic variation . In response to the emergence of MCR-1 colistin resistance in the very late of 2015 [10] , Chinese agricultural government has taken on a more active role in preventing the further spread of MCR-1 colistin resistance by formally banning the use of colistin as a growth booster in China in early 2017 . Moreover , it is also important to reconsider the clinical use of colistin as a final line of refuge against lethal infections with carbapenem-resistant superbugs . Our findings represent a full mechanistic understanding of ICR-Mo , a representative member of a family of chromosomal relatives to transferable MCR-1 colistin resistance . It might provide molecular basis for the rational development of small molecules targeting the reversal of MCR-like resistance to colistin , a last-resort antibiotic .
All the strains used here are derivatives of E . coli MG1655 ( S1 Table ) . As described with mcr-2 [22] , the full-length icr-Mo ( AXE82_07515 ) was synthesized in vitro . To generate hybrid versions of icr-Mo and mcr-1 ( and/or eptA ) , overlapping PCR experiments were conducted [22] . The point-mutants of icr-Mo were produced using the Mut Express II Fast Mutagenesis Kit V2 ( Vazyme Biotech Co . , Ltd ) with suitable primers ( S1 Table ) . The pET21a expression vector was used for protein production , and arabinose-inducible plasmid pBAD24 was utilized for functional complementation ( S1 Table ) . All the constructs were confirmed with direct DNA sequencing . Luria-Bertani ( LB ) broth ( either liquid culture or solid agar plates ) was applied and appropriate antibiotics such as ampicillin and colistin were supplied . The minimum inhibitory concentration ( MIC ) of colistin was determined using a liquid broth dilution test as recommended by EUCAST with Cation-adjusted Mueller-Hinton Broth ( CAMHB ) [22] . When necessary , 0 . 2% arabinose was added into CAMHB media to induce the expression of pBAD24-borne icr-Mo and its derivatives in E . coli . Also , the viabilities of E . coli carrying icr-Mo and its derivatives were judged with solid LBA broth dilution test [14 , 62 , 63] . As descried with MCR-2 [22] with little change , ICR-Mo and its chimeric versions were overexpressed . The membrane fraction containing the protein of interest was solubilized in buffer B ( 20 mM Tris-HCl [pH 8 . 0] , 100 mM NaCl , 5% glycerol , 1% DDM ( M/V ) ) and then centrifuged at 38000 r . p . m for 1 . 5 h at 4°C . The resultant supernatant was subjected to affinity purification of protein with pre-equilibrated Ni-NTA agarose beads . The expected integral membrane proteins of ICR-Mo and its derivatives were eluted using an elution buffer ( 20 mM Tris-HCl [pH 8 . 0] , 100 mM NaCl , 100 mM imidazole , 5% glycerol , 0 . 03% DDM ( M/V ) ) [14 , 22] . Following determination of the purified protein with 12% SDS-PAGE , the protein band with expected size was cut from gels and subjected to MS-based identification . The data of acquired polypeptides was confirmed by BLAST against NCBI NR database . Circular dichroism ( CD ) tests were conducted to characterize secondary structures of ICR-Mo and its chimeric variants in the Tris buffer [20 mM Tris-HCl [pH 8 . 0] , 100 mM NaCl , 5% glycerol , 0 . 03% DDM] . The CD spectra were recorded on a Jasco Model J-1500 spectrometer ( Jasco Corp . , Tokyo , Japan ) through continuous wavelength scanning ( in triplicate ) from 190 to 240 nm at a scan rate of 50 nm/min [64] and smoothed with a Savitsky-Golay filter [65] . CD spectra were analyzed using SELCON3 program ( protein basis set 10 ) in the CDPro software package ( http://lamar . colostate . edu/~sreeram/CDPro/ ) , developed at the Department of Biochemistry and molecular Biology of Colorado State University [66] . The different percentages were measured , which corresponded to four major types of secondary structure motifs ( α-helices , β-sheet , η-turn and coils ) [67] . To further probe whether or not zinc ions are occupied in ICR-Mo and their derivatives , inductively coupled plasma mass spectrometry ( ICP-MS ) was performed . The protein samples ( ~0 . 2 mg/ml ) were subjected to the NexION 300X ICP-MS instrument ( PerkinElmer , USA ) with helium as carrier gas [68] . The mass-to-charge ratio ( m/z ) was measured in the mode of kinetic energy discrimination ( KED ) mode . In total , seven proteins were examined , which include 3 parental enzymes ( ICR-Mo [AXE82_07515] , EptA and MCR-1 ) and 4 hybrid versions ( namely TM ( AXE82 ) -EptA , TM ( EptA ) -AXE82 , TM ( AXE82 ) -MCR-1 and TM ( MCR-1 ) -AXE82 ) . As described with EptA by Anandan et al . [34] , we tested the enzymatic activity of ICR-Mo in vitro . In this enzymatic reaction system , the fluorescent substrate , is abbreviated as NBD-PEA from 1-acyl-2-{12-[ ( 7-nitro-2-1 , 3-benzoxadiazol-4-yl ) amino] dodecanoyl}-sn-glycero-3-phosphoethanolamine ( Avanti Lipids , USA ) . The reaction system ( 50 μl in total ) consists of 50 mM HEPES ( pH 7 . 50 ) , 100 mM NaCl , 0 . 03% of DDM , 0 . 2 mM NBD-PEA and 40 μM ICR-Mo [MCR-1/MCR-2 and derivatives] . The reaction proceeded at 25°C for around 24 hrs [36 , 37] . Thin layer chromatography ( TLC ) was performed to detect the presence of the NBD-glycerol product hydrolyzed from its substrate NBD-glycerol-3-PEA . Of note , fluorescence signal of NBD-glycerol-3-PEA ( and/or NBD-glycerol ) separated with TLC plates was visualized by Epi blue light ( 455–485 nm ) and a corresponding filter of a ChemiDoc MP imaging system ( Biorad , CA , USA ) [34] . The identity of both NBD-glycerol-3-PEA and NBD-glycerol was determined by Liquid Chromatography Mass Spectrometry ( LC/MS ) system ( Agilent technologies 6460 Triple Quad LC/MS ) [69] . Mass spectrometry was coupled with electrospray ionization ( ESI ) source , in which neutral loss ion ( m/z 141 ) mode was set for scanning of the positive ion . The samples were eluted with the solution of methanol/0 . 1% methanoic acid ( 95:5 ) at 0 . 3 ml/min and separated with an analytical chromatographic column of Zorbax SB C18 ( 2 . 1*50 mm , 3 . 5 μm ) LPS-lipid A was extracted as described by Liu et al . [70] . Following separation with SDS-PAGE ( 10% ) , the purity of LPS specimens was judged with silver staining [36 , 37] . The lipid A species that satisfied the purity criteria were then subjected to structural identification with MALDI-TOF-MS ( Bruker , ultrafleXtreme ) in negative ion mode with the linear detector [70] . Mid-log phase cultures of E . coli strains were used in the challenge by colistin ( 4 μg/ml , 0 . 5 h ) . The oxidant sensor dye DCFH2-DA ( i . e . , 2′ , 7′-dichlorodihydrofluorescein diacetate , Sigma ) was utilized to detect intra-cellular accumulation of reactive oxygen species ( ROS ) . The fluorescent dichlorofluorescein ( DCF ) , the oxidation product of DCFH2-DA , was visualized using a Zeiss LSM 510 Meta confocal laser scanning microscope [71] . Totally , four types of E . coli strains used here correspond to FYJ796 [MG1655 with empty vector] which served as the negative control , FYJ832 [MG1655 carrying pBAD24::eptA] , FYJ795 [MG1655 carrying pBAD24::mcr-1] and FYJ968 [MG1655 with pBAD24::axe82_07515] , respectively ( S1 Table ) . Using two ROS inhibitors: bipyridine ( a ferric chelator ) and L-cysteine ( a ROS scavenger ) , chemical rescue experiments were carried out as described by Collins and coauthors [72 , 73] with minor modifications . Following different challenges , bacterial viability of E . coli strains ( with or without icr-Mo ( or mcr-1/2 ) ) was recorded . The four different kinds of challenges employed here were i ) colistin alone; ii ) colistin combined with 2 , 2’-dipyridine [74]; iii ) colistin mixed with L-cysteine [75] and iv ) the mixture consisting of colistin , 2 , 2’-dipyridine [74] and L-cysteine [75] . In this assay , compounds were supplemented as follows: 20 μg/ml for colistin ( Sigma ) , 500 μM for 2 , 2’-dipyridine ( Sangon Biotech ) and 10 mM for L-cysteine ( Sangon Biotech ) . After 30 mins of incubation at 37°C , cultures were serially diluted . The selected dilutions ( 10−4~10−7 ) were dropped onto LB agar plates ( 5 μl each ) . Colony forming units ( CFU ) were enumerated [76] . Sequence alignment of ICR-Mo ( AXE82_07515 ) with its paralogues was conducted with Clustal Omega ( http://www . ebi . ac . uk/Tools/msa/clustalo/ ) and processed via the program ESPript 3 . 0 ( http://espript . ibcp . fr/ESPript/cgi-bin/ESPript . cgi ) [77] . The trans-membrane region of ICR-Mo was predicted using TMHMM server v2 . 0 ( http://www . cbs . dtu . dk/services/TMHMM/ ) . Using the N . meningitis EptA ( PDB accession number , 5FGN ) as structural template [34] , architecture of ICR-Mo was modeled via Swiss-Model [35] . The value of GMQE ( Global Model Quality Estimation ) is 0 . 68 and the score of QMEAN ( which provides a global and local absolute quality estimate on the modeled structure [78] ) is -3 . 62 . Thus , it suggests a reliable qualified structural prediction . The ready-to-dock chemical structure of PE ( ID: ZINC32837871 ) and head group of PE ( ID: ZINC02798545 ) was sampled from ZINC database [79] . The UCSF DOCK 6 software ( version 6 . 7 ) was applied to predict binding patterns of PE molecule vs EptA and head group of PE to ICR-Mo [80] . LigPlot+ was used to illustrate the diagrams for possible ligand-protein interaction [81] . Concretely , protein structure was processed for molecular docking using UCSF Chimera software [82] . Solvent molecules were removed . Hydrogens were added and charges were assigned using chimera tool Dock Prep . Preferred orientation of PE in EptA was searched in 20 Å space around complexed ligand dodecyl-β-D-maltoside ( DDM ) in 5FGN . A set of spheres located in 20 Å space around DDM was calculated by the use of the program Sphere_select . The accessory program GRID in DOCK6 software package was used to compute van der Waals potential grid and electrostatic potential grid for energy scoring . Given that PE molecule contains many rotatable bonds and is thus more flexible , anchor-and-grow algorithm was set for conformational search in docking studies . This type of flexible ligand docking allows the ligand to structurally rearrange in response to the receptor . The nucleotide sequence of AXE82_07515 ( designated as icr-Mo ) from M . osloensis was used as a query to perform a nucleotide BLAST with options enabled to exclude models and uncultured environmental samples . The blastn algorithm parameters were modified to display 1000 target sequences with the query word size set to 7 , allowing it to return ‘somewhat similar sequences’ to the query . All sequences with a minimum of 30% identity and greater than 70% query coverage were selected and exported . In addition , members of the distantly related Sulfatase family , if not identified in the blast search , were included . Unique nucleotide sequences were identified using Uniqueseq ( https://www . ncbi . nlm . nih . gov/CBBresearch/Spouge/html_ncbi/html/fasta/uniqueseq . cgi ) and aligned using MUSCLE ( https://www . ebi . ac . uk/Tools/msa/muscle/ ) . A total of 54 unique nucleotide sequences were utilized for the subsequent phylogenetic analysis . The 5’ end of the aligned sequences was trimmed to improve conservancy and then utilized for phylogenetic analysis . jModeltest ( via MEGA 7 [83] ) was used to identify the best-fit nucleotide substitution model . Using MEGA 7 [83] , the best model obtained from before was used to generate a maximum-likelihood tree with 1000 bootstrap replicates . Initial tree ( s ) for the heuristic search were obtained automatically in MEGA 7 by applying Neighbor-Joining and BioNJ algorithms [84] to a matrix of pairwise distances estimated using the Maximum Composite Likelihood ( MCL ) approach , and then selecting the topology with superior log likelihood value . A GTR model with discrete Gamma distribution and Invariant sites [85] was used to model evolutionary rate differences among sites ( 5 categories ( +G , parameter = 1 . 1625 ) ) . | This work represents the first functional definition of ICR-Mo , a member of a family of MCR-1/2 like genes from Moraxella . As inferred by phylogeny , this is the closest family of chromosomally encoded determinants of colistin resistance . We show evidence that ICR-Mo exploits a possible “ping-pong” mechanism to catalyze the transfer of PEA from its donor PE to the 1 ( or 4’ ) -phosphate position of lipid A via an adduct of ICR-Mo-bound PEA . Structure-guided functional studies reveal that ICR-Mo possesses a PE lipid substrate recognizing cavity having a role in enzymatic catalysis and the resulting bestowal of antibiotic resistance against colistin . The expression of icr-Mo in E . coli prevents the generation of ROS in response to treatment with colistin . The fact that the two domains of ICR-Mo cannot be functionally exchanged with their counterparts in MCR-1/2 validates its standing as a distinct phylogenetic entity . In summary , the results define a closer chromosomal relative of MCR-1/2 variants than EptA and highlight its role in the evolution of transferable colistin resistance . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"enzymes",
"drugs",
"microbiology",
"enzymology",
"organisms",
"planar",
"chromatography",
"antibiotic",
"resistance",
"antibiotics",
"pharmacology",
"plants",
"thin-layer",
"chromatography",
"research",
"and",
"analysis",
"methods",
"legumes",
"polymyxins",
"lipids",
"antimicrobial",
"resistance",
"proteins",
"transferases",
"chemistry",
"peas",
"mass",
"spectrometry",
"matrix-assisted",
"laser",
"desorption",
"ionization",
"time-of-flight",
"mass",
"spectrometry",
"chromatographic",
"techniques",
"biochemistry",
"analytical",
"chemistry",
"eukaryota",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"spectrum",
"analysis",
"techniques"
] | 2018 | Defining ICR-Mo, an intrinsic colistin resistance determinant from Moraxella osloensis |
Chikungunya virus ( CHIKV ) , an arbovirus , is responsible for a two-stage disabling disease , consisting of an acute febrile polyarthritis for the first 10 days , frequently followed by chronic rheumatisms , sometimes lasting for years . Up to now , the pathophysiology of the chronic stage has been elusive . Considering the existence of occasional peripheral vascular disorders and some unexpected seronegativity during the chronic stage of the disease , we hypothesized the role of cryoglobulins . From April 2005 to May 2007 , all travelers with suspected CHIKV infection were prospectively recorded in our hospital department . Demographic , clinical and laboratory findings ( anti-CHIKV IgM and IgG , cryoglobulin ) were registered at the first consultation or hospitalization and during follow-up . Among the 66 travelers with clinical suspicion of CHIKV infection , 51 presented anti-CHIKV IgM . There were 45 positive with the serological assay tested at room temperature , and six more , which first tested negative when sera were kept at 4°C until analysis , became positive after a 2-hour incubation of the sera at 37°C . Forty-eight of the 51 CHIKV-seropositive patients were screened for cryoglobulinemia; 94% were positive at least once during their follow-up . Over 90% of the CHIKV-infected patients had concomitant arthralgias and cryoglobulinemia . Cryoglobulin prevalence and level drop with time as patients recover , spontaneously or after short-term corticotherapy . In some patients cryoglobulins remained positive after 1 year . Prevalence of mixed cryoglobulinemia was high in CHIKV-infected travelers with long-lasting symptoms . No significant association between cryoglobulinemia and clinical manifestations could be evidenced . The exact prognostic value of cryoglobulin levels has yet to be determined . Responsibility of cryoglobulinemia was suspected in unexpected false negativity of serological assays at room temperature , leading us to recommend performing serology on pre-warmed sera .
Chikungunya fever is an emerging arboviral disease characterized by a brief fever , headache and myalgias , occasional evanescent rash , inflammatory polyarthralgias , arthritides or tenosynovitis that can last for months to years [1]–[4] . Chikungunya virus ( CHIKV ) was identified in the 1950s in Africa [2] , and soon after in Asia [5] . It can be responsible for major epidemics , sometimes separated by silent periods [6] . From 2004 to 2006 , a giant CHIKV outbreak successively swept out Kenya and most islands of western Indian Ocean [7] . In Réunion Island , the outbreak was explosive at the beginning of 2006 with a pick of 45 , 000 cases per week . Up to 2006 June 1st , about one third of the 770 , 000 residents had been infected [8] . Another huge outbreak recently stroke India with 2 to 7 million estimated cases [9] , and is currently spreading to Southeastern Asia [10] . During this period , Chikungunya fever was also identified in more than 1 , 000 travelers returning from the epidemic areas to European countries [1] , [11] , [12] and the USA [13] , [14] . CHIKV-infected travelers included viremic patients who returned home to countries where competent vectors are present , raising serious concern for the globalization of the disease [7] , [15] . The Italian outbreak in August 2007 has demonstrated the reality of this threat [16] . During the recent CHIKV outbreaks , previously described clinical features [3] , [4] as well as the low rate of asymptomatic infections were confirmed [7] , [17] . The disease was also responsible for unusual and unfrequent complications , including severe newborn infections after peripartum mother-to-infant transmission , meningo-encephalitis , hepatitis , myocarditis , severe epidermolysis [17] , [18] and lead to surmortality [8] . Transitory peripheral vascular disorders ( PVD ) , mostly Raynaud syndrome , were also observed few weeks after the disease onset [1] . Considering persistent arthralgias and occasional PVD , we hypothesized that cryoglobulin could be involved in the pathophysiology of the disease , as described for hepatitis C infection . In this infection high level of mixed cryoglobulinemia is commonly detected in patients with chronic viral replication and is strongly associated with liver damage and peripheral neuropathies [19] , [20] .
From April 2005 throughout May 2007 , all patients with possible imported CHIKV infection ( recent travel in Indian Ocean islands and presence or history of fever and/or arthralgias ) were prospectively recorded at Laveran Military Hospital in Marseille , France . The criteria for confirmed cases were i ) presence of specific anti-CHIKV IgM and/or positive RT-PCR and/or isolation of CHIKV from blood and/or specific anti-CHIKV IgG , ii ) recent clinical feature consistent with CHIKV infection , iii ) no other etiology identified . Demographic , clinical and laboratory findings were registered for all patients at their first consultation or hospitalization and during follow-up . For patients seen more than 10 days after the onset of illness , early clinical features were identified using a retrospective questionnaire . The clinical status of each patient was actively monitored by the same physician ( FS ) during consultations and/or by phone calls every 2–3 months . The early stage is defined as the first 10 days of clinical disease while second stage is defined as symptoms and signs persisting more than 10 days after disease onset [1] . The same physician ( FS ) orally informed all patients about the requirement of blood samples to diagnose the aetiology of their polyarthralgia . The script for obtaining oral consent was accepted by the Institutional Review Board at Laveran Hospital . All patients gave their oral consent , which was noted in their individual medical file . A single blood sample of less than 50 cc was taken after consultation , when clinically required; no supplementary blood samples were taken for research . At each consultation , patients were tested for CHIKV serology using in-house IgM-capture and IgG-sandwich enzyme-linked immunosorbent assays [6] . Serologies were performed first on sera kept at 4°C before analysis and second , on sera kept at 37°C until analysis . IgG and IgM for West Nile , dengue and Rift Valley fever viruses were also assayed . On acute sera , RT-PCR and isolation of CHIKV in Vero-E6 cells were attempted [21] . Cryoglobulinemia was screened using the following procedure: i ) collection of 14 ml of blood in pre-warmed tubes at the hospital's biochemistry laboratory , ii ) clotting at 37°C for 3 hours , iii ) centrifugation at 37°C ( 3 , 000 RPM for 10 minutes ) , iiii ) freezing of the pellet at 4°C for 10 days , to allow generation of cryoprecipitates . Purification , characterization and quantification of cryoglobulins were performed by the same operator ( MO ) , as previously described [22] . Cryoglobulins were classified as type I for monoclonal component , type II for one monoclonal component associated with polyclonal immunoglobulins , type II–III for more than one monoclonal component associated with polyclonal immunoglobulins , and type III for polyclonal immunoglobulins [23] , [24] . Type III cryoglobulinemia was considered positive when level was higher than 5 mg/L [22] , whereas type II and type II–III cryoglobulins were considered positive regardless of concentration level [22] . CHIKV RNA in cryoprecipitates was searched using RT-PCR , as described elsewhere [21] . Patients also underwent immunological tests: ( i ) determination of C3 and C4 complement components using an immunonephelemetric method ( Dade Behring Paris France ) , ( ii ) determination of total haemolytic complement using the total haemolytic complement kit ( The Binding Site Saint Egreve France ) , ( ii ) search for rheumatoid factors , by latex immunoagglutination ( Biomérieux Marcy l'Etoile France ) , ( iv ) detection of antinuclear antibodies by indirect immunofluorescence ( Eurobio Paris France ) . Serological screening for hepatitis C virus ( HCV ) infection was systematically performed using sandwich enzyme-linked immunosorbent assays ( Biorad Marne La Coquette France ) . After 6 months , blood testing for asymptomatic patients was restricted to CHIKV serology and detection of cryoglobulinemia . Chi 2 , Kruskall-Wallis and exact Fisher's , two-tailed probability tests were used with a significant p value of 0 . 05 ( Stata 9 . 0 Software , StataCorp College Station , Texas , USA ) .
During the 25 month-study , 66 French patients with clinical suspicion of CHIKV infection were prospectively included . Among them , 51 presented with anti-CHIKV IgM ( see below ) . Fifteen patients remained seronegative for both anti-CHIKV IgM and IgG . No patients had detectable CHIK virus in sera ( RT PCR and CHIK V isolation on Vero cell were negative ) . Sex ratio ( M/F ) and median age in seropositive patients were respectively 1 . 04 and 54 years ( range: 21 y–78 y ) , versus 0 . 67 and 47 years ( 21 y–75 y ) in seronegative patients ( no statistical difference ) . Six CHIKV-infected patients were lost to follow-up after the first consultation . The 45 other patients have been followed in our medical unit for a median duration of 14 months ( range: 54 days–25 months ) , through 2 to 6 consultations per patient; resulting in a total of 138 consultations for CHIKV patients performed up to May 2007 . Table 1 summarizes the demographic , epidemiological and clinical events of the 51 confirmed cases . Ninety-eight percent of the seropositive cases suffered at least once with arthralgia , 71% with tenosynovitis and 20% with transitory PVD . Thirteen patients experienced at least one clinical relapse during follow-up , i . e . became symptomatic again after at least one symptom free month , mostly with subacute arthralgias in hands and feet . Thirteen patients developed de novo transitory PVD , mainly in fingers and sometimes in toes , within the second and third months after the disease onset . No other aetiology ( drug , auto-immune disorder , Buerger disease , local trauma ) than CHIKV was identified for relapses or PVD . No morphological changes were observed in the 3 patients for whom a digital capillaroscopy was performed . PVD was not significantly associated with gender . Conversely , tenosynovitis was significantly more frequent in women , both on the day of consultation ( p = 0 . 01 ) or during the previous month ( p = 0 . 04 ) . Among the 51 CHIKV-infected patients , cryoglobulinemia was screened for 48 , i . e . 42 patients on first consultation and 6 later during follow-up . Cryoglobulinemia screening was repeated 2 to 5 times for 38 patients during the study period . The median delay between CHIKV acute symptomatology onset and first cryoglobulinemia testing was 45 days ( range: 5–640 d ) . On first consultation , 37/42 patients ( 88% ) were cryoglobulinemic . Among the five patients without cryoglobulinemia: i ) three symptomatic patients respectively free of cryoglobulinemia on days 5 , 16 and 36 after disease onset , developed cryoglobulinemia on days 27 , 45 and 69 , respectively; ii ) one symptomatic patient free of cryoglobulinemia on his first test at day 124 ( a week after a short-term corticotherapy ) became positive with a type II on day 204; iii ) one patient had been symptom-free for weeks and was not cryoglobulinemic when tested on day 107 after disease onset . Among the 6 patients first tested at the second consultation , 4 were cryoglobulinemic and two remained negative for cryoglobulin on day 143 and day 355 , respectively , while they were both symptom-free . Finally , 94% of the 48 CHIKV-infected tested patients were positive for cryoglobulinemia at least once during their follow-up . During the study period , 118 cryoglobulin assays were performed , 83 were positive . There were 61% ( 51/83 ) type II , 30% ( 25/83 ) type II–III and 8% ( 7/83 ) type III . Monoclonal IgM-κ and IgM-λ components were found in 56/83 cryoglobulins ( 68% ) and 48/83 ( 58% ) , respectively; polyclonal IgM components in 83/83 cryoglobulins ( 100% ) . Monoclonal IgA was found in only 2/83 cryoglobulins . Thus , all cryoglobulins associated with CHIKV infection were mixed cryoglobulins . The cryoglobulin concentration could be determined for 79/83 cryoglobulinemias . The median cryoglobulin level was 8 mg . L−1 ( range: 3–165 ) . No CHIKV was detected by RT-PCR in 24 tested cryoprecipitates . Neither antinuclear antibody nor rheumatoid factors were found during follow-up . C3 , C4 complement fractions and haemolytic complement levels remained normal . All patients were seronegative for HCV .
Our study was an empiric prospective study of symptomatic CHIKV-infected travelers coming back from Western Indian Ocean . Thus , follow-up for cryoglobulinemia was not performed in seronegative patients in the absence of ethic committee consent . Up to now , pathogenesis of CHIKV-induced rheumatism remains unknown . Fourie et al . identified low titers of rheumatoid factor in CHIKV infection [3] , whereas no antinuclear antibodies or rheumatoid factors were detected here like in previous studies [4] . We identified for the first time the association between CHIKV infection and cryoglobulinemia among infected travelers . As no type I cryoglobulin ( composed of one monoclonal immunoglobulin ) was detected in our patients , CHIKV seems to induce only mixed cryoglobulinemia ( MC ) , either type II , II–III or III . MC has already been described in miscellaneous acute and chronic infections [19] , [20] . In these cryoglobulinemia seemed to rapidly decrease and disappear within a few weeks of pathogen clearance , although persisting MC has been observed in few chronic infections , the most common being chronic hepatitis C virus ( HCV ) ( MC prevalence range: 20–55% ) [19] , [20] . In our study , the strength of association between CHIKV infection and MC was much higher 94% . A biased high prevalence of CHIKV-MC due to use of a highly sensitive method can be ruled in regard of the frequent negativity of cryoglobulinemia detection over time during follow-up . The persistence of CHIKV-associated MC ( CHIKV-MC ) was unexpected , as the CHIKV genome has never been detected in blood after 12 days of disease evolution , even in chronic symptomatic patients [25] . Consistently , we failed to detect CHIKV-RNA in the cryoprecipitates , as described in chronic HCV infection [26] , [27] . The absence of chronic viremia could also explain the much lower level of CHIKV-MC than those observed in chronic HCV infection [22] . Ozden et al . recently showed the presence of CHIKV in muscle satellite cells of a non-immunocompromised patient who was still painful three months after disease onset [28] . Moreover , Jaffar-Bandjee et al . , using molecular tools , evidenced persistence of CHIKV in perivascular macrophages of synovial tissue in a chronic elbow hygroma of a patient infected for one year [29] . This discovery suggests that CHIKV can induce a chronic infection with active replication , conversely to most other arboviruses . Persistence of CHIKV also seemed to stimulate host immunity , as inflammation with macrophages and T cells was locally concomitantly observed . Thus , viral replication could be involved in the prolonged persistence of anti-CHIKV IgM and , as a consequence , of CHIKV-MC in patients with long-persisting symptoms . CHIKV-MC should be the possible “missing link” between CHIKV and some symptoms of the chronic stage . The second main finding is the existence of false seronegativity in one out of three and a half CHIKV-infected patient with typical clinical presentation of CHIKV infection when using “classical” ELISA assay at room temperature . This phenomenon can be responsible for non-recognition of CHIKV infection in individuals with chronic rheumatism , leading to useless explorations , as well as underestimation of seroprevalence in an endemic/epidemic area . Considering the high prevalence of CHIKV-MC in our cohort , we assume that cryoglobulinemia may be a significant factor in misdiagnosing the disease when using “conventional” serology . Specific anti-CHIKV antibodies , i . e . IgM and IgG , could all be trapped in the cryoprecipitate , as already described in chronic HCV infection [30] . Therefore , when facing a patient with a clinical suspicion of CHIKV infection and with paradoxical seronegativity , we recommend the following procedure: at least pre-warmed the serum at 37°C before serology testing or , preferably , manage the blood sample as required for any cryoglobulin research: sampling and centrifugation at 37°C , decantation and serum pre-warming before the ELISA assays . However , considering the high prevalence of MC among our CHIKV-seronegative patients ( 40% ) and the unexpected persistent seronegativity in a few patients with a high clinical suspicion of CHIKV-infection , we cannot exclude some residual false negative results due to the precipitation of CHIKV-MC before blood centrifugation . Chronic MC is commonly associated with various symptoms , mostly arthralgias , purpura , weakness and Raynaud syndrome [19] , [20] . In our study , no CHIKV-infected patient presented the complete clinical triad associated with MC , but arthralgias were present in all and combined with PVD in 24% . The concomitancy of arthralgias and cryoglobulinemia in more than 90% of the cases is consistent , although not demonstrative , with an involvement of cryoglobulin in arthralgias . Frequent exacerbation of rheumatic pain and handicap , and the occasional incidence of transitory PVD have been recently described within the 2nd and/or 3rd months after CHIKV-infection onset [1] . These clinical manifestations are synchronous with increasing cryoglobulin levels . The low cryoglobulin blood levels -when compared with chronic HCV infection- could explain the lack of vascular purpura in our cohort . No significant association between PVD and CHIKV-MC prevalence was observed , possibly due to the small number of patients , although there was a positive trend between presence of PVD and CHIKV-MC levels ( p = 0 . 06 ) . We failed to find any significant association between CHIKV-MC type or level and any other sign or symptom . Most patients self-declared improvements after short-term general corticotherapy . In few of them we could evidence concomitant CHIKV-MC disappearance . Corticosteroids could interfere through two actions: immunosuppression of B lymphocyte activation and/or decrease of joint and tendon inflammation . Randomized studies are needed to confirm the benefits and risks of general corticotherapy and to specify therapeutic modalities in chronic CHIKV-associated rheumatism . Complementary investigations are also required to determine whether antiviral drugs , such as chloroquine or interferon could be helpful in stopping CHIKV replication in muscle satellite cells and MC production in symptomatic chronic patients , as suggested elsewhere [31] . Finally , the present work also shows the evolution of CHIKV-MC types and levels over time . Schematically , type III or type II cryoglobulin appear first ( median delay: around 40 days ) , followed by type II–III ( median delay: around 2 months ) , conversely to what is usually described in HCV infection [32] . Cryoglobulin levels reached their acme within the 3rd month after disease onset and remained stable up to the 6th , before a slow decline . Regarding the parallel evolution of CHIKV-MC and anti-CHIKV IgM antibodies over time and the systematic detection of IgM in CHIKV-MC , the direct involvement of these antibodies in cryoprecipitates can be suspected . After 6 months of evolution , CHIKV-MC levels decrease for most patients , in parallel with obvious clinical improvement . No cryoglobulin was detected in most of patients after they became symptom-free , whatever the delay of recovery . However the evolution was not linear . About one quarter of the patients underwent clinical relapses with distal arthralgias and a concomitant increase or reappearance of cryoglobulin . The disappearance of CHIKV-MC could be predictive for a clinical cure in chronic CHIKV disease , as described in HCV infection after interferon and ribavirin or rituximab treatment [19] . Further studies with long time follow-up are required to determine if type or level of early CHIKV-MC has a real prognostic value . In Brighton's experience , 87 . 9% of 107 CHIKV-infected patients self-declared cured 3 years after disease onset , while 12 . 1% mentioned persistent symptoms including occasional discomfort , persistent joint stiffness , or stiffness and pain and effusion ( 3 . 7% , 2 . 8% , and 6% respectively ) [4] . CHIKV infection is currently spreading in Africa , Indian Ocean , India and Southeastern Asia and threats many others areas where Aedes spp . is present . It is responsible for long-persisting symptoms , which severely impair quality of life of CHIKV-infected patients although natives or travelers . Thus , the identification of very high prevalence of CHIKV-MC is of importance . First , its presence can induce false negativity in serology performed at room temperature; leading to the recommendation of using pre-warmed sera for serology both for individual diagnosis , and seroprevalence estimation in endemic areas . Second , CHIKV-MC may be involved in the pathogenesis of the chronic stage , mainly CHIKV-associated chronic rheumatism and PVD . Third , its prolonged disappearance could be a marker of the definitive clinical cure . | Chikungunya virus is present in tropical Africa and Asia and is transmitted by mosquito bites . The disease is characterized by fever , headache , severe joint pain and transient skin rash for about a week . Most patients experience persisting joint pain and/or stiffness for months to years . In routine practice , diagnosis is based upon serology . Since 2004 there has been an ongoing giant outbreak of Chikungunya fever in East Africa , the Indian Ocean Islands , India and East Asia . In parallel , more than 1 , 000 travelers were diagnosed with imported Chikungunya infection in most developed countries . Considering the clinical features of our patients ( joint pain ) , we hypothesized that cryoglobulins could be involved in the pathophysiology of the disease as observed in chronic hepatitis C infection . Cryoglobulins , which are immunoglobulins that precipitate when temperature is below 37°C , can induce rheumatic and vascular disorders . From April 2005 through May 2007 , we screened all patients with possible imported Chikungunya infection for cryoglobulins . They were present in over 90% of patients , and possibly responsible for the unexpected false negativity of serological assays . Cryoglobulin frequency and levels decreased with time in recovering patients . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"infectious",
"diseases",
"infectious",
"diseases/viral",
"infections"
] | 2009 | Persisting Mixed Cryoglobulinemia in Chikungunya Infection |
The simultaneous expression of the hunchback gene in the numerous nuclei of the developing fly embryo gives us a unique opportunity to study how transcription is regulated in living organisms . A recently developed MS2-MCP technique for imaging nascent messenger RNA in living Drosophila embryos allows us to quantify the dynamics of the developmental transcription process . The initial measurement of the morphogens by the hunchback promoter takes place during very short cell cycles , not only giving each nucleus little time for a precise readout , but also resulting in short time traces of transcription . Additionally , the relationship between the measured signal and the promoter state depends on the molecular design of the reporting probe . We develop an analysis approach based on tailor made autocorrelation functions that overcomes the short trace problems and quantifies the dynamics of transcription initiation . Based on live imaging data , we identify signatures of bursty transcription initiation from the hunchback promoter . We show that the precision of the expression of the hunchback gene to measure its position along the anterior-posterior axis is low both at the boundary and in the anterior even at cycle 13 , suggesting additional post-transcriptional averaging mechanisms to provide the precision observed in fixed embryos .
During development the different identities of cells are determined by sequentially expressing particular subsets of genes in different parts of the embryo . Proper development relies on the correct spatial-temporal assignment of cell types . In the fly embryo , the initial information about the position along the anterior-posterior ( AP ) axis is encoded in the exponentially decaying Bicoid gradient . The simultaneous expression of the Bicoid target gene hunchback in the multiple nuclei of the developing fly embryo gives us a unique opportunity to study how transcription is regulated and controlled in a living organism [1 , 2] . Despite many downstream points where possible mistakes can be corrected [1 , 3 , 4] , the initial mRNA readout of the maternal Bicoid gradient by the hunchback gene is remarkably accurate and reproducible between embryos [5 , 6]: it is highly expressed in the anterior part of the embryo , quickly decreasing in the middle and not expressed in the posterior part . This precision is even more surprising given the very short duration of the cell cycles ( 6–15 minutes ) during which the initial Bicoid readout takes place and the intrinsic molecular noise in transcription regulation [7–9] . Even though most of our understanding of transcription regulation in the fly embryo comes from studies of fixed samples , gene expression is a dynamic process . The process involves the assembly of the transcription machinery and depends on the concentrations of the maternal gradients [10] . Recent studies based on single-cell temporal measurements of a short lived luciferase reporter gene under the control of a number of promoters in mouse fibroblast cell cultures [11 , 12] and experiments in E . Coli and yeast populations [13–16] have quantitatively confirmed that mRNAs are generally produced in bursts , which result from periods of activation and inactivation . In early fly development , what are the dynamical properties of transcription initiation that allow for the concentration of the Bicoid gradient and other maternal factors to be measured in these short intervals between mitoses ? In order to quantitatively describe the events involved in transcription initiation , we need to have a signature of this process in the form of time dependent traces of RNA production . Recently , live imaging techniques have been developed to simultaneously track the RNA production in all nuclei throughout the developmental period from nuclear cycle 11 to cycle 14 [17 , 18] . In these experiments , an MS2-binding cassette is placed directly under the control of an additional copy of a proximal hunchback promoter . As this reporter gene is transcribed , mRNA loops are expressed that bind fluorescent MCP proteins . Their accumulation at the transcribed locus gives an intense localized signal above the background level of unbound MCP proteins ( Fig 1C ) [19] . By monitoring the developing embryo , we obtain for each nucleus a time dependent fluorescence trace that is indicative of the dynamics of transcription regulation at the hunchback promoter ( Fig 1B , 1D and 1F ) . However the fluorescent time traces inevitably provide an indirect observation of the transcription dynamics . The signal is noisy , convoluting both experimental and intrinsic noise with the properties of the probe: the jitter in the signal is not necessary indicative of actual gene switching but could simply result from a momentarily decrease in the recording of the intensity . To obtain a sufficiently strong intensity of the signal to overcome background fluorescence , a long probe with a large number of loops is needed , which introduces a buffering time . In the current experiments the minimal buffering time is the time needed to transcribe a fluorescent probe made of 24 loops . It is τ m i n buff = 72 s and it prevents direct observation of the activation events [19] . To understand the details of the regulatory process that controls mRNA expression we need to quantify the statistics of the activation and inactivation times , as has been performed in cell cultures [11 , 12 , 14 , 15] . However the very short duration of the cell cycles ( 6-15 minutes for cell cycles 11-13 ) in early fly development prevents accumulation of statistics about the inactivation events and interpretation of these distributions . Direct observation of the traces suggests that transcription regulation is not static but displays bursts of activity and inactivity . However the eye can often be misleading when interpreting stochastic traces . In this paper we develop a statistical analysis of time dependent gene expression traces based on specially designed autocorrelation functions to investigate the dynamics of transcription regulation . This method overcomes the analysis difficulties resulting from naturally short traces caused by the limited duration of the cell cycles that make it impossible to infer the properties of regulation directly from sampling the activation and inactivation time statistics . Combining our analysis technique with models of transcription initiation , estimates of the precision of the transcriptional readout and high resolution microscopy imaging of the MS2 cassette under the control of the hunchback promoter in heterozygous flies , we find evidence suggesting bursty transcription initiation in cell cycles 12-13 . For the switching timescales we observe experimentally , the autocorrelation function analysis alone is not able to reliably distinguish between different models for promoter activation and we use information about the precision of the transcriptional readout to conclude that transcription is most likely bursty . Based on the analysis of the time traces , we show that the precision of the transcriptional readout in each cell cycle is relatively imprecise compared to the expected precision of the mRNA measurement obtained from fixed samples , both in terms of cell-to-cell variability [5] and embryo-to-embryo variability [6] . We discuss the limitations of the inference for models of different complexity in different parameter regimes .
Before we present our results , we first analyze the traces and present a new analysis technique . We study the transcriptional dynamics of the hunchback promoter ( depicted in Fig 1A and 1B ) by generating embryos that express an MS2 reporter cassette under the control of the proximal hunchback promoter ( Fig 1C ) , using previously developed tools [17 , 18] , with an improved MS2 cassette [20] ( see Materials and Methods for details ) . The MS2 cassette was placed towards the 3’ end of the transcribed sequence and contained 24 MS2 loop motifs . While the gene is being transcribed , each newly synthesized MS2 loop binds MCP-GFP ( expressed at low levels and freely diffusing in the embryo ) . In each nucleus , where transcription at this reporter is ongoing , we observe a unique bright fluorescent spot , which corresponds to the accumulation of several MCP-GFP molecules at the locus ( Fig 1C ) . We assume that the fluorescent signal from a labeled mRNA disappears from the recording spot when the RNAP reaches the end of the transgene . With this setup we image the total signal in four fly embryos using confocal microscopy , simultaneously in all nuclei ( Fig 1D ) from the beginning of cell cycle ( cc ) 11 to the end of cell cycle 13 . In each nucleus we obtain a signal that corresponds to the temporal dependence of the fluorescence intensity of the transcriptional process , which we refer to as the time trace of each spot . A cartoon representation of such a trace resulting from the polymerase activity ( Fig 1E ) dictated by the promoter dynamics ( Fig 1B ) is shown in Fig 1F . We present examples of the traces analyzed in this paper in S1 Fig and the signal preprocessing steps in the Materials and Methods and SI Section A . To characterize the dynamics of the hunchback promoter we need to describe its switching rates between ON states , when the gene is transcribed by the polymerase at an enhanced rate and the OFF states when the gene is effectively silent with only a small basal transcriptional activity ( Fig 1A and 1B ) . Estimating the ON and OFF rates directly from the traces is problematic due to the buffering time and to the high background fluorescence levels coming from the unbound MCP-GFP proteins that make it difficult to distinguish real OFF events from noise . To overcome this problem , we consider the autocorrelation function of the signal . To avoid biases from different signal strengths from each nucleus , we first subtract the mean of the fluorescence in each nucleus , F ( ti ) − 〈F ( ti ) 〉 and then calculate the steady state connected autocorrelation function of the fluorescence signal ( equivalent to a normalized auto-covariance ) , C ( τ ) , at two time points separated by a delay time τ , F ( ti ) and F ( ti + τ ) , normalized by the variance of the signal over the traces , according to Eqs 12 and 13 in Materials and Methods . We limit our analysis to the constant expression part of the interphase ( which we call “steady state”—we discuss this assumption at the end of the Simulated data Results section ) by taking a window in the middle of the trace to avoid the initial activation and final deactivation of the gene between the cell cycles ( see Materials and Methods and S2 Fig ) . We will always work with the connected autocorrelation function , which indicates that the mean of the signal is subtracted from the trace . The autocorrelation function is a powerful approach since it averages out all temporally uncorrelated noise , such as camera shot noise or the instantaneous fluctuations of the fluorescent probe concentrations . Fig 2A compares the normalized connected autocorrelation functions calculated for the steady state expression in the anterior of the embryo ( excluding the initial activation and final deactivation times after and before mitosis ) in cell cycles 12 and 13 of varying durations: ∼ 3 and ∼ 6 minutes . Fig 2B shows the same functions for traces that have been curtailed to all have equal length . The steady state signal from cell cycle 11 did not have enough time points to gather sufficient statistics to calculate the autocorrelation function . As expected , the functions decay showing a characteristic correlation time , then reach a valley at negative values before increasing again . Since the number of data points separated by large intervals is small the uncertainty increases with τ . Autocorrelation functions calculated for very long time traces have neither the negative valley nor the increase at large τ . For example , the long-time connected autocorrelation functions calculated from the simulated traces ( Fig 2C ) of the process described in Fig 1 that are shown in Fig 2D , differ from the short time connected autocorrelation function in Fig 2E calculated from the same trace ( see SI Section G for a description of the simulations ) . As the traces get longer the connected autocorrelation function approaches the longtime results ( S4 Fig ) . The connected autocorrelation function of a finite duration trace of a simple correlated brownian motion ( an Ornstein-Uhlenbeck process ) displays the same properties ( see S5 Fig ) . The dip is thus an artifact of the finite size of the trace . We also see that the autocorrelation functions shift to the left for short cell cycles ( Fig 2A ) , resulting , for earlier cell cycles , in shorter directly read-off correlation times , defined as the value of τ at which the autocorrelation function decays by e . However , calculating the autocorrelation functions for time traces of equal lengths for all cell cycles ( Fig 2B ) shows that the shift was also a bias of the finite trace lengths , and after taking it into account , the transcription process in all the cell cycles has the same dynamics ( although we note that the dynamics directly read out from this truncated trace is not the true long time dynamics ) . This preliminary analysis shows that to extract information about the dynamics of transcription initiation we will need to account for the finite time traces . Additionally , a direct readout of even effective rates from the correlation time is difficult , because the autocorrelation coming from the underlying gene regulatory signal ( Fig 1B ) is obscured by the autocorrelation due to the timescale needed for the transcription of the sequence containing the MS2 cassette ( Fig 1E ) —the gene buffering time , τbuff . The observed time traces are a convolution of these inputs ( Fig 1F ) . The analysis is thus limited by the buffering time of the signal ( τbuff = 72s in our system ) , given as the length of the transcribed genomic sequence that carries the fluorescing MS2 loops divided by the polymerase velocity . A direct readout of the switching rates is only possible if the autocorrelation time of the promoter is larger than the buffering time . The form of the autocorrelation function and our ability to distinguish signal from noise also depends on the precise positioning and length of the fluorescent gene [19] . A construct with the MS2 transgene placed at the 3’ end of the gene ( Fig 3A ) gives a differentiable readout of the promoter activity even for two sets of fast switching rates between the active and inactive states . However , in this case the weak signal is hard to distinguish from background fluorescence levels . Conversely , a 5’ positioning of the transgene ( Fig 3B ) is insensitive to background fluorescence . However it only differentiates autocorrelation functions calculated from very slow switching processes [19] . In summary , a construct with the MS2 placed at the 3’ end of the gene allows for a direct readout of the transcriptional kinetics in a much wider range of switching rates than a 5’ construct , although the autocorrelation function of a 3’ construct is more sensitive to background fluorescence . The promoter activity we are interested in inferring can in principle be described by models of varying complexity ( see Fig 1A ) . We consider and compare three types of models in this paper . We note this is a small subset of possible models . In particular , we do not consider models with multiple levels of transcription as was considered in [21] or reversible promoter cycles . In the simplest case , the gene is consecutively yet noisily expressed . The RNAP starts transcribing following a Poisson distribution of discrete ON-activation ( or firing ) events—this has previously been called a static promoter ( not represented in Fig 1A ) . After the polymerase binds , the next polymerase cannot bind before the promoter is cleared ( a timescale estimated to be τblock ∼ 6s in our experiments ) . The effective firing rate of this model is the Poisson rate , r , shifted by a deterministic τblock ∼ 6s , reff = ( τblock + r−1 ) −1 , and we call this discrete time model a Poisson-like promoter . Although the promoter dynamics would be uncorrelated in this case , the gene buffering would still produce a finite correlation time ( see SI Section F ) . Alternatively , the promoter could have two well defined expression states: an ON state during which the polymerase is transcribing at an enhanced level and an OFF state when it transcribes at a basal level . This situation can be modeled by stochastic switching between the two states with rates kon and koff ( left panel in Fig 1A and Materials and Methods ) . However , as was previously observed in both eukaryotic and prokaryotic cell cultures [11 , 12 , 14 , 15] , once the gene is switched off the system may have to progress through a series of OFF states before the gene can be reactivated . Recently these kinds of cycle models have been discussed for the hunchback promoter [22] . The intermediate states can correspond to , for example , the assembly of the transcription initiation complex , opening of the chromatin or transcription factor cooperativity . These kinds of situations can either be modeled by a promoter cycle ( middle panel in Fig 1A and Materials and Methods ) , with a number of consecutive OFF states , or by an effective two state model that accounts for the resulting non-exponential , but gamma function distribution of waiting times in the OFF state ( right panel in Fig 1A and Materials and Methods ) . The time the polymerase spends transcribing the DNA does not dependent on the promoter model . In both the two-state and promoter cycle model the gene switches from the ON to the OFF state with exponentially distributed waiting times described by a rate koff ( Fig 1A ) . In the two-state model the jumps from the OFF to the ON state are also exponentially distributed with a switching rate kon ( Fig 1A ) . In the three state cycle model considered in this paper , an inactive gene can be in two different OFF states . The gene leaves these states with different switching rates , k1 and k2 , respectively . The ordering of k1 and k2 is impossible to detect in the current experiment . In the three state cycle model we can define an effective on-switching rate k on eff = ( 1 / k 1 + 1 / k 2 ) - 1 . k on eff corresponds to the inverse of the average waiting time in the overall OFF state , and the waiting times for exiting this effective OFF state are not exponentially distributed . The gamma function distributed switching time is an approximation of this effective rate . We present our method for all of these models and consider all but the gamma function distributed switching time model to learn about the dynamics of hunchback promoter dynamics . To infer the transcription dynamics from the data we built a mathematical model that calculates the autocorrelation functions accounting for the experimental details of the probes , incorporating the MS2 loops at various positions along the gene and correcting for the finite length of the time traces . The basic idea behind our approach is that while the initiation of transcription is stochastic and involves switching between the ON and possibly a number of OFF states ( X ( t ) in Fig 1B denotes the binary gene expression state ) , if we assume a constant elongation velocity the obscuring of the signal by the probe design is completely deterministic [18 , 23] , which results in the random variable a ( i , t ) ∈ {0 , 1} that describes the presence or absence of the polymerase at position i at time t ( Fig 1E ) . We count the progression of the polymerase in discrete time steps , where one time step corresponds to the time it takes the polymerase to cover a distance of 150 base pairs equal to its own length ( Fig 4A ) . The promoter dynamics can thus be learned from the noisy autocorrelation function of the fluorescence intensity normalized by the intensity coming from one MS2 loop , F ( t ) = ∑ i = 1 r L i a ( i , t ) ( Fig 1F ) , even for switching timescales smaller than the fluorescent probe buffering time τbuff , provided the parameters of the probe design encoded in the loop function Li ( positioning of the probe etc . ) are known ( Fig 1C ) and the intensity signal is calibrated knowing the fluorescence intensity coming from one MS2 loop [18] . Broadly , our model assumes that once the promoter is in an ON state the polymerase binds and deterministically travels along the gene producing MS2 loops containing mRNA that immediately bind MCP and result in a strong localized fluorescence ( Fig 4 ) . The presence or absence of a polymerase at position i at time t , a ( i , t ) is simply a delayed readout of the promoter state at time t − i , a ( i , t ) = X ( t − i ) where t is measured in polymerase time steps ( Fig 1B ) . We assume that polymerase is abundant and that at every time step a new polymerase starts transcribing , provided the gene is in the ON state ( Fig 1B and 1E ) . The amount of fluorescence produced by the gene at one time point is determined by the number of polymerases on the gene ( Fig 4A ) . The amount of fluorescence from one polymerase that is at position i on the gene depends on the cumulated number of loops that the polymerase has produced Li , where 1 ≤ i ≤ r . r corresponds to the maximum number of polymerases that can transcribe the gene at a given time and Li = 1 corresponds to one loop fluorescing , as depicted in the cartoon in Fig 4B . The known loop function Li depends on the build and the position of the MS2 cassette on the gene , it is input to the model and does not necessarily take an integer value since the polymerase length and the loop length do not coincide ( Fig 4B ) . Given the average fraction of time the transcription initiation site is occupied by the polymerase , Pon , the average fluorescence in the steady state is: 〈 F 〉 = P on ∑ i = 1 r L i . ( 1 ) Since we assume the polymerase moves deterministically along the gene , seeing a fluorescence signal both at time t and position i , and at time s and position j means the gene was ON at time t − i and s − j , which is determined by how many loops ( i and j ) the polymerase has produced . Taking the earlier of these times , we need to calculate the probability that the gene is also ON at the later time . The autocorrelation function of the fluorescence can thus be written as: 〈 F ( t ) F ( s ) 〉 = ∑ i = 1 r ∑ j = 1 r L i L j P gene was ON at time min ( t - i , s - j ) · A ( | t - i - s + j | ) , ( 2 ) where A ( n ) is the probability that the gene is ON at time n given that it was ON at time 0 , and time is expressed in polymerase steps . The precise form of Pon , P ( gene was ON at time min ( t − i , s − j ) ) and A ( |t − i − s + j| ) depends on the type of the promoter switching model . We assume that the polymerase moves at constant speed along the gene and that there is no splicing throughout the transcription process . We give explicit expressions for all the models used in the Materials and Methods section and the Supplementary Information . Knowing the design of the construct ( length of the probe and number of loops that have been transcribed at each position ) and calibrating the signal , we use Eq 1 to directly learn Pon from the data . In the two and multi-state models Pon provides us with the ratio of switching rates and we then use Eq 2 to obtain their particular values ( see Materials and Methods ) . To avoid biases coming from nucleus to nucleus variability , we calculated the normalized connected correlation function defined in Eqs 12 and 13 in Materials and Methods . The theoretically calculated connected autocorrelation function , Cr ( Eq 14 , which corresponds to the longtime correlation function in Fig 2C and 2D ) , differs from the empirically calculated connected autocorrelation function from the traces , c ( r ) ( Eqs 12 and 13 in Materials and Methods , which correspond to the short time correlation function in Fig 2C and 2E ) , due to finite size effects coming from spurious correlations between the empirical mean and the data points . Since by definition the mean of a connected autocorrelation function is zero ( see Eqs 12 and 13 in Materials and Methods ) , the area under the autocorrelation function must be zero . For short traces this produces the artificial dip discussed in Fig 2 , which for long traces is not visible , as it is equally distributed over long times . To compare our theoretical and empirical correlation functions we explicitly calculate the finite size correction and include this correction in our analysis ( Materials and Methods and SI Section H and I ) . In this paper , we have analyzed data from fly embryos with 3’ promoter constructs only , limiting ourselves to the steady state part of the trace ( see S2 Fig ) . However the method can also be applied to non-steady state systems ( see SI Section C ) and to other constructs , including cross-correlation functions calculated from signals of different colors inserted at different positions along the gene ( see SI Section J ) . We use simulated data to show that prediction and inference are possible for cross-correlation functions of a two-colored signal ( see S6 Fig ) , but that the accuracy of inference is limited by the use of the 5’ probe . To check that the inference method correctly infers the parameters of the model , we first tested the autocorrelation based inference on simulated short-trace data with underlying molecular models with different levels of complexity ( Fig 3C ) for a construct with the MS2 probe placed at the 3’ end of the gene ( Fig 3A ) . In Fig 3D we compare autocorrelation functions for the three state model for constructs with the MS2 loops positioned at the beginning of the transcribed region ( 5’ , Fig 3B ) and at the end of the transcribed region ( 3’ , Fig 3A ) , and the cross-correlation function calculated from a two-colored probe construct ( Fig 3E ) . The analytical model correctly calculates the short trace autocorrelation function and is able to infer the dynamics of promoter switching for all models . It can also be adapted to infer the promoter switching parameters for any intermediate MS2 construct position , given the limitations of each of the constructs discussed above [19] . The autocorrelation function based inference reproduces the underlying parameters of the dynamics with great accuracy not just for switching timescales longer than the gene buffering time , τbuff , that obscures the signal ( Fig 3F ) , but also for smaller timescales that are within an order of magnitude of the gene buffering time . In Fig 3F we show the results of the inference for the 3’ two state model for different values of the ON and OFF rates , kon and koff . For switching timescales much shorter than the gene buffering time , the autocorrelation function coming from the length of the construct dominates the signal and the precision of the inference goes down . For very fast switching rates ( kon + koff > 0 . 12s−1 ) , increasing the length of the traces or the number of nuclei ( red vs blue curve for values of kon + koff larger than 0 . 1s−1 in Fig 3F ) does not help estimate the properties of transcription . In this regime , the inferred value of kon + koff disagrees with the true parameters even when the inference uses long time traces and a large number of nuclei . For intermediate switching rates ( 0 . 07 − 0 . 12s−1 ) , increasing the trace length or increasing the number of nuclei extends the inference range ( black and green dashed lines vs blue solid line in Fig 3F ) , and in all cases increasing the number of nuclei decreases the uncertainty as can be seen from the smaller error bars ( shown only for the red and blue lines for figure clarity ) . Using two colored probes attached at different positions along the gene gives two measurements of transcription and allows for an independent measurement of the speed of the polymerase—one of the parameters of the model that currently must be taken from other experiments . While the estimates of polymerase speed in the fly embryo are reliable [18] , this parameter has been pointed out as a confounding factor in other correlation analyses [24] . The autocorrelation approach also correctly infers the parameters of transcriptional processes when applied to traces that are out of steady state ( see SI Section C ) . However , since the process is no longer translationally invariant more traces are needed to accumulate sufficient statistics . For this reason , in the current analysis of fly embryos we do not analyze the transient dynamics at the beginning and end of each cycle and we restrict ourselves to the middle of the interphase assuming steady state is reached ( see S2 Fig for details ) . We do not know whether the underlying dynamics is completely in steady state . We limit our analysis to a time frame window where the intensity of the fluorescence signal plateaus ( see S2 Fig for an example ) . We can motivate the steady state assumption a posteriori: the inferred switching timescales ( smaller than 50s ) are small enough for the system to relax to steady state within one cell-cycle . However we cannot fully rule out other mechanisms that could keep the system out of steady state ( such as changes in the Bicoid concentration ) . We divided the embryo into the anterior region , defined as the region between 0% and 35% of the egg length ( the position at 50% of the egg length marks the embryo midpoint ) , where hunchback expression is high , and the boundary region , defined as the region between 45% and 55% egg length , where hunchback expression decreases . The mean probability for the gene to be ON during a given cell cycle Pon ( restricted to the times excluding the initial activation and deactivation of the gene , which we will call the steady state regime ) , given by Eq 1 , is consistent between the four embryos in cell cycle 12 and 13 , both in the anterior region and at the boundary ( Fig 5A ) . The probability for the gene to be ON is over three fold higher in the anterior region than in the boundary and does not change with the cell cycle . Pon ∼ 0 . 5 in the anterior indicates that in each nucleus the polymerase spends about half the steady state expression time transcribing the observed gene . At the boundary the gene is transcribed on average during about 10% of the steady state part of the cell cycle . The estimates for Pon in the earlier cell cycles were not reproducible between the four embryos , likely because the time traces were too short to gather sufficient statistics to accurately calculate the maximum and average of the signal . We concentrated on cell cycle 12 and 13 for the remainder of the analysis . At the boundary , neighboring nuclei have dramatically different expression levels of the Hunchback protein . From measurements of the Bicoid gradient , Gregor and collaborators estimated that for two neighboring nuclei to make different readouts , they must be able to distinguish Bicoid concentrations that differ by 10% [27] . Following the Berg and Purcell [28] argument for receptor accuracy , and using measurements of diffusion constants for Bicoid proteins from cell cycle 14 , the authors showed that , based on protein concentrations , the hunchback gene is not able to read-out the differences in the concentrations of Bicoid proteins to the required 10% accuracy in the time that cell cycle 14 lasts . Even considering the revised higher values of Bicoid’s diffusion coefficient measured in a subsequent study [5] , the precision of the Bicoid gradient read-out remains difficult to explain . The authors invoked spatial averaging of Hunchback proteins as a possible mechanism that achieves this precision . Spatial averaging can increase precision , but it can also smear the boundary . Erdmann et al calculated the optimal diffusion constant Hunchback proteins must have for the averaging argument to work [29] and showed it is consistent with experimental observations [5 , 25] . However precision can already be established at the mRNA level and , using measurements on fixed embryos , Little and co-workers found that the relative intrinsic nuclei-to-nuclei variability of the mRNA transcribed from a hunchback locus is ∼ 50% [6] . Measurements of cytoplasmic mRNA reduced this variability to ∼ 10% [6] . Here we go one step further and use our direct measurements of transcription from the hunchback gene to directly estimate the precision with which the hunchback promoter makes a readout of its regulatory environment in a given cell cycle in a given region of the embryo , δPon/Pon . δPon/Pon is the relative error of the probability of the gene to be ON averaged over the steady state part of a cell cycle . Since the total number of mRNA molecules produced in a given cycle is proportional to Pon ( shown in S9A Fig as a function of embryo length ) , the precision at the level of produced mRNA in a given cycle is equal to the precision in the expression of the gene , δmRNA/<mRNA> = δPon/Pon . The accuracy of transcription activation is encoded in the stochasticity of gene activation . In the two state model , the gene randomly switches between two states: active and inactive , making a measurement about the regulatory factors in its environment and indirectly inferring the position of its nucleus . Since no additional information is provided by a measurement that is strongly correlated to the previous one , the cell can only base its positional readout on a series of independent measurements . Two measurements are statistically independent , if they are separated by at least the expectation value of the time τi it takes the system to reset itself: τ i ∼ 1 k on eff + k off , ( 3 ) where in a two state model k on eff = k on . A more detailed estimate obtained by computing the variance of the time spent ON by the gene during the interphase ( see SI Section K ) shows that Eq 3 underestimates the time needed to perform independent measurements . We find that for a two state model the accuracy of the readout of the total mRNA produced is limited by the variability of a two state variable divided by the estimated number of independent measurements within one cell cycle: δ mRNA <mRNA > = 2 τ i ( 1 - P on ) T P on , ( 4 ) where T is the duration of the cell cycle and the factor 2 is a prefactor correction to the naive estimate . Eq 4 is valid in the limit of T >> τi ( the exact result if given in SI Section K ) . Using the rates inferred from the autocorrelation analysis ( Fig 5D ) we see that the precision of the gene readout is much lower at the boundary than in the anterior , does not change with the cell cycle and is reproducible between embryos ( blue points on the ordinate in Fig 7A ) . In the anterior part of the embryo it reaches ∼ 50% , while at the boundary , it is very large , ∼ 150% , even at cell cycle 13 . In the Poisson-like promoter model , to calculate the relative error in the total mRNA produced , the polymerase arrival times are described by an effective firing rate of reff = ( τblock + 1/r ) −1 . Within this model , the fraction of the total time the polymerase cannot bind , because the binding site is occupied is P on P = τ block · r eff ( see SI Section F ) . The total produced mRNA is then proportional to the time the gene is transcribed , and the relative error in the total produced mRNA depends on the relative error of the firing times of this modified Poisson process and the number of independent measurements , nP = T/ ( τblock + 1/r ) ( see SI Section K ) : δ mRNA < mRNA > = τ block ( 1 - P on P ) T . ( 5 ) Using the rates for P on P inferred from the data ( Fig 5A ) , the relative error in the total mRNA produced ( blue points on the ordinate in Fig 7A ) is slightly higher in the boundary region ( ∼ 15% ) than in the anterior of the embryo ( ∼ 10% ) and does not change with the cell cycle . We can compare both of these theoretical estimates with direct estimates of the relative error of the total mRNA produced during a cell cycle , δmRNA/mRNA , from the data . We divide the embryo into anterior and boundary strips , as we did for the inference procedure and calculate the mean and variance of Pon . These empirical estimates of the gene measurement precision agree with the theoretical estimates ( Fig 7A ) for the two state model , but disagree with the predictions of the Poisson-like promoter model . For completeness , we also calculated the relative error for the three state model ( see S9B Fig ) , which shows better agreement than the Poisson-like promoter model but slightly worse than a two state model . We verified that our conclusions about the scale of our empirical estimates are the same for all embryos ( S9C Fig ) and do not depend on the definition of the boundary and anterior regions ( S9D Fig ) . Since the predicted relative error for the Poisson-like promoter model is lower than the relative error calculated directly from the data , we could imagine that the data estimate is more susceptible to additional sources of experimental noise . However the very large disagreement between the Poisson-like promoter prediction and the data at the boundary suggests the Poisson-like promoter model is not an accurate description . This higher experimental variability also cannot be explained by variable expression levels within the regions we are considering: S9D Fig shows that the experimental relative error does not significantly decrease if we take smaller windows and taking the Pon in the boundary region to range from 0 . 35 to 0 . 5 ( Fig 3A ) would translate into a , at best , two fold increase in the relative error predicted from the Poisson-like promoter model , which is not enough to reach the experimentally observed relative error . While we are unable to rule out the Poisson-like promoter model based on the fit to the autocorrelation function , a different statistic—the relative error in the produced mRNA—suggests that the promoter is most likely well described by a two state model , and possibly a three state cycle . To see whether temporal integration of the mRNA produced can increase precision , we compared the empirical estimate of the steady state mRNA production ( red line in Fig 7B ) to the relative error of the total mRNA produced in cell cycle 13 ( blue line in Fig 7B ) and the total mRNA produced from cell cycle 10 to 13 ( green line in Fig 7B ) averaged over embryos . Assuming that the mRNA molecules are equally divided between daughter cells during division , and they are all kept in the cell throughout cell-cycles 10–13 ( which is incorrect but provides a best case estimate ) , then each nuclei has the total mRNA produced in cell cycle 13 , 1/2 of the total mRNA produced by its mother in cell cycle 12 , 1/4 of the mRNA produced by its grandmother in cell cycle etc . While we see about a 1/3 increase in the precision at the boundary from integrating the mRNA produced in different cell cycles , the estimate in the anterior region is not helped by integration over the cell cycles . For completeness of the discussion of the relative errors in the different models , we calculated the relative error assuming the same k off + k on eff for a three state cycle ( koneff=k1−1+k2−1 ) −1 as for a two state model ( k on eff = k on ) for different values of koff and k on eff ( Fig 7C ) . We found that the relative error is always lower for the three state cycle model and the error decreases regardless of the duration of the cell cycle . As expected from Eq 4 , the relative error is decreased by increasing kon and decreasing koff . However the increase in precision from a three state cycle model in the parameter regime we inferred for the two state model in the the fly embryo is relatively modest ( from ∼ 74% for the two state model to ∼ 67% for the three state model ) . Similarly , in Fig 7D we compared the prediction for the relative errors for the Poisson-like promoter model to two state models with the same probability of the gene to be transcribed , Pon , but different switching rates between the two states ( kon and koff ) . Faster switching increases the precision of the two state promoter , since the number of independent measurements increases . The Poisson-like promoter is always more accurate than the two state promoter . Many previous analysis of precision from static images calculated the relative error of the distribution of a binary variable , which in each nucleus was 1 if the nucleus expressed mRNA in the snapshot , and 0 if it did not express [30 , 31] . We analyzed our data using this definition of activity ( see S9E Fig for mean activity as a function of position ) and found that for most embryos the relative error in the anterior drops to zero ( S9F Fig ) , indicating that all nuclei in a given region show the same expression state , but at the boundary the precision is still ∼ 50% , in agreement with previous reports about the total mRNA in the nucleus [6] . This provides additional evidence for the bursty nature of transcription in the anterior of the embryo , in agreement with previous results that showed a relationship between Bicoid concentration and transcriptional burst of downstream genes [32] .
In contrast to previous studies [17 , 18] , including ours , which failed to show evidence for bursty switching of the hunchback promoter , by developing more advanced analysis techniques we show that the promoter has distinct periods of enhanced polymerase transcription followed by identifiable periods of basal polymerase activity . Our conclusions are based on combining a new autocorrelation based analysis approach , applied to live imaging MS2 data to infer switching parameters , with an analysis of the precision of readout of promoter . The data we used in this paper was generated with a modified MS2 cassette [20] ( see the Experimental procedures section in Materials and Methods ) compared to the previously published data [17] . However the difference in our conclusions mainly comes from a detailed analysis of the traces . Quantification of transcription from time dependent fluorescent traces in prokaryotes and mammalian cell cultures has shown that the promoter states cycle through at least three states [11 , 12] . In one of these states the polymerase transcribes at enhanced levels , while in most of the remaining states the transcription machinery gets reassembled or the chromatin remodels . We find that in the living developing fly embryo , the hunchback promoter also cycles through at least two states , although based on the parameter inference alone we cannot conclusively rule out the possibility of a static promoter with a Poisson-like firing rate or of a more complex promoter with more effective states when the gene is inactive . Only a combination of the inferred parameters using the autocorrelation function with another statistic ( the relative error in the produced mRNA ) allows us to favor the two state model ( or more complex models ) over the other considered mode of transcription activation . The main impediment to distinguishing different types of transcriptional models comes from the very short durations of the interphase in the early cell cycles when the hunchback gene is expressed . We showed using simulations that increasing the number of embryonic samples would not help us distinguish between two and three state models , however looking at longer time traces would be informative ( Fig 6 ) . Since cell cycle 14 lasts about 45 minutes , our analysis shows that the steady state part of the interphase provides enough time to gather statistics that can inform us about the detailed nature of the bursts . Unfortunately , other transcription factors , such as the other gap genes regulating hunchback expression in cell cycle 14 , could possibly change the nature of the transcriptional dynamics in a time dependent manner . We showed that the transcriptional dynamics is constant and reproducible in the earlier cell cycles ( 12-13 ) ( Fig 2 ) , so independently of the question of the nature of the bursts , it would be very interesting to see whether and how it changes when the nature of regulation changes . In the parameter regime of relatively fast switching that we inferred from our data , the autocorrelation function for the two state model and the Poisson-like promoter model are very similar . In this parameter regime , the form of the autocorrelation function is governed by the autocorrelation of the fluorescent probe . So while the autocorrelation function approach is able to disentangle the real promoter switching from the buffering of the construct to determine the parameters assuming an underlying model , we cannot conclusively discriminate between these two models , without looking at other statistics . Using simulations ( S7 Fig ) , we showed that for promoters with slower switching characteristics , this discrimination task is possible and the autocorrelation function approach alone can reliably discriminate between different models . In the parameter regime inferred for the hunchback promoter , having longer traces would not be helpful for this discrimination task and we have to look for other statistics ( S10 Fig ) . However using new constructs with MS2 binding sites that have higher binding affinity to MCP and decrease the noise from the binding/unbinding of MCP to the RNA would make it possible to use shorter MS2 cassettes without increasing background fluorescence . These cassettes would decrease the buffering time and extend the parameter regime in which we can distinguish between the Poisson-like and two state promoter models . Alternatively to focussing on longer traces , a construct with two sets of MS2 loops placed at the two ends of the gene that bind different colored probes could be used to learn more about transcription dynamics [33] . We do not have access to data coming from such a promoter , but our analysis approach can be extended to calculate the cross-correlation function between the intensities of the two colored probes . Such cross-correlation analysis has previously been used to study transcription in cell cultures [34] , transcriptional noise [35] and regulation in bacteria [36 , 37] . Our theoretical prediction for such a cross-correlation function agrees with simulation results ( Fig 3C ) . Unfortunately , the cross-correlation function with one set of probes inserted at the 5’ end and the other at the 3’ shares the same problems of a 5’ construct . For fast switching rates , such a cross-correlation function suffers from the large buffering time ( τbuff ∼ 300s in [18] ) drawback of the 5’ design and can only be used for inferring large switching rates [38] ( see S6 Fig ) . Similarly , the cross-correlation function cannot discriminate between a two state and Poisson-like promoter for relatively fast switching . However , it does gives us access into dynamical parameters of transcription such as the speed of polymerase and it is able to characterize whether mRNA transcription is in fact deterministic and identify potential introns . Possibly , cross-correlations from two colored probes both inserted closer to the 3’ end could be optimal designs . Our method requires knowing the design of the experimental system ( number and position of the loops ) , the speed of polymerase as input and calibrating the maximal fluorescence from one gene . While the polymerase speed is an important parameter and erroneous assumption could influence the inference , we have shown that our inference is relatively insensitive to polymerase speeds ( see S11 Fig ) . In the current experiments we do not have an independent calibration of the maximal fluorescence coming from one gene , which could introduce potential errors in our analysis . However the reproducibility of our results suggests that these potential errors are small . We assumed an effective model that describes the transcription state of the whole gene and does not explicitly take into account the individual binding sites . As a result all the parameters we learn are effective and describe the overall change in the expression state of the gene and not the binding and unbinding of Bicoid to the individual binding sites . For concreteness we presented our model assuming a change in the promoter state and constitutive polymerase binding , but our current model does not discriminate between situations where the transcriptional kinetics are driven by polymerase binding and unbinding and promoter kinetics . The presented formalism can be extended to more complex scenarios that describe the kinetics of the individual binding sites and random polymerase arrival times . Since we already have little resolution power to discriminate between these effective models , we chose to interpret the results of only these effective models . The exact contribution of the individual transcription binding sites could be inferred from the activity of promoters with mutated binding sites . Similarly , other more complex models , such as a reversible three state model , or a model with many ON states , have not been ruled out by our current analysis but are possible within the current framework . The time traces we had to analyze are very short and finite size effects are pronounced . Unlike in cell culture studies , where long time traces are available , we could not collect enough ON and OFF time statistics to characterize the promoter dynamics from the waiting time distributions . In this paper we show that simple statistics , the auto- and cross-correlation functions are powerful general tools that can be used in these kinds of challenging circumstances . To reach our final conclusion we had to combine different kinds of statistics , which is also a useful strategy when limited by data . The approach we propose is a general method that can be used for any type of time trace analysis . However it becomes very useful when studying in vivo biological processes , where the biology naturally limits the available statistics . In our case the number of ON and OFF events is naturally limited by the short duration of the cell cycles . Our method explicitly calculates correlation functions for short traces , correcting for the finite size effects , and can be also used without making steady state assumptions about the dynamics ( although this requires collecting sufficient statistics about two time points , which may be hard for short traces ) . With these corrections we see that while an effective two state model of the underlying dynamics of transcription regulation holds in the anterior and boundary regions of the embryo in all of the early cell cycles , the rates are different in the boundary and anterior regions , showing a strong dependence on position dependent factors such as Bicoid or maternal and zygotic Hunchback concentrations . More statistics will make it possible to build more explicit models of Bicoid dependent activation . While our method is able to deconvolute the effects of the fluorescent probe and infer rates below the buffering limit of the probe ( in our case τbuff ∼ 72s , see Fig 3F ) , in all cases , the rates that we can infer from time dependent traces are naturally limited by the timescales at which the polymerase leaves the promoter , which in our case is estimated to be τblock ∼ 6s . If the switching rates are faster than this scale , even a perfect , noiseless and infinitely accurate sampling of the dynamics will not be able to overcome this natural limit . The inferred rates are reproducible between nuclei and embryos and the inter-embryo variability is similar to the inner-embryo variability ( Fig 5A , 5C and 5D ) . The embryo-to-embryo variability can come from Bicoid variability , which is ∼ 10% [27] , so we do not expect the observed expression variability to be less , variability in growth rate and RNAP availability and external environmental factors . Additional sources of noise are experimental noise and most importantly problems with data calibration of what is the maximal level of fluorescence intensity . We used the obtained results to estimate the precision of the transcriptional process from the hunchback promoter . We found that even in the anterior region , the variability in the mRNA produced in steady state by the different nuclei is large , with a relative error of about 50% ( Fig 7A ) . This variability further increases to 150% of the mean mRNA produced at the boundary . These empirical estimates are completely explained for a two state promoter model by theoretical arguments , which treat the gene as an independent measuring device that samples the environment , correcting for the number of independent measurements during a cell cycle . In both cases , the precision at the level of the gene readout is not sufficient to form the precise Hunchback boundary up to half a nuclear width [39] . Even extending our argument to the total mRNA produced in the early cell cycles ( Fig 7B ) does not help . Having an irreversible promoter cycle could increase the theoretical precision , but only slightly in the parameter regime we have inferred and it would not change the quantitative conclusions about low precision backed by the empirical results . A Poisson-like promoter , while not compatible with the observed error rates , does have a significantly smaller error . The construct we used here was limited to the 500 bp of the proximal hunchback promoter , which is known to recapitulate the hunchback endogenous expression observed in Fluorescent In Situ Hybridization ( FISH ) [20] . It is possible that the boundary phenotype is recovered by averaging of mRNAs and proteins produced by the real gene or the transgenes in other nuclei . In the latter case , this would point towards a robust “safety” averaging mechanism that relies on the population . Alternatively , we have to be aware that the sharp boundaries were only detected on fixed samples and that having access to the dynamics of the transcription process likely provides a more accurate view on the process . We calculated and estimated from the data the precision of the gene readout based on the variability of the transcription process between nuclei . We find that the transcriptional process at a given position is quite noisy . Previous estimates of precision were based on data from fixed samples and did not consider the probability of the gene to be ON , but assumed a binary representation where each nuclei is either active or inactive . By analyzing the full dynamic process we show that the gene is bursty and the transcriptional process itself is much more variable . Reducing the information contained in our traces to binary states , we find precise expression in the anterior , but still large variability at the boundary , similarly to previous results from Fluorescent In Situ Hybridization ( FISH ) aiming to detect all mRNAs [6] . Assuming that the precision in determining the position of the nuclei is encoded in the precision of the gene readout , a gene with the dynamics characterized in this paper needs to measure the signal ∼ 200 times longer at the boundary to achieve the observed ∼ 10% precision . A gene in the anterior would need to integrate only ∼ 25 times longer . These results again suggest that the precision in determining the position of the nuclei is not only encoded in the time averaged gene readout , but probably relies either on spatial averaging mechanisms [27 , 29 , 40] or more detailed features of the temporal information encoded in the full trace [32] . In summary , the early developing fly embryo provides a natural system where we can investigate a functional setting the dynamics of transcription in a living organism . In our data analysis we are confronted with the same limitations that natural genes face: an estimate of the environmental conditions must be made in a very short time . Analysis of dynamical traces suggests that transcription is a bursty process with relatively large inter-nuclei variability , suggesting that simply the templated one to one time-averaged readout of the Bicoid gradient is unlikely . Comparing mutant experiments can shed light on exactly how the decision to form the sharp hunchback mRNA and protein boundary is made . | The fly embryo provides a natural laboratory to study the dynamics of transcription and its implications for the developing organism . Using live imaging experiments we investigate the nature of transcription regulation of the hunchback gene—the first to read out the maternal Bicoid gradient . While traditional time trace analysis methods based on OFF time distributions or autocorrelation functions fail for short signals , our tailored autocorrelation function overcomes these limitations revealing bursty dynamics that is reproducible between cell cycles and embryos . The inferred rates result in a lot of variability in the readout of nuclei sensing similar Bicoid concentrations , suggesting additional readout mechanisms than a one-to-one mapping of the input onto the output . | [
"Abstract",
"Introduction",
"Results",
"Discussion"
] | [
"engineering",
"and",
"technology",
"gene",
"regulation",
"signal",
"processing",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"messenger",
"rna",
"dna-binding",
"proteins",
"dna",
"transcription",
"autocorrelation",
"developmental",
"biology",
"mathematics",
"polymerases",
"statistics",
"(mathematics)",
"embryos",
"research",
"and",
"analysis",
"methods",
"embryology",
"proteins",
"mathematical",
"and",
"statistical",
"techniques",
"gene",
"expression",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"statistical",
"methods"
] | 2016 | Precision of Readout at the hunchback Gene: Analyzing Short Transcription Time Traces in Living Fly Embryos |
Synaesthesia is an unusual perceptual experience in which an inducer stimulus triggers a percept in a different domain in addition to its own . To explore the conditions under which synaesthesia evolves , we studied a neuronal network model that represents two recurrently connected neural systems . The interactions in the network evolve according to learning rules that optimize sensory sensitivity . We demonstrate several scenarios , such as sensory deprivation or heightened plasticity , under which synaesthesia can evolve even though the inputs to the two systems are statistically independent and the initial cross-talk interactions are zero . Sensory deprivation is the known causal mechanism for acquired synaesthesia and increased plasticity is implicated in developmental synaesthesia . The model unifies different causes of synaesthesia within a single theoretical framework and repositions synaesthesia not as some quirk of aberrant connectivity , but rather as a functional brain state that can emerge as a consequence of optimising sensory information processing .
A broad distinction made in the synaesthesia literature is between acquired and developmental forms and it is presently unclear whether a single model or mechanism can account for them both . Developmental forms of synaesthesia have no known triggering event . The typical explanation is that genetic differences in these individuals give rise to structural and functional differences in their brains [5] . Genetic differences linked to synaesthesia have been identified and synaesthesia is known to run in families [e . g . 6] . However , the exact synaesthetic associations themselves do not appear to be inherited , despite being stable within individuals . Thus , a mother may perceive ‘A’ as red and her daughter may perceive it as blue [7] . One of the earlier ways of describing synaesthesia is in terms of a breakdown in modularity [8] . In effect , a given brain region ( e . g . that responsible for colour perception ) responds to multiple inputs in synaesthetes but not others ( e . g . responding to sounds or achromatic letters as well as colours ) . The evidence from functional imaging generally supports this idea [4] . Ramachandran and Hubbard [9] suggest that adjacent regions of cortex may be particularly predisposed to pair as synaesthetic inducers and concurrents in developmental synaesthesia . This may explain why combinations such as grapheme-colour synaesthesia are particularly prevalent [10]; i . e . because of anatomical proximity within the visual ventral stream of grapheme recognition and colour perception . Computational models in general have suggested that a high degree of local clustering is an optimal solution for cortico-cortical connectivity [11] . Although some cases of developmental synaesthesia appear to have derived their associations from , say , alphabet books/blocks this is not the norm [12 , 13] . Similarly , most people exposed to coloured alphabets do not develop synaesthesia . Moreover , for some synaesthetes the spoken or written word “red” may even be synaesthetically blue , or some other colour [14] . As such , associative learning does not seem a plausible general mechanism . However , the mapping between inducers and concurrents is not random . Monotonic mappings have been reported in a variety of types of synaesthesia: increasing pitch is associated with increased luminance in auditory-visual synaesthesia [15]; increased weight is associated with decreased luminance in tactile-visual synaesthesia [16]; and increasing numerosity of digits is linked to decreasing saturation and luminance in number-colour synaesthesia [17] . In the case of letter-colour synaesthesia , there appear to be multiple influences: colours depend on the shapes of letters and their frequency in the alphabet [18] . Synaesthesia tends to be unidirectional such that , for example , a sound may trigger a colour but a colour doesn’t trigger a sound . However , there is some evidence that bidirectionality may occur implicitly ( e . g . a colour may speed up detection of a subsequent grapheme ) , and a few cases in which it has been documented to occur explicitly [19] . When bidirectional synaesthesia is present it need not be symmetrical; for instance , a given sound may trigger a red colour , but seeing a red colour triggers a very different sound [20] . With regards to acquired synaesthesia , there is a known triggering event that leads to the onset of synaesthesia . Synaesthesia can be acquired in two different ways—as a result of sensory impairments [e . g . blindness 21] or as a result of taking certain drugs such as LSD ( lysergic acid diethylamide[22] ) . The latter tends to be temporary and occurs quickly ( minutes , hours ) , whereas the former can occur either quickly ( days ) or slowly ( months or years ) and lasts for long or indefinite periods . Superficially , acquired synaesthesia appears to have somewhat different characteristics from developmental forms of synaesthesia . The nature of the inducer tends to be a sensory stimulus: there are no instances on record of acquired grapheme-colour synaesthesia , for example . This faster acting mechanism is consistent with unmasking ( i . e . removal of inhibition ) of pathways that are already established or enhancement of existing excitatory interactions . For instance , after blindfolding for a few days the ‘visual’ cortex responds to inputs from touch and audition [23] . Although this is not strictly synaesthesia , it represents an example of an inducer triggering a concurrent in neurophysiological terms if not in terms of perceptual experience . In addition to changes in inhibition/excitation , there may be slower-acting structural changes [e . g . synaptogenesis along multi-sensory pathways 24] that lead to acquired synaesthesia and explains why the onset of synaesthesia can occur up to a year after sensory loss . There is only one known computational model of synaesthesia [25] . This model is based on a self-organising Kohonen network and was established to account for one very specific type of synaesthesia: a tendency of some people to experience the sequence of numbers in a spatial configuration . The approach taken in the present study is very different in that it aims to offer a general account of the kinds of scenarios in which synaesthesia might evolve from a neural network and is not seeking to model any particular variety of synaesthesia . The basic architecture of the model below contains two sets of units that can be construed as different modalities ( or , rather , features within a modality ) . The two different sets of input neurons connect to additional layers of output neurons ( Fig 1 ) . The neurons in each output layer are connected by recurrent connections and additional recurrent connections connect the two output layers with one another . In order for synaesthesia to evolve in the first place it would require connections to already be in place between the two modalities , although not necessarily functional . This is developmentally plausible [26] . In our model the difference between synaesthetes and non-synaesthetes lies in whether these connections become functional as a result of the learning process . The presence of synaesthesia is thus operationalised as stable non-zero cross-talk connections between modalities 1 and 2 ( or vice versa ) , together with the observation that stimulating one set of inputs activates both modalities ( i . e . 1→1+2 and/or 2→1+2 ) . The evolution of the recurrent connections in the network , both internal and cross-talk , is governed by learning rules that optimise the information representation of the external inputs into the modalities [27 , 28] . More specifically , the quality of the representation is measured by the mutual information [29] between the input to the network and the neuronal output . Here the input corresponds to the total input to both modalities and similarly the output corresponds to the total output of both modalities after reaching steady state . In our context , the mutual information reflects the ability of the network to discriminate between two similar inputs or , in other words , its sensitivity to changes in the external inputs . In the beginning of the learning process , the cross-talk connections are set to near zero . During learning , the network is presented with input samples of certain statistical characteristics . A major question relates to the role of statistical correlations between the inputs to both modalities . If the inputs are statistically correlated , it is not surprising that cross-talk connections will evolve . From a computational point of view the network can take advantage of these correlations and improve the quality of the representation . However , it seems that in most real-world cases no such correlation underlies synaesthesia . Thus , we try to examine the conditions under which synaesthesia develops despite the fact that there are no correlations between the inputs . In our network model , when the inputs to the two modalities are uncorrelated , typically no cross-talk connections evolve . However , as we show in the following sections , under certain conditions they develop and synaesthesia emerges .
We first analyze a network where each modality contains a single input neuron and a single output neuron ( Fig 2 ) . The simplicity of this network model makes it amenable to analytical investigation in addition to computational simulations . The input and output neurons in each modality are connected in a feed-forward manner . The input to each modality is taken to be normally distributed with zero mean , and the two one-dimensional distributions are statistically independent . There are additional recurrent ( cross-talk ) connections between the two output units . Synaesthesia evolves when the cross-talk connections between the two units increase and become functional . In order to determine the evolution of synaesthesia , we first identify the conditions under which zero cross-talk connectivity ( K12 = K21 = 0 ) is a fixed-point of the learning dynamics , and then look for the conditions under which this fixed-point becomes unstable . In other words , the question is what will happen to a small perturbation to the connections . If both connections go back to zero , the no-cross-talk state is a stable state . The interesting case is when this state becomes unstable and the cross-talk connections develop . The information maximization learning rules for the connections K12 and K21 form a set of two coupled nonlinear equations . We linearized these equations around the point K12 = K21 = 0 and explored the discrete time dynamics by analysing the corresponding eigenvalues . The details of the investigation appear in the Supporting Information and the Results are summarised in Fig 3 . We first analysed how the stability of the fixed-point depends on the variances of the two output neurons . These variances are determined by the variances of the Gaussian distributions at the input . The higher the input variance , the higher the output variance , but due to the bounded non-linearity of the output neurons , the output variance is constrained to be between 0 and 0 . 25 ( see Supporting Information ) . Fig 3A shows the phase diagram of the stability as a function of the two output variances . For pairs of variances in the central region , the no-cross-talk state is stable . Outside this region cross-talk connections evolve ( i . e . synaesthesia occurs ) . There are various scenarios in which a network can be driven outside the no-cross-talk region . For instance , consider a situation in which the variance of the second unit is decreased ( shown by the green arrow ) . This situation is analogous to sensory deprivation at the second unit . At the same point the network develops cross-talk connections from the non-deprived unit to the deprived unit , which increase the output variance at the deprived unit . Similarly , cross-talk connections evolve when the variance of the first unit is increased ( shown by the blue arrow ) . This situation is analogous to sensory flooding at the first unit . Fig 3B shows the same phase diagram together with a surface which describes the critical learning rate , ηcritical , as a function of the variances . Above this surface , synaesthesia appears ( although the statistical variances alone give a "normal" state , without synaesthesia ) . This reflects instability of the learning dynamics due to the high plasticity . The interpretation is that people with high synaptic plasticity are more likely to develop synaesthesia . It cannot be seen in the graph ( in order to have a satisfying resolution for the z-axis ) , but when both variances approach 0 . 25 , the critical learning rate approaches infinity . This means that close to these variance values and in the main regime ( of no cross-talk ) , the learning rate must be very large to result in cross-talk , or synaesthesia . Fig 3C and 3D represent two specific examples of end points within this model space . Fig 3C represents the more typical scenario of no cross-talk such that s1 is sensitive to inputs from x1 alone and s2 is sensitive to inputs from x2 alone . Fig 3D represents an example of the state of the model after the evolution of cross-talk under a sensory flooding scenario . In this model , s2 is activated by inputs from both x2 and x1 ( i . e . a case of modality 1→ modality 2 synaesthesia ) . Note also how s2 has become more sensitive to its own inputs; that is , synaesthesia has increased unimodal sensitivity within the concurrent modality ( modality 2 ) . By contrast , the cross-modal inputs from modality 2 to modality 1 are negative ( inhibitory ) ; i . e . the synaesthesia is not bidirectional . We next verified that results of the analytical investigation using numerical simulations of the corresponding network . The input to each modality was random and normally distributed . The range of variances was sampled in a resolution of about 0 . 01 , and the total amount of simulations was 729 ( 27x27 ) . In each simulation the learning process was run with a different pair of variances ( of both units ) . The initial values for the cross-talk connections were randomly chosen in a ring around the origin ( K12 = K21 = 0 ) . In this case , we checked whether the network converged back to the no cross-talk state or diverged . The results ( Fig 4 ) are consistent with the analytical calculations . The "leaking" of stable-points into the theoretical unstable-area and vice versa , and the asymmetry in respect to the major diagonal is the result of insufficient accuracy or not enough learning-steps in the simulation . The simple model reveals a number of scenarios in which cross-talk may emerge between recurrently connected units , receiving different inputs , based on the principle of maximising the overall sensitivity of the network model . Decreased variance of the input is analogous to sensory deprivation , which is the known aetiology in most ( if not all ) cases of acquired synaesthesia . Sensory flooding ( increased variance of one input ) is another possible cause for synaesthesia . Synaesthetes also have better perceptual discrimination within the concurrent modality [30] . Interestingly , it has recently been found that autism , which is linked to sensory flooding , is also co-morbid with synaesthesia [e . g . 31] . Another finding is related to the learning rate . As the analysis shows , there is a critical value above which the network may develop synaesthesia . This prediction is consistent with the established fact that developmental synaesthesia usually occurs at an early age , when the brain is more plastic . It may also be related to the fact that developmental synaesthesia is linked to enhanced memory abilities [32] . The analysis of the simple model shows that the evolution of cross-talk connections occurs in several scenarios; namely sensory deprivation , sensory flooding and high plasticity . However , the existence of cross-talk in itself does not necessarily reflect synesthetic behavior , since synesthesia also requires a systematic mapping of inducers to concurrents . The aim of this section is to extend these findings in a more complex model containing a population of output units in each modality . In this scenario , each unit has the potential to represent a particular feature of the input and , therefore , it enables us to explore how features in one modality are mapped to features in the other modality . For instance , do monotonic mappings between features in different modalities emerge ? Are they entirely idiosyncratic ? Under which conditions do the mappings fluctuate or become stable ? In synaesthesia , the mappings tend to be consistent within an individual . The mappings tend to differ across individuals but are not strictly random: for instance , synaesthetes tend to show monotonic relationships between pitch and luminance [15] . In this model , the input to each modality is two-dimensional characterized by an angle and a distance from the origin ( Fig 5 ) . The angle , φ , represents a one-dimensional perceptual space ( e . g . the pitch of a sound , the luminance of a colour ) and the distance from the origin , r , represents intensity . The magnitudes , r , of the input samples were drawn from a normal distribution ( with standard deviation proportional to the mean ) and the angles were drawn from a uniform distribution ( Fig 5B; blue dots ) . Altogether , there are four input-neurons , and the inputs to the two modalities are uncorrelated ( Fig 5A ) . The network was presented with random inputs and the recurrent synaptic connections were updated according to the gradient-based learning rules . The feed-forward connections were set to be unit vectors with different angles , θi , which spanned all possible angles from 0° to 360° ( Fig 5B; red radial lines ) . Thus , the weighted input to each neuron in the output layer is: r cos ( θi−φ ) . In this sense , the angle θi can be referred to as the preferred angle of the i'th neuron . An external stimulus at a given angle φ elicits a 'hill' of activity around the neuron with the closest preferred angle . Each modality in this model is similar to a visual hypercolumn , the basic functional unit of the primary visual cortex , which contains a representation of all possible orientations . Analysis of the behaviour of a single hypercolumn network model with these properties and the same information maximization approach appears in [28] . Here , we analyse the case of two coupled networks of this type . In the simulations , we explored the effect of the mean input magnitude and of the plasticity ( learning rate ) . In this model , like in the simple network , the cross-talk connections were initially set to near-zero . We assumed for simplicity that the level of plasticity is the same for all recurrent interactions in the network , and therefore used a single learning rate . The network showed various types of behavior depending on the learning rates and input statistics . An example is shown in Fig 6 . In this simulation , the characteristic magnitudes of the inputs were r1 = 0 . 2 and r2 = 2 . This situation is analogous to sensory deprivation of modality 1 . The recurrent interaction matrix has a block structure , where the diagonal blocks ( Fig 6A ) correspond to the interactions within each modality and the off-diagonal blocks ( Fig 6B ) correspond to the cross-talk interactions . The cross-talk interactions are much weaker compared to the interactions within each modality , as evident by the corresponding scale bars . The interactions within each modality are symmetric and they are excitatory for neurons with similar preferred angles and inhibitory for neurons with more distant preferred angles [28] . However , the strength of the interactions is much stronger in modality 1 , the deprived one , reflecting stronger amplification of its direct inputs ( Fig 6C ) . The cross-talk interactions from modality 2 to modality 1 are mainly excitatory , whereas the cross talk interactions from modality 1 to modality 2 are mainly inhibitory ( Fig 6B ) , resembling the behaviour of the simple model ( Fig 3D ) . We also checked the existence of synesthetic behavior by directly stimulating one modality and testing the response of the other . Fig 6D shows the response of modality 1 to stimulation of modality 2 at an angle of 30° . A compact representation of the response is provided by the magnitude and angle of the population vector ( Methods; [28] ) . The magnitude of the population vector of modality 1 in response to stimulation of modality 2 at different angles is finite ( Fig 6E , red ) . In contrast , the magnitude of the population vector of modality 2 in response to stimulation of modality 1 is effectively zero ( Fig 6E , blue ) . The angle of the population vector of modality 1 in response to stimulation of modality 2 shows a clear systematic mapping ( Fig 6F ) . The fact that the mapping is phase-shifted and decreasing is not important since the values are arbitrary , but the fact that there is a monotonic relationship at all is not trivial ( given that no such mapping was present in the input ) Fig 7 summarizes the results from 5 simulations and demonstrates the different scenarios that can lead to the evolution of synaesthesia . The values of the input magnitudes and the level of plasticity appear inside each panel . The first 3 simulations ( Fig 7A–7C ) describe conditions under which no synaesthesia evolved , resulting in population vectors with zero magnitude . The simulation in Fig 7D had the same input statistics as in Fig 7A ( r1 = r2 = 0 . 2 ) , but a slightly higher level of plasticity . The magnitude of the population vectors is finite in both directions , reflecting a bi-directional synaesthesia ( Fig 7D , left panel ) . This is not surprising as there was complete symmetry between the two modalities in terms of the input statistics . Nevertheless , the mapping from modality 1 to modality 2 is monotonic , whereas the mapping in the opposite direction is non-monotonic ( Fig 7D , right panel ) . This reflects some arbitrary symmetry breaking in the evolution of the cross-talk connection pattern . This may have been caused by small differences in the realization of the random inputs to the modalities . Naively , we would expect the network to be symmetrical , since the properties of both modalities are the same . However , this behavior shows that other extrema of the objective function may exist , extrema which do not preserve the symmetry between the modalities . The simulation in Fig 7E serves as another example of how high plasticity can lead to synaesthesia , when comparing it to the simulation in Fig 7B . Again both had the same input statistics but different plasticity levels . It also demonstrates how sensory deprivation can lead to synaesthesia when comparing it to the simulation in Fig 7C . The simulations in Fig 7C and 7E had the same learning rate , but the magnitude of the inputs to modality 1 was reduced in the simulation of Fig 7E , resulting in a clear monotonic mapping ( Fig 7E , right panel ) . The high-dimensional model produces synaesthesia-like behaviour in response to the same kinds of parameter changes identified in the simple model: namely an increase in learning rate ( analogous to high plasticity ) and if one modality becomes more or less sensitive to its direct input relative to the other ( sensory deprivation/flooding ) . This model also enabled us to explore the relationship between the inducer and concurrent . Although there was no correlated input during learning , the relationship between the inducer and concurrent tended to be monotonic , as is found in many naturally occurring forms of synaesthesia . This is not a trivial outcome , and suggests that such mappings are an emergent property of this kind of neural architecture .
For the last twenty years , theories of synaesthesia have been dominated by two general models: disinhibited feedback from multi-sensory regions to uni-sensory regions , and cross-talk theories which have emphasised the presence of atypical ( and direct ) structural connectivity between modalities [33] . Whereas the former explanation has tended to be favoured for explaining acquired synaesthesia , the latter has dominated explanations of developmental synaesthesia . The approach taken in our computational model represents a significant departure from this current status quo , and has generated novel insights . Our model repositions synaesthesia not as some quirk of aberrant connectivity but rather as a functional brain state that emerges , under certain conditions , as a consequence of optimising sensory information processing . In short , this model goes beyond others by offering an account not only of how synaesthesia emerges but also of why synaesthesia emerges . It offers a unifying account of acquired and developmental forms of synaesthesia insofar as it explains how the same outcome can emerge under different conditions within the same model . Acquired synaesthesia is often associated with sensory deprivation due to damage to the sensory organs or pathways . Our model proposes that the same learning process that optimizes information representation naturally causes neurons in the deprived modality to enhance incoming inputs from intact modalities , leading to synaesthesia . To provide some intuition , we note that our model maximizes the output entropy of the network , which depends on two factors: one is the entropy of each single neuron , i . e . how variable the activity of single neurons is , and the other is the correlations among the neurons . Maximizing this entropy favours high single neuron entropy and low correlations among the neurons . The cross-talk connections induce correlations between the two modalities , which in general tend to reduce the output entropy . However , when one modality is deprived of input , it may be beneficial to have cross-talk connections from the intact modality to the deprived modality . The increase in the single neuron entropy due to the cross-talk connections can compensate for the higher correlations and result in a total increase of the output entropy . Loosely speaking , the deprived neurons seek for other neuronal sources of variability and enhance their connections with them . This mechanism , which emerges naturally in our computational framework , can also be useful for modelling the changes in neural representation that take place in other conditions such as phantom-limb [34] . Although functional accounts for acquired synaesthesia have been proposed in the past , no such comparable account has been put forward for developmental synaesthesia . Our model suggests that it arises from instability in the learning process due to high plasticity . It implies that synaesthetes have higher plasticity compared to non-synaesthetes or a relatively prolonged period of high-plasticity during childhood . Later on , as plasticity in the relevant brain areas decreases , the evolved cross-talk connections become stable . In line with this idea , whole-genome studies link some forms of synaesthesia to genes involved in plasticity , which have higher expression during early childhood [35] . Furthermore , developmental synaesthesia does not appear to be linked to sensory impairments and , if anything , is linked to increased perceptual sensitivities ( notably within the concurrent modality ) . For instance , grapheme-colour synaesthetes show enhanced colour discrimination abilities [36] . In the proposed model , the recurrent connections within the concurrent modality amplify both its direct inputs and the ones from the inducer modality . Thus , an association between synaesthesia and increased perceptual sensitivity is an emergent property of the model , at least under certain scenarios , and it is important to explore the extent to which the presence of synaesthesia ( cross-modal sensitivity ) necessarily goes hand-in-hand with changes in intra-modal sensitivity . In terms of the underlying neurobiological mechanisms , the increased amplification by the recurrent interactions in our model is consistent with findings that indicate increased excitability and elevated glutamate concentration in the relevant cortical areas in synaesthetes [37 , 38] . Traditionally , synaesthesia has not been linked to theories of learning and memory because it has been considered to reflect an innate ( in its developmental form ) cross-wiring of the senses . This view has been challenged on several fronts [e . g . 39 , 40] . Firstly , many of the stimuli that induce synaesthesia ( e . g . graphemes ) are themselves learned . Secondly , for some synaesthetes the particular associations have been influenced by childhood coloured letter sets [13] . Moreover , some general cross-modal correspondences ( e . g . between pitch and vertical positions ) thought to reflect innate vestiges of synaesthesia have been shown to occur as statistical regularities in the environment [41] . Finally , synaesthetes ( at least for grapheme-colour synaesthesia ) are known to have better acquisition of new memories , and this may be related to increased plasticity during learning [32] . Future simulations of the model could use partially correlated inputs to the two modalities to model childhood exposure to coloured letter sets ( they are not fully correlated given that most literacy exposure is with achromatic letters ) . It may well be the case that there is an interaction between learning rate ( an innate parameter within the synaesthete brain ) and these partial associations ( in the environment ) , which explains why most people do not go on to develop synaesthesia after exposure to these stimuli . An interesting hypothesis that emerges from this study regards the relationship between synaesthesia and the concept of critical brain dynamics [28 , 42 , 43] . The goal of the learning process in our model is to find the pattern of recurrent interactions that maximizes the sensitivity of the network to changes in its external inputs . In analogy to physical systems , in which the sensitivity ( often termed susceptibility ) to external inputs diverges near a critical point [44] , here , as the network maximizes its sensitivity , it also tends to approach a critical point [28] . This critical point represents the border between normal amplification of external inputs and a regime governed by attractor dynamics . In the context of sensory processing , the super-critical attractor phase can be thought of as hallucinations that reflect the learned pattern of interactions . A useful measure for identifying critical dynamics is the time it takes the recurrent network to reach steady-state . When close to critical points , many dynamical systems display the phenomenon of critical slowing down [28 , 45] . Interestingly , in simulations of the complex model in which synaesthesia evolved , when the learning process approached the optimal pattern of interactions , the dynamics of the recurrent network became substantially slower ( the number of iterations required to process each input sample until reaching steady-state was ~35000–45000 compared to ~1000–4000 in the beginning of the learning process ) . This observation suggests that in the proposed model networks that developed synaesthesia operate closer to a critical point compared to networks that did not develop synaesthesia . The prediction is that there may be evidence of the neural signatures of critical dynamics in synaesthetes [46 , 47] , particularly as synaesthesia is developing . In terms of its similarities to other models , our model resembles the direct cross-talk ( or cross-activation ) models proposed by others [48] primarily to account for developmental forms of synaesthesia . Although the model represents a direct form of cross-talk , it is an open question as to whether the model would produce similar patterns if neurons from modalities 1 and 2 were not directly connected but were themselves both connected via a third pool of neurons that receives no direct input from 1 and 2 . There is some evidence for both direct and indirect types of neural architecture in synaesthesia as assessed via fMRI effective connectivity [49] . The addition of an interconnecting hub area in future modelling attempts would give the model top-down representations that could be adapted to the ( Bayesian ) predictive coding framework . Unlike the present ( bottom-up ) model , the predictive coding approach describes perception as top-down inference that is constrained and altered by sensory signals . A non-computationally explicit account of synaesthesia in terms of predictive coding has been articulated [50] . Moreover , the kinds of learning algorithms employed in our model are compatible with this approach [51] . The gradient-based learning rules used in this study are not local and are thus expected to reflect the long-term evolution of the system rather than mimicking the moment-by-moment dynamics of real neural circuits . In addition , the neurons in the model are described by simplified rate dynamics which do not capture the complex dynamics of real neurons . An important direction for future modelling work would be the examination of more biologically realistic networks that also optimize information representation . The scenarios for the evolution of synaesthesia described in this study are very general and we believe that similar scenarios would appear also in more realistic networks . In summary , these computational models permit new ways of thinking about synaesthesia both in terms of causal mechanisms and in terms of optimising perceptual function . It generates non-trivial outcomes ( e . g . generating monotonic mappings not found in the input characteristics ) and non-trivial predictions ( e . g . relating to learning , unimodal perceptual sensitivity , hallucinatory tendencies ) .
The general architecture of the model is described in Fig 1 . It involves an input layer with N neurons and an output layer with M neurons . We consider here only overcomplete representations , in which M ≥ N . In the simple model M = N = 2 , and in the more complex model N = 4 and M = 142 . The feedforward interactions are described by the M x N matrix W and the recurrent interactions by the M x M matrix K . During the presentation of each input sample , the input components xi are fixed . The output neurons obey the following dynamics τdsidt=−si+g ( ∑j=1Nwijxj+∑k=1Mkiksk ) , i=1 , … , M . where g is some nonlinear squashing function and τ is a characteristic time scale . The steady-state responses are given by si=g ( ∑j=1Nwijxj+∑k=1Mkiksk ) , i=1 , … , M . The representation of the external inputs is evaluated using the mutual information between the input and the steady-state output of the network [52] . The mutual information can be expressed as the difference between the entropy of the output and the conditional entropy of the output given the input . The conditional entropy represents the entropy of the output noise . Because the network response is a deterministic function of the input , the mutual information is functionally only dependent on the entropy of the outputs . As shown in [27] , maximizing the output entropy ( and therefore the mutual information ) is equivalent to minimizing the following objective function: ε=−12〈ln det ( χTχ ) 〉x=−12Tr〈ln ( χTχ ) 〉x , where χij=∂si∂xj is the Jacobian matrix of the transformation and reflects the sensitivity of the output units to changes in the input units . We also refer to this matrix as the susceptibility matrix as it is analogous to the susceptibility of physical systems to external fields . The adaptive parameters of the algorithm are the sets of feedforward and recurrent interactions , Wij and Kij . The learning rules for these parameters are derived from this objective function using the gradient decent method , as shown in [27] . Here we focus only on the recurrent interactions . The gradient descent learning rule for the recurrent interactions is ΔK=−η∂ε∂K=η〈 ( χΓ ) T+ϕTasT〉 , where η is the learning rate , the matrix ϕ is given by ϕ = ( G−1−K ) −1 and satisfies χ = ϕW , the matrix G is defined as Gij = g′i δij , the matrix Γ is defined as Γ = ( χTχ ) −1χTϕ and the components of the vector a are given by ak=[χΓ]kkg″k ( g′k ) 3 . The triangular brackets denote averaging over the input samples . During the learning process , the evolving networks can approach a critical point in their dynamics ( see Discussion ) . In such cases , the objective function becomes very sensitive to changes in the pattern of interactions . In some cases the objective function may even increase rather than decrease . One way to avoid this is to gradually reduce the learning rate to very small magnitudes . However , to minimize the number of free parameters and make the interpretation clearer , we chose to leave the learning rate fixed across the learning process . Rather , we saved the interaction patterns in the course of the learning process and if a substantial increase in the objective function was identified , we simply chose the interaction pattern associated with the minimal value of the objective function , namely the optimal pattern . To estimate the convergence time of the recurrent network and identify critical slowing down , we defined a criterion for stability of the neuronal activities and measured the time it takes the network to satisfy this criterion . A substantial increase in the convergence time suggests that the network operates close to a critical point . We indeed observed such substantial slowdown of the network dynamics , in particular in the simulations that developed synaesthesia when they approached the optimal pattern of interactions . As a consequence , the simulations could be very long ( up to a couple of weeks on a standard PC station ) . | Synaesthesia is a remarkable form of altered perception , where one attribute of a stimulus ( e . g . sound ) leads to the conscious experience of an additional attribute ( often colour ) . Despite being known about for 200 years , there is no commonly agreed upon model for how and why synaesthesia emerges . This study presents a new model of synaesthesia based on computational principles that accounts for the emergence of different types of synaesthesia ( acquired and developmental ) as well as many of its key characteristics . The model describes how two independent neuronal systems can evolve to interact with one another even though their inputs are statistically uncorrelated . Specifically , synaesthesia arises as a result of instability in the learning process that shapes the network , which can be caused by heightened plasticity or due to sensory deprivation of one of the systems . The model unifies different aspects of synaesthesia and generates novel insights and predictions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"learning",
"neural",
"networks",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"cognitive",
"psychology",
"sensory",
"deprivation",
"network",
"analysis",
"thermodynamics",
"computer",
"and",
"information",
"sciences",
"entropy",
"animal",
"cells",
"human",
"learning",
"physics",
"cellular",
"neuroscience",
"psychology",
"cell",
"biology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"sensory",
"perception",
"physical",
"sciences",
"cognitive",
"science"
] | 2016 | The Emergence of Synaesthesia in a Neuronal Network Model via Changes in Perceptual Sensitivity and Plasticity |
Human brain anatomy and function display a combination of modular and hierarchical organization , suggesting the importance of both cohesive structures and variable resolutions in the facilitation of healthy cognitive processes . However , tools to simultaneously probe these features of brain architecture require further development . We propose and apply a set of methods to extract cohesive structures in network representations of brain connectivity using multi-resolution techniques . We employ a combination of soft thresholding , windowed thresholding , and resolution in community detection , that enable us to identify and isolate structures associated with different weights . One such mesoscale structure is bipartivity , which quantifies the extent to which the brain is divided into two partitions with high connectivity between partitions and low connectivity within partitions . A second , complementary mesoscale structure is modularity , which quantifies the extent to which the brain is divided into multiple communities with strong connectivity within each community and weak connectivity between communities . Our methods lead to multi-resolution curves of these network diagnostics over a range of spatial , geometric , and structural scales . For statistical comparison , we contrast our results with those obtained for several benchmark null models . Our work demonstrates that multi-resolution diagnostic curves capture complex organizational profiles in weighted graphs . We apply these methods to the identification of resolution-specific characteristics of healthy weighted graph architecture and altered connectivity profiles in psychiatric disease .
Noninvasive neuroimaging techniques provide quantitative measurements of structural and functional connectivity in the human brain . Functional magnetic resonance imaging ( fMRI ) indirectly resolves time dependent neural activity by measuring the blood-oxygen-level-dependent ( BOLD ) signal while the subject is at rest or performing a cognitive task . Diffusion weighted imaging ( DWI ) techniques use MRI to map the diffusion of water molecules along white matter tracts in the brain , from which anatomical connections between brain regions can be inferred . In each case , measurements can be represented as a weighted network [1] , [2] , [3] , [4] , [5] , [6] , [7] , in which nodes correspond to brain regions , and the weighted connection strength between two nodes can , for example , represent correlated activity ( fMRI ) or fiber density ( DWI ) . The resulting network is complex , and richly structured , with weights that exhibit a broad range of values , reflecting a continuous spectrum from weak to strong connections . The network representation of human brain connectivity facilitates quantitative and statistically stringent investigations of human cognitive function , aging and development , and injury or disease . Target applications of these measurements include disease diagnosis , monitoring of disease progression , and prediction of treatment outcomes [8] , [9] , [10] , [11] , [12] . However , efforts to develop robust methods to reduce these large and complex neuroimaging data sets to statistical diagnostics that differentiate between patient populations have been stymied by the dearth of methods to quantify the statistical significance of apparent group differences in network organization [13] , [14] , [15] , [16] . Network comparisons can be performed in several ways . In one approach , network comparisons are made after applying a threshold to weighted structural and functional connectivity matrices to fix the number of edges at a constant value in all individuals [4] , [13] . Edges with weights passing the threshold are set to a value of while all others are set to a value of ( a process referred to as ‘binarizing’ ) . In some cases results are tested for robustness across multiple thresholds , although this increases the probability of Type I ( false positive ) errors from multiple non-independent comparisons . More generally , this procedure disregards potentially important neurobiological information present in the original edge weights . A second approach involves examination of network geometry in the original weighted matrices without binarizing . However , because the values of weighted metrics can be influenced by both the average weight of the matrix and the distribution of weights , this approach presents peculiar complications for the assessment of group differences [14] , [15] . Critically , neither of these two approaches for network comparison allow for a principled examination of network structure as a function of weight ( strong versus weak connections ) or space ( short versus long connections ) . Disease-related group differences in network architecture that are present only at a particular edge weight range or at a specific spatial resolution can therefore remain hidden . In this paper , we employ several techniques to examine the multi-resolution structure of weighted connectivity matrices: soft thresholding , windowed thresholding , and modularity resolution . A summary of these techniques is presented in Table 1 , and each method is discussed in more detail in the Methods section . We apply these techniques to two previously published data sets: structural networks extracted from diffusion tractography data in five healthy human subjects [17] and functional networks extracted from resting state fMRI data in people with schizophrenia and healthy controls ( N = 58 ) [15] . As benchmark comparisons , we also explore a set of synthetic networks that includes a random Erdős-Rényi network ( ER ) , a ring lattice ( RL ) , a small-world network ( SW ) , and a fractal hierarchical network ( FH ) . While multi-resolution techniques could be usefully applied to a broad range of network diagnostics , here we focus on two complementary mesoscale characteristics that can be used to probe the manner in which groups of brain regions are connected with one another: modularity and bipartivity . Modularity quantifies community structure in a network by identifying groups of brain regions that are more strongly connected to other regions in their group than to regions in other groups [18] , [19] . Communities of different sizes , nested within one another , have been identified in both structural and functional brain networks [20] , [21] , [22] , [23] , [24] and are thought to constrain information processing [25] , [26] , [27] . Bipartivity quantifies bipartite structure in a network by separating brain regions into two groups with sparse connectivity within each group and dense connectivity between the two groups . The dichotomous nature of bipartitivity is particularly interesting to quantify in systems with bilateral symmetry such as the human brain , in which inter- and intra-hemispheric connectivity display differential structural [28] and functional [29] network properties . We describe these techniques , diagnostics , and null models in detail in the Methods section and illustrate their utility in addressing questions in systems and clinical neuroscience in the Results section .
We examine two separate neuroimaging data sets acquired noninvasively from humans . The first is a set of structural networks constructed from diffusion spectrum imaging ( DSI ) data ( taken from 5 subjects; one subject was scanned twice ) and the second is a set of functional networks constructed from resting state functional magnetic resonance imaging ( fMRI ) data . For comparison , we generate 4 types of synthetic network models that range from an Erdős-Rényi random graph to models that include more complex structure , including hierarchy and clustering . Each network is described by an adjacency matrix whose entry describes the weight of the edge connecting nodes and . For the empirical brain networks , the edge weights are determined by neuroimaging measurements . For the synthetic networks , we construct weights to mimic the structural organization of the network , as described below . In this paper we focus on two , complementary network characteristics that can be used to probe the manner in which groups of brain regions are connected with one another: modularity and bipartivity . These methods , however , are more generally applicable , and could be used to evaluate weight dependence of other metrics and data sets as well [38] . We quantify organization of weighted networks across varying ranges of connection weights ( weak to strong ) and connection lengths ( short to long ) , using three complementary approaches: soft thresholding , windowed thresholding , and multi-resolution community detection . Each method employs a control parameter that we vary to generate network diagnostic curves , representing characteristics of the network under study . We refer to these curves as mesoscopic response functions ( MRF ) of the network [34] . A summary of the methods and control parameters is contained in Table 1 .
In this first section , we examine Problem 1: Probing Drivers of Weighted Modularity . We begin our analysis with measurements that illustrate the sensitivity of mesoscale network diagnostics to the edge weight organization and distribution . We employ the weighted modularity to characterize the community structure of the network and ask whether the value of this diagnostic is primarily driven by the organization of strong edges or by the organization of weak edges . To answer this question , we will utilize two methods: network rewiring and soft thresholding . In this section , we examine Problem 2: Determining Network Differences in Multiresolution Structure . Determining differences in multiresolution network structure requires a set of techniques that quantify and summarize this structure in mesoscopic response functions of network diagnostics . Windowed thresholding is a unique candidate technique in that it resolves network structure associated with sets of edges with different weights . The technique decomposes a weighted adjacency matrix into a family of graphs . Each graph in this family shares a window size corresponding to the percentage of edges in the original network retained in the graph . A family of graphs is therefore characterized by a control parameter corresponding to the mean weight of edges within the window . As we illustrate below in the context of network modularity and bipartivity , this control parameter can be optimized to uncover differences in multiresolution structure of networks . In this section , we examine Problem 4: Identifying Physical Correlates of Multiresolution Structure . Windowed thresholding enables us to separately probe the organization of edges according to specific properties that define their weights ( e . g . , weak versus strong , long versus short , etc . ) . This prevents strong edge weights from dominating the measurements . However , an important limitation of this method is that network diagnostics are restricted to a particular class of nodes within each window , and thus this method potentially misses important structure associated with the topology connecting different geometrical scales . Here we illustrate methods to combine these sources of information , and demonstrate that small communities in human anatomical brain networks tend to be geographically localized while large communities tend to be geographically distributed . To probe community structure at different geometrical scales , we employ a complementary approach . By tuning the resolution parameter in the optimization of the modularity quality function ( Equations 1 and 2 ) , we can identify partitions of the network into both many small ( high ) and a few large ( low ) communities ( see ) . Compared to our other control parameters , the resolution parameter tunes the output of a diagnostic ( e . g . the number of communities ) , rather than acting directly in the edge weights themselves . Varying allows us to probe several features of the network . First , we can uncover the fragmentation profile of a network , as illustrated in Fig . 6A . For example , in brain DSI networks , the number of non-singleton communities peaks at approximately , after which the network fragments into isolated nodes . Second , we can probe the relationship between community structure and physical network embedding . We observe a polynomial relationship between community radius and the resolution parameter ( see Fig . 6B ) and by extension the number of nodes in the community ( Fig . 6C ) , highlighting the interdependence of geometrical and spatial structure in the brain . Small communities tend to be geographically localized while large communities tend to be geographically distributed , suggestive of efficient embedding [54] . Modular networks that are not efficiently embedded into physical space would demonstrate no such relationship . The interaction between space and topology could enhance the organization of information transmission and computing: smaller information processing tasks could be completed by local circuits while larger tasks might make use of more extensive connectivity patterns . In this section , we examine Problem 5: Demonstrating Clinical Relevance . A primary goal of the analysis of human brain networks is to identify changes in network architecture that relate to neurodegenerative diseases or mental disorders . In this section , we investigate the potential applications of multi-resolution techniques to group comparisons by comparing the MRFs for functional brain networks extracted from healthy controls to those extracted from people with schizophrenia . We demonstrate surprisingly that group differences can be located in graphs composed of weak edges . Functional brain network architecture at rest [15] , [55] , [56] and during task performance [8] , [57] , [58] is known to be altered in schizophrenia [9] , [59] , supporting the hypothesis that schizophrenia stems from large-scale brain dysconnectivity [60] , [61] , [62] . Indeed , we observe that the weighted modularity ( obtained at the default resolution of ) is significantly higher in people with schizophrenia than it is in healthy controls ( nonparametric permutation test: ; see Fig . 7A ) . MRFs can be used to probe these dysconnectivies in novel ways . By varying , we demonstrate that this group difference is evident over a range of resolutions , corresponding to partitions with both somewhat smaller and somewhat larger communities ( nonparametric permutation test of group differences between these curves: ; see Fig . 7B ) . The MRF also shows that the community structure over very small and very large communities is not different between the two groups , indicating that signatures of dysconnectivity can be constrained to specific resolutions . To determine whether the patterns of relatively weak or relatively strong edges are most relevant to disease-related alterations in mesoscale brain network architecture , we use the windowed thresholding technique ( see Fig . 7C ) . In both groups , graphs composed of strong edges display higher modularity than graphs composed of weak edges . However , group differences are predominantly located in graphs composed of either the strongest ( high modularity ) or weakest ( low modularity ) edges . No group difference is evident in the bulk of graphs constructed from edges with medium weights , which display modularity values similar to those of the Erdös-Rényi model at the same connection density . The surprising utility of weak connections in uncovering dysconnectivity signatures has been noted previously in the context of schizophrenia [15] and could potentially be of use in the study of other neuropsychiatric disorders and brain injury . The organization of weak and strong edges is further elucidated by the MRFs for bipartivity as a function of the connection weight ( see Fig . 7D ) . In both groups , bipartivity values are smallest for edges with large weights . This finding is consistent with those presented in Fig . 7C , indicating that strong edges tend to be well localized within functional communities . In general , we observe significantly larger values of bipartivity in people with schizophrenia than in controls ( nonparametric permutation test: ) . This difference is spread approximately evenly across the entire range of connection weights , in contrast to the group differences in modularity which were localized to a specific resolution range .
Mesoscale structures—including modularity and bipartivity—display organization that is dependent on the weight of connections included in the network . The strongest connections in high resolution structural brain networks display stronger community structure , more lateralization of communities to the two hemispheres , and less bipartivity than the set of weakest connections . However , even the weakest connections display a modularity that is greater than expected in a random graph , suggesting the presence of nontrivial structure that might provide important insights into brain organization and function . These results are interesting in light of the fact that such connections have previously been thought to be driven purely by noise and are often removed from the network for statistical reasons [65] . Our results that weak connections retain structure are consistent with recent evidence demonstrating the potential utility of studying the topology ( binary ) and geometry ( weighted ) of weak connections for diagnostic purposes [15] . Non-random structure in weak connections could stem from multiple factors – some more biologically interesting than others . First , experimental noise that is preferentially located in particular brain regions ( e . g . , fronto-temporal susceptibility artifacts ) could lead to non-homogeneous network structure . Second , weak connections could be driven by different neurophysiological mechanisms than those driving strong connections ( e . g . , phase-lagged interactions would be measured as weak connections in correlation-based functional networks ) . In this case , the mesoscopic response functions would map out a transition between one mechanism for weak connections and another for strong connections . This latter possibility is potentially interesting from a clinical perspective because it could help to disambiguate the role of multiple mechanisms that could drive the altered connectivity patterns evident in many disease states [8] , [66] , [67] . In this work , we have simply noted the presence of non-random structure in weak connections but we cannot disambiguate the role of these factors . In addition to weight dependence , mesoscale structures also display organization that is dependent on the length of connections included in the network . Networks composed of relatively long connections show weaker community structure than networks composed of relatively short connections . This distributed nature of long connections is consistent with their hypothetical role in connecting disparate functional modules [68] . Furthermore , our results show that communities composed of long connections are more likely to span both hemispheres and display high bipartivity while those composed of shorter connections tend to be more lateralized and display weaker bipartivity , suggesting their role in both inter-hemispheric and inter-lobe communication . To complement these analyses , we investigated the relationship between community structure and connection distance by employing the structural resolution parameter in the optimization of the modularity quality function . Our results demonstrate that modules composed of few nodes ( i . e . , community structure at fine-scale structural resolutions ) have small spatial radii while those composed of more nodes ( i . e . , larger-scale structural resolutions ) have larger spatial radii . Importantly , this mapping between structural and spatial resolutions would not be expected from a network randomly embedded into physical space [54] . Furthermore this relationship between structural resolution and spatial dimension of modules suggests a non-random but rather hierarchical organization within modules , since we find the sub-modules of each module to be have a smaller radius than the super-module . This would not be expected if the sub-structure of a module was randomly organized . These results highlight the relationship between the geometry of a network ( based on edge weights ) and the physical embedding of that network into 3-dimensional space , a relationship which is of interest in a wide variety of complex systems [69] . Such a relationship is consistent with a large body of prior work demonstrating that brain networks extracted from a range of species tend to have near-minimal wiring [5] , [20] , [55] , [70] , [71] , [72] , implicating a density of connections in local geographic neighborhoods and a sparsity of connections bridging those neighborhoods . In contrast to these previous studies that link physical distance to global ( path-length and efficiency ) and local ( clustering coefficient and local efficiency ) network diagnostics , our work uncovers the complementary influence of physical space on meso-scale structures ( modularity and bipartivity ) . The interaction between space and topology in brain systems could be driven by energetic and metabolic constraints on network development [3] , [20] , [54] , [73] . Such constraints might also play a role in the fine-grained spatial geometry of white matter fiber tracts , which cross one another at degree angles [74] , thereby potentially minimizing electromagnetic interference . Moreover , such constraints likely have important implications for system function , where short connections are potentially easier to maintain and use than long connections [75] . If such a functional consequence of physical constraints existed , it might partially explain the functional deficits observed in disease states associated with large-scale disconnectivity [59] , [64] , [75] , [76] . However , converse evidence from normal human development indicates that some distributed processing based on long distance connections is necessary for healthy cognitive function [77] . Future work is necessary to better understand the role of physical constraints on brain development and organization . While community structure in functional and structural brain networks has been examined in a number of studies [22] , [23] , [54] , [78] , other types of mesoscale properties have been less studied . Here we employ two diagnostics – bipartivity and community laterality – that could capture signatures of one of the unusual properties of the brain compared to other complex systems: the physical symmetry between the two hemispheres . Our data suggest that such symmetry is observable in the organization of brain networks . The lateralization of communities is greatest for strong , local connections and smallest for weak , long-range connections , consistent with the preference for small modules to contain nodes in a local geographic neighborhood . Conversely , the bipartivity is greatest for mid-strength and long-range connections respectively and the two parts of the bipartite structure appear to map out both anterior-posterior and left-right axes of brain development . The putative functional role of these network symmetries is at least preliminarily supported by our finding that bipartivity of resting state functional brain networks in people with schizophrenia is significantly higher than that in healthy controls . Such whole-brain network signatures could derive from local asymmetries in white matter microstructure [79] and decreased interhemispheric white matter connectivity previously observed in schizophrenia [80] . In this paper , we have described several complementary techniques for uncovering multi-resolution structure in weighted networks . Each method has advantages and disadvantages and enables one to capture different features present in the network data . Here , for example , we have illustrated these techniques in probing the geometry ( weak versus strong edges ) , embedding ( short versus long edges ) , and structural resolution ( small versus large communities ) of network architecture . However , in some cases one might wish to probe multiple features of the network architecture simultaneously . In Figure 8 , we illustrate the combined use of the structural resolution parameter and windowed thresholding to elucidate the fragmentation profiles of empirical and synthetic network models . In the fractal hierarchical network , we observe that the banding of the number of communities as a function of the mean connection weight is evident for small but not large values of the structural resolution parameter , indicating that the structural organization of the network can be hidden if one studies communities of small size . The remainder of the models and the empirical network data display a relatively constant number of communities over a broad range of mean connection weights for a given value of the structural resolution parameter . This consistency suggests , not surprisingly , that the number of communities is driven more by than by the underlying network topology , which for each of these networks ( with the exception of the Erdös-Rényi model ) varies over different mean connection weights . The band structure in the ER model is similar to the band structure in the SW model for weak edges ( less than approximately ) because both sets of edges are randomly distributed: for the ER model , all connections are randomly distributed , and for the SW model , long-range connections ( which are the weak edges in our weighting scheme ) are randomly distributed . Therefore , both networks display the same geometry for weak edge weights but differ in the geometry of their strong edge weights . Unlike both the SW and ER models , the brain DSI network shows a decreasing number of communities as for stronger edge weights over a range of the resolution parameter . More similar to the fractal hierarchical network , this finding suggests that the human brain network displays hierarchical structure , with a few strong and large communities become composed of many more weaker communities [54] . All networks completely fragment ( zero non-singleton communities at high values of the structural resolution parameter ) at similar but not identical values of , making comparisons across topologies for fixed difficult . We plan to investigate the combined use of these multi-resolution methods in more detail in future work . Our work demonstrates several benefits to the employment of multiresolution network analysis techniques . Such techniques enable statistically meaningful assessments of the organization of weighted spatial networks as a function of edge density , length , and location in Euclidean space . The ability to resolve mesoscale properties – like modularity and bipartivity – over spatial and geometric scales facilitates a deeper understanding of network organization than is possible in the examination of any single resolution alone . Moreover , it enables more focused mechanistic hypotheses for altered connectivity profiles in clinical states where network organization might be perturbed in one resolution ( weak or local connections ) more than in another ( strong or long connections ) . | The human brain is a fascinating organ full of exquisite anatomical and functional detail . A striking feature of this detail lies in the presence of small modules nested within one another across hierarchical levels of organization . Here we develop and apply computational analysis tools to probe these features of brain architecture by examining network representations in which brain areas are treated as network nodes and links between areas are treated as network edges . The class of methods that we describe are referred to as “multi-resolution techniques” and enable us to identify and isolate neural structures associated with different edge properties . Our methods lead to multi-resolution curves of these network diagnostics over a range of spatial , geometric , and structural scales . For statistical comparison , we contrast our results with those obtained for several benchmark null models . Our work demonstrates that multi-resolution diagnostic curves capture complex organizational profiles in weighted graphs . We apply these methods to the identification of resolution-specific characteristics of healthy weighted graph architecture and altered connectivity profiles in psychiatric disease . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"systems",
"biology",
"theoretical",
"biology",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"biophysics",
"neuroscience"
] | 2014 | Resolving Anatomical and Functional Structure in Human Brain Organization: Identifying Mesoscale Organization in Weighted Network Representations |
Stochastic fluctuations in gene expression give rise to cell-to-cell variability in protein levels which can potentially cause variability in cellular phenotype . For TRAIL ( TNF-related apoptosis-inducing ligand ) variability manifests itself as dramatic differences in the time between ligand exposure and the sudden activation of the effector caspases that kill cells . However , the contribution of individual proteins to phenotypic variability has not been explored in detail . In this paper we use feature-based sensitivity analysis as a means to estimate the impact of variation in key apoptosis regulators on variability in the dynamics of cell death . We use Monte Carlo sampling from measured protein concentration distributions in combination with a previously validated ordinary differential equation model of apoptosis to simulate the dynamics of receptor-mediated apoptosis . We find that variation in the concentrations of some proteins matters much more than variation in others and that precisely which proteins matter depends both on the concentrations of other proteins and on whether correlations in protein levels are taken into account . A prediction from simulation that we confirm experimentally is that variability in fate is sensitive to even small increases in the levels of Bcl-2 . We also show that sensitivity to Bcl-2 levels is itself sensitive to the levels of interacting proteins . The contextual dependency is implicit in the mathematical formulation of sensitivity , but our data show that it is also important for biologically relevant parameter values . Our work provides a conceptual and practical means to study and understand the impact of cell-to-cell variability in protein expression levels on cell fate using deterministic models and sampling from parameter distributions .
Variability in the responses of tumor cells to biological stimuli is often ascribed to genetic differences . However , it has become increasingly clear that even genetically identical cells growing in a homogenous environment respond differently to ligands , drugs , or other stimuli . Non-genetic variability at the single-cell level has been demonstrated in the activation of immune responses [1] , [2] , [3] , [4] , viral infectivity [5] , [6] , [7] , developmental fate [8] , [9] , [10] , [11] , antibiotic resistance [12] , and sensitivity to therapeutic drugs [13] , [14] , [15] . Such variability can arise from relatively long-lasting “epigenetic” changes that have their origins in stable and heritable programs of gene expression [16] and can be sensitive to histone deactylase inhibitors that disrupt the histone code [14] . Substantial phenotypic variability also arises from fluctuation in the levels or activities of proteins ( or other biomolecules ) that control cell fate; the current paper is concerned with this type of variability . Two sources of non-genetic variability can be distinguished . The first , often called “intrinsic noise” , arises when the copy number of molecules participating in a reaction under study is sufficiently small that probabilistic fluctuations in protein-protein interactions or biochemical reactions have observable effects [17] . Such processes are modeled using stochastic methods . The second source of variation , often called “extrinsic noise , ” arises when protein concentrations in individual cells are high enough that single-cell reaction trajectories are well approximated by mass-action kinetics , but “external” or pre-existing cell-to-cell differences in the activities or concentrations of biomolecules have an effect [17] . With either intrinsic or extrinsic noise , phenotypes vary from one cell to the next but the processes that cause cells to differ are either part of or external to the biological process under study . When clonal cell populations are treated with TNF-related apoptosis inducing ligand ( TRAIL ) , their response is dramatically different from cell to cell: some cells die with 45 min , some die after as long as 12 hr , and some do not die at all [15] , [18] . We have investigated the contributions of intrinsic and extrinsic noise to this variability by studying sister cells [15] . Were cell-to-cell variability to arise predominantly from intrinsic noise , we would expect sister cells to be no more correlated phenotypically than two cells selected at random from a population: intrinsic noise cannot be inherited . However , time-lapse microscopy has shown that the time and probability of TRAIL-induced cell death are highly correlated in newly born sister-cell pairs . The correlation in time of death between sister cells decays on a time scale of hours to days so that older sister cells are ultimately no more similar to each other than are pairs of cells selected at random from the population . Were cell-to-cell variability in phenotype to arise from differences in protein levels or activities , we would expect phenotypes to be transiently heritable ( as observed with TRAIL ) because binomial partitioning of cellular contents at division causes sisters to inherit similar numbers of high abundance biomolecules [19] , [20] . Subsequent decorrelation in protein levels , and thus in time and probability of death , is also expected because fluctuations in protein synthesis and degradation ( processes that exhibit significant intrinsic noise [6] ) have an increasing impact as time progresses . The time required for sister cells to diverge and recapitulate the steady state distribution is known as the “remixing time” . Factors that determine remixing times are not fully understood [8] , [9] , [21] but translation rates are one contributor . Because cell-to-cell variability in responses to TRAIL can be ascribed primarily to differences in protein concentrations existing at the time of ligand addition and not to intrinsic noise in signal transduction reactions , deterministic mass-action modeling is appropriate [15] . Indeed , attempts to reproduce observed variability in cell death dynamics using conventional stochastic simulations have not succeeded , probably because proteins that regulate apoptosis are abundant [22] . TRAIL-mediated apoptosis involves binding of TRAIL ligand to transmembrane DR4/5 receptors and consequent activation of effector caspases . To simulate these processes we have developed a series of mass-action models based on networks of ordinary differential equations ( ODEs; referred to as extrinsic apoptosis reaction models , or EARMs ) that have been validated in single-cell studies using small molecule drugs , pathway-wide RNAi , and protein overexpression [18] . EARM describes the dynamics of death in single cells with good accuracy , particularly when cells are exposed to low-dose cycloheximide that blocks de novo protein synthesis ( from the perspective of modeling , use of cycloheximide obviates the need to model TRAIL-induced transcription and translation and reduces the number of model parameters ) . Upon TRAIL stimulation , death-inducing signaling complexes ( DISCs ) assemble on the cytoplasmic tails of TRAIL-bound DR4/DR5 receptors , activating initiator pro-caspases-8 and -10 ( hereafter referred to as caspase-8 or C8 for simplicity , Figure 1 ) . Active caspase-8 directly cleaves effector pro-caspases-3 and -7 ( hereafter simplified to caspase-3 or C3 ) but in most cell types , including those studied here , caspase-3 activity is held in check by XIAP until mitochondrial outer membrane permeabilization ( MOMP ) takes place . MOMP is controlled by members of the Bcl-2-family of proteins , which includes both positive and negative regulators . Active caspase-8 cleaves Bid into tBid which then induces a conformational change in Bax . Active Bax translocates to the mitochondria , where it ( or its homolog Bak ) multimerizes and form transmembrane pores . Pore assembly is antagonized by anti-apoptotic Bcl-2 proteins present in the cytosol and outer mitochondrial membrane . Only when levels of active Bax/Bak exceed those of inhibitory Bcl-2 proteins does pore formation begin and MOMP take place , releasing cytochrome c and Smac into the cytosol in a sudden , all-or-none process . Cytochrome c forms an apoptosome complex that also contains Apaf-1 and activates caspase-9 , thereby creating an additional factor capable of processing pro-caspase-3 . Smac binds to XIAP , which prevents XIAP from associating with active caspase-3 , freeing caspase-3 so it can cleave substrates such as the inhibitor of caspase-activated DNase ( ICAD ) and poly ( ADP-ribose ) polymerase ( PARP ) , and thereby promote fragmentation of the genome and proteome . The current work aims to evaluate the impact of changes in the concentrations of apoptosis regulators on the dynamics of effector caspase activation in cells treated with TRAIL , particularly for changes that arise from natural variation in protein levels from one cell to the next . Because we observe such variation to be constant across a continuously growing cell population ( that is , to be quasi-static ) time-invariant distributions of initial protein concentrations can be used as inputs for ODE modeling . We measured protein concentration distributions in asynchronous cell populations using flow cytometry and microscopy and , when suitable reagents were available we also measured correlations in the concentrations of different proteins . Variability in the dynamics of apoptosis was then simulated by sampling from these distributions . In principle , we expect variation in the levels of some proteins to matter more than variation in others , and we show that this is indeed the case . We explore the impact of correlations in the levels of one or more model species and also ask how many species must be measured to accurately predict phenotype . Finally , we show that the phenotypic consequences of variation in particular proteins are affected by changes in the concentrations other apoptotic regulators , demonstrating contextual dependency in the contributions made by regulatory molecules to the timing and probability of cell death .
Cell death is represented in EARM by caspase-mediated proteolysis of PARP to generate cleaved PARP ( cPARP ) : previous studies have shown that HeLa cells are inviable when cPARP exceeds ∼10% of initial PARP levels ( [cPARP]>0 . 10*[PARP]0 ) [23] . Under normal circumstances , cleavage of effector caspase substrates constitutes a “snap-action” switch that is well described by a sigmoidal Boltzmann trajectory in which an extended delay is followed by sudden and complete PARP proteolysis [18] , [24] . The delay time is long ( 45 min to 12 hr ) , varies from cell to cell , and increases as the dose of TRAIL decreases . In contrast , the rate of PARP cleavage is rapid once begun , and dose-invariant ( the time between the first detectable cleavage of PARP and its completion is typically 20–25 min ) . However , RNAi-mediated depletion of regulatory proteins such as XIAP or treatment of cells with proteasome inhibitors such as MG132 changes the dynamics so that PARP is cleaved more slowly and may not go to completion . We have previously argued that these qualitative changes create a pathological state in which caspase-activated DNase is active , genomic DNA damaged , but some damaged cells do not die [18] , [23] . Such cells have been proposed to play a role in tumor initiation [25] , [26] , [27] , [28] . Thus there is a fundamental difference between natural variation in the timing of apoptosis and the breakdown associated with slow and incomplete execution of the apoptosis program . The impact of changes in the initial protein concentrations or other parameters on model output is determined using sensitivity analysis . For extrinsic apoptosis we can distinguish between normal and pathological behaviors using four features of cPARP and cytosolic Smac trajectories: 1 ) tPARP , the time between ligand exposure and 50% PARP cleavage ( i . e . time of cell death ) , 2 ) tMOMP , the time between ligand exposure and MOMP ( defined experimentally as the first image in which a fluorescent MOMP reporter appears diffuse in the cytoplasm and in the model as the time at which 50% of Smac has translocated to the cytosolic compartment ) , 3 ) tswitch , the time between the start and finish of PARP cleavage ( a measure of the rate of PARP cleavage ) , and 4 ) fPARP , the fraction of PARP cleaved at the end of the simulation or experiment ( a measure of the completeness of apoptosis ) . In unperturbed HeLa cells exposed to 50 ng/ml TRAIL and 2 . 5 µg/ml cycloheximide , tPARP and tMOMP varied from 2–6 hr , tPARP occurred ∼10 min after tMOMP ( at a single cell level ) , tswitch was ∼25 min , and fPARP was ∼1 . 0 [23] , [29] . The sensitivities of these features to changes in parameter values are related to but distinct from conventional sensitivities . Analytical expressions for feature sensitivities are described in Box 1 but the calculations in this paper actually involve numerical methods that account for complex non-local effects ( see Text S2 and Box 2 for further details ) . In the numerical approach , sensitivities are calculated by Monte Carlo sampling of initial protein concentrations thereby repeatedly evaluating the local slope of the response curve for a feature . We evaluated feature sensitivities with respect to 16 non-zero initial protein concentrations individually by simulating PARP cleavage and Smac translocation dynamics in cells exposed to 50 ng/ml TRAIL while sampling uniformly in the exponent over a range of 102–107 proteins per cell ( all other parameters remained at their nominal values ) . This range of concentrations reflects the range of possible concentrations for proteins in a mammalian cell [30] , [31]; for apoptotic regulators it is a reasonable approximation to the concentration range achieved by protein overexpression or RNAi-mediated protein depletion . EARM has been validated using such methods and it performs well under these conditions [15] , [23] , [29] , although most analysis has been performed for protein levels present in unperturbed HeLa cells . The sensitivity of model features to changes in protein levels broadly conformed to expectation: for tMOMP , higher levels of pro-apoptotic proteins lying upstream of pore formation decreased its value ( Receptor , caspase-8 , Bid , and Bax; Figure 3A , green curves show responses , blue curves show sensitivities ) whereas higher levels of upstream anti-apoptotic proteins ( FLIP , Bar , Mcl-1 , Bcl-2; Figure 3A ) increased its value . In contrast , changes in the levels of downstream proteins had little effect ( e . g . caspase-3 , caspase-9 , Apaf1; Figure 3A , bottom row ) . The sensitivities of other features are shown in Figures S2 , S3 , S4 in Text S1 . They reveal that different features exhibit different sensitivity with respect to protein initial concentrations . These sensitivities varied considerably in magnitude with position in parameter space: over some concentration ranges , small changes in the levels of proteins such as FLIP , Mcl-1 or Bcl-2 had a large impact on model output but over other ranges the impact of small changes was minimal ( Figure 3A ) . For example , when [Bcl-2] lay between 104 and 105 molecules per cell , tMOMP changed rapidly whereas with [Bcl-2] between 102 and 104 molecules per cell , tMOMP changed very little . Because sensitivity is a local property of a model , this result is expected from a mathematical perspective but is often overlooked from a biological perspective . To estimate mean concentrations of apoptotic regulators in HeLa cells , we used calibrated immunoblotting and recombinant protein standards ( Figure S5 in Text S1 ) ; to estimate variation in protein levels from one cell to the next we used flow cytometry ( [15]; Figures S6 , S7 , S8 in Text S1 ) . The selectivity of antibodies was validated using siRNA-mediated protein knockdown and/or protein over-expression ( Figure S7 in Text S1 ) . Bcl-2 , for which good antibodies are available , was assayed using both immunofluorescence microscopy and flow cytometry and we observed excellent agreement between the two types of measurement ( Figure 3B ) . Both measurements rely on immunodetection and we wanted to exclude the possibility that variability in antibody-antigen binding might have a significant impact on the measured distributions . We therefore transfected cells with a construct expressing GFP-Bcl-2 and quantified the intensities of GFP and of anti-Bcl-2 immunofluorescence in the same cells ( Figure 3C ) . In a two-dimensional scatter plot of these measurements , variation along the diagonal represents real cell-to-cell differences in Bcl-2 concentration whereas off-diagonal variation represents differences between antibody-based and GFP-based estimates of Bcl-2 abundance . Off-diagonal variation can arise from antibody binding , instrument error , the presence of some immature and non-fluorescent GFP-Bcl-2 molecules , or variability in levels of endogenous Bcl-2 . Off-diagonal variation therefore represents an upper-bound estimate of measurement error arising from antibody binding and detection . We observed that off-diagonal variation was significantly smaller than natural cell-to-cell variation in endogenous Bcl-2 levels , suggesting that estimates of variability in Bcl-2 levels do indeed reflect real differences in protein abundance from cell to cell ( Figure 3D ) . Across a set of five proteins for which we could demonstrate antibody selectivity ( only a subset of commercially available antibodies are suitable ) , measured distributions of protein abundance were unimodal and long-tailed as has been observed previously in mammalian cells [21] , [32] . All were well fit by log-normal distributions , although we cannot exclude the possibility that gamma distributions or other long-tailed distributions are also appropriate representations of the data ( Figure S8 in Text S1 ) . The coefficients of variation ( CV; standard deviation divided by mean ) were between 0 . 43 and 0 . 47 but some of this variation is expected to arise from differences in cell size . We therefore selected cells with similar forward and side scatter measurements , which reduced CVs to 0 . 28±0 . 02 to 0 . 30±0 . 03 depending on the protein . Given the relatively narrow range of values for the CV , it seemed reasonable to assume similar variance for those proteins we could not assay experimentally . We therefore set CV = 0 . 25 for proteins whose distributions were assumed rather than measured ( to err on the conservative side ) . In Figure 3A we relate the sensitivity of tMOMP to endogenous protein concentrations in HeLa cells by positioning double vertical bars at the 5th and 95th percentiles of the protein concentration distributions ( see also Figures S1 , S2 , S3 , S4 for other features ) . In general , data and simulations predicted HeLa cells to be more sensitive to increases than to decreases in the levels of anti-apoptotic proteins across the endogenous range; the opposite is true for pro-apoptotic proteins . The endogenous distributions of Bcl-2 and Bax were particularly interesting because they suggested that HeLa cells lie close to a region of parameter space in which small changes are expected to have a large impact on tMOMP , a finding we analyze in greater detail below . To compute the impact of natural variation in protein levels on tMOMP , tPARP , fPARP , and tswitch , we used Monte Carlo methods that sample from log-normal distributions of initial protein concentrations ( based on measured or assumed values for the mean and variance; Figure 4A ) . The predicted distributions of all four features were unimodal regardless of whether we varied each protein concentration individually ( with all others fixed at their nominal values; Table S2 in Text S3 ) or varied all proteins simultaneously ( while sampling independently; Figure 4B–E ) . In no case did natural variation in the concentration of a single protein generate as much variation in tMOMP as simultaneous variation in all proteins ( CV = 0 . 11 for tMOMP when Bax alone was varied as compared to CV = 0 . 24 for tMOMP when all proteins were varied independently ) . Results for tPARP were similar except that variation in XIAP levels ( and to a lesser extent , variation in other proteins downstream of MOMP ) also had an impact ( Figure 5C ) . This is expected since XIAP is a direct negative regulator of caspase-3 , and caspase-3 is the enzyme that cleaves PARP . From these data we conclude that experimentally observed variation in the time of MOMP or PARP cleavage is controlled in a multi-factorial manner and that total phenotypic variation cannot be explained by measured variation in any single protein concentration . A mathematical explanation of this effect is presented in Box 2 . In contrast , variability in fPARP was dominated by variability in the levels of XIAP , and the impact of varying XIAP alone was almost as great as that of varying all proteins simultaneously ( Box 2 , Equation 3 shows that it cannot be greater; Figure 5E ) . The situation was similar for tswitch , except that factors downstream of MOMP also had an effect on this feature . These observations help to explain previous RNAi and over-expression data showing that forced changes in the levels of proteins that impact fPARP also affect tswitch , and that both features are particularly sensitive to changes in the levels or activity of XIAP [18] , [23] , [33] . Indeed , the most potent way to reduce the efficiency of apoptosis experimentally in HeLa cells ( i . e . increase tswitch or reduce fPARP ) and generate “half-dead” cells , appears to be to interfere with the levels or activity of XIAP [23]; the same is true in HCT116 human colon carcinoma cells [34] . Taken together , these results make the point that the “robustness” of a cell to variation in any single parameter is strongly dependent on the feature being evaluated: tMOMP is robust to variation in [XIAP] but tswitch and fPARP are particularly sensitive to it . We find that virtually all of the proteins in the model are determining factors ( sensitive parameters ) for at least one physiologically important variable . Correlation in the levels of regulatory proteins is expected to alter the relationship between variability in protein concentration and variability in phenotype ( Box 2 , Equation 2 ) . Using two-color flow cytometry , we measured correlations in the concentrations of Bax , Bcl-2 , Bid , caspase-3 and XIAP across all ten pairwise combinations ( suitable antibodies pairs were not available for other regulatory proteins ) . Gating on forward and side scatter was used to select for cells of similar size since positive correlation is expected simply based on cell volume . With stringent gating , we observed positive linear correlation coefficients that ranged from R∼0 . 4 for caspase-3 and Bcl-2 to R∼0 . 7 for Bax and Bcl-2 ( Figure 5A–B ) . No negative correlations were observed , consistent with results from bacteria showing that extrinsic noise is expected to cause all protein levels to be positively correlated unless they are specifically counter-regulated [35] . To determine the impact of correlation in protein expression on model features we constructed a five-dimensional joint distribution based on pair-wise measurements and performed Monte Carlo sampling . Protein pairs whose co-variance is unknown were assumed to be uncorrelated . Including the measured correlations in initial protein concentrations reduced the predicted variability in tPARP from a CV = 0 . 23 to CV = 0 . 19 , a statistically significant improvement in the match to experimental data ( for which CV = 0 . 18; Figure 5C ) . We conclude that measured co-variation in protein levels has a significant impact on variability in the timing of death . To investigate the impact of correlations in protein concentrations in cases in which experimental data could not be collected , we performed simulations considering each protein pair and assuming either independent distributions or a single joint pairwise distribution with R = 0 . 7 ( the highest correlation observed experimentally; Figure 5B ) . All other proteins were sampled independently from their respective log-normal distributions . tPARP is the feature whose variance was affected by the greatest number of parameters ( Figure 4B–E ) and we therefore focused on it . For each pair of proteins in the model , we computed the ratio between the variance in tPARP expected under assumptions of independence or positive correlation ( Figure 5D ) . In many cases effects were relatively modest . The largest single difference involved the anti-apoptotic XIAP protein and its pro-apoptotic binding partner Smac whose assumed correlation reduced dispersion in tPARP two-fold . This represents one example of a general phenomenon: positive correlations in the concentrations of pairs of proteins having opposing roles in apoptosis reduced the spread in death times ( Figure 5D , red shading; see Box 2 for further explanation ) . Although correlations in protein levels had a modest impact on variability in tPARP , it significantly altered the phenotypic consequences of variation in individual proteins . This can be seen by sampling from the experimentally determined joint distributions for Bax , Bcl-2 , Bid , caspase-3 and XIAP ( allowing all other proteins to vary independently ) and using the Pearson correlation coefficient ( R ) to score the relationship between tPARP and the initial concentration of each model species ( Figure 5E and Figure S9 in Text S1; similar results were obtained using Spearman's rank correlation coefficients , not shown ) . As expected , the R value for pro-apoptotic proteins was negative and for anti-apoptotic proteins it was positive ( see also Figure S10 in Text S1 for an analysis yielding similar results using the slope of the regression line ) . Unexpectedly , Bcl-2 had virtually no correlation with tPARP ( R = 0 . 002 , Figure 5E ) when concentrations were sampled from joint distributions even though it was significantly correlated when protein levels were assumed to be independent ( R = 0 . 37 ) . To try to explain this , we removed only the correlation between Bcl-2 and Bax from the joint distribution and re-computed R values for tPARP . This restored the impact of variance in Bcl-2 on tPARP , demonstrating the contextual dependency of feature sensitivities ( Figure 5E ) . At the mechanistic level , this result can be explained by the fact that cells that have higher Bcl-2 levels ( which should cause them to die later , on average ) also have more Bax ( causing them to die earlier , on average ) because of positive correlation between Bcl-2 and Bax concentrations . Thus , the correlation between Bax and Bcl-2 dampens the variability in the [Bcl-2]0:[Bax]0 ratio and masks the impact of Bcl-2 on tPARP . More generally , correlations in protein concentrations can substantially alter individual parameter sensitivities and such effects could be strong enough to alter apparent mechanistic relationships between proteins and phenotypes . Conceivably , these effects could also be sufficient to mask the impact of forced changes in protein expression induced by RNAi or overexpression . To determine the practical consequences of multi-factorial control over tPARP , we asked how many measurements are required to accurately predict time of death in a single cell . We performed 105 simulations of PARP cleavage in cells treated with 50 ng/ml TRAIL and low-dose cycloheximide assuming Bax , Bcl-2 , Bid , caspase-3 and XIAP concentrations to co-vary and all other proteins to vary independently . We modeled the process of measuring one to eight proteins at a single-cell level assuming experimental error of ±12 . 5% and then computed how accurately tMOMP could be determined . In actual microscopy experiments tMOMP is typically sampled at 3 min intervals , resulting in a mean squared error ( MSE ) of ∼0 . 03 . When we assumed knowledge only of [Bid]0 , the MSE in predicting tMOMP was 0 . 26 , a poor estimate given experimental error ( reflected in the scatter of points around the trend line in Figure 6A , left ) . Next , we prioritized measurements by ranking them based on their contributions to variation in tPARP ( as judged by CV or R2 values with respect to tPARP as shown in Figures 4C and 5E ) or to variation in tMOMP ( as judged by R2 values with respect to tMOMP , as shown in Figure S10 in Text S1 ) . We observed that ability to predict tMOMP increased progressively ( Figure 6B and Figure S11 in Text S1 ) and that knowledge of the seven or eight most sensitive protein concentrations was necessary to achieve an MSE approaching experimental error ( i . e . 0 . 03; Figure 6A–B ) . In contrast , selecting proteins for measurement at random from the full set of 16 species having non-zero initial concentrations was ineffective in reducing the MSE . We conclude that accurate prediction of time of death from initial protein concentrations requires data on many proteins concentrations even in the best case . Single-cell measurement of 30 or more protein levels is now possible by mass cytometry [36] , but these measurements destroy cells and it is therefore impossible to use the method to link multiplex measurement of protein concentration to events , such as cell death , that occur many hours later . In this sense , identifying the factors that determine the time of death of a single cell is not yet achievable experimentally , even though determining distributions of death times is straightforward . HeLa cells are predicted to be highly sensitive to even modest increases in Bcl-2 concentrations above endogenous levels ( Figure 3 ) . To test this prediction , we expressed variable levels of GFP-Bcl-2 in HeLa cells and then monitored tMOMP using a live-cell reporter ( IMS-RP , [23] ) in cells exposed to 50 ng/ml TRAIL plus low-dose cycloheximide . At wild-type levels of Bcl-2 , all cells died within 5 hr , but as GFP-Bcl-2 levels increased above 4×105 molecules/cell ( ∼13-fold above wild-type ) , a sudden transition was observed in tMOMP such that cell death was blocked indefinitely ( Figure 7A ) . Between 2×105 and 4×105 Bcl-2 molecules/cell we observed a region of variable fate , with a subset of cells undergoing MOMP and others surviving ( Figure 7A , gray shaded region ) . To determine how this variability arises , we ran a series of simulations using correlated initial conditions and sampled GFP-Bcl-2 levels over the experimentally observed range of 104 to 1 . 2×106 molecules per cell . Simulations were run for 12 hr , well past the time at which the last cells died in experiments; any simulated cells that had not undergone MOMP by 12 hr were assumed to have survived . As in real cells , we observed a region of variable fate at intermediate levels of Bcl-2 ( Figure 7B , gray shaded region ) although Bcl-2 levels bounding the region of variable fate were ∼3–4-fold lower in simulations than in experiments , a discrepancy we attribute either to error in the measurement of absolute protein abundance or to imprecise model calibration . When we used simulation to compare initial protein concentrations in cells that are predicted to survive vs . cells that are predicted to die ( within the region of variable fate ) , we observed that [Bax]0 differed the most: cells that died had higher [Bax]0 and those that survived had lower [Bax]0 ( Figure 7C , double asterisks ) . Thus , whereas multiple proteins play a role in controlling variability in death time under wild-type conditions ( Figure 4B–C ) , under conditions of moderate Bcl-2 over-expression , [Bax]0 becomes the primary regulator of cell fate ( Figure 7C ) . Given these findings , we asked whether simply doubling Bcl-2 levels would change which proteins were most influential in controlling variability in tMOMP . Strikingly , simply doubling the average initial Bcl-2 concentration from 3×104 to 6×104 proteins per cell was sufficient to alter the sensitivity of tMOMP to variation in the levels of other apoptosis regulators proteins . At 3×104 Bcl-2/cell , no single protein had a dominant effect on variability in tMOMP ( Figure 4B ) whereas at 6×104 Bcl-2/cell , Bax and Bcl-2 had nearly three times greater impact than any other protein , and varying either Bax or Bcl-2 yielded nearly as much variability in tMOMP as did varying all proteins simultaneously ( sampling independently; Figure 7D ) . This demonstrates that protein over-expression ( and presumably also protein depletion ) changes the relative importance of other proteins in control of tMOMP . We also note that doubling [Bcl-2]0 significantly increased the variability of tMOMP: CV increased from 0 . 25 to 0 . 36 ( black bars in Figures 4B and 7D , and Figure S12 in Text S1 ) . Thus , the contribution made by variation in the level of particular proteins to variability in outcome is not necessarily a constant: contributions of individual proteins can vary dramatically over biologically plausible concentration ranges . Such contextual dependence of protein sensitivity also shows how protein over-expression can be misleading with respect to identifying factors that regulate a phenotype under endogenous conditions . To further explore this context sensitivity , we focused on the joint control of tMOMP by three proteins that most influenced its variability in the model: Bcl-2 , Bax , and Bid ( Figure 4B ) . We changed [Bax]0 and [Bid]0 above and below default values in discrete three and ten-fold steps , respectively ( the magnitude of these steps was chosen based on the sensitivity of tMOMP to the initial protein concentrations when evaluated under baseline conditions ) , while computing the relationship between tMOMP and [Bcl-2]0 . We observed that changing [Bax]0 shifted the tMOMP vs . [Bcl-2]0 curves along the x-axis whereas increasing [Bid]0 shifted the curves along the y-axis and also changed the sharpness of the curves ( Figure 8A ) . The net result was that changes in [Bax]0 affected the concentration of Bcl-2 at which cell fate switched from death to survival , whereas changes in [Bid]0 affected the mean and variance of tMOMP . The impact of natural variability in Bcl-2 expression ( measured in HeLa cells; orange shading ) on variance in tMOMP was greater at lower Bax levels ( Figure 8B ) . For cells with normal [Bid]0 , lowering [Bax]0 shifted the cells such that endogenous variability in [Bcl-2]0 created a huge spread in death time , with some cells surviving even at endogenous [Bcl-2]0 ( Figure 8A–B , left ) . Taken together , these results demonstrate that the sensitivity of tMOMP to Bcl-2 levels is itself sensitive to the levels of two interacting proteins ( a form of second-order sensitivity ) .
In this paper we examine the impact of naturally occurring variability in protein levels on variability in TRAIL-induced apoptosis . Our approach builds on previous work showing that variability in the timing and probability of death arises from cell-to-cell differences in protein levels that exist prior to TRAIL exposure and that this variability can therefore be modeled within a deterministic framework [15] . We make use of four features of apoptosis dynamics to explore the contributions made by variability in regulatory molecules to variability in the timing and efficiency of cell death . These dynamics were simulated by sampling from either independent or joint distributions whose variances were determined experimentally or estimated to represent the range of endogenous protein expression . Model parameters were not adjusted in this study to reproduce observed variability in responses to TRAIL; rather our ability to reproduce experimental data simply arose from substituting single values for initial conditions with log-normal distributions centered on previously determined EARM protein concentrations [15] , [23] , [29] , [34] . Using sampling and feature-based sensitivity analyses , we find that multiple upstream proteins control the timing of death ( tMOMP or tPARP ) in HeLa cells , but that XIAP is the primary determinant of the rate and extent of death ( tswitch and fPARP , respectively ) . We also find that co-variation in protein levels reduces variability in death time , particularly when activator-inhibitor pairs are assumed to co-vary positively ( e . g . Bcl-2 and Bax ) . Finally , we show through simulation and experiment that HeLa cells reside near a region of extreme sensitivity to Bcl-2 such that modest Bcl-2 over-expression causes cells to enter a region of parameter space associated with variable fate in which the primary determinants of phenotypic variability are quite different from those pertaining to normal conditions . Correlations in the levels of different proteins across a population of single cells changes the apparent importance of specific proteins in controlling cellular phenotypes . Simulations based on measured correlations in protein concentrations appear to better represent the biology of real cells . However , we have found that such correlations can also have unexpected results because correlations can mask the biochemical roles of specific proteins . To date , we have only measured ten pairwise correlations ( creating a joint distribution for five protein species ) but in real cells , extrinsic noise will correlate all proteins to some degree unless they are actively regulated otherwise [35] . Using Equation 2 ( Box 2 ) , it is straightforward to estimate the potential impact of correlated protein expression ( for pairs or larger sets of proteins ) on model output and to prioritize measurement of those correlations with the greatest potential impact . Under wild-type conditions , variability in the time at which a cell dies arises from variability in the concentrations of multiple regulatory molecules . Even perfect knowledge of the concentration of the model species that most strongly influences phenotype is only partially predictive of time-to-death because variability in other proteins makes a substantial contribution . In wild-type HeLa cells , knowledge of the eight most sensitive proteins is required to achieve a level of predictive ability ( R2∼0 . 8 ) that can be achieved by experimentally measuring the rate of the single reaction corresponding to cleavage of the initiator caspase reporter protein [15] . This reflects the fact that a dynamic measurement reporting on a complex reaction has significantly more “information content” than a series of static measurements . The ability of Bcl-2 to block apoptosis is well known [37] , [38] , [39] but our analysis sheds light on the precise mapping between the levels of Bcl-2 in individual cells and time of death . A relatively modest increase in Bcl-2 concentration ( 6-fold to 13-fold over endogenous Bcl-2 levels ) causes cells to enter a region of parameter space associated with variable fate; in this region , Bax becomes the primary factor determining whether a cell lives or dies . A further increase in Bcl-2 over-expression ( >13-fold ) causes MOMP to be blocked indefinitely and corresponds approximately to the degree of Bcl-2 over-expression found in leukemic cells [40] . We note that in other cell types , this degree of over-expression might have no effect due to compensatory changes in the levels of other proteins in the apoptotic network . In Type I cells , for example , over-expression of Bcl-2 does not block death because MOMP is not needed to trigger apoptosis [41] . We have previously suggested that tPARP in HeLa cells is primarily determined by proteins controlling the rate of initiator caspase activation [15] , but the results in this paper suggest that in other cells ( e . g . those with slightly higher Bcl-2 levels than HeLa cells , Figure 7D ) , time-to-death may be primarily determined by other proteins , such as those that control the MOMP threshold . Whether the particular sensitivity of HeLa cells to natural variation in Bcl-2 and Bax levels confers a selective advantage or whether it is accidental cannot yet be determined . The context dependence of classical and feature-based sensitivities is obvious mathematically but it is generally under-appreciated: sensitivity is not simply a function of network topology but also of position in parameter space . We show that “context dependence” is relevant over the natural range of protein concentrations found in populations of human cells . In HeLa cells , for example , variation in the levels of six proteins contributes roughly equally to variability in time of death under normal conditions , but when Bcl-2 levels are raised just two fold , only two proteins exert a significant impact . A clear implication is that we cannot consider experiments in which proteins levels are altered one at a time ( by RNAi or over-expression ) to represent univariate explorations of regulatory mechanism . Instead , protein over-expression and protein depletion shift cells in parameter space such that different proteins are dominant in controlling phenotype as compared to a wild-type context . This is true even if we consider only the immediate properties of the regulatory network and leave out the undoubtedly significant compensatory effects that occur at the level of other cellular pathways . Thus , the importance of specific proteins in a regulatory pathway is likely to be mis-estimated based on univariate and qualitative assessment of experimental perturbation; quantitative , system-level approaches promise to be more accurate in this regard . Non-genetic heterogeneity has recently emerged as an important topic in a variety of fields and it has become increasingly clear that cell-to-cell variability in protein expression is a key factor in a wide range of cellular decisions [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] . We expect the approach described here for analyzing cell-to-cell variability in receptor-mediated apoptosis to be generally useful in the analysis of other signaling systems in which phenotypic variability is observed . This is particularly true in those cases in which pre-existing variation in protein levels is a dominating influence and it is appropriate to use deterministic modeling coupled to Monte-Carlo procedures for sampling parameter distributions .
HeLa cells were maintained in DMEM ( Mediatech , Inc ) supplemented with L-glutamine ( Gibco ) , penicillin/streptomycin ( Gibco ) and 10% fetal bovine serum ( FBS , Mediatech , Inc ) . FuGENE 6 ( Roche ) was used to transfect HeLa cells expressing IMS-RP [23] with a pExchange vector ( Stratagene ) into which we cloned a cDNA for EGFP-Bcl-2 . Stable EGFP-Bcl-2 transfectants were isolated by selecting with neomycin and sorted on a FACSAria ( BD Biosciences ) to sample expression levels across a wide range . HeLa cells expressing IMS-RP and GFP-Bcl-2 were plated in a 96-well glass bottom plate ( Matrical ) . For all live-cell microscopy experiments , cells were treated with 50 ng/ml Superkiller TRAIL ( Alexis Biochemicals ) and 2 . 5 ug/ml cycloheximide ( Sigma-Aldrich ) and imaged on a Nikon TE2000E at 20× magnification with frames every 5 min in a 37°C humidified chamber in phenol-red free CO2-independent medium ( Invitrogen ) supplemented with 1% FBS , L-Glutamine , and Penicillin/Streptomycin ) . GFP-Bcl-2 fluorescence was quantified at t = 0 ( time of treatment ) by manually outlining the cell and measuring the average fluorescence intensity within the outline . MOMP was scored manually by monitoring cytosolic translocation of IMS-RP . To convert the x-axes in Figure 6A to proteins/cell , the average GFP fluorescence intensity at t = 0 was set equal to the average number of GFP-tagged proteins per cell as measured by quantitative immunoblotting ( Figure S5 in Text S1 ) . Distributions of initial protein levels were measured in untreated HeLa cells ( fixed with 4% paraformaldehyde and permeabilized with methanol ) on a FACSCalibur ( BD Biosciences ) . Antibodies were carefully validated as described in Figures S6 , S7 in Text S1 and the following antibodies were found to be suitable for measurement of total protein levels: α-Bid ( HPA000722 , Atlas Antibodies ) , α-Bax ( MAB4601 , Chemicon International ) , α-Bcl-2 ( SC7382 and SC783 , Santa Cruz Biotechnology ) , α-XIAP ( 610717 , BD Biosciences ) , α-caspase-3 ( SC7272 , Santa Cruz Biotechnology ) . Correlations in protein levels were measured by combining pairs of antibodies generated in different species or pairing fluorophore-conjugated version of the primary antibodies listed above ( α-caspase-3-AF488 , α-Bax-PE , α-XIAP-AF647 ) . Cells were gated in both forward scatter and side scatter to select a population of cells of similar size and the data analyzed in MatLab ( Mathworks ) . Simulations were run in Jacobian ( RESgroup ) using the EARM1 . 3 ordinary differential equation model . EARM1 . 3 is an extension of the original EARM1 . 1 [29] , modified to include general protein synthesis and degradation as described previously [15] . This model had been manually calibrated to represent the response of a single HeLa cell to TRAIL treatment [18] . Lists of reactions , initial protein concentration and parameter values are included in Tables S1 , S2 , S3 , S4 in Text S3 . When sets of simulations called for sampling from distributions of initial protein levels , we used a custom Perl script to generate series of random numbers that were sampled from a multivariate normal distribution with specified variances and co-variances ( see Tables S2 and S3 in Text S3 ) . These number series were then transformed to achieve the final log-normally distributed series with appropriate means and coefficients of variation . To calculate feature sensitivities numerically , series of simulation pairs were run for each protein by sampling its initial concentration uniformly in the exponent for values between 102 to 107 proteins per cell and then running simulations using 100% and 101% of this value , setting all other initial protein concentrations and rate constants at their default value . Sensitivities were then calculated using:where k100 and k101 are the values of the sampled initial protein concentration for the simulation pair ( 100% and 101% , respectively ) and similarly q100 and q101 are the values of feature q using 100% and 101% of the sampled initial protein concentration , respectively . | Variability among members of a clonal cell population is increasingly recognized as a near-universal characteristic of prokaryotic and eukaryotic cells . Variability can arise from random fluctuations in the biochemical reactions that control gene transcription , protein synthesis or signal transduction networks . For variability in receptor-mediated signaling responses ( in the current work , those activated by the death-inducing ligand TRAIL ) , we can often distinguish between the influence of stochastic processes that occur prior to ligand exposure and those that occur subsequently . One manifestation of prior variability is cell-to-cell differences in protein concentrations , and this paper uses a combination of modeling and experimentation to ask how these differences impact variability in phenotype , specifically with respect to the timing and probability of cell death . We find that fluctuations in multiple proteins contribute jointly to phenotypic variability , that the contributions of specific proteins to phenotypic variability are highly sensitive to the concentrations of other proteins , and that correlations in protein levels ( detectable experimentally ) also have a measurable impact on phenotype . Our work provides insight into the regulation of apoptosis and also represents a general approach for understanding cell-to-cell variability in signal transduction pathways . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"death",
"signaling",
"networks",
"apoptotic",
"signaling",
"signaling",
"in",
"cellular",
"processes",
"biochemistry",
"simulations",
"biology",
"antiapoptotic",
"signaling",
"molecular",
"biology",
"systems",
"biology",
"biochemistry",
"biochemical",
"simulations",
"signal",
"transduction",
"apoptotic",
"signaling",
"cascade",
"cell",
"biology",
"computational",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics",
"signaling",
"cascades"
] | 2012 | Exploring the Contextual Sensitivity of Factors that Determine Cell-to-Cell Variability in Receptor-Mediated Apoptosis |
A better description of the extent and structure of genetic diversity in dengue virus ( DENV ) in endemic settings is central to its eventual control . To this end we determined the complete coding region sequence of 187 DENV-2 genomes and 68 E genes from viruses sampled from Vietnamese patients between 1995 and 2009 . Strikingly , an episode of genotype replacement was observed , with Asian 1 lineage viruses entirely displacing the previously dominant Asian/American lineage viruses . This genotype replacement event also seems to have occurred within DENV-2 in Thailand and Cambodia , suggestive of a major difference in viral fitness . To determine the cause of this major evolutionary event we compared both the infectivity of the Asian 1 and Asian/American genotypes in mosquitoes and their viraemia levels in humans . Although there was little difference in infectivity in mosquitoes , we observed significantly higher plasma viraemia levels in paediatric patients infected with Asian 1 lineage viruses relative to Asian/American viruses , a phenotype that is predicted to result in a higher probability of human-to-mosquito transmission . These results provide a mechanistic basis to a marked change in the genetic structure of DENV-2 and more broadly underscore that an understanding of DENV evolutionary dynamics can inform the development of vaccines and anti-viral drugs .
Dengue viruses ( DENV ) are vector-borne RNA viruses of the family Flaviviridae and are classified as either four distinct viruses or different serotypes ( DENV-1 to DENV-4 ) , each of which can infect humans . Infection with DENV can cause a spectrum of outcomes , ranging from asymptomatic infection through to clinically significant disease . Severe disease in children is commonly characterised by increased vascular permeability , thrombocytopenia and a bleeding diathesis that leads to life-threatening dengue shock syndrome . Each year , approximately 40 million clinically apparent dengue cases occur globally and an estimated two-thirds of the world's population live in areas of risk [1] . There are currently no licensed dengue vaccines or specific interventions to treat the disease . The error prone RNA-dependent RNA polymerase responsible for genomic replication is central to the generation of DENV genetic diversity , upon which natural selection can act . Although viral fitness is clearly multi-faceted , selection during DENV evolution is likely to be driven in part by the underlying immune status and DENV-infection history of the human population in which the viruses are circulating . For example , patterns of DHF incidence within endemic populations such as Thailand exhibits complex wave-like dynamics with the four viral serotypes co-circulating in a single population and with each serotype dominant on a 8–10 year periodic cycle [2] . An additional layer of complexity exists in the dynamics of serotype oscillations in that there are multiple , genetically distinct viral lineages or genotypes within each virus serotype [3] . Intriguingly , these viral lineages experience a process of ongoing birth and death , possibly because of fitness differences in the human or mosquito host that allow some lineages to survive better than others . For example , in Thailand a turnover of DENV-2 strains was observed between 1980 and 1987 [4] and of DENV-3 strains in the 1990s [5] . Similarly , within-serotype lineage turnover has been documented in Sri Lanka ( DENV-3 ) [6] and Myanmar ( DENV-1 ) [7] . On a wider scale , the introduction of a genetically distinct Asian/American DENV-2 strain into the Americas resulted in the replacement , and possibly extinction , of the resident American DENV-2 lineage [8] , with the process particularly well described in Puerto Rico [9] . Lineage replacement can also have epidemiological significance . For example , the lineage replacement events in the Americas ( DENV-2 ) and Sri Lanka ( DENV-3 ) were associated with increases in disease incidence and severity [6] , [8] and thereby implying that intrinsic differences in virulence exist between viruses of the same serotype , a phenomenon consistent with earlier field observations [10] , [11] . Although the occurrence of lineage replacement is one the most intriguing aspects of DENV molecular epidemiology , its mechanistic basis is unclear . In particular , it is uncertain whether such lineage replacement events reflect ( i ) large-scale epidemiological processes that are independent of viral genotype , such as random population bottlenecks , perhaps caused by large-scale declines in mosquito numbers during the annual dry season , or ( ii ) because the viral lineages in question differ in fitness such that one is able out-compete another , perhaps because they possess mutations that allow them to evade cross-protective herd immunity . Choosing between these two models is of central importance because it enables predictions to be made as to what viral lineages are likely to proliferate in the near future , and in so doing allows the more accurate design of vaccines and anti-viral drugs . The aim of this study was to reveal , for the first time , the changing transmission patterns of DENV serotypes ( and genotypes ) in southern Viet Nam and their relationship to disease incidence . Rather than focusing on a single gene in isolation , we undertook an expansive genomic approach , resulting in the largest sample of DENV genome ( complete coding region ) sequences generated to date . Our results reveal a major lineage replacement event that has occurred within Viet Nam specifically , and in South-East Asia more generally , and which has resulted in the dominance of one viral genotype .
The dengue patients from whom virus genome sequences were determined were enrolled into an ongoing ( since 2001 ) prospective , descriptive study at the Hospital for Tropical Diseases in Ho Chi Minh City , Viet Nam . Patients ( or their parents/guardians ) gave written informed consent to participate in the study . The study protocol was approved by the Hospital for Tropical Diseases and the Oxford University Tropical Research Ethical Committee . Serological investigations ( IgM and IgG capture ELISAs ) were performed using paired plasma samples using methods described previously [12] . Serology was interpreted as suggestive of secondary infection if DENV-reactive IgG was detected in the capture ELISA in the first week of illness . DENV viraemia levels were determined using an internally-controlled , serotype-specific , real-time RT-PCR assay that has been validated and described previously [13] . Diagnostic culture for DENV was performed by inoculation of 50 µl of plasma onto C6/36 mosquito cells in plastic 75×12 mm tubes . Cultures were incubated at 30°C for 7 days , then blind passaged twice in C6/36 cells for 7 days each . DENV genomic consensus sequencing was conducted at the Sanger Institute ( Cambridge , UK ) and the Broad Institute ( Cambridge , MA , USA ) . Two approaches were employed for genomic sequencing . In the first approach ( Sanger Institute ) ( n = 45 viruses ) , PCR amplification of the DENV-2 genome ( C6/36 cells , ≤3 passages ) was performed in 3 overlapping amplimers ( 2 . 4 kb , 4 . 5 kb and 4 . 5 kb , primer sequences available on request ) . These amplimers were pooled in approximately equimolar concentrations , then randomly sheared by sonication . Sheared DNA fragments were separated according to size by gel electrophoresis and fragments corresponding to a size range of 0 . 7 to 1 . 0 kb were removed and shot-gun cloned . From each shotgun library , between 96 and 192 clones were sequenced by dideoxy sequencing using universal and reverse primers . Regions of low or no coverage we filled by specific PCR amplification and sequenced . In a second approach , viral genomes ( Viet Nam , n = 142 and Cambodia n = 39 ) were sequenced using the Broad Institute's capillary sequencing ( Applied Biosystems ) directed amplification viral sequencing pipeline ( see http://www . broadinstitute . org/annotation/viral/Dengue ) . This sequencing effort was part of the Broad Institute's Genome Resources in Dengue Consortium ( GRID ) project . Viral RNA was isolated from diagnostic plasma samples ( QIAmp viral RNA mini kit , Qiagen ) and the RNA genome reverse transcribed to cDNA with superscript III reverse transcriptase ( Invitrogen ) , random hexamers ( Roche ) and a specific oligonucleotide targeting the 3′ end of the target genome sequences ( nt 10701–10723 for DENV-2 ) . cDNA was then amplified using a high fidelity DNA polymerase ( pfu Ultra II , Stratagene ) and a pool of specific primers to produce 14 overlapping amplicons of 1 . 5 to 2 kb in size for a physical coverage of 2X . Amplicons were then sequenced in the forward and reverse direction using primer panels consisting of 96 specific primer pairs , tailed with M13 forward and reverse primer sequence , that produce 500–700 bp amplicons from the target viral genome . Amplicons were then sequenced in the forward and reverse direction using M13 primer . Total coverage delivered post amplification and sequencing was ∼8-fold . Resulting sequence reads were assembled de novo and annotated sing the Broad Institute's in-house viral assembly and annotation algorithms . All genome sequences newly determined here have been deposited in GenBank and assigned accession numbers ( Table S1 ) . In brief , the virus nucleotide sequence from position 817–2520 was amplified in 5 overlapping amplimers of between 403 and 503 nt in length . Amplification was by RT-PCR using a high fidelity polymerase and RNA extracted directly from patient plasma samples as template . Each amplimer was sequenced on both strands by conventional capillary sequencing on an Applied Biosystems 3130 genetic analyzer using specific primers . The resulting sequence reads were assembled and the E gene sequence annotated in AlignX , a software program in the vectorNTI suite ( Invitrogen , USA ) . Primers used for sequence amplification are available from the authors on request . All E gene sequences newly determined here have been deposited in GenBank and assigned accession numbers; GU211738-GU211764 , GU434146-GU434159 and GU908494- GU908520 ) . A sequence alignment was manually constructed ( in Se-Al , v2 . 0 ) for the complete coding regions ( 10173 nt ) of 187 DENV-2 genome sequences upon which phylogenetic analysis could be undertaken . In addition , to place these data in a wider phylogenetic context , we extracted the envelope ( E ) gene region from these genomic sequences and combined these data with all those E gene sequences sampled globally and available on GenBank . This resulted in a data set of 941 E gene sequences , 1485 nt in length . For the 941 E gene sequences we utilized the maximum likelihood ( ML ) approach available within the PHYML package [14] and incorporating the GTR+Γ4 model of nucleotide substitution . A bootstrap resampling process ( 1 , 000 replications ) using the neighbor-joining method available in the PAUP* package [15] was used to assess the robustness of individual nodes on the phylogeny , also employing the GTR+Γ4 substitution model and using the parameter settings taken from PHYML . A very similar phylogeny containing the same major groupings was obtained using the ML method within the GARLI package and assuming the simpler HKY85+Γ4 model [16] ( tree available from the authors on request ) . To place the evolution of the Vietnamese DENV-2 in real time we inferred a Maximum Clade Credibility ( MCC ) tree of the 187 complete genome sequences using the Bayesian Markov Chain Monte Carlo ( MCMC ) method available in the BEAST package [17] and employing the exact day of viral sampling . This analysis utilized a strict molecular clock , a GTR+Γ4 model of nucleotide substitution , a different substitution rate for each codon position , and a Bayesian skyline prior as the latter is clearly the best descriptor of the complex population dynamics of DENV . Similar results , with no major differences in topology or coalescent times , were obtained under a relaxed ( uncorrelated lognormal ) molecular clock and different substitution models ( results available from the authors on request ) . All chains were run for a sufficient length ( usually 300 million generations ) and multiple times to ensure convergence with 10% removed as burn-in . This analysis also allowed us to estimate times to common ancestry ( TMRCA ) for key nodes on the DENV-2 phylogeny . The degree of uncertainty in each parameter estimate is provided by the 95% highest posterior density ( HPD ) values , while posterior probability values provide an assessment of the degree of support for each node on the tree . The BEAST package was also used to infer Bayesian skyline plots for both the Asian I ( n = 139 ) and Asian/American ( n = 46 ) genotypes . This analysis enabled a graphical depiction of changing levels of relative genetic diversity through time ( Neτ , where Ne is the effective population size and τ the host-to-host generation time ) . Low passage ( <4 ) virus stocks of DENV-2 isolates were prepared on C6/36 cell cultures and quantified by plaque assay on BHK-21 cells with the titre calculated in plaque forming units per ml . All virus inocula for mosquito studies were prepared by infection of C6/36 cells ( MOI = 0 . 1 ) followed by culture for 4 days at 28°C . Cultures were centrifuged and the culture supernatant used to spike artificial blood meals as described below . Ae . aegypti from immatures ( larvae/pupae ) sampled from 5 discrete locations in Ho Chi Minh City ( HCMC ) , Viet Nam during May 2009 were used for generation of low filial number mosquitoes for oral infection studies . Adults were allowed to emerge in the laboratory , mate randomly , and feed on fresh defibrinated sheep blood meal held at 37°C through pieces of parafilm stretched over water-jacketed glass feeders . Adults were allowed to lay eggs on wet filter papers which were then collected and stored on dry pieces of filter paper maintained under high humidity . Larvae were reared on a diet of commercial dog food . Pupae were transferred to screened cages , and adults were fed 10% sucrose in PBS ad libitum . All mosquitoes were maintained in an insectary at 27°C , with a relative humidity of 80% and a 12∶12 light-dark cycle in an environmental chamber . In order to obtain a large number of F2 females , egg batches from multiple gonotrophic cycles were combined and hatched simultaneously . Artificial blood meals consisted of fresh , defibrinated sheep blood to which serial , ten-fold dilutions of culture supernatant from individual DENV-2 infected C6/36 cells was added at a final dilution of 1/5 per blood meal . Cohorts of 40 , 7 day-old female mosquitoes starved of sucrose for 18–24 hrs were offered an infectious blood meal containing either neat or diluted ( 10−1 , 10−2 , 10−3 ) virus culture supernatant for 45 min via membrane feeding as described above . Fully engorged mosquitoes were collected and incubated for 12 days . After 12 days mosquitoes were killed by incubation at −20°C for 1 hr and the mosquito carcass from each individual placed into 0 . 7 ml of mosquito diluent ( MD ) consisting of 20% heat-inactivated fetal bovine serum in PBS with 50 µg/ml penicillin/streptomycin , 50 µg/ml gentamicin , and 2 . 5 µg/ml fungizone . All samples were stored at −80°C before processing . Samples were thawed on ice and homogenized . Infected bodies were detected by quantitative RT-PCR ( qRT-PCR ) on 140 µl of homogenate ( experiment 1 ) or by NS1 detection using a commercial antigen capture ELISA ( BioRad ) and 50 µl of homogenate ( experiment 1 and 2 ) . NS1 detection was 100% specific and 97% sensitive relative to qRT-PCR for detection of infected carcasses ( Simmons et al , unpublished results ) . The 50% infectious dose values were determined using a three parameter dose-response curve fitted in Prism software ( Version 5 , Graphpad Software Inc , La Jolla , CA ) . All statistical analysis was performed using Intercooled STATA version 9 . 2 ( StataCorp , TX ) . Significance was assigned at P<0 . 05 for all parameters and were two-sided . Uncertainty was expressed as 95% confidence intervals . The Mann Whitney test was used for continuous variables . In addition , we compared log10-viremia levels between the two-genotypes using a multiple linear regression model . In addition to the genotype , we adjusted for the following potential confounding factors: age , sex , infection status ( primary , secondary , or unknown ) , and day of illness .
Dengue is hyperendemic in southern Viet Nam . A Ministry of Health coordinated clinical and virological surveillance system operates in the 20 southern provinces of Viet Nam and has revealed oscillations in disease incidence and serotype prevalence in this region since 1996 ( Fig . 1A ) . During 2003–2006 , DENV-2 was the most prevalent serotype detected by surveillance and its circulation was temporally associated with increased disease incidence relative to the period 1999–2002 ( Fig . 1A ) . Correspondingly , between 2003–2004 , and to a lesser extent 2005–2006 , the Hospital for Tropical Diseases in Ho Chi Minh City ( HCMC ) experienced an increased dengue case burden relative to previous years , most of which was attributable to DENV-2 infection ( Fig . 1B ) . To understand the evolutionary background to the emergence and decline of DENV-2 , we determined the complete coding region ( consensus ) sequence of this virus from 187 hospitalized patients with residential addresses in or around HCMC and sampled between 2001 and 2008 ( accession numbers in Table S1 ) . In total , 143 ( 77% ) coding region sequences were obtained directly from patient plasma , 2 ( 1 . 0% ) from cerebrospinal fluid and 42 ( 22% ) from cultured virus ( all with ≤3 passages in C6/36 cells ) . Phylogenetic analysis revealed that three major lineages of DENV-2 in Viet Nam , representing the Cosmopolitan ( n = 2 ) , Asian/American ( n = 46 ) and Asian 1 ( n = 139 ) genotypes , were present in the sampled population ( Fig . 2 ) . The complete coding region phylogeny was also marked by a strong temporal structure , such that viruses belonging to the Asian/American genotype were only sampled between 2001 and 2006 , and not thereafter ( Fig . 2 ) . In contrast , Asian 1 viruses were only sampled from 2003 onwards but quickly became the dominant genotype in the DENV-2 population ( Fig . 2 ) . Bayesian molecular clock analysis suggested the most recent common ancestor of Asian 1 viruses in southern Viet Nam existed between 1996–1998 . In contrast , the most recent common ancestor of Asian/American lineage viruses may have originated more than a decade earlier , with a credible range of dates spanning 1988–1991 . The dramatic emergence and then dominance of Asian 1 viruses is also apparent on Bayesian skyline plots of viruses from each genotype ( Fig . S1 ) . The skyline plot for the Asian/American genotype is characterized by a down turn in relative genetic diversity in recent years , indicative of a decline of incidence , while the equivalent skyline plot for Asian I genotype exhibits both continued growth and a higher level of relative genetic diversity , itself compatible with elevated fitness ( Fig . S1 ) . Genome-wide rates of evolutionary change were similar in both viruses at ∼1×10−3 nucleotide substitutions per site , per year . Importantly , the temporal structure of the DENV-2 tree was not related to spatial differences in sampling , with viruses belonging to both clades being sampled from the same geographical area ( Fig . S2 ) . To take a more expansive view of the DENV-2 virus population , we performed a phylogenetic analysis of all globally sampled E gene sequences currently available . This included 68 new E gene sequences generated in the course of this study from DENV-2 viruses sampled from patients in HCMC , Viet Nam between 1995–2009 . Furthermore , we included 39 sequences from DENV-2 isolates sampled as part of the Cambodian national dengue surveillance program and also generated in this study ( Table S1 ) . Phylogenetic analysis indicated that all available Vietnamese DENV-2 E gene sequences sampled between 1995 to 2002 ( n = 45 ) belonged to the Asian/American genotype ( Fig . 3 ) , with Asian 1 viruses being first sampled ( by this study ) in 2003 and dominating after 2006 . The temporal structure in the both the genome ( Fig . 2 ) and E gene trees ( Fig . 3 ) , together with the Bayesian molecular clock analysis , supports the contention that Asian 1 genotype viruses were a relatively recent introduction into southern Viet Nam , or circulated at a level below the sensitivity of sampling prior to 2003 . Strikingly , there are distinct similarities in the temporal structure of the Vietnamese and Cambodian E gene phylogenies , with Asian 1 genotype viruses also being first sampled in Cambodia in 2003 and apparently displacing the resident Asian/American genotype to the extent that only Asian 1 viruses were sampled in Cambodia after 2005 ( Fig . 3 ) . A similar pattern is evident in Thailand , for which sampling is most intense ( n = 225 ) , and where both Asian 1 and Asian/American genotype viruses apparently co-circulated in the 1980's , only for Asian 1 viruses to then entirely dominate the sampling since 1991 ( Fig . 3 ) . Overall , it is striking that Asian 1 genotype viruses now dominate the sampled DENV-2 population in these three high burden countries . The temporal pattern by which Asian/American and Asian 1 viruses were sampled in Viet Nam , Cambodia and Thailand is summarized in Figure 4 . Given the timing and high frequency with which Asian I viruses are sampled in Thailand it is possible that this represents the source population for this genotype , although this will need to be confirmed on a more geographically balanced sample of sequences . More broadly , this E gene phylogenetic analysis reveals that no DENV-2 from South-East Asia sampled after 2006 belonged to the Asian/American genotype , clearly indicating that this viral lineage has suffered a major decline in population frequency in this region . Finally , it is striking that although only four Vietnamese viruses fell into the Cosmopolitan genotype on the basis of E gene sequence data , three of these , sampled from HCMC during the period 2006–2009 , clustered together on the tree ( Fig . 3 ) . Hence , despite the dominance of Asian I genotype , “refugia” of other viral lineages can persist in localities with large numbers of susceptible hosts . The branch separating the Vietnamese Asian/American and Asian 1 viruses was characterized by 668 nucleotide and 73 amino acid differences , of which 16 amino acid differences can be considered non-conservative ( Table 1 ) . Of the 73 amino acid differences , only three ( NS3142K-R , NS3186R-K and NS5563K-R ) occurred in previously mapped T cell epitopes [18] , suggesting these viruses could have very similar antigenic characteristics for T cells . Two non-conservative amino acid replacements in the E protein – at E226 ( T to K ) and E228 ( G to E ) – are of particular note as they had not been previously detected in any of >2000 DENV E gene sequences available on GenBank and occur at amino acid sites that are largely invariant at the global scale . Such an evolutionary pattern is strongly suggestive of a major impact on fitness and it is notable that these changes fall in a surface exposed loop in the E protein [19] . Experimentally determining the fitness effects of these mutations , and perhaps of other that define the Asian I genotype , is clearly an important avenue of future research . To explore fitness differences in the mosquito host , we first compared the growth dynamics of three viruses from each of the Asian 1 and Asian/American genotypes in Ae . albopictus C6/36 cells in vitro . At multiplicities of infection of 0 . 1 and 0 . 01 we did not detect measurable differences in the rate of virus replication or the peak viral titre attained after 6 days of culture ( Fig . S3 ) . The 50% infectious dose ( ID50 ) of 3 low-passage virus isolates from each DENV-2 lineage was then measured in low filial cohorts of Ae . aegypti to determine if Asian 1 viruses had a measurable advantage in “infectiousness” . The Ae . aegypti mosquitoes used in these experiments were F2 generation derived from immature forms collected in HCMC , Viet Nam . Twelve days after ingesting a spiked-blood meal , DENV infection was determined by qRT-PCR and detection of NS1 in homogenates of individual mosquito carcasses . The ID50 values determined for each virus were consistent between independent experiments , but there was substantial variation between viruses from the same lineage and there was no apparent trend towards lower ID50 values for Asian 1 lineage viruses ( Table 2 ) . These data suggest that infectiousness per se for local Ae . aegypti might not account for the epidemiologically observed fitness differences between these viral lineages . We hypothesized that the rapid displacement of Asian/American viruses by Asian 1 viruses in Viet Nam reflected a relative fitness advantage that might manifest as a measurable virological feature in patients with dengue . To explore this further , the genotype of DENV-2 virus in 389 pediatric inpatients with DENV-2 infections and admitted to the Hospital for Tropical Diseases between 2001 and 2008 was determined by sequencing of nucleotides 9938-10115 ( accession numbers GU211349-GU211737 ) and alignment with reference sequences . All patients were in a prospective study of dengue and admitted to the same pediatric ward . The serological and demographic and features were not significantly different between the two groups ( Table 3 ) . Fever clearance times ( a surrogate of duration of viral infection ) did not differ significantly between patients with Asian 1 ( n = 289 ) or Asian/American ( n = 100 ) DENV-2 infections ( P = 0 . 27 , log-rank test ) , nor did platelet nadirs ( P = 0 . 07 , Mann Whitney test ) or maximum haematocrit levels ( P = 0 . 16 , Mann Whitney test ) . Intriguingly , however , viremia levels were significantly higher in patients with Asian 1 DENV-2 infections compared to Asian/American DENV-2 infections at the time of study enrolment ( Fig . 5 ) . In a multivariate analysis , log10-viremia levels were significantly higher for the Asian 1 genotype at the time of study enrolment: the adjusted mean difference between the two genotypes was 0 . 94 ( 95% CI 0 . 60 to 1 . 27; p<0 . 001 ) . Higher viremias in Asian 1 DENV-2 infections could plausibly lead to an increased rate of human-to-mosquito transmission and hence an elevated fitness .
Oscillations in dengue incidence and serotype prevalence are a characteristic feature of the epidemiology of this virus . Although changes in genotype composition through time are similarly a relatively common observation in studies of DENV evolution , to date there has been generally insufficient data to determine ( i ) whether changing patterns of genotype prevalence have any association with viral phenotype including the virological manifestation of DENV infection , and ( ii ) the evolutionary basis of these lineage replacement events , and specifically the respective roles of natural selection versus genetic drift . Here we demonstrate a rapid and apparently complete lineage ( genotype ) replacement event within DENV-2 in southern Viet Nam that was temporally associated with an increase in disease incidence . Strikingly , a functional basis for the displacement of resident Asian/American lineage viruses was suggested by higher viremia levels in pediatric patients with Asian 1 DENV-2 infections . The presence of higher viremias in children hospitalized with Asian 1 DENV-2 infections relative to Asian/American DENV-2 infections would likely increase the probability of human-to-mosquito transmission and hence facilitate greater population diffusion . Another possible outcome of higher viraemia levels is an increased incidence of more severe disease . We did not detect significant differences in the extent of capillary permeability or thrombocytopaenia between patients with Asian 1 or Asian/American viruses in the cohort of hospitalized patients ( n = 389 ) here . This suggests that Asian 1 DENV-2 infections were not overtly associated with more severe disease . However , this was a relatively small sample size to detect differences in clinical outcomes and a larger cohort of symptomatic patients , including non-hospitalised individuals , would most likely be needed to answer this question definitively . Our best estimates suggest the Asian 1 genotype was first introduced into southern Viet Nam in the late 1990's . Our sampling illuminated the replacement of the previously dominant Asian/American genotype by Asian 1 viruses during 2003–2007 , a period in which there was an almost doubling of dengue incidence , mostly associated with DENV-2 . A large number of susceptible hosts in the population , and an associated increased force of infection , could help explain the seemingly short period in which genotype replacement occurred . Whether the apparent fitness advantage of Asian 1 viruses could be attributable to “antigenic fitness” in the face of the population-wide immune landscape during this period is unknown . However , there is a precedent for biologically relevant antigenic differences between genotypes of DENV-2 . For example , South-East Asian DENV-2 viruses are less susceptible than American lineage viruses to cross-neutralization by antibodies elicited by DENV-1 infection [20] . Population wide seroepidemiology , coupled with a better understanding of correlates of immunity , are clearly needed to understand serotype and genotype replacement in all endemic regions . Virus traits in the mosquito host might also explain the difference in fitness between the Asian I and Asian/American genotypes . As an example , previous studies have demonstrated “SE Asian genotype” viruses ( there was no analysis of differences between Asian 1 or Asian/American genotypes ) are more infectious and disseminate faster in Ae . aegypti mosquitoes , and replicate more efficiently in human dendritic cells , than American genotype DENV-2 viruses [21] , [22] , [23] . However , we were unable to detect a measurable difference between Asian 1 or Asian/American viruses in terms of overall replication rates in C6/36 mosquito cells , or infectiousness for local Ae . aegypti mosquitoes . Other features of the virus-mosquito interaction ( e . g . extrinsic incubation time ) could be equally or more important than the infectious dose . However , published data from Armstrong et al . suggested that Asian 1 and Asian/American viruses had similar dissemination rates in Ae . aegypti mosquitoes [22] . Further studies are therefore needed to understand the importance of the mosquito as a site for the expression of the fitness differences between these two virus lineages . Intriguingly , the displacement of Asian/American lineage viruses by Asian 1 viruses has also seemingly occurred in Thailand and Cambodia . In Thailand , the Asian/American genotype most likely co-circulated with the Asian 1 genotype for at least a decade prior to 1991 , but is then absent from amongst the 139 Thai DENV-2 viruses sampled between 1992 and 2006 , which all belong to the Asian 1 lineage . In Cambodia , despite a smaller sample size , lineage replacement appears to have occurred along a remarkably similar time-frame to that seen in Viet Nam , with only Asian 1 viruses being sampled after 2005 . Both Thailand and Cambodia have considerable transport links with Viet Nam and it's conceivable these are sources for the introduction of Asian 1 viruses into Viet Nam . In sum , we show that a lineage replacement event in DENV is highly likely to be linked to an underlying difference in fitness . This suggests that natural selection may play a more important role in shaping viral dynamics than previously realized . The virus genetic traits associated with the fitness of Asian 1 viruses are difficult to identify definitively on the basis of sequence data alone . Elucidation of the possible functional consequences of the 16 amino acid differences that characterize the Asian I viruses will clearly require complex reverse genetic experiments . However , we predict that the amino acid changes at E226 and E228 will be of particular importance given that they occur at sites that are invariant across all DENV sequences sampled to date ( and which suggest that the vast majority of mutations at these sites are strongly deleterious because they have a major impact on fitness ) . Our documentation of a major lineage replacement event , coupled with the current dominance of Asian I viruses , suggests that this genotype will continue to dominate DENV-2 infections in Thailand , Cambodia and Viet Nam unless there is a major change in the host environment , such as that brought on by changes in serotype ( and which themselves exhibit complex population dynamics [24] ) . The prevalence of Asian I DENV-2 viruses has multiple implications . First , it is paramount that the DENV-2 component of future dengue vaccines ( reviewed in [25] ) be competent at eliciting immunity to viruses belonging to this genotype . Similarly , programs to develop anti-viral drugs for dengue should include Asian 1 DENV-2 viruses in their pre-clinical discovery and development programs [26] . Furthermore , we would predict that Asian 1 viruses will continue to outcompete Asian/American DENV-2 viruses . A likely future setting for this event is in the Americas where currently Asian/American DENV-2 viruses predominate , having themselves displaced the American genotype . | Dengue virus ( DENV ) is a mosquito transmitted RNA virus . One of the most characteristic patterns of DENV evolution is that viral lineages , including whole genotypes , are born and die-out on a regular basis . However , the precise cause of such lineage turnover is unclear . To address this issue we explored the genome-scale genetic diversity and evolution of dengue serotype 2 virus ( DENV-2 ) in Viet Nam , a country with a high burden of dengue disease . We observed that DENV viruses assigned to the Asian 1 lineage were likely introduced into southern Viet Nam in the late 1990's ( perhaps from Thailand ) and subsequently displaced the resident Asian/American lineage serotype 2 viruses . A similar scenario seems to have occurred in Thailand and Cambodia , such that there appears to have been a region-wide proliferation of Asian 1 lineage viruses . Investigation of Vietnamese patients experiencing DENV-2 infection revealed that Asian 1 viruses also attain higher virus levels in the blood than viruses of the Asian/American lineage . This difference in virus titre is likely to have a profound impact on viral fitness by increasing the probability of mosquito transmission , and therein providing Asian 1 lineage viruses with a selective advantage . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"virology/virus",
"evolution",
"and",
"symbiosis"
] | 2010 | Emergence of the Asian 1 Genotype of Dengue Virus Serotype 2 in Viet Nam: In Vivo Fitness Advantage and Lineage Replacement in South-East Asia |
Mosquito-borne Rift Valley fever virus ( RVFV ) causes acute , often severe , disease in livestock and humans . To determine the exposure factors and range of symptoms associated with human RVF , we performed a population-based cross-sectional survey in six villages across a 40 km transect in northeastern Kenya . A systematic survey of the total populations of six Northeastern Kenyan villages was performed . Among 1082 residents tested via anti-RVFV IgG ELISA , seroprevalence was 15% ( CI95% , 13–17% ) . Prevalence did not vary significantly among villages . Subject age was a significant factor , with 31% ( 154/498 ) of adults seropositive vs . only 2% of children ≤15 years ( 12/583 ) . Seroprevalence was higher among men ( 18% ) than women ( 13% ) . Factors associated with seropositivity included a history of animal exposure , non-focal fever symptoms , symptoms related to meningoencephalitis , and eye symptoms . Using cluster analysis in RVFV positive participants , a more severe symptom phenotype was empirically defined as having somatic symptoms of acute fever plus eye symptoms , and possibly one or more meningoencephalitic or hemorrhagic symptoms . Associated with this more severe disease phenotype were older age , village , recent illness , and loss of a family member during the last outbreak . In multivariate analysis , sheltering livestock ( aOR = 3 . 5 CI95% 0 . 93–13 . 61 , P = 0 . 065 ) , disposing of livestock abortus ( aOR = 4 . 11 , CI95% 0 . 63–26 . 79 , P = 0 . 14 ) , and village location ( P = 0 . 009 ) were independently associated with the severe disease phenotype . Our results demonstrate that a significant proportion of the population in northeastern Kenya has been infected with RVFV . Village and certain animal husbandry activities were associated with more severe disease . Older age , male gender , herder occupation , killing and butchering livestock , and poor visual acuity were useful markers for increased RVFV infection . Formal vision testing may therefore prove to be a helpful , low-technology tool for RVF screening during epidemics in high-risk rural settings .
Rift Valley fever virus ( RVFV ) is a mosquito-borne zoonotic disease that poses a significant risk to human health in endemic regions of Africa and the Middle East [1] . Epizootics usually precede epidemics and can result in large-scale abortion storms in local livestock populations [2] . These RVFV outbreaks in human and animal populations result in significant economic damage from trade embargos and significant livestock losses in affected areas [3] . Recent data also demonstrate that RVFV can be transmitted to humans during interepidemic periods [4–6] . RVFV infection is categorized as a neglected tropical disease due to the fact that RVFV disproportionately affects resource-limited semi-nomadic herding communities , is poverty promoting , and has long-lasting sequelae [5] . Additionally , RVF is expanding its range , threatening other areas of the world as an emerging infectious disease; notably , both Europe and the United States have the necessary vectors and livestock reservoirs to sustain autochthonous RVFV transmission [7 , 8] . The severity of RVFV manifestation , its devastating economic and public health effects , and its potential to be sustained in new regions make the study of RVFV transmission and disease a high priority . Clinically , most often RVFV causes no symptoms or a mild illness manifesting with fever and liver abnormalities [4] . More rarely , RVFV is known to cause cases of retinitis , encephalitis , or hemorrhagic diathesis with hepatitis during epidemics [9] , but these manifestations are variable and currently unpredictable . Most primary infections are thought to cause only self-limited febrile illness followed by complete recovery . It is not yet clear why severe cases occur- these consist of patients with neurologic dysfunction ( up to 8% ) , and hemorrhagic cases ( up to 1% , which is then associated with mortality of up to 50% ) [2 , 4] . Furthermore , RVFV causes visual disturbances including reversible anterior uveitis ( up to 30% of cases ) , and permanent retinitis ( up to 20% ) [10] . This broad spectrum of human RVF disease has been most recently confirmed in investigations of the 2006–2007 epidemics in East Africa [9] . Outbreaks in NE Kenya ( Garissa County ) were reported during the last epidemic [2] , but RVFV activity in nearby Ijara constituency ( Masalani and Sangailu ) , was not specifically monitored . Other reports have shown evidence of interepidemic human RVFV transmission in Ijara constituency ( Masalani ) [5–6] . It has been suggested that clinical phenotype of disease may be in part determined by the route of RVFV transmission , with animal-related transmission likely to be more severe than mosquito borne disease [1] . To expand this knowledge , the goal of the present study was to identify the exposures and other risk factors associated with human RVFV transmission and disease severity in a typical East African endemic setting , the Ijara constituency , Sangailu location , Kenya .
All participants provided written consent under a protocol approved by the Human Investigations Review Board of University Hospitals Case Medical Center ( No . 11–09–01 ) and the Ethical Review Committee of the Kenya Medical Research Institute , Nairobi , Kenya ( Non-SSC Protocol No . 195 ) . Before participation , written informed consent was obtained from adult study subjects , and parents provided written informed consent for their participating children . Children over 7 years of age also provided individual assent . This study was performed in the semi-arid Sangailu Location of Ijara constituency , Kenya . Six villages ( Golabele , Sabenale , Gedilun , Matarba , Korahindi , and Tumtish ) were sampled for demographic , epidemiological , and health information during area-wide household surveys performed from August through November of 2011 , five years after the last known RVF epidemic in the area ( 2006–2007 ) . The villages are located off a main road across a span of approximately 40 Km , in a transect running southwest to northeast ( Fig . 1 ) centered around coordinates 1 deg . 19 min S , 40 deg . 44 min E . The northern-most village , Tumtish , is located 39 Km from the border with Somalia . The participating populations studied were comprised predominantly of herders and semi-nomadic pastoralists of Somali ethnicity . A typical household landscape is shown in Fig . 2 . For unique identification and subsequent analysis of the spatial distribution of RVFV serostatus , participating household locations were geo-referenced by Global Positioning System with the use of a Garmin eTrex handheld device ( Garmin , Schaffhausen , Switzerland ) . Study recruitment began after consultation and approval by local leaders and administrators . After an initial demographic census was performed to determine the current local population and its distribution , a systematic survey of the total populations of six Northeastern Kenyan villages was performed . The villages were systematically surveyed in sequence to reach the desired sample size of >1000 enrolled individuals . All residents were eligible for inclusion , except that those residing in the area for <2 years , and children less than 1 year of age were excluded . The study sample was representative of the local ethnic mix of 99% Kenyan Somali and < 1% Bantu , Indian , or other Asian . Participants had formal interviews to detail their demographics , occupation , mosquito exposure , and animal exposure within the last 6 months , and any previous nonspecific symptoms related to RVF in the last 30 days , or severe symptoms at any time ( see Supporting S1 Text ) . Visual acuity testing by use of the Snellen chart and physical exams were performed , with an emphasis on signs of recent or remote RVF . Children under 5 yrs . old ( N = 358 ) were excluded from visual acuity testing . Poor visual acuity was defined as a score less than or equal to 6/9 meters in either eye by standard Snellen eye chart testing . A stratified subset ( 118 ) of visually symptomatic and asymptomatic subjects also underwent dilated fundoscopic eye exam and imaging using a retinal camera . Peripheral blood was then collected for serological testing for anti-RVFV IgG . Sera were tested for IgG against RVFV by standardized ELISA protocol [11 , 12] and confirmed by plaque reduction neutralization testing ( PRNT ) , as previously described [5 , 6 , 12–14] . Specimens having an ELISA OD of >0 . 25 and a PRNT titer of ≥1:20 were considered positive . The confirmatory plaque reduction neutralization testing ( PRNT ) was performed at the University of Texas , Medical Branch at Galveston . Statistical analysis examined the association of subject demographic and exposure factors with two primary outcomes: i ) odds of seropositivity and ii ) odds of having had the more severe symptoms of RVF . Initial chi-square tests were performed to identify the association of categorical factors with RVFV seropositivity and t-tests were used for continuous variables . A series of nested multivariable logistic regression models were next developed that initially included predictors significant in bivariate comparisons , as well as those considered of biological relevance prior to conduct of the study . Bivariate results and stepwise regression models were used to aid in the selection of variables to be included in the final models . Non-significant variables ( p > = 0 . 10 ) were removed in stepwise fashion to help identify the variables with the greatest multiply adjusted links to RVFV seropositivity or symptom score . For this analysis , statistical significance was set at the 0 . 05 level . Following analysis of individual symptoms , a relative RVF severity score was developed using the two-step cluster algorithm in SPSS v . 21 ( IBM , Armonk NY , USA ) to empirically define significant constellations of milder , moderate , and more severe symptom states among RVFV seropositive subjects [15] . The severe disease phenotype was defined as having somatic symptoms of acute fever plus eye symptoms and possibly one or more meningoencephalitic or hemorrhagic symptoms ( S1 Fig ) . Mild disease phenotype was defined as RVFV seropositivity with few to no symptoms . Multivariate logistic regression models were run using severe vs . mild disease categories and significant variables from bivariate analysis , excluding those variables used to define disease severity . These statistical models were performed using SAS software ( SAS Institute Inc . version 9 . 3 , Cary , NC , USA ) . Statistical analysis of spatial patterns of seropositivity among the participating households was performed with the use of Point Pattern Analysis software [16] and Clusterseer 2 . 0 software ( Biomedware , Ann Arbor , MI )
Of the 1134 participants enrolled in the study , 1082 completed all phases of the examination and were tested for RVFV infection . Of these , 164 were RVFV seropositive ( 15%; CI95% 13–17% ) . Males were more likely to be RVFV infected: 18% ( 79/433 ) were seropositive compared to 13% ( 85/646 ) of females ( P = 0 . 023; Table 1 ) . Adults ( ≥16 years old ) were also more likely than children to be RVFV infected: Thirty-one percent ( 152/487 ) of adults were seropositive compared to 2% ( 12/595 ) of children ( P < 0 . 001 ) . The average age of seropositive people was 42 ± 19 . 5 years ( range 6–85 years ) vs . 17 ± 17 for seronegatives ( range 1–84 years ) . No significant differences in seropositivity were seen among the six villages studied in the Sangailu region: Golabele ( 17 . 6%; 15/85 ) , Korahindi ( 17 . 0%; 49/288 ) ; Sabenale ( 15 . 6%; 10/64 ) ; Matarba ( 14 . 3%; 33/231 ) ; Gedilun ( 13 . 9%; 32/230 ) , and Tumtish ( 13 . 6%; 25/184 ) . This was not surprising given the uniformity of landscape and environmental features , and the socioeconomic homogeneity within pastoralist communities of this region . From initial bivariate analysis , RVFV seropositivity was significantly associated with multiple environmental exposures , as well as certain physical signs and reported symptoms ( see Table 1 ) . After multivariable adjustment , our most parsimonious logistic model of seropositive status found that older age ( 4% increase per year CI95% 2–9% , p<0 . 0001 ) , male gender ( adjusted Odds Ratio ( aOR ) = 1 . 8 , CI95% 1 . 2–2 . 7 , P < 0 . 01 ) , poor measured visual acuity ( aOR = 1 . 7 , CI95% 1 . 01–3 . 0 , P < 0 . 05 ) , a history of malaise ( aOR = 1 . 6 , CI95% 1 . 06–2 . 5 , P < 0 . 03 ) , and a history of killing livestock ( aOR = 2 . 0 , CI95% 1 . 4–3 . 3 , P < 0 . 001 ) were each independently associated with seropositivity ( Table 2 ) . Ten percent ( 112/1081 ) of people surveyed self-reported poor vision . Those who were RVFV-infected were more likely to report poor vision: 26% ( 43/164 ) , as compared to 7 . 5% ( 69/917 ) among uninfected ( P < 0 . 0001 ) . Sixteen percent ( 111/710 ) of tested subjects had poor measured visual acuity . Those who were anti-RVFV positive were more likely to have poor measured visual acuity , 36% ( 57/159 ) compared to 9 . 8% ( 54/551 ) of seronegatives ( P < 0 . 0001 ) . Fundoscopic exams were performed on 118 study participants . Here , objective eye disease was defined as having uveitis , retinitis , retinal scar or retinal hemorrhage . Overall , 26% ( 30/118 ) were found to have eye disease defined as retinitis or retinal hemorrhage . Of those serologically tested for RVFV seropositivity , 21% ( 6/28 ) were RVFV seropositive . Across the study landscape , we did not find any significant global pattern of clustering for anti-RVFV serostatus beyond the underlying distribution of households in each village ( using Ripley’s weighted K-function testing [17] over a range from 50 to 850 meters ) . However , within Korahindi , Matarba , Tumtish , and Sabenale there was evidence ( using the Getis G-statistic [18] ) of significant local clustering , at the 25–100 m scale , of greater per-household density of cases within the certain sections of these communities . Most remarkably , it was noted that all of the confirmed RVFV-positive subjects in Sabenale came from just 3 of 20 houses sampled ( Fig . 3 , left panel ) . One of these three seropositive houses in Sabenale had 7 seropositive people out of 11 total household residents . Fig . 3‘s right panel indicates the clustering pattern within the Sangailu area’s central village of Matarba . One hundred sixty-four people were RVFV exposed and included in this analysis . Ninety-three percent ( 152/164 ) were adults and 52% ( 85/164 ) were female . Thirty percent ( 49/164 ) were from Korahindi , 20% ( 33/164 ) from Matarba , 20% ( 32/164 ) from Gedilun , 15% ( 25/164 ) from Tumtish , 9% ( 15/164 ) from Golalbele , and 6% ( 10/164 ) from Sabenale . Those in the moderate/severe group ( N = 111 ) were more likely to be older ( p = 0 . 007; mean age 44 . 9 years vs . 36 . 13 years ) than those in the mild group ( N = 53 ) . In bivariate analysis , the most severe of the RVF symptom-cluster phenotypes was associated with older age , village , recent illness , and death of a family member ( Table 3 ) . In a multivariable logistic model controlling for age and village , it was found that those who sheltered livestock or disposed of livestock fetuses were at significantly greater risk for having this more severe illness complex ( Table 4 ) . Sheltering livestock put one at three and a half times the risk for more severe disease ( aOR = 3 . 5 CI95% 0 . 93–13 . 61 , P = 0 . 064 ) . Disposing of livestock abortus put one at four times risk of having a severe disease phenotype ( aOR = 4 . 11 , CI95% 0 . 63–26 . 79 , P = 0 . 140 ) .
A significant proportion of the population in the semi-arid areas of northeastern Kenya have been infected with RVFV . Other than older age , most of the factors significantly associated with anti-RVFV seropositivity and RVF disease severity were related to pastoralist lifestyles and animal exposures . These included the common practices of livestock shelter at home and livestock fetus disposal , as typically observed in this region . During epizootics , RVFV-infected herds will experience abortion storms and affected virus-contaminated abortus is often handled by herders , significantly increasing their risk for RVFV infection by aerosol and direct contact , and possibly , their risk for more severe RVF disease [1] . Similarly , infected animals brought to slaughter provide potential avenues for transmission via direct blood contact or aerosolization . Animal husbandry exposures were similar in each of the studied villages , which may explain the lack of difference in seropositivity among villages . No empiric global clustering effect was observed for household anti-RVFV seroprevalence across the study landscape , but within some villages , significant local clustering of seropositive households was documented and severe disease manifestations were more common in certain villages than others . It seems that , within certain communities , a few high-risk households carry the burden of RVFV infection , perhaps defined by combined eco-social landscape factors . Between the identified high-risk households , differences in animal husbandry practices could not be determined , but their ( unmeasured ) animal herd seropositivity could have differed , leading to greater individual household exposures . Other factors such as socioeconomic status and local landscape ( vegetation and soil ) may have also played a role in exposure risk variation as seen in the last RVFV Kenyan outbreak [19] , but these were not measured in our study . As noted in other studies , RVFV seropositivity rates were much lower among children [5 , 6 , 12] . Although there is a built-in age/time bias , in that older people have had a longer time to be exposed to RVFV infected mosquitoes and livestock , the present study and past studies suggest a significant step in RVFV infection risk over the age of 15 years [5 , 6] . Cultural practices may have limited children’s exposure , as they may be less likely to directly handle infectious materials before the age of 16 years . Of the subject symptoms we elicited , backache was the most strongly associated with RVFV seropositivity . Among confirmed , hospitalized RVF patients , an initial syndrome consisting of severe headache , fever , arthralgias , and general malaise has been described that occurs prior to the onset of delirium and mental confusion and/or hemorrhagic manifestations [9] . Among encephalitis-related symptoms , photophobia , mental confusion , and meningismus were all associated with evidence of past RVFV infection ( Table 1 ) . This is significant because longitudinal case-series in West Africa have noted that RVF may result in long-term schizophrenic or dementia-like manifestations [20] . Our observed association between mental confusion symptoms and RVFV seropositivity is consistent with this previously documented RVFV-related finding . Self-reported visual impairment and reduced measured visual acuity were both correlated with RVFV seropositivity . Uveitis and/or retinitis are two of the most common sequelae of human RVF [6] . In our study , a history of eye pain , red eyes , or photophobia ( eyes sensitive to light ) were significantly associated with RVFV seropositivity , which may have been due to RVF uveitis at the time of acute infection . Of the subset of subjects who had retinitis on fundoscopic examination , only one quarter were seropositive , demonstrating a larger burden of unrelated retinal disease in this community . The analysis of RVF disease severity ( Tables 3 and 4 ) suggests that exposure factors have only a minimal impact on the risk for disease severity , even though they increase risk for infection ( as seen in Table 1 ) . This agrees with a recent paper from Anyangu et al [1] , in which four exposure factors were associated with severe disease versus nonsevere RVF disease during bivariate analyses ( animal contact/herding animals , caring for animals during birthing , touching an aborted animal fetus , and being a herdsperson ) ; however , only one factor ( touching an aborted animal fetus ) was associated with disease severity in the multivariable model . Our study found that sheltering livestock and disposal of a livestock fetus were associated with severity of disease , but these were not statistically significant in our model . The lack of statistical significance may be due to the small study number ( 164 ) in this analysis and could be affected by the remote timing of our study in relation to the last outbreak . It is likely that individual level factors ( genes , co-morbidities ) also determine risk of disease severity , and only not mode of infection . This study was limited by the self-reported nature of the exposures and symptom data , which are subject to recall bias . Other prevalent infections , such as malaria , may have accounted for the reported fever-related symptoms . In particular , other arbovirus infections , such as West Nile virus , chikungunya virus , and dengue , might have accounted for reported ocular symptoms , as they , too , are known to cause uveitis and retinitis [21–23] . Because those local residents who had experienced severe RVFV disease in 2006–2007 could have had up to a 50% chance of dying , a survival bias is inherent in our study , and the factors associated with the most severe RVFV disease phenotype ( hemorrhagic fever and death ) may not have been identified . Also we performed our analysis of severe disease with a small sample size of only 164 participants . This limited sample size may have prevented the elucidation of some factors associated with severe RVF disease . In conclusion , RVFV infection in northeastern Kenya is significantly associated with older age , male gender , livestock harvesting , and poor vision . Spatial analysis suggests that very high-risk households exist within at-risk communities , which appear to harbor most of the RVFV infection burden . Animal exposure factors were linked to severity of human RVF disease symptoms , as suggested in previous studies [1] . Finally , the prominence of vision-related symptoms and ocular findings suggests that these may prove to be useful indicators of active or recent RVF disease in at-risk settings where serological or PCR RVFV testing is not available . | Rift Valley fever virus ( RVFV ) causes serious disease in both animals and humans . Large-scale outbreaks result in devastating economic losses and create many urgent public health concerns . Among humans , the symptoms of RVF are variable , having a broad spectrum of disease that ranges from mild to severe fever symptoms , and can include ocular complications , encephalitis , and sometimes hemorrhagic disease . In this study , 1082 at-risk Kenyan subjects were serum antibody-tested for evidence of prior RVFV infection and their demographic , health , and exposure data were collated . Seroprevalence was moderately high across the study area ( 15% ) but did not differ significantly among villages across the study region . Age , gender , and herding occupation were all significantly associated with being RVFV seropositive . Older age , village and certain animal husbandry activities were associated with more severe disease . Poor visual acuity was more likely in the seropositive group . This better definition of risk factors and associated symptom complexes should prove helpful for RVF screening during future outbreaks in high-risk rural settings . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Factors Associated with Severe Human Rift Valley Fever in Sangailu, Garissa County, Kenya |
Following introduction of antimycobacterial treatment of Buruli ulcer disease ( BUD ) , several clinical studies evaluated treatment outcomes of BUD patients , in particular healing times , secondary lesions and functional limitations . Whereas recurrences were rarely observed , paradoxical reactions and functional limitations frequently occurred . Although systematic BUD control in Togo was established as early as 2007 , treatment outcome has not been reviewed to date . Therefore , a pilot project on post-treatment follow-up of BUD patients in Togo aimed to evaluate treatment outcomes and to provide recommendations for optimization of treatment success . Out of 199 laboratory confirmed BUD patients , 129 could be enrolled in the study . The lesions of 109 patients ( 84 . 5% ) were completely healed without any complications , 5 patients ( 3 . 9% ) had secondary lesions and 15 patients ( 11 . 6% ) had functional limitations . Edema , category III ulcers >15cm , healing times >180 days and a limitation of movement at time of discharge constituted the main risk factors significantly associated with BUD related functional limitations ( P<0 . 01 ) . Review of all BUD related documentation revealed major shortcomings , in particular concerning medical records on adjuvant surgical and physiotherapeutic treatment . This study presents the first systematic analysis of treatment outcome of BUD patients from Togo . Median times to healing and the absence of recurrences were in line with findings reported by other investigators . The percentage of functional limitations of 11 . 6% was lower than in other studies , and edema , category III ulcers , healing time >180 days and limitation of movement at discharge constituted the main risk factors for functional limitations in Togolese BUD patients . Standardized treatment plans , patient assessment and follow-up , as well as improved management of medical records are recommended to allow for intensified monitoring of disease progression and healing process , to facilitate implementation of therapeutic measures and to optimize treatment success .
Buruli ulcer disease ( BUD ) , caused by Mycobacterium ulcerans , is a chronic , necrotizing skin disease which has been reported from more than 30 countries worldwide with a focus in West Africa [1] . BUD predominantly affects impoverished inhabitants of remote rural areas , approximately 50% of the cases are children <15 years [1–2] . Initially BUD manifests as painless nodule , plaque , papule , or edema followed by large , painless ulcerations with characteristically undermined edges [1–3] . Also cases with osteomyelitis occur [1–2 , 4–5] . Lesions are divided into three categories ( I: single lesions , <5 cm diameter; II: single lesions , 5–15 cm diameter; III: single lesions , >15 cm diameter , multiple lesions , lesions at critical sites , osteomyelitis ) [2] . As a result of scarring and contractures emerging during the healing process , especially patients who are not treated early suffer long-term functional disability [1 , 6] . As the mode of transmission of BUD has not been elucidated to date , proven strategies of prevention do not exist [1] . Early diagnosis and treatment are therefore core elements of BUD control which requires strong commitment of health workers at community level , laboratory confirmation of 70% of suspected cases by standardized diagnostic methods ( preferably IS2404 PCR ) , and standardized antimycobacterial treatment ( rifampicin [R] in combination with streptomycin [S] , alternatively clarithromycin [C] for eight weeks ) , if necessary complemented by surgery and/or physiotherapy [2 , 7–9] . The WHO classified BUD as one of the currently five neglected tropical diseases ( NTDs ) in line for the “innovative and intensified disease management ( IDM ) ” approach , demanding a major scaling up of active case detection , treatment , monitoring and surveillance [10] . Since the introduction of antimycobacterial combination therapy a number of clinical studies investigated the treatment outcome of BUD patients , in particular healing times , secondary lesions and functional limitations . Whereas several authors observed healing of lesions of more than 90% of patients receiving various antimycobacterial treatment regimens ( RS8 , RS4/RC4 , RS2/RC6 ) within twelve months [11–13] , information on the time to healing varies . Nienhuis et al . reported median healing times of category I lesions of 18 weeks , and 30 weeks for category II and III lesions respectively [12] . Sarfo et al . further specified median healing times for nodules of 8 weeks , for ulcers of overall 12 weeks ( category I: 12 weeks; category II: 11 weeks; category III: 15 . 5 weeks ) , and edema ranging from 2–48 weeks [11] , Phillips et al . described median healing times of 14 weeks ( RS8 ) and 16 weeks ( RS2RC6 ) [13] , and Vincent et al . observed median healing times of 12 . 6 weeks [5] . Available data from various studies also suggest that healing of up to two thirds of patients occurs within about 25 weeks after onset of treatment [5 , 12–14] . Whereas proven recurrences were non-existent [11–13] or below 2% [15] , paradoxical reactions in terms of deterioration of lesions on antibiotic treatment or the appearance of secondary lesions during or after treatment , were described for individual patients [16–18] and for larger patient cohorts . Nienhuis et al . found an increase in lesion size in up to 80% , and secondary lesions in 6% of the patients participating in the BURULICO antimicrobial trial in Ghana [12 , 19] , O’Brien et al . described paradoxical reactions in 21% of an Australian patient cohort [20] , and Phillips et al . reported 9% of paradoxical reactions in a Ghanaian patient cohort participating in a recent antimicrobial trial ( RS2/RC6 ) [13] . Increases in lesion size were commonly observed during the first three months after onset of treatment [19–20] , but also delayed paradoxical reactions in terms of new lesions occurring up to thirteen months after the end of antibiotic treatment are known [17–18] . Functional limitations were frequently observed . Data from two cohorts of laboratory confirmed BUD patients from Ghana treated between 2003 and 2005 ( surgery with or without concomitant antibiotic treatment ) , and between 2004 and 2009 ( antimycobacterial treatment with or without surgical intervention ) , suggested functional limitations in 27% and 33 . 3% of the patients [21–22] . A comparison of two patient cohorts from the Democratic Republic of the Congo treated between 2002 and 2004 ( surgical treatment only ) and 2005–2007 ( the majority of patients underwent surgery , more than 50% also received antimycobacterial therapy ) showed that 23 . 4% and 19 . 5% of the patients healed with complications [4] . A recent study from Benin analyzed a cohort of more than 1000 BUD patients treated between 2005 and 2011 with antimycobacterial combination therapy and surgery if required , and reported 22% permanent functional limitations one year after treatment [5] . Since the early 2000s , several investigators conducted in-depth assessments of functional limitations and identified important risk factors for their development , in particular location on joints and extremities of limbs , lesions >15 cm , and lesions at head and neck [21–28] . Beyond that , Vincent et al . recently established a specific profile of risk factors for BUD patients from Benin ( edema , osteomyelitis , lesions >15 cm in diameter , multifocal lesions , healing times >107 days ) and introduced the operational definition “severe Buruli ulcer” to earmark patients at risk for functional limitations for specific disability prevention measures [5] . In Togo , systematic BUD control was initiated in 2007 . Whereas case finding , laboratory confirmation and antimycobacterial treatment have been fully implemented [29–30] , accompanying POD ( prevention of disability ) measures as outlined by the WHO [6] are not yet sufficiently embedded in routine procedures , and treatment outcome has not been monitored . This study presents the first analysis of treatment outcome of BUD patients in Togo , critically reviews procedures with a possible impact on the occurrence of complications , and provides recommendations for optimization of treatment success .
Ethical clearance was obtained through the national Togolese ethics committee ( ‘‘Comité de Bioéthique pour la Recherche en Santé” ) at the University of Lomé ( 14/2010/CBRS ) and the study was approved by the ‘‘Ministère de la Santé de la République Togolaise” Lomé , Togo ( Ref . No . 0009/2011/MS/DGS/DPLET ) . Written informed consent ( IC ) was obtained in French , if necessary translated into local languages , from all study participants and/or their legal representatives if aged below 18 years . In Togo , BUD control mainly operates through a network of district based CLTs ( “Contrôleurs Lèpre-TB-Buruli” ) and community based ASCs ( “Agent Santé Communitaire” ) . CLTs regularly conduct sensitization activities in villages and schools , furthermore perform active case finding supported by ASCs who report patients with suspected BUD lesions to their corresponding CLT . Due to extended coverage of sensitization activities , self-referrals of patients to the nearest health post ( “Unité de Soins Périphérique [USP]” ) are on the rise . CLTs as well as USP head nurses ( “[ICP] Infirmier Chef Poste” ) refer clinically suspected BUD cases to the regional reference hospital ( “Centre Hospitalier Régionale [CHR] Tsévié” ) . At CHR-Tsévié a specifically trained medical assistant ( “point focal” [PF] ) is in charge of further proceedings , such as physical examination , documentation on the WHO recommended BU 01 . N and BU 01 . R forms in case of suspected recurrences [31] , sample collection and initiation of laboratory confirmation at the national hygiene institute ( “Institute Nationale d’Hygiène [INH]” ) according to standardized procedures [29–30] . Whereas most patients with category I ( partially also category II ) lesions are referred for outpatient treatment to USPs , the majority of patients with severe category II and category III lesions , and children <15 years in general are treated at CHR-Tsévié . Antimycobacterial treatment follows WHO recommendations and is complemented by surgical intervention if required [2] . Wound management at CHR-Tsévié is performed by nurses , at the USPs by the ICP , in both cases according to instructions of the PF . Patients who , according to the judgment of the PF , require physical therapy are referred to physiotherapists who provide treatment at the reference hospital , the USPs and also the patients homes . Currently 7 state examined physiotherapists are seconded to treatment of BUD patients . Eligible for the study were 199 PCR confirmed BUD patients originating from regions “Maritime” , “Savanes” , “Plateaux” and “Centrale” who were treated within the period from September 2007 to November 2013 with more than 6 months elapsed since the end of treatment . Inclusion and exclusion criteria are shown in Table 1 . Clinical , epidemiological and treatment data were retrospectively compiled from existing databases of previous studies which contained information retrieved from standardized WHO BU 01 . N and project specific laboratory data entry forms , and cross-checked with original paper forms [29–30] . To collect standardized data on treatment outcome a study specific form ( S1 Form ) was employed which consisted of several sections: A-D ) clinical/epidemiological baseline and treatment data ( taken from existing databases , prefilled prior to follow-up visits ) ; E-F ) information on location of suspected secondary lesions ( recorded in the field ) and clinical samples collected for laboratory diagnosis; G-H ) assessment of limitations of movement ( LOM ) and impairment in daily activities ( based on the questionnaire developed by Simonet V [32]; documented in the field ) ; I ) recommendations for further medical treatment ( issued after data analysis ) . In addition , for patients with open wounds at follow-up the BU 01 . R form and a clinical record form ( as used routinely in Togo; S2 Form [30] ) were filled in the field . S1 Table summarizes all parameters collected for analysis . For patients with secondary lesions and functional limitations case histories were retrospectively retrieved from medical records , where available . The distance from the patients’ location of residence to CHR was obtained from logbooks of DAHW-T cars . In addition , documentation on physiotherapy was retrospectively reviewed , as far as accessible ( “Fiche de bilan des patients atteints de l’ulcère de Buruli” and “Prévention des incapacités liées à l’UB—formulaire de base”; see S3 and S4 Forms ) . A total of 25 follow-up visits to 29 USPs ( corresponding to the catchment area of 61 villages ) were conducted in January-April 2013 ( 110 patients ) and May-June 2014 ( 19 patients ) . In advance , patients were grouped according to location of residence and accessibility of the nearest USP and summoned by the responsible ICP upon instructions of the PF at a specific date . A field team ( surgeon , physiotherapist , medical assistant and PF ) enrolled study participants at the USPs and performed clinical examination and questioning according to the above described study form ( S1 Form ) if informed consent was provided . For patients unavailable to attend an attempt was made to retrieve them in their villages at a later date . Patients whose lesions were healed without complications were discharged . From patients with open wounds at examination ( in that context referred to as secondary lesions ) lesions were measured and categorized according to WHO guidelines , and clinical samples were collected for microbiological analysis . Patients with anatomical impairment ( including excessive scars and open wounds ) were subjected to goniometric measurements according to the sagittal , frontal , and transverse rotation method ( SFTR ) [33] and scars were measured by the surgeon and medical assistant of the field team . Furthermore , these patients were questioned about the functional impairment in daily life according to the questionnaire in part H of the study form ( S1 Form ) . Functional limitations were defined as BUD related anatomical impairment as determined by goniometry and/or measurement of scars and were classified in type I ( i . e . anatomical impairment not hampering daily activities ) and type II ( i . e . anatomical impairment hampering daily activities ) . Collection of swab samples , fine-needle aspirates or 3mm punch biopsies as well as microscopy and IS2404 qPCR followed standardized procedures at the laboratories of the INH , Lomé accompanied by external quality assurance conducted at the Department for Infectious Diseases and Tropical Medicine , Munich , Germany , as recently described [30] . For regular bacteriological analysis , conducted in the accredited ( COFRAC , ‘‘Comité Français d’Accréditation” according to NF EN ISO/CEI 17025 [version 2005] ) bacteriology unit of the INH , swab samples were inoculated on Chapman ( mannitol-salt [MSA] ) agar , blood agar and nutrient broth ( BioRad , Munich , Germany ) . Colonies indicative for Staphylococcus aureus were isolated from MSA agar , analyzed by Gram staining , catalase and coagulase test , and subjected to susceptibility testing using the Kirby-Bauer disc diffusion method ( 15 antimicrobials ) on Müller-Hinton agar ( BioRad ) [34] . The study design was a non-randomized clinical cohort study . Statistical analysis ( chi-square test , including Fisher exact test ) was carried out by EPIINFO 3 . 3 . 2 . ( CDC , Atlanta , GA , USA ) . The results of statistical analyses were presented by P-values . Significant differences were defined as P-values below 0 . 05 .
Out of 199 BUD patients eligible for the study , 129 ( 64 . 8% ) could be retrieved and enrolled as follow-up patients in the study . Among the 129 follow-up BUD patients , 46 . 5% were male . At the time of initial diagnosis 90 of 129 follow-up patients ( 69 . 8% ) were below 15 years of age ( range 2–68 years , median 10 years , interquartile range [IQ] 7–16 years ) . The patients originated from 6 districts of region “Maritime” . The distance from the place of residence to CHR-Tsévié was known for 120 patients ( 93 . 0% ) and was 1–23 km for 47 patients ( 39 . 2% ) and 24–135 km for 73 patients ( 60 . 8% ) . The duration of disease before clinical diagnosis was known for 128 patients ( 99 . 2% ) and was 0–42 days for 81 patients ( 63 . 3% ) , and 43–3 . 600 days for 47 ( 36 . 7% ) patients . Baseline data and details on statistical analyses are provided in S2 Table . Out of 199 BUD patients eligible for the study , 70 patients ( 35 . 2% ) could not be enrolled ( drop-outs ) . Forty-three patients ( 61 . 4% ) had moved to an unknown address , 24 ( 34 . 3% ) were not found and 3 were deceased ( 4 . 3% ) . Among the 70 drop-out patients , 52 . 9% were male ( no significant difference with the follow-up patients ) . At the time of initial diagnosis 36 of 70 drop-out patients ( 51 . 4% ) were above 15 years of age ( range 2–65 years , median 15 . 5 years , IQ 8 . 3–28 years ) and significantly older than the follow-up patients ( P<0 . 01% ) . The drop-out patients originated from 6 districts of region “Maritime” , 2 districts of region “Plateaux” , 1 district of region “Centrale” and 1 district of region “Savanes” . The distance from the place of residence to CHR-Tsévié was known for 60 drop-out patients ( 85 . 7% ) and was 1–23 km for 34 patients ( 56 . 7% ) and 24–135 km for 26 patients ( 43 . 3% ) ; the drop-out patients lived significantly closer to CHR Tsévié than the follow-up patients ( P = 0 . 03% ) . The duration of disease before clinical diagnosis was known for 69 drop-out patients ( 98 . 6% ) and was 0–42 days for 21 patients ( 30 . 4% ) and 43–3 . 600 days for 48 ( 69 . 6% ) patients; the drop-out patients had a significantly longer duration of disease than the follow-up patients ( P<0 . 01 ) . Baseline data and details on statistical analyses are provided in S2 Table . At the time of initial diagnosis 73 of the 129 follow-up patients ( 56 . 6% ) had ulcers and 56 patients ( 43 . 4% ) had non-ulcerative lesions ( nodule , n = 19 [33 . 9%]; plaque , n = 26 [46 . 4%]; edema , n = 11 [19 . 6%] , 10 edemas evolved into an ulcer ) . Fifty-nine patients ( 45 . 7% ) had category I lesions , 44 patients had category II lesions ( 34 . 1% ) and 26 patients ( 20 . 2% ) had category III lesions . Four of the patients with category III lesions had multiple lesions [multiple ulcers , n = 2; ulcer and nodule , n = 1; ulcer and plaque , n = 1] ) . The localization of lesions was as follows: upper limbs , n = 51 ( 39 . 5% ) ; lower limbs , n = 50 ( 39 . 8% ) ; trunk/head , n = 28 ( 21 . 7% ) . Lesions of 45 patients ( 34 . 9% ) involved joints ( category I , n = 23 [51 . 1%]; category II , n = 18 [40 . 0%]; category III , n = 4 [8 . 9%] ) . LOM at time of initial diagnosis were not documented . Fourty-nine patients ( 38 . 0%; category I , n = 13; category II , n = 17; category III , n = 19 ) received antimycobacterial treatment at CHR-Tsévié , 35 of these patients ( 71 . 4% ) underwent also surgery ( excision and grafting , n = 11 [31 . 4%]; grafting , n = 23 [65 . 7%]; reconstructive surgery , n = 1 [2 . 9%] ) . Out of these 35 patients , 10 had category I lesions , 8 had category II lesions , and 17 had category III lesions . Eighty patients ( 62 . 0% ) were referred to USPs ( category I , n = 46; category II , n = 27; category III , n = 7 ) for antimycobacterial therapy . Among the 129 follow-up BUD patients , 126 patients completed antibiotic therapy ( 97 . 7% ) , three patients ( 2 . 3% ) did not ( two patients were incompliant and for the third the reason was not known ) . LOM after the end of treatment were documented for 126 patients ( 97 . 7% ) . Out of them , 17 patients ( 13 . 5%; category I , n = 2; category II , n = 6; category III , n = 9 ) were discharged with LOM . The time to healing was known for 124 patients ( 96 . 1% ) and ranged from 1–146 days for 63 patients ( 50 . 8%; significant correlation with category I lesions [P<0 . 01] ) , and 147–784 days for 61 patients ( 49 . 2%; significant correlation with category III lesions [P<0 . 01] ) . Stratified into categories of lesions , 57 patients ( 46 . 0% ) with category I lesions had a median healing time of 108 days ( IQR: 93 . 5–149 . 5 ) , 42 patients ( 33 . 9% ) with category II lesions had a median healing time of 151 days ( IQR: 125 . 8–208 ) , and 25 patients ( 20 . 1% ) with category III lesions had a median healing time of 256 days ( IQR: 177–314 ) . Among 41 patients with healing times of more than 180 days , we also observed a correlation with functional limitations ( P<0 . 01 ) . According to the BU 01 . N forms , 117 out of the 129 follow-up BUD patients ( 90 . 7% ) received physiotherapy; however for 76 of these patients ( 65 . 0% ) detailed documentation and physiotherapeutic treatment protocols were not available . Nine follow-up BUD patients ( 7 . 0% ) did not receive physiotherapy for unknown reasons and for three patients ( 2 . 3% ) this information was not available . Eighteen patients ( 23 . 1% ) were treated at CHR-Tsévié only , 60 patients ( 76 . 9% ) at USPs/patients homes and 39 patients ( 33 . 3% ) at both locations . The number of sessions was documented for 95 patients ( 81 . 2% ) : 24–99 sessions , n = 46 ( 48 . 4% ) ; 100–520 sessions , n = 49 ( 51 . 6% ) . Detailed physiotherapeutical treatment protocols , however , did not exist . Among the 129 follow-up BUD patients , the lesions of 109 patients ( 84 . 5% ) were completely healed without any complications . Five patients ( 3 . 9% ) had secondary lesions ( 2 of them in combination with functional limitations ) . Whereas M . ulcerans DNA was not detected in any of the lesions , strains of S . aureus were isolated from two patients ( one of them revealed a methicillin resistant S . aureus [MRSA] ) , in four cases the etiology of secondary lesions remained unclear . Fifteen patients ( 11 . 6% ) had functional limitations ( type I , n = 5 [3 . 9%]; type II , n = 10 [7 . 8%]; two of them in combination with secondary lesions ) . From 80 patients ( 62 . 0% ) scars were measured . Out of them , 22 patients ( 27 . 5% ) had scars with a diameter of <5cm , 33 ( 41 . 3% ) had scars with a diameter of 5–15cm , and 25 ( 31 . 3% ) had scars with a diameter of >15cm . Among the clinical findings , functional limitations were significantly associated with healing times >180 days ( P<0 . 01 ) , edema ( P<0 . 01 ) , and category III lesions ( ulcers >15cm or multiple lesions; P<0 . 01 ) , and a documented LOM at time of discharge ( P<0 . 01 ) . Treatment related factors significantly associated with functional limitations were surgery ( P<0 . 01 ) and hospitalization at CHR-Tsévié ( P<0 . 01 ) . S2 Table provides detailed risk factor analyses .
This study provides the first analysis of treatment outcome of BUD patients in Togo . The median times to healing as determined for various categories of patients lie within the range of values reported by other authors . Likewise , our data also suggest that the lesions of approximately two third of the patients healed within about 25 weeks as reported by other authors [5 , 11–14] . The absence of proven recurrences in our study is also in line with the low or non-existing recurrence rate as observed by other investigators [11–13 , 15] . As previously published , one patient of our study cohort had developed a delayed paradoxical reaction 10 months after the end of antimycobacterial treatment [18] . At the time of follow-up initial and secondary lesions were completely healed , the patient was therefore not included in the group of patients with complications . Five patients had secondary lesions at the time of clinical examination which may be related to delayed type paradoxical reactions—this is however purely speculative as the patients could not precisely indicate time of occurrence and clinical course of the lesions . From the lesions of two of these patients S . aureus strains , one of them MRSA , were isolated . Although this is the first reported case of MRSA from Togo , this finding was to be expected as investigators from the neighboring countries Ghana and Benin have recently shown that a high proportion of BUD lesions are colonized with S . aureus , and MRSA is frequently isolated [35–37] . The Togolese MRSA patient was treated with vancomycin and reportedly healed under antibiosis . It became however apparent that follow-up procedures for identification of such complications are lacking , furthermore , a concept for antibiotic management of super-infected BUD lesions does not exist . A drawback of this study was that almost 35% of laboratory confirmed patients treated with standardized antimycobacterial treatment could not be retrieved at follow-up visits . According to medical records or BU 01 . N forms , 60 of the drop-out patients ( 85 . 7% , out of them 58 patients without LOM [96 . 7%] ) were completely healed at discharge . We could however not assess long-term sequelae among the drop-out cohort . To avoid these lost to follow-ups , which are likely to occur in mobile populations such as in Togo , we strongly recommend the introduction of standardized follow-up procedures for BUD patients in Togo . Among the cohort of BUD patients retrieved for follow-up the percentage of functional limitations of 11 . 6% was lower than in other studies [4–5 , 21–22] . However , we need to mention that , in the absence of formal definitions , we introduced an operational definition of type I and II functional limitations , therefore a direct comparison between our data and other studies may not be possible without restrictions . Our data suggest that edema and category III ulcers , a healing time >180 days as well as LOM at discharge constitute the main risk factors for functional limitations in Togolese BUD patients . The finding that hospitalization and surgical treatment at CHR Tsévié were also associated with functional limitations can be explained by the fact that 73% of patients with category III lesions were hospitalized at the reference center and 89% of them underwent surgery . In analogy with the operational definition of a “severe Buruli ulcer” as established by Vincent et al . [5] , we suggest to introduce criteria for the systematic identification of patients with increased risk for functional limitations also into clinical management of BUD in Togo . We propose to draw special attention to patients initially presenting with edema and category III ulcers , furthermore—although our data did not show a significant correlation—joint involvement as shown by other authors [21–23 , 25–26] . A recent study emphasized the special importance of wound care for the prevention of BUD related functional limitations [38] . Although according to our data most Togolese patients with “severe Buruli ulcer” have already been hospitalized in CHR-Tsévié and received advanced wound management , we recommend making it a general rule . Optimal wound management should consist of daily cleansing with saline solution ( in cases of severe exudation twice a day ) , removal of necrotic tissue , and use of vaseline dressing for prevention of drying of the wound . In addition , consistent implementation of the POD related essential health interventions as outlined by the WHO are required and necessitate intensified training programs for hospital staff , CLTs , ICPs and physiotherapists [6] . This study provided an excellent opportunity to review all BUD related documentation . Clinical , epidemiological and treatment records on BU 01 . N forms were for the most part complete . The status of LOM at admission was however not documented , and information on evolution of wounds during treatment was not available . For that reason we were not able to retrospectively analyze the prevalence of early paradoxical reactions in terms of enlargement of wounds . Likewise , extensions of lesions were only known for the time of admission and it was impossible to keep track of lesions expanding over joints subsequent to initial diagnosis , which may explain the absence of a significant correlation between lesions over joints and functional limitations in our study cohort . Concerning surgery , operation reports were not available , and information on indication , type and frequency of surgical interventions was largely retrieved from handwritten notes and oral reports of PF and surgeons . For more than 60% of the patients who allegedly had received physical therapy , written documentation was absent , and treatment protocols indicating the type of exercises performed did not exist . Therefore , conclusions on the impact of physical therapy on prevention and clinical improvement of functional limitations could not be drawn in this study . In view of these findings , optimization of procedures accompanying or following antimycobacterial treatment are highly recommended . Improvement of documentation of surgical and physiotherapeutic interventions is required and shall be facilitated through filing maps . Furthermore , to standardize concomitant physiotherapeutic measures , at the time of admission each patient should be seen by a physical therapist to decide on the general requirement of physical therapy and to prepare a treatment schedule , if applicable . Upon completion of antimycobacterial treatment , the PF at CHR-Tsévié and specially trained CLTs at the USPs respectively , should conduct a standardized assessment for each patient to decide on discharge and/or further therapeutic measures . The individual package of measures for each patient shall be defined in a treatment schedule which is regularly monitored by PF and CLTs . As a general rule , all patients should be followed until complete healing of the wound , afterwards at least once per year for a five year period , thus facilitating timely recognition of two further risk factors for functional limitations , i . e . prolonged healing times and LOM at/after discharge , as well as delayed paradoxical reactions . Regular feedback on fulfillment of treatment measures and results of follow-up visits to the PNLUB-LP ( “Programme National de Lutte contre l’Ulcère de Buruli , la Lèpre et le Pian” ) is considered mandatory to enhance the transparency of the system and to allow for further evaluation and improvement . | Buruli ulcer disease ( BUD ) is a mycobacterial skin disease which leads to large ulcerations and causes disabilities in approximately 25% of the patients . Treatment consists of antimycobacterial drugs , complemented by surgery and physiotherapy if necessary . Available data on treatment outcome of BUD patients suggest that recurrences are rare; paradoxical reactions and functional limitations , however , frequently occur . BUD control in Togo was introduced already in 2007 , but treatment outcome has not yet been reviewed . Therefore , a clinical follow-up study assessed a cohort of 129 BUD patients at least six months after the end of treatment . The lesions of 84 . 5% of the patients were healed without complications , 3 . 9% had secondary lesions , and 11 . 6% , a lower proportion than in other studies , had functional limitations . Hereby , edema , category III ulcers , healing times >180 days , and limitation of movement at discharge constituted the main risk factors . Review of all BUD related documentation revealed a number of shortcomings , in particular concerning medical records . In view of these findings , standardization of procedures for creating of therapy plans , patient assessment and follow-up , as well as improved management of medical records are recommended to facilitate implementation of therapeutic measures to optimize treatment outcome and to allow for further evaluation . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Treatment Outcome of Patients with Buruli Ulcer Disease in Togo |
In fast-transcribing prokaryotic genes , such as an rrn gene in Escherichia coli , many RNA polymerases ( RNAPs ) transcribe the DNA simultaneously . Active elongation of RNAPs is often interrupted by pauses , which has been observed to cause RNAP traffic jams; yet some studies indicate that elongation seems to be faster in the presence of multiple RNAPs than elongation by a single RNAP . We propose that an interaction between RNAPs via the torque produced by RNAP motion on helically twisted DNA can explain this apparent paradox . We have incorporated the torque mechanism into a stochastic model and simulated transcription both with and without torque . Simulation results illustrate that the torque causes shorter pause durations and fewer collisions between polymerases . Our results suggest that the torsional interaction of RNAPs is an important mechanism in maintaining fast transcription times , and that transcription should be viewed as a cooperative group effort by multiple polymerases .
Transcription is the first , and often the key , step in the control of gene expression . The process of transcription has several important phases . First , an RNA polymerase ( RNAP ) binds to a promotor sequence of a gene and initiates elongation . Next , the RNAP elongates down the DNA generating a single-stranded copy of RNA , and finally terminates , releasing the nascent copy of RNA . If the resulting RNA is mRNA , it is then translated by a ribosome to a chain of amino acids that fold to produce a protein . If the resulting RNA is rRNA or tRNA , it is not translated but provides a scaffold to facilitate binding of other proteins to form RNA-protein complexes such as ribosomes . In prokaryotes , both transcription and translation happen in the cytoplasm of a cell and can occur simultaneously . Therefore regulation of gene expression in bacteria , such as E . Coli , primarily happens at the transcriptional level [1 , 2] . Elongation of RNAP along the DNA strand is not uniform , but is interrupted by frequent pauses . There are at least three different types of pauses; backtracking pauses , hairpin pauses , and ubiquitous pauses [3 , 4] . Backtracking pauses and hairpin pauses have been shown to have a higher probability of occurring during transcription of specific sequences [5–7] . On the other hand , ubiquitous pauses are thought to have no dependence on DNA sequence and are equally likely to occur at any position along the DNA strand . It has been theorized that ubiquitous pauses are caused by a restructuring of the polymerase [4] , but the exact cause remains an open question . These pauses are short ( 1–5 seconds ) and occur approximately every 100 base pairs ( bp ) [4] . There has been substantial interest to understand the effect of the presence of pauses on the average transcription time and therefore output of the RNA , for highly transcribed genes . Presence of pauses may lead to traffic jams of RNAPs when one polymerase stops , affecting the trailing polymerases [8–10] . According to Klumpp et . al . [10] , in their stochastic model RNAPs experienced a 40% reduction in the average elongation rate in dense traffic , amplifying the pause effect . This is similar to the results of a PDE model previously published by the authors [8 , 9] . In less traffic , the RNAPs in the model experienced only a 12% reduction of the average elongation rate [10] . An ODE model was developed in 2008 that studied the interactions of simultaneously transcribing RNAPs [11] . In this paper , Tripathi et . al . were able to incorporate the mechanochemical cycles of each RNAP into their model using two main states; when the pyrophosphate PPi is bound to an RNAP and when PPi is not bound . Using a mean-field approximation , they calculated the average rate of RNA production . With highly transcribed genes , the interactions between RNAPs can have a large impact on elongation efficiency . A prototypical example of a highly transcribed gene is an rrn operon in E . coli . Each E . coli genome has seven rrn operons whose transcription produces ribosomal RNA ( rRNA ) which provides a scaffold for a ribosome [12–14] . During conditions of rapid growth there are as many as 70 , 000 ribosomes in a cell . To keep up with high demand for ribosomes , 90% of transcription in fast growing E . coli produces rRNA and tRNA , and only 10% produces mRNA [15] . As a result , there is a high density of RNAPs on all rrn operons and a high transcription completion rate is imperative . Experimental measurements have shown that approximately 31% of an rrn operon is covered by RNAPs ( about 51 RNAPs ) [16] during high growth conditions , which strongly suggest that the polymerases interact either directly , or indirectly during transcription . This interaction appears to be cooperative . In vivo and in vitro experiments have demonstrated that presence of multiple RNAPs in close proximity can assist in increasing the average elongation rate . A trailing RNAP can help a paused RNAP to re-enter translocation , thereby decreasing the delay caused by pauses [17] . The magnitude of the cooperativity effect has not been firmly established . However , it is worth noting that the average elongation rate of RNAPs on the rrn operon is 90 nucleotides per second ( nt/s ) [10 , 12 , 18–20] , which is about double the in vivo elongation velocity on protein coding genes [10 , 21–23] . While the elegant paper of Epstein et . al . [17] firmly established the cooperativity effect , it did not propose a mechanistic explanation for this phenomena . In this paper we propose that the torsional force between the elongating RNAP and the DNA , caused by the helical structure of the DNA , may provide the mechanical underpinning for the interaction between elongating polymerases . The basis for our model is a set of experimental measurements by Ma and co-authors [24] . Using in vitro single-molecule experiments , they measured both the magnitude of torque exerted by elongating RNAP on DNA , and the effect of supercoiled DNA on RNAP velocity , pause density and pause duration . We construct a model for transcription that substantially extends a basic stochastic model referred to as a Totally Asymmetric Simple Exclusion Process ( TASEP ) , [2 , 8 , 10 , 26–32] . In the most basic TASEP model each individual RNAP enzyme , hops along the DNA strand with a predetermined mean hopping rate provided that the forward site is unoccupied . The entire enzyme , spanning 35 nucleotides , translocates forward one nucleotide at a time as a unit , with the position of the RNAP being determined by the front of the enzyme . In addition to the basic TASEP , a mechanism for RNAP pausing can also be implemented . When pauses are included in the TASEP implementation , both the mean pause frequency and mean pause duration are constant and chosen a priori . In our model , Elongation with Torque Assisted Motion ( ETAM ) , the rate of hopping depends on the torque between the polymerase and its closest two neighboring polymerases . The amount of torque is , in turn , the result of the relative motion of RNAPs on the DNA strand . Transcriptional pauses are included in the model as well . The mean hopping rate , mean pause frequency , and mean pause duration are dynamically updated within the model , and these parameters vary as the amount of torque varies for each RNAP . For any given RNAP that is translocating , the hopping rate and pause information depends upon the torque that is currently being experienced , and the subsequent motion is determined by sampling from the respective probability distribution functions . We base our model of the torque effects on experimental results by Ma [24] which experimentally measured the effect of torque on translocation ( hopping ) rate , pause duration and pause frequency . Over-twisting of DNA by shortening the distance between the polymerases increases ( decreases ) the translocation rate of the leading ( trailing ) polymerase , decreases ( increases ) the probability of entering a paused state for leading ( trailing ) polymerase and shortens ( lengthens ) the pause duration of the leading ( trailing ) polymerase . ETAM simulation results show that the torque-based interaction between RNAPs results in a substantial cooperation effect between RNAPs . As a trailing RNAP approaches a leading RNAP , the resulting torque increases the effective elongation rate and reduces the likelihood and duration of pauses of the leading RNAP . At the same time , the effective elongation rate of the trailing polymerase decreases , while the likelihood and duration of pauses increases . As a result of this interaction , the duration of pauses decreases , and the average number of completed transcription events increases . The effect of this interaction is not unlike that of autonomously driven and communicating vehicles ( “google” cars ) on the road . By automatically adjusting velocity and helping each other to maintain proper spacing and shorten pauses , the collective motion of polymerases becomes more efficient with an average transcription time that is 37 . 5% shorter than that produced by the TASEP model . In this sense , the RNAPs are collaborating in order to transcribe the strand more efficiently than they would if they were traveling at a constant rate .
DNA double helix structure makes one full rotation in approximately 10 . 5 base pairs [34] . RNAPs are large molecules that translocate along this twisted structure , which places constraints on the mutual motion of DNA and RNAP . If the DNA strand were fixed in space , an RNAP would have to rotate around DNA during translocation . The size of the RNAP and the packed environment inside the cell precludes this motion and a localized rotation of DNA has been observed [35 , 36] . If the DNA strand were free to rotate , it could spin along its long axis as it enters a stationary RNAP . However , if DNA is fixed upstream of the elongating RNAP and the RNAP elongates without rotation , it applies torque to the DNA . This torque is stored within the portion of the DNA strand between the fixed end and the RNAP , and if the amount of torque is large enough , it can either preclude or facilitate the forward motion of itself and its neighboring RNAP . The effect of torque on DNA—RNAP interaction was experimentally quantified in recent work of Ma et . al . [24] where a single-molecule optical trap experiment was employed in order to measure the effect of twist in a DNA strand on transcribing RNAP . In particular , they first applied a predetermined value of torque to a strand of DNA and then measured the elongation rate , the pause frequency and the pause duration of an RNAP elongating on the strand . Two types of twisting mechanisms are used to describe the applied torque . The first is over-twisting , and it is characterized by applying twist in a manner that shortens the length of the full rotation of the helix ( measured in base pairs ) . Likewise , a twist that increases the length of the full rotation of the helix is termed under-twisting . Ma and collaborators observed that over-twisting decreases the elongation rate of the RNAP and increases both the likelihood and the duration of pauses . On the other hand , under-twisting was found to increase the elongation rate and to decrease both the likelihood and the duration of the pauses . To illustrate how these results can have an effect on the transcription of DNA by multiple polymerases , consider three consecutive polymerases on a DNA strand labeled as Pi − 1 , Pi and Pi+1 in Fig 1 . This notation is defined in detail in the section Incorporating Torque into Stochastic Model . We model the small segment of the DNA strand between two neighboring RNAPs as an elastic rod , and the elongation motion of one of the RNAPs imparts a twist within the elastic rod . The torque that results from this twisting motion is calculated using classical elasticity theory , and this calculation is detailed in the section Torque Between RNAP and DNA . A brief schematic overview of the motion of the RNAPs is presented below . With respect to the motion of Pi along the strand , the RNAP represented by Pi − 1 is referred to as the leading RNAP , and the RNAP labeled Pi+1 is referred to as the trailing RNAP . If we assume that individual RNAPs never move at exactly the same time , when Pi moves forward , both Pi − 1 and Pi+1 provide anchors for the DNA strand . This movement imparts a torque to the portion of the DNA strand ( one elastic rod ) between Pi − 1 and Pi , as well as to the DNA strand between Pi and Pi+1 ( another elastic rod ) . The portion of the strand between Pi − 1 and Pi will over-twist , and the portion of the strand between Pi and Pi+1 will under-twist . Note that the over-twist will increase the elongation rate of Pi − 1 ( i . e . Pi − 1 receives a “push from the back” ) and under-twist will also increase elongation rate of Pi+1 ( i . e . Pi+1 receives a “pull from the front” ) . It is noted in a later section that both of these effects tend to synchronize the motion of all three polymerases . Our simulation results report the following quantities: the average transcription time , the average pause duration , the average collision duration time , the number of pauses and collisions , and the average transcriptional delay time experienced by an RNAP . Each of the above quantities is calculated over a range of initiation rates α ⋅ β where β is the average elongation rate of 90 nt/s and α ∈ [0 . 0001 , 0 . 0115] using 11 discrete values within this interval . An α value of 0 . 0001 corresponds to an initiation every 111 seconds on average , while α = 0 . 0115 would have an initiation every 0 . 96 seconds on average . For each value of α , we performed 50 simulations of both ETAM and TASEP and ran the simulation for 10 , 000 simulated seconds . This ensures a sufficient amount of data is collected for accurate results when compiled and averaged together . In each of the fifty simulations we record the start and end time of each RNAP transcription process , and we also record both the beginning and the end time of each pause and collision . Experimental results from physicists and biologists give an average transcription time for the rrn gene of approximately 60 seconds [4 , 16] , and in [16] , the authors also assert that the rrn gene is , on average , approximately 31% covered . This corresponds to an average velocity of 90 nt/s . With the physical parameters used for both the TASEP and ETAM model simulations , we attempt to mimic the biological case of transcription of this gene . In order to obtain the average transcription time per RNAP , the transcription time for each RNAP within the simulation is recorded , and these values are averaged over all of the RNAPs for each specific initiation rate . The results can be seen in Fig 2A where data is presented for two situations . In addition to the simulations for both TASEP and ETAM models with pauses , we also performed numerical simulations for both models in the case of no pauses for comparison purposes . We refer to the case without pauses as the “baseline” case for each model . While the RNAPs could still experience collisions for the baseline case , the number of collisions was significantly lower , see the curves with the dotted lines in Fig 2A . This baseline model allows us to calculate the transcription time for an RNAP without any transcriptional delays caused by pauses , and we note that for both models , the average transcription time is close to 60 seconds and agrees well with numbers reported in the literature . For α = 0 . 0115 , we observe a 61 . 23 second average transcription time for the baseline TASEP model and a slightly faster average transcription time of 54 . 7 seconds for the baseline ETAM model . Examining the case where transcriptional pauses are introduced into each of the models , we see very different effects , and these are shown in the solid curves of Fig 2A . For α = 0 . 0115 , the average transcription time for TASEP is approximately 156 . 25 seconds , which for a DNA strand of 5450 nucleotides corresponds to an average velocity of 34 . 88 nt/s . This rate is significantly lower than the 90 nt/s resulting from an experimental average transcription time of 60 seconds for the rrn gene reported in [4 , 16] . In contrast , for the ETAM model at α = 0 . 0115 , the average transcription time was approximately 97 . 73 seconds , which corresponds to an average velocity of 55 . 77 nt/s . This average velocity agrees much better with the experimental values . The significant difference in average transcription times between the two models is attributable to differences in the average number of pauses and their durations , as well as differences in the average number of collisions between RNAP’s in the two models . In the following sections we carefully examine the effect of torque on these quantities . In the TASEP model simulations , each RNAP experienced , on average , 70 pauses with an average duration of 0 . 55 seconds . As expected , these results are constant for the TASEP model over the entire range of initiation rates as seen for the magenta curves in Fig 3A and 3B respectively . For the ETAM model , the average number of pauses experienced by an RNAP varies significantly over the range of initiation rates included in the simulations . The number of pauses increases monotonically from 202 . 69 for α = 0 . 0001 to 2169 . 32 for α = 0 . 0115 . This is a 970 . 3% increase over the range of initiation rates; moreover , for the entire collection of these simulations , the average number of pauses was consistently larger than that of the TASEP model , see Fig 3A . For the largest value of α , RNAPs in ETAM experienced 2999% more pauses than RNAPs in TASEP . This result can be intuitively explained based on the construction of the ETAM model discussed in the Methods Section . If there are more RNAPs on a DNA strand , the RNAPs will be more likely to interact and the interaction they experience will likely be stronger than if there were fewer RNAPs . This is because the distance between RNAPs is smaller , on average , for a strand with a higher percentage of RNAP coverage . Therefore one would expect RNAPs to experience more pauses in an environment with more coverage ( higher values of α ) than they would in an environment with less coverage ( lower values of α ) . An important and interesting observation that accompanies the preceding results is the data for the average duration time of these transcriptional pauses . Our results indicate that in the ETAM model , the average duration time of the pauses decreases significantly for higher initiation rates , see Fig 3B . The duration time decreased 91 . 7% over the range of initiation rates , from 0 . 24 seconds when α = 0 . 0001 to 0 . 02 seconds when α = 0 . 0115 . Note that the values of pause duration on the order of 0 . 02 seconds would not be detectable in experiments , and so very likely the motion of polymerases at these coverages will experience fewer observable pauses . At the highest value for α , the average pause duration time is 96 . 4% lower in ETAM than in TASEP . We propose that the effects of the torque mechanism on the average duration time of pauses can be summarized as follows . While an RNAP is more likely to pause in regimes with higher initiation rates , the effects of torque , that is , the “push from the back” and the “pull from the front , ” are stronger when the neighboring elongating RNAPs are closer to the paused RNAP . Therefore , the paused RNAP can be pushed or pulled “out” of its pause state and into active elongation by means of these torsional effects much more quickly than in a regime where the pause duration time is determined by purely stochastic effects . We hypothesize that we can explain results of Fig 3C by the fact that the torque mechanism of the ETAM model allows transcribing RNAPs to maintain their spacing relative to their neighbors and to decrease the number of collisions ( as described in the Methods Section under Incorporation of Collisions ) that occur among polymerases . In order to investigate this hypothesis , we monitor and record the average number of collisions that occur and the time durations of those collisions for both the TASEP and the ETAM models . The initial intuitive expectation is that one would observe an increase in the number of collisions as the initiation rates increase ( or as the percent of coverage increases ) . That is , for a larger coverage of RNAPs on the DNA strand , collisions should become more likely . Likewise , if there are very few RNAPs on the DNA simultaneously , collisions are unlikely . As shown in Fig 3C , the data for the average number of collisions behaves as expected . However , the number of collisions increases much more rapidly in the case of the TASEP model than for the ETAM model . For the highest initiation rate simulated , RNAPs in the TASEP model experienced 1026 . 5% more collisions than RNAPs in the ETAM model . In particular , results show that approximately 550 more collisions per RNAP occur within TASEP; we observe an average of 53 . 18 collisions per RNAP in the ETAM model and an average of 599 . 09 collisions per RNAP in the TASEP model . For each of the models , the data sets shown in Fig 3C were fit with a linear least squares model , and the average number of collisions experienced by an RNAP has approximately linear growth over the range of α values for both cases . The linear fits can be seen graphically in Fig 4 , and the equations for the lines Cτ ( α ) for ETAM and C ( α ) for TASEP are given by C τ ( α ) = 4609 . 65 α C ( α ) = 52175 . 4 α Computing a ratio of the slopes of these two lines , we observe that the number of collisions for the ETAM results is growing at approximately 9% of the rate of the number of collisions of the TASEP results . Given that RNAPs in the TASEP model collide much more often , the average duration time of each collision becomes critical . Fig 3D shows that for the largest initiation rate , α = 0 . 0115 , the average duration time of a collision is significantly longer in TASEP , 0 . 104 seconds , than in ETAM , 0 . 011 seconds . Fig 3D indicates that RNAPs in the TASEP model tend to experience collisions that last approximately five times as long as those in the ETAM model . Moreover , this effect is consistent over the entire range of initiation rates included in the study . We believe this indicates that the torque is contributing to the RNAP’s ability to maintain a degree of spacing and distance with its neighbors , thereby reducing the number of collisions that occur for the ETAM case . In addition , when collisions occur in the ETAM model , they tend to be very short in duration . In summary , the inclusion of torque effects into the basic stochastic model seems to allow transcribing RNAPs to dynamically manage elongation velocity and spacing so as to avoid collisions and traffic jams . While the RNAPs experienced more pauses in the ETAM simulations , these pauses were significantly shorter in duration than those of the TASEP simulations , with far fewer collisions and shorter collision durations . To investigate more closely the effect of torque on pauses , collisions and their duration we attempt to summarize these effects by computing average transcriptional delay experienced by a polymerase in each model . During the transcription process , an RNAP experiences both pauses and collisions , each with a certain time duration . These interruptions cause active elongation of the RNAP to cease until the RNAP is able to move again . The amount of time that an RNAP is unable to translocate contributes to the delay that the RNAP experiences . To quantify this concept , we define the average total delay to be sum of both the average delay due to pauses and the average delay due to collisions . The average total delay per RNAP is computed with the formula Delay total = Delay pause + Delay collision = ( ave # of pauses per RNAP ) · ( ave pause duration ) + ( ave# of collisions per RNAP ) · ( ave collision duration ) the results of which can be seen in Fig 2B . Specific values of the total delay over the range of initiation rates for each of the ETAM and TASEP models can be found in Tables 1 and 2 , respectively . The average delay experienced by an RNAP in the TASEP model increased from 38 . 52 seconds when α = 0 . 0001 , to 100 . 61 seconds when α = 0 . 0115 . Conversely for the ETAM model , the average delay decreased from 48 . 63 seconds to 40 . 82 seconds over the same range of initiation rates . For the ETAM model , the decrease in delay values for increasing initiation rates ( and thus increasing coverage ) , is evidence to suggest that the torque contributes to an increase in efficiency with multiple RNAPs transcribing simultaneously . Moreover , the comparison of these two models allows us to observe that , in the TASEP model , the phenomena that drives the total delay is that of the delays due to collisions ( as opposed to the delays due to transcriptional pauses experienced ) . Table 2 clearly demonstrates that , for the TASEP model , the delay which the RNAPs experience due to pauses remains virtually constant over the entire range of initiation parameters and that the fact that the total delay is increasing , is almost exclusively attributable to the delays due to collisions . In contrast , Fig 3C and Table 1 demonstrate that the torque mechanism included in the ETAM model prevents many collisions from happening , and it also decreases the amount of delay that RNAPs incur from those few collisions that actually do occur . This is especially apparent with the higher initiation rates . The torsional interaction between RNAPs leads to far more efficient transcription , especially in the case of high coverage of the DNA strand . The torque provides a mechanism for the RNAPs to interact and to cooperatively prevent collisions from occurring . If an RNAP pauses , the trailing RNAP will slow down and/or enter a pause with a much higher probability due to the increasing torque applied to it as its elongation continues . The trailing RNAP will likely enter a pause or it may push the leading RNAP into active elongation before a collision occurs . The evidence of this interaction can be seen in the difference in the number of collisions experienced per RNAP in the two models in Fig 3C and Tables 1 and 2 . The comparison of the ETAM and TASEP models leads us to propose that torque is an important mechanism in the regulation of transcription . Neighboring RNAPs may interact with each other using torque to optimize their elongation efficiency as a group effort as opposed to an individual process as suggested in [17] . In this paper , Epshtein et . al . show that transcription times are faster with multiple RNAPs present on the strand as opposed to the case of a single molecule transcribing . The ETAM simulations for α = 0 . 0001 have an average transcription time per RNAP of 109 seconds , with an average time between initiations of 111 seconds , see Table 1 . With this parameter setting , the simulation is essentially a model of transcription by a single polymerase . The average transcription time for α = 0 . 0115 decreased nearly 12 seconds from the lower value of α . This is largely due to the paused RNAPs being pushed back into active elongation by their neighboring polymerases , a phenomenon suggested by Epshtein and Nudler [17] . Thus far , our results are being presented in terms of the average values , we now consider the variance of the average transcription time , the average pause duration , and the average collision duration; these are the quantities that have received careful consideration for model comparison . The coefficient of variation and the variance to mean ratio are presented in Fig 5 ( A ) and 5 ( B ) respectively . In both of these variance measures , the variability from the mean of the ETAM model decreases significantly and becomes very small in as the initiation rates increase while the TASEP model variability continues to grow . The baseline models , simulated with no pauses , are included in order to show that the variance is very close to zero for both models in the absence of pauses . With the addition of pauses , the possibility of prolonged traffic jams and collisions immediately increases the variance in both models . For the case of ETAM , as more RNAPs are added to the DNA strand , the torque interaction between polymerases drives the variance in transcription time with respect to the mean down towards zero again , as shown by the coefficient of variation and the variance to mean ratio . For both the pause duration and the collision duration , the variance for ETAM decreases as α increases . Therefore when there are fewer RNAPs on a DNA strand , the higher variance in pause duration indicates that the pause duration for the RNAPs is somewhat variable . However , as more RNAPs are transcribing simultaneously , the case for the higher α values , this variance goes to zero for both the pause duration and collision duration . The torque interaction between RNAPs is much stronger when RNAPs are closer together . This allows the RNAPs to communicate , and it stabilizes the elongation process . As a result of the decrease in pause durations and number of collisions , the RNAPs do not experience large traffic jams that can cause a great deal of variability in transcription time . In higher densities , the RNAPs can also push a paused RNAP back into elongation which causes uniformly short pause durations with very little variability . The variance for the TASEP model is significantly different than for ETAM . Since the mean hopping rate , mean pause frequency as well as the mean pause duration for TASEP are all prescribed in the model , independent of the parameter α and constant throughout the simulation process , it is expected that the variance for the pause duration remains constant over the range of initiation rates in the case of TASEP . However , the variance for the collision duration slightly increases , and the variance measures for the transcription time increase considerably . Although the variance for the collision duration only slightly increases , the average number of collisions increases significantly as α increases , see Fig 3C , and this results in the increased variability of the average transcription time at higher values of α . Without torque , the RNAPs in the TASEP model are unable to work together . Therefore , when there are many RNAPs elongating on the same DNA strand , an individual RNAP is more likely to experience a traffic jam . The increased frequency with which these traffic jams occur and the variability of the length of these traffic jams can cause a large difference in transcription times for individual RNAPs .
By incorporating the torque mechanism into a basic TASEP model of transcription we are able to see a cooperative effect among transcribing RNAPs . This effect was noted experimentally in 2003 by Epshtein and Nudler [17] . At the time , the mechanism causing this behavior was unclear . After the recent developments by Ma et al [24] , and results from our model simulations , we propose that the torsion on the DNA caused by RNAP transcription is allowing the RNAPs to communicate with each other in order to maintain proper spacing , thereby avoiding collisions , and to increase the rate of transcription . A theoretical examination of the effect of the torque mechanism proposed here can be found in the Methods Section under Mean Field Approximation Model . A classical car following model that includes forces from both the leading and trailing RNAPs yields a steady state density-velocity relationship that qualitatively explains why the torque mechanism generates cooperative behavior among the neighboring RNAPs . The cooperation between RNAPs is clearly seen from the results of our stochastic model , ETAM , which incorporates torque into a basic TASEP model . We compare the results of this model with those of the TASEP model to isolate the effect of the torque . There are two results that clearly demonstrate this cooperative effect: the average number of collisions each RNAP experiences and the average pause duration . With a high initiation rate of RNAPs onto the DNA , each polymerase experiences on average 550 fewer collisions with neighboring RNAP during the course of transcription when the elongation is regulated by torque . This is a direct result of the torque allowing the RNAPs to communicate with the polymerases that are closest to them . During a simulation of TASEP , an RNAP will elongate until it either pauses or is stopped because the next nucleotide downstream is occupied by a paused polymerase . With ETAM , as an RNAP elongates close to a paused RNAP , the resisting torque experienced by the elongating RNAP makes it much more likely to pause . At the same time , the paused RNAP experiences an assisting torque from the elongating RNAP which can push it out of a pause and into active elongation . This interaction prevents an average of 550 extra collisions from occurring in high coverage regime . The average pause duration is the other quantity most affected by the torque . With ETAM the pause duration is not a fixed quantity but can be dynamically recalculated to account for the actions of neighboring RNAPs . Pause durations can be shortened when the polymerases surrounding the paused RNAP elongate . Simulation of ETAM produced , on average , 0 . 02 second pauses as opposed to 0 . 55 second pauses simulated in TASEP . When comparing the average transcription time per polymerase we see an even more striking difference . The ETAM model shows polymerases experiencing an average transcription time of 97 . 73 seconds as opposed to 156 . 25 seconds in TASEP . The delay a polymerase experiences as a result of collisions and pauses has the largest effect on the overall transcription time . With fewer collisions and shorter pause durations , the RNAPs simulated in ETAM have significantly less delay resulting in far more efficient transcription . As promising as these results are , they depend on how we fit limited data for velocity , pause frequency , and pause duration into functions of torque as discussed in the Methods Section under Incorporating Experimental Data to Determine the Effect of Torque . With a small amount of data points available and no data for values near the stall torque and melting torque , the models that we use to fit these functions near those end points are somewhat arbitrary . These high and low values for torque are calculated quite often with a high density of RNAPs on a DNA strand since the torsional interactions are so strong . As a result , the performance of the ETAM model depends on the choice for these functional descriptions when the torque values are sampled from regions where no experimental data is currently available . This issue is explored within the Methods Section under Results from Different Pause Frequency Functions . As nanotechnology continues to improve , our hope is that data will become available for velocity , pause duration , and pause frequency at both very low and very high torque values . This will allow us to better fit our model to the data without needing to make assumptions for the extreme cases . Even with the limited data , the cooperation effect is evident in the ETAM results , with shorter transcription times in the simulations for the range of high initiation rates . With more polymerases transcribing DNA simultaneously , each RNAP experiences less delay than RNAPs transcribing with a smaller amount of polymerases . In the case of the highly transcribed genes , transcription can be viewed as a group effort , with torsional interactions allowing all of the RNAPs to transcribe more efficiently .
TASEP is a stochastic model that has been used to describe the process of both transcription and translation [2 , 8 , 10 , 26–32] . In TASEP , each RNAP enzyme hops forward with a constant rate β on a 1-dimensional strand with a discrete and finite number of sites . RNAPs cannot occupy the same site and therefore will cease active elongation if the next site is occupied ( we will refer to this as a collision ) . Only when the next site is vacated will elongation resume . We have implemented open boundary conditions where the RNAP enzymes enter the strand at a given rate and exit the strand once they reach the opposite end . Specifically , each RNAP enters our simulation ( initiates ) with a rate of ( α ⋅ β ) , and leaves the simulation ( terminates ) with a rate of ( γ ⋅ β ) . This is consistent with a differential equation model for transcription proposed originally in the late 1960s [33] . Therefore α and γ are scalars that multiply the elongation rate to obtain initiation and termination rates , respectively . For ease of simulation , the enzyme enters the simulation one nucleotide at a time , and exits the simulation by hopping off as a unit . The torque interaction between RNAPs described in the Results section above , allows the polymerases to work together using the over-twisting and under-twisting in the DNA strand . The following section makes this process more precise by deriving a mathematical expression for the amount of torque that an RNAP imparts to DNA during translocation . The role of a mean field approximation model , as for any other model , is to provide insight into a more detailed stochastic model . In deriving such a model , we inevitably simplify the fine model and therefore the deterministic model will preserve only some aspects of the detailed model . Some portions of the analysis given here use techniques borrowed from the treatment of various traffic flow models , and we point out a connection between this discussion and a particular traffic flow model in a remark at the end of this section . Our starting point in deriving the mean field approximation model is the expression for torque given by Eq ( 9 ) . Our goal is to capture behavior of small perturbations around the stochastic equilibrium when this torque is zero . This implies that there exists a preferred distance L0 such that L0 , i = L0 , i+1 = L0 for all i and where Li = Li+1 = L0 . Assume that the elongation velocity at zero torque is β . It follows that the polymerase Pi is elongating with a velocity d P i ( t ) d t = V ( τ i ( t ) ) = V μ π 2 r 4 10 . 5 ln L i + 1 L i , ( 10 ) where the torque τi at time t is given by Eq ( 9 ) , and the expression inside the logarithm is simplified by our assumption that L0 , i = L0 , i+1 = L0 for all i . The velocity function V is computed using one of the polynomial functions given by either Eqs ( 19 ) or ( 22 ) introduced in the section Incorporating Experimental Data to Determine the Effect of Torque . In particular , simulation results presented in the section Linear Fit to the Data show that a ( piecewise ) linear relationship between torque and velocity is sufficient to observe the cooperative behavior among RNAPs that is the subject of this work . Using a linear relationship between torque and velocity , say V ( τ ) = A ˜ τ + B ˜ , where A ˜ , B ˜ represent arbitrary constants , the form of the velocity described above becomes . d P i ( t ) d t = A ln L i + 1 L i + B , ( 11 ) where A , B are constants that are determined by the equation in Eq ( 9 ) as well as the piecewise linear velocity relationship used here which is found in Eq ( 22 ) . Note that the constant A < 0 since the relationship between torque and velocity is a strictly decreasing one . In order to work with all positive constants , we let C = −A and using basic rules of logarithms , we arrive at the relationship d P i ( t ) d t = C ln L i L i + 1 + B , ( 12 ) where C , B > 0 . Next we examine the quotient Li/Li+1 in view of our assumption that both distances Li and Li+1 are near the preferred spacing L0 . Note that if both RNAPs are exactly at the preferred spacing , then Li = Li+1 = L0 , and the velocity simplifies to the constant B . Hence we prescribe B = β . Now suppose that the RNAPs are near the preferred spacing but not precisely achieving that spacing . Noting that these distances represent consecutive distances surrounding the RNAP on the DNA strand , if the RNAP elongates one nucleotide , then one distance Li is perturbed just slightly from the steady state , and the other is perturbed by that same amount but with the opposite sign . That is , the quotient can be expressed as L i L i + 1 = L 0 + s L 0 - s = ( L 0 + s ) L 0 1 1 - s L 0 = 1 + s L 0 1 1 - s L 0 where |s|< < 1 is a small parameter . Using a series expansion for the term in the square brackets , we have L i L i + 1 = 1 + s L 0 1 + s L 0 + s L 0 2 + ⋯ = 1 + 2 s L 0 + O s L 0 2 Returning to the ODE system above , we make the linear approximation of the quotient in terms of s so that the RNAP velocity is approximately d P i ( t ) d t = C ln L i L i + 1 + β = C ln 1 + 2 s L 0 + β ( 13 ) Next we proceed to relate the expression s L 0 to the steady state density of the system so that the left side of Eq ( 13 ) is replaced by the steady state RNAP velocity . We use a typical convention from the traffic flow literature and assume that the density is inversely proportional to distance between polymerases ρ = 1 L . Assuming that the RNAP density , ρ is either at steady state or is a slight perturbation from steady state , we write ρ = 1 L = 1 L 0 + s = 1 L 0 1 1 + s L 0 , Using a series expansion for the term in parentheses , one obtains ρ = 1 L 0 1 - s L 0 + O s L 0 2 . Approximating the density with the linear term , we have 1 L ≈ 1 L 0 1 - s L 0 = 1 L 0 - s L 0 2 which then simplifies to s L 0 ≈ L 0 1 L 0 - 1 L = L 0 ( ρ 0 - ρ ) . Using this relation in Eq ( 13 ) , when ρ = ρ 0 = 1 L 0 , then we have that the average velocity v ( ρ ) = C ln 1 + 2 L 0 ( ρ 0 - ρ ) + β . ( 14 ) Observe , that when ρ = ρ0 the RNAP velocity is β . However , for small perturbations of ρ so that ρ > ρ0 then the argument in the logarithm is less than 1 and the particle velocity is lower than β , and if ρ < ρ0 then the argument in the logarithm is greater than 1 and the particle velocity is faster that β . This self-adjustment of the velocity is a negative feedback that rejects local perturbations away from preferred density ρ0 and preferred velocity β . That is , v ′ ( ρ ) = - 2 C L 0 1 + 2 L 0 ( ρ 0 - ρ ) , and v′ ( ρ0 ) <0 , which can be observed graphically in Fig 8 . Recent single molecule experiments by Ma and co-authors [24] attempt to advance our understanding of the relationship between the torque and the movement of the RNAP . The data obtained in those experiments suggests a relationship between torque and the elongation velocity of the RNAP . It has also been known for many years that during elongation , RNAPs experience short , frequent pauses where active elongation is stalled or arrested . Ma [24] also suggests that the nature of these pauses is tied to the amount of torque that an RNAP is experiencing at a given moment of time . The data reported in [24] was limited , as can be seen in Fig 9A–9C , and we explored various ways to develop accurate representations of these relationships that could be used in a mathematical model . The following subsections give an overview of various approaches to incorporating the data into the current model , and simulation results using these different choices are discussed later in the paper . We simulate the transcription process using a Kinetic Monte Carlo algorithm [49 , 50] which is often used for simulation of coupled Poisson processes . First , we provide an outline of the computations performed at each step of the simulation following an elongation event . The incorporation of the pauses and the recording of the collisions is detailed in this section , and we also describe the basic structure and order of events within the simulation algorithm . The small number of data points given in [24] require us to extrapolate in order to characterize the relationships between the torque and the various physical quantities ( elongation velocity , pause frequency and pause duration ) for the ETAM model in the extreme cases where the torque values are near the melting or the stall cases , that is , where the absolute value of the torque is large . For those two cases , the functional relationships are constructed based on experimental biological data from [24] as well as other literature . The results presented in the Results Section at the beginning of this paper focused on that combination of information to inform the mathematical model; however , the choices made for the functional relationships were still somewhat arbitrary . The results of this section illustrate that the choices made for the case of torque values near either melting or stall are very crucial to the results of the mathematical model . The most important result shows that the cooperative effect discussed previously is observed for some choices of data fit constructions but not for others . Hence , the results reported in this section demonstrate the need for more experimental data over a larger range of torque values in order to produce a realistic mathematical model . Although the average transcription time and total delay functions displayed for the ETAM model in the Results Section are largely concave up , the results are very different when using the piecewise linear curve fit discussed in the section Linear Fit to the Data for ETAM simulations . With the piecewise linear fit , the average transcription time and total delay have the same behavior as the nonlinear fit for small values of α , but the results continue to increase and remain concave down for the larger initiation rates . This is in sharp contrast to the behavior of the data generated using the nonlinear fit in the case of larger initiation rates . This can be seen in Fig 12 . For α = 0 . 0001 the average transcription time is approximately 108 seconds with a total delay of 45 seconds . This increases over the range of α , to a final average transcription time of 147 seconds and a delay of 78 seconds . This is a 36% increase in transcription times and a 73% increase in total delay . If we compare these to the values calculated for its nonlinear counterpart , we see that for the largest value of α = 0 . 0115 , there is a 50% increase in transcription time and a 91% increase in delay for the piecewise linear fit . While the result with the piecewise linear fit is much closer to a 60 second average transcription time than TASEP , the cooperative effect of the nonlinear fit from the Results Section diminished for this particular choice of piecewise linear fit . The cooperative effect of the ETAM in the Results Section shows a decrease in transcription time as the coverage of the DNA strand increases to biologically relevant situations , and that effect is different than both the piecewise linear fit for ETAM and the TASEP case where transcription time monotonically increases as the coverage of the DNA strand increases . Because of the difference in results between the nonlinear fit of the data and that of a piecewise linear fit , we investigate which function or combination of functions drives this difference . Below we explore several possible combinations of piecewise linear and nonlinear data fit choices , and we find that the behavior is being driven by the pause duration function , specifically the pause duration for very low torque values near the melting threshold , see Fig 13 . This figure compares the ETAM transcription time and delay results using the nonlinear fit for the pause duration function ( Fig 9 ) , and the piecewise linear fit for the pause duration function ( Fig 10 ) with two other curves . The other curves show results when the pause duration function is constructed in a piecewise manner by combining portions of the piecewise linear and nonlinear fits in order to define other composite piecewise functions for the pause duration function . The curve labeled “Nonlinear Left” shows the results for average transcription time and average delay where the piecewise linear fit is used for all functions except the pause duration . The pause duration , instead , is a new piecewise defined function that is a combination of the nonlinear fit for α ∈ [-10 , 5] and the piecewise linear fit for α ∈ [5 , 11] . Similarly the curve labeled “Nonlinear Right” represents the results obtained using the nonlinear fit for α ∈ [5 , 11] and the piecewise linear fit for α ∈ [-10 , 5] for the pause duration function . As one can observe , the results for the piecewise linear fit and the nonlinear right fit are qualitatively similar . The most surprising are the results for the nonlinear left fit . For this curve fit , with large values of α , the RNAPs experience nearly 60 second transcription times with virtually no delay . The reason for this difference in results can be seen as we investigate the graph of pause duration for torque values near -10 pN⋅nm . Fig 14 depicts the nonlinear pause duration and the piecewise linear pause duration for torque values in the interval [-10 , -5]; it is essentially a “zoom in” of the graph in Fig 10C . It’s important to note that pause durations are not fixed but are calculated upon entering a pause and then recalculated when neighboring RNAPs elongate . However , when an RNAP is paused , the recalculated pause duration will always be smaller than the original because it is either being pulled from in front or pushed from behind by the elongation of a neighbor . Therefore even if the original pause duration assigned to the RNAP is large , the recalculated pause duration is likely to be small , as evidenced by the extremely small average pause durations in high coverage environments . Hence , the range of pause duration values for the case of small torque is very important . In the case of the nonlinear data fit with low torque values , the pause duration can be set as low as zero , in which case , the RNAP is released from the pause state and is free to elongate . However for the case of the piecewise linear data fit , the lowest pause duration possible is approximately 0 . 3 seconds . With the large amount of pauses experienced by an RNAP , the difference in pause duration for low torque values is driving the difference in behavior between the nonlinear and piecewise linear data fit . To finish our discussion on pause duration we again consider Fig 13 . We concentrate on the difference in the results of the “Nonlinear” data fit and the “Nonlinear Left” data fit whose pause duration is linear for α ∈ [5 , 10] and nonlinear for α ∈ [-10 , 5] . If the behavior of the results is driven by the pause duration for low torque values , how do we account for this difference , as they have the same pause duration for those torque values ? This difference can be attributed to the pause frequency function . Recall , the “Nonlinear Left” data fit has the piecewise linear fit for velocity and pause frequency . The piecewise linear fit for large torque values would give a pause frequency of 0 . 049 when the torque is 11 pN⋅nm , as opposed to a pause frequency of 1 for the nonlinear fit . The RNAPs will experience significantly fewer pauses with the linear pause frequency , encountering on average 146 pauses when α = 0 . 0115 , as opposed to 2169 pauses . Regardless of how short the duration , a very large number of these minor interruptions in elongation ( corresponding to a large value of the frequency function ) can have a large effect on the overall transcription time of an RNAP . With this in mind , we investigate how the value of the pause frequency function when torque is 11 pN⋅nm can affect the results of the model . As mentioned earlier , the choices for data fit in the cases of very high torque values and very low torque values were somewhat arbitrary with the limited data points . One choice which has proven to be crucial is to use a pause frequency of 1 when torque is 11 pN⋅nm , the approximate value for stall torque . Here we illustrate the impact of that choice on our results . Using the nonlinear data fit for pause duration and velocity , as in Eqs ( 21 ) and ( 19 ) , we perform a set of calculations for various choices for the function value when torque is 11 pN⋅nm . We continue to fit the pause frequency function using quadratic functions similar to Eq ( 20 ) ; however the value of the pause frequency function when the torque value is 11 pN⋅nm is allowed to range over a variety of values smaller than 1 in order to compare the results , see Fig 15 . The values for pause frequency used when torque is 11 pN⋅nm are {0 . 2 , 0 . 5 , 0 . 6 , 0 . 7 , 0 . 8 , 0 . 9} . We also report the results for the original pause frequency function that takes on the value of 1 at the stall torque . Also included are the results for the case where the pause frequency function data is fit with one quadratic function which has a value of approximately 0 . 05 at stall torque . The average transcription time and total delay for all of these different choices for data fit can be seen in Fig 16A and 16B respectively . Pause frequency increasing to 0 . 05 and 0 . 2 at stall torque have the fastest transcription times , both being very close to a 60 second transcription time and , in the case of the lowest pause frequency , actually being faster than a 60 second transcription time . These two pause frequency functions give results that agree extremely well with experimental data for large values of α . If we consider the difference in the average number of pauses per RNAP under these different pause frequency functions , we can understand the difference in delay . Table 4 shows the average number of pauses per RNAP , average pause duration , and the corresponding delay due to pauses when using the various pause frequencies for the value of α = 0 . 0115 , as shown in Fig 16 . By examining the delay values in Table 4 and comparing these to Fig 16B , one can see that the pause delay contributes the majority of the time toward the total delay each polymerase experiences . The pause delay is influenced mostly by the number of pauses per polymerase which is a direct result of the choice of the pause frequency function . Another interesting result is the shift in behavior from frequency 0 . 5 to frequency 0 . 6 . The results for lower frequency values are concave up over the range of α . However with pause frequency up to 0 . 6 we begin to see a global maximum when α = 0 . 0004 . For pause frequency greater than or equal to 0 . 8 the global maximum is when α = 0 . 0007 . For values of α larger than 0 . 0007 , the coverage of the DNA strand is large enough that the RNAPs feel a substantial effect from their neighboring polymerases and begin to experience a cooperative effect . It is in this range that while the polymerases experience more pauses , the decrease in the pause duration is enough to shorten the total delay . We believe this is the range of initiation values where the RNAPs are now close enough for the torsional interaction to push a paused RNAP back into elongation , as proposed by Epshtein et . al . [17] , substantially quicker than they had been previously . As the values of α increase beyond 0 . 0007 , this cooperation becomes even more pronounced . Regardless of pause frequency , the overall cooperative behavior is clear from the decrease in delay and transcription times for the larger coverages . Mathematically , choosing a value for pause frequency equal to 1 , or even close to 1 , when torque is 11 pN⋅nm is the natural choice as this is the stall torque . However , if one is attempting to fit the experimental data with one function , one would choose a pause frequency close to 0 . 2 . For that case , we have 30% coverage of the DNA strand with an average transcription time of just under 62 seconds at the highest initiation rate . This agrees very well with the results presented by Neuman et . al . [4] . In order to properly tune the ETAM model proposed here , more data is necessary for the extreme cases near the stall torque and melting torque . In order to illustrate the importance of accurate measurements near the stall torque and the melting torque for the ETAM model , we investigate how often these torque values are sampled during the simulations of the ETAM model over the range of α . Fig 17 depicts the number of times each torque value is calculated in a simulation as a percentage of the total number of torque values that are computed , displayed as a histogram . We show the results for both the baseline simulation ( no pauses ) , and the pause simulation for α = 0 . 0001 ( Fig 17A ) and α = 0 . 0115 ( Fig 17B ) . In these simulations , we count the number of times the torque values of 10 pN⋅nm and -10 pN⋅nm are computed , and we plot this percentage as histogram bars at -10 and 10 respectively . In between these values we tabulated the number of times a torque value fell in the interval ( -10 , -9] and plotted that percentage in the histogram bars at -9 . The percentage of torque values in ( -9 , -8] were plotted in the histogram bars at -8 , and so on , up to the torque value of 9 . The percentage of values in the interval ( 9 , 10 ) were plotted under the label of <10 . As shown in Fig 17 , for our lowest initiation rate ( α = 0 . 0001 ) the algorithm computes a torque of 0 pN⋅nm approximately 50 percent of the time in the baseline simulation and just under 40 percent for the pause simulation . This is to be expected as many of the RNAPs for this value of α are transcribing as single molecules and therefore would not generate torque values away from 0 pN⋅nm . For α = 0 . 0115 , the results are very different . The interaction between RNAPs causes the 0 pN⋅nm to be computed only 10% of the time . The extreme values of -10 pN⋅nm and 10 pN⋅nm are calculated the most often . In the pause simulation -10 pN⋅nm is computed approximately 25% of the time and 10 pN⋅nm about 35% of the time . This high percentage is due to the fact that the RNAPs are very close together at this high density and therefore the interactions between them are extremely strong . As a result , the torque values related to the extreme cases of stall and melting along with the choices one uses to construct the functional relationships between torque and all three of velocity , pause duration , and pause frequency at these extreme torque values combine to create a large impact on the ETAM results . | Transcription of DNA by RNA polymerases is the first step of gene expression . It has been known that multiple RNA polymerases copying the same gene help each other to move faster , but the mechanism of this interaction is not known . We propose that the torque imposed by polymerase on helically twisted DNA and transmitted to the neighboring polymerases may play a central role in the observed cooperative behavior of polymerases . We incorporated the torque between polymerases into a basic stochastic elongation model and found that transcription times in this model match experimental data better than those of the same stochastic model without the torque effects . Using torque as the interacting mechanism of polymerases leads to significantly fewer collisions and traffic jams of polymerases . The resulting motion of polymerases resembles the motion of velocity-synchronized driverless cars on the highway . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"velocity",
"classical",
"mechanics",
"dna-binding",
"proteins",
"dna",
"transcription",
"simulation",
"and",
"modeling",
"polymerases",
"research",
"and",
"analysis",
"methods",
"curve",
"fitting",
"mathematical",
"functions",
"proteins",
"mathematical",
"and",
"statistical",
"techniques",
"gene",
"expression",
"torque",
"physics",
"biochemistry",
"biochemical",
"simulations",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"motion"
] | 2016 | A Mechanistic Model for Cooperative Behavior of Co-transcribing RNA Polymerases |
A fundamental strategy for organising connections in the nervous system is the formation of neural maps . Map formation has been most intensively studied in sensory systems where the central arrangement of axon terminals reflects the distribution of sensory neuron cell bodies in the periphery or the sensory modality . This straightforward link between anatomy and function has facilitated tremendous progress in identifying cellular and molecular mechanisms that underpin map development . Much less is known about the way in which networks that underlie locomotion are organised . We recently showed that in the Drosophila embryo , dendrites of motorneurons form a neural map , being arranged topographically in the antero-posterior axis to represent the distribution of their target muscles in the periphery . However , the way in which a dendritic myotopic map forms has not been resolved and whether postsynaptic dendrites are involved in establishing sets of connections has been relatively little explored . In this study , we show that motorneurons also form a myotopic map in a second neuropile axis , with respect to the ventral midline , and they achieve this by targeting their dendrites to distinct medio-lateral territories . We demonstrate that this map is “hard-wired”; that is , it forms in the absence of excitatory synaptic inputs or when presynaptic terminals have been displaced . We show that the midline signalling systems Slit/Robo and Netrin/Frazzled are the main molecular mechanisms that underlie dendritic targeting with respect to the midline . Robo and Frazzled are required cell-autonomously in motorneurons and the balance of their opposite actions determines the dendritic target territory . A quantitative analysis shows that dendritic morphology emerges as guidance cue receptors determine the distribution of the available dendrites , whose total length and branching frequency are specified by other cell intrinsic programmes . Our results suggest that the formation of dendritic myotopic maps in response to midline guidance cues may be a conserved strategy for organising connections in motor systems . We further propose that sets of connections may be specified , at least to a degree , by global patterning systems that deliver pre- and postsynaptic partner terminals to common “meeting regions . ”
Understanding the organisational logic of a neuronal network is a necessary step towards unravelling the mechanisms that underlie its specification and assembly . For many sensory systems , axon terminals are arranged in the central nervous system ( CNS ) to form neural representations of the topography or modality of the sensory neurons in the periphery [1] . This straightforward link between neuronal anatomy and function has fuelled the remarkable progress in identifying the underlying cellular and molecular mechanisms . In the visual system , for example , retino-topic connections are specified by matching gradients of axon guidance molecules in the retina and its target , the tectum/superior colliculus ( for review see [2] ) . For motor systems in contrast , much less is known . A central organisational principle that we recently discovered in the Drosophila embryonic nerve cord is that the input structures of motorneurons , the dendrites , are distributed in the antero-posterior axis so that they form a neural “myotopic” representation of the body wall musculature in the periphery [3] . In vertebrates motor pool–specific differences of dendrite distributions have also been observed [4]–[6] , suggesting that myotopic dendritic maps may constitute a conserved organisational framework for motor systems . Other manifest regularities of vertebrate motor systems , such as the grouping of motorneuron cell bodies into pools and columns , are thought to reflect primarily ontogenetic rather than functional relationships [7] , [8] . The idea of myotopic maps implies that different dendritic territories represent , at least to a degree , different patterns of connectivity with presynaptic neurons . Support for this notion has been found in the mouse spinal cord where the expression of the transcription factor Pea3 in certain motor pools correlates with a particular dendritic distribution . Loss of Pea3 leads to ectopic expansion of these motorneuron dendrites into the central grey matter and also aberrant innervation by Ia afferents that would normally synapse only with Pea3 negative motor pools [6] . The way in which motorneuron dendrites attain their particular morphologies and territories so as to form myotopic maps has not been resolved . Much of what we know about dendrite development derives from work on sensory systems , and in general , the shapes of dendritic trees emerge as the product of intrinsic cell specification programmes and interactions with extrinsic cues and neural activity ( reviewed in [9] , [10] ) . How dendrites are positioned in particular layers or territories remains incompletely understood . In the mammalian retina , for example , layer-specific innervation of retinal ganglion cell dendrites relies on activity-dependent dendritic pruning and consolidation [11] , [12] , while in zebrafish most retinal ganglion cells put their dendrites directly into appropriate target laminae [13] and stratification of the inner plexiform layer occurs in the absence of neural activity [14] . Such “hard-wiring” is also evident in the Drosophila olfactory system , where the graded expression of Semaphorin-1a in projection neuron dendrites contributes to the formation of a sensory map in the antennal lobe [15] , [16] . In this work , we have studied the mechanisms in a motor system that underlie the generation of different dendritic morphologies and the targeting of dendrites to distinct territories . We used the locomotor system of the Drosophila embryo , currently the only model system in which an explicit myotopic distribution of dendrites has been demonstrated [3] . First , we show that the internal muscle motorneurons fall into three morphological classes that have dendrites in distinct territories with respect to the ventral midline . These medio-lateral dendritic domains are arranged to form a myotopic representation of the muscle field . Second , we demonstrate that this myotopic map is generated by dendritic targeting and that it forms in the absence of excitatory neurotransmission and when presynaptic partner terminals have been displaced . Third , we have identified Robo and Frazzled signalling in the motorneurons as the key mechanism for dendritic targeting and map formation . Fourth , a detailed quantitative analysis of dendritic trees reveals that the programmes specifying dendritic growth and targeting are separable . Neurons generate a cell type–specific amount of dendritic length and number of branch points , while the combinatorial action of Robo and Frazzled implements the distribution of the available dendritic material in response to the midline derived cues Slit and Netrin , respectively . Just as we have demonstrated here for postsynaptic dendrites , global guidance has previously been shown to also position presynaptic sensory terminals in distinct neuropile regions [17] , [18] . We therefore suggest that such global patterning cues organise connectivity in that they coordinate the delivery of pre- and postsynaptic partner terminals to common regions .
We set out to investigate how different dendritic morphologies and territories are generated in a motor system , using the neuromuscular system of the Drosophila embryo as a model . Its principal components are segmentally repeated arrays of body wall muscles ( 30 per abdominal half segment ) , each innervated by a specific motorneuron [19] , [20] . The motorneuron dendrites are the substrate on which connections with presynaptic cholinergic interneurons form [21]–[25] . We labelled 180 cells ( on average 11 . 25 for each identified motorneuron and a minimum of five ) and charted the dendritic morphologies and territories of the motorneurons that innervate the internal muscles ( see Table S1 and Figure S1 ) , using retrograde labelling with the lipophilic tracer dyes “DiI” and “DiD . ” We did so in the context of independent landmarks , a set of Fasciclin 2-positive axon bundles [26] , at 18 . 5 h after egg laying ( AEL ) , when the motor system first becomes robustly functional [21] , [27] and the geometry of motorneuron dendritic trees has become sufficiently invariant to permit quantitative comparisons [25] . We find that there are three classes of motorneurons based on dendritic arbor morphology and territory with respect to the ventral midline: i ) motorneurons with dendrites in the lateral neuropile ( between the lateral and intermediate Fasciclin 2 tracts ) , ii ) in the lateral and intermediate neuropile ( between the intermediate and medial Fasciclin 2 tracts ) , and iii ) in the lateral , intermediate plus medial neuropile ( posterior commissure ) ( Figure 1 ) . Moreover , the medio-lateral positions of motorneuron dendrites correlate with the dorsal to ventral locations of their target muscles in the periphery . Motorneurons with dorsal targets ( DA1 , DA3 , DO1–5 ) have their dendrites in the lateral neuropile , while those innervating ventral and lateral muscles ( LL1 , VL2–4 , VO1–2 ) also have dendrites in the intermediate neuropile . Coverage of the medial neuropile is particular to motorneurons innervating the most ventral group of muscles ( VO3–6 ) . These dendritic domains are arranged in the medio-lateral axis of the neuropile in such a way that they form a neural , myotopic representation of the distribution of body wall muscles in the periphery ( Figure 1 ) . Only a single motorneuron deviates from this clear-cut correlation between dendritic medio-lateral position and target muscle location: MN-DA2 has dendrites not only in the lateral neuropile , like other motorneurons with dorsal targets , but also in the intermediate neuropile ( see also Figure S2 for all internal motorneurons ) . Having identified an experimental framework with clear , reproducible distinctions between dendritic morphologies and territories , we sought to identify the mechanisms that underlie the generation of these differences . Neurons can acquire characteristic dendritic geometries by different strategies . Dendrites might grow out radially , in a random fashion , so that the dendritic territory emerges as some branches are maintained while other , inappropriately targeted segments are pruned back . This process of radial exploration is thought to involve selective stabilisation of branches by synaptic contact and/or transmission [10]–[12] . Alternatively , dendritic growth may be biased towards a particular direction or area in response to guidance cues [16] , [28] , [29] . To distinguish between these two alternatives , we established a developmental time line , comparing dendritic territories at different developmental stages: i ) 15 h AEL , 1 h before synaptic connections first become functional; ii ) 18 . 5 h AEL , when the motor system is first robustly operational; and iii ) 21 h AEL ( hatching ) , when the system is mature [21] . At 15 h AEL dendritic trees are more variable though less extensive than at 18 . 5 h AEL . In the majority of cases ( 12/16 cell types ) motorneuron dendritic arbors have already generated the morphology and have invaded the neuropile territories that are characteristic of their more mature 18 . 5 h counterparts ( Figure 2 ) . For instance , MN-DA1 ( aCC ) , MNs-DO3–5 , and MN-DA3 have predominantly laterally located dendrites at 15 h ( 87 . 5% have entirely laterally located dendrites; n = 24 labelled cells ) , as at 18 . 5 h AEL ( 100% with entirely laterally positioned dendrites for MN-DA1 , n = 9 fills; 77 . 8% for MNs-DO3–5 , n = 18 fills; 20% for MN-DA3 , n = 24 fills , Figure 2 ) . Dendritic innervation of lateral and intermediate ( MN-DA2 , MN-LL1 , MN-VO1 [RP4] , MN-VO2 [RP1] , MN-VL2 , MN-VL3/4 [RP3] ) or lateral to medial ( MN-VO4–6 ) neuropile territories is already apparent at 15 h AEL ( n = 38 fills ) and consistently still present at 18 . 5 h AEL ( n = 95 fills; Figure 2 ) . Changes in dendritic territories between 15 h and 18 . 5 h AEL were manifest for only 4/16 of the motorneurons . Disappearance of dendritic branches transiently located in the intermediate neuropile was evident for MN-DO1 and MN-DO2: at 15 h AEL 64% of the two cells ( n = 11 ) had dendritic branches in the intermediate neuropile , while at 18 . 5 h AEL all MN-DO2 and all but one of the MN-DO1 dendrites were confined laterally ( n = 15 ) . Late exploration of the midline neuropile was seen for two other motorneurons: at 18 . 5 h AEL MN-VO3 and MN-VO4/5 have characteristic midline-targeting dendrites ( n = 19 ) , which are never present earlier , at 15 h AEL ( n = 8; Figure 2 ) . We next asked what dendritic changes might occur between 18 . 5 h AEL and hatching at 21 h AEL . Based on a subset of nine representative motorneurons ( MN-DO1 [n = 5] , MN-DO2 [n = 1] , MN-DO3 [n = 2] , MN-DA3 [n = 8] , MN-LL1 [n = 13] , MN-VL2 [n = 1] , RP3 [n = 2] , MN-VO4/5 [n = 11] , MN-VO4–6 [n = 5] ) we see no substantial change in the overall morphology of dendritic trees or their medio-lateral territories between 18 . 5 h and at 21 h AEL ( Figure 2 ) . In summary , we find that already at 15 h AEL , before the onset of synaptic input , 75% ( 12/16 ) of the internal muscle motorneuron types have their dendrites located in and confined to the territories that are characteristic of the functional ( 18 . 5 h AEL ) and mature system ( 21 h AEL , hatching ) . This suggests , at least for the majority of motorneurons , that synaptic activity is probably not required for targeting dendrites to particular domains . To test directly whether synaptic transmission is indeed dispensable for the formation of the dendritic myotopic map , we visualised motorneurons in chal13 mutant embryos [30] at 18 . 5 h AEL . These embryos fail to synthesize acetylcholine , which is the main , and at this stage probably exclusive , excitatory neurotransmitter for motorneurons [21] , [22] . Cholinergic synaptic input onto motorneurons normally commences at 16 h AEL [21] and negatively regulates dendritic growth [25] . We analysed internal muscle motorneurons representative of the three modes of establishing dendritic territories: i ) “late refining” MN-DO1 ( n = 10 ) and MN-DO2 ( n = 3 ) transiently put some dendritic branches “inappropriately” into the intermediate neuropile before confining these to the lateral neuropile by 18 . 5 h AEL; ii ) “late exploring” MN-VO4/5 ( n = 9 ) has dendrites that invade the midline neuropile relatively late , during the 15–18 . 5 h interval; iii ) MN-DA3 ( n = 11 ) , MN-LL1 ( n = 10 ) , and MN-VO4–6 ( n = 6 ) have already attained their characteristic dendritic territories by 15 h AEL . In the absence of acetylcholine we detected changes in the dendritic arbors in a fraction of MN-DA3 and MN-LL1 cells , as compared to wild-type . 4/11 MN-DA3 had larger and 3/10 MN-LL1 had smaller than normal dendritic arborisations in the intermediate neuropile . However , all other motorneurons studied in 18 . 5 h chal13 mutant embryos had dendritic morphologies and innervated territories that were comparable to wild-type ( Figure 3 ) . This shows that excitatory synaptic input is not essential for the development of normal overall motorneuron dendrite morphology or the formation of the myotopic map . The dendritic trees of motorneurons form within an existing scaffold of interneuron axons and we previously found evidence for dendritic growth being regulated by contact with presynaptic partner terminals [25] . We therefore asked if the presynaptic partner terminals might provide patterning information for the dendritic myotopic map . To this end , we displaced the presynaptic partner axons , contained in the set of cholinergic interneurons [22] , by expression of two potent chimeric axon guidance receptors: UAS-Fraextracellular-Robointracellular-myc ( UAS-FraexRoin ) , shown to shift axons away from the midline , and UAS-Roboextracellular-Fraintracellular-myc ( UAS-RoexFrain ) , which can mediate the opposite effect [31] . As expected , expression of UAS-FraexRoin with Cha-GAL4 leads to a severe depletion or absence of cholinergic axons in the commissures and expression of UAS-RoexFrain to a thickening of commissural cholinergic tracts ( Figure 4B–4D′; see also [25] ) . In addition , we find that expression of either chimeric construct efficiently displaces cholinergic axon terminals out of the dorsal motor neuropile . Under these conditions contact of motorneuron dendrites with cholinergic interneuron terminals is severely reduced and potentially absent at 18 . 5 h AEL , unlike in the wild-type ( Figure 4E–4G ) ; yet the overall organization of the neuropile , including the distribution of Fasciclin 2-positive tracts , is not obviously affected ( see also [25] ) . We find that these manipulations do not obviously affect midline targeting of dendrites by motorneurons that innervate ventral oblique muscles ( n = 4 , unpublished data ) . Dendritic intermediate ( MN-LL1; white curved arrows ) and lateral ( MN-DA3 ) territories also remain distinct , though appear more variable as compared to controls ( Figure 4H–4J ) . We conclude that the presynaptic partner axons do not provide positional information necessary for the myotopic map of motorneuron dendrites . However , the increased variability observed under these experimental conditions suggests interactions with presynaptic partners may influence finer aspects of dendritic arbors as previously shown [25] . Next we asked whether the different distributions of motorneuron dendrites with respect to the ventral midline resulted from different responses to midline derived guidance cues . Obvious candidates are Slit and Netrins and their respective receptors , Robo and Frazzled . These have been shown to regulate the position of outgrowing axons by attraction ( Netrin-Frazzled ) and repulsion ( Netrin-Unc5 , Slit-Robo; reviewed by [32] ) and to gate midline crossing of dendrites in the Drosophila embryo , larva , and adult [28] , [33] , [34] . We focused on three motorneurons , each representing one of the three principal classes of dendritic morphology and medio-lateral territories: i ) MN-VO4–6 has dendrites in the lateral , intermediate , and midline neuropile ( Figure 5A ) ; ii ) MN-LL1 has only lateral and intermediate dendrites ( Figure 5B ) ; and iii ) MN-DA3 dendrites are located in the lateral neuropile ( Figure 5C ) . These neurons meet two additional criteria: first , their axons do not traverse the dorsal neuropile , so that one can clearly differentiate between dendrite morphogenesis and axonal ( collateral ) outgrowth; secondly , they can be manipulated genetically with great specificity using the CQ-GAL4 expression line . Use of this expression line does not obviously interfere with motor axon pathfinding or target recognition , as assayed by differential labelling of multiple motorneurons in a segment and the presence and distribution of neuromuscular junctions in the periphery ( unpublished data ) . In the first instance , we investigated dendritic targeting to the midline territory . We tested if exclusion of MN-LL1 and MN-DA3 dendrites from the midline neuropile was implemented by the presence of Robo in these cells . To this end we down-regulated Robo in MN-LL1 and MN-DA3 through targeted expression of UAS-comm [35] , [36] and found this to induce ectopic dendritic innervation of the midline neuropile ( Figure 5E , 5F , and 5Q; 100% penetrance for MN-LL1 , n = 14; 62 . 5% for MN-DA3 , n = 8 ) . Loss of Robo in embryos entirely mutant for robo ( robo1/roboGA1112 ) also generates a comparable ectopic midline innervation phenotype ( Figure 5H , 5I , and 5Q; 100% penetrance for MN-LL1 , n = 4; 89% for MN-DA3 , n = 9 ) , and this can be rescued by reinstating robo selectively in MN-LL1 and MN-DA3 using CQ-GAL4 ( Figure 5K , 5L , and 5Q ) ( rescue efficiency: 75% for MN-LL1 , n = 8; 100% for MN-DA3 , n = 4 ) . Conversely , we find that expression of the Slit receptor Robo in MN-VO4–6 can abolish dendritic targeting to the midline ( 20% penetrance , n = 15 ) ( Figure 5D and 5Q ) , a phenotype that we have never observed in the wild-type at 18 . 5 h AEL ( n = 17 ) . We suspect that the low penetrance in this particular case is due to low GAL4 activity in MN-VO4–6 at the time when its dendrites first explore the midline , and perhaps endogenous comm expression , which would normally permit dendritic growth to the midline and antagonize the effects of ectopically expressed Robo . In embryos mutant for the Netrin receptor Frazzled ( fra3/fra4 ) , MN-VO4–6 shows a comparable phenotype albeit at high penetrance ( as does MN-VO4/5; 100% penetrance , n = 11 for these two cells ) , and this can be rescued by reinstating Frazzled in MN-VO4–6 ( 86% penetrance , n = 7 ) ( Figure 5G , 5J , and 5Q ) . These results show that cell-autonomous expression of guidance cue receptors is necessary ( for Frazzled ) and sufficient ( for Robo ) to gate the growth of motorneuron dendrites to the midline neuropile . The data further suggest that dendrite growth to the midline requires not only a lack ( or low levels ) of Robo-mediated repulsion but also attraction mediated by Frazzled . We confirmed an absolute requirement for Frazzled . MN-DA3 and MN-LL1 dendrites fail to innervate the midline neuropile upon down-regulation of Robo when Frazzled is also absent ( 100% penetrance , n = 10; Figure 5M , 5N , and 5Q ) but ectopically target the midline when Frazzled expression is selectively reinstated in these cells ( 66% penetrance , n = 18; Figure 5O , 5P , and 5Q ) . We then examined how the distinction between the intermediate and lateral dendritic territories is specified at a distance ( approx . 5–10 µm ) from the ventral midline . For instance , the distinguishing feature between MN-DA3 and MN-LL1 is that MN-LL1 has an additional dendritic sub-arbor in the intermediate neuropile ( Figure 6A and 6A′; white curved arrow ) . To test whether Robo-Slit signalling in dendrites might also define this distinction between intermediate and lateral neuropile territories , we expressed Robo selectively in MN-LL1 and MN-DA3 . We find that increasing the levels of Robo in MN-LL1 reliably converts its dendritic tree to a MN-DA3-like morphology in 15/19 cases ( Figure 6B , 6B′ , and 6G ) . For MN-DA3 , this manipulation enhances the characteristic lateral confinement of its dendrites in 8/10 cases ( Figure 6G ) . We next tested the role of Frazzled-mediated attraction in generating the distinction between intermediate ( MN-LL1 ) and lateral ( MN-DA3 ) dendritic territories . To this end we removed , reinstated , and overexpressed frazzled in MN-LL1 and MN-DA3 . In fra3/fra4 mutant embryos , 63% of MN-LL1 dendritic arbors lack the normally pronounced intermediate dendritic arborisation ( n = 20; Figure 6C , 6C′ , and 6G ) , while targeting of MN-DA3 dendrites is not significantly affected ( n = 8; Figure 6G ) . Reinstating frazzled selectively in MN-LL1 in fra3/fra4 mutant embryos efficiently rescues dendritic targeting to the intermediate neuropile ( n = 11 ) . Moreover , this manipulation leads to a greater proportion of dendritic branches innervating the intermediate neuropile ( Figure 6D and 6D′; black curved arrows ) , as does overexpression of Frazzled in an otherwise wild-type background ( 57% of cases , n = 14; Figure 6E and 6E′; black curved arrows ) . For MN-DA3 expression of UAS-fra leads to ectopic innervation of the intermediate neuropile in 50% of cases , converting the dendritic arbor to a MN-LL1-like morphology ( n = 18; Figure 6F and 6F′; white curved arrows ) . Frazzled overexpression in MN-LL1 or MN-DA3 never led to ectopic midline targeting of dendrites ( n = 44 ) . Last , we tested the requirement for Robo and Frazzled signalling in motorneurons for setting up dendritic medio-lateral territories at an earlier stage , when dendritic domains first become recognizably distinct [3] . At 15 h AEL we find the same requirement as at 18 . 5 h AEL for the combinatorial action of Robo and Frazzled in motorneurons in the three dendritic territories with respect to the ventral midline ( Figure S3 ) . These results suggest that Robo and Frazzled are the key factors , whose relative levels in motorneurons determine the distinction between lateral and intermediate dendritic territories . Lateral confinement of dendrites ( e . g . , MN-DA3 ) can be achieved either by high levels of Robo and/or low levels of Frazzled expression . Targeting to the intermediate ( but not midline ) neuropile ( e . g . , MN-LL1 ) requires relatively high levels of Frazzled and low Robo activity . Our data show that Frazzled is absolutely required for dendritic growth to the midline . However , other Netrin receptors , such as Unc5 [37] and Dscam [38] , [39] , have been identified , as well as Netrin independent midline guidance systems [40] , [41] . To determine whether Frazzled is the main Netrin receptor for dendritic targeting and Netrin the sole attractant , we asked if loss of Frazzled produced the same dendritic phenotype as the loss of Netrin . This is indeed the case . In netABΔ mutants [40] MN-VO4–6 dendrites fail to target the midline neuropile ( 100% penetrance , n = 6 ) precisely as in fra3/fra4 mutants ( 100% penetrance , n = 4; MN-VO4/5 has the same mutant phenotype , n = 7 ) ( Figure 7A , 7C , and 7E ) . Similarly , MN-LL1 has a clearly reduced innervation of the intermediate neuropile in 63%–64% of cases in both netABΔ and fra3/fra4 mutant embryos ( n = 14 for netABΔ; n = 20 for fra3/fra4 ) ( Figure 7B , 7D , and 7F–7H ) . These observations suggest that in the embryonic nerve cord attraction of motorneuron dendrites to the ventral midline is mediated primarily , if not exclusively , by a Frazzled-containing receptor complex in response to Netrin . At the same time , we cannot entirely rule out that other ligand/receptor pairs might also contribute , though in more subtle ways , to positioning dendrites to the intermediate or midline neuropile . Previous studies have shown that Frazzled/DCC-Netrin signalling can promote axonal growth [42] , [43] . Others have implicated Robo-Slit signalling in regulating axonal and dendritic branching , the extension of axons , and the formation of dendrites [33] , [44]–[47] . We therefore wanted to know how Frazzled and Robo signalling affects the growth and branching of motorneuron dendrites as it regulates their distribution in the neuropile . To this end , we quantified [48] , [49] overall dendritic lengths and number of tips of MN-DA3 and MN-LL1 arbors under different experimental conditions ( Figure 8 ) . We find that the wild-type MN-DA3 and MN-LL1 reproducibly generate dendritic arbors with characteristically different total lengths ( MN-DA3: 221 . 7 µm±47 . 7 versus MN-LL1: 183 . 1 µm±35 . 7; t test: p = 0 . 02 ) and tip numbers ( MN-DA3: 53 . 0±10 . 0 versus MN-LL1: 42 . 8±7 . 6; t test: p = 0 . 005 ) . Cell-specific loss- and gain-of-function of Robo ( UAS-comm and UAS-robo ) as well as overexpression of Frazzled reproducibly generates clear dendritic targeting phenotypes . However , these manipulations do not lead to statistically significant changes in overall dendritic length or tip number ( Figure 8G ) . This indicates that Robo and Frazzled regulate the positioning of dendritic trees without noticeably affecting overall growth and branching . To test this hypothesis we focused on MN-LL1 and quantified the lengths of dendritic arbors located in the lateral neuropile as a percentage of total arbor length . Indeed , we find that changing the levels of Robo changes the proportion of the MN-LL1 dendritic tree that is put into the lateral neuropile . In the wild-type ( n = 15 ) , the average proportion of the dendritic tree in the lateral domain is 67 . 0%±4 . 3% ( 122 . 4 µm±24 . 2 ) of total tree length ( 183 . 1 µm±35 . 7 ) ( Figure 9B and 9D ) . Down-regulation of Robo by UAS-comm expression ( n = 13 ) induces ectopic dendritic growth towards the midline with a concomitant reduction of the lateral arbor to 52 . 5%±9 . 5% ( 101 . 6 µm±29 . 0; total length: 193 . 3 µm±39 . 3 ) ( Figure 9A and 9D ) . Conversely , expression of UAS-robo ( n = 13 ) reduces innervation of the intermediate neuropile and leads to a greater proportion of the arbor to be located in the lateral territory , namely 93 . 8%±7 . 3% ( 156 . 2 µm±32 . 1 ) of the total length ( 166 . 3 µm±29 . 8 ) ( Figure 9C and 9D ) . These quantifications suggest that , at least in the embryo , central neurons move towards generating a cell type–specific amount of dendritic length with a particular frequency of branching events . Guidance cue receptors act to distribute “available” dendrites , probably by locally modulating the rate of growth and/or stability of individual branches . At the same time , our data point to the existence of mechanisms that integrate such local changes across the entire arbor , since we do not observe statistically significant changes in overall tree length under different conditions . Finally , we asked what the significance might be of partitioning the neuropile into distinct dendritic domains . It is reasonable to suppose that muscles of similar position and orientation exert related functions , and so might operate in concert during locomotion . The myotopic segregation of motorneuron dendrites might therefore reflect differences in connectivity . Since individual presynaptic partner neurons have not yet been identified , we sought to address this issue by asking whether motorneuron dendrites targeted to lateral , intermediate , or midline territories received presynaptic contacts . We visualised presynaptic active zones of cholinergic interneurons with ChaB19/7 . 4-GAL4 driving UAS-bruchpilot-mRFP and labelled dendritic trees of MN-DA3 ( n = 5 ) , MN-LL1 ( n = 6 ) , and VO4/5-MN ( n = 4 ) with DiD/DiO in newly hatched larvae ( 21 h AEL ) . Using custom-made image analysis software [25] , [48] , [50] we reconstructed the dendritic arbors of motorneurons and assayed for putative presynaptic specialisations on these ( based on apposition within light microscopic resolution , approximately a 400 nm radius of the reconstructed dendrites ) . We find putative presynaptic sites on dendrites in the lateral , intermediate , and midline neuropile , though at present we cannot determine if these actually represent type-specific patterns of connectivity ( Figure 10 ) . In the light of reports that suggest that different motorneurons in the Drosophila embryo receive different inputs [22] , we interpret these observations as an indication that the segregation of motorneuron dendrites into distinct myotopic domains might be an underlying feature , or perhaps mechanism , of motorneuron class-specific patterns of connectivity .
Neural maps are manifestations of an organisational strategy commonly used by nervous systems to order synaptic connections . The view of these maps has been largely axonocentric and focused on sensory systems , though recent studies have challenged the notion of dendrites as a “passive” party in arranging the distribution of connections [13] , [16] , [51] , [52] . Here , we have demonstrated that motorneuron dendrites generate a neural , myotopic map in a motor system and that this manifest regularity can form independently of its presynaptic partner terminals . An essential feature of neural maps is the spatial segregation of synaptic connections . In the Drosophila embryonic nerve cord , there is some overlap between dendritic domains in the antero-posterior neuropile axis . Overlap of dendritic territories is also evident in the medio-lateral dimension , since all motorneurons have arborisations in the lateral neuropile , though distinctions arise by virtue of dendrites in additional intermediate and medial neuropile regions . The combination of myotopic mapping in both dimensions may serve to maximise the segregation between dendrites of different motorneuron groups . For example , the dendritic domain of motorneurons with dorsal targets differs from the territory innervated by ventrally projecting motorneurons in the antero-posterior location and the medio-lateral extent . Myotopic mapping in two dimensions could also provide a degree of flexibility that could facilitate wiring up in a combinatorial fashion . For instance , muscle LL1 lies at the interface between the dorsal and ventral muscle field; its motorneuron , MN-LL1 , has one part of its dendritic arbor in the lateral domain that is characteristic for dorsally projecting motorneurons , while the other part of the dendritic tree innervates the intermediate neuropile precisely where ventrally projecting motorneurons put their dendrites . Myotopic dendritic maps might constitute a general organisational principle in motor systems . In insects , a comparable system of organisation has now been demonstrated also for the adult motor system of Drosophila ( see companion study by Brierley and colleagues [53] and [54] ) and a degree of topographic organisation had previously been suggested for the dendrites of motorneurons that innervate the body wall muscles in the moth Manduca sexta [55] . In vertebrates too , there is evidence that different motor pools elaborate their dendrites in distinct regions of the spinal cord in chick , turtle , and mouse [4]–[6] . Moreover , elegant work in the mouse has shown that differences in dendritic territories correlate with and may determine the specificity of proprioceptive afferent inputs [6] . The neural map that we have characterised here is composed of three morphological classes of motorneurons with dendrites innervating either i ) the lateral or ii ) the lateral and intermediate or iii ) the lateral , intermediate , and medial/midline neuropile ( Figure 1 ) . We have shown that the motorneuron dendrites are targeted to these medio-lateral territories by the combinatorial , cell-autonomous actions of the midline guidance cue receptors Robo and Frazzled . The formation of dendritic territories by directed , targeted growth appears to be an important mechanism that may be more widespread than previously anticipated [56] , though the underlying mechanisms may vary . Global patterning cues have been implicated in the vertebrate cortex ( Sema3A [29] ) . In the zebrafish retina , live imaging has shown that retinal ganglion cells put their dendrites into specific strata of the inner plexiform layer , but the roles of guidance cues and interactions with partner ( amacrine ) cells have not yet been studied [13] . Slit/Robo and Netrin/Frazzled mediated gating of dendritic midline crossing has been previously documented in Drosophila embryos [28] and zebrafish [57] . Here , we demonstrate for the first time that dendrites are targeted to distinct medio-lateral territories by the combinatorial , opposing actions of Robo and Frazzled and that this is the main mechanism underlying the formation of the myotopic map . Strikingly , the same signalling pathways also regulate dendritic targeting of adult motorneurons in Drosophila , suggesting this to be a conserved mechanism ( see companion paper [53] ) . Robo gates midline crossing of dendrites and in addition , at progressively higher signalling levels , restricts dendritic targeting to intermediate and lateral territories . Frazzled , on the other hand , is required for targeting dendrites towards the midline into intermediate and medial territories . Our data argue that Frazzled is expressed by representatives of all three motorneuron types ( see also [58] ) . Recently , Yang and colleagues [59] showed that expression of frazzled leads to a concomitant transcriptional up-regulation of comm , thus linking Frazzled-mediated attraction to the midline with a decrease in Robo-mediated repulsion . While this has been demonstrated for midline crossing of axons in the Drosophila embryo , we found that , at least until 18 . 5 h AEL , expression of UAS-frazzled alone was not sufficient to induce midline crossing of dendrites in MN-LL1 and MN-DA3 ( see Figure 6 ) . It is conceivable that differences in expression levels and/or timing between CQ-GAL4 used here and egl-GAL4 used by Yang et al . might account for the differences in axonal and dendritic responses to UAS-frazzled expression . Moreover , the widespread expression of Frazzled in motorneurons and other cells in the CNS may point to additional functions , potentially synaptogenesis , as recently shown in C . elegans [60] , [61] . Strikingly , neither synaptic excitatory activity nor the presynaptic ( cholinergic ) partner terminals seem to be necessary for the formation of the map . The map is already evident by 15 h AEL , before motorneurons receive synaptic inputs ( Figure 2 ) . It also forms in the absence of acetylcholine , the main ( and at that stage probably exclusive ) neurotransmitter to which motorneurons respond ( Figure 3 ) [22] . Moreover , motorneuron dendrites innervate their characteristic dendritic domains when the cholinergic terminals have been displaced to outside the motor neuropile ( Figure 4 ) . However , interactions with presynaptic partners seem to contribute to its refinement . First , we find that dendritic mistargeting phenotypes show a greater degree of penetrance earlier ( 15 h AEL ) than later ( 18 . 5 h AEL ) in development ( Figure S3 ) . Secondly , when interactions with presynaptic partner terminals are reduced or absent , dendritic arbor size increases [25] and the distinction between dendritic territories is less evident than in controls ( Figure 4H–4J ) . Fine-tuning of terminal arbors and sets of connections through contact and activity-dependent mechanisms is a well-established feature of neural maps in sensory systems ( for review see [10] , [62] ) and our observations suggest that this may also apply to motor systems . The formation of the myotopic map is the product of dendritic targeting . It is therefore intimately linked with the question of how cell type–specific dendritic morphologies are specified . For instance , changing the balance between the Robo and Frazzled guidance receptors in motorneurons is sufficient to “convert” dendritic morphologies from one type to another ( Figures 5 and 6 ) . The importance of target territories for determining dendritic arbor morphology has recently been explored in a study of lobula plate tangential cells in the blowfly , where the distinguishing parameter between the dendritic trees of four functionally defined neurons were not growth or branching characteristics but the regions where neurons put their dendrites [63] . Because Slit/Robo and Netrin/Frazzled signalling have been reported to affect dendritic and axonal branching as well as axonal growth , respectively , we asked what the effect was on motorneuron dendrites of altered Robo and Frazzled levels [33] , [42]–[47] . We find that in the wild-type different motorneurons generate characteristically different amounts of dendritic length and numbers of branch points ( MN-DA1/aCC and MN-VO2/RP1 [25] , RP2 [34] , MN-DA3 and MN-LL1 , this study ) . In the Drosophila embryo and larva , Slit/Robo interactions have been suggested to promote the formation of dendrites and/or branching events [44] , [45] , similar to what had previously been shown for cultured vertebrate neurons [47] . Our data on embryonic motorneurons are not compatible with this interpretation . First , when altering the levels of Robo ( or Frazzled ) in individual motorneurons and mistargeting their dendrites , we could not detect statistically significant changes in total dendritic length or number of branch points . Instead , for MN-DA3 and MN-LL1 , we observed that dendritic arbors respond to changes in the expression levels of midline cue receptors by altering the amount of dendritic length distributed to the medial , intermediate , and lateral neuropile ( Figures 8 and 9 ) . Secondly , in nerve cords entirely mutant for the Slit receptor Robo we see an increase in dendrite branching at the midline ( Figure S4 ) . Our observations suggest that for Drosophila motorneurons Slit/Robo interactions negatively regulate the establishment and branching of dendrites and thus specify dendritic target territories by defining “exclusion” zones in the neuropile . The quantitative data from this and a companion study [53] suggest that dendritic morphology is the product of two intrinsic , genetically separable programmes: one that specifies the total dendritic length to be generated and the frequency of branching; the other implements the distribution of these dendrites in the target territory , presumably by locally modulating rates of extension , stabilisation , and retraction of branches in response to extrinsic signals . Observations from a previous study [34] and other systems , e . g . , insect sensory neurons [64] and vertebrate cortical neurons [65] , are compatible with this model . The question of how neural circuits are generated remains at the heart of developmental neurobiology . At one extreme , one could envisage that every synapse was genetically specified , the product of an exquisitely choreographed sequence of cell-cell interactions . At the other extreme , neural networks might assemble through random cell-cell interactions and feedback processes enabling functional validation . The latter view supposes that neurons inherently generate polarised processes , have a high propensity to form synapses , and arrive at a favourable activity state through homeostatic mechanisms . Current evidence suggests that , at least for most systems , circuits form by a combination of genetic specification and the capacity to self-organise ( for reviews see [10] , [62] , [66] ) . In this study we have demonstrated that the postsynaptic structures of motorneurons , the dendrites , form a neural map . We have also shown that dendrites are closely apposed to cholinergic presynaptic specialisations in their target territories ( Figure 10 ) , suggesting that the segregation of dendrites may be a mechanism that facilitates the formation of specific sets of connections . Strikingly , this map of postsynaptic dendrites appears to be “hard-wired” in that it can form independently of its presynaptic partners and it is generated in response to a third party , the midline guidance cues Slit and Netrin ( see also companion paper [53] ) . A comparable example is the Drosophila antennal lobe , where projection neurons form a neural map independently of their presynaptic olfactory receptor neurons , though in this sensory system the nature and source of the cue ( s ) remain to be determined [15] , [16] . With this study we complement previous work that demonstrated the positioning of presynaptic axon terminals by midline cues , also independently of their synaptic partners [17] , [18] . Together , these results suggest that global patterning cues set up the functional architecture of the nervous system by independently directing pre- and postsynaptic partner terminals towards common “meeting” areas . Clearly , such global guidance systems deliver a relatively coarse level of specificity and there is ample evidence for the existence of codes of cell-adhesion molecules and local receptor-ligand interactions capable of conferring a high degree of synaptic specificity [67]–[75] . Therefore , one has to ask what the contribution is of global partitioning systems in establishing patterns of connections that lead to a functional neural network . A recent study in the Xenopus tadpole spinal cord has addressed this issue . Conducting patch clamp recordings from pairs of neurons , Li and colleagues [76] found that the actual pattern of connections in the motor circuit reveals a remarkable lack of specificity . Furthermore , the segregation of axons and dendrites into a few broad domains appears to be sufficient to generate the connections that do form and to enable the emergence of a functional network [76] . The implication is that neurons might be intrinsically promiscuous and targeting nerve terminals to distinct territories by global patterning cues , as we have shown here , is important to restrict this synaptogenic potential and thereby confer a degree of specificity that is necessary for the emergence of network function .
The following fly stocks were used: Oregon-R , Fasciclin2-GFP on X ( always used in a heterozygous condition in females [w− , Fasciclin2-GFP/w−] [77] ) , amorphic fra3/fra4 [78] , amorphic netABΔ [40] , UAS-framyc on III [78] , amorphic robo1/roboGA1112 [79] , UAS-robo two insertions 2B , 3D on III [79] , ChaB19/7 . 4-GAL4 on II [80] , UAS-commissureless on X [36] , UAS-RoboexFrain-myc and UAS-FraexRoboin-myc [31] , UAS-bruchpilot-mRFP on III ( generously provided by S . Mertel and S . Sigrist ) , chal13 [30] , and UAS-myr-mRFP1 ( generated by Henry Cheng , obtained from the Bloomington Stock Center ) . Lethal mutations/insertions were kept over FM7 , CyO , and TM3 balancer chromosomes that are additionally marked with Kr-GAL4 , UAS-GFP [81] . Selective expression in MN-DA3 , MN-LL1 , and MN-VO4–6 was achieved using CQ-GAL4 with insertions on chromosomes II and III . This line expresses in the five CQ/U-motorneurons MN-DO1 , MN-DO2 , MN-DA2 , MN-DA3 , and MN-LL1 [3] as well as MN-VO4–6 in approximately 30% of half segments , always confirmed in expression experiments by the presence of an additional reporter , UAS-bruchpilot-mRFP , at respective NMJs . Rarely , up to eight cells per half segment can be seen expressing with CQ-GAL4 , indicating potential expression in one or two interneurons , though we have no evidence of these having terminations in the motor neuropile . robo mutant: w− , Fas2GFP/w−; robo1/roboGA1112 Down-regulation of Robo: w− , Fas2GFP/w− , UAS-comm; CQ-GAL4/+; CQ-GAL4/+ Robo expression: w− , Fas2GFP/w−; CQ-GAL4/+; CQ-GAL4 , UAS-brpRFP/UAS-robo2B , 3D ( for MN-VO4–6 ) w− , Fas2GFP/w−; CQ-GAL4/+; CQ-GAL4; UAS- robo2B , 3D ( for MN-LL1 and MN-DA3 ) robo mutant with cell-autonomous rescue: w− , Fas2GFP/w−; robo1/roboGA1112; CQ-GAL4; UAS-robo2B , 3D frazzled mutant: w− , Fas2GFP/w−; fra3/fra4 , CQ-GAL4; CQ-GAL4/+ Frazzled expression: w2 , Fas2GFP/w2; CQ-GAL4/+; CQ-GAL4 / UAS-fra-myc frazzled mutant with cell-autonomous rescue of frazzled: w− , Fas2GFP/w−; fra4 , CQ-GAL4/fra3; CQ-GAL4 , UAS-brpRFP/UAS-fra-myc ( for MN-VO4–6 ) w− , Fas2GFP/w−; fra4 , CQ-GAL4/fra3; CQ-GAL4/UAS-fra-myc ( for MN-LL1 ) Down-regulation of Robo in a frazzled mutant: w− , Fas2GFP/w− , UAS-comm; fra4 , CQ-GAL4/fra3; CQ-GAL4/+ Down-regulation of Robo in a frazzled mutant with cell-autonomous rescue of frazzled: w− , Fas2GFP/w− , UAS-comm; fra4 , CQ-GAL4/fra3; CQ-GAL4/UAS-fra-myc netrin mutant: w− , Fas2GFP , netABΔ/y Labelling of cholinergic presynaptic sites: w−; ChaB19/7 . 4-GAL4/+; UAS-brpRFP/+ or UAS-brpRFP Labelling and shifting of cholinergic terminals: w− , Fas2GFP/w−; ChaB19/7 . 4-GAL4 , UAS-myr-mRFP/+; +/+ w− , Fas2GFP/w−; ChaB19/7 . 4-GAL4 , UAS-myr-mRFP/UAS-Fraex-Roboin-myc; +/+ w− , Fas2GFP/w−; ChaB19/7 . 4-GAL4 , UAS-myr-mRFP/UAS-Roboex-Frain-myc; +/+ Embryos 15 h AEL were dissected as described in [19] , though without collagenase treatment; embryos 18 . 5 h AEL ( onset of trachea filling ) were dissected as described in [21] . Embryos were then fixed with 3 . 7% formaldehyde in saline for 2 . 5 min and rinsed . Retrograde labellings were done as described by [19] , and in addition neuromuscular junctions were visualised by FITC-conjugated anti-horseradish peroxidase incubation for ∼3–6 min ( Jackson ImmunoResearch , West Grove , PA , United States; 1∶50 dilution in saline ) , followed by saline washes . Neuro-DiO ( Biotium ) , DiD , and DiI ( Molecular Probes , Eugene , OR , United States ) were used at 2 mg/ml , 2 mg/ml , and 4 mg/ml , respectively , dissolved in vegetable oil . Anterograde Lucifer Yellow ( Invitrogen ) labellings were done as in [18] . Labelled neurons were imaged with a Yokagawa CSU-22 confocal field scanner mounted on an Olympus BX51WI Spinning Disc microscope , using a 63×/1 . 2 NA ( Olympus ) water immersion objective . Image z-stacks were acquired using MetaMorph software ( Molecular Devices ) and processed using ImageJ 1 . 39 s software ( U . S . National Institutes of Health , Bethesda , MD , USA , http://rsb . info . nih . gov/ij/ ) . Cumulative dendrite plots: z-projections of dendritic trees were scaled and aligned isometrically onto a common reference grid with Photoshop CS2 ( Adobe Systems , San Jose , CA , USA ) , using the position of the motorneuron axon in one channel as the antero-posterior and the outer and inner Fasciclin2-positive axon tracts as medio-lateral reference points . Silhouettes ( intensity information was discarded ) of dendritic trees of each experimental condition were summed using ImageJ . Dual channel confocal image stacks were generated ( z-step size: 300 nm ) of Neuro-DiO labelled dendrites and a presynaptic marker expressed in cholinergic neurons ( w−; ChaB19/7 . 4-GAL4/+; UAS-bruchpilot-mRFP/+ or w−; ChaB19/7 . 4-GAL4/+; UAS-bruchpilot-mRFP/UAS-bruchpilot-mRFP ) . Dendrites were reconstructed using a custom-made module [48]–[50] for AMIRA software ( version 4 . 1 ) . Relative probabilities of synaptic contact on reconstructed dendrites were calculated based on both the distance and signal intensity of presynaptic mRFP-puncta , and represented by a colour code , ranging from blue ( indicating a relatively low probability of contact ) to red ( <400 nm distance , indicating a relatively high probability of synaptic contact ) . Geometrical data from dendritic “skeleton trees” were exported from AMIRA as csv-files , analysed , and plotted using “R project” ( R Foundation for Statistical Computing , Vienna , Austria , 2005 . http://R-project . org ) . Data were analysed statistically using the Shapiro-Wilk test to assess for normality followed by a Student's t test or a Wilcoxon rank-sum test as appropriate . | How neural networks governing locomotion are organised is less well understood than those governing sensory systems . In the Drosophila embryo dendrites form the input structures of motorneurons , and are arranged along the anterior-posterior axis in the central nervous system to reflect the distribution of body wall muscles in the periphery . Here we examine how a motorneuron dendritic map develops . We find that motorneurons target their dendrites also to distinct medio-lateral territories . This map appears to be “hard-wired” in that its formation does not require synaptic input or the proper positioning of partner terminals . Instead , dendritic targeting is determined by the responsiveness of individual motorneurons to midline guidance cues , mediated by the Slit receptor Robo and the Netrin receptor Frazzled . These findings complement and mirror similar results by others on the positioning of presynaptic axon terminals , and together they suggest a central role for global guidance cues in generating connectivity by delivering partner terminals independently of one another to common “meeting regions . ” | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] | [
"neuroscience/motor",
"systems",
"neuroscience/neurodevelopment",
"developmental",
"biology/neurodevelopment"
] | 2009 | Midline Signalling Systems Direct the Formation of a Neural Map by Dendritic Targeting in the Drosophila Motor System |
The human oxidative phosphorylation ( OxPhos ) system consists of approximately 90 proteins encoded by nuclear and mitochondrial genomes and serves as the primary cellular pathway for ATP biosynthesis . While the core protein machinery for OxPhos is well characterized , many of its assembly , maturation , and regulatory factors remain unknown . We exploited the tight transcriptional control of the genes encoding the core OxPhos machinery to identify novel regulators . We developed a computational procedure , which we call expression screening , which integrates information from thousands of microarray data sets in a principled manner to identify genes that are consistently co-expressed with a target pathway across biological contexts . We applied expression screening to predict dozens of novel regulators of OxPhos . For two candidate genes , CHCHD2 and SLIRP , we show that silencing with RNAi results in destabilization of OxPhos complexes and a marked loss of OxPhos enzymatic activity . Moreover , we show that SLIRP plays an essential role in maintaining mitochondrial-localized mRNA transcripts that encode OxPhos protein subunits . Our findings provide a catalogue of potential novel OxPhos regulators that advance our understanding of the coordination between nuclear and mitochondrial genomes for the regulation of cellular energy metabolism .
Mitochondrial oxidative phosphorylation ( OxPhos ) is central to energy homeostasis and human health by serving as the cell's primary generator of ATP . The core machinery underlying OxPhos consists of approximately 90 distinct protein subunits that form five complexes residing in the inner mitochondrial membrane . Complexes I through IV comprise the oxygen-dependent electron transport chain responsible for driving the generation of ATP by complex V . OxPhos is the only process in the mammalian cell under dual genetic control: thirteen essential structural subunits are encoded by mitochondrial DNA ( mtDNA ) while remaining subunits are encoded by nuclear genes , and are imported into mitochondria [1] . The biogenesis of OxPhos requires many accessory factors responsible for replicating mtDNA as well as transcribing and translating the mitochondrial mRNAs ( mtRNA ) [2] , [3] . Furthermore , the mtDNA-encoded subunits must be coordinately assembled with the nuclear-encoded subunits and metal co-factors to form functional complexes , a process likely requiring far more assembly factors than are currently known [4] . Dysfunction in any of these processes or in the OxPhos machinery itself may result in a respiratory chain disorder , a large class of inborn errors of metabolism [5] . For approximately 50% of patients with respiratory chain disorders , the underlying genetic defect remains unknown , despite excluding obvious members of the OxPhos pathway [4] , [6]–[8] . Many of these disorders are likely due to genetic defects in currently uncharacterized OxPhos assembly or regulatory factors . The OxPhos structural subunits exhibit tight transcriptional regulation that offers a strategy for identifying its non-structural regulators based upon shared patterns of co-expression in microarray experiments [9] , [10] . In fact , our laboratory used this approach to identify the gene LRPPRC , which encodes a critical regulator of mtRNA and when mutated is the underlying cause of a respiratory chain disorder called Leigh Syndrome French-Canadian variant [11] . However , while successful in identifying LRPPRC , this previous analysis used only one data set interrogating tissue-specific gene expression [12] , [13] . Such co-expression analyses that rely upon individual contexts are not ideal for functional prediction because they are subject to inherent limitations of microarray experiments including technical artifacts , experimental bias and real but confounding correlations with functionally distinct pathways [14] . To overcome these limitations and to generalize our previous approach , we reasoned that large-scale integration across many independent microarray experiments , each surveying a different biological context , would help distinguish genuine co-regulation from random co-expression by identifying genes that consistently co-express with OxPhos . In the yeast Saccharomyces cerevisiae , several groups have performed such expression data integration studies to predict protein function [15]–[17] . With the recent availability of large repositories of mammalian microarray data , it is now possible to apply similar approaches to functionally classify uncharacterized human proteins [18] , [19] . Studying mammalian data is especially important for OxPhos given that the mammalian OxPhos pathway differs significantly from the yeast counterpart . For example , S . cerevisiae lacks a proton pumping complex I , the largest OxPhos complex in human cells consisting of forty-five distinct protein subunits [20] and a common target of respiratory chain disease [21] , [22] . Furthermore , mammalian mtDNA is circular , whereas yeast mtDNA can form linear concatemers [23] . Moreover , mammalian mtRNA processing differs markedly from S . cerevisaie as mammalian mtRNA does not contain introns and is polyadenylated [24] . In the present paper , we introduce a computational methodology , called “expression screening” , that takes advantage of the growing wealth of freely available mammalian microarray data to search for genes that exhibit consistent co-expression with a given “query” gene set . Applying this procedure to the mammalian OxPhos pathway revealed a number of putative regulators that now emerge as attractive candidate genes for OxPhos disorders . We experimentally validated two genes , CHCHD2 ( coiled-coil-helix-coiled-coil-helix domain containing 2 ) and SLIRP ( SRA-stem loop interacting RNA-binding protein; also known as C14orf156 ) as essential for OxPhos function . We further characterized SLIRP as a RNA-binding domain containing protein necessary for the maintenance of mtRNA protein-encoding transcripts and whose robust co-expression with the nuclear OxPhos subunits provides a putative regulatory link between nuclear and mitochondrial gene expression .
We next applied expression screening to the OxPhos pathway using the 1427 microarray dataset compendium , and a manually curated gene set of nuclear-encoded structural OxPhos subunits ( Figure 2A , Table S4 ) . We excluded the mtDNA-encoded subunits from the query set since these were not well measured by the Affymetrix platforms . The resulting co-expression matrix ( Figure 2B ) reveals the robust coordination of OxPhos gene expression in a large variety of biological contexts . The OxPhos gene set exhibits robust intra-correlation ( weight wd>0 . 75 ) in nearly 10% of microarray datasets present in the compendium ( Table S1 ) . The data set weights enable us to spotlight biological contexts in the compendium for which the modulation of OxPhos gene expression may play an important role . Experiments with large weights include expected conditions such as exercise ( GSE1659 ) , Alzheimer's disease ( GSE5281 ) and Pgc1α over-expression ( GSE4330 ) as well as lesser-studied contexts including down-regulation of OxPhos followed by recovery during time-courses of skeletal muscle regeneration ( GSE469 , GSE5413 ) . We applied the data integration procedure to identify genes that are consistently co-expressed with OxPhos in the microarray compendium ( Table S5 ) . As with the cholesterol biosynthesis pathway , data integration better predicts known genes involved in the OxPhos pathway when compared to the most predictive data set alone ( Figure 2C ) . At a specificity of 99 . 4% , we were able to recover 85% of the OxPhos pathway ( Figure 2C ) . The integration procedure also lessens confounding correlations with functionally distinct pathways . For example , OxPhos is frequently co-expressed with other genes encoding mitochondrial proteins during mitochondrial biogenesis and turnover , regardless of their specific role in oxidative phosphorylation [30]–[32] . Additionally , OxPhos gene expression may correlate with the expression of other functionally distinct “house-keeping” pathways , especially the cytosolic ribosome , since genes involved in both pathways share a similar set of conserved promoter elements and are controlled by an over-lapping set of transcriptional regulators [33] , [34] . In agreement with these findings , we observed significant co-expression ( median integrated probability pg = 0 . 70 ) of the cytosolic ribosome with the OxPhos subunits ( Figure 2D ) . However , integrating co-expression across all data sets in the microarray compendium clearly distinguished the OxPhos pathway from other mitochondrial genes and components of the cytosolic ribosome , demonstrating the specificity of expression screening ( Figure 2D ) . We next examined the non-OxPhos genes exhibiting the highest co-expression scores . To ensure that co-expression is conserved among mammals , we required that a gene is co-expressed with OxPhos when analyzing human and mouse microarray datasets independently ( pg>0 . 70 in both species ) . The top 20 non-OxPhos genes meeting this criterion are shown in Figure 3 . Several of the non-OxPhos genes listed in Figure 3 have known metabolic roles in oxidative metabolism such as genes encoding Kreb's cycle enzymes , MDH2 and SUCLG1 , as well as several mitochondrial ribosomal subunits necessary for translation of the OxPhos subunits encoded by mtDNA . Other high-scoring genes have never been functionally associated with OxPhos and most lack orthologues in S . cerevisiae . It is notable that recent mass spectrometry studies of highly purified mammalian mitochondria have localized every protein present in Figure 3 to the mitochondria with the exception of HINT1 , TCEB2 and MDH1 , which are primarily cytosolic proteins [35]–[37] . Recently , two candidates identified by our expression screen , C14orf2 and USMG5 , have been co-purified with complex V , having been previously missed in purifications of OxPhos due to their small size and biochemical properties ( ∼7 kD ) [38] , [39] . While the functions of these two proteins are still unknown , their physical association with complex V further supports the specificity of the expression screening results for identifying OxPhos-related genes . Interestingly , two other uncharacterized proteins presented in Figure 3 , C1orf151 and C12orf62 , are also less than 10 kD in size ( 8 . 6 kD and 6 . 4 kD , respectively ) and contain a single-pass transmembrane domain similar to C14orf2 and USMG5 . These molecular similarities suggest that C1orf151 and C12orf62 may also physically associate with OxPhos . To validate the results from the OxPhos expression screen , we selected five functionally uncharacterized mitochondrial candidates from Figure 3 for which we could obtain reliable shRNA reagents to experimentally test their role in OxPhos function ( C14orf2 , USMG5 , CHCHD2 , SLIRP and PARK7 ) . For each of the five candidate genes , we identified at least two independent , non-toxic shRNAs that deplete mRNA abundance by more than 85% ( Figure S2 ) . We were unable to obtain high quality shRNA reagents for other candidates including C1orf151 , C19orf20 and C12orf62 . We first silenced each candidate gene in immortalized human fibroblasts and measured the live-cell oxygen consumption rate ( OCR ) as a general parameter of basal OxPhos activity . Silencing of two candidates , CHCHD2 and SLIRP , significantly reduced cellular OCR by approximately 40% compared to control cells ( P< . 05; Figure 4A ) . Additionally , a single shRNA targeting C14orf2 reduced OCR by 35% ( P< . 05 ) ; however , this result may be due to an off-target effect since a second hairpin targeting C14orf2 did not substantially affect OCR ( Figure 4A ) . Inherited or acquired mutations causing OxPhos dysfunction often destabilize or cause the misassembly of one or more of the five complexes comprising OxPhos . We therefore assessed whether knock-down of any of the five candidates affected complex stability by blotting for “labile” OxPhos subunits whose stability depends on their respective complex being properly assembled in the mitochondrial inner membrane ( Figure 4B ) . Again , we noted that knock-down of SLIRP and CHCHD2 clearly affected OxPhos as both dramatically reduced the abundance of the complex IV subunit , COX2 , and to a lesser extent , NDUFB8 , a component of complex I . To ensure that CHCHD2 and SLIRP are responsible for maintaining the activities of OxPhos complexes I and IV in native form , we measured the activity of immuno-captured preparations of these complexes ( Figure 4C and 4D ) [40] . Reducing the expression of both candidates reduced cellular CIV activity ( P< . 05 ) while only SLIRP significantly affected CI ( P< . 05 ) . The SLIRP protein contains an RNA-binding domain and was previously reported to associate with steroid receptor RNA activator ( SRA ) , a nuclear non-coding RNA , and thereby repress the ability of SRA to activate nuclear receptors [41] . However , SLIRP is predominantly mitochondrial [35] , [41] . Since the protein is localized to the mitochondria and is able to bind RNA , we hypothesized that it might affect OxPhos activity by directly modulating the level of mtRNA , either through expression , processing or stability of the mitochondrial transcripts . mtRNA is transcribed from mtDNA in two continuous poly-cistronic transcripts ( one from each mtDNA strand ) , which are subsequently processed to produce eleven OxPhos protein-encoding mRNAs , two ribosomal RNAs ( rRNA ) and a full complement of tRNAs . The processed mtRNAs are individually regulated by mtRNA stability factors , many of which remain to be identified [42] . To determine whether SLIRP acts in the mtRNA processing pathway , we designed a full panel of qPCR assays to measure the abundance of each protein-coding and ribosomal mtRNA transcript ( Table S5 ) . We again used shRNA to reduce SLIRP expression and measured the resulting effect on each mtRNA transcript . Knock-down of SLIRP significantly reduced the abundance of all eleven protein-encoding mtRNA transcripts ( Figure 5A ) , while the mtDNA copy-number was unaffected ( Figure 5B ) . The most pronounced mtRNA reduction was seen for transcripts encoding complex IV subunits as well as the bi-cistronic transcript encoding the ND4 and ND4L subunits of complex I , which is concordant with the specific complex I and IV biochemical defects shown in Figure 4 . The effect of SLIRP depletion upon mtRNA appears specific to the protein-encoding mtRNA transcripts since it did not affect the expression of the 12S or 16S mitochondrial rRNAs ( Figure 5A ) , even though these rRNAs are encoded on the same primary poly-cistronic transcript that contains all but one of the mitochondrial mRNAs . To assess whether this regulation of mtRNA by SLIRP is conserved among mammals , we also silenced the gene encoding the mouse ortholog of SLIRP in C2C12 myoblasts . We again observed down-regulation of all three complex IV-encoding mtRNAs ( Figure S3 ) . Since SLIRP is proposed to be alternatively localized to the nucleus , we wondered whether it might affect mtRNA expression indirectly by regulating the nuclear expression of known mtDNA transcription factors or mtRNA regulators . However , shRNA targeting SLIRP did not significantly alter the expression of known nuclear-encoded mtRNA regulators TFAM , TFB1M , and TFB2M , nor did it affect the expression of the nuclear-encoded OxPhos subunit UQCRC1 , further suggesting that SLIRP acts within the mitochondria to regulate mtRNA abundance ( Figure S4 ) . Finally , we investigated whether over-expression of SLIRP would be sufficient to boost mtRNA abundance in the cell . Over-expressing SLIRP for 48 hours did not alter mtRNA abundance , but over-expression did rescue the down-regulation of mtRNA resulting from knock-down of SLIRP in human cells . Besides demonstrating that the over-expression construct is functional and that the shRNA is on-target , this indicates that SLIRP is not a limiting factor for mtRNA abundance in wild-type cells ( Figure 5C ) . Since adequate expression of SLIRP is essential for maintaining mtRNA levels , we asked whether SLIRP is transcriptionally regulated in response to a depletion of mtRNA . We used ethidium bromide ( EtBr ) , a DNA-intercalating agent that is selectively absorbed by mitochondria and reduces mtDNA copy-number and mtRNA expression in the cell [43] . Following treatment with EtBr for four days , we did not observe any compensatory increase in SLIRP expression as mtDNA and mtRNA were depleted in a concentration-dependent manner ( Figure 6A ) . Surprisingly , however , we did observe a dramatic reduction in SLIRP at the protein level , in a manner depending on the concentration of EtBr ( Figure 6B ) . This suggests that the stability of SLIRP depends upon either mtDNA copy-number or mtRNA abundance . A similar phenomenon has been previously reported for TFAM , a critical regulator of both mitochondrial DNA and RNA [30] . TFAM coats the mtDNA to protect it from degradation but TFAM is also dependent upon mtDNA for its own protein stability [44] . We wondered whether SLIRP , being an RNA-binding protein , depends exclusively upon mtRNA for its stability . To assess whether mtRNA rather than mtDNA quantity is important for stabilizing SLIRP we used shRNA to deplete cells of LRPPRC , a mitochondrial protein necessary for maintaining mtRNA expression but not mtDNA copy-number [45] ( Figure 6C ) . We again observed a substantial drop in SLIRP protein abundance , likely indicating a mutual partnership between SLIRP and mtRNA where each is responsible for the other's stability within the mitochondria ( Figure 6D ) .
To predict novel OxPhos regulators we have developed a method called expression screening , which utilizes the inherent strong transcriptional co-expression of the known OxPhos structural subunits to identify other genes sharing similar expression profiles in a compendium of microarray data . While co-expression analysis alone cannot fully distinguish functionally relevant co-regulation from mere correlation , we have demonstrated that the integration of evidence from hundreds of biological contexts significantly enhances predictive power . In this manner , we were able to build a reliable classifier for membership in the OxPhos pathway that will be a useful resource for prioritizing candidate genes in patients with respiratory chain disorders . Expression screening predicted several functionally uncharacterized genes as novel regulators of the OxPhos pathway . Of these , CHCHD2 and SLIRP resulted in clear OxPhos deficits when targeted by shRNA . CHCHD2 is a member of a family of proteins containing the CHCH domain ( coiled-coil-helix-coiled-coil-helix; PFAM#06747 ) . Conserved cysteines within this motif have previously been implicated in metal coordination or transport suggesting that CHCHD2's role in stabilizing complex IV may be related to regulation of the complex IV copper centers [46] . Interestingly , the CHCH domain is also found in an OxPhos complex I subunit , NDUFA8 , and in another Complex IV assembly factor , COX19 [46] , [47] . Further study of this important domain should lend insight to the assembly and function of OxPhos . Silencing of some high-scoring expression screening candidates including USMG5 , C14orf2 and PARK7 did not result in an oxidative phenotype . These outcomes may reflect common caveats with shRNA experiments including insufficient protein knock-down , functional redundancy or lack of the proper experimental context . For example , our expression screen implicates the gene PARK7 , named for causing Parkinson's disease when mutated , as a key player in OxPhos biology . Currently , there is no established role for PARK7 in the OxPhos pathway [48] . While we did not observe an effect on basal oxygen consumption when perturbing PARK7 expression in our cell line , PARK7 protein may still be an important OxPhos regulator that acts in a context-dependent manner . Others have reported PARK7 to act as an antioxidant that scavenges mitochondrial radical oxygen species , a harmful by-product of an active OxPhos system [49] . Additionally , cells depleted of PARK7 are hyper-sensitive to rotenone treatment , a potent complex I inhibitor [50] . SLIRP is consistently co-expressed with the nuclear OxPhos machinery and regulates the abundance of the mitochondrial protein-encoding transcripts . These properties raise the interesting possibility that SLIRP is co-regulated with the nuclear OxPhos genes in order to coordinate nuclear and mitochondrial OxPhos gene expression . This phenomenon has also been previously reported for genes encoding the mtDNA transcription factors: TFAM , TFB1M and TFB2M [51] , [52] . In certain biological contexts , these genes have been noted to be co-expressed with nuclear OxPhos genes [51] , [52] . In our expression screen , we did observe co-expression of these factors with OxPhos in certain microarray experiments , but this co-expression was not frequent enough to generate a high score in the overall data integration . In contrast , SLIRP scored among the top 20 genes in the genome for its co-expression with OxPhos ( Figure 3 ) , strongly implicating a role for SLIRP in synchronizing nuclear and mitochondrial gene expression . The precise molecular mechanism by which SLIRP maintains mtRNA is not yet clear . To date , most studies of mtRNA maintenance has focused upon the core transcriptional machinery responsible for transcribing the primary poly-cistronic transcripts . This essential machinery includes the mitochondrial polymerase , POLRMT and its partner MRPL12 , as well as the transcription factors TFAM , TFB1M and TFB2M [3] . However , key factors in mammalian mtRNA post-transcriptional processing and stability remain unknown [24] . For example , a human mitochondrial poly ( A ) polymerase ( mtPAP ) has been recently identified [53] , but this protein does not contain an obvious RNA-binding domain , suggesting that it requires one or more currently unidentified RNA-binding partners [24] . Additionally , in S . cerevisiae , several mitochondrial RNA-binding proteins stabilize mtRNA transcripts , but proteins with similar functions have not been found in mammals [54]–[56] . Given its involvement in maintaining mtRNA it is tempting to speculate that SLIRP fulfills one or more of these roles in mammalian mtRNA biology . SLIRP joins a small cast of RNA-binding mitochondrial proteins that are responsible for maintaining steady-state mtRNA in human cells: LRPPRC , SUPV3L1 and PTCD2 [45] , [57] , [58] . LRPPRC is of central importance for OxPhos function , and patients harboring mutations in LRPPRC develop the French-Canadian variant of Leigh syndrome , a devastating hepatocerebral metabolic disorder [11] . In our study , shRNA targeting SLIRP phenocopies mutations in LRPPRC by causing loss of mtRNA and a significant reduction in complex IV activity . Given these similarities , SLIRP should be considered a candidate gene for respiratory chain disorders . Many of the proteins encoded by the human genome are still functionally uncharacterized and methods such as expression screening will be useful in closing the gaps in our knowledge of cellular pathways [59] . While we have applied expression screening to the cholesterol biosynthesis and OxPhos pathways , this technique is readily extendible to any gene set exhibiting transcriptional co-regulation . In a separate study we show the utility of this method in identifying new mitochondrial proteins important for heme biosynthesis [60] . Expression screening is not the only tool that should be considered for functional prediction . Network-based integration of protein-protein interaction data and other data integration methods have also been successfully applied to predict the functions of uncharacterized proteins [19] , [61] , [62] . Combining expression screening with these or other methods could possibly yield more accurate predictions , especially in cases where transcriptional regulation may not be the dominant mode of regulatory control . Still , microarray gene expression data has several advantages: it is by far the most abundant data source available , offers a more unbiased approach than most techniques , and permits investigating gene function in specific biological contexts . As public repositories of microarray data continue to grow at an accelerating pace , we anticipate that expression screening will become an increasingly important tool for discovering gene function .
A total of 2 , 052 mouse and human microarray data sets were downloaded from the NCBI Gene Expression Omnibus [18] in March 2008 for the Affymetrix platforms HG-U133A ( GEO GPL96 ) , HG-U133+ ( GEO GPL570 ) , MG-U75Av2 ( GEO GPL81 ) , Mouse430A ( GEO GPL339 ) and Mouse430 ( GEO GPL1261 ) . We discarded data sets with less than 6 arrays as well as data sets containing multiple tissues . We then merged overlapping data so that no two data sets shared identical arrays , resulting in a final compendium of D = 1 , 427 data sets . For a given query gene set S , each data set d and each gene g , we calculated the vector of Pearson correlation coefficients rgj between g and all other genes j . We then define the correlation between g and S as the GSEA-P enrichment score ( ES ) statistic [63] with g as the “phenotype” variable and N representing the total number of genes . We next calculated randomized enrichment scores ES0 by randomly permuting values ( arrays ) for gene g , re-calculating rgj and applying the above formula . We pooled N0 = 100 , 000 permuted ES0 values from all N genes to estimate the marginal null distribution of enrichment scores . From this we estimated the global false discovery rate ( FDR ) of each actual ES value [63] , [64] as the ratio of tail probabilities: We take qgd = 1−FDRgd to represent the probability of co-expression of gene g with the query gene set S in data set d . The data set weights wd were defined as the average of qgd across the query genes . We then integrated these probabilities using a robust Bayesian formula [25] to obtain a final probability pg of co-expression of gene g with the query gene set , where is the average of qgd in data set d . This method of data integration assumes conditional independence between data sets given the co-expression hypothesis , which allows concurring evidence from multiple data sets to reinforce each other in calculating the integrated probability . Incorporating the prior p0 affords some robustness to outliers in terms of qgd values close to 0 or 1 , which can arise from the permutation-based FDR estimation . For the OxPhos expression screen the prior p0 was set to 5% , roughly corresponding to the fraction of mitochondrial genes in the genome [35] . Since the query genes are used to calculate the weights wd , the sensitivity and specificity was estimated using leave-one-out cross-validation , with one gene withheld from the weights calculation in each iteration . We mapped Affymetrix probesets to NCBI Homologene identifiers using a previously described method [65] . For the query gene sets ( OxPhos and cholesterol pathways ) we validated each gene's mapping by Blast . It is important that each query gene is represented only once in the gene set . For cases in which multiple Affymetrix probesets map to a gene in the query gene set , we chose the probeset with the least potential for cross-hybridization according to Affymetrix probeset annotations . Specifically , we used the following Affymetrix probeset suffix hierarchy ( ‘at’>‘a_at’>‘s_at’>‘x_at’ ) . In cases where there were ties , we chose the lower numbered probeset to represent a gene . We performed expression screening as described above separately for each microarray platform , using Affymetrix probeset-level data . When integrating across all array platforms , we chose for each Homologene identifier the probeset with maximal pg in the platform-specific screen . We then re-integrated across all data sets , ignoring missing values , to produce the final list of probabilities , pg ( Figure 3 , Table S3 , Table S5 ) . MCH58 immortalized human fibroblasts were kindly donated by Eric Shoubridge [66] . 293T and C2C12 cells were received from the American Type Culture Collection ( CRL1772 & CRL11268 ) . Unless otherwise indicated , all experiments were carried out in DMEM , 4 . 5 g/L glucose , 10% FBS ( Sigma #2442 ) supplemented with 2 mM glutamine , 100 I . U Penicillin and 100 ug/ml Streptomycin . Lentiviral vectors for expressing shRNA ( pLKO . 1 ) were received from the Broad Institute's RNAi Consortium or from Open Biosystems . Unique identifiers of each shRNA construct can be found in Figure S1 . Procedures and reagents for virus production are adapted from the Broad Institute's RNAi Consortium protocols [67] . Briefly , 400 , 000 293T cells were seeded in a 24-well dish and 12 hr later triple-transfected with pLKO . 1-shRNA , a packaging plasmid ( pCMV-d8 . 91 ) and a pseudotyping plasmid ( pMD2-VSVg ) using Fugene6 reagent at 3∶1 ( reagent∶DNA ) ( Roche #11815091001 ) . Media was refreshed 18 hr post-transfection , supplemented with 1% BSA and virus collected 24 h later . For infection , 30 , 000 cells were seeded in a 24-well dish . 30 ul viral supernatant was added to cells for a final volume of 500 ul media containing 8 ug/ml polybrene ( Sigma #H9268 ) . The plates were spun 800rcf for 30 min at 32°C , returned to 37°C and 24 h post-infection were selected for infection with 2 ug/ml puromycin ( Sigma #P9620 ) . RNA for assessing knock-down efficiency was isolated 7–10 days post-infection . RNA was isolated using the RNeasy system ( Qiagen #74106 ) with two repetitions of DNAse digestion to remove mtDNA and genomic DNA from the sample . 1 ug of RNA was used for 1st-strand cDNA synthesis using a mix of poly ( dT ) and random hexamer primers ( SuperScript III , Invitrogen , #18080 ) . Genomic DNA used for the analysis of mtDNA quantity per cell was isolated using Qiagen DNeasy system . 1 ng genomic DNA was used for multiplex qPCR analysis to simultaneously measure nuclear DNA and mtDNA ( See Table S6 and [68] ) . qPCR of cDNA and genomic DNA was performed using the 96-well ABI7500 qPCR system in 20 ul reactions prepared with 2× master-mix ( ABI #4369510 ) , the appropriate 20× ABI taqman assay ( Table S6 ) and diluted cDNA sample . A pDONR-221 Gateway clone for SLIRP , used previously by our laboratory for cellular localization studies [35] , was sequence verified and cloned into a pcDNA destination vector in-frame with a C-terminal V5-His tag ( pDEST40 , Invitrogen #12274-015 ) . 293T cells were infected with validated shRNAs targeting either SLIRP or GFP as described above . Seven days post-infection , 400 , 000 cells were seeded in a 24-well dish and transfected with either pcDNA-LacZ-V5-His or pcDNA-SLIRP-V5-His . 48 hrs post-transfection , cells were harvested for downstream analyses . Live-cell oxygen consumption readings ( OCR ) were performed using a 24-well Seahorse XF24 Bioflux analyzer . MCH58 cells were seeded at a density of 30 , 000/well in unbuffered media ( 4 . 5 g/L glucose , 4 mM glutamine , 1 mM pyruvate in DMEM , pH 7 . 4 ) . The XF24 analyzer was set to read OCR/well as an average of four 3 min measurements with a 1 min mixing interval between measurements . Each plate contained four different samples of five replicates each with one sample always being the shRNA vector control ( pLKO . 1 ) that was used for normalization and comparison between experiments . Each “batch” of four samples was measured on four different plates summing to 20 replicates per sample . Samples were randomized within each plate to avoid plate-position effects and experiments were repeated multiple times to ensure reproducibility . To correct for cell seeding errors , we measured the total protein content per well after each experiment ( BCA method ) and normalized the OCR per well by dividing by the corresponding protein concentration . 5 ug of cleared whole cell lysate isolated in RIPA buffer was used per lane on a 4–12% Bis-Tris gel ( Invitrogen , #NP0321 ) and blotted on a PVDF membrane ( Invitrogen , #LC2005 ) using a semi-dry transfer apparatus ( Bio-Rad ) , 15 V , 20 min . Membranes were blocked for 2 hr at room temperature in tris-buffered-saline solution ( Boston BioProducts #BM300 ) with . 1% Tween-20 and 5% BSA ( TBS-T-BSA ) . Primary antibodies were incubated with the membrane overnight at 4°C in TBS-T-BSA at the dilutions reported in Table S7 . Secondary sheep-anti-mouse ( GE-HealthCare , #na931v ) or sheep-anti-rabbit ( GE-HealthCare , #na934v ) antibodies were incubated with the membrane at a 1∶5 , 000 dilution in TBS-T-BSA for 45 min at room temperature . The membrane was developed using Super Signal West Pico ( Pierce , #34077 ) . Complex I and Complex IV Dipstick activity assays were performed on 20 ug and 25 ug cleared whole cell lysate , respectively , according to the manufacturer's protocol ( Mitosciences #MS130 and #MS430 ) . 30 ul of lysis buffer A was used per 500 , 000 cells for lysis and solubilization . | Respiratory chain disorders represent the largest class of inborn errors in metabolism affecting 1 in every 5 , 000 individuals . Biochemically , these disorders are characterized by a breakdown in the cellular process called oxidative phosphorylation ( OxPhos ) , which is responsible for generating most of the cell's energy in the form of ATP . Sadly , for approximately 50% of patients diagnosed , we do not know the molecular cause behind these disorders . One possible reason for our limited diagnostic capability is that these patients harbor a mutation in a gene that is not known to act in the OxPhos pathway . We therefore designed a computational strategy called expression screening that integrates publicly available genome-wide gene expression data to predict new genes that may play a role in OxPhos biology . We identified several uncharacterized genes that were strongly predicted by our procedure to function in the OxPhos pathway and experimentally validated two genes , SLIRP and CHCHD2 , as being essential for OxPhos function . These genes , as well as others predicted by expression screening to regulate OxPhos , represent a valuable resource for identifying the molecular underpinnings of respiratory chain disorders . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"biology/post-translational",
"regulation",
"of",
"gene",
"expression",
"genetics",
"and",
"genomics/gene",
"discovery",
"biochemistry",
"genetics",
"and",
"genomics/functional",
"genomics",
"genetics",
"and",
"genomics/gene",
"expression",
"molecular",
"biology/mrna",
"stability",
"computational",
"biology/metabolic",
"networks",
"computational",
"biology/genomics",
"genetics",
"and",
"genomics/medical",
"genetics"
] | 2009 | A Computational Screen for Regulators of Oxidative Phosphorylation Implicates SLIRP in Mitochondrial RNA Homeostasis |
The molecular mechanisms that translate drug treatment into beneficial and unwanted effects are largely unknown . We present here a novel approach to detect gene-drug and gene-side effect associations based on the phenotypic similarity of drugs and single gene perturbations in mice that account for the polypharmacological property of drugs . We scored the phenotypic similarity of human side effect profiles of 1 , 667 small molecules and biologicals to profiles of phenotypic traits of 5 , 384 mouse genes . The benchmarking with known relationships revealed a strong enrichment of physical and indirect drug-target connections , causative drug target-side effect links as well as gene-drug links involved in pharmacogenetic associations among phenotypically similar gene-drug pairs . The validation by in vitro assays and the experimental verification of an unknown connection between oxandrolone and prokineticin receptor 2 reinforces the ability of this method to provide new molecular insights underlying drug treatment . Thus , this approach may aid in the proposal of novel and personalized treatments .
A drug can modulate its targets directly or indirectly ( e . g . via modulation of the gene expression ) and only a small proportion of these protein targets are known [1–3] . Due to this incomplete understanding of drug mode of action , current drug treatment often suffers from unwanted effects [4] . In addition , the promiscuity of many drugs , that is the tendency of drugs to modulate multiple targets [5] , hampers the anticipation of drug response and adverse effects in clinical practice . This is furthermore complicated by the genomic heterogeneity in the population , which produces a large variability of efficacy and adverse effects among patients [6] . Pharmacogenomic studies fortified the important role of gene sequence polymorphisms in drug efficacy and adverse effects [7–9] . Understanding each individual’s drug response is , thus , an additional challenge in the treatment of diseases and has a huge impact on attrition rates in drug discovery . Therefore , in order to personalize medication and to improve drug efficacy as well as drug safety , it is necessary to develop novel approaches expanding the knowledge of the molecular mechanisms underlying drug treatment . Several experimental techniques have been developed to detect molecular associations of drugs [10] . However , limitations on the identifiable drug targets and their indirect effects , the high cost and low throughput of those experiments have hindered the elucidation of molecular determinants of many drugs . Classical approaches to detect drug-target interactions are based on biochemical affinity purification [11] . This method is time consuming and can only detect abundant high-affinity binding proteins , hampering its applicability to detect indirect and low affinity associations as well as interactions with protein complexes . Chemical proteomics approaches that typically combine affinity chromatography and proteomic techniques [12] have the advantage of finding interactions on a large scale . Yet , the challenge persists to detect interactions with proteins expressed at low levels without including unspecific bindings . Expression-cloning-based methods , like phage display or yeast three-hybrid [13] , can circumvent the low protein abundance issue [14] , but they cannot always capture the complexity of molecular and chemical interactions in the human organism [15] . Computational methods are arising as alternative and complementary approaches to propose novel molecular drug interactions . Methods relying e . g . on structural similarity of compounds [4 , 16] or side effect similarity have been successfully applied to reveal drug-target relationships and also to provide mechanistic insights into adverse effects [5 , 17 , 18] . Recently , the comparison of side effects of drugs and phenotypic traits of perturbed genes in mouse models has also been proposed as an option to identify drug targets [19] . Interestingly , this approach has the advantage of not relying on established drug-target relationships , offering the potential to discover novel drug-target interactions . This method follows the idea that the manipulation of a target by genetic or pharmacological means should consistently lead to phenotypic changes that are aligned with the desired therapeutic effect [20] . In this aspect , it has been shown that phenotypes resulting from knock-out mice correlate well with known phenotypes of drug response [21] . However , to detect single gene perturbations in mice that share similar phenotypes with drugs in a sensitive manner , several methodological challenges need to be overcome . These challenges arise from the large number of side effects of drugs stemming from their polypharmacological potential [22 , 23] as well as physiological differences between mice and humans . In this work , we develop a new scoring scheme to evaluate the similarity of phenotypic traits from gene perturbations in mice and side effect profiles of drugs , which is able to cope with the polypharmacological property of drugs . Our approach reveals molecular associations of drugs and genes including direct and indirect drug targets , gene-drug links involved in pharmacogenetic associations as well as causal protein-side effect relationships . We moreover provide experimental evidence of the capability of our phenotypic similarity scoring scheme to detect novel drug-target interactions .
In order to learn more about drug mode of action and molecular mechanisms underlying side effects , we devised an extended semantic similarity scoring system to identify drugs and mouse genes that share similar phenotypes and are , thus , likely to be molecularly related [24] . As we aimed to find perturbations of mouse genes reproducing the side effects of drugs in human , we encoded the phenotypic traits in mice and the drug side effects with the Medical Dictionary for Regulatory Activities ( MedDRA ) [25] . MedDRA is a highly standardised medical terminology used for the annotation of adverse side effects in clinical trials that contains terms related to human health . Besides , an adaptation of the hierarchical organization of MedDRA ( see Materials and Methods for details ) enables the assessment of the semantic distance between phenotypes of different perturbations in mammalian organisms [26] . To determine drug—mouse gene pairs with similar phenotypes , we conceived a symmetric score that averages the drug-gene and gene-drug phenotypic similarities ( see Materials and Methods for details ) and devised an approach accounting for the well-known tendency of drugs to bind multiple targets [4] , also known as polypharmacological property . To compute the gene-drug phenotypic similarity , we averaged the scores of the most similar side effect-trait pair for all gene phenotypic traits , thereby prioritizing drugs that share a large proportion of their side effects with traits of mouse models . In contrast , for the computation of the drug-gene phenotypic similarity , we averaged only a subset of the 20 side effects with the highest phenotypic trait-side effect scores ( see S2 and S3 Figs for cut-off evaluation ) . In this way , we disregarded side effects likely to be unrelated to a single target , thereby correcting for the polypharmacology of drugs ( Fig 1a ) . In addition , as associations of drugs with many side effects to genes with low phenotypic information are more likely to score high by chance , we scaled the resulting score with the number of mouse phenotypic traits and downweighted drugs with many side effects ( see Materials and Methods for details ) . This phenotypic similarity measurement allowed us to calculate the similarity of over 8 million gene-drug pairs involving 1 , 667 drugs and 5 , 834 single gene perturbations in mice . In order to assess if our method can give insights into drug mode of action as well as associations between drug targets and adverse effects , we compared the phenotypic similarity of the gene-drug pairs to different data sets providing information of relationships between drugs , genes , and side effects ( see Fig 1b ) . We tested if our approach detects known direct and indirect drug-target relationships , including those mediated through drug targets via protein-protein interactions , pharmacogenetic variations , and causal connections between proteins and side effects . First , we evaluated if our method is able to detect known drug-target associations from the STITCH database [27] . This dataset includes direct interactions , that is , physically interacting drug-target pairs ( 863 , 074 pairs ) and drugs that modulate the targets indirectly ( 4 , 118 , 052 pairs ) , e . g . by altering the expression pattern of a gene via DNA binding ( Fig 2a ) . We assessed the performance of our model using precision , ROC and lift measurements ( Fig 2b and S3 Fig as well as S5 Fig ) . The lift value is a measurement of the performance of a method evaluated against a random choice model that estimates the precision of a scoring scheme in relation to the probability of obtaining a true value by chance ( see Materials and Methods for details ) . We found that phenotypically related pairs are strongly enriched in both direct and indirect gene-drug molecular interactions ( see Fig 2b ) . We also compared the performance of our model with a previously proposed gene-drug semantic similarity scoring system [24] and observed that our semantic similarity approach outperforms it significantly ( see S4 Fig ) . For a close manual inspection of the most confident gene-drug links detected by our approach , we focused on the set of high scoring gene-drug associations where the precision in the direct associations exceeds 10% and the lift reaches a value of 20 ( Fig 2b ) ( 1338 associations connecting 214 genes and 394 drugs ) . For indirect targets , we obtained an enrichment over random ( lift value ) of over 7 at this cutoff . We provide the full list of high scoring pairs in S1 Table . In addition , examples mentioned herein are shown in Table 1 . Instances of direct connections among the high phenotypic similarity relationships include the antipsychotic drugs aripiprazole and risperidone linked to their direct target dopamine receptor 2 ( DRD2 ) , the Vitamin D Receptor ( VDR ) connected to its ligand ergocalciferol ( Vitamin D2 ) . In addition , we found the Estrogen Receptor 1 ( ESR1 ) related to the steroid estradiol and its derivatives estradiol acetate and estradiol cypionate ( see Table 1 and S1 Table ) . We also detected a high semantic similarity between steroids that activate the androgen receptor ( AR ) ( e . g . oxymetholone , oxandrolone , nandrolone , fluoxymesterone , oxymetholone , and testosterone ) and the AR mouse gene as well as a connection between testosterone and the mouse gene coding for the testosterone transforming enzyme aromatase ( cytochrome P450 family 19 subfamily A member 1 , CYP19A1 ) . An indirect relationship in which the drug increases the expression of the gene is exemplified by the high phenotypic similarity between testosterone and the follicle-stimulating hormone receptor ( FSHR ) [28] . Indirect gene-drug associations might involve different molecular actions ranging from distant ( e . g . gene expression ) to closer ( e . g . an interaction with a gene product physically bound to the drug target ) mechanisms . Hence , we decided to investigate whether our method is able to capture close molecular associations between a drug and the gene product . To that aim , we analysed the distance between the known drug target and the gene product in a protein-protein interaction network . We found that phenotypically similar gene-drug pairs tend to share close molecular mechanisms ( see Supplemental Information for details ) . Taken together , these results indicate that our scoring scheme enables the detection of shared molecular links between drug targets and phenotypically similar gene products . These molecular connections include direct physical binding of the drug to its target as well as indirect effects such as those influencing transcriptional regulation . Pharmacogenomics studies highlighted the important role of genetic polymorphism in drug efficacy and adverse effects [7] . Whereas some pharmacogenetic associations involve genetic variations in the physical interacting drug targets ( e . g . DRD2 connected to aripiprazole and risperidone ) , others appear to be mediated by more complex and indirect functional associations . The observation of gene-drug pairs involved in pharmacogenetic relationships among the top scoring associations prompted us to investigate in a systematic way if gene-drug pairs involved in pharmacogenetic interactions are also enriched in phenotypically similar gene-drug pairs . To test this hypothesis , we utilized the associations annotated in Pharmacogenomics Knowledge Base ( PharmGKB ) [29] . PharmaGKB contains different types of pharmacogenomics relationships , including the “pheno” and “clinical” associations ( see Materials and Methods for details ) . The “pheno” connections link a gene variant and an affected phenotype , whereas the “clinical” links are manually curated annotations of clinically relevant pharmacogenetic variant—drug pairs . We tested if our scoring scheme can detect gene-drug links resulting from these types of annotations by calculating the enrichment over random ( lift ) for each relation type . The gene-drug pairs with high phenotypic similarity showed a strong enrichment over random ( over 100 ) for the phenotype annotations and even stronger enrichment of 250 for the clinical associations ( Fig 2c ) . A known gene-drug pair involved in pharmacogenetic associations among phenotypically similar pairs includes for example the dopamine receptor 2 ( DRD2 ) connected to antipsychotics . In the DRD2 gene , the SNP rs1799978 is a significant predictor for the response to the antipsychotic risperidone [30] . Molecularly more distant pairs related to genetic variations on drug response are exemplified by the link between the leptin receptor ( LEPR ) and antipsychotics . In particular , the LEPR Q223R polymorphism is significantly associated to obesity in women treated with atypical antipsychotic drugs [31] . Pharmacogenetics is a newly evolving field and consequently , many pharmacogenetic interactions are not known . This might cause an underestimation of the performance of the method to retrieve these interactions . In this context , a literature survey revealed the presence of pharmacogenetic associations among highly similar gene-drug pairs . For example , mutations in the TP53 gene , which exhibited a high phenotypic similarity to an antagonist of the estrogen receptor tamoxifen , have been shown to be a significant predictor of poor response in tamoxifen-treated patients [32] . This example illustrates the potential of the method to reveal novel genes involved in pharmacogenetic interactions . Since drugs and genes sharing similar phenotypes are likely to be molecularly related [24] , we reasoned that these shared phenotypes might give valuable insights into protein-side effect relationships and explain the molecular causes of drug adverse effects . For example , the cancer related phenotypes ( e . g . “adenocarcinoma” , “malignant soft tissue neoplasm“ , and “sarcoma“ ) linked to tamoxifen treatment as well as to mice with impaired TP53 protein function might be due to the effect of the drug on the function/activity of p53 protein , for example , via a cooperative effect with its target , the estrogen receptor [33] . Similarly , the metabolic effects shared by drugs and mice harbouring defective leptin ( LEP ) or leptin receptor ( LEPR ) could be related to the action of these drugs on the proteins coded by LEP and LEPR genes . For example , “obesity“ , “insulin resistance” , and “glucose tolerance impaired” are phenotypes that atypical antipsychotics as well as antivirals share with LEP and LEPR . These effects might be caused by the direct or indirect influence of the drugs on these proteins ( Fig 3a ) . To systematically test if the set of side effects most similar to the gene phenotype reveal causative connections between the gene product and the side effects , we compared these terms to a manually curated dataset of previously reported side effect-protein relationships [18] . In particular , we compared the fraction of known gene-side effect relationships among the shared side effects ( see Materials and Methods for details ) of high semantic similarity pairs to the corresponding fraction of low scoring pairs . We found that this fraction is significantly ( 5 . 88E-10 , Wilkoxon ranksum ) larger in the gene-drug pairs having a high semantic similarity ( Fig 3b ) . These findings show that investigating the specific side effects leading to a high overall phenotypic similarity can reveal causative connections between the gene product and the side effects . Thus , our method is capable of providing hypotheses about the molecular mechanisms leading to adverse drug effects . Biologicals are recently gaining the attention of pharmaceutical companies as they open new avenues for targeting non-druggable proteins , that is , proteins that do not bind chemicals naturally . Thus , we tested if our method can also be applied to detect molecular associations of biologicals and shed light onto their side effects . For that , we manually analysed the relationships of the 51 biologicals with mouse genes in our set of top-scoring associations ( see S2 Table ) . Among these associations we found known relationships between biologicals and their protein interaction partners including the link of follitropin beta , a recombinant form of follicle stimulating hormone ( FSH ) to its receptor FSHR ( see zoom-in of Fig 4 and Table 1 ) . We also detected obvious indirect associations between biologicals and the human orthologues of phenotypically similar mouse genes . For example , coagulation related proteins ( e . g . alpha and gamma chain of fibrinogen ( FGA , FGG ) and coagulation factor XIII ( F13A1 ) are connected to lepirudin recombinant , a recombinant hirudin derived from yeast cells and a direct thrombin inhibitor ( Fig 4 zoom-in and Table 1 ) . All the evidence presented above show that our method can detect drug-target associations as well as gene-drug links resulting from pharmacogenetic studies and provides hypotheses on the molecular mechanisms behind adverse drug effects . To seek for experimental evidence on the phenotypically similar drug-target associations , we first compared our high scoring associations with the hits of in vitro assays of ToxCast project [34] , which recently systematically tested more than 1 , 800 compounds in 821 in vitro assays . This database contains in vitro activity information for 14 , 238 gene-drug pairs analyzed here , comprising the activity of 141 drugs on different assays related to 263 genes ( see Materials and Methods for details ) . We found experimental information for 38 of the high scoring pairs ( S3 Table ) . These pairs are clearly enriched in associations with experimental support ( 4-fold , see S8 Fig ) , with 50% of these pairs appearing as hits of in vitro assays . These results therefore provide experimental support for the phenotypically similar gene-drug associations . In order to present also experimental evidence on the newly discovered drug-target associations involving proteins that have not been extensively screened for ligands , we sought for unknown connections of a drug and a mouse gene for which a functional assay for the encoded protein is commercially available . Among these connections , we found the intriguing link between two derivatives of male hormone dihydrotestosterone , oxandrolone , and oxymetholone and the gene encoding for prokineticin receptor 2 ( PROKR2 ) . Prokineticin receptor 2 has recently been implicated for the first time in the binding of small molecules [35–38] . As oxandrolone shows a slightly higher phenotypic similarity to PROKR2 than oxymetholone , we decided to investigate if oxandrolone has a functional effect on prokineticin receptor 2 protein ( PKR2 ) . We experimentally tested the possible activity of oxandrolone on the receptor in an agonist and an antagonistic functional assays of HEK-293 cells expressing PKR2 [39] ( see Materials and Methods ) . We first tested a single concentration of oxandrolone ( 1 . 0E-05M ) on both assays and observed activity on the antagonist assay . We then determined the dissociation constant ( Kb ) value ( 9 . 5E-06 M ) of the oxandrolone antagonistic activity in a dose-response curve ( Fig 5 ) . To rule out the possibility of a non-specific effect of oxandrolone on Ca2+ concentrations , oxandrolone was tested at high concentrations on HEK-293 cells expressing PKR2 in the absence of the known stimulant PK2 ( S9 Fig ) . In these conditions , oxandrolone showed no interference with Ca2+ mobilization , demonstrating its specificity on PKR2 signalling . The experimental validation of the antagonistic effect of oxandrolone on PKR2 proves that our method is able to detect novel drug-target interactions , further reinforcing the applicability of our method to elucidate drug action . In summary , we have shown the potential of our novel extended similarity scoring system to unravel molecular mechanisms underlying drug treatment leading to unwanted side effects . Not only known genes encoding for drug targets exhibit a high semantic similarity to their associated drugs , but also protein interaction partners of the known targets do so . Our approach is moreover able to discover gene-drug connections involved in pharmacogenetic interactions , protein-side effects links and associations with biologicals . We furthermore proved experimentally that our method can detect novel drug-target interactions .
In this work , we have analysed the similarity of phenotypes resulting from drug treatment ( side effects ) and from gene perturbations in mouse models . We showed that our extended semantic similarity approach can elucidate molecular mechanisms that translate drug influence into phenotypic effects giving insights into intended ( on-targets ) and unintended ( off-targets ) interactions . The detailed analysis of the phenotypes shared by gene-drug pairs exhibiting a high similarity revealed valuable insights into causative connections between drug targets and side effects . We validated the potential of our approach to detect known drug targets , genes involved in pharmacogenetic associations as well as connections to biologicals and experimentally proved the applicability of our method to detect novel drug-target physical interactions . Phenotypically similar gene-drug pairs exhibit a strong enrichment in known direct as well as in indirect drug-target relationships ( Fig 2b ) . The proximity of the human orthologue of the mouse gene to the known drug target in the PPI network confirms that our approach detects links between drugs targets and functionally related proteins ( S6 Fig ) . This implies that proteins encoded by genes that share similar phenotypes with a drug are likely to physically interact with that drug’s target or cooperate with it in the same pathway . The high phenotypic similarity between LEP/LEPR and drugs for the nervous system including the dopamine agonist aripiprazole and the selective serotonin inhibitors paroxetine , pramipexole and escitalopram further shows the potential of our method to detect non-obvious associations . The shared phenotypes of aripiprazole and mice with impaired LEP/LEPR function like “obesity” , “insulin resistance” and “glucose tolerance impaired” ( Fig 3a ) point to alterations of leptinergic signalling by the antipsychotic aripiprazole . The pharmacogenetic association of LEPR and LEP [31] to obesity caused by antipsychotics additionally confirms this . Similarly , the side effects that LEP and LEPR share with selective serotonin inhibitors such as “food craving” , “binge eating” , “hyperphagia” and “obesity” ( Fig 3a ) suggest a connection between serotonin signalling and leptin . Indeed , recent studies show that the regulation of appetite by leptin takes place for the most part through inhibition of serotonin synthesis and release by brainstem neurons [40 , 41] . These results show that our scoring scheme gives valuable insights into complex mechanisms leading to adverse effects of drugs . We have demonstrated that our method predicts also molecular associations with biopharmaceutical drugs , where the market has gained a very high and still growing value in recent years [42] and whose adverse effects are often not fully understood . For example , the recombinant form of follicle stimulating hormone ( FSH ) , follitropin beta , shows a high phenotypic similarity with mouse models with perturbed FSHR gene ( Fig 4 , Table 1 ) . This association is further confirmed by pharmacogenetic associations in PharmGKB . Consistent with the role of gonadotropin releasing hormone ( GnRH ) in the release of FSH , we find FSHR also associated to goserelin , an injectable GnRH superagonist , and to leuprorelin , a GnRH analog ( Table 1 ) . In line with this association , the stimulation of FSH by GnRH leads to increased levels of testosterone , progesterone , and estradiol . Interestingly , this mechanism is also clearly reflected by the phenotypic traits that FSHR shares with leuprorelin and goserelin that include "Blood testosterone decreased" , "Oestradiol decreased" , "Progesterone abnormal" and "Blood oestrogen abnormal" . Intriguingly , the steroid oxandrolone exhibits a high semantic similarity to its known receptor AR as well as to PROKR2 ( Fig 5 and Table 1 ) . We hypothesized that the observed phenotypic similarity between oxandrolone and PROKR2 could result from indirect effects e . g . via a functional connection between AR and PROKR2 or , alternatively , from the effects of the drug on the PKR2 pathway . To precisely determine the link between oxandrolone and PROKR2 , we tested the activity of the drug on in vitro functional assays and observed an antagonistic effect of oxandrolone on PKR2 signalling . We showed that this effect is specific of PKR2 signalling ( see S9 Fig ) . Furthermore , it is independent from the AR target , since HEK-293 cells do not express AR . These results indicate that the detected antagonistic activity is either mediated via a physical interaction of oxandrolone with PKR2 or through the interference with downstream proteins of the PKR2 pathway . The direct binding of oxandrolone to the G-protein coupled receptor ( GPCR ) protein PKR2 is plausible , since several research studies support the involvement of GPCRs on non-genomic effects of androgens and other steroids [43] . These fast effects of steroids are not mediated by the classical transcriptional effects , but through interaction of steroids with their target receptors , which also include GPCRs . Despite the relative low affinity of oxandrolone to PKR2 , its plasma concentration is compatible with a clinically relevant interaction of oxandrolone and PKR2 . At oxandrolone therapeutic dosages ( 10 mg ) , its average plasma concentration ( 417 ng/ml ) is in the micromolar range ( 1 . 6 μM ) [44] and is expected to be even higher in weightlifters and bodybuilders who chronically administer it at supraphysiologic doses [45] . In addition , the physiological effects of a PKR2 antagonist as well as those observed in individuals carrying PROKR2 gene mutations are in concordance with the clinical effects of oxandrolone being mediated via PKR2 signalling . The clinical effects observed after oxandrolone treatment include low levels of luteinizing hormone ( LH ) , GnRH and testosterone [46 , 47] . Similarly , a PKR2 antagonist ( 3Cl-MPL ) has been shown to blunt circulating luteinizing hormone ( LH ) levels in mice [48] . Moreover , mutations in PROKR2 lead to GnRH-deficiency and more specifically to Kallmann syndrome , a disease characterized by hypogonadism , a decreased functional activity of the gonads [49] . Interestingly , mutations in PROKR2 linked to Kallmann syndrome have been shown to impair Ca2+ release in HEK293 cells [50 , 51] , which is consistent with the effect of oxandrolone on PKR2 in the same cell lines detected herein . However , the effects of oxandrolone are not clearly distinguishable from those expected as AR agonist , which also include the influence on LH and GnRH levels [52] . Whether the clinical effects of oxandrolone are mediated by the antagonistic activity of oxandrolone on PKR2 , via its activity on the potent target AR or via collaborative effects clearly requires further experimental investigation . Interestingly , oxandrolone causes less virilising effects and also less adverse effects than the natural AR agonist testosterone [53] . Although these effects have been ascribed to the inability of oxandrolone to aromatise to estradiol [54] , in the light of this new finding , the contribution of PKR2 activity of the drug to the decrease of steroid levels deserves further investigation . Taken together , we could experimentally verify a formerly unknown ( direct or indirect ) interaction between the GPCR PKR2 and oxandrolone , which further sheds light into the clinical effects of oxandrolone treatment . The potential of our method was further confirmed experimentally using the results of in vitro assays of the ToxCast project . We observed that phenotypically similar gene-drug pairs are strongly enriched in pairs where the drug is active on in vitro assays of the protein target encoded by the gene . The results of these assays confirmed associations annotated in the drug target datasets analyzed herein such as DRD2 and haloperidol or VDR and ergocalciferol and provide experimental support for new ones including TP53 and azathioprine ( see S3 Table ) . We have demonstrated that our approach is also able to detect genes responsible for the variation on drug response as indicated by the strong enrichment over random of gene-drug associations from PharmGKB ( Fig 2c ) . Recent studies revealed the strong impact of genetic variations on an individual’s response to drugs [55] . However , the majority of the genetic variants responsible for the observed variability on drug response in the population remain to be elucidated . Our method could guide future pharmacogenomic studies by proposing a prioritized list of candidate genes involved in drug response in an analogous manner to gene prioritization used for genome-wide association studies of ( rare or multifactorial ) diseases [19 , 56–58] . This may have important implications in personalized treatment decisions helping to improve drug efficacy and safety . Although we have shown that our approach can detect many meaningful gene-drug connections based on in vivo phenotypic information , it has also limitations inherent to the cross-species comparison . Mutations in genes in mice do not necessarily have the same effect in human and associations involving species-specific ( mouse or human ) genes or gene families may not be detected by our method . In addition , drugs can act differently in different species , for example due to differences in drug metabolizing enzymes [59] . Terms commonly used as side effect descriptions , such as headache , may not be detected as phenotypic feature in mice . Moreover , the mapping of MPO-annotated phenotypic traits to the MedDRA vocabulary leads to a loss of information among the phenotypes linked to mouse genes . As a consequence , the number of the genes for which we could detect a sufficient phenotypic similarity to drugs is reduced . Using a stringent mapping procedure , we could translate with high confidence 26% of the mouse phenotypic descriptors utilized as gene annotations in MGI to MedDRA ( S1–S4 Files ) . Interestingly , the coverage of terms mapped to MedDRA terms per mouse gene was higher ( 29% ) than the total number of unique MPO terms mapped to MedDRA . This is explained by the higher likelihood of MPO terms representing frequently observed mouse phenotypic traits to be translated into MedDRA ( S1b Fig ) . This reduces the impact of the loss of phenotypic information in the approach . We chose to code the phenotypes in the MedDRA terminology over other widely used human clinical ontologies such as HPO to optimally capture information of the effects of drugs in human . MedDRA is intended to be used in the pre- and post-marketing phases of the medicines regulatory process covering also adverse drug reactions [25] . Thus , MedDRA meets our aims better than HPO , which is tailored to cover phenotypic abnormalities of human diseases [60] . Since drug side effect information is recorded from populations with heterogeneous genetic backgrounds , we opted to aggregate phenotypic traits resulting from different genetic perturbations in mice of the same gene ( e . g knock-outs , knock-ins or SNPs ) and did not differentiate between distinct genetic backgrounds . This has the advantage of enriching the number of phenotypic descriptions linked to genes . In the future , more detailed phenotypic data of genes and drugs would allow the development of more discriminative tools to predict for example agonistic or antagonistic effects of the drug on the gene product and specific genetic variants modulating drug response . This information could include the frequency of occurrence of a side effect per specific drug , additional readouts of the drug action or gene attributes such as metabolic profiles . Taken together , the results presented here show that our semantic similarity approach is suited to detect mice models mimicking drug phenotypes with high precision and accuracy ( see S3c–S3f Fig ) , allowing determining 1338 phenotypic associations connecting 214 genes and 394 drugs of diverse indication areas . The results of this work demonstrate that the phenotypic similarity between drugs and genes gives valuable insights into molecular mechanisms of drug treatment . The knowledge about relationships between drugs and genes has important implications in personalized treatment decisions , as considering drug mode of action and the genetic predisposition of a patient could circumvent drug inefficacy and adverse effects . Even effects resulting from treatment with combination of drugs may be anticipated if a better understanding of drug mode of action is obtained . Moreover , this information could be used for drug repurposing , because novel drug-target interactions may provide insights for the application of marketed drugs to new indications . We prove that comparing drug side effects and mouse phenotypic traits reveals insights into drug mode of action . Gene-drug pairs exhibiting a high phenotypic similarity are enriched in known direct and indirect drug-target relationships . Our systems biology approach moreover extends the knowledge about the molecular mechanisms leading to unwanted side effects and about genetic variation influencing drug response . We furthermore provide in vitro evidence for the potential of our approach to detect drug-target associations . The experimental validation of a novel drug-target interaction enabled us in addition to get insights into molecular mechanisms of oxandrolone treatment . Thus , this analysis improves drug therapy by advancing the understanding of modes of drug action , adverse effects and genes involved in pharmacogenetic interactions . This may help to find new therapeutic applications for drugs or aid in personalized treatment decisions .
We extracted drug phenotypic information from our in-house drug repository which contains 3987 unique side effects associated to 1667 drugs ( 155 , 973 pairs ) as previously described [26] . The side effect data was parsed from public documents directed at health care professionals or the public such as drug labels , monographs or assessment reports . We annotated the phenotypic information employing the Medical Dictionary for Regulatory Activities ( MedDRA ) , a medical terminology intended to describe e . g . diagnoses , symptoms and signs , adverse drug reactions and therapeutic indications [25] . Terms of this terminology were collected from diverse sources like the World Health Organization's ( WHO ) adverse reaction terminology , Coding Symbols for a Thesaurus of Adverse Reaction Terms ( COSTART ) and International Classification of Diseases ( ICD ) 9 and are maintained , further developed and distributed by the Maintenance Support Services Organisation . MedDRA is organized hierarchically which allows us to compute the semantic relatedness of the terms in this ontology . We used an adapted four-level hierarchy of MedDRA suited for semantic reasoning described previously [26] . In this adapted version of MedDRA , we merged the fifth and the fourth level of MedDRA , because there is no clear hierarchical relationship between these two levels . In addition , as described in [26] , we integrated 59 Standardized MedDRA Queries ( SMQs ) , which represent groups of terms across the entire ontology to a defined medical condition . These modifications make MedDRA applicable to measure the semantic similarity of phenotypes . To evaluate the performance of the presented phenotypic similarity scoring scheme we used Receiver Operating Characteristic ( ROC ) , precision , lift and accuracy plots . For all performance plots , the R package "ROCR" [66] was utilized . In the following paragraph we explain the different performance measurements . Let Y be a random variable representing the known information about the relatedness of a drug-gene pair and Ŷ a random variable representing the classification according to the scoring scheme for a randomly drawn sample . φ denotes the positive class and φ¯ denotes the negative class , respectively . Further , P describes the number of ( empirical ) positives , N the number of negatives , TP the number of true positives and FP the number of false positives . We benchmarked our results with known human drug-targets interactions from the STITCH database [27] . We only used associations from curated databases , excluding e . g . relationships based only on textmining and applied a high confidence cutoff of 0 . 7 for the drug-target relationships . We mapped the drugs by name , including synonyms , to our in house drug dictionary . Moreover , we distinguished direct physical interactions from indirect ones as described in [26] . We created a benchmark set of direct interactions and another one of indirect associations by selecting those genes/drugs where at least one physical or indirect interaction has been reported , respectively . In these benchmark sets we constructed the positive set with all known associations involving those genes and drugs and the negative set by all possible combinations of these drugs and genes . This resulted in 863 , 074 and 4 , 118 , 052 drug gene pairs , respectively . We divided our set of drug-gene associations into high and low scoring ones where the precision in direct associations exceeded 10% , which was true at a score higher or equal to 0 . 354 . This lead to 1338 high scoring associations linking 214 genes to 394 drugs . We compared the developed scoring scheme to the one proposed by Hoehndorf et al . [24] by evaluating the performance of the algorithm of Hoehndorf and collaborators using our data . The semantic similarity scoring scheme proposed by Hoehndorf and collaborators provides two asymmetric scores corresponding to drug-gene and gene-drug pairs . In order to compare our approach with the previous published one [24] using the same annotation framework , we applied the algorithm ( https://code . google . com/p/phenomeblast/source/browse/trunk/phenotypenetwork/SimGIC-twosides . cc ) proposed by Hoehndorf to our MedDRA framework and phenotypic annotations ( see S4 Fig ) . To calculate the Hoehndorf non-symmetric scores , we utilized our phenotypes of drugs and genes encoded in MedDRA and the information content of the most informative common ancestor within MedDRA as input of the given code . We did not change the code of the provided algorithm , implying a pre-set threshold of at least 7 phenotypes defined in the code ( define MINPHENOTYPES 7 ) . Subsequently , we compared the resulting non-symmetric similarity scores ( drug-gene and gene-drug pairs ) to the scores of our scoring scheme , where we also applied the threshold of 7 phenotypes . We compared the performance of the two approaches using lift and ROC plots ( S4 Fig ) and our benchmark set of direct drug-target interactions ( see section drug-target interactions above ) . Both performance measurements showed that our approach outperforms the method proposed by Hoehndorf and collaborators . To calculate the significance of the increase of the ROC plots over the method proposed by Hoehndorf et al . , we used the function roc . test from the R package pROC to compare the differences in ROCAUC . We tested the hypothesis that our ROC plot has significantly higher AUC values than each of the assymetric scores proposed by Hoehndorf et al . We observed significantly higher ROCAUC values for our method . Also the lift plot illustrates that the approach presented here performs remarkably better classifying true positives at high scoring regions . We noticed that the application of Hoenhdorf algorithm to our benchmark set of physically interacting human drug-targets results in lower AUC values than those published by Hoehndorf , who use different benchmark sets [24] . These differences suggest that the benchmark sets have an influence on the performance as measured by the ROC plots . In order to confirm this hypothesis , we evaluated the performance of the scores of our MedDRA-based semantic similarity measurement on two benchmark sets provided by Hoehndorf ( in http://phenomebrowser . net/drugeffect-data . tar . bz2 ) . In particular , we benchmarked our approach with the DrugBank ( drugeffect-data/positive/drugbank-targets . txt ) and STITCH ( drugeffect-data/positive/stitch-human-targets-0 . 7 . txt ) datasets , which contain associations between gene MGI IDs and drug STITCH IDs reported in the DrugBank [67] and STITCH databases [27] , respectively . We first mapped the drugs and genes for which we have phenotypic information to MGI and STITCH identifiers . To map the genes in our dataset to the MGI IDs in these benchmark sets , we utilized the file mousephenotypes-names . txt provided by Hoehndorf et al . containing the MGI marker ID and gene name ( in the folder drugeffect-data/input-phenotypes/ ) . We mapped the drug names to STITCH IDs ( STITCHORIG: ) using a file downloaded from the STITCH website ( chemical . aliases . v3 . 1 . tsv , accessed in Dec . 2013 ) . This resulted in 7 , 262 , 444 MGI ID-STITCH ID pairs associated with scores from our algorithm . Following the descriptions of Hoehndorf and collaborators , we constructed ROC plots utilizing as positive sets the DrugBank and STITCH benchmark files and all the remaining possible associations of drugs and genes as negative set . The negative set , thus , includes drugs with side effects linked to genes with phenotypic information never occurring in the evaluation dataset of DrugBank or STITCH . This results in a bigger dataset with a higher proportion of true negatives in low similarity regions leading to a better performance in the ROC curve ( see S5 Fig ) , although the interesting top-scoring cases only contribute very little to the overall AUC . These results demonstrate the influence of the different benchmark sets on the ROC plot evaluation performance . Biologicals or biopharmaceuticals are “recombinant therapeutic proteins and nucleic acid based products and in the broader sense also engineered cell or tissue-based products”[68 , 69] . In order to identify biologicals within highly similar drug-gene relations , we manually curated the list of drugs that are part of these relationships . In total , we identified 51 biologicals associated with a high phenotypic similarity to 95 genes via 226 connections by our phenotypic similarity measurement . We extracted protein-protein interactions from the String database [70] . In order to guarantee high confident interactions we selected protein-protein associations using a cutoff of 0 . 7 [71] . Using this network , we calculated the shortest path for each human orthologue of the mouse gene product in our data set to all drug targets from drugs in our data set . We used the target information provided by STITCH where a direct interaction is reported and annotated for each drug-gene pair the closest distance between a known target of the drug and the human gene product . To test if our method is able to detect genes involved in pharmacogenetic associations , we compared our results to the data collected in the Pharmacogenomics Knowledge Base ( PharmGKB ) [29] . PharmGKB includes connections of genes to single drugs , groups of drugs as well as therapeutic classes of drugs . If multiple drugs were annotated simultaneously to one gene , we split these associations and treated each drug-gene pair individually . In order to analyse therapeutic classes of drugs , we classified the drugs in our dataset using the class “pharmacological subgroup” from the Anatomical Therapeutic Chemical ( ATC ) classification system . Subsequently , we mapped the resulting set via the drug name ( including synonyms ) , if possible , or via pharmacological subgroup otherwise to the files “clinical_ann” and “pheno_ann” of PharmGKB . The “clinical” links are manually curated annotations of clinically relevant pharmacogenetic variant—drug pairs and “pheno” connections link a gene variant and an affected phenotype . The positive set of our investigation consisted of all drug-gene associations from PharmGKB and the negative set of all possible drug-gene combinations of this mapping where no pharmacogenetic interaction was reported in PharmGKB . Altogether we analysed 616 . 955 and 4 . 401 . 175 drug-gene pairs by calculating the enrichment over random ( lift ) of clinical and phenotypic pharmacogenetic associations in these pairs , respectively . We moreover checked the distribution of the quantity of data points by binning according to the phenotypic similarity score and calculating the natural logarithm of the number of drug-gene pairs per bin ( S7 Fig ) . We checked if the side effects most similar to gene phenotypic traits of the most phenotypic similar drug-gene pairs are enriched in a manually curated dataset of known side effect-protein relationships published recently [19] . For every drug-gene pair , we calculated the fraction of known gene-adverse effect associations among the side effects sharing similar phenotypic traits to the mouse gene under consideration ( maximal 20 side effects ) . Subsequently , we compared the resulting fraction of causal gene-side effect relationships of the high-scoring pairs to the low-scoring ones . To provide experimental evidence for interactions of phenotypically similar drug-gene pairs , we compared our results to the hits of the ToxCast project [34 , 72] . We collected the hit annotations from the file AllResults_hitc_Matrix_141121 . csv , where 1860 compounds are tested in 822 assays . Subsequently , we annotated the assays to their intended target and mapped the targets to our set of genes and the compounds to our drug data set . Then , we calculated the precision of the resulting 14 , 238 drug-gene pairs in relation to their phenotypic similarity score . Furthermore , we manually investigated the 38 high scoring drug-gene pairs , which included 19 experimentally validated associations . In order to test the hypothesis that oxandrolone interacts with prokineticin receptor 2 ( PKR2 ) , the antagonistic and agonistic effects of oxandrolone were investigated in functional assays of PKR2 activity [39] . All the experiments were performed by the company CEREP . Oxandrolone was purchased from Sigma-Aldrich and Ehrenstorfer GmbH . In these assays the activity of PKR2 was traced measuring the intracellular Ca2+ by fluorimetry in HEK-293 cells expressing PKR2 . In the agonist experiments , Ca2+ mobilization after oxandrolone stimulation was measured and then the agonist effect of oxandrolone was calculated as a % of control response to the known reference agonist PK2 ( used at 10 nM concentration ) . The activity of oxandrolone on the antagonist assay was tested after the stimulation of PKR2 with the control reference agonist PK2 at a concentration of 2 nM . The antagonist effect was then calculated as a % inhibition of control reference agonist response . Oxandrolone was initially tested in both assays at a concentration of 1 . 0E-05M . At this concentration , we only detected activity on the antagonist assay ( 32 . 3% inhibition of control agonist response ) . We subsequently quantified the antagonistic effect of oxandrolone in a dose-response curve and determined the Kb ( dissociation constant ) value . We fitted the dose-response curve using the R function drm from the package drc ( cran . r-project . org/web/packages/drc/index . html ) with the four-parameter log-logistic function LL . 4 . To rule out the possibility of a non-specific effect of oxandrolone on Ca2+ concentrations on the antagonist assays , we additionally measured Ca2+ using fluorimetry in PKR2 expressing HEK-293 cells employing high concentrations of oxandrolone ( 1 . 0E-05 , 3 . 0E-05 , 1 . 0E-04 , 3 . 0E-04 ) in the absence of the known stimulant PK2 ( S9 Fig ) . | In order to avoid unwanted effects of current drug interventions , it is necessary to expand the knowledge of the molecular mechanisms related to drug action . Side effects offer insight into drug action , as for example similar side effects of unrelated drugs can be caused by their common off-targets . Moreover , the phenotypes of systematic single gene perturbation screenings in mice strongly contribute to the comprehension of gene function . Here , we present a novel approach that detects molecular interactions of drugs based on the phenotypic similarity of drugs and mouse models . The method is benchmarked with diverse data sets including drug-target interactions as well as gene-drug links of pharmacogenetic associations and validated by in vitro assays . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"drug",
"therapy",
"medicine",
"and",
"health",
"sciences",
"model",
"organisms",
"drug",
"research",
"and",
"development",
"pharmaceutics",
"phenotypes",
"genetics",
"adverse",
"reactions",
"pharmacology",
"biology",
"and",
"life",
"sciences",
"mouse",
"models",
"pharmacogenetics",
"drug",
"interactions",
"drug",
"information",
"drug",
"discovery",
"research",
"and",
"analysis",
"methods"
] | 2016 | A Novel Drug-Mouse Phenotypic Similarity Method Detects Molecular Determinants of Drug Effects |
Dengue viruses are endemic across most tropical and subtropical regions . Because no proven vaccines are available , dengue prevention is primarily accomplished through controlling the mosquito vector Aedes aegypti . While dispersal distance is generally believed to be ∼100 m , patterns of dispersion may vary in urban areas due to landscape features acting as barriers or corridors to dispersal . Anthropogenic features ultimately affect the flow of genes affecting vector competence and insecticide resistance . Therefore , a thorough understanding of what parameters impact dispersal is essential for efficient implementation of any mosquito population suppression program . Population replacement and genetic control strategies currently under consideration are also dependent upon a thorough understanding of mosquito dispersal in urban settings . We examined the effect of a major highway on dispersal patterns over a 2 year period . A . aegypti larvae were collected on the east and west sides of Uriah Butler Highway ( UBH ) to examine any effect UBH may have on the observed population structure in the Charlieville neighborhood in Trinidad , West Indies . A panel of nine microsatellites , two SNPs and a 710 bp sequence of mtDNA cytochrome oxidase subunit 1 ( CO1 ) were used for the molecular analyses of the samples . Three CO1 haplotypes were identified , one of which was only found on the east side of the road in 2006 and 2007 . AMOVA using mtCO1 and nuclear markers revealed significant differentiation between the east- and west-side collections . Our results indicate that anthropogenic barriers to A . aegypti dispersal exist in urban environments and should be considered when implementing control programs during dengue outbreaks and population suppression or replacement programs .
Anthropogenic assisted invasions by non-indigenous insect vectors of human disease have and will continue to have profound effects on global health [1] . In addition , anthropogenic land use changes can represent primary drivers of infectious disease epidemics and significantly alter disease transmission dynamics [2] . The dengue and yellow fever vector mosquito , Aedes aegypti , is a remarkably successful invasive species . A highly anthropophilic form likely emerged in North Africa within the past 2–4 thousand years and has subsequently been transported via human efforts to most subtropical and tropical regions worldwide [3] , [4] . Approximately two-fifths of the world's population is at risk for dengue infection and an estimated 500 , 000 people are affected by dengue hemorrhagic fever ( DHF ) annually , with fatality rates exceeding 20% when proper treatment is unavailable [5] . Dengue is widely distributed in the tropics , occurring in Central and South America , South and Southeast Asia , Africa , the Caribbean , and Pacific regions [6] . During the past 45 years the incidence of dengue infection has steadily increased throughout the globe as greater numbers of people permanently migrate to cities with continued growth and urbanization [7] . Aedes aegypti population dynamics in urban areas is subject to daily as well as seasonal meteorological variability [8] . The interaction between temperature , relative humidity and rainfall impact adult survival and availability of oviposition sites . The goal of A . aegypti control programs is to reduce the population density of adult mosquitoes below a critical threshold where epidemic dengue transmission is unlikely to occur [9] . Vector population suppression programs most often involve the elimination or insecticide treatment of larval habitats that are typically man-made containers located within or around houses . During epidemic outbreaks , ultra low volume ( ULV ) spraying of insecticides is often used as an emergency control measure to reduce the adult mosquito population [10] . Critical to the long-term success of any A . aegypti population suppression method is the influence of dispersion patterns of adult mosquitoes . A greater understanding of factors limiting adult dispersal would allow health agencies to be more efficient in allocating resources to vector control programs . Moreover , considerable interest exists in developing novel dengue control strategies through the development of genetically modified A . aegypti incapable of transmitting dengue virus ( DENV ) and their subsequent introduction into the field as part of a population replacement program [11] , [12] . A thorough understanding of dispersal behavior in urban environments is essential to successful implementation of any control strategy . Population structure of A . aegypti is complex , varies by region and scale , and can be influenced by environment and geography [13]–[20] . Urban estimates of genetic differentiation have varied in part due to environmental conditions and dispersal patterns [21]–[24] . Typically , adult A . aegypti mosquitoes travel relatively short distances of up to ∼100 m , although longer dispersal estimates of ∼800 m have been observed [25]–[29] . In Queensland , Australia , a mark-release-recapture study reported that A . aegypti would readily cross smaller , quieter roads , but significantly fewer crossed a major highway near the release point , and concluded that busy roads may have impeded dispersal [30] . Similar observations were made with bumblebees ( Bombus impatiens and B . affinis ) , where Bhattacharya et al . [31] reported high fidelity between the bumblebees and their foraging sites and that they would rarely cross nearby roads or railways . These observations indicate the possibility that habitat fragmentation due to roads or other anthropogenic environmental manipulations may act as significant barriers to migration of A . aegypti and other insects . In urban environments anthropogenic landscape features can result in habitat fragmentation and thereby influence dispersal patterns of mosquitoes . Ultimately , these features can affect the flow of genes conditioning vector competence and insecticide resistance . In the current study , we report evidence of limited A . aegypti movement across an expansive 4 lane split highway in an urban environment in Trinidad , West Indies , as evidenced by mitochondrial and nuclear molecular analyses .
A . aegypti larvae were collected within urban breeding sites along a 900 m length of Uriah Butler Highway ( UBH ) on the east and west sides in the Charlieville neighborhood of Chauguanas . UBH is the major north-south highway in western Trinidad , extending from San Fernando in the south to east of Port of Spain ( Figs . 1 & 2 ) . The distance between buildings on the east and west sides of UBH ranged from ∼80 m to ∼130 m and the two sides were connected by a walking overpass and on-off ramps on both ends of the sampling area . Charlieville is a diverse urban neighborhood with mixed commercial , industrial , and residential buildings clustered closely together . Larvae were collected during October of 2006 and 2007 with assistance from field technicians within the Insect Vector Control Division at the Ministry of Health . In 2006 , larvae were collected from 18 larval habitat sites and in 2007 collections were taken from nine sites . Our samples were collected from diverse , but typical , larval habitats including water storage drums , folded sheets of commercial plastic , small buckets , and neglected and disused auto parts . Larvae were preserved in ethanol , and carried to the University of Notre Dame for genotyping . DNA was extracted from the mosquito samples using a standard phenol-chloroform method [32] . In 2006 , 147 larvae were genotyped; 58 from the western side of UBH and 89 on the eastern side while in 2007 , DNA was extracted from a total of 83 larvae; 33 from the west and 50 from the east . A 710-basepair region of the cytochrome c oxidase subunit I gene ( CO1 ) was amplified using CO1-specific universal primers [33] . Five µl of template DNA was amplified by polymerase chain reaction ( PCR ) in a 25 . 0 µl reaction containing 1X Taq buffer ( 10 mM KCl , 2 mM Tris , pH 9 . 0 , 0 . 02% TritonX ) , 1 . 5 mM MgCl2 , 0 . 4 mM each dATP , dCTP , dGTP , dTTP , 5 pmoles of each primer , and 1 U Taq DNA polymerase . The thermocycle conditions were 94°C for 1 min , 4 cycles at 94°C for 1 min , 45°C for 1 . 5 min , followed by 34 cycles at 94°C for 1 min , 50°C for 1 . 5 min , and 72°C for 1 min with a final extension period at 72°C for 5 minutes . Haplotypes were identified by examining banding patterns using Single Strand Conformation Polymorphism ( SSCP ) as per [34] . Briefly , 5 . 0 µl of PCR product was mixed with 3 . 0 µl of Denaturing Loading Mix ( DLM ) , that consisted of 0 . 1 ml of 1N NaOH , 9 . 5 ml formamide , 0 . 005 g of bromophenol blue , 0 . 005 g of xylene cyanol , and brought up to 10 ml with ddH20 . The mixture was denatured at 95°C for 5 min and snap cooled on ice . Approximately 7 . 0 µl of PCR product-DLM were used for electrophoresis on 42×33 cm 5% polyacrylamide gels for 3–4 hours at 30 milliamps . Gels were stained using silver nitrate solution with protocol adapted from Promega's GenePrint® STR Systems ( Promega U . S . , Madison , WI ) to visualize DNA banding patterns . Confirmation of haplotypes was accomplished by sequencing 5–10 individuals of each haplotype from both strands using the same CO1 primers . A total of 11 markers including nine microsatellite loci [35] , [36] and two Single Nucleotide Polymorphism ( SNP ) loci [37] were used for genotyping . PCR amplification was performed with genomic DNA isolated from individual mosquitoes in 25 µl volumes as described above . PCR reactions for microsatellite loci AC2 , AG2 , AG7 , A10 , B19 , CT2 , H08 , M201 , M313 were performed under the following conditions: 94°C for five minutes , followed by 30 cycles of 94°C for 1 min , 60°C anneal for 1 min , 72°C extension for 2 min , and a final 72°C extension for 10 min . PCR conditions for SNP loci LF178 and RT6 were performed at: 94°C for 5 minutes , followed by 39 cycles at 94°C for 20 sec , 55°C for 20 sec , and 72°C for 30 sec , and a final 72°C extension for 10 min . SNP products LF178 and RT6 were digested with RsaI and MnlI respectively and size fractionated in 3% agarose gels and visualized with ethidium bromide under UV light . Polymorphisms in microsatellite loci were resolved and analyzed using a Beckman-Coulter CEQ8000 and Beckman-Coulter CEQ8000 software . Briefly , microsatellites were amplified using dye-labeled primers ( Sigma Proligo , Sigma-Aldrich Inc . , St . Louis , MO ) and pooled into groups of 3 loci . Pools consisted of 0 . 4 µl of 400 bp standard , and 30 . 0 µl of standard loading solution ( Beckman-Coulter Inc . , Fullerton , CA ) with 1 . 0 µl of diluted amplified product added to each well . Conformance with Hardy-Weinberg equilibrium ( HWE ) , gametic disequilibrium between pairs of loci in each population , and inbreeding coefficients ( FIS ) were computed on FSTAT version 2 . 9 . 3 . 2 [38] . Presence of null alleles was examined using Micro-Checker [39] . Mitochondrial sequences were aligned using SEQMAN from the Lasergene package ( DNASTAR Inc . , Madison , WI ) and analyzed using DnaSP [40] . Population structure was examined using locus by locus Analysis of Molecular Variance ( AMOVA ) for nuclear markers and mtDNA haplotypes were analyzed using standard AMOVA on Arlequin version 3 . 1 [41] .
The mitochondrial CO1 gene showed sequence polymorphisms among the samples . A total of 3 haplotypes were identified within the 177 individuals in our test samples ( Fig . 3 ) . Mitochondrial CO1 haplotypes-1 and 2 were the most common across populations and years , accounting for ∼42% and ∼52% of total known haplotypes . Haplotype-3 comprised ∼5% of the total , but was unique to the eastern side of the road in 2006 and 2007 samples and was never detected on the western side of UBH . Haplotype-2 was the only haplotype detected from the 2007-west population ( Fig . 3 ) . Haplotype frequencies were compared spatially ( east compared to west ) and temporally ( 2006 compared to 2007 ) . There was significant differentiation between mosquitoes collected on the east side of the road and those collected on the west side of the road in 2006 and 2007 ( Table 1 ) . Temporally , the 2006-west population differed from the 2007-west population , however on the east side of the road there was no genetic differentiation between years . Estimated FST values were moderate to large , ranging from 0 . 042 to 0 . 390 . Our data suggested relatively lower FST values for spatial samples than temporal samples . The east and west side samples show FST 0 . 172 ( year 2006 ) and 0 . 390 ( year 2007 ) and the 2006 and 2007 samples show FST 0 . 249 ( east ) and 0 . 49 ( west ) . Five out of 44 tests ( 11% ) were significant for deviation from Hardy-Weinberg equilibrium after Bonferroni correction ( Table 2 ) . Deviations in expected heterozygosity were due to heterozygosity deficits in locus B19 ( 2006-east and 2006-west collections ) , and locus M313 ( 2006-east and 2007-east collections ) . Deviation at locus M201 in the 2007-west population was a result of heterozygote excess . Null alleles were identified at loci LF178 and RT6 in the 2006-east and 2007-east populations , respectively and in locus A10 in both 2007 populations , but the loci were in HW equilibrium . Evidence for null alleles was present in locus M313 in all but the 2007-west population , and locus B19 had high levels of null alleles in all populations , and both B19 and M313 deviated from HW equilibrium . Gametic disequilibrium analysis revealed significant disequilibrium between loci H08 and A10 and loci H08 and AG7 . The H08 and AG7 loci are physically linked on chromosome II , while locus A10 is located on chromosome III . Due to the presence of null alleles at loci B19 , M313 , and A10 , and gametic disequilibrium with locus H08 , these markers were removed prior to AMOVA . Four of the five microsatellites had private alleles that were specific to either the 2006-east , 2006-west , 2007-east or 2007-west collections . Neither SNP locus had private alleles . Markers AG2 and CT2 had alleles that were present in a population at a frequency between 5–10% ( Fig . 4 & 5 ) . In marker AG2 , 9 of the 16 alleles were private to at least one collection , however most had a frequency <3% . Allele AG2-O was present only in the 2006-west collection at 6 . 5% . Allele AG2-D was present in 2006-east , 2006-west , and in 2007 was only present in the 2007-east collection at a frequency of 10 . 9% ( Fig . 4 ) . Similarly allele AG2-N was present in both 2006 collections ( east and west ) , but unlike allele AG2-D was absent in the 2007-east collection and present in the 2007-west collection at 10 . 3% ( Fig . 4 ) . Six alleles were found in CT2 , 3 were private to at least one collection ( Fig . 5 ) . Allele CT2-D was present in the 2006-east ( 5 . 4% ) collection , 2007-east ( 4 . 2% ) , and 2007-west ( 3 . 9% ) , but was not present in the 2006-west collection ( Fig . 4 ) . In the remaining 2 microsatellites ( AC2 and AG7 ) alleles that were private in at least one collection existed , but the frequency ranged from ∼2% to ∼4% ( Fig . 5 & 6 ) . Overall , the amount of variation contained between populations was small , ranging from 0 . 2% to 1 . 4% of the total variation . AMOVA found small but significant FST estimates between collections on the east and west side of UBH in 2006 and 2007 , ranging from 0 . 011 to 0 . 021 ( Table 3 ) .
Aedes aegypti is highly adapted to a peridomestic environmental niche which has enabled it to spread throughout most large tropical cities in the world [42] . After confirmation that A . aegypti was the primary vector of yellow fever virus and dengue virus , disease prevention programs focused on the control of the mosquito vectors [43] . The pinnacle of dengue control began with the decision by PAHO to eradicate A . aegypti from the western hemisphere using a top down control structure and effective use of insecticides [42] , [44] , [45] . Incidences of A . aegypti transmitted diseases were greatly reduced along with the distribution of A . aegypti in the late 1950s to mid-1970s . However , in the late-1970s control programs were disbanded in part to financial considerations and the realization that unless a global campaign to eradicate A . aegypti was undertaken any attempts to eliminate the mosquito from the Americas would be unsuccessful due to the increased frequency and speed of air travel and other transportation options capable of transporting A . aegypti eggs and adults [46] , [47] . Further complicating eradication efforts was the emergence of resistance to insecticides in the mid-1950s [46] . As a result in areas where A . aegypti was once eliminated , reinfestation and outbreaks of dengue eventually followed where vigilant vector surveillance and control was not implemented [43] , [47] , [48] . In addition , the availability of a highly effective vaccine for yellow fever likely contributed to a decline in active mosquito surveillance programs in areas certified as A . aegypti free . Our micro-geographic analysis of genetic variability in A . aegypti from Trinidad was studied using microsatellite and SNP nuclear markers , and mitochondrial CO1 sequences . The distribution of the 3 CO1 haplotypes is strongly indicative of UBH acting as a barrier to dispersal . We were unable to detect haplotype-3 on the west side of the road in either 2006 or 2007 . This is interesting because the distance between collections on the east and west side of UBH ranged from ∼80 m to ∼130 m , which given dispersal estimates ranging from 100 m to 800 m , should not have limited adult dispersal potential and mosquitoes with haplotype-3 would likely be expected to have colonized the west side of UBH unless they were unable to successfully transect the highway . FST estimates from AMOVA also revealed significant differentiation between populations collected on the east and west side of UBH in 2006 and 2007 . Results from nuclear marker analysis therefore showed the same pattern of differentiation as the mitochondrial sequence data; however the magnitude of the FST and amount of variation was much lower ( Table 3 ) . Two explanations , neither mutually exclusive may explain the discrepancy in magnitude between the class of markers . The first is due to the preservation of diversity as a result of A . aegypti utilizing heterogeneous larval habitats in Trinidad . A large number of alternative habitats that are suitable for mosquito production may have gone undetected by surveyors in the Charlieville neighborhood as indicated by the large numbers of alleles detected ( Fig . 4 , 5 , 6 & 7 ) . Of the 25 containers from which we collected larvae , 10 were from containers other than water storage drums . In Trinidad water storage drums are the main source of A . aegypti production , however surveys have shown that A . aegypti mosquitoes will utilize a wide range of permanent and semi-permanent containers and the types of containers used can depend upon the degree of urbanization [49] , [50] . The second explanation is the existence of homoplasy in the microsatellites which has been proposed as an explanation for observed differences in differentiation between SNP and mtDNA marker in A . aegypti populations in Venezuela [13] . If present , homoplasy would underestimate the amount of differentiation between the collections [51] . We also observed temporal differences in haplotype frequency between 2006 and 2007 , although no significant population structure was found between 2006 and 2007 on the east side of UBH ( Table 1 ) . In all four groups of mosquitoes there was a decrease in the frequency of CO1 haplotype-1 and increase in haplotype-2 ( Fig . 3 ) . One possible explanation for this observation is that the resulting change in haplotype frequencies is a consequence of vector control efforts conducted periodically by the Ministry of Health in Trinidad . Changes in haplotype frequency could also be due to normal temporal fluctuations in the mosquito population . If the former is the primary cause affecting haplotype frequency one would expect to see reduced heterozygosity in the nuclear markers , which was not observed . Although there was not a physical barrier separating the 2006 and 2007 collections the significant differentiation observed between the 2006 west and 2007 west population was not unexpected , as previous examinations of temporal variation in A . aegypti populations have also reported changes in genetic differentiation over seasons , which were most likely due to changes in mosquito density , availability of oviposition sites , and the type of environment [21] , [23] , [52] . In Phnom Penh , Cambodia , genetic differentiation in A . aegypti populations was influenced by seasonality and environment type ( urban vs . suburban vs . rural ) , with significant differentiation occurring within the city [23] . The authors suggested that ideal urban conditions , including an abundant supply of hosts and oviposition sites , limited the need for dispersal by adult mosquitoes . In suburban areas differentiation was in part dependent upon the physical environment with variations in human density , availability of running water , and rural versus residential developments impacting the population structure . Dispersal range is an important aspect of dengue transmission and much research has been conducted attempting to determine how far A . aegypti adults travel , however large variations in daily and lifetime dispersal rates have been reported . Larger estimates of dispersal have reported mosquitoes traveling >800 m [29] , [53] . Many studies using mark-release-recapture methods have reported a shorter flight range of A . aegypti [25] , [54]–[56] . Examining mean distance traveled ( MDT ) and the flight range within which 50% ( FR50 ) and 90% ( FR90 ) of mosquitoes travel , as opposed to maximum distance traveled may be a more epidemiologically important parameter [29] . In a Kenyan village , McDonald and others [25] recaptured a majority of mosquitoes within the house they were released over 12 days . Marked mosquitoes released in a tire dump in New Delhi , India dispersal ranged between 50–200 m , but most were recaptured within 50 m of the release point [57] . Similarly , Muir and Kay [54] reported females having a MDT of 56 m and FR90 of 108 m . Released mosquitoes tended to cluster around houses with some dispersal towards adjacent houses and mosquitoes released on the perimeter of villages moved towards the center of the village [26] , [29] , [58]–[60] . The relatively large numbers and duration of DENV infected females captured in houses with confirmed dengue cases in Merida , Mexico may further indicate high fidelity between A . aegypti mosquitoes and place of pupal emergence [61] . Results from both classes of makers show strong evidence of limited gene flow across UBH , effectively fragmenting the populations on the east and west side of the highway . Mosquito dispersal patterns are nonrandom and influenced by environmental factors as reported by Sheppard et al . [62] and Hausermann et al . [63] in A . aegypti mosquitoes using mark-release-recapture methods . Furthermore , Chadee [64] indicated that prevailing weather patterns may potentially influence dispersion . Range of dispersal is dependent upon a mosquito's ability to remain in flight and the availability and abundance of shelter , food sources , hosts for blood meals and suitable oviposition sites [62] . Suitable host availability may reduce dispersal as reported by Suwonkerd et al . [65] where fewer A . aegypti mosquitoes exited a hut when a human host was present than with controls consisting of a dog or no host . Edman et al . [66] reported that when an abundance of suitable oviposition sites were available dispersion of female A . aegypti mosquitoes was reduced . Although the distance across the highway is well within dispersal estimates for A . aegypti , lack of cover and shade may have made UBH a harsh environment for mosquitoes to transect . This is supported by Tun-Lin et al . [67] who reported shade as a significant factor impacting the presence of A . aegypti in premise surveys and Russell et al . [30] reported that released A . aegypti dispersal patterns were nonrandom with more mosquitoes being recaptured along a corridor with heavy shading from trees and vegetation . Furthermore , oviposition sites were most likely minimal , even along peripheral ditches and nonexistent blood meal hosts may have dissuaded migration across UBH and prevented a stepping stone model of colonization from occurring over UBH . Ecological features including accessible water and availability of oviposition sites , vegetation patterns , humidity , and housing density contribute to determining the distribution of A . aegypti mosquitoes . The effects of topographic features of urban environments are not fully understood , however Reiter et al . [27] noted that buildings were not an impediment to A . aegypti flight . Our results indicate that urban landscape features do contain barriers to dispersal , and thereby affect the population structure of mosquitoes . This information could be used by vector control agencies to more efficiently target mosquito populations for suppression . Control programs can divide an urban area into zones of control along landscape features that are large enough to impede dispersal . This technique allows for the possibility of local elimination of A . aegypti moquitoes , barring or at least minimizing re-infestation due to the active transportation of the mosquito . Furthermore , during dengue outbreaks control agencies can more accurately target areas of higher risk along these same control zones . Difficulties in vaccine development [68] and the sequencing of the entire A . aegypti genome [69] have shifted some research efforts to preventing illness by developing applications that make use of transgenic mosquitoes incapable of transmitting the virus . Central to the successful use of transgenic mosquitoes to replace competent vector populations or to effect population suppression/elimination is a thorough understanding of A . aegypti bionomics , answering the basic questions of how many mosquitoes need to be released , where is the best place for them to be released , and when should they be released [70] . Results from early efforts using the sterile insect technique ( SIT ) to eliminate mosquito populations could have been improved or were negatively influenced by incomplete knowledge of adult mosquito dispersal behavior in A . aegypti and Culex fatigans [71] , [72] . Anthropogenic landscape features may therefore have profound effects on the implementation of traditional as well as proposed novel genetic mosquito control programs . Yakob et al . [73] explored the dynamics of population suppression dynamics with SIT and insects engineered to carry a dominant lethal gene ( RIDL ) . Mathematical models indicated that dispersion parameters for A . aegypti were fundamental in the success of replacement efforts and that enhanced connectivity treatment and peripheral populations could result in increased densities of wild-type mosquitoes as a consequence of SIT programs . Natural and anthropogenic barriers may actively influence , either positively or negatively depending on the strategy , the success of population replacement or population reduction by limiting the effective range of gene flow . Understanding the role of landscape features on population dispersal is likely critical to achieving success with any A . aegypti control strategy . | Worldwide , 2 . 5 billion people are at risk for dengue infection , with no vaccine or treatment available . Thus dengue prevention is largely focused on controlling its mosquito vector , Aedes aegypti . Traditional mosquito control approaches typically include insecticide applications and breeding site source reduction . Presently , novel dengue control measures including the sterile insect technique and population replacement with dengue-incompetent transgenic mosquitoes are also being considered . Success of all population control programs is in part dependent upon understanding mosquito population ecology , including how anthropogenic effects on the urban landscape influence dispersal and expansion . We conducted a two year population genetic study examining how a major metropolitan highway impacts mosquito dispersal in Trinidad , West Indies . As evidenced by significant differentiation using both nuclear and mitochondrial DNA sequences , the highway acted as a significant barrier to dispersal . Our results suggest that anthropogenic landscape features can be used effectively to enhance population suppression/replacement measures by defining mosquito control zones along recognized landscape barriers that limit population dispersal . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"ecology/behavioral",
"ecology",
"infectious",
"diseases/viral",
"infections",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"ecology/spatial",
"and",
"landscape",
"ecology",
"ecology/population",
"ecology",
"genetics",
"and",
"genomics/population",
"genetics"
] | 2010 | Influence of Urban Landscapes on Population Dynamics in a Short-Distance Migrant Mosquito: Evidence for the Dengue Vector Aedes aegypti |
Modern genetic mapping is plagued by the “missing heritability” problem , which refers to the discordance between the estimated heritabilities of quantitative traits and the variance accounted for by mapped causative variants . One major potential explanation for the missing heritability is allelic heterogeneity , in which there are multiple causative variants at each causative gene with only a fraction having been identified . The majority of genome-wide association studies ( GWAS ) implicitly assume that a single SNP can explain all the variance for a causative locus . However , if allelic heterogeneity is prevalent , a substantial amount of genetic variance will remain unexplained . In this paper , we take a haplotype-based mapping approach and quantify the number of alleles segregating at each locus using a large set of 7922 eQTL contributing to regulatory variation in the Drosophila melanogaster female head . Not only does this study provide a comprehensive eQTL map for a major community genetic resource , the Drosophila Synthetic Population Resource , but it also provides a direct test of the allelic heterogeneity hypothesis . We find that 95% of cis-eQTLs and 78% of trans-eQTLs are due to multiple alleles , demonstrating that allelic heterogeneity is widespread in Drosophila eQTL . Allelic heterogeneity likely contributes significantly to the missing heritability problem common in GWAS studies .
Uncovering the genetic basis of quantitative phenotypes is a central , yet unresolved problem in biology . There is a major discrepancy between the heritability estimates of most quantitative traits and the amount of heritable variation accounted for by all variants localized to a causative site . This phenomenon is often referred to as the “missing heritability” problem . Several hypotheses have been offered as possible explanations , including widespread epistasis [1] , the infinitesimal model ( many , very small effect loci influencing the phenotype of interest that are difficult to detect statistically ) [2]–[4] , rare alleles of large effect , that are also statistically difficult to detect [5]–[7] , and widespread allelic heterogeneity ( many independent effects segregating at each causative locus ) [7] . This quest to understand the genetic basis of complex traits has given rise to a community-based strategy of creating freely-available genetic resource populations in model organisms such as mice [8]–[10] , Arabidopsis thaliana [11] , [12] , maize [13]–[16] , and Drosophila melanogaster [17]–[20] . Those organisms with the greatest genetic resources and with a community of researchers focused on a single system provide a logical starting point toward finding the missing heritability associated with quantitative phenotypes . In addition , the experimental designs of some of these resources are well suited to test different hypotheses for the sources of missing heritability . For example , Bloom et al . [21] used a large segregant pool from a two line yeast cross to demonstrate that epistasis is not a major contributor to the heritability of most traits . In particular , resources that have a well-defined multi-haplotype structure can be used to identify the extent of allelic heterogeneity [22] owing to the ability to estimate trait means for each haplotype at each mapped QTL . By focusing effort on these community resources , the hope is that we will gain a better understanding of the causes of missing heritability problem . Much of the genetic variation underlying whole organism phenotypes is thought to be due to regulatory variation , i . e . , variants influencing gene expression [23]–[26] . Causative loci are linked to whole organism phenotypes through the transcriptome , an interrelated network of transcripts whose abundances influence the resulting phenotype . The transcript abundances of most genes are quantitative traits themselves and have heritabilities comparable to typical whole-organism phenotypes [24] , [26] , [27] . Increasingly , expression quantitative trait locus ( eQTL ) mapping is being used to identify the source of genetic variation in transcript abundances with the ultimate goal of linking variation at the nucleotide level to variation in gene expression and to variation in visible phenotypes . Expression QTL studies have shown that most genes have local ( cis ) eQTL that tend to be located near the transcription start site and to be of fairly large effect . Distant regulatory effects ( trans-eQTL ) are more difficult to identify , likely because they are more numerous and are of smaller average effect , leaving a great deal of variation in transcript abundance unexplained [23] , [24] , [26] , [27] . There is a growing movement toward identifying the causative quantitative trait nucleotides ( QTN ) underlying cis-eQTL , often with the assumption there is a single causative site [28]–[30] . However , if most eQTL harbor allelic heterogeneity [31] , identifying a single causative variant will cause researchers to miss a significant portion of the genetic variation [7] . Here we describe transcriptome-wide mapping in female head tissue in the Drosophila Synthetic Population Resource ( DSPR ) [17] , [18] , one of the major genetic reference panels in the Drosophila model system . Our goals are two-fold . First , we aim to provide a comprehensive map of cis- and trans-eQTL for female head tissue in the DSPR . A key advantage of genetic reference panels is the potential to integrate phenotypes measured at multiple levels on genetically identical individuals . Incorporating eQTL data with visible phenotype data can increase mapping power and help users identify candidate genes [9] , [23] , [25] , [32] . Second , we use the large set of discovered eQTL to quantify the number of alleles segregating at each causative locus , providing an evaluation of the degree of allelic heterogeneity at both cis- and trans-eQTL . The hypothesis that allelic heterogeneity is prevalent in quantitative traits has not been tested directly , in part because it is difficult to do so using a genome-wide association ( GWAS ) framework . Within loci , linkage disequilibrium makes it very difficult to distinguish between two SNPs tagging two independent causative sites versus a single causative site . In addition , the step-wise regression approaches used , for example [2] , [33] , to identify multiple SNPs in a gene region associated with a phenotype lack power . The result is that the majority of GWAS that have identified multiple SNPs at a single locus using conditional analysis rarely identify more than two such SNPs despite very large sample sizes e . g . [2] but see [33] . In contrast , mapping in the DSPR and other multi-parental advanced generation intercross mapping panels take a haplotype based approach , providing a natural way to distinguish between multiple alleles at each QTL and a way to ascertain the potential contribution of allelic heterogeneity to the missing heritability problem .
We identified a total of 7922 eQTLs corresponding to 7850 transcripts out of a total of 11064 transcripts tested ( Figure 1 ) . Details for all eQTLs are in Table S1 . Of these , 7704 transcripts were associated with a single cis-eQTL , 71 were associated with both cis- and trans-eQTL , and 75 were associated with only trans-eQTL . A small percentage of eQTLs ( ∼7%; Table 1 ) were associated with only a single recombinant inbred line ( RIL ) population ( pA or pB; see methods ) , but for most eQTL fitting both pA and pB was necessary to explain the eQTL signal , indicating that causative variants were present in both populations . The amount of variation explained by our mapped eQTLs was high ( Figure 2 ) , though our stringent , experiment-wise permutation-based correction for multiple tests severely limits our ability to detect QTL of small effect . Not surprisingly , the variance explained by cis-eQTLs was higher than trans-eQTLs [24] . Our cis-eQTLs explained a median of 24% of the phenotypic variance , and 855 eQTL explained more than 50% of the phenotypic variance . Using our heritability estimates for each transcript abundance , we estimated the percentage of the heritability each eQTL explained . The median for the percent heritability explained by each eQTL was 73% . Our trans-eQTLs explained lower levels of variance , the median phenotypic variance explained was 15% , and the median percent heritability explained was 38% . However , if heritability values are underestimated , and/or we overestimate the effects of eQTLs ( which is likely due to the Beavis effect [34] ) , these values will be inflated . This effect is obvious for the set of eQTL estimated to explain greater than 100% of the heritability ( Figure 2A ) . Our mapping resolution was high ( Figure 3 ) . We used two methods for estimating confidence intervals , a 3 LOD drop and the Bayesian credible interval . We excluded confidence intervals that spanned centromeres or occurred near telomeres , because these tend to cover very large physical distances ( 7% of eQTLs ) . The Bayesian credible intervals tended to be narrower than 3 LOD drops ( median BCI = 110 kb , 0 . 25 cM; median 3 LOD drop = 240 kb , 0 . 51 cM ) , but the range was larger for BCIs ( BCI: 0–4 . 5 Mb , 0–6 . 5 cM; 3 LOD drop: 20 kb–4 . 0 Mb , 0 . 001–3 . 9 cM ) . The median number of genes within cis-eQTL CIs was 32 ( range 1–551 ) , and within trans-eQTL CIs , the median was 44 ( range: 5–479 ) . We have provided a comprehensive map of eQTLs for female head tissue in the Drosophila model system within the constraints of our statistical power . There is little doubt many smaller effect eQTLs exist that we were not able to identify given our conservative statistical threshold . Our use of trans-heterozygote individuals means that we not only avoid the effects of inbreeding depression , but we have also obtained estimates for all eQTL for both pA and pB DSPR populations . Overall , our results confirm what many other researchers have observed , widespread large effect cis-eQTLs and smaller effect trans-eQTLs [23] , [24] , [26] , [27] . One of the major advantages of a stable genetic panel is the ability to measure multiple traits at multiple levels on genetically identical individuals , which allows for the potential to combine these sources of data to identify causative genes [9] , [23] , [25] , [32] . We expect this dataset to be very useful to DSPR users , particularly those interrogating phenotypes measured in females with relevance to neuroanatomy or behavior . All of the raw and analyzed data are freely available at http://FlyRILs . org/Data . The data have also been deposited in NCBI's Gene Expression Omnibus [35] and are accessible through GEO Series accession number GSE52076 . We identified regions of the genome associated with a high trans-eQTL density to identify eQTL regulating the expression of several other genes ( trans hotspots ) . There were two regions of high trans-eQTL density , TQTLA and TQTLB ( Figure 4; Table 2 ) . These clusters regulate several genes distributed throughout the genome , as is apparent in Figure 1 . We used a gene ontology term finder [36] to determine whether the sets of genes regulated by these trans-eQTL were related to a common process . The set of 16 genes regulated by TQTLA showed enrichment for circadian rhythm of gene expression ( 2 of the 16 genes regulated by TQTLA; P = 0 . 0007 ) . We used principal components analysis on the set of 16 genes to create a composite variable . All 16 genes load fairly evenly on the first principal component ( absolute value range: 0 . 08–0 . 20 ) . We then correlated this composite variable with expression measures for each gene in the TQTLA region to identify possible candidate genes . The gene timeless ( tim ) was highly correlated with the TQTLA composite variable ( r = 0 . 90 ) , and it does have a significant cis-eQTL . All other genes in the interval had a correlation with an absolute value of less than 0 . 5 . Additionally , after correlating the expression of each of the 16 transcripts regulated by TQTLA with the expression of all genes in the TQTLA region , timeless showed the maximum pairwise correlation in all 16 cases ( absolute value of correlation range:0 . 35–0 . 84 ) . The estimated haplotype means follow this pattern and are correlated with the estimated effects for the timeless cis-eQTL in most cases ( average absolute value correlation: 0 . 65; min: 0 . 03; max: 0 . 99 ) . The gene timeless ( tim ) is expressed in the adult central nervous system [37] and is involved in transcriptional regulation of circadian rhythm [38] . Not all genes in the TQTLA interval are included in our expression set . For example , some genes may have been dropped due to the presence of SNPs in probes , or were not included in the Nimblegen probe set to begin with . For TQTLA , 23 genes in the interval are not represented in the expression set . However , none of these genes are associated with any terms involving circadian rhythm , regulation of gene expression , or transcription ( http://FlyBase . org ) [39] , and we therefore do not consider any of these likely candidate genes . The genes associated with TQTLB are enriched for several GO terms including compound eye pigmentation ( 2/11 genes; P = 0 . 005 ) , the umbrella term: single-organism metabolic process ( 6/11 genes; P = 0 . 007 ) , and several specific metabolic process terms: tryptophan metabolic process ( 2/11 genes; P = 0 . 008 ) , indolalkylamine metabolic process ( 2/11 genes; P = 0 . 0008 ) , indole-containing compound metabolic process ( 2/11 genes; P = 0 . 002 ) , aromatic amino acid family metabolic process ( 2/11 genes; P = 0 . 006 ) . Once again we performed PCA to create a composite variable . sugarbabe ( sug ) was the gene most highly correlated with the TQTLB composite variable ( r = −0 . 63 ) and does have a significant cis-eQTL . All other genes in the interval had a correlation with an absolute value of less than 0 . 4 . Loadings were again fairly even ( absolute value range for all other genes: 0 . 08–0 . 39 ) . Pairwise correlations between the transcripts associated with TQTLB and the expression measures in the interval showed sugarbabe to be most highly correlated in all cases except two: gene CG5321 and gene CG6834 ( absolute value of correlation range for all other genes: 0 . 40–0 . 52 ) . These two genes were also the two with the lowest loading values on the composite variable . The correlation between the estimated haplotype effects for the cis-eQTL for sugarbabe , and the effects for the trans-eQTLs were moderate ( mean absolute value correlation: 0 . 24; min: 0 . 005; max: 0 . 44 ) . The gene sugarbabe ( sug ) is expressed in the adult head [37] , is involved in regulation of transcription [40] , is involved in regulation of response to starvation [41] , and is part of the insulin-like growth factor signaling pathway [41] . The 21 genes not included in the interval are not associated with any terms involving metabolism , regulation of gene expression , or transcription ( http://FlyBase . org ) [39] . We have identified two trans hotspots , and , in both cases , we were able to use our expression dataset to narrow the causative gene to a single likely candidate gene . Previous eQTL studies have identified many more trans hotspots that regulate many more genes ( hundreds or thousands ) than our two identified hotspots ( TQTLA: 16 genes; TQTLB: 11 genes; e . g . [27] , [42] , reviewed in [24] , [26] ) . However , while some of these global regulators of gene expression have been confirmed as true signals , most notably in yeast [43] , [44] , Kang et al . [43] show how hotspots can result from confounding factors such as batch effects . In our dataset , we employed PCA to correct for possible batch effects [45] . This method has been shown to increase power to detect eQTL [29] , [45] , [46] , however , it makes identifying even true trans global regulators impossible . The signal that results from a global regulator is statistically indistinguishable from an unmeasured batch effect . In addition , even true global regulators can confound the detection of other true eQTLs , and correcting for these true global regulators increases the power to detect these other associations [43] , [45] . It is possible to distinguish true trans hotspots from batch effects using biological replicates [43] , but for our study we chose to maximize the number of RILs rather than increase replication to maximize our statistical power to map eQTL . As a result , we are unable to detect many trans hotspots in this study . However , our stringent statistical correction does give us increased confidence that the eQTL we do identify are indeed true signals . The vast majority of our eQTLs appear to be multiallelic ( Figure S1 ) . In 95% of cases , the number of alleles estimated at cis-eQTL was 3 or greater . For trans-eQTL this percentage was somewhat lower , at 78% . Figure 5 shows an example of an eQTL where the best model is a two allele model and of an eQTL where the full haplotype model is the best model . In cases where we estimated multiple alleles , we were able to explain additional phenotypic variance compared to the best two allele model ( Figure S2 ) , sometimes as much as an additional 27% . We investigated our ability to accurately estimate the number of alleles by performing a simulation designed to provide the highest power to distinguish between different alleles ( see methods ) . Our simulation revealed that our estimator underestimates the number of alleles in 63% of cases , correctly estimates the true number of alleles in 26% of cases , and overestimates the number of alleles in 10% of cases ( Figure 6 ) . This bias toward underestimating the number of alleles gets increasingly severe as the true number of alleles increases . Our simulations with a lower effect size ( 5% ) and normally distributed allelic effects both resulted in an even stronger bias toward underestimating the true number of alleles . Our allele number distribution for cis-eQTLs is no doubt composed of a mixture of eQTLs of varied numbers of true alleles . Overall , it is closest to the distribution we obtain for a simulation of ∼5 alleles . So while most of our estimates for cis-eQTL are for 3–4 alleles , many may be determined by many more alleles . Our results indicate widespread allelic heterogeneity for both cis- and trans-eQTLs . The focus of mapping studies is often to identify the single causative variant underlying a significant signal , the implicit assumption being that the causative loci are biallelic . cis-eQTL in particular , with their large effects , are thought to be more likely than other traits to have a simple genetic architecture and be biallelic [22] , [28]–[30] . Baud et al . [22] found some support for this idea when comparing a two allele model to the full haplotype model in hippocampus eQTLs in the heterogeneous stock mouse resource [32] . They found that in 97% of cases , the two allele model was superior for cis-eQTLs while trans-eQTLs were more likely to be multiallelic [22] . However , in contrast to these findings , cis-eQTLs have been found to be multiallelic in Drosophila [47] , Arabidopsis [42] , and humans [31] , [48] . Our results strongly confirm the result of multiallelism in Drosophila with 95% of cis-eQTLs estimated to be due to 3 or more alleles . This result indicates that in Drosophila , widespread allelic heterogeneity exists at one of the most basic levels of genetic variation: cis-regulatory variation . Widespread allelic heterogeneity is one potential explanation for the missing heritability problem in the study of complex traits . Allelic heterogeneity presents a statistical challenge for GWAS [7] . GWAS utilize natural populations and interrogate each SNP ( or other specific variant ) for association with the phenotype of interest . At the single gene level , it is difficult to distinguish between simple linkage disequilibrium between a single causative variant and other , nearby neutral SNPs , and multiple independent causative SNPs . If GWAS focus only on the strongest association at a locus , in the presence of allelic heterogeneity that individual variant will account for less of the variation than the entire gene , causing the effect of the locus to be underestimated [7] . In this respect , haplotype-based mapping approaches , such as the one described here , have an advantage because entire haplotypes ( and thus an entire set of causative variants associated with a single gene ) are tested together . The effect size associated with the causative gene will tend to be larger and easier to detect in this framework . This effect , combined with the more favorable frequencies of alleles in linkage based panels could explain why these studies tend to explain very large proportions of the heritable variation [9] , [21] , [49] , while GWAS grapple with large amounts of missing heritability . However , one drawback of current haplotype-based methods is that they do not have single gene resolution and therefore identifying the causative gene within the QTL interval can be a significant challenge . Furthermore , while identifying the causative loci under allelic heterogeneity is easier with haplotype based methods , the subsequent identification of the causative SNPs within the loci is made much more complicated by heterogeneity [17] , [18] , [50] . Allelic heterogeneity is typical for Mendelian diseases ( http://www . omim . org/ ) and it has been suggested as the likely model for quantitative traits [51] . There is a growing body of empirical [2] , [17] , [22] , [31] , [42] , [47] , [50] and theoretical [7] support for this idea . For example , one of the largest GWA studies found support for allelic heterogeneity for human height by identifying several cases of multiple SNPs likely associated with the same gene [2] . Even age related macular degeneration , the first successful GWA study [52] , has subsequently been shown to harbor multiple functional alleles [53]–[56] . Our results should therefore not be surprising . However , they do suggest the community should focus on developing experimental designs and analytical methods , e . g . , [7] , that function well under a model of allelic heterogeneity .
We used RILs from the DSPR ( http://FlyRILs . org ) to map genome-wide expression variation . The DSPR has been described in detail previously . Complete details of the development of the DSPR , founder whole genome re-sequencing , and RIL genotyping are described in [17] . The development of the hidden Markov model to infer the mosaic structure of the RILs and the power and mapping resolution of the DSPR for QTL mapping are described in [18] . Briefly , the DSPR is a multi-founder advanced intercross panel consisting of a set of over 1700 RILs of Drosophila melanogaster . Two 8-way synthetic populations ( pA and pB ) were created from two independent sets of 7 inbred founder lines ( A1–A7 or B1–B7 ) with one additional line ( AB8 ) shared by both populations . Each synthetic population was maintained as two independent replicate subpopulations ( pA1 and pA2 or pB1 and pB2 ) , kept at a large population size , and allowed to freely recombine for 50 generations . At generation 50 , each subpopulation gave rise to ∼500 RILs via 25 generations of full-sib mating . The genomes of the original fifteen inbred founder lines have been completely re-sequenced , and the complete underlying founder haplotype structure of all RILs in the panel has been determined via Restriction-Associated DNA ( RAD ) sequencing along with a hidden Markov model ( HMM ) . In order to avoid potentially mapping QTL for inbreeding depression , we phenotyped trans-heterozygote F1 individuals from crosses between pA females and pB males . The crosses were done to maintain the subpopulation structure by crossing pA1 to pB2 and pA2 to pB1 . In both cases , we arbitrarily crossed pA and pB RILs with the same line number ( i . e . , pA11*pB21 , … , pA1n*pB2n , pA21*pB11 , … , pA2n*pB1n ) . For each of 596 crosses , we generated 4–6 replicate cross vials containing 10 virgin pA females and 10 pB males and cleared the adults after 24–48 hours to maintain roughly equal larval density across experimental vials . Both the inbred RIL parents and the experimental trans-heterozygous cross progeny were raised on standard cornmeal-yeast-molasses media at 25°C , 50% relative humidity , and on a 12∶12 light∶dark regime . Progeny from each cross vial were allowed to emerge and mate in the source vial for 2–4 days . Then 250–300 females were harvested over CO2 from the multiple replicate vials . Since we did not isolate virgin females on eclosion , females are very likely mated . These experimental females were kept for 24 hours in fresh vials to minimize any effects of the anesthesia before the heads were isolated ( 3–5 days old ) . Heads were removed by transferring the females without anesthesia to a 50 ml conical bottom centrifuge tube , freezing in liquid nitrogen , vigorously vortexing , and sieving using dry ice-chilled brass analytical sieves ( mesh sizes 0 . 0165 and 0 . 0278 inches ) , separating heads from bodies and from legs and wings . Head samples were stored at −80°C until RNA isolation . We did not have any technical or biological replicates aside from the effect of pooling 250–300 individuals , collected from multiple source vials , for each sample . This was intentional because we are mainly interested in the variance among RILs . There were two exceptions to this lack of replication . Crosses A1 . 299×B2 . 299 and A1 . 350×B2 . 350 were prepared independently twice . RNA was isolated using TRIzol Reagent ( Life Technologies ) , cleaned up using RNeasy Mini spin columns ( Qiagen ) , concentrated—if necessary—using a vacuum centrifuge , and shipped to the Carver Center for Genomics Microarray Center at the University of Iowa for cDNA synthesis and array hybridization . We used Nimblegen 12×135 K arrays to assay genome-wide gene expression . These arrays assay 16 , 637 transcripts with eight 60 bp probes per transcript . Each array holds 12 different crosses . All data analysis was performed in R [57] . Initially , we performed standard quantile normalization and corrected for background effects using the normalize and backgroundCorrect functions in the oligo package to correct for any overall array effects [58]–[61] . We then created a custom probe-to-transcript map using the most recent version of the CDS file available at FlyBase ( v . 5 . 48 ) . We blasted all probe sequences against the CDS , requiring an exact match [62] , [63] . We eliminated any probe sequences without an exact 60 bp match to a transcript ( 6842 probes ) . We did not require a unique match given many transcripts from the same gene share portions of their sequences . Thus a single probe can correspond to multiple transcripts . Single nucleotide polymorphisms in probe sequences are known to affect array hybridization and thus expression measurement [64]–[68] . We took advantage of the availability of full genome sequences for all 15 founder lines to identify SNPs within probe sequences . We first updated the alignment and SNP calling for the founder re-sequencing data using the Burrows-Wheeler Aligner ( BWA ) [69] with the following switches: -m 50000000 -R 5000 , followed by the SAMtools [70] mpileup command ( the initial alignment used Mosaik and a custom SNP caller , see [17] ) to obtain an accurate , comprehensive list of SNPs in the founder lines ( http://FlyRILs . org/Data , Release 3 ) . We also applied the following filters: 1 ) at least one founder was fixed for the minor allele and at least three founders were fixed for the major allele ( given a coverage of 10× ) , 2 ) minimum overall coverage of 90 ( 5 per sample ) , and 3 ) maximum overall coverage of 3600 . A large proportion of our probe sequences contained SNPs segregating in the set of DSPR founder lines . Because we have the full genome sequences in silico of all RILs in the panel , we were able to identify all positions in probes that are SNPs in our RIL panel and test for the effect of each SNP on the expression measurement . We discarded any probes containing multiple SNPs ( 22018 probes ) . For probes containing a single SNP , we used the haplotype probabilities from the hidden Markov model to infer the probability each RIL harbored the minor allele and assigned a genotype value to each cross by adding the paternal and maternal probabilities . In the case of perfect certainty , genotype values are: 2 = AA , 1 = Aa , and 0 = aa . We then tested for the effect of the SNP on the expression measurement by fitting the following model:where y is the expression measurement , S is subpopulation , M is the cross genotype at the marker , and βs and βm are the corresponding effect estimates . We then eliminated all probes with a p-value less than 0 . 05 ( 21141 probes ) . Following re-mapping of probes and elimination of probes with SNPs affecting expression , transcripts were associated with a variable number of probes instead of each transcript being associated with exactly 8 probes as in the original NimbleGen array design . We eliminated any transcript associated with fewer than four probes . Next , we performed standard RMA using the basicRMA function in the oligo package [61] to combine probe-specific data and generate a single expression measure per transcript . Many genes are associated with multiple transcripts . Whether the expression of different transcripts can be independently assessed is dependent on how many probes uniquely map to each transcript . We calculated pairwise correlations between each transcript in each set of transcripts associated with a single gene . If all of the pairwise correlations between the set of transcripts were > = 0 . 95 , we used the average expression for the gene . Otherwise , we mapped each transcript separately . We will refer to all expression measures ( including those averaged across transcripts for a single gene ) simply as transcripts for clarity . We followed the methods of [29] , [46] and used principal components analysis ( PCA ) to minimize batch effects [45] and increase our power to detect QTL . Following quantile normalization of each transcript to coerce each transcript distribution to be normal , we performed PCA on the entire set of transcripts . We selected the first 10 principal components to correct our expression measurements . The percentage of the variance explained by each remaining principal component was below 1% ( Figure S3 ) . We then fit the following modelwhere yi is the ith expression measurement , S is subpopulation , xj is the jth principal component , and βs , i and βj are the corresponding effect estimates . We used the resulting residuals for the remaining analyses . We performed an additional round of quantile normalization on these residuals to ensure normality . We estimated the narrow-sense heritabilities for all transcripts by fitting a linear mixed model using the polygenic function in the GenABEL package [71] . Briefly , the model includes a random effect polygenic term whose variance is determined by the kinship matrix between RIL crosses . We calculated the kinship matrix using the genome-wide haplotype assignments resulting from the HMM . At each position spaced every 0 . 025 cM , we calculated the probability of identity by decent and averaged these across the genome to obtain the relationship coefficient . Our kinship matrix is thus estimated over genetic distance . We then used the polygenic function to calculate heritabilities for each transcript [71] . To map eQTLs , we first selected transcripts expressed above background levels . We utilized the two replicated samples , A1 . 299×B2 . 299 and A1 . 350×B2 . 350 , to identify the point where measurements were less repeatable and excluded all transcripts with expression levels below this point ( Figure S4 ) . This cutoff excluded approximately 23% of transcripts . For all included transcripts , we performed haplotype-based genome scans by fitting the following model at regularly spaced positions every 10 KB across the genome ( 11768 positions; http://FlyRILs . org/Data , Release 3 ) . where yr , i is the ith transcript , μ is the grand mean , GA , j are the genotype probabilities for the jth paternal RIL , GB , j are the genotype probabilities for the jth maternal RIL , and βA , j , and βB , j are the corresponding effect estimates . Because we assayed only females , the model for the X chromosome is the same as for the autosomes . At each position , we calculated the F-statistic for the overall effect of genotype and obtained LOD scores . To identify the statistical significance threshold , we performed 1000 permutations of the expression measures [72] . The entire set of expression measures was permuted together to maintain the correlation structure in the dataset . We used these permutations to determine a conservative genome-wide , experiment-wise 5% significance threshold ( threshold = 14 . 99 ) . We also determined a separate threshold for cis-eQTL . We defined cis-eQTL as QTL occurring within 1 . 5 cM of the transcription start [18] site for each transcript ( 1 . 5 cM is our typical confidence interval width ) . To define a cis-only threshold , we only included the LOD scores for the positions within 1 . 5 cM of the transcription start for each gene ( threshold = 14 . 4 ) . We identified all peaks with LOD scores exceeding the above-defined thresholds . When multiple nearby peaks were identified , we determined whether their 3 LOD drop intervals overlapped , and , if so , only the peak with the highest LOD score was retained . We expect 3 LOD drops to be a conservative estimate of the 95% confidence interval . Standard 2 LOD drops have been shown to be overly narrow for pA×pB cross designs [18] . It should be noted however , that confidence intervals on QTL locations are not true 95% confidence intervals and effect size , sample size , and the number of haplotypes in the model affect the degree of coverage . We also calculated Bayes credible intervals , for which 95% coverage tends to be more consistent [73] , [74] . In a pA×pB cross , a mapped QTL may be due to genomic variation at that position in only one population or in both . We identified peaks associated with only a single population using Akaike's Information Criterion ( AIC ) . We calculated the AIC for three models: pA alone , pB alone , and pA & pB . The smallest AIC indicates the model with the best fit . Thus any cases in which the lowest AIC resulted from a reduced model , the QTL peak was concluded to be due to variation in a single population . We identified trans-eQTLs influencing multiple transcripts by estimating the trans-eQTL density across the genome using a 500 kb sliding window with a step size of 1 kb . Our estimate of density included only unique genes , not transcripts to avoid counting multiple transcripts associated with a single gene as independent events . If trans-eQTL density in a window exceeded the density expected by chance under a Poisson distribution , we concluded it was a significant trans hotspot . This threshold for a Poisson distribution given the total number of trans-eQTLs ( 147 ) , the window size ( 500 kb ) , the size of the genome tested ( 118 Mb ) and the Bonferonni corrected P-value threshold ( 117 , 741 tests; P = 4 . 2×10−7 ) is a trans-eQTL density greater than 6 . We delineated the size of these hotspot regions as the lowermost and uppermost confidence interval bound for any trans-eQTL peak included in a window exceeding a density of 6 . Our initial scan identified 3 trans hotspots but upon further investigation , we determined one to be a false signal resulting from a single gene family . All of the eQTL peaks associated with this hotspot represent 13 members of a single gene family located on the X chromosome: Stellate ( Ste ) . In addition , members of this family also occur at an unlocalized region in the heterochromatin on the X chromosome . The “trans-” eQTL we map regulating this family is located at the very tip of the X chromosome , making it very likely we are tagging this heterochromatic location of Stellate members , and it is in fact an additional cis effect . In fact , all thirteen members show two peaks , one cis peak and a second “trans” peak at the tip of the X , indicating most of our probes for these genes are tagging multiple members of this gene family . In addition , Stellate is expressed in adult males and involved in spermatogenesis ( http://FlyBase . org ) [39] . It is likely we are seeing high expression due to large numbers of copies of gene family members ( ∼200 copies ) [75] . We therefore excluded this trans hotspot . We estimated the number of alleles at each eQTL using a model comparison technique similar to the method employed by Yalcin et al . [76] and Baud et al . [22] The major difference in our approach is that we consider models with more than 2 alleles and do not restrict our analysis to specific SNPs in the QTL interval . The merge analysis employed by Baud et al . [22] considered all two allele models associated with a single SNP within the QTL interval . We simply assign different alleles to different haplotypes without those necessarily corresponding to SNPs in the interval . This method also allows us to consider models with several alleles . For each eQTL , at the peak position , we fit all possible models for different numbers of alleles , fitting a maximum of 11337 models at each eQTL . We first estimated the haplotype means at the peak , sorted these means , and then fit all possible models that did not change the order of the haplotype means for 2 , 3 , 4 , 5 , 6 , 7 , 8 , and 16 ( the full model allowing different estimates for AB8 in pA RILs and AB8 in pB RILs ) alleles ( Figure S5 ) . We only included haplotypes at the peak that occurred at least 5 times ( at a probability of greater than 95% ) in our set of crosses . Haplotypes at lower frequencies lead to inaccurate estimates of haplotype means with large standard errors . For each possible allele grouping , individual founder haplotype probabilities in each allele group were summed to obtain a probability each RIL harbored each allele group . For example , if haplotypes A3 and A5 are grouped as a single allele named allele 1 , and the probabilities a given RIL cross harbors the A3 or A5 haplotype are 0 . 90 and 0 . 03 respectively , then the probability that RIL cross harbors allele 1 is 0 . 93 ( i . e . , the probability the RIL cross harbors either A3 OR A5 and thus allele 1 ) . Alleles were only combined within pA and within pB given that the pA and pB sets of probabilities are independent . The model fit was as follows:where yr , i is the ith transcript , μ is the grand mean , na is the number of pA allele groupings , nb is the number of pB allele groupings , GA , c are the genotype probabilities for the cth paternal allele group , GB , d are the genotype probabilities for the dth maternal allele group , and βA , c , and βB , d are the corresponding effect estimates . The model with the lowest P-value was chosen as the best model and the number of alleles associated with this model was recorded . We also explored using Akaike's information criterion ( AIC ) to choose the best model , however simulations revealed a higher error rate using AIC ( see below ) . Table S3 provides hard coded genotype assignments for all RIL crosses at all significant eQTL . To test our method of estimating the number of alleles associated with QTL , we simulated QTL stemming from between 2 and 15 different alleles and subsequently estimated the number of alleles using the model comparison methodology described above . We intentionally set up this simulation to make distinguishing different alleles as easy as possible . We performed 1000 iterations for each of 2 , 3 , 4 , 5 , 6 , 7 , 8 and 15 alleles ( the full model assuming the same effect for AB8 in the pA and pB panels ) . For each iteration , we randomly selected 600 pA RILs and 600 pB RILs from the DSPR panel and randomly paired them to create pA-pB crosses . We then simulated a QTL in this set of RIL crosses at a randomly selected position in the genome with the chosen number of alleles . We assigned the different alleles equal effects , because we found equal effects gave higher power to distinguish different alleles compared to pulling effects from a normal distribution ( Figure S6 ) . For example , for a four allele model each founder haplotype was randomly assigned an effect of 1 , 2 , 3 , or , 4 . We assumed an additive model to calculate a genetic effect for each cross . We generated a set of random normal deviates N ( μ = 0 , ) to correspond to environmental variance where z = the percent of the phenotypic variance explained by the QTL and is the genetic variance at the QTL . The percent of the total phenotypic variance explained by the QTL was randomly chosen from our observed distribution of phenotypic variance explained by cis-eQTLs . These effects tend to be quite large , however , we found large effects lead to higher power to distinguish different alleles ( Figure S7 ) . We then estimated the number of alleles at our simulated QTL as described above . We used two methods to determine the best model: 1 ) the model with the lowest P-value , and 2 ) the model with the lowest AIC . Our results showed the method using P-values had a greater accuracy ( P-value method: 26% accuracy; AIC method: 19% accuracy ) . More importantly , the AIC method overestimates the true number of alleles more often , estimating more than two alleles in 83% of cases when the true number of alleles is two ( Table S2 ) . We prefer the method that is more conservative , meaning it has a greater tendency to underestimate rather than overestimate the number of alleles , and we therefore use the P-value method in all subsequent analysis ( Figure S8 ) . Complete sensitivity information for the different methods and the different simulation models can be seen in Figures S5 , S6 , S7 and in Table S2 . | For traits with complex genetic inheritance it has generally proven very difficult to identify the majority of the specific causative variants involved . A range of hypotheses have been put forward to explain this so-called “missing heritability” . One idea—allelic heterogeneity , where genes each harbor multiple different causative variants—has received little attention , because it is difficult to detect with most genetic mapping designs . Here we make use of a panel of Drosophila melanogaster lines derived from multiple founders , allowing us to directly test for the presence of multiple alleles at a large set of genetic loci influencing gene expression . We find that the vast majority of loci harbor more than two functional alleles , demonstrating extensive allelic heterogeneity at the level of gene expression and suggesting that such heterogeneity is an important factor determining the genetic basis of complex trait variation in general . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Methods"
] | [
"model",
"organisms",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"research",
"and",
"analysis",
"methods"
] | 2014 | Genetic Dissection of the Drosophila melanogaster Female Head Transcriptome Reveals Widespread Allelic Heterogeneity |
There is much interest in characterizing the variation in a human individual , because this may elucidate what contributes significantly to a person's phenotype , thereby enabling personalized genomics . We focus here on the variants in a person's ‘exome , ’ which is the set of exons in a genome , because the exome is believed to harbor much of the functional variation . We provide an analysis of the ∼12 , 500 variants that affect the protein coding portion of an individual's genome . We identified ∼10 , 400 nonsynonymous single nucleotide polymorphisms ( nsSNPs ) in this individual , of which ∼15–20% are rare in the human population . We predict ∼1 , 500 nsSNPs affect protein function and these tend be heterozygous , rare , or novel . Of the ∼700 coding indels , approximately half tend to have lengths that are a multiple of three , which causes insertions/deletions of amino acids in the corresponding protein , rather than introducing frameshifts . Coding indels also occur frequently at the termini of genes , so even if an indel causes a frameshift , an alternative start or stop site in the gene can still be used to make a functional protein . In summary , we reduced the set of ∼12 , 500 nonsilent coding variants by ∼8-fold to a set of variants that are most likely to have major effects on their proteins' functions . This is our first glimpse of an individual's exome and a snapshot of the current state of personalized genomics . The majority of coding variants in this individual are common and appear to be functionally neutral . Our results also indicate that some variants can be used to improve the current NCBI human reference genome . As more genomes are sequenced , many rare variants and non-SNP variants will be discovered . We present an approach to analyze the coding variation in humans by proposing multiple bioinformatic methods to hone in on possible functional variation .
Genetic variation in the protein-coding portion of genes is of significant interest in the study of human health . The focus on coding exons , or the ‘exome’ , is due to the common belief that the exome harbors the most functional variation [1] . This is based on the observation that mutations that cause Mendelian diseases occur primarily in genes [1] , [2] . Mutations that cause amino acid substitutions , including changes to nonsense codons , in their respective genes are the most frequent type of disease mutation ( ∼60% ) [1] . In addition , small indels in genes account for almost a quarter of the mutations in Mendelian disease [1] , [2] . Meanwhile , less than 1% of Mendelian disease mutations have been found in regulatory regions . For complex diseases , such as Alzheimer's , obesity , and heart disease , it is unknown how much variation in genes will contribute to disease , compared to variation in regulatory regions [3] , [4] . There have been many efforts to re-sequence genes to identify and characterize gene variation in humans [5]–[11] . One of the proposals of the 1000 Genomes Project , an international collaboration that aims to sequence one thousand genomes , focuses specifically on re-sequencing coding exons [12] . Additionally , many groups are developing technology for high-throughput resequencing of exons [13]–[15] . Because there has been much progress in sequencing individual human genomes [16]–[20] , our understanding of functional variation is an important step towards an era of personalized medicine , where a doctor could inform patients' of their disease susceptibilities based on their genome sequences . If the exome harbors much of the functional variation responsible for a person's phenotype , then identification and characterization of the individual's variation in the exome could enable individualized genomics . In this study , we focus our analysis on the exome of an individual human by providing a detailed characterization of the variants in protein-coding regions . We present the analysis of the coding variants in an exome from a diploid human genome assembly , which was termed HuRef [16] . This paper analyzes the different types of nonsilent coding variants . There are ∼12 , 500 coding variants that change protein sequence in the HuRef genome . We show that most of the variation in this individual is common and appears to be functionally neutral . Furthermore , we are able to reduce the ∼12 , 500 coding variants down to ∼1 , 600 variants that potentially affect protein function and may be involved in phenotypic effects . To the best of our knowledge , this is the first analysis of the exome of an individual human , and may serve as a benchmark for future studies on the variation in human exomes . There are several aspects to this study . The first is to describe a snapshot of what personalized genomics means today . If a person was to have his genome sequenced today , we show what insights about the individual could be gleaned from the protein coding component alone . Another aspect is that we can use the HuRef variants to improve the current NCBI reference genome , which will simplify future analysis of additional genomes . The final aspect is how one could characterize the variation obtained from sequencing many individual human genomes , and the approaches that could be developed to mitigate these studies . This is our first glimpse of an individual's human genome and as additional genomes are sequenced , many rare variants and non-SNP variants will be discovered . From this first exome , we can see what challenges one might encounter and propose approaches to face these challenges . Thus we present one approach for the analysis of coding variation in a human by detecting different trends for each variant type and demonstrating what phenotypes can be interpreted with our current knowledge .
This individual has 10 , 389 nsSNPs , of which 5 , 604 are heterozygous and 4 , 785 are homozygous ( Table 1 ) , where homozygous variants are loci where the alleles differ from the NCBI reference genome , but are the same within the HuRef assembly . It has been previously estimated that the number of heterozygous nsSNPs in an individual ranges from 24 , 000 to 40 , 000 [6]; our observed value of 5 , 604 is much less than this . This estimate was based on the nonsynonymous substitution rate based on a small number of genes , and extrapolating this value across all genes . The overestimate is partly due to the assumption of the human genome having 45 , 000–100 , 000 genes , but even if we assume the human genome has 20 , 000–30 , 000 genes , the estimate remains 1 . 5–2× higher than what we report . Possible explanations for the discrepancy is that the substitution rate in genes is extremely variable due to differences in local rates of mutation and recombination [6] . Thus we believe our number to be more accurate because it examines all genes rather than extrapolating from a small gene set . The nsSNPs account for a little more than half of the coding SNPs in the diploid genome ( Table 1 ) . The 1∶1 ratio of nonsynonymous SNPs to synonymous SNPs agrees with previously published reports [6] , [8] . Approximately 7% of the nsSNPs were not found in dbSNP and are thus novel . We expect novel SNPs to be rare [21]–[23] and hence observed on a single chromosome in an individual . This was affirmed with the observation that 72% of the novel nsSNPs are heterozygous . We wanted to find nsSNPs in this individual that may be undergoing negative selection . We use allele frequency in the human population as an indicator that a variant might be under negative selection . According to the theory of natural selection , functionally neutral mutations can reach high minor allele frequencies , whereas deleterious mutations will be selected against and remain rare in a population . Rare variants do not necessarily have to be deleterious; they can be recent mutational events . Variants that are neutral , slightly deleterious , or under positive selection can become common in a population . To see what proportion of nsSNPs may be undergoing negative selection , we retrieved the allele frequencies of these nsSNPs from the HapMap Project [24] , [25] ( Figure 1 ) . The majority of HuRef nsSNPs with known allele frequencies are common ( > = 0 . 05 ) . For 79% of the homozygous nsSNPs , the NCBI human genome has the minor allele ( MAF<0 . 5 ) . Therefore , the homozygous nsSNPs in HuRef tend to represent the major alleles in the human population and it is likely that the majority of these homozygous nsSNPs are neutral because they have reached such high frequencies . Also , 19% of the homozygous alleles in HuRef have an allele frequency of 1 , which suggests that NCBI contains a rare or erroneous allele at these positions . The majority of HuRef heterozygous SNPs are also common with only 9% of the nsSNPs with known allele frequencies being rare ( MAF<0 . 05 ) . Since we do not have allele frequencies for all the HuRef nsSNPs , we must estimate the proportion of rare nsSNPs ( MAF<0 . 05 ) in this individual . A previous simulation has estimated that ∼28% of heterozygous SNPs in an individual are rare , but that study made assumptions about human population size and its growth [26] . For 67% of the nsSNPs with known HapMap allele frequencies , we know the exact number of rare nsSNPs ( 56 homozygous and 326 heterozygous ) . For the remaining 33% of nsSNPs with unknown allele frequencies , we can estimate the proportion of rare nsSNPs based on the sequencing of a subset of heterozygous novel nsSNPs and the fraction of rare homozygous SNPs with known allele frequencies ( see Methods ) . Using this approach , we estimate ∼1 , 600–2 , 000 rare nsSNPs in this individual's genome , the lower bound takes into account the ∼25% false positive rate for novel SNPs ( see Methods ) . We conclude that ∼15–20% of the nsSNPs in an individual are rare , and ∼95% of the rare nsSNPs are heterozygous ( see Methods ) . The number of rare variants found in this individual may guide our expectations when we sequence additional genomes in the future . We wanted to identify the nsSNPs that may affect protein function and possibly be involved in human health and undergoing negative selection . Algorithms exist that predict whether an amino acid substitution affects protein function based on sequence conservation and/or structure [27]–[34] . When applied to human nsSNPs from re-sequencing projects , 0–30% of nsSNPs are predicted to affect function [9] , [31]–[33] , [35] . This range is based on datasets containing a relatively small number of nsSNPs ( ∼50–600 ) in a small number of genes ( ∼100–200 ) . What distinguishes our analysis from previous reports [9] , [31]–[33] , [35] is that we examine a single individual , rather than a population of individuals – thus we are establishing a benchmark for individualized genomics , as opposed to population genetics . Furthermore , we study all genes , unlike the previous studies that focused on certain classes of genes that were selected for their possible relevance in human health . For our study , we use the algorithm , SIFT ( Sorting Intolerant From Tolerant ) to determine if a nsSNP may affect protein function [33] . SIFT takes into account whether the amino acid change resulting from a nsSNP lies in the conserved region of the protein and the type of physiochemical change , and outputs a prediction to whether a nsSNP may affect protein function . We note that SIFT and other amino acid substitution prediction algorithms [27]–[34] only predict whether a nsSNP affects protein function . These algorithms do not predict whether a variant alters the processing or stability of transcripts . Approximately 75% of the HuRef nsSNPs had SIFT predictions ( see Methods ) , and 14% were predicted to impact protein function ( Figure 2 ) . This suggests that the majority of nsSNPs in this individual are functionally neutral . It also indicates that an individual has ∼1 , 500 ( 14% of 10 , 389 ) nsSNPs that affect protein function with deleterious effects , and we are able to confirm a previous estimate [34] . This previous estimate was obtained by taking the average nonsynonymous nucleotide substitution rate ( based on a small number of genes ) and extrapolating it for all genes . Meanwhile our estimate is based on the actual number of observed nsSNPs in an individual . The ∼1 , 500 nsSNPs predicted to affect protein function are deleterious in the evolutionary sense with each nsSNP having a selection coefficient s≈10−3 , on average [34] . A small selection coefficient suggests that the deleterious nsSNP either has negligible effects on health , has effects late in life after reproduction has occurred , or causes a disadvantage in certain environments . We term the nsSNPs predicted to affect protein function as predicted-protein-affecting nsSNPs . The other ∼9 , 000 nsSNPs are effectively neutral mutations with 0<s<10−4 , assuming an effective population size of 10 , 000 [36] . Thus , the effects of nsSNPs span a spectrum ranging from neutral to mildly deleterious , and the predicted-protein-affecting nsSNPs tend to be more detrimental . Heterozygous nsSNPs are two times more likely to be predicted as protein-affecting compared to the homozygous nsSNPs ( p<0 . 001; Figure 2 ) . We reason that predicted-protein-affecting nsSNPs are expected to be selected against and therefore , less likely to both reach high allele frequencies and be observed in homozygote form . Rare nsSNPs are also two times more likely to be predicted as protein-affecting compared to common nsSNPs ( p<0 . 001; Figure 2 ) and this trend has been reported previously [9] , [10] , [35] . This suggests that a higher proportion of the rare nsSNPs are deleterious , and are more likely undergoing negative selection . Also , a higher percentage of novel nsSNPs and nsSNPs with unknown allele frequencies are predicted to be protein affecting compared to all the nsSNPs ( Figure 2 ) , but this difference only holds for the heterozygous SNPs and not the homozygous nsSNPs ( Figure S1; Text S1 ) . Therefore , novel , rare , and heterozygous nsSNPs are more likely to affect protein function and cause phenotypic effects . Yet rare and novel nsSNPs are difficult to characterize because they are underpowered in whole-genome association studies [4] . Thus , one of the major challenges in the future of genomics is how to correlate rare and novel nsSNPs with phenotypes . There are 105 HuRef SNPs that result in premature termination codons , or stop codons , in 103 genes , hereafter referred to as PTC-SNPs . This corresponds to 0 . 5% of coding SNPs . These SNPs are expected to result in the loss of their respective proteins and hence be under strong negative selection . Yet when we retrieved allele frequencies for PTC-SNPs , all of the 36 PTC-SNPs with known allele frequencies were common , which shows that not all PTC-SNPs are under strong purifying selection . We investigate possible reasons for why some PTC-SNPs are not under strong negative selection and are able to reach high allele frequencies in the human population . Thirty percent ( 31/105 ) of the PTC-SNPs occur in segmental duplications in the human genome compared to 9 . 8% for synonymous SNPs . We assume gene redundancy would rescue these mutations , although loss of one copy of a gene can still have quantitative effects [37] . It is also possible that these PTC-SNPs have been mistakenly mapped due to the difficulty of assembling highly duplicated regions . We remove the PTC-SNPs in segmentally duplicated regions from consideration and 74 PTC-SNPs in 73 genes remain . A substantial fraction ( 42% ) of the remaining genes with PTC-SNPs are hypothetical . Hypothetical genes containing common PTC-SNPs may not be important to the human population , and using these variants may improve annotation of the human genome . If the PTC-SNPs in hypothetical genes are removed from consideration , 43 PTC-SNPs in 42 genes remain , and we sought to characterize these further . Because we saw that three times as many PTC-SNPs occur in segmental duplications than expected , we postulated that multiple copies of a gene may permit the existence of a PTC-SNP . We examined the size of the gene family for the remaining 42 genes , and found that the median gene family size is 6 , which is higher than the median gene family size for all genes , which is 2 ( p<0 . 001 ) . Thus , PTC-SNPs tend to occur in genes that have other homologues present in the genome , which may rescue the full or partial loss of a related gene [38] . There are only 9 PTC-SNPs in 8 genes that are non-hypothetical and unique members of their gene family . In general , PTC-SNPs tend to occur in hypothetical proteins , segmentally duplicated regions , and gene families with multiple members . In the future , we may be able to use these trends to prioritize which PTC-SNPs are most likely to have functional consequences . All the PTC SNPs in HuRef can be found in Table S1 , and none are in genes known to be involved in disease . Indels are the second most abundant type of genetic variation , following single nucleotide substitutions and account for almost a quarter of the genetic variation implicated in disease [2] . Coding indels can significantly impact their corresponding genes if they introduce frameshifts that lead to unfinished protein products . The HuRef genome contains a total of 739 coding indels , which consists of 281 heterozygous indels and 458 homozygous indels . To the best of our knowledge , this is the largest set of human coding indel variants identified to date [39] . We find an enrichment of indels that have sizes that are multiples of 3 in the HuRef coding indel set ( Figure 3 ) . We will refer to indels that have lengths divisible by 3 as 3n indels and indels with lengths not divisible by 3 as non-3n indels , where n is an integer . In coding regions , a non-3n indel would cause a frameshift that usually leads to a truncated protein product whereas a 3n indel would cause deletion or insertion of amino acid ( s ) . By comparing the diversity rates between coding indels and indels genome-wide ( Figure S2 ) , we find that 94% of non-3n coding indels have been eliminated by natural selection . In contrast , only 46% of 3n coding indels have been eliminated . This signifies that 3n coding indels are not as strongly selected against as non-3n indels . Many of the indels are due to polymorphism in tandem repeat sequence . Only 6% of coding regions are classified as tandem repeats , yet 52% ( 381/739 ) of all coding indels occur in tandem repeats . The majority ( 73% ) of the tandem repeats in coding regions have a periodicity of 3 and account for 66% ( 252/384 ) of the 3n bp coding indels . This suggests that local regions in a protein coding sequence can be prone to mutation that will either remove or insert amino acids into the protein product . In contrast to SNPs , many indels are not validated and their allele frequencies are unknown due to difficulty in their ascertainment using either sequencing or genotyping technologies [39] , [40] . To validate these indels , we verified if the indel was confirmed in the chimpanzee genome sequence [41] . We determined that 24% ( 181/739 ) of the coding indels correspond to the chimpanzee sequence . These indels are likely to be common in the human population if the indels occurred before the divergence of chimpanzee and human , although an alternative possibility is that some of these mutations are recurring events . This signifies that at least 24% of the HuRef coding indels are real . We sought explanations of how an individual human genome could have 739 indels affecting 607 genes , and yet the individual appears to lack severe phenotypic effects . Small coding indels ( < = 30 bp ) account for 84% ( 621/739 ) of the indels . In the following section , we analyze the 621 small coding indels ( < = 30 bp ) . Large indels are discussed in a later section . Many of the small coding indels were located at the N- and C-termini of their respective proteins . We calculated the relative position of the indel in the protein by dividing the indel's position by the total protein length . With this metric , one would expect that the indel's position would be uniformly distributed across the protein . Instead , indels tend to occur at the N- and C-termini of their proteins ( Figure 4 ) . If a coding indel occurs at the C-terminus of the protein , it may not affect the function of the protein because most of the protein product has been translated successfully . If a coding indel occurs at the N-terminus of the protein , this may be rescued by a downstream start codon in the coding region ( see Figure S3 for an example ) . This suggests that indels at the N- and C-termini of their proteins are functionally neutral and a future study using these indels to propose alternate start and stop sites could improve human gene annotation . Furthermore , a high proportion of homozygous coding indels were located near an exon boundary . A large proportion of the small homozygous coding indels were within 10 bp of the exon boundary: 27% ( 101/344 ) compared to 12% ( 34/277 ) for small heterozygous coding indels . Close inspection of the homozygous indels near exon boundaries showed that these indels were near small introns and the HuRef allele corrects the NCBI reference genome to provide a better gene model . Figure 5 shows an example where a 1-bp homozygous coding insertion borders a 2 bp intron , so rather than causing a frameshift , the small intron is replaced by an amino acid . Incorporating the homozygous indel from HuRef likely produces the correct protein sequence . Therefore , it is very likely that the HuRef assembly has the correct sequence , and could potentially be used to correct gene structures that were based on the NCBI human genome . Many of these small homozygous indels within 10 bp of an exon boundary were also found in chimp ( 61% ( 62/101 ) ) , further evidence that these indels are the ancestral alleles and likely to be accurate . We re-sequenced seven homozygous non-3n indels that were either near exon boundaries ( 4 ) , and/or confirmed by chimp ( 4 ) , and/or at the N-terminus of a protein ( 1 ) ( see Methods ) . These non-3n indels would supposedly cause frameshifts , yet all seven indels were found to be common ( MAF>0 . 05 ) and four were determined to have an allele frequency of 1 ( Table S2 ) . This indicates that NCBI has a rare or erroneous allele at these positions and suggests that homozygous indels located at certain locations may correct the NCBI genome sequence . We assume that indels near protein termini and near exon boundaries are functionally neutral . Removing these indels from the indel set reduces the entire set by 45% , with 342 indels remaining . Of the remaining indels , the fraction of indels that have length 3n increases from 49% ( 303/621 ) to 60% ( 205/342 ) . This suggests that while the termini of a protein may be able to tolerate the presence of non-3n bp indels , elsewhere in a protein there is a greater preference for 3n indels . We categorize the remaining indels by their 3n and non-3n lengths . Whereas 35% ( 71/205 ) of 3n indels are found in hypothetical proteins , 56% ( 77/137 ) of non-3n indels are found in hypothetical proteins ( p<0 . 001 ) . This suggests that non-3n indels occur in genes that can tolerate deleterious mutations , which may be pseudogenes or genes under weak selective constraints and we may be able to use these variants to identify genes that are likely not important for human health . We also noticed that the 3n coding indels that are not in tandem repeats tend to avoid regions of the protein that are highly conserved ( Figure S4 ) . In summary , many of the indels are located at exon boundaries or protein termini and these are likely to be functionally neutral . The remaining coding indels typically have sizes of 3n , and those that do not tend to occur in hypothetical proteins . To the best of our knowledge , these trends have not been previously reported , and this suggests that a substantial fraction of coding indels are functionally neutral . In the future , we can use the HuRef indels to improve gene annotation and our observations to develop a method that distinguishes between functional and neutral indels . We sought to understand genes missing large regions of coding sequence , leading to a gene and/or exon deletion that would render the gene non-functional . Therefore , we focused on deletions of coding sequence in this analysis . We will discuss newly observed genes and insertions of coding sequence in existing genes in a future manuscript . We identified 1 , 454 exons in 1 , 046 genes where at least half of the coding exon's sequence was missing from the HuRef assembly . Further investigation showed that genes with “missing” exons are most likely due to low coverage or assembly issues with repetitive regions rather than the human individual truly missing part of a gene ( see Methods and Text S1; Figure S5 ) . This was confirmed by resequencing a subset of the missing exons in the HuRef sample and validating that in fact , most of the “missing exons” are present in the HuRef sample ( see Methods and Text S1 ) . We examined the HuRef variants in genes known to be involved in disease ( based on the OMIM database [42] ) to make correlations between the HuRef variation and possible phenotypes . There are a total of 682 nsSNPs in 443 disease genes ( Table S3 ) . The allele frequency distribution and fraction of predicted-protein affecting SNPs are similar to non-disease genes ( Figure S6 ) . We examined the nsSNPs from dbSNP [43] that were found in the disease database OMIM [42] ( see Text S1 ) . Using this approach , seven HuRef nsSNPs were found to be associated with disease ( Table 2 ) . The HuRef individual is heterozygous for all seven SNPs and most of these disease-associated SNPs are common in the population . It may be considered surprising that these nsSNPs are common since they were found in the OMIM disease database . This is due to the fact that we looked for overlap with nsSNPs that are in both OMIM and dbSNP , and dbSNP tends to be biased for common SNPs [21] . From these seven variants , one could simplistically infer that the HuRef individual has an increased risk for eating disorder ( BDNF ) , 1 . 5-fold reduced risk to multiple myeloma ( LIG4 ) , an increased risk to prostate cancer which can be rescued by taking vitamin E supplements ( SOD2 ) , and allergic tendencies ( SPINK5 ) ( Table 2 ) . Some of the SNPs have known interactions with environmental factors ( SOD2 and BDNF in Table 2 ) . However , the published risks for these variants are from population-based studies , and may not apply to this specific individual because it does not take into account other interacting genetic loci and his environment [44]–[47] . Therefore , Table 2 does not show exact risks for this individual and predicting the phenotype or lack of phenotype of this individual is premature . These variants , like many of the risk variants being uncovered by genome-wide association studies , have low risks and we may not have a clear understanding of their clinical utility until all the relevant factors ( both genetic and environmental ) for a particular disease have been elucidated [44]–[47] . These seven well-studied examples demonstrate the complexity of trying to interpret variants and their impact , but there are still many variants in this individual that are uncharacterized . For the 682 nsSNPs in 443 disease genes , 27 are rare ( MAF<0 . 05 ) , 18 are novel , and 81 are predicted to affect function ( Table S3 ) . Interpretation of these variants is difficult because of the absence of literature for many of the observed variants . One challenge is that these variants , even if they affect protein function , could be phenotypically neutral in certain contexts ( see SOD2 and BDNF in Table 2 and [48] ) . Also , even if there is evidence that a nsSNP is under negative selection ( e . g . predicted to affect function and/or rare ) , it is not straightforward to interpret a possible phenotype because mutations at different locations in the same gene can have different effects [49] . The difficulty of inferring phenotypic consequences from a variant is depicted in the following example . rs562556 , which is homozygous in HuRef and has an unknown minor allele frequency , introduces the amino acid substitution V474I in PCSK9 , and this amino acid substitution is predicted to affect protein function . Because defects in PCSK9 cause familial hypercholesterolemia ( OMIM∶607786 ) , one could speculate that this SNP could affect the donor's cholesterol levels . However , extensive functional studies of this variant and others are necessary before any conclusions can be made . Because only 1% ( 7/682 ) of the nsSNPs in disease genes in this individual human have been well-characterized , this indicates that we are only at the beginning of relating genotypes to phenotypes , even for the well-characterized disease genes . There were 28 indels in 26 disease genes ( Table S4 ) . Only 3 out of the 28 indels have lengths that are not multiples of 3n , and would cause frameshifts . Two of these indels appear to be annotation issues and are likely to be functionally neutral ( see Text S1 ) . The third indel is in ACOX2 . This protein is involved in lipid metabolism , and patients with Zellweger syndrome lack this protein [50] . Because the indel is heterozygous in the HuRef individual and one functional copy is present , the individual may not be adversely affected by this mutation . For the remaining 25 indels with length 3n in 23 disease genes , 84% were in tandem repeats . Some of these genes are known to cause disease due to polyglutamine and polyalanine repeat expansions ( AR , ATXN2 , ATXN3 , HD , TBP ) . For these genes , we confirmed that the number of repeats in the HuRef genome falls within the range of what is observed for unaffected individuals . In addition , this individual is heterozygous for a 24-bp duplication in the CHIT1 gene , which activates a cryptic 3′ splice site and causes chitinase deficiency [51] . Even though the indel produces a nonfunctional protein , the indel is observed at an allele frequency of 23% in the general population . One possibility for its high incidence is that this indel may provide a selective advantage against fungal pathogens [51] . To examine selective constraints on different regions across the genome , we estimated the diversity rate θ ( see Methods ) . We calculate the diversity rates for regions throughout the entire genome as well as calculating the diversity values for genic regions ( Table 3 ) . Values of θ tend to be negatively correlated with the strength of selection , where low values of θ tend to indicate strong selective pressures , while high values do not . The order of θ is: coding in disease genes<coding<conserved noncoding intronic<splice sites<3′UTR≈5′UTR≈conserved noncoding intergenic<promoter<introns<repeats . This indicates that coding regions in disease regions are the most selectively constrained regions and repeats are the least . Some diversity values and trends have been reported in previous publications [6]–[8] , [52] , [53] , and our results are in agreement with these values . However , to our knowledge this is the first study with such an extensive list of regions . We also find indels are significantly under-represented in coding regions; there is a 43∶1 SNP∶indel ratio for coding regions compared to a 7∶1 SNP∶indel ratio genome-wide . It is reasonable to observe stronger selection against indels in coding regions where they can introduce frameshifts . We further explored if the observed HuRef variation could indicate whether certain genes were under stronger selective constraints than others . We identified 538 genes containing common nsSNPs located in conserved regions of the protein and are predicted to affect protein function . We called this set of genes the Commonly-Affected genes . In parallel , we identified 79 genes containing rare nsSNPs predicted to affect protein function , and we termed this set the Rarely-Affected genes . We hypothesized that Commonly-Affected genes would be under weaker selective constraints since predicted-protein-affecting SNPs , common in the human population , were found in these genes . Ka/Ks , is a metric used to quantify selection of a gene and is calculated as the ratio of the nonsynonymous ( amino-acid affecting ) substitution rate to the synonymous substitution rate . A high Ka/Ks ratio indicates that a gene is undergoing weak selection , although positive selection is also a possibility if Ka/Ks is >1 across the entire gene or part of the gene [54] . We observe that the Commonly-Affected genes tend to have higher Ka/Ks ratios than the Rarely-Affected genes ( p = 0 . 09; Figure 6 ) . This suggests that Commonly-Affected genes may not be under strong selective constraints . Presumably , mutations affecting gene function in the Commonly-Affected genes will not have significant consequences on human health , which is why these predicted-protein-affecting nsSNPs can rise to high allele frequencies . In support of this hypothesis , the Commonly-Affected genes are dominated by olfactory receptors ( 78/538 ) which is consistent with the previous observation that humans do not depend on olfaction to the same extent as other species [55] . In the future , extending this type of analysis to nsSNPs from HapMap or additional human genomes may allow identification of additional Commonly-Affected genes . This could help improve the scientific community's knowledge of the human genome by identifying which genes may not play an important role in human health if mutated . We also compared the genome properties of Dr . James Watson [20] to those of Dr . Craig Venter . SNP diversity rates based on Dr . Watson's genome are slightly elevated compared to Dr . Craig Venter's ( Table 3 ) , but in general , follow a similar trend as stated previously . The one exception is simple repeats , which have a 2-fold higher SNP diversity rate in Dr . Watson compared to Dr . Venter; this may be due to differences between sequencing technologies . Of the 3 . 1 million SNPs in Dr . Venter's genome that map to the NCBI genome , 56% are shared with Dr . Watson . There are 606 , 719 and the 288 , 723 novel SNPs that map to the NCBI human genome in Dr . Watson's and Dr . Venter's genomes , respectively . One possible reason for the smaller number of novel SNPs in Dr . Venter's genome is because Dr . Venter's genome is partially represented in the Celera human genome assembly [16] , [56] , and variants have been mined from the Celera assembly and subsequently deposited in dbSNP [57] . Of the novel SNPs , only 32 , 528 SNPs are shared between the two individuals . This demonstrates the value of sequencing additional human genomes to discover novel variants . We compare the number of nsSNPs and coding indels between the two exomes ( Table 4 ) . Similar numbers of nonsynonymous variants are detected in both individuals . However , 14% and 20% of nsSNPs are predicted to affect function in Dr . Venter's and Dr . Watson's exomes , respectively . One possible reason for this difference could reside in the different prediction algorithms employed [20] , [30] , [33] . Future analyses that use a standardized approach may clarify this apparent difference . The number of indels in Dr . Watson's genome is less than half of the number of indels we observe in HuRef ( Table 4 ) , most likely because 1 bp indels were discarded [20] . As more individuals are sequenced , scientists will be able to establish general trends and the average values for metrics that characterize a human individual's genome .
This is the first study of an individual's exome and we establish what one may expect to observe from the variation in the exome of an individual . We show that the majority of coding variants in a human are neutral or nearly neutral . This is not unexpected , since we know that this genome creates an individual who has survived past 60 years of age . We also find that within an individual , the basic principles of genetics are followed . Additionally , we examined the variation in genes known to be involved in disease , and found no indication that the individual should have a severe disease , which matches the phenotype currently known . Despite having a human's complete genome sequence , we are only at the tip of the iceberg for understanding how an individual's genotype and phenotype are related . One significant challenge is that the phenotypic effects of the majority of genes are unknown . Currently , only 7% of genes are annotated with OMIM disease associations so that it is difficult to predict the phenotypic effects of variants for a large proportion of genes . If one is to rank genes by importance and effect on phenotype , then based on the results of this study , one might consider that genes containing PTC-SNPs , frameshifting indels , and damaging nsSNPs that are common in the human population to be under weak selection , and variants in these genes may not be as relevant to human health . Several groups have used gene ontology , literature , and other sources to predict potential disease genes [67] , [68] , we propose that one can also use observed human variation to increase our understanding of the human genome . Even if a gene is known to be involved in disease , it is difficult to understand if a variant in the gene will have a phenotypic effect . We found that 99% of the nsSNPs in disease genes could not be characterized by current literature . Different mutations in the same gene can cause different phenotypic effects [49] , thus making it difficult to interpret possible phenotypes . Furthermore , some variants have phenotypic effects only under certain environments ( see SOD2 and BDNF in Table 2 and [48] ) . Also , when looking at complex phenotypes , multiple variants in coding and non-coding regions are likely to be involved [63]–[66] . This genetic complexity , as well as exposure to various environmental factors , will need to be taken into account in assessing risk for various diseases . How can geneticists start to grasp the significance of phenotype-genotype correlations ? This question is especially relevant to companies offering personalized genomics to their consumers ( e . g . 23andMe , Navigenics , deCODE Genetics ) . When looking amongst the human population , there are many rare SNPs , but when looking at a single human individual , the majority of the SNPs are common [6] , [69] , [70] , and in this study we estimate that >80% of the nsSNPs in an individual are common . Therefore , understanding which common variants are involved in common disease will greatly benefit an individual , because common variants account for a significant fraction of the variation in each human . Recent genome-wide association studies have identified common variants implicated in disease [71]–[73] and these studies will continue to find common disease-associated variants in the near future . These discoveries will be valuable for interpreting a large proportion of an individual's genome . However , one should be cautious in the interpretation of these variants because variants with low associated risks may not necessarily have good predictability in the clinical setting [44] , [45] . In contrast , rare variants are harder to study because genome-wide association studies are insufficiently powered to detect rare variants [4] . We found that a higher fraction of rare nsSNPs were predicted to affect protein function compared to common nsSNPs , in agreement with previous studies [9] , [10] , [35] . This suggests that a small proportion of a large number of common variants and a larger proportion of a small number of rare variants will contribute to the health of a human individual . Genome-wide association studies tend not to have the power to detect rare etiological variants [4] so that predicting whether a rare mutation in an individual causes disease without any other phenotypic information is extremely difficult . Therefore , one of the major future challenges in personalized genomics is the interpretation of the effects of rare variants found in an individual , especially if this information will be relayed back to the individual and could impact the person's lifestyle . In addition to interpretation and analysis , much effort was expended in ensuring that a variant was authentic ( see Methods ) . There could be unintended negative consequences for telling a person that they have a disease variant , when in actuality the variant could be detected in error . Common reasons for our false positive variant calls were technical sequencing error , low sequence coverage , and low-complexity sequence . When interpreting an individual's genome that can potentially impact a person's lifestyle , manual curation , editing , and verification by other technologies seems prudent . The human genome sequence has had a significant impact on research since its availability in 2001 [56] , [74] , but our analysis suggests that rare or erroneous alleles may have been incorporated into the NCBI human genome sequence , and that this sequence can be corrected and improved . Because one of the goals of the 1000 Genomes Project is to improve the human reference sequence [75] , our study points to where such improvements can be made . We find that ∼80% of the homozygous SNPs in HuRef tend to be the major allele in the population ( Figure 1 ) , and ∼20% of the homozygous HuRef nsSNPs had allele frequencies equal to 1 . Thus , at the majority of HuRef homozygous positions , the NCBI human reference sequence has the minor allele . If the scientific community sequenced many individuals , it could determine the major allele at each position in the genome . If the major allele was incorporated into the coding sequence , and this was used in subsequent gene prediction models , then the predominant form of the protein in the human population would be represented instead of a rarer form . Also , by using the common allele instead of the sometimes rarer NCBI allele , the number of perceived variation would be reduced when comparing human genomes . For example , if the NCBI human genome sequence were to incorporate the major allele and HuRef was then compared to this modified NCBI genome sequence , then we estimate the new number of HuRef homozygous SNPs genome-wide would be ∼300 , 000 ( 0 . 2 * 1 . 45 million [16] ) and the number of total variants between HuRef and the modified NCBI would reduce from 4 . 1 million [16] to 3 million . Similarly , if we remove the homozygous nsSNPs that correspond to the common allele in European population ( AF>0 . 5 ) , then the fraction of rare nsSNPs increases by 6% . This demonstrates the importance of the human genome reference sequence for the evaluation of variation in individual human genomes . The scientific community could also make use of coding indel variation to correct and improve gene annotation . Indels near exon boundaries appear to provide the correct gene models to give the appropriate protein product ( Figure 5 , for example ) [76] . Gene annotation for translation starts and stops can be further refined based on our observation that coding indels are found frequently at the N- or C-termini of their proteins . Frameshifting-indels at the N-termini of proteins could indicate that a translation start site further downstream may be the true start codon , or at least an alternative start codon can still yield a functional protein . Indels such as these may be polymorphic and so accounting for these indels could simplify future analyses on exomes as they could be quickly regarded as functionally neutral and reduce the total number of indels that need to be analyzed . We also observed a trend where common predicted-protein-affecting nsSNPs , PTC-SNPs and frameshift-inducing indels tend to occur in hypothetical genes . This suggests that these genes are not under strong selective pressures and mutations in these genes may not be relevant in the human population . Future studies with more human sequences identifying additional nonfunctional mutations in genes would help us confirm whether these genes are essential . Our exome analysis is currently limited to one individual . However , there will be significant benefits from sequencing many individuals whose phenotypes are known . One can envision collecting the genetic variation from these genomes and grouping individuals based on their respective phenotypes . Then for each phenotype , one may discover the genes which are involved in disease by looking collectively at the rare and common variants [11] , [77] , [78] . The analysis can also be strengthened by analyzing pathways instead of individual genes [78] . Furthermore , whole genome sequencing means we need not be limited to the exome . Using whole genomes , one could look for clustered mutations in conserved regulatory elements , especially since many association studies have found disease-associated loci in non-coding regions [63]–[66] . To assess the role of noncoding variation , we examined HuRef variation in and around genes involved in the melanoma pathway because the HuRef donor has reported a case of melanoma ( Figure S7 ) . We found that the majority of the variants occurred in conserved noncoding regions . This suggests that it may be insufficient to sequence just the exome and it is important to understand all types of variation , coding and noncoding , as well as interactions with the environment when studying phenotypes . We have filtered the initial set of ∼12 , 500 coding variants that affect protein sequence to a substantially smaller set that are most likely to have major effects on gene function ( Figure 7 ) . The trends that we have detected suggest we can reduce the number of putative functional coding variants by ∼8-fold and will provide a future guide for how one can analyze coding variants when additional human genomes are sequenced in the future . Additionally , the variants found here and in future studies may be used to improve our understanding of the human genome by correcting gene annotation and identifying genes not likely to be relevant to human health . We anticipate that this study will help guide the scientific community's expectations and experimental design in future genome sequencing projects .
We used the filtered variant set as described in [16] . We define homozygous variants as loci where the alleles differ from the NCBI reference genome , but are the same within the HuRef assembly . This variant set was used to generate the diversity values in Table 3 . For all other sections in this manuscript , quality inspection of variants was performed . To assure the quality of novel coding variants , we inspected manually the sequencing traces of novel heterozygous nsSNPs , all heterozygous PTC-SNPs , and all coding indels less than 20 bp in length . The sequence traces for these coding variants were extracted and three people independently reviewed the traces , by examining the quality of the traces and determining whether the variant was correctly called . If at least two people confirmed the existence of the variant , the variant was deemed acceptable , otherwise the variant was discarded . 35% ( 424/1196 ) of the novel heterozygous nsSNPs , 12% ( 9/73 ) of the heterozygous PTC-SNPs , and 33% ( 355/1088 ) of the coding indels were discarded . This may suggest that a significant fraction of the variants reported in [16] are dubious . However , this analysis is restricted to coding variation which is known to be under strong selection compared to the rest of the genome . Hence , there will be fewer real variants in coding regions and a higher proportion of the novel coding variants will be false positives . 20% ( 123/611 ) of the homozygous indels were reclassified as heterozygous because there was trace evidence for a second allele . The importance of filtering is demonstrated with the following observation . Manual inspection reduces the number of novel nsSNPs by a third , but especially filters out a higher proportion of predicted-protein-affecting nsSNPs . Prior to filtering , the number of novel heterozygous predicted-protein-affecting nsSNPs is 195 , after filtering this is more than halved to 89 novel nsSNPs predicted to affect protein function . SNPs not found in dbSNP v . 126 [43] were designated as novel . All allele frequencies were based on the CEU samples genotyped from the HapMap Project [24] , [25] , unless otherwise stated . Allele frequencies for 72% ( 3429/4785 ) of the homozygous nsSNPs and 63% ( 3544/5604 ) of the heterozygous nsSNPs were obtained . For heterozygous SNPs , we report the minor allele frequency ( MAF ) . For homozygous SNPs , we report the CEU allele frequency of the allele observed in the HuRef genome . We define common SNPs as SNPs with allele frequencies > = 0 . 05 and rare SNPs with allele frequencies <0 . 05 . The amino acid changes resulting from coding variants were determined by SNPClassifier , an internally developed software tool . The HuRef variants , their alleles , and positions in genomic coordinates , are provided as input into SNPClassifier . Annotation is automatically retrieved from Ensembl and is used to assign variants to defined gene categories . Variants in or near genes can be subtyped as: promoter ( 1 kb upstream of the transcription start site ) , intronic , 5′ UTR , 3′UTR , coding , or downstream of the transcript ( 1 kb ) . Coding SNPs are designated either as synonymous or nonsynonymous and coding indels are designated as either frameshift or amino acid insertions/deletions . The resulting protein product from coding indels that introduce frameshifts is also output . SIFT predictions for nonsynonymous SNPs were obtained by using SIFT 2 . 1 . 1 [33] . The protein sequences containing nonsynonymous SNPs were searched against SwissProt-Trembl 54 . Confidence in predictions is measured by the median sequence information , we used a cutoff of 3 . 5 for confidence . Approximately 75% ( 7 , 781/ 10 , 389 ) of the nsSNPs had SIFT predictions , the remaining 25% did not have a sufficient number of homologous sequences that are needed for prediction . We estimate the number of rare nsSNPs with allele frequency ( AF ) <0 . 05 in an individual . For the 67% nsSNPs with known AFs from the HapMap Project , there are 56 rare homozygous nsSNPs and 326 rare heterozygous nsSNPs . For the 1 , 356 homozygous nsSNPs with unknown AFs , the percentage predicted to affect function is similar to that seen for homozygous SNPs with known AFs ( Figure S1 ) . If the homozygous nsSNPs had a higher proportion rare SNPs , then a higher fraction should be predicted-protein-affecting but because they are similar , we assume that the homozygous nsSNPs with unknown AFs have a similar proportion of rare SNPs as the homozygous nsSNPs with known AFs . Because 1 . 6% ( 56/3429 ) of the homozygous nsSNPs with known AFs are rare , we estimate ∼22 ( 1 . 6% * 1 , 356 ) of the homozygous nsSNPs with unknown AF are rare , so in addition to the 56 rare homozygous nsSNPs with known AF , there is a total of ∼80 rare homozygous nsSNPs in this individual . For heterozygous nsSNPs , there are 326 heterozygous rare nsSNPs with known MAF , and 2 , 060 heterozygous nsSNPs with unknown MAF . From sequencing , as much as a quarter of the heterozygous nsSNPs with unknown MAF could be false positives , although this estimate is likely to be an upper bound ( see Sequencing for Variant Validation in Methods ) . Therefore the range of novel heterozygous nsSNPs falls within ∼1 , 550–2 , 060 . We also ascertained from sequencing that ∼75% of the heterozygous novel nsSNPs are rare . Therefore , we estimate that ∼1 , 200–1 , 550 of the heterozygous nsSNPs with unknown MAFs are rare and in total , there are ∼1 , 500–1 , 900 rare heterozygous nsSNPs . Thus , we estimate ∼1 , 600–2 , 000 rare nsSNPs in this individual's genome , and ∼95% of the rare nsSNPs are in heterozygous state . For an indel's location , we calculated the relative position of the indel in the protein by taking the first amino acid position affected by the indel . We divided the position by the total length of the protein , so that a relative protein position value close to 0 indicates that the indel affects the N-terminus of the protein , and a relative protein position value close to 1 indicates that the indel affects the C-terminus of the protein . We designate that an indel affects the N-terminus of a protein if the relative protein position is between 0 and 0 . 1; an indel affects the C-terminus of a protein if the relative protein position is between 0 . 9 and 1 . 0 . Thus , an indel is said to affect the N-terminus or C-terminus of the protein if it lies within the first 10% or last 10% of the open reading frame , respectively . To examine whether an indel occurs in a conserved region of the protein , the sequence alignment of the protein sequence with homologues from other organisms were retrieved from Ensembl . At every position in the protein alignment , sequence conservation was calculated [79] . The conservation value at the indel's position is compared with all other positions , and the percentile rank is calculated . If the number of sequences in the alignment was less than 10 , the data point was removed . The HuRef assembly was mapped by an assembly-to-assembly comparison to the NCBI build 36 human reference genome [16] . Regions in NCBI reference that were missing in the human diploid assembly were identified . We intersected the missing regions with coding exons greater than 50 bp in length and ensured that at least 50% of the exon was missing from the HuRef assembly in order to consider the exon . To double-check that the missing sequence was not in unassembled sequence , we searched the exonic sequence using MEGABLAST [80] against the HuRef assembled sequence and the unassembled singletons . MEGABLAST hits greater than 95% identify and with 50 bp minimum length were kept . We decided exons were not truly missing if >90% of its length were covered by these MEGABLAST hits . The final set consisted of 1 , 454 exons in 1 , 046 genes . We removed the genes located on sex chromosomes because the sex chromosomes are known to have low coverage [16] . After removing these genes , there were 719 genes with 880 missing or partial exons . To investigate read depth for this set of exons , we re-mapped all untrimmed reads from [16] to the set of exons using ‘snapper’ ( http://kmer . wiki . sourceforge . net/ ) , a seed-and-extend mapper . All 20-mer seeds were extended , and any alignments over 94% identity were reported . As a control , we also remapped reads to a set of exons that were randomly selected from all exons . Whereas the control exons show a normal distribution with the median number of reads centering at 7 . 6 , the missing exons show a bimodal distribution with either very few reads or many reads ( Figure S5b ) . This reflects that genes with “missing” exons are most likely due to assembly issues with repetitive regions or low coverage . 66% of the “missing” exons have an average read depth of less than 2 reads , which emphasizes the importance of adequate coverage in a human genome . We generated PCR primers to 15 regions in 12 genes with ‘missing exons’ , 9 PTC-SNPs , 15 coding indels , and 26 novel heterozygous nsSNPs ( Table S2 ) . These 65 PCR primers consistently amplified their cognate genomic regions in 46 unrelated CEU individuals and the HuRef sample . The DNA for the HuRef sample was extracted from whole blood ( see Methods in [16] ) . We sequenced the PCR products using Sanger dideoxy sequencing ( see [16] for sequencing protocol ) . The 46 unrelated CEU individuals were part of the HapMap CEU panel , and their Coriell identifiers are provided in Table S2 . Of the 26 heterozygous novel nsSNPs , 6 failed to be confirmed in the HuRef sample and instead matched the NCBI allele . There was also 1 nsSNP that failed to be confirmed in the HuRef sample but was observed in other samples and this was considered to be a false negative . This suggests that the false positive error for HuRef's novel nsSNPs is ∼25% ( 23% = 6/26 ) . This estimate is likely to be an upper bound due to the following reason . The 26 nsSNPs occurred in non-hypothetical genes , and nsSNPs in hypothetical genes may be under little or no selective pressures compared to nsSNPs in non-hypothetical proteins and the former can reach high allele frequencies . Hence , this false positive error may be inflated . For the novel nsSNPs that we could confirm in HuRef , the mean MAF of the novel SNPs was 0 . 09 and 74% ( 14/19 ) of the SNPs were rare ( MAF<0 . 05 ) . For the PCR products spanning missing exons , 14 regions from 11 genes were successfully amplified in the HuRef sample and this confirmed that HuRef is not missing exons for these genes ( Table S2 ) . In the 12th gene PRED58 , a 66 bp coding deletion in HuRef was observed but this was seen in all other DNA samples , suggesting that NCBI has the rare or erroneous allele . All genome coordinates are with respect to NCBI build 36 and all gene designations are with respect to Ensembl v . 41 . A gene was considered hypothetical if in its gene description , it had no description or was described as an “open reading frame” , an “orf” which signifies an open reading frame , a cDNA clone , putative , probable , uncharacterized , “similar to” another protein , a pseudogene , a fragment , hypothetical , a novel protein , novel transcript , or if it was invalid as described in Clamp et al . [81] . Under this classification , there were 20 , 561 non-hypothetical genes and 10 , 624 hypothetical genes consisting of 29 , 401 , 727 bp and 6 , 176 , 706 bp respectively . Ka/Ks is the ratio of the nonsynonymous substitution rate to the synonymous substitution rate . Ka/Ks values based on human-mouse orthologous gene pairs were retrieved from Ensembl Biomart . In Table 3 , constitutive exons are those coding exons that are expressed in 100% of the transcripts for its given gene . If a coding exon was present in <50% of the transcripts , it was designated as an alternative exon . Splice sites include the 20 bp within exon boundaries ( 10 bp intronic , 10 bp exonic for each exon boundary ) . Segmental duplication regions were taken from the UCSC genomicSuperDups file ( >1 kb length , >90% identity ) . TandemRepeatFinder [82] was used to designate tandem repeats using the parameters match = 2 , mismatch = 5 , delta = 5 , PM = 75 , PI = 20 , minscore = 35 . Non-genic conserved regions were taken from phastConsElements17way that were > = 50 bp in length . If any part of the PHAST region intersected with coding , 5′UTR , and/or 3′UTR , the PHAST conserved element was removed . Therefore , the conserved regions in Table 3 were not overlapping or bordering coding , 5′UTR , or 3′UTR regions . We estimate diversity θ [83] as θ = K/aL , where . K is the number of variants identified , L is the number of base pairs , and n is the number of alleles . For indels , K is the number of indel events . In the case of a single diploid genome , n = 2 , so a reduces to 1 . Then θ = K/L which is simply the number of heterozygous variants divided by the length sequenced . The 95% confidence interval for θ is [0 , θ+2θ] or [0 , 3θ] , as calculated in [16] . Diversity values were calculated for the various types of regions listed in Table 3 and the counts for these values can be found in Table S5 . We also attempted to look at the diversity values for gene ontology categories , but were unable to do so because of the low numbers of coding variants per gene ( data not shown ) . Diversity values for Dr . James Watson's genome were calculated using the 1 . 86 million heterozygous SNPs reported in [20] . For the denominator L , we assumed the entire chromosome was covered by reads and used the chromosome lengths from the UCSC genome browser . If this assumption is not true , then an inflated L will underestimate θ . Diversity values for indels were not calculated because indel data was not available for Dr . Watson's genome . | Characterizing the functional variation in an individual is an important step towards the era of personalized medicine . Protein-coding exons are thought to be especially enriched in functional variation . In 2007 , we published the genome sequence of J . Craig Venter . Here we analyze the genetic variation of J . Craig Venter's exome , focusing on variation in the coding portion of genes , which is thought to contribute significantly to a person's physical make-up . We survey ∼12 , 500 nonsilent coding variants and , by applying multiple bioinformatic approaches , we reduce the number of potential phenotypic variants by ∼8-fold . Our analysis provides a snapshot of the current state of personalized genomics . We find that <1% of variants are linked to any known phenotypes; this demonstrates the dearth of scientific knowledge for phenotype-genotype associations . However , ∼80% of an individual's nonsynonymous variants are commonly found in the human population and , because phenotypic associations to common variants will be elucidated via genome-wide association studies over the next few years , the capability to interpret personalized genomes will expand and evolve . As sequencing of individual genomes becomes more prevalent , the bioinformatic approaches we present in this study can be used as a paradigm to pursue the study of protein-coding variants for the genomes of many individuals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/genome",
"projects",
"genetics",
"and",
"genomics",
"genetics",
"and",
"genomics/bioinformatics",
"genetics",
"and",
"genomics/medical",
"genetics"
] | 2008 | Genetic Variation in an Individual Human Exome |
The interplay between autophagy and intracellular pathogens is intricate as autophagy is an essential cellular response to fight against infections , whereas numerous microbes have developed strategies to escape this process or even exploit it to their own benefit . The fine tuned timing and/or selective molecular pathways involved in the induction of autophagy upon infections could be the cornerstone allowing cells to either control intracellular pathogens , or be invaded by them . We report here that measles virus infection induces successive autophagy signallings in permissive cells , via distinct and uncoupled molecular pathways . Immediately upon infection , attenuated measles virus induces a first transient wave of autophagy , via a pathway involving its cellular receptor CD46 and the scaffold protein GOPC . Soon after infection , a new autophagy signalling is initiated which requires viral replication and the expression of the non-structural measles virus protein C . Strikingly , this second autophagy signalling can be sustained overtime within infected cells , independently of the expression of C , but via a third autophagy input resulting from cell-cell fusion and the formation of syncytia . Whereas this sustained autophagy signalling leads to the autophagy degradation of cellular contents , viral proteins escape from degradation . Furthermore , this autophagy flux is ultimately exploited by measles virus to limit the death of infected cells and to improve viral particle formation . Whereas CD150 dependent virulent strains of measles virus are unable to induce the early CD46/GOPC dependent autophagy wave , they induce and exploit the late and sustained autophagy . Overall , our work describes distinct molecular pathways for an induction of self-beneficial sustained autophagy by measles virus .
Measles is a highly infectious human disease caused by infection with measles virus ( MeV ) , one of the most contagious human pathogens [1] . Measles infection takes place by the respiratory route and clinical symptoms include respiratory infection , fever , cough , coryza , conjunctivitis and the appearance of a generalized maculopapular rash , the hallmark of measles . Although MeV infection results in lifelong immunity , a transient but profound immunosuppression occurs by one to two weeks of infection and persists for several weeks [2] . MeV infection-induced complications essentially include secondary infections but also post-infectious encephalitis and subacute sclerosing panencephalitis ( SSPE ) [3] , [4] . Despite the existence of an efficient vaccine based on attenuated MeV strains , recent measles outbreaks highlighted that this disease is still an important cause of mortality , especially among children in developing countries [1] , [5] . The understanding of the biological interplay between attenuated versus virulent MeV strains and cellular components remains a challenge if we aim at developing tools to counter infection and use MeV as a vector for therapies . MeV is a negative-stranded non-segmented RNA enveloped virus that belongs to the genus Morbillivirus of the Paramyxoviridae family [6] . Its genome encodes six structural proteins , the nucleoprotein MeV-N , a phosphoprotein MeV-P , an RNA polymerase MeV-L , the haemagglutinin MeV-H , the fusion protein MeV-F and the matrix protein MeV-M , and for two non-structural proteins not present within the virion , MeV-V and MeV-C . To date three different host-cell receptors for MeV-H attachment have been identified: CD46 which is expressed on all human nucleated somatic cells [7]; CD150 ( also known as SLAM ) , expressed on immature thymocytes , activated B and T lymphocytes , macrophages and mature dendritic cells [8] , and NECTIN-4 expressed on epithelial cells [9] , [10] . Whereas MeV-H of clinical/virulent strains of MeV attach to CD150 or NECTIN-4 , MeV-H of vaccine/attenuated-laboratory strains bind to either CD46 , CD150 or NECTIN-4 [11] . Following attachment , conformational modification of the MeV-F protein leads to membrane fusion and virus entry at the host-cell plasma membrane , but also to syncytia formation , resulting from fusion of infected cells with uninfected cells [12] . Subsequently to fusion , the MeV ribonucleoprotein is delivered into the host-cell cytosol and viral mRNAs start to accumulate linearly as soon as 5 to 6 hours post infection before being exponentially increased until 24 hours [13] . MeV-V and MeV-C were shown to contribute to MeV replication by foiling innate antiviral immune responses [14] , and newly generated ribonucleoproteins are assembled to bud from cell surfaces . Among cell-host functions modulated upon MeV infection , we recently described macroautophagy , thereafter referred to as autophagy [15] , [16] . Autophagy is a catabolic process essential for the maintenance of cellular homeostasis , through the elimination of otherwise deleterious cytosolic components , and for the recycling of metabolites [17] , [18] . During autophagy , a cup-shaped isolated membrane , the phagophore , elongates within the cytosol up to generate an autophagosome vesicle that sequesters large portions of the cytoplasm . This content is ultimately degraded within autolysosomes , following maturation of autophagosome by fusion with a lysosome . At the molecular level , autophagy is a highly regulated process that involves numerous proteins including those encoded by autophagy-related ( ATG ) genes [19] . The autophagy machinery is used as a universal cell defence against intracellular microbes since it allows their delivery to degradative lysosomes [17] . Autophagy may also contribute to activate antiviral innate immunity [20] , [21] , as well as adaptive immune response by delivering virus-derived peptides for presentation by major histocompatibility complex ( MHC ) molecules to T lymphocytes [22] , [23] , [24] . Although upon a wide range of viral infections autophagosome formation is observed , few molecular details of viral interaction with the autophagy machinery are known [21] . Nevertheless , numerous viruses have evolved molecular strategies to counteract autophagy in order to escape this process , or even to exploit it to improve their own infectivity [25] . We have previously reported that the binding of an attenuated strain of MeV on CD46 induces the flux of autophagy . This induction relies on the molecular connection of one isoform of CD46 , CD46-Cyt-1 , to BECLIN-1 via the scaffold protein GOPC [16] , [26] . Furthermore , we have shown that attenuated MeV can exploit autophagy through the physical interaction of the MeV-C protein with the autophagy-associated protein IRGM ( immunity-related GTPase M ) [27] . We wondered here how MeV regulates autophagy in the course of infection and how the virus exploits autophagy to its own advantage . We report that MeV infection induces successive autophagic signalling leading to a sustained increase of the autophagy flux . Whereas only attenuated MeV strains induce an early wave of autophagy dependent on CD46-Cyt1/GOPC , both attenuated and virulent strains induce and exploit a late but sustained productive autophagy wave .
We have previously reported that HeLa cell infection with the attenuated Edmonston ( Ed ) -MeV strain induces autophagy as soon as 1 . 5 hours post infection via the engagement of the CD46-Cyt-1/GOPC pathway [16] . We have also shown that autophagy was still detected by 24 hours of Ed-MeV infection [27] . Therefore , we examined autophagy kinetics upon Ed-MeV infection . GFP-LC3-HeLa cells were infected with Ed-MeV and autophagy was assessed at different time points by numerating GFP-LC3-labeled structures representing LC3-II-containing autophagosomes [28] . Strikingly , we found that Ed-MeV infection induces two successive waves of autophagosome accumulation ( Fig . 1A ) . The first wave is induced as soon as 1 . 5 hours post infection , but this wave is transient and returned to basal level by 3 to 6 hours of infection . At these time points , autophagy was however still functional since the treatment of MeV-infected cells with chloroquine , which inhibits autophagosome recycling , still allowed the detection of an accumulation of autophagosomes ( Fig . 1B and S1 ) . This result suggests that following the transient early wave of autophagy , MeV does not actively inhibit autophagosome formation . Moreover , a second wave of autophagy is then induced , 9 hours after infection , and is sustained up to 48 hours post Ed-MeV infection ( Fig . 1A ) . Similar waves of autophagy induction were observed by tracking LC3-II by western blot ( Fig . S2 ) . Of note , in the late time points , autophagosome accumulation was measured essentially in multinucleated cells , as MeV infection induces the massive formation of syncytia . Thus , for each experiment , the number of autophagosomes was normalized to the number of nuclei within a cell . To investigate the molecular pathways underlining these two waves of autophagosome accumulation we used small interfering ( si ) RNAs to reduce the expression of putative key proteins . As expected , we found that the reduced expression of GOPC ( Fig . S3 ) disrupted the early autophagosome accumulation , showing that early autophagy requires the expression of this scaffold protein ( Fig . 1C , 1 . 5 h p . i . ) . In contrast , late autophagy was not affected by the reduced expression of GOPC , although as expected , the treatment of cells with siATG5 prevented autophagosome accumulation ( Fig . 1D , 24 h p . i . and Fig . S3 ) . Since the timing of MeV-induced late wave of autophagy correlated with the beginning of an efficient MeV replication ( 9 hours post infection ) [13] , we wondered whether virus replication was important for the induction of the late GOPC-independent autophagy wave . We found that infection of GFP-LC3-HeLa cells with a non-replicative ultraviolet-treated Ed-MeV ( UV-Ed-MeV ) still induced the early autophagosome accumulation ( Fig . 1E ) . However , UV-Ed-MeV infection did not lead to the induction of the late wave of autophagy ( Fig . 1F ) . Altogether these results indicated that infection with an attenuated strain of MeV induces two successive waves of autophagy through distinct and uncoupled molecular pathways . An early one involves the cellular receptor CD46 and the scaffold protein GOPC , and a late one requires viral replication . The non-structural protein MeV-C is synthesised during the course of MeV infection . As we have recently shown that the single overexpression of MeV-C was sufficient to induce autophagy via an IRGM-dependent pathway [15] , [27] , we asked whether measles viral proteins synthesised during viral replication were indeed involved in late autophagy induction . To this end , we used 0 . 5 µg/ml cycloheximide which strongly prevented protein translation in the course of infection ( Fig . S4 ) , and autophagosome accumulation was monitored in GFP-LC3-HeLa cells infected with infectious Ed-MeV . Interestingly , whereas cycloheximide treatment did not modulate the early autophagy wave ( Fig . 2A ) , it totally abolished the induction of the later one ( Fig . 2B ) . The absence of autophagosome accumulation in 24 hours Ed-MeV-infected cells was not due to a limited availability of autophagy proteins due to cycloheximide treatment . Indeed , cells treated with rapamycin , an autophagy inducer , still induced autophagy in the presence of cycloheximide for the same period of time ( Fig . 2A/B ) . In order to determine whether MeV-C contributes indeed to autophagy induction during the course of infection , we then used a recombinant virus of the attenuated strain of MeV Schwarz ( Sch-MeV ) deficient for MeV-C expression ( Sch-MeVΔC ) . First , we found that wild type Sch-MeV behaved as Ed-MeV since Sch-MeV infection induced two successive waves of autophagy , with similar kinetics than Ed-MeV infection ( Fig . 2C ) . However , we found that in contrast to the wild type strain , Sch-MeVΔC did not induce autophagy 24 hours post infection ( Fig . 2D ) . Thus , this result suggested that MeV-C expression is an essential prerequisite for an efficient induction of a second MeV-induced autophagy wave . MeV infection massively induces syncytia formation . However , very few syncytia were detected upon Sch-MeVΔC infection ( data not shown ) . We therefore wondered whether , beyond MeV-C expression , the formation of syncytia might contribute to the induction of autophagy upon MeV infection . We analysed autophagosome accumulation in either mononucleated or multinucleated cells 24 hours after infection with either Sch-MeV or Sch-MeVΔC . Autophagy was analysed in MeV-infected cells by numerating GFP-LC3+ vesicles exclusively in infected cells detected by MeV-N staining . Interestingly , whereas autophagy was detected in both mononucleated or multinucleated cells infected with Sch-MeV , autophagy was only detected in the rare multinucleated cells formed upon Sch-MeVΔC infection ( Fig . 3A ) . Altogether , these results indicate that MeV-C is required to induce the late autophagy wave in infected cells that have not yet fused with other cells , but is dispensable in syncytia . To further evaluate a role of syncytia formation in autophagy induction , wild type Ed-MeV-infected GFP-LC3-HeLa cells were cultured in the presence of a fusion inhibitory peptide ( FIP ) which inhibits syncytia formation without preventing individual infectious MeV particles entry within host cells [14] . Whereas the FIP treatment completely abolished MeV-induced syncytia formation ( not shown ) , autophagy was still detected in 24 hours-infected mononucleated cells , although with a 50% reduction when compared to untreated infected cells ( Fig . 3B ) . This reduction was neither due to a general effect of FIP on the autophagy process , since accumulation of autophagosomes by rapamycin was not affected by FIP treatment ( Fig . 3B ) , nor to a decrease of viral protein synthesis as monitored by detection of similar levels of expression of MeV-N and MeV-P in the presence or not of FIP ( Fig . S5 ) . To determine whether MeV replication within infected-syncytia was required to induce autophagy , we analysed autophagy in multinucleated cells in absence of MeV replication . First , MeV-H/F-co-transfected HeLa cells ( Fig . S6A [29] ) were co-cultured with GFP-LC3-HeLa cells leading to the cell-cell fusion via a viral H/F proteins dependent fusion process . In GFP+ multinucleated cells ( those resulting from H+/F+ GFP-LC3− cells which have fused with H−/F− GFP-LC3+ cells ) , we found an increased number of autophagosomes , when compared with GFP+ mononucleated cells ( Fig . 3C ) . The reduced expression of GOPC did not modulate autophagosome accumulation in these multinucleated cells ( Fig . 3D ) . Moreover , multinucleated cells resulting from cell-cell fusion forced by polyethylene glycol ( PEG ) displayed an increased number of GFP-LC3+ autophagosomes ( Fig . 3E ) . Thus , all together these results strongly suggested that in the course of MeV infection , the syncytia formation mediated by Ed-MeV-H/F viral proteins is sufficient to induce an autophagic signalling in multinucleated cells , independently of both viral replication and of the CD46/GOPC-dependent pathway . Autophagy is a dynamic , multi-step process that can be modulated at several levels . An accumulation of autophagosomes can either reflect an increase of the autophagy flux , through the formation of de novo autophagosomes , or a reduced turnover of autophagosomes recycling , due to an inhibition of their fusion with lysosomes . We previously reported that the early CD46-Cyt-1/GOPC dependent induction of autophagy upon MeV infection resulted from an increase of the autophagy flux [16] . We confirmed this observation here by showing that the level of expression of the long-lived protein p62 , a target of autophagy degradation , was reduced during the early wave of autophagy induced 1 . 5 hours post Ed-MeV or Sch-MeV infection ( Fig . 4A/B ) . However , p62 expression was equivalent in 3–6 hours infected cells and in uninfected cells ( Fig . 4A/B ) , the time points during which autophagy was not detected in MeV-infected cells ( Fig . 1A ) . We then asked whether the late and sustained accumulation of autophagosomes also resulted from an increased autophagy flux . We found a strong reduction of p62 expression in 24 hours Ed-MeV-infected cells suggesting an increase of the autophagy flux at this late time point ( Fig . 4C ) . To further determine the reason of the late autophagosome accumulation upon infection , we used stably expressing mRFP-GFP-LC3-HeLa cells that allow for the distinction between autophagosomes ( GFP+RFP+ puncta ) and autolysosomes ( GFP−RFP+ puncta ) due to the quenched signal of the GFP in acidic compartments [30] , [31] . Although we found an increased number of total autophagy vesicles in Ed-MeV-infected cells compared to control cells , an equivalent ratio of number of autophagosomes and autolysosomes was measured in infected cells and in control or rapamycin-treated cells ( Fig . 4D ) . As expected , much more autophagosomes than autolysosomes were numerated in chloroquine-treated control cells , an inhibitor of autolysosome acidification ( Fig . 4D ) . Thus , both the early and the late induction of autophagy following attenuated MeV infection induce de novo autophagosome formation and leads to the achievement of the autophagy process , without inhibition of autophagosome maturation . We therefore wondered whether this sustained autophagic flux degrades MeV proteins during the course of infection . We determined the level of expression of two viral proteins , the nucleoprotein MeV-N and the phosphoprotein MeV-P , in control cells and in cells treated with siATG5 in order to prevent autophagy . We found that autophagy competent or deficient cells expressed similar levels of MeV protein after 24 hours or 48 hours of infection ( Fig . 4E/F ) . This result suggested that MeV proteins escape from autophagic degradation induced by MeV replication . Whereas several viruses take advantage of autophagy to replicate while inhibiting autophagosome maturation , MeV infection induces an increase of the autophagy flux . We wondered whether the autophagy flux was indeed required to improve viral particle production . First , we performed kinetics studies and , in accordance with our previous report [15] , we found a strong impairment in Ed-MeV particle formation 24 hours to 72 hours post-infection in cells with a reduced expression of ATG5 , as compared to control cells ( Fig . 5A ) . Autophagy involvement in MeV particle formation was confirmed with the reduced expression of another autophagy essential gene ATG7 ( Fig . S7A ) , which also compromised Ed-MeV production ( Fig . S7B ) . We then looked at the incidence of experimental autophagy induction on MeV replication . We found that HeLa cells treated with the autophagy inducer rapamycin produced twice more infectious Ed-MeV particles than untreated cells ( Fig . 5B ) . This result was corroborated when we analysed the expression of MeV proteins . We found that rapamycin treatment led to an increased level of expression of the two MeV proteins MeV-N and MeV-P ( Fig . 5C/D ) . Importantly , the inhibition of autophagosome maturation with chloroquine decreased the viral particle production , when compared to control cells , highlighting the importance of the autophagic flux for an efficient MeV replication in HeLa cells ( Fig . 5B ) . However , chloroquine treatment did not have impact on the level of expression of MeV protein ( Fig . 5C/D ) , confirming that these proteins are not targeted towards autophagic degradation . Altogether , these results indicated that MeV infection leads to a productive autophagy which is required for an efficient production of MeV infectious particles . As autophagy has no direct impact on viral protein synthesis , we wondered whether MeV could take benefit from autophagy by extending infected cell survival . We found that autophagy inhibition using siATG5 increased death of cells infected by Sch-MeV as measured by cell numeration ( Fig . 6A ) . We confirmed this result by looking at the percentage of late apoptotic cells/dead cells determined by the double labeling for Annexin V and 7-Amino-Actinomycin ( 7AAD ) and found twice more Annexin V+/7AAD+ in siATG5-treated cells than in control cells upon Sch-MeV infection ( Fig . 6B ) . In contrast , the experimental induction of autophagy with rapamycin prior infection was found to protect Sch-MeV-infected cells from cell death ( Fig . 6C/D ) . Moreover , the reduced expression of ATG5 increased Sch-MeVΔC-infected cell death ( Fig . S8 ) , but more slightly than wild type Sch-MeV ( Fig . 6A ) . However , autophagy promotion with rapamycin protected efficiently Sch-MeVΔC-infected cells from death ( Fig . S8 ) . Taken together , these results strongly suggested that autophagy induced by MeV could contribute to protect infected cells from death what contributes to improve infectivity . Corroborating this hypothesis , we found that the inhibition of apoptosis , using the pan-caspase inhibitor Z-VAD , improved MeV particle production by 48 hours-infected cells ( Fig . 6E ) . Furthermore , the prevention of autophagy in Z-VAD-treated cells did not modulate MeV production ( Fig . 6E ) , suggesting that the main function of autophagy in the course of MeV infection is to delay MeV-induced apoptosis . We then investigated whether a virulent strain of MeV , which does not bind CD46 but CD150 to infect cells [8] , might also induce the two waves of autophagy during infection , and exploit autophagy in HeLa cells . As expected , we did not detect autophagy in HeLa cells incubated with the virulent G954-MeV strain , since these cells do not express CD150 ( Fig . S9 ) [8] . However , interestingly we found only a marginal accumulation of GFP+ dots in CD150-GFP-LC3-HeLa cells , 1 , 5 hours after infection ( Fig . 7A and Fig . S10 ) . By contrast , a strong increase of the number of autophagosomes was observed 24 hours after G954-MeV infection ( Fig . 7B ) . These dots corresponded to autophagosomes since their accumulation was abolished in cells with a reduced expression of the essential gene for autophagy ATG5 ( Fig . S11 ) . Additionally , we did not observe a decrease of p62 expression 1 , 5 hours post G954-MeV infection , whereas its expression drops off by 24 hours of infection ( Fig . 7C/D ) . Moreover , tracking RFP-GFP-LC3 in infected CD150-HeLa cells confirmed that the autophagy flux is increased in 24 hours G954-MeV-infected cells ( Fig . 7E ) . Thus , these results suggested that virulent CD150-dependent CD46-independent MeV strain infection does mostly induce the productive late wave of autophagy . Additionally , we found that UV-G954-MeV and cycloheximide treatment prevented autophagosome accumulation in 24 hours G954-MeV-infected CD150-HeLa cells , indicating that , as for attenuated strains , neo-synthesized viral proteins are required for autophagy infection by a virulent MeV strain ( Fig . 7F and S12 ) . Moreover , virulent H/F co-expression on HeLa cells ( Fig . S6B ) co-cultured with CD150-GFP-LC3 HeLa cells was sufficient to induce an autophagic signalling in multinucleated cells , independently of viral replication ( Fig . 7G ) . Finally , we found that virulent MeV replicated less efficiently in siATG5-treated cells than in control cells ( Fig . 7H ) . Altogether , these results indicated that a virulent MeV strain unable to induce the first CD46-dependent autophagic wave , induces and exploits the late autophagic wave to replicate .
We found that infection with attenuated MeV induces two successive waves of autophagy via distinct molecular pathways . We have previously described the attenuated MeV receptor CD46-Cyt-1 as a pathogen receptor able to induce an early autophagy flux , subsequently to pathogen detection , via its association with the scaffold protein GOPC , which relies to the autophagosome formation complex BECLIN 1/VPS34 [16] . We show here that this autophagy induction is very transient as it probably stops after CD46/GOPC-mediated autophagic signalling following virus entry . Interestingly , virulent strains do not bind to CD46 and therefore are unable to induce the CD46-dependent early autophagy wave . Instead , virulent strains bind to CD150 . CD150 was reported to recruit the autophagy-associated molecules BECLIN 1 and VPS34 to the phagosome of macrophages which uptake gram-negative bacteria [32] . However , whether CD150 engagement regulates the autophagy process has not been observed . Our results indicate that CD150 does not induce autophagy ( or very marginally ) upon virulent MeV entry . Thus , an attractive hypothesis would be that the first autophagy wave induced by attenuated/vaccinal MeV strains , but not triggered by virulent/clinical ones , could contribute to the attenuation of MeV infectivity , an issue that remains to be investigated . An immune function resulting from CD46-induced autophagy could be to facilitate intravesicular TLR engagement . Indeed in plasmacytoid dendritic cells , TLR7 gains access to viral-replication intermediates through autophagy to induce antiviral type I interferon ( IFN-I ) production [20] . Interestingly , in B cells , TLR9 is recruited to autophagosomes upon B cell receptor ( BCR ) internalization to enhance B cell activation [33] . Furthermore , it was shown that CD46-binding adenoviruses are routed in a pathway that allows TLR9-dependent IFN-I induction , whereas adenoviruses using CAR as cellular receptor does not [34] . Thus , the immediate autophagy induction mediated by CD46 might confer an early response towards CD46-binding MeV strains , protecting cells from massive replication by producing IFN-I prior to extended MeV replication , and contributing to their attenuation . HeLa cells being poor producers of IFN-I , in our experimental setup attenuated MeV can replicate efficiently . It will be of interest to determine whether autophagy induced by CD46-binding viruses is involved in early IFN-I induction . After few hours of infection , a new autophagy flux is induced , which is sustained over time . This process is independent of a CD46-Cyt-1/GOPC signalling , but requires viral replication and relies on the expression of the C protein of MeV . Importantly , in HeLa cells , for attenuated and virulent strains of MeV , we demonstrated that this second wave of autophagy is manipulated by MeV to increase infectious viral particle formation . We described recently the autophagy-associated protein IRGM as being a cellular target of MeV for both the induction of autophagy 24 hours post-infection ( what correspond to the second autophagy wave ) , and to improve MeV infectivity [15] . Moreover , we have reported that the single expression of MeV-C , which can interact with IRGM , was sufficient to induce autophagy , through an IRGM dependent pathway [15] . Here , we extended these observations by showing that MeV deficient for the expression of the C protein ( MeVΔC ) does not induce autophagy in infected mononucleated cells . These results confirmed a prominent role of the MeV-C protein in the induction of the second wave of autophagy , subsequently to MeV replication . The exact role of MeV-C in MeV-induced autophagy requires further investigations . We have previously shown that the MeV-C partner IRGM can interact with several other human autophagy-associated proteins , ATG5 , ATG10 , LC3C and BIF-1 , which could be involved in MeV-induced autophagy , consequently to C expression [15] , [35] . Interestingly , MeV-C was also shown to interfere with host defence mechanisms by dampening antiviral IFN-I activation through the downregulation of viral RNA synthesis in order to limit their detection by IFN-I-inducing cytosolic receptor such as RIG-I or MDA5 [36] , [37] . We highlight here an unrevealed role of the C protein in autophagy induction upon MeV infection , which is ultimately used by the virus to improve its infectivity . Strikingly , whereas the expression of MeV-C is required for the induction of the late autophagy wave in infected mononucleated cells , its expression is not critical for the induction of autophagy in syncytia . Indeed , we found that whereas Sch-MeVΔC does not induce autophagy in mononucleated cells , autophagy was still observed in syncytia . Interestingly , it has recently been suggested that viruses of the Morbillivirus genus including MeV , could induce autophagy through a fusogenic dependent mechanism which requires the coexpression of MeV-F and MeV-H proteins [38] . Using an alternative approach to analyse autophagy , we found an increase of autophagosomes in multinucleated cells formed through a MeV-H/F proteins-mediated process . Furthermore , we found that the forced fusion between cells promoted by PEG-treatment induces autophagy . Thus , the plasma membrane perturbations/damages resulting from cell-cell fusion could be sufficient to trigger an autophagy signalling , with functions to eliminate/recycle excessive/redundant/damaged cytosolic materials within newly formed multinucleated cells . The upstream signals inducing autophagy subsequently to the fusion between cells remain to be investigated . Interestingly , it was proposed that autophagy could play a role in syncytia formation [38] . Thus , an attractive hypothesis would be that during MeV infection , the MeV-C protein induces autophagy in infected cells through an IRGM-dependent pathway , which could contribute to the facilitation of the syncytia formation . In turn , through a fusogenic process , syncytia formation could further increase autophagy in multinucleated cells leading to a positive feedback loop of autophagy induction , maintained by MeV-C expression . In accordance with this hypothesis , the number of syncytia is strongly reduced upon infection with Sch-MeVΔC compared with wild type Sch-MeV ( data not shown ) . Thus , whereas the first wave of autophagy triggered by the CD46-Cyt-1/GOPC pathway would concern exclusively the primary infected cells by infectious attenuated viral particles , the second wave is induced following MeV-C expression and would be maintained over time by syncytia formation . MeV-induced syncytia were reported to be dynamic entities with an unusual extended life span [39] . Interestingly , we found that autophagy contributes to protect MeV-infected cells from apoptotic cell death . Thus autophagy induction in syncytia could delay MeV-induced cell death , and contribute to an efficient spreading of the virus . Indeed , we report that inhibition of apoptosis strongly facilitates MeV replication , what is not anymore modulated by additional autophagy inhibition . Furthermore , in support of this hypothesis MeVΔC , which does not induce autophagy over basal level in mononucleated cells , presents a growth defect and induces more apoptosis in infected cells than wild type virus [40] . Thus , one way by which MeV exploits autophagy is by protecting infected cells from cell death which otherwise could limit viral replication and propagation of numerous infectious viral particles . Other viruses were recently described to manipulate autophagy in order to prevent cell death , as the human flavivirus dengue virus type 2 and the Chikungunya virus [41] , [42] . MeV infection induces an increase of the autophagy flux . However , MeV proteins are not degraded by autophagy what suggests that MeV proteins would escape from targeting towards autophagosomes . Indeed , colocalization between GFP-LC3 and the MeV-N protein was not observed . Furthermore , it has been recently reported that the Morbillivirus replication complex and autophagosomes do not colocalize [38] . However , we have shown that maturation of autophagosomes is necessary to promote the formation of infectious viral particles . Beyond delaying death of infected cells , productive autophagy could contribute to the generation of an excess of metabolites used by MeV to optimize its replication . Autophagy induced by Dengue virus infection is used to regulate cellular lipid metabolism to generate ATP and to promote viral replication [43] . As we found that MeV proteins synthesis is not modulated by autophagy , nucleic acids and/or lipids generated from autophagy-mediated recycling could be primary metabolites exploited to improve MeV replication and/or assembly . Alternatively , the different autophagy signalling triggered in the course of MeV infection might be required to coordinate different steps of the virus cycle . Indeed , productive autophagy has been shown to facilitate poliovirus replication , which gradually gets benefit of the complete autophagic process : whereas autophagosome formation is involved in the viral RNA synthesis and early phases of the virus cycle , the acidification of the autolysosomes contributes to the final maturation of virus particles [44] . Similarly , autophagy impacts on the final maturation of infectious Dengue virus particles formation , as inhibition of autophagy leads to the production of noninfectious particles [45] . The contribution of the different autophagy signallings in MeV maturation remains to be investigated . Our work describes the induction of successive molecular pathways contributing to autophagy in response to an infection by attenuated MeV . Whereas an early wave of autophagy induction is triggered by the host recognition of the pathogen , via CD46-Cyt-1/GOPC , a later one seems to be the result of a direct interaction of viral proteins with the autophagy machinery , and is sustained within syncytia . The sustained wave is exploited by attenuated as well as virulent strains of MeV to promote the production of infectious viral particles . The complete understanding of the complex relationship between autophagy and MeV might allow a better understanding of the contribution of autophagy for the attenuated or virulent characters of a pathogen , and a better usage of MeV-derived vaccine for therapy .
The experiments in this article were performed at Biological Safety Level 2 in accordance with the regulations set forth by the national French committee of genetics ( commission de génie génétique ) . HeLa , GFP-LC3-HeLa and mRFP-GFP-LC3 HeLa cells were maintained in RPMI 1640 , Vero cells were maintained in DMEM . All the media were supplemented with 50 µg/mL gentamicin and 10% fetal bovine serum ( FBS ) . Ed-MeV was obtained from ATCC and G954-MeV was kindly provided by B . Horvat ( INSERM U1111 , France ) . HeLa cells were infected with MeV at the indicated MOI . After the indicated period of infection , cells were submitted to 5 cycles of freezing at −80°C and defrosting at 37°C and infectious viral particles were quantified by limiting dilution on confluent Vero cells . Smartpool siRNA targeting ATG5 , ATG7 and GOPC , as well as control siRNA ( AUACCUAACUGAUGAGACCUU ) were from Dharmacon ( Perbio , Brebières , France ) . 0 . 1×106 HeLa , GFP-LC3 HeLa cells were plated in 6-well plates 24 hours prior to transfection with 100 pmol si-RNA using Lipofectamine RNAiMAX ( Invitrogen ) according to manufacturer's instructions . Protein expression level was assessed by western-blot three days post transfection . HeLa , GFP-LC3-HeLa and mRFP-GFP-LC3 HeLa cells were transfected with the plasmid pCAGGS-CD150 , kindly provided by Y . Yanagi ( Japan ) [8] , using lipofectamine 2000 ( Invitrogen ) according to manufacturer's instructions . 24 h post transfection , CD150 expression was checked by FACS analysis after a staining using an anti-CD150-PE antibody and as a negative control an isotype antibody . The lentiviral peGAET-ires-puro expression plasmids was a gift of P . Mangeot ( ENS-Lyon , INSERM U1111 , Lyon , France ) . To generate peGAET-cd150-ires-puro plasmid , cd150 sequence was amplified by PCR and replaced the tTA sequence in peGAET-ires-puro into the EcoR1 and XhoI restriction sites . Viral particles were commonly produced by cotransfection of 293T cells with lentiviral peGAET-cd150-ires-puro plasmid and the helper plasmids encoding the proteins required for vector packaging ( Plateau AniRA Vectorologie UMS3444 , US8 ) . Supernatant was collected at days 2 post-transfection , filtered and concentrated by ultracentrifugation . GFP-LC3-HeLa and HeLa were transduced with concentrated viral particles in the presence of 8 µg/ml polybrene ( Sigma ) . 48 h post transduction , cells were treated with 1 µg/ml puromycin for 10 days . Surviving clones were expanded in 1 µg/ml puromycin and analysed for stable integration of the transgene and expression of CD150 protein by flow cytometry . HeLa cells were co-transfected with the plasmid pCXN2-F ( Edmonston strain ) , and either with the plasmid pCXN2-H ( Edmonston strain ) or pCXN2-H ( KA strain ) [37] using lipofectamine 2000 ( Invitrogen ) according to manufacturer's instructions . 24 h post transfection , MeV-H and MeV-F expression were checked by FACS analysis using specific primary antibodies . GFP-LC3-HeLa cells were seeded in 25 cm2 flask to be confluent the next day , in RPMI 10% FBS in absence of antibiotics . A PEG solution was prepared by mixing 10 g of autoclaved PEG 6000 ( Sigma ) with 10 ml of RPMI without FBS . The cells were washed twice with warmed PBS and then incubated 10 min at 37°C with PEG . The PEG was then removed and RPMI free of serum was added progressively onto the cells . This RPMI was then removed and replaced by 5 ml of RPMI 10% FBS for 4 hours . Finally , the cells were detached with Versene ( Invitrogen ) and plated in 24 well-plates on cover slips coated with Poly-L-Lysine and treated with rapamycin for 2 hours . Cells were lysed with lysis buffer ( PBS-1× , NP40 1% or 0 . 5% and protease inhibitor ( Roche ) ) . Soluble proteins were separated by SDS-PAGE ( NuPAGE Novex Tris-Acetate Mini Gels , Invitrogen ) and transferred to nitrocellulose membranes with iblot Gel Transfer System ( Invitrogen ) . Primary antibodies , anti-Actin ( mouse monoclonal ) , anti-MeV-N ( mouse monoclonal ) , anti-MeV-P ( mouse monoclonal or rabbit polyclonal ) , anti-p62 ( mouse monoclonal ) , secondary HRP linked anti-rabbit or HRP linked anti-mouse antibodies were used and antigen-antibody complexes were visualized by enhanced chemiluminescence . After the indicated treatments , GFP-LC3-HeLa or mRFP-GFP-LC3 HeLa cells were fixed with 4% paraformaldehyde . For Measles nucleoprotein staining , an anti-MeV-N antibody ( mouse monoclonal , clone 120 produced in the lab ) was used at 10 µg/mL , followed by secondary antibody conjugated to Alexa Fluor 568 . Cells were analysed using a Confocal Axioplan2 LSM510 microscope ( Zeiss , Göttingen , Germany ) equipped with the LSM 510 META system ( Zeiss ) and mounted with an Axioscope 63× oil immersion lens ( Zeiss ) . The number of GFP + was numerated from one single plan section per cell and normalized to the number of nuclei . In the legend , a cell profile means “per nucleus” because for syncytia , number of GFP+ vesicles was normalized to the number of nuclei . In each case , number of GFP+ vesicles was numerated from 100 to 200 cells for each experiment . Cells were mixed with 0 . 4% trypan blue ( Gibco ) and unstained ( viable ) and stained ( nonviable ) cells were numerated on a Bürker microscope slide ( Marienfeld ) . For each condition , at least 100 cells have been numerated . Cells were stained with the PE Annexin V Apoptosis Detection Kit I ( BD Pharmingen ) according to manufacturer's instructions and cells were analysed by using an Accuri C6 flow cytometer and the C Flow software . Anti-MeV-N ( mouse monoclonal , clone 120 ) , MeV-H ( mouse monoclonal , clone 55 ) , anti-MeV-F ( mouse monoclonal , clone Y503 ) , and anti-GOPC ( GOPC-GST rabbit anti-serum ) were produced in the lab , anti-MeV-P ( mouse monoclonal 49 . 21 or rabbit polyclonal J37171 ) were kindly provided by D . Gerlier ( INSERM U1111 , France ) . Anti-Actin ( A2066 ) , anti-ATG5 ( A0856 ) , anti-MAP1LC3B ( L7543 ) and anti-ATG7 ( A2856 ) were from Sigma ( St Louis , Mo , USA ) . Anti-rabbit HRP ( NA 934 ) or anti-mouse HRP ( NA 931 ) were from Amersham Biosciences ( Uppsala , Sweden ) . Anti-p62 ( SQSTM1 ( D-3 ) : sc-28359 ) was from Santa Cruz Biotechnology ( California , USA ) . Anti-mouse Alexa Fluor 568 was purchased from Invitrogen ( Molecular Probes ) . Anti-CD150-PE ( 559592 ) was from BD Pharmigen and isotype ( IOTest IgG2a ( Mouse ) -FITC/IgG1 ( Mouse ) -PE ) was from Beckman Coulter ( Immunotech SAS , Marseille , France ) . Pharmacological reagents used were cycloheximide ( C4859 , Sigma ) , rapamycin ( Calbiochem ) , chloroquine ( C6628 , Sigma ) and Z-Val-Ala-DL-Asp ( OMe ) -fluoromethylketone ( Z-VAD ) ( N-1560 , Bachem ) . | Autophagy is an evolutionarily conserved lysosomal dependent degradative pathway for recycling of long-lived proteins and damaged organelles . Autophagy is also an essential cellular response to fight infection by destroying infectious pathogens trapped within autophagosomes and plays a key role in the induction of both innate and adaptive immune responses . Numerous viruses have evolved strategies to counteract autophagy in order to escape from degradation or/and to inhibit immune signals . The kinetic and molecular pathways involved in the induction of autophagy upon infections might determine if cells would be able to control pathogens or would be invaded by them . We showed that measles virus ( MeV ) infection induces successive autophagy signallings in cells via distinct molecular pathways . A first autophagy wave is induced by the engagement of the MeV cellular receptor CD46 and the scaffold protein GOPC . A second wave is initiated after viral replication by the expression of the non-structural MeV protein C and is sustained overtime within infected cells thanks to the formation of syncytia . This sustained autophagy is exploited by MeV to limit the death of infected cells and to improve viral particle formation . We describe new molecular pathways by which MeV hijacks autophagy to promote its infectivity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] | [] | 2013 | Sustained Autophagy Contributes to Measles Virus Infectivity |
The Scc2-Scc4 complex is essential for loading the cohesin complex onto DNA . Cohesin has important roles in chromosome segregation , DSB repair , and chromosome condensation . Here we report that Scc2 is important for gene expression in budding yeast . Scc2 and the transcriptional regulator Paf1 collaborate to promote the production of Box H/ACA snoRNAs which guide pseudouridylation of RNAs including ribosomal RNA . Mutation of SCC2 was associated with defects in the production of ribosomal RNA , ribosome assembly , and splicing . While the scc2 mutant does not have a general defect in protein synthesis , it shows increased frameshifting and reduced cap-independent translation . These findings suggest Scc2 normally promotes a gene expression program that supports translational fidelity . We hypothesize that translational dysfunction may contribute to the human disorder Cornelia de Lange syndrome , which is caused by mutations in NIPBL , the human ortholog of SCC2 .
Cohesion between sister chromatids generates the force that holds sister chromatids together until the onset of anaphase [1] . Cohesion is generated by cohesin , an evolutionarily conserved multi-subunit protein complex consisting of four core subunits: Smc1 , Smc3 , the α-kleisin subunit Mcd1/Scc1/Rad21 , and the HEAT repeat-containing protein Scc3/SA1 or SA2 . The complex forms a ring-like structure that entraps DNA [2 , 3] . Smc1 and Smc3 belong to the structural maintenance of chromosome ( SMC ) ATPase superfamily [1 , 4] . This family also includes subunits of the condensin and Smc5/6 complexes . Cohesin and condensin loading are facilitated by Scc2-Scc4 [5–7] . In addition to its role in chromosome segregation , cohesin promotes DNA damage repair [8] and regulates gene expression [9 , 10] . It has been proposed that the human ortholog of SCC2 , NIPBL , could regulate gene expression [11] . In budding yeast , Scc2 binds to Pol II transcribed genes encoding ribosomal proteins and small nuclear and nucleolar RNAs ( snRNA and snoRNAs ) , Pol III transcribed genes encoding tRNAs and other noncoding RNAs , as well as pericentric domains [5] . These same regions are bound by condensin [5] . Defects in the association of Scc2 with these highly transcribed regions could potentially affect the expression of these genes . For instance , a recent report suggested Scc2 may help maintain a nucleosome-free region [12] , which could potentially promote both SMC complex loading and gene expression . Since Scc2 target genes contribute to translation , a decrease in their expression may negatively affect translation . Mutations in SCC2/NIPBL and genes encoding the cohesin subunits result in a developmental syndrome known as Cornelia de Lange syndrome ( CdLS ) . The causative mutations are spread throughout the NIPBL gene , most often resulting in a partial loss of function [13] . How mutations in NIPBL cause developmental defects remains largely unknown . Examination of cells from affected CdLS probands suggests differential gene expression might cause the developmental defects observed in CdLS patients , rather than chromosome missegregation [14 , 15] . Recent work using zebrafish to model CdLS is consistent with the idea that partial loss of function of nipbl reduces translation [16] , but the molecular mechanisms are unclear . In this study , we examined how the cohesin loader Scc2 regulates gene expression in budding yeast . We utilized a temperature-sensitive partial loss of function mutation in SCC2 ( scc2-4 ) that has previously been shown to ( 1 ) reduce cohesin and condensin association with chromosomes at 37°C [5 , 7] , ( 2 ) delay DSB repair [8] , and ( 3 ) disrupt nucleolar morphology at 30°C [5] . We found that the mutant protein is expressed at normal levels and displays a normal binding profile at centromeres , but has reduced association with genic regions including ribosomal DNA ( rDNA ) , tDNAs , snoDNAs , and ribosomal protein genes . To examine the biological processes and RNAs regulated by Scc2 , we performed RNA sequencing at permissive temperature . Consistent with a previous report [17] , we observed differential expression of hundreds of genes . Gene expression signatures suggested both ribosome biogenesis and mitochondrial function would be impacted in the mutant and functional analysis confirms they are both negatively affected . One group of down-regulated genes were the H/ACA snoRNAs , which guide the site-specific pseudouridylaton of rRNAs , tRNAs , mRNAs , and noncoding RNAs [18–20] . The production of H/ACA snoRNAs appears to be facilitated by Scc2-dependent recruitment of the RNA Pol II-associated factor ( Paf1 ) complex . The scc2-4 mutant showed defective rRNA production and modification with a mild reduction in global protein synthesis . More strikingly , translational fidelity was reduced as shown by decreased internal ribosomal entry site ( IRES ) usage , increased frameshifting , and decreased resistance to translational inhibitors . Our results in budding yeast strongly suggest that the Scc2 regulated gene expression program promotes translational fidelity .
SCC2 is a large gene encoding a protein with several domains ( Fig 1A ) . Human and budding yeast NIPBL/SCC2 do not align very well . However , the function of Scc2 in loading SMC complexes is evolutionarily conserved from yeast to humans . Evolutionarily conserved HEAT repeats are located at the C-terminus . By chemical mutagenesis , a temperature sensitive mutant was isolated ( E534K ) , named scc2-4 [1] . The mutated amino acid is evolutionarily conserved from yeast to human ( Fig 1B ) . The E534K mutation falls in a central region of the protein with unknown function . However , the surrounding amino acids are highly conserved . At 30°C , the scc2-4 mutant grows more slowly than WT ( Fig 1C ) . Western blot analysis shows that the level of the mutant protein is similar to WT ( Fig 1D ) . In order to understand how the mutant Scc2 protein associates with the genome , we performed ChIP-seq . Strains were cultured in YPD+CSM ( complete supplement mixture ) at 30°C until mid-log phase , fixed in formaldehyde for 2 hrs , and chromatin was harvested for ChIP-seq . The basic pattern of binding for the WT Scc2 protein was similar to previous reports [5 , 21] . Enrichment for different regions of the genome was further characterized using metagene plots ( Fig 2 ) . On the y-axis is the mean count per million ( cpm ) and on the x-axis is the position in base pairs . No apparent difference in centromere enrichment was observed for Scc2E534K when compared to the WT ( Fig 2A ) . In contrast , we observed reduced enrichment for rDNA , snoDNAs ( Box H/ACA and Box C/D ) , tDNAs , and the small and large ribosomal protein genes in the Scc2E534K ChIP ( Fig 2 ) . Thus , the mutation appears to compromise genic association without compromising centromere association . To investigate the biological processes regulated by Scc2 , RNA sequencing was performed for WT and the scc2-4 mutant . Each strain was grown in triplicate , at 30°C in YPD+CSM to mid- log phase . The distribution of gene expression can be viewed in the MA plot , defined as a plot of log-intensity ratios ( M-values ) versus log-intensity averages ( A-values ) ( Fig 3A ) . The log2 ratio of scc2-4/WT is shown on the y-axis and the geometric mean of the reads per kilobase of transcript per million reads mapped ( RPKM ) values on the x-axis . Using an adjusted p-value of 0 . 05 as criteria , there are 2644 differentially expressed genes , with 1285 up-regulated and 1359 down-regulated in scc2-4 mutant . Further applying a fold change threshold of 1 . 5 ( corresponding to an absolute log2 value of 0 . 6 ) returns 823 up-regulated and 760 down-regulated genes for GO analyses and general comparisons . This is many more genes than are bound by Scc2 , suggesting that many of the changes in gene expression are due to indirect effects . GO term analyses for up-regulated genes shows enrichment for genes involved in ribosome biogenesis and rRNA processing ( Fig 3B and S1A and S1B Fig ) . The upregulation of ribosomal protein genes ( RPs ) and the processome in general suggests these mRNAs are not rate limiting in the scc2-4 mutant for ribosome biogenesis , although ribosome assembly does appear to be affected ( see below ) . This upregulation is in direct contrast to our previous analysis of gene expression in the cohesin acetyltransferase mutant ( eco1-W216G ) , which shows downregulation of these gene groups [22] . GO term analysis for the down-regulated genes shows enrichment for genes required for oxidative phosphorylation ( Fig 3C and S1C and S1D Fig ) . We further explored specific aspects of the gene expression profile in the scc2-4 mutant . The downregulation of genes involved in oxidative phosphorylation suggested the mutant might be hypersensitive to a drug that inhibits mitochondrial function , such as chloramphenicol which blocks protein synthesis by the mitochondrial ribosome . Consistent with their respective gene expression profiles , the scc2-4 mutant , but not the eco1-W216G mutant , showed very poor growth on plates with a sublethal dose of chloramphenicol ( S2 Fig ) . Another group recently published the gene expression profile of the scc2-4 mutant [17] . However , these experiments were conducted at non-permissive temperature whereas our experiments were all conducted under permissive conditions . The two data sets have some differences , as expected , but the gene expression profile from Lindgren et al suggests mitochondrial function and translation would be negatively impacted in the scc2-4 mutant , although this was not tested . However , they did show that cohesin binding locations were not significantly affected in the mutant background , helping to rule out this effect as an explanation for the differential gene expression . Scc2 binds to genes that encode RNA components of ribonucleoprotein complexes such as NME1 , a component of RNAse MRP that modifies ribosomal RNA , SCR1 , part of the signal recognition complex , and SNR6 , the U6 component of the spliceosome . The mutant protein binds less well to these regions , correlating with significantly lower expression in the mutant ( p<0 . 05 ) . Given the reduction in U6 , we decided to analyze splicing in the mutant compared to WT . By measuring splicing efficiency ( computing the mean nucleotide coverage over every spliced junction divided by the mean nucleotide coverage for reads falling within the exonic splicing unit ) , we were able to detect a modest but reproducible defect in splicing ( S3 Fig ) . As a control we used the eco1-W216G mutant , where we did not find a splicing defect ( S3 Fig ) . Analysis of the correlation between genome-wide enrichment of Scc2-bound sequences and gene expression revealed contrasting findings . While loss of binding of Scc2 correlated with lower gene expression at certain regions of the genome ( e . g . Box H/ACA , some Box C/D snoDNAs ) , there was increased expression at RP genes , and no correlation at other regions . Overall the gene expression pattern is likely the product of a combination of direct effects from changes in Scc2 binding and indirect effects . It is also possible that Scc2 may play different roles in gene expression at different groups of genes . The reduced enrichment for the ribosomal DNA repeats in the scc2-4 mutant ( Fig 2C ) , combined with our previous work demonstrating that cohesion is required to form a nucleolus [23] and produce normal levels of ribosomal RNA [22] , drove us to examine rRNA production and nucleolar morphology in the scc2-4 mutant . rRNA synthesis was measured by pulse-labeling cells for 5 min with 3H-uridine . This approach is a reliable method for estimating RNA Pol I activity since most RNA synthesis during exponential growth is Pol I derived . Incorporation of uridine into total RNA was reduced by approximately 3-fold in the scc2-4 mutant ( Fig 4A ) , consistent with the observed decrease in growth rate in the scc2-4 mutant . rRNA production was further investigated by pulse-labeling cells with 3H-methylmethionine for 5 min , chased with excess cold methionine for 5 min and extracting RNA . Incorporated 3H was quantified in 25S and 18 rRNAs ( Fig 4B ) . rRNA is methylated cotranscriptionally in yeast , making this approach a reliable method for quantifying the production of the methylated forms of the 25S and 18S rRNAs [24] . By this method , we also observed a 3 . 6-fold reduction in the production of 25S rRNA ( Fig 4B , right ) and a 4-fold reduction in 18S rRNA production ( Fig 4B , bottom right ) . We further examined the processing of the initial rRNA transcript into the 25S and 18S forms over time . We found no defect in the processing rate for 25S and 18S rRNAs in scc2-4 mutant ( S4 Fig ) . In summary , the scc2-4 mutant has reduced rRNA production . Nucleolar morphology has been reported to be aberrant in the scc2-4 mutant based on visualization of the nucleolar protein Net1 fused to GFP [5] . We extended this observation using electron microscopy ( S5 Fig ) . The dense staining nucleolar material did not adopt the normal crescent shaped structure in the mutant , consistent with the idea that reduced binding of Scc2 and reduced loading cohesin and/or condensin at the rDNA could result in a failure to gather the rDNA repeats into a normal nucleolar structure . Ribosomes are partially assembled in the nucleolus and transported into the nucleus and then cytoplasm to be fully assembled and engaged in translation . Since mutation in SCC2 results in differential expression of messages involved in ribosome biogenesis , including ribosomal protein genes and genes involved in processing and assembling ribosomes , we further investigated ribosome production in the scc2-4 mutant . GFP tagged protein components of the large ( Rpl25 ) and small ( Rps2 ) ribosomal subunits were expressed in the WT and scc2-4 mutant strains and examined by microscopy . Rather than the normal even distribution in the cytoplasm , we observed accumulation of ribosomal proteins in the scc2-4 mutant in the nucleus/nucleolus that suggested that ribosome assembly/export was affected ( Fig 5A and 5D ) . Using flow cytometry , we found higher mean fluorescence for both 40S and 60S reporters in the scc2-4 mutant ( Fig 5B and 5E ) . We further analyzed the data by generating a cumulative distribution function for each sample , calculating the distance between samples , and testing for statistical significance using the Kolmogorov—Smirnov ( KS ) test , see Materials and Method and [22] ( Fig 5D and 5F ) . The KS test was statistically significant for both the 40S and 60S subunit reporters in the scc2-4 mutant compared to WT . This analysis suggests ribosome assembly may be deficient in the mutant . The defects in nucleolar morphology , rRNA production , and ribosome distribution in the scc2-4 mutant are quite similar to those reported for the cohesin acetyltransferase mutant , suggesting cohesion is important for these aspects of nucleolar structure and function . One notable group of Scc2 bound genes that was down-regulated in the scc2-4 mutant was the box H/ACA snoRNA genes ( Fig 6A ) . Some box C/D RNAs were also downregulated , but not all . Box H/ACA snoRNA genes stand alone whereas box C/D snoRNA genes lie in the introns of ribosomal protein genes which may make them subject to the regulation of their host genes . The differential expression of snoRNAs was not observed in the eco1-W216G mutant [22] . These small nucleolar rRNAs guide site-specific modification of rRNAs and other RNAs . Box C/D RNAs guide the methylation of RNAs in the context of a ribonucleoprotein particle . The box H/ACA snoRNAs are also part of ribonucleoprotein particles ( snoRNPs ) along with the essential proteins ( Nhp2 , Cbf5 , Nop10 and Gar1 ) that together catalyze site-specific pseudouridylation of RNAs . Pseudouridylation is important for RNA stability and interactions with other RNAs and proteins [25] . Mutation of the CBF5 component of this snoRNP leads to reduced pseudouridylation of rRNA and reduced translational fidelity [26 , 27] . These defects are thought to contribute to cases of dyskeratosis congenita caused by mutations in DKC1 , the human ortholog of CBF5 . Downregulation of box H/ACA snoRNAs could reduce pseudouridylation of RNAs . To test this idea , rRNA was isolated from WT and the scc2-4 mutant and the distribution of pseudouridines was examined by the CMC-primer extension method . Mapping of pseudo-sites on rRNA revealed downregulation of pseudouridylation at positions Ψ1003 ( guided by SNR5 ) and Ψ2258/60 ( guided by SNR191 ) ( Fig 6B ) . Reduced pseudouridylation may derive from the reduced expression of Box H/ACA snoRNAs observed in the scc2-4 mutant . snoRNAs are transcribed by RNA polymerase II . The Tbf1 transcription factor has been identified at snoRNA genes [28] . We examined whether the enrichment for snoRNA genes was compromised in ChIPs performed for either of these two baits in the scc2-4 mutant . The qPCR signal for snoRNA genes for both Tbf1 and RNA Pol II ( all forms ) is similar in WT and scc2-4 mutant samples ( S6 Fig ) , suggesting defective localization of these proteins cannot account for the lower levels of snoRNAs observed in the mutant . The Paf1 complex , which is comprised of Paf1 , Ctr9 , Leo1 , Rtf1 , and Cdc73 , is important for transcription elongation . [29–31] . The Paf1 complex has been shown to contribute to the production of snoRNAs [29 , 32] . To assess whether Paf1 recruitment was affected , we tagged Paf1 at its endogenous locus and examined the enrichment for snoRNA genes by ChIP . In contrast to RNA Pol II and Tbf1 , enrichment for snoDNAs ( Box C/D and H/ACA snoRNAs ) in the Paf1 ChIP was significantly reduced in the scc2-4 mutant ( Fig 7A ) . In addition to the snoDNAs , enrichment for rDNA and RP genes was also reduced in Paf1 ChIP in the scc2-4 mutant ( S7 Fig ) . Given the global decrease in Paf1 recruitment , we next asked about mRNA and protein levels of Paf1 in the scc2-4 mutant . While mRNA levels for the subunits of the Paf1 complex were not significantly affected in our RNA seq data , we observed less Paf1 protein in the scc2-4 mutant by Western blot ( Fig 7B ) . The protein level of Ctr9 , another member of the Paf1 complex was also reduced ( Fig 7C ) , implying that the scc2-4 mutation compromises the steady state level of Paf1 complex subunits . Interestingly , deletion of PAF1 caused a reduction in protein levels of Scc2 ( Fig 7B ) . Therefore , the simplest explanation for the decreases in ChIP/qPCR signal is that Paf1 and Scc2 depend on each other for stable expression . The Paf1 complex has previously been shown to promote transcription elongation by RNA Pol I [24] . We explored whether Scc2 recruitment to snoDNAs , rDNA , and RP genes was affected in the absence of Paf1 . Deletion of PAF1 resulted in reduced enrichment for snoDNAs in Scc2 ChIP ( Fig 7D ) , implying that Scc2 and Paf1 are dependent on each other for full recruitment at these genes . While enrichment for the rDNA and RP genes was compromised in the Paf1 ChIP in the scc2-4 mutant ( S7A and S7B Fig ) , deletion of PAF1 did not affect enrichment for these loci in the Scc2 ChIP ( S7C and S7D Fig ) , suggesting that the low recruitment of Scc2E534K cannot be attributed to less Paf1 . In addition , PAF1 deletion did not compromise Scc2 binding to other Paf1 target genes , implying that the requirement of Paf1 for Scc2 recruitment was unique to snoDNAs . These results suggest that Paf1 and Scc2 cooperate uniquely in their co-recruitment to snoDNAs ( Fig 7E ) . Since mutation of SCC2 affects rRNA biogenesis , we examined actively translating ribosomes by polysome analysis . Polysome analysis of WT and scc2-4 mutant strains showed nearly identical profiles ( S8A Fig ) . To more carefully investigate the effects of the scc2-4 mutation on global protein synthesis , 35S-incorporation was examined . A mild reduction in protein synthesis in the scc2-4 mutant was observed ( S8B Fig ) . Consistently , the mutant failed to grow on plates with a sublethal concentration of cycloheximide , an inhibitor of protein translation ( S8C Fig ) . We hypothesized that a downregulation of box H/ACA snoRNAs , as observed in the scc2-4 mutant , might phenocopy a CBF5 mutant . Mutations in CBF5 show limited effects on translational efficiency , but significant effects on translational fidelity [26] . We decided to examine translational fidelity in the scc2-4 mutant . We first tested the ability of the mutant to translate IRES-dependent mRNAs . IRES elements are used by both viral and endogenous genes for cap-independent translation initiation [27] . The Cricket Paralysis virus ( CrPV ) IRES has been shown to be active in both yeast and mammalian cells [26 , 33] and able to initiate translation by directly recruiting ribosomes to the initiation sites without initiation factors . Defects in ribosome biogenesis can impair cap-independent translation initiation directed by CrPV IGR IRES elements [26 , 33] and in cellular RNAs that utilize the IRES mechanism [27] . We monitored translation initiation from the CrPV IGR IRES elements using a dual luciferase reporter which is able to measure translation in vivo in yeast cells ( Fig 8 ) . While cap-dependent translation was very weakly reduced in the scc2-4 mutant , consistent with the results of 35S-methionine incorporation ( S8B Fig ) ; IRES usage was reduced by more than 40% in the scc2-4 mutant when compared to WT ( Fig 8A ) , implying that Scc2 promotes cap-independent translation . Faithfully maintaining the translational reading frame is one of the important functions of ribosomes . Viral mRNA signals that can induce ribosomes to shift frame by one base in either -1 or +1 direction have been tremendously useful for examining translational fidelity [26 , 34] . We examined frameshifting in the scc2-4 mutant . Dual luciferase reporters able to detect -1 frameshifting ( L-A ) or +1 frameshifting ( Ty1 ) were employed . The scc2-4 mutant had an approximately 3-fold increase in L-A mediated -1 frameshifting compared to WT ( 3 . 4% to 1 . 1%; p = 0 . 006 ) ( Fig 8B ) . However , no change in +1 frameshifting was observed . We further examined the efficiency of stop codon recognition using dual luciferase reporters ( Fig 8C ) in the scc2-4 mutant . The scc2-4 mutant was able to recognize stop codons as efficiently as WT . Since paf1Δ mutants also have defects in snoRNA expression , we examined whether a paf1Δ mutant had defects in translational fidelity . Translation initiation was examined by transforming the CrPV IGR IRES dual luciferase construct in a paf1Δ mutant . Similar to the scc2-4 mutant , use of the IRES sequence to initiate translation was severely compromised in the paf1Δ mutant ( S9A Fig ) . We also observed a significant increase in -1 frameshifting in the paf1Δ mutant when compared to the WT strain ( S9B Fig ) . However , +1 frameshifting was unaffected . Overall the paf1Δ and scc2-4 mutants show deficits in translational fidelity similar to those caused by mutation of CBF5 [26] . Changes in ribosome-tRNA interactions can affect translational fidelity and therefore influence the sensitivity of yeast to translational inhibitors . Translational inhibitors that specifically bind to ribosomes serve as useful tools for examining changes in ribosome function . Anisomycin , which inhibits -1 frameshifting by interfering with binding of aa-tRNA to the A-site , was used to probe for defects in this region of the ribosome [35] . Similarly , paromomycin , which promotes translational misreading , and increases programmed -1 ribosomal frameshifting in a mof2-1 strain [36] , was used to probe interactions at the decoding center on the small subunit involving the aa-tRNA [37] . While the scc2-4 mutant was hypersensitive to paromomycin ( Fig 8D ) , as might be expected given the tendency for -1 frameshifting , inhibition of -1 frameshifting with anisomycin [35] did not rescue the growth defect observed in the mutant ( S10 Fig ) . The failure of anisomycin to rescue the growth of the scc2-4 mutant probably reflects the fact that Scc2 contributes to growth in multiple ways . Overall , our results are consistent with Scc2 normally promoting the production of rRNA and translational fidelity .
Studies in human , fly , and yeast genomes show that Scc2 binds to the promoter regions of active genes [5 , 11 , 14 , 38] . However , how Scc2 contributes to the expression of active genes is not well understood . From ChIP-seq analysis , we observed that Scc2 not only binds to the promoter regions of genes , but also in the gene body of snoRNA and tRNA genes ( Fig 2 ) . Binding of Scc2 to both promoter and genes bodies ( in particular the snoDNAs ) suggest that its role in regulating transcription might not be restricted to initiation . In support of this hypothesis , we observed that Paf1 , a protein required for transcription elongation by RNA Pol II [29] , is significantly reduced in the scc2-4 mutant . Studies in budding yeast have shown that Paf1 is recruited to the promoter and coding regions of genes and aids in the processing of snoRNA transcripts [30] . The observation that Scc2 and Paf1 recruitment depend on each other at snoDNAs suggests the hypothesis that Paf1 and Scc2 together promote production of snoRNAs . This hypothesis is further supported by the observation that paf1Δ mutants have translational fidelity defects similar to the scc2-4 mutant . Paf1 levels are reduced in the scc2-4 mutant and there is a global reduction in ChIP signal , suggesting this could contribute to differential gene expression . The translation of at least one subunit of the Paf1 complex could be compromised in the scc2-4 mutant , and this could then impact the stability of the other subunits [39] . Alternatively , the mutation in Scc2 could directly impact the stability of the Paf1 complex , for instance , at shared binding sites . The cooperation of Paf1 and Scc2 could have wider implications in humans where the Paf1 complex is overexpressed in a wide range of cancers including prostate , breast , renal , and gastric cancers [40 , 41] . Different groups of Scc2 target genes show different responses to the loss of binding . This could be due in part to other factors that operate at those gene groups . For instance , Scc2 binding to snoDNAs appears to depend on Paf1 , but this is not the case at other regions of the genome such rDNA and RP genes . Instead , RP genes are expressed at higher levels with reduced Scc2 binding . The elevation may be due to competition between cohesin binding and RNA Pol II binding at these promoters . As the scc2-4 mutant decreases cohesin loading , it may allow RNA Pol II more access ( Patrick Grant , in prep ) . The scc2-4 mutant has translation defects . Some phenotypes are similar to what we previously reported in the acetyltransferase mutant , eco1-W216G [22] , such as defects in nucleolar morphology , ribosome distribution , and rRNA production . In both cases these phenotypes could result from reduced cohesin and condensin at the rDNA . Therefore cohesin may be important for nucleolar structure and function [23] . However , a closer look reveals significant differences in both gene expression and translational processes in the two mutants . For example , only the scc2-4 mutation is associated with reduced levels of box H/ACA snoRNAs and poor translational fidelity as well as elevated levels of ribosomal protein gene messages . The acetyltransferase mutation has reduced levels of ribosomal protein gene messages , increased levels of the stress induced transcription factor Gcn4 , and a pronounced effect on translational efficiency . Reduced translational efficiency is observed in cells derived from RBS patients and in zebrafish models for CdLS , and these cohesinopathies models are partially rescued by stimulating translation [16 , 42] . We speculate that the differences in gene expression between scc2 and eco1 mutants might be based in the molecular mechanisms by which they influence gene expression . While Scc2 may maintain nucleosome free regions and promote cohesin and condensin loading , Eco1 helps to maintain a high level of stable acetylated cohesin . Loss of function mutations in these two genes may have different effects on gene expression . This is consistent with mutations in these two genes causing distinct human syndromes . Ribosomopathies are diseases caused by mutations that affect ribosome biogenesis . One such disease is dyskeratosis congenita ( DKC ) . Reduced pseudouridylation of RNA is part of the etiology of DKC , a ribosomopathy characterized by bone marrow failure , skin abnormalities and increased susceptibility to cancer [27 , 43] . Mutations in a pseudouridine synthase that modifies ribosomal RNAs and other RNAs ( DKC1/CBF5 ) cause some cases of DKC . Studies from X-DC patient lymphoblasts and fibroblasts show downregulation of Box H/ACA snoRNAs [44] , a characteristic feature in scc2-4 mutant . Interestingly , the scc2-4 mutant mimics cbf5 mutants which have an impaired ability to translate IRES containing genes . Our results suggest that the scc2-4 mutation reduces rRNA production as well as modification , possibly in part by affecting the production of H/ACA snoRNAs , which are important for pseudouridylation , resulting in reduced translational fidelity . Reduced levels of some box C/D RNAs could also affect translation due to impaired methylation of rRNA . We previously reported that human Roberts syndrome ( RBS ) cells and zebrafish models of RBS have defects in ribosome biogenesis and protein synthesis [22 , 42] , suggesting that RBS is at least in part a ribosomopathy . Zebrafish models for CdLS also show reduced translation [16] . Our findings suggest a shared feature of cohesinopathies and ribosomopathies is defective translation . Translation might therefore serve as a therapeutic target for CdLS .
All strains were derived from BY4741 ( MATa his3Δ0 leu2Δ0 met15Δ0 ura3Δ0 ) ( S1 Table ) . The scc2-4 mutations was originally isolated by the Nasmyth laboratory [1] . We constructed scc2-4 mutant strains in the laboratory by PCR amplifying the mutation fused to a drug resistance marker from yeast genomic DNA , transforming the PCR product into the desired strain background , and isolating temperature sensitive colonies . The presence of the mutation was confirmed by sequencing . Ribosomal RNA production was examined as previously described with some modifications [45] . Briefly , duplicate cultures of WT and scc2-4 mutant strains were grown to mid-log phase ( OD600 = 0 . 3 ) in SD-ura medium supplemented with 6 . 7 ng/μl uracil . 3H-uridine ( 5 μCi ) was added to 500 μL of each culture and incubated at 30°C for 5 min with aeration . Samples were treated with 2 . 5 mL 10% trichloroacetic acid ( TCA ) and 2 . 5 mg/ml of uridine . After filtration through nitrocellulose , each membrane was washed with 5% TCA , dried , and counted in a Beckman LS 6500 multipurpose scintillation counter . WT and scc2-4 strains were grown till mid-log phase YPD+CSM medium . Cells were pelleted , washed in PBS and resuspended in a similar volume of pre-warmed SD-met medium . Aliquot for zero time point was taken . Samples were then supplemented with 27 . 5 μCi 35S-methionine and 1mg/ml unlabeled-methionine . Samples were withdrawn at 15 min intervals for 2 hr . Amount of incorporated 35S-methionine in proteins was measured by adapting Kang and Hershey approach [46] . Briefly , cells were lysed in 1 . 8 N NaOH buffer containing 0 . 2 M β-mercaptoethanol . Proteins were precipitated with hot 10% trichloroacetic acid and precipitates washed twice in acetone . The precipitates were dissolved in 1% SDS and boiled for 10 min . Aliquots of samples were counted for 35S-methionine incorporation using a scintillation counter . WT and scc2-4 mutants strains were grown in SD-methionine medium to mid-log phase ( OD600 = 0 . 3 ) Cells were pulse-labeled for 2 min with 250 μCi/ml 3H-methylmethionine and chased with 5 mM cold methionine . Samples were removed at 0 , 2 , 5 and 15 min intervals and flash-frozen in liquid nitrogen . RNA was extracted from cells , run on 1% formaldehyde agarose gel , transferred to a Hybond-N+ nylon membrane ( GE Healthcare ) , and detected by autoradiography . WT and scc2-4 mutant strains were grown in triplicate to mid-log phase ( ~OD600 = 0 . 8 ) in YPD+CSM medium . Cells were pelleted and RNA was extracted . RNA integrity was checked following isolation using an Agilent Technologies 2100 Bioanalyzer and a 1 . 2% formaldehyde agarose gel . Samples were depleted of ribosomal RNA using the Epicentre Ribo-zero Gold kit ( human/mouse/rat , cat# RZG1224 ) according to manufacturer’s protocol . Libraries for sequencing were prepared using the Illumina Truseq RNA library preparation kit ( cat# RS-930-2001 ) according to the manufacturer’s protocol . Libraries were pooled and run on an agarose gel , size selected from 200-400bp ( with adapters ) , and run on the Hi-seq . Reads were mapped to saccCer2 using tophat to generate bam alignment files . Gene coordinates from ensembl biomart were then iterated over to count the number of reads mapping to each feature . These counts were then used to generate gene expression coefficients and statistics using the DESeq package in R . The GO term enrichment analysis was carried out using the hypergeometric test as implemented in the GeneAnswers package from Bioconductor . For the metagene analysis , reads were extended to 150 base pairs before calculating coverage . Coverage was adjusted to reads per million . For a group of genes , RPM coverage was extracted for 600 bp on either side of the TSS and the values were averaged per base pair for the group . The generated fastq files from RNA sequencing were aligned using tophat [47] against the S . cerevisiae genome assembly EF4 and gene annotation from the Ensembl release 69 ( Oct . 2012 ) with the default options . Between 17 . 4M and 26 . 0M reads were aligned with 82 . 2% to 85 . 1% uniquely mapped reads per library . Splicing efficiency ( expressed as Percent Spliced Out or PSO ) was measured by computing the mean nucleotide coverage over every spliced junctions divided by the mean nucleotide coverage for reads falling within the exonic splicing unit ( i . e . counting all the reads overlapping within the two exons flanking the splice junctions and dividing by the length of the two exons ) . Significant changes in PSO were computed by performing ANOVA for each splicing unit . The computed probabilities of the difference between the means were adjusted for multiple testing using the Benjamini and Hochberg method . Genome-wide chromatin immunoprecipitation sequencing ( ChIP-seq ) was performed as previously described [48] . Briefly , yeast strains were cultured in duplicate in YPD to mid-log phase ( OD600 = 0 . 8 ) . Strains were crosslinked for either 15 min or 2 hrs ( for Scc2 ChIP ) with 1% formaldehyde at room temperature with occasional swirling . Crosslinnking was quenched with 2 . 5M Glycine . Cultures were pelleted , washed with cold PBS , and resuspended in PBS and placed at 4°C overnight . Cells were spheroplasted with 2 . 5 mg/ml zymolyase . Spheroplasts were washed and sonicated in SDS lysis buffer ( 1 mM PMSF , 0 . μg/ml pepstatin A , 0 . 6 μg/ml leupeptin ) to between 300–1000 bp in length . Chromatin extracts were diluted in immunoprecipitation ( IP ) buffer , debris was pelleted , and the supernatant was decanted into conical tubes . This chromatin solution was aliquoted for IPs . Chromatin extracts were immunoprecipitated overnight with antibody at 4°C followed by the addition of IgG beads . Beads were washed in several steps , and DNA was recovered in 1% SDS/0 . 1 M NaHCO3 elution buffer . Crosslinking was reversed by incubation at 65°C overnight , followed by protease treatment , phenol chloroform extraction , and ethanol precipitation of the recovered DNA . IPs were performed in duplicate for each sample . Sequence libraries were constructed and validated . Reads were mapped to saccCer2 using tophat to generate bam alignment files . Antibodies used for ChIP experiments are as follows; Myc antibody from Cell signaling ( Cat #2276 ) , HA antibody ( 12CA5 ) from Roche and RNA Pol II CTD4H8 antibody from Millipore ( 4 ul per 1mg protein IP ) . Accumulation of ribosomal protein GFP fusion proteins in strains was measured with spinning disc confocal microscopy ( Zeiss ) . The GFP intensity in WT and scc2-4 mutant strains was quantified as previously described [22] . Briefly , peak GFP fluorescence intensity in each cell was calculated by measuring the pulse height in the cell using a B1 detector ( 525/50 emission ) . For each sample approximately 10 , 000 cells were measured . To quantify the distance the empirical distribution function of two samples , the Kolmogorov-Smirnov ( KS ) statistic was used . This statistic enables us to calculate the distance between two biological replicates ( same genotype ) and between samples with different genotypes since the distribution of fluorescence intensity among GFP positive cells is non-Gaussian . Using KS-distance enables us to determine whether the distance between WT and mutant samples is significantly greater than between replicates of the same genotype . The IRES and frameshifting assays were performed as previously described [33 , 49] . Briefly , yeast strains were transformed with the indicated reporter plasmids . To examine IRES activity , overnight cultures of strains harboring CrPV IGR IRES and mutant cc-gg constructs were subcultured in SD-leu medium to mid-log phase ( approx . OD600 = 0 . 5 to 1 . 0 ) . Cells were lysed with 100 μl 1X passive lysis buffer ( PLB ) for 2 min . A dual luciferase assay kit ( Promega ) was used to measure luminescence according to the manufacturer’s protocol . IRES activity was measured using Firefly/Renilla ratio normalized to the Firefly/Renilla ratio of the wild type strain . The frameshifting and readthrough assays were conducted with dual luciferase reporters [33] . Translational fidelity assays employed the dual-luciferase reporter system using pJD375 ( the 0-frame control ) , pJD377 ( containing the Ty1 +1 PRF signal ) , pJD376 ( containing the yeast L-A virus -1 PRF signal ) . Luciferase activities were determined as previously described [50 , 51] . To measure frameshifting , firefly/renilla luciferase ratio generated from the 0-frame control reporter ( pJD375 ) was divided into that from the frameshifting signal-containing constructs ( pJD377 and pJD376 ) and multiplied by 100% to obtain the frameshifting efficiencies for each construct . For the readthrough assay , overnight cultures of strains transformed with dual luciferase construct were subcultured and grown to mid-log phase , pelleted , lysed with 1X passive lysis buffer and dual luciferase assay performed in triplicates according to manufacturer’s protocol ( Promega ) . Percentage readthrough was then calculated as previously described [33] . Readthrough occurred at a stop codon if firefly luciferase was translated following the Renilla luciferase ORF . Values obtained from firefly luciferase were normalized to Renilla luciferase activity as an internal control . The value obtained was then divided by the luciferase activity normalized to Renilla luciferase from a reporter with no stop codon and theoretically could be 100% readthrough for each reporter . Thus the percentage readthrough is expressed as Firefly/Renilla luciferase activity ratio divided by the firefly/Renilla luciferase activity ratio of the sense codon reporter and multiplied by 100 . Growth and temperature sensitivity assays were performed using 10-fold serial dilutions of cells spotted onto the indicated medium and incubated at 30°C for 2–3 days . Maximal growth rates and sensitivity to translational inhibitors were determined using the TECAN machine in YPD medium containing 5 μg/mL anisomycin , or 800 μg/mL paromomycin . Western blot experiments were conducted using trichloroacetic acid ( TCA ) precipitation as previously described [52 , 53] . Briefly overnight cultures were subcultured in YPD medium until mid-log phase ( OD600 = 0 . 6 ) . Cell were pelleted and washed 2 times in 20% TCA . Cells were resuspended in residual TCA and glass beads added . Cells were vortexed for 4 min at 3000 g and spun for 5 min at 14000 rpm . To each sample , 100 μl sample loading buffer was added and vortexed . Tris-HCl pH 8 . 8 or more was added and vortexed until the color changed to purple blue . Samples were boiled at 95°C for 5 min and loaded onto an SDS-PAGE gel . Pseudouridylation assay was performed as previously described with a few modifications [54] . The experimental procedure is summarized below . | The structure of chromosomes contributes to the production of RNAs . Chromosome structure is maintained in part by an evolutionarily conserved group of proteins known as structural maintenance of chromosomes proteins . These proteins are loaded onto chromosomes by a second evolutionarily conserved protein complex known as Scc2-Scc4 . The Scc2 component is often mutated in a human developmental disorder known as Cornelia de Lange syndrome . We find that Scc2 in budding yeast is important for the production of a group of RNAs known as snoRNAs . These RNAs play a critical role in ribosome production; without them the fidelity of ribosomes suffers . Ribosomes are the molecular machines that translate mRNAs to proteins . Scc2 is also important for the production of the RNA components of ribosomes . Our findings suggest Scc2 normally promotes the production of RNAs that support translational fidelity . We hypothesize that defects in protein synthesis may contribute to Cornelia de Lange syndrome . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | The SMC Loader Scc2 Promotes ncRNA Biogenesis and Translational Fidelity |
The bacterial second messenger bis- ( 3′–5′ ) cyclic dimeric guanosine monophosphate ( c-di-GMP ) has emerged as a central regulator for biofilm formation . Increased cellular c-di-GMP levels lead to stable cell attachment , which in Pseudomonas fluorescens requires the transmembrane receptor LapD . LapD exhibits a conserved and widely used modular architecture containing a HAMP domain and degenerate diguanylate cyclase and phosphodiesterase domains . c-di-GMP binding to the LapD degenerate phosphodiesterase domain is communicated via the HAMP relay to the periplasmic domain , triggering sequestration of the protease LapG , thus preventing cleavage of the surface adhesin LapA . Here , we elucidate the molecular mechanism of autoinhibition and activation of LapD based on structure–function analyses and crystal structures of the entire periplasmic domain and the intracellular signaling unit in two different states . In the absence of c-di-GMP , the intracellular module assumes an inactive conformation . Binding of c-di-GMP to the phosphodiesterase domain disrupts the inactive state , permitting the formation of a trans-subunit dimer interface between adjacent phosphodiesterase domains via interactions conserved in c-di-GMP-degrading enzymes . Efficient mechanical coupling of the conformational changes across the membrane is realized through an extensively domain-swapped , unique periplasmic fold . Our structural and functional analyses identified a conserved system for the regulation of periplasmic proteases in a wide variety of bacteria , including many free-living and pathogenic species .
Bacterial biofilms arise from planktonic microbial cells that attach to surfaces and form sessile multicellular communities , a process relevant to their survival in hostile habitats and for bacterial pathogenesis [1] . Recent work has identified biofilm formation as a multiphase process with strict temporal and spatial regulation , often accompanied by adaptational strategies such as phenotypic variation , development of antibiotic resistance , and virulence gene expression [2] , [3] . On the cellular level , functional differentiation events including changes in motility , cell adhesion , and secretion are among the many processes driving bacterial biofilm formation . Such a plethora of physiological responses inevitably poses the question of how regulation is achieved , and a nucleotide unique to bacteria , bis- ( 3′–5′ ) cyclic dimeric guanosine monophosphate ( c-di-GMP ) , has emerged as a key signaling molecule in this process [4] , [5] . c-di-GMP is a monocyclic RNA dinucleotide that functions as an intracellular second messenger exerting control at the transcriptional , translational , and posttranslational levels [6] . It is generated from two guanosine triphosphate ( GTP ) molecules by GGDEF domain–containing diguanylate cyclases , and degraded by phosphodiesterases containing either EAL or HD-GYP protein domains [7]–[10] . The majority of cellular c-di-GMP appears to be bound to protein , eliciting localized , rather than more diffusive , signals [5] . To date , only a few c-di-GMP receptors have been identified , but they are strikingly diverse , including a class of riboswitches [11] . Protein domains involved in c-di-GMP signal recognition include PilZ domains [12] , [13] , a non-canonical receiver domain in VpsT of Vibrio cholerae [14] , the AAA σ54 interaction domain–containing transcription factor FleQ of P . aeruginosa [15] , and the cyclic nucleotide monophosphate–binding domain in Clp of Xanthomonas campestris [16] . In other cases , c-di-GMP turnover domains can also serve as sensors for the dinucleotide . For example , in GGDEF domain–containing proteins , an RxxD motif can serve as a c-di-GMP-binding inhibitory site either to regulate the activity of active enzymes ( e . g . , PleD of Caulobacter crescentus and WspR of P . aeruginosa ) [17] , [18] or to mediate protein–protein interactions in degenerate homologs ( e . g . , PelD of P . aeruginosa and CdgG of V . cholerae ) [19] , [20] . Bacterial proteins that mediate c-di-GMP turnover and signal transduction are often composed of multiple domains , allowing for a variety of regulatory inputs , signaling events , and/or physiological responses [21] . For example , a large number of these proteins contain both GGDEF and EAL domains in the same polypeptide chain . These proteins fall into three main categories based on their catalytic activity: tandem domain–containing proteins with both diguanylate cyclase and phosphodiesterase activity; proteins with only one active domain , in which the degenerate , inactive domain exhibits a regulatory function; and proteins in which both domains are degenerate and likely to work as c-di-GMP receptors [22] , [23] . Despite the frequent occurrence of this signaling module in bacterial genomes , structural and mechanistic insight regarding their function and regulation is sparse . The transmembrane protein LapD belongs to the last group . It contains degenerate GGDEF and EAL domains that lack catalytic activity , but it is capable of c-di-GMP binding via its divergent phosphodiesterase domain [24] . LapD is required for stable cell attachment and biofilm formation in P . fluorescens and P . putida [25]–[27] . It responds to changes in cellular c-di-GMP levels modulated by the availability of inorganic phosphate , an essential nutrient that is limiting in many ecosystems [24] , [28] . Under phosphate starvation conditions , the expression of the phosphodiesterase RapA is upregulated , reducing cellular c-di-GMP levels and cell attachment . Increased phosphate availability yields an inactive Pho regulon , reduced RapA expression , and , as a consequence , a rise in cellular c-di-GMP concentration . As c-di-GMP levels change LapD switches between two states: the dinucleotide-unbound off state that retards stable biofilm formation by facilitating the secretion of the cell surface adhesin LapA , and the c-di-GMP-bound on state that supports cell adhesion by preventing the release of LapA from the outer membrane [24] , [26] . Binding of c-di-GMP to the LapD EAL domain is relayed to the periplasmic output domain through an inside-out signaling mechanism that utilizes a juxtamembrane HAMP domain , a relay module often found in bacterial transmembrane receptors [24] . Accompanying work by Newell et al . [29] reveals the complete c-di-GMP signaling circuit by which LapD controls cell attachment in response to phosphate availability . For wild-type LapD , c-di-GMP binding appears to induce a conformational change , which activates the receptor . As a consequence , the affinity of the periplasmic domain for the cysteine protease LapG increases , limiting its access to LapA . Perturbations in the HAMP domain by deletion of some key elements yield a constitutively active receptor , independent of dinucleotide binding . However , it has remained unclear what prevents LapD from adopting an active conformation and how dinucleotide binding translates into an output signal . Here , we present three crystal structures of LapD from P . fluorescens that provide models for the c-di-GMP-unbound cytoplasmic domain lacking only the HAMP domain , a c-di-GMP-bound EAL domain dimer , and the periplasmic domain . Together these structures span almost the entire receptor and elucidate molecular mechanisms that regulate LapD function . The crystal structure of the cytoplasmic module containing the GGDEF–EAL tandem domains reveals the presence of an autoinhibitory motif formed by a helical extension of the HAMP domain . In this inactive state , the GGDEF domain restricts dinucleotide access to the EAL domain module . The crystal structure of dimeric , c-di-GMP-bound EAL domains provides insight into the conformational changes resulting from dinucleotide binding . Based on the crystal structure of the periplasmic output domain of LapD , we identify functionally important residues and propose a model for the regulation of LapD activity in inside-out signal transduction . Finally , our structural studies highlight many conserved features that allow us to identify similar signaling systems in a variety of bacterial strains including common pathogens such as V . cholerae and Legionella pneumophila .
In order to elucidate the molecular mechanism that regulates LapD function , we determined the crystal structure of the intracellular module of P . fluorescens LapD , comprising a HAMP–GGDEF domain linker segment and the degenerate GGDEF–EAL domain module ( LapDdual; residues 220–648 ) ( Figure 1 ) . Based on secondary structure predictions , the linker forms a continuation of the second HAMP domain helix ( Figure S1 ) . We will refer to this motif as the signaling helix ( S helix ) in analogy to helical extensions found in association with other HAMP domains , where they are involved in transducing signals through the HAMP domain to the adjacent signaling modules [30]–[32] . The structure of LapDdual ( space group P32 , one molecule in the asymmetric unit ) was solved by single-wavelength anomalous dispersion phasing using selenomethione-substituted protein crystals ( Table S1 ) . We also obtained a second crystal form involving different crystal packing contacts ( space group I23 , one molecule in the asymmetric unit ) , yet the overall structure of LapDdual in the two crystals is identical ( root mean square deviation [rmsd] of 0 . 9 Å over all atoms; Figure S2A and S2C ) . In both cases , the biologically significant unit was predicted to be a monomer , based on energetic and geometric estimations [33] . The overall fold of the GGDEF and EAL domains in LapD is very similar to those of other active or inactive diguanylate cyclases and phosphodiesterases , respectively ( GGDEF domains: PleD , WspR , and FimX , average Cα–rmsd of 1 . 2 Å; EAL domains: YkuI , BlrP1 , FimX , and TDB1265 , average Cα–rmsd of 1 . 6 Å; Figures S3 and S4 ) [17] , [18] , [34]–[37] . The GGDEF signature motif in LapD consists of residues RGGEF , placing a glycine residue at the position of the active site residue that coordinates a divalent cation important for catalysis in active cyclases [38] ( Figure S3 ) . In addition , a non-conservative substitution introduces a charge change in another metal-coordinating residue in PleD ( D327 ) , which is an arginine residue in LapD ( R273 ) . Other significant changes that affect activity concern positive residues in PleD that interact with the phosphate moiety of GTP ( K442 and K327 ) . In LapD , these residues are glutamates ( E388 and E392 ) . In general , changes rendering LapD inactive for cyclase activity are comparable to those observed in FimX [36] . Similarly , the EAL domain of LapD contains non-conservative changes in residues important for catalysis ( Figure S4 ) . Most strikingly , the first residue of the signature EAL motif , which is involved in the coordination of a metal ion , is mutated to a lysine residue in LapD ( KVL motif ) [34] , [35] , [37] , [39] . LapDdual adopts a compact , bilobal conformation ( Figure 1A ) . The GGDEF domain , comprising the N-terminal lobe , caps the dinucleotide-binding pocket of the EAL domain , which forms the C-terminal lobe of the tandem domain structure . The EAL domain buttresses the N-terminal S helix via predominantly hydrophobic interactions , burying 1 , 170 Å2 ( Figure 1 ) . The binding groove on the EAL domain , which accommodates the S helix , consists of the helix α6 and an adjacent loop . The latter has been identified as a conserved motif in catalytically active EAL domain–containing phosphodiesterases , in which it is involved in dimerization and catalysis [34] , [40] . In LapD , the consensus sequence of the loop determined for active phosphodiesterases is not conserved [40] . This loop was referred to as loop 6 in SadR/RocR [40] and β5-α5 loop in the light-regulated phosphodiesterase BlrP1 [34] . We will refer to this motif as the switch loop of LapD , in analogy to the switch regions in G proteins . In addition to the S helix–EAL domain interaction , the GGDEF domain contacts the dinucleotide-binding surface of the EAL domain at multiple points , forming a loosely packed interface that buries 1 , 620 Å2 of surface area ( Figures 1A , S5A , and S5B ) . One such contact , the salt bridge between an arginine residue ( R450 ) and a glutamate residue ( E262 ) , forms a particularly close interaction ( Figure S5A ) . R450 is located just downstream of the signature EAL motif ( KVL in LapD ) at the center of the c-di-GMP-binding site . E262 is presented by a loop of the GGDEF domain . While E262 directly occupies the dinucleotide-binding site , the loop itself is located at its periphery , partially blocking access of c-di-GMP to the EAL domain ( Figure S5B ) . Although the conformation of apo-LapD observed in the crystal structure is incompatible with c-di-GMP binding , the binding site is not completely occluded ( Figure S5B ) , and there may be a sufficient proportion of accessible EAL domains in solution to respond to increasing c-di-GMP concentrations , competing with the inhibitory interactions . In addition , there may be cooperative effects within the dimeric , full-length receptor that are not apparent from the structures of the isolated domains . The loop that connects the S helix to the GGDEF domain adopts a conformation that is identical to the linkage between active diguanylate cyclase domains and their regulatory domains ( Figure S5C ) . The conformation is stabilized by a salt bridge between two strictly conserved residues that are located at the beginning of the connecting loop and just upstream of the signature GGDEF motif ( 318RGGEF322 in LapD ) , respectively: D239 in the loop and R316 in the GGDEF domain of LapD , D174 and R249 in WspR , and D292 and R366 in PleD [17] , [18] , [38] , [41] . This interaction likely constrains the loop conformation , restricting the overall rotational freedom of the GGDEF domain relative to its associated regulatory module , the S helix in the case of LapD and the response receiver domain in the case of PleD and WspR . In summary , the structural analysis of the cytoplasmic domain of LapD reveals that in the absence of c-di-GMP , the protein resides in a conformation incompatible with dinucleotide binding , with the GGDEF domain restricting access of c-di-GMP to the EAL domain . Dinucleotide binding would be accompanied by a major conformational change disrupting the conformation observed in the crystal structure . The crystal structure of LapDEAL bound to c-di-GMP ( residues 399–648; LapDEAL•c-di-GMP; Figure 2 ) was solved by molecular replacement using the EAL domain from apo-LapDdual as the search model ( Table S1 ) . We obtained crystals in two independent conditions , yielding two different crystal forms ( space group C2221 , two molecules per asymmetric unit; and space group P6522 , one molecule per asymmetric unit; Figure S2B and S2C ) . While the majority of the crystal packing contacts were different , both crystal forms maintained a common dimer of EAL domains , and the resulting structures superimposed almost perfectly ( rmsd of 0 . 6 Å over all atoms ) . Structures of the apo-EAL domain or c-di-GMP-bound LapDdual could not be obtained to date , and the structural comparison will be made between the isolated EAL domain bound to c-di-GMP and apo-LapDdual . c-di-GMP binding did not alter the overall conformation of the EAL domain observed in the apo-LapDdual structure ( rmsd of 0 . 6 Å over all atoms ) ( Figure 2 ) , consistent with the lack of major conformational changes upon dinucleotide binding to the EAL domains of YkuI , TDB1265 , and FimX [35]–[37] . Minor changes in the dinucleotide-binding pocket are confined to four c-di-GMP-coordinating residues that adopt an alternate side chain rotamer conformation ( Figure 2A ) . The most notable conformational change in LapDEAL upon c-di-GMP binding occurs in the switch loop ( Figure 2B ) . Dinucleotide binding and the absence of the S helix in the isolated EAL domain allow the loop to restructure , resulting in the switching of the conserved phenylalanine residue F566 ( Figure 2B ) . In apo-LapDdual , the side chain of F566 faces inward and is located at the center of the S helix–binding interface ( Figure 1B ) . In contrast , the switch loop adopts a conformation in the c-di-GMP-bound structure positioning F566 so that it can participate in homodimerization ( Figures 2B and 3 ) . Whether this change is due to the flexibility of the loop , adjusting its conformation to accommodate the S helix–bound and dimeric states , or depends on dinucleotide binding awaits further structural analysis . The symmetric LapDEAL domain dimer is reminiscent of the oligomeric state in active EAL domain–containing phosphodiesterases , such as in P . aeruginosa SadR/RocR , Bacillus subtilis YkuI , Thiobacillus denitrificans TDB1265 , and the BLUF domain–regulated photoreceptor BlrP1 from Klebsiella pneumoniae , where dimerization is involved in positioning an aspartate residue that in the active protein coordinates a cation for efficient catalysis ( Figure S4 ) [34] , [35] , [37] , [40] . Most importantly , dimerization of the c-di-GMP-bound EAL domains is incompatible with the conformation observed in the crystals of apo-LapDdual ( Figure 3C ) . The surface occupied by the S helix overlaps significantly with the homodimerization interface , which indicates that dinucleotide-induced conformational changes will include the displacement of the GGDEF domain and the S helix . More generally , the preservation of EAL domain dimerization in LapD and the conformational change of the switch loop upon c-di-GMP binding suggest their importance for signaling and regulation in GGDEF–EAL domain–containing proteins . Based on the crystallographic data , a simple model would suggest that LapD is subject to an autoinhibition mechanism . In contrast to other c-di-GMP receptors with known structures , in which the dinucleotide-binding site is freely accessible in the apo state ( Figure S6 ) , intramolecular interactions restrict dinucleotide access to the EAL domain in LapD . c-di-GMP binding would disrupt these interactions , resulting in a change in conformation of the receptor . Alternatively , mutations in the regulatory features predicted to destabilize the interaction should relieve the autoinhibition and alter the shape and activity of the receptor . To test this model , structure-guided mutations were introduced into LapD to assess the functional relevance of the autoinhibitory conformation and EAL domain dimerization ( Figure 4A ) . Site-directed mutations were introduced into the S helix that were predicted to weaken its interaction with the EAL domain without affecting EAL domain dimerization propensity ( F222A , F222E , S229D , E230A , or L232E; Figure 1B ) . Another set of mutations targeted the GGDEF–EAL domain interface , focusing on changes in the GGDEF domain that would not interfere with EAL domain function ( M252E , E262A , or E333A; Figure S5A ) . Finally , A602 was targeted for mutation . A602 was identified as a residue at the center of the EAL domain dimerization interface ( Figure 3B ) . The structure of apo-LapDdual showed A602 at the periphery of the S helix–EAL domain interaction , suggesting that perturbations at this site may maintain the autoinhibited state ( Figure 1B ) . Mutations were introduced into LapDdual , the EAL domain , and the full-length receptor . It is important to note that LapD is a dimeric receptor via its HAMP and output domains , and therefore EAL domain dimerization ( and dinucleotide binding ) represents a conformational change within the receptor , rather than a change in its oligomeric state . The comparative analyses described below reveal the basic properties of the cytoplasmic module of LapD , especially the correlation between c-di-GMP binding and dimerization ( Figures 4–7 ) . However , the specific interaction energies will likely be enhanced in the context of the full-length receptor compared to those of the isolated domains . Cell-based assays elucidate the functional relevance of these properties in intact LapD ( Figures 8 and 9 ) . We employed two methods to assess c-di-GMP binding to LapD . A gel-filtration-based assay essentially measures the off rate of nucleotide from a preformed complex . The filter binding assay is a semi-quantitative assay that allows for higher throughput and the generation of titration curves yielding an apparent dissociation constant ( Kd ) [24] ( Table 1 ) . Mutations in the regulatory motifs and dimer interface have a measurable effect on c-di-GMP binding to LapD . A single-point mutation in the S helix increased the overall dinucleotide binding and the apparent affinity of LapDdual for c-di-GMP by almost 2-fold ( S229D; Figure 4B and 4C ) . Removal of the glutamate side chain in residue 262 that occludes the dinucleotide binding site in the LapDdual structure ( E262A ) has a similar effect . In contrast , replacing A602 with a glutamate residue reduced c-di-GMP binding to LapDdual both in the gel-filtration-based binding experiment and in a filter binding assay , suggesting an interdependence of dinucleotide binding and EAL domain dimerization . We next analyzed the oligomerization state of LapDdual protein variants in solution , using static multi-angle light scattering ( MALS ) ( Figure 5 ) . This method provides the population-averaged absolute molecular weight and hence quaternary state of proteins eluting from a gel filtration column . The technique measures the intensity of scattered laser light from a particle at multiple angles , which is proportional to the product of the molecular weight and the concentration of the particle , permitting rapid and facile comparison of oligomeric equilibria across a series of mutants [42] . The wild-type LapDdual protein elutes in a single peak from the size exclusion column with a molecular weight of 43 . 5 kDa , indicating a monomeric state in solution ( Figure 5A , left column ) . Incubation of the protein with c-di-GMP shifted the peak elution volume and increased the molecular weight slightly to 54 . 5 kDa . While being monomeric in the absence of dinucleotide , both the S helix–EAL and the GGDEF–EAL interface mutants ( S229D and E262A , respectively ) showed more distinct shifts in molecular weight towards dimeric species upon c-di-GMP binding ( 77 . 5 kDa and 71 . 4 kDa , respectively; Figure 5A , left column ) . As predicted on the basis of the structural analysis , LapDdual variants containing a glutamate substitution in place of A602 ( A602E and S229D/A602E ) are monomeric in solution , independent of the presence of dinucleotide and unaffected by the additional mutation S229D . In general , the intermediate molecular weights and non-Gaussian peak shapes observed for wild-type LapDdual and the mutants S229D and E262A , predicted to be less inhibited , incubated with c-di-GMP prior to gel filtration , may indicate a fast exchange between monomeric and dimeric species relative to the data acquisition time and/or instability of the complex . To further investigate this phenomenon , we conducted concentration-dependent experiments by subjecting LapDdual to light scattering measurements at concentrations between 20 and 320 µM with or without incubation in c-di-GMP . All samples eluted as single peaks from the gel filtration column and showed no signs of unspecific protein aggregation . Protein concentration determination across the peak volume indicated that samples were diluted consistently ∼15-fold during the chromatography . All LapDdual variants were monomeric in the absence of c-di-GMP across the entire concentration range ( Figure 5B ) . LapDdual proteins with a mutation at the dimerization interface ( A602E or S229D/A602E ) were insensitive to c-di-GMP addition and remained monomeric . Wild-type LapDdual showed signs of oligomerization only at the highest concentrations tested . In contrast , the molecular weight of LapDdual variants with single-point mutations S229D or E262A , predicted to disrupt autoinhibitory features , increased in a concentration-dependent manner in the presence of c-di-GMP , indicative of dimerization of the isolated cytoplasmic domain in solution . Considering the modest dinucleotide-binding affinities ( Figure 4B and 4C ) , dissociation of c-di-GMP from LapD during gel filtration may also contribute to a destabilization of the dimeric state . To investigate this possibility , we repeated the experiments at the highest protein concentration with c-di-GMP present in the mobile phase ( Figure 5 ) . While proteins containing the A602E mutation ( A602E or S229D/A602E ) remained monomeric , wild-type LapDdual and the mutants S229D and E262A exhibited more pronounced dimerization in the same assay , with molecular weights close to the theoretical values for dimers calculated based their sequence . The observation that wild-type LapDdual displayed only a moderate , c-di-GMP-induced dimer formation when c-di-GMP was omitted from the mobile phase , but robust dimerization when the dinucleotide was present throughout the experiment , distinct from the behavior of the mutants S229D or E262A , indicates that the c-di-GMP-induced conformational changes and dimerization are reversible and underscores the interdependence of dinucleotide binding and EAL domain dimerization ( Figure 5 ) . In order to investigate the propensity for dimer interface formation under equilibrium conditions , we performed analytical ultracentrifugation experiments on wild-type LapDdual and on the S229D mutant . As expected based on the light scattering analysis , the concentration profiles of the c-di-GMP-free proteins could be well fit by a monomeric model , assuming a fixed molecular weight equivalent to the calculated value ( Figure 6A and 6B ) . When allowed to refine against a monomer:dimer equilibrium , the Kd for dimerization ( Kddimer ) refined to values of 400 µM ( 95% confidence interval: 0 to 6 , 300 µM ) and 420 µM ( 95% confidence interval: 600 to 4 , 700 µM ) , respectively . Thus , each construct exhibits only a minimal propensity for dimerization , and apo-LapDdual is statistically indistinguishable from a pure monomer population . However , and again consistent with the light scattering data , in the presence of c-di-GMP , the concentration profiles of both proteins were poorly modeled unless the bound state was allowed to form dimers ( Figure 6C and 6D ) . In this case , the refined Kddimer values were 670 nM ( 95% confidence interval: 370 to 1 , 000 nM ) and 180 nM ( 95% confidence interval: 80 to 270 nM ) , respectively . It is thus clear that in the presence of c-di-GMP , the propensity of the intracellular domain to form a dimer interface is several orders of magnitude stronger than that of the apo states of both proteins . Based on the nonoverlapping confidence intervals , it also appears that there may be a slight , but statistically significant , enhancement in the dimerization propensity of the S229D mutant , paralleling its increased affinity for c-di-GMP and the results from the light scattering experiments . By and large , comparable results were obtained for the isolated EAL domain ( Figure 7; Table 1 ) . The wild-type domain bound c-di-GMP with an apparent Kd of 13 . 1 ± 0 . 9 µM , whereas the A602E mutant showed a decreased affinity , with an apparent Kd of 36 . 3 ± 5 . 4 µM ( Figure 7A ) . Similar to LapDdual , the isolated EAL domain showed concentration-dependent oligomerization in light scattering experiments only upon incubation with c-di-GMP ( Figure 7B ) . The presence of c-di-GMP in the mobile phase stabilized the dimeric species further , although to a lesser extent then observed with LapDdual . In contrast , the EAL domain containing the A602E mutation remained monomeric even in the presence of c-di-GMP . In general , LapDdual shows a higher propensity for dimer formation than LapDEAL ( Figures 5 and 7 ) , and this behavior correlates with the stability of the nucleotide-bound complex ( see above ) . The mutation A602E severely affects dimer formation of LapDdual , and hence nucleotide binding . Additionally , the contribution of the GGDEF domain to dimerization in LapDdual would also be consistent with a larger apparent impact of the A602E mutation on dimerization . The effect of the A602E mutation is less pronounced for LapDEAL since this construct forms weaker dimers overall . Together , these data suggest a similar mode of dimerization of LapDEAL and LapDdual . However , in comparing the light scattering results in the presence and absence of c-di-GMP in the mobile phase , the greater discrepancy in residual dimerization observed for the LapDdual construct suggests that in the tandem domain the autoinhibited structure reassembles as nucleotide is withdrawn . In summary , LapD appears to be autoinhibited for efficient dinucleotide binding by structural features involving the S helix and occupancy of the c-di-GMP-binding site by the GGDEF domain . Based on the observation that the A602E mutation , located in the EAL domain homodimer interface and outside of the c-di-GMP-binding site , renders the protein monomeric and reduces dinucleotide binding , we propose that dimerization and c-di-GMP binding are interdependent events in LapDdual and LapDEAL . An additional conformational change in the cytoplasmic domain of LapD , accompanied by the release of the inhibitory S helix and/or nucleotide binding , is likely to occur as well . Stable biofilms of P . fluorescens require LapD expression and the presence of c-di-GMP [24] . To examine the contribution of inter-domain interactions to LapD's function in vivo , full-length LapD variants were assessed for their ability to promote biofilm formation in a ΔlapD mutant strain ( Figure 8 ) . We observed a range of phenotypes , from a slight reduction in biofilm formation relative to the wild-type , to strong hyper-adherent phenotypes comparable to that observed when LapD is constitutively activated by mutations in the HAMP domain [24] ( Figures 8 and 9 ) . The mutation that we predict to disrupt the S helix–EAL interface in the autoinhibited conformation , S229D , caused an “activated” phenotype , consistent with its increased dinucleotide binding and dimerization propensity in vitro ( Figures 4–6 ) . Similar results were obtained with the mutant F222E , whereas a less disruptive alanine substitution was tolerated at this position . In the apo-LapDdual structure , the E262 residue is positioned such that it would occlude binding of c-di-GMP to the EAL domain ( Figure S5B ) . Consistent with this and its increased binding of c-di-GMP ( Figure 4 ) , the E262A mutation results in an increase in biofilm formation relative to the wild-type allele ( Figure 8A ) . Yet , the E262A mutant phenotype is not as extreme as that exhibited in the case of the S229D mutation , despite comparable increases in c-di-GMP binding and dimerization by these proteins in vitro ( Figures 4 and 5 ) . This suggests that the E262A mutant is still subject to autoinhibition in vivo , albeit with higher sensitivity for c-di-GMP than the wild-type protein . Structurally , this may be explained by removal of the side chain that directly occupies the c-di-GMP-binding site without disturbing the S helix–EAL domain interaction . Other mutations showed intermediate ( L232E and M252E ) or no significant changes ( F222A , E230A , and E333A ) in phenotype , roughly corresponding to their surface exposure in the autoinhibited state structure ( Figures 1B , 8A , and S5A ) . The A602E mutation , which disrupts the dimerization interface of the EAL domain and reduces steady state c-di-GMP binding in vitro ( Figures 4 , 5 , and 7 ) , led to a small but significant decrease in biofilm formation relative to the wild-type allele ( Figure 8 ) . The observation that the A602E mutant showed a minor loss of function in vivo , distinct from the more pronounced loss of function observed with mutants in the dinucleotide binding pocket [24] , argues that dimerization increases the stability of the dinucleotide-bound state rather than being required for c-di-GMP binding per se . While this modest reduction in function in vivo seemed incongruous with the severe defect in dimerization and binding exhibited by the dual-domain and EAL domain construct in vitro , we further tested its significance by introducing the A602E mutation into activated alleles of LapD , S229D , and F222E . The reduction in biofilm formation in the double mutants was significant , corroborating that EAL domain dimerization plays a role in LapD function in vivo ( Figure 8B ) . The single mutants were also tested for their response to phosphate starvation , a physiological input for LapD-mediated signaling that leads to a reduction of cellular c-di-GMP concentration [24] , [28] . At low c-di-GMP concentration , wild-type LapD activity is downregulated , which results in the release of the adhesin LapA from the cell surface and thus a reduction in biofilm formation ( Figure 9A , top ) [24] . Mutations in the S helix–EAL domain interface ( F222E and S229D ) failed to respond to phosphate starvation efficiently , showing little to no reduction in biofilm formation ( Figure 9A ) . The effect was comparable to a deletion mutant described previously , in which a helical segment of the HAMP domain was removed , yielding a constitutively active , deregulated receptor ( Figure 9A ) [24] . In contrast , mutation of the residue in the GGDEF domain that occupies the c-di-GMP-binding site ( E262A ) showed an intermediate response to phosphate starvation , suggesting that mutant receptor function is still controlled by c-di-GMP , albeit not as effectively as in wild-type LapD ( Figure 9A ) . Similar to the trends observed in the static biofilm assay ( Figure 8A ) , other mutations in LapD showed more subtle effects in the phosphate starvation experiments ( Figure 9B ) . Collectively , these results suggest that the S helix–EAL domain interface stabilizes the off state . The interaction is the dominant autoinhibitory feature responsible for positioning the GGDEF domain to occlude the c-di-GMP-binding pocket and therefore ensure appropriate control of LapD activation in vivo . In addition , EAL domain dimerization via a conserved mode of interaction is likely to contribute to the efficiency of the signaling system by stabilizing the activated conformation , although it appears to be a secondary component of the activation mechanism . In order to shed light on how changes in the cytosolic domain are sensed in the periplasm , we determined the structure of the entire output domain ( residues 22–151; Figure 1A ) . Crystals grown with selenomethionine-derivatized protein diffracted X-rays to a maximum resolution of 1 . 8 Å ( Table S1 ) . The structure was solved by single-wavelength anomalous dispersion phasing . The final model consists of two molecules per asymmetric unit spanning residues 23–150 ( Figures 10A and S7A ) . The periplasmic output domain of LapD forms an extensively interwoven , domain-swapped dimer sharing 3 , 429 Å2 interfacial surface area between the protomers ( 1/3 of LapD's output domain molecular surface ) ( Figures 10 and S7B ) . The dimer adopts an overall V-shaped conformation . Each arm of the fold consists of two α-helices and two β-strands contributed by one of the two protomers , complemented by two β-strands flanked by helical segments from the other . The N- and C-terminal helices of LapD's output domain presumably connect directly to the transmembrane helices and the HAMP domains . The two half sites are linked via a long connecting segment that crosses over at the center of the dimer . The two protomers superimpose well except for a subtle rigid body rotation around the linker ( Figure S7A ) . A DALI ( distance-matrix alignment ) search comparing LapD's output domain to proteins in the RSCB Protein Data Bank ( PDB ) revealed structural similarity of its domain-swapped arms to the periplasmic domain of the sensor histidine kinase CitA ( Z-score = 5 . 4 , rmsd of 2 . 5 Å ) [43]–[45] . The periplasmic modules of CitA and related proteins show some homology to PAS domains and have been classified as PDC ( PhoQ-DcuS-CitA ) protein domains [46] , [47] . Such domains occur in many other bacterial transmembrane proteins , but unlike LapD's output domain , they are found to form a variety of regular , non-swapped dimers [44] , [47] , [48] . A sequence alignment of 18 sequences was constructed , including LapD homologs from other Pseudomonas strains and extending to more distantly related sequences from other bacterial genera ( Figure S1; Table 2 ) . Mapping sequence conservation onto the accessible molecular surface revealed a few potentially important motifs ( Figures 10C and S8 . The PxWF and LW segments ( residues 103–106 and 144–145 of LapD , respectively ) form a continuous surface at the bottom of the dimer . While the LW segment is part of the surface that accommodates the long N-terminal helix of the adjacent protomer , the PxWF is likely to interact with the inner membrane . The other striking feature is a strictly conserved loop connecting the strands β3 and β4 formed by the conserved GWxQ motif ( residues 124–127 of LapD ) . W125 forms the most distal point of the periplasmic domain located at the center of the loop , and its side chain is in an outward-facing rotamer conformation ( Figure 10C ) . Given its strict conservation and peculiar conformation , we targeted W125 in a site-directed mutagenesis study , replacing its side chain non-conservatively with a glutamate residue . The mutant output domain expressed and purified indistinguishably from the wild-type protein but had distinct functional properties . In a purified system using hexahistidine ( His6 ) –tagged LapG , a periplasmic cysteine protease that binds to LapD's output domain in a c-di-GMP-dependent manner ( see Newell et al . [29] ) , we could efficiently pull down the untagged wild-type output domain ( Figure 10D ) . Luminescent detection-based quantification indicates a binding stoichiometry of two LapG molecules per output domain dimer at saturating conditions . This result indicates that in the absence of the transmembrane and cytoplasmic domains , the output domain adopts a LapG-binding-competent state . In contrast , the output domain mutant W125E failed to interact with LapG in this assay . Consistent with these results , a full-length allele harboring the W125E mutation failed to restore LapD-dependent biofilm formation in a ΔlapD genetic background ( Figure 10E ) . The periplasmic loss-of-function mutation is also dominant over the highly activating S229D mutation when introduced in the same allele , underlining the functional importance of W125 in transmitting cytosolic signaling events to the periplasm . Our structural analyses of LapD revealed an autoinhibited conformation of the cytosolic domains in the absence of c-di-GMP , a dimeric state of c-di-GMP-bound EAL domains in the active state , and a domain-swapped dimer of the periplasmic output domain that is competent for LapG binding . The HAMP domain was modeled based on available structural information for this relay module , with the S helix forming a continuous extension of the HAMP domain's second helix [49] , [50] . In conjunction with the biochemical and genetic analyses described in an accompanying manuscript , we propose the following model for the activation of LapD and its mechanism of inside-out signaling across the inner bacterial membrane ( Figure 11 ) . The S helix and GGDEF domain function as a physical lock , gating access of c-di-GMP to the EAL domain . In this conformation , LapD's output domain is held in a LapG-binding-incompetent state , and hence LapG gains access to and cleaves LapA , releasing this critical biofilm adhesin from the cell surface . An increase in the cellular c-di-GMP level , concomitant with a sampling of a c-di-GMP-binding-competent conformation of LapD , will outcompete the inhibitory interactions in the cytoplasmic domains , likely accompanied by a large conformational change allowing EAL domain dimerization . Coupling between dimerization and c-di-GMP binding may further contribute to the efficiency of the activation switch , by preventing reversal to the autoinhibited state . Many mutations in the cytoplasmic module including the HAMP domain lead to aberrant , constitutive activation of LapD ( Figures 8 and 9 ) [24] . These data suggest that intrinsic autoinhibitory interactions are indeed necessary to prevent the system from adopting a constitutively active conformation . Based on the primary sequence and secondary structure predictions , the HAMP domain is directly linked to the GGDEF–EAL domain module via the S helix . HAMP domains occur in a large number of predominantly transmembrane sensor proteins that transmit signals from the environment across the cell membrane to elicit an intracellular response ( outside-in signaling ) [21] . Rotation of the helices in HAMP dimers has been described as the main mechanism for signal transmission [49] . It is conceivable that the EAL domain–S helix interaction stabilizes the off state , and that the release of the EAL domain from the S helix will allow the receptor to relax . The disengagement may trigger a rotation in the HAMP domain in a similar fashion to in other HAMP domains [49] , [50] , yielding a conformational change in the output domain and allowing the periplasmic domain of LapD to sequester LapG . What is the relevance of the unusual fold of LapD's output domain ? Unlike CitA and related sensor proteins , which bind small molecules in the periplasm and relay this information to the inside of the cell , LapD sequesters a periplasmic protein upon receiving a cytosolic signal . We speculate that a domain-swapped fold would respond more efficiently and precisely in coupling conformational changes in the cytosolic domains across the membrane than canonical dimeric periplasmic domains . One may consider the periplasmic domain of LapD as a single domain given the extensive sharing of structural elements and a negligible monomer–dimer transition . Given the functional importance and the particular position of W125 , we hypothesize that the output domain may act as a molecular ruler , with the tryptophan residues forming the tips of a caliper . Varying the angle between the arms of the V-shaped fold upon c-di-GMP-triggered HAMP domain rotation could form the basis for modulating binding of LapG in the periplasm , assuming that both tryptophan residues of the dimeric , periplasmic fold interact with LapG ( monomers or dimers ) . Although competent for specific LapG binding , the isolated LapD output domain failed to compete for LapG sequestration with the full-length c-di-GMP-bound receptor ( P . D . N . , unpublished data ) . It is likely that the intracellular and transmembrane domains facilitate the formation of a stable , high-affinity state . In addition , removal of the domain from its native context may alter its conformation . The observation that the isolated output domain can bind LapG is consistent with a model in which the dinucleotide-free , intracellular domains hold the receptor in an autoinhibited conformation that relaxes into a LapG-binding state upon activation . Consequently , deletion of the regulatory domains would allow for the output domain to adopt the active , LapG-binding conformation . In addition , potential higher-order oligomerization of LapD into lattices may contribute to sequestering LapG over larger membrane surfaces and to the fine-tuning of the signaling system . Two crystal structures described here , of the output domain and the c-di-GMP-bound EAL domain , show some potentially relevant higher-order interactions ( Figure S7C and S7D ) . Further experiments will be required to determine the oligomeric state of full-length LapD in the absence and presence of c-di-GMP . Based on sequence conservation , LapD homologs in other Pseudomonas strains , including P . putida and P . aeruginosa , are likely to function in a similar fashion ( Figure S1; Table 2 ) [24] , [27] . While LapD and LapG from P . aeruginosa ( PA1433 and PA1434 , respectively ) show a high degree of sequence conservation and functionally rescue deletions in these genes in P . fluorescens , no biofilm phenotype has been associated with this signaling system in their native strain [23] , consistent with the absence of an obvious LapA homolog in this species . In contrast , we identified similar effector systems and targets in more distant genera including Legionella and various Vibrio strains . In all these bacteria , lapD and lapG homologs with conserved , functionally important residues exist within the same operon ( Figure S1; Table 2 ) . LapD from V . cholerae El Tor represents a special case since its EAL domain is encoded by a second gene , separated from the transmembrane receptor containing the output , HAMP , and GGDEF domains . While the relevance of this finding requires further investigation , these genes have been found upregulated in rugose strains of V . cholerae , associated with increased biofilm formation [51] . The bioinformatic analysis also detected the presence of associated ABC transporters in genomes encoding LapD homologs , as in the case of P . fluorescens . Putative substrates of the cysteine protease LapG may fall into one of two categories . Newell et al . [29] identified the large adhesin LapA as a LapG substrate , involved in biofilm formation and stability in P . fluorescens . Based on the cleavage site sequence , other LapA homologs were identified in a variety of strains . In addition , we predict that LapG homologs may have different substrates in systems for which no clear LapA-type proteins could be identified . Regions with homology to the LapG-cleavage site of LapA have been identified in RTX-like bacterial toxins , and for the majority of such candidate substrates , these proteins are encoded in close genetic proximity to lapD and lapG homologs . The GGDEF–EAL domain–containing proteins described here are degenerate with respect to their active sites , lack catalytic activity , and function as c-di-GMP receptors . A similar system has been previously described in Escherichia coli . Unlike LapD , the transmembrane HAMP–GGDEF–EAL domain–containing protein CsrD regulates degradation of regulatory RNAs , but we speculate that the cytosolic module may be autoregulated in a similar fashion [52] . Other proteins containing the tandem domain module with a higher degree of conservation at the putative enzyme active sites exist in association with a HAMP domain in some bacterial genomes ( e . g . , V . cholerae ) . The mechanism described for LapD may also be applicable to these systems , in which the HAMP domain and S helix could be regulatory features to control the phosphodiesterase and/or diguanylate cyclase activity in the outside-in signaling mechanism , thus leading to changes in cellular c-di-GMP levels . Here , we elucidated the molecular mechanism underlying the function and regulation of P . fluorescens LapD , a transmembrane receptor essential for biofilm formation in this strain . Similar receptors are conserved in many bacteria where they control a LapG-type , periplasmic protease . LapD is autoinhibited with regard to c-di-GMP binding by interactions of the EAL domain with the S helix and the GGDEF domain . Receptor activation requires the concurrent release of the EAL domain from these interactions and the binding of c-di-GMP , which triggers a conformational change in the output domain from an incompetent to a competent state with regard to LapG binding [29] . Mutations in the regulatory features that weaken the autoinhibitory interactions render LapD constitutively active even under phosphate starvation ( low c-di-GMP levels; Figure 9 ) . This is in contrast to other c-di-GMP receptors with known structure , such as PilZ domain–containing proteins [53] , [54] , VpsT [14] , and the GGDEF–EAL domain–containing protein FimX [36] . In all these cases , the c-di-GMP-binding site appears to be readily accessible in the apo states ( Figure S6 ) . In PlzD , dinucleotide binding introduces a conformational change that changes the relative orientation of its two domains [53] . In FimX , the EAL domains form the distal tips of an elongated , dimeric protein [36] . c-di-GMP binding to the isolated EAL domain or the full-length protein is indistinguishable , and no major conformational change has been observed for FimX upon dinucleotide binding , suggesting a mode of signal transmission that may rely on partner proteins [36] , [55] . Given the occurrence of the HAMP–GGDEF–EAL domain module in many other proteins from different free-living and pathogenic bacterial species , the results discussed here will have broad implications for receptors predicted to mediate either inside-out or outside-in signaling involving the bacterial second messenger c-di-GMP .
The dual GGDEF–EAL domain module ( LapDdual; residues 220–648 ) , the EAL domain ( LapDEAL; residues 399–648 ) , and the periplasmic output domain ( LapDoutput; residues 22–151 ) of P . fluorescens Pf0-1 LapD were produced following standard molecular biology and liquid chromatography techniques . Crystals were obtained by hanging drop vapor diffusion , and datasets were collected using synchrotron radiation at the Cornell High Energy Synchrotron Source ( Ithaca , New York ) . Detailed protocols are provided in Text S1 . For MALS measurements , purified proteins ( 20–320 µM , injected concentration ) were subjected to size exclusion chromatography ( SEC ) using a WTC-030S5 ( for LapDdual ) or WTC-015S5 ( for LapDEAL ) column ( Wyatt Technology ) equilibrated in gel filtration buffer ( 25 mM Tris-HCl [pH 8 . 4] and 250 mM NaCl ) . Where specified , wild-type or mutant LapD protein variants were incubated with c-di-GMP ( 500 µM ) , produced enzymatically ( see Text S1 ) , for 30 min at room temperature prior to SEC . The SEC system was coupled to an 18-angle static light scattering detector and a refractive index detector ( DAWN HELEOS-II and Optilab T-rEX , respectively , Wyatt Technology ) . Data were collected at 25°C every second at a flow rate of 1 . 0 ml/min and analyzed with the software ASTRA , yielding the molecular weight and mass distribution ( polydispersity ) of the samples . For data quality control and normalization of the light scattering detectors , monomeric bovine serum albumin ( Sigma ) was used . Ultracentrifugation experiments were performed at 20°C in a Beckman ProteomeLab XL-A centrifuge equipped with an AN-60 rotor and absorbance optics . Sedimentation equilibrium data were recorded for 12–15 h each at speeds of 10 , 000 , 14 , 000 , and 20 , 000 rpm . Scans were taken at 1-h intervals with a 0 . 001-cm step size along the radial axis and five replicates per data point . Attainment of sedimentation equilibrium was verified using the program WinMATCH ( D . A . Yphantis and J . W . Lary; www . biotech . uconn . edu/auf ) . Six-sector cells were loaded with 1× , 2× , and 4× dilutions of ∼12 µM stock solutions of either wild-type or S229D LapDdual in 25 mM Tris ( pH 7 . 5 ) and 150 mM NaCl , either neat or supplemented to a final concentration of 20 µM c-di-GMP . Curves collected at all three speeds for all three channels were globally fit . Protein partial specific volume ( ) and buffer density and viscosity ( ρ , η ) were calculated using the program SEDNTERP [56] . Sedimentation equilibrium data were analyzed using the program SEDANAL [57] , using either single-species models or models including dinucleotide binding and protein dimerization . Proteins ( 250 µM ) were preincubated with excess c-di-GMP ( 500 µM ) at 4°C and separated from unbound dinucleotide via SEC . SEC-eluted protein peaks were collected , concentrated to a final concentration of 200 µM to normalize for protein content , heat-denatured , and filtered through Microcon Centrifugal Filter Units ( Millipore , 10 kDa cutoff ) . Dinucleotide content in the resulting samples was analyzed on a C18 reverse-phase HPLC column by using a methanol-phosphate gradient ( buffer A: 100 mM monobasic potassium phosphate [pH 6 . 0]; buffer B: 70% buffer A , 30% methanol ) [18] . Purified nucleotides were used for standardization . Integrated areas of the c-di-GMP peaks from three independent experiments were plotted relative to those for the wild-type LapDdual and LapDEAL protein constructs . Binding of [32P]-c-di-GMP to purified LapDdual or LapDEAL ( 1 µM ) was assessed by filter binding assays as described before [24] , [25] . Unspecific background binding was determined by using bovine serum albumin , and was subtracted from the data obtained for LapD-containing samples . Data were fitted to a one-site-specific binding model Y = Bmax·X/ ( Kd + X ) in GraphPad Prism ( Bmax , maximum specific binding; Kd , apparent binding constant ) . His6-tagged LapG was incubated with NiNTA superflow resin ( Qiagen ) in low-salt binding buffer ( 25 mM Tris-HCl [pH 8 . 4] , 75 mM NaCl , 25 mM KCl , and 40 mM Imidazole ) . After removal of any unbound protein in consecutive wash steps , untagged LapD output domain variants were added to the reaction and incubated for 1 h at 4°C under nutation . The resin was extensively washed in low-salt binding buffer . The remaining affinity-bound proteins or protein complexes were eluted from the slurry in elution buffer ( 25 mM Tris-HCl [pH 8 . 4] , 500 mM NaCl , and 300 mM Imidazole ) and visualized using standard denaturing gel electrophoresis ( SDS-PAGE ) . For quantification , gels were stained with SYPRO Ruby gel stain ( Molecular Probes ) following the manufacturer's directions , and imaged on a VersaDoc MP system ( Bio-Rad ) . Routine culturing of P . fluorescens Pf0-1 and E . coli was done in lysogeny broth at 30°C and 37°C , respectively . When appropriate , antibiotics were added to the medium at the following concentrations: E . coli , 10 µg/ml gentamicin; P . fluorescens , 20 µg/ml gentamicin . Plasmids were introduced into P . fluorescens by electroporation as described previously [58] . K10T medium for biofilm assays was prepared as described previously [59] . K10T-π is 50 mM Tris-HCl ( pH 7 . 4 ) , 0 . 2% ( wt/vol ) Bacto tryptone , 0 . 15% ( vol/vol ) glycerol , and 0 . 61 mM Mg2SO4 . K10T-1 medium is K10T-π amended with 1 mM K2HPO4 . A list of strains and plasmids used in the cell-based assays is provided in Table S2 . To quantify biofilm formation , strains were grown statically for 6 h in K10T-1 medium as described previously [24] . Biofilm biomass was stained with 0 . 1% crystal violet for 15 min , the stain was dissolved , and the biofilm quantified by spectrophotometry , measuring the optical density at 550 nm . We analyzed the effects of inorganic phosphate starvation on attachment by comparing biofilm levels in high-phosphate ( K10T-1 ) and low-phosphate ( K10T-π ) media over time , as done previously [24] . LapD proteins expressed in P . fluorescens Pf0-1 were visualized by Western blot as described previously [24] , with the following modifications . Blots were probed for the His6 epitope with a rabbit anti-His6 antibody ( Genscript ) . Samples consisted of clarified cell lysates prepared by harvesting cells from 3 ml of overnight culture , sonicating 3×10s in 500 µl of buffer ( 20 mM Tris [pH 8] and 10 mM MgCl2 ) , and pelleting debris at 15 , 000g for 12 min . Samples were normalized to protein concentration using the BCA kit ( Pierce ) . Atomic coordinates and structure factors have been deposited in the RCSB Protein Data Bank ( http://www . pdb . org ) under the ID codes 3pjt , 3pju , 3pjv , 3pjw , and 3pjx . | Bacteria have the ability to form surface-attached communities , so-called biofilms , in both free-living environmental habitats and during pathogenic colonization in infectious diseases . Many of the cellular processes contributing to biofilm formation , for example , changes in motility , cell adhesion , and secretion , are regulated by the nucleotide-based second messenger c-di-GMP , which is unique to bacteria . In Pseudomonas fluorescens , there are high levels of c-di-GMP within the bacterial cell when there is plentiful nutrient availability inside the cell , and the c-di-GMP levels determine stable biofilm formation outside the cell . LapD , a transmembrane receptor for intracellular c-di-GMP , communicates changing c-di-GMP levels to the outside of the cell by controlling the stability of the large adhesin protein LapA , which keeps bacteria attached to a surface or to other cells . We conducted X-ray crystallographic analyses of the structure of the intracellular and periplasmic modules of LapD that , in combination with functional studies , including those shown in an accompanying study by Newell et al . , reveal the molecular mechanisms regulating receptor function . When phosphate availability is severely restricted , intracellular c-di-GMP levels are low and LapD is in held in an “off” state by an autoinhibitory interaction , which permits the proteolytic processing of LapA , its release from the cell surface , and consequently biofilm dispersal . Conversely , when there are higher phosphate levels in the growth medium , c-di-GMP increases and binds to a cytoplasmic domain of LapD , disrupting the autoinhibitory state and triggering a conformational change that sequesters the periplasmic protease responsible for cleavage of LapA , ultimately yielding stable cell attachment . By revealing key motifs for the regulation of LapD , we have identified similar systems in many other bacterial strains that may control periplasmic protein processing events in a similar fashion . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"biochemistry/cell",
"signaling",
"and",
"trafficking",
"structures",
"biochemistry/biomacromolecule-ligand",
"interactions",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines",
"biophysics/protein",
"chemistry",
"and",
"proteomics",
"biophysics/membrane",
"proteins",
"and",
"energy",
"transduction",
"infectious",
"diseases/bacterial",
"infections",
"microbiology",
"microbiology/microbial",
"growth",
"and",
"development",
"biophysics/cell",
"signaling",
"and",
"trafficking",
"structures",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] | 2011 | Structural Basis for c-di-GMP-Mediated Inside-Out Signaling Controlling Periplasmic Proteolysis |
Segments are fundamental units in animal development which are made of distinct cell lineages separated by boundaries . Although boundaries show limited plasticity during their formation for sharpening , cell lineages make compartments that become tightly restricted as development goes on . Here , we characterize a unique case of breaking of the segment boundary in late drosophila embryos . During dorsal closure , specific cells from anterior compartments cross the segment boundary and enter the adjacent posterior compartments . This cell mixing behaviour is driven by an anterior-to-posterior reprogramming mechanism involving de novo expression of the homeodomain protein Engrailed . Mixing is accompanied by stereotyped local cell intercalation , converting the segment boundary into a relaxation compartment important for tension-release during morphogenesis . This process of lineage switching and cell remodelling is controlled by JNK signalling . Our results reveal plasticity of segment boundaries during late morphogenesis and a role for JNK-dependent developmental reprogramming in this process .
Patterning of tissue progenitors through specific gene expression precedes tissue morphogenesis . Once cells are committed to a particular lineage , they generally keep to it throughout development . Nonetheless , plasticity of segmental lineages is commonly observed during the stages of boundary sharpening , like for example during Drosophila segmentation [1] , [2] , [3] , [4] and rhombomere formation in the vertebrate hindbrain [5] , [6] , [7] , [8] , [9] , [10] . In contrast , during later development , reprogramming of patterned cells is mostly associated with pathological conditions ( e . g . regeneration ) [11] or experimental procedures ( e . g . cloning , grafting , or overexpression of selector genes ) [12] . Rare cases of fate switching have nonetheless been reported during somitogenesis and hindbrain segmentation in the chick embryo [5] , [13] , [14] and during Caenorhabditis elegans embryogenesis [15] . Still , whether patterning can be re-adjusted during late tissue morphogenesis remains elusive . Dorsal closure in Drosophila embryos is a powerful model of epithelial morphogenesis and wound-healing [16] , [17] , [18] . It proceeds through cell stretching and a zipping mechanism that lead to the convergence and suture of the lateral leading edges ( LE ) at the dorsal midline ( see Video S1 ) . This cell movement is believed to be collective and uniform . By looking at dorsal closure in live Drosophila embryos , we reveal a highly stereotyped pattern of cell reprogramming and intercalation , resulting in the remodelling of segment boundaries during late epithelial morphogenesis .
Tracking of the dorsal ectoderm cells using confocal live imaging revealed several unexpected cell rearrangements taking place within the leading edge ( Figure 1A–D and Video S2 ) . First , we observed that in abdominal segments , one cell from each anterior compartment mixes with the posterior compartment by the end of dorsal closure . We designate these versatile cells the mixer cells ( MCs; yellow in Figure 1B–D ) . These cells have been noticed recently and have been qualified as an aberration in patterning [19] . Second , we show that two cells from the ventral ectoderm intercalate into the leading edge , posterior to each MC ( Figure 1C , D ) . The two intercalating cells , one from the anterior compartment ( anterior intercalating , AI; green in Figure 1B–D ) and the other from the posterior compartment ( posterior intercalating , PI; red in Figure 1B–D ) , thus establish new segment boundaries dorsally ( Figure 1D and Video S3 ) . This striking pattern of remodelling is spatially and temporally regulated along the leading edge , with a degree of fluctuation , from embryo-to-embryo , in the timing and number of intercalating cells ( Figure 1E ) . To investigate the mixing mechanism , we analysed the origin and identity of the MCs during dorsal closure . Originally , the MCs occupy the dorsal-anterior corner of each anterior compartment ( Figure 1C , D ) . They are clearly identifiable as part of a single row of cells , known as the groove cells , which form a morphological furrow that marks each segment border , perpendicularly to the leading edge [20] , [21] . Like other groove cells , the MCs express higher levels of the actin anti-capping protein Enabled ( Ena ) ( Figure 2A; Figure S1A ) . The anterior nature of the MCs was confirmed by looking at endogenous Patched ( Ptc ) expression , which is indeed present throughout the process of cell mixing ( Figure 2A; Figure S1B ) . Thus , both its initial position as well as the expression of Ena , Ptc , and of compartment specific drivers ( ptc-gal4 positive and en-lacZ negative; see Figure S1C ) show that the MC is the dorsal-most anterior groove cell . The MC behaviour challenges the compartment boundary rule stating that cells from different compartments cannot mix due to different cell affinities that sort them out [2] , [3] , [4] , [10] , [22] , [23] . One possible explanation for the violation of this law is that the MCs may be re-programmed to acquire posterior identity . Strikingly , the analysis of endogenous protein levels revealed that the MCs start expressing the selector protein Engrailed ( En ) [24] prior to their shifting towards the posterior compartment ( Figure 2A , B; Figure S1 data ) . The profile of En accumulation in the MCs is distinct from bona fide posterior En-expressing cells present in the neighbouring posterior compartment ( Figure 2B; Figure S2 ) , suggesting that En expression in the MCs is controlled by a different mechanism . Double staining for endogenous En and Ptc shows that the MCs express both markers ( Figure 2A; Figure S1B ) , Ptc first then both , which supports the idea that the MCs were originally anterior cells that subsequently acquired posterior identity . Consistent with previous work showing that ectopic expression of En in anterior cells is sufficient to determine posterior-type cells [25] , [26] , these results suggest that the MCs undergo anterior-to-posterior reprogramming through de novo expression of the En posterior determinant , thus favouring their mixing into the posterior compartment . To demonstrate a direct role of En in MC formation , we inhibited its function in the anterior compartment by inducing en RNAi using the ptc-gal4 driver . These embryos showed a decrease of En expression in the MC ( Figure 3A ) . In addition , they exhibited a significant number of segments ( 40% ) with aberrant cell mixing , i . e . with partial or no mixing at all ( Figure 3B ) . These results indicate that de novo expression of En in the MCs is essential for their reprogramming and mixing behaviour . The differentiation of the dorsal leading edge , to which MCs belong , is under the control of the conserved JNK pathway . Embryos lacking the activity of the JNKK/hemipterous ( hep ) gene do not express the LE reporter line puckered-lacZ ( puc-lacZ ) , fail to undergo dorsal closure and die later in development [27] . Interestingly , JNKK mutant embryos are completely lacking cell intercalation and MC shifting ( Figure 4A ) . The expression of Ptc and Ena is normal in these embryos , showing that the identity of the groove cells is not affected in JNKK mutants ( Figure 4A; Figure S3A ) . In contrast , expression of En could not be detected in the MCs ( Figure 4A top; Figure S3A ) , which indicates that JNK signalling is essential for de novo En expression . To distinguish compartment specific activities , a dominant negative form of Drosophila JNK/Basket ( BskDN ) was expressed either in the anterior or in the posterior compartment using the ptc-gal4 or en-gal4 driver , respectively . The extinction of JNK activity was assessed by the loss of puc-lacZ expression ( Figure 4A ) . Embryos expressing BskDN in the posterior compartment ( en>bskDN ) showed no phenotype ( Figure 4A ) . In contrast , expression of BskDN in the anterior compartment ( ptc> bskDN ) led to the complete absence of MC intercalation , as is observed in JNKK mutant embryos ( Figure 4A , C; Figure S4 and Video S4 ) . The same result was obtained when blocking JNK signalling through overexpression of the JNK phosphatase Puckered ( Puc ) ( Figure 4C; Figure S3B ) . Absence of JNK activity in ptc>bskDN or ptc>puc embryos ( but not with the en-gal4 driver ) also led to the abolition of En expression in the MCs ( Figure 4A; Figure S3B ) . Interestingly , although most ( 87% ) ptc>bskDN embryos were able to complete dorsal closure , 92% of them showed a high degree of segment mismatching at the dorsal midline ( 53% of A1–A6 segments showed defects; Figure 4D ) . This suggests that MC formation and intercalation play a role in segment adjustment at the time of suture , consistent with a previous hypothesis [19] . In contrast , matching was normal in en>bskDN embryos . Together , these results indicate that JNK signalling is essential in the anterior compartment , most likely in the MCs , to promote anterior-to-posterior reprogramming through de novo expression of En , compartment mixing , and segment adjustment . In order to address the effect of excess JNK activity in the process , we ectopically expressed either a wild type ( Hep ) or an activated form of DJNKK ( Hepact ) in the anterior compartment using the ptc-gal4 driver . These gain-of-function conditions induced a dramatic increase in the number of intercalated cells and the formation of ectopic MCs at the segment boundaries ( Figure 4B , C; Figures S3C , S5 , and Video S5 ) . These ectopic MCs express Ptc , Ena , and En like normal MCs . These results show that more lateral groove cells are competent for reprogramming , but they are restricted by the field of JNK activity in the leading edge . Each MC has a mirror-image counterpart at the LE parasegment ( PS ) boundary ( MC* in Figure 5A ) that never develops into a MC . Interestingly , the asymmetry of the MC pattern correlates with Wingless ( Wg ) activity across the segment [28] and the presence of the groove at the segment boundary ( Figure 5A ) [20] , [21] . In addition , in JNK gain-of-function embryos , extra MCs only appear along the segment boundary ( Figure 4B ) , suggesting that only groove cells can differentiate into MCs . To test this hypothesis , we made use of specific wg mutant embryos in which an ectopic groove is formed at the PS boundary [20] . In this context , MCs* were transformed into ectopic MCs at the PS boundary ( Figure 5B ) . Like genuine MCs , transformed MCs* express Ptc , Ena , and most importantly En , which suggests that Wg suppresses the MC pathway at the PS boundary . To test whether Wg itself can repress MC formation , Wg was expressed ectopically in the MCs ( ptc>wg ) where it is not normally active [28] . This blocks MC reprogramming and cell remodelling ( Figure 5C; Video S6 ) . Consistently , En expression is no longer detected in MCs . These results indicate that Wg has a non-permissive function at the PS boundary through the blocking of groove cell differentiation , thus restricting the MC pathway to the segment boundary ( Figure 5D ) . Therefore , only dorsal groove cells are competent for MC formation ( Figure 5D ) . Dorsal closure is characterised by dramatic cell elongation ( 3-fold in the DV axis ) accompanied by the formation of a LE supracellular actin cable and amnioserosa contraction , all of which contribute to tissue tension ( see Video S1 ) [29]–[33] . To test the effect of tension on the intercalation process , we applied laser ablation to live embryos expressing βCatenin-GFP . The tension in tested segments was assessed in three conditions ( control , amnioserosa ablation , and cable ablation ) by measuring the ectoderm recoil after a single cell ablation at the leading edge ( Figure 6B ) . Increase in LE tension was induced by ablation of the pulling amnioserosa , while its release was induced through a double ablation of the actin cable on each side of a test segment ( Figure 6A ) . We next compared the dynamics of LE insertion in the controls and in embryos mechanically challenged by laser . In control embryos , the PI cell ( red in Figure 6C ) takes , on average , 14 min to complete insertion in the leading edge . This time increases dramatically when tension is reduced in the cable ( cable ablation condition; 60 min , Figure 6C middle panel , 6D ) , while it is shortened ( 4 min ) in conditions of higher tension generated by amnioserosa ablation ( Figure 6C bottom panel , 5D ) . These data show that the dynamics of intercalation depends on local tissue tension and suggest a role of intercalation in tension modulation . Improper tension release along the leading edge , in the absence of cell intercalation , could therefore explain the reduced ability of segments to match with their counterparts , as observed in JNK mutant conditions ( Figure 4D ) . To perturb tension genetically , we analyzed the pattern of intercalation in zipper ( zip , encoding MyoII ) mutants , in which a reduced tissue tension has been reported [34] . Interestingly , these embryos show a reduced level of cell intercalation , supporting our model of a link between tissue tension and the rate of intercalation ( Figure 4C ) . Based on these results , we propose that the MC pathway provides an adaptive response to tissue tension by allowing an increase of cell number in the leading edge . Indeed , one major consequence of boundary remodelling is the addition of intercalating cells ( AI and PI ) , which increases the cellular number of the leading edge by approximately 10% ( Figure 1E ) . The adaptive nature of cell intercalation is reflected by the flexibility in the number ( from 0 to 3 ) of intercalating cells ( Figure 1E ) , which contrasts with the robustness of MC reprogramming assessed by de novo expression of En . In our model , MC formation would weaken the segment boundary ( i . e . , through a change in cell affinity ) , making it a preferred site competent for tension-dependent intercalation . Our data and work published by Peralta et al . [35] indicate that the width remains constant on average with only slight oscillations during dorsal closure . Therefore , for constant width the increase in the number of cells implies that each cell is less stretched , thus inducing tension relaxation in individual cells . MC formation and associated local cell intercalation thus provide each segment with a tuneable relaxation compartment , important for tension release during morphogenesis ( Figure 6E ) . In this study we unravel the mechanism of a unique case of breaching of the segment boundary during late morphogenesis , i . e . post-patterning and post-boundary sharpening . This process is shown to be highly stereotyped and developmentally regulated through JNK signalling . Our data indicate that it takes place through a two-step mechanism , involving first MC formation , which is then followed by cell intercalation . Indeed , de novo expression of En in the dorsal groove cell always precedes intercalation ( Figure 2B ) . Furthermore , we can observe MC formation and mixing without intercalation like in the thoracic segments ( Figure 1E ) , but intercalation was never observed in the absence of MC formation: for example , when MC reprogramming is blocked in JNK loss-of-function conditions , no intercalation occurs ( Figure 4A ) . Cell mixing thus takes place through a novel morphogenetic mechanism involving plasticity of the segment boundary and compartment relaxation via patterned intercalation . It would be interesting to see if plasticity of boundaries can be a general mechanism for fine tuning late morphogenesis . Intriguingly , late expression of En in anterior cells has been reported at the anterior-posterior boundary in the wing imaginal disc . But contrary to the MC process , the so-called “S . Blair cells” do not mix with the posterior En-expressing cells [36] , and their function remains elusive . It would be interesting to reinvestigate their late behaviour using time lapse approaches [37] . Interestingly , the JNK pathway has been shown to be involved in transdetermination of injured imaginal discs [38] , reminiscent of the MC reprogramming described here . Hence , JNK signalling represents a fundamental morphogenetic and cell reprogramming pathway essential for developmental and regenerative sealing . Work on MC boundary violation and reprogramming provides a novel model to understand the molecular basis of cell plasticity .
The following fly lines were used: βcatenin-GFP ( 8556 ) , UAS-h-actin-CFP ( 7064 ) , UAS-myr-RFP ( 7119 ) , UAS-Dαcatenin-GFP , UAS-hepact ( 9306 ) , UAS-bskDN ( 6409 ) , zip1 ( 4199 ) , UAS-lamGFP ( 7377 ) ( all from the Bloomington stock center ) , UAS-en-RNAi ( VDRC#35697 ) , ptc-gal4 ( gift from N . Perrimon ) , en-gal4 ( gift from A . Brand; see Figure S6 for en-gal4 expression pattern in MCs ) , pucE69 ( puc-lacZ [39] ) , UAS-puc2a [40] , UAS-wg [41] , hep1 , hepr75 and UAS-hep4E [27] , wgcx4 , en-gal4 , and wgcx4 , armS10 [20] . The following recombined lines were used for video time-lapse of dorsal closure in various genetic backgrounds ( this study ) : ( 1 ) w*; βcatenin-GFP , en-gal4/UAS-h-actin-CFP; ( 2 ) w* , ptc-gal4 , UAS-Dαcatenin-GFP; ( 3 ) w*/UAS-bskDN; ptc-gal4 , UAS-Dαcatenin-GFP; ( 4 ) w*; βcatenin-GFP , ptc-gal4; UAS-hep4E; ( 5 ) w*; βcatenin-GFP , ptc-gal4/UAS-lacZ; ( 6 ) w*/UAS-bskDN; βcatenin-GFP , ptc-gal4; ( 7 ) w*; βcatenin-GFP , ptc-gal4; UAS-puc2a; ( 8 ) w*; βcatenin-GFP , en-gal4/UAS-lacZ; ( 9 ) w*; βcatenin-GFP , en-gal4; UAS-wg; ( 10 ) w*; βcatenin-GFP , ptc-gal4;UAS-wg; and ( 11 ) w*/UAS-bskDN; βcatenin-GFP , ptc-gal4; UAS-myr-RFP . Removal of late wg function was obtained using wgcx4 , en-gal4/UAS- armS10 embryos [20] . Embryos were dechorionated in 1 . 6% bleach , fixed for 15 min in heptane and 4% paraformadehyde diluted in PBS ( 50∶50 mix ) , devitellinised in heptane and methanol ( or chilled 70% ethanol when presence of GFP ) ( 50∶50 mix ) for 2 min using a vortex ( or incubated at −20°C for 7 min before vortexing when GFP ) , rinsed 3 times in methanol , then 3 times in ethanol , rehydrated sequentially in ethanol/PBS 0 . 1% triton solutions ( 70/30 , 50/50 , 30/70 , 0/100 ) for 5 min each time , then blocked in PBS 0 . 1% triton 1% BSA for a minimum of 2 h at room temperature before applying primary antibodies for overnight incubation at 4°C . Primary antibodies were washed 6×10 min with the blocking solution at room temperature before adding secondary antibodies for a minimum of 2 h at room temperature . Finally , embryos were treated with DAPI ( 10 µg/ml , Biochemika ) for 5 min at room temperature . 6×10 min washing in PBS 0 . 1% triton preceded mounting in Mowiol® 4-88 Reagent ( Calbiochem ) . Antibodies used: mouse anti-Ena 5G2 ( 1/500 ) , mouse anti-Ptc apa I ( 1/50 ) , anti-Wg 4D4 ( 1/500 ) ( Developmental Studies Hybridoma Bank ) , rabbit anti-En ( 1/200; Santa Cruz ) , chicken anti-β-Galactosidase ( 1/1000; Genetex ) , anti-mouse Al488 ( 1/400; Molecular Probe ) , anti-rabbit cy5 ( 1/100 ) , and anti-chicken cy3 ( 1/400 ) both from Jackson . Images were taken with a Zeiss LSM 510 Meta confocal microscope using ×40 1 . 3 NA or ×63 oil immersion objectives . Embryos were dechorionated in bleach , then staged and placed dorsal side down on a coverslip . Embryos were then coated with halocarbon oil and covered with a hermetic chamber containing a piece of damp paper for hydration . This mounting system ensures normal development of 95% of embryos . Videos last from 2 to 5 h with stacks of 25 images ( thickness from 30 to 40 µm ) taken every 5 min . Image and video assembly was done using ImageJ . Stacks are projected using either a maximal intensity or an average projection . Cell intercalations were analysed by tracking manually each cell with ImageJ . Graphs were made using Microsoft Excel . Video S3 was made using Microsoft PowerPoint and Alcoosoft PPT2Video converter . Ablations were performed using a two-photon pulsed Spectraphysic's Tsunami laser combined with a Zeiss LSM 510 Meta confocal microscope for imaging . The power was calibrated in each experiment using a test embryo and ablations were performed with the Zeiss “bleach” macro to control the size and timing of each cut . For MC ablation , the actin cable on the dorsal side was targeted , while for amnioserosa ablation , the laser beam was focused on the apical area to destroy adherens junctions and the cytoskeleton in a region of interest of 10–30 micrometers long , parallel to the AP axis . To determine cable tension , we used the classical definition of tension in a purse string as the magnitude of the pulling force exerted by the string [29] . The application of Newton's second law under the conditions of low Reynolds number ( viscous fluid ) shows that the initial recoil speed of a cable after the cut is proportional to the contribution of the suppressed force , i . e . tension [32] , [42] . ImageJ was used to quantify En and β-Galactosidase levels on projections of non-saturated stacks of images . For a given segment , the absolute intensity of En in the MCs was normalised to the average absolute intensities of the bona fide En-expressing cells of the leading edge . An average of these relative intensities was calculated for stages of intercalation as shown in Figure 2 . For each embryo , only segments A2 , A3 , and A4 were considered as they are most representative of mixing and intercalation . Relative intensity in the MCs is the ratio of absolute MC intensity/average of absolute intensities in bona fide En cells . All analyses were performed using the Mann-Whitney non-parametric test , which does not assume any condition on the distribution and is adapted to independent experiments and small sample sizes . p values were computed using the statistics toolbox from the Matlab software . | Multicellular organisms are assembled from different cell types , each following a particular fate depending on their history and location . During development , cells are organized into compartments , which are essential for the correct formation of organs . Within the compartments , cells follow two general rules: ( i ) cells that have acquired a given fate cannot change their differentiation state and ( ii ) cells from one compartment stay together and never mix with cells from other compartments . In this work , we identified a group of unique cells in Drosophila melanogaster embryos called mixer cells which move from one compartment of the epidermis to another , breaking the compartment boundary rule . Our data show that this unique behaviour depends on the nuclear reprogramming of the mixer cells , which change their fate and acquire the identity of the destination compartment de novo . We show that the shift in identity and compartment mixing are due to the expression of a single gene ( Engrailed ) , under the control of JNK , a signalling pathway that is conserved across species . Interestingly , this process of reprogramming and mixing provides a mechanism of tension relaxation to the tissue during morphogenesis that allows dorsal closure of the Drosophila embryo ( an event that resembles wound healing ) . This work reveals a novel model of cell plasticity that is amenable to genetic study , with potential application in the field of regenerative medicine . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology/cell",
"differentiation",
"developmental",
"biology",
"developmental",
"biology/pattern",
"formation",
"developmental",
"biology/morphogenesis",
"and",
"cell",
"biology"
] | 2010 | JNK Signalling Controls Remodelling of the Segment Boundary through Cell Reprogramming during Drosophila Morphogenesis |
The ventral striatum ( VS ) , like its cortical afferents , is closely associated with processing of rewards , but the relative contributions of striatal and cortical reward systems remains unclear . Most theories posit distinct roles for these structures , despite their similarities . We compared responses of VS neurons to those of ventromedial prefrontal cortex ( vmPFC ) Area 14 neurons , recorded in a risky choice task . Five major response patterns observed in vmPFC were also observed in VS: ( 1 ) offer value encoding , ( 2 ) value difference encoding , ( 3 ) preferential encoding of chosen relative to unchosen value , ( 4 ) a correlation between residual variance in responses and choices , and ( 5 ) prominent encoding of outcomes . We did observe some differences as well; in particular , preferential encoding of the chosen option was stronger and started earlier in VS than in vmPFC . Nonetheless , the close match between vmPFC and VS suggests that cortex and its striatal targets make overlapping contributions to economic choice .
Making beneficial choices about rewarding options is a major function of our brains and is critical for our survival . Consequently , understanding the mechanisms of reward-based choice is a major goal of psychology , microeconomics , animal behavior , and psychiatry [1–7] . Recent empirical and theoretical work has begun to uncover the basic underpinnings of reward-based choice ( reviewed in [8–11] ) . Research into this topic is directly inspired by the success of the perceptual decision-making research program [12 , 13] . One reason why we currently lack a correspondingly detailed understanding of reward-based choice is that the full set of brain structures involved in this process , and their specific functions , has yet to be established . In particular , it remains unclear whether reward-based choice takes place in a single core structure that has a dedicated value comparison function , or whether it occurs more broadly , as comparison steps are implemented in unison across different brain regions [14] . Among brain regions associated with reward-based choice , we are particularly interested in the ventral striatum ( VS ) and the ventromedial prefrontal cortex ( vmPFC ) [15] . Both regions are associated with option evaluation and with value comparison in neuroimaging and lesion studies [16–24] . On the one hand , this similarity in response properties suggests that they may play similar roles in reward-based choice . On the other hand , much evidence points to distinct roles for the vmPFC and VS . Specifically , VS , like other striatal regions , is generally linked to learning , including habit learning , and to action selection , while vmPFC , like other prefrontal regions , is associated with executive control and flexible , online regulation of behavior [25–40] . Of course , there is a sizable literature on the contributions of ventral striatum to reward-based choice , including action selection [37–40] . These include learning , or action-selection–centered approaches ( e . g . , actor-critic models , in which VS learns to predict future rewards , while PFC formulates a choice policy designed to maximize reward [41 , 42] ) , and gating or modulation theories , wherein the ventral striatum facilitates motor plans by disinhibiting motor plans [43 , 44] . Indeed , one recent paper found value coding in VS precedes choice but only follows choice in orbitofrontal cortex ( OFC; a structure that is adjacent to vmPFC ) , suggesting that it is VS , and not cortex , that directs the choice [45] . One common trend among these models is that they generally attribute ventral striatum and cortical areas different , and generally complementary , functions . The different roles assigned to cortical and striatal reward signals may reflect true functional differences between these areas , but it is difficult to know for certain without data collected in the same tasks with the same methods in the two areas . We recently examined the function of Area 14 of the vmPFC in a simple reward-based choice task [15] . We found that neuronal responses encode offers’ subjective values using a single value scale . That is , they integrate across dimensions to form an abstract value variable and then gradually come to encode the value of the chosen—as opposed to unchosen—option . When two offers are made , neurons show opposed tuning for their values—suggestive of a mutual inhibition choice process [15 , 16 , 46 , 47] . Finally , these neurons also showed choice probability correlations , suggesting that their activity may contribute directly to selection [15 , 48] . We argued that these responses implicate vmPFC in a mutual inhibition process that implements value comparison . On the one hand , vmPFC ( possibly along with OFC ) may be relatively unique in its role , and other connected brain areas , like VS , dorsolateral prefrontal cortex , and anterior cingulate cortex may play complementary roles less central to choice [8 , 10 , 49–53] . On the other hand , such regions may play roles similar to that of vmPFC as part of a larger , multi-site comparison process [14] . Among these areas , we are especially interested in the VS because of the widespread assumption that cortex and striatum have strongly distinct roles in cognition . The main cells of the striatum are medium spiny neurons , inhibitory ( GABA-ergic ) cells that receive inputs from cortex and that transmit information to the pallidum [44–56] . The question of just how much interaction there is within the striatum has been one of the most important ones in striatal anatomy and function over the past three decades [57] . Nonetheless , it is very likely that there is at least some within-striatum processing going on [58] . First , there is some ( but not decisive [59 , 60] ) evidence for lateral inhibition effects within the striatum [61–69] . The extent of these functional connections is quantifiable in vitro [70 , 71] . Indeed , some of these connections are reciprocal [58] . There are also more esoteric possibilities for intrastriatal interactions as well , such as nitric oxide communication through gap junctions [72–75] . In any case , there is clear support for the idea of within-striatum processing , supporting the idea that some mutual inhibition may be occurring within the striatum itself . We found that VS , like vmPFC , represents abstract values and value differences , suggestive of a process of mutual inhibition [15 , 16 , 24] . ( Note that competitive interaction within the striatum is anatomically plausible [58–60] . ) We also observed preferential selectivity for chosen , as opposed to unchosen options and choice probability correlates . Both areas encoded outcomes of gambles more strongly than other task variables . Relative to vmPFC , the effects in VS were observed at roughly the same frequency , although they were slightly more common in VS , and preferential representation of the chosen option occurred earlier in time . Aside from these differences , we did not observe any major functional differences in vmPFC and VS response properties . These findings suggest that the basic microcomputations supportive of choice processes can be observed in both cortical and subcortical reward areas . More broadly , they provide tentative support for the idea that choice cannot be localized to one specific region of the brain , but instead reflects the outcome of comparison processes occurring in multiple brain regions .
Our dataset consists of responses from 124 VS neurons ( 55 neurons in monkey B , 69 neurons in monkey C ) . Our published vmPFC dataset contains 156 vmPFC neurons recorded in the same task ( 106 in monkey B , 50 in monkey H; [15] ) . In this VS study , we recorded an average of 510 . 2 trials per neuron ( range: n = 168 to n = 813 trials ) . Neurons were localized to the ventral striatum ( Figs 2A and S4 ) . We defined three task epochs for analysis . ( To make comparison with our earlier study easier , we use the same epochs and names for epochs we used in our earlier study; [15] . ) Epoch 1 began with the presentation of offer 1 , epoch 2 began with the presentation of offer 2 , and epoch 3 began with the gamble outcome . Each epoch lasted 500 ms . We favor a 500 ms time window because ( 1 ) it allows us to detect even sluggish responses and ( 2 ) by using the same epoch across studies , we reduce the chance of inadvertent p-hacking . We found that 55 . 65% of neurons ( n = 69/124 ) showed some sensitivity to task events , as indicated by individual cell ANOVAs of firing rate against epoch for the three task epochs and a fourth 500 ms inter-trial epoch ( p < 0 . 0001 , binomial test ) . For comparison , we found that 46 . 15% of neurons ( n = 72/156 ) in vmPFC showed some sensitivity to task events . These results indicate that VS neurons are slightly , but not significantly , more likely to respond to the task than vmPFC neurons ( chi-square test , X2 = 2 . 4896 , p = 0 . 1146 ) . All proportions refer to all recorded neurons , not just ones that produced a significant response modulation . We first examined coding of economic variables ( i . e . , probability and reward size ) in the first offer epoch ( epoch 1 ) . Fig 2B shows responses of an example neuron to offers separated by their probability and reward size . In epoch 1 this neuron’s firing rates encoded the probability of offer 1 ( linear regression; β = 0 . 1602 , p = 0 . 0013 ) and the reward size of offer 1 ( β = 0 . 1667 , p = 0 . 0008 ) . We found that the firing rates of 14 . 5% of cells ( n = 18/124 ) were correlated with ( and thus , in our parlance , “encoded” ) the probability of winning ( linear regression , α = 0 . 05; Table 1 ) . This proportion of encoding neurons is much greater than expected by chance , and so is unlikely to reflect random noise ( binomial test , p < 0 . 0001 ) . In this first epoch , the same proportion of neurons ( 14 . 5% ) encoded the potential reward size available ( i . e . , gamble stakes ) . These proportions are similar to but slightly higher than the analogous proportions observed in vmPFC neurons ( during epoch 1 , 7 . 7% of vmPFC neurons encoded probability and 11 . 5% encoded reward size; [15] ) . Note that safe offers , which occurred on 12 . 5% of trials , have a fixed 100% reward probability and a relatively small reward . Therefore they make high probability offers more likely to have small reward sizes than not . This introduces a negative correlation between reward size and probability , and as a result , trials with safe offers are excluded from this analysis . We asked whether VS neurons carry an integrated reward signal , as is the case in vmPFC [15] . The alternative is that they preferentially encode either probability or reward , or even both orthogonally; such coding schemes may be used in area 13 of OFC [53 , 76] . To address this question , we compared regression coefficients for firing rate versus probability with coefficients from the regression of firing rate versus reward size ( again , limiting ourselves to epoch 1 ) . If neurons encode an abstract value form of offer value [8 , 77] then their separately calculated regression coefficients for probability and reward size should themselves be positively correlated [15] . Such abstract value coding would be consistent with the use of a single value scale to encode reward amount . Indeed , we found a significant positive correlation between these coefficients ( R = 0 . 24 , p = 0 . 007; Fig 2C ) . These data are consistent with the idea that the ensemble of VS neurons represents value in an abstract format . Moreover , they mirror those found in vmPFC ( R = 0 . 25 , p = 0 . 0023; [17] ) , and suggest that abstract reward value representation occurs in both cortex and ventral striatum [8 , 27 , 78] . Do VS neurons have qualitatively different response latencies for value coding than neurons in vmPFC ? First , we separated trials into high or low offer 1 value categories . Then , using a sliding t test , we found the first 20 ms period after offer 1 presentation where a t test on firing rates in those two sets of trials was significant at p < 0 . 05 . This process found a significant difference sometime in the 500 ms after offer 1 presentation for 23/124 VS cells and 17/156 VM cells . We found no significant difference between these VS cell response latencies ( mean: 78 . 52 ms ) and vmPFC cell latencies: ( mean: 75 . 73 ms ) for offer 1 value coding after offer 1 presentation ( t test , t ( 38 ) = 0 . 2588 , p = 0 . 7972 ) . In order to look for offer value signals in VS neurons , we operationalized an offer’s value as its gamble’s expected value , that is , its reward magnitude multiplied by its reward probability . Fig 3A shows responses of an example neuron to offers separated by their relative expected values ( value of offer 1 minus value of offer 2 ) . Its firing rates encoded the expected value of offer 1 in epoch 1 ( linear regression; β = 0 . 1705 , p < 0 . 0001 ) and in epoch 2 ( β = -0 . 0985 , p = 0 . 0149 ) . The sign flip indicates that the direction of tuning for offer 1 was reversed for the second epoch . This neuron also encoded the expected value of offer 2 in epoch 2 ( β = 0 . 1698 , p < 0 . 0001 ) , meaning that during epoch 2 it coded both values simultaneously . This propensity to code two values simultaneously was observed across the population ( Fig 3B ) . In epoch 2 , significant proportions of neurons encoded both offer value 1 ( n = 16/124 , 12 . 9% ) and offer value 2 ( n = 27/124 , 21 . 8% ) . These value signals are robust to time window changes , for example , +/- ~100 ms around the responses seen in Fig 3B . In our example neuron , tuning directions for expected values 1 and 2 have opposed signs . This anti-correlation is consistent with antagonistic coding of these offers , i . e . , representations of the two values interact competitively to influence the firing rate of this neuron ( cf . [16] ) . This antagonistic pattern is observed at the population level as well . Overall regression coefficients for offer value 1 in epoch 2 are anti-correlated with coefficients for offer value 2 in the same epoch ( R = -0 . 2919 , p = 0 . 0010 , Fig 3C ) . To match the criteria used in the above analyses , this analysis does not include trials with safe options; however , if we repeat the analysis with the safe offer trials as well , we find the same anti-correlation ( R = -0 . 2766 , p = 0 . 0019 ) . This finding of antagonistic coding may be a signature of comparison through mutual inhibition and is also observed , at a slightly weaker strength , in vmPFC neurons ( R = -0 . 218 , p = 0 . 006; [15] ) . We next looked at response latencies for antagonistic value coding during epoch 2 by separating trials by which of the two offers had a higher value . Using a sliding t test , we found the first 20 ms period after offer 2 presentation in which a t test on firing rates in those two sets of trials was significant at p < 0 . 05 . We found a significant difference sometime in the 500 ms after offer 2 presentation for 22/124 VS cells and 15/156 VM cells . We found no significant difference between VS cell response latencies ( mean: 119 . 35 ms ) and vmPFC cell latencies: ( mean: 92 . 42 ms ) for antagonistic value coding after offer 2 presentation ( t test , t ( 35 ) = 0 . 5014 , p = 0 . 6192 ) . We found 14 fast-spiking interneurons ( FSIs ) and 66 medium spiny neurons ( MSNs ) in our 124 neuron population , using waveform criteria as delineated by Jin et al . [79] . Four of the fourteen FSIs ( 28 . 6% ) and 10 of the 66 MSNs ( 15 . 2% ) significantly encode the difference between the offered values during epoch 2 ( correlation , p < 0 . 05 ) . Although small , these proportions are both greater than what would be expected by chance ( binomial tests , FSIs: p = 0 . 0004; MSNs: p = 0 . 0004 ) . The ratio of FSIs that show antagonistic coding is not significantly different from that of the MSNs ( chi-square , X2 = 1 . 4408 , p = 0 . 2300 ) . The data presented so far are consistent with the idea that VS contains two distinct memory buffers for reward value , one for currently presented options and the other for previously presented options stored in working memory ( cf . [80] ) . To further test this idea , we examined the relationship between a vector of regression coefficients for option 1 in epoch 1 and option 2 in epoch 2 for all cells . We found a significant positive correlation between these vectors ( R = 0 . 6363 , p < 0 . 0001; see Fig 3D ) . This suggests that whatever effect a larger offer 1 had on firing rates during epoch 1 in each neuron ( excitatory or suppressive ) , the same effect was observed for those neurons to a larger offer 2 in epoch 2 . This finding suggests that VS neurons use a single coding framework consistently across time to code the currently offered option ( cf . [24] ) . This consequently suggests that neurons do not use a single format to represent a single option’s value over the course of a trial . Instead , the format used is different in the context in which the option is currently offered and in the context in which the option was offered in the past and presumably remembered . This context-dependent coding pattern for offered options across the two epochs was also observed in vmPFC neurons [15] . After determining that neurons in VS encode the values of both offers simultaneously and antagonistically in epoch 2 , we next examined whether they preferentially signal the chosen one . Fig 4A shows responses of an example neuron to offers separated by the expected value of the chosen offer . Its firing rates encoded the expected value of the chosen offer in epoch 1 ( knowing offer 1 gives the monkey partial information as to his eventual chosen offer; β = 0 . 2298 , p < 0 . 0001 , linear regression ) and on into epoch 2 ( β = 0 . 2765 , p < 0 . 0001 ) and epoch 3 ( β = 0 . 2420 , p < 0 . 0001 ) . Fig 4B shows the proportion of neurons in each dataset whose activity is significantly modulated by chosen offer values ( VS: dark blue line; vmPFC: light blue line ) and by unchosen offer values ( VS: dark red line; vmPFC: light red line ) in a sliding 500 ms time window . Note that Fig 5A and 5B both show a peak during epoch 3 that is even larger than the peak in epoch 2 because the value of the chosen offer was highly correlated with the value of the outcome , the coding of which was stronger than other effects; see below . We found the same coding frequency for the value of the chosen and unchosen offers during epoch 1 ( 12 . 9% of cells for both chosen and unchosen options , n = 16/124 ) suggesting that VS does not distinguish chosen from unchosen options at this point , or if it does so , does it too weakly to detect with this analysis . In the first 500 ms of epoch 2 , we again found no difference in coding of chosen and unchosen offers ( 9 . 7% coded chosen and 12 . 9% of cells coded unchosen ) , but by the end of this epoch ( the last 500 ms ) , we saw the gradual emergence of a preference for a chosen offer . Specifically , we saw stronger coding for chosen offers than for unchosen offers ( 15 . 3% and 6 . 5% of cells , respectively; 15 . 3% is significantly more than chance , p < 0 . 0001; 6 . 5% is not , p = 0 . 1689 ) . Note that this change in coding frequency , 6 . 5% to 15 . 3% , is itself significant , χ2 = 16 . 168 , p < 0 . 0001 . As shown in Fig 4C , we observed a gradual increase in the proportion of cells whose firing rates significantly encoded this difference ( between chosen value and unchosen value ) . The horizontal lines show when the proportion of cells shown reaches significance ( gray: 5%; purple: binomial test at α = 0 . 05 on VS dataset; green: binomial test at α = 0 . 05 on vmPFC dataset ) . The proportion of VS neurons’ first significant bin was 1 . 42 s after offer 1 presentation , while the proportion of VM neurons’ first significant bin was 2 . 39 s after offer 1 presentation . These results suggest that VS has access to information about the choice process before vmPFC . Indeed , in VS the preferential coding occurred before the saccade that implements the choice , but this was not the case vmPFC . Another way to look at the timing of choice-related signals is to look at decodability of chosen offer as a function of time in the trial . In other words , we examine how accurate an ideal observing decoder would be at decoding eventual choice ( offer 1 or offer 2 ) from firing rates as the trial progresses ( see Methods ) . Fig 4D shows decodability of chosen offer in a 500 ms sliding boxcar for VS neurons ( purple ) and vmPFC neurons ( green ) . The horizontal lines show when the proportion of trials correctly classified reach significance ( gray: 50%; purple: binomial test at α = . 05 on VS dataset; green: binomial test at α = . 05 on vmPFC dataset ) . Both VS and vmPFC cells showed peaks of significant choice decodability during epoch 1 ( VS onset: 240 ms after offer 1 presentation; vmPFC onset 316 ms ) and again in epoch 2 ( VS onset: 180 ms after offer 2 presentation; vmPFC onset 440 ms ) . It is important to note that peaks of significant choice decodability are quite transient in vmPFC during both epochs and in VS during epoch 1 , and therefore may be due to statistical noise . These results suggest that choices may be more quickly and more reliably decodable in VS than in vmPFC . Note that a more sensitive test of choice-related variance did show significant encoding of choice prior to the choice saccade in vmPFC ( [15]; see below ) . To further investigate the connection between neural activity in VS and choice , we made a calculation similar to a choice probability [81] . For each neuron , we first regressed firing rate in epoch 1 onto offer 1 value , probability , and reward size , and determined the residuals . We then examined whether the values of the residuals from this regression predicted choice ( offer 1 versus offer 2 ) for each neuron . To associate the residuals with choice , we simply ran a binomial regression on choice as a function of the ( continuous ) residual variable . ( We confirmed that a simple correlation test produces similar results . ) In other words , we computed the residual variance in firing rate after accounting for the factors that influence value . We found a significant relationship between residual firing rate variance and choice in 8 . 87% of cells ( n = 11/124 ) , which is more than is expected by chance ( p = 0 . 0218 , binomial test ) . Given that offer 2 has yet to be presented and the choice is not yet made , it may seem odd for choice selectivity to be present in epoch 1 . However , we believe this result is expected in the case that response to offer 1 represents value and that value representation in turn influences choice . Likewise , residual variation in firing rate in response to offer value 2 during epoch 2 predicted choice in 12 . 90% of cells ( n = 16/124 , p < 0 . 0001 , binomial test ) . Correlations between residual firing rate variance and choice following the second offer reveal ( mean absolute value correlation coefficient: 0 . 0629 ) were stronger than following the first ( mean absolute value correlation coefficient: 0 . 0456; t test of correlation coefficients between residual firing rate variance and choice; t ( 246 ) = 2 . 40 , p = 0 . 0171 ) . This result is consistent with the idea that the population of VS neurons gradually comes to encode the chosen offer value more than the unchosen offer value as the decision emerges ( see Fig 4B and 4C ) . In our earlier study , we also observed choice probability correlates in vmPFC; the preponderance of the effects was similar in both cases [15] . These shared patterns suggest that fluctuations in VS responses , like those in vmPFC , relate , however indirectly , to ongoing choice processes in a similar way . Outcome monitoring is a prominent aspect of vmPFC responses [15] . Fig 5A shows responses of an example VS neuron with trials separated by gamble outcome . This neuron encoded received reward size in epoch 3 ( R = -0 . 8402 , p < 0 . 0001 , linear regression ) . We observed a significant encoding of gamble outcome in 33 . 9% of cells ( n = 42/124; Fig 5B ) . Of these cells , 50% ( n = 21/42 ) showed negative tuning , while the other 50% showed positive tuning . Outcome coding continued across the delay between trials , and previous trial outcome was a major influence on firing rates during both epochs 1 ( 28 . 2% of cells , n = 35/124 ) and 2 ( 16 . 1% of cells , n = 20/124 , p < 0 . 0001; Fig 5C ) . Previous trial outcome even influenced responses during the current trial’s outcome epoch in 12 . 9% of cells ( n = 16/124 ) . Above , we reported that coding for offer values 1 and 2 use a single value scale coding format . We next looked at whether the coding format for outcome was similar to that of the coding format for offer value 1 and 2 . We did so by comparing tuning profiles for outcome and offer value 1 . In particular , we asked whether regression coefficients for offer value 1 in epoch 1 were correlated with regression coefficients for received reward size in epoch 3 . Alternatively , these coefficients could be uncorrelated , which would indicate that neurons that fire preferentially for larger offer 1 values are not also the same neurons that fire preferentially for large outcomes . We found a significant correlation between these regression coefficients ( R = 0 . 2712 , p = 0 . 0023 ) . This suggests that VS neurons use a single , or at least similar , coding scheme to represent offer values and represent outcomes . This finding matches that observed in vmPFC as well ( R = 0 . 22 , p = 0 . 0054; [15] ) . Dopamine neurons do not provide a labeled line representation of reward size; instead their reward encoding is normalized by reward prediction ( i . e . , it is a reward prediction error [82] ) . Despite prominent dopaminergic inputs to VS , previous investigations into reward prediction error coding in VS neurons have had a mixture of positive [83] and negative [84] results . We performed a stepwise regression to determine whether , after accounting for reward size ( first step of the regression ) , post-outcome responses in VS are related to probability of that reward ( second step ) . Because many neurons have negative tuning , we flipped the values for neurons that had negative individual tuning profiles , as measured by regression coefficient , whether their regression coefficients were significant or not . Starting with just risky trials ( i . e . , with no small , safe offers ) , the gamble outcome regressor met the criteria for model inclusion ( β = 0 . 2017 , p < 0 . 0001 ) , but the reward probability of the chosen offer did not ( β = -0 . 0082 , p = 0 . 6926 ) . We then repeated these analyses for the medium- and high-reward size trials separately . We find similar results when examining only trials in which a medium-reward option was chosen ( gamble outcome: β = 0 . 2362 , p < 0 . 0001; chosen option reward probability: β = -0 . 0140 , p = 0 . 6743 ) and when examining only trials in which a high-reward option was chosen ( gamble outcome: β = 0 . 2224 , p < 0 . 0001; chosen option reward probability: β = -0 . 0460 , p = 0 . 0806 ) . This finding indicates that pure outcome is a better descriptor of VS outcome-related responses in this task than reward prediction error . These response patterns align VS with vmPFC in the same task , and distinguish them from dACC in a similar ( albeit not identical ) task [85] .
Here we examined the responses of neurons in VS during a gambling choice task we previously used to study the function of vmPFC [15] . We wondered whether these two areas have similar or distinct contributions to choice . We found that task-related response properties of VS neurons are strikingly similar to those in vmPFC . Specifically , we found that neurons in VS show five major response patterns . First , responses to offered gambles encode offer value ( i . e . , they signaled expected value , not probability and stakes separately ) . Second , neurons code the difference in value of the two offers ( i . e . , tuning for the two offers is antagonistic ) , suggesting a mutual inhibition or tug-of-war-like process . Third , neurons initially signal the values of both options and then gradually come to signal chosen values and not unchosen values . Fourth , residual variation in firing rate after regressing out value coding signals predicts choices . Fifth , neurons show prominent outcome-related responses . We did find some differences: there is clearer evidence that choice-related signals precede choice in VS than in vmPFC ( although these signals still reach significance before choice in vmPFC by some measures [15] ) . Overall , however , the strong overlap in functions of VS and vmPFC suggests that these two regions have similar functions , at least in the context of a straightforward economic choice task . We hypothesized that neurons in VS participate in the choice process and that they do so as part of an anatomically distributed mutual inhibition process . Specifically , following Hunt and Behrens ( 2012 ) , we hypothesized that value representations of the two offers in VS compete for control of neuronal activity [15 , 16] . This hypothesis can be contrasted with two major alternative hypotheses: ( 1 ) that choice occurs through a horse-race type process , in which there is no competition and thus no mutual inhibition , and ( 2 ) that choice does not occur in this brain region . If our hypothesis is correct , then firing rates should reflect this competition . Specifically , firing rates in neurons that represent values of the offers should be antagonistically affected by those two values ( negative correlation between regression coefficients ) , and not additively ( positive correlation ) or orthogonally ( no correlation ) , nor should separate sets of neurons represent the values of each option . Thus , we predict that firing rates should reflect the difference in values of the two offers . Our results support this hypothesis: at the population level , regression coefficients for offer 1 value are anti-correlated with regression coefficients for offer 2 value , suggesting a tendency towards antagonistic value coding . One alternative possibility is that the VS contains two groups of neurons whose firing rates represent either the value of offer 1 or of offer 2 , and that they do not respond to the other offer . This situation would be consistent with both of our alternative hypotheses . Fortunately , we can test for this possibility by comparing the absolute values of the correlation coefficients . We find that these are significantly positively correlated , supporting the idea that neurons are drawn from a single , competitively tuned population and inconsistent with the idea that they are drawn from two different populations . In other words , it does not appear to be the case that individual neurons are specialized for one of the two offers; instead , the competition takes place either in the neurons themselves , or in their inputs . At the least , we show that vmPFC and its striatal target VS both carry three classes of signals related to choices: choice inputs , correlates of middle stages of choice , and choice outputs . These matching patterns of results are consistent with the idea that vmPFC and VS play fundamentally similar , rather than contrasting , roles in choice . However , the existence of these signals does not prove that choices occur within both ( or even either ) area; one or both may receive copies of this information from other regions . Nonetheless , these data can exclude the hypothesis that choice occurs and is complete prior to information entering into VS . Another possibility is that VS , and not vmPFC , is the site at which comparison occurs . Indeed , one recent paper reported effects consistent with this hypothesis in rats [45] . Future work will be needed to resolve this question . The similarity between VS responses and those in vmPFC [15] and even OFC [77] may not appear surprising , given that OFC and vmPFC are two major cortical afferents to the region of VS in which we recorded [56] . On the other hand , most theories emphasize the contribution of VS to learning and other processes distal to the evaluation and comparison processes that directly implement reward-based choices [25–32] . In contrast , neural studies of choice processes generally focus on the cerebral cortex . The present results suggest that this view is too narrow , and that VS , like its cortical afferents , may participate directly in computations that are critical for reward-based decisions . In any case , the present results do not imply that prominent accounts of VS function in control of learning are incorrect; quite the contrary , we suspect that cortical regions may have some of the same functions generally ascribed to striatum [86] . Indeed , our task expressly minimizes the importance of learning in its design . Previous studies of ventral striatum function in a reward-based choice context have generally attributed ventral striatum and cortical areas different , and generally complementary , functions , such as actor-critic models [41] or gating/modulation theories [43 , 44] . Several research groups have shown that neurons in the dorsal striatum ( both caudate and putamen ) respond differently depending upon the reward expected from an action , suggesting they encode action values [30 , 87–89] . These neurons may directly influence choice by providing a bias signal over specific actions [37–40 , 90 , 91] . The present results are consistent with idea that VS neurons also encode action value , although they suggest it participates in comparison as well . This comparison process may contribute to the evaluation process critical in actor-critic and actor-director-critic models [29] . These models see the job of the VS to calculate the value associated with actions or abstract choice states and to drive learning accordingly . Interestingly , Samejima et al . found a conspicuous absence of value difference encoding in the dorsal striatum . These results suggest that the dorsal striatum may play a role in reward-based choice , but are generally silent on the question of the function of VS . In one recent study using a delay discounting task , Cai , Kim , and Lee [27] found prominent value sum signals in VS neurons ( 20% of cells ) and no significant encoding of value difference ( 5% of cells , the number expected by chance ) . Nor did they find significant chosen value signals in VS . From these results , they concluded that VS participates in signaling task state but does not contribute to value comparison . We find the results of the two studies to be strikingly different . We suspect that the difference is most likely due to differences in task design—specifically , the use of asynchronous presentation in our task . Hunt and colleagues have demonstrated that asynchronous and simultaneous presentation of offers in reward-based choice task can lead to differential involvement of different structures [46] . If so , by using an asynchronous presentation , we may have uncovered a comparison role that was masked by the task design of this earlier study . Despite our findings , we are reluctant to abandon localizationism . In regards to value comparison , Rushworth and colleagues have provided some evidence that lateral structures in the OFC , for example , do not participate in comparison , but instead mediate it [10 , 51 , 92] . Likewise , Wilson et al . have argued that OFC participates in state signaling but does not directly implement evaluation and comparison processes [93] . Meanwhile , we have shown some evidence that dorsal anterior cingulate cortex ( dACC ) , another reward region that provides inputs to ventral striatum , may not contribute directly to choice , at least under somewhat different choice conditions [50] . Around the time of choice , neurons in dACC carry signals that depend on outcomes of decisions , but not related to value comparison per se . These findings are consistent with other ideas linking dACC to regulation of strategic adjustments and executive control . Further afield , posterior cingulate cortex ( PCC ) shows blood-oxygen-level dependent correlates of value and salience , but does not appear to implement choice , either [94 , 95] . Instead , it seems to detect long-term changes that necessitate deeper strategic shifts , including implementation of long-term learning [96–98] . Our data do raise the possibility that multiple brain regions perform similar computations at roughly the same time . If so , then how does the output system—the motor system—adjudicate between competing decisions in order to select the single best course of action ? Our data do not provide much guidance on this topic although we hope to pursue this area in the future . Our best guess is that the brain weights different systems based on reinforcement learning principles [11 , 99–101] . One result that surprised us is the lack of prominent reward prediction error ( RPE ) signals in VS neurons . This finding differentiates VS from its dopamine inputs , which show clear and prominent RPE signals [82] . One possibility is that , unlike dopamine neurons , VS only carries RPE signals when they lead to adjustments , learning , or changes in strategy . ( We have previously argued that dACC neurons have this property . ) In this case , the lack of RPE signals may reflect the specific nature of the gamble task we used: no trial-to-trial learning was required , nor was any trial-to-trial adjustment observed . If so , it would suggest that VS neurons use RPE signals from dopamine neurons to construct a gated adjustment or learning signal that changes based on task context ( cf . [85] ) . Regardless , this finding suggests that the VS does not simply copy the RPE signals of its dopaminergic afferents , and that its responses ( at least in this task ) are more strongly accounted for by its cortical inputs . Like reward-based choices , perceptual decisions are traditionally linked to the cerebral cortex . However , recent work by Ding and colleagues clearly demonstrates the role of the striatum in perceptual decisions [102–104] . Indeed , it is striking how reliably the classic perceptual decision correlate can be found in striatal structures [102] . Our results here suggest that similar arguments may apply to the striatum’s role in economic choice , as well . Indeed , we conjecture that many classic prefrontal functions , including choice and executive control , may depend on corticostriatal circuits . Recent work , for example , has demonstrated that the striatum may be involved in executive functions such as planning and cognitive flexibility [105–107] . Taken alongside these findings , our work suggests that the function of the striatum in human decision-making may overlap more with that of the cortex than previously thought .
Animal procedures were approved by the University Committee on Animal Resources at the University of Rochester and conducted in observance of the Public Health Service’s Guide for the Care and Use of Animals . Two water-restricted male rhesus macaques ( Macaca mulatta ) were trained to perform oculomotor tasks for liquid reward . For each animal , a small prosthesis for maintaining head position was used , and a single Cilux recording chamber with a standard recording grid ( Crist Instruments ) was placed over the ventral striatum . All recorded neurons were analyzed and reported; no neurons were excluded from analysis . Position was verified by magnetic resonance imaging and a Brainsight system ( Rogue Research Inc . ) . Brainsight is a commercially available system ( Rogue Research , Montreal , QC ) designed to facilitate intracranial navigation in living animals . The general principle of the system is to combine presurgical placement of magnetically opaque fiducial markers with structural MRI scans , followed by generation of a computerized representation of the cranium and brain . When used by a trained technician ( our technician , Marc Mancarella , was formally trained by Brainsight ) , Brainsight allows placement of electrode tips with ~1 mm precision in the X and Y planes ( the Z-plane is affected by standard recording variables , but variability is reduced through careful calibration of the microdrive system ) . Animals received appropriate analgesics and antibiotics after all procedures . Throughout both behavioral and physiological recording sessions , the chamber was kept sterile with regular antibiotic washes and sealed with sterile caps . No animals were killed and histology was not conducted over the course of this study . We defined VS as the coronal planes situated between 28 . 02 and 20 . 66 mm rostral to interaural plane , the horizontal planes situated between 0 to 8 . 01 mm from ventral surface of striatum , and the sagittal planes between 0 to 8 . 69 mm from medial wall ( Figs 2A and S4 ) . Our recordings were made from a central region within this zone , which was selected on a voxel-by-voxel basis and in reference to the Paxinos macaque brain atlas [108] . We confirmed recording sites before each recording session using Brainsight with structural magnetic resonance images taken prior to the experiment . Neuroimaging was performed at the Rochester Center for Brain Imaging , with a Siemens 3T MAGNETOM Trio Tim using 0 . 5 mm voxels . Single electrodes ( Frederick Haer & Co . , impedance range 0 . 8 to 4M Ω ) were lowered using a microdrive ( NAN Instruments ) until neuronal waveforms were isolated on a Plexon system ( Plexon ) . Neurons were selected for study solely on the basis of the quality of isolation; we never pre-selected based on task-related response properties or excluded any neurons that surpassed our isolation criteria . Eye position was sampled at 1 , 000 Hz by an infrared eye-tracking system ( SR Research ) . The task was controlled by a computer running Matlab ( Mathworks ) with Psychtoolbox [109] and Eyelink Toolbox [110] . A computer monitor was placed 57 cm from the animal and centered on its eyes ( Fig 1A ) . A standard solenoid valve dispensed water rewards . Monkeys performed a two-option gambling task identical to the one we used in a previous investigation ( Fig 1A ) [15] . Two offers were presented on each trial . Each offer was represented by a rectangle 300 pixels tall and 80 pixels wide ( 11 . 35° of visual angle tall and 4 . 08° of visual angle wide ) . Options offered either a gamble or a safe ( 100% probability ) bet for liquid reward . Gamble offers were defined by both reward size and probability , which were randomized independent to one another for each trial . Each gamble rectangle had two sections , one red and the other either blue or green . The size of the blue or green portions indicated the probability of winning a medium ( 165 μL ) or large reward ( 240 μL ) , respectively ( Fig 1B ) . These probabilities were drawn from a uniform distribution between 0% and 100% . Safe offers were entirely gray , and selecting one would result in a small reward ( 125 μL ) 100% of the time . Offers were separated from the central fixation point by 550 pixels ( 27 . 53° of visual angle ) . The sides of the first and second offer ( left or right ) were randomized each trial . Each offer appeared for 400 ms followed by a 600 ms empty screen . After the offers were presented one at a time , a central fixation point appeared and the monkey fixated on it for 100 ms . Then both offers appeared simultaneously and the animal indicated its choice by shifting gaze to its preferred offer , maintaining fixation on it for 200 ms . Failure to maintain gaze for 200 ms would return the monkey to a choice state; thus monkeys were free to change their mind if they did so within 200 ms ( although they seldom did ) . Following a successful 200-ms fixation , the gamble was immediately resolved and a liquid reward was delivered . Trials that took more than 7 s were considered inattentive and were excluded from analysis ( this removed <1% of trials ) . Outcomes that yielded rewards were accompanied by a white circle in the center of the chosen offer ( see Fig 1A ) . Each trial was followed by an 800-ms inter-trial interval with a blank screen . Probabilities were drawn from uniform distributions with resolution only limited by the size of the screen’s pixels , which let us present hundreds of unique gambles . Offer reward sizes were selected at random and independent of one another with a 43 . 75% probability of blue ( medium reward ) gamble , a 43 . 75% probability of green ( large reward ) gambles , and 12 . 5% probability of safe offers . Note that this means two offers with the same reward size could be presented in the same trial . PSTHs were constructed by aligning spike rasters to the start of each trial and averaging firing rates across multiple trials . Firing rates were calculated in 20-ms bins , but generally were analyzed in 500 ms epochs . For display , PSTHs were smoothed with a 200-ms running boxcar . Some statistical tests of neuronal activity were only appropriate when applied to neurons one-at-a-time because of variations in response properties across the population . In such cases , a binomial test was used to determine if a significant portion of individual neurons reached significance on their own , which would allow conclusions about the neural population as a whole . These animals had previously performed other tasks where the same color hierarchy was maintained ( green > blue > gray ) , but with a different sets of precise amounts . Because of this , we reasoned that the animals would encode reward size ordinally in our task . To account for this , our analyses consistently make use of an ordinal coding of reward size , with gray , blue , and green offers offering 1 , 2 , and 3 water units , respectively . The confidence intervals in Figs 2B , 3C and 3D are fit to the data by estimating confidence intervals on regression parameters ( betas and intercepts ) using a least squares method . The area highlighted in red in each of these figures lies between lines calculated using betas and intercepts from the parameter CI upper and lower bounds . Fig 4D made use of a decoding analysis . We first separated trials by choice . We required the same number of trials both across neurons and across conditions ( offer 1 versus offer 2 ) . Therefore , for each analysis , we first found the lowest number of trials in either of the two conditions across all of the neurons , and used this as the number of trials we would give to our classifier . Although neurons were not recorded simultaneously , we treated them as if they were and grouped trials together across neurons as if they were a single trial . Thus , each of these pseudo-trials was paired with values from each neuron , giving us an n by m matrix ( where m is the minimum number of trials in each condition across neurons and n is the number of neurons ) . The only criterion for grouping trials together was that they fell in the same condition ( choose offer 1 or choose offer 2 ) , and thus the trials used differed in terms of other task variables ( reward size and probability ) . We took the mean firing rate of each neuron in each of these trials as input into to a Euclidean nearest-neighbor , leave-one-out classifier . This treats each trial as a point in n-dimensional space ( where n is the number of neurons , and the position in a given dimension was the mean firing rate of one neuron ) . To classify each trial , we took the mean position of the two groups ( choose offer 1 or choose offer 2 ) excluding the trial to be classified . We then took the Euclidean distance between the current trial and the mean position of the two groups—whichever distance was smaller was the group the trial was classified as . We performed one analysis to investigate how variance in firing related to variance in choice preference . We started by determining the best-fit curve for firing rate in epoch 1 as a function of the expected value of the first offer . In separate analyses , we fit to a line and to the best-fit second-order polynomial . We then classified each trial based on whether the observed firing rate in epoch 1 was greater or lower than a value predicted by the best-fit function . Finally , we correlated choice with whether firing rate was higher or lower than expected for each trial . We tested for a significant relation within each individual neuron using Pearson’s correlation test of these two sets of variables trial-by-trial . We then repeated this analysis for epoch 2 . | The neural calculations underlying reward-based choice are closely associated with a network of brain areas including the ventral striatum ( VS ) and ventromedial prefrontal cortex ( vmPFC ) . Most theories ascribe distinct roles to these two structures during choice , but these differences have yet to be confirmed at the level of single neurons . We compared responses of VS neurons to those of vmPFC neurons recorded in rhesus macaques choosing between potential gambles for water rewards . We found widespread similarities in the way that VS and vmPFC neurons fire during the choice process . Neurons in both areas encoded the value of the offered gamble , the difference in value between offered gambles , and the gamble outcome . Additionally , both areas showed stronger coding for the chosen gamble than for the unchosen one and predicted choice even when we controlled for offer value . Interestingly , preferential encoding of the chosen option was stronger and started earlier in VS than in vmPFC . Nonetheless , similarities between vmPFC and VS suggest that cortex and its striatal targets make overlapping contributions to reward-based choice . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Signatures of Value Comparison in Ventral Striatum Neurons |
Pathogenic microbes rely on environmental cues to initiate key events during infection such as differentiation , motility , egress and invasion of cells or tissues . Earlier investigations showed that an acidic environment activates motility of the protozoan parasite T . gondii . Conversely , potassium ions , which are abundant in the intracellular milieu that bathes immotile replicating parasites , suppress motility . Since motility is required for efficient parasite cell invasion and egress we sought to better understand its regulation by environmental cues . We found that low pH stimulates motility by triggering Ca2+-dependent secretion of apical micronemes , and that this cue is sufficient to overcome suppression by potassium ions and drive parasite motility , cell invasion and egress . We also discovered that acidification promotes membrane binding and cytolytic activity of perforin-like protein 1 ( PLP1 ) , a pore-forming protein required for efficient egress . Agents that neutralize pH reduce the efficiency of PLP1-dependent perforation of host membranes and compromise egress . Finally , although low pH stimulation of microneme secretion promotes cell invasion , it also causes PLP1-dependent damage to host cells , suggesting a mechanism by which neutral extracellular pH subdues PLP1 activity to allow cell invasion without overt damage to the target cell . These findings implicate acidification as a signal to activate microneme secretion and confine cytolytic activity to egress without compromising the viability of the next cell infected .
Infectious microorganisms experience diverse microenvironments during infection of a host . Such pathogenic microbes utilize specific cues to assess the environment and trigger an appropriate response that aids in their survival . For example , malaria parasites ( Plasmodium spp ) sense xanthurenic acid and a drop in temperature upon infecting a mosquito to trigger male gamete exflagellation for sexual reproduction [1] . The related apicomplexan parasite Toxoplasma gondii is thought to utilize the rapid build up of self-made abscisic acid as an intrinsic cue to exit from infected cells [2] , an event termed egress . However , other studies have shown that T . gondii can respond to diverse environmental changes including the loss of host cell viability [3] , [4] or an increase in reduction potential [5] to trigger egress . Most apicomplexan parasites including , malaria parasites and T . gondii , replicate within a membrane bound compartment termed the parasitophorous vacuole ( PV ) . The PV microenvironment is presumed to be similar to that of the host cell cytoplasm based on the detection of a hypothetical pore that permits fluorescent dyes of <∼1 , 300 Da freely pass across the PV membrane [6]–[8] . However , the identity of this pore has not been reported for malaria parasites or T . gondii . Also , the malaria PV has much higher Ca2+ concentrations ( ∼40 µM ) than the cytosol of infected erythrocytes ( 100 nM ) [9] , indicating a restricted flow of Ca2+ . Another report suggested that a membrane potential across the T . gondii PV membrane is maintained by proton and potassium P-type ATPases [10] . The limited and , in some cases , apparently discrepant findings for the PV highlight the need to better understand this microenvironment and how it changes during key events in the life cycle . T . gondii persists as a chronic infection in an estimated one third of the global human population , causing opportunistic disease in a subset of those infected . It also produces disease in domestic livestock , wild mammals and birds . In humans , the parasite is especially virulent when acquired congenitally or in reactivated disease , which occurs when the host becomes immune-suppressed . Pathogenesis is driven by iterations of the tachyzoite lytic cycle , which includes host cell invasion , replication within the PV , host cell egress and migration to infect a neighboring cell . Parasite motility and host cell invasion require the coordinated action of parasite proteins secreted from apical secretory organelles called rhoptries and micronemes . Micronemes are Ca+2-regulated secretory organelles that are controlled by phosphorylation-based signaling pathways ( reviewed in [11] ) . Potassium and Ca+2 ion fluxes have been shown to influence parasite motility and egress ( reviewed in [12] ) . High K+ concentrations , mimicking the intracellular state , inhibit microneme secretion and motility , and a drop in external K+ triggers microneme secretion [3] . Although the precise mechanisms of K+ sensing by the parasite are still emerging , it is known that intracellular Ca2+ , phospholipase C and at least two Ca2+-dependent protein kinases ( CDPKs ) are involved [13] . Other studies have shown that Ca2+ release from intracellular stores regulates parasite motility by activating the glideosome and apical secretion of transmembrane micronemal adhesins , which engage the motor to transduce power into motion [14]–[16] . Parasite sensing of environmental K+ is thought to ensure that the motility system is in neutral during intracellular replication , but is available for engagement to rapidly exit from an infected cell . Malaria sporozoites also respond to K+ fluxes [17] , implying a conserved mechanism for environmental sensing and regulation of motility . Earlier investigators of tachyzoite motility additionally found that motility is pH-dependent [18] . Alkaline conditions inhibited motility and acidic buffers induced motility . Here we demonstrate that pH-dependent motility involves the activation of microneme secretion . We also implicate acidification as an enhancer of egress both by promoting microneme secretion and enhancing cytolysis by perforin-like protein 1 ( PLP1 ) , a pore forming protein required for efficient egress . Our findings suggest that pH-dependent microneme secretion and activation of PLP1 is another layer of regulating parasite behavior to promote parasite success in changing environments .
Early studies on tachyzoite motility showed that parasite gliding is inhibited by high concentrations of K+ and alkaline pH and promoted by acidic pH [18] . We revisited the effect of pH on gliding by purifying parasites in high K+ , alkaline buffer ( pH 8 . 4 ) , switching to the same buffer at neutral or low pH and observing motility over time . Low pH stimulated motility in >90% of observed parasites and motility was sustained in a majority of parasites for at least 15 min ( Figure 1A ) , confirming previous findings [18] . Switching from alkaline to neutral pH led to no significant change in motility . Since parasite motility is linked to microneme secretion , we tested the effect of pH on microneme secretion by incubating parasites in buffer of varying pH and probing the secreted fraction via immunoblots for the microneme adhesive protein MIC2 , a galactose-binding protein MIC4 and PLP1 . More microneme secretion was detected at low pH compared to neutral and alkaline pH ( Figure 1B–C ) . In contrast , secretion of dense granule proteins GRA1 and GRA4 was largely independent of pH . Low pH induction of microneme secretion was more effective than stimulation with 1% ethanol , which is commonly used to activate microneme discharge ( Figure 1C ) [19] . Low-pH induced microneme secretion was sensitive to the Ca2+ chelator , BAPTA-AM , indicating dependence on intracellular Ca2+ ( Figure 1D ) . We also noted pH-dependent differences in the proteolysis of MIC2 and MIC4 , which are processed by the micronemal serine protease SUB1 . Processing was inhibited by low pH , suggesting the protease SUB1 functions optimally at neutral pH ( Figure 1D , E ) . Acidic pH stimulated microneme secretion despite the presence of high K+ , whereas minimal secretion occurred at neutral-alkaline pH in a high K+ environment ( Figure 1E ) . These findings reveal that low pH can overcome the normally microneme suppressive effects of a high K+ environment , which the parasite experiences within an infected cell . The results further suggest that low pH activation of microneme secretion contributes to pH dependent motility . Having established that low pH overcomes K+ suppression of microneme secretion , we reasoned that exposure of intracellular parasites to low pH should activate microneme secretion and egress despite a high K+ environment . We tested this by permeabilizing infected cells with digitonin in high K+ buffer with varying pH . Minimal egress was observed under alkaline ( pH 8 . 4 ) or neutral ( pH 7 . 4 ) conditions ( Figure 2A ) , consistent with a previous report [3] . On the other hand , acidic pH induced parasite egress in a manner related to the degree of acidification . Together with the above data , these findings suggest that low pH is sufficient to stimulate microneme secretion and initiate motility and parasite egress . Next we tested the extent to which the pH of the PV changes during induced egress and replication by expressing a pH sensitive GFP variant , superecliptic pHluorin , in the parasitophorous vacuolar space of plp1ko parasites ( plp1kosepH ) . plp1ko parasites were used instead of WT parasites to avoid losing the probe from the vacuolar space upon secretion of PLP1 during egress . Superecliptic pHluorin is highly fluorescent at neutral pH but is quenched at low pH [20] . We verified the pH-dependence of fluorescence by measuring the signal in superecliptic pHluorin plp1ko infected cells lysed at low or neutral pH and in live infected cells . Fluorescence was completely quenched in cells lysed at low pH ( Figure 2B ) . A modest but significant drop in fluorescence was detected upon treating infected cells with Ca2+ ionophore ( Figure 2C ) . This decrease in fluorescence was reversed with NH4Cl , a weak base , which accumulates in acidic compartments , raising luminal pH . DCCD , a P-type ATPase inhibitor also partially reversed the drop in fluorescence ( Figure 2D ) . These findings suggest a moderate decrease in vacuolar pH occurs upon egress induction . To observe changes in PV pH during parasite replication and spontaneous egress , we measured the fluorescent signal of live superecliptic pHluorin plp1ko infected cells over the course of intracellular replication , . If the PV pH is neutral , NH4Cl treated cells are expected to have a similar amount of fluorescence as cells in buffer alone . Conversely , if vacuolar pH is acidic , NH4Cl treated cells should show a stronger signal than untreated cells due to the pH neutralizing effect of treatment . Super-ecliptic pHluorin signals increased identically during parasite replication until ∼30 h post-inoculation . At this time point , however , the curves began to diverge , with a substantial suppression of fluorescence that was reversed by NH4Cl treatment ( Figure 2D ) . Microscopic examination of the infected monolayers indicated that most of the parasites remained intracellular or at least within spherical structures representing failed egress events until after 44 h post-infection . These findings imply a population-scale decrease in vacuolar pH occurs late in the replication cycle , prior to or during spontaneous parasite egress . Next we reasoned if an acidic pH contributes to parasite egress , induced egress should be sensitive to pH neutralization . Consistent with this , we found that parasite egress is suppressed by NH4Cl treatment upon stimulation with Ca2+ ionophore or dithiothreitol ( DTT ) , an egress inducer that activates a PV nucleotide triphosphatase [5] ( Figure 3A–B ) . PV acidification could occur through passive accumulation of metabolic wastes , or active delivery of protons by a proton-pump . We tested the latter by determining the effect of H+-ATPase inhibitors on induced egress . We found that parasite egress was sensitive to the P-type ATPase inhibitor , DCCD ( Figure 3C–D ) , but not the V-type ATPase inhibitor bafilomycin or the H+/K+ exchange inhibitor omeprazole ( Figure S1 ) . Egress induced by the phosphodiesterase inhibitor Zaprinast , which triggers microneme secretion and egress via activation of protein kinase G [21] , was also sensitive to NH4Cl or DCCD treatment ( Figure 3D ) . The above pH-neutralizing agents did not significantly alter motility ( Figure S2A , B ) or microneme secretion ( Figure S2C ) induced by the Ca2+ ionophore in a neutral buffer , rendering it unlikely that that they affected egress by impairing microneme secretion or the parasite motor system . Together these findings suggest a role for acidic pH and a P-type ATPase in egress . Next we tested the extent that pH-neutralizing agents affect the activity of important egress effectors during induced egress . Previous work has shown the microneme protein PLP1 to be crucial for rapid parasite egress [22] . To determine whether the reduced parasite egress was due to an inhibition of PLP1 activity occurring with pH-neutralization , we tested the effect of these treatments on egress-associated membrane permeabilization . Since PLP1 is necessary for membrane damage during egress [22] , we used the membrane impermeable dye propidium iodide ( PI ) to assess membrane permeabilization upon treating infected cells with Ca2+ ionophore to induce microneme secretion . Parasites were immobilized with the F-actin inhibitor cytochalasin D ( CytD ) to prevent membrane damage due to actin-myosin dependent parasite motility . Whereas vehicle-treated cells maintained intact membranes , ionophore treatment lead to the permeabilization of the majority of WT-infected cells . This activity was PLP1-dependent since no significant permeabilization was observed in ionophore-treated plp1ko-infected cells ( Figure 4B , D ) . Ionophore-induced membrane permeabilization was also sensitive to both NH4Cl and DCCD , suggesting that pH-neutralization suppresses PLP1 activity ( Figure 4A , C ) . As indicated above , NH4Cl and DCCD did not inhibit parasite gliding motility or microneme secretion , ruling these out processes as possible off-targets of treatment ( Figure S2 ) . Since several members of the protein superfamily to which PLP1 belongs are regulated by pH , we used recombinant PLP1 to determine the extent that its activity is pH-dependence . Using hemolysis of erythrocytes as a measure of PLP1 activity , we observed increased PLP1 cytolytic activity beginning at pH 6 . 4 and peaking at pH 5 . 4 ( Figure 5A ) . Approximately 7 times more PLP1 activity was seen at pH 5 . 4 than pH 7 . 4 . The pH dependent profile of PLP1 lytic activity was similar to that of listeriolysin O ( LLO ) , which Listeria monocytogenes uses to escape from the acidifying primary vacuole after cell entry [23] . The PLP1 cytolytic profile was distinct from streptolysin O ( SLO ) , which displayed a similar amount of lytic activity across a broad range of pH values . LLO and SLO are cholesterol dependent cytolysins ( CDC ) , which are members of the CDC/membrane attack complex-perforin superfamily that includes PLP1 [24] . We next tested the effect of pH on PLP1 membrane binding using erythrocyte ghosts as a model membrane . We observed that PLP1 membrane binding activity mirrors the lytic activity with more binding at acidic pH than neutral pH ( Figure 5B ) . Previous work demonstrated that PLP1 N- and C-terminal domains both contain membrane-binding activity [25] . Subsequently , we tested for pH-dependent membrane binding of full-length , mature PLP1 and the N-and C-terminal domains by sucrose density gradient membrane flotation . At low pH the majority of mature PLP1 was membrane bound ( fractions 1–6 ) , whereas at neutral pH most of the PLP1 remained unbound ( fractions 7–12 ) . Membrane binding by the N-terminal domain was not significantly affected by pH , whereas the C-terminal domain showed increased binding at low pH ( Figure 5C ) . These findings are consistent with a previous report showing a dominant role for the C-terminal domain in membrane binding [25] . Together the findings suggest that PLP1 cytolytic activity is pH-dependent at the membrane-binding stage . We next investigated the effect of low pH on parasite invasion . Although our above findings predict that the low pH stimulation of microneme secretion should augment parasite invasion , it should also enhance PLP1 activity if invading parasite secretes it . Parasites were purified in high K+ buffer and allowed to settle on host cells prior to switching the buffer to either DMEM , or DMEM buffered to pH 5 . 4 or 7 . 4 and incubating at 37°C for 2 min , then washed and fixed . We found that low pH substantially increases the amount of attached parasites and invaded parasites ( Figure 6A ) , likely due to increased microneme secretion and motility . Interestingly , acidic conditions appeared to reduce the proportion of invaded parasites relative to attached parasites To test if PLP1- and pH-dependent membrane damage occurs during invasion we pre-loaded host cells with calcein-AM , and WT and plp1ko parasites were allowed to settle on host cells in high K+ Endo buffer to block microneme secretion and invasion . Endo buffer was subsequently removed and replaced with DMEM-like buffer at pH 5 . 4 or 7 . 4 . After 10 minute's incubation , supernatant was collected and calcein release from host cells was quantified by fluorometry . Results were normalized to detergent lysed cells set at 100% . Host cells incubated with WT parasites at neutral pH did not show a density dependent increase in calcein release , consistent with parasite invasion in the absence of overt damage to host cells ( Figure 6B ) . However , host cells incubated with WT parasites at low pH resulted in a significant parasite density-dependent increase in calcein release , a trend that was not seen for plp1ko parasites at neutral or acidic pH , indicating a requirement for PLP1 in pH-dependent cell damage . Together these findings suggest that low pH stimulates invasion but also promotes PLP1-dependent wounding of target host cells .
Microneme secretion and motility contribute to host cell invasion and egress , but are quiescent during intracellular replication . Earlier work showed that a high K+ environment mimicking the intracellular milieu suppresses parasite motility and egress [3] , [18] . A previous report also suggested that T . gondii tachyzoite motility is regulated by pH , with moderately acidic conditions strongly enhancing motility [18] . Here we show that an acidic environment promotes microneme secretion , suggesting a mechanism for pH augmentation of motility . We also found that parasite exposure to an acidic environment overcame high K+ inhibition of microneme secretion , indicating that acidification of the intracellular environment can trigger microneme secretion and egress even if the high K+ intracellular milieu is intact . Consistent with this notion , we found that low pH can drive parasite egress in a high K+ environment . Using the pH-sensitive probe superecliptic pHluorin , we observed that egress induction with A23187 leads to a modest decrease in signal , reflecting a shift towards acidic pH . It remains unclear if A23187-induced PV acidification is due to effects of Ca2+ flux on proton pumps , due to signaling or secretion events downstream of Ca2+ flux , or both . Regardless , using the same probe we also detected a reduction in PV pH late in the replication cycle , possibly near the time of egress . Although these experiments were done with plp1ko parasites to minimize loss of the probe from PLP1 membrane damage , experiments showing inhibitory effects of pH neutralization on egress were done with WT parasites , indicating findings are not strain specific . Earlier investigations determined that upon host cell entry , the PV avoids fusion with endosomes , preventing acidification and degradation of the invaded parasite [26] . Previous efforts to examine vacuolar homeostasis also found a free-flow of small molecules ( <1 , 300 Da ) between the PV and host cytosol [7] , implying an equivalent neutral pH in both sites . However , these studies were carried out at 24 h or less post-inoculation . Our results concur with the vacuolar pH being neutral at this time point during replication , but suggest that acidification occurs later during the replicative cycle , perhaps immediately prior to egress . Our attempts to directly measure the intravacuolar pH during spontaneous egress were compromised by low expression , poor signal-to-noise and toxicity of ratiometric pHluorin , and difficulties capturing rare egress events with high spatiotemporal resolution . Breaching these obstacles will require substantial improvements in genetically encoded , ratiometric pH biosensors and imaging technologies . Further support for a role of pH in egress came from showing that treating infected cells with pH-neutralizing agents reduced parasite egress regardless of the egress inducer . Treatment with the P-type ATPase inhibitor , DCCD , impaired induced egress , implying a role for active proton flux during egress . P-type ATPases , also called E1-E2 ATPases , are an evolutionarily conserved group of cation or lipid transporters . The parasite expresses two P-type ATPases ( TgPMA1 and TgPMA2 ) of the Type IIIA subfamily , which includes H+-ATPases of plants and fungi [27] . TgPMA1 ( TGME49_252640 ) localizes to the parasite plasma membrane and is substantially upregulated in bradyzoites during the chronic stage of infection [27] . Genetic ablation of TgPMA1 led to a reduction in tachyzoite to bradyzoite conversion in vitro [27] . TgPMA1 deficient parasites had no reported growth defect as tachyzoites; however , egress was not assessed . TgPMA2 ( TGME49_284598 ) is expressed in tachyzoites and bradyzoites [27] , though its function remains to be characterized . The orientation of TgPMA1 and TgPMA2 is such that they are expected to pump protons from the parasite cytoplasm into the PV , consistent with a possible role in acidification of the PV . It is also possible that a DCCD-sensitive host P-type H+-ATPase contributes to the putative acidification of the PV . At least one plasma membrane multipass protein is known to occupy the nascent PV membrane after parasite invasion [28] , raising the possibility that other such host proteins including ion pumps could reside in the PV membrane . The accumulation of excreted metabolic waste products in the PV or infected cell late during intracellular replication could also contribute to acidification . Also , since it remains unclear how abscisic acid triggers egress , the possibility that it contributes to PV acidification warrants consideration . Other pore forming proteins including LLO are regulated by pH . LLO pH-sensitivity is mediated by three acidic amino acids in the transmembrane helices , which are part of domain 3 [29] , [30] . These residues are thought to repel one another at neutral pH leading to protein denaturation and loss of activity . Our findings suggest that PLP1 uses a distinct mechanism , however , since low pH augments PLP1 membrane binding via the C-terminal domain , which is positionally and functionally equivalent to domain 4 of LLO . Although the N-terminal domain also has membrane-binding activity , its membrane binding was not significantly affected by pH . Our findings also indicate that PLP1 membrane binding is pH-dependent , but it remains possible that subsequent steps such as oligomerization and membrane insertion are also pH regulated . The pore forming activity of human perforin is pH-dependent at a step following membrane binding [31] . Thus , pore-formation may be regulated by different environmental conditions at multiple steps of pore-formation . Future structural comparisons of PLP1 with LLO and perforin might illuminate divergent features that modulate the activity of these proteins in the varied environments in which they function . The increased microneme secretion at low pH also leads to dramatically higher parasite attachment and invasion . However , this augmentation of parasite association with the host cell comes at the expense of increased cell wounding . It is difficult to distinguish the extent to which increased microneme secretion versus increased PLP1 activity contributes to the membrane damage observed . The effect is probably due to a combination of the two factors . Regardless , membrane permeabilization was entirely PLP1-dependent since it was not detected in the PLP1-deficient strain at either pH . Collectively , our findings suggest a working model in which acidification of the PV during late stage parasite replication or immediately preceding egress augments both microneme secretion and PLP1 activity . That pH 5 . 9–6 . 4 is sufficient to promote microneme secretion and PLP1 membrane binding implies that moderate acidification is adequate to enhance microneme- and PLP1-dependent egress . PV acidification is capable of overcoming K+-dependent suppression of microneme secretion and motility , thus it can act as a primary trigger in the high K+ intracellular environment of a viable cell . Initial pH dependent release of PLP1 is expected to cause membrane damage resulting in a loss of K+ from the infected cell , thus further promoting Ca2+signaling , motor activation and microneme secretion to accelerate membrane damage , motility and egress . It should be noted , however , that while motility independent disruption of the PVM during egress is strictly PLP1-dependent [22] , the contribution of PLP1 to disruption of the host plasma membrane remains unknown . The model further predicts that the low K+ , neutral pH environment experienced by extracellular parasites after egress augments microneme secretion but suppresses PLP1 activity . Thus the extracellular environment is conducive to microneme-based motility , parasite attachment and invasion , but it suppresses PLP1-dependent damage to the membrane of the target cell , thereby permitting correct formation of the PV during invasion . This model does not exclude other levels of regulation e . g . , differential accessibility of PLP1 receptors or functionally distinct subpopulations of micronemes [32] , which could occur in parallel to ensure maximal PLP1 activity during egress while minimizing membrane damage during invasion . It is also expected that the proposed pH regulatory mechanism functions in parallel with other sensory and signaling pathways to coordinate egress under different circumstances .
Student's t-tests were used to assess differences in quantitative experiments , which were performed at least three times , with technical replicates within each experiment in some cases . Qualitative experiments were performed at least twice and often three to four times . T . gondii tachyzoites were maintained in human foreskin fibroblasts ( HFF ) as previously described [33] . Gliding experiments were conducted using a Zeiss temperature/CO2/humidity modulation system on a Zeiss Axio inverted microscope equipped with an Axiocam MRM CCD camera . For pH-dependent gliding , parasites were filter purified in high K+ buffer ( 145 mM KCl , 5 mM NaCl , 1 mM MgCl2 , 15 mM MES , 15 mM HEPES , pH 8 . 4 ) and allowed to settle in a glass-bottom petri dish . After parasites had settled , images were collected every 100 ms for 6 min . Then buffer was exchanged for the same buffer adjusted to pH 7 . 4 or 5 . 4 and images were collected every 100 ms for 30 min . For inhibitor gliding experiments , parasites were filter-purified in PBS , resuspended and allowed to settle in HBSSC ( Hanks buffered salt solution , 10 mM HEPES , 1 mM CaCl2 , 1 mM MgCl2 ) . After initial images were collected , buffer was exchanged for 40 mM NH4Cl or 40 µM DCCD in HBSSC and parasite motility was observed as above . Maximum projection images and videos were examined for motile parasites and the percent of motile parasites graphed over time . For inhibitor treatment assays , values were normalized to the percent motile parasites at time zero and the fold change in motility over time was graphed . Microneme secretion induced with A23187 or ethanol as described previously [14] in the presence of vehicle , 10 , 40 mM NH4Cl , or 10 , 40 µM DCCD [10] . Low-pH induced secretion was tested by purifying parasites in Endo buffer ( 44 . 7 mM K2SO4 , 106 mM sucrose , 10 mM MgSO4 , 20 mM Tris-H2SO4 ( pH 8 . 2 ) , 5 mM glucose , 3 . 5 mg/ml BSA ) [18] , and re-suspending in invasion buffer ( 110 mM NaCl , 0 . 9 mM NaH2PO4 , 44 mM NaHCO3 , 5 . 4 mM KCl , 0 . 8 mM MgSO4 , 1 . 8 mM CaCl2 ) or Endo buffer of the indicated pH at 37°C for 2 min . Calcium-dependence of secretion was tested by pre-incubation with BAPTA-AM as described previously with the following modifications: freshly egressed RH parasites were purified in Endo buffer and pre-treated with BAPTA-AM in Endo buffer at 37°C , pelleted , and then resuspended in 37°C Endo buffer of pH 5 . 4 or 7 . 4 with or without BAPTA-AM , incubated for 2 min at 37°C and then placed on ice [14] . To test the effect of potassium inhibition of microneme secretion , freshly egressed RH parasites were purified in Endo buffer at room temperature , pelleted , and resuspended in 37°C Endo buffer with either 45 mM KCl/5 mM NaCl or 5 mM KCl/45 mM NaCl , incubated at 37°C for 2 min , and then placed on ice . For all secretion assays , after incubating on ice , parasites were pelleted at 4°C , the secreted fraction ( supernatant ) was removed and spun again and the parasite pellet was washed in cold PBS prior to resuspending in boiling SDS-PAGE loading buffer . Secreted fractions and parasite pellets were examined by immunoblotting with the indicated antibodies . Low-pH induced egress was tested by inoculating HFF in an 8-well chamber slide with 3 µl of freshly egressed RH parasites/well and incubating for 30 h . The slide was then washed twice with warm high K+ buffer ( pH 8 . 4 ) and the buffer was replaced with buffer of the same composition and varying pH with and without 15 µM digitonin . The slide was incubated at 37°C for 3 min , fixed with 8% formaldehyde and occupied vacuoles were enumerated as previously described [25] . Superecliptic and ratiometric pHluorin vectors were kindly provided by Dr . Gero Miesenbock by material transfer agreement ( University of Michigan , SSP no . 13477; Memorial Sloan-Kettering Institute , SK# 19367 ) [20] . The genes were subcloned into the DsRed vacuolar expression vector [22] . Parasites were transfected with plasmid , transformed parasites were selected for with chloramphenicol and cloned by limiting dilution . Superecliptic pHluorin was highly expressed by parasites in the PV . pH-sensitivity was tested by inoculating HFF with varying concentrations of super-ecliptic pHluorin expressing plp1ko parasites ( plp1kosepH ) in a 96-well plate , and incubating for 30 h at 37°C . plp1ko parasites were used to retain the fluorescent signal in the parasitophorous vacuole upon egress induction . Wells were washed twice with warm PBS and PBS without Triton-X-100 was used to measure fluorescence in live , infected cells . PBS ( pH 5 . 4 or 7 . 4 ) with 0 . 1% Triton-X-100 was used to measure fluorescence in lysed cells . A23187-induced changes in superecliptic pHluorin signal in live , infected cells were tested by inoculating HFF ( grown in phenol red-free DMEM ) in a 96-well plate with plp1kosepH parasites ( 2 . 5×106 parasites/ml , 100 µl/well ) , incubating for 30 h , washing the wells twice with warm PBS , and adding HBSSC with 2 µM A23187 or DMSO , without or with 20 mM NH4Cl or 40 µM DCCD . The time between the addition of the compounds and fluorescence measurement was approximately 5 min . The majority of plp1ko parasites are unable to egress in this time period . Fluorescence over time in live , infected cells was observed by inoculating HFF with plp1kosepH ( 1×105 parasites/ml , 100 µl/well ) in a 96-well plate . At the indicated time points , 2 sets of triplicate wells were washed twice with warm PBS . One hundred µl warm HBSSC with or without 20 mM NH4Cl was added and fluorescence was read in a pre-warmed plate reader . Following the fluorescence reading , the plates were reincubated and a new set of wells was used for each time point . At each time point , wells were briefly examined microscopically prior to the fluorescent reading to check if the majority of parasites were intracellular . The experiment was terminated beyond 44 h due to the progression of endogenous egress . Fluorescence was measured at excitation 485/20 nm and emission 530/25 nm in a BioTek Synergy HT microplate reader at 37°C . Background from uninfected cells was subtracted from the total fluorescence for each condition . Data points represent the average and standard deviation of 2 or 3 independent experiments consisting of triplicate wells in each experiment . All assays were conducted in clear plastic 96-well plates since optimization experiments indicted they performed equally well as black-sided well plates . We also attempted to measure the pH of the PV using ratiometric pHluorin [20] . Transfection of the original ratiometric pHluorin sequence in the DsRed vacuolar expression vector [22] into T . gondii tachyzoites followed by drug selection and isolation of clones revealed that ratiometric pHluorin was transcribed but not translated as demonstrated by production of mRNA by RT-PCR and lack of detection by fluorescence microscopy , immunoblot or pulse chase 35S-methionine/cysteine metabolic labeling and immunoprecipitation . Ratiometric pHluorin was subsequently codon-optimized , chemically synthesized ( GenScript Inc ) and subcloned into the DsRed vacuolar expression vector as above . Codon-optimized ratiometric pHluorin was expressed by the parasites , and fluorescent parasites recovered upon drug selection and cloning by limited dilution . Imaging was performed with Zeiss filter sets 21 HE ( excitation 340/30 nm +387/15 nm , emission 510/90 nm ) and 38 HE ( excitation 470/40 nm , emission 525/50 nm ) . Fluorescent parasites were lost upon prolonged passage despite continuous drug selection , thus limiting the experiments to transiently transfected parasites . However , due to low expression and fluorescence , the exposure times required for signal detection were longer than those needed to measure rapid changes in pH during induced egress . Additionally , PVs of mock transfected parasites were noted to have varying degrees of autofluorescence in the 390 nm channel , which is the pH-sensitive wavelength , giving low confidence in the ability of ratiometric pHluorin to accurately reflect changes in PV pH especially as the total amount of signal was low late in the endogenous replication cycle . Egress assays with A23187 were conducted as previously described [33] with the following modifications: HFF in an 8-well chamber slide were inoculated with 3 µl of freshly egressed wild type ( RH ) parasites and incubated for 30 h at 37°C . Wells were washed twice with warm PBS and a 2 minute pretreatment at 37°C of the inhibitor ( NH4Cl , bafilomycin , dicyclohexylcarbodiimide ( DCCD ) , omeprazole ) in HBSSC was applied prior to addition of 4 µM A23187 ( 2 µM final concentration ) or DMSO with or without the indicated treatment in HBSSC for 2 minutes and fixation in 8% formaldehyde . Inhibitors were purchased from Sigma and tested at the indicated concentrations . Zaprinast-induced egress was tested at 250 µM ( final concentration ) in the same manner as A23187-induced egress . Immunofluorescence was performed and enumerated for SAG1 and GRA7 and occupied/unoccupied vacuoles as previously described [33] . Egress-associated membrane permeabilization was tested by inoculating HFF cells in an 8-well chamber slide with 2 µl of freshly egressed RH or plp1ko parasites and incubating for 30 h at 37°C . Following 2 washes with warm PBS , wells were treated with 100 µl HBSSC+1 µM CytD with or without NH4Cl or DCCD for 3 min at 37°C . Then 100 µl/well was added of 4 µM A23187/DMSO , 1 µM CytD , 12 . 5 µg/ml propidium iodide ( PI ) with the indicated final concentrations of NH4Cl and DCCD and incubated for 3 min . Following the incubation , cells were washed twice with warm PBS , fixed with 4% formaldehyde and stained with DAPI . Membrane permeabilization was quantified by the number of infected cells with PI-positive nuclei . Recombinant PLP1 and LLO were generated as previously described [33] . Streptolysin O ( SLO ) was handled according to the manufacturer's instructions ( Murex Diagnostics ) . pH-dependent hemolysis was assessed by washing erythrocytes in PBS ( pH 7 . 4 ) , pelleting the RBC , and re-suspending in PBS of indicated pH ( prepared by mixing sodium mono- and diphosphate in different amounts and adjusting pH with HCl or NaOH ) with 100 nM recombinant protein . RBC and recombinant protein were incubated at 37°C for 1 h , pelleted , and hemolysis was measured by absorbance at 540 nm of the supernatant . pH-dependent binding was tested by incubating erythrocyte ghosts , prepared as previously reported , with recombinant protein in PBS of the indicated pH [33] . RBC ghosts and recombinant protein was incubated at 37°C for 30 min; cells were pelleted and washed three times with cold PBS at neutral pH and bound samples were analyzed by SDS-PAGE and immunoblot . Parasite invasion was tested as previously described [34] with the following modifications: parasites were purified in Endo buffer and allowed to settle on HFF in an 8-well chamber slide at room temperature for 20 min . Then Endo buffer was replaced with either DMEM or DMEM-like buffer ( invasion buffer ) ( 110 mM NaCl , 0 . 9 mM NaH2PO4 , 44 mM NaHCO3 , 5 . 4 mM KCl , 0 . 8 mM MgSO4 , 1 . 8 mM CaCl2 ) at pH 5 . 4 or 7 . 4 and the slide was incubated at 37°C for 2 min . The buffer or media was removed , wells were washed twice with room temperature PBS and the slide was fixed with 0 . 4% formaldehyde in PBS . Immunofluorescence staining and parasite quantification was conducted as previously described for attached and invaded parasites . Cell wounding was tested by pre-loading host cells with 1 µM calcein-AM in phenol-red free DMEM and incubating for 1 h at 37°C , followed by two washes with warm PBS . Parasites were filter-purified in Endo buffer and applied to host cells in a 96-well plate by centrifuging at 500 g for 5 min . Supernatant was removed and replaced with 100 µl/well DMEM-like buffer , pH 7 . 4 or 5 . 4 . Plates were incubated for 10 min at 37°C and centrifuged as above . Fifty µl of supernatant was transferred to another plate and calcein fluorescence was read in a 96-well plate reader . | Toxoplasma and related parasites including those that cause malaria are obligate intracellular pathogens that replicate within a specialized compartment termed the parasitophorous vacuole . To infect new host cells these parasites must first escape from the parasitophorous vacuole and other limiting membranes of the currently infected cell . Escape , or egress as it is often called , depends on the timely release of adhesive proteins and lysis factors from secretory organelles called micronemes . Although this secretory event is crucial for egress , the natural environmental cues that trigger microneme secretion remain poorly defined . Here we discover that acidification of the parasitophorous vacuole is sufficient to trigger microneme secretion and promote the activity of a lysis factor called PLP1 . We also show that pH-neutralizing drugs inhibit egress and provide evidence of parasitophorous vacuole acidification approximately coinciding with parasite egress from infected host cells . The findings support a working model in which acidification activates microneme dependent motility and lytic activity to execute egress and destruction of infected cells . The results also provide insight into how PLP1 lytic activity is stimulated during egress in an acidic environment and subsequently suppressed by the neutral extracellular environment , thus permitting cell invasion with minimal damage to the next target cell . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"biology",
"and",
"life",
"sciences",
"protozoan",
"infections",
"parasitic",
"diseases",
"parasitology",
"toxoplasmosis"
] | 2014 | Acidification Activates Toxoplasma gondii Motility and Egress by Enhancing Protein Secretion and Cytolytic Activity |
Identifying latent structure in high-dimensional genomic data is essential for exploring biological processes . Here , we consider recovering gene co-expression networks from gene expression data , where each network encodes relationships between genes that are co-regulated by shared biological mechanisms . To do this , we develop a Bayesian statistical model for biclustering to infer subsets of co-regulated genes that covary in all of the samples or in only a subset of the samples . Our biclustering method , BicMix , allows overcomplete representations of the data , computational tractability , and joint modeling of unknown confounders and biological signals . Compared with related biclustering methods , BicMix recovers latent structure with higher precision across diverse simulation scenarios as compared to state-of-the-art biclustering methods . Further , we develop a principled method to recover context specific gene co-expression networks from the estimated sparse biclustering matrices . We apply BicMix to breast cancer gene expression data and to gene expression data from a cardiovascular study cohort , and we recover gene co-expression networks that are differential across ER+ and ER- samples and across male and female samples . We apply BicMix to the Genotype-Tissue Expression ( GTEx ) pilot data , and we find tissue specific gene networks . We validate these findings by using our tissue specific networks to identify trans-eQTLs specific to one of four primary tissues .
Cellular mechanisms tightly regulate gene transcription . Gene transcription is not independently regulated across genes: many of the mechanisms regulating transcription affect multiple genes simultaneously . Functional gene modules consist of subsets of genes that share similar expression patterns and perform coordinated cellular functions [1 , 2] . This cluster-based description of gene expression fails to capture the informative co-expression patterns among genes within a gene module . If we consider each gene as a vertex in a network , then pairs of genes within a gene module for which the correlation in expression levels cannot be explained by other genes may be connected by an undirected edge . Across all genes , these pairwise relationships constitute gene co-expression networks . Constructing these undirected gene networks , as compared to clustering genes into gene modules [3–6] , provides rich detail about pairwise gene relationships . An even richer structure capturing these pairwise relationships would be a directed network of genes , but currently directed networks are computationally intractable to construct relative to undirected gene networks [7–10] . Our work describes a rigorous approach to recover undirected gene co-expression networks from gene expression data that uses a probabilistic latent factor model to quantify the relationships between all pairs of genes . Several methods have been proposed to construct gene co-expression networks by partitioning a set of genes ( and , in some cases , samples ) into gene modules from which an undirected graph is elicited [11–14] . In most cases , gene partitioning creates disjoint sets of genes , implying that genes only participate in a single gene module . Biologically this assumption does not hold; the impact is that the gene networks built from methods that assume disjoint clusters are modular . These approaches are not probabilistic , and thus uncertainty in the network edges is not well characterized . Alternatively , statistical latent factor models are often used to identify groups of co-regulated genes in gene expression data [15–18] . In particular , latent factor models decompose a matrix Y ∈ ℜp×n of p genes and n samples into the product of two matrices , Λ ∈ ℜp×K , the factor loadings , and X ∈ ℜK×n , the latent factor matrix with K latent factors , assuming independent Gaussian noise . Because it is costly to obtain and assay genome-wide gene expression levels in a single sample , most expression studies include observations of many more genes p than samples n [19 , 20] . This so-called p ≫ n scenario limits our ability to find latent structure in this expansive , underconstrained space . High dimensional data suggests the use of strong regularization on the latent space to provide sufficient structure for the optimization to reach a robust solution . For example , we may regularize a latent space to discourage all but a few genes from contributing to a latent factor through a sparsity-inducing prior or penalty on the loading vectors [15 , 21 , 22] . Non-disjoint clusters of genes can be extracted from the resulting fitted sparse loading matrix by recovering all genes with non-zero loadings on the same factor [16] . Sparse latent factor models are more interpretable than their non-sparse counterparts because of this clustering effect . Besides encouraging sparsity in the factor loading matrix , which results in non-disjoint clusters of genes that co-vary across all samples , one can also induce sparsity in the factor matrix , which results in non-disjoint subsets of samples within which subsets of genes uniquely exhibit co-variation . Statistically , this corresponds to regularizing both factor and loading matrices using priors that encourage zero-valued elements . Biologically , such a model recovers components that identify small numbers of correlated genes , where the correlation among the genes is exclusive to , for example , female samples . This statistical model encodes a general framework known as biclustering [23–35] . A biclustering model decomposes a matrix into clusters that each correspond to a subset of samples and a subset of features that exhibit latent structure unique to those subsets . Gene expression levels have been shown to be sensitive to a number of environmental , biological , and technical covariates including experimental batch , sex , ethnicity , smoking status , or sample tissue heterogeneity [36 , 37] . Methods to adjust the observation matrix to control the effects of these covariates without eliminating signals of interest have been proposed . Most attempts have been limited to correcting for confounding effects in a two-stage approach [20 , 38 , 39] or controlling for confounding effects jointly with association testing [40 , 41] . The two-stage approach applied to estimates of co-expression networks has not been successful: often variation in expression levels of groups of co-expressed genes are captured in the estimates of confounding effects and controlled in the first stage , leading to false negatives [42] . In this paper , we develop a Bayesian statistical model for biclustering called BicMix . Our motivation behind developing this method was to identify large numbers of subsets of co-regulated genes capturing as many sources of gene transcription variation as possible within arbitrary subsets of the samples . Our biclustering model also includes non-sparse components to represent sources of transcription variation that affect all genes or all samples , which includes many types of confounding effects . We developed a simple but principled statistical method to reconstruct gene co-expression networks based on the regularized covariance matrices estimated using our biclustering model . This method recovers different types of gene co-expression networks , categorized by quantifying the contribution of each sample to the latent components: i ) ubiquitous co-expression networks , ii ) co-expression networks specific to a sample context , and iii ) networks with differentially co-expressed genes across a sample context . In this paper , we motivate and describe our Bayesian model for biclustering , BicMix . We validate our biclustering model on extensive simulations and compare biclustering with a number of state-of-the-art methods . We then apply our model to gene expression data without correcting for known or unknown confounders . In particular , we apply our biclustering model to gene expression levels measured in heterogeneous breast cancer tissue samples to recover a co-expression network that is differentially expressed across estrogen receptor positive and negative ( ER+ and ER- ) samples [43 , 44] . Next , we apply our biclustering model to gene expression levels measured in lymphoblastoid cell lines ( LCLs ) from a cohort of patients in a cardiovascular disease study to identify co-expression networks with differential co-expression across males and females , and across smoking status [45] . Finally , we apply BicMix to the Genotype-Tissue Expression ( GTEx ) pilot data to elucidate tissue specific gene co-expression networks [46] . We validate the recovered networks by identifying tissue specific trans-eQTLs using the recovered tissue specific co-expression networks .
Biclustering was first introduced to detect clusters of states and years that showed similar voting patterns among Republicans in national elections [47] and was later referred to as biclustering in the context of identifying co-expressed genes in subsets of samples [23] . Biclustering has also been referred to as two mode clustering [48] , subspace clustering [49 , 50] , or co-clustering [51] in various applied contexts . Biclustering was used successfully to explore latent sparse structure in different applied domains [52] , including gene expression data [23 , 24 , 53–56] , neuroscience [57] , time series data [54] , and recommendation systems [58] . Refer to comprehensive biclustering reviews for details [27 , 59] . Biclustering approaches fall into four general categories . The first category assumes that each observed gene expression level for one sample is a linear combination of a mean effect , a row ( gene ) effect , and a column ( sample ) effect , some of which may be zero [23] . One approach in this category , Plaid , captures gene expression levels as a sum of many sparse submatrix components , where each submatrix includes non-zero values only for a subset of genes and subset of samples [30 , 31] . The second category of biclustering methods uses hierarchical clustering to group together similar samples and features [3] . For example , samples may be clustered by considering some measure of feature similarity [24 , 25 , 28 , 32 , 35] . The third category of biclustering methods builds up biclusters by iteratively grouping features in a greedy way—e . g . , identifying all genes that have correlated expression levels with a selected gene—and then removing samples that do not support that grouping [26] . The last category of biclustering methods uses Bayesian sparse factor analysis models [33] . These models decompose a gene expression matrix into two sparse matrices . Sparsity-inducing priors , such as the Laplace prior , are imposed on elements of both the loading and the factor matrices to induce zero-valued elements . K biclusters are recovered as the non-zero feature and sample components for each of the K latent components . Our approach falls into this last category of a sparse latent factor model for biclustering . We validated our biclustering model using simulated data sets and compared results from five state-of-the-art biclustering methods . We simulated data from an alternative generative model for observation matrix Y = Λ X + ϵ , where Y has dimension p = 500 by n = 300 and ϵ i , j ∼ N ( 0 , ν − 1 ) . Within this model , we simulated sparsity as follows: for each loading and factor , a number m ∈ [5 , 20] of elements were randomly selected and assigned values drawn from N ( 0 , 2 ) ; the remaining elements were set to zero . We allowed components to share as many as five elements . Simulation 1 ( Sim1 ) had ten sparse components . Simulation 2 ( Sim2 ) had ten sparse components and five dense components , for which the loadings and factors were drawn from a N ( 0 , 2 ) distribution . The components were shuffled so that a sparse loading may correspond to a dense factor , and vice versa . For both simulations we considered low and high noise scenarios: the residual variance parameter in the low noise ( LN ) setting was ν−1 = 1 and , in the high noise ( HN ) setting , was ν−1 = 2 . Ten matrices Y were generated from each simulation scenario . We ran BicMix and five other biclustering methods–Fabia [33] , Plaid [30] , CC [23] , Bimax [27] , and Spectral biclustering [34] . For all simulations , we ran BicMix by setting a = b = 0 . 5 , c = 1 , d = 0 . 5 , e = f = 1 and ν = ξ = 1 to promote substantial sparsity at the local level ( i . e . , the horseshoe prior [66] ) , weaker sparsity at the factor specific level ( i . e . , the Strawderman-Berger prior [67 , 68] ) , and a uniform prior at the global level of the hierarchy [60] . The algorithm was initialized with warm start values by running MCMC for 500 iterations and using the final sample as the initial state for variational EM . For BicMix results , components that were classified as sparse have each element thresholded at 10−10 , because our parameter estimation methods converged to values near , but not exactly , zero . All other methods were run using their recommended settings ( see Methods ) . For Sim2 , we corrected the simulated data for the dense components by controlling for five principal components ( PCs ) before all other methods were run; without this initial correction for dense components , results from all five other biclustering methods were uninterpretable . For all runs , BicMix was initialized with K = 50 latent factors; all other methods were initialized with the correct number of sparse factors K = 10 . For Fabia , we ran the software in two different ways . The results from running Fabia with the recommended settings are denoted as Fabia . We also set the sparsity threshold in Fabia to the number ( from 100 quantiles of the uniform distribution over [0 . 1 , 5] ) that produced the closest match in the recovered matrices to the number of non-zero elements in the simulated data; we label these results Fabia-truth . We used the recovery and relevance score ( R&R score ) [27] to measure the false discovery rate ( FDR ) and sensitivity of each method in recovering true biclusterings . Let the true set of sparse matrices be M1 and the estimated set of sparse matrices be M2; then the R&R score is calculated as: Rec = 1 | M 1 | ∑ b 1 ∈ M 1 max b 2 ∈ M 2 b 1 ∩ b 2 b 1 ∪ b 2 , ( 2 ) Rel = 1 | M 2 | ∑ b 2 ∈ M 2 max b 1 ∈ M 1 b 1 ∩ b 2 b 1 ∪ b 2 . ( 3 ) Recovery quantifies the proportion of true clusters that are recovered ( i . e . , recall ) ; relevance refers to the proportion of true clusters identified in the recovered clusters ( i . e . , precision ) . For BicMix , we applied this R&R score to the components for which both the loading and the factor vectors were sparse , which indicates a bicluster . For the doubly-sparse latent factor models , Fabia and BicMix , we also calculated a sparse stability index ( SSI ) [60] to compare the recovered and true matrices; SSI is invariant to label switching and scale , and falls between [0 , 1] with 1 indicating perfect recovery . For Sim1 , we found that BicMix recovered the sparse loadings , sparse factors , and the biclusters well in the low noise scenario based on both R&R ( Fig 2a ) and SSI ( Fig 2b ) . Fabia-truth had the second best performance based on R&R . For comparison , Fabia-truth achieved better R&R scores than Fabia ( Fig 2a ) ; the clustering results from BicMix dominated those from Fabia-truth , although there was only a small gain in relevance in the low noise Sim1 results for BicMix . Plaid showed high relevance for the recovered biclusters regardless of the noise level for Sim1 , but at the expense of poor recovery scores . The remaining methods did not perform well in these simulations with respect to the R&R score for both low and high noise simulation scenarios . For Sim2 , BicMix correctly identified the sparse and dense components ( Fig 2a ) , where a threshold of 〈zk〉 > 0 . 9 was used to determine when a loading k was dense . The performance of Fabia on Sim2 deteriorated substantially relative to its performance on Sim1 , although the confounders were removed using principal components ( PCs ) and the correct number of factors was given . For both BicMix and Fabia , additional noise in the simulation made bicluster recovery more difficult , as shown in deterioration of the recovery score for both methods; however , unlike Fabia , the relevance score of the biclustering from BicMix was unaffected by additional noise in the Sim2 high noise scenario . The other methods show inferior performance relative to BicMix and Fabia on Sim2 . CC assumes that genes in each bicluster have constant expression values , which limits its ability to cluster genes with heterogeneous expression levels . Bimax assumes binary gene expression values ( i . e . , over- or under-expressed ) , which limits its utility for heterogeneous expression levels . Spectral biclustering imposes orthogonal constraints on the biclusters; this orthogonality assumption is violated in these simulations and also in gene expression data , where correlated sources of variation may impact similar subsets of genes . We now turn to the application of BicMix to gene expression data from three studies . For each application , we first describe the biclustering results from BicMix . Then we show the interpretable networks that were recovered from the BicMix results and discuss network validation using cis- and trans-eQTLs .
In this work , we developed a statistical approach to biclustering based on a Bayesian sparse latent factor model . We included a two-component mixture distribution to allow both sparse and dense representations of the features or samples , which captures heterogeneous sources of structured variation within the gene expression data . We used the regularized covariance matrix estimated from the latent factor model to build a Gaussian graphical model with the features represented as nodes in the undirected network . By extracting covariance matrices corresponding to subsets of components , we were able to identify gene co-expression networks that were shared across all samples , unique to a subset of samples , or differential across sample subsets . We applied our methodology to breast tumor tissue gene expression samples and recovered co-expression networks that are differential across ER+ and ER- tumor types . We applied our methodology to gene expression data from the CAP project and recovered sex-differential , sex specific , and smoking status specific gene co-expression networks . We applied our methodology to the GTEx gene expression pilot data and recovered tissue specific networks for four tissues , which we validated by identifying tissue specific trans-eQTLs . Factor analysis methods , including the biclustering approach presented here but extending to many other well-studied models , are statistical tools developed for exploratory analyses of the data . In this work , we have exploited the latent structure in both the factor and the loading matrix to estimate the covariance matrix that is specific to sample subsets . We considered tissue type and tumor types , but these methods can be used for any observed binary , categorical , integer , or continuous covariate ( e . g . , case-control status , batch , sex , age , EBV load ) . We showed in the Results that the recovered latent structure has substantial context specific biological meaning . Our results on the GTEx data show that a number of genes were identified as part of multiple tissue specific networks . While individual genes may overlap across networks , the interactions of those genes did not . Genes that co-occurred in multiple tissue specific networks are good candidates to test for differential function across tissues . We also used this approach to study sexual dimorphism , extracting gene networks specific to one sex or differential across the sexes . We showed the potential of this approach to improve statistical power to identify sex specific trans-eQTLs . In this version of BicMix , extracting a covariance matrix specific to a subset of the samples was performed post hoc: the linear projection to the latent space was performed in a mostly unsupervised way , although our three layer sparsity inducing priors add additional structure above SFA-type approaches . As described in the Results , there were multiple categories of gene-interactions that we recovered . These categories included: gene interactions that existed across context , gene interactions that were unique to specific contexts , and gene interactions that were present across contexts but differentially interact in different contexts . However , this approach currently does not use known context to directly inform the projection . Indirectly , we saw that the sparsity structure on the samples allowed small subsets of the samples to inform projection , but this still relied on a post hoc interpretation of those sample subsets to recover specific network types . Correlation between contexts , such as tissue and batch , or smoking status and sex , would confound these results; here we checked for these correlation among observed covariates ( S2 and S6 Figs ) and also validated our results using literature searches and trans-eQTL recovery . Furthermore , it may be the case that , for a sample context of interest ( e . g . , age , sex ) , there is insufficient signal that is uniquely attributable to those samples ( e . g . , female ) to identify a covariance matrix corresponding to the values of interest in this unsupervised framework . We are currently extending this approach so that the linear projection is explicitly informed by the context of interest .
We consider the following factor analysis model: Y = Λ X + ϵ , ( 4 ) where Y ∈ ℜp×n is the matrix of observed variables , Λ ∈ ℜp×K is the loading matrix , X ∈ ℜK×n is the factor matrix , and ϵ ∈ ℜp×n is the residual error matrix for p genes and n samples . We assumed ϵ ⋅ , i ∼ N ( 0 , Ψ ) , where Ψ = diag ( ψ1 , … , ψp ) . In previous work [60 , 62] , a three parameter beta ( T P B ) [61] prior was used to model the variance of Λ . The three parameter distribution has the form f ( x : a , b , ϕ ) = Γ ( a + b ) Γ ( a ) Γ ( b ) ϕ b x b - 1 ( 1 - x ) a - 1 { 1 + ( ϕ - 1 ) x } - ( a + b ) , ( 5 ) for x ∈ ( 0 , 1 ) , a > 0 , b > 0 and ϕ > 0 . We used T P B to induce flexible shrinkage to both Λ and X . Specifically , we included three layers of shrinkage—global , factor specific , and local—for both the factors and the loadings . Next we describe the sparsity-inducing structure for Λ and X . Because of the large dimension of matrices to which we apply BicMix , we used approximate methods for parameter estimation . In particular , we used variational expectation maximization ( VEM ) to estimate values for latent variables and parameters directly from the data in an approximate way . Extending previous work [60] , the posterior probability P = p ( Λ , X , z , o , Θ | Y ) is written as: P ∝ p ( Y | Λ , X ) p ( Λ | z , Θ Λ ) p ( X | o , Θ X ) p ( z | Θ Λ ) p ( o | Θ X ) p ( Θ Λ ) p ( Θ X ) ∝ p ( Y | Λ , X ) P ( Λ ) P ( X ) , ( 30 ) where we used ΘΛ and ΘX to denote the set of parameters related to Λ and X , respectively . Then , P ( Λ ) = p ( Λ | z , Θ Λ ) p ( z | Θ Λ ) p ( Θ Λ ) = ∏ j = 1 p ∏ k = 1 K N ( Λ j , k | θ j , k ) G a ( θ j , k | a , δ j , k ) G a ( δ j , k | b , ϕ k ) 1 z k = 1 × ∏ j = 1 p ∏ k = 1 K N ( Λ j , k | ϕ k ) 1 z k = 0 ∏ k = 1 K B e r n ( z k | π ) B e t a ( π | α , β ) × ∏ k = 1 K G a ( ϕ k | c , τ k ) G a ( τ k | d , η ) G a ( η | e , γ ) G a ( γ | f , ν ) ( 31 ) and P ( X ) = p ( X | o , Θ X ) p ( o | Θ X ) p ( Θ X ) = ∏ k = 1 K ∏ i = 1 n N ( x k , i | σ k , i ) G a ( σ k , i | a X , ρ k , i ) G a ( ρ k , i | b X , ω k ) 1 o k = 1 × ∏ k = 1 K ∏ i = 1 n N ( x k , i | ω k ) 1 o k = 0 ∏ k = 1 K B e r n ( o k | π X ) B e t a ( π X | α X , β X ) × ∏ k = 1 K G a ( ω k | c X , κ k ) G a ( κ k | d X , χ ) G a ( χ | e X , φ ) G a ( φ | f X , ξ ) ( 32 ) To derive the variational EM algorithm , we expanded the posterior probability ( Eq ( 30 ) ) and wrote the expected complete log likelihood for parameters related to Λ: Q ( ΘΛ ) = 〈ℓc ( ΘΛ , Λ|z , X , Y ) 〉 as: Q ( Θ Λ ) ∝ ∑ j = 1 p ∑ i = 1 n log p ( y j , i | Λ , X , Θ Λ , z ) + ∑ j = 1 p ∑ k = 1 K p ( z k | Θ Λ ) log p ( Λ j , k | Θ Λ , z k ) + log p ( Θ Λ ) ∝ - p 2 ln | Ψ | - ∑ j = 1 p ∑ i = 1 n y j , i - ∑ k = 1 K Λ j , k x k , i 2 2 ψ j , j + ∑ j = 1 p ∑ k = 1 K 1 - z k - 1 2 ln ϕ k - Λ j , k 2 2 ϕ k + ∑ j = 1 p ∑ k = 1 K z k - 1 2 ln θ j , k - Λ j , k 2 2 θ j , k + a ln δ j , k + ( a - 1 ) ln θ j , k - δ j , k θ j , k + ∑ j = 1 p ∑ k = 1 K z k b ln ϕ k + ( b - 1 ) ln δ j , k - ϕ k δ j , k + ∑ k = 1 K z k ln π + ( 1 - z k ) ln ( 1 - π ) + ∑ k = 1 K c ln τ k + ( c - 1 ) ln ϕ k - τ k ϕ k + d ln η + ( d - 1 ) ln τ k - η τ k + e ln γ + ( e - 1 ) ln η - γ η + f ln ν + ( f - 1 ) ln γ - ν γ + α ln π + β ln ( 1 - π ) , ( 33 ) where we used 〈X〉 to represent the expected value of X . Similarly , the expected complete log likelihood for parameters related to X takes the following form: Q ( Θ X ) ∝ ∑ j = 1 p ∑ i = 1 n log p ( y j , i | Λ , X , Θ X , O ) + ∑ k = 1 K ∑ i = 1 n p ( o k | Θ X ) log p ( x k , i | Θ X , O k ) + log p ( Θ X ) ∝ - p 2 ln | Ψ | - ∑ j = 1 p ∑ i = 1 n y j , i - ∑ k = 1 K Λ j , k x k , i 2 2 ψ j , j + ∑ k = 1 K ∑ i = 1 n 1 - o k - 1 2 ln ω k - x k , i 2 2 ω k + ∑ k = 1 K ∑ i = 1 n o k - 1 2 ln σ k , i - x k , i 2 2 σ k , i + a X ln ρ k , i + ( a X - 1 ) ln σ k , i - ρ k , i σ k , i + ∑ k = 1 K ∑ i = 1 n o k b X ln ω k + ( b X - 1 ) ln ρ k , i - ω k ρ k , i + ∑ k = 1 K o k ln π X + ( 1 - o k ) ln ( 1 - π X ) + ∑ k = 1 K c X ln κ k + ( c X - 1 ) ln ω k - κ k ω k + d X ln χ + ( d X - 1 ) ln κ k - χ κ k + e X ln φ + ( e X - 1 ) ln χ - φ χ + f X ln ξ + ( f X - 1 ) ln φ - ξ φ + α X ln π X + β X ln ( 1 - π X ) . ( 34 ) To simplify the calculation , we assumed that the joint distribution p ( ok , xk , i ) factorizes as p ( ok ) p ( xk , i ) , implying that corresponding factors’ and loadings’ sparsity statuses are independent . We computed the MAP estimates for the parameters that encourage sparsity in the Λ matrix , Θ Λ ^ = a r g max Θ Λ Q ( Θ Λ ) . Specifically , we solved equation ∂ Q ( Θ Λ ) ∂ Θ Λ = 0 to find the closed form MAP estimates . The MAP estimate for the jth row of Λ , Λj , ⋅ , in matrix form , is: Λ ^ j , · = Y j , · ψ j , j - 1 X T X ψ j , j - 1 X T + Z Θ j - 1 + ( I - Z ) Φ - 1 - 1 , ( 35 ) where Θ j = θ j , 1 0 ⋯ 0 0 θ j , 2 ⋯ 0 ⋮ ⋮ ⋱ ⋮ 0 0 ⋯ θ j , K , Φ = ϕ 1 0 ⋯ 0 0 ϕ 2 ⋯ 0 ⋮ ⋮ ⋱ ⋮ 0 0 ⋯ ϕ K , ( 36 ) and Z = z 1 0 ⋯ 0 0 z 2 ⋯ 0 ⋮ ⋮ ⋱ ⋮ 0 0 ⋯ z K ( 37 ) and I is the identity matrix . We computed the expected value of X , 〈X〉 , which has the following form: X · , i = ( Λ T Ψ - 1 Λ + O Σ i - 1 + I - O Ω - 1 ) - 1 Λ T Ψ - 1 Y · , i , ( 38 ) where Σ i = σ 1 , i 0 ⋯ 0 0 σ 2 , i ⋯ 0 ⋮ ⋮ ⋱ ⋮ 0 0 ⋯ σ K , i , Ω = ω 1 0 ⋯ 0 0 ω 2 ⋯ 0 ⋮ ⋮ ⋱ ⋮ 0 0 ⋯ ω K , ( 39 ) and O = o 1 0 ⋯ 0 0 o 2 ⋯ 0 ⋮ ⋮ ⋱ ⋮ 0 0 ⋯ o K . ( 40 ) We computed the expected value of X ψ j , j − 1 X T: X ψ j , j - 1 X T = ψ j , j - 1 X X T + Σ X ( 41 ) where ΣX denotes the covariance matrix of X . Parameter θj , k has a generalized inverse Gaussian conditional probability [60 , 61] , and its MAP estimate is: θ ^ j , k = 2 a - 3 + ( 2 a - 3 ) 2 + 8 Λ j , k 2 δ j , k 4 δ j , k . ( 42 ) Similarly , the MAP estimate for σk , i has the following closed form: σ ^ k , i = 2 a X - 3 + ( 2 a X - 3 ) 2 + 8 x k , i 2 ρ k , i 4 ρ k , i . ( 43 ) The MAP estimate for δj , k is: δ ^ j , k = a + b - 1 θ j , k + ϕ k . ( 44 ) Correspondingly , ρ ^ k , i = a X + b X - 1 σ k , i + ω k . ( 45 ) Parameter ϕk generates both sparse and dense components , and its MAP estimate takes the form: ϕ ^ k = H + H 2 + M T M , ( 46 ) where H = p b 〈 z k 〉 + c - 1 - p 2 ( 1 - 〈 z k 〉 ) ( 47 ) M = 2 〈 z k 〉 ∑ j = 1 p δ j , k + τ k ( 48 ) T = ∑ j = 1 p Λ j , k 2 . ( 49 ) Correspondingly , ω ^ k = H X + H X 2 + M X T X M X , ( 50 ) where H X = n b X 〈 o k 〉 + c X - 1 - n 2 ( 1 - 〈 o k 〉 ) ( 51 ) M X = 2 〈 o k 〉 ∑ i = 1 n ρ k , i + κ k ( 52 ) T X = ∑ i = 1 n x k , i 2 . ( 53 ) The following parameters have similar updates to δj , k , with simple closed forms because of the conjugacy of the distributions: τ ^ k = c + d - 1 ϕ k + η ( 54 ) η ^ = K d + e - 1 γ + ∑ k τ k ( 55 ) γ ^ = e + f - 1 η + ν ( 56 ) The corresponding parameters related to X have similar forms: κ ^ k = c X + d X - 1 ω k + χ ( 57 ) χ ^ = K d X + e X - 1 φ + ∑ k κ k ( 58 ) φ ^ = e X + f X - 1 χ + ξ . ( 59 ) The expected value of zk|Θ is computed as: 〈 z k | Θ Λ 〉 = p ( z k = 1 | Θ Λ ) = π ∏ j = 1 p N ( Λ j , k | θ j , k ) G a ( θ j , k | a , δ j , k ) G a ( δ j , k | b , ϕ k ) ( 1 - π ) ( ∏ j = 1 p N ( Λ j , k | ϕ k ) ) + π ∏ j = 1 p N ( Λ j , k | θ j , k ) G a ( θ j , k | a , δ j , k ) G a ( δ j , k | b , ϕ k ) ( 60 ) The prior on the indicator variable for sparse and dense components , π , has a beta distribution , and its geometric mean is the following: 〈 ln π 〉 = ψ ∑ k = 1 K 〈 z k 〉 + α - ψ K + α + β ( 61 ) where ψ is the digamma function . The corresponding parameters related to X are 〈 o k | Θ X 〉 = p ( o k = 1 | Θ X ) = π ∏ i = 1 n N ( x k , i | σ k , i ) G a ( σ k , i | a X , ρ k , i ) G a ( ρ k , i | b X , ω k ) ( 1 - π ) ( ∏ i = 1 n N ( x k , i | ω k ) ) + π ∏ i = 1 n N ( x k , i | σ k , i ) G a ( σ k , i | a X , ρ k , i ) G a ( ρ k , i | b X , ω k ) . ( 62 ) 〈 ln π X 〉 = ψ ∑ k = 1 K o k + α X - ψ K + α X + β X ( 63 ) Assuming that the residual precision has a conjugate ( gamma ) prior , 1 ψ j , j ∼ G a ( 1 , 1 ) , then we have Ψ = diag YY T - 2 Y X T Λ T + Λ XX T Λ T + 2 I n + 2 . ( 64 ) In estimating the factors , we invert a K × K matrix for each sample , and , in estimating the loading matrix , we invert a K × K matrix for reach gene , so the main source of computational complexity in our algorithm is O ( ( n + p ) K 3 ) . To summarize the description above , we write the complete VEM algorithm for parameter updates: Algorithm 1: Variational expectation maximization for BicMix Data: Y , K , n_itr , a , b , c , d , e , f , aX , bX , cX , dX , eX , fX , α , β , αX , βX Initialize starting values: Sample η , γ , χ , φ ← G a ( 1 , 1 ) Sample π ← ℬ e t a ( α , β ) , πX ← ℬ e t a ( α X , β X ) , for j ← 1 to p do Sample ψj , j ← G a ( 1 , 1 ) for k ← 1 to K do Sample zk ← ℬ e r n ( π ) , ok ← ℬ e r n ( π X ) Sample ϕk , τk , ωk , κk ← G a ( 1 , 1 ) for j ← 1 to p do Sample Λj , k ← N ( 0 , 1 ) , Sample θj , k , δj , k ← G a ( 1 , 1 ) for i ← 1 to n do Sample xk , i ← N ( 0 , 1 ) , Sample σk i , ρk , i ← N ( 0 , 1 ) for t ← 1 to n_itr do for j ← 1 to p do Update Λj , ⋅ ← Eq ( 35 ) for k ← 1 to K do Update θj , k ← Eq ( 42 ) , δj , k ← Eq ( 44 ) for k ← 1 to K do Update ϕk ← Eq ( 46 ) , τk ← Eq ( 54 ) , zk ← Eq ( 60 ) Update η ← Eq ( 55 ) , γ ← Eq ( 56 ) , π ← Eq ( 61 ) for i ← 1 to n do Update X⋅ , i ← Eq ( 38 ) for k ← 1 to K do Update σk , i ← Eq ( 43 ) , ρk , i ← Eq ( 45 ) for k ← 1 to K do Update ωk ← Eq ( 50 ) , κk ← Eq ( 57 ) , ok ← Eq ( 62 ) Update χ ← Eq ( 58 ) , φ ← Eq ( 59 ) , πX ← Eq ( 63 ) for j ← 1 to p do Update ψj , j ← Eq ( 64 ) Output Λ , X , z , o Across hundreds of runs , we found that the number of recovered factors was stable within each application with low variance . As a rule of thumb , we initialized the number of latent factors to K = 2 × min ( n , p ) , and , if we find that the number of factors recovered is not reduced by approximately a third , we increased the number of initial factors until we saw this substantial reduction in K . We derive below the conditional distributions that capture the MCMC approach that we implemented for BicMix . In our manuscript , we used MCMC to compute the warm start parameter settings in the simulations . We updated the loading matrix Λ one row at a time , where each row consists of values across the K components . The jth row of the loading matrix , Λj , ⋅ , has the following posterior distribution , Λ j , · | Y j , · X , Θ i , ψ j , j ∼ N Y j , · ψ j , j - 1 X T X ψ j , j - 1 X T + V j - 1 - 1 , X ψ j , j - 1 X T + V j - 1 ( 65 ) where Vj is a K × K diagonal matrix . If we used Vj , k , k to denote the ( k , k ) th element for Vj , then we sampled Vj , k , k and its related parameters as follows: V j , k , k = θ j , k if z k = 1 ; ϕ k if z k = 0 . ( 66 ) Similarly , each column of X consists of values across the K components; the ith column of the factor matrix , X⋅ , i , has the following posterior distribution , X · , i | Y · , i , Λ , Σ i , Ψ ∼ N ( Λ T Ψ - 1 Λ + W i - 1 ) - 1 Λ T Ψ - 1 Y · , i , Λ T Ψ - 1 Λ + W i - 1 , ( 67 ) where Wi is a K × K diagonal matrix . If Wi , k , k denotes the ( k , k ) th element of Wi , then we sampled the value of Wi , k , k as follows: W i , k , k = σ k , i if o k = 1 ; ω k if o k = 0 . ( 68 ) We sampled values for the parameters conditional on sparse and dense state as follows . If zk = 1 , θ j , k | Λ j , k , δ j , k ∼ GIG a - 1 2 , 2 δ j , k , Λ j , k 2 ( 69 ) δ j , k | θ j , k , ϕ k ∼ G a ( a + b , θ j , k + ϕ k ) ( 70 ) ϕ k | δ j , k , τ k ∼ G a p b + c , ∑ j = 1 p δ j , k + τ k . ( 71 ) If zk = 0 , ϕ k | τ k , Λ j , k ∼ GIG c - p 2 , 2 τ k , ∑ j = 1 p Λ j , k 2 . ( 72 ) Correspondingly , the following parameters related to X were sampled as follows . If ok = 1 σ k , i | x k , i , ρ k , i ∼ GIG a X - 1 2 , 2 ρ k , i , x k , i 2 ( 73 ) ρ k , i | σ k , i , ω k ∼ G a ( a X + b X , ρ k , i + ω k ) ( 74 ) ω k | ρ k , i , ω k ∼ G a n b X + c X , ∑ i = 1 n ρ k , i + ω k . ( 75 ) If ok = 0 ω k | κ k , x k , i ∼ GIG c X - n 2 , 2 κ k , ∑ i = 1 n x k , i 2 . ( 76 ) The following parameters are not sparse or dense component specific; they each have a gamma conditional distribution because of conjugacy: τ k | ϕ k , η ∼ G a c + d , ϕ k + η ( 77 ) η | γ , τ k ∼ G a K d + e , γ + ∑ k = 1 K τ k ( 78 ) γ | η , ν ∼ G a ( e + f , η + ν ) . ( 79 ) Parameters related to X were sampled as , κ k ∼ G a c X + d X , ω k + χ ( 80 ) χ ∼ G a K d X + e X , φ + ∑ k κ k ( 81 ) φ ∼ G a e X + f X , χ + ξ . ( 82 ) The conditional probability for zk has a Bernoulli distribution: p ( z k = 1 | Θ Λ ) = π ∏ j = 1 p N ( Λ j , k | θ j , k ) G a ( θ j , k | a , δ j , k ) G a ( δ j , k | b , ϕ k ) ( 1 - π ) ( ∏ j = 1 p N ( Λ j , k | ϕ k ) ) + π ∏ j = 1 p N ( Λ j , k | θ j , k ) G a ( θ j , k | a , δ j , k ) G a ( δ j , k | b , ϕ k ) . Let pz = p ( zk = 1|ΘΛ ) ; then the conditional probability for zk is z k | p z ∼ B e r n ( p z ) . ( 83 ) The mixing proportion π has a beta conditional probability: π | α , β , z k ∼ B e t a α + ∑ k = 1 K 1 z k = 1 , K - ∑ k = 1 K 1 z k = 0 + β ( 84 ) where 1 is the indicator function . Similarly for X , the conditional probability for ok has a Bernoulli distribution: p ( o k = 1 | Θ X ) = π ∏ i = 1 n N ( x k , i | σ k , i ) G a ( σ k , i | a X , ρ k , i ) G a ( ρ k , i | b X , ω k ) ( 1 - π ) ( ∏ i = 1 n N ( x k , i | ω k ) ) + π ∏ i = 1 n N ( x k , i | σ k , i ) G a ( σ k , i | a X , ρ k , i ) G a ( ρ k , i | b X , ω k ) . ( 85 ) Let pX = p ( ok = 1|ΘX ) ; then the conditional probability for ok is o k | p X ∼ B e r n ( p X ) . ( 86 ) The mixing proportion π has a beta conditional probability: π X | α X , β X , O k ∼ B e t a α X + ∑ k = 1 K 1 o k = 1 , K - ∑ k = 1 K 1 o k = 0 + β X ( 87 ) where 1 is the indicator function . Finally , we have , ψ j , j ∼ IG n 2 + 1 , ∑ i = 1 n y j , i - ∑ k = 1 K Λ j , k x k , i 2 2 + 1 . ( 88 ) We implemented the following MCMC algorithm for sampling the parameters of the BicMix model . Algorithm 2: MCMC algorithm for BicMix Data: Y p × n gene expression matrix , K , n_itr Initialize parameters for t ← 1 to n_itr do for j ← 1 to p do for k ← 1 to K do Sample Vj , k , k according to Eq ( 66 ) Sample Λj , ⋅ according to Eq ( 65 ) for k ← 1 to K do if zk = 1 then Sample ϕk according to Eq ( 71 ) for j ← 1 to p do Sample θj , k according to Eq ( 69 ) , δj , k according to Eq ( 70 ) , if zk = 0 then Sample ϕk according to Eq ( 72 ) for k ← 1 to K do Sample τk according to Eq ( 77 ) , zk with Eq ( 83 ) Sample η according to Eq ( 78 ) , γ with Eq ( 79 ) , π with Eq ( 84 ) for i ← 1 to n do for k ← 1 to K do Sample Wi , k , k according to Eq ( 68 ) Sample X⋅ , i according to Eq ( 67 ) for k ← 1 to K do if ok = 1 then Sample ωk according to Eq ( 75 ) for i ← 1 to n do Sample σk , i according to Eq ( 73 ) , ρk , i according to Eq ( 74 ) , if ok = 0 then Sample ωk according to Eq ( 76 ) for k ← 1 to K do Sample κk according to Eq ( 80 ) , ok according to Eq ( 86 ) Sample χ according to Eq ( 81 ) , φ with Eq ( 82 ) , πX with Eq ( 87 ) for j ← 1 to p do Sample ψj , j according to Eq ( 88 ) Output Λ , X , z , o Using these simulated data , we compared BicMix to five other methods: Fabia , Plaid , CC , Bimax , and Spectral biclustering . We ran these methods using the following settings . For Sim1 , we set the number of components to the correct values , and ran each method as follows . For Sim2 , we corrected the simulated data for the dense components by controlling for five PCs in the simulated gene expression data , and we ran the methods on the residual matrix as in Sim1 , setting the number of components to 10 . We calculated a simple statistic to check the redundancy of the multiple components recovered across multiple runs as follows . For every component in all runs , we counted the number of genes with non-zero values , denoted as ng , and the number of samples with non-zero values , denoted as ns for each component . We then grouped the components that share the same ng and ns . For each pair of components in the same group , we counted how many components have non-zero values for the same genes and the same samples ( i . e . , the ℓ0 norm between components , or the Manhattan distance ) . Redundant components corresponded to pairs of factors and loadings for which the Manhattan distance is zero . We write out the algorithm we used to build the gene co-expression networks using the fitted BicMix model . Note that the sparsity-inducing prior on the covariance matrix of the factors increases the difficulty of computing the gene-wise covariance matrix relative to the common identity matrix covariance in the prior of the factors; however , all of the elements necessary to compute an estimate of the factor covariance matrix have been explicitly quantified in the VEM algorithm already . Algorithm 3: Algorithm to construct gene co-expression network Data: p × K loading matrix and K × n factor matrix; Ψ; net_type , rep , c , ( d ) . for i ← 1 to n_runs do for k1 ← 1 to K − 1 do for k2 ← K1 + 1 to K do if cor ( Λk1 , Λk2 ) × cor ( Xk1 , Xk2 ) > 0 . 5 then discard run if net_type = subset specific then Add component to A when non-zero factors are only in context c if net_type = subset differential then Compute Wilcoxon signed-rank test for non-zero factor values across contexts c , d Add component to A when p < 1 × 10−10 Construct the covariance matrix for X as Σ ← 〈 X XT 〉 − 〈 X 〉 〈 X 〉T ( Eqs 41 and 38 ) Calculate the variance for the residual as Ψ← Eq ( 64 ) Construct the precision matrix for subset A as Δ i = ( Λ A i Σ A , A i Λ A i T + Ψ ) − 1 Run GeneNet [63] on Δi to test significance of edges Store edges with probability of presence ≥ 0 . 8 Count number of times each edge is found across all runs Keep edges that are found ≥ rep times Output the nodes and edges Draw graph using Gephi [117] We computed the ( approximate ) expected number of edges to appear r or more times at random using our ensemble method as follows . Let R represent the total number of runs , r represent the threshold for number of runs an edge must appear , E represent the total number of possible edges , and e ^ represent the average number of edges per run . Note that this is an approximate expectation because we are using e ^ instead of the true number of edges recovered in each run; this expectation has a noticeable impact on the result when the variance in edges per run is large ( this was the case only in the breast cancer data set ) , but otherwise lead to a reasonable approximation . We computed the probability of a single edge occurring in at least r networks as follows: P r ( | e i | ≥ r ) = 1 - ∑ j = 1 r - 1 P r ( | e i | = j ) P r ( | e i | = j ) = Rj e ^ E j 1 - e ^ E R - j . Then the expected number of edges that will occur r or more times at random was approximated as follows , assuming edge independence: E r [ | e | ] = E · P r ( | e i | ≥ r ) . ( 89 ) | Recovering gene co-expression networks from high-throughput experiments to measure gene expression levels is essential for understanding the genetic regulation of complex traits . It is often assumed for simplicity that gene co-expression networks are static across different contexts—e . g . , drug exposure , genotype , tissue , age , and sex . The biological reality is that , along with differences in gene expression levels , there are differences in gene interactions across contexts . In this work , we describe a model for Bayesian biclustering , or recovering non-disjoint clusters of co-expressed genes in subsets of samples using gene expression level data . Using results from our biclustering model , we build gene co-expression networks jointly across all genes by computing the full regularized covariance matrix between all pairs of genes instead of testing each possible edge separately . Because biclustering recovers structure in subsets of the samples , we are able to recover gene co-expression networks that occur across all samples , that are differential across contexts ( e . g . , up-regulated in males and down-regulated in females ) , and that are unique to a context ( e . g . , only co-expressed in lung tissue ) . We illustrate the robustness of our approach and biologically validate the networks recovered from three different gene expression data sets . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"genetic",
"networks",
"breast",
"tumors",
"cardiovascular",
"anatomy",
"gene",
"regulation",
"cancers",
"and",
"neoplasms",
"random",
"variables",
"covariance",
"simulation",
"and",
"modeling",
"oncology",
"mathematics",
"arteries",
"network",
"analysis",
"genome",
"analysis",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"blood",
"vessels",
"gene",
"expression",
"breast",
"cancer",
"probability",
"theory",
"centrality",
"anatomy",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"computational",
"biology"
] | 2016 | Context Specific and Differential Gene Co-expression Networks via Bayesian Biclustering |
It is well established that individuals age differently . Yet the nature of these inter-individual differences is still largely unknown . For humans , two main hypotheses have been recently formulated: individuals may experience differences in aging rate or aging timing . This issue is central because it directly influences predictions for human lifespan and provides strong insights into the biological determinants of aging . In this article , we propose a model which lets population heterogeneity emerge from an evolutionary algorithm . We find that whether individuals differ in ( i ) aging rate or ( ii ) timing leads to different emerging population heterogeneity . Yet , in both cases , the same mortality patterns are observed at the population level . These patterns qualitatively reproduce those of yeasts , flies , worms and humans . Such findings , supported by an extensive parameter exploration , suggest that mortality patterns across species and their potential shapes belong to a limited and robust set of possible curves . In addition , we use our model to shed light on the notion of subpopulations , link population heterogeneity with the experimental results of stress induction experiments and provide predictions about the expected mortality patterns . As biology is moving towards the study of the distribution of individual-based measures , the model and framework we propose here paves the way for evolutionary interpretations of empirical and experimental data linking the individual level to the population level .
Aging can be generally defined as age-related changes in a set of variables , from growth rate to reproductive effort , which influence the fitness of an organism . Aging is a multiscale process which can be measured at almost every level of the individual organism . Individual metrics of aging include a broad range of processes from damage to DNA and proteins [1] , [2] to tissue loss of functionality [3] . Complementary to characterization of aging through individual metrics , there is a long history of demographic studies of aging going back to Gompertz's seminal studies [4]–[6] studying age-specific mortality . Age-specific mortality is arguably one of the most documented measure of aging . Since Gompertz' seminal work on human data [4] , such mortality curves have been obtained for a large variety of species [7] , in a broad range of environmental conditions ( e . g . , [8] , [9] ) . These curves play a central role in understanding aging processes and predicting the dynamics of population growth and human life expectancy [10] , [11] . Age-specific mortality curves allow the comparison of aging processes between species as the same measure can be applied from unicellular organisms to humans , as long as the death of the individual is clearly defined . All of these contribute to make changes in mortality over age a well-accepted definition of aging from the demographic perspective . This is the definition of aging considered throughout this paper . One of the striking observations resulting from inter-species comparison is that yeast , fly , worm and human mortality patterns share common properties . They all exhibit exponential increase and decrease with age [7] , but also differ in the timing and magnitude of these exponential phases and some species even exhibit mid-age or late-age plateaus [7] . In this article , we study the evolution of mortality patterns to investigate the nature of inter-individual differences in aging . The nature of inter-individual differences in aging is crucial to study and forecast a population's life expectancy , but is as yet still largely unknown . For humans , two main hypotheses have been proposed recently [11]: individuals may differ in aging rates or timing . Namely , from one individual to the other , the aging processes may be slowed or delayed . Addressing this issue is import to understand the potential and limits of individual medical treatments . Fundamental questions about inter-individual differences , i . e . individual heterogeneity , can be tackled studying age-specific mortality patterns . Because of these exponential patterns , differences in aging rates and timing can be visualized respectively as changes in slope and shift in mortality patterns [11] . Such changes have been documented for Drosophila melanogaster populations: when facing a diet restriction , a change in aging timing occurs while the response to fluctuations in temperature is a change in aging rate [12] , [13] . Previous work has used age-specific mortality to show that the aging rate seemed to be conserved in different human populations [14] and differed between different strains of Caenorhabditis elegans [15] . Both shift in level and differences in rate have been reported between baboon populations [16] or between male and females flies [17] . These mortality curves are population measures: they result from the aggregation of individual's aging . Biodemographic studies of aging have shown that specific population heterogeneity can reproduce the main features of qualitatively different mortality patterns , such as late-age mortality plateaus [18] , [19] . To do so , these studies make ad-hoc assumptions about ( i ) population heterogeneity , e . g . , a Gamma distribution , and ( ii ) the nature of inter-individual differences in aging . In this paper , we address these issues in the light of evolution , we do not assume a specific population heterogeneity and we explore different types of inter-individual differences in aging . In our model , we include Gompertz aging , that is exponential increase in mortality hazard with age , and implement heterogeneity in aging rates and timing . Population heterogeneity is allowed to evolve over generations under the influence of life-history trade-offs . Following the disposable soma theory [20] , [21] and empirical observations made in a broad range of species [22] , we incorporate a trade-off between reproduction and aging . In this framework , investing in maintenance mechanisms ( rather than reproduction ) results either in slowing or delaying aging depending on the nature of inter-individual differences . In section ‘Transitions in mortality curves’ , we discuss predictions concerning population heterogeneity and the corresponding mortality patterns under different assumptions about inter-individual differences . We compare our results to mortality patterns of yeast , flies , worms and humans respectively . In section ‘Influence of mutation rate’ , we test the robustness of our results with respect to mutation rates and highlight new features for the distribution of heterogeneity . Finally , in section ‘The notion of subpopulation’ , we exploit the predictions of the model to shed light on the notion of subpopulations which is invoked in numerous experimental studies [23] , [24] .
The model we propose describes evolving populations in which individuals age , die and compete for reproduction while they are alive . The offspring they produce fill the next non-overlapping generation until the desired population size is reached . Survival and reproductive success of each individual are connected following the disposable soma theory . The more an individual invests in reproduction , the shorter its lifespan . For each individual , a single parameter describes its resource allocation strategy between maintenance and reproduction . Different individuals may have different strategies ( i . e . , different ) and one of the key features of these models is that this distribution of strategy in the population evolves over generations . For each generation , all the individuals are synchronized , starting with age zero . While an individual is alive , at each time step it has a probability to be chosen for reproduction , proportional to its parameter , normalized by the sum of the of alive individuals: . The normalization factor implements a competition for reproduction because the ability for each individual to reproduce depends on the composition of the whole population . The higher are more likely to reproduce than the lower , but if the population is reduced to one single individual , the probability it has to reproduce is one , independently of its parameter . Competition for reproduction has been documented in a broad range of species and can take numerous forms , from limited access to habitat [25] to dependence on external resources [26] . Once an individual is chosen for reproduction , it produces one offspring which inherits its value if no mutation occurs . We use the term mutations here to describe a process introducing variability in the inheritance process . In Text S1 , we explore different implementations of such a process and show that the conclusions of this paper do not rely on specific modeling choices . Each reproductive event has a fixed mutation probability : if a mutation occurs , the offspring value of is replaced by a random number drawn from the uniform distribution between zero and one . The result is that is a heritable trait , subject to mutations and influences the reproductive success of the individual . The model we propose simulates evolving populations with a fixed size and non-overlapping generations . We start studying individuals at maturity , so that they are able to reproduce since time . In the initial population each individual is assigned a randomly chosen drawn from a uniform distribution between zero and one . We then let population heterogeneity change over generations until it reaches a quasi stationary distribution . Building the generation from generation occurs as follows . First , at time , all the individuals are alive . The higher have a higher probability to be chosen for reproduction and therefore the first offspring added to the initially empty generation are likely to have high . Then , as time goes on , individuals with high face a faster aging process than their counterparts with low . As the high die out , lower have more and more access to reproduction , thus producing on average low offspring . The ‘low ’ phenotype experiences a positive frequency-dependent selection as its fitness increases as their frequency in the population increases . The reproduction process is iterated until the generation is filled . Some individuals might have inherited a very low because of the mutation process . They live much longer than the time required to fill a generation and do not reproduce . These individuals we will refer to as the ‘oldest old’ and present the influence they have on mortality patterns in section ‘Transitions in mortality curves’ . This reproduction process is asexual , as one individual is either cloned , if no mutations occur , or modified in the other case . In Text S1 , we show that introducing sexual reproduction does not alter the conclusions of the paper . Following the disposable soma theory , we implement a trade-off at the individual level between survival and reproduction . For each individual , the hazard of death between age and age is given by a function which increases with and . The dependence on depicts heterogeneity in the population as each individual has its own . The increase in mortality over age depends on the interactions between ( i ) the species biological characteristics and ( ii ) its environment [12] , [13] . This interaction is captured by a parameter in our model which influences the rate of the individual's exponential aging , differing from one species to the other . This baseline aging is then modulated by the individual strategy of resource allocation between maintenance and reproduction , captured by the parameter . Whether modulates the rate or the timing of the baseline aging is an open question [11] . To address it , we therefore study two distinct individual mortality functions . First , we implement which leads to heterogeneity in aging rates , described afterwards as the Heterogeneity in aging Rate Model ( HRM ) . The parameter reflects the initial mortality and the parameter the interactions between the biology of the species and its environment . These two parameters are constant over generations and identical for all the individuals , thus defining a baseline aging . Second , to derive a mortality function for the Heterogeneity in aging Timing Model ( HTM ) , let us now consider linear dependence on for the individual mortality function and two different individuals , respectively defined by and . Their respective mortality function and are: and . Therefore , a change in initial mortality corresponds to a delay in aging , i . e . , a shift of the mortality curve along the time axis , while both individuals still share the same aging rate . In contrast , in the case of the mortality function presented in the previous paragraph , two different individuals would start from the same initial mortality but differ in their aging rates . The two hypotheses concerning heterogeneity in aging postulate different underlying trade-off mechanisms . In the case of heterogeneity in aging rates , investing in maintenance mechanisms slows down the aging process whereas in the case of heterogeneity in aging timing , the same investment delays the aging process . In Text S1 , we also present and discuss the influence of age-independent factors , usually referred to as extrinsic mortality , on the evolution of mortality patterns . Over generations , the evolutionary process reshapes the distribution of -strategies according to the environmental parameter and the choice of the individual mortality function . For instance if , in the ( extreme ) case of , the optimal strategy is because the trade-off has no effect . On the other hand , high values of are likely to draw the distribution of towards lower values so that individuals have to survive to get the opportunity to reproduce . We present the evolution of these distributions in section ‘Evolution of heterogeneity’ . After several hundreds of generations ( 400 in the simulations below ) , the distribution stabilizes and mortality curves can be estimated . The probability for a whole population to disappear in the course of evolution is strictly positive . Indeed , each individual has a non-zero probability to die at each time step and this could happen before it reproduces at all . Population size can theoretically decrease over generations until extinction . Yet , for a broad range of parameters , this does not occur and a quasi-stationary distribution of strategies is reached . In Text S1 , we provide a mathematical formulation of the model which would allow one to study the existence and uniqueness of these quasi-stationary distributions from a formal standpoint . Here , in the following sections , we focus on the properties of these post-evolution quasi-stationary distributions . In this paper , we focus on adult mortality patterns and , as such , we do not explicitly address the heterogeneity arising from developmental processes . We assume that development leads to the expression of the inherited phenotype . In Text S1 , we explore the effects of a maturation period before the onset of reproduction and provide mathematical tools to address the issue of additional development-induced heterogeneity in more depth . This additional information could be used to study the known effects of the environment during development on population heterogeneity [27] . In this section , we present the results of evolution in different environments under two distinct assumptions , heterogeneity in aging rates ( HRM ) and heterogeneity in aging timing ( HTM ) , and compare the outcome with empirical and experimental data . In figure 1 , first row , we present mortality patterns of C . elegans , D . melanogaster , humans and yeasts . From left to right , the C . elegans mortality curve presents a two-stage exponential increase , usually referred to as a ‘kink . ’ Human mortality data also exhibits a two-stage exponential , but is separated by a slowing down or even a slight decrease . The D . melanogaster mortality curve has an exponential-plateau-exponential pattern; in contrast , the one for yeasts display a clear U-shape . Simulation results with both the HRM and HTM , presented in figure 1 , second and third row respectively , exhibit the same set of mortality curves . Both models lead to a transition from an exponential-exponential ( ‘kink’ ) pattern under small values for ( on the left ) , to an exponential-slowing down-exponential pattern when evolution occurs under higher values for . Increasing further leads to an exponential-plateau-exponential pattern , then to an exponential-decrease-exponential pattern . Finally , the highest value for presented here corresponds to an exponential-U-shape-exponential pattern . These transitions in shape reproduce the transitions observed in experimental and empirical data . Obtaining all these experimental patterns only requires the modification of the single parameter and letting population heterogeneity evolve . In Text S1 , we show that these mortality patterns are also robust to an extensive exploration of parameter space . The transitions in mortality patterns presented in figure 1 are calculated without accounting for the oldest-old individuals , i . e . , those living longer than the time it takes to fill a generation . In figure 2 , we present the influence of these individuals on mortality patterns . Including these individuals does not modify the qualitative shapes presented in figure 1 but leads to a decrease in mortality at late ages . Simulated mortality patterns of both the HRM and the HTM resemble medfly mortality which presents two peaks and a late-age decrease . In both models , population heterogeneity is shaped by natural selection , via a life-history trade-off between survival and reproduction . The transitions observed in the two versions of the model rely on a change of the underlying heterogeneity in in response to the parameter . Figure 3 shows the quasi-stationary distribution of as a function of in the case of heterogeneity in aging rates . Under low basal damage accumulation ( low ) , the most prevalent strategy consists of investing in reproduction ( close to one ) . As increases , being able to survive in order to reproduce leads the distribution of strategy to move towards more investment in maintenance mechanisms ( low values of ) . Mortality patterns presented in figure 1 , second row , directly derive from these evolved distributions of heterogeneity . In the case of heterogeneity in aging timing , population heterogeneity evolves with the parameter as depicted in figure 4 . The same qualitative understanding holds: under low damage accumulation , investment in reproduction is favored while maintenance becomes prevalent if damage accumulates faster . Here , the distributions are bimodal and increasing mainly changes the ratio between the two peaks , while also slightly shifting the left peak towards more maintenance . In the previous case ( heterogeneity in aging rates ) , the distributions are unimodal and the main effect of is the shift . In our models , individuals are competing with each other to access reproduction . Different , that is different resource allocation strategies , have different prevalence , as shown in figures 3 and 4 . As the parameters and are fixed initially and kept constant over generations , the fact that a whole distribution of strategies is maintained in the population , without one taking over , could be surprising . Previous models of competition have shown that in many set-ups , one strategy invades the population and the distribution of heterogeneity after several generations is a single peak corresponding to the optimal strategy or species [28] . Polymorphism can be maintained in density-dependent selection models , but it usually requires negative density-dependence [29] ( that is the fitness of a trait increases as the frequency of the trait in the population decreases ) . Interestingly , in this model , heterogeneity is maintained over generations even though all the experience a positive density-dependent selection . In Text S1 , we account for this sustained heterogeneity in both the HTM and the HRM . Mutations occurring on influence the heterogeneity distributions in evolved populations . A high mutation rate tends to make the distribution of closer to a uniform distribution . In this section , we first present the effects of high mutation rates on populations heterogeneity as well as the corresponding effects on the mortality patterns . First , we find that the properties of asymmetry in heterogeneity described in figure 3 and 4 are maintained despite high mutation rates , as shown in figures 5 and 6 . The left side of the distribution is not altered by the mutation process: the shape of the distribution close to zero remains unchanged even under high mutation rates . This derives from the fact that individuals below a certain threshold have negligible opportunities to reproduce and are therefore absent from the population . More details about the asymmetry between low and high are provided in Text S1 . Results in sections ‘Transitions in mortality curves’ and ‘Evolution of heterogeneity’ suggest that changes in population heterogeneity have a considerable influence on mortality patterns , providing the transitions in shape previously described . Yet , we find here that if the left-side of the distribution is preserved , then mortality patterns are not qualitatively changed , as shown in figure 7 and 8 . In this case , the exponential-plateau-exponential pattern remains unchanged . These findings imply that the presence of the strongest individuals dramatically influences the shape of mortality patterns . Whether stress induction experiments induce adaptation which modifies population heterogeneity is a burning issue [27] . If adaptation occurs when is modified , mortality patterns presented in figure 1 emerge . In this section , we complement the previous results with predictions concerning the mortality patterns expected when there is no adaptation of population heterogeneity . In figures 11 and 12 , we show the qualitative changes in mortality patterns when is modified but not the distribution of in the HRM and the HTM respectively . In both figures , the distribution of presented in the right column almost always led to exponential-exponential ( kink ) mortality patterns . The only exception occurs for the HRM and high combination because individuals die too quickly to exhibit the two-stage exponential increase . In all the other cases , the kink shape is preserved but the time at which the kink occurs as well as the corresponding mortality levels are modified by . These findings are in agreement with experimental results for C . elegans observed under different diets [9] . The results presented in the other columns provide predictions concerning the expected mortality patterns for other population heterogeneity . The interests of this exploration are twofold . First , it allows one to determine whether a given stress induction experiment will modify population heterogeneity . If the same distribution of cannot produce the mortality pattern with and without stress by only modifying , our model predicts that adaptation as occurred . Second , if adaptation does not occur , the mortality patterns under perturbation present significant differences between the HRM and the HTM . One of the main observation which stands out from the comparison between figure 11 and 12 is that the mortality rates are more modified in the HRM than in the HTM in response to changes in . Therefore , this prediction can provide a way to decide whether inter-individual differences in aging rely on rate of timing .
The simplicity of the framework we propose also enables the formalization of intuitive notions , such as a ‘subpopulation’ and the extensive exploration of mortality functions provides predictions of possible curves for organisms yet to be studied , along with expected distributions of heterogeneity . The robustness of mortality patterns observed suggests that aging is itself a robust process , relying on similar processes across species . We hope that this work paves the way for ( i ) faster understanding and classification of heterogeneity distributions across species which are not model organisms and ( ii ) opens up new prospects in terms of understanding the evolution of aging and its robustness . The set-up allows easy changes and explorations , as well as creating space to make the interactions between aging and reproduction more complex . In that sense , it complements previous approaches combining evolutionary theories and heterogeneity , as it provides a framework to explore yet to be explained aging dynamics .
We have simulated our model with both a continuous time and a discrete time framework , in both a stochastic and a deterministic manner , in order to ensure the robustness of our results . All models have been implemented in C and this section provides numerical methods for the algorithm used in our simulations . We have also implemented a deterministic model which corresponds to the stochastic models presented in the main text in the case of infinite populations . The purpose is twofold: ( i ) it ensures the robustness of our results and ( ii ) it paves the way for an analytical analysis of the evolutionary algorithm we present in this paper . For instance , the formalisation presented in this section allows one to study from a formal standpoint the convergence to a stationary distribution with infinite populations . The deterministic model is defined at the population level and describes the changes in , the distribution of in the population at generation . This distribution changes over generations through the same principle: individuals reproduce based on their and die according to the individual mortality function chosen . The distribution of changes over time in one generation , as individuals die , following: . is therefore defined aswhere accounts for the probability that an offspring inherits a parameter given that its parent had a parameter and for the finite size of the population . Indeed , as the rate of reproduction is constant over time , limiting the population size is equivalent to limiting the overall time allowed for reproduction . In sum , where is the operator on distributions described in the equation above . The evolutionary process described in the main text converges towards a stationary distribution of after several hundreds of generations when a fixed point for the operator is reached . In Text S1 , we show that the deterministic model and the stochastic models produce the same results . | Aging is a widespread phenomenon across the tree of life . From yeast to humans , mortality changes over age have been widely documented . Interestingly , all individuals are not equal with respect to the aging process: large variability in individual life span has been reported , even in clonal populations . Understanding the nature of these differences is of great interest for medical research . So far , two hypotheses have been proposed: individuals may differ in their aging rate or timing . Here , we show that these two hypotheses can reproduce experimental and empirical mortality patterns as a result of natural selection . We also predict the corresponding population heterogeneity in aging . Many studies define subpopulations ad hoc , the work we present provides insight into a more accurate description of inter-individual differences in aging . Finally , our analysis also predicts the modifications of these mortality patterns under stressful conditions . This exploration reproduces experimental data obtained with heat shocks and permits to foresee new mortality patterns that could be observed with other perturbations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"aging",
"evolutionary",
"biology",
"evolutionary",
"modeling",
"biology",
"computational",
"biology"
] | 2013 | How Evolving Heterogeneity Distributions of Resource Allocation Strategies Shape Mortality Patterns |
Few studies have estimated prevalence of neurocysticercosis ( NCC ) among persons with epilepsy in sub-Saharan Africa . While the limitations of serological testing in identification of NCC are well known , the characteristics of persons who are misdiagnosed based on serology have not been explored . The first objective of this pilot study was to estimate the prevalence of NCC in epilepsy outpatients from an area of South Africa endemic for cysticercosis . The second objective was to estimate the accuracy of serological testing in detecting NCC in these outpatients and characterize sources of disagreement between serology and neuroimaging . All out-patients aged 5 or older attending the epilepsy clinic of St . Elizabeth's Hospital in Lusikisiki , Eastern Cape Province , between July 2004 and April 2005 were invited to participate . Epidemiological data were collected by local study staff using a standardized questionnaire . Blood samples were tested by ELISA for antibody and antigen for Taenia solium . Four randomly chosen , consenting participants were transported each week to Mthatha for brain CT scan . The proportion of persons with epilepsy attending St . Elizabeth clinic with CT-confirmed NCC was 37% ( 95% CI: 27%–48% ) . Using CT as the gold standard , the sensitivity and specificity of antibody testing for identifying NCC were 54 . 5% ( 36 . 4%–71 . 9% ) and 69 . 2% ( 52 . 4%–83 . 0% ) , respectively . Sensitivity improved to 78 . 6% ( 49 . 2%–95 . 3% ) for those with active lesions . Sensitivity and specificity of antigen testing were considerably poorer . Compared to false negatives , true positives more often had active lesions . False positives were more likely to keep pigs and to have seizure onset within the past year than were true negatives . The prevalence of NCC in South African outpatients with epilepsy is similar to that observed in other countries where cysticercosis is prevalent . Errors in classification of NCC using serology alone may reflect the natural history of NCC .
Neurocysticercosis ( NCC ) results when the central nervous system ( CNS ) is invaded by the larval stage of Taenia solium , a zoonotic cestode transmitted between humans and pigs . Humans infected with the adult tapeworm , a disease called taeniasis , contaminate the environment with the parasites' eggs contained in the proglotids that are shed in their feces . Pigs become infected with the eggs and the larvae migrate to several tissues , including muscles . Humans acquire taeniasis when eating insufficiently cooked , infected pork meat . In areas where sanitation and personal hygiene are poor , humans can also become infected with the eggs present in feces of people infected with taeniasis . Eggs develop to larvae , enter the CNS and can lodge in the parenchyma , subarachnoid space , or the ventricular system of the brain and are sometimes seen in the eye [1] . Since each proglotid contains about 40 , 000 eggs and infected humans may shed a few proglotids per day , a single carrier can be responsible for the infection of multiple individuals and pigs in a community [1] , [2] . The manifestations of NCC are protean , but approximately 80% to 90% of persons with symptomatic NCC experience seizures at some time in the course of their illness [3] , [4] , making this manifestation a reasonable starting point for identifying NCC cases in populations at risk . Acute symptomatic seizures ( seizures occurring only in the presence of an acute central nervous system ( CNS ) process ) in NCC are thought to result from inflammatory responses in the brain in response to a degenerating cyst [5] . Calcified intraparenchymal cysts may serve as a focus for remote symptomatic seizures as late sequelae of NCC [5] , but there is not always concordance between the location of calcified cysts and a seizure focus identified by electroencephalogram [6] . The proportion of persons with acute symptomatic seizures associated with NCC who go on to develop epilepsy ( recurrent , unprovoked seizures ) is unknown , and determining the proportion of seizures attributable to NCC is problematic in the absence of repeated serological testing and concurrent diagnostic neuroimaging . In spite of these difficulties , NCC is considered to be the leading preventable cause of epilepsy in developing counties [7]–[9] . Some of the studies used to support this contention have not discriminated between acute symptomatic seizures and epilepsy or have counted as NCC cases persons with seizures and positive serology for T . solium ( for example , Garcia et al . , 1993 [10]; Nsengiyumva et al . 2003 [11] ) . In the absence of neuroimaging evidence of brain lesions compatible with NCC , it is not possible to determine whether infection with T . solium may be the cause of the epilepsy . This is especially of concern in areas endemic for cysticercosis where many people may be exposed to the eggs of the parasite , including exposure after the onset of epilepsy . The relation between number , location and stage of NCC lesions and the presence and type of clinical manifestations further complicates the discrimination between acute symptomatic seizures and epilepsy in the field [10] , [12]–[15] . Alternatives to serology for identifying lesions of NCC in the brain are computerized tomography ( CT ) or magnetic resonance imaging ( MRI ) of the brain . The results of the imaging combined with epidemiological and serological results have been suggested by an international panel of experts for the diagnosis of NCC [16] . Reliance on neuroimaging studies for the diagnosis of NCC is not generally feasible , however , especially in developing countries where the disease is likely to be most prevalent and CT and MRI are often not available . In addition , it is not a reasonable method for case identification in community-based field studies of the epidemiology of NCC . Development of alternative , valid strategies for identifying persons with NCC would greatly facilitate such studies . The first objective of this pilot study was to estimate the proportion with NCC in persons attending an outpatient clinic for epilepsy in an area of South Africa endemic for cysticercosis . A second objective was to estimate the accuracy of antibody and antigen serological testing in detecting cases of NCC in patients with epilepsy .
Persons eligible for the pilot study were out-patients who were attending the epilepsy clinic of St . Elizabeth's Hospital in Lusikisiki , Eastern Cape Province , South Africa . These patients were either referred for seizures from a rural clinic or presented themselves directly to the hospital for seizure diagnosis and care . All individuals aged 5 years or older seen at St . Elizabeth's Hospital between July 2004 and April 2005 with possible new onset epilepsy ( incident cases ) and those returning for continuing care of epilepsy ( prevalent cases ) were invited to participate . Clinical and epidemiological data were collected by local study staff for all eligible consenting patients . Attempts were also made to obtain blood samples from all participants in order to test for the presence of antibody and antigen for T . solium . A standardized questionnaire , adapted from Preux [9] and developed by the Cysticercosis Working Group of Eastern and Southern Africa ( CWGESA ) , was administered in either English or isi-Xhosa . Isi-Xhosa is spoken by the vast majority ( 95% ) of the black population in the Eastern Cape Province [17] . Translation from English to isi-Xhosa and back translation to English was done by two independent local residents who were fluent in both languages . The interview included questions designed to screen for epilepsy as well as information about potential risk factors for and knowledge of cysticercosis and taeniasis , pig-keeping practices of the household , and questions on their expenditures for medical care for seizures . ( Copies of the questionnaire are available on request . ) Mothers or primary care takers completed the interview for children . Clinical data for each participant , including the diagnosis of epilepsy and the types of seizure manifestations , were provided by the local co-investigator ( G . S . -O . ) . The International League Against Epilepsy ( ILAE ) classification of seizure types was used to designate primary type of seizure based on descriptions of seizure manifestations in the clinical data [18] , [19] . For these analyses , epilepsy was defined as a clinical diagnosis of seizures in combination with self-report from the questionnaire of ever having had more than one seizure . At least two seizures separated by at least 24 hours had to have occurred for a patient to be diagnosed with epilepsy or recurrent unprovoked seizures . An incident case of epilepsy was defined based on questionnaire response as occurrence of first seizure within the past year . Serological testing for T . solium antibody and antigen was offered to all participants . Serological testing for the presence of antibody to T . solium was done using the ELISA method with purified antigens ( RIDASCREEN® Taenia solium IgG commercial diagnosis kit ( R-Biopharm , Darmstadt , Germany ) ) [20] . Blood serum samples were collected and transported from Lusikisiki to Mthatha where serum was removed , aliquoted and frozen . One frozen serum aliquot was transported to East London to detect the presence of antibodies in the serum using an ELISA test . The second serum aliquot was transported to the Medical Research Council in Cape Town to detect the presence of antigens in the serum using an ELISA . The test used to detect antigens was the B158/B60 monoclonal antibody-based ELISA developed by Dorny et al . , 2000 [21] . For this Ag-ELISA the calculation of the cut off is based on the results of eight negative controls . The optical density ( OD ) of each serum sample is compared with the mean OD of a series of eight samples from non-infected individuals at a probability level of P<0 . 001 to determine the result in the test [22] . The calculation is based on Student's t-test . In some analyses , the results of the antigen-ELISA and antibody-ELISA were considered together to define a positive or negative result . If both measures were positive , the combined measure was considered positive; if both were negative , the combined measure was negative . A positive result to either test was considered positive , even if the other result was missing . Similarly , a negative result to either test was considered negative , even if the other result was missing . These assumptions were considered reasonable since among patients with both antigen and antibody results available , 95% ( 38/40 ) of those with a negative antibody result also had a negative antigen result . Only one person had data missing for both antigen and antibodies . Each week , four randomly chosen patients with clinically diagnosed seizure disorder and who gave written consent were transported for CT scan of the brain ( Toshiba Asteion TSX-021A ) to the Nelson Mandela Academic Hospital ( NMAH ) , a teaching hospital affiliated with the Walter Sisulu University for Technology and Science in Mthatha . The transportation , per diem and cost of the CT were covered by the research project and did not incur any costs to the patients . When estimating the proportion of cases associated with NCC , it is essential that seizure patients be sampled randomly for CT in order to avoid persons suspected of having NCC being preferentially referred . The random sampling was done by first randomly selecting a day of the week ( Monday through Friday ) . The first four eligible patients at St . Elizabeth's seen on the randomly chosen day were referred to the NMAH . If there were not four eligible patients on that day , the remaining number required to reach a total of four were chosen from the first eligible patients seen the following day . If the random day was Friday and there were not four eligible patients on that day , then the remainder were selected on the following Monday . CT scans were done with and without radiographic contrast material for all referred patients unless contrast was contraindicated ( e . g . , allergy to contrast media , pregnancy ) . Scans were read by radiologists ( IT , MAA ) at the NMAH as part of usual clinical care . Readings were done without knowledge of serological results . For this pilot study , results of the CT scans were used as the “gold standard” for classification of NCC . We recognized that imaging is not a true gold standard for the diagnosis of NCC but there were no other imaging facility available to the patients in Mthatha . It is reasonable , however , to assume that the CT-scan will be a much better test for identifying lesions in the brain and a serological test . Using the CT report results , cases were classified as “NCC” ( either evidence of visible cysticerci , colloidal cysts or presence of cerebral calcifications with a specific CT diagnosis of NCC ) or “not NCC” ( no diagnosis of NCC and no evidence of cerebral calcifications of any kind ) . Serology results were not used in making these designations . Persons with CT scans interpreted as “cerebral calcifications” but without mention of NCC and who had no evidence of active cysticerci were included in one set of estimates of the proportion of epilepsy associated with NCC but were excluded from analyses of NCC in order to reduce the possibility of misclassification . All interview data were independently entered in an Access® database by two individuals . Questions for which there was a disagreement of entries of more than 5% were reviewed and corrected . Questions with disagreements of less than 5% were considered acceptable . Sixty percent of all entries had a mismatch of 2% or less . Prevalence proportion ratios , sensitivity , specificity and predictive values were estimated . The 95 percent confidence intervals ( 95%CI ) of the field-based diagnosis of NCC were calculated assuming a binomial distribution and using the exact Clopper-Pearson method [23] . Data were examined for confounding and interaction for all covariates , but since neither was present , univariate results are reported . All analyses were conducted using SAS version 9 . 1 ( SAS Institute , Inc . , Cary , North Carolina ) . Two signed consent forms were used in this study . The first consent form asked participants for their agreement to be interviewed and have a blood sample drawn for the serological analysis . A second consent was used for those patients selected at random to have a CT-scan in Mthatha . This project was approved by the Institutional Review Boards of the University of Oklahoma Health Sciences Center , the Walter Sisulo University , and St . Elizabeth Hospital .
A total of 296 individuals with suspected epilepsy were seen at the outpatient clinic in St . Elizabeth of whom 281 were diagnosed with at least one seizure according to the clinical report . Socio-demographic and clinical characteristics as well as serological results from these patients are shown in Table 1 . Ages ranged from 5 to 76 years with 31 ( 11 . 0% ) cases being children ( i . e . , <16 years of age ) . The seroprevalence of antibodies and antigens to the larval stage of T . solium were 32 . 6% ( 95%CI: 27 . 1%–38 . 5% ) and 7 . 9% ( 95%CI: 4 . 5%–12 . 8% ) , respectively . Generalized seizures were reported slightly more often than partial seizures . Of the generalized seizures , tonic-clonic were most often reported , and the most commonly observed partial seizures were either complex partial seizures with motor manifestations ( n = 79 , 28 . 1% ) or partial seizures secondarily generalized ( n = 40 , 14 . 3% ) . A total of 244 patients ( 86 . 8% , 95%CI: 82 . 3%–90 . 6% ) met our definition of epilepsy . Among those with epilepsy , the onset of seizures was within the past year in 28 . 9% ( 95%CI: 23 . 0%–35 . 4% ) of cases , while another 31 . 2% ( 95%CI: 25 . 1%–37 . 8% ) had been having seizures for 10 years or longer at the time of this study . Since the referral of patients for CT was done randomly , the characteristics of those with and without CT are expected to be similar . There were more patients sent for CT for whom seizure frequency was unknown ( 13 . 5% vs . 2 . 9% ) and more children aged less than 16 ( 16 . 3% vs 7 . 2% ) , but no other statistically significant differences were observed in demographic , seizure type or duration , or serological results between those who were or were not referred for CT . The distributions of demographic and clinical characteristics of those referred as compared to those not referred for a CT scan are given in Table S1 . A total of 92 patients who met the study's definition of epilepsy received a CT scan . Of those , 34 ( 37 . 0% , 95%CI: 27 . 1%–47 . 7% ) had a CT diagnosis of NCC; 20 ( 21 . 7%; 95%CI: 13 . 8%–31 . 6% ) of these showed calcification with a diagnosis of NCC and 14 ( 15 . 2%; 95%CI: 8 . 6%–24 . 2% ) had either visible cysticerci or colloidal lesions ( subsequently referred to as “active” NCC ) . In these two groups , the proportions with new-onset seizures were 31 . 6% ( 95%CI: 12 . 6%–56 . 6% ) and 46 . 2% ( 95%CI: 19 . 2%–74 . 9% ) , respectively . If persons with epilepsy who had CT evidence of cerebral calcifications not diagnosed as NCC ( n = 19 ) were included as possible cases of NCC , the proportion of epilepsy associated with NCC would be 43/92 or nearly 47% . These 19 uncertain cases were excluded from further analysis of NCC as these lesions may be due to other cerebral infections such as tuberculosis . Some cases of NCC may have been missed by this approach , which would under-estimate the proportion of seizure cases associated with NCC . Excluding them has the benefit , however , of reducing the chances of misclassification within the NCC group , which increases the validity of the comparisons between NCC and non-NCC epilepsy cases . Characteristics of those with calcified NCC lesions , ‘active’ NCC , and those with no CT abnormalities are summarized in Table 2 . Those with ‘active’ lesions were less often female and more often seropositive than those with calcified lesions or those with no CT abnormalities . Prevalence proportion ratios for selected characteristics were calculated comparing those with: 1 ) any definite NCC diagnosis to those with no CT abnormality , and 2 ) those with ‘active’ NCC to those with no abnormality ( Table 3 ) . The small number of cases in each group is reflected in the wide 95% CI . When compared to those with epilepsy who had no CT abnormalities , the proportion of incident cases of epilepsy tended to be higher in those with NCC ( PPR = 1 . 97 , 95%CI: 0 . 87–4 . 42 ) and even higher in those with ‘active’ NCC ( PPR = 2 . 39 , 95%CI: 0 . 97–5 . 89 ) . The prevalence of a positive antibody test was higher in those with NCC ( PPR = 1 . 77 , 95%CI: 1 . 01–3 . 12 ) , especially in those with ‘active’ NCC ( PPR = 2 . 55 , 95%CI: 1 . 48–4 . 40 ) as compared to those without NCC . Compared to those with calcified NCC lesions , the prevalence of seropositivity for antibody to T . solium was 2 . 55 times higher ( 95%CI: 1 . 48–4 . 40 ) in those with ‘active’ NCC . The validity of serological testing for antibodies or antigens as diagnostic tools for NCC in persons with epilepsy was estimated separately ( Table 4 ) . The predictive value for NCC of a positive antibody test was 60 . 0% ( 95%CI: 40 . 6%–77 . 3% ) . Hence , 40% ( 95%CI: 22 . 7%–59 . 4% ) of those testing positive to the antibody ELISA but without NCC would be falsely attributed to NCC ( false positives ) based on serology alone . When this analysis was confined to those who had ‘active’ NCC , the sensitivity of serology in epilepsy patients was higher ( 78 . 6% , 95%CI: 49 . 2%–95 . 3% ) . However , since the prevalence of ‘active’ NCC is lower than that of all NCC , the predictive value of a positive antibody test in these patients was reduced to only 47 . 8% ( 95%CI: 26 . 8%–69 . 4% ) , therefore leading to 52 . 2% false positives . The predictive value of a positive antigen test was similar to that for antibody at 66 . 7% ( 95%CI: 22 . 2%–95 . 7% ) but the sensitivity was very poor at 17 . 4% ( 95%CI: 5 . 0%–38 . 8% ) . When the validity of antigen testing was examined only among those with ‘active’ NCC lesions , the predictive value for a positive antigen test was 50 . 0% ( 95%CI: 6 . 8%–93 . 2% ) . Serological studies might be more useful in identifying cases of NCC if only those with new onset ( incident ) seizures were assessed . Unfortunately , there were only 14 patients with new onset seizures who had a CT-scan and either antigen or antibody serology performed . Nonetheless , analysis in this small group suggested that being seropositive for T . solium may be a good field diagnostic tool for NCC among incident seizure cases ( predictive value of positive serology = 80% , 95%CI: 44%–98% ) , although 40% ( 95%CI: 5%–85% ) of those whose seizures are not related to NCC will be falsely categorized by serology as NCC . A series of comparisons was made between those who were correctly and incorrectly classified as NCC by serology ( antibody ELISA or antigen ELISA ) in order to see whether there were characteristics that distinguished them . Serological true positives ( CT positive for NCC and seropositive ) and false negatives ( CT positive for NCC but seronegative ) as well as serologically true negatives ( CT negative for NCC and seronegative ) and false positive ( CT negative for NCC and seropositive ) are compared in Table 5 . The small numbers in each group are reflected in the wide 95% CI . More true positives had ‘active’ lesions ( PPR = 3 . 06 , 95%CI: 1 . 04–8 . 97 ) than false negatives . Persons with epilepsy who are falsely identified as cases of NCC based on serology were more often present or past owners of pigs ( PPR = 1 . 35 , 95%CI: 1 . 08–1 . 69 ) and had a higher prevalence of incident seizures ( PPR = 1 . 23 , 95%CI: 1 . 02–1 . 47 ) compared to those correctly identified by serology as epilepsy not due to NCC .
The source population for this study included persons seeking medical care for symptomatic seizures ( acute seizures ) or recurrent asymptomatic seizures ( epilepsy ) at an outpatient clinic of St . Elizabeth Hospital , a rural hospital in Lusikisiki , Eastern Cape Province , South Africa . While the study includes consecutive , unselected epilepsy patients , this group does not represent a community-based sample of persons with epilepsy since only those seeking care are represented . The characteristics of these patients with respect to age distribution and seizure types were similar , however , to those observed in community-based studies of epilepsy conducted in sub-Saharan Africa [9] , [11] , [24] . Nearly one-third of epilepsy patients were seropositive to T . solium antibodies . This proportion is within the range of that reported from endemic countries of Sub-Saharan Africa using the antibody-ELISA test [25] , [26] . When CT results are used to identify NCC , the proportion of epilepsy patients with evidence of NCC was 37% , which is higher than what would have been identified based on either ELISA test . This proportion is also comparable to many reports based on neuroimaging findings from other endemic regions [25] , [27] . Only 15% of those with seizures had lesions with either a demonstrated scolex or that were described as colloidal or cystic; seizures in these patients may be more correctly classified as acute symptomatic seizures . One of the aims of this pilot study was to estimate the validity of serological testing for T . solium , as a readily available classification tool in the field , in identifying persons with NCC among those with epilepsy or recurrent seizures . When using CT-scan results as the “gold standard” , only 60% of those seropositive to the antibody to T . solium truly had NCC and nearly one-third of those truly without NCC would be wrongly identified as NCC based on serology alone . Also , more than one-third of those seronegative to the antibodies of T . solium had NCC , and would thus not have been diagnosed using this field-based criterion . Some of the relatively poor sensitivity and specificity of serology in correctly identifying persons with NCC may be a result of the antibody ELISA test used in this study , which is based on purified antigens . However , in a study conducted using 67 sera and 53 CSF samples from confirmed cases of NCC in Chile , the sensitivity and specificity of the ELISA using a purified antigen was 97% and 98 . 3% for the serum samples , respectively [28] . We have used a commercially available kit which may have slightly different accuracy values . Other antibody ELISA tests using crude extract of the metacestodes have shown some limitations . In a double blind , head-to-head comparison in the same population in which only 6% of NCC cases had calcified lesions , EITB was more sensitive than antibody-ELISA based on crude extract of the metacestode in detecting antibody to T . solium ( 86% vs . 41% respectively ) but approximately equal in specificity ( 93% v . 96% , respectively ) [13] . In populations in which the proportion of NCC with calcified or single lesions is high , both the ELISA test using crude extract of the metacestode [13] or the EITB [29] , [30] have significantly reduced sensitivity . We found that antigen testing was very poor for identifying persons with NCC , even among those with active disease . This is likely due to the natural history of NCC in which the incubation period for CNS manifestations is not clearly understood but appears to be extremely variable; symptoms can occur many years after the primary infection [31] . While the breakdown of cysts in the brain and subsequent release of parasite antigen is thought to trigger an inflammatory response which results in acute CNS manifestations in NCC [5] , antigens in the CNS may not be present at detectable levels in the serum . This would be contrary , however , to what has been demonstrated for antibodies detectable by EITB , in which samples from sera were more often positive than samples from cerebral spinal fluid [13] . Unfortunately , this pilot study was too small to examine the validity of positive serology for antigen in those with new onset seizures . False negatives , i . e . , those with NCC by CT but with negative serology , were less likely than true positives to have new onset seizures . Failure to identify these cases by antibody or antigen serology as possible NCC is compatible with the reported decline in detectable levels of antibody to T . solium over time [10] , [32] and is a potential source of significant misclassification of NCC in all prevalence studies that use serology to detect antibodies , either with ELISA or EITB , to identify NCC in persons with epilepsy [33] . In addition , the performance of antibody-detection test ( EITB ) has been shown to be reduced among NCC cases showing single , small enhancing lesions [30] . False positives ( no NCC by CT but seropositive ) likely include persons whose epilepsy is from other causes who are seropositive for T . solium because they live in an endemic area . People who are positive to the presence of antibodies may have been infected in the past or may have been exposed to the cysts and developed sufficient antibodies to prevent the cysts from establishing in the body . People who are positive to the presence of antibodies or antigens may be infected with the cyst elsewhere in their body . These groups contribute to the over-estimation of the proportion of epilepsy due to NCC when only serological testing is used to identify cases . In this study , neuroimaging by computerized tomography was used as the “gold standard” against which serology was compared; no MRI facilities were available in the study area . Misclassification of NCC based on CT would occur if the scan missed existing lesions ( false negatives ) or if lesions were misidentified as NCC ( false positives ) . If present , such misclassification in the “gold standard” would impact the apparent performance of serology . The potential for false positives was reduced by excluding those with calcifications of uncertain nature and also by assessment of the validity of serology to the subset of cases with CT evidence of cysterci . This latter group is unlikely to be misclassified . CT and MRI are equally valid for parenchymal lesions , such as those in the present study [34] , [35] , while CT is considered superior to MRI for identifying calcified lesions , which represented 59% of NCC cases in the present study [15] , [34] . Misclassification of serological results was reduced by requiring that results be definitely positive; those with intermediate results were not counted as seropositive . The data available for the present study made it difficult to separate symptomatic seizures from epilepsy . We did include as epilepsy only those who self-reported having had more than one seizure episode and clearly , the duration of seizures was very long for many individuals , but the presence of some acute symptomatic seizures cannot be ruled out . It is interesting to note that the predictive value of positive serology for antibody to T . solium was highest in those with either visible cysteci or colloidal lesions on CT , suggesting that antibody testing might be useful for identifying acute symptomatic seizures , but not epilepsy . The use of serological testing as a method for identifying cases of NCC in the field will continue to be inadequate because of the natural history of NCC . Although not fully understood because of the absence of cohort data , available cross-sectional and case-control information has helped to define some aspects of the disease which impact on the validity of any attempt to measure it in population settings . These characteristics include the evolving nature of the CNS lesions and how the location , stage and number of lesions may affect the clinical manifestations of NCC [5] , [10] , [12] , [14] , [15]; the unknown nature of the balance between intensity of infection and measurable antibody levels vs . CNS invasion; the duration of measurable antibody and antigen levels in relation to the presence of observable CNS involvement [10] , [32] . Given these limitations , it is unlikely that even improved serology will alone be sufficient for the valid identification of persons with NCC . | Epilepsy is a significant contributor to morbidity world-wide in persons of all ages . Little is known , however , about its causes . In the developing world , parasitic infections of the brain , in particular Taenia solium neurocysticercosis ( NCC ) are thought to be important factors . Determining whether or not there is infection in the brain is difficult since to be certain , specialized imaging studies , such as CT scans , are required . These are expensive and not widely available . In addition , they are not appropriate for use in large , population-based studies . Thus , blood tests for evidence of infection with T . solium are often done instead to estimate the presence of NCC . In this study's population of persons with epilepsy being seen at a hospital out-patient clinic in South Africa , 37% had CT evidence of NCC , a percentage similar to that reported in other developing countries . The study also found that blood tests were not generally useful compared to CT for correctly identifying those persons who did or did not have NCC , and thus , they cannot be relied upon for field studies of NCC . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"public",
"health",
"and",
"epidemiology/global",
"health",
"neurological",
"disorders/epilepsy"
] | 2009 | Accuracy of Serological Testing for the Diagnosis of Prevalent Neurocysticercosis in Outpatients with Epilepsy, Eastern Cape Province, South Africa |
An increasing amount of evidence indicates that developmental programs are tightly regulated by the complex interplay between signaling pathways , as well as transcriptional and epigenetic processes . Here , we have uncovered coordination between transcriptional and morphogen cues to specify Merkel cells , poorly understood skin cells that mediate light touch sensations . In murine dorsal skin , Merkel cells are part of touch domes , which are skin structures consisting of specialized keratinocytes , Merkel cells , and afferent neurons , and are located exclusively around primary hair follicles . We show that the developing primary hair follicle functions as a niche required for Merkel cell specification . We find that intraepidermal Sonic hedgehog ( Shh ) signaling , initiated by the production of Shh ligand in the developing hair follicles , is required for Merkel cell specification . The importance of Shh for Merkel cell formation is further reinforced by the fact that Shh overexpression in embryonic epidermal progenitors leads to ectopic Merkel cells . Interestingly , Shh signaling is common to primary , secondary , and tertiary hair follicles , raising the possibility that there are restrictive mechanisms that regulate Merkel cell specification exclusively around primary hair follicles . Indeed , we find that loss of Polycomb repressive complex 2 ( PRC2 ) in the epidermis results in the formation of ectopic Merkel cells that are associated with all hair types . We show that PRC2 loss expands the field of epidermal cells competent to differentiate into Merkel cells through the upregulation of key Merkel-differentiation genes , which are known PRC2 targets . Importantly , PRC2-mediated repression of the Merkel cell differentiation program requires inductive Shh signaling to form mature Merkel cells . Our study exemplifies how the interplay between epigenetic and morphogen cues regulates the complex patterning and formation of the mammalian skin structures .
The skin epithelium is an excellent model system to study mechanisms of stem cell maintenance and differentiation [1] . During skin development , a single layer of embryonic epidermal stem cells gives rise to multiple lineages , including the interfollicular epidermis ( IFE ) , the hair follicles , and the Merkel cells [1 , 2] . The precise patterning of the skin suggests that there is crosstalk between different skin epithelial lineages . However , the precise mechanisms coordinating the development of skin structures are largely unknown . Merkel cells are a subtype of mechanoreceptor cells involved in light touch sensations . Merkel cells are often organized in structures called touch domes . Touch domes are composed of Merkel cells and specialized keratinocytes , and are innervated by sensory neurons [2–8] . In humans , Merkel cell touch domes are localized in regions of high tactile acuity , either in glabrous skin or associated with hair follicles [2 , 9] . Similarly , in mice , Merkel cells are present in the glabrous epidermis of the paws , as well as in touch domes in the dorsal skin and in the outer root sheath of whisker hair follicles [2 , 9] . Much of our knowledge of the molecular mechanisms controlling Merkel cell development and homeostasis comes from the analysis of murine dorsal skin , where touch domes are organized in polarized , crescent-shaped , highly patterned structures that are located exclusively around primary hair follicles , which correspond to 1–3% of the mouse hair coat [2–8] . Previous studies have shown that the Wnt and Shh signaling pathways are essential for hair follicle development [10–21] . Mice in which Wnt signaling has been abrogated in the skin fail to develop hair follicles [15 , 16 , 22] . Wnt-dependent mesenchymal-epithelial signaling events induce hair follicle development , which is initiated by the formation of dense patches of epidermal cells , called hair placodes , and their associated dermal condensates ( Fig 1A ) . Shh ligand is expressed in the hair placodes and Shh signaling is required for hair follicle down-growth , as in Shh KO mice , hair follicles remain arrested at the hair placode stage [11 , 20 , 21 , 23 , 24] . Other signaling pathways , such as Eda/Edar , FGF , and BMP signaling , also function to regulate hair follicle morphogenesis , acting downstream of Wnt or Shh signaling [25] . The importance of these signaling pathways for Merkel cell specification is largely unknown . There are three types of hair follicles , but only primary hairs contain Merkel cells . It is unknown , however , if the invariant localization of Merkel cells is mediated by primary hair follicle-associated signals controlling the setting up of the Merkel cell lineage and , if so , what those signals are . Furthermore , it is unknown whether the signaling mechanisms controlling Merkel cell formation in hairy skin also regulate Merkel cell development in the glabrous skin . Once Merkel cells are specified , Merkel cell differentiation occurs through a maturation process characterized by the temporally regulated progressive expression of Merkel cell-specific genes , including those encoding transcription factors essential for Merkel differentiation ( Atoh1 , Sox2 , and Isl1 ) , cytoskeletal proteins ( Krt8 , Krt18 , and Krt20 ) , and components of the synaptic machinery [26] . Merkel cells and their associated neurons appear to develop independently , with Merkel cells only becoming innervated postnatally [27 , 28] . Furthermore , recent studies have shown that Polycomb repressive complex 2 ( PRC2 ) restricts the Merkel cell differentiation program by repressing the Sox2 and Isl1 genes [29 , 30] . Here , we aimed to elucidate the molecular mechanisms regulating Merkel cell development . We show that hair follicle development is essential to Merkel cell formation , as abrogation of either Wnt or Shh signaling , which disrupts hair follicle formation , prevents Merkel cell specification . We find that Shh ligand produced by developing hair follicles is required to activate Shh signaling in the epidermis and for Merkel cell formation . Moreover , Shh overexpression in epidermal progenitors leads to ectopic Merkel cells , signifying the importance of Shh signaling for Merkel cell development . Interestingly , we show that in mice lacking PRC2 in the skin epidermis , Merkel cell specification is no longer restricted to primary hair follicles , but is now associated with all hair follicle types , indicating that all hair follicle types have the potential to induce Merkel cells . PRC2 does not affect essential hair follicle signaling pathways but instead restricts Merkel cell differentiation by repressing the expression of critical Merkel cell differentiation transcription factors in epidermal progenitor cells . Importantly , PRC2-null ectopic Merkel cells fail to form in the absence of Shh signaling . Finally , we show that inductive Shh signaling and repressive PRC2 mechanisms also control Merkel cell formation in the glabrous paw skin , indicating the global role of these signaling pathways in Merkel cell specification . Together , our data show that Polycomb-mediated repression imposes a restriction on the inductive Shh signaling that is essential for Merkel cell formation . Our study demonstrates how epigenetic and cell signaling cues interact to specify cell fate in a mammalian developmental system .
In murine wild type ( WT ) dorsal skin , Merkel cells are located in touch domes and found exclusively around primary hairs [2 , 31 , 32] . The primary hairs are formed during the first wave of hair follicle development , starting at embryonic day ( E ) 14 ( Fig 1A ) [2 , 31 , 32] . Analysis of Merkel cell formation in the dorsal skin revealed that early , immature Merkel cells are first detected at E15 and are localized around follicles that are in the hair germ stage . Merkel cells are not detected in the IFE , nor are they found in hair follicles at the earlier hair placode stage ( Fig 1B ) [26] . The spatial and temporal relationship between hair follicle and Merkel cell development prompted us to hypothesize that the developing hair follicle provides a local microenvironment required for Merkel cell formation . To test this , we started by analyzing the presence of Merkel cells in mice where Wnt/β-catenin signaling was conditionally ablated in the epidermis , abrogating hair follicle development prior to the formation of the hair placodes ( S1A Fig ) [15 , 16 , 18 , 25 , 34] . To generate these mice , we crossed mice carrying the conditional null allele of β-catenin ( Ctnnb1flox ) with mice expressing Cre recombinase under the control of the Krt 14 promoter , which is active in embryonic epidermal stem cells starting at E12 ( β-cat cKO ) [29 , 35] . Immunofluorescence staining for the β-catenin protein confirmed that it was lost in the β-cat cKO epidermis ( S1B Fig ) in neonate ( P0 ) skin . Importantly , analysis of multiple Merkel cell markers revealed a complete absence of these cells in P0 β-cat cKO skin , while they were present in the control skin ( Figs 1C–1E and S1C ) . We tested whether there was an increase in apoptosis by staining for Activated Caspase 3 and by Terminal deoxynucleotidyl transferase dUTP nick end labeling ( TUNEL ) staining , which labels fragmented nuclear DNA , detecting the extensive DNA degradation that occurs in apoptotic cells . TUNEL also labels cells undergoing cornification in the suprabasal layers of the interfollicular epidermis [33] . Combining Activated Caspase 3 and TUNEL analyses , we observed no increase in apoptosis in the P0 β-cat cKO epidermis , compared to control epidermis ( Figs 1F and S1E ) . We measured proliferation by quantifying the number of cells positive for phosphorylated Histone H3 ( PH3 ) in the epidermis . No alteration in proliferation was observed in the P0 β-cat cKO epidermis , compared to control epidermis ( S1D Fig ) . Since abrogation of Wnt/β-catenin signaling abolishes Merkel cell formation , we analyzed whether Merkel cells have active Wnt signaling . To do this , we used TCF/Lef1:H2B-GFP reporter mice , in which H2B-GFP is expressed in cells with active Wnt signaling [36] , and analyzed H2B-GFP levels in Merkel cells at E16 , during the early stages of Merkel cell formation . Interestingly , we observed that the H2B-GFP fluorescence signal in Sox2 ( + ) Merkel cells was low and close to the background level , indicating that Merkel cells have low , if any active Wnt signaling ( Fig 1G ) . In contrast , and consistent with previous reports , dermal papilla cells and some Krt14 ( + ) epidermal cells have strong levels of Wnt signaling activity ( Fig 1G ) [22 , 37] . These data indicate that while Wnt/β-catenin signaling is required for Merkel cell formation , it likely functions indirectly in the control of this process . Loss of β-catenin in the epidermis abrogates a cascade of signaling pathways that are normally activated during hair growth [25] . We speculated that one of the downstream pathways functions directly to specify Merkel cells . Shh signaling is one of the downstream pathways that are activated by Wnt signaling in the epidermis [25] . We therefore investigated whether Shh signaling is required for Merkel cell specification . We started by analyzing the presence of Merkel cells in the skin of sonic hedgehog ( Shh ) -null mice ( Shh KO ) . To generate these mice , we crossed mice carrying the ShhEGFPcre allele , a null Shh allele [38] . As Shh KO mice have severe developmental defects , including an absence of fully developed limbs , and die soon after birth [21 , 39] , we performed our analysis at embryonic time points . In situ hybridization analysis of Shh mRNA confirmed the loss of Shh expression in Shh KO mice ( S2A Fig ) and , as previously reported [20 , 21 , 25 , 34] , in E18 Shh KO mice , hair follicle development is arrested at the placode stage ( S2B Fig ) . Immunofluorescence analysis of Merkel cell markers revealed a complete absence of Merkel cells in E18 Shh KO skins ( Fig 2A–2C ) . No increase in apoptosis or changes in cell proliferation were observed in Shh KO skins ( Figs 2D and S2C and S2D ) . Analysis of an earlier time point , E16 , also revealed loss of Merkel cells ( Figs 2E and S2E ) and no increase in the number of apoptotic cells ( Figs 2F and S2F ) , indicating that Merkel cells are not specified in Shh KO embryos and are not dying by apoptosis . In the skin , there are multiple sources of Shh expression , including the hair follicles , and the neurons innervating the skin [23 , 24 , 40] . Since our data show that Shh is required for Merkel cell formation , we analyzed whether the Shh ligand that is produced by the hair placodes is required for this process . To test this , we generated Shh epidermis-conditional knockout mice ( Shh cKO ) by crossing Shhflox mice with Krt14-Cre mice . In situ hybridization of Shh mRNA confirmed loss of Shh expression in P0 Shh-null hair follicles compared to control epidermis ( S2G Fig ) . Similar to those of Shh KO mice , Shh cKO hair follicles were arrested at the placode stage ( Fig 2G ) , and there were no significant alterations in apoptosis or proliferation between P0 Shh cKO and control epidermis ( Figs 2H , S2I and S2J ) . Importantly , by performing immunofluorescence analysis of multiple Merkel cell markers , we detected a complete absence of Merkel cells in P0 Shh cKO dorsal skin ( Figs 2G , 2I , 2J and S2H ) , indicating that the Shh produced by developing hair follicles is required for Merkel cell formation during development . The Shh produced by the hair placodes activates Shh signaling both in the mesenchymal dermal papilla and in the epidermis and , in both of these compartments , active Shh signaling is required for proper hair follicle formation [11 , 20 , 21 , 25] . Since the Shh produced by the developing hair follicle is essential for Merkel cell development , we tested whether Shh signaling in the epidermis is required for the specification of the Merkel cells . To address this , we generated epidermis-conditional knockout mice of Smoothened ( Smo cKO ) , an essential Shh-signaling signal transducer , by crossing Smoflox mice with Krt14-Cre mice . Consistent with previous reports , Smo-null hair follicles failed to develop properly , presenting abnormal growth and morphogenesis ( Figs 3A and S3A ) [41] . Importantly , immunofluorescence analysis of Merkel cell markers revealed a drastic reduction in the number of Merkel cells in neonatal Smo cKO skin compared to control skin ( Fig 3A–3C ) , although a few fully differentiated Merkel cells were observed in P0 Smo cKO skin ( Fig 3A–3C ) . Analysis of cell proliferation and apoptosis did not reveal significant alterations between control and Smo cKO skins at P0 ( Figs 3D , S3B and S3C ) . Analysis of Smo cKO epidermis 9 days after birth ( P9 ) revealed that hair follicle morphogenesis is grossly affected ( S3D Fig ) [41] . While there was no increase in apoptosis in these mice ( S3E Fig ) , Merkel cells were nearly completely absent from the P9 Smo cKO dorsal skin ( S3F Fig ) . This suggests that the observed phenotype in Smo cKO skin was not due to delay of Merkel cell formation . While hair follicle development is severely affected in Smo-null epidermis , the initial steps of hair follicle specification at E15 are not affected in the Smo cKO ( Fig 3E ) [41] . Importantly , no Merkel cells could be detected in the Smo cKO epidermis at this time point , when hair follicles are normally specified ( Fig 3E and 3F ) . Analysis of apoptosis did not reveal significant alterations between control and Smo cKO skins at E15 ( S3G and S3H Fig ) . Thus , we concluded that intraepidermal Shh signaling is required for Merkel cell formation . Gli1 is an effector of Shh signaling that can be used to detect Shh signaling activity in cells [11 , 42] . By analyzing β-galactosidase staining in the developing skin of Gli1-LacZ reporter mice [42] and Gli1 in situ hybridization in WT skins , we observed that Gli1 expression is strong in the developing hair follicles and is also detected in the epidermis near hair follicles , as well as in the dermis ( S3I–S3J’ Fig ) . Next , we analyzed whether Merkel cells have active Shh signaling activity by performing Gli1 in situ hybridization , followed by immunofluorescence staining for an early Merkel cell marker , Krt8 . We found that 14% of Merkel cells were Gli1 ( + ) ( Fig 3G ) , suggesting that while the Shh pathway might be active in some Merkel cells , Shh signaling activity is not detectable in most Merkel cells . As lineage tracing analysis has shown that Merkel cells arise from Krt14 ( + ) epidermal stem cells during development [43 , 44] , we analyzed Gli1 expression and found that 46% of the Krt14 ( + ) cells that were in the vicinity of the Merkel cells express Gli1 ( Fig 3G ) , while Gli1 was rarely detected in the Krt14 ( + ) cells located in the epidermis between the hair follicles ( Fig 3H ) . Merkel cells are localized not only in hairy skin but also in glabrous skin , such as paw epidermis . To test whether the intraepidermal Shh signaling is globally required for Merkel cell development , we analyzed the presence of Merkel cells in the glabrous paw epidermis of Smo cKO mice . In order to quantify the number of Merkel cells in the paws , we measured the length of the Krt14 ( + ) epidermis and counted the number of Merkel cells localized in the different regions of the paws ( S4A Fig ) . The analysis of paws of P0 Smo cKO mice revealed a very significant reduction in the number of Merkel cells compared to control ( Fig 3I and 3J ) . No increase in apoptosis was detected in the glabrous paw epidermis of Smo cKO mice ( S4B Fig ) . The fact that the number of Merkel cells was reduced to about half in the Smo cKO paws suggests that intraepidermal Shh signaling is important for Merkel specification in the glabrous skin , but that other signaling pathways are also likely to play a role in Merkel cell formation in the glabrous skin . Finally , we investigated whether epidermis-derived Shh is important for Merkel cell formation in the paw epidermis , as we observed in the hairy skin . We thus analyzed paws of P0 Shh cKO mice and found that while Merkel cells where still present , there was a very significant reduction in the number of Merkel cells in Shh cKO , compared to control paws ( Fig 3K and 3L ) . Since intraepidermal Shh signaling is required for Merkel specification , we wanted to investigate whether increased expression of Shh in the developing epidermis results in increased production of Merkel cells . Previous studies have shown that overexpression of Shh in Krt14 ( + ) epidermal progenitors during development results in the formation of Basal cell carcinoma ( BCC ) -like epidermal hyperplasia and leads to death of the newborn mice [12] . In order to overexpress Shh in the epidermis , we injected high-titer lentiviruses expressing a doxycycline inducible Shh-PGK-H2B-RFP construct into amniotic liquids of E9 Rosa26-rtTA embryos to infect the epidermis ( Fig 4A , left ) . We administered doxycycline to the pregnant females starting at E12 and collected the injected embryos at E17 ( Fig 4A , right ) . Similar to previous reports [12] , overexpression of Shh , identified by RFP expression , resulted in BCC-like epidermal hyperplasia , while the control , uninfected epidermis was not altered ( Fig 4B and 4C ) . Consistent with the observed epidermal alterations , the overexpression of Shh resulted in increased proliferation ( Fig 4F and 4F’ ) , while there was no significant increase in apoptosis ( Fig 4G and 4G’ ) . Immunofluorescence staining for Merkel cell marker Krt8 revealed that the regions with increased Shh expression had a drastic increase in the number of Krt8 ( + ) cells ( Fig 4C , 4C’ and 4E ) compared to control regions ( Fig 4B , 4B’ and 4D ) . Shh overexpression resulted in disruption of the Krt14 ( + ) layer ( Fig 4E ) , and , interestingly , the Krt8 ( + ) cells were often localized to these areas ( Fig 4E ) . To determine whether the ectopic Krt8 ( + ) cells were indeed Merkel cells , we performed immunofluorescence stainings for early Merkel cell markers Sox2 ( Fig 4H and 4I’ ) and Isl1 ( Fig 4J and 4K’ ) and found that the ectopic Krt8 ( + ) cells also expressed Sox2 ( Fig 4I and 4I’ ) and Isl1 ( Fig 4K and 4K’ ) , suggesting that overexpression of Shh results in the production of ectopic Merkel cells . Together , our data indicate that epidermal Shh signaling promotes Merkel cell specification during development . Our data show that epidermal Shh signaling is required for Merkel cell specification . Interestingly , while Merkel cells only form around the primary hairs of the murine dorsal skin , Shh signaling is not unique to primary hairs , as Shh ligand is produced by all types of hair follicles and Shh signaling is globally required for hair follicle down-growth . We therefore aimed to understand how Merkel cell formation is restricted to primary hair follicles . Our previous studies have shown that when any essential PRC2 subunits are conditionally ablated from the epidermis , there is an increase in the number of Merkel cells [29 , 30] . To further analyze the role of PRC2 in Merkel cell development , we further analyzed skin-conditional knockouts for either the histone methyltransferases Ezh1 and Ezh2 ( Ezh1/2 2KO ) or EED ( EED cKO ) . We performed whole-mount immunofluorescence analysis of the Ezh1/2-null and EED-null back skin and , surprisingly , found that the ectopic Krt8 ( + ) Merkel cells were not uniformly distributed in the knockout epidermis . Instead , Ezh1/2-null and EED-null ectopic Merkel cells were clustered , and we observed an increase in the number of these clusters in Ezh1/2-null epidermis ( Fig 5A ) . The particular patterning of the PRC2-null Merkel cell clusters ( Fig 5A ) led us to hypothesize that the ectopic Merkel cells in the PRC2-null epidermis are associated with hair follicles and , importantly , with different types of hair follicles . We have previously shown that hair formation is not affected in neonate Ezh1/2-null or EED-null skin , with the hair follicles being specified in normal numbers and with correct developmental timing ( S5A Fig ) [29 , 30 , 45] . Thus , we investigated whether PRC2-null ectopic Merkel cells are indeed associated with different types of hair follicles . While different types of hair follicles vary in structure , there are no known molecular markers to distinguish between them . Interestingly , recent studies have shown that the dermal papilla cells differ between some hair types [46–48] . For example , the transcription factor Sox2 is absent from the dermal papillae of tertiary hairs , but is present in the dermal papillae of the primary and secondary hair follicles [46 , 47] . The different hair follicle types also differ in their developmental timing: primary hairs , or first wave follicles , are the first hairs to appear in the back skin at E14 , secondary hairs are formed during the second wave of hair follicle development starting at E16 , and tertiary hair follicles are formed during the third wave , appearing just before birth ( Fig 1A ) [2 , 31 , 32] . Thus , we were able to identify the different hair types by their developmental timing , as well as co-staining for Sox2 and integrin α8 ( α8 ) , a general dermal papilla marker [47 , 48] . In newborn mice , the 1st wave hairs are Sox2 ( + ) /α8 ( + ) , contain all differentiated layers , and are the longest follicles in skin; the 2nd wave hairs are Sox2 ( + ) /α8 ( + ) , but are not yet fully formed , and are still growing downwards; the 3rd wave hairs are Sox2 ( - ) /α8 ( + ) , small , and just starting to grow downwards ( Fig 5B ) . Consistent with previous reports , we confirmed that in the control dorsal skin , Krt20 ( + ) /Sox2 ( + ) Merkel cells only appear around the primary hairs and not the other hair types ( Fig 5B ) [2] . Intriguingly , in PRC2-null epidermis , we identified Merkel cells around both 2nd wave and 3rd wave hair follicles ( Figs 5B’ and S5B ) , indicating that the ectopic Merkel cells can be found around all hair types in PRC2-null skin . We concluded that the loss of PRC2 repression results in the formation of ectopic Merkel cells that are associated with all hair follicle types . Next , we wanted to determine whether the formation of PRC2-null ectopic Merkel cells is also linked to hair follicle development , as with control Merkel cells . To do this , we analyzed Ezh1/2-null and EED-null skins at E16 , as this time point catches follicles of different waves at different developmental stages . For both control and PRC2-null skins at E16 , the first wave hair follicles were at hair peg or bulbous peg stages , the second wave follicles were at placode or hair germ stages , and the third wave follicles were yet to be formed ( Figs 1A and 5C and 5D” ) [34] . Immunofluorescence staining for early Merkel cell marker Krt8 revealed that , not surprisingly , the first wave hair follicles in both control and PRC2-null skin contained Merkel cells ( Fig 5C–5C” ) , while control second wave follicles completely lacked Merkel cells ( Fig 5D ) . Importantly , many PRC2-null second wave hair follicles contained Merkel cells ( Fig 5D’ and 5D” ) . PRC2-null Merkel cells were associated with hair follicles at the hair germ stage , and even some at the placode stage ( Fig 5D’ and 5D” ) , but they were not detected in the epidermis between hair follicles . Finally , no alterations in proliferation were observed in PRC2-null hair follicles or IFE at E16 ( S5C Fig ) . Taken together , these data indicate that loss of PRC2 results in the formation of ectopic Merkel cells that are associated with all types of hair follicles , and , similar to WT Merkel cells , the development of PRC2-null Merkel cells coincides with the early stages of hair follicle formation . As PRC2 functions as a transcriptional repressor [26 , 49 , 50] , we hypothesized that the increased Merkel cell formation observed in the PRC2-null epidermis could be due to increased Shh signaling activity , resulting in the expansion of the signaling required for Merkel cell formation . Interestingly , RT-qPCR analysis did not reveal derepression of critical components of the Shh signaling pathway or activation of its downstream target genes in neonatal PRC2-null epidermis ( Fig 6A , left ) . Shh and Gli1 in situ hybridization performed on PRC2-null skins further confirmed the RT-qPCR analysis , indicating that there was no detectable increase in Shh signaling in the developing epidermis ( Fig 6D and 6E ) . Additionally , RT-qPCR analysis for critical components of the Wnt signaling pathway and its downstream targets did not reveal derepression of these genes in the PRC2-null epidermis ( Fig 6A , right ) . Immunohistochemistry staining for β-catenin in PRC2-null skin at E16 indicated that the loss of PRC2 did not result in increased nuclear β-catenin staining in first or second wave hair follicles , or in the IFE , when compared to control skin ( Fig 6B and 6C ) . Finally , expression profiling of the PRC2-null neonate epidermis revealed that the major singling pathways involved in hair follicle development [25] are not affected in EED-null skin epithelium ( S6A Fig ) . If PRC2 does not alter the hair follicle microenvironment , how does loss of PRC2 result in ectopic Merkel cell formation ? To start addressing this question , we first analyzed how PRC2-mediated repression is regulated in Merkel cells . We performed fluorescent activated cell sorting ( FACS ) and purified Atoh1-GFP ( + ) Merkel cells and Atoh1-GFP ( - ) epidermal cells ( Fig 6F ) . We next subjected the isolated cells to chromatin immunoprecipitation assay followed by semi-quantitavie PCR ( ChIP-qPCR ) analysis to investigate the presence of PRC2-dependent repressive histone mark H3K27me3 at key Merkel cell differentiation genes , Atoh1 , Sox2 , and Isl1 [26 , 29 , 43 , 52 , 53] . The analysis revealed the presence of H3K27me3 at Atoh1 , Isl1 , and Sox2 in epidermal cells ( Fig 6G ) , where these genes are silenced ( Fig 6H ) . The level of this histone mark at the Merkel-specific genes was significantly reduced in Merkel cells ( Fig 6G ) , where Atoh1 , Isl1 , and Sox2 were expressed ( Fig 6H ) . Interestingly , the reduction in PRC2-mediated repression in Merkel cells was not global . Indeed , while the expression of PRC2 subunits Ezh2 , EED , and Suz12 was reduced in Merkel cells , compared to the epidermis ( S6B Fig ) , immunofluorescent studies revealed that at global levels , the H3K27me3 mark was not reduced in Merkel cells either during embryonic development , at E16 , ( S6C Fig ) or at birth ( Fig 6I ) . Thus , PRC2-mediated repression is relieved from the Atoh1 , Sox2 , and Isl1 genes in Merkel cells . Transcriptional profiling of EED-null neonatal epidermal cells revealed that loss of PRC2 results in derepression of Atoh1 , Sox2 , and Isl1 ( Figs 6J and S6A ) . This is in accordance with our previous data , which showed that loss of PRC2 results in derepression of Sox2 and Isl1 in the epidermis [29 , 30] . ChIP-qPCR confirmed the loss of PRC2-dependent H3K27me3 mark in the EED-null epidermis ( Fig 6K ) . Together , these data indicate that loss of PRC2-mediated repression in the epidermis leads to upregulation of key Merkel cell differentiation genes . These findings led us to hypothesize that PRC2 regulates the competent field of epidermal cells capable to differentiate into Merkel cells , but which requires inductive Shh signaling to fully differentiate into Merkel cells . To test this hypothesis , we generated EED Smo epidermis-conditional double-knockout mice ( EED Smo 2cKO ) . If PRC2 regulates the competence of the epidermis to produce Merkel cells in response to Shh-dependent inductive signals , we predicted that the EED Smo 2cKO would have a significant reduction in Merkel cell specification . Indeed , neonate EED Smo 2cKO mice exhibited a highly significant decrease in the number of Merkel cells formed when compared to control and EED cKO epidermis , as in the Smo cKO ( Fig 7A and 7B ) . There were no significant alterations in proliferation or apoptosis in the EED Smo 2cKO mice ( Figs 7C , S7A and S7B ) . Finally , we analyzed the glabrous paw epidermis of EED Smo 2cKO mice . We observed that , while there are more Merkel cells in the EED Smo 2cKO mice than in the control and Smo cKO mice , there is a highly significant reduction in the number of Merkel cells when compared to EED cKO paws , which exhibit a dramatic increase in the number of Merkel cells ( Fig 7D and 7E ) . This significant reduction in the number of Merkel cells in the EED Smo 2cKO paw epidermis was not due to increased apoptosis ( S7C Fig ) . Together , these data indicate that the loss of PRC2 leads to derepression of critical Merkel cell differentiation transcription factors and results in the expansion of the competent field of cells capable of producing Merkel cells . In the dorsal skin , ectopic Merkel cells arise only around hair follicles , which provide the local Shh signaling necessary for Merkel cell formation . While specification of Merkel cells in the glabrous paw epidermis has not yet been well characterized , intraepidermal Shh does appear to play a similar role in Merkel cell specification in this region as well ( Fig 7F ) .
In this paper , we set out to determine the molecular mechanisms controlling the development of Merkel cells , mechanoreceptor cells involved in touch sensations . Our data show that in hairy skin , the primary hair follicles function as an essential niche required for Merkel cell formation . We show that defects in hair follicle development in β-Catenin cKO and Shh KO mice result in a complete absence of Merkel cells . Furthermore , we found that intraepidermal Shh signaling , initiated by the production of Shh ligand in the developing hairs , is required for Merkel cell specification . The importance of the signals produced by one skin lineage for the specification of another elegantly demonstrates the complex interplay between the developmental processes that assures proper skin patterning . Since Shh is produced by all developing hairs and Shh signaling is associated with all hair follicle types , it is likely that there are developmentally regulated restrictive signals that function to ensure that the establishment of Merkel cells occurs solely around the primary hairs . Our data show that PRC2-mediated repression performs this function and restricts Merkel cell production to primary hairs . Indeed , loss of PRC2 in the epidermis results in ectopic Merkel cells that are localized around all hair follicle types . We show that loss of PRC2 does not alter hair follicle development [29 , 30] or the levels of Shh signaling in the skin epithelium . Instead , it leads to derepression of Sox2 , Atoh1 , and Isl1 , which encode critical Merkel cell-specific transcription factors essential for Merkel cell differentiation [26 , 29 , 43 , 52 , 53] . These data indicate that the loss of PRC2 expands the field of epidermal cells that are capable of differentiating into Merkel cells . Accordingly , concomitant loss of PRC2 and Smo significantly reduces Merkel cell formation in comparison to loss of PRC2 alone , indicating that the enhanced potential to form Merkel cells observed in the PRC2-null skin requires inductive Shh signaling . How does PRC2 restrict Merkel cell formation exclusively to primary hair follicles ? One possibility is that a primary-hair-specific signal relieves PRC2-mediated repression from the Merkel cell differentiation genes , allowing for Merkel cell formation . It has been shown that abrogation of Eda/Edar signaling or Fgf20 in the skin specifically disrupts primary hair follicle development , suggesting that Eda/Edar- and/or Fgf20-dependent pathways could play a role in this process [54–57] . Although disruption of Eda/Edar signaling has been shown to result in Merkel cell loss [28] , loss of expression of Merkel cell markers observed in Edar loss of function can be rescued by activation of Shh signaling ( personal communication , I . Brownell ) , suggesting that Eda/Edar only regulates Merkel cell formation indirectly by activating Shh signaling . Additionally , Merkel cell formation is not significantly affected in mice that lack Fgf20 ( personal communication , I . Brownell ) . Furthermore , while it has been proposed that the dermal papillae of different hair follicles are intrinsically different [58] , recent studies have suggested that the hair types are not intrinsically different and can switch in the adult , in association with changes in the number of cells in the dermal papilla [59 , 60] . Together , these data suggest that a primary-hair-specific signal is not likely to be involved in relieving PRC2-mediated repression during development . Alternatively , as the primary hair follicles are the first follicles to form , their unique association with Merkel cells might be due , instead , to temporal signals that modulate Polycomb repression during embryogenesis . As primary hair follicles form around E14 , it is possible that PRC2-mediated repression of the Merkel cell differentiation program is more promiscuous at this time point , whereas at E16 and E18 , when other hair types are formed , PRC2 repression is strong and represses the Merkel cell differentiation genes . The PRC2 complex has been shown to interact with the PRC1 complex , DNA methylation enzymes , and other repressive chromatin regulators [49 , 61] . Thus , PRC2 might recruit additional repressive complexes to Merkel cell differentiation genes at later developmental stages and lock them in the constrictively silenced state . The mechanisms controlling PRC2 activity in epidermal progenitors during development will be an exciting area of future investigation . The identified role of PRC2 in repressing a differentiation program in epidermal progenitors is somewhat surprising , as the majority of studies of somatic stem cells have shown that PRC2 promotes proliferation and survival [49 , 61 , 62] . In embryonic stem cells , PRC2 has been shown to repress key differentiation genes; however , in contrast to our findings in epidermal progenitors , PRC2-null embryonic stem cells maintain their pluripotent state and do not undergo differentiation [49 , 61 , 63] . We show that PRC2 restricts the Merkel cell program in epidermal progenitors by targeting and repressing Atoh1 , Sox2 , and Isl1 genes . Similar to what we find in epidermal progenitors , the Atoh1 gene is targeted by the PRC2-dependent H3K27me3 mark in prosensory domain progenitors , which give rise to inner ear cells . Importantly , upon differentiation of these progenitors in nascent hair cells , there is a reduction of H3K27me3 at the Atoh1 locus , which accompanies upregulation of Atoh1 expression [26] . In this regard , it will be interesting to evaluate the role for PRC2 in controlling inner ear formation and to determine whether loss of PRC2 leads to upregulation of Atoh1 expression and , possibly , ectopic hair cells . Our studies showing that Merkel cell specification requires the developing hair follicles , which provide necessary inductive signals , illustrates the essential function of the hair follicle as a niche for Merkel cell formation . Stem cell niches have long been known to play essential roles in stem cell maintenance and survival , and regulation of stem cell activity . Importantly , niche dysfunction has been associated with stem cell aging and cancer [64] . Additional signaling pathways , including BMP and FGF , amongst others , that regulate the growth of the hair follicle [64] might also promote Merkel cell formation . Crosstalk between neurons and Merkel cells should also be further investigated . While innervation of hair follicles occurs at early stages of hair development and coincides with the appearance of Merkel cells [65] , it has been shown that Merkel cells only become innervated after birth [27] . Furthermore , neuronal-derived Shh protein is not required for Merkel cell specification during development [66] and touch dome innervation occurs in the absence of Merkel cells [52] . In light of recent data showing that nerve-secreted Shh is required for the self-renewal and maintenance of adult touch domes [66 , 67] , the interdependence of neuronal and Merkel cell morphogenesis should be further investigated . While most of our studies were performed on murine dorsal skin , our findings on the importance of epidermal Shh signaling in promoting Merkel cell development , and of PRC2 in repressing Merkel cell formation hold true in glabrous skin , such as paw epidermis . Importantly , loss of Smo partially represses the expansion of Merkel cells observed in PRC2-null paws , suggesting that the two signaling events are interacting in similar ways in the glabrous to what we observe in the hairy skin . The identified general role for these molecular pathways in controlling Merkel cell formation might be important for understanding the development and homeostasis of Merkel cells in human skin , where Merkel cells are present in the interfollicular and glabrous epidermis , as well as in the outer root sheath of hair follicles [68] . Our data also show that while the epidermal Shh signaling is clearly important for the formation of Merkel cells in the paw epidermis , loss of Smo only partially reduces the number of Merkel cells in the glabrous skin , suggesting that other signaling pathways also regulate Merkel cell specification in this tissue . Finally , understanding Merkel cell development and how this process is regulated by the local microenvironment might shed light on the mechanisms of the formation of Merkel cell carcinoma , a deadly disease for which there is no effective treatment [69] . Understanding how Merkel cell formation is regulated at a molecular level can , in the future , contribute to therapeutic strategies to treat Merkel cell carcinoma .
All mice were housed and cared for according to Icahn School of Medicine at Mount Sinai ( ISMMS ) - and Institutional Animal Care and Use Committee ( IACUC ) -approved protocol LA11-0020 . Dr . Millar and Dr . Cohen from The Center for Comparative Medicine and Surgery ( CCMS ) oversee all animal care at the ISMMS . For research purposes and in cases of veterinarian-monitored illness , we use carbon dioxide in accordance with the Panel on Euthanasia of the American Veterinary Medical Association to euthanize animals . All mice were housed and cared for according to MSSM and IACUC approved protocols . Ezh1deleted and Ezh2flox mice were previously reported [45] . EEDflox mice were provided by Weipeng Mu and Terry Magnuson [70] . Tcf/Lef:H2B-GFP mice were provided by Anna-Katerina Hadjantonakis [36] . Gli1LacZ , Ctnnb1flox , Smoflox , ShhEGFPCre , Atoh1-GFP , R26-rtTA , and Krt14-Cre mice were obtained from The Jackson Laboratory . Wild type C57BL/6 mice were obtained from Charles River Laboratories . Mice were genotyped by PCR using DNA extracted from tail skin . BrdU was administered as previously reported [51] . Briefly , BrdU was administered ( 50μg BrdU per 1g mouse weight ) to mice or pregnant females 3–5 h before sacrificing [45] . For immunofluorescence , tissues were collected from mice , embedded fresh into OCT , and subsequently cut into 10μm sections using a Leica Cryostat . Embryos collected after lentiviral infection for the Shh overexpression experiment were pre-fixed for in 4% PFA 1h at RT . Slides were fixed for 10 min ( or 7 min for slides of Shh overexpression embryos ) in 4% PFA and blocked for 1h or overnight in PBS-Triton with BSA/NDS . Primary antibodies were diluted in blocking solution and incubations were carried out for 1h or overnight , followed by incubation in secondary antibodies for 1h at room temperature . Slides were then counterstained with DAPI and mounted using antifade mounting media . Whole-mount immunofluorescence was performed as previously described [26] . Briefly , back skins were collected from newborn mice and placed in Dispase for 1h at 37°C , after which the epidermal portion was peeled from the dermis and fixed in 4% PFA for 2h . Skins were blocked overnight in PBS-Triton with BSA/NDS . Primary antibodies were diluted in blocking solution and incubations were carried out for 4h at room temperature , followed by incubation in secondary antibodies for 4h at room temperature . Skins were then counterstained with DAPI and mounted in antifade mounting media for imaging . TUNEL stainings were performed using the Roche TUNEL kit ( Roche , 11684795910 ) according to the manufacturer’s instructions . β-Catenin immunohistochemistry was performed as previously described [71] . Briefly , after rehydration , paraffin-embedded tissues were subjected to antigen retrieval with 10mM citrate buffer pH 6 . 0 and then incubated overnight at 4°C with primary antibodies , using the M . O . M . kit ( Vector Laboratories ) . Secondary anti-mouse coupled to HRP ( 1:100 ) ( Vectastain ABC Kit Mouse IgG , Vector Laboratories PK-4002 ) was used for 1h at room temperature , and the staining was developed with the DAB Peroxidase Substrate Kit ( Vector Laboratories SK-4100 ) . Tissues were collected from mice , embedded fresh into OCT , and subsequently cut into 10μm sections using a Leica Cryostat . For detection of β–Galactosidase activity , slides were fixed for 10 min in 4% PFA , washed with PBS , and incubated with 1 mg/ml X-gal substrates in PBS with 100mM NaPO4 1 . 3 mM MgCl2 , 3 mM K3Fe ( CN ) 6 , and 3 mM K4Fe ( CN ) 6 overnight at 37°C . In situ hybridization for Shh and Gli1 was performed using RNAscope probes and assays , according to the manufacturer’s instructions ( Advanced Cell Diagnostics ) . Slides were imaged using a Leica DM6000 slide microscope and either 10x , 20x , or 40x objectives . Confocal microscopy was performed using a Leica SP5 DM and either 20x or 40x objectives . The quantification of Merkel cells per mm of skin was performed as described [29] . Briefly , the length of each section was measured and the number of positively stained cells was counted . Typical section lengths were between 7–14 mm . We counted a large number of Merkel cells in the controls conditions ( >300 Krt8 ( + ) cells ) and then counted the number of Merkel cells in a similar length of skin for the each mutant line . Typically , at least 100 mm of skin were counted for each condition . The quantification of Merkel cell clusters was done using Leica LAS AF software; tile scan images of the dorsal skin where acquired and the number of clusters of ≥3 Krt8 ( + ) cells per mm2 was quantified for different areas of the dorsal skin of at least 3 different individuals . Merkel cells in glabrous paw skins were quantified using Leica LAS AF software; 10x tile scan images of sections of the paws where acquired . The number of Merkel cells was quantified on the zoomed-in images and the length of the Krt14 ( + ) epidermis was measured on the large-scale tiled images , using the Leica LAS AF software and ImageJ software . PH3 ( + ) cells were quantified using a fluorescence microscope . The length of each section was measured and the number of positively stained cells was counted . Whenever it was ambiguous whether a PH3 ( + ) cells was in the epidermis , an image with Krt14 counter staining was acquired . For the Shh overexpression experiment , PH3 ( + ) cells where quantified using Leica LAS AF software; the number of PH3 ( + ) cells were quantified on images and the area of Krt14 ( + ) epidermis was measured using ImageJ software . BrdU ( + ) cells where quantified using Leica LAS AF software; nuclear DAPI staining was used to quantify the total number of cells and the % of BrdU ( + ) cells was calculated . Gli1 ( + ) cells were quantified in the same manner . Fluorescence intensity was calculated from at least three raw , single-channel greyscale images per condition using Leica LAS AF software . Staining in different skin cell populations was compared to background levels of fluorescence , which were measured as non-nuclear areas of the skin and areas outside of the skin in the same image . In all column bar graphs , mean value ± one standard deviation is presented . Box-and-whisker plots show first to third quartiles around the median , with whiskers showing 5%-95% range and outliers presented as individual data points . All quantifications were performed on multiple cell populations from different animals . To determine the significance between control and EED cKO in the ChIP-qPCR experiment in Fig 6K , an unpaired t-test was used . To determine the significance between two groups in all other experiments , the Mann-Whitney test was performed ( GraphPad Prism 5 ) . To determine the significance between more than two compared groups , the Kruskal-Wallis test was performed with the Dunn’s Multiple Comparisons post test ( GraphPad Prism 5 ) . For all statistical tests , the p<0 . 05 level of confidence was accepted for statistical significance , and actual p-values ( to four decimal places ) were provided in the figure legends . Antibodies were used as follows: Krt14 ( generous gift of Julie Segre , National Human Genome Research Institute , MD , USA , 1:20 , 000 ) ; Krt8 ( Developmental Studies Hybridoma Bank , TROMA-1 , 1:500 ) ; Krt18 ( abcam , ab668 , 1:100 ) ; Krt20 ( Dako , M7019 , 1:70 ) ; Sox2 ( Stemgent , 09–0024 , 1:150 ) ; Isl1 ( abcam , ab109517 , 1:250 ) ; GFP ( abcam , ab13970 , 1/1000 ) ; Krt5 ( generous gift of Elaine Fuchs , The Rockefeller University , NY , USA , 1:500 ) ; Phospho-histone H3 ( Upstate , 06–570 , 1/1000 ) ; BrdU ( abcam , ab1893 , 1:250 ) ; Integrin α8 ( Santa Cruz , sc-30982 , 1:100 ) ; Activated Caspase 3 ( R&D , AF835 , 1:250 ) ; β-Catenin ( BD Biosciences , 610153 , 1:100 ) ; H3K27me3 ( Millipore , 07–449 , 1:300 ) . For IF , secondary Abs coupled to Alexa 488 , 549 or 649 , were from Jackson Laboratories ( 1:1000 ) ; E-cadherin ( Invitrogen , 131900 , 1/2000 ) . For immunohistochemistry , secondary anti-mouse coupled to HRP was used ( Vectastain ABC Kit Mouse IgG , Vector Laboratories PK-4002 ) . Newborn back skins from Atoh1-GFP mice were dissected and incubated overnight in 1 . 26U/ml dispase ( Invitrogen ) at 4°C . The epidermis was gently peeled from the dermis and incubated for 15 min with 0 . 25% Trypsin , 2 . 21 mM EDTA ( CORNING ) . Epidermal cells were washed with PBS and stained with 1:400 EpCAM-APC Antibody ( BioLeged 118214 ) at room temperature for 15 min . Merkel cells were sorted at low pressure using the ARIA sorter ( BD ) as Atoh1-GFP ( + ) EpCAM ( + ) . As a control , EpCAM ( + ) GFP ( - ) epidermal cells were collected . DAPI was used as a live/dead exclusion dye . Newborn back skins were incubated overnight in 1 . 26U/ml dispase ( Invitrogen ) at 4°C . The epidermis was gently peeled from the dermis and washed with phosphate-buffered saline ( PBS ) . Epidermal cells were dissociated by treatment with 0 . 25% Trypsin , 2 . 21 mM EDTA ( Corning ) for 15 minutes , and dissociated keratinocytes were washed with PBS and lysed in RLT Plus buffer ( QIAGEN ) . FACS purified Merkel and epidermal cells were collected in RLT Plus buffer ( QIAGEN ) . RNA was purified with the RNeasy Plus mini Kit ( QIAGEN ) according to the manufacturer's instructions . For semi-quantitative analysis , reverse transcription was performed using qScript ( Quanta ) Superscript Supermix , and qPCR was performed using Roche SYBR green reagents and a Lightcycler480 machine . All primers are listed in S1 Table . RNA samples were amplified and labeled using the Low Input Quick Amp Labeling Kit ( Agilent Technologies , USA ) according to manufacturer's instructions . In brief , 100ng of total RNA from each sample was used to prepare Cyanine-3 ( Cy3 ) -labeled cRNA for hybridization . The Universal Mouse Reference RNA ( Agilent Technologies ) was Cyanine-5 ( Cy5 ) -labeled and used as an internal control , with dye swaps between samples and reference . The RNA Spike-In kit ( Agilent Technologies ) was used as an external control to monitor the microarray workflow and accuracy . For microarray hybridization , 300ng of labeled sample was fragmented and hybridized against 300ng of labeled Universal Mouse Reference RNA on SurePrint G3 Mouse GE 8X60K microarrays ( Agilent ) for 17 hours at 65°C in a rotating hybridization oven ( Agilent ) . Following hybridization , microarrays were washed with GE wash buffer 1 ( Agilent ) for 1 min at room temperature and with GE wash buffer 2 ( Agilent ) for 1 min at 37°C . Microarrays were scanned using a SureScan Microarray Scanner ( Agilent Technologies ) . The scanned images were analyzed with Agilent Feature Extraction v12 . 0 . 1 . 1 and GeneSpring v13 . 0 software ( Agilent Technologies ) . Statistical analysis was carried out using an unpaired t-test and genes with a p-value <0 . 05 and an absolute fold change ≥2 were considered significantly differentially expressed . For the construction of the heatmaps , Log2 normalized expression values generated by the GeneSpring software were used for control and EED cKO samples . Fold changes were then generated between the control and EED cKO average expression levels to demonstrate differences in the expression values between the two conditions . Known genes of signaling pathways involved in hair follicle morphogenesis and skin development ( Wnt , Shh , FGF , BMP , and Notch signaling ) [72] , as well as Merkel cell-specific genes were mined from the KEGG database . Expression values for these genes were selected from the data and used for the construction of the heatmaps . Individual heatmaps were built with the R package ggplot2 . The microarray data for this publication were deposited in NCBI’s Gene Expression Omnibus [73] and can be viewed using GEO Series accession number GSE83244 ( http://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE83244 ) . ChIP assays were performed as described [74] . Merkel cells and epidermal control cells were obtained by FACS purification and total epidermal cells were obtained from P0 back skin after dispase treatment , followed by tripsinization of the epidermis . Cells were cross-linked with 1% formaldehyde/PBS solution for 10 min at room temperature . After fixation was stopped with 125mM Glycine , nuclei were extracted in lysis buffer [50mM HEPES ( pH 7 . 5 ) , 140mM NaCl , 1mM EDTA , 10% glycerol , 0 . 5% NP-40 , 0 . 25% Triton X100] with protease inhibitors ( Complete , Roche ) . Before ChIP , nuclei were resuspended in sonication buffer [10mM Tris-HCl ( pH 8 ) , 200mM NaCl , 1mM EDTA , 0 . 5mM EGTA , 0 . 1% Na-deoxycholate , 0 . 5% N-laurylsarcosine , 1% Triton X100] , lysed , and sonicated with Bioruptor ( Diagenode , UCD-200 ) according to a 30x regimen of 30 sec . sonication followed by 30 sec . rest at 2 . 7°C to solubilize and shear cross-linked DNAs . After centrifugation , the supernatant was incubated overnight at 4°C with 40μl of Dynal Protein G magnetic beads ( Invitrogen ) , which had been pre-incubated with 4μl of the H3 or H3K27me3 , antibodies . After ChIP , samples were washed with low salt , high salt , LiCl , and Tris-EDTA buffers for 10 min at 4°C . Cross-linking was reversed by overnight incubation at 65°C , followed by RNase A and proteinase K treatment . Samples were purified with ChIP DNA Clean & Concentrator columns ( Zymo Research ) . All qPCR was performed using Roche SYBR green reagents and a Lightcycler480 machine , and the percentage of input recovery was calculated . For histone marks , the histone modifications signal was calculated based on the bulk histone signal . All primers are listed in S1 Table . Ultrasound-guided lentiviral injection procedures have been previously described [75] . In brief , R26-rtTA male mice were mated to CD1 female mice to generate R26-rtTA embryos for Shh lentiviral injection . A high-titer ( >109 CFU ) inducible Shh overexpression lentiviral construct ( LV-TRE-Shh-PGK-H2B-RFP ) [76] was used to perform microinjections into the amniotic cavities of E9 embryos . E9 timed pregnant mice were anesthetized with 2 . 5% isoflurane and 1% oxygen , and then positioned on a mouse platform ( Integrated Rail System , VisualSonics ) . A midline incision was performed , and uterine horns were gently exteriorized through the incision and carefully drawn through a ParafilmH flap in the bottom of a sterilized petri dish . Four to five embryos were injected using a micropipette , with 1 μl of Shh overexpression lentivirus for each embryo , and the uterine horn was reinserted into the peritoneal cavity . The abdominal wall and skin were closed with sutures . Expression of the viral construct was induced at E12 by gavage treatment of 200μl doxycycline ( 10 mg/ml in sterile water , Sigma-Aldrich ) to the mice that were pregnant with the injected pups . The pregnant mice were then fed doxycycline chow ( 200 mg/kg , Bio-Serv ) for 5 days . Shh o/exp and control embryos were collected at E17 and fixed with 4% PFA for 1 h . Fixed embryos were washed five times with PBS and embedded in OCT for further analysis . Microarray data have been deposited into the GEO database with accession number GSE83244 . | Merkel cells are innervated touch-receptor cells that are responsible for light touch sensations . They originate from embryonic epidermal stem cells and , in hairy regions of skin , are organized in touch domes . Touch domes are highly patterned structures that form exclusively around primary hair follicles . Strikingly , the mechanisms controlling Merkel cell formation are largely unknown . Here , we show that the hair follicle functions as a niche required for Merkel cell formation . We find that intraepidermal Sonic hedgehog ( Shh ) signaling , initiated by the production of Shh in the developing hair follicles , is required for Merkel cell specification , whereas Shh overexpression in embryonic epidermal progenitors leads to ectopic Merkel cells . Interestingly , Shh signaling is common to all hair types , suggesting that there are restrictive mechanisms that allow Merkel cell specification to occur exclusively around primary hairs . Indeed , we find that loss of Polycomb repressive complex 2 ( PRC2 ) in the epidermis leads to the formation of ectopic Merkel cells around all hair types . We show that PRC2 loss expands the field of epidermal cells competent to differentiate into Merkel cells through derepression of key Merkel-differentiation genes; however , inductive Shh signaling is still required for the formation of mature Merkel cells . Our study illustrates how the interplay between epigenetic and morphogen cues functions to establish the complex patterning and formation of the mammalian skin . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"skin",
"cell",
"death",
"merkel",
"cells",
"medicine",
"and",
"health",
"sciences",
"integumentary",
"system",
"cell",
"processes",
"cell",
"differentiation",
"epithelial",
"cells",
"developmental",
"biology",
"epidermis",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"hair",
"animal",
"cells",
"biological",
"tissue",
"hair",
"follicles",
"signal",
"transduction",
"cell",
"staining",
"cell",
"biology",
"anatomy",
"apoptosis",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"wnt",
"signaling",
"cascade",
"cell",
"signaling",
"signaling",
"cascades"
] | 2016 | Polycomb-Mediated Repression and Sonic Hedgehog Signaling Interact to Regulate Merkel Cell Specification during Skin Development |
West Nile virus ( WNV ) is a highly pathogenic flavivirus transmitted by Culex spp . mosquitoes . In North America ( NA ) , lineage 1 WNV caused the largest outbreak of neuroinvasive disease to date , while a novel pathogenic lineage 2 strain circulates in southern Europe . To estimate WNV lineage 2 epidemic potential it is paramount to know if mosquitoes from currently WNV-free areas can support further spread of this epidemic . We assessed WNV vector competence of Culex pipiens mosquitoes originating from north-western Europe ( NWE ) in direct comparison with those from NA . We exposed mosquitoes to infectious blood meals of lineage 1 or 2 WNV and determined the infection and transmission rates . We explored reasons for vector competence differences by comparing intrathoracic injection versus blood meal infection , and we investigated the influence of temperature . We found that NWE mosquitoes are highly competent for both WNV lineages , with transmission rates up to 25% . Compared to NA mosquitoes , transmission rates for lineage 2 WNV were significantly elevated in NWE mosquitoes due to better virus dissemination from the midgut and a shorter extrinsic incubation time . WNV infection rates further increased with temperature increase . Our study provides experimental evidence to indicate markedly different risk levels between both continents for lineage 2 WNV transmission and suggests a degree of genotype-genotype specificity in the interaction between virus and vector . Our experiments with varying temperatures explain the current localized WNV activity in southern Europe , yet imply further epidemic spread throughout NWE during periods with favourable climatic conditions . This emphasizes the need for intensified surveillance of virus activity in current WNV disease-free regions and warrants increased awareness in clinics throughout Europe .
West Nile virus ( WNV; family Flaviviridae , genus Flavivirus ) is an important mosquito-borne human pathogen associated with febrile illness , which may develop into severe neuroinvasive disease and death [1] . The pathogenic isolates of WNV can be classified into two lineages . Lineage 1 WNV strains have long been endemic in Africa , Australia , the Middle East , Asia and southern Europe [2 , 3] . In the 1990s , lineage 1 WNV re-emerged in southern Europe and the Middle-East [4–6] . In 1999 , lineage 1 WNV was unintentionally introduced into New York City from where it spread rapidly across the United States where it is now endemic [7] . With an accumulated 17 , 463 cases of neuroinvasive disease and 1 , 668 reported deaths between 1999 and 2013 , this outbreak quickly evolved into the largest outbreak of neuroinvasive disease to date [8] . Lineage 2 WNV strains have been endemic in sub-Saharan Africa and Madagascar and were previously considered to be of low pathogenicity [2 , 3] . In 2010 , a highly pathogenic lineage 2 WNV isolate caused a large outbreak in Greece [9] , which resulted in 262 cases of human disease and 35 deaths . Lineage 2 WNV then quickly became endemic in South-East Europe and with annual outbreaks to date WNV disease in the region has increased seven-fold [10–12] . At present , WNV disease does not extend into north-western Europe ( NWE ) [11] . During enzootic transmission , WNV circulates primarily between mosquitoes of the Culex genus and birds . Many avian species in North America ( NA ) [13] and Europe [14 , 15] are suitable reservoirs/amplifying hosts and can produce high viral titres upon WNV infection . Infected mosquitoes also blood feed on other vertebrate hosts , which leads to frequent infections in humans and horses [16] . In Europe , the main Culex species found positive during WNV surveys is the common house mosquito Culex pipiens [17] . In NA the most prevalent and effective vector species for WNV are Culex pipiens , Culex tarsalis and Culex quinquefasciatus [18 , 19] . Laboratory experiments show that NA Culex pipiens mosquitoes are competent vectors for NA isolates of lineage 1 WNV [20] . The vector competence of European mosquitoes to lineage 1 WNV has not been intensively studied nor has it been compared directly to competent vectors from NA [21] . The vector competence of NA and European mosquitoes for transmission of novel European lineage 2 WNV isolates has not yet been determined , but this is of high importance now that a highly pathogenic lineage 2 WNV has emerged in Europe , which appears to be as neuroinvasive as WNV isolates from lineage 1 [9] . As the global activity of these pathogenic WNV lineages has significantly increased over the past two decades , we set out to assess the potential for virus transmission in areas that are currently free of lineage 1 and/or lineage 2 WNV strains . The results show that European mosquitoes from an area free of WNV disease have the intrinsic capability to transmit both lineage 1 and lineage 2 WNV . However , comparing transmission rates at varying temperatures provides evidence that the differences in climatic conditions currently restrict the spread of WNV throughout Europe .
C6/36 ( ATCC CRL-1660 ) and Culex tarsalis cells ( CDC , Fort Collins , CO ) were grown on Leibovitz L15 and Schneiders ( Gibco ) medium supplemented with 10% fetal bovine serum ( FBS; Gibco ) . Hela ( ATCC CCL-2 ) , DF-1 ( ATCC CRL-12203 ) and Vero E6 ( ATCC CRL-1586 ) cells were cultured with DMEM Hepes ( Gibco , Bleiswijk , The Netherlands ) buffered medium supplemented with 10% FBS containing penicillin ( 100IU/ml ) and streptomycin ( 100μg/ml ) . When Vero E6 cells were incubated with mosquito lysates or saliva the growth medium was supplemented with fungizone ( 2 , 5μg/ml ) and gentamycin ( 50μg/ml ) . P2 virus stocks of the NY’99 and Gr’10 isolates were grown on C6/36 cells and titrated on Vero E6 cells . The NWE Culex pipiens colony originated from Brummen , The Netherlands ( °05'23 . 2"N 6°09'20 . 1"E ) and was established in 2010 and maintained at 23°C . The NA Culex pipiens colony [20] was maintained at 26°C . Both mosquito colonies were kept in Bugdorm cages with a 16:8 light:dark ( L:D ) cycle and 60% relative humidity ( RH ) and were provided with 6% glucose solution . Bovine or chicken whole blood ( KemperKip , Uden , The Netherlands ) was provided through the Hemotek PS5 ( Discovery Workshops ) for egg production . Egg rafts were allowed to hatch in tap water supplemented with Liquifry No . 1 ( Interpet Ltd . , Dorking , UK ) . Larvae were fed with a 1:1:1 mixture of bovine liver powder ( Sigma-Aldrich , Zwijndrecht , The Netherlands ) , ground rabbit food and ground koi food . 2–5 day old mosquitoes were infected either via ingestion of an infectious blood meal or via intrathoracic injections . Infectious blood meals: Whole chicken blood was mixed with the respective P2 virus stock to a final concentration of 1 . 4*108 WNV infectious particles per ml . Mosquitoes were allowed to membrane feed , using the Hemotek system and a parafilm membrane , in a dark climate controlled room ( 24°C , 70% RH ) . After 1 hour , mosquitoes were sedated with 100% CO2 and the fully engorged females were selected . Injections: Mosquitoes were sedated with CO2 and placed on a semi-permeable pad , attached to 100% CO2 . Mosquitoes were infected by intrathoracic injection using the Drummond nanoject 2 ( Drummond scientific company , United States ) . Infected mosquitoes were incubated at their respective temperatures with a 16:8 L:D cycle and fed with 6% sugar water during the course of the experiment . Legs and wings of sedated mosquitoes were removed and their proboscis was inserted into a 200ul filter tip containing 5 ul of salivation medium ( 50% FBS and 50% sugar water ( glucose , W/V 50% ) ) . Mosquitoes were allowed to salivate for 45 minutes . Mosquito bodies were frozen in individual Eppendorf tubes containing 0 . 5 mm zirconium beads ( Next Advance , New York , USA ) at -80°C . The mixture containing the saliva was added to 55 ul of fully supplemented growth medium . Frozen mosquito bodies were homogenized in the bullet blender storm ( Next Advance New York , USA ) in 100 μl of fully supplemented medium and centrifuged for 90 s at 14000 rpm in a table top centrifuge . 30ul of the supernatant from the mosquito homogenate or the saliva containing mixture was incubated on a monolayer of Vero cells in a 96-wells plate . After 2–4 hours the medium was replaced by 100 μl of fresh fully supplemented medium . Wells were scored for WNV-specific cytopathic effects ( CPE ) , confirmed with immunofluorescence assay ( IFA ) against WNV E [22] at three days post infection ( dpi ) . WNV titres were determined using 10μl of the supernatant from the mosquito homogenate in end point dilution assays on Vero E6 cells . WNV infection was scored by CPE , confirmed with IFA at three dpi . Maps displaying the mean diurnal temperature during July and August of the indicated year [23] . Human cases of WNV in Europe , during 2011 , 2012 or 2013 were projected on the location where they were reported [11] . To eliminate potential imported cases , WNV cases were only considered when a country reported more than one case for that year . WNV infections in mosquito bodies and saliva were scored positive or negative and significant differences were calculated using the Fisher’s exact test ( P<0 . 05 ) . Differences in WNV titres ( TCID50/ml ) in infected mosquito bodies and heads were calculated using the Mann Whitney test ( P<0 . 05 ) .
Culex pipiens mosquitoes from NWE ( The Netherlands ) , and a NA Culex pipiens colony [20] were infected with either the novel pathogenic lineage 2 WNV isolate ( WNV-lin2 ) from Greece’10 or lineage 1 isolate ( WNV-lin1 ) , New York ‘99 . The vector competence of NA mosquitoes for WNV-lin1 has been well-described [20] and serves as a reference for the infection and transmission rates of WNV . The WNV-lin2 and WNV-lin1 isolates displayed similar growth kinetics in human , avian and mosquito cell cultures ( Fig 1A–1C ) . Infectious blood meals containing 1 . 4*108 TCID50/ml of either WNV-lin2 or WNV-lin1 isolates were fed to the NWE and NA Culex pipiens mosquitoes . Fully engorged females were selected and kept at an ambient temperature of 23°C . Immediately after completion of the blood meal , a subset of fully engorged females was tested for the presence of infectious WNV to confirm that both mosquito populations had ingested equal amounts of infectious virus particles ( Fig 2A ) . Infection with either WNV isolate did not influence mosquito survival during the course of the experiments ( Fig 2B ) . After 14 days , saliva was collected from a large and random subset of the mosquitoes . Both the isolated saliva as well as all the mosquito bodies were examined for the presence of WNV ( schematic representation of the experiment , Fig 3A ) . The combined results from five independent experiments are summarized in Table 1 ( infection rates , bodies ) and 2 ( transmission rates , saliva ) . Both the NWE and NA mosquitoes were equally susceptible to infection , but with significant differences in the infection rates between the WNV-lin2 and WNV-lin1 isolates ( Table 1 , P<0 . 05 ) . Dissemination of WNV into the saliva of a vector is a prerequisite for successful transmission . After consuming a blood meal that contained WNV-lin1 , 22% and 19% of respectively the NWE and NA mosquitoes had detectable levels of WNV in their saliva ( Fig 3B ) . In contrast , the WNV-lin2 isolate was detectable in the saliva of 24% of the NWE mosquitoes , but only in 8% of the NA mosquitoes ( Fig 3B , P<0 . 05 ) indicative of a strongly reduced susceptibility of the latter for WNV-lin2 . The observed differences in transmission rates between WNV-lin2 infected NWE and NA mosquitoes are underscored when only the percentage of WNV-infected mosquitoes are considered . From the population of WNV-infected mosquitoes , successful replication and dissemination of WNV-lin2 into the saliva was found in 59% of the NWE , compared to only 24% of the NA mosquitoes ( Table 2 ) . In an effort to understand these differences in transmissibility we first determined the tissue culture infectious dose of WNV present in each positive mosquito body by end point dilution assays . The viral titres in individual mosquito bodies were highly variable and could reach up to 109 TCID50/ml for WNV-lin2 infected NWE mosquitoes , compared to a three logs lower maximum titre of only 106 TCID50/ml in NA mosquitoes ( Fig 3C ) . Comparison between the WNV titres of saliva-positive and saliva-negative mosquitoes within the same sample population showed that significantly more infectious WNV particles were present in the bodies of mosquitoes with positive saliva than with negative saliva ( Fig 3C ) . This indicates that the level of WNV replication in the mosquito body determines the dissemination into the salivary glands . Efficient infection and escape from the midgut epithelial cells is necessary for dissemination of the virus to other tissues , including the salivary glands [24–26] . When the midgut was circumvented by injecting WNV-lin2 directly into the thorax , all mosquitoes from both NWE and NA became readily infected ( Fig 4A , open symbols ) and up to 100% of injected individuals were able to transmit WNV at day eight post injection ( Fig 4B , open symbols ) . In contrast , infectious blood meals resulted in differential proportions of the NWE and NA mosquitoes being able to transmit WNV-lin2 ( Fig 4B , closed symbols ) , again with NWE as a more competent vector . Strikingly , both eight and eleven days post infection , the WNV-lin2 isolate was detected in the saliva of 14% of NWE mosquitoes , compared to <3% of NA mosquitoes ( Fig 4B , closed symbols , P = 0 . 1076 and P = 0 . 0259 respectively ) . Thus , oral infection with the WNV-lin2 isolate results in better dissemination and a shorter mean extrinsic incubation period , suggesting that WNV-lin2 escapes more effectively from the midgut epithelial cells in mosquitoes from NWE compared to those from NA . Taken together , transmission of both WNV lineages is intrinsically possible in NWE whereas there is no evidence to suggest that WNV-lin2 can utilize NA mosquitoes as effective vectors due to limited dissemination to the salivary glands . As the mosquito colonies used in this laboratory study are representatives of their respective populations from the described areas , the experiments presented here show that highly WNV-competent Culex pipiens mosquitoes are present in NWE . Vector competence is , however , not only attributed to intrinsic factors , but also subjective to extrinsic factors , most notably the ambient temperature [20 , 27] . Because indigenous WNV activity is currently absent in NWE [11 , 28] , but competent European bird species are present [14 , 15] , we hypothesized that temperature limits the vector competence of European mosquitoes for WNV transmission . To test this hypothesis , we infected both NWE and NA mosquitoes with the WNV-lin2 isolate via a WNV-containing blood meal and incubated the mosquitoes at three different temperatures for 14 days post oral infections . The first temperature represented the average summer temperatures in large parts of NWE , including the origin of our NWE mosquito colony ( The Netherlands; 18°C ) . The second temperature was an intermediate temperature ( 23°C ) , while the third temperature matched the average summer temperature of the area where WNV-lin2 was isolated ( Greece; 28°C ) [29] . The warmest period of the year ( July and August ) also corresponded with the peak in WNV amplification and transmission [8] . Higher temperatures significantly increased the percentage of WNV-infected mosquito vectors , with no apparent difference between NWE and NA mosquitoes ( Fig 5 , P<0 . 05 ) . At 18°C , 17% ( n = 29 ) and 19% ( n = 41 ) of mosquitoes were infected with WNV-lin2 , whereas incubation at 28°C increased the infection rates to 58% ( n = 36 ) and 52% ( n = 25 ) for NWE and NA mosquitoes , respectively . Comparison between the spatial arrangement of recent WNV outbreaks in Europe per annum and the corresponding mean temperature during peak transmission season strengthens this hypothesis by displaying a strong correlation between WNV outbreaks and the mean diurnal summer temperature throughout Europe ( Fig 6A and 6B and 6C ) . The mean temperatures at which WNV outbreaks occurred in 2011 , 2012 and 2013 were 24 . 6°C , 25 . 3°C , and 23 . 5°C , with standard deviations of 2 . 4°C , 2 . 7°C , and 2 . 1°C , respectively ( Fig 6D ) . Together , the mean temperatures at the respective locations of individual outbreaks give an indication of the average summer temperatures at which there is an elevated risk for WNV activity .
Here we show that both pathogenic lineages of WNV can effectively infect mosquitoes from NWE . Our finding that two geographically separated Culex pipiens populations ( NWE and NA ) have a markedly different vector competence for WNV-lin2 , suggests a degree of genotype-genotype specificity in the interaction between virus and vector . Alternatively , the presence of certain endosymbionts or insect-specific flaviviruses can have an effect on the vector competence as well [30 , 31] . As the differential infection rate is only apparent when WNV is infected orally and not via intrathoracic injections , this suggests that WNV-lin2 escapes more effectively from the midgut epithelial cells in mosquitoes from NWE compared to those from NA . The presence of highly competent vector species in WNV disease-free areas suggests that extrinsic factors such as temperature play an essential role in the current distribution of WNV . In NWE , the lower average summer temperature ( <20°C ) may provide a possible explanation for the current WNV epidemics , which remain restricted to southern Europe . However , other extrinsic factors can shape the vectorial capacity and may compensate for a reduced vector competence at low temperature by facilitating larger mosquito populations . The recent resurgence of WNV disease in the United States was most likely fuelled by climatic conditions that were favourable for local vector populations [32 , 33] . In addition , hybrids between two closely related Culex pipiens forms may increase the incidence of human WNV disease , as these ‘bridge-vectors’ are considered less ornithophilic and more likely to feed on other vertebrates , including humans . These hybrids are relatively common in North America , but not in north-western Europe [34] . How effective different European Culex pipiens populations are in transmitting WNV is currently unknown , but this can be investigated by using wild caught Culex pipiens mosquitoes from a variety of sources and regions . Additionally , viral adaptations that increase the replication efficiency at lower temperatures could further facilitate and enhance transmission of WNV throughout Europe . Indeed , the WNV-lin1 isolate has already proven to be able to adapt to the different climatic conditions in the Americas [20] , while other flaviviruses , including a close relative of WNV , Usutu virus , are already endemic in parts of NWE [35] . Finally , travel and trade continuously ( re- ) introduce WNV to areas free of overt WNV-disease . In addition , WNV transmission in the absence of noticeable disease has been suggested based upon serological surveys in ( sentinel ) birds [36 , 37] , which may spark a WNV epidemic in NWE . The presence of vectors that are intrinsically capable of transmitting WNV increases the chances for novel outbreaks of WNV-disease , especially when global warming or temporary weather extremes will favour the vectorial capacity of Culex pipiens . Based on our results and experimental evidence by others that European birds are suitable amplifying hosts [14] , we propose that WNV surveillance in mosquitoes and birds should be intensified , especially in areas where climatic conditions are more favourable , to allow early detection and the implementation of effective mitigation and intervention strategies . Furthermore , awareness by clinicians throughout Europe is warranted in order to more effectively diagnose cases of human WNV ( neurological ) disease . | West Nile virus ( WNV ) is on the rise in Europe , with increasing numbers of human cases of neurological disease and death since 2010 . However , it is currently unknown whether or not WNV will continue to spread to north-western Europe ( NWE ) , in a fashion similar to the WNV epidemic sweep in the United States ( 1999–2004 ) . The presence of competent mosquitoes is a strict requirement for WNV transmission , but no laboratory studies have been conducted with the new European lineage 2 WNV outbreak strain . Our study is the first to investigate transmissibility in NWE Culex pipiens for lineage 2 WNV in a systematic , direct comparison with North American Culex pipiens and with the lineage 1 WNV strain . We demonstrate that European mosquitoes are highly competent for both WNV lineages , which underscores the epidemic potential of WNV in Europe . However , the transmission rate for lineage 2 WNV was significantly lower in North American mosquitoes , which indicates different risk levels between both continents for lineage 2 but not lineage 1 WNV . Based on our result , we propose that WNV surveillance in mosquitoes and birds must be intensified in Europe to allow early detection , timely intervention strategies and prevent outbreaks of WNV neurological disease . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | West Nile Virus: High Transmission Rate in North-Western European Mosquitoes Indicates Its Epidemic Potential and Warrants Increased Surveillance |
The type VI secretion system ( T6SS ) is a widespread protein secretion apparatus used by Gram-negative bacteria to deliver toxic effector proteins into adjacent bacterial or host cells . Here , we uncovered a role in interbacterial competition for the two T6SSs encoded by the marine pathogen Vibrio alginolyticus . Using comparative proteomics and genetics , we identified their effector repertoires . In addition to the previously described effector V12G01_02265 , we identified three new effectors secreted by T6SS1 , indicating that the T6SS1 secretes at least four antibacterial effectors , of which three are members of the MIX-effector class . We also showed that the T6SS2 secretes at least three antibacterial effectors . Our findings revealed that many MIX-effectors belonging to clan V are “orphan” effectors that neighbor mobile elements and are shared between marine bacteria via horizontal gene transfer . We demonstrated that a MIX V-effector from V . alginolyticus is a functional T6SS effector when ectopically expressed in another Vibrio species . We propose that mobile MIX V-effectors serve as an environmental reservoir of T6SS effectors that are shared and used to diversify antibacterial toxin repertoires in marine bacteria , resulting in enhanced competitive fitness .
The type VI secretion system ( T6SS ) is a protein secretion apparatus found in Gram-negative bacteria [1] . While it was originally described as a bacterial virulence determinant [2–4] , subsequent findings demonstrated that many T6SSs are used as antibacterial determinants in interbacterial competition [5–8] . This tightly regulated macromolecular secretion apparatus functions similarly to a contractile phage tail but in a reverse orientation [1] . Upon perception of an extracellular signal , the secreted tail tube complex , composed of an inner tube made of stacked hexameric rings of Hcp that are capped by a trimer of VgrG and a PAAR repeat-containing protein , is propelled outside of the cell and into an adjacent recipient cell [1 , 9 , 10] . This tail tube is decorated with effector proteins containing toxic activities , either as domains fused to components of the tail tube or as proteins that bind to them [11] . Several T6SS effectors have been identified and found to cause toxicity through various mechanisms such as actin cross-linking [3] , nuclease activity [12 , 13] , and pore-forming [14] . In addition , two effector superfamilies with antibacterial peptidoglycan-hydrolase and phospholipase activities have been described [6 , 7] . Several proteins containing Rearrangement hotspot ( Rhs ) repeats were also suggested to be T6SS effectors [12 , 15] . We recently identified a widespread class of polymorphic T6SS effectors called MIX-effectors [16] . These effectors share an N-terminal motif named MIX ( Marker for type sIX effectors ) and have polymorphic C-terminal domains with diverse predicted antibacterial or anti-eukaryotic activities [16] . Notably , T6SS effectors that possess antibacterial activities are encoded in bicistronic units together with a gene that encodes for their cognate immunity protein that protects the cell against self-intoxication [6 , 7] . Up to six T6SSs can be encoded within a single bacterial genome [17] , and each system can be differentially regulated [18–21] . Vibrio alginolyticus , a Gram-negative , halophilic marine pathogen associated with wound infections , otitis and gastroenteritis , is one of the most commonly reported disease-causing Vibrio species in the United States [22] , and was also recently found to be a cause of coral diseases [23 , 24] . It encodes two T6SSs ( VaT6SS1 and VaT6SS2 ) [25] . Sheng et al . previously reported several transcription factors and regulators that control the activation of V . alginolyticus T6SS1 ( VaT6SS1 ) in the EPGS strain [25 , 26] . More recently , we found that VaT6SS1 of the V . alginolyticus 12G01 strain functions as an antibacterial determinant , and identified a MIX-effector , V12G01_02265 ( hereafter we will use the prefix Va instead of the locus prefix V12G01_ , thus the aforementioned protein is Va02265 ) , that mediated antibacterial toxicity and is paired with an immunity protein , Va02260 [16] . In a previous study , we characterized the environmental conditions and cues that activate the two T6SSs found in the marine pathogen V . parahaemolyticus [20 , 27] , and identified secreted effectors that mediate the antibacterial activity of the V . parahaemolyticus T6SS1 ( VpT6SS1 ) [16] . However , we found no role for VpT6SS2 [20] . The two V . alginolyticus T6SS gene clusters , encoding VaT6SS1 and VaT6SS2 ( S1A Fig ) , are similar to the V . parahaemolyticus T6SS clusters in both gene content and organization [20] . However , the environmental conditions that activate the V . alginolyticus T6SSs and whether they differ from the conditions that regulate the T6SSs in V . parahaemolyticus , the activity of VaT6SS2 , and the V . alginolyticus T6SSs effector repertoires , remain unknown . In this work , we set out to characterize the T6SSs in V . alginolyticus . We found that the V . alginolyticus T6SSs are differentially regulated by salinity and temperature , and that both systems can mediate bacterial killing during interbacterial competition . Using comparative proteomics , we identified several T6SS effectors , including MIX-effectors , that mediate antibacterial killing . Finally , we found a subset of mobile T6SS MIX-effectors that are shared between marine bacteria via horizontal gene transfer , and showed that such a mobile MIX-effector from V . alginolyticus can be transferred into V . parahaemolyiticus and retain the toxic activity as a secreted T6SS effector . These results indicate that a subset of MIX-effectors are found on mobile genetic elements and can be horizontally transferred between bacteria .
Upon analysis of its genomic sequences , the V . alginolyticus 12G01 strain was found to have two T6SSs that are similar to those previously reported for V . alginolyticus EPGS [25] and for V . parahaemolyticus RIMD 2210633 [20] ( S1A Fig ) . VaT6SS1 contains two putative transcriptional regulators , Va01475 and Va01550 , which are homologs of the V . parahaemolyticus T6SS1 positive regulators VP1407 and VP1391 , respectively [20] . A homolog of the V . parahaemolyticus MIX-effector VP1388 , Va01565 , is found at the beginning of the VaT6SS1 gene cluster ( S1A Fig ) [16] . The VaT6SS2 gene cluster does not appear to encode any effectors or transcriptional regulators ( S1A Fig ) . As a preliminary step to characterize the T6SSs of V . alginolyticus , we first generated V . alginolyticus 12G01 derivative strains in which the T6SSs were inactivated by deletions in the genes encoding the necessary inner tube component Hcp of VaT6SS1 ( Δhcp1 ) , VaT6SS2 ( Δhcp2 ) , or both systems ( Δhcp1/Δhcp2 ) . Next , we tested whether inactivation of the VaT6SSs affected growth . To this end , we monitored the growth of the wild-type and Δhcp strains in MLB media at 30°C by measuring the OD600 of the cultures over time . No difference in growth was detected ( S1B Fig ) , indicating that the T6SSs do not affect V . alginolyticus growth . We previously reported that VaT6SS1 mediates bacterial killing on LB agar plates at 30°C [16] . As V . alginolyticus is a marine bacterium that thrives during warm months under various conditions in the environment and the host [22 , 28] , we set out to determine how environmental conditions such as salinity and temperature affect the activity of VaT6SS1 . To this end , we monitored the viability of E . coli before and after co-culture with wild-type V . alginolyticus , a Δhcp1 derivative in which VaT6SS1 is inactive , or alone on LB or MLB plates ( containing 1% and 3% NaCl , respectively ) , at 30°C or 37°C . When co-cultured at 37°C , V . alginolyticus was unable to kill E . coli on LB plates ( Fig 1A ) , but it was able to kill E . coli on MLB plates ( Fig 1B ) . Surprisingly , whereas deletion of hcp1 largely abrogated the antibacterial toxicity of V . alginolyticus on LB at 30°C , it had only a marginal effect when co-cultures were grown on MLB plates at 30°C ( Fig 1 ) . This result suggested that there is another antibacterial determinant other than VaT6SS1 that can mediate interbacterial competition under high salt conditions . We hypothesized that VaT6SS2 can also mediate antibacterial toxicity . To test our hypothesis , we repeated the E . coli competition assays with a Δhcp2 derivative in which VaT6SS2 is inactive and with a Δhcp1/Δhcp2 derivative in which both VaT6SS1 and VaT6SS2 are inactive . Whereas deletion of hcp2 had only a marginal effect on the antibacterial toxicity of V . alginolyticus on LB at 30°C , it had a considerable effect on MLB at 30°C ( Fig 1 ) . Consistent with these observations , the Δhcp1/Δhcp2 derivative had no antibacterial toxicity under the tested conditions ( Fig 1 ) . These results indicated that both V . alginolyticus T6SSs possess antibacterial activities , yet they are active under different salinity and temperature conditions . VaT6SS1 is more active under low salt conditions ( on LB plates ) , whereas VaT6SS2 is more active under high salt conditions ( on MLB plates ) . Both T6SSs are active at 30°C , but only VaT6SS2 is also active at 37°C ( Fig 1 ) . After uncovering a role for VaT6SS2 in interbacterial competition , we next sought to identify the secreted effectors that mediate this antibacterial activity . To this end , we used comparative proteomics to find proteins that are secreted in a VaT6SS2-dependent manner . We used mass spectrometry ( MS ) to analyze the secretomes of V . alginolyticus Δhcp1 ( with an active VaT6SS2 ) and Δhcp1/Δhcp2 ( with an inactive VaT6SS2 ) strains grown under VaT6SS2-inducing conditions ( i . e . high salt media at 30°C ) ( see S1 Dataset ) . The strains used were deleted for hcp1 to detect only proteins secreted by VaT6SS2 . We identified 7 proteins that were differentially found in the supernatant of the Δhcp1 strain in which VaT6SS2 was active ( Table 1 ) . Va07588 is the VaT6SS2 tail tube secreted component VgrG2 and served to validate our VaT6SS2 secretome analysis ( Table 1 ) . We predicted that three of the VaT6SS2 secreted proteins: Va16922 , Va18287 , and Va03175 , were antibacterial effectors as they possess predicted nuclease ( Va16922 and Va18287 ) or pore-forming colicin-like ( Va03175 ) domains ( according to HHPred analysis [29] ) that can mediate antibacterial toxicity . Moreover , the genes encoding for these three proteins were immediately upstream of small open reading frames ( ORFs ) that could encode for their cognate immunity proteins ( Va16927 , Va18282 , and Va03180 , respectively ) . These three putative effector/immunity pairs were encoded outside of the VaT6SS2 gene cluster ( Fig 2A ) . To test whether Va16922/7 , Va18287/2 , and Va03175/80 are VaT6SS2 effector/immunity pairs , we monitored the ability of a V . alginolyticus wild-type strain , which is immune against self-intoxication , to kill strains with deletions in the putative effector/immunity gene pairs . Indeed , the wild-type strain was able to kill strains with deletions in va16922-7 and va18287-2 when co-cultured under VaT6SS2 inducing conditions ( Fig 2B and 2C ) , indicating that deletion of these bicistronic units resulted in loss of immunity against self-intoxication . However , a strain deleted for va03175-80 was still immune against self-intoxication ( Fig 2D ) , suggesting that either Va03175/80 are not an effector/immunity pair , or that there is an additional immunity gene . Using Va03180 as template , we performed a BLAST search to look for possible redundant immunity proteins in V . alginolyticus 12G01 . We identified Va03170 , encoded by the gene immediately upstream of the putative effector Va03175 , as a homolog of Va03180 ( 65% identity ) . Therefore , we generated a strain deleted for the putative effector and the two homologous putative immunity genes , Δva03170-80 . As predicted , the wild-type strain was able to kill the Δva03170-80 strain when co-cultured under VaT6SS2 inducing conditions ( Fig 2D ) , indicating that va03170 and va03180 encode for redundant immunity proteins against Va03175-medited toxicity . In all cases , inactivation of VaT6SS2 by deletion of hcp2 , or deletion of the effector/immunity pair in the attacking strain , resulted in loss of self-intoxication indicating that the toxic effectors were delivered by VaT6SS2 and encoded within these bicistronic units . Moreover , exogenous expression of the putative immunity proteins from a plasmid in the prey strains deleted for the effector/immunity pairs restored immunity against self-intoxication ( Fig 2 ) . Importantly , expression of either Va03170 or Va03180 from a plasmid restored immunity against self-intoxication in the Δva03170-80 strain indicating they are indeed redundant immunity proteins against Va03175-mediated toxicity ( Fig 2D and S2 Fig ) . Taken together , these results demonstrate that Va16922/7 , Va18287/2 , and Va03175/80/70 , are effector/immunity pairs of VaT6SS2 . Three additional proteins , Va03300 , Va04801 , and Va13639 , were secreted in a VaT6SS2-dependent manner ( Table 1 ) . However , we could not confidently determine whether they are VaT6SS2 antibacterial effectors or not , as either we did not identify an adjacent putative immunity gene ( for Va13639 ) , the short adjacent ORF was not associated with the gene encoding the secreted protein in other bacterial genomes and is thus not predicted to encode for its cognate immunity protein ( for Va03300 ) , or deletion of the adjacent ORF did not result in loss of immunity against self-intoxication and we did not find additional homologs of the putative immunity proteins encoded by V . alginolyticus 12G01 that could provide redundant immunity ( for Va03300 and Va04801 ) . Taken together , our results indicate that VaT6SS2 delivers at least three effectors into recipient cells to mediate antibacterial toxicity . To gain a more comprehensive understanding of the V . alginolyticus T6SS effector repertoires , we next set out to identify the effectors of VaT6SS1 using comparative proteomics . We were unable to detect VaT6SS1 activity under liquid growth conditions similar to those in which we saw VaT6SS1 antibacterial activity during competition experiments on agar plates ( i . e . LB medium at 30°C ) ( S3 Fig ) . Nevertheless , we recently reported that H-NS , a bacterial histone-like nucleoid structuring protein , serves as a repressor of the V . parahaemolyticus VpT6SS1 and that its deletion results in activation of VpT6SS1 even under non-inducing conditions [27] . We hypothesized that H-NS can also act as a repressor of VaT6SS1 and that its deletion may lead to activation of the system . Indeed , when we used a V . alginolyticus strain deleted for hns , expression and secretion of Hcp1 that was endogenously tagged at the C-terminus with a FLAG tag ( Hcp1-FLAG ) were readily detected ( S3 Fig ) , indicating that H-NS is a repressor of VaT6SS1 activity . This finding allowed us to analyze the VaT6SS1 secretome even when V . alginolyticus were grown in liquid . With Δhns strains , we used comparative proteomics to find proteins that are secreted only when VaT6SS1 is active . Again , we used MS to analyze the secretomes of V . alginolyticus Δhcp2/Δhns ( with an active VaT6SS1 ) and Δhcp1/Δhcp2/Δhns ( with an inactive VaT6SS1 ) strains ( see S2 Dataset ) . The strains used were also deleted for hcp2 to detect only proteins secreted by VaT6SS1 . We identified 6 proteins that were differentially found in the supernatant of the strain in which VaT6SS1 was active ( Table 2 ) . One of these proteins , Va01440 , is the VaT6SS1 tail tube secreted component PAAR1 and served to validate our VaT6SS1 secretome analysis . As predicted by the presence of the MIX motif , two of the identified secreted proteins , Va16152 and Va01565 , were previously classified by us as putative T6SS MIX-effectors [16] . Va16152 contains an N-terminal MIX motif belonging to the MIX IV clan and a C-terminal pore-forming colicin-like domain ( according to HHPred analysis [29] ) , and is encoded outside of the VaT6SS1 gene cluster . Va01565 , which is encoded at the beginning of the VaT6SS1 gene cluster , contains a MIX motif belonging to the MIX I clan and is a homolog of the V . parahaemolyticus MIX-effector VP1388 [16] . Another VaT6SS1 secreted protein , Va01435 , is encoded at the end of the VaT6SS1 gene cluster and is predicted to contain an N-terminal LysM peptidoglycan-binding domain followed by a peptidoglycan ( PG ) hydrolase domain and a lysozyme-like domain ( according to HHPred analysis [29] ) . Moreover , the genes encoding for these three proteins were immediately upstream of small ORFs that could encode for their cognate immunity proteins ( Va16147 , Va01560 , and Va01430 , respectively ) ( Fig 3A ) . To test whether Va16152/47 , Va01565/0 , and Va01435/0 are VaT6SS1 effector/immunity pairs , we monitored the ability of a V . alginolyticus wild-type strain to kill strains with deletions in the putative effector/immunity gene pairs . As shown in Fig 3B–3D , the wild-type strain was able to inhibit the growth of strains with deletions in va16152-47 , va01565-0 and va01435-0 when co-cultured under VaT6SS1 inducing conditions , and inactivation of VaT6SS1 by deletion of hcp1 or deletion of the effector/immunity pairs in the attacking strains resulted in increased growth of the prey strains . Moreover , exogenous expression of the putative immunity proteins Va16147 and Va01560 from a plasmid , but not of Va01430 , also resulted in increased growth of the prey strains deleted for the cognate effector/immunity pairs . It is possible that the inability of the plasmid encoding for Va01430 to complement the deletion resulted from poor expression of Va01430 under the tested conditions . We were surprised by the minor deleterious effect of the wild-type attacker strains on the effector/immunity deletion strains , as we expected that the VaT6SS1 would mediate bacterial killing based on the toxic effects observed for VaT6SS1 when E . coli was used as prey ( Fig 1 ) . We reasoned that perhaps the toxic effects were minor because we failed to properly activate the VaT6SS1 under the tested conditions and thus were not observing the full toxic effect of the individual effectors . To test whether this is the case , we used the Δhcp2/Δhns strain ( with a constitutively active VaT6SS1 ) as an attacker strain in competition assays with the effector/immunity deletion strains . As predicted , upon de-repression of VaT6SS1 by deleting H-NS the attacking strain was able to kill all tested effector/immunity deletion strains ( Fig 3E ) . This killing was VaT6SS1-mediated as a Δhcp1/Δhcp2/Δhns attacking strain ( with an inactive VaT6SS1 ) was not toxic to the prey strains . Notably , the constitutive activation of VaT6SS1 in the attacking strain did not result in bacterial toxicity simply because of over-expression or delivery of effectors , as it was not toxic to a Δhcp1 prey strain that did not have an active VaT6SS1 but still had all of the genes encoding for the putative immunity proteins ( Fig 3E ) . Two additional proteins , Va01555 and Va17542 , were secreted in a VaT6SS1-dependent manner in our comparative proteomics analysis ( Table 2 ) . However , we concluded that they were most likely not bona fide antibacterial VaT6SS1 effectors . Va01555 is a homolog of the V . parahaemolyticus VP1390 which we previously identified as secreted by VpT6SS1 but ruled out as an antibacterial effector because it had no detectable immunity protein [16] . Va17542 is a homolog of the V . parahaemolyticus VopQ , a virulence effector protein of the Type III Secretion System 1 ( T3SS1 ) [30–32] . It is possible that the detection of Va17542 in our secretome was an artifact resulting from hyper-activation of T3SS1 by deletion of hns [33] . Taken together with our previous identification of the VaT6SS1 MIX-effector Va02265 [16] , our results indicate that VaT6SS1 delivers at least four effectors , three of which are MIX-effectors , into recipient cells to mediate antibacterial toxicity . We next examined whether the VaT6SS1 effector/immunity pairs that we identified in the 12G01 strain were also found in other V . alginolyticus strains . Homologs of the effector Va01435 , as well as of the two MIX-effectors Va01565 and Va16152 , and their cognate immunity proteins were encoded in the genomes of other sequenced V . alginolyticus strains ( i . e . NBRC 15630 , E0666 , and 40B ) in the same synteny as in strain 12G01 . However , the bicistronic unit encoding the MIX-effector/immunity pair Va02265/0 was not found in the genomes of other sequenced V . alginolyticus strains ( Fig 4A ) , suggesting that it was recently acquired by the 12G01 strain . Interestingly , we found bicistronic units encoding homologs of the Va02265/0 MIX-effector/immunity cassette in genomes of other Vibrio species ( e . g . V . anguillarum NB10 ) , albeit at a synteny distinct from the one it had in V . alginolyticus 12G01 ( Fig 4A ) . These results suggested that this MIX-effector/immunity cassette might be transmitted horizontally between Vibrio species . In our previous work , we reported that MIX-effectors group into five distinct clans named MIX I-V based on the sequences of their MIX-containing regions [16] . The V . alginolyticus MIX-effector Va02265 belongs to the MIX V clan and does not neighbor other T6SS components on the genome [16] . Therefore , we classified it as an "orphan" MIX-effector . To test whether other MIX V-effectors are "orphan" , we next examined the genome neighborhoods of genes encoding other MIX V-effectors . Remarkably , we found that most members of the MIX V clan are “orphan” MIX-effectors that do not neighbor any other T6SS component ( Only 1 out of 124 identified MIX V-effectors was found to neighbor T6SS components; 35 out of the 124 remain uncertain as not all neighboring genes were identified; See S3 Dataset and S1 File ) . Furthermore , MIX V clan members were only found in marine γ-proteobacteria , with the vast majority distributed among Vibrionales ( 117 ) , and a few found in Alteromonadales ( 4 ) , Aeromonadales ( 2 ) , and Oceanospirillaales ( 1 ) ( S4 Fig ) . We also noticed that some MIX V members are encoded adjacent to transposable elements such as transposases and integrases . Thus , we hypothesized that MIX-effector that belong to the MIX V clan are mobile and can be shared between marine bacteria . In support of this notion , the “orphan” MIX V-effector that we previously identified in V . parahaemolyticus RIMD 2210633 , VPA1263 [16] , is encoded within the V . parahaemolyticus island-6 ( VPaI-6; vpa1254-vpa1270 ) that contains a transposase and an integrase . This VPaI-6 was suggested to be a mobile element acquired by pandemic strains [34] . The bicistronic unit encoding VPA1263 and its cognate immunity Vti2 [16] can be found in distinct locations on the genomes of other Vibrios ( e . g . Vibrio campbelli ATCC BAA-1116 ) flanked by transposase genes ( Fig 4B ) . Furthermore , MIX V members can be located on plasmids . In Aliivibrio salmonicida , one MIX V member ( VSAL_p840_46 ) is encoded adjacent to an IS insertion element ( VSAL_p840_45 ) on one of four plasmids that also includes T6SS components ( i . e . VgrG: VSAL_p840_36 and Hcp: VSAL_p840_35 ) and viral conjugative transfer genes ( VSAL_p840_1 –VSAL_p840_21 ) . A second MIX V-effector that is more closely related to the plasmid copy than to other MIX V members resides in the chromosome ( VSAL_I0031 ) near a noted transposase ( VSAL_I0029 ) . These findings further support our hypothesis that “orphan” MIX-effectors belonging to the MIX V clan are mobile T6SS effectors that can move between Gram-negative marine bacteria via horizontal gene transfer and be used as effectors by MIX-secreting T6SSs . Whereas previous reports demonstrated that homologous effectors from different species can be secreted by T6SSs [35] , the possibility that bacteria can use their T6SSs to secrete newly acquired effectors which are not naturally encoded in their genome has not been addressed . To directly test whether bacteria can use a newly acquired mobile MIX V-effector to increase their competitive fitness , we used the V . alginolyticus MIX V-effector/immunity pair Va02265/0 and asked whether a V . parahaemolyticus that contains a MIX-secreting T6SS ( VpT6SS1 ) but no homologs of Va02265 can use Va02265 as a T6SS effector to gain competitive advantage over its parental kin ( which are otherwise immune against self-intoxication ) . Indeed , the V . parahaemolyticus POR1 strain was able to kill a POR1 parental prey when the Va02265/0 effector/immunity cassette was introduced and expressed from a plasmid in the attacking strain ( Fig 5A ) . This Va2265-mediated self-intoxication was dependent on VpT6SS1 activity , as an attacker strain expressing the effector/immunity cassette that had an inactive VpT6SS1 ( POR1Δhcp1 ) was no longer able to kill the parental prey . Moreover , expression of the Va02260 immunity protein from a plasmid in the parental prey strain resulted in immunity against the Va02265-mediated intoxication . Thus , these results demonstrate that an “orphan” MIX-effector belonging to the MIX V clan can be used by another Vibrio strain as a T6SS effector and provide competitive advantage . However , it appears that there is some degree of specificity towards T6SSs , as the V . cholerae V52 strain was unable to use Va02265 as an antibacterial effector in a self-intoxication assay ( Fig 5B ) under conditions in which this strain mediated T6SS-dependent killing of E . coli [36] .
In this work , we used genetic and proteomic analyses to characterize the environmental conditions that activate the two T6SSs found in the marine pathogen V . alginolyticus , identify their functions , and determine their effector repertoires . We found that both T6SSs mediate interbacterial competition although they are active under different salinity and temperature conditions , suggesting they are utilized by this bacterium in different environments . Surprisingly , even though the V . alginolyticus T6SS gene clusters are similar to those of V . parahaemolyticus in terms of gene content and organization , and both bacteria reside in similar habitats , the regulation of their T6SSs differs . For example , whereas the V . parahaemolyticus VpT6SS1 is active under high salt conditions , it appears that its V . alginolyticus counterpart , VaT6SS1 , is active under low salt conditions . Moreover , in our previous studies we were unable to detect antibacterial activity for VpT6SS2 due to repression of the system by surface-sensing activation under our bacterial competition assay conditions [20] . However , the results shown here indicate that VaT6SS2 is not inhibited by surface-sensing as antibacterial activity was readily detectable in competition assays performed on agar plates . This differential regulation of the V . parahaemolyticus and V . alginolyticus T6SSs allowed us to identify the antibacterial activity of VaT6SS2 and its effector repertoire . In this work we identified six new T6SS effectors in V . algnolyticus 12G01 . While we are currently investigating their biochemical activities , our results demonstrate that all six effectors mediate antibacterial toxicities . The presence of antibacterial T6SS effectors in bicistronic units together with genes that encode for their cognate immunity proteins is well documented [7 , 11 , 16 , 37] . Indeed , we showed that the proteins encoded downstream of the six effectors that were secreted in a T6SS-dependent manner in our comparative proteomics analyses do provide immunity against T6SS-mediated intoxication . Taken together with the various putative toxin domains found in the six T6SS-secretd proteins , we conclude that they are antibacterial T6SS effectors . Based on the three VaT6SS2 effectors that we identified in this work , we hypothesize that the V . parahaemolyticus VpT6SS2 also mediates antibacterial activity under conditions we have yet to uncover . This hypothesis is supported by the presence of a close homolog to the VaT6SS2 effector Va18287 , a member of the Rhs class of T6SS effectors , in the V . parahaemolyticus genome ( Table 1 ) [12] . The V . parahaemolyticus homolog , VP1517 , contains an RhsA domain and a predicted C-terminal nuclease domain of the HNH/ENDO VII superfamily which is often found in bacterial toxins and could thus serve as an antibacterial T6SS effector [15] . Notably , the impact of our previous discovery of the MIX motif , which enabled us to identify hundreds of effectors belonging to the MIX-effector class in various bacterial species [16] , is further underscored in this work . While we predicted the presence of three MIX-effectors in V . alginolyticus 12G01 , we only found two secreted MIX-effectors in our comparative proteomics analysis ( Va01565 and Va16152 ) . Thus , the third MIX-effector , Va02265 , would not have been identified if not for the presence of the MIX motif in its sequence , as it was not encoded close to other T6SS components on the genome ( and therefore labeled as an "orphan" MIX-effector ) . Furthermore , our finding that the V . alginolyticus VaT6SS1 can secrete three MIX-effectors that belong to different MIX clans ( Va01565 to MIX I , Va16152 to MIX IV , and Va02265 to MIX V ) implies that T6SSs have a certain degree of freedom in the MIX-effectors they can secrete . Our observation that most members of the MIX V clan are "orphan" MIX-effectors that often neighbor transposable elements led us to hypothesize that they are mobile and shared between marine bacteria . Remarkably , Borgeaud et . al . recently reported that the T6SS is part of the competence regulon in V . cholerae [38] . They showed that T6SS-mediated killing allows V . cholerae to uptake the released DNA of the lysed bacterial competitor and incorporate it into its own DNA , thus fostering horizontal gene transfer and driving evolution [38] . It is therefore compelling to speculate that similar mechanisms are found in other marine bacteria , and that bacteria can use their T6SSs to prey on DNA from their competitors and acquire new mobile MIX V-effector/immunity cassettes that will provide increased fitness in future competitions as they diversify their T6SS effector repertoires . Another mechanism that drives evolution of virulence factors is the presence of non-integrated conjugative plasmids like in Aliivibrio salmonicida LFI1238 . Codon usage analysis showed that the genomic chromosomal copy of the MIX V-effector in Aliivibrio is more closely related to the plasmid copy than to the genome background [39] , suggesting the chromosomal gene that neighbors transposases originated from the plasmid . The observed high occurrence of transposable elements in the Aliivibrio salmonicida LFI1238 genome is thought to mediate this gene transfer and represent a mechanism for driving diversity in the chromosome . It is therefore possible that similar mechanisms are found in other Vibrios to enable horizontal gene transfer of mobile MIX V-effectors . We found that MIX V-effectors are only present in marine bacteria , mostly in members of the Vibrionales family . These bacteria can interact with each other in the same aquatic habitats , thus providing access to various MIX V-effectors from competing species . A similar phenomenon was recently reported in Xanthomonads , where Tn3-like transposons play a role in spreading virulence effectors of the T3SS via horizontal gene transfer [40] . In conclusion , we propose that mobile MIX V-effectors serve as an environmental reservoir of polymorphic antibacterial toxins that can be shared between marine bacteria via horizontal gene transfer and used to enrich the versatility of T6SS effector repertoires , thus increasing competitive fitness .
The Vibrio alginolyticus 12G01 strain and the Vibrio parahaemolyticus RIMD 2210633 derivative strain POR1 ( RIMD 2210633 ΔtdhAS ) [41] and their derivatives were routinely cultured in Marine Luria-Bertani ( MLB ) broth ( Luria-Bertani broth containing 3% sodium chloride ) or on Marine minimal media ( MMM ) agar [42] at 30°C . The Vibrio cholerae V52 strain and its derivative V52ΔvipA ( a gift from Dr . J . Mekalanos , Harvard Medical School ) [4 , 43] were routinely cultured in LB broth at 37°C . E . coli DH5α and S17-1 ( λ pir ) were routinely cultured in 2×YT broth at 37°C . The medium was supplemented with kanamycin or chloramphenicol where necessary . Arabinose was added to solid or liquid media at a final concentration of 0 . 1% ( w/v ) when induction of expression from a plasmid was required . For ectopic expression of va01560 , va01430 , va16147 , va03180 , va03170 , va18282 , and va16927 , the genes' coding sequences were amplified and cloned into the MCS of the arabinose-inducible expression vector pBAD33 ( Invitrogen ) containing chloramphenicol resistance . For ectopic expression of the effector/immunity gene pair , the coding regions of effector and immunity genes va02265-va02260 were amplified together , including the stop codons , and cloned into the MCS of the pBAD/Myc-His vector ( Invitrogen ) in which the antibiotic resistance was changed from ampicillin to kanamycin . The resulting plasmids were conjugated into V . alginolyticus , V . cholerae , or V . parahaemolyticus using tri-parental mating . For the generation of in-frame Vibrio deletion strains , the nucleotide sequences 1 kb upstream and 1 kb downstream of the effector/immunity pairs , the hcp genes ( va01540 and va07583 are hcp1 and hcp2 , respectively ) or hns ( va20201 ) were amplified and cloned together into pDM4 , a CmROriR6K suicide plasmid . For the generation of the C-terminal FLAG tagged Hcp1 strain , a C-terminal FLAG tagged version and the nucleotide sequences 1 kb downstream of the V . alginolyticus hcp1 were amplified and cloned together into pDM4 . The resulting pDM4 plasmid was conjugated into V . alginolyticus from E . coli S17-1 ( λ pir ) and transconjugants were selected on media containing 25 μg/ml chloramphenicol . Bacteria were counter-selected by growing on media containing 15% sucrose . Deletions and insertions were confirmed by PCR . Assay was performed as previously described [20] and was repeated twice with similar results . Results of a representative experiment are shown . Vibrio strains were grown overnight in MLB . Cells were washed and re-suspended in 5 ml of LB media . Cultures were incubated with agitation for 5 hours at 30°C . Expression and secretion were determined by immunoblot as previously described [20] with anti-FLAG antibodies ( Sigma Aldrich ) . Equal loading of total protein lysates was confirmed by analysis of representative bands using Ponceau S staining of the immunoblot membrane . Bacterial strains were grown over-night in MLB ( V . alginolyticus and V . parahaemolyticus ) , LB ( V . cholerae ) , or 2xYT ( E . coli ) . Bacterial cultures were mixed and spotted on LB or MLB plates as previously described [20] . CFU of the prey spotted at t = 0h were determined by plating 10-fold serial dilutions on selective media plates . Bacterial spots were harvested from plates after 4 hours incubation and the CFU of the surviving prey cells were determined . Assays were repeated at least twice with similar results , and results of a representative experiment are shown . For analysis of the VaT6SS2 secretome , V . alginolyticus cultures of Δhcp1 ( with an active VaT6SS2; T6SS2+ ) and Δhcp1/Δhcp2 ( with an inactive VaT6SS2; T6SS2- ) strains were grown in triplicate in 50 ml MLB media at an initial OD600 = 0 . 54 for 5 h at 30°C . For analysis of the VaT6SS1 secretome , V . alginolyticus cultures of Δhcp2/Δhns ( with an active VaT6SS1; T6SS1+ ) and Δhcp1/Δhcp2/Δhns ( with an inactive VaT6SS1; T6SS1- ) strains were grown in triplicate in 50 ml LB media at an initial OD600 = 0 . 18 for 5 h at 30°C . Media were collected and proteins precipitated as previously described [44] . Protein samples were run 10 mm into the top of an SDS-PAGE gel , stained with Coomassie Blue , and excised . Overnight digestion with trypsin ( Promega ) was performed after reduction and alkylation with DTT and iodoacetamide ( Sigma—Aldrich ) . The resulting samples were analyzed by tandem MS using either a QExactive or Orbitrap Elite mass spectrometer ( Thermo Electron ) coupled to an Ultimate 3000 RSLC-Nano liquid chromatography system ( Dionex ) . Peptides were loaded onto either a 180 μm i . d . , 15-cm long , self-packed column containing 1 . 9 μm C18 resin ( Dr . Maisch , Ammerbuch , Germany ) or a 75 μm i . d . , 50-cm long Easy Spray column ( Thermo ) and eluted with a gradient of either 0–28% buffer B for 40 min or 0–28% buffer B for 60 min . Buffer A consisted of 2% ( v/v ) acetonitrile ( ACN ) and 0 . 1% formic acid in water . Buffer B consisted of 80% ( v/v ) ACN , 10% ( v/v ) trifluoroethanol , and 0 . 08% formic acid in water . To ensure accurate Label-free quantification , control and experiment samples ( i . e . T6SS1+/T6SS1- or T6SS2+/T6SS2- ) were run on the same column with the same gradient using the same instrument . The mass spectrometer acquired up to 10 fragment spectra for each full spectrum acquired Raw MS data files were converted to peak list format using ProteoWizard msconvert ( version 3 . 0 . 3535 ) [45] . The resulting files were analyzed using the central proteomics facilities pipeline ( CPFP ) , version 2 . 1 . 0 [46 , 47] . Peptide identification was performed using the X ! Tandem [48] and open MS search algorithm ( OMSSA ) [49] search engines against a database consisting of V . alginolyticus 12G01 sequences from UniProt KnowledgeBase , with common contaminants and reversed decoy sequences appended [50] . Fragment and precursor tolerances of 20 ppm and 0 . 1 Da were specified , and three missed cleavages were allowed . Carbamidomethylation of Cys was specified as a fixed modification , and oxidation of Met was specified as a variable modification . Label-free quantitation of proteins across samples was performed using SINQ normalized spectral index software [51] . To identify statistically significant differences in protein amount between T6SS1+/T6SS1- and T6SS2+/T6SS2− strains , SINQ quantitation results for three biological replicates per strain were processed using the power law global error model ( PLGEM ) package in R [52 , 53] . Protein identifications were filtered to an estimated 1% protein false discovery rate ( FDR ) using the concatenated target-decoy method [50] . An additional requirement of two unique peptide sequences per protein was imposed , resulting in a final protein FDR<1% . Spectral index quantitation was performed using peptide-to-spectrum matches ( PSMs ) with a q-value≤0 . 01 , corresponding to a 1% FDR rate for PSMs . The datasets of tandem MS results were uploaded to the MassIVE repository ( http://massive . ucsd . edu/ProteoSAFe/status . jsp ? task=f4d7613b8ee2414985a8e0bbfcf905fe; MassIVE ID: MSV000078946 ) . To identify MIX V-effectors , we queried the nr database from NCBI ( Feb 24 , 2015 ) with the N-terminal sequence of VPA1263 ( gi| 28901118 , 1–320 ) that includes MIX V using PSI-BLAST [54] with default values ( E-value cutoff 0 . 005 , 5 iterations ) . We limited the resulting hits to include only refseq sequences from complete genomes , and clustered the bounded sequence hits using CLANS [55] . Sequences that cluster together with VPA1263 and cover the MIX V sequence motifs are represented in S3 Dataset . The Va02265 MIX V sequence was detected by PSI-BLAST , but is omitted from the resulting database , as it is not part of refseq . Taxonomic distributions of the corresponding sequences were generated with batch entrez on the NCBI website . To locate the genome neighborhoods of MIX V effectors , we located their NCBI gene identifiers ( gi ) among nucleotide records of completely sequences genomes . We noted the gis corresponding to protein sequences located adjacent to MIX V ( +/- 3 genes ) . The completely sequenced genomes represent various stages of assembly completeness linking shorter contigs into ordered genomes . For those MIX V that reside near the ends of shorter contigs , we designated any missing neighbors with an “x” ( see S3 Dataset and S1 File for additional information ) . We collected all protein sequences corresponding to MIX V-effectors and their neighbors and defined the domain content from the COG database using batch CD-search [56] on the NCBI server ( default cutoff E-value 0 . 01 ) . To determine whether a MIX V gene neighbors T6SS components , we manually searched identified domains for those that correspond to the 13 core T6SS components , as previously identified by Boyer et al . [17] . Protein sequence databases were generated for each species using coding sequence collected from NCBI nucleotide records for Vibrio parahaemolyticus RIMD 2210633 chromosome 1 ( NC_004603 . 1 ) and chromosome 2 ( NC_004605 . 1 ) , Vibrio parahaemolyticus BB22OP chromosome 1 ( NC_019955 . 1 ) and chromosome 2 ( NC_019971 . 1 ) , Vibrio alginolyticus 12G01 scaffold ( CH902589 . 1 ) , Vibrio alginolyticus NBRC 15630 = ATCC 17749 chromosome 1 ( NC_022349 . 1 ) and chromosome 2 ( NC_022359 . 1 ) , Vibrio alginolyticus 40B scaffold ( ACZB01000054 . 1 ) , Vibrio anguillarum NB10 chromosome 2 ( LK021129 ) , and Vibrio campbellii ATCC BAA-1116 chromosome 1 ( NC_022269 . 1 ) . To define gene correspondence between species , MIX V and surrounding genes from Vibrio parahaemolyticus RIMD 2210633 and Vibrio alginolyticus 12G01 were used as BLAST queries against each database , keeping only top hits ( E-value cutoff 0 . 001 ) . Conserved ordering of top BLAST hits in the chromosomes ( of scaffold ) was considered to define gene synteny between the various species . Domain annotations were defined according to NCBI conserved domain database ( cdd ) . | The bacterial type VI secretion system ( T6SS ) is a contact-dependent protein secretion apparatus that is emerging as a major component of interbacterial competition in the environment . The bacterium Vibrio alginolyticus is a pathogen of marine animals and a causal agent of wound infections , otitis , and gastroenteritis in humans . In this study , we provide a comprehensive characterization of the environmental regulation , antibacterial activities , and secreted effector repertoires of the two T6SSs found in this pathogen . We also identify a subset of T6SS effectors that appear to be mobile and shared between marine bacteria that can interact with each other in aquatic environments . Our findings suggest that bacteria can incorporate T6SS effectors from competitors in the environment . These newly acquired toxins may be used to expand and diversify T6SS effector repertoires and enhance bacterial fitness . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Type VI Secretion System Toxins Horizontally Shared between Marine Bacteria |
Networks are becoming a ubiquitous metaphor for the understanding of complex biological systems , spanning the range between molecular signalling pathways , neural networks in the brain , and interacting species in a food web . In many models , we face an intricate interplay between the topology of the network and the dynamics of the system , which is generally very hard to disentangle . A dynamical feature that has been subject of intense research in various fields are correlations between the noisy activity of nodes in a network . We consider a class of systems , where discrete signals are sent along the links of the network . Such systems are of particular relevance in neuroscience , because they provide models for networks of neurons that use action potentials for communication . We study correlations in dynamic networks with arbitrary topology , assuming linear pulse coupling . With our novel approach , we are able to understand in detail how specific structural motifs affect pairwise correlations . Based on a power series decomposition of the covariance matrix , we describe the conditions under which very indirect interactions will have a pronounced effect on correlations and population dynamics . In random networks , we find that indirect interactions may lead to a broad distribution of activation levels with low average but highly variable correlations . This phenomenon is even more pronounced in networks with distance dependent connectivity . In contrast , networks with highly connected hubs or patchy connections often exhibit strong average correlations . Our results are particularly relevant in view of new experimental techniques that enable the parallel recording of spiking activity from a large number of neurons , an appropriate interpretation of which is hampered by the currently limited understanding of structure-dynamics relations in complex networks .
Analysis of networks of interacting elements has become a tool for system analysis in many areas of biology , including the study of interacting species [1] , cell dynamics [2] and the brain [3] . A fundamental question is how the dynamics , and eventually the function , of the system as a whole depends on the characteristics of the underlying network . A specific aspect of dynamics that has been linked to structure are fluctuations in the activity and their correlations in noisy systems . This work deals with neuronal networks , but other examples include gene-regulatory networks [4] , where noise propagating through the network leads to correlations [5] , and different network structures have important influence on dynamics by providing feedback loops [6] , [7] . The connection between correlations and structure is of special interest in neuroscience . First , correlations between neural spike trains are believed to play an important role in information processing [8] , [9] and learning [10] . Second , the structure of neural networks , encoded by synaptic connections between neurons , is exceedingly complex . Experimental findings show that synaptic architecture is intricate and structured on a fine scale [11] , [12] . Nonrandom features are induced by neuron morphology , for example distance dependent connectivity [13] , [14] , or specific connectivity rules depending on neuron types [15] , [16] . A number of novel techniques promise to supply further details on local connectivity [17] , [18] . Measured spike activity of neurons in such networks shows , despite high irregularity , significant correlations . Recent technical advances like multiple tetrode recordings [19] , multielectrode arrays [20]–[22] or calcium imaging techniques [23] , [24] allow the measurement of correlations between the activity of an increasingly large number of neuron pairs in vivo . This makes it possible to study the dynamics of large networks in detail . Since recurrent connections represent a substantial part of connectivity , it has been proposed that correlations originate to a large degree in the convergence and divergence of direct connectivity and common input [8] and must therefore strongly depend on connectivity patterns [25] . Experimental studies found evidence for this thesis in a predominantly feed-forward circuit [26] . In another study , only relatively small correlations were detected [27] and weak common input effects or a mechanism of active decorrelation were postulated . In recent theoretical work recurrent effects have been found to be an important factor in correlation dynamics and can account for decorrelation [20] , [22] . Several theoretical studies have analysed the effects of correlations on neuron response [28] , [29] and the transmission of correlations [30]–[34] , also through several layers [35] . However , the description of the interaction of recurrent connectivity , correlations and neuron dynamics in a self-consistent theory has not been presented yet . Even in the case of networks of strongly simplified neuron models like integrate and fire or binary neurons , nonlinear effects prohibit the evaluation of effects of complex connectivity patterns . In [36] , [37] correlations in populations of neurons were studied in a linear model that accounted for recurrent feedback . With a similar model , the framework of interacting point processes developed by Hawkes [38] , [39] , we analyse effects of different connectivity patterns on pairwise correlations in strongly recurrent networks . Spike trains are modeled as stochastic processes with presynaptic spikes affecting postsynaptic firing rates in a linear manner . We describe a local network in a state of irregular activity , without modulations in external input . This allows the self-consistent analytical treatment of recurrent feedback and a transparent description of structural effects on pairwise correlations . One application is the disentanglement of the explicit contributions of recurrent input on correlations in spike trains in order to take into account not only effects of direct connections , but also indirect connectivity , see Figure 1 . We find that variations in synaptic topology can substantially influence correlations . We present several scenarios for characteristic network architectures , which show that different connectivity patterns affect correlations predominantly through their influence on statistics of indirect connections . An influential model for local neural populations is the random network model [40] , [41] , possibly with distance-dependent connectivity . In this case , the average correlations , and thereby the level of population fluctuations or noise , only depend on the average connectivity and not on the precise connectivity profile . The latter , however , influences higher order properties of the correlation distribution . This insensitivity to fine-tuning is due to the homogeneity of the connectivity of individual neurons in this type of networks . The effect has also been observed in a very recent study , where large-scale simulations were performed [42] . In networks with more complex structural elements , like hubs or patches , however , we find that also average correlations depend on details of the connectivity pattern . Part of this work has been published in abstract form [43] .
In order to study correlations in networks of spiking neurons with arbitrary connectivity we use the theory derived in [38] , which we refer to as Hawkes model , for the calculation of stationary rates and correlations in networks of linearly interacting point processes . We only summarise the definitions and equations needed in the specific context here . A mathematically more rigorous description can be found in [38] and detailed applications in [44] , [45] . We will use capital letters for matrices and lower case letters for matrix entries , for example . Vectors will not be marked explicitly , but their nature should be clear from the context . Fourier transformed quantities , discrete or continuous , will be denoted by , for example . Used symbols are summarised in Table 1 . Our networks consists of neurons with excitatory and inhibitory neurons . Spike trains of neurons are modeled as realisations of Poisson processes with time-dependent rates . We have ( 1 ) where denotes the mathematical expectation , in this case across spike train realisations . Neurons thus fire randomly with a fluctuating rate which depends on presynaptic input . For the population of neurons we use the spike train vector and the rate vector . Spikes of neuron influence the rate of a connected neuron by inducing a transient rate change with a time course described by the interaction kernel , which can in principle be different for all connections . For the sake of simplicity we use the same interaction kernels for all neurons of a subpopulation . The rate change due to a spike of an excitatory presynaptic neuron is described by and of an inhibitory neuron by . The total excitatory synaptic weight can then be defined as and the inhibitory weight accordingly as . Connections between neurons are chosen randomly under varying restrictions , as explained in the following sections . For unconnected neurons . The evolution of the rate vector is governed by the matrix equation ( 2 ) The effect of presynaptic spikes at time on postsynaptic rates is given by the interaction kernels in the matrix and depends on the elapsed time . Due to the linearity of the convolution , effects of individual spikes are superimposed linearly . The constant spike probability can be interpreted as constant external drive . We require all interactions to respect causality , that is for . The Hawkes model was originally defined for positive interaction kernels . Inhibitory kernels can lead to negative values of at certain times , so strictly one should use the rectified variable as a basis for spike generation . We assume further on that becomes negative only rarely and ignore the non-linearity introduced by this rectification . The effects of this approximation are illustrated in Figure 2 . In the equilibrium state , where the expectation value for the rates does not depend on time , we then have where we denoted the expectation of the fluctuating rates by for notational simplicity . An explicit expression for the equilibrium average rates is ( 4 ) where refers to the identity matrix . We describe correlations between spike trains by the covariance density matrix . For point processes it is formally defined as the inverse Fourier transform of the spike cross-spectrum , but can in analogy to the case for discrete time be written as ( 5 ) and corresponds to the probability of finding a spike after a time lag , given that a spike happened at time , multiplied by the rate . The term represents chance correlations such that for uncorrelated spike trains for . Due to the point process nature of spike trains , autocovariance densities have a discontinuous contribution . This discontinuity is separated explicitly from the continuous part using the diagonal rate matrix with the constant elements ( here denotes the Kronecker delta ) . For independent spike trains so that one recovers the autocorrelation density function of Poisson processes , . A self-consistent equation that determines the covariance density matrix is ( 6 ) for . A key result in [38] is that , if the Fourier transform of the kernel matrix ( 7 ) is known , ( 6 ) can be solved and the Fourier transform of the cross covariance density is given by ( 8 ) The definition of the Fourier transform implies that and accordingly , where we introduced the shortcuts and for the integrated covariance density matrix and kernel matrix , respectively . They are , from ( 8 ) , related by ( 9 ) The rate Equation ( 4 ) becomes with these definitions ( 10 ) Equation ( 8 ) describes the time-dependent correlation functions of an ensemble of linearly interacting units . In this work we concentrate on purely structure-related phenomena under stationary conditions . Therefore we focus on the integrated covariance densities , which are described by Equation ( 9 ) . Differences in the shape of the interaction kernels which do not alter the integral do not affect our results . One example is the effect of delays , which only shift interaction kernels in time . Furthermore we restrict ourselves to systems where all eigenvalues of satisfy . This condition guarantees the existence of the matrix inverse in ( 9 ) and ( 10 ) . Moreover , if the real part for any , no stable equilibrium exists and network activity can explode . For further details see Section 1 of Supporting Text S1 . The matrix elements and have an intuitive interpretation . The integrated impulse response corresponds to the average number of additional spikes in neuron caused by an extra spike in neuron . The integrated cross-correlations , in the following simply denoted as correlations , equal , for asymptotically large counting windows , the covariances of spike counts and between spike trains and , ( 11 ) see for example [20] , [46] . On the population level one finds for the population count variance normalised by the bin size , , that ( 12 ) Strictly this is only true in the limit of infinitely large bin size . However , the approximation is good for counting windows that are large with respect to the temporal width of the interaction kernel . In this sense , the sum of the correlations is a measure for the fluctuations of population activity . Another measure for correlations that is widely used is the correlation coefficient , . In this context it is not convenient , as the normalisation over the count variance destroys the simple relation to the population fluctuations . Even worse , as count variances are , just as covariances , influenced by network structure , for example global synchrony is not captured by this measure . We simulated networks of linearly interacting point processes in order to illustrate the theory , Figure 2 . In this network connections between all nodes are realised with constant probability . Parameters were chosen such that net recurrent input is inhibitory . The full connectivity matrix was used for the rate and correlation predictions in Equations ( 9 ) and ( 10 ) and the population count variance , Equation ( 12 ) . Further simulation details are given below . This figure demonstrates that the approximation that fluctuating rates stay largely above zero gives good results even in effectively inhibitory networks with strong synapses . There are nonetheless slight deviations between prediction and simulation . On the one hand , fluctuations of the variable around a positive mean can reach below zero . This factor is especially relevant if rate fluctuations are high , for example because of strong synapses and low mean input . On the other hand , strongly inhibitory input can result into a negative mean value of for some neurons . This can happen only for wide rate distributions and strong inhibition , since the ensemble average of is always positive . In Figure 2C it is shown that only few neurons have predicted rates below zero , and that deviations between predicted and simulated rate distributions are significant primarily for low rates . The correlations in panel D are hardly affected . We found that for a wide range of parameters Hawkes' theory returns correct results for most of the rates and correlations even in effectively inhibitory networks . Simulations of linearly interacting point processes were conducted using the NEST simulator [47] . Spikes of each neuron were generated with a rate corresponding to the current value of the intrinsic variable . Negative values of were permitted , but resulted in no spike output . Neurons received external drive corresponding to a constant rate of . Incoming spikes resulted in an increase/decrease of of amplitude for excitatory/inhibitory spikes , which decayed with a time constant of . This corresponds to exponential interaction kernels with total weights and . Synaptic delay was . Simulation time step was for the correlation and rate measurement and for spikes shown in the raster plot . In Figure 2 total simulation time was . Data from an initial period of was dropped . Correlograms were recorded for the remaining time with a maximum time lag of ( data not shown ) . The value for the correlations was obtained from the total number of coincident spikes in this interval . The total number of spikes was used for the measurement of the rates , while population fluctuations were determined from bins in the first .
In this section we address how recurrent connectivity affects rates and correlations . Mathematically , the kernel matrix is the adjacency matrix of a weighted directed graph . Single neurons correspond to nodes and connections are weighted by the integrated interaction kernels . With the shorthand Equation ( 9 ) becomes ( 13 ) where the rates are given by ( 10 ) , . For simplicity we normalise the external input , . The matrix describes the effect of network topology on rates and correlations . Under the assumptions stated in the methods section , can be written as a geometric series , The terms of this series describe how the rates result from external and recurrent input . The matrix relates to the part of the rates resulting directly from external input . For , each of the single terms corresponds to indirect input of other nodes via paths of length . The element consists of the sum over all possible weighted paths from node to node in steps via the nodes ( note that ) . Since , the elements of describe the influence of neuron on neuron via all possible paths . Similarly ( 14 ) with . The first term accounts for the integral of the autocorrelation functions of independent stationary Poisson processes , given by their rates . Higher-order terms in this series describe recurrent contributions to correlations and autocorrelation . The matrix elements of are ( 15 ) In these expressions , a term like describes the direct effect of neuron on , taking into account the interaction strength and the rate of the presynaptic neuron . For example , in the term with and the elements describe indirect input of to via all . For , counts the common input of neurons and from all . Altogether , the series expansion of the correlation equation describes how the full correlation between neurons and results from the contributions of all neurons , weighted by their rate , via all possible paths of length to node and length to node , for all and . These paths with two branches are the subgroup of network motifs that contribute to correlations . Further examples are given in Figure 3 . The distribution of correlation coefficients depends on the distributions of these motifs . Note that larger motifs are built from smaller ones , hence distributions of different motifs are not independent . As mentioned before , the sum ( 14 ) converges only if the magnitude of all eigenvalues of is smaller than one . This ensures that the feedback by recurrent connections does not cause runaway network activation . Both too strong recurrent excitation and too strong recurrent inhibition can lead to a divergence of the series . In such cases , our approach does not allow correlations to be traced back to specific network motifs . Under this condition , the size of higher-order terms , that is the collective influence of paths of length and , decreases with their total length or order . This can be stated more precisely if one uses as a measure for the contribution the operator norm . After diagonalising we have ( 16 ) where denotes the eigenvalue with the largest absolute value . If it is close to one , contributions decay slowly with order and many higher-order terms contribute to correlations . In this dynamic context the network can then be called strongly recurrent . The average correlation across all pairs can be computed by counting the weighted paths between two given nodes . The average contribution of paths of length is ( 17 ) Let us separate the contributions from rates to the autocorrelations and define the average correlation by ( 18 ) The population fluctuations are determined by , ( 19 ) As a first approximation let us assume that every neuron in a given subpopulation projects to a fixed number of neurons in each subpopulation , denoted by . Furthermore , each neuron receives the same number of input connections from neurons of the two subpopulations , denoted by and . Synaptic partners are chosen randomly . These networks are called regular in graph theory , since the number of outgoing and incoming connections of each neuron , called the out- and in-degree , is identical for all neurons . This restriction can be relaxed to approximate certain types of networks , as we discuss in the respective sections . We set the external input . Then the total input to each neuron is . The shortcut corresponds to the average input each neuron receives from a potential presynaptic neuron . Since input is the same for all neurons , all rates are equal . Their value can be obtained as follows by the expansion of ( 10 ) , In a similar manner , analytical expressions for the average correlations can be obtained . Explicit calculations can be found in Section 2 of Supporting Text S1 . In particular , the average correlation and hence the population fluctuations only depend on the parameters and . Closed expressions can be derived in the special case where there is a uniform connection probability between all nodes , i . e . ( 20 ) With and one finds for the individual contributions ( 21 ) and the average correlation ( 22 ) Here , can be interpreted as the average direct interaction between two nodes and as the average common input shared by two nodes . Average correlations are determined by mean input and mean common input . Equation ( 22 ) can be used as an approximation if the degree distribution is narrow . In particular this is the case in large random networks with independent connections , independent input and output and uniform connection probabilities . These conditions ensure that deviations from the fixed out- and in-degrees balance out on average in a large matrix . Numerical examples can be found in the following section . Instead of purely random networks we now consider networks of nodes arranged in a ring with distance dependent connectivity . The type of each neuron is determined randomly with probabilities and , such that on average excitatory and inhibitory neurons are distributed over the ring . Outbound connections of each neuron to a potential postsynaptic neuron are then determined from a probability profile or , depending on the mutual geodesic distance on the ring . The average interaction between two randomly picked neurons at a distance is A sketch for this construction scheme is depicted in Figure 6A . For the connection probabilities we use a boxcar profile , and , where denotes the Heaviside step function . Neurons with a distance smaller than are connected with a probability , where and depend on the type of the presynaptic neuron . The stability of such a network depends on the radius of the bulk spectrum . Additionally and in contrast to the random network , besides the eigenvalue corresponding to the mean input of a neuron , a number of additional real eigenvalues exist outside the bulk spectrum . A typical spectrum is plotted in Figure 6B . These eigenvalues are particularly pronounced for locally strongly connected rings with large and belong to large scale oscillatory eigenmodes . The sign of these eigenvalues depends on the shape of the interaction profile . For short-range excitation and long-range inhibition ( 6C ) , that is a hat-like profile , these eigenvalues are positive and tend to destabilise the system . For the opposite , or inverted-hat case ( 6D ) , these eigenmodes do not affect stability , therefore stability is determined by the radius of the bulk spectrum . This can be seen as an analogue to the case of net inhibitory input in random networks . As in a random network , the degree distribution of nodes in a ring network is narrow , hence Equation ( 22 ) is a good approximation for the average correlation if the total connection probability is independent on the neuron type , In this case the average correlation does not depend on the specific connectivity profile . However , the full distribution of correlations depends on the connection profile , Figure 6E and F . For localised excitation the eigenvalues of oscillatory modes get close to 1 , rendering the network almost unstable , and many longer paths contribute to correlations . Since for ring networks neighbouring nodes can share a lot of indirect input , while more distant ones do not , this leads to more extreme values for pairwise correlations . We found that in networks with narrow degree distributions average correlations are determined by global parameters like the population sizes and overall connectivity , see Equation ( 22 ) . In networks with broad degree distribution however , the regular-graph approximation is no longer valid . Thus , in such networks the fine structure of the connectivity will , in general , play a role in determining the average correlation . To elucidate this phenomenon , we use a network model characterised by a geometric degree distribution . The fine structure can then be manipulated without altering the overall connectivity . Specifically , the connection statistics of a given node will depend on the out-degree . The network model is defined as follows ( compare Figure 8A ) . Out-degrees of excitatory and inhibitory neurons are chosen from a geometric distribution with a probability where the parameter corresponds to the mean out-degree . The resulting distribution has a mean connection probability of and a long tail . Excitatory neurons are then divided into classes according to their out-degree . We will call neurons with out-degree hubs and the rest non-hubs to distinguish the classes in this specific example . Postsynaptic neurons for non-hubs and inhibitory neurons are chosen randomly from all other neurons . For each hub we fix the fraction of connections that go to other hubs . The number of connections to excitatory neurons is chosen from a binomial distribution with parameter . A number of the postsynaptic neurons are randomly chosen from other hubs , outputs go to non-hub excitatory neurons and connections to randomly chosen inhibitory neurons . By varying between 0 and 1 , excitatory hubs can be chosen to form a more or less densely connected subnetwork . From the cumulative geometric distribution function , , the expected fraction of hubs is , which is about 0 . 35 for . If hubs are preferentially connected to non-hubs , otherwise hubs are more likely connected to each other . By construction the parameters do not depend on . Hence terms with , including common input , are also independent of . The statistics for longer paths are however different . If excitatory hubs preferentially connect to hubs , the number of long paths within the excitatory population increases . The effects on correlations are illustrated in Figure 8 . Densely connected hubs increase average correlations . While the contributions of smaller motifs do not change significantly , from the larger motifs all but the pure chain motif contributions are affected . Different effects can be observed in networks of neurons with patchy connections and non-homogeneous spatial distribution of neuron types . A simple network with patchy connections can be constructed from neurons arranged in a ring . We consider two variants: one where all inhibitory neurons are situated in the same area of the ring , compare Figure 9A , and one where they are randomly distributed over the ring . For each neuron , postsynaptic partners are chosen from a “patch” , a population of neighbouring neurons which is located at a random position , with a probability . If neuron populations are not uniformly distributed , this leads to large variations in single neuron , even if average values are kept fixed . We compare networks where excitatory and inhibitory neurons are spatially separate , Figure 9A , versus randomly mixed populations . In Figure 9B average correlations are compared to correlations in networks with random connectivity . If excitatory and inhibitory neurons are distributed randomly , no significant increase is seen , but if populations are separate , correlations are increased strongly when patches are smaller . In Figure 9C is depicted which network motifs are responsible for the increase of correlations . It can be observed that the difference in correlation is mainly due to differences in contributions of symmetric common input motifs with , and to some extent of nearly symmetric ones ( ) . The reason is that if neurons of the same type receive common input , firing rates of their respective postsynaptic targets will be correlated . If their types differ , their targets receive correlated input of different signs , inducing negatively correlated rate fluctuations . Patchy output connections lead to an increased fraction of postsynaptic neurons of equal type if populations are spatially separated . In this case average correlations are increased . This effect is a direct consequence of the spatial organisation of neurons and connections . The same effect could however be achieved by assuming that single neurons preferentially connect to a specific neuron type . A comparison of motif contributions to correlations , Figures 8C and 9C , shows that different architectures increase correlations via different motifs . Asymmetric motifs play a role in the correlation increase for hubs , but almost none for patchy networks .
The framework of linearly interacting point processes in [38] provides a transparent description of equilibrium rates and correlations . It has been used previously to infer information about direct connectivity from correlations in small networks [44] , as one amongst many other methods , see for example [49] , [50] and references therein . Another application was the study of spike-time dependent plasticity [45] , [51] and , in an extended framework , the description of spike train autocorrelations in mouse retinal ganglion cells [52] . An approach using linearised rate dynamics was applied to describe states of spontaneous activity and correlations in [53] . Correlations in populations of neurons have been studied in a rate model in [36] and in a point process framework in [37] . Hawkes' point process theory allows the treatment of correlations on the level of spike trains as well as the understanding of the relation of complex connectivity patterns to the statistics of pairwise correlations . Although Hawkes' equations are an exact description of interacting point processes only for strictly excitatory interactions , numerical simulations show that predictions are accurate also for networks of excitatory and inhibitory neurons . Hence correlations can be calculated analytically even in effectively inhibitory networks in a wide range of parameters , as has already been proposed in [39] . One should note , however , that for networks with strong inhibition in combination with strong synaptic weights and low external input , low rates are not reproduced well . The activity of cortical neurons is often characterised by low correlations [27] , and can exhibit near-Poissonian spike train statistics [54] with a coefficient of variation near one . In theoretical work , similar activity has been found in balanced networks [41] in a certain input regime [40] . The level and time dependence of external input influences the general state of activity as well as pairwise correlations . In this study we are only concerned with an equilibrium resting state of a local network with asynchronous activity where external input is constant or unknown . We use Poisson processes as a phenomenological description for such a state and do not consider the biophysical mechanisms behind spiking activity , nor the reasons for asynchronous spiking on a network level . However , we found in simulations of networks of integrate and fire neurons of comparable connectivity parameters in an asynchronous-irregular state that correlations can be attributed to a large degree to linear effects of recurrent connectivity , although single neuron dynamics are nonlinear and spike train statistics are not ideally Poissonian ( data not shown ) . Thus , although a linear treatment may seem like a strong simplification , this suggests that Hawkes' theory can be used as a generic linear approximation for the spike dynamics of complex networks of neurons . A similar point has been made in [53] . We quantified correlations by integrated cross-correlation functions in a stationary state . The shape of the resulting correlation functions , which has been treated for example in [30] , [37] , [55] , was not analysed . The advantage is that our results are independent of single neuron properties like the shape of the linear response kernel . Specific connectivity properties that can be described by a graph , as for example reviewed in [3] , can be directly evaluated with respect to their effects on correlations . In Hawkes' framework , taking into account contributions to pairwise correlations from direct interactions , indirect interactions , common input and interactions via longer paths is equivalent to a self-consistent description of correlations . This interpretation helps to derive analytical results for simple networks . Furthermore it allows an understanding of the way in which recurrent connectivity influences correlations via multiple feed-back and feed-forward channels . In particular , we showed why common input and direct input contributions are generally not sufficient to describe correlations quantitatively , even in a linear model . We showed that average correlations in networks with narrow degree distributions are largely independent of specific connectivity patterns . This agrees with results from a recent study [42] , where conductance based neurons in two-dimensional networks with Gaussian connectivity were simulated . There , the degree of single neurons was kept fixed and population averaged correlations were shown to be invariant to different connectivity patterns . For net-inhibitory networks , indirect contributions to correlations effectively reduce average correlations . A similar effect has been described in [20] and in [36] for a rate model . In networks with strong recurrence , characterised by eigenvalues of the connectivity matrix close to one , correlation distributions are strongly influenced by higher-order contributions . In these networks broad distributions of correlations arise . In contrast , in very sparsely connected networks correlations depend mainly on direct connectivity . Can we estimate the importance of recurrence from experimentally accessible parameters ? In [56] the probability of a single extra input spike to generate an additional output spike , corresponding to , has been measured in rat barrel cortex in vivo as 0 . 019 . Additionally , the number of connections made by each neuron was estimated to be about 1500 . We now consider a local network with a fraction of inhibitory neurons of 20% . We assume an inhibitory synaptic weight to balance the excitation , such that . The estimated mean degree is consistent with many different topologies . Let us consider the case of a uniform random network of 15000 neurons with connection probability 0 . 1 . For comparison we also look at a densely connected subnetwork of just 2500 neurons with a connection probability of 0 . 6 . The first model results in a spectral radius for the connectivity matrix , hence falling in the linearly unstable regime . In contrast , the second network displays a spectral radius slightly below one , which indicates linear stability . What can we conclude from this discussion ? In the first place , this crude estimate of the spectral radius suggests that a value in the order of one is not an unrealistic assumption for real neural networks . This would call for a consistent treatment of long-range , higher-order interactions . This view is also supported by simulations of integrate and fire networks [31] , which can yield similarly values for the spectral radius close to one . Our second example , although biologically less realistic , shows the range in which the spectral radius can vary , even if certain network parameters are kept fixed . This highlights the importance of the connectivity structure of local neural networks , as different network architectures can strongly affect the stability of a certain activity state . We addressed ring networks with distance-dependent connection probability . Here , average correlations do not depend on the connectivity profile . However , for densely coupled neighbourhoods very broad correlation distributions can arise . A Mexican hat-like interaction has especially strong effects , since in that case higher-order contributions amplify correlations . This is not surprising since it is known that Mexican hat-like profiles can support large-scale activity patterns [57] . This implies that local inhibition increases network stability and leads to less extreme values for correlations . Distributions of correlations and distance dependence of correlations have been measured experimentally [20] , [21] , but they have not yet been related directly to anatomical connectivity parameters . In [19] , the distance dependence of pairwise correlations as well as higher-order correlations has been measured experimentally . A generalisation of Hawkes' correlation equations in conjunction with the framework of cumulant-correlations discussed in [58] presents a promising route to study structure dependence also of higher-order correlations . A generalisation to two-dimensional networks with distance dependent connectivity could be used to further investigate the relation between neural field models which describe large-scale dynamics [59]–[61] and random networks . However , the analysis using the full connectivity matrix allows to incorporate effects of random connectivity beyond the mean field limit . One example is that stability of networks is not only determined by mean recurrent input , but also by input variance . Pairwise correlations affect activity in pooled spike trains [62] . We found that distance dependence of connectivity creates strongly coupled neighbourhoods and that population signals therefore depend on the connectivity statistics of the network . Such population signals could for example be related to local field potentials . If the degree distribution is wide , networks can be constructed where connection probability depends on the out-degree of postsynaptic neurons . We considered networks where excitatory hubs , defined by a large out-degree , form a more or less densely connected subnetwork . Similar networks have been studied in [63] . In graph-theoretic terms , the connectivity between these hubs influences the assortativity of the network . A commonly used measure is the assortativity coefficient , which is the correlation coefficient between degrees of connected nodes . We calculated a generalised version for weighted networks , the weighted assortativity coefficient [64] . It can assume values between -1 and 1 . Our networks have values between −0 . 22 and −0 . 05 . Negative assortativity values are a consequence of the geometric degree distribution , but networks with more densely connected hubs have a higher coefficient . In our model , more assortative networks exhibit larger correlations than more disassortative ones . This illustrates how differences in higher-order statistics of connectivity can influence correlations , even if low order statistics do not differ . In networks with patchy connections , an increase of correlations can be observed when populations of neurons are spatially non-homogeneous . Some information about how network architecture influences correlations can be obtained from examining contributions of individual motifs . In patchy networks mainly the contributions of symmetric motifs are higher , when excitatory and inhibitory neurons are separated , and therefore responsible for the correlation increase . In networks with hubs also asymmetric motifs play a role . We found that fine-scale structure has important implications for the dynamics of neural networks . Under certain conditions , like narrow degree distributions , local connectivity has surprisingly little influence on global population averages . This suggests the use of mean-field models . On the other hand , broad degree distributions or the existence of connected hubs influence activity also on the population level . Such factors represent , in fact , major determinants of the activity state of a network and , therefore , should be explicitly considered in models of large scale network dynamics . As considerable efforts are dedicated to the construction of detailed connection maps of brains on multiple scales , we believe that the analysis of the influence of detailed connectivity data , possibly with more refined models , has much to contribute to a better understanding of neural dynamics . | Many biological systems have been described as networks whose complex properties influence the behaviour of the system . Correlations of activity in such networks are of interest in a variety of fields , from gene-regulatory networks to neuroscience . Due to novel experimental techniques allowing the recording of the activity of many pairs of neurons and their importance with respect to the functional interpretation of spike data , spike train correlations in neural networks have recently attracted a considerable amount of attention . Although origin and function of these correlations is not known in detail , they are believed to have a fundamental influence on information processing and learning . We present a detailed explanation of how recurrent connectivity induces correlations in local neural networks and how structural features affect their size and distribution . We examine under which conditions network characteristics like distance dependent connectivity , hubs or patches markedly influence correlations and population signals . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"circuit",
"models",
"physics",
"mathematics",
"neural",
"networks",
"computational",
"neuroscience",
"biology",
"biophysics",
"neuroscience"
] | 2011 | How Structure Determines Correlations in Neuronal Networks |
Pathogenic mycobacteria reside in macrophages where they avoid lysosomal targeting and degradation through poorly understood mechanisms proposed to involve arrest of phagosomal maturation at an early endosomal stage . A clear understanding of how this relates to host defenses elicited from various intracellular compartments is also missing and can only be studied using techniques allowing single cell and subcellular analyses . Using confocal imaging of human primary macrophages infected with Mycobacterium avium ( Mav ) we show evidence that Mav phagosomes are not arrested at an early endosomal stage , but mature to a ( LAMP1+/LAMP2+/CD63+ ) late endosomal/phagolysosomal stage where inflammatory signaling and Mav growth restriction is initiated through a mechanism involving Toll-like receptors ( TLR ) 7 and 8 , the adaptor MyD88 and transcription factors NF-κB and IRF-1 . Furthermore , a fraction of the mycobacteria re-establish in a less hostile compartment ( LAMP1-/LAMP2-/CD63- ) where they not only evade destruction , but also recognition by TLRs , growth restriction and inflammatory host responses that could be detrimental for intracellular survival and establishment of chronic infections .
Both Mycobacterium tuberculosis ( Mtb ) and pathogenic non-tuberculous mycobacteria like M . avium ( Mav ) have developed mechanisms to hijack the normal trafficking of phagosomes and use macrophages as a natural habitat and tools of spread in the host [1–3] . However , the spatiotemporal dynamics of how intracellular trafficking of mycobacteria relates to recognition by pattern recognition receptors ( PRRs ) and elicitation of antibacterial responses by host macrophages is not fully understood , in particular for Mav . Mav can establish chronic infections and clinical disease that is hard to treat such as pulmonary disease , lymphadenitis or disseminated infection [4 , 5] . In search of new therapeutic strategies to shorten the treatment of mycobacterial diseases and meet the increasing drug resistance it is thus crucial to understand the minimum infectious unit , the mycobacterium-infected macrophage . Macrophages initiate different destruction programs depending on the particular intruder encountered . The process begins with the engagement of pathogen-recognition receptors ( PRR ) like Toll-like receptors ( TLRs ) and C-type lectins at the plasma membrane , initiating inflammatory signaling followed by phagocytosis and the gradual maturation of the phagosome into a phagolysosome , where bacteria are attempted digested [6] . Different mycobacterial ligands are displayed and activate distinct PRRs present in the various cellular compartments [6 , 7] . TLR2/1/6 are expressed on the plasma membrane where they recognize mycobacterial lipoproteins , proteins and glycolipids [8 , 9] . TLR3 , 7 , 8 and 9 depend on the protein Unc-93 homolog B1 ( UNC93B1 ) to traffic from the endoplasmic reticulum to endolysosomal compartments where they detect nucleic acids [10 , 11] . TLR9 has previously been shown to recognize mycobacterial DNA [12 , 13] . TLR7 and TLR8 recognize single-stranded RNA and RNA degradation products from viruses and bacteria , and polymorphisms in TLR7 and 8 have been associated with increased susceptibility to pulmonary tuberculosis [11 , 14–22] . However , whether TLR7/8 play a role in recognizing mycobacteria is unclear . When activated , TLR-ligand complexes are bridged by sorting adaptors like toll-interleukin 1 receptor ( TIR ) domain-containing adaptor protein ( TIRAP ) or TRIF-related adaptor molecule ( TRAM ) to signaling adapter molecules such as Myeloid differentiation primary response gene 88 ( MyD88 ) and TIR-domain-containing adapter-inducing interferon-β ( TRIF ) in raft domains of various sub-cellular membranes . Signaling cascades culminate in the nuclear translocation of transcription factors like nuclear factor ( NF ) -κB and interferon regulatory factors ( IRFs ) and subsequent production of inflammatory mediators , type I interferons ( IFNs ) and antimicrobial programs [6 , 23 , 24] . The innate inflammatory response is central in anti-mycobacterial host defenses . Mice deficient in MyD88 are susceptible to infections with either Mtb or Mav [25 , 26] , and we have recently demonstrated that negative regulation of inflammatory signaling by a cellular stress sensor , Keap1 , facilitates intracellular growth of Mav in human primary macrophages [27] . However , mycobacteria are intracellular pathogens that are capable of evading most antimicrobial strategies and killing through poorly characterized mechanisms that involve arrest of phagosomal maturation and lysosomal delivery , while access to essential nutrients is retained [1 , 2] . Phagosomal maturation arrest is often suggested to happen at an early endosomal stage for several reasons: Mtb and Mav maintain the pH of their compartment at 6–6 . 5 by excluding vATPase and preventing fusion with lysosomes [28–30] . The mycobacterial compartments also stain positive for transferrin receptor and retain the early endosomal marker Rab5 , but not the late endosomal marker Rab7 , although Rab7 was shown by Koide's group to be transiently present on Mtb phagosomes [31–33] ) . We have previously shown that the Mav compartment ( MavC ) interacts with the Rab11+ recycling endocytic pathway , from where Mav can access transferrin-iron and avoid the antibacterial protein lipocalin 2 [34 , 35] . To dissect from which compartment ( s ) Mav elicits inflammatory signaling , the mechanism ( s ) involved and how these relate to phagosomal maturation arrest and mycobacterial survival , we investigated the spatiotemporal recognition of Mav infection at a sub-cellular level in human primary macrophages using confocal microscopy . We found that Mav-containing phagosomes mature to LAMP1+ late endosomes/phagolysosomes ( MavPLs ) where growth is prevented and inflammatory signaling is activated through TLR7/8 . Furthermore , we found that a fraction of Mav re-established in a less hostile environment , the LAMP1- MavCs , thereby evading TLR activation and subsequent inflammatory signaling , cytokine production and growth restriction . Thus , the MavC represents a safe house for Mav , shielding it both from lysosomal destruction and detection by TLRs .
It is well known that mycobacterial pathogens escape killing by host macrophages and reside in a compartment where they can access nutrition [7 , 30 , 34 , 36] . However , most studies rely on few time points of infection with incomplete and often incongruent results regarding the spatiotemporal location/movement of mycobacteria inside macrophages , leaving the exact trafficking pathways to be elucidated . We infected human monocyte-derived macrophages ( MDMs ) with Mav expressing CFP ( Mav-CFP ) for 10 minutes ( pulse ) and chased for several time points between 5 minutes and 3 days . Confocal microscopy revealed that bacteria are first internalized in early endosomes , as shown by early endosome antigen 1 ( EEA1 ) immuno-labeling ( Fig 1A , bottom-to-top projection of 3D-stack from boxed area is shown in lower panels ) . Profiling the fluorescence intensity along the line ( FAL ) of Mav phagosomes confirmed that mycobacteria were enclosed by EEA1+ membranes ( Fig 1B ) . Quantification over time showed a significant decrease of EEA1 associated with Mav phagosomes already 5 to 15 minutes after uptake ( P <0 . 005 ) , indicative of rapid maturation of Mav phagosomes that remained EEA1 negative from 30–60 minutes onwards ( Fig 1C ) . Pathogenic mycobacteria were previously thought to arrest phagosomal maturation at an early endosomal stage based on the lack of Rab7 recruitment or retainment [31–33] , v-ATPase exclusion and lack of acidification [28] . Mycobacterial phagosomes are however , heterogeneously harboring markers of late endosomes/lysosomes like lysosomal-associated membrane protein 1 ( LAMP1 ) , LAMP2 and LAMP3/CD63 [28 , 29 , 37 , 38] . Correspondingly , we found that Mav phagosomes temporarily acquired LAMP1+ staining ( Fig 1D and 1E , bottom-to-top projection of 3D-stack from boxed area is shown in lower panels , and corresponding FAL profiling in Fig 1F ) . Quantification over time showed that 70% of Mav resided in LAMP1+ late endosomes/phagolysosomes as early as 15 minutes post infection , increasing to more than 90% 4 hours to 1 day post infection , suggesting Mav phagosomes are not arrested at an early endosomal stage but mature normally to a late endosomal/lysosomal stage ( Fig 1G , P <0 . 005 for both time points ) . Interestingly , 2–3 days post infection there was a significant decrease of Mav phagosomes associated with LAMP1 ( Fig 1G ) , further suggesting sorting and escape of an increasing fraction of the mycobacteria from phagolysosomes ( Fig 1E with FAL profiling in Fγ and quantifications over time in G ) . We confirmed the results by staining for LAMP2 and for CD63/LAMP3 , two additional established markers of late endosomes/lysosomes , in combination with LAMP1 ( S1 Fig ) . Mav stays compartmental as it lacks the ESX-1 secretion system suggested to facilitate translocation of Mtb and M . marinum to the cell cytosol [39–41] . These observations are also consistent with our previous findings that Mav is associated with transferrin and Rab11a after 4 days of infection [34] . Mycobacteria in general and Mav in particular are known to survive acidic conditions but may not replicate in fully matured phagolysosomes [1 , 42 , 43] . To assess the acidity of Mav phagosomes we did live-cell imaging of Mav-infected macrophages loaded with Cresyl violet , a membrane-permeant fluorophore that localizes to lysosomes and acidic vacuoles [44] . The fraction of acidic Mav phagosomes increased 1 day and decreased 2 days post infection , following the pattern of LAMP staining ( S2 Fig ) . We then asked whether MavPLs or MavCs would support replication of Mav by quantifying of the number of Mav in LAMP1+ and LAMP1- compartments over time ( Fig 2 ) . The number of Mav in MavPLs did not change whereas the number of Mav in MavCs steadily increased over the 3 days of infection , clearly demonstrating that MavCs , but not MavPLs , support growth of Mav . Infection with live Mav will over time yield a mixture of live and dead bacteria residing in different subcellular compartments , whereas dead Mav should all be routed to LAMP1+ phagolysosomes for degradation . To test this hypothesis we next assessed trafficking of paraformaldehyde ( PFA ) -killed Mav in human primary macrophages ( Fig 3A–3C and S3 Fig ) . Phagosomes with dead Mav rapidly stained positive for LAMP1 as depicted in the bottom-to-top projection of a 3D-stack from the boxed area in Fig 3A and the corresponding FAL profiling ( Fig 3B ) . Quantification over time demonstrated that , contrary to live Mav , PFA-killed Mav was retained in LAMP1+ phagolysosomes throughout the 3-day infection and significantly different from live Mav 3 days post infection ( Figs 3C and 1G ) . We observed no difference in the uptake of live and PFA-killed Mav ( S3B Fig ) . Our results thus suggest that live Mav temporarily reside in phagolysosomes before re-establishing in a new compartment allowing replication . Macrophage inflammatory responses to mycobacterial pathogens are well described [2 , 7 , 8] , but it is largely unknown from which subcellular compartments mycobacteria are recognized and signaling originates . NF-κB is one of the key transcription factors driving inflammatory responses , and we and others have demonstrated that IRF-1 is activated in various infections including mycobacteria [27 , 45–47] . Thus , we first monitored the nuclear translocation of NF-κB ( phospho-p65 ) and IRF-1 4 hours to 3 days post Mav infection of human primary macrophages using confocal microscopy . Both pathways were significantly activated by Mav with IRF-1 slightly delayed compared to NF-κB ( S4 Fig ) . Even single cell analyses may not be sufficient to fully relate trafficking to inflammatory signaling if the cell is infected with more than one microbe . We thus conducted 3D confocal microscopy of macrophages harboring only single bacteria and assayed for nuclear translocation of IRF-1 and Mav association with LAMP1 ( Fig 4A–4D ) . Every cell was first scored as positive or negative for nuclear translocation of IRF-1 , whereafter the location of the mycobacterium was determined ( LAMP1+ or LAMP1- ) . IRF-1 was chosen over NF-κB simply because the NF-κB Ab that works for immunofluorescence and nuclear translocation assays is not compatible with the protocol we need to use for LAMP1 staining , whereas the IRF-1 Ab is . More than 90% of the cells that were positive for nuclear IRF-1 harbored Mav associated with LAMP1+ membranes at all time points measured over the 3 days infection ( Fig 4A , 4B and 4E ) . Conversely , non-activated cells defined by absent nuclear IRF-1 harbored Mav in compartments devoid of LAMP1 ( Fig 4C and 4D ) , suggesting that Mav induces inflammatory signaling from LAMP1+ phagolysosomal compartments , and not from MavCs . Since PFA-killed Mav was retained in LAMP1+ phagolysosomes ( Fig 3 ) the inflammatory response in cells with only dead Mav should reflect signaling mainly from this compartment . We thus compared the temporal activation by live and PFA-killed Mav of IRF-1 nuclear translocation and cytokine secretion from macrophages ( Fig 4F and S4D Fig ) . PFA-killed Mav induced nuclear translocation of IRF-1 and the response was sustained over the 3-day time course in comparison to infection with live Mav , where the percentage of Mav-infected cells with nuclear IRF-1 declined with time ( Fig 4F , P<0 . 005 ) . There was no difference in uptake of live and PFA-killed Mav ( S3B Fig ) . The nuclear translocation assay quantifies the frequency of activated Mav-infected macrophages but does not necessarily inform about the nature or magnitude of the various responses that usually involve multiple transcription factors , e . g . for cytokine production . Multiplex ELISA on cell supernatants revealed that , despite donor variations , PFA-killed Mav induced increased levels of secreted TNF , IL-6 , IL-10 and IP-10 , but not IL-8 , compared to live Mav ( S4D Fig ) . Taken together , our results suggest that PFA-killed Mav remains in phagolysosomes and induces strong and sustained inflammatory signaling whereas the fraction of live Mav residing in phagolysosomes decreases over time , coinciding with reduced inflammatory responses induced by Mav incapable of escaping . This also strengthens the evidence that Mav induces inflammatory signaling from LAMP1+ phagolysosomes and not LAMP1- MavCs in human primary macrophages . To further validate the compartmental origin of Mav recognition and induction of host responses we next investigated the temporal activation and localization of the TLR signaling adaptor , MyD88 . Immuno-staining of MyD88 in macrophages harboring live or PFA-killed Mav revealed the formation of aggregates ( Myddosomes [48] as shown in Fig 5A ) . Quantification of Myddosomes by image segmentation based on particle size and fluorescence intensity displayed a pattern similar to IRF-1 activation: live Mav increased Myddosome formation over the first 24 hours followed by a decrease 2 and 3 days post infection , whereas the number of Myddosomes per cell increased substantially at day 3 in macrophages harboring PFA-killed Mav ( Fig 5B ) . Co-localization analysis revealed that about 20% of Mav were associated with MyD88 4 hours post infection , decreasing over time for live Mav and , conversely , increasing over time for PFA-killed Mav ( Fig 5C , P <0 . 005 ) . Concomitant staining for LAMP1 revealed a specific recruitment of MyD88 to LAMP1+ Mav phagolysosomes as depicted in the bottom-to-top projection of the 3D-stack marked with an asterisk in Fig 5D ( box with solid lines ) and corresponding FAL profiling ( Fig 5E ) . Importantly , we were not able to detect MyD88 associated with LAMP1- MavCs ( Fig 5D , bottom-to-top projection of a 3D-stack from the dashed box , and corresponding FAL profiling ( Fig 5F ) . Moreover , in LAMP1+MyD88+ phagolysosomes the morphology of Mav often appeared coccoid rather than rod-shaped , indicative of partial degradation , whereas mycobacteria residing in LAMP1-MyD88- MavCs appeared intact and forming elongated chains with visible septa indicating growth and replication ( Fig 5D ) . This is supported by results in Fig 2 showing that Mav numbers increased in MavCs over time , but not in MavPLs . Nuclear translocation of IRF-1 was markedly reduced over the first 3 days of infection in macrophages treated with siRNA to MyD88 , suggesting MyD88 is conveying Mav-induced responses from the LAMP1+ phagolysosome ( Fig 5G and S6 Fig ) . Our data thus suggest that Mav induces MyD88 recruitment and signaling only from LAMP1+ phagolysosomes , and by re-establishing in LAMP1- compartments Mav not only escapes lysosomal destruction but also evades recognition and MyD88-mediated host responses . To confirm a functional role for MyD88 in Mav-induced responses we next assessed secretion of inflammatory cytokines 4 hours to 3 days post infection of macrophages treated with siRNA to MyD88 or non-targeted control ( NTC ) ( Fig 6A , S6 and S7 Figs ) . Mav uptake was similar in siMyD88 and siNTC treated macrophages ( S8A Fig ) . Despite variability between donors we found that TNF-α , IL-6 and IL-10 secretion was significantly decreased in siMyD88 treated cells ( Fig 6A ) . We calculated the area under the curve ( AUC ) for each cytokine/donor couple to obtain an estimate of the total amount of secreted cytokines induced by the infection . Knocking down MyD88 considerably reduced the AUCs for TNF-α , IL-6 and IL-10 in 8/10 , 7/10 and 8/10 of the donors , respectively , whereas IL-8 secretion was not affected ( S7 Fig , cutoff set at 25% variation ) . The results further support that MyD88 is conveying Mav-induced responses from the phagolysosome , possibly initiated by Mav engaging endolysosomal TLRs . Of the endolysosomal TLRs only TLR9 has been shown to respond to mycobacteria [21 , 49] . We also found TLR7 and 8 to be expressed in human primary macrophages ( S6 Fig and [19] ) . Both TLR7 and 8 recognizes ssRNA in a species-specific manner [50] and we and others have shown that human TLR8 responds to bacterial RNA [11 , 15 , 19 , 20 , 51] . We next used siRNA to knock down TLR7 , TLR8 , and UNC93B1 in human primary macrophages , achieving an average reduction in mRNA levels of 80% ( S6 and S7 Figs ) . UNC93B1 is needed for transport to and activity of TLR7/8 in endolysosomes [11] . Similar to what we observed with MyD88 , knocking down TLR8 consistently and significantly reduced Mav-induced secretion of TNF-α , IL-6 and IL-10 by macrophages in 8/10 , 7/10 and 8/10 of the donors ( Fig 6B with accompanying AUCs in S7 Fig ) . A more variable reduction in cytokine levels was seen when knocking down TLR7 ( 6/9 , 5/9 , 7/9 ) or UNC93B1 ( 5/9 , 6/9 , 7/9 ) , with IL-10 responses most consistently reduced . Interestingly , siTLR8 more consistently reduced cytokine responses compared to siTLR7 despite similar knockdown efficiencies , thus our data indicate non-redundancy between TLR7 and 8 in responding to Mav . As for MyD88 , IL-8 responses were not affected by knocking down TLR7 , 8 or UNC93B1 ( Fig 6B and S7 Fig ) . A similar trend was also seen in response to PFA-killed Mav , with reduced cytokine responses when knocking down MyD88 , TLR7/8 or UNC93B ( S5 Fig ) . There was no difference in Mav uptake in target and control siRNA treated cells , indicating that none of these proteins are central for Mav phagocytosis ( S8A Fig ) . Knockdown of the NOD1 or STING did not impact on Mav-induced secretion of TNF-α , IL-6 , IL-10 or IL-8 ( Fig 6B and S8D Fig ) , which was not surprising since Mav lacks the ESX-1 secretion system and does not translocate into the cytosol where these PRRs are located [6 , 40] . We confirmed that each siRNA was specific by assessing the expression of both targeted and non-targeted genes ( S6D and S8C Figs ) , and functional by challenging knockdown macrophages with pure ligands expected to activate the particular pathway ( s ) or act through other PRRs ( S8E and S8F Fig ) . As expected , LPS-induced TNF/IL-6 was strongly impaired in cells treated with siTLR4 , but largely unaffected by siTLR7 , siTLR8 or siUNC93B1 ( S8E Fig ) . Correspondingly , responses to CL264 were reduced when knocking down TLR7 and normal in siTLR8 and siSTING treated macrophages ( S8E Fig ) . Mav-induced nuclear translocation of NF-κB and IRF-1 was substantially reduced by knocking down TLR7 , 8 or UNC93B1 throughout the time course of 4 hours to 3 days ( Fig 6C ) . The TLR8 ligand CL75 also induced IRF-1 nuclear translocation , confirming IRF-1 is activated by TLR8 engagement ( S8F Fig ) . We have so far shown evidence that nuclear translocation of transcription factors is only activated from LAMP1+ MavPLs , MyD88 is only recruited to MavPLs , and MyD88 is involved in activating cytokine production . We next stained for TNF-α and LAMP1 in Mav-infected macrophages and found that 80% of the infected cells produced TNF-α , and of these , more than 90% harbored Mav in LAMP1+ MavPLs whereas less than 10% of the TNF-producing cells had Mav in LAMP1- MavCs ( Fig 7A–7C ) . To detect intracellular TNF-α we used a transport inhibitor cocktail that blocks transport from the endoplasmic reticulum ( ER ) to Golgi , explaining the accumulation of TNF-α in perinuclear ER ( Fig 7A ) . Taken together these results suggest that Mav engages TLR8 and also TLR7 in phagolysosomes , resulting in recruitment of MyD88 , inflammatory signaling and secretion of cytokines . TLRs signal via the canonical IKK-complex and we recently showed the importance of IKK-β in controlling intracellular growth of Mav in human primary macrophages [27] . To examine if TLR7/8/MyD88 signaling affects intracellular Mav growth we next quantified Mav-CFP intensities across infected macrophages at the single cell level 4 hours and 3 days post infection ( Fig 8 ) . The CFP-level and thus the total number of Mav increased ten-fold over 3 days of infection ( Fig 8B and 8C ) . 4 hours post infection there was no difference in CFP-levels between target or control siRNA treated cells , whereas the CFP-levels were significantly higher in cells treated with siMyD88 , siTLR7 or siTLR8 compared to non-targeting control 3 days post infection , suggesting that TLR7/8/MyD88 signaling is central in controlling intracellular growth of Mav in human primary macrophages ( Fig 8B and 8C ) . A working model summarizing our findings in the paper is shown in Fig 8D .
Despite decades of research on mycobacterial phagosome biogenesis the process is incompletely understood , including how mycobacteria interact with and evoke host responses from the various subcellular compartments inside macrophages . Both MavCs and MtbCs are generally found to retain characteristics of early endosomes with elevated pH , unprocessed cathepsins and access to transferrin , suggesting that phagosomal maturation is arrested at an early stage [28–31 , 36 , 37 , 52] . Using confocal microscopy on single-infected human primary macrophages we have revealed here that Mav phagosomes are not arrested at an early endosomal stage but mature normally to LAMP1+ phagolysosomes where Mav is degraded and recognized by TLR7/TLR8 , eliciting host responses by recruitment of MyD88 to the phagolysosomal compartment . A fraction of Mav escapes phagolysosomal degradation by re-establishing in a new compartment where it replicates and , most importantly , to where MyD88 is not recruited and inflammatory signaling is not evoked . This way Mav eventually evades both degradation and inflammatory host responses that could be detrimental for intracellular survival and establishment of chronic infection . Our findings are not all contradictory to published data: several studies show LAMP1/LAMP2 or CD63 staining of mycobacterial compartments to varying degrees [28–31 , 36–38 , 52 , 53] . These have often been interpreted to contain dead bacteria , or in the cases where the proton-ATPase is also lacking , not considered to be fully phagolysosomal in nature . However , spearheaded by electron microscopy studies of macrophages infected with Mav , Mtb or M . bovis BCG , compelling evidence indicates that mycobacteria transit into late endosomes/lysosomes before re-establishing a new compartment [33 , 38 , 54–57] . This new compartment would be equivalent to our MavC where we see no LAMP1/LAMP2/CD63 , no MyD88 recruitment or signaling , but mycobacterial division . It has been observed that when mycobacteria divide , the vacuole appears to divide with them , sequestering each bacillus in a separate tight vacuole [52 , 54 , 58] . Tight membrane apposition may facilitate their control of fusion with other vacuoles within the cell and mycobacterial cell wall lipids are shown to be critical in maintaining the MtbCs [29 , 38] . Our previous studies showed that MavCs are positive for transferrin and Rab11a suggesting communication with recycling endosomes [34] , also in agreement with literature [36 , 53] . The lack of MyD88 recruitment and inflammatory signaling from the MavC suggest this compartment is key to survival and growth of Mav inside macrophages , and thus an attractive therapeutic target . However , several questions remain to be answered: How are mycobacteria sorted into MavCs and prevent them from fusing with lysosomes ? Moreover , given McDonough and others are right in their observations that MavCs and MtbCs bud off from the phagolysosome: are TLRs or sorting TLR adaptor TIRAP [59] excluded from the membrane of forming MavCs , or is the absence of Mav recognition and MyD88 recruitment to the MavC due to lack of either ligand exposure ( no RNA or DNA release ) or ligand binding to TLRs ? For the first question our results suggest Mav needs to be viable to establish in MavCs since PFA-killed Mav were retained in phagolysosomes . This is also well described in literature and stress factors like low pH , oxidative stress and nutrient deficiency in the maturing phagosome may act as important cues to the bacteria enabling them to mount counteractive mechanisms and escape [1 , 29] . Several virulence factors are needed for intraphagosomal survival [57 , 60–66] , but only a few secreted phosphatases ( PtpA , SapM [67–69] ) and a phospho-inositide 3 kinase mimic ( MAV_2941 [70] ) are so far shown to directly impact on the fusogenicity of the compartment . How these virulence factors access the cytosol is unclear , although ESAT-6 , an Mtb antigen that is secreted through the ESX-1 secretion system , is shown to have pore-forming activity . ESX-1 is also needed for the release of Mtb DNA and translocation of Mtb into the cytosol during infection [39 , 71] . Mav lacks the ESX-1 secretion system and does not translocate into the cytosol [28 , 40 , 72] . However , Bermudez' group recently identified an Mav oligopeptide transporter , OppA , that delivers MAV_2941 into the macrophage cytosol to interfere with phagosomal maturation [70 , 73] . We did not find STING or NOD1 to be involved in Mav inflammatory responses , nor did we observe any macrophage cell death indicative of inflammasome activation and pyroptosis , suggesting little to no Mav ligands were translocated into the cytosol to activate the cytosolic surveillance system . Our results thus support a scenario where Mav phagosomes mature to a late endosomal/lysosomal stage where decreasing pH and other intraphagosomal cues induce stress-responses in the mycobacteria facilitating their re-establishment in the MavC . Our discovery that pathogenic mycobacteria hide and divide in a compartment that is devoid of TLR signaling is significant . For endosomal TLR signaling to happen , the receptors need to be trafficked to the right compartment and proteolytically cleaved to bind ligands . Activated ligand-TLR dimer complexes are recognized by sorting adaptors TIRAP or TRAM that recruit signaling adaptors MyD88 or TRAM to form a supramolecular organizing center ( e . g . Myddosomes ) conveying signaling [11 , 23 , 24] . UNC93B1 mediates the trafficking of TLR3 , 7 , 8 and 9 to endosomes and the absence of UNC93B1 abrogates signaling [11] , which is what we also observed for Mav in siUNC93B1 treated macrophages . The mechanism by how TLR7/8 are recruited to Mav containing phagolysosomes is elusive , if it is induced by Mav recognition on the plasma membrane or during phagocytosis , and if it involves regulated trafficking and fusion of preformed TLR7/8 endosomes with the Mav containing phagosomes . Human TLR7 and TLR8 are processed at neutral pH and may be present in MavCs , but receptor engagement most likely happens in lysosomal compartments since both TLR7 and 8 respond to ssRNA degradation products ( ribonucleosides and oligoribonucleotides ) and not intact ssRNA , indicating RNA processing should be required [14–18] . Regarding the second question , this could explain why Mav does not signal from the MavC , which is not acidified and thus have limited proteolytic activity , together with the fact that live and intact Mav probably does not release much nucleic acids . We did not observe Myddosome formation on MavCs , and work is in progress to verify if signaling is not emerging from MavCs because TLRs are not activated or if TLRs or sorting/bridging adaptors are excluded from the compartment . Instead we found that Mav was degraded and engaged TLR7/8/MyD88 in the phagolysosomal compartment , triggering Myddosome formation , cytokine release and Mav growth restriction . Concomitant engagement by both TLR7 and TLR8 is not surprising since one would expect most microbes to contain ligands for both receptors . TLR7 and 8 recognize ssRNA degradation products together with guanosine or uridine nucleosides , respectively , in a species-specific manner [14–18 , 50] . However , it is unclear why two ssRNA receptors have evolved and , whereas some cells like human pDCs predominantly express TLR7 , we found both TLR7 and TLR8 to be present in human primary macrophages ( here and [19] ) . Furthermore , although TLR8 was more reliably involved in activating Mav cytokine responses , knocking down either receptor resulted in substantially reduced cytokine responses by most donors and increased intracellular growth of Mav . Our findings are partly in agreement with a study by Mancuso et al . describing activation by group B streptococci of a TLR7/MyD88/IRF1 pathway driving IFNβ expression in mouse conventional DCs , and where MyD88 was recruited to LAMP1+ compartments containing partly digested bacteria [22] . We found that Mav engagement of TLR7/8/MyD88 activated nuclear translocation of NF-κB and IRF-1 , and induction of inflammatory cytokines . Our previous studies have indicated a possible involvement of IRF-1 and IRF-5 in Mav-induced IFNβ-responses [27] , and Staphylococcus aureus RNA induces IFNβ in a TLR8/IRF-5 dependent manner [19] . However , Mav is a weak inducer of type I IFNs in human primary macrophages despite pronounced IRF-1-activation , and discrepancies could thus arise from differences in species , cell types , and the pathogens investigated . IRF-1 is strongly induced by interferons but can also be directly recruited to promoter elements of TNF , IL-6 and IL-12β by LPS stimulation [47 , 74] . IRF-1 is also involved in TLR2-mediated induction of CCL5 in mouse macrophages [75] and viral induction of type III IFNs in epithelial cells [76] . We show activation of IRF-1 nuclear translocation by the TLR8 ligand CL75 and have previously shown that Mav-induced IRF-1 involves IKKβ , but its role in mycobacterial infections remains to be established . Heterogeneity is usually reported for mycobacterial phagosomes regarding membrane markers and organelle identity , underlining the need for careful spatiotemporal studies of mycobacterial trafficking in the subcellular space of single cells . These are hampered by the scarcity of good antibodies for use in primary cells and the dangers associated with overexpression of fluorescently tagged proteins in cell lines . Moreover , it is notoriously difficult to probe the viability of single mycobacteria: slow growth rates , thick and impermeable cell wall and apparent persistence of reporter proteins even post mortem is commonly experienced , making it difficult to discriminate recently dead from live bacteria . Based on our findings we expect the phagolysosome to contain a mixture of live and dead Mav at all times . Over time we observed morphological changes indicative of mycobacterial degradation in LAMP1+ phagolysosomes and , conversely , elongated mycobacteria in division in LAMP1- MavCs , but never the other way around . We also confirmed that Mav replicates in MavCs and not in MavPLs by counting bacteria in LAMP-stained compartments over time . We thus feel confident that the MavC supports mycobacterial growth . Still it remains to be proved exactly how Mav modulates host membranes to prevent fusion with lysosomes and either excludes TLRs or interferes with recognition and signaling to facilitate survival .
The following primary antibodies were used: anti-EEA1 ( Santa-Cruz , H300 ) , anti-IRF-1 ( Santa-Cruz , C20 ) , anti phospho-p65 ( CST , XP ( R ) D14E12 ) , anti-LAMP1 mouse monoclonal ( Santa-Cruz , H4A3 ) and rabbit polyclonal ( Abcam , ab24170 ) , anti-CD63 ( Abcam , ab59479 ) and anti-MyD88 ( Abcam , ab2064 ) . Secondary antibodies Alexa Fluor 555 or Alexa Fluor 647–conjugated goat anti-rabbit or anti-mouse IgG and the nuclear dye Hoechst 33342 were from Life Technologies . Ultrapure LPS ( E . coli 0111:B4 ) , TLR7 ligand CL264 and TLR8 ligand CL75 were purchased from Invivogen . Human peripheral blood mononuclear cells ( PBMCs ) were isolated from buffy coats obtained from the Blood Bank , St Olav’s Hospital , Trondheim , Norway , by density centrifugation using Lymphoprep ( Axis-shield ) . Monocyte-derived macrophages ( MDMs ) were generated by plastic adherence for 1h in complete RPMI 1640 ( 680 μM L-Glutamine and 10 mM Hepes , GIBCO ) supplemented with 5% pooled human serum ( The Blood Bank , St Olavs hospital ) at 37°C and 5% CO2 . After three washing steps with Hank’s Balanced Salt solution ( GIBCO ) , monocytes were cultivated for 6 days with a change of medium at day 3 in RPMI 1640/10% human serum and 10 ng/ml recombinant M-CSF ( R&D Systems ) . At day 6 the medium was replaced with RPMI 1640/10% human serum and used for experiments on day 7 . Transfection with siRNA was performed using siLentFect Lipid Reagent ( Bio-Rad ) for RNAi according to the manufacturer’s protocol . Gene knockdown was evaluated by reverse transcription quantitative PCR ( RT-qPCR ) . MyD88 , UNC93B1 , TLR4 , TLR7 , TLR8 , NOD1 and STING pooled ON-TARGETplus human siRNAs ( Dharmacon/Thermo Scientific ) were used to target MyD88 , UNC93B1 , TLR4 , TLR7 , TLR8 , NOD1 and STING . MDMs were treated with 20 nM siRNA two times ( day -4 and day -2 ) before changing to fresh medium ( RPMI 1640/10% human serum ) , rest for 1–2 hours and challenge with TLR ligands or Mav . MDMs were washed with cold PBS and lysed in buffer RLT ( Qiagen ) with 1% β-mercaptoethanol . Total RNA was extracted using RNeasy Mini kit and QIAcube according to the manufacturer’s protocol ( Qiagen ) , including DNase I digestion ( RNase-free DNase set ) . The samples included in the study presented an OD260/280 ratio ~ 2 assessed using a ND-1000 spectrophotometer ( NanoDrop ) . cDNA was synthetized from normalized amounts of RNA using the High-Capacity RNA-to-cDNA kit according to manufacturer’s recommendations ( Applied Biosystems ) . qPCR reactions were performed in 20 μl total volume with 10 ng cDNA input , PerfeCta qPCR FastMix , UNG , ROX ( Quanta Biosciences ) and TaqMan Gene Expression Assays ( Applied Biosystems ) : GAPDH ( Hs99999905_m1 ) , TLR7 ( Hs019333259_s1 ) , TLR8 ( Hs00152972_m1 ) , MyD88 ( Hs00182082_m1 ) , UNC93B1 ( Hs00276771_m1 ) , STING ( Hs00736958_m1 ) , NOD1 ( Hs00196075_m1 ) . The targeted genes were amplified with a StepOnePlus Real-Time PCR System and relative quantities of gene expression were calculated using the comparative CT method with GAPDH gene expression as endogenous control . Mav clone 104 expressing CFP was cultured in liquid Middlebrook 7H9 medium ( Difco/Becton Dickinson ) supplemented with 0 . 5% glycerol , 0 . 05% Tween 80 and 10% albumin dextrose catalase . Cultures were maintained at log phase growth ( optical density between 0 . 3 and 0 . 6 measured at 600 nm , OD600 ) in a 180 rpm shaking incubator at 37°C for a maximum of 5 days . At the day of infection , bacteria were washed with PBS , sonicated and passed through a Gauge 15 needle to ensure single-cell suspension before challenging day 7 MDMs for 10 minutes at a multiplicity of infection of 10 . MDMs were subsequently washed and maintained in culture for the appropriate time . Time course of infection was carried out in a backward manner ( i . e . first infection corresponding to day 3 post infection , second infection corresponding to day 2 post infection etc ) in order to process all samples for immunostaining at the same time . In some experiments MDMs were challenged with TLR ligands for 4 hours at the following concentration: ultrapure LPS ( TLR4 , 25 ng/ml ) , CL75 ( TLR8 , 1 μg/ml ) , CL264 ( TLR7 , 2 . 5 μg/ml ) . Human MDMs cultivated on glass-bottomed 96 well plates ( IBL ) were fixed and permeabilized using a standard protocol as previously described [77] . Briefly , cells were fixed in 4% PFA for 10 min and then incubated in NH4Cl for 10 min to quench PFA-induced auto-fluorescence prior to permeabilization with PBS/0 . 05% Saponin . Cells were next incubated for at least 90 min in PBS/0 . 05% Saponin/20% human serum to reduce non-specific binding before staining with primary antibodies ( 1 μg/ml ) in PBS/0 . 05% Saponin/1% human serum over night at 4°C . Cells were washed with PBS/0 . 05% Saponin/1% human serum and incubated with secondary antibodies for 45 min at room temperature , washed again and stored at 4°C in PBS containing Hoechst for nuclear staining . For resolution limited imaging , MDMs cultivated on glass-bottomed 96 well plates were imaged either with a Zeiss LSM510 confocal microscope with 63x NA = 1 . 4 objective ( Carl Zeiss Micro-imaging Inc . ) or with a Leica SP8 confocal microscope ( Leica Microsystem ) with 63x NA = 1 . 4 objective . Emissions were collected using PMT detectors ( Zeiss LSM510 ) and HyD detectors ( Leica SP8 ) . In both cases the following acquisition parameters were used: numerical zoom set to reach a pixel size of approximately 90 nm , frame averaging 4 , and 3D acquisition to collect the entire cell with a z step of 0 . 3 μm . Each fluorophore was recorded using sequential acquisition to minimize cross excitation and channel bleed through . On Zeiss LSM510 Hoechst was excited with a 405 nm diode laser and emission was collected through a 420–480 IR band-pass filter . CFP was excited with a 458 nm Argon laser and emissions were collected through a 470–500 band-pass filter . Alexa-555/633 fluorophores were excited with 565/633 nm HeNe lasers and emissions were collected through a 575–615 band-pass filter and a 650 long-pass filter , respectively . On the Leica SP8 microscope , Hoechst and CFP were excited using a 405 nm diode laser and emissions collected by adjusting detector windows as follows: Hoechst 420–470 nm and CFP 480–540 nm . Alexa-555/633 fluorophores were excited with 553/638 nm diode lasers and collection windows were set between 560–630 nm and 650–750 nm , respectively . The same parameters were used for non-resolution limited imaging with the following changes: 20x NA = 0 . 5 or 40x NA = 1 . 3 objectives , no numerical zoom was applied , frame averaging 2 , 3D acquisition was recorded with a z step of 0 . 7 μm . For phagosomal acidification , MDMs previously infected with Mav-CFP were loaded with Cresyl Violet ( Sigma-Aldrich ) 1 μM diluted in complete media for 5 min at 37°C and then washed with fresh media . Cells were imaged with Zeiss LSM510 confocal microscope with 63x NA = 1 . 4 objective ( Carl Zeiss Micro-imaging Inc . ) equipped with a 37°C encagement . Cresyl Violet was excited with 547 laser diode and emissions were collected through a 575 long-pass filter . Images were analyzed with Image J ( NIH ) and Leica Analyzing Suite ( Leica Microsystems ) . High-resolution images were screened in 3D , Mav localization in early endosomes ( EEA1 ) and lysosomes ( LAMP1 , CD63 ) was assessed by quantifying fluorescence levels along a line manually placed to intersect bacterial phagosomes using the “Plot Profile” tool in Image J . For quantification of nuclear translocation , single fluorescence images and merged images were synchronized and nuclear localization of the transcription factors assessed based on Hoechst staining . Cells on edges of the field of view ( i . e . not completely acquired in the FOV ) were taken into account only if bacteria and nuclei were detected . Cells were classified as follows: the total number of cells , number of infected cells and number of activated cells ( i . e . with transcription factor in the nucleus ) . The percentage of each category for one specific time point and one specific donor was calculated for each field of view and compared to the control . For the MyD88 aggregation assay , images were treated as follows: MyD88 nuclear staining , if any , was removed; aggregates were segmented with the “Threshold” algorithm from Image J using Max “Entropy option” and with the fluorescence intensity defined by the control condition . Aggregates were counted using the “Find Maxima” application . 3D stacks were projected using the “average” setting . Cells containing Mav as well as background regions were manually encircled and CFP levels ( integrated density ) were measured . Corrected Total Cell Fluorescence ( CTCF ) was calculated using the following formula: CTCF = Integrated Density– ( Area of selected cell x Mean fluorescence of background ) . A minimum of 250 infected cells per condition and per donor were counted . Supernatants from human MDMs challenged with Mav or TLR ligands were collected and cytokine secretion profiles for GM-CSF , GRO-α , IFN-α/β/γ , IL-1α/β , IL-1 Receptor Antagonist , IL-2 IL-4 , IL-5 , IL-6 , IL-8 , IL-10 , IL-12p70 , IL-17A , IL-18 , IL-23 , IL-31 , IP-10 , MCP-1 , RANTES , SDF1-α and TNF-α/β were analyzed according to the manufacturer's instruction using the ProcartaPlex Human Cytokine and Chemokine panel ( Affimetrix , eBioscience ) . For each donor , the Fisher Exact Test was used to compare measurements at different time points with the control , or to compare live with PFA-killed Mav responses at a given time point . For nuclear translocation experiments , two-tailed t-test was used on donor average values . For intracellular localization exact and single donor analysis P values were calculated using the Fisher Exact Test . Areas under the curve were calculated for each cytokine/donor couple using the Prism software ( Graph Pad ) and P values were calculated using Wilcoxon matched-pairs rank test . Significant P values were set as follows: * <0 . 05 , ** <0 . 01 and *** <0 . 005 . | Mycobacterium avium is increasingly reported as a causative agent of non-tuberculous disease in immunocompromised patients and in individuals with underlying disease or using immunosuppressant drugs , with prevalence often higher than the more pathogenic M . tuberculosis in developed countries . Both M . avium and M . tuberculosis cause persistent infections by surviving inside host macrophages . Here , we identify from which compartment M . avium evoke inflammatory signaling in human primary macrophages , and the pattern-recognition receptors involved . In essence , we present three key findings: 1 ) M . avium phagosomes are not arrested at an early endosomal stage , but rather mature normally into phagolysosomes from where a fraction of the bacteria escape and re-establish in a new compartment . 2 ) In addition to avoiding degradation in phagolysosomes , by escaping M . avium also evade inflammatory signaling . 3 ) M . avium unable to escape is degraded in phagolysosomes and recognized by Toll-like receptors 7 and 8 . Our results can contribute to new understanding of intracellular infections , and thus have vital clinical implications for development of novel anti-microbial strategies and host-targeted therapy to mycobacterial and other infectious diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"vesicles",
"pathology",
"and",
"laboratory",
"medicine",
"immune",
"physiology",
"cytokines",
"gene",
"regulation",
"immunology",
"light",
"microscopy",
"immune",
"receptor",
"signaling",
"developmental",
"biology",
"signs",
"and",
"symptoms",
"phagosomes",
"membrane",
"receptor",
"signaling",
"microscopy",
"molecular",
"development",
"confocal",
"microscopy",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"immune",
"system",
"proteins",
"white",
"blood",
"cells",
"inflammation",
"animal",
"cells",
"proteins",
"gene",
"expression",
"immune",
"response",
"immune",
"system",
"toll-like",
"receptors",
"biochemistry",
"rna",
"signal",
"transduction",
"diagnostic",
"medicine",
"cell",
"biology",
"nucleic",
"acids",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"immune",
"receptors",
"macrophages",
"non-coding",
"rna",
"cell",
"signaling"
] | 2017 | Persistent mycobacteria evade an antibacterial program mediated by phagolysosomal TLR7/8/MyD88 in human primary macrophages |
Flubendazole , originally developed to treat infections with intestinal nematodes , has been shown to be efficacious in animal models of filarial infections . For treatment of filarial nematodes , systemic exposure is needed . For this purpose , an orally bioavailable amorphous solid dispersion ( ASD ) formulation of flubendazole was developed . As this formulation results in improved systemic absorption , the pharmacokinetic and toxicological profile of the flubendazole ASD formulation have been assessed to ensure human safety before clinical trials could be initiated . Safety pharmacology , toxicity and genotoxicity studies have been conducted with the flubendazole ASD formulation . In animals , flubendazole has good oral bioavailability from an ASD formulation ranging from 15% in dogs , 27% in rats to more than 100% in jirds . In in vivo toxicity studies with the ASD formulation , high systemic exposure to flubendazole and its main metabolites was reached . Flubendazole , up to high peak plasma concentrations , does not induce Cmax related effects in CNS or cardiovascular system . In repeated dose toxicity studies in rats and dogs , flubendazole-induced changes were observed in haematological , lymphoid and gastrointestinal systems and in testes . In dogs , the liver was an additional target organ . Upon treatment cessation , at least partial recovery was observed for these changes in dogs . In rats , the No Observed Adverse Effect Level ( NOAEL ) was 5 mg ( as base ) /kg body weight/day ( mg eq . /kg/day ) in males and 2 . 5 mg eq . /kg/day in females . In dogs , the NOAEL was lower than 20 mg eq . /kg/day . Regarding genotoxicity , flubendazole was negative in the Ames test , but positive in the in vivo micronucleus test . Based on these results , in combination with previously described genotoxicity and reproductive toxicity data and the outcome of the preclinical efficacy studies , it was concluded that no flubendazole treatment regimen can be selected that would provide efficacy in humans at safe exposure .
Onchocerciasis is a neglected tropical disease caused by the parasitic worm species Onchocerca volvulus , which spreads through bites from infected black flies . The presence of larvae in the skin causes several symptoms , including intolerable itching . Larvae can migrate to the eye and cause decreased vision and blindness . Adult worms live in nodules in the skin , can survive for 10–15 years and produce thousands of larvae per day . Current treatment of onchocerciasis relies on three drugs , albendazole , ivermectine and diethylcarbamazine , agents working primarily against microfilarial stages . In tropical medicine , there is a need for a macrofilaricidal drug that safely kills adult filarial worms . [1 , 2] Flubendazole , originally discovered and developed by researcher Dr . Paul Janssen and his team , was first approved for human use in 1980 to treat soil transmitted helminths ( STH ) , also known as intestinal worms . This methylcarbamate benzimidazole anthelmintic has been shown to have a marked macrofilaricidal effect on many filarial species . [2] Oral flubendazole formulations commercialized for the treatment of STH ( trademarks Flubenol , Fluvermal and others ) are very poorly absorbed . For treatment of STH , flubendazole acts locally in the gut . In vivo activity of flubendazole against a variety of filariid species has been reported in animals and man after parenteral administration indicating systemic exposure is needed for the treatment of onchocerciasis . For this purpose , an orally bioavailable amorphous solid dispersion ( ASD ) formulation was developed . ASD is an approach to formulate poorly water-soluble drugs in the amorphous form , for the enhancement of dissolution rate and bioavailability . Because of the significantly higher systemic exposure anticipated from the ASD formulation , the safe evaluation of this formulation in clinical trials could not be supported by nonclinical safety studies performed for the marketed oral flubendazole formulation . In the nonclinical safety studies performed in support of flubendazole for STH , systemic exposure was very low . The set of nonclinical studies performed for the safety evaluation of the ASD formulation and their design including number of animals used , were based on international standards described in guidelines of the International Conference on Harmonisation ( ICH ) of Technical Requirements for Registration of Pharmaceuticals for Human Use . ICH guidelines promote safe and ethical development of pharmaceuticals and reducing the use of animals in accordance with the 3R ( reduce/refine/replace ) principles is part of their objectives . With the ASD formulation , two-week repeated dose toxicity studies followed by 1-month recovery were performed in rats and dogs as well as a cardiovascular safety study in dogs and a CNS safety study in rats . Flubendazole amorphous solid dispersion had been tested in an in vivo micronucleus test [1] and an explorative oral embryofetal developmental toxicity study in the rat [3] . Flubendazole was also tested in an in vivo micronucleus test in rats as a solution/suspension in polyethylene glycol 400 ( PEG400 ) and hydroxypropyl-β-cyclodextrin . An Ames test was performed with flubendazole . The results of the evaluation of flubendazole and its main metabolites in the Ames test and the in vitro micronucleus test have been described [1] . This paper describes the results of these preclinical safety studies and their impact/implications for the potential use of this new orally bioavailable ASD formulation in humans , for the conduct of clinical trials and finally for the risk/benefit associated with the use of orally bioavailable flubendazole for the treatment of onchocerciasis in the field .
After intravenous administration , flubendazole ( FBZ ) showed a low plasma clearance in rodents ( rats and jirds ) and high clearance in dogs ( Table 1 ) , a medium volume of distribution and a short half-life in the 3 species . After oral administration , the bioavailability is > 100% in jirds , 27% in rats and 15% in dogs assuming the same clearance after intravenous and oral administration . Two metabolites of flubendazole were measured in plasma: the hydrolyzed flubendazole ( H-FBZ ) and the reduced flubendazole ( R-FBZ ) as these 2 metabolites might be potentially active . ( Table 2 ) The plasma exposure ratios between parent and H-FBZ were 0 . 5 in jirds , 0 . 9 in rats and 1 . 7 in dogs . The plasma exposure ratios between parent and R-FBZ were 0 . 04 in jirds , 0 . 35 in rats and 2 . 8 in dogs . Data available from the Dryad digital repository: https://doi . org/10 . 5061/dryad . 5vv774m [4]
Flubendazole as currently formulated for the treatment of gastrointestinal parasites has poor systemic availability when given orally . Due to its poor bioavailability and solubility , the original oral formulation induces limited systemic or gastrointestinal toxicities and is negative in the bone marrow micronucleus tests . [1] Flubendazole has been shown to be highly efficacious in animal models of filarial infections when dosed subcutaneously . The clinical efficacy of flubendazole for treatment of onchocerciasis has also been reported after intramuscular administration of a suspension in humans . However , intramuscular administration resulted in serious injection site reactions . [5] Consequently , it is preferred to increase systemic availability of flubendazole by improved oral formulations . [6] Several orally bio-available formulations were developed . The oral ASD formulation was selected for preclinical development . As the new oral ASD formulation resulted in improved systemic absorption , preclinical development with this formulation did not only include extensive evaluation of efficacy and pharmacokinetics but also a re-evaluation of the toxicity of flubendazole including potential genotoxicity . It was anticipated that increased systemic exposure to flubendazole would result in toxicity related to the tubulin binding , the mechanism of action of flubendazole . Therefore , the purpose of the additional efficacy , pharmacokinetic and toxicity studies with the ASD formulation was to evaluate whether a treatment regimen could be identified that would result in an acceptable risk/benefit profile . In vivo oral toxicity studies with the ASD formulation described above , clearly show relevant systemic exposure to flubendazole and its main metabolites , reduced flubendazole and hydrolyzed flubendazole . ( Tables 3 , 4 , 12 and 18 ) In the rat and dog , the highest exposures reached at the end of 2-week repeated dosing ( data in Tables 12 and 18 ) were multiple times higher when compared with systemic exposure in humans ( less than 1 ng/ml Cmax in plasma ) following treatment with the marketed flubendazole formulation at the usual dosage ( 100 mg once or twice a day for 3 days ) . [1] In the preclinical in vivo filarial efficacy studies described by MP Hübner et al . , 100% efficacy could not be achieved with oral treatment . Highest efficacy achieved with oral treatment with the ASD formulation was 90% in the Litomosoides jird model at 15 mg eq . /kg/day for 10 days [7] . One hundred % efficacy was only observed with subcutaneous treatment in several models of infection at 10 mg/kg/day for 5 days . This subcutaneous treatment was associated with a Cmax of 50 ng/ml and an AUC0-24h of 1100 ng . h/ml , both on day 5 ( data of Litomosoides sigmodontis study in female jirds [7] ) . Notwithstanding lower efficacy , oral treatment at 15 mg eq . /kg/day for 10 days was associated with much higher exposures . On day 10 , Cmax was 3300 ng/ml and AUC0-24h was 20000 ng . h/ml ( data of Litomosoides sigmodontis study in female jirds [7] ) . The pharmacokinetic profile after a single subcutaneous and a single oral dose was very different . After an oral dose at 15 mg eq . /kg , the plasma concentration decreased very rapidly , dropping below the limit of quantification after 72 hours while after a single subcutaneous injection at 10 mg/kg , sustained plasma levels were observed for at least 3 months [7] . These data indicate that prolonged exposure at low plasma concentration is probably more important for efficacy than reaching high plasma concentrations . Comparing exposures required for efficacy and those achieved in toxicity studies , we can conclude that the exposure in the oral toxicity studies with the ASD formulation , did not fully cover the exposure after oral treatment at 15 mg eq . /kg/day for 10 days in the Litomosoides jird model which was associated with 90% efficacy . However , when comparing with Cmax and AUC 0-24h of the subcutaneous treatment regimen associated with 100% efficacy in the Litomosoides jird model , exposures in the oral toxicity studies with the ASD formulation were higher . In the CNS safety study in rats and the cardiovascular safety study in dogs , a single dose up to 60 mg eq . /kg in the rat and 120 mg eq . /kg in the dog , did not result in relevant CNS findings in the rat or cardiovascular findings in the dog . In the rat , the dose of 60 mg eq . /day was associated with a peak exposure of 4330 ng/ml . ( Table 4 ) In the dog , peak exposure was not determined but plasma concentrations measured 6 . 5 hours after dosing demonstrated exposure to flubendazole and its main metabolites . Mean plasma concentration of flubendazole at 120 mg eq . /day 6 . 5 hours after dosing was 186 ng/ml . ( Table 3 ) This is however an underestimation of the peak exposure since Cmax at 50 mg eq . /kg/day b . i . d . in the 2-week repeated dose toxicity study in male dogs on day 1 was 1660 ng/ml . These data indicate that flubendazole up to high peak plasma concentrations , does not induce Cmax related effects in the central nervous system or cardiovascular system . In the 2-week repeated dose toxicity studies in rats and dogs , main flubendazole-related changes were observed in haematological and lymphoid systems , gastrointestinal system and testes . In the dog , the liver was an additional target organ of toxicity of flubendazole . ( rat data in Tables 7 , 8 , 9 , 10 and 11; dog data in Tables 14 , 15 , 16 and 17 ) Flubendazole-induced changes in haematological , lymphoid and gastrointestinal systems were considered the consequence of its binding to tubulin . Like many benzimidazoles , flubendazole binds to the tubulin protein to the same site as colchicine . This binding results in inhibition of the polymerization of tubulin and thus disruption of the mitotic spindle . [1] For colchicine it is described that disruption of the microtubular network results in arrest of mitosis in metaphase because chromosome separation depends on microtubular function , thus inhibiting cell division . The organ systems with the highest turnover rates such as bone marrow and the gastrointestinal tract are the most vulnerable and most readily affected . [8] These organ systems have also been described to be affected by treatment with other benzimidazoles which is in line with their similar mechanism of action ( tubulin binding ) . [9 , 10] Dosing with other benzimidazoles like albendazole , mebendazole and oxfendazole in preclinical toxicity studies also resulted in testicular changes . [9 , 10] For oxfendazole , the mechanisms underlying the testicular toxicity are most probably disruption of the microtubules , and degeneration of the Sertoli cells . [11] Regarding the liver toxicity observed in dogs treated with the ASD-formulation of flubendazole , there is no obvious link with the binding to tubulin by flubendazole . These flubendazole-induced changes were at least partially reversible as shown in the dog 2-week toxicity study after 1-month recovery . From a quantitative point of view , flubendazole-related changes were observed in the rat 2-week toxicity study at the mid dose and above with the low dose , i . e . 5 mg eq . /kg/day for males and 2 . 5 mg eq . /kg/day for females being the NOAEL ( exposures associated with NOAEL in Table 12 ) . In the dog 2-week toxicity study , flubendazole-related changes were observed at the low dose and the NOAEL could not be established . Since we do not know the NOAEL in the dog study , in case of clinical trials , careful evaluation for potential flubendazole-induced changes would be warranted by monitoring for gastrointestinal symptoms , haematology and clinical chemistry changes for potential effects on bone marrow and liver and by sperm evaluation for potential testicular toxicity . Contraception for males during 3 months after dosing should be recommended because of the risk of potential impact of testicular toxicity on offspring . Since the testicular toxicity is probably , at least partially , related to disruption of the microtubular network and hence potential disruption of chromosome separation , it can be associated with chromosomal damage . For DNA damage , it has been described that it can be transmitted from fathers to offspring [12] . Results of an explorative oral embryofetal developmental toxicity study in the rat have been published by M . Longo et al . [2] They also used a flubendazole-ASD formulation ( developed and provided by Abbvie ) at dose levels of flubendazole of 2 , 3 . 46 and 6 . 32 mg eq . /kg/day . Pregnant female Sprague-Dawley rats were dosed on gestational day ( GD ) 9 . 5 and 10 . 5 and embryos/fetuses were evaluated on GD 11 . 5 , 12 . 5 or 20 . At 2 mg eq . /kg/day , flubendazole did not interfere with rat embryofetal development ( Cmax = 389 ng/ml , AUC0-24h = 2190 ng . h/ml after single administration ) . From 3 . 46 mg eq . /kg/day ( Cmax = 546 ng/ml , AUC0-24h = 4830 ng . h/ml after single administration ) onward , flubendazole markedly reduced embryonic development by GD 12 . 5 . On GD 20 , 80% of fetuses showed malformations . [2] Based on these results , flubendazole is considered a strong teratogen . Inclusion of women of childbearing potential should therefore , only be considered in well controlled clinical trials of limited size with appropriate contraceptive measures . Flubendazole is a potent aneugen in vitro . This aneugenicity is the consequence of the binding of flubendazole to the tubulin protein , inhibiting polymerization of tubulin and consequently causing mitotic spindle disruption . Spindle poisons all have the potential to induce polyploidy and aneugenicity . [1] The hydrolyzed metabolite of flubendazole is negative in the in vitro MNT , but the reduced metabolite ( R- and S-forms ) shows both aneugenic and clastogenic activity . Like flubendazole itself , both metabolites are negative in the Ames test . [1] The in vivo micronucleus test with the ASD formulation of flubendazole published by Tweats et al . also showed evidence of induced aneugenicity . [1] This is in line with the results of the in vivo micronucleus test with an aqueous solution/suspension of flubendazole in demineralised water containing polyethylene glycol 400 , hydroxypropyl-β-cyclodextrin and HCl , described in this article . In this study , flubendazole induced structural and/or numerical chromosome aberrations in erythrocytes of rat bone marrow from a dose of 5 mg/kg/day onwards , at similar exposure levels as with the ASD formulation . ( Table 20 ) [1] Aneugens are accepted as having threshold dose responses with a clear mode of action , which in the case of flubendazole would be inhibition of tubulin polymerization . [1] However , clastogens are not considered to have a threshold dose response and since the reduced metabolite of flubendazole shows clastogenic activity , a threshold approach cannot be applied for the positive effects in the in vivo micronucleus test . Because of the risk of carcinogenicity linked to the aneugenicity and clastogenicity , no clinical trials in healthy volunteers are allowed and the duration of clinical trials in patients should be limited to single dose only . However , the preclinical efficacy studies showed highest activity upon prolonged exposure after a subcutaneous administration . Such prolonged exposure to a compound with the formation of a potential clastogenic metabolite is associated with a risk for carcinogenicity . Based on the results of the preclinical toxicity and genotoxicity studies described and discussed in this article , it was concluded that the risk/benefit associated with the use of orally bioavailable flubendazole for the treatment of onchocerciasis in the field is unfavorable . Especially the clastogenicity of the reduced metabolite of flubendazole in combination with the need for exposure beyond one day for efficacy based on the preclinical efficacy studies described by MP Hübner et al . [7] resulted in risk ( for carcinogenicity ) which outweighed the benefit .
All work was conducted in accredited laboratories and according to international guidelines . Safety pharmacology studies in animals were carried out by Charles River Laboratories France Safety Assessment SAS ( previously WIL Research Europe-Lyon , France ) . The study design was in general compliance with the following animal health and welfare guidelines: Guide for the care and use of laboratory animals , 2011; Decree n° 2013–118 relating to the protection of animals used in scientific experiments described in the Journal Officiel de la République Française on 01 February 2013; Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes . The Test Facility is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International . Both rat and dog repeated dose toxicity studies were carried out by Charles River Laboratories Den Bosch B . V . ( previously WIL Research Europe B . V . , The Netherlands ) according to their internal Standard Operating Procedures . The protocols were reviewed and agreed by the Animal Welfare Officer and the Ethical Committee ( DEC 14–59 ) as required by the Dutch Act on Animal Experimentation ( February 1997 ) . The rat in vivo micronucleus study was performed in an AAALAC-approved laboratory of Johnson & Johnson In Belgium . Johnson & Johnson vivarium facilities meet inspection agency standards , and all animals are treated humanely and cared for in accordance with the European [13] and Belgian [14] guidelines , and with the principles of euthanasia as stated in the Report of the American Veterinary Medical Association Panel . [15] Flubendazole has a molecular weight of 313 . 28 , and a molecular formula of C16H12FN3O3 . For the Ames assay , flubendazole was prepared as a solution in dimethyl sulfoxide ( DMSO ) . For in vivo pharmacokinetic studies with intravenous administration , a solution of flubendazole was made in polyethylene glycol 400 ( PEG400 ) with 20% of hydroxypropyl-β-cyclodextrin pH = 4 . 2 for the rat . For the intravenous pharmacokinetic studies in the jird and the dog , a solution of flubendazole was made in 10 v/v % polyglycol/PEG400 ratio 1–1 and 35 v/v % hydroxypropyl-β-cyclodextrin at pH = 4 . For oral pharmacokinetic studies and two-week repeated dose toxicity studies and safety pharmacology studies , the test article/drug product was an amorphous solid dispersion ( ASD ) powder with a potency of 94 . 8 mg active ingredient flubendazole per g ASD powder . Because the active ingredient flubendazole only constituted 9 . 48% of the ASD powder , dose levels were expressed as mg eq . /kg body weight meaning mg base/kg body weight . To calculate the amount of ASD powder equivalent to the mg base , conversion factor was as follows: 1 mg base = 10 . 5 mg drug product . The test article was prepared as a suspension containing 0 . 5% w/v Methocel A4M ( Sigma-Aldrich ) in Elix water . For the rat in vivo micronucleus test , an aqueous solution/suspension in demineralised water containing 10% polyethylene glycol 400 , 20% hydroxypropyl-β-cyclodextrin and HCl to pH 1 . 5± 0 . 1 was prepared . The pharmacokinetic studies were performed after single intravenous administration to determine key parameters such as plasma clearance , volume of distribution and half-life , and after single oral administration to determine exposure of flubendazole and its metabolites . The toxicokinetics of flubendazole and its metabolites , which is the determination of the exposure within the toxicology studies were also performed . In both studies , the plasma samples were analyzed individually for flubendazole , H-FBZ and R-FBZ using a qualified LC-MS/MS method . After preparation and addition of the internal standard , samples were precipitated with acetonitrile , mixed and centrifuged . The supernatant was evaporated to dryness under nitrogen flow at 50°C and reconstituted with a mixture of 0 . 1% formic acid and acetonitrile ( 90/10 , v/v ) . The extract was injected onto an Acquity UPLC BEH C18 column ( 50x2 . 1 mm , 1 . 7μm particles ) ( Waters , Milford , USA ) . The chromatographic system consisted of a Shimadzu SIL30ACMP autosampler and Shimadzu LC30 pumps ( Shimadzu , Kyoto , Japan ) . The mobile phase was a mixture of 1% formic acid and acetonitrile with a flow rate of 0 . 6 ml/min and a 2 . 5-minute gradient going from 20 to 60% acetonitrile followed by a 1-minute step gradient to 98% acetonitrile . Mass spectrometric detection was performed on an API4000 triple quadrupole mass spectrometer ( Sciex , Framingham , USA ) with Turbo Ion Spray ionization operated in positive ion mode . Flubendazole , hydrolyzed flubendazole ( H-FBZ ) and reduced flubendazole ( R-FBZ ) were quantified against calibration samples and quality control samples , prepared in the same matrix as the study samples , by means of a qualified analytical method with the lowest limit of quantitation of 0 . 2 , 0 . 4 and 0 . 2 ng/ml respectively and an upper limit of quantitation of 3000 ng/ml for all three analytes across the different studies . Flubendazole was administered intravenously by bolus at 1 and 2 mg/kg in male jirds and male rats , respectively or by a short infusion ( 15 minutes ) in male dogs at 0 . 5 mg/kg . Flubendazole was administered orally by gavage at 20 mg eq . /kg in male jirds and male rats and at 35 mg eq . /kg in male dogs . The jird and the rat were not fasted . The dogs were fasted overnight . 2h post dose , dogs were given their regular food . After intravenous administration , blood samples were collected at 0 . 05 , 0 . 25 , 0 . 5 , 1 , 3 , 7 hours after dosing in male jirds ( bolus ) , at 0 . 0117 , 0 . 333 , 1 , 2 , 4 , 7 and 24 hours after dosing in male rats ( bolus ) , at 0 . 125 , 0 . 25 ( end of infusion ) , 0 . 33 , 0 . 42 , 0 . 50 , 0 . 75 , 1 . 25 , 2 . 25 , 4 . 25 , 7 . 25 and 24 . 25 hours after the start of infusion in male dogs . After oral administration , blood samples were collected at 0 . 5 , 1 , 2 , 4 , 7 and 24 hours in male jirds , rats and dogs . Plasma concentrations of flubendazole and its 2 metabolites were determined . Several pharmacokinetic parameters were determined: the maximum concentrations ( Cmax ) , the time to reach the Cmax ( tmax ) , the half-life ( t1/2 ) , the area under the curve from time zero to infinity ( AUC0-inf ) , the clearance ( CL ) and the volume of distribution ( at steady state Vdss or Vd ) In the safety pharmacology studies , potential undesirable effects of flubendazole on physiological functions in relation to exposure in the therapeutic range and above were investigated . Organ systems evaluated were cardiovascular system , respiratory system ( evaluation included in cardiovascular safety study ) and central nervous system ( CNS ) . Genotoxicity assessments were performed to detect if flubendazole induces genetic damage . | This article describes pharmacokinetic profiles and results of safety pharmacology , toxicity and genotoxicity studies with an oral ASD formulation of flubendazole with improved bioavailability . Flubendazole administered as ASD formulation has good oral bioavailability in animals ranging from 15% to more than 100% . In in vivo toxicology studies , increased systemic exposure does not induce Cmax-related effects in CNS and cardiovascular systems . Increased exposure upon repeated dosing results in changes in haematological , lymphoid and gastrointestinal systems and in testes . In dogs , the liver was an additional target organ . These changes were at least partially reversible . Flubendazole is negative in the Ames test but positive in the in vivo micronucleus test . Because of the carcinogenic risk associated with this positive effect in the in vivo micronucleus test , exposure duration in patients should not exceed one day . Flubendazole-induced toxicity and associated risk is monitorable and controllable in patients if stringent precautions are applied in view of testicular toxicity and previously described teratogenicity . Considering both , treatment regimen needed for efficacy and outcome of toxicity and genotoxicity studies , it was concluded that the risk/benefit associated with the use of orally bioavailable flubendazole for the treatment of onchocerciasis in the field does not support further development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"micronuclei",
"atrophy",
"microbiology",
"vertebrates",
"dogs",
"animals",
"mammals",
"toxicology",
"toxicity",
"routes",
"of",
"administration",
"red",
"blood",
"cells",
"metabolites",
"signs",
"and",
"symptoms",
"pharmacology",
"oral",
"administration",
"animal",
"cells",
"metabolism",
"blood",
"plasma",
"biochemistry",
"eukaryota",
"diagnostic",
"medicine",
"blood",
"anatomy",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"protozoology",
"cellular",
"types",
"amniotes",
"organisms"
] | 2019 | Preclinical toxicity and pharmacokinetics of a new orally bioavailable flubendazole formulation and the impact for clinical trials and risk/benefit to patients |
Influenza A virus ( IAV ) infection in the respiratory tract triggers robust innate and adaptive immune responses , resulting in both virus clearance and lung inflammation and injury . After virus clearance , resolution of ongoing inflammation and tissue repair occur during a distinct recovery period . B7 family co-stimulatory molecules such as CD80 and CD86 have important roles in modulating T cell activity during the initiation and effector stages of the host response to IAV infection , but their potential role during recovery and resolution of inflammation is unknown . We found that antibody-mediated CD86 blockade in vivo after virus clearance led to a delay in recovery , characterized by increased numbers of lung neutrophils and inflammatory cytokines in airways and lung interstitium , but no change in conventional IAV-specific T cell responses . However , CD86 blockade led to decreased numbers of FoxP3+ regulatory T cells ( Tregs ) , and adoptive transfer of Tregs into αCD86 treated mice rescued the effect of the blockade , supporting a role for Tregs in promoting recovery after virus clearance . Specific depletion of Tregs late after infection mimicked the CD86 blockade phenotype , confirming a role for Tregs during recovery after virus clearance . Furthermore , we identified neutrophils as a target of Treg suppression since neutrophil depletion in Treg-depleted mice reduced excess inflammatory cytokines in the airways . These results demonstrate that Tregs , in a CD86 dependent mechanism , contribute to the resolution of disease after IAV infection , in part by suppressing neutrophil-driven cytokine release into the airways .
Influenza A virus ( IAV ) infects and replicates in the respiratory tract , triggering a robust immune response . There is a large body of evidence to support the concept that fine control of the immune response to IAV is essential to prevent excessive tissue injury; while an immune mediated inflammatory response is required to effectively clear the virus , an exaggerated host response can result in bystander tissue damage and a subsequent decrease in lung function [1]–[3] . Therefore , many studies have attempted to determine whether inflammation can be controlled without compromising virus clearance [4]–[6] . Towards this goal , several components of the host immune system have been identified that enhance or reduce inflammation and injury; however , these studies have focused on the regulation of the initiation and/or effector phases of adaptive T cell responses to IAV . Thus , although significant inflammation and tissue damage are evident in the respiratory tract after infectious virus clearance , the factors involved in regulating the resolution of inflammation and injury after IAV clearance remain poorly characterized . In a mouse model of IAV infection , recovery following virus clearance involves both tissue repair and the resolution of inflammation [7] . A number of studies have highlighted key factors and cell types during this complex process . In large part , local tissue repair within the respiratory tract is mediated by respiratory epithelial stem cells and fibroblasts [7] , but there is provocative evidence to suggest that the host immune response is an important regulator of this process . For example , innate lymphoid cells and natural killer cells have been demonstrated to promote epithelial proliferation after IAV infection through the production of the growth factor amphiregullin [8] and the pro-wound healing cytokine IL-22 [9] , respectively . In contrast to tissue repair , the resolution of inflammation has been considered to be primarily a consequence of the reduced IAV-specific effector T cell activity and decreased numbers of effector T cells following elimination of viral antigen ( i . e . removal of the stimulus for inflammation ( antigen ) leads to a gradual return to homeostasis ) [7] . However , in other models of acute lung injury ( ALI ) , there is increasing evidence to support the importance of pro-resolving molecules such as Resolvin D1 and TGFβ in regulating clearance of inflammatory cells from the respiratory tract [10] , [11] . Understanding the factors that control resolution of lung disease after IAV clearance could help identify therapeutic targets to promote faster recovery from severe respiratory viral infections . CD86 and CD80 are members of the B7 family serving as co-stimulatory molecules . These ligands interact almost exclusively with the receptors CTLA4 and CD28 , which in turn , are primarily expressed on CD4+ and CD8+ T cells . The CD86 and CD80 co-stimulatory molecules have been shown to enhance tissue inflammation by augmenting the response of conventional T effector cells [12] as well as to suppress inflammation through their ability to augment/sustain regulatory T cell ( Treg ) activity [13] . In the influenza model , CD80 and CD86 have been analyzed primarily in their capacity to support the induction of pro-inflammatory T effector responses and the expression of pro-inflammatory mediators by the T effector cells . During induction of T cell responses within the draining lymph node , CD28 engagement on naïve T cells by CD80/86 on antigen presenting cells ( APCs ) is required for a robust IAV-specific T cell response and efficient virus clearance [14]–[16] . In contrast , disruption of CD80/86 signaling to receptors on T cells after effector T cell generation dramatically reduces IAV-specific T cell effector activity in the respiratory tract ( e . g . cytokine production and proliferation ) but does not impact virus clearance or morbidity of mice [17] , [18] . However , the role of CD80/86 co-stimulation in the recovery phase after IAV infection has not been previously evaluated . Furthermore , it is known that CD28 is required for the development , homeostatic maintenance , and proliferative expansion of Tregs , and CTLA4 is required for both expansion and expression of Treg suppressive function [13] . Importantly , although the ligand function of CD86 and CD80 demonstrates significant overlap , these two co-stimulatory molecules have been demonstrated to perform distinct roles under certain conditions . For example , in the NOD mouse model , CD80 and CD86 differ in their capacity to stimulate effector versus regulatory T cells [19] , [20] . Treg cells are a subset of CD4+ T cells that express FoxP3 and can suppress inflammation in a number of disease models [21] , [22] . During IAV infection , Treg cells have been shown to suppress innate and adaptive immune responses during the induction of primary [23] , [24] and memory responses [25]; however , little enhanced pathology was noted in Treg depleted mice during primary IAV infection . Only in a model of memory responses to IAV infection did the enhanced effector CD8+ T cell response , resulting from elimination of Tregs at the time of secondary IAV challenge , cause increased morbidity in mice . The role of Treg cells , though , has not been evaluated during resolution of disease after virus clearance . In models of ALI ( e . g . LPS-induced lung injury ) , Treg cells do promote resolution of injury in the respiratory tract , in large part by limiting accumulation of innate immune cells in the lung [26] and promoting non-inflammatory clearance of apoptotic neutrophils [11] , [27] . Because of their immune-regulatory capabilities , Tregs are an obvious candidate to play a pivotal role in the resolution phase of infection after clearance of IAV . In the present study , we initially evaluated the potential role of B7-family molecules CD80 and CD86 during resolution of disease ( i . e . after virus clearance ) by in vivo administration of a blocking monoclonal antibody directed to CD80 or CD86at day 8 post infection ( p . i . ) , a time point shortly after IAV clearance within the respiratory tract [17] . We found that while in vivo administration of blocking monoclonal antibody directed to CD80 did not affect recovery , CD86 blockade at this time point lead to a delay in recovery as measured by regain of weight following virus clearance . Within the respiratory tract , this delay was associated with an increased number of leukocytes , augmented inflammatory cytokines , and a dramatic loss of the Treg population , with no corresponding impact on effector T cell activity , and this loss of Treg numbers correlated with a reduction in Treg proliferation . Importantly , adoptive transfer of Tregs into αCD86 treated mice rescued the effect of the blockade while acute depletion of Tregs during recovery mimicked the late αCD86 blockade model , supporting a role for Tregs in promoting recovery after virus clearance . The role of CD86 co-stimulation and Treg function during recovery from influenza infection is discussed .
Previous work by our and other laboratories has established a role for the co-stimulatory molecules CD80 and CD86 in regulating IAV-specific T cell effector activity in the influenza-infected respiratory tract [17] , [18] . Hufford et al . showed that simultaneous in vivo blockade of CD80 and CD86 ( by monoclonal antibody administration ) after initiation of T cell responses but before virus clearance within the infected lungs ( i . e . day 5–6 p . i . ) led to a dramatic reduction in T cell-derived cytokines , without any significant impact on virus control or morbidity of the mice . After infectious virus clearance and elimination of most virus-infected cells ( i . e . day 8–10 p . i . ) , inflammation in the respiratory tract was gradually resolved , as evidenced by the reduction of inflammatory cells and cytokines , as well as the contraction of antigen specific effector T cells . However , the role of these residual immune cells and co-stimulatory ligands , such as CD80/86 , during recovery and resolution of lung inflammation remained ill-defined . We wanted to determine what part , if any , co-stimulation ( i . e . engagement of the CD80 and/or CD86 ligands ) played in the regulation of inflammation during recovery following IAV clearance . To this end , we infected mice with a sublethal dose of the mouse adapted A/PR/8/34 ( PR8 ) IAV strain , and we administered 200 µg αCD80 or αCD86 blocking antibodies intra-peritoneal ( i . p . ) at day 8 p . i . , that is after virus clearance but before resolution of inflammation and recovery . CD80 blockade did not have any impact on recovery as reflected in weight gain during the recovery phase ( data not shown ) . However , we found that CD86 blockade during the recovery phase led to prolonged morbidity as evidenced by a delay in weight gain ( Figure 1A ) although mice eventually regained 90–95% of their starting body weight by day 25 p . i . ( data not shown ) . Importantly , this effect of in vivo CD86 blockade on morbidity was only evident during a narrow time frame following virus clearance and was not seen either when the blocking antibody was administered during active virus clearance from the lungs ( day 6 p . i . ) or when it was given late in the recovery phase ( day 12 p . i . ) ( Figure 1B ) . Although CD80 and CD86 share receptors and can have comparable effects , several studies have demonstrated that these molecules can orchestrate distinctly different T cell functions [28] , [29] . These non-redundant functions may reflect differences in binding affinities for their receptors and/or differences in pattern or distribution of ligand and receptor display at any given point during the T cell response [30] . Importantly , we did not see any impact of day 9p . i . CD86 blockade on the tempo of virus clearance as determined either by infectious virus titer or viral genome copy number ( Figure 1C–D ) , consistent with an earlier report from our lab showing that CD86 blockade early during infection ( i . e . day 5 p . i . ) did not affect virus clearance [17] . This data suggested a unique role for CD86 during the recovery phase of infection following IAV clearance , which is distinct from its previously defined roles during the induction and effector phases of T cell responses . Since CD86 blockade led to increased morbidity without any change in viral clearance , we wanted to determine if CD86 blockade during the resolution phase of infection delayed recovery by affecting the extent or characteristics of pulmonary inflammation . In order to assess changes to immune responses after αCD86 treatment during the recovery phase ( on day 9 p . i . ) , we sampled the bronchial alveolar lavage ( BAL ) fluid and immune cells within the respiratory tract at various times post-CD86 blockade . We detected an increase in the accumulation of neutrophils in the lung interstitium following CD86 blockade . However , the percentage and absolute numbers of other abundant lung infiltrating innate immune cells including Ly6Chi monocytes , Ly6Clo monocytes , and eosinophils were unchanged after αCD86 administration ( Figure 2A ) . By Luminex 30-plex cytokine array analysis , we detected elevated levels of only three innate immune associated cytokines , G-CSF , LIF , and eotaxin in the BAL fluid following blockade ( Figure 2B ) . All of these cytokines have chemotactic potential , and G-CSF , in particular , is associated with the capacity to support the generation and survival of neutrophils [31] . We did not detect any significant impact of CD86 blockade on total lung CD4+ or CD8+ T cell numbers within the IAV infected lungs , nor on BAL cytokines typical of effector T cell origin ( e . g . IFNγ ) ( Figure 2C–D ) . If anything , there was a trend toward fewer CD4+ and CD8+ T cells , which is counterintuitive in mice with increased morbidity and inflammation . To further probe the potential impact of αCD86 treatment on antigen-specific T cells , we prepared lung cell suspensions from day 12 p . i . infected mice that had received αCD86 treatment on day 9 p . i . , and we stimulated these cells ex vivo with PR8-infected BMDCs . Flow cytometry analysis of intracellular cytokine staining revealed that in vivo αCD86 blockade had no impact on the quantity of antigen specific T cells capable of producing IFNγ ( Figure S1 ) , suggesting that antigen-specific effector T cells were not overtly altered by αCD86 treatment during the recovery phase . To identify the potential cellular targets of CD86 blockade , we evaluated expression of CD86 and its receptors , CD28 and CTLA4 , within the respiratory tract and the lung draining lymph nodes at day 10 p . i . We found that expression of CD28 and CTLA4 , as expected , was restricted to T cells , with CD4+ T cells having the highest frequency of receptor-positive cells ( Figure 3A–B ) . In contrast , CD86 had very promiscuous expression , and in addition to its display on the surface of traditional antigen presenting cells , including CD11b+ monocytes/macrophages and B220+ B cells , CD86 was also abundantly expressed on T cells in the recovering lungs ( Figure 3C–D ) . The numbers of CD86+ CD4+ and CD8+ T cells increased during the course of infection , with maximal numbers around day 10 p . i . ( Figure 3E ) . It is perhaps noteworthy that peak expression of CD86 includes the timeframe when in vivo blockade of this co-stimulatory ligand is most effective in promoting excess morbidity . Importantly , neutrophils expressed neither the CD86 ligand nor its receptors at this time point . This suggests that the increase in lung neutrophil accumulations observed after CD86 blockade is not due to a direct effect of blockade on the migration , accumulation or function of the neutrophils infiltrating the lungs . In view of the T cell-restricted expression of the CD86 receptors , CD28 and CTLA-4 , it was likely that CD86 blockade was interfering with CTLA4 and/or CD28 signaling in T cells following virus clearance , ultimately leading to increased inflammation and increased neutrophil accumulation during the recovery phase of infection . Although we observed no change in overall T cell numbers in the respiratory tract following CD86 blockade in vivo ( Figure 2C ) , we did observe that there was a transient but significant decrease in CD25 expression on lung CD4+ T cells one day following CD86 blockade , that is on day 10 p . i . following CD86 blockade on day 9 p . i . ( Figure 4A ) . Since CD25 serves as a marker for regulatory T cells ( as well as activated effector T cells ) , we evaluated the impact of in vivo CD86 blockade on Treg cells in the lungs by flow cytometry ( using FoxP3+ expression as a marker for the Treg cell compartment ) . We found a significant reduction in both the frequency and absolute number of CD4+Treg cells in the lungs on day 14 p . i . after CD86 blockade administered on day 9 p . i . ( Figure 4B ) . This could explain the slight trend toward lower CD4+ T cells seen in Figure 2 since FoxP3+Treg cells are embedded in that data . Interestingly , although we saw a decrease in CD25 expression on total CD4+ T cells at day 10 ( one day after CD86 blockade ) , we did not see a reduction in CD25 expression on the remaining Tregs cells on day 14 p . i . , suggesting that although Tregs were diminished in numbers , their per cell activation state was not altered by αCD86 blockade ( Figure S2A ) . To further evaluate Treg function after CD86 blockade , we harvested lung cell suspensions from day 12 p . i . mice ( that had received αCD86 on day 9 ) , used PR8 infected BMDCs to re-stimulate lung T cells , and then measured IL-10 production in FoxP3+Tregs by intracellular cytokine staining . We found that although the total percentage of Tregs was beginning to drop within the total CD4+ population ( Figure S2B ) , the frequency of IL-10+ Treg cells was not reduced within the FoxP3+ population ( Figure S2C ) , suggesting that Treg function in remaining cells was not altered by αCD86 . Importantly , using the same day 12 p . i . lung cell suspensions , we included αCD86 in vitro in separate BMDC co-cultures to evaluate if the presence of the blocking antibody in the cultures could reduce Treg IL-10 production . We saw no impact of αCD86 on the frequency of IL-10+Tregs under these culture conditions , further supporting the idea that αCD86 did not directly alter Treg function on a per cell basis ( Figure S2D ) . We next wanted to determine if the Treg cells in the infected respiratory tract expressed CD86 and/or its receptors CD28 and CTLA4 . We found that Tregs did express high levels of both CD28 and CTLA4 , as well as lower levels of CD86 over the course of infection ( Figure S3 A–C ) . These data suggest that CD86 blockade could prevent signaling on Treg cells through CD28 and/or CTLA4 , which are required for Treg homeostatic maintenance and proliferative expansion [13] . It is unlikely CD86 blockade affected Tregs intrinsically since the majority of Treg cells in the respiratory tract were CD86 negative , and there is no known signaling role for CD86 on T cells . Because CD28 and CTLA4 can promote Treg proliferation , we wanted to determine whether CD86 treatment caused a defect in Treg proliferation , which could plausibly explain the loss of Treg numbers in the respiratory tract . We examined Ki67 expression in both conventional CD4+FoxP3− T effector cells and FoxP3+Treg cells in the lung after CD86 blockade ( administered at day 9 p . i . ) . We found that a sizable percentage of Tregs in the recovering lung were Ki67+ , and this high percentage was largely maintained from day 10 to day 12 p . i . In contrast , the CD4+ effector T cells had a much lower frequency of proliferating cells that continued to drop to near baseline from day 10 to day 12 p . i . ( Figure 4C ) . Importantly , when we compared the frequency and numbers of proliferating Tregs and CD4+ effector T cells in the lung after αCD86 treatment , we saw a significant decrease in the quantity of Ki67+Tregs , suggesting that CD86 co-stimulation promotes Treg proliferation after IAV infection ( Figure 4D–E ) . There was a trend toward reduced Ki67+ effector CD4+ T cell proliferation after αCD86 treatment , but this did not reach statistical significance , likely because only a small fraction of these cells were proliferating at this late time after infection . This observation is consistent with the previous findings from our lab that the effector T response rapidly wanes after day 7 p . i . in the absence of influenza antigen [17] , and it provides a plausible explanation as to why the late CD86 blockade selectively impacts Treg cells , not T effector cells . Since Treg cells play an anti-inflammatory role in many disease models , a loss of Treg cells following in vivo CD86 blockade could , in part at least , explain the observed increase in lung inflammation . Treg cells have been reported to suppress CD8+ T cell responses as well as innate immune responses at the induction of IAV-specific immune responses or during IAV clearance [23] , [25] , [32]; however , to our knowledge , no role for Treg cells has been established during the recovery phase of IAV infection , that is after infectious virus clearance . To explore the potential role of Treg cells in the resolution of inflammation during recovery from influenza infection , we first determined the kinetics of Treg cell accumulation in the respiratory tract during IAV infection . To this end , we infected wild type mice with IAV and determined the number of Tregs in the lungs and draining lymph nodes at various times p . i . Consistent with earlier reports [25] , [33] , we found that Treg cells are recruited to and accumulate in the lungs following infection , and , of note , Treg cell numbers also persist well into the recovery phase ( Figure 4F ) . Consequently , these cells have the potential to participate in limiting excess residual inflammation even after infectious virus has been cleared and virus-infected cells largely eliminated . Furthermore , we found that even in wild type infected mice not undergoing CD86 blockade the number of Treg cells detected in the lung after virus clearance positively correlated with recovery of body weight ( Figure 4G ) . This difference in Treg numbers among infected untreated wild type animals could , in part , contribute to the variability in tempo of weight gain that is typically observed during the recovery/resolution phase of IAV infection . The above findings suggested that one effect of in vivo CD86 blockade was to limit the accumulation and/or persistence of Treg cells necessary to limit inflammation following IAV clearance from the lung . We therefore wanted to independently evaluate the role of Tregs during recovery and in particular whether depletion of Treg cells in vivo at this late time point following infection ( i . e . day 8–10 p . i . ) would mimic the effect of CD86 blockade . To explore this possibility , we employed the DEREG ( Depletion of REGulatory T cells ) mouse model to deplete Tregs after virus clearance but before the onset of recovery . DEREG mice express the diphtheria toxin ( DT ) receptor ( and GFP ) driven off of the FoxP3 promoter [34] . Because of reports indicating that FoxP3 may be expressed on cell types other than Tregs , in particular epithelial cells from multiple organs , including the lung [35] and to ensure that DT-mediated depletion only affected immune cells , for this analysis we employed bone marrow chimeric mice in which DEREG bone marrow was used to reconstitute irradiated wild type animals ( Figure 5A ) . Following irradiation and reconstitution , we infected these mice and administered DT at days 8 and 10 p . i . We found that DT administration lead to a rapid decrease in total Treg cells in the lungs ( Figure S4A ) . While loss of GFP+ Treg cells was extensive and prolonged , the small number of GFP−Tregs did preferentially expand following DT treatment ultimately resulting in the restoration of total Treg numbers within one week ( Figure S4A ) . We evaluated innate immune responses in the respiratory tract after Treg depletion and found that Treg depletion in DEREG mice mimicked the inflammatory signature of CD86 blockade at days 8–10 p . i . There was a significant increase in neutrophil numbers ( Figure 5B–C ) , but not Ly6Chi monocytes ( Figure 5D ) , and the innate cytokines G-CSF , eotaxin , and LIF were also increased in the airways ( BAL fluid ) ( Figure 5E ) . Also similar to the CD86 blockade model , Treg cell depletion delayed regain of weight in mice ( Figure 5F ) , without changing the kinetics of virus clearance ( Figure S4B ) . Furthermore , although effector T cells have been implicated as targets of Treg suppression early during infection [23] , late Treg cell depletion had no impact on the total numbers of conventional CD4+ and CD8+ T cells in the lungs ( Figure 5G ) , or the levels of prototypical effector T cell cytokines ( e . g . IL-10 , IFN-γ , and TNF ) in the BAL fluid ( Figure 5H ) . Treg depletion did not completely mimic the effect of CD86 blockade . For example , we noted a large increase in the innate cytokine IL-6 following acute Treg cell depletion ( Figure 5E ) which was not detected following CD86 blockade . Also , the defect in weight gain was delayed in the DEREG model compared to the αCD86 model , but this defect coincided with the time post-CD86 blockade when Treg numbers are detectably reduced . We did not , however , carry out an extended time course in these mice to determine when or if Treg depleted mice eventually recover full body weight . Thus the αCD86 blockade and Treg depletion models showed significant overlap , but unlike αCD86 treatment which produced a gradual decrease in lung Treg cell numbers over several days , acute Treg depletion produced by DT administration likely had additional consequences ( e . g . elevated IL-6 production ) . To determine if the delay in recovery and excess inflammation triggered by CD86 blockade was at least in part attributable to the loss of Treg cells in the lungs following CD86 blockade , we administered αCD86 to infected mice on day 9 p . i . , and then adoptively transferred 2×106Tregs by the i . v . route on day 11 p . i . We found that the Treg transfer following CD86 blockade lead to accelerated weight gain ( Figure 5I ) . Importantly , we observed a decrease in total neutrophils in the respiratory tract ( Figure 5J ) following Treg cell transfer . These findings suggest that loss of Tregs resulting from CD86 blockade contributed to delay in recovery from IAV infection , and that following infectious virus clearance , Treg cells may play a critical role in controlling residual inflammation and facilitating recovery from infection . We chose this CD86 blockade and transfer schedule because the αCD86 antibody potently masks the CD86 epitope for only approximately 48 hours post injection in the respiratory tract ( data not shown ) . Therefore , transferring cells two days post αCD86 treatment minimizes effects of residual antibody on the transferred Treg cells , although CD86 blockade did not suppress Treg function as measured by CD25 expression and IL-10 production ( Figure S2 ) and the majority of isolated Treg cells did not express CD86 ( Figure S5 ) . However , since some transferred Tregs cells , isolated from naïve spleen , did express CD86 , we cannot completely rule out a possible functional contribution of CD86 expressed on Tregs in facilitating recovery from influenza virus infection . To determine the possible direct targets of Treg suppression during recovery , we took a closer look at the kinetics of inflammatory cells and mediators after Treg depletion in DEREG mice , this time using a single dose of DT at day 9 p . i . We found that IL-6 , LIF , and neutrophil numbers were not statistically significantly impacted until several days after Treg depletion , suggesting that they were not directly suppressed by Tregs ( Figure 6A–B ) , although there was a trend that IL-6 appeared to be elevated shortly after Treg depletion . Importantly , G-CSF was immediately and dramatically increased , suggesting that the cellular source ( s ) of G-CSF was under direct Treg control ( Figure 6A ) . We analyzed the mRNA content of various cell populations from the lung on day 15 p . i . and found that the majority of CSF3 mRNA was found within the neutrophil compartment , with lower but still detectable levels in the CD45− and macrophage/monocyte compartments ( Figure 6C ) . This suggests that Tregs may directly suppress neutrophil-derived G-CSF , which could contribute to pulmonary inflammation through a feed-forward circuit of neutrophil recruitment and/or persistence in the respiratory tract . Since neutrophils were identified as a potential source of G-CSF and target of Treg suppression , we wanted to determine if neutrophil depletion would rescue the effects of Treg depletion . We used the neutrophil-depleting αLy6G ( 1A8 ) antibody to reduce the number of neutrophils after Treg depletion . Using the DEREG bone marrow chimeras , we infected mice with a sublethal dose of PR8 , injected DT at day 9 p . i . , and administered 50 µg of αLy6G mAb on days 11 and 13 p . i . We then measured parameters of lung inflammation on day 14 . We found that neutrophil depletion significantly reduced excess cytokines in the BAL after Treg depletion , suggesting that neutrophils are a major source or stimulator of inflammatory cytokines after Treg depletion ( Figure 6D ) . Neutrophil depletion failed , however , to rescue weight loss after Treg depletion ( Figure 6E ) , suggesting that Treg cells may regulate other features or aspects of the recovery process following IAV clearance . Alternatively , subsets of neutrophils exist that can possess anti-inflammatory functions; therefore , total neutrophil depletion may lead to loss of both pro-inflammatory neutrophils and putative beneficial neutrophils .
In this report , we described a previously unappreciated positive role for the co-stimulatory ligand CD86 in the resolution of inflammation following virus clearance during recovery from IAV infection . We observed that blocking CD86 ( but not CD80 ) engagement in vivo in a narrow time window ( i . e . between days 8 and 12 p . i . ) resulted in prolonged morbidity ( i . e . delayed regain of weight ) . This delay in recovery was associated with a sustained innate inflammatory response reflected in elevated levels of several cytokines/chemokines in the BAL fluid and excess accumulation of neutrophils in the lung tissue compared to the experimental controls . This “late” blockade of CD86 in vivo did not alter the frequency or activity of activated CD8+ and CD4+ T effector cells present in the recovering lungs; however; CD86 blockade did result in a significant decrease in the FoxP3+ Treg cell population present in the respiratory tract , due at least in part to a reduction in the proliferation of Treg cells . Targeted Treg cell depletion during the aforementioned time late in infection in the DEREG mouse model mimicked the inflammatory signature of the CD86 blockade phenotype , and adoptive transfer of Treg cells , in part , rescued the effects of the CD86 blockade . Finally , we identified neutrophils as a target for regulation/suppression by Treg cells since neutrophil depletion reversed the effect of Treg cell depletion on inflammatory cytokines in the BAL . In the present study , we demonstrated a novel role for the co-stimulatory B7 family member CD86 to resolve ongoing inflammation following IAV clearance through maintenance of the Treg cell response . Resolution is defined as a return to homeostasis , and this process encompasses both the dampening of the immune system and initiation and completion of repair processes , ultimately leading to the regain of normal physiological function [7] , [10] . Previous studies investigating the co-stimulatory B7 family molecules during influenza infection have demonstrated their critical role in activating naïve T cells in the draining lymph nodes and triggering IAV-specific T cell effector activity within the respiratory tract [15] , [16] , [36] . In the context of both induction and the expression of effector T cell responses to influenza infection , CD80 and CD86 appeared to have overlapping , essentially equivalent roles acting in concert with viral antigen presented on the surface of an APC to trigger naïve T cell activation and/or expression of T cell effector activity [17] . However , our report indicated that during the recovery/resolution phase of IAV infection , CD86 displayed a unique activity ( i . e . not displayed by CD80 ) to control excess inflammation in the lungs following virus clearance without any apparent impact of this co-stimulatory ligand on residual lung effector T cells . Non-redundant roles for CD80 and CD86 have been reported in other disease models where these two co-stimulatory molecules have been demonstrated to display different roles in the promotion of responses by either Treg cells or T effector cells . This phenomenon has been most well-studied in the NOD mouse model of type I diabetes . Counter to what we observed in the IAV infection model , in NOD mice CD80 is critically important for Treg maintenance , whereas CD86 promotes the autoreactive effector T cell response [19] , [37] . However in other models , CD86 , not CD80 , has been demonstrated to preferentially promote Treg responses [29] , [38] . In a comprehensive review of the regulation of Treg cells by CD80/86 co-stimulation , Bour-Jordan et al . argues that these seemingly contradictory findings were likely due to differences in expression levels of CD80 and CD86 in both time and space across different model systems [13] . As we demonstrate in this report , in the IAV infection model , CD86 plays a unique role in controlling excess inflammation following virus clearance through its ability to sustain Treg cell numbers and/or function . To our knowledge , this study is the first to evaluate the role of Tregs during resolution of influenza-induced disease , although Treg cells have been implicated in inflammation/injury resolution in models of acute lung injury ( ALI ) [26] , [27] . Models of ALI share certain characteristics with influenza-induced lung injury , including infiltration of neutrophils and other leukocytes into the respiratory tract , epithelial damage , increased vascular permeability , and hypoxemia . In ALI , Treg cells become activated by either pattern recognition receptors ( e . g . TLR4 for LPS-induced ALI ) or injury-related self-antigens to promote neutrophil apoptosis and reduce cytokine production by phagocytic macrophages [11] , [39] . A similar environment potentially exists during the recovery phase after IAV clearance: viral antigen is low , injury-related self-antigens are abundant , and Treg cells are required to limit neutrophil accumulation and innate cytokine production in the respiratory tract . It remains to be determined how Treg cells function to limit neutrophil numbers and promote weight gain after influenza infection , but it may be possible to draw parallels with the ALI model , where Treg-derived TGF beta and adenosine ( through the ecto-enzyme CD73 ) have been shown to diminish inflammation and promote resolution of injury [11] , [27] . An increasing body of evidence and emerging consensus suggests that IL-10 is not a major mechanism by which Tregs promote resolution of lung injury [39] , although Tregs are capable of producing IL-10 during the resolution phase . This view is consistent with our findings on the level of IL-10 expression during the recovery/resolution phase of influenza infection . IL-10 is an important anti-inflammatory molecule during virus clearance by effector T cells and is primarily produced by effector T cells themselves [40] , but as we showed in Figure 2 , the level of IL-10 in the BAL during the recovery phase is very low and does not depend on Tregs . It is also notable that effector T cells do not appear to be targets of Treg suppression during the recovery phase after IAV infection , in contrast to a previously described role for Tregs to suppress the CD8+ T cell response during the induction phase of IAV infection . It is likely that at this late stage after infection , IAV antigen availability limits the T cell responses without the need for Treg suppression . This concept is supported by previous work from our lab: Hufford et al . showed that the in vivo CD8+ T cell IFNγ response , which is dependent on antigen presentation , decreased dramatically from day 7 to day 8 p . i . , correlating with clearance of infectious virus from the BAL fluid [17] . Therefore , it appears that by day 8 or 9 post infection , at the time of Treg depletion , effector T cell responses are already past their peak and are likely regulated by antigen availability , with no requirement for control by Tregs . The mechanism ( s ) by which CD86 sustains Treg proliferation and total Treg numbers during recovery from IAV infection still needs to be more fully defined , but several possibilities can be considered . CD28 and CTLA4 are expressed on Treg cells , and these receptors have been shown to be critical for Treg development , homeostasis , proliferation , and function [13] , [41] . Consequently , disruption of CD86 ligand engagement for one or both of these receptors directly on Treg cells could result in reduced proliferation , adversely affecting the numbers of Treg cells as suggested by our findings . However , additional studies would be required to determine if CD86 blockade could be impacting Treg survival and/or function in addition to proliferation . Alternatively , although we did not detect any changes in prototypical IAV-specific T cell IFNγ production , it is possible that αCD86 treatment could impact the Treg response indirectly through other less well studied effector T cell derived functions . For example IL-2 , a critical regulator of Treg viability and maintenance , could potentially help sustain Treg responses in the late phase of influenza infection , with CD4+ T effector cells serving as a source of IL-2 through a CD86 engagement dependent mechanism [42] , [43] . It is important to note that although there is some evidence that CD86 itself can signal upon ligation with its receptors in certain contexts on dendritic cells and B cells [30] , [44] , CD86 has not been described to signal in T cells and has no known signaling domain , suggesting that αCD86 likely does not impact T cells through blockade of an intrinsic CD86-derived signal . It is important to acknowledge that although we see dramatic effects of αCD86 blockade in the lung , we cannot rule out that the antibody may be acting in another location ( e . g . draining lymph node ) to reduce Treg cell numbers , ultimately resulting in loss of Treg cells in the lung and enhanced inflammation in the respiratory tract . However , although the blockade is not restricted to the lung , the blocking antibody does at least block CD86 in the respiratory tract , since labeled αCD86 does not bind to lung cells harvested from mice within 48 post in-vivo blockade , indicating that CD86 epitope is masked . Due to high levels of capillary leak and alveolar injury at the time of administration ( day 10 post infection ) , lung specific blockade of CD86 would not be practicable since intranasal administration of αCD86 would readily enter the systemic circulation . One interesting observation from this report is the expression of CD86 on both Teff ( Figure 3C–E ) and Treg cells ( Figure S3 ) during the recovery phase . Although it is unclear if CD86 on T cells contributes to Treg maintenance in our system , Taylor et al reported that T cell-T cell interactions ( i . e . CD86 on T cells triggering receptors on other T cells ) may preferentially occur through CD86 , and these interactions may be important to ameliorate disease in a model of graft-versus-host disease [45] . Alternatively , Tregs may require CD86 in the context of a classic APC , and we have observed that MHC II+ mononuclear cells in the respiratory tract late after infection may express either CD80 or CD86 or both ( data not shown ) . Furthermore , differences in distribution of CD80 and CD86 expression on various APCs , including B cells , monocytes/macrophages , and dendritic cells , may explain why CD86 alone promotes resolution of disease in this model . Finally , if Treg cells are indeed responding to injury derived self-antigens and/or other damage associated signals to promote resolution of inflammation after IAV clearance , then CD86 may be critical to provide co-stimulation for these self-specific responses . Yadav et al demonstrated that CD86 preferentially primed self-antigen specific Teff and Treg cells in the NOD model , and CD80 was unable to compensate for this function in CD86 KO mice [37] . Unfortunately , dissecting the critical cellular source of CD86 during recovery is difficult because we would need to target CD86 disruption in both time ( recovery period ) and space ( specific cell populations ) , but these questions could possibly be answered with the development of a cell-specific inducible CD86 knock out mouse . Understanding the cellular interactions required for CD86-mediated recovery and Treg cell maintenance would provide more targets to manipulate this novel pro-resolution pathway . The contribution of neutrophils to inflammation and disease during influenza infection is complex and controversial [46]–[48] . Our data suggests that in the absence of Treg cells , lung infiltrating neutrophils may be detrimental during the recovery phase after influenza infection possibly by producing and/or by promoting release of pro-inflammatory cytokines by other cell types ( e . g . respiratory epithelial cells ) into the recovering lungs . This possibility is supported by evidence that neutrophils express a detrimental inflammatory signature during highly pathogenic influenza infections [49] and that Treg cells suppress innate cell dependent inflammation in influenza infected Rag−/− mice [32] . The mechanism by which neutrophils promote cytokine production and how this impacts overall recovery after virus infection remains to be defined . Because we found that neutrophils expressed high mRNA levels of the pro-survival cytokine G-CSF , one hypothesis is that neutrophil accumulation fuels a feed-forward inflammatory circuit , and continued inflammatory functions of neutrophils drive cytokine production by nearby immune and epithelial cells . However , neutrophils are capable of pro-resolution functions in addition to their more well-characterized pro-inflammatory roles [50] , [51] , so it is possible that Tregs may promote a pro-resolution function in neutrophils , in addition to preventing accumulation of inflammatory neutrophils . This could possibly explain why neutrophil depletion did not rescue all aspects of recovery since neutrophil depletion may eliminate pro-resolution factors in addition to detrimental pro-inflammatory factors . Finally , even though other cell numbers in the respiratory tract were not increased , Tregs may target and functionally alter other cells during recovery , including dendritic cells , monocytes/macrophages , and CD45− respiratory epithelial cells . A potential alternative Treg target is the cellular source of IL-6 . We observed an apparent increase in IL-6 immediately following Treg depletion although this trend did not reach statistical significance until 6 days after Treg depletion . Based on our gene expression analysis from sorted lung cells late after infection , BAL IL-6 could be derived from either CD11b+ monocytes/macrophages or CD45− cells , which include lung epithelial cells , endothelial cells , and other stromal cells such as fibroblasts , and these cells could be novel targets of Treg suppression during the recovery phase . Furthermore , IL-6 is a potent pro-inflammatory cytokine which could have local effects to promote neutrophil survival in the lung [52] , as well as systemic effects on eating/weight gain as it enters the circulation from the damaged respiratory tract [53] , which could contribute to continued weight loss . Furthermore , our studies of the resolution phase after IAV infection has been limited to analyses of cytokine mediators and inflammatory cells , but recent studies have highlighted the importance of lipid mediators ( e . g . resolvins and lipoxins ) [10] and growth factors ( e . g . amphireggulin ) [8] in the resolution of lung inflammation and injury . Future studies of Tregs in the control of resolution after respiratory infection should include analyses of other classes of soluble mediators , which would perhaps help identify alternative Treg targets . Interestingly , both growth factors and lipid mediators are largely derived from epithelial cells , which could represent a novel Treg target in resolution . Taken together , this report describes a novel pro-resolution role for CD86 co-stimulation late after IAV infection . CD86 is required for a robust Treg response during the recovery phase after IAV clearance . Treg cells control the extent of pulmonary neutrophilia during the resolution phase of infection and regulate the expression of several cytokines of innate immune origin released into the recovering lungs . Treg cells promote resolution of inflammation , at least in part , by suppressing neutrophil-dependent cytokine production in the respiratory tract after virus clearance . Finally , since this late onset Treg cell response does not impact IAV clearance , the CD86-dependent Treg response could be a viable therapeutic target to suppress excess inflammation/injury without interfering with virus clearance .
BALB/c , C57BL/6 , and DEREG mice were purchased from the National Cancer Institute ( NCI ) . All mice were housed at the University of Virginia in a pathogen-free environment . Mice used in experiments were between 8–12 weeks old and matched for age and sex . Type A influenza virus A/PR/8/34 ( H1N1 ) was grown in day 10 chicken embryo allantoic cavities as described previously [54] . Mice were infected with 300 egg infectious doses ( EID50 ) of A/PR/8/34 i . n . ( corresponding to a 0 . 1 LD50 dose ) unless otherwise noted . Treg cells were depleted from DEREG mice by administration of 40 ug/kg diphtheria toxin i . p . at the indicated day post influenza infection . Mice were sacrificed by cervical dislocation . Lungs were perfused through the right heart with 10 mL PBS to remove cells from the vasculature . To prepare single cell suspension , lungs were minced and digested in media containing 183 U/mL collagenase D ( Worthington ) for 45 minutes at 37°C . Lung tissue was then disrupted through a steel screen , and red blood cells were lysed with ACK lysis buffer . Live cells were determined by trypan blue exclusion and counted with a hemocytometer . To prepare total lung RNA , lungs were processed with an electric homogenizer in 1 mL TRIzol ( Invitrogen ) and stored at −80°C . RNA was extracted from TRIzol ( Invitrogen ) homogenized samples , and cDNA was generated using Superscript III ( Life Technologies ) according to the manufacturers' protocols . qPCR was performed on a Life Technologies StepOne instrument using SYBR Green ( Life Technologies ) according to manufacturer's instructions . Relative gene expression is calculated by the following formula: 2∧ ( ΔCt ) , where ΔCt = Ct ( HPRT ) - Ct ( gene of interest ) . PCR primer sequences are as follows: Flu PA ( Fw 5′-CGG TCC AAA TTC CTG CTGCTG A-3′ and Rev 5′-CAT TGG GTT CCT TCC ATC CA-3′ ) , Flu NP ( Fw 5′-AGG GTC GGT TGC TCA CAA GT-3′ and Rev 5′-TGC TGC CAT AAC GGT TGT TC-3′ ) , HPRT ( Fw 5′-CTC CGC CGG CTT CCT CCT CA-3′ and Rev 5′-ACC TGG TTC ATC ATC GCT AAT C-3′ ) , LIF ( Fw 5′-ATG TGC GCC TAA CAT GAC AG-3′ and Rev 5′-TAT GCG ACC ATC CGA TAC AG-3′ ) , CSF3 ( Fw 5′-ATG GCT CAA CTT TCT GCC CAG-3′ and Rev 5′-CTG ACA GTG ACC AGG GGA AC-3′ ) , IL6 ( Fw 5′-ACG GCC TTC CCT ACT TCA CA and Rev 5′-TCC AGA AGA CCA GAG GAA ATT TT-3′ ) , and Eotaxin ( Fw 5′-CAG ATG CAC CCT GAA AGC CAT A-3′ and Rev 5′-TGC TTT GTG GCA TCC TGG AC-3′ ) . The following mAbs were purchased from BD or eBioscience ( unless otherwise stated ) , as conjugated to FITC , Alexa-488 , PE , PE-Cy7 , PerCP-Cy5 . 5 , APC , Alexa Fluor 647 , APC–Alexa Fluor 780 , or biotin: CD4 ( GK1 . 5 ) , CD4 ( L3T4 ) , CD8α ( 53–6 . 7 ) , CD11b ( M1/70 ) , CD11c ( HL3 ) , CD19 ( 1D3 ) , CD25 ( PC61 ) , CD25 ( 7D4 ) , CD45 ( 30-F11 ) , CD80 ( 16-10A1 ) , CD86 ( GL-1 ) , CD90 . 2 ( 53–2 . 1 ) , Gr-1 ( RB6-8C5 ) , SigLecF ( E50-2440 ) , Ly6G ( 1A8 ) , Ly6C ( AL-21 ) , I-Ad ( AMS-32-1 ) , CTLA-4 ( UC10-4F10-11 ) , FoxP3 ( FJK-16s ) , and CD28 ( 37 . 51 ) . Anti–mouse CD16/32 used for Fc receptor blocking was isolated and purified in the University of Virginia Hybridoma Core Facility . Cells were suspended in buffer containing PBS , 2% FBS , 10 mM EDTA , and 0 . 01% NaN3 . Fc receptors were blocked with anti-mouse CD16/32 , and then cells were incubated with specific monoclonal antibodies or fluorescence minus one controls for 30 minutes at 4°C . Where indicated , after surface staining , intracellular FoxP3 staining was performed using the FoxP3 fixation/permeabilization kit ( eBiosciences ) . Flow cytometry was performed on FACS Canto flow cytometers ( BD ) , and data were analyzed using FlowJo ( Tree Star , Inc . ) . Bronchial alveolar lavage ( BAL ) fluids were harvested by cannulating the trachea and injecting , then withdrawing 0 . 5 mL PBS into the airways three times . Cells were removed by centrifugation and supernatants were stored at −80°C until analyzed . Cytokines were quantified by a multiplex Luminex assay ( University of Virginia Flow Cytometry Core Facility ) . For in vivo CD86 and CD80 blockade , mice were given 200 µg anti-CD86 ( clone GL-1 from Bio X Cell ) or anti-CD80 ( clone 16-10A1 from Bio X Cell ) via i . p . injection at noted day post influenza infection . For neutrophil depletion , mice were injected i . p . with 50 µg anti-Ly6G ( clone 1A8 from Bio X Cell ) at the indicated days post influenza infection . Matching concentration of isotype antibody was injected for control mice . We used a tissue culture infectious dose 50 ( TCID50 ) assay followed by a hemagglutination assay to quantify infectious virus in BAL fluid , as previously described [17] . In brief , we infected Madin-Darby canine kidney cells with 10-fold dilutions of BAL fluid from infected mice , then incubated the cultures for 3 days at 37°C . Supernatants were collected and mixed with a half volume of 1% chicken red blood cells ( Charles River Spafas ) in PBS , and TCID50 units were calculated from hemagglutination patterns . Mice were irradiated with 1050 rads and , within 24 hours , received an i . v . infusion of red blood cell-lysed bone marrow cells ( 1×106 cells ) from uninfected DEREG mice . DEREG bone marrow chimeras were allowed to reconstitute for 8 weeks before IAV infection . CD25+CD4+ Treg cells were isolated from spleens of uninfected mice using the MACS Regulatory T cell Isolation kit ( Milltenyi ) according to the manufacturer's instruction . A total of 2×106Tregs or CD25−CD4+ control cells were injected i . v . at indicated days post influenza infection . PR8-infected BMDCs were generated as follows: total bone marrow was isolated from uninfected mice and cultured in the presence of GM-CSF for one week to generate BMDCs . On day 7 of culture , BMDCs were harvested and incubated overnight with 1 mL PR8 virus stock at 37°C , thus loading both class I and class II MHC molecules with viral antigen . The following day , BMDCs were washed to remove free virions and subsequently cultured with lung cell suspensions in a 3∶2 ratio for 5 hours in the presence of GolgiStop ( BD ) . After culture , cells were stained for surface markers then fixed and permeablized ( BD Cytofix/Cytoperm ) and stained for intracellular cytokines . Unless otherwise noted , a student T test was used to compare two treatment groups . Groups larger than two were analyzed with the one-way analysis of variance test . Comparisons of two or more groups over time were analyzed with the two-way analysis of variance test followed by the Bonferroni post-test . These statistical analyses were performed using Prism5 software ( for Windows; GraphPad Software , Inc . ) . Results are expressed as means ± SEM . Values of P<0 . 05 were considered statistically significant ( * ) . All animal experiments were conducted in accordance with the Animal Welfare Act ( Public Law 91–579 ) and the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health ( OLAW/NIH , 2002 ) . All animal experiments were carried out in accordance with protocols approved by the University of Virginia Animal Care and Use Committee ( Protocol Number 2230 ) . For anesthesia , a mixture of Ketamine ( 20 mg/ml ) /Xylazine ( 2 mg/ml ) was injected intraperitoneally . Mice were euthanized by cervical dislocation . | Influenza A virus ( IAV ) infection can cause severe inflammation and injury in the respiratory tract , which must be resolved and repaired for the host to fully recover after virus clearance . Evidence is emerging that host immune responses may regulate tissue repair and resolution of inflammation after IAV infection . Early in IAV infection , the co-stimulatory molecules CD80 and CD86 promote inflammation through triggering IAV-specific T cell responses , but no role for CD80/86 in recovery after virus clearance has been previously established . By in vivo antibody-mediated blockade of CD80 or CD86 after virus clearance , we found that engagement of CD86 ( but not CD80 ) was required for optimal recovery after influenza infection . Furthermore , we determined that CD86 was essential for maintaining the FoxP3+regulatory T cell ( Treg ) population in the respiratory tract , and CD86-dependent Tregs promoted recovery by suppressing pulmonary inflammation and supporting regain of weight after virus clearance . In addition , we demonstrated that Tregs suppress neutrophils late after infection , preventing neutrophils from driving excess inflammatory cytokine release into the airways . Taken together , we propose a novel role for CD86 engagement late after IAV infection to promote resolution of inflammation and host recovery through a Treg-dependent mechanism . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"inflammation",
"infectious",
"diseases",
"cytokines",
"clinical",
"immunology",
"immunity",
"molecular",
"development",
"influenza",
"biology",
"and",
"life",
"sciences",
"immunology",
"pulmonary",
"immunology",
"immunoregulation",
"viral",
"diseases",
"immune",
"response",
"immune",
"system"
] | 2014 | Late Engagement of CD86 after Influenza Virus Clearance Promotes Recovery in a FoxP3+ Regulatory T Cell Dependent Manner |
We have previously evaluated the vaccine efficacies of seven tetraspanins of Echinococcus multilocularis ( Em-TSP1–7 ) against alveolar echinococcosis ( AE ) by subcutaneous ( s . c . ) administration with Freund's adjuvant . Over 85% of liver cyst lesion number reductions ( CLNR ) were achieved by recombinant Em-TSP1 ( rEm-TSP1 ) and -TSP3 ( rEm-TSP3 ) . However , to develop an efficient and safe human vaccine , the efficacy of TSP mucosal vaccines must be thoroughly evaluated . rEm-TSP1 and -TSP3 along with nontoxic CpG ODN ( CpG oligodeoxynucleotides ) adjuvant were intranasally ( i . n . ) immunized to BALB/c mice and their vaccine efficacies were evaluated by counting liver CLNR ( experiment I ) . 37 . 1% ( p<0 . 05 ) and 62 . 1% ( p<0 . 001 ) of CLNR were achieved by these two proteins , respectively . To study the protection-associated immune responses induced by rEm-TSP3 via different immunization routes ( i . n . administration with CpG or s . c . immunization with Freund's adjuvant ) , the systemic and mucosal antibody responses were detected by ELISA ( experiment II ) . S . c . and i . n . administration of rEm-TSP3 achieved 81 . 9% ( p<0 . 001 ) and 62 . 8% ( p<0 . 01 ) CLNR in the liver , respectively . Both the immunization routes evoked strong serum IgG , IgG1 and IgG2α responses; i . n . immunization induced significantly higher IgA responses in nasal cavity and intestine compared with s . c . immunization ( p<0 . 001 ) . Both immunization routes induced extremely strong liver IgA antibody responses ( p<0 . 001 ) . The Th1 and Th2 cell responses were assessed by examining the IgG1/IgG2α ratio at two and three weeks post-immunization . S . c . immunization resulted in a reduction in the IgG1/IgG2α ratio ( Th1 tendency ) , whereas i . n . immunization caused a shift from Th1 to Th2 . Moreover , immunohistochemistry showed that Em-TSP1 and -TSP3 were extensively located on the surface of E . multilocularis cysts , protoscoleces and adult worms with additional expression of Em-TSP3 in the inner part of protoscoleces and oncospheres . Our study indicated that i . n . administration of rEm-TSP3 with CpG is able to induce both systemic and local immune responses and thus provides significant protection against AE .
Alveolar echinococcosis ( AE ) , caused by E . multilocularis , is known as a very important zoonotic disease , which is endemic in the large areas of the Northern Hemisphere [1] and is often life-threatening . E . multilocularis infection of intermediate hosts ( humans and rodents ) occurs after oral ingestion of mature oncosphere-containing eggs . In the small intestine , the oncospheres hatch out and then migrate via the hepatic vein to the liver , where they form cyst masses and increasingly transform into multiple vesicles filled with fluid and protoscoleces . The parasitic vesicles are lined with a germinal layer ( GL ) and a laminated layer ( LL ) , which are immediately surrounded by an exuberant granulomatous response generated by the host immune system [2] , [3] . Development/infection of E . multilocularis larvae in host intestine , blood and liver is characterized by systemic and/or mucosal immune responses . However , it doesn't mean that all the immune responses are protection-associated . To the contrary , some are modulated by the parasites and are thus susceptibility-associated . In particular , during the chronic stage of infection , protective immune responses are down-regulated by Echinococcus parasites using some molecules for benefiting their long-term survival in the intermediate host liver [4]–[7] . Studies of immunological profiles showed that , in the infected intermediate host , early Th1-polarized cytokine production , which can kill the metacestodes at the initial stages of development , shifts to a predominantly Th2 response during the chronic stage [4] , [6] , [8] . It is believed that in Echinococcus infection , Th2 responses are mainly associated with susceptibility to Echinococcus infection , whereas Th1 responses contribute to protection [5] , [6] , [8]–[13] . As was shown , some of the proteins expressed on the surface of , or excreted by cestode parasites are involved in immunoregulations , whereby the parasites escape host immune attack and survival in the long term [14]– . Therefore , suppressing/interfering with the function of these proteins using specific antibodies or immune-associated cytokines are key points considerable for efficient vaccine design . Much progress has been made in vaccine development against Schistosoma parasite infection using a surface protein , tetraspanin [15]–[17] . In our previous study , seven tetraspanins have been identified in E . multilocularis larvae and are used to develop vaccines against E . multilocularis infection , which induced significant levels of protection when subcutaneously administered with Freund's adjuvant [18] . Remarkably , vaccinations with rEm-TSP1 and -TSP3 were shown to induce strong serum IgG immune responses in immunized BALB/c mice and received an >85% of liver cyst lesion number reductions ( CLNR ) after orally challenged with parasite eggs . However , due to the toxicity of Freund's adjuvant [19] , [20] , an extensive application of this vaccine model in humans is not feasible . Of the adjuvants used to develop anti-helminth vaccines , CpG ODN has been proved an ideal choice for its non-toxicity and ability enable to induce strong systemic and/or local protective immune responses [21] . Many studies developing anti-protozoan and -helminth vaccines used CpG as an adjuvant [22]–[24] . Evaluation of the adjuvant efficacy and safety of CpG in primates , including humans [25] , [26] , made it possible for developing safe human vaccines . CpG was reported to induce strong anti-parasite mucosal immune responses [22] . Mucosal administration is painless and easier than other administration routes and able to induce more specific antibodies , predominantly in local secretions , against pathogens invasion [27] . Intranasal administration ( i . n . ) , the most efficient mucosal delivery route for antigens , has the following properties that make it the priority route in the present study . First , it is thought to confer the highest level of mucosal immunity , which is capable of priming a full range of local immune responses ( so-called ‘common mucosal immune system’ ) as well as systemic immune responses against protective antigens [28] , and only requires a small antigen dose [29] , [30] . Second , it does not require injection and is therefore safe and painless . Third , it does not require trained medical personnel for delivery and is thus more appropriate for mass vaccination programmes , especially in under-developed countries [27] . To date , i . n . immunization of antigens against helminth infection has achieved much in Ascaris suum [31] , [32] , Trichinella spiralis [33] , Schistosoma mansoni [34] and E . granulosus [35] . As a pilot study on evaluation of i . n . vaccine efficacy of TSPs , two independent experiments were performed by us . In experiment I , we compared the i . n . vaccine efficacy between rEm-TSP1 and -TSP3 which showed >85% of liver CLNR in our previous study , when used as an s . c . vaccine [18] . In experiment II , we evaluated the vaccine efficacy of rEm-TSP3 plus Freund's adjuvant ( s . c . ) versus rEm-TSP3 plus CpG adjuvant ( i . n . ) under the same experimental conditions . The systemic ( serum ) and local ( nasal cavity , intestine and liver ) immune responses induced by both administration routes were investigated and their possible roles in protection discussed .
This study was carried out in strict accordance with the recommendations set out in the Guidelines for Animal Experimentation of the Japanese Association for Laboratory Animal Science and the protocol for the animal experiments was approved by the ethics committee of Hokkaido University ( Permit Number: 09-0144 ) and the Hokkaido Institute of Public Health ( Permit Number: K20-6 ) . All surgery was performed under isoflurane anesthesia , and all efforts were made to minimize suffering . The regions encoding the LEL ( large extracellular loop ) domain of Em-TSP1 and Em-TSP3 were amplified from the full-length enriched cDNA library of E . multilocularis metacestode , subcloned into the pBAD/Thio-TOPO vector ( Invitrogen , USA ) , expressed and purified as previously described [18] . Briefly , E . coli TOP10 cells ( Invitrogen , USA ) were transformed with a recombinant plasmid according to the manufacturer's instructions ( pBAD/TOPO® ThioFusion™ Expression Kit , Invitrogen , USA ) . Recombinant proteins were purified from E . coli lysates using a HisTrap affinity column under nondenaturing conditions ( HisTrap FF crude 1 ml , GE Healthcare , USA ) and stored at −80°C . A pBAD/Thio-TOPO vector without inserts was also expressed and thioredoxin ( TRX ) was purified as a negative control . Genbank accession number for tetraspanins used in this study and those referred in the text were listed in Table S1 . Cyst tissue and protoscoleces of Hokkaido isolate were derived individually from infected Mongolia gerbils and cotton rats . Adult worms were isolated from the intestine of an infected dog . Samples were transferred to freezing medium ( Tissur-O . C . T Compound , Miles , USA ) and stored at −80°C . Cryosections ( 5 µm ) were cut on a Leica CM1900 rotatory microtome ( Leica , Germany ) at −20°C , mounted on slides , and then fixed in acetone for 10 min . After air drying , the slides were re-hydrated in phosphate buffered saline ( PBS ) and endogenous peroxidase was inactivated by incubating for 10 min in 0 . 3% hydrogen peroxide ( H2O2 ) ( in methanol ) . Samples were washed with PBS for 5 min and incubated with rabbit anti-rEm-TSP1 and -TSP3 antibodies respectively , at a dilution of 1∶600 in 3% BSA/PBS for 1 h . After an additional washing step as above , the slices were incubated with Alexa Fluor R488 goat anti-rabbit IgG ( H+L ) ( Invitrogen , USA ) at a dilution of 1∶2 , 000 in 3% BSA/PBS for 1 h . Following three washing steps , the stained samples were embedded in glycerol/phosphate buffer ( v/v , 9∶1 ) and viewed under an Olympus BX50 fluorescence microscope ( Olympus , Japan ) . All procedures were carried out at room temperature . Pre-immune rabbit serum was used as a negative control . Five-week-old BALB/c mice ( male ) were maintained in cages in a P3 animal room at 23–25°C with a 12 h light/12 h dark cycle . Litter was cleaned weekly . They were provided with food and water ad libitum and immunized at the age of 6 weeks . E . multilocularis ( Hokkaido isolate ) eggs were collected from the feces of an experimentally-infected dog . Eggs were microscopically observed to confirm their morphological integrity before challenge ( classical morphology of an egg isolated from the dog is shown in Figure S1 ) . The data were analyzed using one-way ANOVA followed by a multiple comparison Tukey's test . Differences were considered statistically significant at p<0 . 05 , very significant at p<0 . 01 and extremely significant at p<0 . 001 .
Staining revealed immunolocalization of Em-TSP1 and Em-TSP3 at the surface ( germinal layer/tegument ) of both forms of E . multilocularis larva ( cyst and protoscolex ) ( Figure 1A , B ) and both of the antigens were also detected on the tegument of the adult worms ( Figure 1C ) . Interestingly , expression of Em-TSP3 was also detected on the sucker and rostellum of the protoscoleces ( Figure 1B ) and on the oncospheres in adult eggs ( Figure 1C , D ) . Pre-immunized serum , used as a negative control , had no obvious reaction with any of the fixed tissues ( Figure 1 ) . Microscopic images of H&E-stained larvae , adults and eggs are also shown as a reference ( Figure 1 ) . In experiment I , vaccination of mice with rEm-TSP3 plus CpG significantly reduced the liver cyst lesion numbers ( 62 . 1% , p<0 . 001 ) , compared with that in the PBS control group . The liver CLNR after immunization with rEm-TSP1 plus CpG was significant ( p<0 . 01 ) , but lower ( 37 . 1% ) than that after immunization with rEm-TSP3 . A mixture of rEm-TSP1 and -TSP3 also resulted in a significant , but lower ( 28 . 7% , p<0 . 05 ) , rate of liver CLNR than either rEm-TSP1 or rEm-TSP3 . The CLNRs in the CpG and TRX control groups were 10 . 9% and 22 . 1% ( p>0 . 05 ) , respectively ( Table 3 ) . The difference in the vaccine efficacy between rEm-TSP1 and -TSP3 was statistically significant ( p<0 . 05 ) . In experiment II , s . c . immunization with rEm-TSP3 significantly reduced the number of cyst lesions ( 82% , p<0 . 001 ) compared with that in the control group . The liver CLNR in the TRX and CpG control groups were 44% ( p<0 . 05 ) and 27% ( p<0 . 05 ) , respectively . I . n . immunization with rEm-TSP3 plus CpG resulted in a CLNR of 61% ( p<0 . 01 ) , whereas rEm-TSP3 alone , and the TRX and CpG controls showed reductions of 22% ( p>0 . 05 ) , 38% ( p<0 . 05 ) and 37% ( p<0 . 05 ) , respectively ( Table 4 ) . There was a significant difference between the vaccine efficacies of the two administration routes ( p<0 . 05 ) . Antibody responses against rEm-TSP3 evoked by s . c . and i . n . administration were detected by ELISA . Compared with the PBS control , significant IgG responses were detected in the groups immunized with rEm-TSP3+CFA/IFA ( p<0 . 001 ) and rEm-TSP3+CpG ( p<0 . 001 ) ( Figure 2A ) . A significant difference was observed between the two administration routes ( p<0 . 001 ) . In the i . n . group , a significant difference was observed between rEm-TSP3+CpG and rEm-TSP3 ( p<0 . 001 ) alone . Strong IgG1 and IgG2α subclass antibody responses were induced by both s . c . and i . n . administration of rEm-TSP3 with adjuvants ( p<0 . 001 ) , while rEm-TSP3 alone induced neither significant IgG1 nor IgG2α responses ( Figure 2B , C ) . Significant IgM antibody responses were detected in both s . c . ( p<0 . 001 ) and i . n . ( p<0 . 05 ) groups immunized with rEm-TSP3 plus adjuvant , but there was no significant difference observed between them . rEm-TSP3 alone did not induce a significant IgM antibody response ( Figure 2D ) . A relatively strong serum IgA antibody response was detected in the group immunized with rEm-TSP3+CpG ( p<0 . 001 ) ( Figure 2E ) . In the nasal cavity , an extremely high IgA response ( p<0 . 001 ) was only detected in the group i . n . immunized with rEm-TSP3+CpG ( Figure 3A ) . I . n . immunization with rEm-TSP3+CpG induced higher intestinal IgA responses ( p<0 . 001 ) than the other groups ( Figure 3B ) . In the liver extracts , significant IgA responses were detected in the groups immunized with rEm-TSP3+CFA/IFA ( s . c . ) ( p<0 . 001 ) and rEm-TSP3+CpG ( i . n . ) ( p<0 . 01 ) , with the former being slightly higher ( Figure 3C ) . Th1 and Th2 cell responses were assessed according to the IgG1/IgG2α ratio ( Table 5 ) . Two weeks post-immunization , s . c . immunization induced a mixed Th1/Th2 response , with the Th2 response predominating . This shifted to a reduced IgG1/IgG2α ratio 3 weeks post-immunization ( a Th1 tendency ) . Inversely , i . n . immunization resulted in a tendency to shift from a Th1-dominant response to Th2-dominant one . Complement-mediated lysis of protoscoleces began 2 h after the addition of untreated sera ( from both non-immunized and immunized groups ) and most were lysed at 8 h ( Figure 4A , D ) ; however , no lysis of protoscoleces treated with complement-inactivated sera ( Figure 4B , E ) was observed . Also , there was no visible protoscolex lysis after inactivation of C3 proactivator at 50°C for 30 min ( Figure 4C , F ) .
As a transmembrane protein , tetraspanin is abundantly expressed on the tegument or body wall of some helminthes , including Schistosoma mansoni [15] , [36] , S . japonicum [16] and Caenorhabditis elegans [37] , and is believed to play very important roles in signal transduction , cell proliferation , adhesion , migration , fusion and host-parasite interactions [38] , [39] . Recently , we cloned seven tetraspanins ( Em-TSP1–7 ) from E . multilocularis [18] , of which only Em-TSP5 ( E24 ) was confirmed its location on the surface of the larvae ( cysts and protoscoleces ) [40] . In this study , immunohistochemical analysis showed that both Em-TSP1 and -TSP3 are expressed on the surface of larvae and adult worms . Notably , the expression of Em-TSP3 on sucker , rostellum and inner tegument of protoscoleces and oncospheres was also observed . From these results we proposed that these proteins play important roles in Echinococcus-host interactions and that using them as vaccines may interfere with the parasite survival strategy [15] , [17] , [41]–[44] . Moreover , as reported previously , tetraspanins showed their potential protective effects against different stages of Schistosoma infections [15] , [41]–[43] . Thus , vaccination of Em-TSP1 and -TSP3 proteins is believed to provide ‘broad-spectrum protection’ against the different stages infection by E . multilocularis . We previously showed the high protective efficacy of seven tetraspanins ( TSP1–7 ) against E . multilocularis infection in BALB/c mice after s . c . administration with Freund's adjuvant [18] . However , the toxicity of Freund's adjuvant limits the application of this vaccine model to human vaccines [19] , [20] . Therefore , in this study , we used nontoxic CpG ODN as a vaccine adjuvant , which induces both systemic and mucosal immune responses in immunized animals [21] , and evokes strong protective immune responses when used as mucosal adjuvant [22] , [25] , [45]–[47] . Studies on evaluation of CpG adjuvant efficient in primates ( including humans ) undoubtedly is an important step in human mucosal vaccine development [25] , [26] . Recently , CpG was used as a mucosal adjuvant in developing anti-protozoan and -helminth vaccines [22]–[24] . Of the different mucosal administration routes , i . n . delivery is the most appropriate , mainly because it induces the full range of local immune responses ( so-called ‘common mucosal immune system’ ) [28] and induces strong immunity after only small dose of vaccine [29] , [30] . Based on the above properties , i . n . immunization and the CpG adjuvant has many advantages; therefore , we evaluated the efficacy of i . n . delivery of the rEm-TSP1 and -TSP3 vaccines using CpG as an adjuvant , and further investigated the systemic and mucosal immune responses mounted against rEm-TSP3 , compared with those induced by s . c . immunization using Freund's adjuvant . In experiment I , significantly high liver CLNRs , resulted from i . n . immunization with rEm-TSP1 plus CpG ( 37 . 1% , p<0 . 01 ) and rEm-TSP3 plus CpG ( 62 . 1% , p<0 . 001 ) , were observed compared with the PBS control . However , our previous study reported the efficacies of 87 . 9% and 85 . 1% , respectively , for the rEm-TSP1 and -TSP3 vaccines , when administered s . c . with Freund's adjuvant [18] . Taken together , our own results and those of others [48] , conclude that different antigen delivery routes greatly affect vaccine efficacy . In experiment II , we focused on the rEm-TSP3 protein , which showed higher liver CLNR than rEm-TSP1 in experiment I . We investigated systemic and mucosal immune responses associated with protection in BALB/c mice immunized s . c . or i . n . Under the same conditions , the vaccine efficacies induced by the two immunization routes were similar to those observed in Experiment I and our previous study [18] . ELISA data showed that , after the third immunization , extremely significant serum IgG immune responses were induced by both the administration routes ( p<0 . 001 ) , with the former stronger than the latter ( p<0 . 001 ) . Meanwhile , significant levels of IgG1 , IgG2α and IgM antibodies were detected . Th1 and Th2 cell responses were evaluated according to the IgG1/IgG2α ratio . After the second immunization , a mixed Th1/Th2 cell response was evoked in the s . c . -immunized group , dominated by a Th2 response , whereas the IgG1/IgG2α ratio reduced after the third immunization ( a Th1 tendency ) . Inversely , i . n . immunization resulted in a shift from a Th1 to a Th2 response . As previously reported , antibodies form a critical part of the immune response against taeniid metacestodes , with IgG1 , IgG2α , IgG2β , and IgE play a major role in oncosphere killing , although the involvement of other mechanisms can not be ruled out [4] . Antibody-dependent , complement-mediated lysis is a pivotal characteristic during the early stages of taeniid cestodes infection of intermediate hosts [4] , [49] . Interestingly , our in vitro complement assay showed the lytic effect on protoscoleces by both normal/non-immunized and rEm-TSP3-immunized serum from BALB/c mice . Notably , treatment of sera at 56°C for 30 min ( complement inactivation ) or at 50°C for 30 min ( C3 proactivator inactivation ) abolished protoscolex lysis , suggesting that at least an alternative pathway exists in BALB/c mice for complement activation in the initial infection of Echinococcus metacestode . Growing evidences suggest the importance of a Th1/Th2 balance during parasite infections , such as infections by Trypanosoma [50] , Schistosoma [15] , [17] and Echinococcus [4]–[8] , [10]–[13] . In Echinococcus infection , early Th1-polarized cytokine production , which can kill the metacestodes at the initial stages of development , shifts to a Th2 cytokine response during the chronic stage [4] , [6] , [8] . The characteristic of Th1 profile , mainly induced by IL-12 , is the secretion of IL-2 , TNF and especially IFN-γ , which lead to the recruitment and activation of the cellular effector phase of immunity [51] , [52] . A shift from a Th1 to a Th2 cytokine profile , mainly induced by an increased secretion of IL-10 , a cytokine typically associated with immunoregulation of effector responses , is thought to limit and ultimately terminate inflammatory responses [53]–[55] . Echinococcus metacestode has developed a number of strategies for escaping host immune attack [4]–[7] . The shift from a protective immune response ( Th1 response ) to a nonprotective one ( Th2 response ) is thought to be one of the most important mechanisms , whereby Echinococcus metacestodes regulate host immune responses to benefit their long-term survival in the hosts , by using some molecules like antigen B [56] . Tetraspanins in the tegument of schistosomula and adult worms are suggested to act as receptors for host ligands , including MHC molecules , by which parasites mask their nonself status , and thereby escape host immune responses [15] . In the present study , because CpG was expected to ( but actually failed to ) elicit significant Th1 cell responses as previously reported [21] , we hypothesize that the tetraspanins used also share the same immunomodulatory mechanism as antigen B . This would offset of the adjuvant activity of CpG ( and even Freund's adjuvant ) , which would provide a reasonable explanation for the induction of predominant Th2 cell responses in this study . It is clear that systemic immune responses , as mentioned above , play a crucial role in protection against Echinococcus parasite infections . However , since the early , natural infection of eggs/oncospheres begins at the gastrointestinal membrane and the invasion , rooting and development/proliferation of Echinococcus metacestodes occurs in the host organs ( mainly the liver ) , the immunological events occurring at the local mucosa should not be neglected . Secretory IgA is a critical component of the mucosal immune system and plays an important role as the first lines of defense against many parasite infections , such as Giardia [57] , [58] and Echinococcus [59] . IgA responses in the mucosa might be hypothesized to target the parasite by neutralizing parasite ES products , attenuating the parasite-host interaction and interfering with parasite feeding and survival [60]–[62] . Induction of eosinophil degranulation by IgA is another important characteristic of mucosal immunological events as observed by Abu-Ghazaleh et al . [63] . Additionally , mucosal immunity could play a role in tolerance induction against E . multilocularis that may be a prerequisite for the subsequent development of the larvae in the liver , and for the occurrence of alveolar proliferation [59] . In the present study , a remarkably strong intranasal IgA response was induced ( p<0 . 001 ) even though only a very low level of serum IgA was induced , by i . n . immunization with rEm-TSP3 plus CpG; however , only a limited intestinal IgA response was detected . rEm-TSP3 alone ( without any adjuvant ) failed to induce a high IgA response in either i . n . -or s . c . -immunized groups . It is noteworthy that high levels of IgA antibodies were detected in liver extracts after immunization by both immunization routes . Although the exact mechanisms underlying IgA anti-parasite response in the liver are unknown , infection of this organ by E . multilocularis larvae is characterized by a chronic process; antibody responses in the liver , including IgA , as well as other immune-associated factors should not be neglected [6] , [64] , [65] and further studies regarding their roles are required . Taken together , the results of the present study suggest that i . n . administration of rEm-TSP3 plus CpG , which induces systemic and local immune responses , is a prospective model for the development of nontoxic human vaccines . It is also proposed that , in our vaccine model , protection is mainly provided by serum antibodies including antibody-dependent , complement-mediated lysis against early infection by E . multilocularis ( oncospheres in blood ) . Although intestinal IgA is believed to play an important role in inhibiting invasion of hatched oncospheres , our vaccines did not induce satisfactory levels of IgA antibodies; this will be an important consideration in our further work . Moreover , the extensive expression of Em-TSP3 on the surface/tegument of E . multilocularis at different developmental stages suggests the possibility of developing TSP-based vaccines with ‘broad-spectrum protection’ against the worm infections at different stages ( oncosphere , cyst , protoscolex and even adult ) . On the other hand , the surface/tegument tetraspanin proteins may have crucial functions in protecting the Echinococcus parasite by shifting the host protective immune response ( Th1 response ) to non-protective one ( Th2 response ) , especially during the stages of liver infection . Therefore , further work is needed to develop new strategies for offsetting this ‘undesired effect’ of tetraspanin proteins , such as enhancement of the biological activity of the CpG ODN adjuvant by modifying its backbone chemistry , and the use of different delivery methods , including mixing or cross-linking of CpG ODN to other carrier compounds as reviewed by Mutwiri et al . , elsewhere [66] . | Humans and rodents become infected with E . multilocularis by oral ingesting of the eggs , which then develop into cysts in the liver and progress an endless proliferation . Untreated AE has a fatality rate of >90% in humans . Tetraspanins have been identified in Schistosoma and showed potential as the prospective vaccine candidates . In our recent study , we first identified seven tetraspanins in E . multilocularis and evaluated their protective efficacies as vaccines against AE when subcutaneously administered to BALB/c mice . Mucosal immunization of protective proteins is able to induce strong local and systemic immune responses , which might play a crucial role in protecting humans against E . multilocularis infection via the intestine , blood and liver . We focused on Em-TSP3 , which achieved significant vaccine efficacy via both s . c . and i . n . routes . The adjuvanticity of nontoxic CpG OND as i . n . vaccine adjuvant was evaluated . The widespread expression of Em-TSP3 in all the developmental stages of E . multilocularis , and the strong local and systemic immune responses evoked by i . n . administration of rEm-TSP3 with CpG OND adjuvant suggest that this study might open the way for developing efficient , nontoxic human mucosal vaccines against AE . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"veterinary",
"pathology",
"veterinary",
"diseases",
"pathology",
"veterinary",
"microbiology",
"microbiology",
"veterinary",
"science",
"veterinary",
"medicine"
] | 2012 | A Pilot Study on Developing Mucosal Vaccine against Alveolar Echinococcosis (AE) Using Recombinant Tetraspanin 3: Vaccine Efficacy and Immunology |
Many microorganisms exhibit high levels of intragenic recombination following horizontal gene transfer events . Furthermore , many microbial genes are subject to strong diversifying selection as part of the pathogenic process . A multiple sequence alignment is an essential starting point for many of the tools that provide fundamental insights on gene structure and evolution , such as phylogenetics; however , an accurate alignment is not always possible to attain . In this study , a new analytic approach was developed in order to better quantify the genetic organization of highly diversified genes whose alleles do not align . This BLAST-based method , denoted BLAST Miner , employs an iterative process that places short segments of highly similar sequence into discrete datasets that are designated “modules . ” The relative positions of modules along the length of the genes , and their frequency of occurrence , are used to identify sequence duplications , insertions , and rearrangements . Partial alleles of sof from Streptococcus pyogenes , encoding a surface protein under host immune selection , were analyzed for module content . High-frequency Modules 6 and 13 were identified and examined in depth . Nucleotide sequences corresponding to both modules contain numerous duplications and inverted repeats , whereby many codons form palindromic pairs . Combined with evidence for a strong codon usage bias , data suggest that Module 6 and 13 sequences are under selection to preserve their nucleic acid secondary structure . The concentration of overlapping tandem and inverted repeats within a small region of DNA is highly suggestive of a mechanistic role for Module 6 and 13 sequences in promoting aberrant recombination . Analysis of pbp2X alleles from Streptococcus pneumoniae , encoding cell wall enzymes that confer antibiotic resistance , supports the broad applicability of this tool in deciphering the genetic organization of highly recombined genes . BLAST Miner shares with phylogenetics the important predictive quality that leads to the generation of testable hypotheses based on sequence data .
Rapidly evolving genes are among the most biologically intriguing , yet they are also among the most difficult to analyze . The arms race between host and pathogen , fueled by strong selection pressures , can yield very high levels of sequence variation in certain microbial genes . Diversifying selection is often imposed on the microbial pathogen by the adaptive immune response of the host or through antibiotic therapy . Genetic change can arise through mutation and recombination . These processes often lead to insertions and deletions ( indels ) . Intragenic recombination between divergent sequences can lead to mosaic-like structures . Unless genetic regions of extensive sequence heterogeneity are interspersed with sufficiently long stretches of highly conserved sequence , multiple-sequence alignments can be difficult or impossible to attain . Yet , an accurate and reliable alignment is an essential starting point for many of the analytic tools that provide fundamental insights on gene structure and evolution , such as phylogenetics . Among the many microorganisms that exhibit extensive genetic diversity are pathogenic streptococci , including S . pneumoniae and S . pyogenes . These organisms cause a wide variety of diseases , ranging from mild to severe , and infect humans throughout the world . A hallmark feature of S . pneumoniae is the widespread emergence of resistance to penicillin within the past 40 years . Resistance is conferred by enzymes known as penicillin-binding proteins ( PBPs ) , which primarily function as transpeptidases active in cell wall biosynthesis , but have evolved a lower binding affinity for the inhibitory drug [1 , 2] . Until the recent advent of sequence-based typing [3] , strains of S . pyogenes were defined by serological-based–typing schemes targeting proteins present on the bacterial cell surface in numerous , antigenically distinct forms [4] . Included among these is serum opacity factor ( SOF ) , which also plays a role in virulence . Both streptococcal species are characterized by relatively high rates of genetic recombination resulting from horizontal gene transfer events [5] . Highly diversified microbial genes often play a key role in the pathogenesis of infectious diseases . Many alleles of pbp genes have been identified , and for the most part their sequences can be readily aligned , allowing for detailed structural analysis . However , attempts to produce a multiple sequence alignment of sof alleles have been confounded by their extensive sequence heterogeneity . In an effort to achieve a better understanding of the underlying structural organization of sof , a new bioinformatics tool—BLAST Miner—was developed . The well-studied pbp2x alleles of S . pneumoniae were also assessed by BLAST Miner , in order to test the broader application of this software to highly diversified genes .
Multiple-sequence alignments of alleles corresponding to the pbp2x locus of S . pneumoniae and the sof locus of S . pyogenes are graphically depicted in Figure S1; the pbp2x and sof loci encode PBP-2x and SOF proteins , respectively . Although pbp2x genes display evidence of intragenic recombination and a mosaic-like structure [1 , 2] , the 41 distinct partial alleles of pbp2x can be aligned by ClustalW without a single gap . In striking contrast , 45 . 6% of the positions in the sequence alignment of 139 partial sof alleles consists of gaps . The finding of poor sequence alignment of sof is not limited to ClustalW; the Muscle and MAFFT programs resulted in an even higher proportion of gaps ( Figure S1 ) . The inability to generate a reliable multiple-sequence alignment for sof severely limits application of the many tools that are used to assess gene structure , selection , recombination , and phylogeny . The BLAST Miner program was developed to better address problems encountered by the lack of accurate sequence alignments for highly diversified genes , such as sof . In general terms , this BLAST-based method seeks to identify segments of genes displaying high-sequence homologies , independent of their relative positions along the entire sequence length . To demonstrate the validity and broader application of BLAST Miner , the pbp2x gene was first selected for analysis . The pbp2x input data for BLAST Miner consists of 53 partial sequences that are 1 , 578 bp in length [2] , in a FASTA format . The partial alleles encompass the transpeptidase-coding domain , and together the genes encode PBPs that span a wide range of binding affinities for penicillin . BLAST Miner uses the MegaBLAST hit results of an all-ways pairwise comparison ( Figure S2 ) , for which the 53 pbp2x sequences yielded 10 , 489 BLAST hit records ( Table 1 ) . The parameters for the MegaBLAST step include a word size that is set to 16 bases for the pbp2x analysis . Removal of duplicate sequences and their attendant BLAST hit records reduced the total to 6 , 174 , corresponding to the 41 unique pbp2x alleles present in the multiple-sequence alignment of Figure S1 . In the next step , stringency filtering was performed to remove BLAST hit records that fall below a user-defined percent identity threshold , which was set to 90% for pbp2x and yielded 3 , 220 BLAST hit records ( Table 1 ) . An additional process , denoted as the scanning step , is introduced to reduce the bias that can occur with an extended alignment between two sequences . An extended alignment generates only one BLAST hit record , as opposed to several BLAST hit records having shorter alignment lengths but which together span the same region . The number of bases within the BLAST hit starting region that is used for scanning is user-definable , and was set at 24 for pbp2x analysis . The nucleotide ( nt ) sequences comprising the 3 , 220 BLAST hit records were used to scan the entire pbp2x database for additional exact matches against the starting regions of all original BLAST hits . The scanning step led to an increase in the number of BLAST hit records to 13 , 252 ( Table 1 ) . Since only exact matches ( i . e . , 100% nt sequence identity ) are added , no reduction in stringency results from this process . In effect , this adjustment makes BLAST Miner less sensitive to the size and composition of the sequence dataset . The fundamental unit that is derived from BLAST Miner is referred to as the module . A module is a dataset comprising >1 segment of nt sequence , corresponding to >1 pbp2x allele , whereby each segment is >16 nt in length and has >90% nt identity to at least one other sequence segment assigned to that module . The 16-nt length is based on the word-size setting in MegaBLAST , and the 90% nt identity value reflects the percent identity threshold used for stringency filtering . Segments of sequences that are not included in any BLAST hit record are not assigned to a module and thereby are excluded from further BLAST Miner analyses . Excluded sequences are characterized by their presence in only a single allele and in a nonduplicated form . It is important to emphasize that modules are defined only by their starting position , and that the sequence segments that define a module can vary in length , so long as they exceed the minimum length set by the word size . Module length is indeterminate , but for practical purposes , it is useful to consider the end of a module being the first point at which a subsequent module is defined . The module correlation algorithm ( further described in Materials and Methods , and in Figure S2 ) , was applied to the 13 , 257 BLAST hit records and yielded 38 modules for pbp2x ( Table 1 ) . This algorithm assigns discrete segments of nt sequence to a module based on an iterative process that recognizes related BLAST hit records . The extent of nt identity between any two sequence segments that are assigned to the same module can range from 100% to as low as 90% following the first iteration ( i . e . , equal to percent identity threshold ) , 81% following the second iteration ( i . e . , 90% of 90 ) , 72 . 9% following the third iteration , and so forth . The majority of pbp2x modules ( 33 of 38 , or 87% ) were assembled by only one or two iterations of the module correlation algorithm ( Figure 1A ) , signifying that all sequence segments assigned to these modules were >81% identical to one another . The maximum number of occurrences of any one module among the 41 pbp2x alleles was 41 ( Figure 1B ) . None of the modules occurred more than once within a given pbp2x allele ( unpublished data ) , indicating a lack of intragenic duplications in this dataset . The nt sequence of each pbp2x allele can be converted into a module map . Module maps showing the nt start positions for each module are depicted for three pbp2x alleles ( Figure 2 ) . The module maps for all 41 pbp2x alleles were combined and graphically represented as one interconnected network of modules ( Figure 3 ) . Modules are indicated by the nodes ( circles ) of the network , whereby the diameter of the node is directly proportional to the number of pbp2x alleles harboring that module . Figure 3A shows all 38 modules that were defined for the 41 partial pbp2x alleles . The relative position of each node along the x-axis reflects the average position of the module start site within the alleles in which the module occurs . Each pbp2x allele represents a walk through the module network graph . Arrows depict the connections between contiguous modules , and the thickness of the arrow is proportional to the number of times that the two connected modules lie adjacent when the complete dataset of 41 pbp2x alleles is considered . The direction of the arrowhead depicts the relative order of the two adjacent modules . For the pbp2x network , all arrows point in the 5′ to 3′ direction , indicating that there are no rearrangements in the relative order of modules within any of the pbp2x alleles . This BLAST Miner finding is consistent with the high-quality multiple-sequence alignment obtained using Clustal W ( Figure S1 ) . The module network structure of pbp2x can be used to infer intragenic recombination; however , a systematic method for quantifying recombination remains to be developed . A past history of recombination is strongly suggested by pairs of alleles that follow different pathways into or out of the same node . For example , Module 15 has a single incoming pathway , but five distinct outgoing pathways of connections ( Figure 3A , near 5′ end of graph ) . BLAST Miner is less sensitive at detecting recombination between highly similar sequences as compared with alignment-based methods , since it may fail to detect recombination between pairs of sequences that exceed the percent identity threshold setting . Thus , BLAST Miner provides a conservative estimate of recombination . Figure 3B depicts a simplified pbp2x network , showing only the modules that occur in >10% of the pbp2x alleles and connection pathways that occur at least twice . Although even this simplified network is complex , there are several striking features . The central zone of presumed low recombination shows relatively few distinct types of module-to-module connections within the area bounded by Modules 23 and 8 . The central region is flanked on both sides by a much higher density of distinct module connections , possibly signifying higher levels of recombination . That the central region encoding the transpeptidase domain may have undergone lower rates of recombination is supported by MaxChi analysis , a standard alignment-based recombination detection method ( Figure S3 ) [6] . In MaxChi , recombination breakpoints are detected near the 3′ end of a conserved sequence , whereas in BLAST Miner the module start site is placed near the 5′ end of a highly homologous region . The actual crossover sites probably lie somewhere in between these two sites , and flank the central portion of the transpeptidase domain . The central zone of reduced recombination reveals three major pathways of module connections: Module 25 to 18 ( pathway I ) , Module 25 to 28 ( pathway II ) , and Module 29 to 28 ( pathway III ) ( Figure 3B ) . Of biological relevance is the finding that the three discrete pathways of module connections correlate with drug resistance phenotype , whereby alleles conferring drug susceptibility tend to follow the upper pathway connections ( I ) , alleles conferring resistance tend to follow the lower pathway ( III ) , and alleles that are intermediate in their resistance profile tend to follow pathway II . This finding has further support in a phenogram that is constructed based on the module maps of all 41 pbp2x alleles ( Figure S4 ) . The pbp2x data indicate that BLAST Miner can uncover key biological relationships between genotype and phenotype . Since an accurate alignment can be readily generated for pbp2x alleles , numerous tools are available for gaining an increased understanding of pbp2x gene structure , and BLAST Miner does not necessarily provide additional insights on important structural features of the pbp2x locus . Instead , the well-studied pbp2x genes have served to validate the BLAST Miner application . Unlike pbp2x , the sof alleles of S . pyogenes yield a poor quality alignment that is rich in sequence gaps ( Figure S1 ) . BLAST Miner was developed with the goal of providing an analytic tool for otherwise intractable sequence data , such as that found for sof . The input data for BLAST Miner analysis of sof consists of 249 complete or partial sof sequences that had been previously deposited in Genbank or were generated specifically for this study . The input sequences ranged in length from 319 bp for partial sof sequences , to 6 , 386 bp containing the complete sof gene and flanking sequence ( Table S1 ) . The all-ways pairwise MegaBLAST analysis of sof sequences resulted in ~200 , 000 BLAST hit records ( Table 2 ) . Removal of duplicate records and stringency filtering reduced the number of records to ~78 , 000 , corresponding to 152 unique sof sequences . As done for pbp2x analysis , the word size was set to 16 and the percent identity threshold was set to 90 . The BLAST hit records were used to scan the entire sof sequence dataset for additional exact matches , whereby the module alignment length was set to 24 nt . The scanning process increased the number of BLAST hit records to ~228 , 000 . The module correlation process led to the initial assignment of 269 sof modules . One of the attractive features of BLAST Miner is that the input sequences of the FASTA file do not require prior trimming to a specified length . However , to perform further analysis beyond the initial module assignment , the region of interest within the gene needs to be specified by the user . Anchor modules that define the genetic region of interest are selected via a simple trial-and-error process within BLAST Miner , until the optimal region and/or targeted sequences are bracketed as desired . The anchor modules chosen for sof analysis are proximal to the hybridization sites of the oligonucleotide primers used for PCR amplification of the hypervariable region encoding the amino-terminus of the mature SOF protein [7] . The 5′ end anchor module is composed of a single module , designated Module 1; the sequence segments assigned to Module 1 correspond to a portion of the highly conserved signal peptide-coding region . The 3′ end anchor module of sof , designated Module End , is a composite of three modules—Modules 10 , 51 , and 4—because no single module can satisfactorily bracket the region of interest . Modules 10 and 51 are mutually exclusive , whereby either is present in nearly all sof alleles . Module 4 was added to the composite anchor Module End to capture the last remaining sof sequence . Of the 269 modules that were identified before designation of the anchor modules , ~64% were positioned outside of the anchor-bracketed region . Modules that were not contained within the bracketed region were discarded from the dataset . This reduced the total number of modules for sof to 97 ( Table 2 ) . Both the 5′ end anchor Module 1 and the 3′ end anchor Module End were present in 151 of the 152 sof sequences remaining in the dataset; one sof sequence was truncated and , therefore , was dropped from the dataset . The nt sequences within the anchor-bracketed regions of the remaining 151 sof sequences were compared with one another , and 12 duplicate sequences were identified and removed from the dataset ( Table 2 ) . The final sof dataset contained 139 unique , anchor-bracketed partial sof alleles , ranging in length from 329 to 472 bp ( Table S1 ) . Each sof module comprises segments of nt sequence , whereby each segment is >16 nt in length and has >90% nt identity of at least one other sequence assigned to that module . The sequence similarity between any two sequence segments assigned to the same module may drop below 90% as the number of iterations of the module correlation algorithm increases . However , all sequences assigned to the same module are related through a network of highly similar intermediates . The distribution of the number of iterations used to define each of the 97 modules of sof is depicted in Figure 4A ( the complete dataset is presented in Table S2 ) . Most sof modules were defined by relatively few iterations , with 78% of modules assembled through <3 iterations , and 42% defined in a single iteration . Modules with the highest number of iterations were Module 6 with 28 iterations , and Module 13 with 19 iterations . Although each sof sequence segment assigned to a module has high similarity to more than one other sequence segment within its 5′ end region of 16 nt ( i . e . , the BLAST hit starting region ) , the downstream region can vary widely in the extent of sequence similarity . For sof , it was observed that high levels of sequence similarity typically extended to >24 nt . The sequence segments of a selected module can be trimmed to 24 bp in length and aligned . Figure 5 provides an example of Clustal W alignments of trimmed sequence segments corresponding to four representative modules ( Figure 5A–5D ) , whereby each module was compiled via a different number of iterations of the module correlation algorithm . Module 12 sequence segments are defined by seven iterations of the module correlation algorithm , yielding 32 unique 24-mers ( Figure 5D ) . In theory , based on seven iterations , the sequence identity between any two segments assigned to Module 12 can be as low as 47 . 8% . In actuality , the level of similarity between the 24-mers appears to be much higher . Alignment of the sequence segments corresponding to Module 12 also indicates that the module start site had migrated during the iterative module correlation process by 1 nt to 8 nt , relative to the initial BLAST hit record used in the first iteration . When coupled with the alignment and trimming process for defining 24-mers , the result is two overlapping sets of 24-mer sequence segments offset by 8 nt . Module start site slippage is a common property of high-iteration modules . By definition , each of the 97 sof modules are represented at least twice within the dataset of 139 partial sof alleles . The majority of modules ( 96% ) occur >3 times within the dataset , and 38 modules ( 40% ) occur >20 times ( Figure 4B; Table S2 ) . Fifteen modules occur multiple times within the same allele ( Table S2 ) ; these represent module duplications . The most highly prevalent modules within the sof dataset are Modules 6 and 13 , with 583 occurrences of Module 6 , and 716 occurrences of Module 13 . The mean occurrence of Modules 6 and 13 per sof allele is 4 . 2 and 5 . 2 , respectively . Only four of the 97 modules are present within all 139 partial sof sequences: the anchor Modules 1 and End ( as expected , by definition ) , and the highly prevalent Modules 6 and 13 ( Figure 4C , Table S2 ) . More than 80% of the 97 modules are distributed among at least five sof alleles . The frequency of co-occurrence of modules within individual sof alleles was assessed for all possible module pairs . The observed co-occurrence frequency was compared with the null hypothesis ( expected ) , which states that module co-occurrence is random . A chi-squared goodness-of-fit test with the Yates' correction was used to identify module pairs that tend to be either tightly linked or mutually exclusive . Significance values were Bonferroni-corrected for the 4 , 650 pairwise comparisons performed; the 49 module pairs with adjusted p-values < 0 . 05 are listed in Table S3 . Significant , positive values in the deviation from expected co-occurrence were observed for 43 module pairs , indicative of positive linkage . Positive linkage may be the consequence of tight physical linkage or epistatic interactions leading to increased fitness . Negative values in the deviation from expected co-occurrence were observed for six module pairs , indicative of negative linkage or mutual exclusivity . Unlike pbp2x , there are numerous examples whereby the relative 5′ to 3′ order of specific modules differ among sof alleles . This finding is suggestive of a genetic rearrangement involving co-occurring modules . Since some modules are present multiple times within a single sof allele ( e . g . , Modules 6 and 13 ) , a conservative approach was taken for estimating the number of module pairs having undergone rearrangements in their relative order , whereby only the 5′-endmost position of each module in a given sof allele is used to ascertain the relative module order . The relative order of modules in co-occurring pairs was assessed for all 139 partial sof alleles . Eighty-seven module pairs displayed a switch in their relative order ( unpublished data ) . Figure 6 ( right panel ) shows module maps for two sof alleles , and illustrates a reversal in the relative order of two blocks of modules . Thus , the data provide evidence for a history of aberrant recombination involving sof genes . The module maps of all 139 sof alleles are graphically summarized as an interconnected network of modules ( Figure 7 ) . Figure 7A depicts all 97 modules that were identified among the 139 partial sof alleles , whereas Figure 7B depicts only the 49 modules that occur in >10% of the sof alleles . Modules 1 , 6 , 13 , and End are present in all 139 ( 100% ) sof alleles ( Figure 4C ) , and , therefore , these four modules correspond to the largest nodes . The relative position of each node along the x-axis reflects the average distance of the module start position from the 5′ and 3′ ends of the anchor-bracketed region , whereby distance is based on the first occurrence of the module within each sof allele containing that module . In contrast to pbp2x ( Figure 3 ) , the sof network graph displays backward-oriented arrowheads ( Figure 7 ) . This finding is indicative of a genetic rearrangement that changes the relative order of two modules within a single sof allele . Alternatively , backward-facing arrows can arise if a module occurs multiple times within an allele . Since each module is depicted by a single node , a module that is positioned between multiple copies of another module , or is involved in a rearrangement , can yield a series of arrows that form a closed loop . This is readily observed in graphic depictions of the module networks for individual sof alleles ( Figure S5 ) . In the module network graph for the complete sof dataset ( Figure 7A ) , nodes with forward- and backward-facing arrows include Modules 6 , 8 , 13 , 18 , and 26 , each of which are also highly prevalent among sof alleles ( Table S2 ) . The high number of both forward- and backward-facing arrows impinging on Module 13 suggests that the corresponding sequence segments have undergone frequent recombination . This idea is further supported by the observation that Module 13 often flanks other modules that have undergone rearrangements in their relative order ( Figure 6 ) . Figure 7B shows a less complex network structure for sof , whereby rare modules and connections have been removed . The central portion of the network is relatively devoid of backward-facing arrows ( box ) . This observation suggests that the majority of rearrangements and/or duplications , resulting in shifts in relative module order , have occurred closer to the 5′ or 3′ ends of the sof region of interest . Furthermore , within the boxed central region are two dominant pathways of module-to-module connections . The module content of the upper versus lower pathway correlates with module pairs that tend to be mutually exclusive . For example , Modules 12 and 28 have a large negative value in their deviation from their expected co-occurrence ( Table S3; corrected p < 0 . 000006 ) , and they occupy similar relative positions along the x-axis ( Figure 7B ) . As observed in the central zone of restricted recombination within the pbp2x network graph ( Figures 3B and S3 ) , the upper and lower pathways of module connections in sof may represent two major lineages of sof having different selectable phenotypes . Comparison of the individual module network paths taken by single sof alleles highlights the allelic differences ( Figure S5A versus Figure S5B ) . For example , in the 3′ half of the graphs , the upper panel shows a path containing multiple closed loops that involve Module 26 , whereas the lower panel shows a continuous linear path through Module 26 to the end of the graph . Closed loops and backward-facing arrows are suggestive of an intragenic rearrangement or duplication . Sequence segments corresponding to Modules 6 and 13 , as defined by BLAST Miner , exhibit several structural features that may help to explain the biology of sof . Along with the anchor Modules 1 and End , Modules 6 and 13 are the only modules that occur in all of 139 sof alleles . In addition , Module 6 and 13 sequence segments occur multiple times per sof allele ( mean of 4 . 2 and 5 . 2 times , respectively ) . However , the 716 occurrences of Module 13 ( Table S2 ) are limited to 289 discrete blocks , many of which contain multiple copies of Module 13; Module 6 displays similar properties . The high occurrence rate for Modules 6 and 13 is probably a consequence of their composition , which is rich in short tandem direct-sequence repeats ( Figures 6 and 8 ) . In addition to tandem duplications , the sequence segments composing Modules 6 and 13 contain inverted repeats ( Figure 8 ) . The inverted repeats within Module 13 sequences give rise to complementary pairings both with itself , and in conjunction with Module 6 sequences . The REPuter software tool [8] , which computes repeats and palindromes within a single sequence , was used to locate and identify all perfect and near-perfect inverted repeats within the anchor-bracketed region of sof sequence AF139751 . Table 3 lists all 36 perfect or near-perfect inverted repeats >8 nt in length . Nearly all of the inverted repeats ( 34 of 36 ) involve Modules 6 and/or 13 sequences . Of the 18 inverted repeats that involve Module 6 sequences , 16 of the repeats are complementary to Module 13 sequences . The 33 inverted repeats of Module 13 sequence segments form palindromes either with Module 6 segments or with another Module 13 sequence . Another notable feature of both Module 6 and 13 sequence segments is that they display strong codon usage bias relative to the complete S . pyogenes genome [9] , or when compared with the complete collection of sequences of the 139 partial sof alleles ( Table S4 ) . For example , the AGU codon for Ser accounts for 79% of Ser-specific codons within the in-frame Module 13 sequences , but accounts for only 23% of the Ser-specific codons present in the complete S . pyogenes genome . Module 13 sequences also have a predicted amino acid sequence that is highly enriched in just a few residues . Seven amino acids account for 96% of the residues within Module 13 sequence segments ( Table S4 ) . Top-ranking among these are codons for Thr , Ser , Ala , and Gly , representing 36% , 23% , 17% , and 13% of the codons within Module 13 sequences , respectively . Importantly , the codons themselves form palindromic pairs . For example , AGU encoding Ser is often paired with ACU encoding Thr . Like Module 13 , the highly prevalent Module 6 also displays strong codon bias , and is enriched in codons for Thr , Ser , Ala , and Gly that form palindromic pairs . In summary , BLAST Miner identified Modules 6 and 13 sequence segments within all sof alleles and highlighted their unusual structural properties . In-depth examination of the sequences that define Modules 6 and 13 revealed a rich source of short tandem duplications and inverted repeats that form palindromes . The strong codon bias within these palindromic sequence segments is consistent with the idea that strong selection acts at the DNA level to maintain critical biological functions that are dependent on nucleic acid secondary structure . The finding for a high frequency of occurrence of Modules 6 and 13 sequences within partial sof alleles , in positions that often flank other modules that have undergone rearrangements or form nontandem direct repeats , suggests that the palindromes may play a mechanistic role in mediating aberrant genetic recombination .
Strong diversifying selection pressures in rapidly evolving microorganisms can yield alleles whose ancestral history is difficult to reconstruct . Although these rapidly evolving genes may represent a small fraction of the genetic content of the microbe , they are disproportionately important when seeking to understand microbial evolution , ecology , and pathogenesis . The inability to obtain an accurate multiple-sequence alignment further heightens the challenge to understand the underlying organization of the gene . The study of such biologically important loci is hampered by the lack of predictive tools that can be used to generate testable hypotheses that address their structure and function . BLAST Miner differs from many of the predictive tools used for data mining of nt sequence information . Mosaic structures arising from intragenic recombination are often demonstrated via alignment of just a few alleles [7 , 10 , 11] . BLAST 2 sequence ( bl2seq ) is a BLAST-based tool that can be used for detection of duplications and indels [12] , but its utility is limited to comparisons of sequences in pairs . There are numerous quantitative methods for detecting probable crossover points [6 , 13] , even when mosaic structures are not obvious . However , these methods use sequence alignments as a starting point . Networked evolution of numerous alleles can be effectively illustrated by splits graphs [14] , but split decomposition analysis is also predicated on an accurate multiple-sequence alignment . BLAST Miner does not rely on sequence alignments and instead uses BLAST to search for smaller segments of high-sequence homology among a large set of sequences . In this report , BLAST Miner was used to analyze genetic regions of ~1 . 5 kb and ~0 . 4 kb in length ( pbp2x and sof , respectively ) . However , it can be applied to a wide size range of sequence length and is limited only by the maximum allowable size of the table generated in the all-ways pairwise MegaBLAST output , which is 2 GB ( ~20 million records ) . In theory , BLAST Miner can be used for analysis of pathogenicity islands , phage or other viral genomes , and entire microbial chromosomes . The BLAST Miner software tool can also be used to assess variant loci in lower- and higher-order eukaryotes . Segmentally variable genes , having highly variable regions interspersed among well-conserved stretches [15] , may be highly amenable to this analysis . One of the primary outputs of the BLAST Miner program is the so-called module . Modules are defined strictly in terms of sequence similarity , and biological or evolutionary processes that yield or preserve regions of relatively high sequence identity will tend to enhance the detection of modules at the more stringent parameter settings . Modules start sites are determined by the point at which sequence similarity begins anywhere in the dataset . Another key output of BLAST Miner is a network graph , which simultaneously depicts the relationships between all modules in the entire dataset . The network graph can be used to rapidly identify patterns of potential biological significance that are otherwise difficult to detect via pairwise sequence comparisons . Regions containing conserved or variable sequences can be identified , and regions of the network containing closed loops and backward-facing arrows are good candidates for sites of gene rearrangements or recombination events . Network branch points arise when two or more allele paths enter or leave a module by different routes; this may be a consequence of recombination between alleles , insertions or deletions of module-containing sequence , or simply a high degree of sequence diversity in the region adjacent to a defined module . The latter is suggestive of either purifying selection in the module sequence , or diversifying selection in the adjoining region . However , when interpreting module networks , it is important to bear in mind that module nodes represent only the start sites of regions of sequence similarity; the length of downstream sequence segments of high similarity can vary widely for each allele . Also , highly divergent sequences that are not identified by a BLAST hit are not assigned to a module , and therefore they are not represented in the module graphs . A central limitation of BLAST Miner lies in the amount of diversity in the test sequences; it is unlikely to be of value for analysis of highly homologous genes . For example , BLAST Miner defined only a single module among 20 partial alleles of groEL from E . coli , wherein the maximal nt divergence is 3 . 5% ( unpublished data ) . It is anticipated that BLAST Miner will not be particularly useful for evaluation of these types of housekeeping genes [16 , 17] . However , its sensitivity and resolution can be appropriately tuned through the user-defined settings for word size , stringency threshold , module alignment length , and the module slip parameter . For example , with the sof dataset , increasing the stringency threshold from 90% to 98% nt identity reduces the number of modules from 269 to 204; the maximum number of iterations used to define a module is also reduced ( unpublished data ) . For less diverse genes , the number of modules can be increased by lowering the stringency threshold below 90% , reducing the module slip parameter below 4 nt , or increasing the module alignment length above 24 nt . The sof gene encodes SOF , a recognized virulence factor of S . pyogenes [7 , 18–23] . SOF is a sortase-anchored surface protein of ~1 , 000 amino acid residues . SOF contains a fibronectin-binding domain located proximal to the cell wall , a large central domain that interacts with apolipoproteins of mammalian serum and leads to its opacification , and a hypervariable amino-terminal region that appears to be a target of the SOF-serotyping scheme . The serological typing scheme is based on neutralization of the serum opacity reaction by specific antibody . The amino-terminal region of SOF was chosen as the focus of this report , as part of an initial step toward developing a sequence-based typing scheme for sof that parallels the widely used approach for emm-typing [24] . Although a biological function for the amino-terminal region has yet to be ascribed , this region appears to be under strong diversifying selection , perhaps mediated by a strong host immune response that provides protection against infection . Identification of Modules 6 and 13 through BLAST Miner may provide a starting point for understanding the genetic mechanisms underlying sof diversification . The sequence segments that constitute Modules 6 and 13 are striking in their richness for direct and inverted repeats , and imperfect ( quasi ) repeats . Both direct and inverted repeats are often sites of genetic rearrangements mediated by DNA mispairing [25] . These misalignments , or slippage errors , are independent of homologous recombination factors such as RecA , and are referred to as aberrant recombination events . Rearrangements mediated by direct repeats can result in genetic duplications or deletions , whereas inverted repeats can lead to inversions of sequence order . These types of genetic changes are evident among the sof alleles . Aberrant recombination has been described for other streptococcal species [26 , 27] . However , the specific mechanisms underlying aberrant recombination in S . pneumoniae and S . suis appear to be distinct from the likely processes governing rearrangements in sof , as inferred based on our new knowledge of sof gene structure . The strong codon usage bias observed within the Module 6 and 13 sequence segments , and the palindromic codons , are indicative of strong selection pressures that act to preserve the machinery that ultimately gives rise to the genetic changes upon which other selection pressures ( e . g . , host immunity ) can act . If the Module 6 and 13 sequence segments are hotspots for genetic recombination , they may also be hotspots for small indels to the degree that recombination is error-prone and subject to DNA strand slippage [25] . In this regard , it may be particularly relevant that a relatively high proportion of possible +1 frame shift mutations within Module 6 and 13 sequence segments are predicted to generate a stop codon ( i . e . , UAA , UAG , or UGA ) . Specifically , one-third of the codons within Module 13 sequence segments have AA , AG , or GA occupying the first two positions , whereas 66% have uracil in the third position ( unpublished data ) . In the event of a single base ( +1 ) frame shift , an average of 15 . 2% of the codons within the anchor-bracketed regions of sof are converted to stop codons . Overall , Module 13 sequences account for 22 . 9% of the anchor-bracketed regions of sof , yet they contain 32 . 9% of the stop codons that would occur in the event of a +1 frame shift . This event , in turn , could lead to phase variation in the SOF phenotype . To our knowledge , there is no documentation of phase variation in SOF expression; however , the BLAST Miner findings provide a rational basis for formulating hypotheses to test this biological property . The experimental approach can take the form of screening bacterial variants for a sof-positive genotype and a SOF-negative phenotype . In general terms , high-frequency phase variation of a microbial surface protein can be a key part of a survival strategy to escape the host immune response and/or to release the organism from its epithelial attachment site ( e . g . , promote transmission ) . A multiple-sequence alignment is an essential starting point for many of the tools that provide fundamental insights on gene structure , selection , recombination , and phylogeny . However , if recombination is sufficiently high , even tools specifically designed to estimate recombination can exceed their limits . BLAST Miner is a bioinformatics tool that can help provide additional new insights on genes that are intractable with the many tools that rely on an accurate multiple-sequence alignment . It also has the important predictive quality that leads to the generation of testable hypotheses based on sequence data . Analysis of sof by BLAST Miner provides evidence for a novel molecular mechanism for generating genetic diversity in S . pyogenes , a pathogen characterized by a very high number of genetically distinct clones [28 , 29] .
The partial nt sequence was determined for sof genes following PCR amplification of purified S . pyogenes chromosomal DNA with Primers F2 ( 5′-GTATAAACTTAGAAAGTTATCTGTAGG-3′ ) and R3 ( 5′-GGCCATAACATCGGCACCTTCGTCAATT-3′ ) , according to [7] . Newly identified sof alleles were deposited in Genbank and assigned accession numbers DQ450100 to DQ450145 . BLAST Miner ( version 1 . 0 ) is a Microsoft Windows 2000/XP–based stand-alone relational database program written in Delphi ( Borland Software , http://www . borland . com ) . It requires that the Borland Database Engine be installed . The BLAST Miner program , complete with a runtime distribution of the Borland Database Engine , installation instructions , and a user guide , is available for download at http://pantheon . yale . edu/~jw343/blastminer . html . A description of the BLAST Miner algorithm and its applications is provided below and outlined in Figure S2 . Figure S6 provides a screenshot of the main window . BLAST Miner is designed to use the BLAST hit results of an all-ways pairwise comparison of ~20–200+ nt sequences ( e . g . , alleles ) as input data . All sequences to be compared via BLAST are compiled in a FASTA file; a convenient feature of the BLAST Miner approach is that the nt sequences in the FASTA file need not be trimmed to equivalent lengths or positions at this stage of analysis . The FASTA file is used to generate a BLAST database , using the formatdb program , which is available for download as part of the stand-alone BLAST suite of applications ( http://www . ncbi . nlm . nih . gov/BLAST/download . shtml ) . The stand-alone version of MegaBLAST ( version 2 . 2 . 12 ) is used to query the BLAST database with the same FASTA file that was used to generate it , yielding an all-ways pairwise comparison of sequences [30] . The BLAST Miner program handles the interface , and the formatdb and MegaBLAST programs remain hidden . Parameters for the MegaBLAST program include specifying the database , specifying the input and output file names , disabling the complexity filtering , setting the word size ( to 16 bases in this study ) , setting the dropoff value to 10 , and specifying a single tab-delimited line per BLAST hit as the output file format . The MegaBLAST output file is automatically converted into a database table , whereby each BLAST hit is represented by a record that contains the following fields: the name of the query sequence ( QueryID ) , the name of the subject sequence ( SubjectID ) , the percent identity ( Identity ) , the length of the BLAST hit ( AlignmentLength ) , the number of mismatches ( Mismatches ) , the number of gap openings ( GapOpenings ) , the starting and ending position of the BLAST hit in the query sequence ( QStart and QEnd ) , and the starting and ending position of the BLAST hit in the subject sequence ( SStart and SEnd ) . As a first step in processing the MegaBLAST output data , a reciprocal of each BLAST hit record is generated by reversing the subject and query names , and the subject and query hit locations ( QHLs ) . The reciprocal records are added to the database of BLAST hits . The reciprocal step corrects an artifact in MegaBLAST , and also ensures that all BLAST hits are defined according to both the query and subject sequence names . A query or subject sequence that has 100% nt identity over its entire length to another sequence is considered to be a duplicate , and all records referring to the duplicate sequence are removed from the dataset . User-defined stringency parameters are also applied , in order to remove records that fall below the percent identity threshold . For the analyses of this report , the minimum percent nt identity threshold was set to 90% . An additional processing step is included as part of BLAST Miner to render it less sensitive to the size and composition of the sequences in the input dataset . The sequences constituting the BLAST hit records are used to scan the entire database of sequences for additional exact matches against the starting region of the original BLAST hit . The number of bases within the BLAST hit starting region that is used for scanning is user-definable ( in the “module length” option of the main window ) . The scanning step reduces the bias that can occur with an extended alignment between two sequences , which generates only one BLAST hit record , as opposed to several BLAST hit records having shorter alignment lengths but which together span the same region . The scanning step may lead to an increase in the number of BLAST hit records . Since only exact matches ( i . e . , 100% sequence identity ) are added , no reduction in stringency results from this process . The fundamental unit that is derived from BLAST Miner is referred to as the module . A module is a dataset consisting of segments of nt sequence that have a high percent identity to at least one other sequence segment assigned to that module . The database table of BLAST hit records is processed by the module correlation algorithm , an iterative process that searches for matches between appropriate parts of the BLAST hit record name assignment , whereby records that share common BLAST hit locations are grouped together . Each BLAST hit record is assigned a QHL and subject hit location ( SHL ) . To assign the QHL value , the BLAST hit records are first sorted by query sequence name , and next sorted by the nt starting position of the BLAST hit within the query sequence . The sorted BLAST hit records are sequentially numbered to yield the QHL assignment , whereby the number is incremented when either ( a ) the starting nt position exceeds the user-defined module slip threshold parameter or ( b ) a new query sequence is encountered . The module slip threshold is simply defined as the number of nt bases between starting positions of the sequential BLAST hit locations within a single sequence . In this report , the module slip threshold parameter is set to 4 nt . Upon completion of assigning the QHL values , the database is sorted by subject sequence name , and subsequently sorted by the starting position of the BLAST hit within the subject sequence . SHL assignments for each BLAST hit record are made according to the numbering process that is described for QHL assignments . An iterative process is used to assign related BLAST hit records to a module . The iterative process is based on associations between the QHL and SHL numerical values , which reflect the high level of percent nt identity between the query and subject sequence , for the sequence segment that lies immediately downstream of the starting position of the BLAST hit . Since an initial starting point for the iterative process must be chosen , for consistency the most prevalent SHL assignment is selected . All BLAST hit records sharing that SHL value are extracted from the database of BLAST hit records and assigned to the same module ( e . g . , Module 1 ) . In the next step , all remaining records within the database of BLAST hits that have a QHL value matching any QHL number present among the BLAST hit records belonging to Module 1 are also assigned to Module 1 . The process of adding new BLAST hit records to Module 1 is repeated , alternating between QHL and SHL assignments , until no new matching records are found . Thus , at each iteration , sequences are added to a module only if they match an existing BLAST hit record of that module , which in turn is dictated by the percent nt identity threshold parameter . The entire iterative process is repeated , starting with the most prevalent SHL value remaining among the BLAST hit records that were not extracted by their assignment to the first module; this next dataset of sequence segments constitutes the second module . This process continues until all BLAST hit records are assigned to a module . Although sequences in the initial FASTA input file need not be aligned or trimmed to a specified length , further sequence analysis requires that homologous regions be compared . Relatively conserved sequence segments that bracket the portion of the gene to be studied are designated as the anchor modules . Due to possible sequence heterogeneity within the bracketing segments , it may be necessary for an anchor module to be defined as a composite of more than one module . Once the anchor-bracketed region is defined , the original input sequences that display 100% identity over the bracketed region are considered duplicates and removed from the dataset . The number of occurrences of each module contained within the dataset of alleles or partial alleles , defined as unique anchor-bracketed sequences , can be calculated . If the starting position of a module spans more than twice the module slip threshold parameter , an additional tandem copy of the module is declared . The relative order of each module within each allele can also be ascertained . Differences in the relative order of modules ( i . e . , rearrangements ) can be identified by comparing the module content of all alleles . If a module occurs more than once within an allele , the module positioned closest to the 5′ end of the anchor-bracketed region is used for the purpose of detecting rearrangements . The frequency of co-occurrence , or linkage , of different module pairs can be calculated , and the observed co-occurrence frequency can be compared with the null hypothesis that module co-occurrence is random . The relationships between all modules of the complete anchor-bracketed dataset can be graphically displayed as a network of modules , defined by nodes that are connected by arrows . The diameter of each node is directly proportional to the percent of sequences in the dataset that contain the module . Arrows represent the connections between contiguous modules . Arrow thickness is proportional to the frequency with which each connection is observed in the dataset . The relative position of each node along the x-axis is the average position of the first occurrence of the module within the anchor-bracketed region . The y-axis position is arbitrary; it can be adjusted by the user to reposition nodes , in order to minimize their overlap and to enhance visual clarity . Arrow color and node color are also user-definable features . Different node colors can be assigned based on the number of iterations that define a given module . Module network diagrams can be saved in bitmap or enhanced metafile formats . A distance matrix can be constructed from pairwise comparisons of module content of all alleles . The distance score between alleles is calculated based on the presence or absence of shared modules , without regard for differences in the order or number of modules . Scores for modules are weighted , based on their relative frequency in the dataset , such that alleles that differ by a single rare module have smaller distance scores than alleles that differ by a high-frequency module . Weighting was implemented due to the observation that many rare modules arose as a consequence of a very small number of nt substitutions between nearly identical alleles . A neighbor-joining algorithm in the NEIGHBOR program [31] was employed to generate phenograms based on the distance matrices . The BioEdit sequence alignment editor version 7 . 0 . 2 [32] was used for manipulating sequence files and generating sequence identity matrices . The ClustalW algorithm was used for multiple-sequence alignments . Multiple-sequence alignments of the sof dataset were also made with the MUSCLE program version 3 . 6 [33] and the MAFFT program version 5 . 861 [34] . Aggregate codon usage statistics were calculated by concatenating sequences ( in frame ) and tabulating the collective codon composition , using the EditSeq version 5 . 52 ( DNASTAR , http://www . dnastar . com ) program .
The GenBank accession numbers for sof sequences discussed in this paper are listed in Table S1 . The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession numbers for pbp2x sequences discussed are AY950507 and AY950558 , and X16367 . Newly identified sof alleles were deposited in Genbank and assigned accession numbers DQ450100 to DQ450145 . | Microbial genes that accumulate large amounts of nucleotide sequence diversity through lateral exchanges with other microorganisms are often central to understanding key interactions between the microbe and an ever-changing host or environment . Proper sequence alignment of multiple gene alleles is an essential starting point for many of the tools that provide fundamental insights on gene structure and evolution , and allow scientists to develop hypotheses on biological processes . However , for some of the most interesting genes , a good quality alignment can be impossible to attain . We introduce a new software program , BLAST Miner , for analyzing genes that cannot be well-aligned . It relies on identifying small gene segments having high levels of sequence homology , irrespective of their relative positions within the different genes . Genes encoding a drug-resistance determinant and a target of host immunity are used as examples to demonstrate the application of BLAST Miner , and a potentially novel mechanism for generating genetic change is uncovered . This new bioinformatics tool provides an avenue for studying genes that are intractable by most other analytic approaches . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"evolutionary",
"biology",
"genetics",
"and",
"genomics/gene",
"function",
"computational",
"biology"
] | 2007 | Detecting Key Structural Features within Highly Recombined Genes |
African trypanosomes are extracellular protozoan parasites causing a chronic debilitating disease associated with a persistent inflammatory response . Maintaining the balance of the inflammatory response via downregulation of activation of M1-type myeloid cells was previously shown to be crucial to allow prolonged survival . Here we demonstrate that infection with African trypanosomes of IL-27 receptor-deficient ( IL-27R-/- ) mice results in severe liver immunopathology and dramatically reduced survival as compared to wild-type mice . This coincides with the development of an exacerbated Th1-mediated immune response with overactivation of CD4+ T cells and strongly enhanced production of inflammatory cytokines including IFN-γ . What is important is that IL-10 production was not impaired in infected IL-27R-/- mice . Depletion of CD4+ T cells in infected IL-27R-/- mice resulted in a dramatically reduced production of IFN-γ , preventing the early mortality of infected IL-27R-/- mice . This was accompanied by a significantly reduced inflammatory response and a major amelioration of liver pathology . These results could be mimicked by treating IL-27R-/- mice with a neutralizing anti-IFN-γ antibody . Thus , our data identify IL-27 signaling as a novel pathway to prevent early mortality via inhibiting hyperactivation of CD4+ Th1 cells and their excessive secretion of IFN-γ during infection with African trypanosomes . These data are the first to demonstrate the essential role of IL-27 signaling in regulating immune responses to extracellular protozoan infections .
African trypanosomiasis is a vector-borne parasitic disease of medical and veterinary importance . It is estimated that 170 , 000 people contract the disease every year , and that approximately 70 million people mainly in sub-Saharan Africa are at the risk of contracting the disease [1 , 2] . In addition , this disease severely limits the agricultural development by affecting domestic animals in the area [2] . The causative agents of this disease are various species of genus of Trypanosoma , which are extracellular protozoan parasites equipped with a flagellum that emerges from the flagellar pocket and provides the parasite with its motility [2] . Upon the bite of the mammalian host by a trypanosome-infected tsetse fly , the parasites enter the blood circulation via lymph vessels and can multiply in the bloodstream and interstitial fluids of the host [3 , 4] . The parasites have evolved very sophisticated evasion mechanisms to survive in the chronically infected host [3–5] , causing a serious disease that is often fatal without treatment [1 , 2] . Due to practical and ethical reasons , mouse models have become an alternative and proven to be a cornerstone for studying African trypanosomiasis of humans and domestic animals [6] . Most of studies have been performed with T . brucei and T . congolense parasites [3 , 6] . Based on mouse models , although the parasites circulate in the blood stream , the liver is the major place for clearance of the parasites [7–9] . Recent studies demonstrated that Kupffer cells efficiently engulf trypanosomes , which is mediated by both IgM and IgG antibodies specific to the parasites [10–12] . IFN-γ , mainly secreted by VSG-specific CD4+ T cells [13–15] following activation by dendritic cells [16 , 17] , has been shown to mediate protection during African trypanosomiasis [13 , 15 , 18–20] . Proinflammatory cytokines such as IL-12 , TNF-α , as well as iNOS produced by M1-type myeloid cells are also critical for host resistance to African trypanosomes [15 , 21–25] . However , excessive secretions of these inflammatory cytokines by hyperactivated myeloid cells and T cells lead to liver pathology and shorten the survival of infected mice [11 , 22 , 26–29] . In this respect , IL-10 has been found to be essential for maintenance of the immunological balance between protective and pathological immune responses during African trypanosomiasis [11 , 20 , 22 , 26 , 27] . Importantly , the role of IL-10 as an anti-inflammatory agent has been more recently confirmed in cattle , primate and human infections with African trypanosomes [30–32] . It remains unknown whether , in addition to IL-10 signaling , another pathway that maintains this immunological balance exists . IL-27 , a recently identified cytokine produced primarily by macrophages and dendritic cells , is a member of the IL-12 super-family [33] . The IL-27 receptor ( IL-27R ) complex consists of the specific IL-27Rα subunit ( WSX-1 ) and the IL-6R subunit ( gp130 ) , and is expressed on numerous subsets of leukocytes including CD4+ T cells , CD8+ T cells , NK cells , monocytes , Langerhans cells , and dendritic cells [34] . Earlier studies have demonstrated that IL-27 , as a proinflammatory cytokine , drives naïve T cells to differentiate into Th1 cells [35–37] . More recent studies have suggested that IL-27 also has the function to inhibit immunopathology via downregulation of active CD4+ T cells during infections , particularly with intracellular protozoan parasites [38–42] . However , the precise mechanism of CD4+ T cell-mediated immunopathogenesis in the absence of IL-27 signaling still remains incompletely understood . In addition , it is not clear so far whether IL-27 plays an important role in regulation of the immune responses during infections with extracellular protozoan parasites such as African trypanosomes . Based on previous data showing that a subset of highly activated pathological CD4+ T cells produces excessive IFN-γ , and leads to immunopathology and early mortality of mice infected with T . congolense [11 , 28 , 29] , we formulate a hypothesis that IL-27 signaling is , besides IL-10 signaling , another novel pathway that prevents the immunopathology and early mortality via down-regulation of the hyperactivity of CD4+ T cells and their excessive secretion of IFN-γ during experimental Africa trypanosomiasis . With this in mind , we examine in this study how IL-27 signaling regulates the immune responses in mice infected with African trypanosomes .
To evaluate the role of IL-27 signaling during African trypanosomiasis , we first determined whether infection led to increased expression of this cytokine or its receptor . Wild-type C57BL/6 mice were infected with T . congolense , a species of African trypanosomes which are unable to leave the circulation and only live in blood vessels , causing fatal disease in cattle [4] . The mice were euthanized at day 0 , 7 , and 10 after infection , as parasitemia usually peaked on day 6–7 after infection [15 , 29] . As the liver is the major organ for clearance of the parasites [7–9 , 11] , the liver was collected for measurement of mRNA levels of IL-27 and its receptor using real-time quantitative RT-PCR . mRNA levels of both subunits of IL-27 ( IL-27p28 and EBI3 ) were upregulated in the liver of mice at day 7 and 10 after infection , compared to uninfected mice ( Fig 1A ) . In contrast , mRNA levels of IL-27 receptor ( WSX-1 ) were not affected by the infection ( Fig 1A ) . Next , we infected IL-27R-/- ( WSX-1-/- ) and wild-type mice with T . congolense to assess whether IL-27 signaling affected the disease progression . Similar to infected wild-type mice , infected IL-27R-/- mice could control the first wave of parasitemia ( Fig 1B ) . However , IL-27R-/- mice succumbed to the infection on day 12 to 20 after infection with a mean survival time of 14 . 5 days ( Fig 1C ) . In contrast , infected wild-type mice survived until day 67 to 138 days after infection with a mean survival time of 123 days ( Fig 1C ) . Compared to infected wild-type mice , the infected IL-27R-/- mice survived significantly shorter ( p<0 . 01 ) . These data demonstrated that IL-27 signaling is required for survival of mice infected with T . congolense . The above results demonstrated that absence of IL-27 signaling led to earlier mortality of mice infected with African trypanosomes . As uncontrolled inflammation causes early mortality of mice infected with African trypanosomes [3 , 4] , we next examined the plasma levels of inflammatory cytokines and their secretions by cultured spleen cells . As shown in Fig 2A , significantly higher amounts of IFN-γ , IL-12p40 , and TNF-α were detected in the plasma of IL-27R-/- mice infected with T . congolense , compared to infected wild-type mice , on day 7 and 10 after infection ( p<0 . 01 ) . Although the plasma level of IFN-γ in IL-27R-/- mice decreased on day 10 after infection probably due to clearance of the first wave of parasitemia , it was still significantly higher than that of the infected wild-type mice ( p<0 . 01 , Fig 2A ) . To evaluate the secretions of cytokines by spleen cells , spleen cells were collected from IL-27R-/- and wild-type mice on day 7 and 10 after infection with T . congolense , and cultured in vitro for 48 h . The production of IFN-γ , IL-12p40 , and TNF-α by spleen cells were significantly elevated in infected IL-27R-/- mice , compared to infected wild-type mice ( p<0 . 01 or <0 . 05 , Fig 2B ) . As recent studies have shown that IL-27 mainly regulates CD4+ T cell activation during infection with intracellular pathogens [38–42] , we further evaluated IFN-γ-producing CD4+ T cells in the spleen cultures using flow cytometry . A limited and similar percentage and absolute number of CD4+ T cells from uninfected wild-type and IL-27R-/- mice produced IFN-γ after 12 h stimulation with Cell Stimulation Cocktail ( containing PMA , ionomycin , and protein transport inhibitors ) . However , by 7 and 10 days post infection both the percentage and the absolute number of IFN-producing CD4+ T cells were significantly enhanced in IL-27R-/- mice when compared to wild-type cohorts ( Fig 2C ) . We and others have previously shown that excessive systemic inflammatory responses of mice infected with African trypanosomes are associated with severe liver damage [11 , 22 , 43 , 44] . In addition , the liver is the primary organ of trypanosome clearance [7 , 9 , 11] . Therefore , we next evaluated effects of IL-27 signaling on liver pathology during the course of infection with the parasites . IL-27R-/- mice , but not wild-type mice , showed extensive pale geographic areas highly suggestive of necrosis on day 10 after infection with T . congolense ( Fig 3A ) . Microscopic examination of the liver of infected IL-27R-/- mice revealed many large areas with loss of hepatocyte cellular architecture and an infiltration of inflammatory cells ( Fig 3B ) . By contrast , these pathological changes were not observed in the liver of infected wild-type mice ( Fig 3B ) . To further characterize the liver pathology , we measured the serum activities of alanine aminotransferase ( ALT ) of mice during T . congolense infection . As shown in Fig 3C , IL-27R-/- mice had significantly higher serum activities of ALT than wild-type mice on both day 7 and day 10 after infection ( p<0 . 05 ) , indicating death of hepatocytes and release of cytosolic enzymes . These results demonstrated that IL-27 signaling played a major role in prevention of the liver pathology that was associated with enhanced systemic inflammatory responses . It has been shown that IL-10 is crucial for survival of mice infected with African trypanosomes through limiting inflammation [11 , 20] . In particular , failure to control inflammatory responses in mice infected with African trypanosomes in the absence of IL-10 signaling is associated with severe liver pathology [11 , 22 , 27] . In this regard , IL-27 has been shown to drive CD4+ T cells to produce IL-10 for downregulation of inflammation [45–47] . The similarity of the cytokine profile and liver pathology of infected mice in the absence of IL-27 signaling and IL-10 signaling [11 , 20] prompted us to examine whether IL-27 signaling prevented early mortality of mice infected with African trypanosomes via IL-10 . We first compared the disease progression in the absence of IL-27 signaling with that in the absence of IL-10 signaling . T . congolense-infected IL-27R-/- mice and wild-type mice showed similar parasitemia and a significantly reduced survival after administration of anti-IL-10 receptor ( IL-10R ) mAb ( p<0 . 01 , Fig 4A ) . Strikingly , infected wild-type mice treated with anti-IL-10R mAb survived significantly shorter than infected IL-27R-/- mice ( p<0 . 01 , Fig 4A ) , suggesting that IL-27 and IL-10 may independently regulate inflammatory responses during African trypanosomiasis . Next we compared the IL-10 levels in plasma , and supernatant fluids of cultured spleen cells or liver leukocytes between IL-27R-/- and wild-type mice infected with T . congolense . There was no significant difference in IL-10 production in plasma and supernatant fluids of the cultures between IL-27R-/- and wild-type mice on day 7 after infection ( Fig 4B ) . Surprisingly , IL-27R-/- mice even showed significantly higher amounts of IL-10 in both plasma ( up to 14 folds ) and supernatant fluids of cultured spleen cells or liver leukocytes on day 10 after infection ( p<0 . 01 or <0 . 05 , Fig 4B ) , demonstrating that secretion of IL-10 was strengthened , rather than impaired in IL-27R-/- mice infected with African trypanosomes , probably due to deficiency of the immune regulation mediated by IL-27 signaling in those infected IL-27R-/- mice . Taken together , these data suggested that early mortality of IL-27R-/- mice infected with African trypanosomes was not due to impaired IL-10 production . Because early mortality of IL-27R-/- mice infected with African trypanosomes was associated with severe liver pathology without impaired secretion of IL-10 as shown above and because IL-27 has been shown to mainly regulate T cell , particularly CD4+ T cell activation during infection with intracellular pathogens [38–42] , we next characterized CD4+ T cell responses in the liver of IL-27R-/- mice during infection with T . congolense . We found that the frequency and the absolute number of activated hepatic CD4+ T cells ( CD44hiCD62Llow ) were significantly higher in IL-27R-/- mice infected with T . congolense , compared to infected wild-type mice ( p<0 . 01 , Fig 5A ) . The production of IFN-γ , IL-12p40 , and TNF-α by cultured liver leukocytes from infected IL-27R-/- mice was significantly higher than production of these cytokines by liver leukocytes from infected wild-type mice ( p<0 . 001 , <0 . 01 or <0 . 05 , Fig 5B ) . In particular , the production of IFN-γ was enhanced by 4–8 folds in the liver leukocyte cultures of infected IL-27R-/- mice ( Fig 5B ) . Thus , we further evaluated the activation of liver CD4+ T cells by examining their secretions of IFN-γ using single cell analysis . A small and similar percentage and absolute number of CD4+ T cells from uninfected wild-type and IL-27R-/- mice secreted IFN-γ after 12 h stimulation with Cell Stimulation Cocktail ( containing PMA , ionomycin , and protein transport inhibitors ) . In contrast , by day 7 and 10 post infection significantly higher percentage and absolute number of IFN-γ-producing CD4+ T cells were detected in IL-27R-/- mice as compared to wild-type cohorts ( Fig 5C ) . Collectively , these data suggested that the early mortality of IL-27R-/- mice infected with African trypanosomes was associated with exacerbated Th1-mediated immune responses with overactivation of CD4+ T cells . As shown above , CD4+ T cells were excessively activated in the liver of IL-27R-/- mice infected with African trypanosomes , raising the possibility that the early mortality of infected IL-27R-/- mice was a consequence of a CD4+ T cell-dependent immune-mediated pathology . To test this , IL-27R-/- mice infected with T . congolense were treated with depleting anti-mouse CD4 mAb , anti-mouse CD8 mAb , or rat IgG as control; and the course of infection , immune responses , and severity of liver damage were assessed . As shown in S1 Fig , administration of the antibodies efficiently depleted CD4+ T cells or CD8+ T cells in the spleen and liver of the infected mice . Infected mice from all three groups could effectively control the first wave of parasitemia , although depletion of CD4+ T cells resulted in a significantly higher parasitemia at some time points of infection ( p<0 . 01 or <0 . 05 , Fig 6A ) . Strikingly , infected IL-27R-/- mice treated with anti-CD4 mAb had two fold increase of survival compared to infected IL-27R-/- mice treated with rat IgG ( p<0 . 01 , Fig 6A ) . In contrast , depletion of CD8+ T cells did not affect the survival of infected IL-27R-/- mice ( Fig 6A ) . These results demonstrated that IL-27 signaling had a crucial role in dampening CD4+ T cell activation in experimental T . congolense infection in mice , allowing for prolonged survival . We next evaluated the effect of CD4+ T cells on weight loss and liver pathology of IL-27R-/- mice infected with T . congolense . Infected IL-27R-/- mice treated with anti-CD4 mAb had significantly less weight loss at the later stage of infection , compared to infected IL-27R-/- mice treated with rat IgG or anti-CD8 mAb ( p<0 . 01; S2A Fig ) . Importantly , infected IL-27R-/- mice treated with rat IgG or anti-CD8 mAb exhibited many large areas with loss of hepatocyte cellular architecture in the liver , whereas these pathological changes were hardly seen in the liver of infected IL-27R-/- mice treated with anti-CD4 mAb ( S2B Fig ) . In addition , depletion of CD4+ , but not CD8+ , T cells significantly reduced the serum activities of ALT in IL-27R-/- mice infected with T . congolense ( p<0 . 05 , Fig 6B ) . These data suggested that CD4+ T cells played a central role in the development of liver pathology in experimental T . congolense infection , and that IL-27 was crucial for dampening this CD4+ T cell-mediated pathology . We further characterized the contributions of CD4+ T cells to secretion of cytokines in IL-27R-/- mice infected with T . congolense . Depletion of CD4+ , but not CD8+ , T cells significantly reduced plasma levels of IFN-γ and TNF-α in infected IL-27R-/- mice ( p<0 . 001 or <0 . 05 ) , although the reduction of IL-12p40 did not reach statistical significance ( Fig 6C ) . In addition , depletion of CD4+ , but not CD8+ , T cells also resulted in significantly less secretion of IFN-γ by spleen cells from infected IL-27R-/- mice ( p<0 . 05 , S2C Fig ) . Interestingly , depletion of CD4+ T cells almost abrogated the production of IL-10 by spleen cells in infected IL-27R-/- mice ( p<0 . 01 , S2C Fig ) , suggesting that IL-10 was predominantly produced by CD4+ T cells . Importantly , the observation that the enhanced survival of infected IL-27R-/- mice treated with anti-CD4 mAb was correlated with very little secretion of IL-10 further suggested that IL-27 signaling inhibited hyperactivation of Th1 cells in an IL-10 independent manner as shown above in Fig 4 . Having demonstrating that IL-27 is crucial for dampening trypanosomiasis-associated CD4+ T cell activation , needed for prolonged survival , we next addressed the mechanism of CD4+ T cell-mediated mortality of infected IL-27R-/- mice . Because the production of IFN-γ , and the frequency and the absolute number of IFN-γ-producing cells were enhanced in infected IL-27R-/- mice compared to infected wild-type mice ( Fig 2 and Fig 5 ) , and also because depletion of CD4+ T cells dramatically reduced the IFN-γ production ( Fig 6; S2 Fig ) , we examined whether the early mortality of infected IL-27R-/- mice was directly attributed to the overproduction of IFN-γ . IL-27R-/- mice infected with T . congolense were treated with neutralizing anti-IFN-γ mAb or rat IgG as a control . Although administration of anti-IFN-γ mAb led to doubled parasitemia in infected IL-27R-/- mice at the peak on day 7 after infection ( P<0 . 05 ) , the infected IL-27R-/- mice treated with anti-IFN-γ mAb efficiently controlled the first wave of parasitemia as infected control mice did ( Fig 7A ) . Importantly , administration of anti-IFN-γ mAb significantly enhanced the survival of infected IL-27R-/- mice ( p<0 . 01; Fig 7A ) , demonstrating that high levels of IFN-γ accelerated the mortality of IL-27R-/- mice infected with African trypanosomes . We next assessed the effects of IFN-γ neutralization on weight loss and liver pathology of IL-27R-/- mice infected with T . congolense . Infected IL-27R-/- mice treated with anti IFN-γ mAb had significantly less weight loss than infected IL-27R-/- mice treated with rat-IgG on the late stage of infection ( p<0 . 01 , S3A Fig ) . Importantly , infected IL-27R-/- mice treated with anti-IFN-γ did not exhibit areas with loss of hepatocyte cellular architecture in the liver whereas these pathological changes were observed in the liver of infected IL-27R-/- mice treated with rat IgG ( S3B Fig ) . Moreover , neutralization of IFN-γ significantly reduced the serum activities of ALT in infected IL-27R-/- mice ( p<0 . 01 , Fig 7B ) . These data suggested that IFN-γ played a critical role in the development of liver pathology in IL-27R-/- mice infected with African trypanosomes . We finally examined cytokine responses of infected IL-27R-/- mice treated with anti-IFN-γ mAb . IFN-γ was almost undetectable in the plasma of IL-27R-/- mice treated with anti-IFN-γ , suggesting the neutralization was successful ( p<0 . 01 , Fig 7C ) . Plasma levels of IL-12p40 and TNF-α were dramatically reduced in infected IL-27R-/- mice treated with anti-IFN-γ mAb , compared to infected IL-27R-/- mice treated with rat IgG ( p<0 . 01 , Fig 7C ) . Neutralization of IFN-γ also significantly reduced the production of IL-12p40 and TNF-α by cultured spleen cells ( p<0 . 01 , or <0 . 05 , S3C Fig ) . Thus , the results indicated that IFN-γ was critically involved in the enhanced inflammatory responses in IL-27R-/- mice infected with African trypanosomes . We finally characterized the role of IL-27 signaling in regulation of immune responses during T . brucei infection . In contrast to T . congolense , T . brucei species have the ability to penetrate the walls of capillaries , invade interstitial tissues , including the brain tissues , thus serving as a model of human African trypanosomiasis [48 , 49] . T . brucei infection also upregulated the mRNA expressions of IL-27p28 and EBI3 , but not IL-27R-/- in the liver of mice ( Fig 8A ) . IL-27R-/- mice infected with T . brucei efficiently controlled the first wave of parasitemia as infected wild-type did , but survived significantly shorter than infected wild-type mice ( 15 days vs . 32 days , p<0 . 01 , Fig 8B ) , demonstrating an essential role of IL-27 signaling in prevention of the early mortality of mice infected with T . brucei . IL-27R-/- mice infected with T . brucei also showed enhanced IFN-γ production in plasma and supernatant fluids of spleen cultures , as well as enhanced serum activities of ALT , compared to infected wild-type mice ( p<0 . 01 or <0 . 05 , Fig 8C ) . Importantly , depletion of CD4+ , but not CD8+ , T cells enhanced the survival of IL-27R-/- mice infected with T . brucei by 3 folds ( p<0 . 01 , Fig 8D ) . Thus , IL-27 signaling is also required for survival of mice via preventing excessive Th1 immune responses during T . brucei infection .
Successful clearance of African trypanosomes in the bloodstream requires induction of inflammatory immune responses; however , failure to control this inflammation leads to immune-mediated pathology [4 , 50] . IL-10 signaling has been previously suggested to be involved in maintaining this immunological balance in African trypanosomiasis [11 , 20] . In the current study , we have identified IL-27 signaling as a novel pathway to maintain this immunological balance in African trypanosomiasis . Our data are the first to demonstrate the essential role of IL-27 signaling in regulating immune responses to extracellular protozoan infections . More importantly , we provided direct evidence , that infection-associated IL-27 signaling served to extend the survival of the infected host by dampening CD4+ T cell activation and their secretion of IFN-γ . Indeed , the early mortality of infected mice lacking IL-27 signaling ( IL-27R-/-mice ) was correlated with exaggerated inflammatory responses and liver immunopathology . The disease similarity of infected mice lacking IL-27 and IL-10 signaling raised the possibility that regulatory function of IL-27 is mediated via the induction of IL-10 secretion , as IL-27 has the capability of promoting CD4+ T cells to secret IL-10 [45–47] . However , the fact that blocking IL-10R further shortened the survival of infected IL-27R-/- mice and the fact that infected mice lacking IL-10 signaling and infected mice lacking IL-27 signaling had distinct survival suggested that IL-27 functions through a mechanism independent of IL-10 . In addition , compared to infected wild-type mice , infected IL-27R-/- mice produced similar or even higher amounts of IL-10 , depending on the time points examined . Furthermore , the enhanced survival of infected IL-27R-/- mice following depletion of CD4+ T cells was correlated with dramatically reduced secretion of IL-10 . These data suggested that a defect of IL-10 signaling is unlikely to contribute to the early mortality of IL-27R-/- mice . Thus , we suggest that IL-27 suppresses the liver pathology and prevents the early mortality of mice infected with African trypanosomes through IL-10-independent mechanisms , possibly by direct modulation of T cell function . It has been previously demonstrated that IL-10 inhibits accumulation and activation of M1-type myeloid cells , in particular , TIP-DCs ( CD11b+Ly6C+CD11c+TNF and iNOS producing DCs ) in the liver during infection with African trypanosomes [22 , 26 , 27] . Accordingly , African trypanosomes-infected CCR2 deficient mice and MIF ( macrophage migrating inhibitory factor ) deficient mice exhibited significantly reduced accumulation of TIP-DCs , which was correlated with remarked diminished liver pathology , and significantly prolonged survival [26 , 44] . Thus , IL-10 signaling suppresses liver pathology , mainly through downregulation of M1-type myeloid cells [3 , 50] . In contrast , IL-27R-/- mice infected with African trypanosomes displayed more activation of T cells , in particular , CD4+ T cells . Moreover , depletion of CD4+ T cells prevented liver pathology and early mortality of infected IL-27R-/- mice . Obviously , IL-27 signaling functions through limiting activation of CD4+ T cells in African trypanosomiasis . Thus , although both IL-10 signaling and IL-27 signaling are crucial for limiting the inflammatory complications associated to African trypanosome in particular in preventing liver pathology , the two signal pathways involve distinct mechanisms . Dampening accumulation of highly activated CD4+ T cells by IL-27 signaling has also been recently observed in infection with other microorganisms , particularly intracellular protozoan and bacterial pathogens [38 , 40–42 , 51] . Our data demonstrate that the same mechanism exists during infections with extracellular protozoan parasites such as African trypanosomes . However , the precise mechanism of CD4+ T cell-mediated early mortality in previous models was not fully elucidated [38 , 42] . One of the most important properties of CD4+ T cells is that they secret a large amount of IFN-γ upon activation . IFN-γ is required to eliminate intracellular parasites , but also has potential to induce immunopathology [52 , 53] . Indeed , early mortality of IL-27R-/- mice infected with Toxoplasma gondii , or Plasmodium berghei is associated with significantly enhanced production of IFN-γ [38 , 42] , suggesting that IFN-γ might be a critical molecule for CD4+ T cell-mediated mortality in the absence of IL-27 signaling . Surprisingly , neutralization of IFN-γ did not prolong the survival , and had no effect on the liver pathology of IL-27R-/- mice infected with T . gondii or P . berghei at all [38 , 54] . Thus , although CD4+ T cell-mediated mortality coincides with significantly elevated secretion of IFN-γ , it still remains inconclusive whether IFN-γ is the direct mediator of CD4+ T cell-dependent mortality in these infections . In contrast , neutralization of IFN-γ significantly enhanced the survival IL-27R-/- mice infected with African trypanosomes accompanied by a major amelioration of liver pathology , providing direct evidence that IFN-γ directly mediated the mortality of infected IL-27R-/- mice . In addition , enhanced survival of infected IL-27R-/- mice depleted of CD4+ T cells was correlated with a dramatically reduced production of IFN-γ . Obviously , either removing of CD4+ T cells or neutralization of IFN-γ got rid of the lethal effect of IFN-γ , leading to the prolonged survival of infected IL-27R-/- mice . Thus , another important finding of this study is that , in the absence of IL-27 signaling , CD4+ T cells mediated mortality directly through their secretion of IFN-γ , at least , during infection with extracellular protozoan parasites African trypanosomes . It is important to point out that our results in no way exclude the protective role of CD4+ T cells and IFN-γ during infection with the parasites . Indeed , early studies have shown that there was a correlation between high IFN-γ levels in serum , low parasitemia , and host resistance during infection with African trypanosomes [18] . Subsequent studies demonstrated that VSG-specific CD4+ T cells mediated protection via secretion of IFN-γ [13 , 55]; and splenic DCs were the primary cells responsible for activating naïve VSG-specific CD4+ T cell responses [16 , 17] . The protective role of CD4+ T cells and IFN-γ in African trypanosomiasis has been recently confirmed by independent groups [14 , 15 , 19] . In support of previous findings , we showed that either depletion of CD4+ T cells or neutralization of IFN-γ resulted in a significantly elevated peak parasitemia level in IL-27R-/- mice infected with T . congolense , confirming the protective role of CD4+ T cells and IFN-γ during the infection . It is likely that IFN-γ promotes M1-type myeloid cells to produce IL-12 , TNF-α and iNOS , which has been shown to be critically involved in lysis or damage of African trypanosomes [15 , 21 , 23 , 25 , 56] . On the other hand , excessive production of IL-12 , TNF-α and iNOS driven by IFN-γ could also mediate immunopathology of mice infected with African trypanosomes [22 , 24 , 26 , 27 , 57] . Further , IL-12 and TNF-α could stimulate T cells to produce more IFN-γ [4 , 21] . Thus , IL-10 is required to down-regulate the production of IL-12 , TNF-α and iNOS possibly by direct modulation of M1-type myeloid cells [11 , 22 , 26 , 27] . In the present study , we identified IL-27 signaling as a novel pathway to down-regulate the secretion of IFN-γ by direct modulation of CD4+ T cells . Obviously , in the absence of IL-27 signaling , excessive secretions of IFN-γ by CD4+ T cells also mediate liver pathology and mortality , although IL-10 signaling still fully functions and the infected mice produce even more IL-10 , in African trypanosomiasis . Thus , both IL-10 signaling and IL-27 signaling are required for survival of mice infected with the parasites via preventing aberrant inflammatory responses , although they function in a distinct manner in African trypanosomiasis . In conclusion , we have described an essential role for IL-27 signaling in preventing early mortality of mice infected with African trypanosomes through dampening IFN-γ secretion by CD4+ T cells , thus identifying , in addition to previously described IL-10 signaling , a novel pathway for maintenance of immunological balance during infection with extracellular protozoan parasites African trypanosomes . These data contribute significantly to our understanding of both immunopathogenesis of African trypanosomiasis and mechanisms underlying IL-27 immunoregulation during infection with extracellular protozoan and bacterial pathogens .
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The animal protocols involving mice were approved by the University of Maryland Institutional Animal Care and Use Committee ( IACUC ) under protocol R-12-60 . Eight- to teen-week-old C57BL/6NCrJ ( C57BL/6 ) mice and five- to six-week-old outbred Swiss white mice ( CD1 ) were purchased from the National Cancer Institute ( Frederick , MD ) . B6N . 129P2-Il27ratm1Mak ( IL-27R-/- , or WSX-1-/- ) mice were purchased from the Jackson Laboratory and bred in-house . All animal experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee and Institutional Bio-safety Committee of the University of Maryland , College Park . T . congolense , Trans Mara strain , variant antigenic type ( VAT ) TC13 was used in this study . The origin of this parasite strain has been previously described [58] . T . brucei AnTat1 . 1E was obtained from the Institute of Tropical Medicine ( Antwerp , Belgium ) . Frozen stabilates of parasites were used for infecting CD1 mice immunosuppressed with cyclophosphamide , and passages were made every third day as described previously [58] . The parasites were purified from the blood of infected CD1 mice by DEAE-cellulose chromatography [59] and used for infecting mice . Purified rat anti-mouse IL-10 receptor ( IL-10R ) mAb ( Clone 1B1 . 3a ) , purified rat anti-mouse CD4 mAb ( Clone GK1 . 5 ) , purified rat anti-mouse CD8 ( Clone 53–6 . 72 ) , and purified rat anti-mouse IFN-γ mAb ( Clone XMG1 . 2 ) were purchased from BioXCell ( West Lebanon , NH ) . Purified anti-mouse CD16/CD32 ( FcγIII/IIR , Clone 2 . 4G2 ) were purchased from BD Biosciences . APC-Cy7 anti-mouse CD3e ( 145-2C11 ) , PE-anti-mouse IFN-γ ( XMG1 . 2 ) , PE-Cy7-anti-mouse CD4 ( GK1 . 5 ) , PE-Cy7-anti-mouse CD4 ( RM 4–4 ) , FITC-anti-mouse CD8 ( 53–6 . 72 ) , FITC-anti-mouse CD8 ( YTS156 . 7 . 7 ) , APC-anti-mouse CD44 ( IM7 ) , PE-anti-mouse CD62L ( MEL-14 ) , and matching controls were purchased from eBioscience or Biolegend . Mice were infected i . p . with 103 T . congolense TC13 [11] or 5×103 T . brucei AnTat1 . 1E [44] . Some groups of infected mice were injected i . p . with rat anti-mouse IL-10R mAb ( 1B1 . 3a; 0 . 4 mg on day 0 , 2 , 4 , and 6 after infection , respectively ) , anti-mouse CD4 mAb ( GK1 . 5; 0 . 5 mg on day 0 , 2 , 4 , and 6 after infection , respectively ) , anti-mouse CD8 mAb ( 53–6 . 72; 0 . 5 mg on day 0 , 2 , 4 , and 6 after infection , respectively ) , anti-mouse IFN-γ mAb ( XMG1 . 2; 0 . 4 mg on day 0 , 2 , 4 , 6 , 8 , 10 , 12 , and 14 after infection , respectively ) , or rat IgG ( as a control ) . Parasitemia was counted at ×40 magnification by phase-contrast microscopy . The survival time was defined as the number of days after infection that the infected mice remained alive . For analysis of mRNA expression , total RNA was extracted from the homogenates of the liver of uninfected wild-type C57BL/6 mice or mice infected with T . congolense or T . brucei , following the manufacturer’s recommendation ( Life Technologies ) . IL-27p28 , EBI3 , and WSX-1 mRNA levels were quantified by real-time quantitative RT-PCR . The cDNA expression for each sample was standardized using the house keeping gene β-actin . Cycling conditions were as follows: initialization 2 min at 50°C and 10 min at 95°C , followed by 40 cycles of 15 s at 95°C and 1 min at 60°C . Primer pair used were: IL-27p28: 5’-CTGGTACAAGCTGGTTCCTG-3’ , 5’-CTCCAGGGAGTGAAGGAGCT-3; EBI3: 5’-CAGAGTGCAATGCCATGCTTCTC-3’ , 5’-CTGTGAGGTCCTGAGCTGAC-3’; WSX-1: 5’-CAAGAAGAGGTCCCGTGCTG-3’ , 5’-TTGAGCCCAGTCCACCACAT-3’ . Splenocytes were collected from mice . Cells were cultured at a concentration of 5 × 106 cells/ml ( 200 μl/well ) in 96-well tissue culture plates in a humidified incubator containing 5% CO2 . The culture supernatant fluids were collected after 48 h and centrifuged at 1 , 500g for 10 min , and the supernatant fluids were stored for cytokine assays at -20°C until used . Liver leukocytes were isolated as described previously [60] . Briefly , the liver was perfused with PBS until it became pale . Thereafter , the gallbladder was removed and the liver excised carefully from the abdomen . The liver was minced into small pieces with surgical scissors and forced gently through a 70 um cell strainer using a sterile syringe plunger . The preparation obtained was suspended in 50 ml RPMI-1640 medium containing 10% FCS . The cell suspension was centrifuged at 30g with the off-brake setting for 10 min at 4°C . The obtained supernatant was centrifuged at 300g with the high-brake setting for 10 min at 4°C . The pellet was resuspended in 10 ml 37 . 5% Percoll in HBSS containing 100 U/ml heparin and then centrifuged at 850g with the off-brake setting for 30 min at 23°C . This new pellet was resuspended in 2 ml ACK buffer ( erythrocyte lysing buffer ) , and incubated at room temperature for 5 min , then supplemented with 8 ml RPMI-1640 medium containing 10% FCS , followed by centrifugation at 300g with the high-brake setting for 10 min at 8°C . Cells were collected and cultured at a concentration of 5 × 106 cells/ml ( 200 μl/well ) in 96-well tissue culture plates in a humidified incubator containing 5% CO2 . The culture supernatant fluids were collected after 48 h and centrifuged at 1 , 500g for 10 min , and the supernatant fluids were stored for cytokine assays at -20°C until used . Recombinant murine cytokines and Abs to these cytokines for use in ELISA were purchased from BD Biosciences or R&D Systems . The levels of cytokines in culture supernatant fluids or plasma were determined by routine sandwich ELISA using Immuno-4 plates ( Dynax Technologies ) , according to the manufacturer’s protocols . To assess the activation of T cells , intrahepatic leukocytes were isolated as described above . The cells were incubated ( 15 min , 4°C ) with purified anti-mouse CD16/CD32 ( [FcγIII/II Receptor] , clone: 2 . 4G2 ) to block nonspecific binding of Abs to FcRs , washed with staining buffer ( eBioscience ) , resuspended in staining buffer , and stained with mAbs specific for various cell surface markers , or the relevant isotype-matched control Abs . For intracellular IFN-γ staining , spleen cells or intrahepatic leukocytes were diluted to 5 × 106 cells/ml and cultured ( 200 μl/well ) in a 96-well plate in the presence of 1x Cell Stimulation Cocktail ( containing PMA , ionomycin , and protein transport inhibitors , eBioscience ) for 12 h . The cells were then harvested and washed twice in staining buffer . The cells were incubated ( 15 min , 4°C ) with purified anti-mouse CD16/CD32 , washed with staining buffer , followed by staining with mAbs specific for cell surface markers . The cells were fixed and permeabilized using Intracellular Fixation & Permeabilization Buffer Set ( eBiosciences ) . Intracellular staining was then performed using mAbs specific for IFN-γ . Samples were resuspended in staining buffer , tested by FACSAria II , and analyzed using FlowJo software . Liver alanine transaminase ( ALT ) activities were determined using EnzyChrom Alanine Transaminase Assay Kit ( BioAssay Systems ) according to the manufacturer’s instructions . For histopathological examination , the liver was taken from mice on day 10 after infection and fixed with 10% formalin in PBS . Sections were stained with Hematoxylin and Eosin . Data are represented as the mean ± SEM . Significance of differences was determined by ANOVA or a log-rank test for curve comparison using the GraphPad Prism 5 . 0 software . Values of p≤0 . 05 are considered statistically significant . | Infection with extracellular protozoan parasites , African trypanosomes , is characterized by a persistent inflammatory immune response . It has been recently shown that maintaining the balance of the inflammatory responses via dampening M1-type myeloid cell activation is critical to guarantee control of the parasites and survival of the host . In this study , we demonstrated that IL-27 receptor-deficient ( IL-27R-/- ) mice infected with African trypanosomes developed an excessive inflammatory response and severe liver immunopathology , resulting in dramatically reduced survival , as compared to infected wild-type mice . The early mortality of infected IL-27R-/- mice was correlated with significantly elevated secretions of inflammatory cytokines , particularly IFN-γ , and enhanced activation of CD4+ Th1 cells . Importantly , IL-10 production was not impaired in infected IL-27R-/- mice . Either depletion of CD4+ T cells , resulting in a dramatically reduced secretion of IFN-γ , or neutralization of IFN-γ , prevented the early mortality of infected IL-27R-/- mice with a significantly reduced inflammatory response and a major amelioration of the liver pathology . Thus , our data identify IL-27 signaling as a novel pathway to prevent the early mortality via inhibiting hyperactivation of CD4+ Th1 cells and their excessive secretions of IFN-γ during experimental infection with extracellular protozoan parasites African trypanosomes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | IL-27 Signaling Is Crucial for Survival of Mice Infected with African Trypanosomes via Preventing Lethal Effects of CD4+ T Cells and IFN-γ |
Leishmaniasis is a major health problem in some endemic areas and yet , no vaccine is available against any form of the disease . Historically , leishmanization ( LZ ) which is an inoculation of individual with live Leishmania , is the most effective control measure at least against cutaneous leishmaniasis ( CL ) . Due to various reasons , LZ is not used today . Several live attenuated Leishmania have been developed but their use is limited . Previously , we developed a transgenic strain of L . major that harbors two suicide genes tk and cd genes ( lmtkcd+/+ ) for use as a challenge strain in vaccine studies . These genes render the parasite susceptible to Ganciclovir ( GCV ) and 5-flurocytosine ( 5-FC ) . The dual drug sensitive strain of L . major was developed using gene targeting technology using a modified Herpes Simplex Virus thymidine kinase gene ( hsv-tk ) sensitive to Ganciclovir antibiotic and Saccharomyces cerevisae cytosine deaminase gene ( cd sensitive to 5-flurocytosine ) that were stably introduced into L . major chromosome . BALB/c mice inoculated with lmtkcd+/+ developed lesions which upon treatment with GCV and 5-FC completely healed . In the current study , the transgenic lmtkcd+/+strain was assessed as a live vaccine model to determine the time necessary to develop a protective immune response . C57BL/6 mice were inoculated with the transgenic lmtkcd+/+strain , and treated at the time of inoculation ( day0 ) or at day 8 after inoculation . Immunized animals were challenged with wild-type L . major , and complete protection was induced in mice that were treated at day 8 . The results show that in contrast to leishmanization , in group of mice inoculated with a dual sensitive L . major development and persistence of lesion is not necessary to induce Th1 response and protection .
Cutaneous leishmaniasis ( CL ) manifests as a localized self-healing lesion ( s ) that in rare cases develops to a non-healing lesion . If non-healing lesions develop , they are extremely difficult to treat with current therapies [1] . Control measures for leishmaniasis such as vector and/or réservoir control are not always practical , especially in remote endemic areas with limited resources . Efficacy of available drugs for leishmaniasis especially for CL is not acceptable and resistant is emerging [2] , [3] , [4] , [5] , [6] . Leishmanization ( LZ ) involves inoculating of individuals with live virulent Leishmania major to induce a single lesion that mimics a natural infection but with the lesion located at a predetermined site . Upon healing , the leishmanized individuals are protected against natural infection . LZ has been shown to be the most effective control measure at least against CL but the practice has been discontinued except on a limited scale in Uzbekistan . Primarily this is due to the development of chronic lesions that require medical intervention [7] , [8] , [9] . Despite ample evidence that development of an effective vaccine against leishmaniasis is possible there is still no vaccine available against any form of human leishmaniasis [10] , [11] , [12] , [13] . One approach is to derive attenuated live vaccine strains of Leishmania through genetic manipulation to develop a parasite strain which has no virulence or a limited pathogenicity . A number of genetically manipulated Leishmania strains have been developed and studied in animal models with controversial results [14] , [15] , [16] , [17] , [18] . Previously , we developed a transgenic strain of L . major ( tk+/+–cd+/+ ) [lmtkcd+/+] harboring two suicide genes tk and cd genes that confer susceptibility to GCV and 5-FC , . as a challenge strain for vaccine studies . When BALB/c mice were inoculated in the flank with lmtkcd+/+ , lesions developed at the site of inoculation , upon treatment with GCV and 5-FC complete healing occurred [16] . To extend these studies lmtkcd+/+ was used to determine whether persistent infection is required for induction of a protective immune response against subsequent L major infection . The lmtkcd+/+ promastigotes were inoculated into C57BL/6 mice and the inoculated mice were treated at set times with GCV to clear the infection . The mice were then challenged with wild type L major . Long term ( 3 months ) complete protection against challenge with wild type L major was achieved with as little as 8 days vaccination time demonstrating that persistent infection is not required for complete protection .
The ethical committee; Institutional Animal Care and Research Advisory Committee of Pasteur Institute of Iran , Education Office dated January , 2008 , based on the Specific National Ethical Guidelines for Biomedical Research issued by the Research and Technology Deputy of Ministry of Health and Medicinal Education of Iran , issued in 2005 , approved the protocol . The L . major promastigotes ( MHOM/IR/76/ER ) used and from which the transgenic lmtkcd+/+parasites were derived , this L . major is the same isolate which was used for mass leishmanization , preparation of old world experimental vaccine and the Leishmania used for the skin test . Promastigotes were cultured in M199 medium ( Life Technologies , Inc . ) supplemented with 10% heat inactivated fetal calf serum ( Gibco BRL ) and 25 mM HEPES ( Gibco BRL ) , pH 7 at 26°C . The parasite virulence was maintained by passage in BALB/c mouse . Female C57BL/6 mice , 6–8 week-old were purchased from the Animal Breeding Facility Centre ( ABFC ) of Pasteur Institute , Karaj , Iran . The animals were maintained in the animal facility of the Pasteur Institute of Tehran . The experiments were carried out according to the guidelines of Ethic Committee for Human use of Laboratory Animals , Pasteur Institute , Tehran , Iran . Mice were inoculated subcutaneously ( SC ) at the right hind footpad with 2×106 stationary phase promastigotes of either L . major ( MHOM/IR/76/ER ) wild type ( WT ) or the transgenic lmtkcd+/+parasites in 50 µl PBS . The mice inoculated with lmtkcd+/+were divided into 3 groups and treated with a combination of GCV and 5Fcyt , 100 mg/Kg , intra-peritoneally ( IP ) either at the time of parasite inoculation ( day 0 ) , at day 8 after inoculation or for the control group which was left untreated . The dosage of the drugs used in this study was based on our previous study ( 17 ) . The lmtkcd+/+ inoculated groups were challenged in the left footpads with 2×106 virulent WT L . major SC at 3 weeks after the end of the treatment period . The lesion development was recorded by weekly measurement of the footpad thickness at the site of inoculation using a metric caliper up to 12 weeks after inoculation . Parasite burden was quantified once at week 10 after inoculation of the mice with either L . major wild type or with lmtkcd+/+ and again 5 weeks after the challenge with wild type L . major ( 2–5 mice per group ) . The parasite burden in the spleen and draining lymph nodes were determined using limiting dilution analysis . To enhance sensitivity , 2-fold dilutions of the samples ( up to 1/100 ) were used . Delayed-type hypersensitivity ( DTH ) reaction was checked prior to challenge by injection of freeze-thawed ( FT ) Leishmania major ( 2×106 promastigotes in 50 µl per injection ) into the contralateral uninfected hind footpad . FT L . major promastigotes were prepared by repeating a freeze ( −196°C ) /thaw ( 37°C ) cycle ten times . Footpad swelling was measured using a metric caliper at 24 , 48 and 72 h after injection . Three mice from each group were sacrificed before and at 5 weeks after challenge inoculation , spleens were removed and cells cultured in complete RPMI-1640 medium in the presence or absence of 20 µg/well of Soluble Leishmania Antigens ( SLA , 107 Leishmania promastigotes/ml equal to 100 µg/ml ) or Concavalin A ( ConA;10 µg/ml ) or without stimulation as a control . The levels of IFN-γ and IL-4 at weeks 5 and 10 post inoculation with lmtkcd+/+ or WT L . major and 5 weeks after challenge were determined in the supernatant collected from spleen cell culture ( 5 mice per group ) . Briefly , single spleen cell suspension was prepared , cultured and re-stimulated either with SLA ( 100 µg/ml ) or Con A ( 10 µg/ml ) . The supernatant was collected at 72 h . Then , the levels of IFN-γ and IL-4 were titrated using ELISA method according to the manufacturer's instruction ( Bender Medsystems , Gmbh , Austria ) . The sensitivity of the ELISA kits was 3 pg/ml for IL-4 and 7 . 5 pg/ml for IFN- γ . At week 5 after challenge , different groups of mice were tail bled and the levels of anti-Leishmania IgG1 and IgG2a Abs were checked by ELISA . All experiments were done in triplicates and the data was expressed as means ± S . E . M . The data was analyzed by one-way ANOVA followed by Tukey's test using SPSS V . 13 software . P value<0 . 05 was considered as statistically significant .
C57BL/6 mice were inoculated SC with live wild type ( WT ) L . major parasites or lmtkcd+/+ parasites and were either left untreated or treated with GCV/5-FCy at day 0 or day 8 . Lesion development was followed by the measurement of footpad thickness . Following challenge with L . major , the protection rate and the immune responses generated were assessed . C57BL/6 mice inoculated with lmtkcd+/+ or WT parasites and left untreated developed a similar lesion size which was cured around week 8–9 . In contrast , no lesion was developed in the group of mice which was inoculated with lmtkcd+/+ and received GCV/5-FCyt treatment at day 0 or day 8 . The group of mice inoculated with WT L . major which was treated with GCV/5-FCyt developed a lesion similar to the untreated group of mice ( Fig . 1A ) . The draining lymph nodes ( LN ) and spleen parasite burden was measured at week 10 post-inoculation ( 5 mice/group ) . The results showed no difference in the number of parasite in spleen and LN's in groups of mice inoculated with WT L . major and the group which was inoculated with lmtkcd+/+ and received no treatment , the parasite burden of spleen at week 10 after inoculation is presented in Fig . 1B and only parasite burden of spleen at week 5 after challenge with WT L . major is presented in Fig . 2B . At weeks 5 , 10 post-inoculation and week 5 post challenge mice ( 5 per group ) were sacrificed and spleens were removed . A single cell suspension of spleen was prepared and cultured in the presence of either SLA ( 100 µg/ml ) , Con A ( 10 µg/ml ) or without additional stimulation , lymphocyte transformation test ( LTT ) was done at 72 hours and the results showed a significantly ( p<0 . 05 ) stronger LTT in group of mice with history of L . major infection and the group which was inoculated with lmtkcd+/+parasites and treated on day 8 than the group of mice inoculated with lmtkcd+/+parasites and treated on day 0 ( Fig . 1C ) . The supernatants were collected and the levels of IFN-γ were titrated ( Fig . 1D ) . Similar levels of IFN-γ were produced in spleen cells of group of mice inoculated with WT L . major and the group of mice inoculated with lmtkcd+/+ . The level of IL-4 production was low and similar in group of mice inoculated with wild-type L . major or inoculated with lmtkcd+/+ at week 16 post infection ( data not shown ) . To assess whether groups of C57BL/6 mice inoculated with lmtkcd+/+ parasites are protected against WT L . major challenge , at week 5–6 post inoculation ( 3 weeks after the end of treatment upon commencing time ) , the groups of mice which received lmtkcd+/+ and were treated on day 0 or 8 were challenged with L . major . As well , a group of mice which had healed spontaneously after L . major infection and a group of naïve mice were inoculated with L . major as controls . The results showed that the group of mice which was inoculated with lmtkcd+/+ parasites and treated with GCV/5-Fcyt on day 8 and then challenged with WT at week 6 , did not develop any lesion or swelling similar to the group of mice challenged with L . major after previously self-healing lesion . In contrast , the group of mice which was inoculated with lmtkcd+/+ and treated at the same time ( Day 0 ) with GCV/5-Fcyt and the group of naïve mice inoculated with L . major developed lesions ( Fig . 2A ) . The parasite burden was quantified in draining LN at week 5 post-challenge with L . major , as shown in Fig . 2B . The number of parasites isolated from the group of mice which was inoculated with lmtkcd+/+ and treated at day 8 with GCV/5-FCyt and the group of mice which had previously self-healed following L . major infection was significantly ( p<0 . 05 ) lower than the group which was inoculated with lmtkcd+/+ and treated at the same time ( day 0 ) and the group of naïve mice which were inoculated with L . major for the first time . The number of parasites was very low in the groups of mice inoculated with either lmtkcd+/+ and treated at day 8 or inoculated with lmtkcd+/+ and not treated or the group of mice with history of L . major infection or the group of mice which were inoculated with lmtkcd+/+ and treated at day 0 , no significant difference was seen between the number of parasite in these groups . DTH was done in different group of mice by injection of freeze-thawed ( FT ) Leishmania major ( 2×106 promastigotes in 50 µl per injection ) into the contra lateral uninfected hind footpad . The results are presented in Fig . 2C , a similar strong DTH response is seen in group of mice inoculated with WT L . major , or inoculated with lmtkcd+/+ and treated with GCV/5-FCy on day 8 or left untreated , a low DTH response was seen in groups of mice inoculated with lmtkcd+/+ and treated with GCV/5-FCy on day 0 or uninfected naïve mice . At week 10 after inoculation ( before challenge ) and 5 weeks after challenge , the splenocytes were cultured , stimulated in vitro with either SLA ( 100 µg/ml ) , or Con A ( 105 µg/ml ) , or left unstimulated . LTT was done and the culture supernatants were collected at 72 hours and the level of IFN-γ and IL-4 was titrated using ELISA method . A significantly ( p<0 . 05 ) stronger LTT was seen in mice with history of L . major infection and the group which was inoculated with lmtkcd+/+ parasites and treated on day 8 than the group of mice inoculated with lmtkcd+/+ parasites and treated on day 0 ( data not shown ) . The level of IFN-γ was significantly higher in groups of mice inoculated with WT L . major or inoculated with lmtkcd+/+ and treated with GCV/5-FCy on day 8 or left untreated ( Fig . 2D ) . The level of IL-4 was similar in all the groups ( Fig . 2E ) . Serum samples were collected at 5 weeks after challenge , the results are presented in Fig . 2F , as shown a significantly ( P = 0 . 002 ) higher anti-L . major IgG antibodies were seen in the group of mice with history of L . major lesion or group of mice inoculated with lmtkcd++ and treated with GCV/5-FCy on day 8 , in comparison with the group of naïve mice or group of mice inoculated with lmtkcd+/+ and treated with GCV/5-FCy on day 0 . IgG1 and IgG2a showed a significant ( P = 0 . 001 ) increase after challenge compared to before challenge in all the groups and no significant difference was seen between the groups .
Cutaneous leishmaniasis manifests as a self-healing skin lesion ( s ) in exposed parts of the body , the healing process for lesions depends upon the Leishmania species involved and the host immune response . Usually healing takes up to 2 years , but CL might not be cured for several years with currently available treatments . Choices of therapeutic treatments for CL are limited and not always effective , often requiring multiple injections , introduce side effects and control measure tools are not always practical and successful [1] , [2] , [3] , [6] , [19] , [20] , [21] . It is well established that individuals with a history of CL are protected against development of further CL lesion . CL lesion ( s ) development is accompanied by the induction of strong immune response shown by in vivo and in vitro tests ( 9 , 21 ) . Despite many studies on leishmaniasis , immunological surrogate marker ( s ) of protection is not well defined in human leishmaniasis [9] , [10] , [22] . There is ample evidence to suggest that development of an effective vaccine against leishmaniasis is possible , but so far no vaccine is available against any form of leishmaniasis . The results of phase 3 clinical trials using crude Leishmania as vaccine were not promising [4] , [12] , [23] , [24] . It has been shown that in vitro CD4+/CD8+ T-cell responses to live Leishmania major are significantly stronger than responses to dead parasites [25] . The only successful protective measure against CL has been shown to be leishmanization . One of the major drawbacks of LZ is the development of a lesion which might not heal during the expected time period and not respond to treatment [7] , [9] , [10] . Research have therefore focused on developing a Leishmania strain which upon inoculation does not induce a lesion or induces a lesion with limited pathogenicity , but at the same time maintains immunogenicity and as such induce protection in which the leishmanized individuals upon natural infection induce no lesion or even a limited fast healing lesion . In this regard attenuated and genetically manipulated Leishmania were developed and showed to induce protection in murine model of leishmaniasis [4] , [15] , [16] , [26] , [27] . Co-inoculation of Leishmania with CpG ODN showed to reduce the pathogenicity , but yet no Leishmania preparation reached to human use [28] , [29] . Previously , the same group developed a recombinant double drug sensitive strain of lmtkcd+/+ by integration of a genetically engineered HSV tk gene to confer sensitivity to GCV , and the S . cerevisiae cd gene to induce sensitivity to 5-fluorocytosine . Inoculation of BALB/c mice with lmtkcd+/+ induces lesion similar to WT L . major , but the lesion was controllable by treatment with GCV/5-FCyt [16] . BALB/c mice does not mimic human CL so in the current study , C57BL/6 strain which is not a perfect model of human CL but more mimic the disease is used . Leishmanization which is an inoculation of virulent L . major in a predetermined part of the susceptible individuals , LZ induces a lesion similar to natural infection , protection against further multiple lesions is usually developed upon cure of the lesion caused by LZ and so far LZ showed to be the most effective preventive measure against CL . The main drawback of LZ is development of lesion [16] . Using drug sensitive Leishmania mimic natural infection similar to LZ and at the same time due to sensitivity of Leishmania to approved drugs assures a controllable lesion . As it is presented in Fig . 1 , C57BL/6 mice inoculated with L . major lmtkcd+/+ showed a lesion similar to WT L . major ( Fig . 1A , Fig . 2A ) with no difference in parasite burden ( Fig . 1B , Fig . 2B ) . A very low number of Leishmania parasite is detected in the group of mice inoculated with lmtkcd+/+ and treated with GCV/5-FCyt , A small number of Leishmania was detected in spleen of C57BL/6 mice long after recovery from L . major infection ( unpublished data ) . A similar Th1 response was induced shown by LTT ( Fig . 1C ) , DTH ( Fig . 2C ) and the cytokine levels of IFN-γ ( Fig . 1D , Fig . 2D ) and IL-4 ( Fig . 2E ) in groups of mice inoculated with WT and group of mice inoculated with lmtkcd+/+and treated with GCV and 5-FCyt on day 8 or left untreated , although in the group of mice inoculated with lmtkcd+/+and treated with GCV/5-FCyt on day 8 , no lesion was developed at the site of inoculation but the reason for small increase in the size of footpad swelling is due to a slight inflammation which induced at the site of inoculation . Upon challenge with L . major , no lesion was developed and strong protection was seen similar to the group of mice cured from L . major infection ( Fig . 2 A ) . The results showed that despite of no lesion development which was due to under control of recombinant L . major with ganciclovir and 5-Flourocytosin , strong Th1 immune response and protection against WT L . major was induced . | Leishmaniasis is still a major health problem in some endemic foci , yet no vaccine is available against any form of leishmaniasis . It is a general belief that recovery from cutaneous leishmaniasis ( CL ) is accompanied with long life protection . An inoculation of live pathogenic L . major into healthy individuals to induce lesion similar to CL is called Leishmanization ( LZ ) . Historically LZ showed to be the most effective control tool against CL . One of the drawbacks and reason for discontinuation of LZ was lesion development , which rarely lasts long . Treatment of CL is not an easy task . One line of development of an effective vaccine against leishmaniasis , a transgenic strain of L . major harboring two suicide genes tk and cd genes ( lmtkcd+/+ ) , was developed and previously checked in BALB/c mice . In this study , C57BL/6 mice were inoculated with transgenic lmtkcd+/+strain; the rate of protection , parasite burden and the type of immune response were checked , and the results showed that complete protection induced by inoculation of lmtkcd+/+strain if treatment is initiated on day 8 post inoculation . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"immune",
"response",
"biology",
"and",
"life",
"sciences",
"immunology",
"immune",
"system"
] | 2014 | A Dual Drug Sensitive L. major Induces Protection without Lesion in C57BL/6 Mice |
In this study we have analysed AtASY3 , a coiled-coil domain protein that is required for normal meiosis in Arabidopsis . Analysis of an Atasy3-1 mutant reveals that loss of the protein compromises chromosome axis formation and results in reduced numbers of meiotic crossovers ( COs ) . Although the frequency of DNA double-strand breaks ( DSBs ) appears moderately reduced in Atasy3-1 , the main recombination defect is a reduction in the formation of COs . Immunolocalization studies in wild-type meiocytes indicate that the HORMA protein AtASY1 , which is related to Hop1 in budding yeast , forms hyper-abundant domains along the chromosomes that are spatially associated with DSBs and early recombination pathway proteins . Loss of AtASY3 disrupts the axial organization of AtASY1 . Furthermore we show that the AtASY3 and AtASY1 homologs BoASY3 and BoASY1 , from the closely related species Brassica oleracea , are co-immunoprecipitated from meiocyte extracts and that AtASY3 interacts with AtASY1 via residues in its predicted coiled-coil domain . Together our results suggest that AtASY3 is a functional homolog of Red1 . Since studies in budding yeast indicate that Red1 and Hop1 play a key role in establishing a bias to favor inter-homolog recombination ( IHR ) , we propose that AtASY3 and AtASY1 may have a similar role in Arabidopsis . Loss of AtASY3 also disrupts synaptonemal complex ( SC ) formation . In Atasy3-1 the transverse filament protein AtZYP1 forms small patches rather than a continuous SC . The few AtMLH1 foci that remain in Atasy3-1 are found in association with the AtZYP1 patches . This is sufficient to prevent the ectopic recombination observed in the absence of AtZYP1 , thus emphasizing that in addition to its structural role the protein is important for CO formation .
Meiotic recombination is initiated by the formation of Spo11-catalysed DSBs during early prophase I [1] . Each break is resected to produce 3′ single-stranded DNA tails which then interact with the RecA homologs Rad51 and Dmc1 to form nucleoprotein filaments . The filament on one side of the break then invades the homologous duplex DNA on either of the two non-sister chromatids resulting in strand displacement to form a displacement loop ( D-loop ) . Extension of the invading strand via DNA synthesis increases the size of the D-loop , thus enabling the capture of the 3′-end of the DNA from the other side of the DSB . Further DNA synthesis and ligation of the DNA ends leads to the formation of two four-way junctions termed a double-Holliday junction ( dHj ) that , on resolution , results in the formation of a CO ( reviewed in [2] ) . An important feature of meiotic recombination is its close coordination with the alignment , pairing and synapsis of homologous chromosomes during prophase I ( reviewed in [3] ) . Mutants that are defective in chromosome axis or SC morphogenesis exhibit a profound effect on recombination and subsequent CO formation . For example , during meiosis in budding yeast IHR predominates over inter-sister chromatid recombination , this bias being in part dependent on the chromosome axis proteins Hop1 and Red1 [4] . The Hop1-related proteins AtASY1 and Hormad1/2 are thought to perform the equivalent role in Arabidopsis and mouse respectively [5]–[7] . It is proposed that AtASY1 is essential for AtDMC1-dependent IHR . In the absence of AtASY1 , the association of AtDMC1 with the early recombination intermediates appears compromised such that virtually all IHR is aborted [8] . Mutation of the budding yeast SC transverse filament gene ZIP1 results in a failure of CO-designated intermediates to progress to form COs [9] . CO formation is also affected , albeit in a distinct manner , in the corresponding Arabidopsis and rice mutants [10] , [11] . In the case of Arabidopsis lacking the Zip1 homolog , AtZYP1 , there is a moderate reduction in CO frequency which is accompanied by the occurrence of recombination between non-homologous chromosomes . In rice , mutation of the ZEP1 gene leads to an apparent increase in CO/chiasma frequency . The interrelationship between recombination proteins and meiotic chromosome organization is further emphasized in a study in the filamentous fungus Sordaria macrospora [12] . This revealed that , in addition to their previously described recombination functions , Mer3 , Msh4 and Mlh1 have roles in ensuring accurate juxtaposition of the homologous chromosomes . Nevertheless , understanding the functional interrelationship between the recombination machinery and the chromosome axes and SC and the extent to which it is conserved between species has proved challenging . Although the structural organization of meiotic chromosomes is conserved , the chromosome axis and SC proteins exhibit a high level of primary sequence divergence . This limits the efficacy of straightforward homology searches as a route to identification of homologues in different species [13] and raises the question of how functionally related these proteins may be . As one approach to overcome this problem we have begun to make an inventory of proteins present in meiocytes from Arabidopsis and the closely related species , Brassica oleracea , in order to analyze novel meiotic proteins and identify meiotic protein complexes ( [14] , KO , KM , ER and FCHF unpublished ) . Used in combination with homology searches this has enabled us to identify an Arabidopsis meiosis-specific coiled-coil protein , AtASY3 . Analysis of AtASY3 has revealed that it is a component of the chromosome axes during meiotic prophase I . We demonstrate that loss of AtASY3 compromises AtASY1 localization leading to a reduced level of CO formation and a defect in chromosome synapsis . Our study provides further insights into the role of chromosome axis-associated proteins and the SC in the control of CO formation .
Analysis of proteins from B . oleracea meiocytes by mass spectrometry ( MS ) resulted in the detection of 4 peptides with homology to the predicted product of the Arabidopsis gene , At2g46980 ( Figure S1A ) . We noted that At2g46980 ( referred to hereafter as AtASY3 ) is predicted to encode an 88 kDa protein with a coiled-coil domain in its C-terminus region ( Figure S1B ) . A database search revealed that the predicted coiled-coil protein was most closely related to a rice meiotic gene PAIR3 ( 25 . 6% identity ) and had weak homology to Red1 from budding yeast ( 16 . 4% identity ) [15]–[18] . RT-PCR analysis indicated that AtASY3 was transcribed in reproductive tissues but not in vegetative tissues suggesting a potential role during the reproductive stage of development ( Figure S1C ) . Thus we decided to characterize AtASY3 further , to determine whether it encoded a meiotic protein and to establish its role . We obtained three T-DNA insertion lines of AtASY3 from the Nottingham Arabidopsis Stock Centre ( NASC ) ( Figure S1B ) . Molecular characterization of each line confirmed the positions of the T-DNA insertions within AtASY3 ( Figure S2 ) . Homozygous plants from each line showed normal vegetative growth but fertility was reduced by around 75% ( Figure S3A ) . The reduction in fertility was consistent with a defect in meiosis . To confirm this , DAPI-stained chromosome spread preparations from pollen mother cells ( PMCs ) were examined by fluorescence microscopy . Since the three lines were cytologically indistinguishable , the analysis of Atasy3-1 is presented in Figure 1 and Atasy3-2 and Atasy3-3 are shown in Supplementary Figure S3B . Chromosome behavior was apparently normal from G2 through early prophase I ( Figure 1A , 1G ) . However , normal pachytene nuclei were not observed ( Figure 1B , 1H ) . As the chromosomes began to condense during late diplotene/diakinesis it became clear that a proportion of the homologous chromosome pairs had failed to form chiasmata ( Figure 1C , 1I ) . This was confirmed by the presence of univalent chromosomes at metaphase I leading to mis-segregation at both meiotic divisions resulting in the formation of aneuploid tetrads ( Figure 1D–1F , 1J–1L ) . That the meiotic phenotype was due to mutation of AtASY3 was confirmed by an allelism test in which a homozygous Atasy3-1/Atasy3-2 double mutant was found to exhibit meiotic defects indistinguishable from the parental lines ( Figure S3A–S3C ) and a complementation test in which a full length AtASY3 cDNA cloned under the control of the AtDMC1 promoter in pPF408 [19] , was found to restore normal meiosis in Atasy3-1 ( Figure S3D–S3I ) . The distribution of AtASY3 was studied by immunolocalization on chromosome spread preparations of wild-type PMCs using anti-AtASY3 antibody ( Figure 2 ) . AtASY3 foci were first detected on the chromatin at late G2 together with accumulation of the protein in the nucleolus ( Figure 2A ) . At leptotene the nucleolar signal of AtASY3 disappeared and the protein was detected along the chromosome axes ( Figure 2B ) . This persisted through zygotene and pachytene during which partial colocalization with the SC transverse filament protein AtZYP1 [10] was observed ( Figure 2C–2F ) . In wild-type meiocytes at the transition from late G2 to leptotene , which is cytologically typified by a prominent centralized nucleolus , numerous chromatin-associated foci and short stretches of AtASY1 staining were observed ( Figure 2G ) . By leptotene these developed into a linear , yet not entirely uniform , signal along the axes . Analysis using deconvolution software ( see Materials and Methods ) revealed that the variation in signal intensity at leptotene arises because AtASY1 is distributed along the axes as a series of diffuse hyper-abundant patches or domains separated by stretches of lower abundance ( Figure 2H ) . This is accentuated at pachytene when the DAPI-stained chromosomes appear as thick , rope-like structures . At this stage the AtASY1 signal is depleted along the axis and the domains become foci-like in appearance ( Figure 2I ) . The foci appear relatively evenly distributed and consistent in number ( Mean number of foci per nucleus = 160; n = 10 ) . The tendency for AtASY1 to form domains along the axis is supported by electron microscopy ( EM ) studies in the plant Crepis capillaris . Immunologold localization of ASY1 in chromosome spread preparations of C . capillaris meiocytes reveals that the gold particles form discrete axis-associated clusters rather than an evenly distributed signal ( Figure 2J ) . Application of anti-AtASY3 antibody to prophase I spread preparations of chromosomes from Atasy3-1 PMCs did not result in any AtASY3 signal ( Figure 3A ) . This confirmed the specificity of the anti-AtASY3 antibody and supported the RT-PCR analysis which indicated that the AtASY3 transcript was absent in the mutant lines ( Figure S1D ) . Since AtASY3 localized to the chromosome axes during prophase I , we investigated the effect of loss of the protein in Atasy3-1 on other axis components . Immunolocalization of the cohesin proteins AtSMC3 and AtSYN1 , the Arabidopsis ortholog of the budding yeast meiotic cohesin Rec8 [20] , [21] , on spread preparations of Atasy3-1 PMCs was indistinguishable from wild-type . Both proteins were detected as linear chromosome axis-associated signals during prophase I ( Figure 3B–3E ) , suggesting global sister chromatid cohesion is present . In contrast , localization of AtASY3 was dependent on the cohesin complex , as it was completely disrupted in an Atsyn1 mutant ( Figure 3F ) . Localization of AtASY1 in Atasy3-1 meiocytes at late G2/early leptotene was similar to wild-type , with numerous chromatin-associated foci and short stretches of signal observed ( Figure 3G ) . As prophase I progressed , AtASY1 co-localized with the axes . However , rather than forming a linear signal with the underlying domain organization observed in wild-type , the protein was detected as discrete , evenly-distributed foci which persisted until the chromosomes began to condense at the end of prophase I ( Figure 3H ) . These appeared rather heterogeneous in shape and size . The mean number of the AtASY1 foci was 69 per nucleus ( n = 30 ) , but there was considerable variation between individual nuclei , with the number ranging from 39–115 . In contrast to the aberrant localization of AtASY1 in Atasy3-1 , that of AtASY3 in an Atasy1 mutant was indistinguishable from wild-type . This suggests that while normal localization of AtASY1 is dependent on AtASY3 , this relationship is not reciprocal ( Figure 3I ) . The same situation has been observed in budding yeast where Hop1 loading requires Red1 but not vice-versa [22] . Previously we have shown that the association of AtASY1 with the chromosome axes is independent of AtSPO11-induced DSB formation [8] . Consistent with this we observed that the axis-associated AtASY1 foci remained in an Atasy3-1/Atspo11-1-4 double mutant ( Figure 3J ) . As anticipated , the double mutant failed to form chiasmata , confirming that those detected in Atasy3-1 are DSB-dependent ( Figure S4B A , B ) . As the initial cytological analysis of Atasy3-1 indicated a defect in chromosome synapsis we investigated this in more detail . In wild-type meiocytes , the SC transverse filament protein , AtZYP1 , polymerized to form the linear central region of the SC , such that at pachytene each pair of homologous chromosomes was fully synapsed ( Figure 3K ) [10] . In Atasy3-1 the linearization of AtZYP1 to form a continuous SC did not occur . In most cases the AtZYP1 signals remained as foci or on occasion formed short stretches that were often abnormally thick and distorted in appearance . In some instances these structures could represent the accumulation of polycomplexes , nucleating at sites where AtZYP1 was unable to polymerize correctly along the lateral elements of the paired homologs ( Figure 3L ) . That SC formation was disrupted in Atasy3-1 was supported by the analysis of silver-stained chromosome spread preparations using electron microscopy . In wild-type , fully synapsed homologous chromosomes were observed at pachytene ( Figure 4A ) . In chromosome spread preparations of the Atasy3 mutants using the same conditions the nuclei were diffuse and the chromosome axes could not be readily discerned ( Figure 4B ) . However , by modifying the chromosome spreading conditions ( see Materials and Methods ) it was possible to detect nuclei where more extensive regions of axis were visible ( Figure 4C ) . In some cases these were aligned , although the spacing often appeared variable . These observations support the immunolocalization studies that indicated that SC formation was disrupted . They also suggest that although chromosome axes are formed in Atasy3-1 , there is likely some structural defect , possibly making them more susceptible to fragmentation by the chromosome spreading procedure . Alternatively , axis formation may be incomplete in the mutant . To quantify the reduction in COs in the Atasy3 mutants we analysed the chiasma frequency and distribution in 50 Atasy3-1 PMCs . This revealed nuclei containing 0–6 chiasmata with an overall mean chiasma frequency of 3 . 3 ( Figure 1J , Figure S4A , Table S1 ) . Similar results were obtained for Atasy3-2 ( 3 . 17 n = 50 ) and Atasy3-3 ( 3 . 32 n = 50 ) . These observations contrast with wild-type nuclei which contain 8–12 chiasmata with an overall mean of 9 . 76 [23] . Inspection of the chiasma distribution in the Atasy3 mutants revealed that 74 . 8% of the residual chiasmata were localized to the distal regions of the chromosomes . This figure is unchanged from that of wild-type ( 73 . 8% n = 50 ) ( Table S1 ) . To further analyze the functional relationship between AtASY3 and AtASY1 we constructed an Atasy3-1/Atasy1 double mutant and compared the effect on chiasma formation to that in the Atasy3-1 and Atasy1 single mutants . This revealed a reduction in the mean chiasma frequency from 3 . 3 ( n = 50 ) observed in Atasy3-1 to 1 . 78 ( n = 50; P<10−7 ) in Atasy3-1/Atasy1 , but no significant difference between the double mutant and Atasy1 ( 1 . 88 n = 50; P = 0 . 681 ) ( Figure S4B C , D ) . Thus AtASY1 is epistatic to AtASY3 with regard to CO formation , whereas the relationship is reversed in terms of protein loading . A similar relationship exists for DSB formation and Red1 and Hop1 loading in budding yeast [22] , [24] . Whereas Hop1 loading is greatly reduced in a red1 mutant , but not vice-versa , a hop1 mutant exhibits a stronger defect in DSB formation . Our data suggest that the higher CO frequency in Atasy3-1 as compared to Atasy3-1/Atasy1 may be attributable to the AtASY1 foci that remain in the single mutant , but that this is insufficient to promote wild-type levels of COs . However , this interpretation assumes that immunolocalization detects all the axis-associated AtASY1 that is present in Atasy3-1 . In budding yeast a group of proteins referred to as ZMM , comprising Zip1 , Zip2 , Zip3 , Zip4 , Msh4 , Msh5 and Mer3 , are crucial for the formation of interference-sensitive COs [9] . Homologs of the ZMM genes have been identified in Arabidopsis where their mutation results in a ∼85% reduction in CO formation [reviewed in 25] . The remaining COs ( ∼15% ) exhibit a random numerical distribution based on a Poisson analysis [26] , [27] . Since loss of AtASY3 resulted in a substantial reduction in chiasmata/COs we surmised that the protein was required for the formation of normal levels of ZMM-dependent COs . To investigate this , an Atasy3-1/Atmsh4 double mutant was constructed and the chiasma frequency determined to establish if loss of AtASY3 resulted in any reduction in chiasmata over that observed in Atmsh4 . In a survey of 30 metaphase I nuclei from the Atasy3-1/Atmsh4 line no chiasmata were detected , whereas the mean chiasma frequency in the Atmsh4 mutant was 1 . 1 ( n = 30 ) ( Figure S4B E , F ) . This indicates that AtASY3 has a role in the formation of all meiotic COs . To address the basis for the reduced chiasma formation in Atasy3-1 , we began by investigating the level of DSB formation in the mutant . The meiosis-specific histone H2AX is rapidly phosphorylated in chromatin surrounding the site of a DSB [28] , [29] , [30] . This phosphorylated form , γH2AX , can be detected by immunolocalization as foci in chromosome spread preparations from meiocytes . In wild-type nuclei a mean of 160 . 8 ( n = 5 ) γH2AX foci were detected at leptotene ( Figure 5A ) . In Atasy3-1 the corresponding number was 114 . 2 ( n = 5 ) ( Figure 5B ) . This suggested that the formation of DSBs was significantly reduced in the Atasy3-1 mutant ( P<0 . 01 ) . However , based on these observations we cannot exclude the possibility that the observed difference in the number of foci in the mutant compared to wild-type was due to an accelerated turn-over of DSBs . It also assumes that H2AX is phosphorylated at all DSBs in the mutant . Immunolocalization of AtDMC1 and AtRAD51 [31] , [32] revealed a significant reduction in the number of foci observed in Atasy3-1 at leptotene relative to wild-type . In the case of AtDMC1 , the figures for Atasy3-1 and wild-type were 106 . 8 ( n = 5 ) and 143 . 8 ( n = 5 ) ( P = 0 . 02 ) respectively ( Figure 5C , 5D ) . For AtRAD51 , the mean number of foci in Atasy3-1 was 99 . 8 ( n = 5 ) compared to 141 ( n = 5 ) ( P<0 . 01 ) in wild-type meiocytes ( Figure S5A , S5B ) . Although the numbers of AtDMC1 and AtRAD51 foci only indirectly reflect the number of DSBs , these figures are consistent with the decrease in γH2AX foci observed in Atasy3-1 . The MutS homolog AtMSH4 is required for the formation of normal levels of interference sensitive COs [23] . In wild-type , AtMSH4 foci accumulate on the chromosomes at leptotene and these gradually reduce in number through zygotene . By early pachytene only a few remain and they continue to dissociate from the chromatin such that by late pachytene they have disappeared [23] . Similarly , in this study we observed a mean of 140 ( n = 5 ) AtMSH4 foci in wild-type nuclei at leptotene which gradually reduced to around 10 foci detectable at early pachytene ( Figure S5C; Figure 5E ) . A similar pattern of turnover of AtMSH4 foci was observed in Atasy3-1 . However , the peak number of foci observed at each stage was lower than in wild-type . At leptotene a mean of 109 . 8 ( n = 5 ) foci were recorded ( Figure S5D ) . The foci reduced in number to between 0–4 at early pachytene and were found to associate with the stretches of AtZYP1 protein present in Atasy3-1 ( Figure 5F ) . The MutL homologue , AtMLH1 , is thought to mark the sites of COs/chiasmata [33] . Dual-immunolocalization of AtMLH1 and AtZYP1 on chromosome spread preparations revealed 8–12 AtMLH1 foci ( mean = 9 . 8 , n = 10 ) in wild-type nuclei at pachytene ( Figure 5G ) . In Atasy3-1 this was significantly reduced to a mean of 3 . 2 AtMLH1 foci ( P<0 . 001; n = 10 ) per nucleus which is consistent with the observed chiasma frequency . These foci invariably co-localized with the patches of AtZYP1 that remained in the Atasy3-1 meiocytes ( Figure 5H ) . These observations reveal a coordinate reduction in the number of γH2AX , AtDMC1 , AtRAD51 and early AtMSH4 foci to around 60% of the number observed in wild-type . However , there must be an additional defect or defects since the overall number of crossovers , based on AtMLH1 foci and chiasma counts , is only around 30% of the wild-type level . Nevertheless , there is no overall repair defect as there is no evidence of chromosome fragmentation . Previously we have reported that loss of AtASY1 also leads to a major defect in CO formation , but with no obvious defect in DSB formation suggesting that DSBs occur at or near the wild-type level [8] . This was based on immunolocalization of γH2AX foci in squash preparations of meiocytes at early prophase I . Since our analysis of Atasy3-1 was carried out using chromosome spread preparations we decided to determine the number of γH2AX foci in Atasy1 using the same approach . This revealed that the mean number of γH2AX foci was 129 . 5 ( n = 10 ) ( Figure S5E , S5F ) . Thus as previously [8] , there is no evidence of a major depletion of γH2AX foci in Atasy1 , but a slight reduction cannot be excluded ( P = 0 . 12 ) . Dual localization of γH2AX and AtASY1 in wild-type nuclei revealed the γH2AX foci were adjacent to the hyper-abundant domains of AtASY1 and showed a slight tendency to overlap ( Figure 5A ) . This observation is consistent with the proposal that the nascent DSB is tethered to the chromosome axis more or less coincident with its formation [4] , [34] , [35] . Similarly dual localization of AtDMC1 and AtASY1 revealed a close association of the AtDMC1 foci with the AtASY1-stained axis , but the signals were largely distinct and did not overlap ( Figure 5C ) . Despite the fact that AtASY1 was present as discrete foci rather than hyper-abundant domains in Atasy3-1 , these were virtually all ( 97 . 4% n = 5 ) found in association with a γH2AX signal ( Figure 5B ) . In a control designed to detect fortuitous co-localization ( see Materials and Methods ) this figure was 39% ( n = 5 ) . However , in the sample of cells analysed the mean number of γH2AX foci was greater than the number of AtASY1 foci ( 114 . 2 versus 87 . 8 respectively , n = 5 ) . Hence , not all of the γH2AX foci co-localize with AtASY1 in Atasy3-1 . The AtASY1 foci in Atasy3-1 were also found in association with the AtDMC1 foci ( 95% , n = 5 ) ( Figure 5D ) . However , in this case , unlike in wild-type , there was a tendency for signals to overlap . Hence it would appear that the two proteins end up localized at the same sites on the chromosomes but their spatial relationship may be perturbed in the absence of AtASY3 . We have previously shown that the stable association of AtDMC1 foci to the chromosome axes during early prophase I requires AtASY1 [8] . Since AtASY1 localization is disrupted in Atasy3-1 , we investigated the localization of AtDMC1 in Atasy3-1 in more detail . The chronology of AtDMC1 localization was determined by carrying out the immunolocalization of the protein together with prior BrdU pulse-labeling of the PMCs during meiotic S-phase as described previously [8] . This revealed that in Atasy3-1 , Atasy1 and wild-type maximum numbers of AtDMC1 foci accumulated on the chromosomes around 12 h following the BrdU-pulse . However , at 24 h while numerous AtDMC1 foci still remained in both Atasy3-1 and wild-type , they were entirely absent in Atasy1 ( Figure S6 ) . These data indicate that the rapid loss of AtDMC1 foci observed in Atasy1 does not occur in Atasy3-1 . That turnover of AtDMC1 foci in Atasy1 is more rapid than in wild-type could suggest that some normal barrier to progression is absent , such that recombination proceeds but CO-designation is defective . Since the initial rate of accumulation of AtDMC1 is normal in Atasy3-1 it would suggest that the reduced number of foci in the mutant may reflect a reduction in DSBs rather than an increase in turnover . Furthermore , it suggests that despite the overall depletion of AtASY1 , the residual protein is sufficient to ensure that the temporal localization of AtDMC1 is similar to that in wild-type meiocytes . The finding that normal axis-association of AtASY1 was dependent on AtASY3 combined with the observation that the two proteins co-localized during prophase I suggested both a functional and possibly a direct physical inter-relationship between them . To obtain evidence that AtASY3 and AtASY1 may be components of a meiotic complex and possibly interact in planta we conducted a co-immunoprecipitation ( CoIP ) experiment using anti-AtASY1 antibody . Since extracting protein in sufficient quantities from Arabidopsis meiocytes is impractical due to the small size of the anthers , we used meiocyte extracts from B . oleracea . In previous studies we have shown that an anti-AtASY1 antibody recognizes the corresponding Brassica protein , BoASY1 , which shares 83% amino acid sequence identity with AtASY1 [36] . Similarly , analysis of BoASY3 from B . oleracea indicated that it encodes a protein with 77% sequence identity with its Arabidopsis counterpart ( Figure S7 ) and that it localized to the chromosome axes during prophase I of meiosis ( Figure S8 ) . Co-immunoprecipitating proteins were identified by MS of tryptic peptides followed by searches against a combined Brassica database containing the BoASY1 and BoASY3 full-length sequences ( see Materials and Methods for details ) . BoASY1 was identified as the top hit with 40 unique peptides corresponding to 397/599 amino acids ( 65% coverage ) and BoASY3 as the second hit with 28 peptides corresponding to 297/776 amino acids ( 38% coverage ) ( Figure S9A ) . Both proteins were absent from the control sample . Thus it would seem probable that BoASY1 and BoASY3 are components of a complex and given the close conservation with Arabidopsis , is likely the case for AtASY1 and AtASY3 . To determine if the proteins can directly interact , the full-length cDNAs of each gene were cloned as in-frame fusions with the GAL4 DNA binding domain ( DBD ) and the GAL4 activator domain ( AD ) in the yeast two hybrid plasmids pGBKT7 and pGADT7 respectively . Plasmids encoding the AtASY1-DBD and AtASY3-AD fusions and the reciprocal pair of constructs were co-transformed into yeast . At the same time control transformations , where one of the constructs was replaced by the corresponding empty vector , were also carried out . All plasmid combinations enabled the transformed yeast to grow on synthetic dropout medium ( -Leu/-Trp ) that selected the auxotrophic markers on the cloning vectors . When the plasmid combinations were tested on low stringency selective plates ( -Leu/-Trp/-His ) the yeast cells containing AtASY1 and AtASY3 as reciprocal DBD/AD fusions enabled significant growth at all dilutions . Some slight growth was also detected in controls containing the AtASY1-DBD and AtASY3-AD plasmids . However , under higher stringency selection ( -Leu/-Trp/-His/-Ade ) growth was entirely restricted to the combinations carrying both genes ( Figure 6 ) . Further experiments with a series of plasmids in which truncated regions of AtASY3 were fused to the GAL4 activator domain indicated that the interaction of AtASY3 with AtASY1 was dependent on amino acid residues 623–793 of AtASY3 which correspond to the predicted coiled-coil region ( Figure 6 ) . These data suggested that AtASY1 and AtASY3 can interact and likely do so in planta .
We have identified AtASY3 , a chromosome axis protein that is required for normal levels of COs and SC formation in Arabidopsis . AtASY3 is predicted to contain a coiled-coil domain towards the C-terminus . This structural feature is found in other meiotic proteins , such as Red1 in budding yeast , SCP3/Cor1 in mammals and OsPAIR3 in rice , that are components of the axial regions of the SC , yet on the basis of amino acid sequence homology are reported to have no close homologs in other species [15] , [16] , [18] , [37] , [38] . AtASY3 appears similar in this respect . Although it shares 77% sequence identity with BoASY3 from the closely related B . oleracea , the level of sequence identity with Red1 and OsPAIR3 is limited , at 16 . 4% and 25 . 6% respectively . This sequence divergence appears to be a recurrent feature of proteins that are components of the chromosome axes . In Arabidopsis the level of DSBs may be inferred from the number of γH2AX and/or AtDMC1 or AtRAD51 foci detected in meiocytes at early prophase I [8] . Our data suggested a consistent reduction in DSBs from 150–160 in wild-type meiocytes to about 100 in Atasy3-1 , an overall reduction of around 33% . However , as DSB detection is indirect , we cannot formally rule out the possibility that not all DSBs are detected in the mutant . Nevertheless , as the number of γH2AX foci is very similar to those of the strand-exchange proteins this seems a reasonable conclusion . The physical association of the recombination machinery and the chromosome axes has been known for many years [3] , [34] , [39] . Recent studies in budding yeast indicate that the DSB machinery becomes tethered to the chromosome axes prior to break formation and that the axis components Red1 and Hop1 are required for this [4] , [35] . Analysis of red1 mutants has revealed a defect in DSB formation [40]–[42] . It is conceivable that AtASY3 may play a similar role in axis organization that is crucial to enable normal levels of DSB formation . The reduction in foci corresponding to γH2AX and the strand-exchange proteins is consistent with a scenario whereby DSB formation occurs in the context of the axis , rather than the recombination complex associating with the axis following break formation , as in the latter instance , a reduction in γH2AX foci would not be anticipated . However , it remains possible that depletion of AtASY3 may induce a change in chromatin structure that in turn results in a reduction in DSBs . Although DSB formation appears to be significantly reduced in Atasy3-1 , any reduction due to the loss of AtASY1 appears more marginal . This is in broad agreement with our earlier study [8] . It is worth noting that this difference mirrors chromosome axis formation in the two mutants . Whereas loss of AtASY3 disrupts the axes , the Atasy1 mutant has clearly defined axes , albeit with some minor discontinuities , seemingly indicating the importance of the chromosome axes for efficient DSB formation [43] . However , it contrasts with observations in budding yeast and mouse , where mutants lacking Hop1 and HORMAD respectively exhibit strongly reduced numbers of DSBs [44] , [45] . At present the basis and significance of this difference remains unknown . It may reflect an underlying difference in the control of DSB formation , although the formation of DSBs in hop1 and red1 mutants in different budding yeast strains also shows some variation [40] , [46] . A reduction in the mean chiasma/CO frequency to around 30% of the wild-type level was observed in all Atasy3 lines . Analysis of the Atmsh4/Atasy3-1 double mutant indicates that loss of AtASY3 compromises the formation of COs which are subject to CO interference , and also non-interfering COs . This would suggest that AtASY3 plays a crucial role at an early stage in the recombination pathway . A reduction in CO formation is also a characteristic of mutants in the axis-associated proteins Red1 and OsPAIR3 [18] , [24] , [40] , [41] . A recent study in budding yeast has indicated that loss of Red1 results in a conversion from inter-homolog bias to inter-sister bias for DSB repair . It is proposed that this is due to a loss of constraint on Rec8 loading at the recombination site , thus favoring inter-sister recombination [4] . It is conceivable that a similar situation arises through the loss of AtASY3 . If so , this could explain why there is a proportionally greater reduction in CO formation in Atasy3-1 than might be expected from the apparent reduction in DSBs . Simple extrapolation would suggest that if the ratio of COs to non-COs was maintained in the mutant , the mean CO/chiasma frequency would be 6–7 rather than the 3 . 5 observed . This implies that there is a loss of inter-homolog bias and that a greater proportion of the DSBs are repaired via another route such as inter sister-chromatid exchange . Alternatively , loss of AtASY3 may result in preferential processing of some or all of the recombination intermediates to favor non-COs rather than COs . At present the data does not enable us to distinguish between these possibilities . The observation that 75 . 7% of the chiasmata that remain in Atasy3-1 are sub-telomeric/distal is consistent with previous studies in Arabidopsis showing that the telomeres cluster on the nucleolus during early prophase I and as a result the sub-telomeric regions of the homologous chromosomes are placed in proximity [47] . A similar situation was observed in Atasy1 where virtually all the residual chiasmata are distal [48] . Deconvolution of the linear AtASY1 signal in wild-type meiocytes at leptotene through early pachytene revealed that it comprises evenly-spaced axis-associated domains of hyper-abundance interspersed with more lightly staining regions . This organization was supported by immunogold studies in meiocytes from C . capillaris . The number of AtASY1 domains appeared quite consistent , at around 160 per nucleus . The organization of AtASY1 is highly reminiscent of that of Hop1 which has been observed to form domains of alternating hyper-abundance and lower-abundance in budding yeast at early pachytene [49] , [50] . Loss of AtASY3 resulted in a significant effect on the distribution of AtASY1 , such that axis-associated foci were observed during prophase I . These were fewer in number than the AtASY1 domains observed in wild-type and this varied from cell to cell ranging from 39–115 in the sample examined . Nevertheless , dual-localization studies with γH2AX and AtDMC1 indicated that these AtASY1 foci coincided with sites of recombination . However , further investigation will be required to determine whether or not localization of AtASY1 at the axial region still occurs with normal spatial specificity in Atasy3-1 . Studies in budding yeast initially reported that localization of the AtASY1 homolog , Hop1 , was dependent on the chromosome axis protein Red1 [22] . However , more recently it was suggested that while normal levels of Hop1 loading require Red1 , some Hop1 is loaded at the sites of DSBs independently [46] . Our observations suggest a similar relationship between AtASY1 and AtASY3 . In wild-type , the AtASY1-abundant regions along the axes at prophase I correlated both spatially and numerically with the γH2AX and AtDMC1 foci , suggesting that the DSBs are positioned within AtASY1-enriched domains . This is consistent with the proposed role of AtASY1 in promoting IHR [8] . Interestingly , whereas the γH2AX and AtASY1 signals overlapped , the AtDMC1 signal was adjacent to , but did not merge with the AtASY1 signal . Further study will be required to explore the significance of this observation . It appears that a similar spatial relationship between DSBs and AtASY1 is maintained in Atasy3-1 as virtually all the AtASY1 foci in the mutant co-localize with γH2AX . Thus , it suggests that recruitment of AtASY1 and the recombination machinery to the axial region is spatially coordinated . However , these may not be interdependent events since localization of AtASY1 in the Atspo11-1-4 null mutant appears , based on immunofluorescence , normal [8] . Similarly in this study , examination of AtASY1 localization in an Atasy3-1/Atspo11-1-4 double mutant suggests that the AtASY1 foci observed in Atasy3-1 are still formed and associated with the axial region . Nevertheless , further studies will be required to establish if the AtASY1 domains observed in Atspo11-1-4 and Atasy3-1/Atspo11-1-4 are identical to those in the wild-type . If so , then it suggests that AtASY1 is initially recruited to predetermined chromosomal regions that also encompass DSB hotspots , possibly establishing a spatial relationship favoring IHR . Alternatively , it is conceivable that the AtASY1 domain formation observed in wild-type may be guided or influenced by DSB formation or the pre-DNA break recombinosome complex . Studies in budding yeast have also led to the proposal that the interhomolog bias is established before DSB formation with enforcement of the bias occurring at the transition from the nascent DSB to a joint molecule recombination intermediate [41] . Although the majority of the AtASY1 foci in Atasy3-1 were associated with γH2AX foci at early prophase I , there were additional γH2AX signals that did not colocalize . If loss of DSBs and the destabilization of AtASY1 localization observed in Atasy3-1 occurs stochastically , then , given that formation of AtASY1 foci and DSBs does not appear to be co-dependent , one would expect to see a similar proportion of γH2AX and AtASY1 foci that were not associated with one another . As this does not appear to be the case , it seems possible that a sub-set of DSBs occur in regions that are not associated with AtASY1 . It may simply be that although the AtASY1 foci that persist in Atasy3-1 correspond to the position of the domains observed in wild-type they are substantially smaller . Hence a proportion of the γH2AX foci no longer co-localize despite the normal spatial recruitment of the recombination complexes to the chromatin . Alternatively , some DSBs may occur in regions of lower AtASY1 abundance . Previously , it has been proposed in budding yeast that some DSBs are formed at random sites that are not associated with DSB hotspots [41] . Hence , if the AtASY1 domains are coincident with hotspots , which would seem logical , this may also be the case in Arabidopsis . Despite amino acid sequence variation it has been suggested that Red1 and SCP3/Cor1 may be structural analogs [15] . Nevertheless , to date a functional ortholog of Red1 in a multi-cellular organism has not been reported . However , the studies described here suggest that AtASY3 has at least some functional similarity to Red1 . The most compelling evidence arises from the close functional interrelationships that both proteins share with their corresponding HORMA domain proteins . Loss of Red1 and AtASY3 proteins results in a disruption of Hop1 and AtASY1 localization respectively , during prophase I [22] . In budding yeast Red1 has been shown to interact with Hop1 in co-immunoprecipitation experiments [46] . Yeast two-hybrid studies have shown that the 290 amino acids at the C-terminus of Red1 are essential for the interaction with Hop1 . This part of the protein is predicted to form a coiled-coil domain . In this study MS analysis of proteins that were co-immunoprecipitated from Brassica meiocytes using an anti-AtASY1 antibody revealed a likely interaction between BoASY1 and BoASY3 , the homologs of AtASY1 and AtASY3 respectively . That AtASY3 and AtASY1 can directly interact was confirmed by yeast two-hybrid analysis . Moreover further study revealed that the C-terminal of AtASY3 that contains a predicted coiled-coil region is essential for the interaction between the two proteins . Studies indicate that Red1 is required for normal levels of DSB formation . This also appears to be the case for AtASY3 , although this effect does not appear as pronounced in Atasy3-1 as that in a red1 mutant where a reduction in DSB formation to ∼25% of wild-type has been reported [42] . It seems likely that PAIR3 in rice may also be a functional homolog of Red1 . While a direct interaction with the rice HORMA domain protein OsPAIR2 has not yet been demonstrated , localization of OsPAIR2 is OsPAIR3 dependent and an Ospair3 mutant has a similar phenotype to that of Atasy3-1 [17] , [18] . In addition to its structural role within the SC , the budding yeast protein Zip1 has been shown to play a key role in meiotic recombination [51] , [52] . Zip1 together with other members of the ZMM group of proteins , Zip2 , Zip3 , Zip4 , Msh4 , Msh5 and Mer3 , is crucial for the formation of interference-dependent COs [51] . In Arabidopsis loss of AtZYP1 only results in a modest reduction in chiasma frequency to about 80% of the wild-type level . However , many of the remaining chiasmata occur between ectopic chromosome regions , possibly between duplicated sequences that amount to around 60% of the Arabidopsis genome [10] . The studies described here revealed that loss of AtASY3 had a profound effect on the formation of the SC . In most nuclei it appeared that alignment of the chromosome axes was extensively disrupted and immunolocalization studies with the transverse filament protein AtZYP1 indicated little evidence of normal SC assembly . In general , nuclei contained a mixture of AtZYP1 foci or occasional short stretches which appeared abnormally thickened and deformed . Although Atasy3-1 is essentially asynaptic , there is no evidence of the non-allelic recombination observed in plants lacking AtZYP1 . The bivalents that remained in the Atasy3-1 meiocytes at metaphase I comprised homologous chromosomes and there was no evidence of multivalent formation . Immunolocalization studies revealed a reduction in AtMLH1 foci that reflected the reduction in chiasmata in Atasy3-1 . These AtMLH1 foci , which are thought to localize to CO sites [33] , were invariably associated with the residual AtZYP1 present in the mutant . Hence , it would appear that the presence of AtZYP1 at the site of recombination is important for the prevention of non-allelic recombination , but extensive SC polymerization is not required . This finding provides further evidence that in addition to SC formation , AtZYP1 plays an important role in the formation of COs in Arabidopsis . In summary , these studies provide further insight into meiotic CO formation . Moreover they emphasize that despite the lack of sequence homology between the chromosome axes components from different species , it seems likely that a close functional relationship remains .
A . thaliana ecotype Columbia ( 0 ) was used for wild-type analysis . T-DNA insertion lines SALK_143676 , SALK_050971 and SAIL_423_H01 were obtained from NASC for mutant analysis [53] . Plants were grown , material harvested and nucleic acid extractions were performed as previously described by Higgins et al . [23] . The T-DNA insertion site of the mutant lines was confirmed as previously described [23] ( Figure S2 ) . Details of the primers used are presented in Table S2 . Primers ASY3-CM-F1 and ASY3-CM-R1 ( Table S2 ) were used to amplify the entire AtASY3 coding sequence with flanking 5′ and 3′ UTR regions from cDNA clone pda 19140 ( Riken , Japan ) . The PCR product was cloned into the binary vector pPF408 [19] using SpeI sites incorporated into the primers . The construct was confirmed by sequencing . The binary plasmid construct was introduced into Agrobacterium tumefaciens LBA 4404 and plants transformed as previously described [23] . RNA extraction and RT-PCR was carried out as previously described [54] . Details of the primers are given in Table S2 . Nucleotide sequencing was carried out by the Genomics and Proteomics Unit , School of Biosciences , University of Birmingham , UK . Primers ASY3-AB-F1 and ASY3-AB-R1 ( Table S2 ) were used to amplify a 702bp fragment comprising amino acid residues 560 to 793 of AtASY3 from cDNA clone pda 19140 ( Riken , Japan ) . The PCR product was cloned into the expression vector pET21b ( Novagen ) using NdeI and Xhol sites incorporated into ASY3-AB-F1 and ASY3-AB-R1 respectively . Recombinant His-tagged protein was isolated from E . coli BL21 ( Novagen ) under native conditions using Ni-agarose following the manufacturer's protocol ( QIAGEN ) . Polyclonal antiserum against the recombinant protein was raised in rabbit ( BioGenes GmbH , Germany ) . Cytological studies were carried out as previously described [23] . The following antibodies were used: anti-AtASY3 ( rabbit , 1/200 dilution ) anti-AtASY1 ( rabbit/rat , 1/1000 dilution ) , anti-AtMSH4 ( rabbit , 1/500 dilution ) , anti-AtZYP1 ( rabbit/rat , 1/500 dilution ) , anti-AtSMC3 ( rabbit 1/500 dilution ) , anti-AtSYN1 ( rabbit 1/500 dilution ) , anti-AtDMC1 ( rabbit 1/500 dilution ) , anti-AtRAD51 ( rabbit , 1/500 dilution ) , anti-AtMLH1 ( rabbit/rat , 1/200 dilution ) and anti-γH2AX ( ser 139 , catalog no . 07-164 Upstate Biotechnology; rabbit , 1/100 dilution ) [10] , [23] , [55] . Microscopy was carried out using a Nikon 90i Fluorescence Microscope ( Tokyo , Japan ) . Image capture , image analysis and processing were conducted using NIS-Elements-F software ( Nikon , Tokyo , Japan ) . Image deconvolution was carried out using the function “Mexican hat” . This allows better discrimination of the signals . This function performs filtration on the intensity component ( or on every selected component - when working with multichannel images ) of an image using convolution with 5×5 kernel . Mexican Hat kernel is defined as a combination of Laplacian kernel and Gaussian kernel it marks edges and also reduces noise . In double-staining experiments the level of chance overlap of foci was assessed using the misorientation method whereby one of the two images is rotated through180 degrees following which colocalizing foci are counted as previously described [56] . Electron microscopy and immunogold labeling was performed as previously described [36] , [57] except for the modified Atasy3 chromosome spread protocol where the detergent was Triton X-100 0 . 1%+Lipsol 0 . 05% and the digestion time was increased to 7 min . Chiasma counts were carried out as previously described [58] . Chromosome spread preparations from PMCs at metaphase I were examined by light microscopy after fluorescence in situ hybridization ( FISH ) using 45S and 5S rDNA probes . The use of FISH enabled the identification of individual chromosomes . The overall shape of individual bivalents allowed the number and position of individual chiasmata to be determined and this was also informed by the position of the FISH signals . The statistical procedures were carried out as described previously [23] . AtASY3 peptides were initially identified in meiocyte extracts prepared from Brassica oleracaea var . alboglabra A12DHd as previously described [14] . Co-IP experiments were conducted on meiocytes extracted from the same material . Protein extraction was under non-denaturing conditions at 4°C . Briefly , tissue was powdered by grinding in liquid nitrogen , resuspended in IP Buffer ( 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 10% glycerol , 2 mM EDTA , 0 . 1% NP40 , protease inhibitor cocktail ( Roche , #11836170001 ) , phosphatase inhibitor ( Thermo Scientific #78420 ) ) and cell debris removed by centrifugation . Protein extracts were used immediately . Antibodies were cross-linked to Affi-Prep Protein A beads ( Bio-Rad , #156-0006 ) using DMP ( Sigma #D8388 ) and pre-eluted with glycine to remove any non-cross-linked antibody . Meiocyte extracts were pre-cleared by incubation with non-specific IgG . Parallel Co-IPs were carried out using affinity purified anti-AtASY1 antibody and an unrelated antibody as a control . Following washing to remove non-specific proteins , bound proteins were glycine-eluted , trypsin-digested and analysed on an LTQ-Orbitrap Velos mass spectrometer ( Thermo Fisher Scientific ) . Since a complete sequence of the Brassica oleracea genome is currently unavailable , protein identification was carried out against a combined Brassica rapa/napus/oleracea database , downloaded from NCBI ( http://www . ncbi . nlm . nih . gov ) in 2010 , into which the BoASY1 and BoASY3 full-length sequences had been manually inserted . For details of the MS analysis see Figure S9B . Yeast two-hybrid screens were performed according to the Yeastmaker Yeast transformation System 2 manual ( Clontech , USA ) . Briefly , Y2HGold yeast cells were co-transformed with pGADT7 and pGBKT7 using the polyethylene glycol/lithium acetate method . The co-transformed yeasts were grown in SD -Leu/-Trp , SD -Leu/-Trp/-His and SD -Leu/-Trp/-His/-Ade for testing protein-protein interaction through the activation of the two reporter genes His3 and Ade2 . The strength of the interaction was assayed by drop test using serial dilutions of mid-exponential-phase cultures . 3 µl drops of undiluted , 10- and 100-fold diluted culture were spotted on the selective agar medium and incubated at 30°C for 2 days . Details of the primers used for plasmid construction are shown in Table S2 . Plasmid constructs are as shown in Figure 6 . | Homologous recombination ( HR ) during prophase I of meiosis leads to the formation of physical connections , known as chiasmata , between homologous chromosomes ( homologs ) . Chiasmata are essential for accurate homolog segregation at the first meiotic division . HR is initiated by the formation of DNA double-strand breaks ( DSBs ) . As DNA replication prior to meiosis results in the duplication of each homolog to form two identical sister chromatids , a DSB in one sister chromatid could potentially be repaired using the other as the repair template rather than one of the two non-sister chromatids of the homolog . If this route were predominant , the formation of chiasmata would be disfavored and chromosome segregation would be compromised . However , during meiosis there is a strong bias towards inter-homolog recombination ( IHR ) . In this study we have identified AtASY3 , a component of the proteinaceous axes that organize the chromosomes during meiosis in Arabidopsis . We find that AtASY3 interacts with AtASY1 , a previously identified axis protein that is essential for crossover formation . We show that loss of AtASY3 disrupts the axis-organization of AtASY1 . This results in a substantial reduction in chiasmata , and there is extensive chromosome mis-segregation . We propose that loss of AtASY3 affects the efficiency of the inter-homolog bias . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"science",
"plant",
"biology",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Inter-Homolog Crossing-Over and Synapsis in Arabidopsis Meiosis Are Dependent on the Chromosome Axis Protein AtASY3 |
In C . elegans , removal of the germline triggers molecular events in the neighboring intestine , which sends an anti-aging signal to the rest of the animal . In this study , we identified an innate immunity related gene , named irg-7 , as a novel mediator of longevity in germlineless animals . We consider irg-7 to be an integral downstream component of the germline longevity pathway because its expression increases upon germ cell removal and its depletion interferes with the activation of the longevity-promoting transcription factors DAF-16 and DAF-12 in germlineless animals . Furthermore , irg-7 activation by itself sensitizes the animals' innate immune response and extends the lifespan of animals exposed to live bacteria . This lifespan-extending pathogen resistance relies on the somatic gonad as well as on many genes previously associated with the reproductive longevity pathway . This suggests that these genes are also relevant in animals with an intact gonad , and can affect their resistance to pathogens . Altogether , this study demonstrates the tight association between germline homeostasis and the immune response of animals , and raises the possibility that the reproductive system can act as a signaling center to divert resources towards defending against putative pathogen attacks .
One of the most significant findings in the aging research field is the realization that lifespan is determined by the aging rate and by deleterious events that limit lifespan . Over the last few decades , a wealth of genes and signaling molecules that determine lifespan and longevity were identified . Genome-wide screens in the model organism C . elegans revealed that genes that influence aging fall into four conserved longevity pathways: the dietary restriction pathway , the insulin/IGF-1 pathway , the mitochondrial pathway and the reproductive pathway [1 , 2] . In general , one common theme of these longevity regulatory pathways is their ability to shift resources towards enhanced maintenance and stress resistance of the soma [3] . In C . elegans and in Drosophila , germ cell removal can increase lifespan and increase resistance to oxidative stress , proteostasis stress and pathogens [4–9] . These beneficial effects of germ cell depletion require the presence of an intact somatic gonad [4 , 5] . In addition , nuclear hormone signaling [5] , autophagy [10–12] , fat metabolism [10 , 12] , proteasome components [9 , 13] , microRNA-regulators [14] and a variety of transcription factors [5 , 9–12 , 15–18] are required for lifespan extension upon germ cell removal . Some of these genes ( for example daf-16 ) are common mediators of several longevity pathways , whereas others are uniquely associated with the longevity of germlineless animals . Notably , many of the genes implicated in the longevity of germlineless animals act specifically in the intestine [19] , which comprises the animal's adipose tissue and liver , and is also the major site of pathogen colonization in C . elegans . Although C . elegans has evolved a dedicated innate immune response to protect the animals from infecting pathogens [20 , 21] , one of the factors that limit C . elegans lifespan is its slightly pathogenic microbial diet , which colonizes and infects the intestine of aging animals [22–24] . Accordingly , mutants deficient in the central innate immunity PMK-1/ATF-7 pathway exhibit a shortened lifespan even on E . coli OP50 , a weakened bacterial strain that is only slightly pathogenic to C . elegans [25] . Conversely , killing or limiting the proliferation of the bacterial food source extends C . elegans lifespan [22 , 26] . Furthermore , many mutations that prolong lifespan also enhance the resistance of animals to pathogenic bacteria [6 , 27–30] . This enhanced resistance to pathogens is thought to be the consequence of constant expression of a variety of innate immunity-related genes [17 , 30–32] , which act in parallel to the PMK-1/ATF-7 innate immunity pathway [6 , 30 , 33] . Together , these imply that pathogen infection ( even by the nonpathogenic OP50 E . coli strain ) and the host immune response are important determinants of C . elegans lifespan . Here we identified the innate immunity related gene irg-7 as a novel integral component of the reproductive longevity pathway , which links between genes implicated in this longevity pathway and pathogen resistance even in animals with an intact reproductive system .
One of the characteristics of the aging process is a decline in cells’ ability to mount cellular stress responses under conditions of perturbed protein homeostasis in the cytosol , in the ER and in the mitochondria [34–36] . The failure to mount these stress responses abrogates the induction of the corresponding chaperones , whose deficiency further perturbs protein homeostasis . We hypothesized that constitutive expression of chaperones may be beneficial for maintaining proteostasis in aged animals , and thus may contribute to their health-span and their longevity . To test this , we assessed the lifespan of irg-7 ( zc6 ) mutants , in which the hsp-4 ER resident chaperone is constitutively expressed in the animals' intestine due to a partially mapped background mutation ( [37] and Fig 1A ) . This strain , which was originally named upr-1 ( zc6 ) , was generated as part of a seminal study by David Ron’s group , to facilitate the identification of mutations in genes required for the induction of this ER resident chaperone [37] . We note that in the absence of information about the zc6 mutation , it was unclear whether these animals induce ER chaperone expression due to deregulation of gene expression or whether the induction is a reflection of ER stress and activation of the unfolded protein response ( UPR ) . Either way , lifespan analysis of the irg-7 ( zc6 ) mutants revealed that they lived up to 30% longer than control animals ( Fig 1B and S1 Table ) . With the exception of a developmental delay of 4 hours , the physiology of irg-7 ( zc6 ) mutants appeared to be similar to that of wild-type animals in terms of pumping rate , fecundity and progeny profile ( S1 Fig ) . If indeed the longevity of irg-7 ( zc6 ) mutants stems from the constitutive expression of ER-resident chaperones , whose expression is regulated by the activation of the unfolded protein response ( UPR ) , then interference with UPR induction should curtail their lifespan . To test this , we introduced mutations in major UPR genes into the irg-7 ( zc6 ) strain and followed their lifespan . First , we introduced the irg-7 ( zc6 ) mutation into pek-1 ( ok275 ) and atf-6 ( ok551 ) mutants ( harboring deletion mutations in the worm homologs of the PERK and ATF6 genes respectively ) . We found that the irg-7 ( zc6 ) mutation extended the lifespan of pek-1 ( ok275 ) and atf-6 ( ok551 ) mutants to a similar extent as it did in animals with an intact UPR ( P>0 . 2 in Cox Regression analysis of 3 independent experiments for each genotype ) ( Fig 1B and S1 Table ) . This implies that neither of these genes is required for the longevity of irg-7 ( zc6 ) mutants . Surprisingly , we could not introduce the irg-7 ( zc6 ) mutation into ire-1 ( ok799 ) and xbp-1 ( tm2457 ) mutants ( harboring deletion mutations in the worm homologs of the IRE1 and XBP1 genes respectively ) . Specifically , we could not detect any homozygous xbp-1 ( -/- ) ; irg-7 ( zc6 ) or ire-1 ( -/- ) ; irg-7 ( zc6 ) double mutants among the viable progeny of heterozygous xbp-1 ( +/- ) or ire-1 ( +/- ) irg-7 ( zc6 ) mutants . This suggests that some activity of the ire-1/xbp-1 pathway is critical for the survival of irg-7 ( zc6 ) mutants . To circumvent this , we examined the longevity of irg-7 ( zc6 ) mutants treated with xbp-1 or ire-1 RNAi , which reduce the levels of their target genes , rather than completely eliminating them . Treatment with xbp-1 or ire-1 RNAi attenuated the expression of the Phsp4::gfp reporter in irg-7 ( zc6 ) mutants and was compatible with the survival of the irg-7 ( zc6 ) strain . Nevertheless , the irg-7 ( zc6 ) mutation still extended the lifespan of ire-1 or xbp-1 RNAi-treated animals to a similar extent as it did in animals with an intact UPR ( P>0 . 3 in Cox Regression analysis of 3 independent experiments ) ( Fig 1B and S1 Table ) . These findings suggest that the longevity of irg-7 ( zc6 ) mutants does not rely on high expression levels of ER resident chaperone . If not due to increased chaperone levels , why are upr-1/irg-7 ( zc6 ) mutants long-lived ? An answer to this enigma may lie in the identity of the irg-7 gene . To associate the irg-7 zc6 mutation with its molecular and genetic identity , we completed the mapping of the irg-7 ( zc6 ) mutation . This mutation was previously characterized as a semi-dominant mutation located on chromosome X [37] . Using one-step whole-genome-sequencing and a SNP mapping strategy [38] we identified several candidate mutations within a mapping interval on chromosome X that may account for the Irg-7 phenotypes . We used RNAi to knockdown these candidate genes and searched for genes whose inactivation phenocopied or suppressed the phenotypes of the irg-7 ( zc6 ) mutants . We found that inactivation of f40f4 . 6 by RNAi suppressed the longevity of irg-7 ( zc6 ) mutants without affecting the lifespan of wild-type animals ( Fig 2A and S2 Table ) . Inactivation of f40f4 . 6 by RNAi also suppressed the expression of the ER stress reporter in irg-7 ( zc6 ) mutants ( Fig 2B ) , suggesting that the mutation in f40f4 . 6 is triggering the UPR in these animals . Since f40f4 . 6 RNAi suppresses rather than phenocopies the phenotypes associated with the irg-7 ( zc6 ) mutation , these experiments suggest that the irg-7 ( zc6 ) mutation is a gain of function ( gof ) mutation in the f40f4 . 6 gene . In some cases gain of function mutations result in hyperactivation of the normal function of the encoded protein , whereas in other cases they may confer new activities . Likewise , over-expression of a protein can enhance its activity as well . Thus , we examined whether over-expression of F40F4 . 6 would extend lifespan similarly to the irg-7 ( zc6 ) gof mutation . To this end , we determined the lifespan of animals carrying a fosmid that includes the f40f4 . 6 gene . We found that these animals lived longer than wild-type animals . In contrast , animals carrying a partially overlapping fosmid that does not include f40f4 . 6 had a normal lifespan ( Fig 2C and S2 Table ) . These experiments suggest that the irg-7 ( zc6 ) mutation is a gain of function mutation that enhances the normal activity of the protein . Not much is known about the irg-7 gene . Microarray studies demonstrated that the expression of irg-7/f40f4 . 6 is induced upon exposure to some pathogens [39] as well as under dietary restriction ( in this study f40f4 . 6 has been referred to as drd-2 [40] ) . In support of these reports , we confirmed that infection by Photorhabdus luminescens subsp Hb and Enterococcus faecalis bacteria ( henceforth referred to as HB and EF bacteria for brevity ) , induces the transcription of a reporter driven by the irg-7 promoter in the animals intestine ( Fig 2D ) . irg-7 encodes a protein of 2214 amino acids with a modular structure . It includes three EGF domains ( that can be used for association with other regulatory proteins ) , a Von Willebrand factor type A domain ( usually involved in adhesion via metal ion-dependent adhesion sites ) and a C-type lectin domain ( CTLD ) ( usually involved in binding to a wide variety of molecules such as sugars , proteins , lipids and inorganic compounds and implicated in pathogen recognition and clearance [41] ) ( Fig 2E ) . In addition , IRG-7 contains a putative signal sequence , but no putative transmembrane domain , and thus is likely to be secreted . Although each of these domains is individually evolutionarily conserved , we failed to identify any human homolog containing this exact combination of domains . Sequencing of the irg-7 gene in animals carrying the zc6 mutation revealed a deletion of 901 base pairs between nucleotides 3241528–3242428 on chromosome X as well as an insertion of 2 nucleotides ( S2 Fig ) . This deletion , which includes exon and intron regions , ultimately removes 221 amino acids from the protein product while preserving its open reading frame . The existence of a transcript encompassing the zc6 mutation was confirmed by RT-PCR and sequencing ( S2 Fig ) . The region included in the deletion mutation encompasses one of the EGF-like domains of the protein ( Fig 2E ) . Since the zc6 mutation is a gof mutation , this indicates that the deleted EGF domain is an inhibitory domain . One of the factors that limit C . elegans lifespan is its pathogenic food [22 , 26] . CTLD-proteins have the capacity to bind carbohydrates that coat pathogens , and are induced upon contact with a variety of fungal and bacterial pathogens [41] . Thus , proteins harboring C-type lectin domains could potentially play a role in pathogen recognition and clearance [41] . Since IRG-7 harbors a CTLD domain , and since its expression is induced upon infection , we wondered whether the lifespan extension of irg-7 gof mutants was associated with an improved ability to deal with pathogens . Hence , we examined whether the extended lifespan of irg-7 gof mutants fed with live OP50 bacteria would persist if the animals were fed with killed bacteria . As previously reported [22 , 26] , we confirmed that the mere feeding of wild-type animals with dead bacteria extends their lifespan ( Fig 3A ) . In contrast , a diet of heat-killed or UV-killed bacteria did not extend the lifespan of irg-7 gof mutants ( Fig 3A and S3 Table ) . This implies that unlike wild-type animals , the lifespan of irg-7 gof mutants is not limited by the viability of its pathogenic food . OP50 E . Coli is only slightly pathogenic to C . elegans . Thus , we next compared the survival of irg-7 gof mutants and wild-type animals exposed to bacteria that is more pathogenic to C . elegans . We focused on HB and EF bacteria , as these pathogenic bacteria induce irg-7 expression ( [41] and Fig 2D ) . We found that the irg-7 gof mutation extended the survival of the animals exposed to HB or EF bacteria ( Fig 3B and S3 Table ) . Furthermore , treatment with irg-7 RNAi during larval development significantly compromised the survival of animals exposed to HB bacteria during adulthood ( Fig 3C and S3 Table ) . This implies that the irg-7 gene normally promotes the survival of wild-type animals exposed to these pathogenic bacteria . Nevertheless , treatment with irg-7 RNAi during larval development did not affect the survival of animals fed with OP50 during adulthood ( Fig 3D and S3 Table ) . The lack of lifespan shortening by irg-7 RNAi in animals fed with OP50 bacteria , as opposed to the curtailed survival of the irg-7 RNAi-treated animals on the HB bacteria , supports the conclusion that the diminished survival is a result of pathogen sensitivity rather than an aging-related phenotype . These results , together with a consideration of its structural molecular domains , implicate irg-7 in pathogen resistance . In principle , the IRG-7 protein could enhance pathogen resistance by directly neutralizing pathogenic bacteria ( for example via its specialized CTLD domain which can directly bind to sugars encoating bacteria and perforate their cell membrane [42] or cause their spatial segregation of microbiota and the host intestine [43] ) . Alternatively , it may do so indirectly , by activating the animals’ immune response . One of the central signaling cascades of the innate immune response in C . elegans is the PMK-1/ATF-7 pathway [20 , 21] . Thus , we examined whether the irg-7 gof mutation recruited this innate immunity pathway to the defense and well-being of the organism . To this end , we followed the expression of a GFP reporter driven by the promoter of T24B8 . 5 , an ATF-7-target gene [20] . We found that the levels of the ATF-7 reporter increased in irg-7 gof mutants compared to animals with wild-type irg-7 ( Mann-Whitney P<0 . 0001 , Fig 4A ) . The level of the ATF-7 reporter in irg-7 gof mutants were the same as in animals exposed to EF ( Mann-Whitney , P = 0 . 98 ) and PA14 pathogens ( Mann-Whitney , P = 1 . 00 ) ( Fig 4A ) . Interestingly , exposure to HB bacteria did not increase the expression of this ATF-7 reporter ( Mann-Whitney , P = 1 . 00 ) ( Fig 4A ) . These findings are consistent with the interpretation that the PMK-1/ATF-7 innate immunity pathway is activated in irg-7 gof mutants . Next , we examined whether the increased expression of the ATF-7 reporter simply correlated with the irg-7 ( zc6 ) mutation or whether it is important for the longevity of irg-7 gof mutants . In support of the latter , we found that the longevity of irg-7 gof mutants on live OP50 was completely dependent upon the integrity of the innate immunity PMK-1/ATF-7 pathway ( Mantel-Cox P>0 . 05 in six independent experiments—Fig 4B , and S4 Table ) . Similarly , the improved survival of irg-7 gof mutants on the pathogenic HB bacteria was also dependent upon the PMK-1/ATF-7 innate immunity pathway ( Mantel-Cox P>0 . 05 in three independent experiments ) ( Fig 4C and S4 Table ) . Altogether , these findings support the conclusion that irg-7 activation acts as an immune-modulator , which engages the nematode's innate immune response and improves the ability of animals to cope with the pathogenic food . One consequence of activation of the innate immune response is an increased load on the ER , presumably due to massive production of secreted antibacterial proteins [44] . Since the ER stress response is constitutively activated in the intestine of irg-7 gof mutants ( Fig 1A , [37] ) , we wondered whether this stress response may be due to the activity of the PMK-1/ATF-7 innate immunity pathway in these mutants . Consistent with this possibility , inactivation of pmk-1 or atf-7 reduced the levels of the Phsp-4::gfp ER-stress reporter in the intestine of the irg-7 gof animals ( Student’s T-test P<0 . 002 , Fig 4D ) . Thus , the PMK-1/ATF-7 innate immunity pathway appears to contribute to the induction of the UPR in irg-7 gof mutants . Many longevity pathways confer increased pathogen resistance [6 , 27–30] . Hence we wondered whether the irg-7 gene is an integral part of any of the known longevity pathways . To this end , we compared the lifespan of wild-type animals and various long-lived mutants upon treatment with control or irg-7 RNAi . Using this approach , we found that irg-7 RNAi suppressed the extended lifespan of glp-1 mutants which lack a germline ( Mantel-Cox P<0 . 0001 in 3 independent experiments , Fig 5 and S5 Table ) , although it had no effect on the lifespan of wild-type animals ( Mantel-Cox P>0 . 2 in 4 independent experiments , Fig 5 and S5 Table ) . irg-7 RNAi did not affect the lifespan of animals that are long-lived due to reduced insulin/IGF1 signaling ( i . e . daf-2 mutants ) , due to reduced mitochondrial respiration ( clk-1 mutants ) or due to a mutation in the eat-2 gene ( a mutation in an acetyl-choline receptor that causes reduced pharyngeal pumping and extended longevity ) ( Mantel-Cox P>0 . 2 in 3 independent experiments for each long-lived strain , Fig 5 and S5 Table ) . The observation that irg-7 inactivation had only mild effects on wild-type lifespan , but almost completely prevented germline loss from extending lifespan , suggests that irg-7 plays a regulatory role specifically in this pathway . After implicating irg-7 in the germline longevity pathway , we wondered whether its expression is regulated by this pathway . To this end , we examined whether the expression levels or expression pattern of irg-7 changes upon germ cell removal . Specifically , we compared the expression of the irg-7 transcription reporter in animals with a normal reproductive system and in animals with no germ cells . In both cases , the reporter driven by the irg-7 promoter was expressed specifically in the posterior intestinal cells , regardless of the presence or absence of the germ cells . However , the fluorescence levels of the transcriptional reporter increased upon germ cell removal ( Student’s T-test P<0 . 005 , Fig 6A ) . This indicates that irg-7 is associated with the germline longevity pathway in at least two ways . First , it is required for the longevity of germlineless animals . Second , its' transcription is regulated by the reproductive tissues . Interestingly , a similar increase in the levels of the transcriptional reporter was also observed in irg-7 ( zc6 ) mutants , suggesting that irg-7 activation indirectly promotes its own transcription as part of an auto-regulatory loop ( Student’s T-test P<0 . 005 , Fig 6A ) . Since the germline longevity pathway engages many transcription factors , and irg-7 is a new factor in the germline longevity pathway , we examined whether the transcriptional network that promotes longevity in germlineless animals is affected by irg-7 depletion . To this end , we compared the induction of a set of reporters known to be induced by different transcription factors upon germ cell removal . We found that the fat-6 and lgg-1 reporters , target genes of the NHR-80 and PHA-4 transcription factors respectively [10 , 15] , were induced upon germ cell removal regardless of treatment with irg-7 RNAi ( Fig 6B ) . In contrast , irg-7 inactivation compromised the induction of the reporters of sod-3 ( a target gene of DAF-16 ) [45] and cdr-6 ( a target gene of DAF-16 and DAF-12 [46 , 47] ) upon germ cell removal ( Fig 6B ) . Since irg-7 inactivation interfered with the induction of two daf-16 target genes in germlineless animals we hypothesized that irg-7 might be required for the activation of this transcription factor upon germ cell removal . A critical step in the activation of DAF-16 upon germ cell removal is its nuclear translocation in the intestine cells [45] . Thus , we used animals expressing fluorescently tagged DAF-16 to explore whether irg-7 was required for this step . We could not detect translocation of DAF-16::GFP to the nuclei of the intestine cells upon germ cell removal in animals treated with irg-7 RNAi , ( Fig 6C ) . The translocation of DAF-16 from the cytoplasm to the nucleus can be triggered by a variety of signals [48] . Hence , we wondered whether irg-7 inactivation globally abrogates DAF-16’s ability to translocate from the cytoplasm to the nucleus , or is it only impaired in response to a subset of signals ( as in the case of germline depletion ) . To this end , we followed the affect of irg-7 RNAi on DAF-16’s intra-cellular localization in animals with reduced insulin/IGF1 signaling . We confirmed that when insulin signaling is perturbed , DAF-16::GFP translocates to the nuclei of all cells . However , in contrast to its interference with DAF-16::GFP’s translocation to the nuclei in germlineless animals , irg-7 RNAi did not affect the accumulation of DAF-16::GFP in the nuclei of insulin/IGF-1 signaling mutants ( Fig 6C ) . This indicates that irg-7 is specifically required for the nuclear translocation of DAF-16 upon germ cell removal . Inactivation of irg-7 also suppressed the induction of the cdr-6 reporter ( Fig 6B ) . The transcription of the cdr-6 reporter is independently and additively induced upon germ cell removal by either the DAF-16 or the DAF-12 transcription factors [47] . In order to uncouple between the two , we examined the expression of the cdr-6 reporter in daf-16-deficient animals , which can still induce cdr-6 expression upon germ cell removal in a daf-12-dependent manner . We found that irg-7 RNAi suppressed the induction of the cdr-6 reporter upon germ cell removal in this background as well ( Fig 6B ) . This indicates that irg-7 is an important regulator of the germline longevity pathway , which alters the transcriptional outputs of the DAF-16 and DAF-12 transcription factors upon germ cell removal . After establishing that irg-7 gof mutants are long-lived , we explored which known lifespan-related genes are required for their longevity . Specifically , we used RNAi or deletion mutants to inactivate the expression of known longevity genes and then examined the ability of the irg-7 mutants to live long . We found that the increased lifespan of irg-7 gof mutants required most of the genes previously associated with the reproductive longevity pathway ( Fig 7A and S6 Table ) . Some of these genes are uniquely associated with the reproductive longevity pathway ( tcer-1 [49] , kri-1 [50] , daf-9 , daf-12 [5] ) . Others are common to the reproductive longevity pathway and other longevity pathways ( e . g . skn-1 [17 , 51 , 52] , daf-16 [5 , 45 , 53 , 54] , hsf-1[2 , 55] , aak-2 [56] and pha-4 [10 , 57] which are also a part of the dietary restriction and/or insulin/IGF1 signaling longevity pathways ) . Silencing of additional genes , required for insulin/IGF-1-induced longevity but not required for the longevity of germline-less animals ( phi-3 , skr-1 [13] ) , did not shorten the longevity of irg-7 mutants by more than 5% ( Fig 7A and S6 Table ) . Likewise , silencing of genes required uniquely for the mitochondrial longevity pathway ( cdr-2 , ubl-5 [58 , 59] ) , and/or the caloric-restriction longevity pathway ( wwp-1 [60] ) , did not shorten the longevity of irg-7 gof mutants by more than 5% ( Fig 7A and S6 Table ) . Thus , irg-7-mediated longevity appears to be related to the longevity mediated by germ cell removal , as the two rely on many common genes . Lifespan extension by germ cell removal requires the presence of the somatic gonad [5 , 47] . Thus , we examined whether lifespan extension by the irg-7 gof mutation requires the somatic gonad as well . We found that instead of extending lifespan , the irg-7 gof mutation shortened lifespan in the absence of a somatic gonad ( Fig 7B and S6 Table ) . Thus , irg-7 activation , similarly to germ cell removal , relies on the somatic gonad to promote longevity . Finally , we examined the effect of the irg-7 ( gof ) mutation on the longevity of animals that lack germ cells but have a somatic gonad . We found that the irg-7 gof mutation did not increase the lifespan of germlineless animals ( Fig 7C and S6 Table ) . This further supports the notion that the irg-7 activation and germ cell removal increase lifespan by a common mechanism , which relies on many common genes and on the somatic gonad . Since many transcription factors are required for the longevity of germlineless animals as well as for the longevity of irg-7 ( zc6 ) gof mutants ( Fig 7A ) , we asked whether the irg-7 gof mutation activates these transcription factors . Specifically , we focused on the DAF-16 and DAF-12 transcription factors , as irg-7 was required for their activity in germlineless animals ( Fig 6B and 6C ) . First , we followed the levels of a cdr-6 transcriptional reporter , which is transcribed both by DAF-16 and by DAF-12 . We found that the irg-7 gof mutation increased the level of the reporter in animals with an intact reproductive system by 1 . 5 fold ( Student’s T-test P<0 . 0001 , Fig 8A ) . This increase was dependent on the daf-12 transcription factor , as it was not observed in its absence ( Student’s T-test P = 0 . 075 , Fig 8B ) . Thus , irg-7 gof activates the longevity-associated transcription factor DAF-12 . Next , we followed the levels of a sod-3 transcriptional reporter , which is a direct target gene of the DAF-16 transcription factor . Surprisingly , we did not detect an increase in the level of the sod-3 reporter in irg-7 gof mutants ( Student’s T-test P = 0 . 067 , Fig 8C ) . Likewise , we did not observe accumulation of DAF-16::GFP in the nuclei of irg-7 ( zc6 ) animals ( Fig 8D ) . This could indicate that irg-7 gof is not sufficient to promote the nuclear translocation of the DAF-16 transcription factor to the nucleus , nor is it sufficient to significantly increase DAF-16 activity . Nevertheless , DAF-16 was required for the lifespan extension in irg-7 gof mutants ( Fig 7A ) . Furthermore , DAF-16 was required for the improved survival of irg-7 gof mutants exposed to pathogenic HB bacteria ( Fig 3E and S3 Table ) . These suggest that the basal activity of DAF-16 , which is below the sensitivity of this assay , but is not completely absent , is critical for the beneficial phenotypes of irg-7 ( zc6 ) gof mutants . Alternatively , it could be that the transcriptional targets of DAF-16 in animals with activated irg-7 are distinct from those examined here . We found that irg-7 is implicated in pathogen resistance on one hand , and is an integral component of the reproductive longevity pathway on the other hand . Hence , we wondered whether these two processes are linked . Therefore , we asked whether germline physiology is modulated in the presence of pathogens . To this end , we examined whether germline homeostasis is modulated by the pathogenic HB bacteria . We found that exposure to HB bacteria increased germ cell apoptosis ( Mann-Whitney P<0 . 0001 , Fig 9A ) and decreased the number of mitotic germ cells ( Mann-Whitney P<0 . 0001 , Fig 9B ) in the gonads of wild-type animals . To see if these changes in germline homeostasis contributed to the resistance of the animals to the pathogen , we decided to block some of these changes and to see if this increases the sensitivity of the animals to the pathogen . To this end , we used egl-1 mutants , which block germline apoptosis induced by Salmonella typhimurium [61] . As in the case of Salmonella typhimurium-induced germline apoptosis [61] , a mutation in the egl-1 gene blocked HB bacteria-induced germline apoptosis ( Mann-Whitney P = 1 . 00 , Fig 9A ) and increased the sensitivity of the animals to pathogen exposure ( Mantel-Cox P<0 . 02 in 3 independent experiments , Fig 9C and S7 Table ) . These findings raise the possibility that the germline may serve as a tissue-level pathogen sensor whose depletion can promote animals’ survival in the presence of pathogens . Exposure to pathogenic bacteria increases the expression of the irg-7 gene and alters germline homeostasis . Hence , we examined whether these events occur consecutively such that pathogen exposure leads to activation of irg-7 , which in turn affects germline homeostasis . To this end , we examined whether activation of irg-7 is sufficient to affect germline homeostasis . irg-7 activation was achieved by introducing the irg-7 gof mutation or by over-expressing multiple copies of a fosmid that includes the wild-type irg-7 gene in animals fed with OP50 bacteria . We found that the levels of germ cell apoptosis and the amount of mitotic germline in these animals were similar to those of wild-type animals ( Mann-Whitney P>0 . 3 , Fig 9A and 9B ) . This indicates that under these conditions , germline homeostasis is not regulated by irg-7 activation . This once again indicates that irg-7 does not mediate its beneficial effects on OP50-fed animals by acting upstream to the germline . Rather , since irg-7 expression is induced upon germ cell removal , it should be placed downstream to the perturbations in germline homeostasis . In addition to the beneficial lifespan affects conferred by irg-7 activation in animals fed with OP50 bacteria , irg-7 activation also improves the survival of animals on pathogenic bacteria . Hence , we decided to examine whether under these conditions ( upon exposure to pathogenic bacteria ) irg-7 activation leads to perturbations in germline homeostasis . We found that that even in the presence of HB bacteria , no increase in germline apoptosis or reduction in the amount of mitotic germ cells was observed in animals with activated irg-7 . Unexpectedly , the mitotic germline of animals with activated irg-7 exposed to HB bacteria was even more expanded than prior to their exposure to pathogenic bacteria ( Mann-Whitney P<0 . 0001 , Fig 9B ) . Likewise , barely any apoptotic germ cells were detected in the gonads of animals with activated irg-7 upon treatment with HB bacteria ( Fig 9A ) . Whereas the reasons for the robustness of germline homeostasis in irg-7 mutants is not known , an appealing speculation is that irg-7 activation successfully recruited the innate immune response to protect the animals from the pathogenic food and its toxic interaction with the animals’ germline .
Previous studies have demonstrated that signals from reproductive tissues influence longevity , yet only a fraction of the underlying genetic network that controls this process has been elucidated . In this study , we identified irg-7 , as an integral downstream component of the germline longevity pathway . irg-7 encodes an intestinally-produced secreted protein , harboring a single CTLD domain ( typically associated with proteins involved in innate immunity anti-microbial activity ) and several EGF domains . Interestingly , removal of one of these EGF domains activates the protein , assigning it as an auto-inhibitory domain . This auto-inhibition of IRG-7 is reminiscent of toxic mammalian CTLD-harboring proteins , which contain intramolecular inhibitory domains that maintain them in an inert inactive state , which can be switched into an active state by proteolytic removal of the inhibitory segment [42 , 62 , 63] . Our initial interest in irg-7 ( zc6 ) mutants was because of the constitutively high expression of the hsp-4 ER resident chaperone in their intestine . At first , it was unclear whether these animals induce ER chaperone expression due to deregulation of gene expression or as a reflection of ER stress and activation of the UPR . Our findings indicate that the hyperactivation of the ER stress response in these animals is a reflection and a consequence of their activated innate immune response . The UPR-mediated coordination between the secretory capacity and secretory load is critical in animals with an activated immune response . In the absence of such coordination , activation of the innate immune response is detrimental [44] . Consistent with this , complete depletion of the UPR genes ire-1 and/or xbp-1 using deletion mutants was lethal for irg-7 gof mutants , whose innate immune response is activated . Nevertheless , irg-7 gof mutants survived ire-1 and/or xbp-1 RNAi treatment . Furthermore , these RNAi treatments did not compromise their longevity . This dichotomy may indicate that a low level of UPR activity is sufficient to provide the basic adaptation required by the over-loaded ER in animals with an activated innate immune response . This residual UPR activity may reflect low levels of UPR activity throughout the animal . Alternatively , it may reflect UPR activity in specific tissues , most likely in neurons , which are relatively resistant to RNAi treatment . This mechanism is consistent with previous findings that neuronal signaling can modulate the innate immune response in the C . elegans intestine [64 , 65] . We consider irg-7 as an integral downstream component of the germline longevity pathway because its expression is increased upon germ cell depletion and because it is essential for the longevity of germlineless animals . In germlineless animals , depletion of irg-7 interferes with the activation of the longevity-promoting transcription factors DAF-16 and DAF-12 . At the same time , irg-7 deficiency does not affect germ cell regulated activity of the transcription factors NHR-80 and PHA-4 . Thus , at least two independent signaling pathways control the transcriptional network that is set upon germ cell depletion; only one of which implicates irg-7 , which acts upstream or in parallel to daf-16 and daf-12 . These perturbations are enough to preclude lifespan extension by germ cell depletion . Although inactivation of irg-7 does not shorten the lifespan of animals with an intact reproductive system , its’ activation extends the lifespan of the animals . The lifespan extension induced by irg-7 activation requires the presence of the somatic gonad , similarly to the longevity conferred by germ cell removal . In addition , it shares many genetic determinants with the reproductive longevity pathway . Although there is an extensive overlap between genes required for the lifespan extension induced by irg-7 activation and those required for longevity induced by germ cell depletion , the lifespan extension induced by irg-7 activation does not perturb germline homeostasis under normal growth conditions . Thus , the lifespan extension by irg-7 activation is independent of germline depletion , yet requires many of the same longevity genes as in germlineless animals . How does irg-7 activation extend lifespan ? Many CTLD proteins , which are abundant in the nematode genome ( ~280 genes ) , show a pathogen-specific induction during infection and are thought to be important for enhancing the nematode's ability to fight off pathogens [66 , 67] . Furthermore , the expression of several CTLD proteins is regulated by innate immunity pathways [6 , 33 , 67 , 68] . Our data implicates irg-7 in regulating the animals’ sensitivity to pathogens in several ways . First , although the lifespan of wild-type animals is limited by its pathogenic food , in irg-7 ( zc6 ) mutants , lifespan was unaltered whether the animals were fed with live or with dead OP50 bacteria . Second , irg-7 activation promotes the survival of animals on pathogenic HB bacteria , whereas its inactivation renders wild-type animals more sensitive to the same pathogen . Third , irg-7 activates and engages the pmk-1/atf-7 innate immune response pathway to the defense and well-being of the animals and requires this innate immunity pathway for lifespan extension . Altogether , these imply that irg-7-associated life extension is tightly associated with its enhanced ability to defend the animals from the pathogenic toxicity of its food . Since irg-7 is induced in germlineless animals , and since its induction can lead to activation of the PMK-1/ATF-7 innate immunity pathway , our findings suggest that the canonical PMK-1/ATF-7 pathway may also be activated in germlineless animals . Consistent with this , an increase of 1 . 3 fold in the transcript levels of several atf-7 target genes has been observed between germlineless animals and animals with an intact reproductive system [17] . However , the transcript levels of the same target genes were reported unaltered upon germ cell removal in a previous study [6] . This dichotomy may be due to different normalization approaches taken by the two studies . Whereas both studies normalized transcript levels to a major housekeeping gene , the first study also took into account the differential number of cells composing animals with an intact gonad and those that lack a germline , whereas the second study did not . Our finding of increased expression of the ATF-7 target gene T24B8 . 5 , obtained by a transcriptional reporter ( and thus insensitive to the presence/absence of the germline in terms of normalization ) yielded a result consistent with the analysis of Steinbaugh et . al . [17] , and suggests that the PMK-1/ATF-7 pathway may also be activated in germlineless animals and may contribute to their immunity . Nevertheless , this pathway may be less critical for the survival of germlineless animals , which can enhance their immune response via alternative transcription factors [6] . In addition to the innate immunity genes pmk-1 and atf-7 , the lifespan extension of irg-7 gof mutants fed with live OP50 depends on the presence of the somatic gonad as well as on a variety of genes required for longevity induced by germ cell removal . This implicates the same systemic endocrine pathways that increase C . elegans lifespan upon germ cell removal in promoting the resistance of animals to pathogenic challenges . This conclusion is consistent with the expression/activation of many of the genes implicated in the reproductive longevity pathway specifically in the intestine , the same site where the host usually encounters the pathogenic bacteria . Furthermore , in addition to their contribution to germ cell-regulated longevity , both the somatic gonad and the transcription factor DAF-16 have been implicated in the enhanced resistance of germlineless animals to pathogens [6 , 69] . Our findings now implicate additional germline-regulated genes such as tcer-1 , kri-1 , daf-12 , skn-1 and pha-4 in this innate immune response pathway . Interestingly , a bile acid biosynthetic pathway has already been implicated in the systemic communication of cellular stress and activation of the MAP kinase innate immunity pathway by stressed germline [70] . Interestingly , one of the genes required for the longevity of germlineless animals did not affect the longevity of irg-7 gof mutants ( Fig 7A ) . This gene encodes nhr-80 , an intestinal transcription factor that transcribes lipid homeostasis-related genes [15] . This differential requirement of nhr-80 is intriguing and may suggest that nhr-80 is specifically required for the well-being of germlineless animals rather than an integral part of the reproductive longevity pathway . Nevertheless , this differential requirement of nhr-80 is important as it genetically rules out the possibility that the irg-7 gene acts upstream to the germ cells themselves , promoting lifespan extension by limiting the amount of germ cells in the animals . This conclusion is also supported by the fact that in irg-7 gof mutants , germline homeostasis and progeny profiles are similar between wild-type animals and irg-7 gof mutants ( S1 Fig and Fig 9 ) . With the exception of nhr-80 , the implication of a significant amount of components of the reproductive longevity pathway in pathogen resistance and innate immunity is thought provoking , as it raises the possibility that the reproductive system can be used as a signaling center to divert resources towards defending against putative pathogen attacks ( See model in Fig 10 ) . Accordingly , perturbations to germ cell homeostasis ( executed via regulation of germline proliferation and/or germline apoptosis ) may serve as a surveillance center , putatively disrupted by pathogens and their toxins . Such a mode of action is consistent with recent studies that indicate that pathogen surveillance can be achieved indirectly by monitoring internal physiological cues that may be altered by pathogens and their toxins [71–74]; effectively diluting the germline . Accordingly , stress-induced perturbation in germ cell homeostasis can trigger a somatic defense response , including an innate immune response [70 , 75] . Consistent with this model , exposure to some pathogens ( as previously demonstrated in the case of Salmonella typhimurium infection [61] and as shown here for HB luminescence infection ) perturbs germ cell homeostasis . In turn , germ cell depletion can activate innate immunity pathways via the expression of the innate immunity-promoting gene irg-7 ( Fig 6A ) , which in turn can activate the PMK-1/ATF-7 innate immunity pathway ( Fig 4A ) . Likewise , germ cell depletion can promote innate immunity pathways by other means ( i . e activation of the transcription factors DAF-16 and SKN-1 in the animals intestine , which in turn induce the expression of a variety of innate immunity-related genes [6 , 17] . Finally , in further support of this model , the ability to relay a signal of distress via perturbations in germline homeostasis appears to be important for mounting an effective immune response , as animals are less likely to survive some infections when germline apoptosis is blocked [61 and Fig 9] . One of the hallmarks of aging both in nematodes and in humans , is a progressively increased sensitivity to external pathogens , reflecting a failure of the immune response in the old . At least in C . elegans , the decline in the ability to combat pathogens is detrimental to the animals and limits their lifespan [76 , 77] . Accordingly , it is not surprising that the same manipulations in the reproductive system that promote longevity also affect the animals' innate immune response and increase their resistance to pathogens [6 , 29 , 30] . However , it is unlikely that the lifespan extension induced by germ cell removal is only the reflection of improved innate immunity . This is because germ cell removal provides additional benefits for the soma and the organism in addition to improving the animals' resistance to pathogens . For example , germ cell depletion maintains proteasome activity [9] , prolongs the responsiveness of multiple stress response pathways with age [8 , 35] and confers resistance to multiple cellular stresses [7 , 78] . Accordingly , the increase in lifespan achieved by the irg-7 gof mutation and by feeding the animals with dead bacteria are not as big as that produced by germ-cell loss . Furthermore , irg-7 gof only partially phenocopies the transcriptional reprogramming that normally occurs upon germ cell removal . Thus , improved innate immunity is only one of several benefits of animals whose germ cells are depleted . Nevertheless , finding new ways to maintain the immune response in aging animals that undergo immunosensesnce is an important goal that may postpone major lifespan limiting events . Although the immune response in mammals is primarily adaptive in its nature , a basal innate immune response , based on a variety of antiseptic proteins including CTLD proteins , similar to those of the nematode , contributes to the immunity of mammals as well . Thus , better understanding of the molecular events that mediate this basic immune response , in the young and in the old , is important and holds great promise for human health .
L4 stage N2 or irg-7 ( zc6 ) worms were incubated at 20°C and transferred to fresh plates twice a day until they stopped producing progeny . Worms that crawled off the plates , bagged or ruptured were removed from the data set . All progeny plates were incubated at 20°C for 2 additional days and the number of worms that developed was determined . The irg-7 promoter sequence ( 690 bp ) was amplified from genomic DNA and cloned into the HindIII and XbaI sites , replacing a myo-3 promoter in the pCF191plasmid ( previously described in [79] ) . Germline transformations were performed by injection of 50ng/μl plasmid with 15 ng/μl of Punc-54::gfp as a co-transformation marker . To follow expression of fluorescent proteins , transgenic animals were anaesthetized on 2% agarose pads containing 2mM levamisol . Images were taken with a CCD digital camera using a Nikon 90i fluorescence microscope . For each trial , exposure time was calibrated to minimize the number of saturated pixels and was kept constant throughout the experiment . Nikon NIS element software was used to quantify mean fluorescence intensity in the selected area encompassing whole worms . RNAi treatments were performed continuously from the time of hatching unless indicated otherwise . In general , eggs were placed on plates seeded with the RNAi bacteria of interest . Lifespan of 90 animals per strain were scored every 1–2 days . Related lifespans were performed concurrently to minimize variability . In all experiments , lifespan was scored as of the L4 stage which was set as t = 0 . Animals that ruptured or crawled off the plates were included in the lifespan analysis as censored worms . For lifespan on dead bacteria , bacteria was killed by 30 minutes of boiling prior to seeding or by UV irradiation . The statistical program SPSS was used to determine the means and the P values calculated using the log-rank ( Mantel-Cox ) method . For quantifying the amounts of mitotic germ cells , gonads of day-1 adults were dissected , fixed and stained with DAPI as previously described [80] . In short , day-1 worms were transferred to unseeded ( without bacteria ) plates and then placed in 20 μl PBSx1 with 0 . 2mM Levamisole for immobilization . Once the worms were paralyzed , animals were decapitated with a needle to allow the removal of the gonads . The dissected gonads were fixed with 10% formaldehyde for 30 minutes . The fixed gonads were then washed twice in M9 and stained with 1 μg/mL DAPI ( 4' , 6-diamidino-2-phenylindole ) solution for 20 minutes . Worms were washed two times with PBSTx1 and observed under the fluorescent microscope . The boundary of the region of mitotic cells was defined by the most distal row of cells containing nuclei with crescent-shaped DAPI morphology . The amount of mitotic nuclei was scored in sequential focal planes through the width of the germline . The number of apoptotic cells in the gonads of day-2 animals was assessed by scoring the number of SYTO12 labeled cells in the gonad . SYTO12 ( Molecular Probes ) staining was performed as previously described [81] . In short , to obtain an estimate of the relative numbers of corpses in different genetic backgrounds , 2-day adult animals were stained with SYTO 12 ( Molecular Probes , Eugene , OR ) , a vital dye that preferentially stains apoptotic germ cells . Animals were stained by incubating them in a 33 μM aqueous solution of SYTO 12 supplemented with OP50 for 4–5 hours at 25°C . Animals were transferred back to new seeded plates to allow stained bacteria to be purged from the gut . After 30–60 minutes , animals were mounted on agarose pads and inspected using a Nikon eclipse 90i , equipped with standard epifluorescence filters and Nomarski optics . Only animals that stained brightly were scored . E . faecalis OG1RF strain ( ATCC 47077 ) was grown at 37 degrees in brain heart infusion ( BHI ) medium supplemented with Gentamicin ( 25–50μg/ml ) . Photorhabdus luminescens subsp . L strain ( ATCC 29999 ) was grown at 30 degrees for 48 hours in Nutrient Broth medium ( BD cat 234000 ) . Pseudomonas aeruginosa was the clinical isolate PA14 strain . Bacteria were seeded on nematode growth medium ( NGM ) plates seeded with bacteria and incubated overnight at room temperature , with the exception of the fast killing experiment by E . faecalis where the bacteria was seeded on brain heart infusion ( BHI ) agar plates . Survival assays were initiated with 80-90nematodes in the late L4/early day 1 stage , grown from eggs on plates seeded with OP50 bacteria or RNAi bacteria of interest . Survival was scored daily thereafter . The SPSS program was used to determine the means and the P values . P values were calculated using the log-rank ( Mantel-Cox ) method . Error bars represent the standard error of the mean ( SEM ) of independent biological replicates unless indicated otherwise . For a simple comparison between two data sets , P values were determined using unpaired Student’s T-test , assuming unequal variances . For multiple comparisons , between multiple data sets , samples were analyzed by the Kruskal-Wallis method , followed by a post hoc analysis using the Mann-Whitney method with Bonferroni correction for multiple comparisons . The analysis was performed in R . To compare pairs of lifespan and pathogen survival analysis ( i . e a total of two genotypes ) , the Kaplan-Meier method was used to estimate survival as a function of time , and survival differences were analyzed by the Mantel-Cox log-rank test using the SPSS program . To compare survival differences between two pairs of genotypes ( i . e . a total set of four genotypes ) , Cox regression analysis of several genotype factors was performed to compare hazard ratios between genotypes . Contrasts were used to determine whether the ( log ) hazard ratio was significantly different under different genotypes . All performed regressions were corrected using the Bonferroni correction for multiple tests . The analysis was performed in R . This analysis is limited to samples with a constant proportional hazard ratio over time . Information provided in S8 Table . | Increased sensitivity to pathogens is one of the hallmarks of aging . Thus , pathways that slow down the aging process should provide a remedy to this challenge . In this study , we have used the model organism C . elegans to identify a new longevity gene whose activation improves the resistance of the animals to pathogenic food and extends the lifespan of the animals . This improved resistance is the result of activation of the innate immune response and requires many genes known to promote longevity in animals whose germline has been removed . Nevertheless , this improved survival in the presence of pathogenic food is achieved without depleting the germline of the animals . This suggests that these longevity-promoting genes , many of which are conserved between worms and humans , can also protect animals with an intact gonad by diverting resources towards defending against putative pathogen attacks . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"rna",
"interference",
"gonads",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"regulatory",
"proteins",
"immunology",
"microbiology",
"dna-binding",
"proteins",
"germ",
"cells",
"physiological",
"processes",
"homeostasis",
"transcription",
"factors",
"epigenetics",
"bacterial",
"pathogens",
"animal",
"cells",
"genetic",
"interference",
"medical",
"microbiology",
"proteins",
"gene",
"expression",
"microbial",
"pathogens",
"immune",
"response",
"immune",
"system",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"genital",
"anatomy"
] | 2017 | Innate immunity mediated longevity and longevity induced by germ cell removal converge on the C-type lectin domain protein IRG-7 |
Little is known about how the size of meristem cells is regulated and whether it participates in the control of meristem size in plants . Here , we report our findings on shoebox ( shb ) , a mild gibberellin ( GA ) deficient rice mutant that has a short root meristem size . Quantitative analysis of cortical cell length and number indicates that shb has shorter , rather than fewer , cells in the root meristem until around the fifth day after sowing , from which the number of cortical cells is also reduced . These defects can be either corrected by exogenous application of bioactive GA or induced in wild-type roots by a dose-dependent inhibitory effect of paclobutrazol on GA biosynthesis , suggesting that GA deficiency is the primary cause of shb mutant phenotypes . SHB encodes an AP2/ERF transcription factor that directly activates transcription of the GA biosynthesis gene KS1 . Thus , root meristem size in rice is modulated by SHB-mediated GA biosynthesis that regulates the elongation and proliferation of meristem cells in a developmental stage-specific manner .
The size of a plant , or part thereof , is determined by combined activity of cell proliferation and growth during development [1] . Cell proliferation in plants occurs mostly in specialized tissues known as meristems , where new cells are produced to ensure that plants continue to grow in height and width throughout their life . Prior to mitosis , cells in the meristem must double in size by undergoing a slow but steady expansion in the direction perpendicular to the previous division plane , which enables them to divide and keeps the size of their daughter cells constant [2 , 3] . A more pronounced growth ( denoted as post-mitotic cell expansion ) , however , is commonly seen in differentiating cells that are displaced from the meristem . The extent of post-mitotic cell expansion is generally well correlated with the magnitude of organ growth [4] . Cell proliferation and growth in plants are influenced by genetic , hormonal , and environmental inputs . While little is known about the molecular mechanisms that regulate the size of meristem cells , numerous molecular players , including members of the AP2/ERF family of transcription factors , have been demonstrated to control either cell proliferation or post-mitotic cell expansion . For instance , the Arabidopsis AP2 transcription factor AINTEGUMENTA ( ANT ) promotes cell proliferation by maintaining the meristematic competence of cells [5] . ANT activity is activated by ARGOS ( for auxin-regulated gene involved in organ size ) , a novel transcription factor acting downstream of auxin signaling [6] . In rice , several AP2/ERF genes including OsEATB ( for ERF protein associated with tillering and branching [7] , SUBMERGENCE 1A ( SUB1A ) [8] , SNORKEL1 ( SK1 ) and SK2 [9] , were reported to have roles in regulating internode elongation , which is primarily post-mitotic expansion of differentiating cells displaced from the intercalary meristem near the node . SK1 and SK2 were suggested to trigger internode elongation via GA in response to rising water level [9] . By contrast , OsEATB was found to restrict GA responsiveness during the internode elongation process by down-regulating the expression of the GA biosynthetic gene OsCPS2 [7]; whereas SUB1A limits GA responsiveness during prolonged submergence by augmenting accumulation of the DELLA family of GA signaling repressors SLENDER RICE 1 ( SLR1 ) and SLR1 Like 1 ( SLRL1 ) , thus restricting underwater internode elongation and enhancing submergence survival [10] . GA plays an important role in the regulation of cell proliferation and growth during plant development [11–13] . It has been recently established that GA modulates both the rate of cell proliferation and the extent of post-mitotic cell expansion [3 , 14–16] . Inhibition of GA biosynthesis , either genetically in the GA biosynthesis mutant ga1-3 , or by means of chemical treatment using paclobutrazol ( PAC ) , an inhibitor of GA biosynthesis [17 , 18] , reduces substantially the rate of cell proliferation in the Arabidopsis root meristem [3 , 14 , 15] . GA was proposed to promote root growth in Arabidopsis by increasing elongation ( expansion along the root axis ) of both dividing and post-mitotic endodermal cells , thereby indirectly controlling division and elongation of other types of root cells and the overall root meristem size [3] . However , how this process is regulated at the molecular level remains unclear . Here we report the discovery of a novel GA-dependent size-control mechanism in the rice root meristem . We show that root meristem size in rice can be regulated by the extent of cell elongation in the root meristem . SHOEBOX ( SHB ) , an AP2/ERF transcription factor , plays a key role in this mechanism . SHB directly binds to and activates transcription of the GA biosynthesis gene KS1 in the root meristem , leading to the local production of GA that promotes elongation of meristem cells following germination , thus ensuring meristem growth and phenotypic plasticity during early stage of meristem development . At a later stage , SHB-dependent and KS1-mediated GA biosynthesis also participates in the modulation of cell proliferation in the root meristem , indicating a developmental stage-specific function of SHB .
In a rice enhancer trap screen we isolated a recessive mutant with a short primary root phenotype ( Fig 1A ) , which we have named shoebox ( shb; based on the shape of the cortical cells in the root meristem ) . Analysis of median longitudinal sections of root apices of 4-day-old wild-type ( WT ) and shb seedlings showed that the root meristem size of shb was shorter than that of the WT ( Fig 1B and 1D and 1H ) . Quantification of cortical cell number and size in the root meristem of WT and shb mutant plants suggested that this was not due to a reduction in the number of meristematic cortical cells ( Fig 1H ) , but was rather caused by a decrease in the length ( but not width ) of meristematic cortical cells ( Fig 1C and 1E and 1I ) . Consistently , EdU staining indicated that the shb mutation did not noticeably alter cell proliferation in the root meristem ( Fig 1F and 1G ) . Moreover , the average lengths of cortical cells in the root elongation and maturation zone did not differ between shb and the WT ( Fig 1J and 1K ) , suggesting that shb has a root meristem-specific cell elongation defect . Notably , root growth rate and cell production rate in shb were not significantly altered in 3- and 4-day-old shb mutants but started to decline at around 5 days after sowing ( Fig 1L and 1M ) . The aerial part of shb mutant plants has typical characteristics of rice GA-deficient or insensitive mutants [7 , 19 , 20] , such as dwarfism and short internode length ( S1 Fig ) . We thus hypothesized that the root phenotype of shb mutant plants might be caused by a defect in GA biosynthesis and/or signaling and examined whether it could be restored to WT by growing the mutants on medium supplemented with bioactive GA ( GA3 ) . 10 μM GA3 had no apparent effect on the WT control but could fully rescue the short-root phenotype of shb mutants ( Fig 2A and 2B and 2G ) . Average length of cortical cells in the root meristem of shb was restored to that of the WT ( Fig 2D and 2F and 2H ) , producing a root meristem with a similar size to that of the WT ( Fig 2C , 2E and 2I ) . These results suggest that shb could properly respond to GA and GA deficiency is the primary cause of mutant phenotypes . In agreement with this suggestion , we found that the levels of GAs were reduced in shb roots as compared to the WT roots ( Fig 2J ) . Particularly two bioactive GAs , GA3 and GA4 , were significantly lower in shb compared to the WT controls ( Fig 2J ) . Because GA-deficient mutants often show delayed seed germination [21 , 22] . We next compared WT and shb seed germination and found that shb germinated approximately 12 h later than the WT ( S2A–S2C Fig ) . As a result , 4-day-old shb had a markedly shorter root length compared to the WT controls when both seeds were sowed on medium at the same time ( Figs 1A and 2A and 2G ) . To exclude the possibility that the short cortical cell phenotype observed in the shb root meristem was caused by delayed seed germination , we next synchronized WT and shb seed germination by sowing shb seeds on medium 12 h earlier before the WT control and performed a time-course analysis of various root phenotypes . We found that the sizes of root meristem ( S2D Fig ) and meristematic cortical cells ( S2E Fig ) were significantly and constantly shorter in shb than in the WT during the period of analysis . By contrast , meristematic cortical cell number ( S2F Fig ) , root length ( S2G Fig ) , root growth rate ( S2H Fig ) and cell production rate ( S2I Fig ) in the WT and shb were essentially identical until around the fifth day after synchronized germination , from which they started to diverge with significantly lower values in shb compared with the WT . Together , these results confirm our earlier observation that shb root meristem contained shorter cortical cells in the root meristem and suggest that fewer cells were produced from around the fifth day after synchronized germination . These phenotypes were not accompanied by changes in the lengths of elongation and maturation zone cells ( S2J and S2K Fig ) , further demonstrating a specific role of SHB in the root meristem . The shb mutant carries a homozygous T-DNA insertion in the 5th intron of the LOC_Os05g32270 gene ( Fig 3A ) , which dramatically reduces the expression level of LOC_Os05g32270 ( Fig 3B ) . A genomic fragment containing the LOC_Os05g32270 gene and its promoter and 3’ UTR regions could fully complement the mutant phenotypes of shb ( Fig 3C–3J ) . We thus concluded that LOC_Os05g32270 is the SHB gene . The SHB gene has been previously termed OsAP2-EREBP-049 , which encodes a putative transcription factor containing one AP2/EREBP DNA binding domain [23] . Based on the sequence similarity of the AP2/EREBP DNA binding domain , SHB was classified into the Group-1a of the AP2/ERF family , although all the other genes in this group have double AP2-EREBP DNA binding domains . A BLASTP search revealed that putative orthologs of SHB are present in both monocots and dicot plants ( S3A Fig ) . The subsequent phylogenetic analysis suggested that SHB is more closely related to its putative orthologs in monocots than to its dicot counterparts ( S3B Fig ) . Mutations in DWARF & IRREGULAR LEAF ( DIL1 ) , a homologous gene of SHB in maize , were reported to affect internode length , leaf shape and possibly root length [24] , implying that SHB and DIL1 have evolutionarily conserved functions . RNA in situ hybridization showed that SHB is expressed in the root meristem ( Fig 4A ) , thus supporting a functional role for SHB in vivo . A fusion protein of SHB and GFP {SHB-GFP; under the control of the Cowpea Mosaic Virus ( CPMV ) promoter [25]} co-localized with a nuclear marker SRT1-RFP [26] in tobacco epidermal cells ( Fig 4B ) , indicating that SHB functions as a nuclear-localized transcription factor . To determine the cause of GA deficiency in shb , we next examined whether the shb mutation decreases transcription of rice GA biosynthetic genes , including CPS1 , KS1 , KO2 , KAO , GA20OX2/SD1 and GA3OX2 , by quantitative real-time PCR ( qPCR ) . Among these genes , only KS1 and GA20OX2/SD1 were found to be significantly down-regulated by the shb mutation ( Fig 4C ) , suggesting that SHB modulates the levels of bioactive GAs in rice roots through transcriptional activation of GA biosynthetic genes KS1 and GA20OX2/SD1 . Notably , the expression of KO2 was weakly up-regulated in shb , perhaps to compensate for reduction of KS1 , which is involved in an earlier step of the GA biosynthesis pathway . AP2/EREBP proteins are able to bind the GCC-box , which is a short cis-acting element containing a core GCCGCC sequence motif [27] . Analysis of the KS1 promoter identified three GCCGCC motifs located at 205 , 184 , and 131 nucleotides upstream to the translation start site ( ATG; Fig 4D ) , whereas no GCC-box was found in the promoter region of GA20OX2/SD1 . Electrophoretic mobility shift assay ( EMSA ) indicated that SHB could bind to GCCGCC motifs located at 205 and 184 nucleotides upstream to the ATG ( S1 and S2; Fig 4D and 4E ) . No binding was detected with the third GCCGCC motif and SHB was not no longer able to bind to S1 and S2 when the two GCCGCC motifs were mutated ( M1 and M2; Fig 4E ) . To confirm the EMSA result in vivo , we performed ChIP-qPCR experiments with transgenic rice plants expressing a functional SHB-GFP fusion protein ( S4 Fig ) . qPCR showed that fragment b , which contains GCCGCC motifs located at 205 and 184 nucleotides upstream to the ATG of KS1 , was greatly enriched by ChIP with an anti-GFP antibody ( Fig 4F ) . On the contrary , DNA fragments covering other regions of the KS1 promoter , as well as the negative controls , were less amplified ( Fig 4F ) . These data indicate that SHB can directly bind to two closely located GCCGCC motifs in the promoter region of KS1 in vivo . RNA in situ hybridization revealed that KS1 was expressed in an overlapping domain with SHB ( Fig 5A ) and that KS1 had reduced expression level and domain in the shb mutant ( S5 Fig ) , in agreement with our finding that KS1 is a direct downstream target of SHB . A null ks1 mutant with severe GA deficiency [19] was found to have shorter root length and smaller root meristem than the WT control ( Fig 5B , 5C , 5E and 5I ) . 10 μM GA3 could fully complement the meristem size phenotype of ks1 ( Fig 5I ) , suggesting that KS1-dependent GA biosynthesis is essential for root meristem size control . Quantification of cortical cell number and size in the ks1 root meristem revealed that there was a significant decrease in cell proliferation compared to the WT ( Fig 5J ) , which was accompanied by an increase in the average length of cortical cells ( Fig 5K ) , indicating cell cycle arrest [28] . EdU staining further confirmed the cell proliferation defect ( Fig 5G and 5H ) and 10 μM GA3 could largely restore the number of cortical cells in the ks1 root meristem ( Fig 5J ) , suggesting that severe GA deficiency in ks1 is the underlying cause . Consistently , levels of most GAs were significantly reduced or undetectable in ks1 seedlings compared to the WT and shb seedlings ( S1 Table ) . Thus , we conclude that: 1 ) KS1-mediated GA biosynthesis is required for dose-dependent effects of GA on cell elongation and proliferation in the root meristem; and 2 ) The different root meristem phenotypes observed in shb and ks1 mutants result from moderate versus severe GA reduction . The latter conclusion was also suggested by much greater up-regulation of KO2 in ks1 ( S6 Fig ) than in shb ( Fig 4C ) and in agreement with this conclusion , a higher concentration of GA3 was needed to restore in 24 hours the root meristem size in ks1 ( 100 μM; S7A Fig ) than in shb ( 50 μM; S7B Fig ) . Our results suggest that cell proliferation in the rice root meristem is regulated by a dose-dependent effect of GA . Consistently , severe inhibition of GA biosynthesis by 10 or 50 μM PAC significantly impaired cell proliferation in the root meristem of shb mutants and WT plants ( Fig 6A–6F ) , resulting in a smaller root meristem ( Fig 6A , 6B and 6G ) with longer cells ( Fig 6A and 6B and 6H ) . By contrast , 0 . 1 μM PAC had no obvious effects on shb mutants and WT plants ( Fig 6A–6H ) , suggesting that PAC has a dose-dependent effect on cell proliferation . PAC-induced phenotypes could be reversed by co-treatment with GA ( S8 Fig ) , confirming that they were caused by inhibition of GA biosynthesis by PAC . Notably , 1 μM PAC markedly reduced elongation of cortical cells in the WT root meristem ( Fig 6H ) without significantly affecting cell proliferation ( Fig 6A–6F ) , whereas the root phenotypes of shb were less affected at the same concentration ( Fig 6H ) . These observations together suggest that GA has a dose-dependent effect on cell elongation in the root meristem , which is regulated by SHB . RNA in situ hybridization and qPCR showed that SHB transcription was induced by GA3 and repressed by PAC ( Fig 7A and 7B ) , suggesting that SHB functions as a positive regulator of GA signaling . Consistently , the level of KS1 transcripts was positively correlated with the level of GA ( Fig 7B ) . Moreover , SHB and KS1 expression was found to be down-regulated in the GA receptor mutant gid1 [29] ( Fig 7C ) , but up-regulated in slr1 ( Fig 7D ) , a constitutive GA response mutant defective in the SLR1 gene , the only DELLA gene in rice [30] . Together , these results suggest that GA regulates its own biosynthesis via positive feedback regulation on the expression of SHB and KS1 .
In this study , we have identified SHB , an AP2/ERF transcription factor , as a novel regulator of root meristem size in rice . We provide conclusive evidence that , during early stage of meristem development , root meristem size in rice can be regulated through SHB-dependent cell elongation , without the necessity to alter the rate of cell proliferation ( Figs 1 and S2 ) . This indicates that meristems , like organs , are able to adjust their size independent of cell number . A similar example can be found during Drosophila melanogaster imaginal disc development , during which the S6 kinase ( dS6K ) regulates cell size in a cell-autonomous manner without impinging on cell number [31] . From around 5 days after sowing , at which cell number in the root meristem , root growth rate and cell production rate start to decline ( Figs 1I , 1J , S2F , S2H and S2I ) , however , SHB appears to regulate both cell elongation and cell proliferation in the root meristem . SHB is required for longitudinal cell elongation but not radial cell expansion in the cortical layers of the rice root meristem ( Fig 1I ) . Moreover , loss of SHB function did not affect the elongation of post-mitotic cells in the rice root ( Figs 1J and S2J ) , indicating that SHB has a root meristem-specific function and that elongation of meristem cells differs in mechanism from rapid elongation of post-mitotic cells . This idea is in agreement with a previous report on the Arabidopsis STUNTED PLANT 1 ( STP1 ) gene , which is required for elongation of post-mitotic cells but not elongation of meristem cells [2] . STP1 was found to mediate the effect of cytokinin on the elongation of post-mitotic cells in the Arabidopsis root [2 , 32] . The identity of this gene remains unknown , but cytokinin has recently been shown to determine root meristem size by controlling cell differentiation [33] . On the contrary , our data suggest that GA controls root meristem size in rice through its dose-dependent effects on cell elongation and proliferation in the root meristem , which are mediated by SHB in a developmental stage-specific manner . Recent studies in the Arabidopsis root have shown that GA controls root meristem size by modulating cell proliferation [3 , 14] . GA deficiency , either in GA biosynthesis mutants or induced by PAC treatment , significantly impairs cell proliferation in the Arabidopsis root , leading to a decrease in cell production rate and meristem size . Expression of a non-GA-degradable DELLA mutant in endodermal cells was sufficient to inhibit cell proliferation and block root meristem enlargement , indicating that GA controls root meristem size in a DELLA-dependent manner . A reduction in cell elongation in the root meristem was hypothesized to cause reduced number of cell division events and thus block the increase in meristem size [3] . However , several evidences from our study suggest that in rice GA could regulate cell elongation in the root meristem independently of its effect on cell proliferation to influence meristem growth: 1 ) shb had reduced levels of GA3 and GA4 in the root ( Fig 2J ) . Exogenous application of GA3 to the shb root could restore the length of meristematic cortical cells and the size of root meristem to WT ( Fig 2H and 2I ) . 2 ) Treating the WT root with 1 μM PAC could phenocopy the effect of the shb mutation , resulting in a shorter root meristem with reduced cortical cell length but unaltered cell number ( Fig 6 ) . 3 ) Higher concentrations of PAC ( 10–50 μM ) significantly impaired cell proliferation in both WT and shb root meristems ( Fig 6 ) . It is also interesting to note that while Arabidopsis root meristem size nearly doubles during the first 4 days after sowing [34] , rice root meristem size shows only a 10–20% increase over the same period ( S2D Fig ) . The increase in Arabidopsis root meristem size was attributed to a proportional increase in meristematic cortical cell number [34] , whereas the size of rice root meristem ( S2D Fig ) appears to be influenced by both the number ( S2F Fig ) and the SHB-regulated length ( S2E Fig ) of meristematic cortical cells . How does SHB regulate GA levels and consequently cell elongation and proliferation in the root meristem ? Our in vitro and in vivo data suggested that SHB directly binds to and activates transcription of the GA biosynthesis gene KS1 , which has overlapping expression domain with SHB in the root meristem ( Figs 4 and 5 ) . In addition , our qPCR analysis indicated that SHB could also act through the GA biosynthetic gene GA20OX2/SD1 to modulate GA production in the root meristem ( Fig 4C ) , but the underlying molecular mechanism remains to be elucidated . Intriguingly , our qPCR analysis showed that both SHB and KS1 expression were induced by GA but repressed by SLR1 , the only DELLA protein in rice ( Fig 7D ) . These findings suggest that GA regulates its own biosynthesis and consequently cell elongation and proliferation in the rice root meristem via positive feedback regulation on the expression of genes involved in the early steps of GA biosynthesis . Taken together , our data suggest a model ( Fig 7E ) in which the root meristem size is modulated through dose-dependent effects of GA on cell elongation and proliferation in the root meristem , which are mediated by a developmental and possibly cumulative process that involves the root meristem- and developmental stage-specific function of the AP2/ERF transcription factor SHB . SHB , whose transcription is induced by GA , promotes GA biosynthesis in the root meristem by directly binding to and activating the GA biosynthesis gene KS1 . SHB-dependent and KS1-mediated GA availability is well correlated with the ability of meristematic cortical cells to elongate beyond the minimal cell length requirement for normal cell proliferation , allowing optimal rice root meristem growth . Loss of SHB function at the early stage of meristem development reduces the root meristem size by reducing elongation of meristematic cortical cells , without causing apparent defect in cell proliferation . This phenotype can be mimicked in WT roots by a moderate reduction of endogenous GAs in the presence of 1 μM PAC . Loss of SHB function at a later stage of meristem development ( from around the fifth day after sowing ) impairs both cell elongation and proliferation and consequently , further reducing the root meristem size . The reduction of cortical cell number in the shb root meristem could be correlated to the increase of cortical cell length , indicating cell cycle arrest . Loss of KS1 function and exposure of WT roots to high concentrations of PAC ( 10 or 50 μM ) , which strongly reduce endogenous levels of GAs , result in a more severe reduction of cortical cell number and a further elongation of cortical cells , suggesting that the degree of cell cycle arrest is related to the endogenous levels of GAs . The lower the GA levels , the more severe the cell elongation and proliferation defects and consequently , the smaller the root meristem size . Because of the importance of plants on the global level to food security and environmental sustainability , exploring molecular mechanisms underlying the control of plant organ size and growth has become a high priority in plant research worldwide [35] . Given that most agriculturally important crop species are monocots , our finding that SHB and its putative orthologs in monocot crop species are closely clustered in phylogenetic tree is of great importance . Future studies on their functions may lead to the identification of evolutionarily conserved mechanisms in cell size control and further our understanding on how meristem growth is modulated without markedly compromising cell proliferation to influence organ and body size . Consequently , rational design of crop plant architecture may be enabled by modulating the activities of SHB and its orthologs , which may improve the ability of crop plants to cope with adverse weather conditions such as rain , wind and hail [36] , ultimately leading to a second GA-dependent ‘Green Revolution’ in crop productivity .
The mutant line 03Z11ER89 ( shb ) was isolated from a rice enhancer trap collection [37] . The ks1 , gid1 and slr1 mutants were reported previously [19 , 29 , 30] . ks1 mutants set no seeds and therefore the mutation were maintained in a heterozygous state . For field studies , rice plants were cultivated under natural long-day conditions during rice cultivation seasons at the experimental field of Huazhong Agricultural University . For the analysis of seedling root phenotypes , seeds were sterilized and sowed simultaneously ( except for S2D–S2K Fig experiments which were conducted by sowing shb seeds 12 h earlier before WT controls ) on petri plates containing half-strength Murashige and Skoog ( MS ) medium ( Duchefa , The Netherlands ) , supplemented with 1% sucrose , 0 . 05% MES and 0 . 8% agar and adjusted to pH 5 . 8 . The plates were then placed vertically and incubated at 28°C in either continuous darkness or a light/dark ( 14/10 h ) regime , depending on the experimental design . For the complementation of shb phenotypes , the genomic sequence of SHB gene , together with 2 . 51 kb promoter and 386 bp 3’-UTR regions , was inserted into the pCAMBIA2301 vector ( http://www . cambia . org ) to generate pCAMBIA2301-SHB . To construct the complementing SHB-GFP fusion , sunlight GFP was fused in-frame to the C terminus of SHB coding sequence , and subcloned into pCAMBIA2301 under the control of the 2 . 51 kb SHB promoter . The empty vector pCAMBIA2301 was used as a negative control . Transgenic rice plants carrying each of these constructs were produced by using Agrobacterium-mediated transformation of callus of shb mutant . To construct CPMV::SHB-GFP , the coding sequence of SHB was fused in-frame to the N terminus of GFP and subcloned into the pEAQ-HT-DEST1 vector by using the GATEWAY recombination system , under the control of the Cowpea Mosaic Virus ( CPMV ) promoter [25] . Primers used for the construction of these vectors are listed in S2 Table . Protein sequences of putative orthologs of SHB from the other plant species were obtained by using blast search against the NCBI database ( http://www . ncbi . nlm . nih . gov ) . Multiple protein sequence alignment was performed using the ClustalX Version 2 . 0 [38] . The alignment was then manually refined . A phylogenetic tree was constructed using the MEGA 4 . 0 program [39] with the following parameters: Poisson correction , pairwise deletion , and bootstrap ( 1000 replicates; random seed ) . Root length was measured with Image J software ( http://rsb . info . nih . gov/ij ) . Root meristem size was determined by measuring the length from the quiescent center to the first elongated epidermal cell . Cell number in the root meristem , average cell length and width in the root meristem were quantified with all cells in the fourth cortical layer of the root meristem . Cell production rate in the rice root was calculated as described previously [3 , 40] . Briefly , time-course analyses of root growth and mature cell length were performed and root growth rate was determined using a five-point equation . Cell production rate was then calculated using the following equation: cell production rate = root growth rate/ mature cell length . Cells in the fourth cortical layer of the root elongation zone and maturation zone were used to obtain the quantification data of average cell length in the root elongation zone and mature cell length , respectively . For each quantification , at least 15 rice plants were analyzed . EdU staining was performed using an EdU kit ( C10310 , Apollo 488 ) from Ribobio , China , according to the manufacturer’s protocol . Briefly , roots of 4-day-old rice seedlings were immersed in 50 μM EdU solution for either 2 h ( Figs 1F , 1G and 5G and 5H ) or 20 h ( Fig 6C and 6D ) , and then fixed for 30 min in 4% paraformaldehyde , followed by 30 min of incubation with Apollo . The samples were next hand-sectioned longitudinally and EdU images of the sections were then captured with a Leica TCS SP2 confocal laser-scanning microscope equipped with a 20× water immersion objective and analyzed with Leica LAS AF software . Quantification of numbers of EdU-stained cells was performed in the fourth cortical cell layer of the rice root meristem , in a selected portion with a length of 360 μm . Quantification of GAs in the WT , shb and ks1 was performed as described previously [41] , using [2H2] GA1 ( 1 . 00 ng/g ) , [2H2] GA3 ( 1 . 00 ng/g ) , [2H2] GA4 ( 1 . 00 ng/g ) [2H2] GA12 ( 2 . 00 ng/g ) , [2H2] GA24 ( 2 . 00 ng/g ) , [2H2] GA19 ( 5ng/g ) , [2H2]GA20 ( 2ng/g ) , [2H2]GA34 ( 2ng/g ) , [2H2]GA44 ( 2ng/g ) and [2H2] GA53 ( 2 . 00 ng/g ) as internal standards . Roots of 7-day-old seedlings were used for comparison between the WT and shb . To compare the levels of endogenous GAs in ks1 , shb and the WT , whole seedlings were used due to the severity of ks1 root phenotype . Rice seeds were sown on medium supplemented with 10 μM GA3 and cultured for 4 days before analyzing the effect of GA on the root phenotypes . To determine the minimum concentration of GA required to rescue shb and ks1 root phenotypes , 3-day-old seedlings were cultured on medium supplemented with 10 , 25 , 50 , 75 or 100 μM GA3 for 24 hours before analysis . To examine the dose-effects of PAC on rice roots , 3-day-old seedlings were cultured on medium supplemented with 0 . 1 , 1 , 10 or 50 μM PAC for 24 hours before analysis . For mock treatments , medium with ethanol at the final concentration as for chemical treatments was used . Total RNA was extracted from WT and mutant roots using TRIzol ( Invitrogen ) reagent , according to the manufacturer’s instructions . qPCR analysis was performed in a 96-well plate with an ABI StepOnePlus Real-Time PCR System ( Applied Biosystems ) . The following thermal profile was used for all reactions: 95°C for 10 min , 40 cycles of 95°C for 15 s and 60°C for 1 min . The melting curve was determined under the following conditions: 95°C for 15 s , 60°C for 1 min , and 95°C for 15 s . The rice ubiquitin1 gene ( Os03g13170 ) was used as the internal control . All primers used are listed in S2 Table . RNA in situ Hybridization was performed as described previously [42] . Briefly , roots of 4-day-old rice seedlings were fixed in FAA ( 50% ethanol , 5% acetic acid and 3 . 7% formaldehyde ) at 4°C for 24 h , dehydrated in an ethanol series , cleared through a xylene series and then embedded in paraffin . 8 to 12 μm sections were mounted on RNase-free glass slides and in situ hybridization was then performed using digoxigenin-labeled RNA probes transcribed with either T7 or SP6 transcriptase from pGEM-T plasmids containing part of the SHB or KS1 coding sequence , which were PCR amplified with gene-specific oligonucleotide pairs SHBinsitu-F/R and KS1insitu-F/R ( S2 Table ) . Subcellular localization analysis of SHB-GFP was performed as previously described [43] . Briefly , lower leaves of N . tabacum . plants were infiltrated with Agrobacterium strains carrying CPMV-SHB-GFP using a syringe . For co-localization analysis with the nuclear marker RFP-SRT1 [26] the bacteria were mixed in appropriate volumes of infiltration buffer prior to injection into the leaves . Expression of fluorescent proteins was captured 2–3 d after infiltration with a Leica TCS SP2 confocal laser-scanning microscope equipped with a 40× water immersion objective and processed with Leica LAS AF software . To detect the binding of SHB protein to the KS1 promoter , EMSA was performed as described previously [42] with recombinant SHB protein produced in E . coli DE3 cells ( Novagen ) . In brief , the recombinant SHB protein was incubated with an [α 32P]-radiolabeled , double-stranded DNA oligonucleotide that covers the region containing the putative SHB binding sequence ( GCCGCC ) in the KS1 promoter . For control EMSA , nucleotide substitutions were introduced into the putative SHB binding site to produce the control probe . DNA binding reactions were carried out at room temperature for 20 min and the separation of protein-DNA complexes from the free DNA probes was done by non-denaturing polyacrylamide gel electrophoresis followed by auto-radiographic detection . Chromatin extraction and immunoprecipitation were performed as previously described [26] . Briefly , roots of the SHB-GFP plants were first vacuum-infiltrated and fixed in formaldehyde . The chromatin was then isolated from the nuclei of root cells and pre-cleared with sheared salmon sperm DNA/protein A agarose ( Invitrogen ) , and immunoprecipitated with or without an anti-GFP antibody ( Abcam; ab290 ) . The protein/DNA complexes were eluted and crosslinks were reversed to free DNA . The immunoprecipitated DNA was then purified and qPCR was performed with primers against KS1 promoter region . qPCR was also performed with input DNA purified from the pre-cleared chromatin , and the rice Actin gene ( Os11g06390 ) was used as the reference gene for normalization of qPCR data . Rice root nuclear proteins were extracted from roots of 7-day-old WT and SHB-GFP transgenic plants . Western blot analysis was performed as described previously [26] . After washing in acetone and dried , the proteins were resuspended in Laemmli sample buffer , then separated on a 12% SDS-PAGE and transferred to an Immobilon-P PVDF transfer membrane ( Millipore ) . The membrane was blocked with 2% bovine serum albumin in phosphate-buffered saline ( pH 7 . 5 ) , and incubated overnight with primary antibodies , such as anti-GFP ( Abcam; ab290 ) , in a 1:5 , 000 dilution at room temperature . After three washes ( 30 min each ) , the secondary antibody ( goat anti-rabbit IgG [SouthernBiotech] ) at 1:10 , 000 dilution was used . Visualization was performed using the Super Signal West Pico kit ( Pierce ) according to the manufacturer’s instructions . | Little is known concerning the identity of regulatory components and signaling pathways that control the growth of meristem cells in plants . Here , we report that rice plants deficient in the AP2/ERF family gene SHOEBOX ( SHB ) exhibited a severe reduction in the root meristem size . These plants had shorter cells in the root meristems following germination and fewer cells from approximately 5 days after sowing , suggesting that SHB regulates root meristem cell size and number in a developmental stage-specific manner and that cell size participates in the control of root meristem size in rice . SHB is positively regulated by GA signaling and encodes a direct transcriptional activator of the GA biosynthesis gene KS1 , indicating that GA modulates its own biosynthesis and consequently the elongation of meristem cells in rice roots via positive feedback regulation on the transcription of SHB and KS1 . Consistently , application of exogenous GA restored the size of root meristem cells to normal in shb and paclobutrazol-treated wild-type plants . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | SHOEBOX Modulates Root Meristem Size in Rice through Dose-Dependent Effects of Gibberellins on Cell Elongation and Proliferation |
Amyotrophic Lateral Sclerosis ( ALS ) is a fatal neurodegenerative disease characterized by selective loss of motor neurons , muscle atrophy and paralysis . Mutations in the human VAMP-associated protein B ( hVAPB ) cause a heterogeneous group of motor neuron diseases including ALS8 . Despite extensive research , the molecular mechanisms underlying ALS pathogenesis remain largely unknown . Genetic screens for key interactors of hVAPB activity in the intact nervous system , however , represent a fundamental approach towards understanding the in vivo function of hVAPB and its role in ALS pathogenesis . Targeted expression of the disease-causing allele leads to neurodegeneration and progressive decline in motor performance when expressed in the adult Drosophila , eye or in its entire nervous system , respectively . By using these two phenotypic readouts , we carried out a systematic survey of the Drosophila genome to identify modifiers of hVAPB-induced neurotoxicity . Modifiers cluster in a diverse array of biological functions including processes and genes that have been previously linked to hVAPB function , such as proteolysis and vesicular trafficking . In addition to established mechanisms , the screen identified endocytic trafficking and genes controlling proliferation and apoptosis as potent modifiers of ALS8-mediated defects . Surprisingly , the list of modifiers was mostly enriched for proteins linked to lipid droplet biogenesis and dynamics . Computational analysis reveals that most modifiers can be linked into a complex network of interacting genes , and that the human genes homologous to the Drosophila modifiers can be assembled into an interacting network largely overlapping with that in flies . Identity markers of the endocytic process were also found to abnormally accumulate in ALS patients , further supporting the relevance of the fly data for human biology . Collectively , these results not only lead to a better understanding of hVAPB function but also point to potentially relevant targets for therapeutic intervention .
Amyotrophic lateral sclerosis ( ALS ) is characterized by the degeneration of upper and lower motor neurons leading to a progressive deterioration of motor function and ultimately death , due to respiratory failure . Both environmental and genetic factors have been shown to contribute to ALS susceptibility . Approximately 10% of patients with ALS have a family history while the majority of ALS cases are classified as sporadic [1] . In rare inherited familial forms of ALS , specific mutations in causative genes have been identified and this has significantly advanced the field [2] . However , the mechanisms by which these mutations cause motor neuron dysfunction and death remain unknown . Several mutations in the human VAMP-associated protein B ( hVAPB ) have been identified as being causative of a type of ALS known as ALS8 [3–5] . hVAPB is a member of a highly conserved protein family involved in a variety of functions including maintenance of endoplasmic reticulum ( ER ) morphology , vesicular trafficking and intracellular lipid transport [6] . In C . elegans , the N-terminal domain of the protein is cleaved , secreted and functions as an extracellular ligand for Ephrin receptors [7] . Secreted fragments also bind to the leukocyte antigen-related ( LAR ) and roundabout ( Robo ) family receptors situated on the muscle cell surface to control actin organization and mitochondrial morphology and function [8] . Drosophila Vap-33-1 , ( hereafter referred as DVAP ) controls synaptic remodeling and composition of post-synaptic glutamate receptors [9 , 10] . The DVAP protein shows only 22% identity with its human orthologue hVAPB but the two proteins share a common structural organization including a transmembrane domain , a coiled-coil domain and a N-terminal region , which has a high degree of similarity to the C . elegans major sperm proteins ( MSP domain ) . More importantly , expression of hVAPB in flies can rescue the mutant phenotype associated with DVAP loss-of-function mutations , indicating that DVAP and hVAPB are functionally interchangeable [10] . DVAP also shows conservation of residues linked to ALS8 in humans ( human P56 is equivalent to fly P58 , human T46 is equivalent to fly T48 and human V260 is equivalent to fly V264 ) and in Drosophila , expression of transgenic alleles carrying pathogenic mutations reproduces major hallmarks of the human disease , including aggregate formation , chaperon up-regulation and locomotion defects [4 , 10 , 11] . A number of studies have implicated the disease-linked alleles in abnormal unfolded protein response ( UPR ) and in the disruption of both the anterograde axonal transport of mitochondria [12] and calcium homeostasis [13] . Moreover , the P56S disease-linked mutation is a loss-of-function allele through a dominant negative mechanism as it antagonizes the endogenous protein function by recruiting hVAPB into cytoplasmic aggregates [14–16] . Consistent with this disease mechanism , hVAPB levels are decreased in a number of sporadic cases as well as in TDP-43 ( transactive response DNA-binding protein 43 ) and in SOD1 ( superoxide dismutase 1 ) models for ALS [17–19] . The ALS causative gene fused in sarcoma ( FUS ) has been shown to bind VAPB mRNA and TDP-43 is associated with VAPB-positive aggregates in a transgenic mouse model expressing the P56S pathogenic allele [20–22] . Taken together , these data underline the importance of VAPB in ALS pathogenesis and explain why , in recent years , much effort has been devoted to generate models of ALS8 in a number of experimental organisms including Drosophila [2] . Despite this , the molecular mechanisms and the cellular processes underlying ALS8 pathogenesis remain elusive . Over the last few years , genome-wide genetic screens in yeast have proven to be extremely powerful in identifying novel genetic risk factors for ALS [23–25] and investigating fundamental pathogenic mechanisms associated with the ALS disease proteins TDP-43 and FUS [26–29] . These studies have led to the identification of important genetic modifiers of FUS and TDP-43 toxicity , including RNA binding proteins and proteins involved in RNA metabolism , ribosome biogenesis and cellular stress responses . To gain insight into fundamental pathogenic aspects of hVAPB-mediated ALS , we adopted a similar approach and by exploiting the power of Drosophila genetics , we conducted a genetic screen for modifiers of DVAP-P58S-dependent phenotypes . From the identified modifiers , a computational network of highly interconnected genes was assembled that encompasses a wide range of functional categories associated with DVAP-P58S toxicity . The Drosophila data were evaluated for their relevance to human biology by mapping the Drosophila genes to their human homologs and showing that the human genes could be integrated into a network of interacting genes largely overlapping with the Drosophila one . Some modifiers clustered within cellular processes that were already known to affect hVAPB protein functions such as proteolysis , intercellular signaling pathways mediated by Robo receptors and phosphoinositide metabolism [4 , 8 , 15] . However , we also identified a number of genes and processes that have not been previously associated with hVAPB function . Modifiers included genes involved in lipid droplet ( LD ) biogenesis and dynamics as well as endocytosis . We further confirmed the role of endocytosis in ALS pathogenesis by showing that the RAB5 GTPase , an early endosomal marker , is altered in post-mortem tissues of ALS patients . Importantly , a considerable number of DVAP-P58S modifiers are members of the Ras and Hippo ( Hpo ) signal transduction pathways , which control cell proliferation and apoptosis . Our data point to a mechanistic link between Ras and Hpo signaling on one hand and DVAP-P58S-mediated ALS8 phenotypes on the other . These results identify genes and cellular processes that widen our understanding of VAP protein function and provide novel therapeutic targets for ALS and perhaps other closely related neurodegenerative diseases such as fronto-temporal dementia .
To identify genes that influence DVAP-P58S-mediated pathogenesis , we undertook a genetic modifier screen and looked for genes that upon overexpression suppress or enhance the DVAP-P58S neurodegenerative phenotype in the eye . Flies , in which the expression of the DVAP-P58S transgene under the control of the yeast upstream activating sequence ( UAS ) was induced by the eyeless-GAL4 ( ey-GAL4 ) driver [30] , exhibit a robust eye phenotype characterized by decreased size , fused ommatidia , missing and occasionally supernumerary , bristles [15] . A transgenic strain ( DVAP-P58S11 ) was identified , whose targeted expression in the eye displays a reduction in eye size to about 30% of the wild-type value making this intermediate phenotype particularly suitable for an enhancer/suppressor screen . This eye phenotype was observed when flies expressing one copy of the DVAP-P58S transgene ( ey , DVAP-P58S ) , were raised at 30°C to maximize the expression of the transgene ( S1F , G , H Fig ) . A similar phenotype was also associated with the expression of a double copy of the same transgene in the eyes of flies ( ey , DVAP-P58S/DVAP-P58S ) raised at 25°C ( S1A , C , H Fig ) . At 28°C flies expressing one DVAP-P58S copy exhibited a phenotype that was less severe than that associated with flies of the same genotype raised at 30°C ( S1A , C , H Fig ) . However , at 28°C and at 30°C flies carrying a double copy of DVAP-P58S failed to hatch possibly due to a strong ectopic expression of the transgene outside the eye . Control flies raised at any of these temperatures did not show any visible eye phenotype ( S1A , D , F , H Fig ) . Taken together , these data clearly indicate that the reduction in eye size results from DVAP-P58S expression , it is sensitive to its dosage and it is not affected by the increase in the temperature ( S1 Fig ) . Additionally , we tested whether other disease features , such as the accumulation of aggregates , were dependent on the dosage of the pathogenic transgene ( S1I , J , K , L , M Fig ) . Stainings of larval eye imaginal discs with an anti-DVAP antibody showed a similar accumulation of aggregates positive to DVAP-immuno-reactivity in both ey , DVAP-P58S flies raised at 30°C and in ey , DVAP-P58S/DVAP-P58S flies raised at 25°C ( S1M , K Fig ) . A similar accumulation of aggregates was observed in eye imaginal discs of ey , DVAP-P58S flies raised at 28°C ( S1L Fig ) and only a few , sporadic aggregates were present in flies of the same genotype raised at 25°C , which exhibit only a small reduction in eye size ( S1J , B Fig ) . Thus , these data show that the extent of inclusion accumulation strictly correlates with the expression levels of the pathogenic transgene and with the severity of the eye phenotype . For simplicity , we decided to use as tester line for the screen , flies carrying a single copy of the DVAP-P58S transgene raised at 30°C . A line is considered to be a suppressor if it ameliorates the DVAP-P58S phenotype making the DVAP-P58S eye size becoming closer to that of the wild-type . By contrast , a line is an enhancer if it induces a decrease in the eye size of the ey-Gal4 , DVAP-P58S tester line . We previously showed that the reduction in size is due to DVAP-P58S-mediated perturbations in cell survival as coexpression of the Drosophila inhibitor of apoptosis 1 ( DIAP1 ) transgene substantially rescues the small eye phenotype , with adult flies displaying nearly normal eye size [15] . For the actual screen , ey-Gal4 , DVAP-P58S flies were crossed to mutants selected from publically available collections . As part of a large-scale gene disruption project , the Berkeley Drosophila genome project ( BDGP ) generated thousands of mutations associated with stable insertions of modified P-element transposons [31] . We focused on the screening of the EP and the EPgy2 collections of genome-wide insertional mutations for dominant modifications of the eye neurodegenerative phenotype associated with the ey-GAL4 , DVAP-P58S strain . These collections consist of UAS elements inserted in the promoters of endogenous genes and therefore they can be used to overexpress the downstream gene if the UAS-binding transcription factor GAL4 is co-expressed [32] . These collections were chosen for several reasons . Both collections were generated by random insertions of P-elements throughout the fly genome thus representing a large , unbiased , genome-spanning set of mutant lines . Additionally , both the EP and the EPgy2 collections comprise mutant lines in largely non-overlapping genes in which location and orientation of P-element insertions were previously mapped making the identification of the affected gene relatively straightforward . We initially focused on the second and the third chromosome and selected 1 , 120 P-element insertions that , based on the position and orientation of P-elements , are predicted to generate a potential overexpression of the affected gene . For 16 genes for which no EP or EPgy2 insertion lines with the potential of inducing overexpression were available , lines from the P{Mae-UAS . 6 . 11} collection were selected [33 , 34] . In addition , 47 EP and EPgy2 lines with P-elements inserted in the opposite direction , and therefore with the potential of disrupting the affected genes , were analyzed . As a result , a total of 1 , 183 lines were tested ( S2 Fig ) . While performing our screen , we realized that at 30°C , the temperature used to raise flies for the detection of suppressors , the identification of enhancers was difficult as the enhancing effect on the eye phenotype was too subtle . We noticed , however , that flies expressing DVAP-P58S under the control of the ey-GAL4 line exhibit decreased viability possibly due to the ectopic expression of the transgene in brain areas other than the eye . Although not carefully quantified , at 30°C expression of DVAP-P58S in the eye results in organism lethality and this effect is sensitive towards genetic modification as enhancers consistently lead to a decrease in the number of eclosed flies . Subsequently , the enhancing effect on the DVAP-P58S eye neurodegenerative phenotype was confirmed by rearing flies at 28°C to make the DVAP-P58S eye phenotype less severe . In this condition , scoring a putative enhancement of the eye phenotype becomes much easier . At 28°C , DVAP-P58S flies also exhibit partial organism lethality with 80% of the expected number of flies , eclosing . The enhancement effect of modifiers was confirmed by assessing the increase in organism lethality at 28°C . To eliminate false positives , all interacting insertions were retested two more times by an experimenter who was blind to the identity of the gene . Since we are interested in genes that primarily modify the DVAP-P58S-dependent phenotype , lines exhibiting an eye phenotype when driven by the ey-GAL4 line in the absence of the DVAP-P58S transgene were discarded . In summary , only those lines that did not show a neurodegenerative phenotype in trans with ey-GAL4 and for which results could be consistently repeated were designated as modifiers ( S2 Fig ) . Fig . 1A depicts light micrographs of representative eye phenotypes of one suppressor and Fig . 1B shows scatter plots of actual surface areas for the same suppressor . Fig . 1C presents average surface areas of every suppressor resulting from a set of three independent experiments . In Fig . 2 , results of a representative enhancer ( Fig . 2A , B ) and data on the effect of the remaining enhancers for viability and eye neurodegenerative phenotypes are presented ( Fig . 2C ) . The complete list of all suppressors is reported in S1 Table together with the percentage of modifying activity of every modifier . S2 Table reports genes with an enhancing effect on both the eye and organismal lethality phenotypes . In summary , this primary screen identified 85 modifiers including 71 suppressors and 14 enhancers ( S2 Fig ) . The 7% hit rate for genetic modifiers among the 1 , 183 tested strains is double the number of 1–4% typically observed in similar genetic screens in Drosophila . This demonstrates that the adopted approach was successful in identifying genes interacting with the DVAP-P58S neurodegenerative phenotype . However , false positive are common in high- throughput screens and therefore the effect of these modifiers was carefully and quantitatively retested by using an independent allele and an additional assay . Insertions of P-elements within or close to a gene could make difficult to assign phenotypes to specific genes with a high degree of confidence . However , if the ability of an EP or EPgy2 line to genetically modify the ey-Gal4 , DVAP-P58S phenotype is due to a disruption of the gene associated with the insertion , then other mutant alleles for that gene should also show a modifying effect . On the basis of previous studies , we identified additional alleles for many modifiers and tested their ability to modify the phenotype associated with the DVAP-P58S construct . For 62 out of the 85 modifiers , we were able to confirm the modifying effect of the specific gene on the ey-Gal4 , DVAP-P58S phenotype using independently generated alleles obtained from other laboratories/stock Centers or by RNAi knock-down of the putative modifier ( S3 Table , S2 Fig ) . More specifically , of the 71 suppressors 39 were confirmed by showing that an RNAi line has an enhancement effect on the eye/lethality phenotypes . Similarly , 3 suppressor lines were shown to be enhancers when loss-of-function alleles were tested to confirm their interacting ability . Conversely , the suppressor effect of 5 genes was confirmed by UAS lines while 3 suppressors were confirmed by using an independent EP or EPgy2 line . For 3 suppressors , the position and the orientation of the P-element suggest that the allele is likely to be a loss-of-function and therefore the suppression effect was confirmed by using an RNAi line ( for HSPC300 and Hippo ) or an independent loss-of-function allele ( for Klarsicht ) . Finally , the suppressive effect of the Syntaxin7/Avalanche gene was validated by using an antibody specific for the corresponding protein that was shown to be abnormally distributed in ey-Gal4 , DVAP-P58S eye imaginal discs when compared to controls ( S3 Fig ) . Of the 14 enhancers , 7 exhibited a suppressor effect with RNAi alleles and 2 were suppressors when loss-of-function mutations were tested . In summary , by using genetic and immune-histochemical approaches , the effect of 63 out of 85 modifiers was confirmed ( S2 Fig ) . For the remaining 22 lines ( 17 suppressors and 5 enhancers ) , it was not possible to confirm their effect by using RNAi lines . Although the relevance of these genes as potential modifiers of the DVAP-P58S phenotype needed to be confirmed by using additional independent alleles , for 21 of them the P-element insertion has been molecularly mapped and supposed to affect a specific gene . It is therefore conceivable that in these cases , RNAi alleles generated only a weak hypomorph so that the partial inactivation of the gene would not be sufficient to induce a visible effect . For the CG15630 suppressor line , however , the localization of the P-element insertion made it difficult to assign the modifying effect to a specific gene and therefore , at the moment , this gene remains a candidate modifier as it has been found to have an effect with only one allele . In conclusion , a total of 63 genes out 85 were confirmed to be modifiers of DVAP-P58S-induced neurodegenerative phenotype in the adult eye since the modifying ability is caused by changes in activity/expression of interacting genes and not by unrelated or genetic background effects . Our main goal with the primary screen was to generate a short list of candidates that could then be tested in a functional context that is more pertinent to the disease such as the motor system . We confirmed and extended our analysis by testing whether the identified modifiers also affect the DVAP-P58S-induced toxicity in neurons other than those of the adult eye and , more specifically , whether they interact with DVAP-P58S to affect motor behavior . Pan-neural expression of the DVAP-P58S construct under the control of the elav-Gal4 line does not result in viable offspring at 30°C [15] . However , when flies of the same genotype are raised at 28°C , about 70% of the expected offspring is viable and surviving flies exhibit a progressive decline in locomotion ability . The motor performance of flies as a function of age was assessed and quantified using a climbing assay . Control flies show no significant decrease in their motor performance until later in life and , in particular , for the entire period during which the motor performance was tested , about 80% of control flies were still able to climb ( Fig . 3A , B , C , blue lines ) . Conversely , flies expressing DVAP-P58S specifically in the nervous system using the elav-Gal4 driver ( elav-Gal4;DVAP-P58S ) , display progressive impairment of their motor performance ( Fig . 3A , B , C , red lines ) . We then analyzed the effect of every modifier on the climbing phenotype associated with the expression of DVAP-P58S in neurons . As shown in Fig . 3A , B for representative examples of two suppressor genes ( black lines ) , climbing ability was dramatically improved compared to flies that expressed only DVAP-P58S . Conversely , Fig . 3C ( black line ) reports an example of an enhancer where flies show a progressive worsening of the phenotype over the same time period . Data for all modifiers are reported in S4 , S5 , S6 , S7 Fig . Suppressors were classified as strong when they exhibited a consistent and robust suppression for at least three out of five time points tested ( S4 Fig ) . Suppressors with an intermediate effect displayed a significant change in two time points out of five ( S5 Fig ) . Finally , suppressors with a weak effect were those in which performance was improved over the entire time period but differences were of high statistical significance in only one time point or , mildly significant , over two time points ( S6 Fig ) . Eight genes identified as enhancers in the primary screen also exhibit a strong enhancing effect on DVAP-P58S-induced motor dysfunction ( S7 Fig ) . Out the 85 modifiers , 58 were found to have a modifying effect on DVAP-P58S motor performance ( S2 Fig ) . To investigate whether the locomotion defects could be attributed to synaptic deficits , we turned to the Neuromuscular Junction ( NMJ ) of third instar Drosophila larvae , a model glutamatergic synapse , the structural and functional properties of which have been extensively characterized [35] . We examined the morphology of larval NMJs located on muscle 12/13 by immunohistochemistry with antibodies against the presynaptic marker HRP . As previously reported , elav-DVAP-P58S NMJs exhibit a reduction in the number of boutons and an increase in their size when compared to controls [10 , 14] . Here we found that co-expression of either Ric or Diap2 , two potent suppressors of the DVAP-P58S-dependent eye and locomotion phenotypes; also rescued the morphological defects of NMJs both in size and number of boutons ( Fig . 3D , E , F , G , I , J ) . Notably , expression of either Ric or Diap2 alone under the control of the same elav-Gal4 driver does not result in any visible structural change at the synapse ( S8 Fig ) . In contrast to what was observed for Diap2 and Ric , co-expression of Smooth ( Sm ) , a gene with an enhancing effect on the DVAP-P58S-mediated eye and motor phenotypes , leads to a severe disruption of synaptic structural integrity at NMJs . Specifically , while in controls synaptic boutons within a branch resemble a string of beads connected by a short neuritic process , the architecture of NMJs co-expressing DVAP-P58S and Sm is profoundly altered and consists of a reduced number of irregularly shaped boutons which are frequently disconnected from the nerve or axonal branches ( Fig . 3D , H , K ) . Disruptions are very rare in controls and in elav , Sm synaptic terminals and are occasionally found in in DVAP-P58S expressing NMJs but their frequency is drastically increased by co-expression of DVAP-P58S and Sm ( Fig . 3K ) . Finally , a small but statistically significant decrease in bouton number was observed when Smooth expression was induced by the driver alone ( S8F Fig ) . However , this defect is very mild compared to the widespread disruption of synaptic structure due to the concomitant expression of DVAP-P58S and Sm . In summary , the strong correlation between modifying effects on locomotion behavior and on synaptic morphology we reported for a subset of DVAP-P58S interacting genes , further supports the functional relevance of the identified interactions . While most modifiers were isolated as suppressors and confirmed to have a suppressive effect in the motor assay , DAP160 was isolated as a suppressor in the eye but exhibited an enhancing effect on the motor phenotype when broadly overexpressed in neurons . This apparent contradiction suggests that perhaps a more complex interaction exists between DVAP-P58S on one hand and the DAP160 gene on the other . Under these specific testing conditions , neuronal expression of the three modifiers ( rhomboid , Src42 and leak ) exhibit an independent developmental lethality thus impairing the testing of motor performance in adult flies . These genes and the remaining modifiers that were not confirmed by the motor assay still represented bona-fide interactors of DVAP-P58S phenotypes , although additional alleles or/and different assay conditions will be required to assess their effects on DVAP-P58S motor performance . Importantly , these results show that our analysis identified 42 of the 85 tested as high-confidence genetic modifiers ( S2 Fig , S6 Table ) . Indeed these genes exhibit a modifying effect on DVAP-P58S disease phenotypes by using more than one allele and two independent assays , the eye neurodegenerative phenotype and the adult fly motor performance . In conclusion , a large set of modifiers was found to affect both phenotypes , eye degeneration and locomotion ability , in the same direction and there is indeed a strong correlation between the severity of the modifying effect on eye neurodegeneration and that on locomotion impairment . A reasonable test of success for any genetic screen would be the isolation of genes that we would predict to interact with the DVAP protein on the basis of prior experiments . Not all the Drosophila genes have been tagged with EP or EPgy2 insertions and we did not screen all the available lines in the collections . Therefore the isolation of even a few such modifiers would confirm that the screen has been successful . Indeed , we isolated several such genes . Most notable among them is rdgBβ . The gene rdgBβ is predicted to encode a phosphoinositide phosphatase that has not been extensively investigated hitherto . Previous studies suggest that rdgBβ may have functions similar to those of the better characterized phosphodiesterase rdgB [36] . rdgB is the homologue of the human Nir2 gene that has been shown to interact with the hVAPB protein to control Golgi transport in human cell lines [37] . In addition , we and others have shown that a primary and conserved function of VAPB proteins is to control the levels of phosphoinositides associated with cellular membranes [15 , 38 , 39] . Another gene isolated as a strong modifier is leak . In C . elegans , the N-terminal MSP domain of VAPB is cleaved and secreted , serving as an extracellular ligand . The secreted MSP fragment binds LAR and Robo family receptors on the muscle cell surface leading to alterations in actin organization and changes in mitochondrial morphology and function [7 , 8] . Leak is the Drosophila homologue of Robo2 , a protein family controlling axonal pathfinding and tracheal system development [40] . In DVAP mutants , muscle phenotypes similar to those associated with the removal of the corresponding gene in C . elegans have been reported [7 , 8] . It is therefore possible that if leak is a Robo2 receptor , the similarity in phenotypes between Drosophila and nematodes corresponds to a similarity in molecular mechanisms underlying VAPB-mediated functions . In addition , among the high-confidence DVAP-P58S interacting genes , our screen identified actin ( S1 Table ) and mutations in profilin 1 , which regulates actin polymerization , were recently linked to ALS [41 , 42] . Our studies also uncover modifiers for which no involvement in VAPB function has been previously reported . Analysis of these genes could potentially illuminate novel and so far unpredicted , functional processes involved in ALS8 pathogenesis . Although we analyzed a large collection of genes , our screen was not saturating as less than 10% of the about 12 , 000 Drosophila genes were screened . Additionally , important modifiers may have induced subtle changes in DVAP-P58S pathology , which would have been excluded based on our strict criterion of robust suppression or enhancement of DVAP-P58S-associated phenotypes . Therefore , we performed an extensive computational analysis of modifier genes in order to extrapolate cellular processes , pathways and other genes that may have a relevant role in DVAP-P58S-associated disease phenotypes . In the list of modifiers we also incorporated DVAP as it has been shown that the wild-type protein binds to the ALS8-causing allele and is sequestered into the aggregates . In addition , it appears that at least some of the interactions of VAPB proteins are retained by the mutant allele [15 , 16] . It is therefore possible that the wild-type version of the protein has an important role to play in the disease pathogenesis . The 85 modifier genes were analyzed using the R/Bioconductor package ‘topGO’ which uses a GO-graph de-correlation method to remove conditional dependence from the annotated GO tree improving interpretation of results by removal of non-specific higher level terms and allowing for multiple testing correction to control the false positive rate [43] . The predominant processes enriched in the modifier list reflect broad effects on a variety of biological processes reinforcing the concept of the pleiotropic nature of disease pathogenesis ( Fig . 4A ) . Analysis of the Drosophila DVAP-P58S network with topGO identified endosomal regulation , lipid particle metabolism , vesicular trafficking and apoptosis as enriched functional categories ( Fig . 4A ) . Some of these biological processes , such as vesicular trafficking , have already been linked to VAPB activity although their role in VAPB-mediated neurodegeneration has never been directly assessed [44 , 45] . Additionally , functional categories such as endocytosis and lipid particle metabolism represent biological processes that have not been previously associated with VAPB activity ( Fig . 4A ) . Based on a recent report showing that VAPB can function as an oncoprotein in humans [46] , it is particularly intriguing that a number of modifiers fall within enriched functional categories related to apoptosis ( Fig . 4A ) . To further extend our analysis , we searched the GeneMania database [47] to determine whether the genetic modifiers are interconnected through physical and genetic interactions as well as colocalization studies . We found that 72 of the 85 genes are linked into a complex network and , more importantly , 7 of the DVAP-P58S genetic modifiers ( CG7324 , Syx7 , Ero1L , Rab5 , CG5118 , rho and Spp ) have previously been reported to be physical interactors of DVAP [48] ( Fig . 4B ) . We also carried out a statistical analysis comparing our GeneMania derived network graph to those produced by repeatedly sampling the GeneMania data set with 1000 randomised gene lists of the same size . To summarise the topographical properties of the resulting groups of network graphs we calculated the mean node-degree and mean node-betweeness , two properties for which biological networks have been shown to have high values [49] and calculated Z-scores for the comparison of both metrics between our graph and the distribution of random gene graphs . We found significantly higher values for both mean node-degree ( 2 . 07x , Z = 4 . 67 , p = 3 . 05x10-6 ) and mean node-betweeness ( 2 . 86x , Z = 2 . 75 , p = 0 . 006 ) consistent with our interpretation that the network of DVAP-P58S modifiers is structurally very different from that produced from random gene lists and shares key topographical features commonly observed in other known biological networks . Together , these findings confirm and expand categories and pathways associated with DVAP function and provide us with a perspective on the diverse molecular functions that can modulate DVAP-induced pathogenesis in vivo . Careful Drosophila to human homology mapping using a recently developed orthologue mapping tool ( Drosophila RNAi Screening Center Integrative Orthologue Prediction Tool , DIOPT at http://www . flyrnai . org/diopt ) shows that of the 85 Drosophila genes at least 77 have human orthologues ( S4 Table ) . We decided to use DIOPT instead of any other specific software because it has the advantage of integrating the results of several orthologue searching tools based on different algorithms [50] . This tool also includes algorithms identifying potential functionally related proteins in humans based on information derived from protein-protein interaction networks . In searching for orthologues using this tool , we applied a filter that removed predicted human proteins with a DIOPT score less than 2 as this would mean that the identification of this orthologue is supported by only one algorithm . To search whether any of the identified modifiers are known to be associated with a disease of the nervous system , we used a tool based on DIOPT named DIOPT-DIST ( http://www . flyrnai . org/diopt-dist ) . The DIOPT-DIST website links Drosophila genes to high-confidence orthologues of disease genes extracting information from the Online Mendelian Inheritance in Man data set and Genome-Wide Association Studies . We found that a considerable number of human orthologues of genetic modifiers have been associated to neurological disorders including neurodegenerative diseases such as Parkinson’s and Alzheimer’s diseases , multiple sclerosis and spinocerebellar ataxias . Other genes were linked to human psychiatric disorders such as schizophrenia , autism spectrum disorder and mental retardation ( S5 Table ) . The recovery of these genes suggests that the genetic network identified by our screen may overlap , perhaps significantly , with the genetic networks associated with other human neurological disorders . If this turns out to be true , the use of Drosophila to explore other neurological disease networks via genetic screens would have the potential to identify common therapeutic targets that could be tested in other disease models . To study the DVAP-P58S genetic circuitry in the human context we generated a human view of the genetic Drosophila interactome taking advantage of the manually curated source of human molecular interactions from the Ingenuity Pathway Analysis ( IPA database , Ingenuity system at www . ingenuity . com ) . This database integrates human gene relationships derived from a variety of experimental approaches including proteomic studies . Using the hVAPB proteins and the 77 human genes homologous to the genetic modifiers identified in Drosophila , we derived the human interaction network of hVAPB . The generated human interactome includes hVAPB and 31 additional genes indicating that these proteins , which represent 40% of the identified modifiers , are interconnected in the human IPA database ( Fig . 5A ) . Importantly , we found that all the genes included in the hVAPB interactome are also components of the Drosophila genetic network further supporting the relevance of the fly modifiers for the elucidation of the function of hVAPB and for ALS8 pathogenesis . To further expand the hVAPB interactome , we identified proteins in the IPA database that interact with hVAPB and incorporated them into the list of human orthologues of Drosophila modifiers . In so doing , we established a new network of interconnected genes that include the initial 31 modifiers and 12 additional ones ( Fig . 5B ) . Our experimental data integrated with our computational analysis support and reinforce the relevance of the Drosophila model in identifying the cellular processes and in dissecting the complex genetic network underlying ALS8 pathogenesis in humans . To perform a further functional characterization of a subset of modifiers , we focused on genes that were present within both the Drosophila modifier network and the expanded IPA human network of hVAPB . A number of these genes include genes known to function in vesicular and endocytic trafficking . Then we generated a sub-network including protein interactions relevant to DVAP-P58S and proteins with Gene Ontology annotations related to vesicular trafficking and/or endocytosis present in the GeneMania data base ( Fig . 6A ) . A similar network was defined for the human orthologues of these genes using IPA ( Fig . 6B ) . Rab7 is among the genes that are predicted to be part of both the fly and human network but that were not identified in the screen . Rab7 is an endocytic marker of late endosomes that results from the maturation of early endosomes labeled by Rab5 . Rab5 is a potent modifier of DVAP-P58S phenotypes and we showed that up-regulation of Rab5 functions as a potent suppressor ( S1 Table ) . To assess whether a similar effect could also be extended to Rab7 , we crossed ey-Gal4 , DVAP-P58S flies with flies expressing either a wild-type or a constitutive active form of Rab7 and , in both cases , a strong suppression of the disease phenotypes was observed . ( Fig . 6C , D , E ) . While Rab5 and Rab7 provide important organelle identity markers for early and late endosomes respectively , Rab11 is instead the identity marker for recycling endosomes . As for Rab5 and Rab7 , overexpression of a wild-type version of Rab11 suppresses the DVAP-P58S eye neurodegenerative phenotype ( S9 Fig ) . These results are particularly striking if we consider that VAPB has been shown to bind to Rab7 and to be required for the proper spreading of late endosomes throughout the cell [51 , 52] . To test the cellular distribution of endosomes , eye imaginal discs expressing the DVAP-P58S transgene were stained with an antibody specific for Rab5 . While in controls Rab5-positive immuno-reactivity forms a granular pattern dispersed throughout the cell , Rab5 abnormally accumulates in cells expressing the DVAP-P58S transgene and partially overlap with DVAP-P58S-induced aggregates ( Fig . 7A , B ) . A similar effect was observed in neurons of larval brains expressing DVAP-P58S under the control of the pan-neural driver elav-Gal4 ( Fig . 7 C , D ) . These data suggest that inclusion of Rab5 proteins into aggregates and their inadequate delivery to the normal site of function may underlie DVAP-P58S-mediated neurodegeneration . Interestingly , we found that in eye imaginal discs simultaneously expressing Rab5 and DVAP-P58S , Rab5 immunoreactivity exhibits , at least in part , a normal distribution ( S10 Fig ) . These data suggest that suppression of DVAP-P58S toxicity by overexpression of Rab5 proteins occurs by replacement of those proteins that have been incapacitated by inclusion into DVAP-P58S-induced aggregates . However , the elucidation of the mechanism whereby Rab5 protein expression confers protection against DVAP-P58S-associated neurodegeneration , requires further investigation . Collectively computational and experimental evidence provide proof that our data set is enriched for proteins such those involved in vesicular trafficking and endocytosis and that these processes are important in DVAP function and patho-biology . We next performed immunohistochemistry on human post-mortem spinal cord tissue to determine whether RAB5 is present in motor neurons and whether its localization is affected in ALS . RAB5 is robustly expressed in spinal cord motor neurons and localizes to punctuate , granular pattern distributed throughout the cellular cytoplasm ( Fig . 8A ) . Conversely , RAB5 localization in autopsy tissues from two sporadic ALS cases reveals an abnormal accumulation and clustering of RAB5 positive vesicles ( Fig . 8B , C ) . To analyze whether there was a co-localization between hVAPB and RAB5 , we immunostained post-mortem cervical sections of spinal cord from ALS patients and age-matched controls . In controls , hVAPB was homogenously distributed throughout the neuronal cytoplasm and showed a granular appearance consistent with the protein being localized in the endoplasmic reticulum ( Fig . 8D ) . In agreement with what previously described [17 , 18] , we also noticed hVAPB expression was abundant in large motor neurons characterized by a prominent nucleus and nucleolus while smaller cells were still positive for Rab5 immunoreactivity but essentially devoid of hVAPB staining . In ALS tissues however , hVAPB staining was more conspicuous than in controls and present as punctuate accumulations that partially overlap with redistributed Rab5 positive clusters ( Fig . 8E , F ) . Previous studies have reported that in a few ALS cases the expression levels of hVAPB protein is decreased . Reasons that could explain this apparent discrepancy are not known but they could include differences in disease characteristics such as disease duration and age of onset . Interestingly , we reported that in Drosophila expression of DVAP wild-type protein is sufficient to trigger many hallmarks of the disease and that a recently identified ALS causing mutation in hVAPB mirrors the effects of DVAP wild-type overexpression and acts as an allele with increased wild-type activity [11] . These data indicate that in humans as well as in fly models both gain and loss of function mechanisms may play an important role in hVAPB-induced ALS . Notably , a recent study has also reported an abnormal accumulation of RAB5 , RAB7 and RAB11 in post-mortem tissues of ALS patients [53] . Taken together , these data further support the notion that endocytic trafficking is important for ALS8 pathogenesis and that aberrant accumulation of endocytic markers can contribute to the human pathology .
A mutation in the hVAPB gene was initially reported to be causative of a range of motor neuron diseases including ALS8 [3] . The disease due to the P56S mutation in a conserved domain of the protein is characterized by a high degree of heterogeneity in age of onset , severity and clinical progression [3 , 54 , 55] . In a fly model of ALS8 , severity of disease phenotypes has been reported to strongly depend on the dosage/activity of the mutant allele suggesting that genetic changes that either directly or indirectly affect DVAP function may be of therapeutic interest [9 , 15] . Though VAPB proteins have been linked to a number of seemingly unrelated functions [6 , 12 , 15 , 16 , 56] , their interacting genes have not been systematically investigated nor has the link between ALS8-causing mutations in the hVAPB gene and the specific loss of motor neurons been uncovered . To systematically explore the genome for genes that are capable of modulating hVAPB-mediated disease phenotypes in vivo , we took advantage of the ALS8 fly model we recently developed [15] . This model offers the possibility of performing large-scale genetic screens by using two phenotypic readouts: the neurodegenerative phenotype and the motor behavior analysis associated with expression of the disease-causing allele in the eye and the nervous system , respectively [15 , this study] . We identified numerous DVAP-P58S genetic modifiers that fall within a variety of different functional categories suggesting a pleiotropic genetic effect of ALS8 mutations in inducing neuronal dysfunction and death . Overwhelming evidence supports a role of protein degradation deficits in neurodegenerative diseases including ALS through disruption of either of the two major protein clearance pathways: the ubiquitin proteasome system and autophagy [57] . Autophagosomes accumulate in the spinal cord of sporadic ALS patients [58] and a decrease in autophagic flux is present in ALS mice and cell lines expressing mutant SOD1 [59] . Here we report that the autophagy-related 7 ( Atg7 ) protein , essential for autophagy , is a high-confidence suppressor of DVAP-P58S as overexpression of Atg7 suppresses and its down-regulation enhances the DVAP-P58S-associated disease phenotypes ( S6 Table ) . These data are consistent with a scenario in which the autophagy process is down-regulated in DVAP-P58S expressing cells and are in agreement with previous experiments reporting that mice lacking Atg5 and Atg7 in the nervous system exhibit neurodegeneration [60 , 61] . The identification of ALS-causing mutations in genes affecting the ubiquitin-proteosome system directly supports a role for this process in ALS pathogenesis . These genes include ubiquilin-2 , p62/SQSTM1 , optineurin , vasolin-containing protein , and charged multivesicular body protein 2B ( [2] and references therein ) . The ubiquitin-mediated protein degradation pathway ensures that only proteins that are properly folded and assembled are transported to their final destination . When mis/unfolded proteins are not degraded , an ER stress response is induced , known as unfolded protein response ( UPR ) . Previous studies have already implicated VAPB proteins in protein homeostasis by showing that they can modulate the activities of the IRE1 and ATF6 arms of the UPR [4 , 62] . Here , we report that several DVAP-P58S interacting genes including Cullin2 , CG9153 and CG4502 , are involved in ubiquitin-mediated proteolysis . The ubiquitin-proteasome system is a selective protein degradation pathway in which a substrate is first tagged with a chain of ubiquitin and the resulting modified protein is then recognized by the proteasome where its proteolysis takes place . The process of ubiquitination involves a three branched enzymatic cascade . First , the chemically inert ubiquitin molecule is activated in an ATP-dependent reaction by binding to an E1 activating enzyme . Second , ubiquitin is transferred to the active site of an E2 ubiquitin conjugating enzyme . In the final step an E3 ubiquitin ligase functions to orchestrate the transfer of ubiquitin to a substrate protein that needs to be tagged for degradation [63] . Cullins organize the largest class of RING-E3 ubiquitin ligases by functioning as molecular scaffolds tethering a substrate target unit with the ring finger component that recruits the E2 ubiquitin conjugating enzyme [64] . Another type of E3 ubiquitin ligases is represented by the Hect type . The single molecule HECT-type E3 ligases are characterized by a Homologous to the E6-AP carboxyl terminus ( HECT ) domain that forms a thio-ester intermediate with ubiquitin as a pre-requisite for ubiquitin transfer to the substrate protein [65] . The CG9153 DVAP-P58S interacting gene encodes a putative HERC3 homologue in Drosophila . HERC3 contains a HECT domain and has been reported to associate with ubiquilin-2 , an ALS causative gene [66] . Finally , another modifier is the CG4502 that encodes a predicted E2 ubiquitin ligase . Collectively these data provide further support to the notion that the ubiquitin-mediated protein clearance system represents a major component of ALS pathogenesis . However , we were surprised to find that our list of modifiers is highly enriched for proteins linked to lipid droplets ( LDs ) , by either functional or proteomic studies . As an example , the Acyl-CoA synthetase long-chain ( Acsl ) gene is a high-confidence modifier of DVAP-P58S phenotypes ( S6 Table ) and its vertebrate homologue ACSL3 has been implicated in LD biogenesis [67] . LDs are ubiquitous organelles that collect , store and supply lipids [67] . The absence of ACSL3 significantly reduces nucleation of emerging LDs , impairs accumulation of neutral lipids and decreases size and numbers of mature LDs [67 , 68] . Here we show that overexpression of Acsl ameliorates , while its down-regulation exacerbates , both neurodegeneration and motor disabilities associated with neuronal expression of DVAP-P58S . These data indicate that Acsl could be down-regulated in DVAP-P58S neurons and that impaired LD biogenesis may represent an important pathological aspect of VAPB-mediated ALS . Importantly , we also found that loss of function mutations in the Drosophila gene Klarsicht controlling motion and distribution of LDs [69] acts as a potent suppressor of DVAP-P58S phenotypes ( S6 Table ) . Aberrant movement of droplets can compromise their dispersion throughout the cytoplasm . In C . elegans and mice , mutations in VAPB induce an aberrant clustering of LDs in striated muscles that causes an energetic unbalance and a decrease in ATP production following starvation [70] . Consistent with these data , large-scale proteomic studies showed that DVAP is an LD-associated protein [71] . The functional role of LDs in the nervous system has been largely neglected , though recent studies have reported that a number of LD proteins are abundantly expressed in the brain [72 , 73] and some of them have been directly linked to motor neuron diseases . One such protein is seipin . Gain-of-function mutations of seipin cause motor neuron diseases known as Silver syndrome/spastic paraplegia 17 and distal hereditary motor neuropathy type V [74] . Seipin is an ER resident protein that regulates LD morphology in yeast while in mammals is essential for the induction of lipogenesis [75–77] . Mice expressing the seipin transgene carrying the disease-causing mutations , develop symptoms of motor neuron diseases along with an up-regulation of ER stress markers and an induction of autophagy [78] . In addition , spartin/SPG20 , a protein mutated in the motor neuron disease known as the Troyer’s syndrome , is found on LDs and is implicated in LD turnover [79] . Interestingly , it has been proposed that LDs may play a protective role as they are large hydrophobic surfaces that sequester un-/mis-folded proteins and prevent their accumulation in large inclusions [80] . This protective function may be of general relevance also for motor neuron diseases for which the link with LD metabolism is not directly evident . Indeed , SOD1 mouse models of ALS have been shown to benefit from a fat rich diet that appears to attenuate disease symptoms [81] . Taken together , these data suggest the intriguing possibility that maintaining normal levels of neutral lipid content may represent a novel and promising therapeutic strategy in ALS . Here we show that computational and experimental analyses support a role for the endocytic pathway in ALS8 pathogenesis . The endosomal system is necessary for regulating , sorting and degrading proteins via autophagy or the ubiquitin proteasome-system . Rab-GTPases are identity markers of endosomes and regulate endocytosis through interactions with vesicular coat components , motor proteins and SNARE proteins . Incoming substances and receptors for the regulation and fine-tuning of many cellular pathways are initially present on Rab5-containing early endosomes that undergo maturation to become Rab7-containing late endosomes while Rab11 regulates recycling of endocytosed proteins via recycling endosomes [82] . Overexpression of several Rab-GTPases proteins including Rab5 , Rab7 and Rab11 , functions as potent suppressor of DVAP-P58S associated phenotypes . Further supporting a link between hVAPB and endocytosis , we identified as modifiers the SNARE proteins VAMP7 and Syntaxin 7 , which are required for late endosome-lysosome fusion events [83] . We also isolated as a modifier the VPS35 gene , coding for a component of a retromer complex , which guides protein sorting from the endosomal-lysosomal degradation pathway retrogradely to the Golgi network [84] . Furthermore , we found an abnormal distribution of RAB5 into clusters overlapping with hVAPB accumulations in ALS patients , suggesting hVAPB-mediated dysregulation of endosomal trafficking in ALS pathogenesis . These data were confirmed by a recent report showing aberrant accumulation of RAB7- and RAB11-positive vesicles in post-mortem tissues of ALS patients [53] . Interestingly , previous studies have shown that DVAP binds Rab7 and controls intracellular positioning and distribution of late endosomes [51 , 52] . These data collectively indicate that VAPB proteins have a broad and conserved function in facilitating endocytic trafficking and that disruption of this process is a prominent cause of ALS pathogenesis . Among the genes our genetic strategy identified as DVAP-P58S modifiers , there are several members of the Ras signaling pathway , indicating that disruption of this pathway may play a role in ALS pathogenesis . The Ras pathway , which enables cells to respond to external cues , controls cell proliferation , differentiation and apoptosis . In metazoans , the protein Ras , Raf and MEK act sequentially to activate ERK [85] . Remarkably , we report that DVAP-P58S-mediated ALS phenotypes were sensitive to the dosage of a number of genes that function upstream of Raf and downstream of Ras including connector enhancer of KSR ( CNK ) , Src42 and 14-3-3ζ . CNK functions as multivalent adaptor protein that cooperates with Ras and Src42 to induce activation of Raf [86 , 87] . Interestingly , we show that Hippo ( Hpo ) may be up-regulated in neurons expressing the DVAP-P58S transgene as reducing the genetic dosage of the tumor suppressor gene Hpo by using two independently generated alleles is sufficient to suppress the effect of DVAP-P58S expression on the eye neurodegenerative phenotype ( S6 Table ) . Additionally , the same genetic manipulation fully rescued the locomotion defects associated with pan-neural expression of the pathogenic transgene . In agreement with these data , overexpression of Hpo represses cell proliferation and induces apoptosis [88] while the mammalian homologue MST1 is hyperactivated in ALS patients and its down-regulation delays disease onset and extends survival in SOD1 mice models for ALS [89] . It has been reported that Hpo promotes apoptosis by reducing the expression levels of a number of downstream targets including the Drosophila inhibitor of apoptosis 1 ( DIAP1 ) [90] . We have previously shown that DIAP1 is down-regulated in DVAP-P58S mutant background and that up-regulation of DIAP1 mitigates DVAP-P58S neurodegeneration . These data provide further support to the hypothesis that in triggering neurodegeneration by apoptosis , DVAP-P58S acts via the Hpo tumor suppressor pathway [15] . Recently , a high throughput protein-protein interaction analysis looking for additional genetic members of the Hpo pathway identified DVAP as a high-confidence member of the Hpo interactome [91] . This study also implicated a fundamental role for the vesicular trafficking process in various aspects of the Hpo signaling and the importance of VAPB protein function in intracellular vesicular trafficking has been well documented [44 , 45] . These data reinforce our genetic evidence supporting a role for the tumour suppressor gene Hpo in ALS pathogenesis and provide a landscape of molecular interactions that can promote the formulation of additional mechanistic hypotheses to be experimentally validated . Intriguingly , hVAPB can act as an oncoprotein as overexpression of hVAPB in mammary epithelial cells induces an increase in cell proliferation and its expression levels are elevated in primary and metastatic tumor specimens [46] . Although further correspondence between the Drosophila model and the human condition remains to be determined , the relationship between DVAP-P58 phenotypes and the Drosophila modifiers , raises the possibility that pharmacological manipulation of these genes may assist in developing treatments to ameliorate or prevent ALS-related abnormalities . Note added in proof . While our manuscript was under revision , Liu and co-workers [92] reported that reactive oxygen species due to mitochondrial dysfunction lead to LD accumulation , which , in turn , promotes neurodegeneration both in Drosophila and mice . These findings are in agreement with the data of our screen showing that one of the largest and most effective categories of DVAP-P58S modifiers is represented by genes involved in LD biogenesis and dynamics . Therefore these two works converge together in designating a role for LD dysfunction in the aetiology of ALS and , possibly , other neurodegenerative diseases .
A stock carrying ey-GAL4 and UAS-DVAP-P58S on the second chromosome was established by conventional recombination methods and used as a tester line for the screen . Individual EP or EPgy2 lines were crossed to this line and the F1 progeny was tested for suppression or enhancement of the DVAP-P58S-derived small and rough eye phenotype . To prevent accidental and progressive accumulation of extragenic modifiers of the DVAP-P58S phenotype , the ey-Gal4 , DVAP-P58S stock was maintained at 22°C or at 18°C degrees and in these conditions no DVAP-P58S-induced eye phenotype was visible in the stock . Putative candidates isolated from the screen were then maintained for the next round of analysis . To quantify the eye surface area images were analyzed with ImageJ software ( ImageJ software , National Institute of Health , Bethesda , MD , USA ) . The phenotypic effect of doubling the dosage of the DVAP-P58S transgene was studied in CyO+ progeny derived from the cross of ey-Gal4 , DVAP-P58S/CyO flies with flies carrying the DVAP-P58S transgene in homozygosity . To test the effect of overexpressing candidate genes in the motor system , EP or EPgy2 individual stocks were crossed to the tester line expressing the elav-GAL4 driver and the DVAP-P58S transgene . For the genetic experiments in which the phenotypic analysis was performed in the larval imaginal discs , the tester line was balanced over a “green balancer” . GFP expression from the balancer chromosome was used to select larvae for dissection that do not contain the balancer and on which the phenotypic analysis was carried out . Individual strains from the EP and EPgy2 collections as well as from the P{Mae-UAS . 6 . 11} collection were tested for their ability to genetically modify the ey-Gal4-DVAP-P58S eye phenotype by mating 8 to 10 males of their strain to 10–15 females of the ey-GAL4 , DVAP-P58S/CyO screening stock . After 2 days , adults were transferred to a fresh vial to create a duplicate cross and to maintain optimal cultural density . For the same reason , adults were discarded from the duplicate vial after an additional 2 days had passed . Embryos from both vials were raised at a temperature of 30°C in a water bath to maximize the expression of the Gal4 . For the analysis of the enhancing effect , modifiers were first identified because , when compared to the tester line , they exhibit a reduction in the eye size and organism viability at 30°C . The potential enhancing effect of these lines was subsequently confirmed by quantifying the decrease in eye size and viability at 28°C . After crossing females of the ey , DVAP-P58S/CyO-GFP line with wild-type Canton S males , the viability of the tester line ( Vtes ) was calculated as a ratio between the number of CyO+-GFP flies ( ey , DVAP-P58S/+ ) over the number of expected flies based on the Mendelian ratio . The viability of the ey , DVAP-P58S in the presence of the enhancer line ( Venh ) was calculated in the same way . The normalized lethality for every enhancer was expressed as ( 1-Venh/Vtes ) x 100 . RNAi lines were acquired from the Vienna Drosophila RNAi Center while the entire overexpressor collection and a few TRIP RNAi lines were from the Bloomington Drosophila stock Center . We performed gene set enrichment analysis ( GSEA ) on the modifier list by first de-correlating annotated terms in the GO-graph using the weighted elimination method of [43] resulting in conditionally independent GO-terms . We tested for term enrichment by hypergeometric test and corrected for multiple testing by controlling the false discovery rate through application of the Benjamini Yekutieli correction ( α = 0 . 05 ) . The Drosophila association network was obtained by querying GeneMania with the modifier gene using all , but gene co-expression derived interactions . In order to determine whether the network produced contained summary features associated with real biological networks we compared the mean node-degree and mean node-betweenness of the modifier network with those of GeneMania networks constructed from 1000 randomly selected gene lists of the same size . We calculated Z-scores comparing these metrics between the modifier network and metric distributions from the random gene list derived networks . Human orthologues were mapped using predictions made by the DIOPT system . The modifier and the expanded hVAPB networks of the human orthologues were generated through the use of IPA ( Ingenuity Systems at www . ingenuity . com ) . All networks were visualised using Cytoscape [93] . Females of genotype elav-Gal4;DVAP-P58S were crossed to males of the mutant strains . Climbing assays were performed on 10 age-matched adult female flies raised at 28°C as described in [94] . The flies placed in a plastic vial were tapped to the bottom of the vial and the number of flies above 8 cm line was counted after 15 seconds . A total of 10 trials were performed every 48 hours . Third instar larvae were dissected in cold 1x PBS and fixed at room temperature ( RT ) for 10 min in Bouin’s fixative ( 15 Picric Acid:5 Formaldehyde:1 Acetic Acid ) . Samples were washed in 0 . 1% Triton-X100 in PBS ( PBT ) and blocked in PBT containing 10% NGS for 2 hours at RT before being incubated overnight at 4°C with primary antibodies . A rabbit Rab5 primary antibody ( Abcam , ab18211 ) was used at a concentration of 1: 1000 and the same concentration was used for the DVAP primary antibody made in guinea pig . The anti-syntaxin 7 antibody was used at a concentration of 1:200 . Primary antibodies were washed off with PBT at RT for two hours by changing PBT every 15 minutes . Samples were incubated with secondary antibodies at RT for 2h and washed for 2 hours in PBT with changes of PBT every 15 minutes . Secondary antibodies were purchased from Jackson ImmunoResearch and used at a concentration of 1:500 . Stainings on NMJs were performed as previously described [15] . Tissues were mounted in Vectashield Mounting Media ( Vector Laboratories ) . Preparations were imaged on a Zeiss Axiovert LSM510 confocal microscope . Human tissue was obtained from the MRC Edinburgh Brain Bank with full ethical approval for research studies by EoSRES ( East of Scotland Research Ethics Service ) , Ref . No . 11/ES/0022 . For our analysis , we used 3 ALS cases and the same number of age-matched controls . The age range of ALS cases was 55–63 , with 2 females and 1 male case . Disease duration from diagnosis to death ranged from 7 months to 22 months with an average of 13 months . All cases were sporadic and none of them showed the C9orf72 expansion . The age-matched controls died from cardiac disease and after full neuro-pathological post-mortem examination , were considered to have no neurological disease during life . Spinal cords from the cervical region were examined in all cases . Human spinal cord tissue fixed in 10% neutral-buffered formalin was processed into paraffin . 7 μm sections were cut and de-paraffinized with xylene before being rehydrated through graded ethanol solutions . Sections were pre-treated using heat-induced epitope retrieval with Novocastra pH6 retrieval buffer in a decloaking chamber by heating to 125°C for 10sec , cooling to 90°C before washing in running tap water . For immuno-enzymatic stainings , slides were stained on a Leica Vision Biosystems Bond robot using the refine polymer detection kit ( Leica ) as follows . Endogenous peroxidase was blocked with 3% hydrogen peroxide in TBST for 10 min . Sections were incubated with rabbit anti-RAB5 ( 1:250 , Bethyl Laboratories ) primary antibody in 0 . 1% TBST for 2 hours at 25°C and then incubated with anti-rabbit HRP polymer for 15 minutes at 25°C . Staining was visualized using 3 , 3’-diaminobenzidine as chromogen . Tissue was finally subjected to haematoxylin staining , dehydrated through graded ethanol , cleared in xylene and mounted in Pertex ( Cellpath ) . For immuno-fluorescence stainings , following dewaxing in xylene and rehydration in graded ethanols , sections were rinsed in water and then subjected to heat-induced epitope retrieval by heating to 125°C in a decloaking chamber ( Biomed ) before passively cooling to 90°C in Novocastra pH 6 retrieval buffer . Sections were washed in running tap water then placed on a Leica Biosystems Bond X robot for dual tyramide staining as follows . Following blocking in 3% H2O2 for 30 mins in Bond wash buffer and 30 mins in 20% normal goat serum , sections were incubated with the RAB5 antibody at a 1:500 dilution for 60 mins . After two washes in Bond buffer for 5 mins , sections were incubated with goat anti-rabbit peroxidase Fab fragments ( Abcam ) for 30 mins . Sections were again washed twice for 5 mins in Bond buffer and incubated with Tyramide-Cy3 for 10 mins ( PerkinElmer ) . After two additional washes in Bond buffer for 5 mins , sections were subjected to heat-induced antigen retrieval for a further 10 mins using Leica Bond ER1 solution on the robot . Following a wash in bond buffer and blocking in 3% H2O2 for 30 mins in Bond wash buffer and 30 mins in 20% normal goat serum , sections were incubated with an anti-hVAPB antibody at 1:1000 dilution for 60 mins . Sections were then incubated with goat anti-mouse peroxidase fab fragments ( Abcam ) for 30 mins and subsequently with Tyramide FITC for 10 mins ( Perkin Elmer ) . Sections were mounted on slides using permafluor ( Thermo Scientific ) and examined on a Zeiss Axiovert LSM510 confocal microscope . Statistical analysis was performed and graphs were generated using GraphPad 5 . 0 . For experiments with more than two samples , a one-way ANOVA test was applied . Tukey’s multiple comparison test was then used as a post-hoc test when a significant difference was found in the ANOVA test . For experiments with only two samples , a two-tailed unpaired Student’s t-test was applied . For the climbing assay data , two-way ANOVA and Bonferroni as a post-hoc test were used to compare differences in motor performance between genotypes at different time points . | Amyotrophic Lateral Sclerosis ( ALS ) is a neurodegenerative disease causing loss of motor neurons and consequently a progressive deterioration of motor functions . ALS is uniformly fatal with death occurring 5 years after onset of symptoms . There is currently no effective treatment for ALS . Several mutations in a gene called hVAPB have shown that this gene is causative of a type of ALS known as ALS8 . In this study we sought to identify genes and cellular processes that are involved in the toxicity conferred by the defective ALS8 allele . By using the power of Drosophila genetics , we performed a large scale genomic screen and identified a number of genes that can affect hVAPB-mediated toxicity . These modifiers cluster into functional pathways known to be involved in ALS as well as novel ones . The relevance of these modifiers and mechanisms for the human disease was confirmed by showing that the human homologues of the fly modifiers can be organized into a network that closely resembles that of the Drosophila genes . Identifying cellular processes and proteins that modulate hVAPB pathological activity can facilitate the discovery of an effective treatment for ALS . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Network Analyses Reveal Novel Aspects of ALS Pathogenesis |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.